‘atfia't ”EV-2:: ‘ ' an , '91-; 334 jfi 3“"! “. 3"" 7:9“. : 1—1—2; fiat!" ""44.- i, fay", ‘ _ ,‘x. .v“'.4 “A‘na‘f \i‘ ' ‘ 33:4?9‘3‘59 < = ... 4 J , 645%}. 3%? 1* L‘:-~’r¢1’:‘-’J3 ‘ .L'..-An:;‘z‘ :..-. 3-1, J um . . 2' - 46‘; -. mpg. . '11. n7" ‘ ”i=3" . 1n. -3 This is to certify that the thesis entitled EFFECTS OF MOLECULAR WEIGHT OF THERMOPLASTIC MATRIX AND PROCESSING CONDITIONS ON INTERFACIAL ADHESION IN CARBON FIBER COMPOSITES presented by VENKATKRI SHNA RAGHAVENDRAN has been accepted towards fulfillment of the requirements for Master Wdegree in Chemical Engineering «ma/7. Major professor DateWfi‘ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution « iiiiiiiiiiiiiiii LIBRARY Michigan State ' University PLACE iN RETURN BOX to remove this checkout from your record. DATE DUE DATE DUE DATE DUE J! L D MSU Is An Affirmative Action/Equal Opportunity institution chimeras-9.1 EFFECTS OF MOLECULAR WEIGHT OF THERMOPLASTIC MATRIX AND PROCESSING CONDITIONS ON INTERFACIAL ADHESION IN CARBON FIBER COMPOSITES By Venkatkrishna Raghavendran A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemical Engineering 1 994 ABSTRACT EFFECTS OF MOLECULAR WEIGHT OF THERMOPLASTIC MATRIX AND PROCESSING CONDITIONS ON INTERFACIAL ADHESION IN CARBON FIBER COMPOSITES By Venkatkrishna Raghavendran The effects of molecular weight of Bisphenol A polycarbonate, a thermoplastic matrix and the effects of the processing conditions on the interfacial adhesion in carbon fiber composites were investigated. It was found that the molecular weight of the matrix medium and the processing temperature have significant effect on the interfacial shear strength. There was an increase in the level of adhesion with increasing molecular weight, with a major increase between the low molecular weight range, which is below the critical molecular weight and the next higher molecular weight grade. Above the critical limit the increase was monotonous but to a lesser degree. Increasing processing temperature gave an increase in the interfacial adhesion. Experimental investigation show that the interface might also affect the segregation of polymer chains by their molecular weight at the interface. It was seen that the lower molecular weight chains tend to segregate to the interface. TO MY PARENTS & MY MATERNAL UNCLE ACKNOWLEDGEMENTS I would like to earnestly thank Dr. Lawrence T. Drzal for his guidance and support throughout the duration of my work, he has been a constant source of inspiration in both my research work and life. It is a pleasure working with him and I have learnt a great deal under his able guidance. I wish to thank Mike Rich, Brian Rook and Dan Hook for their help and guidance in operating the equipments in the Composite Center and for their friendship. I also wish to thank all the members of my research group for their assistance and friendship and making my stay here an enjoyable one. I am grateful to Dr. Jim DeRudder at GE Plastics, for providing me with samples of BPA polycarbonate in different molecular weight grades and also for conducting gel permeation chromatography work on them. I wish to thank all my friends with whom I spent my first two years of life in East Lansing, they are among the best friends a person can ever desire. I especially wish to thank Sanjay Mishra, Gunjan Trivedi, Sreenath Gajulapalli, Sanjeev Walia and Naresh Amineni for their camaraderie and support. I thank my parents and maternal uncle for always encouraging me and supporting me in all my endeavors. I thank my brother Anand for his emotional and financial support. Finally, I thank my wife Prathima for her immense love and support for enabling me to complete this work. She deserves a lot of credit for her ability to quickly adjust to a new lifestyle in this country and at the same time endure my sometime long and irregular work hours. iv TABLE OF CONTENTS List of Tables List of Figures CHAPTER 1 INTRODUCTION CHAPTER 2 CONSTITUENT MATERIAL PROPERTIES 2.1 Intoduction 2.2 Material selection 2.3 Reinforcing Fiber 2.4 Matrix Characteristics 2.4.1 2.4.2 2.4.3 2.4.3.1 2.4.3.2 2.4.3.3 2.4.3.4 2.4.4 General Properties and Applications Structure and Properties Relations Effect of Molecular Weight Processing Temperature Mechanical Properties Viscosity Crystallinity Effect of Molecular Weight Distribution Page viii ix ll 14 17 17 20 22 22 24 26 28 29 2.4.5 Disadvantages of BPA Polycarbonate 2.5 Conclusion CHAPTER 3 CONSOLIDATION AND INTERFACIAL BONDING - THEORY 3.1 Introduction 3.2 Principles of Autohesive Strength Development 3.2.1 WOOl’s Bonding Model 3.3 Principles of Polymer Adsorption 3.3 Conclusions CHAPTER 4 EFFECT OF MOLECULAR WEIGHT ON INTERFACIAL ADHESION 4.1 Introduction 4.2 Experimental Procedures 4.2.1 Specimen Fabrications 4.2.1a Injection Molding 4.2.lb Hot Press Specimen Fabrication 4.2.2 Testing, data Acquisition and Analysis 4.2.3 Failure Mode Image Acquisitions 4.3 Results and Discussions 4.3.1 Tensile Test Results 4.3.2 Adhesion Tests Results vi 29 32 33 33 34 35 39 43 45 45 45 46 52 55 55 57 4.4 Conclusions CHAPTER 5 EFFECT OF INTERFACE ON SEGREGATION BY MOLECULAR WEIGHT 5.1 Introduction 5.2 Theoretical Background 5.3 Specimen Fabrication 5.4 Results and Discussions 5.5 Conclusions CHAPTER 6 CONCLUSIONS AND RECOMIVIENDATIONS 6.1 Conclusions 6.2 Recommendations for Future Work BIBLIOGRAPHY vii 77 79 79 80 83 84 88 9O 9O 92 93 Table 2.1 2.2 2.3 2.4 4.1 4.2 4.3 4.4 5.1 LIST OF TABLES Dependence of interface strength on matrix strength Nomenclature and molecular weights of the BPA polycarbonate grades Typical properties of the A84 carbon fiber Notched Izod impact strength of rigid plastics at 24°C Kinetic parameter data for hot press fabrication Tensile test result Single fiber fragmentation test results Micro indentation test results ITS test results for verifying segregation by molecular weight at interface viii Page 10 12 14 19 50 56 58 6O 84 LIST OF FIGURES Figures 1. 1 Schematic diagram of the fiber-matrix interphase in composites 2.1 Chromatogram of the GPC analysis of the polycarbonate grades 2.2 Schematic sketch of a section of carbon fiber showing its morphology 2.3 Formation of BPA polycarbonate by phosgenation reaction 2.4 Schematic sketch of the effect of molecular weight on the increase in the glass transition and melting temperatures 2.5 Schematic sketch of the effect of molecular weight on the tensile strength and yield point 2.6 Schematic sketch of the effect of molecular weight on the matrix modulus 2.7 Schematic sketch Of the effect of molecular weight on the zero shear viscosity 2.8 Schematic sketch Of the effect of mwd on the viscosity, note the more non-Newtonian behavior of the broader MWD 3. 1 Schematic sketch of polymer chains at the interface in Wool et al. model 4. 1 A schematic sketch showing the assembly for the hot press fabrication technique 4.2 The Consolidation cycle for making the composite specimens at 250°C 4.3 The plot of UT vs log k for from the kinetic data given in table 4.1 4.4 The photograph of the loading jig used in tensile testing of the SFF test specimens 18 23 25 25 27 30 36 47 49 53 Figure Page 4.5 Interfacial shear strength of the polycarbonate grades with two different 59 molecular weight, note the increases due to the molecular weight 4.6 The effect of the processing temperature on the interfacial shear 61 strength for various molecular weight ranges. 4.7 The interfacial shear strength of the four molecular weight 62 grades processed at 250°C for 45 minutes 4.8 The interfacial shear strength of the four molecular weight 62 grades processed at 275°C for 15 minutes 4.9 The interfacial shear strength of the four molecular weight 63 grades processed at 300°C for 8 minutes 4.10 The three dimensional representation of the effects of molecular 63 weight and processing temperature on interfacial shear strength 4. 11 A comparision of the interfacial shear strength obtained from the 64 testing systems. Note the similarity in the trend 4.12 A schematic representation of the polymer chain adsorbing on to 68 the fiber surface 4. 13 A schematic sketch of the polymer chain losing conformation 68 as it nears the interface 4.14 The cracked surface of the PCOQ grade SFF specimen, the fiber 71 pulled out due to the cracking can be clearly seen 4.15 The bifringement pattern of the PCHF grade processed at 260°C 72 for 40 minutes 4.16 The bifringement pattern of the PC100 grade processed at 260°C 72 for 40 minutes 4.17 The bifringement pattern of the PCHF grade processed at 290°C 73 for 15 minutes 4.18 The bifringement pattern of the PC100 grade processed at 290°C 73 for 15 minutes 4.19 The bifringement pattern of the PCHF grade processed at 315°C 74 for 8 minutes 4.20 The bifringement pattern of the PC100 grade processed at 315°C 74 for 8 minutes Figure 4.21 4.22 4.23 4.24 5.1 5.2 5.3 5.4 The micrograph of the ITS specimen surface of the PCOQ grade composite. Note the deep cracks in the fiber plane The micrograph of the ITS specimen surface of the PCHF grade composite. The amount of cracks are less than the ones seen in PCOQ grade The micrograph of the ITS specimen surface of the PCFF grade composite. The crack growth is seen to be decreasing The micrograph of the ITS specimen surface of the PC100 grade composite. Note the low amount of interfacial cracking in this grade A schematic plot of the segregation by molecular weight at the interface. Note the decrease in the volume fraction with increasing MW The interfacial shear strength of low molecular weight PCOQ coated A84 fibers consolidated in higher molecular weight PC100 bulk matrix The interfacial shear strength of higher molecular weight PC100 coated AS4 fibers consolidated in low molecular weight PCOQ bulk matrix A three dimensional plot of the data given in table 5.1, comparing the IFSS of the coated and the uncoated fibers xi Page 75 75 76 76 82 85 85 86 CHAPTER 1 INTRODUCTION The world as we perceive it today is very much different from the way it was at the turn of this century. There has been a rapid growth of technology, more than all the development seen through out the history of mankind. Metals and their processing fuelled the growth of industries in the last century and through most part of this century. In last thirty years development of a new class of material, polymer composite materials having superior characteristics of light weight, versatility of fabrication and use, strength and environmental resistance have increased the rate of progress in our times. Polymers have come to play a very important role in very diverse fields of the world as we perceive today, their applications can be seen in fields ranging from state of art space age industries to the toys. The quest to get optimal properties for different applications of the polymers has caused them to be blended with other polymers, compounded with chemicals, fillers, modifiers and additives. Polymers have been made into composites with metals, ceramics and reinforcing fibers such as carbon, glass fibers to improve the mechanical and thermal properties of the finished products. Composite materials can be defined as a system consisting of two or more physically distinct and mechanically separable materials, which can be mixed in a controlled manner to have a dispersion of one material in another to achieve optimum properties. The properties are superior and possibly unique in some respects to properties of individual components[ 1]. Composites in structural applications can be classified under 2 one of the following: ceramics, metals and alloys and polymeric composites. One or more of these materials can be used to make various kinds of composites. The discussion in this work will be restricted to polymeric composites only. The polymeric composites can be subdivided into the following categories: fibrous composites (consisting of fibers embedded in polymer matrix), laminated composites (consisting of fibrous composites in one or more than one different planes of orientation), particulate composites ( consisting of particles of reinforcing medium in polymer matrix). Fibrous composites are composed of fibers, which are either continuous and aligned or short and randomly dispersed in a polymer matrix medium. The fibers themselves can be of various types, prominent among them being carbon, glass and polymer fibers. The polymer matrix can be either a thermoset or a thermoplastic. The structure, chemistry and the behavior of these two class of polymers are very different. Thermosets resins are usually low molecular weight liquids which form a cross linking polymer network when a curing agent is added and the mixture is cured. The curing of the thermoset polymer is an irreversible process and parts molded once cannot be remolded. Thermoplastic resins are long chain polymers (containing more than fifty repeating monomeric units) which can be thermoformed at an elevated temperature above the glass transition temperature Tg and high pressure into desired parts. The major advantage of thermoplastic polymers lie in the fact that they can be recycled. Though all polymers have potential to be used as the matrix medium, limitations such as processability and end use properties preclude quite a number of them [2]. Hergenrother and Johnston [3] proposed the following requirement for an ideal polymeric system to be used in composite material: easy to prepreg, good shelf life, acceptable tack and drape, acceptable quality control procedures, low processing temperature and pressure, no volatile evolutions, dense void free matrix, good mechanical performance over the desired temperature range, time and environmental conditions, acceptable 3 repairability and cost effectiveness. Since no available matrix conform to all these desired properties, the choice of a matrix usually represents a compromise. Thermoset matrices such as epoxy resins find wide use in composite industries. Due to their low viscosity they are easy to process and also they have low shrinkage, excellent adhesion and good chemical resistance [1], Over the yearsa large amount of research has been conducted to quantify the various factors affecting the adhesion of thermosets to glass, carbon and polymer fibers. Though thermosets have lot of advanta- geous characteristics. there are a few drawbacks of these resins. The final composites take a long time to mold due to long curing cycles, tendency to degrade at high temperatures (175°C and higher), low resin strength and fracture toughness and a high degree of brittleness [2]. Thermoplastic matrices offer many advantages over thermosetting resins. They have higher temperature resistance, very good fracture toughness, good neat resin strengths, shorter molding cycles, reduced storage and handling problems, infinite shelf life of intermediate prepregs, capability to fusion bond, recyclability and repairability. Difficulties in processng have hampered the use of thermoplastic composites on a wider scale. The intractability of these matrices due to their high viscosities even at elevated processing temperatures give rise to a host of problems such as stiff and boardy prepregs due to high resin content, poor fiber wetting, solution devolatization during casting leading to formation of voids and loss of mechanical strength. The reinforcing fiber morphology, surface chemistry and surface treatments have received a lot Of attention and various factors affecting the adhesion of the fibers to polymer matrices have been studied in depth by many researchers[4,5,6,7,8,9,10,11]. Oxidation, plasma treatments and sizing have been shown to improve the adhesion in most of the fiber-matrix systems[12,,13,14]. In order to translate the desirable matrix properties of the thermoplastics to the composite materials a number of researchers[15,16.l7,18] in recent years have started 4 to investigate the effects Of various properties of the thermoplastic matrices on the fiber- matrix interactions. The present work is also a part of this thrust in the direction of full exploitation of the thermoplastic polymers to get stronger and more durable composites. The work under discussion is directed at detemining interfacial structure property relationship between amorphous thermoplastic matrix properties and on the adhesion levels in carbon fiber-thermoplastic composite system. Amorphous matrix system was chosen to preclude the effects of crystallinity in the bulk matrix and formation of transcrystalline region near the fiber surface and the attendant difficulties in characteriz- ing the fiber matrix interphase. Primary objective of this investigation is to examine the effects of matrix properties, particularly the molecular weight of the thermoplastic resin and processing variables like consolidation temperature and residence time on the interfacial adhesion in carbon fiber Bisphenol-A polycarbonate composite system. The interface between fiber-matrix has been shown [19,20,21] to be a three phase boundary rather than the classical two dimensional boundary between the fiber and the matrix. The third region has been named by different authors as mesophase or the interphase. The interphase region between the fiber and the matrix is very complex and varies in thickness ranging from 5 to 5000 A, a schematic model developed by Drzal[22] and shown in fig. 1, gives a very good visualization of various interactions occurring in the interphase region. The structure and composition of the fiber-matrix interface has not been under- stood very well though a lot of advances have been made in experimental methods to probe the interface[23,24,25,26,27]. The importance of the fiber-matrix interface in the development of good mechanical properties in the composite is well known. But no technique has been developed which can examine the interphase region alone as the thickness of the region is near the resolution limits Of the analytical microscopic tools available at present. Therefore the interphase properties have been correlated to the macroscopically measurable properties of the constituent fiber and bulk matrix properties BULK MATRIX POL YMER of DIFFERENT PROPERTIES SIZNGS ADSORBED MATERIAL FIBER CHEMISTRY THERMAL , TOPOGRAPHY CHEMICAL , FIBER MORPHOLOGY MECHANICAL . BULK FIBER ENVIRONMENTS FIBER'MATR’X INTERPHASE armature” Figure 1.1 Schematic diagram of the fiber-matrix interphase in composites 6 using a few simplified assumptions. Many authors have proposed theoretical relationships between interfacial properties and the bulkfiber and matrix properties. The models developed allow the local stress to be computed based upon the constituent properties. Cox [28] and later Cooke [29] have developed models in which an elastic fiber Of length l, embedded in an elastic matrix under general strain a. The model assumes perfect bonding between the fiber and the matrix. Using the assumption of load transfer through the ends of fiber, leads to the following equation [29]: t = Elem Gm 0'5 sinhB(0.5L -x) 1 l R L osh — 2Efln(r) c 52 Where E = Tensile modulus of the fiber, em = strain in matrix, Gm = shear modulus of matrix, R = interfiber spacing, r = radius of the fiber, B = scaling factor. L = length of the embedded fiber, x = radial distance outward 1 = interfacial shear strength at a fixed point It is clearly evident that, if the specimen geometry is fixed and the fiber is unchanged, the theory predicts a direct dependence of the interfacial shear strength on the product of the matrix strain and square root of the shear modulus of the matrix. Rao and Drzal [30] have shown the square root dependence of the interfacial shear strength in both 7 thermoset and thermoplastic composite systems. The shear modulus of a thermoset epoxy DGEBA was changed by using amine curing agents with different molecular weights. The modulus of the Bisphenol-A polycarbonate which is a thermoplastic matrix, was changed by changing the temperature at which the testing was done. A more detailed description of the testing and the results can be found in ref.[31]. It was shown that a decrease in modulus, all other things being equal, causes a corresponding decrease in interfacial shear strength. Other authors [32,33] also have examined the effects Of matrix properties on the performance of the composite materials. Rosen [32] analyzed the shear stress fields along the fibers parallel to the fiber axis during tensile loading of a composite. He proposed a model consisting of a fiber surrounded by matrix embedded in a composite material exhibiting average composite properties. The fiber was assumed to carry only extensional loads and the matrix to transmit the shear stresses only. An equilibrium approach was used then to derive the relationship between the interfacial shear strength and the constituent properties. Dow [33] evaluated the case in which both fiber and matrix carry load and proposed a relationship similar to the one proposed by Cox and Cooke. Though all the studies have the matrix as the medium by which shear stresses are transferred to the fiber, yet most adhesion studies focus entirely on the interfacial interactions, chemical and physical between fiber and matrix and tend to neglect the matrix itself as having no serious effects on the fiber-matrix adhesion. Little experimental verification has been attempted where the interfacial physiochemistry remains unchanged but matrix properties are changed. Rao [31] had done some work in this direction and had found that matrix modulus has profound effect on the interfacial properties. It would be very tenous to extrapolate interfacial properties from one system to another without accounting for the effects of the matrix pTOperties. The present work was conducted to find the dependence or otherwise of interfacial shear strength on the effect of the molecular weight of the matrix medium.The work also will 8 verify experimentally the effects of the interface upon the segregation by molecular weight of the polymer chains in the interphase region. The thesis has been so arranged that it reflects the correlation between the matrix properties and the interfacial properties. Specifically the chapters are arranged in the following order: Chapter 2. discusses the material selection, various properties of the reinforcing fiber and the matrices. It also deals with various factors affecting the adhesion between the fiber and the matrix. Chapter 3. deals with the theories on the interfacial adhesion between the polymer- polymer interfaces and the polymer-fiber interfaces. The autohesive strength development theory and the Polymer adsorption theories will be described and their applicability in the present investigation will be discussed. Chapter 4. will deal with the the single fiber fragmentation and micro-indentation testings of the carbon fiber polycarbonate composites along with the mechanical testings of neat matrix materials. Chapter 5. deals with the theory proposed in the literature on the segregation by molecular weight of polymer chains at the fiber-matrix interface and the experimental investigations done in the present work. Chapter 6. presents the conclusions and will coherently tie together the results Of the work presented in the earlier chapters. A guideline for future work will be provided to have a clearer perspective of the properties governing the interface in composites. CHAPTER 2 CONSTITUENT MATERIAL PROPERTIES 2. 1 Introduction Since a composite is, by definition, a material with at least two distinct phases seperated by an interface, all composites rely upon some degree of interfacial adhesion. Depending upon the desired properties for the given application, the optimal value of this adhesion may be as high as possible or it may be lower to deliberately facilitate debonding [19], for example, in case of polymeric composite materials the interfacial adhesion should be very good to provide for good stress transfer across the interface, but in certain ceramic composites applications the interfacial adhesion is deliberately reduced. Indeed the true measure of adhesion between the reinforcing fiber and the matrix medium in fibrous polymer composite material is the ability of matrix medium to effectively transfer the stress applied to it to the much stronger fiber. As a consequence of the limiting value of the interfacial strength, should be the bulk matrix strength when there is intimate contact between the fiber and the matrix and the adhesion is perfect. But in reality more complex interactions come into play at the fiber-matrix interface as described in chapter 1. In thermoplastic matrices, the polymer chains in the interphase region are constrained by the fiber surface due to possible chemical and physical interactionsas well as the fiber morphology. The presence of additives and impurities also affect these interactions. The resultant properties of the interphase region therefore can be very much different from the bulk properties. A stronger interphase would require a higher modulus material in the interphase but the toughness would be much lower. Constraints on the matrix phase in the interphase may give rise to regions of localized stress concentrations, which can act as a locus for interphase failure. Though the properties of the interphase might be different from the bulk fiber and matrix properties, there seems to be a direct correlation between the bulk properties and the interface properties, as can be seen from table 2.1 [34]. Composites made from higher strength fiber and matrix have higher levels of adhesion compared to composites made from low strength fiber and matrix. In fact composites made from same fibers but with different matrices having very different strength show stronger adhesion with increasing matrix strength. A similar trend is seen with composites made from same matrix but with different fibers. 10 Table 2. 1 Dependence of interface strength on matrix strength Reinforcing Matrix Medium Fiber Strength Matrix InterfaceStrength Fiber Strength "' AS4a Polyphenylene 3587 MPa 48 MPa 14.82 MPa oxide AS4 Polysulfone 3587 MPa 70 MPa 14.80 MPa AS4 Polyetherimide 3587 MPa 105 MPa 19.28 MPa XASb Polyetherimide 3447 MPa 105 MPa 31.35 MPa AS 4 DGEBA°/mPDAd 3587 MPa 127 MPa 31.32 MPa XAS DGEBA/mPDA 3447 MPa 127 MPa 53.86 MPa ' QOO‘B *- Calculated from the mean critical Aspect ratio (lc/d) data in ref. [34]. Hercules Aerospace PAN based carbon fiber Hysol Grafil PAN based carbon fiber Diglycidylether of Bisphenol-A meta-phenylene diamine 1 1 The molecular weight of a thermoplastic matrix affects its other properties. These effects are well documented in the literature. ~But very little is known about it effects on the interfacial behavior in composites. The present investigation addresses this problem. In this investigation, the effects of the molecular weight and the processing conditions on the interfacial adhesion of a typical thermoplastic matrix in carbon fiber composites will be studied. This chapter therefore deals with the selection of the experimental materials and the reasons for their selection. 2.2. Material Selection The reinforcing fiber chosen for the investigation is Hercules Magnamiteo type AS4-6K PAN based carbon fiber supplied with standard electrolytic oxidative surface treatment. The type AS4 fiber is unsized and is categorized by the manufacturer as a high strength carbon fiber with continuous filament having a tow count of 6000. The polymer used in the investigation was a linear chain amorphous thermoplastic polycarbonate (PC) of bisphenol-A (BPA). The polycarbonate was supplied by GE Plastics Inc. in four different molecular weight grades. Three of the grades were supplied in the form of irregular powder ranging in sizes from submicron particles to 2mm agglomerates. These grades were directly taken from the reactors and sent in sample quantities of forty pounds each without any additives being added to them. One other grade was procured in sample quantity in the form of prefabricated sheets having a thickness of 0.02 inches. This grade is commercially available as Lexan’ 8040 - MC112 and is a FDA approved grade without any antioxidant or UV stabilizers. Neat PC resins were procured without any additives to minimize their effects on the interface, even though these additives (antioxidants and UV stabilizers etc.) are present in very small quantities, may tend to concentrate in the interphase due to their low molecular weights. Table 2.2 gives the details of the different grades, their nomenclatures as given by the manufacturer and the names assigned by the author along with the molecular weight data. 12 Table 2.2 Nomenclature and molecular weights of the BPA polycarbonate grades Manufacturers Author’s Weight a’verage Number average Nomenclature Nomenclature Molecular weight Molecular weight Mw/MN Mw MN ML5721-111 PC100 31135 14382 2.165 Lexan 8040 PC FF 24894 11865 2.098 ML5221-111 PC HP 22963 10310 2.227 ML5832-111 PC CO 16152 7623 2.119 These molecular weight averages were determined by gel permeation chromatog- raphy (GPC) based upon light scattering using absolute weight average molecular weight standards for polystyrene done on a Perkin Elmer LC3O Chromatograph. The solvent used for the analysis was HPCL grade chloroform (CHCl3). The analysis was done using a Jordi 10,000 column with a 0.25 w/w% solution Of the polycarbonate samples in chloroform. The rate of injection was kept at 1 ml/ minute and the response was noted at a rate of one point per second. The chromatogram of the GPC analysis is given in figure 2.1 Here the number average molecular weight MN is the total weight of all the polymer chains in a sample divided by the total number of moles present. The weight average molecular weight MW is given by the summation of the product of the weight fraction and the molecular weight of polymer chain having molecular weight Mx NM NM2 =__Z** and MW=waMx=———2“ 2.1 X”. ‘ 2 MM. N 13 PCIN > PCFF 56- PCHF i s ’C°° a szo‘ z . i 4.0 4;. 4e 42 l i 18 to 20 22 24 28 28 o 32 ELUTION VOLUME [ml] Figure 2. l Chromatogram of the GPC analysis of the polycarbonate grades 2.3 Reinforcing Fiber Hercules Magnamite” type AS4 carbon fiber is a typical high strength, high performance fiber. The tensile strength and the modulus are typical of a high strength fiber ranging in values of around 3500 MPa and 250 GPa respectively. The typical fiber properties according to the manufacturer’s data sheet [35] is given below in Table 2.3 14 Table 2.3 Typical properties of the AS4 carbon fiber Typical Fiber Properties U.S.Units S I Units Tensile Strength 520,000 Psi 3,587 MPa Tensile Modulus 34 x 106 Psi 235 GPa Ultimate Elongation 1.53 % 1.53 % Carbon Content 94.0 % 94.0% Filament Diameter 0.315 mils 8 microns Filament Shape Round Round Carbon fibers can be made from different precursors such as rayon, pitch and polyacrylonitrile (PAN), These precursors are drawn into fibers, stretched and sent through high temperature (> 3000 K) furnaces to be carbonized and graphitized. During the carbonization and graphitization most of the other compounds and elements other than carbon are devolatized leaving behind a carbonaceous material with the carbon in the aromatic ring structure. 15 Carbon fibers come in two predominant groups and are known as HT type and HM type. The HT type of fibers are high'strength high tenacity fibers with tensile strength and modulus ranging in values given for the AS4 fiber in the table above. The high modulus HM type fibers have the modulus values of 350 to 500 GPa. The HM type fibers have the disadvantage of low strain to failure below 0.5 % and therefore are of advantage in applications where stiffness is decisive. Consequently 90 % of all the carbon fibers used today are HT type fibers [36]. Adhesion in thermoset matrix composites may involve the formation of primary chemical bonds between the carbon fiber surface, or chemisorbed species and the highly reactive matrix constituents. Therrnoplastics, in contrast have already undergone polymerization and are much less reactive. Primary chemical bond formation is highly unlikely . Adhesion is probably dependent upon physisorption driven by dispersion forces and upon mechanical interlocking. Hence the emphasis in the present section will be on the morphology of the fiber rather than the chemistry, which would be more appropriate for reactive thermosetting systems. The structure of the carbon fiber determines its type and properties. The carbon fiber is made of stacks of aromatic carbon layers lying approximately parallel to the fiber axis [37]. Generally the high strength fibers are low temperature carbonaceous material with carbon content ranging from 92.5 % to 98 %. In the HT type fibers the basic aromatic structural unit (BSU), which resemble a graphite plane is very small (less than 10 A). The high modulus fibers are usually high temperature carbonaceous material with carbon contents greater than 99 % and their layers are larger and perfect although folded. The BSU is also large (ranging from 45 A to 72 A). A schematic sketch of the fiber 4’“ fl momma. EXTERNAL ._> SURFACE SBCI'KN Pouas " \ , fl Figure 2.2 Schematic sketch of a section of carbon fiber showing its morphology 17 surface is shown in figure 2.2 to demonstrate the morphology of the carbon fiber. The small BSU and consequent higher flaws in the aromatic stacks give the HT type fibers considerably more irregular outer planes. The exposed surface may have more micropores and a topography which is rough leading to a better interlocking of the polymer chains with the surface. As mentioned earlier in the section this interlocking along with primary Van der Waal interactions and secondary polar bonding between the surface oxygen and nucleophilic sites on the polymer chain would be the mechanism by which the thermoplastic matrices adhere to the carbon fiber surface. 2.4 Matrix characteristics 2.4.1. General Properties and Applications The Bisphenol-A polycarbonate is the most widely used commercial polycarbonate. BPA polycarbonates are prepared from an ether exchange reaction or by phosgenation of bisphenol-A, the latter being the more widely used method [38,39,40]. The polymer is produced by interfacial polymerization of bisphenol A and phosgene in aqueous alkaline solution and inert organic solvent such as methyl chloride. The polymerization takes place in the presence of an amine catalyst. Figure 2.3 gives the simplified version of the reaction between BPA and phosgene. A more detailed description of the reaction can be found elsewhere [39]. The linear chain amorphous thermoplastic is unique in certain ways from most of the other thermoplastics. The desirable feature of the polycarbonate is its high deflection temperature, which gives good rigidity to the parts made from it upto 140°C. The very good electrical insulation characteristics of the polymer finds its largest single field of 18 CH3 n H0 —©>—(|3 —-©— 0H + nCOC12 I... — _ BISPHENOL A PHOSGENE CH3 .—I mole—w, + m I O CH3 _n POLYCARBONATE Figure 2.3 Formation of BPA polycarbonate by phosgenation reaction 19 Table 2.4 Notched Izod Impact strength of Rigid Plastics at 24°C Plastic Impact Strength ( ft lbs/inch) Polycarbonate (BPA) 12.00 - 18.00 Polystyrene 0.25 - 0.40 J ABS polymers ' 1.00 - 10.00 Polyvinyl chloride ( polyblends) 3.00 - 20.00 Polymethyl methacrylate 0.40 - 0.50 Cellulose acetate 1.00 - 5.60 I Cellulose nitrate 5.00 - 7.00 Ethyl cellulose 3.50 - 6.00 Nylon 6 1.00 - 3.00 Nylon 6 6 1.00 - 3.00 Nylon 6 -12 1.00 - 1.40 Polyethylene (low density) > 16 Polyethylene (high density) 0.50 - 20.00 Polypropylene 0.50 - 2.00 Epoxy resins 0.20 - 5.00 Polysulfone 1.30 - 5.00 20 application in electronics and electrical engineering, comprising of around 55% of the market. Polycarbonate moldings have also been made for computers, magnetic disc pack housings, terminals and contact strips. The polymer is widely used in making coil formers, which is the protective insulation on the outside of the electrical coils. The high molecular weight PC films are used in the manufacture of capacitors. Transparency of the PC is in the range of 87% to 94%. Due to their excellent transmittance characteris- tics, polycarbonates now dominate the audio compact disc market, where material of high transparency and purity is required. The polycarbonates are also used in making lenses for eyeglasses [38] . The resistance of polycarbonate resins to ’creep’ or deformation under load is markedly superior than most of the other thermoplastics, so it is used as a structural engineering plastic and finds applications such as roofing materials, window panes and load bearing partitions walls in buildings. The polycarbonate has one of the highest toughness among the various thermoplastics available in the market. Table 2.4 [41] gives a comparison of the toughness of various polymers. Other polymers are as rigid or as transparent or even both more rigid and as transparent, but polycarbonate of Bisphenol A is the only material that can provide such a combination of properties at such a reasonable price. The excellent toughness properties make the polycarbonates a very desirable material in sports products [39]. 2.4.2 Structure and Pmperties Relations A fairly accurate prediction of bulk properties of the polymer can be made from the molecular structure of the bisphenol A polycarbonate. The molecule has a symmetri- 21 cal structure and therefore it is not stereospecific; even though it is symmetric the bulky phenyl group, methyl group and polar acetate and ether linkages make it very difficult to crystallize the polymer under normal conditions. The carbonate groups are polar but are separated by the aromatic phenyl groups. Unlike the other aliphatic polycarbonates which can be hydrolyzed very easily, BPA polycarbonate is more resistant to hydrolysis which is ascribed to the more protective influence of hydrophobic phenyl group on each side of the polycarbonate. The equilibrium water adsorption of the BPA polycarbonate is also very low (around 0.3%) [40]. This very low water adsorption of bispenol-A polycarbonate contributes to a high degree of dimensional stability [38]. The repeating unit of the molecule is quite long (21.5 A°) and the presence of benzene rings in the chain restricts flexibility of the molecule. The consequences of the above two factors tend to give a rigid molecular backbone to the BPA polycarbonate leading to a high melting temperature and glass transition temperature Tg, ranging from 145-150°C. The reason for these high values is molecular mobility and not polar interaction. This can be inferred from a comparison to poly (ethylene terepthalate) which has similar chemical groups but lower Tg and Tm. The high value of Tg is attributed to the bulky structure and hindered movements of molecular segment [39] and a high free volume [38]. The long repeating unit further more hinders ordered arrangement. The in- chain movements of phenylene isopropylidene and carbonate groups also take place even though the chain motion is frozen. The merging of the secondary transition of these groups give a diffuse band in the range of -200°C to 0°C and a broad loss modulus maximum [38,42]. These factors have also been stated to produce the high toughness of BPA polycarbonate. The limited degree of crystallinity is another factor contributing to 22 the toughness of the polymer. The absence of both secondary and tertiary C-H bonds lead to a high measure of oxidative stability, the'dry polymer may be kept for hours in the molten state upto 310°C and for shorter times at 320-340°C [39] with little decomposi- tion. The bulky structure also contributes to a high entanglement density above a critical molecular weight and leads to a high melt viscosity of the resin (ca. 4000 Pa-s at 250°C). The effect of increasing temperature on viscosity is less marked with polycarbonate than other polymers. The apparent melt viscosity is also less dependent on the rate of shear than usual with thermoplastics. 2.4.3 Effect of Molecular Weight The molecular weight of the polymer affects the other matrix properties. The effects depend upon the structure and the chemistry of the polymer chain. The effect of the molecular weight on the viscosity, mechanical properties, processing temperature and crystallinity are given in the succeeding sections. 2.4.3.1 Processing Temperature A specific linear amorphous polymer can exist in a number of states according to the temperature and the average molecular weight of the polymer [4]. This is shown diagrammatically in figure 2.4. At low molecular weights, the polymer will be solid below some give temperature, above that temperature it will be liquid. At higher molecular weight such a clearly defined melting point no longer exists and a rubbery intermediate zone is observed. For amorphous thermoplastics such as polycarbonates, the 23 Figure 2.4 Schematic sketch of the effect of molecular weight on the increase in the glass transition and melting temperatures 24 upper limit of this diffuse transition zone is considered to be the melting point of the polymer. For polycarbonates this temperature ranges from 220°C to 260°C, increasing with increasing molecular weight. This raises the processing temperature of polycarbonates to 260-320°C requiring special processing conditions, such as the use of inert atmosphere to reduce thermal oxidative degradations. 2.4.3.2 Mechanical Properties The molecular weight of the polymer has significant influence on the mechanical properties below a critical molecular weight Mc , above which the differences are much less. Very low molecular weight polymers are "cheesy" elastomers with low strength and low elongation to break. The low molecular weight polymers are usually brittle below Tg, but bispenol-A polycarbonates show a more ductile behavior [38], though this author found occurrences of brittle fracture to be more common in the lower molecular weight ranges tested. It has been shown that the chain ends act as imperfections in the chain adversely affecting the strength properties. The number of chain ends are higher in concentration in low molecular weight PC and consequently lower the strength. Figure 2.5 gives a schematic sketch of the effect Of molecular weight on the strength. Above the minimum molecular needed to get satisfactory mechanical behavior, the strength and elongation increases towards a limiting value. For amorphous thermoplastics such as polycarbonate, which tend to have a high Tg, the modulii behavior is schematically shown in figure 2.6. The drop in the modulus occurs at around Tg. The molecular weight will have practically no effect on the modulus below Tg for these polymers [43]. The 25 an \\MC 2 “F Ia» .. 2a).. 100‘ l l l l l 0 2 4 6 8 10 12 I4 16 STRAIN; Figure 2.5 Schematic sketch of the effect of molecular weight on the tensile strength and yield point 7 10 e 10. 10’. SHEAR MODULUS (Pa) 10- 333% ‘1. fig 1". TEMPERATURE 3 10 Figure 2.6 Schematic sketch of the effect of molecular weight on the matrix modulus 26 behavior of the polymer above Tg is determined by the entanglements and the viscous flow becomes more important than the rubbery elasticity of the polymer. 2.4.3.3 Viscosity The effect of the molecular weight on the viscosity is well documented in the literature, in general it is seen that below a critical molecular weight MC of the order of about 5000 to 15000, the viscosity is directly proportional to the weight average molecular weight MW. Above MC, viscosity depends upon a higher power and for many polymers a relation of the form given in equation 2.4.3.3 has been found to hold. 3.4-3.5 2.4, , '1. : KMW 33 where no is the zero shear viscosity. A schematic sketch is shown in the figure 2.7. The reptation theory postulated by deGennes [44] predicts a third power relation between 1),, and MW, the differences between the model and experimental work was explained by Graessley [45] to be due to the polydispersity in the molecular weight of the polymer used in the experimental work compared to the monodisperse molecular weight distribution used to arrive at the model. The melt viscosity of the thermoplastic is dependent upon the molecular weight, the higher the molecular weight, greater the entanglements and greater the melt viscosity. In case of bispenol A polycarbonate, this dependency is quite high, the melt viscosity 27 LOG“o Figure 2.7 Schematic sketch of the effect of molecular weight on the zero shear viscosity 28 increases very rapidly with molecular weight: The melt behavior of the polycarbonate is also independent of shearing rate above 280°C and the polymer exhibits Newtonian behavior. The increased melt viscosity with molecular weight puts a cap on the molecular weight of the polymer which can be used in injection molding and extrusion, the most common processing method for manufacture of parts. High molecular weight of the polycarbonate would require the increased temperature for good flow resulting in thermal decomposition [39]. SO a compromise is necessary between polymers with a viscosity sufficiently high to impart optimum mechanical properties and one with a viscosity sufficiently low to impart real good flow characteristics. 2.4.3.4 Crystallinity BPA polycarbonate is an amorphous thermoplastic, but a look at the chemical structure of the polymer shows that it is symmetric and can be crystallized.Though the rigidity and rather high entanglement density leads to an amorphous nature of the polycarbonate, thin films of polycarbonate have been crystallized to some degree by prolonged heating at elevated temperatures ( 180°C for 8 days), more rapidly by treatment with solvent systems such as acetone or by stretching the film and thereby orienting the polymer chains. Crystallinity was also seen in cases where thin films are cast from poor solvents [39]. The amount of crystallization and size Of the spherulite structure decreases with an increase in molecular weight of the polymer, which are no 29 doubt associated with both increasing stiffness of the molecule and the long repeating unit. The X-ray diagram of the crystalline structure show that the molecules pack in such a way that methyl groups attached to pivotal carbon atom extend towards the back Of the carbonate group of the neighboring chain. 2.4.4 Effect of Molecular Weight Distribution. The molecular weight distribution (MWD) affects the mechanical properties and the viscosity of the polymer. The strength decreases with broadening of the distribution. This is evidently due to the smaller chain molecules acting as stress concentrators leading to stress failures. The broader MWD makes the polymer melt viscosity more non newtonian due to differential shearing of the polymer Chains, this leads to difficulties in processing. Figure 2.8. shows the effect of the MWD. The broader MWD is also undesirable in the composite materials as the lower molecular weight chains tend towards the interface reducing the strength and performance[71]. A more detailed description of the effects of segregation by molecular weight will be done in chapter 5. 2.4.5 Disadvantages of BPA Polycarbonate. The desirable characteristics of the polycarbonate mentioned in the earlier sections would lead us to believe that its usage would be very widespread. At one time it was felt that polycarbonate would become one of the important engineering materials. Such hopes have been frustrated by observation that where the parts made out of polycarbonate are 3O A 939% a \:\ 3;: Low MW \ Figure 2.8 Schematic sketch of the effect of mwd on the viscosity, note the more non-Newtonian behavior of the broader MWD 31 subjected to tensile strain of 0.75 % or more, cracking or crazing occurs, under normal static conditions. Elevated temperatures and many chemical environments also increase the amount of crazing in parts which have frozen in stresses due to molding. This puts a limitation on the amount of loading which can be applied to the molded parts, reducing the ability of the polymer to function up to its full potential. The polycarbonates are also susceptible to solvent crazing, a number of materials exist which neither attack the polymer molecules chemically nor dissolve it, such as alcohols, but when used with polycarbonate, cause cracking of the fabricated parts [39]. The likely reason for this crazing is these media have sufficient solvent action to soften the surface of the parts to such a degree that frozen-in stresses tend to be released with consequent cracking of the surface. The rigidity of the molecules also inhibit the polymer chains from completely relaxing in the time it takes a molded part to drop below the glass transition temperature. These disadvantages can be minimized by molding the parts at higher melt temperatures and annealing just below the Tg for a length of time and coating the finished products with certain protective coatings to reduce the risk of coming into contact with the undesirable chemicals. When fabricated polycarbonate parts are exposed to UV light, a progressive dulling is observed on the exposed surface. The dullness is considered to be due to microscopic cracks on the surface and the molecular weight of the polymer on the surface is significantly lower than the parent polymer leading to the conclusion that there is degradation and chain splitting due to UV light. The flexure strength and modulus remain largely unaffected but the toughness gets reduced profoundly making the parts very 32 brittle.To overcome this problem , UV stabilizers are added to the polycarbonate, though the effects of these stabilizing agents on the mechanical properties are not well known, a previous study [19] had noted that the commercial grade of polycarbonate gave a limiting value of interfacial strength in composites made with carbon fibers, the strength value was found to be unchanged unlike the values found for pure polycarbonate grade which did not have any additives in them. 2.5 Conclusion A detailed description of the properties of both the fiber and the matrix was discussed along with background and the motivation for the present investigation. The effect of the molecular weight of the polycarbonate on the other matrix properties of interest were also examined. A rationale for the selection of the fiber and the polymer matrix was evolved from this discussions and the stage was set for examining the effects of these bulk properties on the composite properties in the next three chapters. CHAPTER 3 CONSOLIDATION AND INTERFACIAL BONDING - THEORY 3. 1 Introduction The thermoplastic composite materials are manufactured by either impregnating the reinforcing fiber with polymer melt or by using a polymer solution. The high melt viscosity of the thermoplastic polymer does complicate the process and has been one of the main reasons for the limited use of thermoplastic in composite materials. In the recent years a number Of processes have been developed such as powder prepregging [46,47], commingling of fibers [48] , slurry processing [49] , aqueous dispersions [50] and aqueous foam prepregging [51] to enable the fabrication of composite parts from thermoplastic matrices. Though these processes differ from each Other in coating the fiber surface with the matrix medium, the underlying principle is the achievement of good fiber wetting for subsequent strength development. The adhesion in thermoplastic composites can be looked at in two ways, the development of strength due to the bonding at polymer-polymer interface when the composite prepregs are consolidated and the other approach deals with the increase in the composite strength due to the adsorption of the polymer chain onto the fiber surface. The first approach assumes good fiber matrix bonding implicit in its formulation, the second approach makes use of enthalpic and entropic interactions in the interphase region to explain the strength development. The former approach is based upon the polymer chain mobility and does not take into consideration the surface energetics in the interphase 33 34 region. The later approach is on molecular level and resolves the surface to chain and inter-chain interactions based upon the surface energetic considerations. In the present chapter a brief description of the underlying principles for each approach will be given along with the discussions on a model from each of them. The results of this discussions will be used to arrive at the time requirement for the consolidation of the composite specimen for conducting single fiber fragmentation and microindentation tests. The polymer adsorption approach will also be used later in explaining the results of the above mentioned tests. 3.2 Principles of Autohesive Strength Development. The adhesion between the polymer interfaces during the consolidation process in thermoplastic composites is considered to be the mechanism for the development of composite mechanical properties [52]. When two amorphous polymer surfaces are brought into intimate contact at a temperature above their glass transition temperature Tg , the chains from the two sides will interdiffuse. The interdiffusion causes the strength of the interface to increase with time until it reaches the cohesive strength of the material [53]. The bonding due to interdiffusion was first proposed by Voyutskii [54] and has gained considerable evidence has been developed primarily due to the work of Wool and coworkers [55,56,57,58]. Others [59,60,61] have also studied the interdiffusion phenomenon and proposed models similar to the one proposed by Wool. deGennes [59] and Prager and Tirell [60] based their assumptions on a similar diffusion model and related the degree of bonding to segmental crossing density. Jud et a1. [61] modeled their relation based on the assumptions that the length of the chain penetrating across the interface is proportional to the degree of bonding. All the models proposed for diffusional bonding use reptation theory for describing the dynamics of polymer chains at the interface. Although the models cannot be distinguished from each other in terms 35 of predicted time dependence, they differ in predicted molecular weight dependence. Since we are interested in understanding the molecular weight dependence, Wool’s model provides a useful framework for understanding polymer bonding and therefore is described in more detail in the following section. 3.2.1 Wool’s Bonding Model. According to Wool [56], the autohesive strength development across the interface was shown to occur in five steps as shown in figure 3.1, they are: 1.) surface rearrange- ment; 2.) surface approach; 3.) surface wetting; 4.) interdiffusion and 5.) randomization. Wool and O’ Connor [57] defined a macroscopic healing function R as the ratio of the interfacial strength a to the strength of the virgin material 0,, and equated it to the convolution product of wetting distribution function dO/dt and intrinsic healing function Rh. The relations are as shown in equations 3.1 and 3.2. R = i 3.1 r=+oo (b R = R t - t d— d1: 3.2 f ,,< ) .1. Where 1 is the tube reptation time of the polymer chain based upon the reptation theory proposed by deGennes[59], which is approximately equal to the relaxation time of the polymer chain. The intrinsic healing function Rh is dependent upon pressure, temperature and molecular weight of the polymer chain. Since the interfacial strength is the sum of both wetting strength 00 and diffusion strength ad. 3.3 36 1&2 Figure 3. 1 Schematic sketch of polymer chains at the interface in Wool et a1. model 37 If the wetting is considered to occur instantaneously as would be the case at high processing temperatures, then the diffusion will be the controlling step. The contribution of the wetting strength would also be very low in comparison with the diffusion strength as the polymer would be a viscous liquid at the processing temperature and have poor load bearing capacity. Therefore the wetting strength can be neglected from the analysis without much loss of accuracy. The reptation theory assumes chain motion of a polymer chain to be constrained in a tube. But the chain ends which are obviously less restricted in their motion compared to the middle region due to their location, have a rather more random nature of motion. Wool et al. assumed these random motions to obey Gaussian statistics[57]. They also called the portion of the chain outside the tube as the minor chain. When surface wetting occurs the minor chains interpenetrate across the interface, the interpenetration distance X was related to the time using the diffusivity of the minor length of the chain Alo— x « (0,0 3-4 Where Dc is the self diffusion coefficient. Then assuming the interfacial bond strength a to be proportional to the average interpenetration depth for t < r, , the tube renewal time aux; X=X.. whent=t 3.5 Using equations 3.4 and 3.5, the time dependence of the interfacial strength becomes 38 3.6 AIH o °< (Dct) A similar result was obtained by them when they related the strength at the interface to the number of chains crossing the interface , n(t) [58]. 3.7 n(t) cc 0 °< t”4 The final results of the analysis gives 0 cc t1/4M'5/4 for ( t < r, ) 3'8 The model developed by Wool et al. was based upon isothermal conditions at melt temperature. Loos et al. [62] modeled the bonding in a non—isothermal process by separating the process into small time increment At, assuming the healing of the interface to be essentially isothermal in this short time intervals. They proposed a model shown below in equation 3.9. R = i o, C(Dtl“ 3.9 O C(T) is a self diffusion factor dependent upon the temperature. If the temperature is just above the glass transition temperature Tg , WLF equation gives a good fit for calculating C(T), For temperatures very much above Tg ( > Tg + 50°C ) as it would be for the BPA polycarbonate at processing conditions, the Arhenius equation was found to give a good fit. 39 3.10 -Ea C(T) °< koexp[ RT] where k0 is a constant found by plotting log C(T) versus 1/T, Ea is the activation energy for the polymer to flow. Relating equations 3.8, 3.9 and 3.10, we get. —E R = £— .. :1/4M'5/4exp[—“ 0 3.11 .T] The equation 3.11 relates the strength development to the processing time, temperature and molecular weight of the polymer chains. Polymer relax in the time scales ranging from 0.01 seconds at low molecular weights to 1005 of seconds for high molecular weights. The relaxation time is dependent upon the entanglements and obviously higher molecular weight chains with more entanglement tend to relax at a much slower rate. This relaxation time is very much dependent upon the temperature and at elevated processing conditions the chains relax at a much faster rate. Therefore the autohesive strength model, though useful for the purpose of finding the bonding time between polymer surfaces during the specimen fabrication, but would be on a shorter time scale range for the adsorption of the polymer chains on to the fiber surface. This leads to the analysis in the next section. 3.3 Principles of Polymer Adsorption The autohesive strength development model is based mainly on the strength development due to the healing of polymer-polymer interface, as it was developed primarily to explain the healing process in neat thermoplastic materials. The model was 40 later utilized in case of thermoplastic composites by many researchers [ 62,63,64]. The model does not take into consideration the time requirement for the polymer chain to bond with the fiber surface, instead assumes an intimate contact and good adhesion between them. In a true composite the adhesion between polymer chains and the fiber surface would be the key factor in development of composite strength. The achievement of interfacial strength between the fiber and the matrix is assumed to occur due to the adsorption of the polymer chains on to the fiber surface, the differences in the surface free energies of the surface and the polymer chains result in enthalpic and entropic interactions which influences the interphase composition and the conformations of the polymer chain in the interphase region. Most of the theoretical work in the literature utilize statistical lattice models [65,66,67] and pseudokinetic Monte Carlo simulation techniques [68,69,70]. Through a variety of numerically intensive algorithms, the structure of the interphase at equilibrium is approximated from an ensemble of chains mixed in the presence of a surface. The approach emphasizes individual chain contributions as much as the inter-chain interactions [71]. The lattice model approach is essentially an extension of the Flory- Huggins thermodynamics to the melt Of anisotropic materials such as polymers. In the lattice representation, the molecular constitution Of a mono disperse homopolymer is summarized in three parameters [72]: 1.) The chain length (number of model segments per chain) r; 2.) The characteristic temperatures , - Zw Tb : M 21‘s k where b stands for bulk, S for the surface, IBM is the attractive energy Of cohesion between segments occupying nearest neighbor sites. ”As is the attractive energy between 41 the segment and the surface site, 2 is the, lattice coordination number, k3 is the Boltzmann constant; 3.) The characteristic pressure where v* is the molar volume of the lattice segments and R the gas constant. The enumeration of the conformations within a lattice layer parallel to the interface is assumed to be proportional to the fraction of unoccupied sites in that layer. The calculation of the potential energy due to the segment-segment interactions in an equipotential layer and its adjacent layers is proportional to the occupancy of the layer. A set of coupled non-linear algebraic equations are formed when the equilibrium conditions and self consistency requirements are imposed on the conformational distribution. A model solution is found by solving the non-linear equations using the numerical algorithms mentioned earlier, which would give the structure Of the interphase as a function layer number or the distance Z normal to the surface. The lattice models are useful for explaining the salient features of interfacial polymer structure and thermodynamics, but require intense numerical solution methods based upon complex algebraic equations for solving them McCullogh and coworkers [71,73,741 have characterized the interactions at the interphase in composites using enthalpic and entropic considerations. Their work differs from the field lattice approach, the emphasis is on using random walk theory on the bulk of polymer chains in the interphase and applying Gibbs equation for surface excess to arrive at the structure of the interphase region. The entropic interaction was shown to segregate the chains by their molecular weight and the enthalpic interactions either amplified the effect or mitigated it based upon the energy difference between surface- chain, chain-chain interactions [71]. A more detailed description of this approach is given in chapter 5. 42 Though these models give very elegant numerical solutions for finding the equilibrium conformations Of the polymer chains in the interphase, their use in arriving at a process model would be very tenuous. The use of these models help in understanding the interphase structure and would therefore be used in the later chapters to describe the experimental results obtained on the macroscopic scale at atomistic levels. Brady and Porter [75] conducted interfacial experiments on polycarbonate—carbon fiber composites and concluded that the adsorption rather than the formation of a transcrystalline zone at the fiber matrix interface, is the mechanism for the interfacial strength development. They used a Langmuir type adsorption isotherm to explain the results of their work. Using Arhenius equation to fit the kinetic parameter. Their equations are as follows. AG‘ = ———k‘ ; AG = G - G 3-13 AG (1+kt) ceq Where AGc and AGceq are changes in the transverse toughness due to annealing at a given time t and at equilibrium (t = 00), Gco is the transverse toughness before annealing, k is the kinetic constant given by [(43] 3.14 k = Aex —“ RT Where A is a constant and E3 is the activation energy for the polymer to flow. Their work gives an equations whose parameters are quantifiable and can be determined from experiments. A similar model was employed in the present investigation by modifying the relation to account for the differences in the fiber volume fractions. A detailed description of the modifications will be discussed in the next chapter while dealing with specimen fabrication. 43 3.3 Conclusions _ The theoretical models for the interfacial bonding and consolidation of fibrous composites were presented. The autohesive strength development model for polymer-polymer interfaces predicts a one fourth power relation between the interfacial strength and time. It was seen that the temperature is a very strong factor in the calculations of bonding time. The field lattice model in the polymer adsorption approach, along with the random walk approach using surface energetics, was found to be useful in explaining the interphase structure at equilibrium. A simple Langmuir type adsorption isotherm approach was also discussed and was found to be useful for calculating the time requirements for the consolidation Of the carbon fiber/polycarbonate composite specimens from experimental data. CHAPTER 4 EFFECT OF MOLECULAR WEIGHT ON INTERFACIAL ADHESION 4. 1 Introduction The effects of molecular weight of the polymer matrix on the other matrix properties were discussed in chapter 2. It was seen that, an increase in molecular weight produced an increase in the value of other properties some undesirable ones, such as viscosity and others desirable, such as resistance to solvent attack etc. The rapid increase in melt viscosity with increase in molecular weight is one of the main reasons in limiting the use of high molecular weight resins in the fabrication of finished parts by injection molding and extrusion. It is also one of the reasons for increasing the molecular weight distributions, where in lower molecular weight chains are added to help in improving the flow behavior. The use of novel fabrication techniques for manufacture of thermoplastic composites, mentioned at the beginning of the chapter 3, such as powder prepregging effectively minimize the viscosity problems during the final consolidation by drastically reducing the flow paths for the resins to flow. These techniques facilitate the use of higher molecular weight resins and its beneficial properties to augment the strength and toughness of the final composite parts. In the present chapter, the focus of the investigation will be on the effects of increasing molecular weight of BPA polycarbonate on the interfacial adhesion. The investigation will be coupled with the effects of the processing temperatures at equilibrium times. It will be shown that, the molecular weight and processing 45 temperatures are important parameters in the increase of the interfacial adhesion in the thermoplastic composite materials. 4.2 Experimental Procedures The experimental procedures for specimen fabrication and testings are descried in this section. 4.2.1 Specimen Fabrications In this investigation, ASTM D 638M tensile testing of the neat matrix specimen and single fiber fragmentation (SFF) and microindentation test on Dow Chemical Inc. Interfacial Testing Systems (ITS) of BPA polycarbonate /AS4 carbon fiber composite materials were done. This section deals with the fabrication of the test specimens for these testing methods. 4.2.13 Injection Molding The test specimens for the ASTM D 638M test were fabricated on a New Britain injection molding machine by injection molding type M-I dogbone coupons according to the dimensions specified in the ASTM standards[76]. The polycarbonate powders of each grade weighing around 10 to 15 pounds were dried overnight in forced air oven at a temperature of 125°C to remove the moisture. Even though the equilibrium moisture content of BPA polycarbonate is only 0.3% [77], the tolerable limit is only 0.2% [78] as the polycarbonate is sensitive to hydrolysis at elevated processing temperatures. Following the 125 °C drying, the polycarbonate powder were injection molded between 250 °C and 290 °C and a pressure of 100 MPa. The molded specimens were cooled in a water cooled die for 20 to 25 seconds and later allowed to cool down to room temperature in ambient atmosphere. Around 20 good specimens from each of the molecular weight grades were selected and half of them were annealed at 130°C, which is slightly below the glass transition temperature Tg , in a vacuum oven for 12 hours to 46 allow the polymer chains to relax from the stress buildups which can occur due to the injection molding operation, the other half were covered with aluminum foil and placed in the plastic ziplock bags to minimize moisture adsorption. 4.2. lb Hot Press Specimen Fabrication The single fiber fragmentation test specimens and micro indentation test specimens were fabricated according to the hot press fabrication technique described in ref.[19]. Prior to the fabrication of both SFF and ITS specimens, preform sheets, one half the thickness of the final specimens were produced by hot pressing the dried polycarbonate powder between acetone cleaned, 0.1 mm thick aluminum pressing sheets with the thickness controlled by a surrounding stainless steel dam. The typical dimensions of the preform sheets were 25 cm x 25 cm x 0.8 mm for the SFF specimen fabrications and 25 cm x 25 cm x 1.6 mm for the ITS specimen fabrications. The preform sheets were produced at 250 °C and a pressure of 500 Psi (3.45 MPa) with a residence time of 5 minutes at the consolidation temperature. The times and temperatures for this preliminary step were kept as low as possible to minimize the chances of thermal degradation. Even though the BPA polycarbonate is resistant for a long period of time to thermal oxidative degradation up to a temperature of 310 °C, still the whole assembly was enclosed in aluminum foil to reduce the risk of oxidation. . The thin aluminum pressing sheets were employed to avoid the use of release agents which might cause surface contamination. The thickness of the aluminum sheet was low enough to facilitate its easy peeling of the polycarbonate sheets without deforming them. For the single fiber fragmentation test specimens, two 15 cm x 10 cm preform sheets, which were thoroughly cleaned with isopropyl alcohol to remove any surface contaminations were used. One preform sheet was placed on a smooth surfaced 30pm thick inert high temperature Kaptan" polyimide sheet and surrounded by a 1.6 mm thick 47 Upper Aluminum Foil Upper Matrix Sheet RTV Silico Gasket Aligned Carbon Fibers Lower Matrix Sheet Spacer Strip Lower Aluminum Foil Figure 4.1 A schematic sketch showing the assembly for the hot press fabrication technique 48 stainless steel dam as shown in figure 4.1. Single separated carbon fibers were carefully taken out of a 25 cm long tow having a filament count of 6000. These single fibers were then carefully draped breadth wise across the preform sheet and metal dams, approxi- mately 5 mm apart. Care was taken to avoid fiber contamination and damage during the layup process by holding them only at their extremities. With enough isolated fibers in place, a high temperature polyimide tape was used to immobilize the fibers by taping them on to the metal dam. A razor knife was used to cut the fibers into approximately 25 mm segments with minimal disturbance. This was done to reduce fiber bending due to matrix flow and contractions. 5 mm wide spacer strips of the same matrix material were placed on either side of the lower preform sheet to maintain separation of the upper and lower sheets and avoid fiber damage during the early heating stage before Tg was reached. The upper preform sheet was placed atop the spacer strip followed by the top Kaptan' polyimide sheet. The entire assembly was sealed in a vacuum bag and carefully placed between the hot press platens of the Carver' press without disturbing the settings. Hot pressing was done according to the consolidation cycle depicted in figure 4.2. To ensure uniform heating, a light load equal to the weight of the upper platen was applied during the heating. This ioad was borne by the spacer strips and did not produce any fiber damage. The Polycarbonate was soaked for 1 hour at 125 °C to remove even trace amounts of moisture. Following the 125 °C soak, the temperature was raised at approxi- mately 5 °C per minute. Upon reaching the desired temperature, a load of 475 KPa was applied and maintained throughout the pressing cycle. The specimens were rapidly cooled at approximately 1°C per second after allowing them to reside at the processing temperature for an appropriate amount of time based upon the equation 3.15. The time requirements were calculated by fitting the values of the interfacial shear strength results obtained by Waterbury [19] for polycarbonate carbon fiber SFF specimens to the equation 3.15 using the change in strength in place of transverse toughness. Figure 4.3 49 300 775— _1500 250.3 0° 225— —1250 m— E I75- 1000: 3150.3 a ..750 125_ 100— -500 75— ’ 50— -250 25- 0 0 b I I I I I I I I I 30 60 90 120 150 m 210 240 270 mm Figure 4.2 The Consolidation cycle for making the composite specimens at 250°C 50 and table 4.1 give the values of the kinetic parameter and the time requirements at different temperatures. According to the literature [79] , the time required to reach steady state adsorption, when chains adsorb onto a initially bare surface is independent of the molecular weight of the polymer chain, the rate determining step appears to be diffusion controlled. This allows for use of the same time scales for the consolidation of different molecular weight grades at each of the processing temperatures. Table 4.1 Kinetic parameter data for hot press fabrication Activation Energy Ea Slope Intercept 9.3 x 10‘ KJ/mole "1.102 x 104 1.22 x 108 Temperature Kinetic Consolidation A0 IAa, °C Parameter Time min'l min 250 0.0862 45 0.795 260 0.1280 40 0.837 275 0.2255 15 0.772 290 0.3854 15 0.853 300 0.5424 8 0.813 315 0.8860 8 0.877 The microindentation test specimens were fabricated by placing a 15 cm x 2.5 cm x 1.6 mm thick preform sheet on a 0.1 mm thick aluminum pressing sheet and surrounding it with a 3.1 mm thick stainless steel dam. A single tow of carbon fiber, which was spread manually to a width of 2.5 cm was draped over this lower preform 51 0 7 -0.5I' ~ -1»- -. 2 U: 2 -1.S~ - .2. - -2. . L ‘ 15.65 1.7 1.75 1.8 1.85 1.9 1.95 W K . ( ) x103 Figure 4.3 The plot of UT vs log k for from the kinetic data given in Table 4.1 52 sheet. The upper preform sheet was then placed atop the carbon fiber tow, followed by the top aluminum pressing sheet. The entire assembly was sealed in an outer aluminum foil and carefully placed between the platens of the Carver° press. The processing cycle for the ITS specimen fabrication was similar to the SFF specimen fabrication, except a higher pressure of 6.9 MPa was applied during the consolidation. Sample specimens having the dimension of 2 cm x 1 cm were cut from the fabricated composite laminate. These sample specimens were then mounted in a clear polyester resin medium, ensuring that the fibers in the composite laminate are always perpendicular to the sample surface. The polyester resin is then allowed to cure for 24 hours at room temperature. After the mounting material is cured the sample is polished on the Struer Abramin polisher using 240, 360, 600, 1000, 2400 and 4000 grit sandpapers for 1 minute each at a speed of 300 rpm. After the preliminary polishing step, the samples were polished using 5 pm and 1am grit alumina powder for 1 minute each on Leco GP20 polisher. Buehler Vibromet I polisher was used with 0.05 um grit alumina powder to polish for 12 hours to give a very smooth surface finish to the ITS specimens for the Observations under the high power light microscope. The samples were later dried in an vacuum oven at 30°C for 4 to 5 hours to remove the moisture which might be adsorbed during the polishing operations. 4.2.2 Testing, Data Acquisition and Analysis The single fiber fragmentation tests were done using a specially prepared microstraining machine capable of applying enough load to the tensile coupon. This jig is fitted to the microscope stage, so that the x-y stage controls can be used to manipulate the jig positions. A photograph of the loading jig is shown in figure 4.4. The microscope used for the SFF tests was an Olympus BH-2 transmitted polarizable light microscope. The fragmentation lengths were measured and a two parameter Poisson-Weibull statistical 53 Figure 4.4 The photograph of the loading jig used in tensile testing of the SFF test specimens 54 analysis[80] based on the Kelly, Tyson shear lag model [81] shown in equation 4.1 and 4.2 was performed using the FiberTrack° program developed by Mark waterbury [19] on a Commodore-Amiga PC. On an average 8 specimens were tested to get a statistical set for each molecular weight grade and processing temperature. rzgfg 4.1 21 0 T,- = _L P 1 _ i] 4.2 20 a A detailed description of the single fiber fragmentation tests can be found in ref.[l9]. The microindentation tests were done on a commercially available Interfacial Testing System (ITS) developed by Dow Chemical Co. Inc [82]. This fully automated instrument is designed to test real composite specimens. The ITS is based on a Mitutoyo optical micrOSCOpe. A diamond tipped indenter mounted on the objective lens and is used to indent single fibers in the composite specimens. The micro indentation testings can be done either using automatic debond technique or semi-automatic debond technique. Initiation of the fiber debond is sensed by a load cell attached to the sample holder. The motion of the sample holder is controlled using a precision-controlled motorized stage with three degrees of freedom. A television camera and monitor is used to observe the debonding. The force required to debond the fiber is input to a computer program which utilizes a closed-form algorithm derived from Mandell et al. [83] finite element analysis using the methods of least squares to calculate the interfacial shear strength. The interfacial shear strength is given out as a function of the a, the axial stress in the fiber at debond; Gm, the shear modulus of the matrix; B. , the axial tensile modulus of the 55 fiber; Tm , the distance of the nearest neighbor of the tested fiber; and d, the diameter of the tested fiber. The theory and the analysis of various factors affecting the micoindentation test can be found in ref. [84]. 4.2.3 Failure Mode Image Acquisitions To document the characteristics of the failure modes in the SFF test specimens, stress bifringement patterns were acquired with a cross polarized transmitted light on polaroid films. Reflected light micrographs of the micro indentation test specimens were also acquired using reflected light on polaroid films to document the crazing and stress cracking which occur in the polycarbonate composites and to show the differences in the cracking with molecular weight. 4.3 Results and Discussions In this section the results of the tensile tests of the neat matrix specimens and the SFF tests and microindentation tests will be described and the effects of the molecular weight of the thermoplastic matrix on the interfacial properties will be discussed. 4.3.1 Tensile Test Results The ASTM D 638M tensile tests of the neat matrix specimens were done on the United Testing System UCC SFM 20 apparatus. The tensile tests were conducted at constant crosshead speed of 0.05 inches/minute, using a 1000 pounds load cell preloaded at 20,000 pounds.The strain in the coupon was measured using a laser extensiometer. The results of the tensile tests are shown in Table 4.2 As it can be noted, the tensile strength and the modulus does not vary with different molecular weight grades. The result of the tensile tests follow the predictions made in the chapter 2. The tensile strength and the modulus does not vary with molecular 56 weight, but the yield strain increases with increasing molecular weight. The increase in the strain to yield is due to increase in the entanglement density with increasing molecular weight of the polymer chain. The low molecular weight PC OQ grade showed anomalous behavior during the tensile testing. The non annealed specimens were observed to yield, where as the annealed specimens show brittle failure even before the yield point is reached. The coupons were seen to break into pieces below 1.5 % strain. Table 4.2 Tensile test result Grade Tensile Strength Tensile Modulus Yield Strain PC100 55.98 :I: 1.56 MPa 2373 i 61 MPa 5.9 % PC100-an 57.70 :I: 0.71MPa 2392 j: 63 MPa 5.9 % PC HP 56.07 i 1.87 MPa 2357 j: 122 MPa 5.5 % PC HF-an 58.63 i 0.77 MPa 2552 i 162 MPa 5.5 % PC 0Q 56.53 :t 1.07 MPa 2027 :I; 216 MPa 5.0 % PC OQ—an Variable (< 30 MPa) 2024 j: 163 MPa No Yielding The reported value of the critical molecular weight MC of the BPA polycarbonate in the literature is around 20,000 [39]. The entanglement molecular weight ME , that is the molecular weight for the polymer chains to start entangling, is known to be approximately half the critical molecular weight ( MC = M5 ) [85]. The weight average molecular weight MW Of the of the PC 0Q grade is only 16,152, which is much below the MC and the number average molecular weight MN is only 7,623, which is also below the MB. In the injection molding Operation, the polymer matrix is forced into the die at a very high speed and under a high pressure, the matrix is then cooled down to below 57 the glass transition temperature Tg rapidly.» This operation stretches and orients the polymer chains in a direction parallel to the flow path. The rapid cooling then freeze the chains in that particular orientation. The strength and ductility of brittle polymers can be greatly modified by molecular orientation of the polymer chains. The tensile strength of the rigid polymer is known to increase in the direction parallel to the uniaxial orientation [41]. The polymer is also seen to become ductile and have a yield point and a high elongation, this must have been the case with the PC OQ grade. The 130°C annealing must have allowed the polymer chains to relax and lose the uniaxial orientation attained during the injection molding operation. Due to its low molecular weight and consequent- ly low entanglement density, the annealed specimens would have shown the brittle behavior. 4.3.2 Adhesion Tests Results The single fiber fragmentation tests were conducted on the low molecular weight PCOQ grade and the higher molecular weight PC100 and PCHF grades only. The intermediate molecular weight PCFF grade was obtained at a much later stage and it was decided, based upon the results Obtained with the other three molecular weight grades that, the result of the PCFF SFF testing will only be statistically significant without adding very much to the knowledge of the interfacial behavior as this particular grade has a molecular weight range in between the PC100 and PCHF molecular weight ranges. The PCFF grade was however used in the ITS tests. The SFF tests were done on samples prepared at 500°F (260°C), 550°F (290°C) and 600°F (315°C), while the subsequent ITS tests were done on samples prepared at 250°C, 275°C and 300°C. Though these differences in the processing temperatures hindered one to one mapping of the SFF results and ITS results. they proved to be fortuitous due to the fact that, interfacial shear strength data was now available at more number of 58 processing temperatures, ranging from 250°C to 315°C. The results of the single fiber fragmentation tests for the different molecular weight grades and processing tempera- tures are shown in table 4.3 and figure 4.5 Table 4.3 Single fiber fragmentation test results Grade Molecular Processing Processing IFSS Weight MW Temperature Time MPa PC OQa 16,152 250°C 20 minutes 35.26 :1; 3.17 PC HFa 22,963 260°C 40 minutes 39.48 i 2.76 PC HFb 22,963 290°C 15 minutes 46.82 :1: 3.14 PC HFc 22,963 315°C 8 minutes 45.32 :1; 3.34 PC100a 31,135 260°C 40 minutes 38.61 3: 2.78 PC100b 31,135 290°C 15 minutes 51.66 :I: 3.98 L. PC100c 31,135 315°C 8minutes 51.43 :t 3.60 \ The results of the microindentation tests for all the molecular weight grades and processing temperatures are shown in table 4.4 and figures 4.6 - 4.9. A consolidated three dimensional figure showing the interfacial shear strength (IFSS) with respect to both temperature and molecular weight is shown in figure 4.10. It can be seen from both tables 4.3 and 4.4, that as the molecular weight increases, the interfacial shear strength (IFSS) at each of the processing temperature increases. Figure 4.6 shows the dependence of IFSS on processing temperature based upon the ITS results. The interfacial shear strength calculations in the single fiber fragmentation test is based upon the fiber strength of as can be seen from equations 4.1 and 4.2, while the Dow Chemical’s ITS uses the matrix shear modulus Gm and the fiber tensile modulus Tm to arrive at the interfacial shear strength. It is of interest to note that the results of 59 60 [j PCHF 50 _I_1 + + [:I PC100 40 re ~ , . T 17 E V30 H H m U) U) E 20 ~— — —« 10 H — —« o . . 260°C 290°C 315°C TEMPERATURE Figure 4.5 Interfacial shear strength of the polycarbonate grades with two different molecular weight, note the increases due to the molecular weight 60 the SFF tests and the ITS tests give similarvalues of IFSS for the PC100 and PCHF samples as represented in figure 4.10. Table 4.4 Micro indentation test results Grade Molecular Processing Processing IFSSifl-I Weight Mw Temperature Time MPa PC OQa 16,152 250°C 45 minutes 20.22 :1; 2.51 PC OQb 16,152 275°C 15 minutes 21.23 :I: 2.72 PC OQc 16,152 300°C 8 minutes 21.13 :t 2.76 PC HFa 22,963 250°C 45 minutes 34.07 :I; 5.08 PC HFb 22,963 275°C 15 minutes 37.54 i 4.61 PC HFc 22,963 300°C 8 minutes 42.01 i 5.09 PC FFa 24,894 250°C 45 minutes 34.23 i 3.75 PC FFb 24,894 275°C 15 minutes 41.30 :i: 3.16 PC FFc 24,894 300°C 8 minutes 43.14 :1: 3.71 PC100a 31,135 250°C 45 minutes 36.43 :1: 4.31 PC100b 31,135 275°C 15 minutes 47.80 :I: 5.68 PC100c 31,135 300°C 8 minutes 48.32 :1: 6.83 The increase in the IFSS is very large between the SFF specimens fabricated at 260°C and 290°C, but the value is seen to slightly decrease at 315°C. A similar trend is seen between the ITS specimens fabricated at 250°C and 275°C. The IFSS for ITS specimens fabricated at 300°C doesn’t increase very much over the IFSS values at 275°C for all the molecular weight grades. It can therefore be inferred that, the optimal processing temperature window for the BPA polycarbonate carbon fiber composite lies between 275°C and 300°C, requiring lower times at higher processing temperatures for attainment of the similar levels Of interfacial adhesion. 61 60‘ + M w =31135 40“ / + M w = 24393 " 1 530-: a 3 M =22963 t + i + w 10‘; .4... Mw=l6152 0: 250 275 300 TEMPERATURE °C Figure 4.6 The effect of the processing temperature on the interfacial shear strength for various molecular weight ranges 62 60 50, [:1 Mw=16152 3:40; ’ Mw=22963 E : U230: E3 : Mw=24893 "‘ 20 g .L L 10—2 _. I Mw=31135 0- Figure 4.7 The interfacial shear strength of the molecular weight grades processed at 250 0C for 45 minutes 60 50 T C] Mw=16152 3340: T Mw=22963 z : 3330-: E20_3 g Mw=24893 10‘; I Mw=31135 0- Figure 4.8 The interfacial shear strength of the molecular weight grades processed at 275 °C for 15 minutes 63 60 , 50 E I Cl M w = 16152 a: 40: t a Mw=2ms E : M w = 24893 10‘ i I M w = 31135 0 g Figure 4.9 The interfacial shear strength of the molecular weight grades processed at 300 °C for 8 minutes VA \YSSMY'A v) Figure 4.10 The three dimensional representation of the effects of the molecular weight and processing temperatures on interfacial adhesion 64 A ~.w\\\\\\\\\\\\\\\\\\\\\\s¢ P s\\\\\\\\\\\\\\\\\\\\\\\\\\ testing systems. Note the similarity in the trend Figure 4.11 A comparision of the interfacial shear strength obtained from the two 65 The decrease in the IFSS of SFF specimens processed at 315°C is probably due to some degree of polymer degradation occurring due to the excessively high processing temperature. As mentioned earlier in chapter 2, the polycarbonate starts degrading rapidly above 310°C. The increase in the IFSS value in both the testing methods with the processing temperature between 250°C and 300°C is evidently due to better wetting of the fiber surface by the polymer melt at higher temperatures. The criterion for a liquid to wet a solid surface is the requirement of positive difference between the surface free energies of the solid 71 and liquid 72 [86] y, - yz cos(0) = 712 2 0 4‘4 Larger values of 712 would increase the wetting of the solid by the liquid. The surface free energy of the polycarbonate is lower than the surface free energy of the A84 carbon fiber (71/ 72 = 1.17) [87]. There is also a decrease in the surface free energy of the polymer melt as the temperature increases[88]. It is well known, that the increase in processing temperature would decrease the viscosity of the thermoplastic melt. The BPA Polycarbonate is also known to start exhibiting Newtonian flow behavior above 280°C [38,39] , which would improve its flow characteristics. The surface roughness would also help in increasing the wettability[88], it also plays an important role in better adhesion at higher temperatures. The lowering of the viscosity would allow the polymer chains to relax from their coiled state and conform more closely to the micropores and other surface topographical reliefs present on the carbon fiber surface, thereby helping in increasing the mechanical locking of the chains to the surface. These two factors, viz. the increased surface free energy difference and the lowering of the viscosity tend to increase the wettability of the carbon fiber surface by the polymer melt and the IFSS at higher temperatures. 66 The low molecular weight grade, PCOQ was found to be extremely brittle. Most of the SFF specimens were seen to crack even while they were being punched in the die to form the dogbone coupons. Only a few specimens (less than 50%) fabricated at 250°C and for a much lower amount of time (20 minutes) were able to withstand the loading and arrive at a critical length of the fiber. Specimens prepared at temperatures above 250°C failed by cracking before the first fiber break could occur at around 1.5 % strain.The failed surface of the cracked specimen figure 4.14, shows the reinforcing carbon fiber exceeding more than 1mm in length having been pulled out of the matrix medium. In the few specimens, which were able to reach the critical length in the SFF test the polycarbonate matrix would have transferred the stress to the carbon fiber due to a higher level of the friction between matrix and the carbon fiber due to the tensile loading. Carafagna et al. [89] noted a similar behavior in case of polycarbonate/liquid crystalline polymer composites. The matrix was seen to adhere poorly to the fiber. The reinforcing fibers were seen to fragment only in the necking region of the specimen. The ability of the matrix to transfer the stress in this necking region was attributed to an increase in the level of friction due to compression and shear stress allowing for better mechanical interlocking between the fiber surface roughness and the matrix chains. The micro indentation tests on the ITS specimens also show the low levels of adhesion between the PCOQ grade and the A84 carbon fiber. The increase in the processing temperatures is also seen to have no effect on the IFSS. The increase in the IFSS with increase in the molecular weight can be explained using the dynamics of polymer adsorption onto the carbon fiber surface and the possible dispersive, polar, chemical and entropic interactions between the fiber surface and the matrix. These interactions tend to change the conformation and the structure of the polymer chain in the interphase region, thereby changing its properties. The final structure of the interphase would then govern most of the macroscopically observed 67 behavior of the polymer composite. The enthalpic and entropic interactions between the fiber surface and the polymer melt act as the driving force for the polymer chains to diffuse to the interface. These interaction also tend to control the adhesion between the fiber and the polymer chain. In contrast to the monomolecular species, it is unlikely that all of the monomeric segments of a polymer chain will be simultaneously in contact with the fiber sur- face[67,85]. The measure of adhesion between the polymer chain and the fiber can be thought of as being dependent on the number of contact points between the active sites on the surface and the chain as shown in the figure 4.12. The portions of the polymer chain which adsorb onto the surface are called trains, the parts which protrude into the bulk are called loops and tails. Due to the constraints imposed on the number of conformations a polymer chain can have at the interface, most of the chains tend to flatten out to occupy a lower energy state[69,90] as shown in figure 4.13. This chain segmental orientation and conformations are seen to extend up to 3.5 times the radius of gyration R1; of the longest polymer chain [71]. Even though, the number of polymer chains per unit area of the surface will be lower in case of the higher molecular weight polymer chains due to their longer chains lengths, the number of contact points would be more. The next layer adsorbed on top of these polymer chains would also be able to entangle more with the loops and tails and create a strong adhesive and cohesive bonding between the surface and the bulk matrix medium. The lower molecular weight polymer chains were seen to be much less affected by the presence of the interface and therefore do not lose as many number of conformations as a higher molecular weight chain [71]. Consequently the lower molecular weight chains tend to still retain a major part of their isotropic bulk conformation. This would reduce the number of contact points between the polymer chain and the surface leading to lowering of the interfacial adhesion. The low amount of the entanglement density also would contribute to the lowering the ability of 68 WWMWCEAINS wwwumm oouncrram's Figure 4. 12 A schematic representation of the polymer chain adsorbing on to the fiber surface @ {—— BULK communion m TAIL \ ‘ OAJ‘ m... Figure 4. 13 A schematic sketch of the polymer chain losing conformation as it nears the interface 69 the low molecular weight polymer to transfer the stress applied onto it, to the reinforcing fiber. In the previous discussion it was seen that, the higher molecular polymer chains would tend to adhere at more number of points to the fiber surface compared to the lower molecular weight chains. It was also seen that due to increased level of interfacial influence, longer polymer chains tend to open and flatten out on the surface compared to the small chains. The dispersive interaction tend to be the most important type of enthalpic interaction between many of the thermoplastic matrices and carbon fiber surface[7l]. The other type of interactions which are possible are polar bonds and hydrogen bond formations. According to Fowkes[9l] the work of adhesion between two surfaces is 4.5 d d W12 3 ZVYIYZ Where the superscript d denotes dispersive type of interactions. Kaelble [92] proposed a modification to the Fowkes equation by adding the increase in the work of adhesion due to polar component. 4.6 d d P P W12 2 2\IYI'Yz + 2II'YIYz The BPA polycarbonate contains polar carbonyl groups surrounded by bulky phenyl and methyl groups and therefore tends to behave like a non polar polymer in bulk conditions. The uncoiling and flattening of the longer chains of higher molecular weight at the interface would allow the polar components to interact with the surface functionalities present on the fiber surface and increase the work of adhesion. These 70 interactions tend to increase the level of adhesion in higher molecular weight specimens compared to the lower molecular weight chains. The increase in the interfacial adhesion with the increase in the molecular weight can also be explained using the relation proposed by Cox [28] and Cooke [29] and given in the equation 1.1 of chapter 1, G... °°’ SinhB(0.5L-x) Ef em R L 2Ef ln(—) COShBE 4.7 I' The relation shows the dependence of the interfacial strength on the matrix properties - the strain in the matrix 3m and the matrix shear modulus Gm. Since the matrix shear modulus remains unchanged with increase in the molecular weight of the matrix, the interfacial adhesion would therefore be dependent upon the strain in the matrix. Matrix yield strain increases with increase in molecular weight due to higher entanglement density, therefore the maximum interfacial strength at yield would be proportional to the molecular weight of the matrix medium e «x MW; r o: MW 4,3 m The bifringement patterns of SFF test specimens consolidated at different processing conditions are shown in the figures 4.14 - 4.19 One of the striking aspects was the similarity of the patterns between the two molecular weight grades. The bifringement patterns were seen to be very strong. The intensity of the bifringement pattern indicate interfacial failure with plastic deformation, which is attributed to good interfacial shear strength [19]. The plastic deformation occurs due to a high amount of strain energy release when the fiber cracks inside the coupon due to increasing tensile 71 Figure 4.14 The cracked surface of the PCOQ grade SFF specimen, the fiber pulled out due to the cracking can be clearly seen 72 Figure 4.15 The bifringement pattern of the PCHF grade processed at 260°C for 40 minutes ” Figure 4.16 The bifringement pattern of the PC100 grade processed at 260°C for 40 minutes 73 Figure 4.17 The bifringement patterns of the PCHF grade processed at 290°C for 15 minutes. Figure 4.18 The bifringement pattern of the PC100 grade processed at 290°C for 15 minutes =3: Figure 4.19 The bifringement pattern of the PCHF grade processed at 315°C for 8 minutes Figure 4.20 The bifringement pattern of the PC100 grade processed at 315°C for 8 minutes 75 Figure 4.21 The micrograph of the ITS specimen surface of the PCOQ grade composite. Note the deep cracks in the fiber plane. Figure 4.22 The micrograph of the ITS specimen surface of the PCHF grade composite. The amount of cracks are less than the ones seen in PCOQ grade 76 Figure 4.23 The micrograph of the ITS specimen surface of the PCFF grade composite. The crack growth is seen to be decreasing Figure 4.24 The micrograph of the ITS specimen surface of the PC100 grade composite. Note the low amount of interfacial cracking in this grade 77 stresses. The SFF specimens consolidated at 315°C show a larger area of matrix deforma- tion, this could possibly be due to matrix degradation during the high temperature processing. The micrographs of the ITS specimens are shown in figures 4.20 - 4.23. These figures can be used to visualize the extent of the fiber matrix interactions in the polycarbonate carbon fiber composites. The composite specimens developed surface cracks through the high fiber content region. The formation of these cracks could be due to the stress relaxation of the polymer chains which are constrained by the presence of the carbon fibers. These cracks were seen to occur when the surface of the specimen was wiped with ethanol, which is used to clean the grits and other impurities present on the surface as water cleaning usually does not completely remove these surface impurities. Though ethanol does not dissolve polycarbonate, it is speculated that it might have a slight amount of solvent action to soften the surface [38] , so that the stresses which get built up just under the surface were released with crazing and cracking of the matrix medium. A random set of experiments were carried out to on a few of the specimens to check if the results obtained were true and not artifacts of the solvent effects of the ethanol. It was found that, the specimens whose surfaces were cleaned only with water gave a similar level of adhesion. Indicating that the ethanol did not have any effects on the interfacial adhesion. 4.4 Conclusions The polycarbonate/ carbon fiber composite fabricated from four different molecular weight grades and at different processing temperatures were investigated. It was observed that, the interfacial shear strength increases with increase in both molecular weight and processing temperature. The increase is seen to be huge between the low molecular weight PCOQ grade, which was shown to be close to the critical molecular weight and 78 the next intermediate PCHF grade. The increase is much less profound above the critical molecular weight, though there is a monotonous increase with increasing molecular weight. It was seen that, a process window with the temperature ranging between 275°C and 300°C is required to obtain optimal composite strength in polycarbonate composite materials. A rationale for the increase in the level of adhesion was proposed based on the polymer wettability, reduction in viscosity, the chain size, its conformations in the interphase region and also the interactions between fiber surface and the polymer chains. The bifringement pattern analysis indicates good adhesion between the carbon fiber and the pure polycarbonate grades PC100 and PCHF. The ITS composite micrographs show differences in the crack size and propagation with molecular weight indicating a low level of adhesion between PCOQ grade and AS4 fiber. The cracks were also seen to decrease in size showing better adhesion as the molecular weight is increased. CHAPTER 5 EFFECT OF INTERFACE ON SEGREGATION BY MOLECULAR WEIGHT 5. 1 Introduction. In the previous chapter, the effect of the molecular weight of the polymer matrix on the interfacial properties of the carbon fiber composites were studied. It was found that the interfacial properties are dependent upon the molecular weight of the polymer matrix, the interfacial adhesion increases with increasing molecular weight. It has been shown theoretically [71], that the interface tends to segregate the polymer chains by their chain size, which is related to its molecular weight. Therefore when a polydisperse polymer medium is used the interphase region would have a different molecular weight distribution compared to the bulk. The segregation by molecular weight is a long term phenomenon typically taking hours in the melt condition to reach the equilibrium. This time period is very long compared to the processing time for most composites. Even so the segregation by molecular weight would be very important in cases where the composite has to be used at elevated temperatures or undergo repeated heating cycles. The segregation is also seen to occur at a much faster rate, on the order of minutes when polymer solutions are used. Hence the segregation effects would become immediately evident in solution impregnated composites. In the present investigation, macroscopic scale experiments were conducted to quantitatively verify the segregation by molecular weight at the interface. The BPA 79 80 polycarbonate PCOQ and PC100 grades were used for the investigation. The average IFSS of these two grades at 275°C is 21.23 MPa and 47.80 respectively, a difference which is more than 100%. The microindentation tests using the Dow Chemical Company’s Interfacial Testing Systems was employed to acquire the interfacial shear strength data. 5.2 Theoretical Background The driving force for the polymer chains to wet the fiber surface and subse- quently adhere to it is due to the surface free energy differences between them. The fiber surface has a higher surface free energy than the matrix and the enthalpy of adsorption is exothermic. According to Pangelinan [71], the loss of entropy suffered by a higher molecular weight chain is greater than that of a lower molecular weight chain as it is brought to the fiber matrix interface. He called the surviving fraction of conformations a polymer chain can have at the interface as persistence function ¢(z). Equating the volume fractions of the polymer chains of different molecular weight at the interface v(z) to the bulk v(oo), for a polydisperse matrix distribution gives vj(oo) 5.1 Vm(°°) vj(z) _ ,-(z) vm(z) 43mm = exp (xj—xm) [L _ I] §(z) 34:) Where _ 5.2 x(z) = Z xm,(z) mx(z) is the mole fraction of the chains with x bonds, xj and xmax are the chains with j segments and maximum segments respectively in the distribution. The ratio of the persistence 81 function reflects the difference in the relative number of the conformations available to species. Since a large chain will always have a larger reduction in the conformations available to it than a smaller chain, the ratio will always be greater than unity. Thus the volume ratio will favor the lower molecular weight species at the interface. A schematic plot of the segregation by molecular weight is shown in figure 5.1 Theoretically, the time requirement for the segregational equilibrium to occur would be in hours. The reason for the prediction of the long time for the equilibrium to occur is basically due to the sluggish rate of diffusion of the polymer chains in the melt condition. The slow rate of diffusion is due to the high melt viscosity of the polymer medium. A look at the self diffusion coefficient presented in equation 5.3 [63], would show this dependence upon the melt viscosity. 135 GS M M200(M=Mc) ’3 (.qu M. where GON is the shear modulus of the matrix in the rubbery plateau region, p is the melt density, < r2 > is the average square of end to end distance of the polymer chain, M is the molecular weight of the polymer chain, MC is the critical molecular weight, 170 is the zero shear viscosity of the matrix melt. Typically the melt viscosity of polycarbonates are around 103 Pa-s and the diffusivities of the polymer chains in the melt would around 10'8 to 10‘10 cm2 /sec. If the fibers are however coated with a low concentration solution of the polymer, the viscosity would be about six to seven orders of magnitude less than the melt viscosity. Consequently the diffusivities of the polymer would be very high and the segregational equilibrium can be acheived in minutes in solution. 82 0.75 HIGHER MOLECULAR WEIGHT 0.45 — 0.30 _’ ‘ VOLUME FRACTION ." —‘ o ~-~--' "’ 0 0.5 1 .0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 NORMALIZED DISTANCE C= z/az Figure 5.1 A schematic plot of the segregation by molecular weight at the interface. Note the decrease in the volume fraction with increasing MW 83 5.3 Specimen Fabrication The composite specimens for the microindentation tests were prepared essentially in the same manner as was described in the section 4.2.1b of the chapter 4. The A84 carbon fibers were however coated with either the low molecular weight grade matrix or the higher molecular weight matrix, to be later consolidated by pressing them between a different molecular weight gradepolycarbonate in film form. A single carbon fiber tow was first cut to a length of 45 cm and spread manually to a width of 2.5 cm, without touching the fibers anywhere except the extremities. The spread fibers were then immobilized in place using adhesive tapes, on two 3 cm wide stainless steel strips, which were kept a distance of 30 cm apart and later cutting off the excess 15 cm of the fibers. The spread fiber tow was then placed in a solution bath containing a 2 % w/w solution of polycarbonate in methylene chloride. The whole bath was gently agitated for the solution to completely wet the fiber surface. After agitating for 15 seconds, the fibers were allowed to soak in the bath for upto 15 minutes. After the soaking Operation, excess solution was drained away and the coated fibers were placed in a heated oven at 80°C for 30 minutes, which was found to be sufficient for the solvent to completely devolatize [19]. Fibers coated with the low molecular weight grade (MW = 16,152) were then placed between the higher molecular weight grade (MW = 31,135) preform sheets and consolidated for 5 minutes, 15 minutes, 30 minutes and 45 minutes at 275°C. Similarly, fibers coated with higher molecular weight grade (MW = 31,135) were consolidated between the low molecular weight grade (MW = 16,152) preform sheets for 5 minutes, 15 minutes and 30 minutes at 275°C. Specimens consolidated for 45 minutes using PCOQ grade bulk medium were found to be excessively brittle and unsuitable for testing. 84 5.4 Results and Discussions The results of the experiments conducted to verify the segregation of the polymer chains by molecular weight are shown in table 5.1 and figures 5.3 and 5.4. Figure 5.5 gives a three dimensional representation of the results. Table 5. l ITS test results for verifying segregation by molecular weight at interface Sample Coating Bulk Consolidation Time IFSS Designation MW MW Temperature minutes (MPa) PCOQb 16152 16152 275°C 15 21.23 1; 2.72 PCOQ100a 16152 31135 275°C 5 17.60 i 2.64 PCOQlOOb 16152 31135 275°C 15 21.54 i 3.18 PCOQ100c 16152 31135 275°C 30 23.30 :1; 3.49 PCOQlOOd 16152 31135 275°C 45 23.70 i 3.60 PCIOOOQa 31135 16152 275°C 5 32.63 3: 4.37 PCIOOOQb 31135 16152 275°C 15 28.98 :I: 4.13 PCIOOOQc 31135 16152 275°C 30 23.40 :i: 3.63 _ PC100b 31135 31135 275°C 15 47.80 j:_5__6i It can be clearly seen that, the composite specimens fabricated using low molecular weight (PCOQ) coating on the fiber, show very low values of the IFSS for all the consolidation times, though there is a slight increase in the value between the specimen fabricated for 5 minutes and 15 minutes, which could be possibly due to a better wetting. It is also evident from the values of the IFSS, that the equilibrium was essentially reached between 15 and 30 minutes of consolidation. The IFSS for these composites are also nearly equal to the IFSS of PCOQ grade composite consolidated at 275°C, indicating that the low molecular weight chains tend to stay at the interface. The 85 IFSS OF LOW MW COATED FIBER IN HIGH MW BULK MATRIX 30, 25— 20— 15— IFSS MPa 104 15 30 CONSOLIDATION TIMES (MINUTES) Figure 5 .2 The interfacial shear strength of low molecular weight PCOQ coated AS4 fibers consolidated in higher molecular weight PC100 bulk matrix IFSS OF HIGHER MW COATED FIBER IN LOW MW BULK MATRIX 40 35- I 304 1 I a 25- 1 I :- 52..- I a —115_. 104 5- 0 5 15 30 CONSOLIDATION TIMES (MINUTES) Figure 5.3 The interfacial shear strength of higher molecular weight PC 100 coated AS4 fibers consolidated in low molecular weight PCOQ bulk matrix 86 x + 050505050 5443322115000%% L SEA»? 53...... .A I III L. _ 4L. A??? .hln1 .. / ////VV..;.% 1. I. a .4... d 0 5 In} 4 09.025050150’. 1433 1 7t 71 1 5 Vi mm W/ %%o e av 30»? A s Figure 5.4 A three dimensional plot of the data given in table 5.1, comparing the IFSS of the coated and the uncoated fibers 87 composite specimens fabricated from higher molecular weight grade (PC100) coated fibers, show a decrease in the IFSS with increasing consolidation time. These results indicate a probable segregation by molecular weight at the interface. The results of the present investigation, indicates a time scale of less than an hour for the segregational equilibrium though the samples were fabricated by consolidating under melt conditions. The reasons for the difference in the time scale of the experiment compared to the theory could be due to the following reasons. In the present investigation, the AS4 fibers were coated using a very dilute solution (2% w/w) of the polycarbonate in MeC12 (solvent viscosity 2 44 x 10'5 Pa-s at 20°C). The solution viscosity would be of the order of 10'3 Pa-s , which is six order of magnitude lower than the melt viscosity. The polymer diffusivities would therefore be much higher and the chains would have therefore reached segregational equilibrium during the coating operation itself. The coating of the fiber using a very dilute solution would have resulted in a very thin coating of the order Of 1000 A. If the molecular weight fraction below 4,000 is. considered to be highly mobile, the low molecular weight grade PCOQ (MW = 16,152) is composed of 20 % of this fraction, whereas the high molecular weight grade PC100 (MW = 31135) has only around 7%. This quantity of low molecular weight material would probably be insufficient to completely coat the fiber surface up to a thickness of 1000 A. However, when fibers coated with the higher molecular weight matrix are processed for longer periods of time, the low molecular weight segments in the bulk will have sufficient time to diffuse to the interface and reduce the IFSS. This is supported by considering the ratios of the diffusivities of the polymer chains. D. M j3 5'4 D. 3 2 88 This relation was derived from equation 5.3. Here 0, and 01- are the parameters such as diffusivity and molecular weight of polymer chains with i and j number of monomer units respectively. A relation of this form is possible due to the fact that GON is independent of molecular weight [93] and the zero shear viscosity no is given by [63] M 5.5 no : HAM)? fOI'M 2 Me C From the ratio shown in equation 5.4, it is clear that, the rate of diffusion of the polymer chains is related to the third power of the molecular weight and therefore the lower weight polymer chains would diffuse much faster than the higher molecular weight chains. z 472 5'6 D4000 .. (31135]3 Dams 4000 This ratio indicates that the low molecular weight chains would have sufficient time to diffuse up to the interface in the specimens consolidated for longer times and hence reduce the IFSS. 5.5 Conclusions In the composites fabricated with low molecular weight polymer coating it was observed that the interfacial adhesion is poordue to the low strain to failure of the low molecular weight fraction. On the other hand the higher molecular weight polymer coating although producing a high initial value Of adhesion, was seen to diffuse away from the interface when consolidated with a lower molecular weight medium as the bulk giving rise to lower adhesion. 89 The consequences Of this phenomenon is very important in composite manufacturing. Most of the commercially available thermoplastic polymers have a broad molecular weight distribution, having the average molecular weights high enough to obtain good mechanical properties but with a significant amount of lower molecular weight chains to promote mobility for ease of processing. The segregation of lower molecular weight chains at the interface can produce a weak and brittle layer, which would reduce the interfacial strength of the composites fabricated from these commercial grade polymers. CHAPTER 6 CONCLUSIONS AND RECOMNIENDATIONS 6.1 Conclusions The present work is a part of ongoing intensive research on determining the effects of the interphase properties on the final composite properties. The results of the previous studies have shown that, the matrix properties have significant influence on the interphase behavior. The present investigation dealt with the effects of molecular weight of the thermoplastic matrix and the processing temperature on the interfacial adhesion in carbon fiber composites. The results Of the investigation conducted on four different molecular weight grades at three different processing conditions using two independent testing methods, show that the molecular weight and the processing conditions significantly affect the interphase properties. The mechanical testing of the neat matrix material showed that, the lower molecular weight grade (PCOQ; MW : 16152) was brittle and failed prematurely even before reaching the yield point. It was also observed that all the specimens Of different molecular weight grades had very similar values of modulus. The higher molecular weight grades yielded at slightly higher value of strain. The results followed the predictions in the literature. The increase in the strain to yield with increasing molecular weight was due to the increase in the entanglement density. From the adhesion test results, it was observed that, below a critical molecular weight, the level of adhesion was very poor. The differences in the level of adhesion 9O 91 between the low molecular weight grade and the next higher molecular weight grade, which had an average weight average molecular weight higher than the critical molecular weight, was between 75 % to 100 %. This was a significant increase. Above the critical molecular weight, the interfacial adhesion increased with increasing molecular weight but at a much slower rate. The increase was between 7 % to 40 % based upon the processing conditions. The increase in the levels of adhesion with increase in the molecular weight is possibly due to increase in the number of contact point between the fiber surface and the polymer chains. whose size is dependent upon the molecular weight. The entanglements between the polymer chains also would be higher for higher molecular weight and contribute towards the increase in the adhesion between the chains adsorbing on to the surface and the chains in the bulk. The processing temperature had no effect on the level of adhesion for the low molecular weight grade. The interfacial shear strength showed no significant increase with increase in the processing temperature. Unlike the low molecular weight grade, the other three molecular weight grades show an increase in the interfacial adhesion, when the processing temperature was increased and the composites were consolidated for an amount of time, which was significantly close to the equilibrium. The increase in the strength ranged between 25 % to 35 % for different grades. The increase in IFSS with the increase in the processing temperaturis a result of polymer chains to conforming more closely with the micropores and other topographical features present on the fiber surface, leading to better mechanical interlocking with the surfaces. The interface was also seen to significantly affect the segregation of the polymer chains by their molecular weight. It was observed that, the lower molecular weight chains tend to remain at the interface if placed there and to diffuse to the 92 interface if given sufficient time and temperature and therby decrease the level of adhesion. The consequences of this phenomenon is very important in composite manufacturing. Most of the commercially available thermoplastic polymers have a broad molecular weight distribution. The average molecular weights are high enough to obtain good mechanical properties whith a significant amount of lower molecular weight chains to promote mobility for ease of processing. The segregation Of lower molecular weight chains at the interface would reduce the interfacial strength of the composites fabricated from these commercial grade polymers towards the values associated with the low molecular weight fraction. 6.2 Recommendations for Future Work. The present investigation used polycarbonate grades whose molecular weight composition was similar to the commercially available grades. The work in this thesis is a first step inunderstanding the effects of the molecular weight on the interfacial behavior in thermoplastic composites. The results presented give a trend. To further the understanding and to evolve a model to predict the influence of the molecular weight of the thermoplastic matrices on the interfacial behavior of a composite material, a set of experiments using monodisperse matrix medium should also be conducted. This would provide the experimental data to accurately quantify the interfacial adhesion based upon the molecular weight. In the course of the investigation it was postulated that the increase in the surface energy difference between the matrix and the fiber along with the surface roughness could have enhanced the interacial adhesion. Further studies need to be conducted with fibers having different surface energetics and surface morphology to increase the understanding of the fiber matrix adhesion in thermoplastic composite materials. BIBLIOGRAPHY 10. 11. 94 BIBLIOGRAPHY Hull, D., "An Introduction to Composite Materials", Cambridge University Press, New York, 1981 Iyer, S. R., "Continuous Processing of Unidirectional Prepreg", Ph.D. Dissertation, Michigan State University, Department of Chemical Engineering, November, 1990 Hergenrother, P. M., and Johnston, N. 1., "Status of High-Temperature Laminating Resins and Adhesives", Resins for Aerospace, pp 3-13, ACS 1979 Dorey, G., "Carbon Fibres and their Applications", J. Phys. D: Appl. Phys, (Printed in the UK), Vol. 20, pp 245, 1987 Hughes, J. D. H., "The Evaluation of Current Carbon Fibres," J. Phys. D: Appl. Phys, (Printed in the UK), Vol. 20, pp 276, 1987 Fitzer, E. and Rensch, H. P., "Carbon Fibre Surfaces and their Analysis," in, Controlled Interphases in Composite Materials (Edited by H. Ishida), Third International Conference on Composite Interfaces, Elsevier, New York, 1990 Donnet, J. B. and Bansal, R. C., "Carbon Fibers", Marcel Dekker, Inc., New York, 1984 Hammer, G. E. and Drzal, L. T., "Graphite Fiber Surface Analysis by X-Ray Photoelectron Spectroscopy and Polar/Dispersive Free Energy Analysis", Appl. of Surf. Sci, Vol. 4, pp 340, 1980 Denison, P., Jones, F.R, Watts, IF, "A Quantification and Investigation of Surface Micro-porosity in Carbon-Fibres Using Labelling and XPS", Surface Interface Analysis, Vol. 12, pp. 455-460, 1988 Katagiri, G., Ishida, H., Ishitani, A., "Raman Spectra of Graphite Edge Planes", Carbon, Vol. 26 (4), pp. 565-571, Jan. 1988 Kavanagh, A. and Schlogi R., "The Morphology of Some Natural and Synthetic Graphites", Carbon, Vol. 26 (1), pp. 23-32, Jan. 1988 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 95 Drzal, L.T., Rich, M.J., Camping,_ JD. and Park, W.J., "Interfacial Shear Strength and Failure Mechanisms in Graphite Fiber Composites", Reinforced Plastics/Composites Institute Proc. 1980, Sect. 20-C, pp. 1-7, 1980 DiBenedetto, A.T., "Evaluation of Fiber Surface Treatments in Composite Materials", Pure & Appl. Chem, Vol. 57 (11), pp. 1659-1665, 1985 Jaworske, D.A., Gaier, J.R., Hung, C.C., Banks, B.A., "Properties and Potential Applications of Brominated P-100 Carbon Fibers", SAMPE Quarterly, 18 (1). Pp. 9-14, Oct. 1986 Folkes, M.J. and Wong, W.K., "Determination of Interfacial Shear Strength in Fibre Reinforced Thermoplastic Composites", Polymer, 28, pp- 1309-1314, Jul. 1987 Hsiao, BS. and Chen, E] H "Study of Transcrystallization in Polymer Composites", Mat. Res. Symp. Proc, Vol. 170, pp. 117-121, Jan. 1990 Nardin, M., Asloun, E. M. and Schultz, J ., "Study of the Carbon Fiber- poly(Ether-Ether-Ketone)(PEEK) Interfaces I-IV", Polymer for Advanced Technologies, Vol. 2, pp 109-176, 1991 Waterbury, M. C. and Drzal, L. T., "Interfacial Shear Strengths of Carbon Fibers in Bisphenol A Polycarbonate", ICCI III, Elsevier Science Publishing, New York, May, 1990 Drzal, L.T., "The Interphase in Epoxy Composites", Advances in Polymer Science ( edited by K. D sek), Vol. 75, pp 1-32, 1986 Theocaris, P. S. , "The Mesophase Concept in Composites," Springer-Verlag Press, New York, 1987 Sharpe, L., "The Interphase in Adhesion," J. Adhesion, Vol. 3, pp 51, 1972 Drzal, L.T., " Fiber-Matrix Interphase", Michigan State University, 1992 Madhukar, M. and Drzal, L. T., "Fiber-Matrix Adhesion and its effect on Composite Mechanical Properties: 1. Inplane and Interlaminar Shear Behavior of Graphite/ Epoxy Composites and II. Longitudinal (0°) and Transverse (90°) Tensile and Flexure behavior of Graphite/ Epoxy Composites," J. of Comp. Mails, Vol. 25, August 1991 Piggott, M. R., "The Effect of the Interface/Interphase on Fiber Composite Properties," Polym. Comp., Vol. 8(5), pp 291, October 1987 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 96 Cazeneuve, C., Castle, J. E. and Watts, J. F., "The Structure of the Interface in Carbon Fibre Composites by Scanning Auger Microscopy, " J. Mater. Sci. , Vol. 25. PP 1902, 1990 Herrera-Franco, P. J., Wu, W-L., Madhukar, M. and Drzal, L. T., "Contempo— rary Methods for the Measurement of Fiber-Matrix Interfacial Shear Strength, 46th Annual Conference: Composites Institute, The Society of Plastics Industry, Inc., February 1991 Drzal, L.T., Herrera-Franco, P.J., "Composite Fiber-Matrix Bond Tests", Engineered Materials Handbook, Vol 3: Adhesives and Sealants, pp 391-405, 1991 Cox, H. L. , " The Elasticity and Strength of Paper and other Fibrous Materials," Br. J. Appl. Physics, Vol. 3(1), pp 122, 1952 Cooke, T. F., "High Performance Fiber Composites with Special Emphasis on the Interface: A Review of the Literature," J. of Polymer Processing, Vol. 793, pp 199, 1987 Rao, V. and Drzal, L. T., "The Dependence of Interfacial Shear Strength on Matrix and Interphase Properties," Polym. Comp., Vol. 12(1), 48, February 1991 Rao, V., "Interfacial Changes During the Processing of a Typical Carbon Fiber/ Epoxy Composite Material" , Ph.D. Dissertation, Michigan State University, 1991 Rosen, B. W., "Mechanics of Composite Strengthening in Fibre Composite Materials," Chapter 3 in Fiber Composite Materials, Amer. Soc. for Metals, Vol. 72, 1964 Dow, N. F., "Study of Stresses near a Discontinuity Of a Filament-Reinforced Composite Material, " General Electric Company, Report TIS R635D61, 1963 Bascom, W.D., Yon, K.J., Jensen, R.M., Cordner, L., " The Adhesion of Carbon Fibers to Thermoset and thermoplastic Polymers, " J. Adhesion, Vol 34, pp 79-98, 1991 Hercules product Data Sheet , 1988 Fitzer, E. and Heine, M. , " Carbon Fibre Manufacture and Surface Treatment" in Fiber Reinforcement for Composite Materials (edited by Bunsell, A.R), Series editor R.B. Pipes, V012, Elsevier 1988 37. 38. 39. 40. 41. 42. 43. 45. 46. 47. 48. 49. 50. 97 Oberlin, A. and Guigon, M., "The Structure Of Carbon Fibers", in Fiber Reinforcement for Composite Materials (edited by Bunsell, A.R), Series editor R.B. Pipes, Vol 2, Elsevier 1988 Brydson, J. A., " Plastic Materials," Butterworth, 1989 Kroschwitz, J .I. , ed. , Encyclopedia of Polymer Science and Engineering, Vol 11, Wiley-Interscience, pp 648-718, 1988 Richardson T.L, " Industrial Plastics: Theoery and Applications ," Delmar, 1989 Nielsen L.E., " Mechanical Properties of Polymers and Composites, " Vol 2, Marcel Dekker, 1974 Grayson, M., ed., Kirk-Othmer Encyclopedia of Chemical Technology, Vol 18, Wiley-Interscience, pp 479-494, 1982 Nielsen L.E., " Mechanical Properties of Polymers and Composites, " Vol 1, Marcel Dekker, 1974 deGennes, P.G., " Reptation of a Polymeric Chain in the Presence Of Fixed Obstacles," J. Chem. Physics, Vol. 55(2), pp 572-579, 1971 Graessley, W.W., " Entangled Linear , Branched and Network Polymer Systems - Molecular Theories," Advances in Polym. Sci., Vol. 47, pp 67-117, 1982 Iyer, S.R., Drzal, LT. and Jayaraman, K., "Method for Fiber Coating with Particles," US. Patent Nos. 5,123,373; 5,102,690 and 5,128,199. 1992 Muzzy, J. D., "Processing of Advanced Thermoplastic Composites," ASME symposium on Manufacturing Science of Composites, p27-39, April 1988 Van West, B.P., Pipes, RB. and Advani, S.G., " The Consolidation of Commingled Thermoplastic Fabrics," Polymer Composites, Vol. 12(6), pp 417- 427, 1991 Dyksterhouse, R. , and Dyksterhouse, J . A. , "Method of Improved Pre-Impregnat- ed Material Comprising a Particulate ’Thermoplastic Polymer Suitable for Use in the Formation of a Substantially Void-Free Fiber-Reinforced Article, " US. Patent Application 114362, Nov 1987 Taylor, G. J . , " Method of Impregnating a Fibrous Textile Material With a Plastic Resin," US. Patent 429105, Sept 1981 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 98 Chary, RR. and Hirt, D.E., " Coating Carbon Fibers with Thermoplastic Polymers Using Aqueous Foam," Proceedings of ANTEC 92, pp 1181-1183 Loos, A.C., and Dara, P.H., "Processing of Thermoplastic Matrix Composites," in Review of progress in Non-Destructive Evaluations. Plenum Press, Vol. 63, pp 1257-1265, 1987 Brown, HR, "The Adhesion between Polymers," Annu. Rev. Mater. Sci ., Vol. 21. PP 463-489, 1991 Voyutskii, S.S., "Autohesion and Adhesion of High Polymers," Wiley- Interscience, 1963 Wool, RP. and O’ Connor, K.M., "Theory Of Crack Healing in Polymers," J. Appl. Phys, Vol. 52(10), pp 5953-5963, 1981 Wool, RP. and O’ Connor, K.M., "Time Dependence of Crack Healing", J. Polym.Sci.: Poly. Lett. Ed, Vol 20, pp 7-16, 1982 Wool, R.P., " Molecular Aspect of Tack, " Rubber. Chem. and T echnol. Vol. 57, pp 307-319, 1983 Wool, R.P., Yuan, BL and McGarel, 0]., "Welding of Polymer Interfaces," Polym. Engr. and Sci., Vol 29(19), pp 1340-1367, 1989 deGennes, P.G., "Mechanical Properties of Polymer Interfaces," in, Physics of Polymer Surfaces and Interfaces (edited by LC. Sanchez), Butterworth- Heinemann, 1992 Prager, S. and Tirell, M. , "The Healing Process at Polymer-Polymer Interfaces," J. Chem. Phys, Vol. 75(10), pp 5194-5198, 1983 Jud, K., Kausch, H.H. and Williams J.G., "Fracture Mechanics Studies of Crack Healing and Welding of Polymers," J. Mater. Sci. , Vol. 6 pp 204-210. 1981 Loos, A.C., Howes, J.C. and Dara, P.H., " Thermoplastic Matrix Composite Processing Model," Adhesion Science Review, Proceedings of the 5th Annual Program/Review Workshop - 1987, pp 263-281 Agarwal, V. " The Role of Molecular Mobility in the Consolidation and Bonding of Thermoplastic Composite Materials," Ph.D dissertation, University of Delaware, 1991 Bastien, L. J. and Gillespie Jr., J.W., " A Non-Isothermal Healing Model for Strength and Toughness of Fusion Bonded Joints of Amorphous Thermoplastics. " Polym. Engr. and Sci., Vol. 31(24), 1720 -1730, 1991 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 99 Scheutjens, J .M. and Fleer, G.J. , "Interactions Between Two Adsorbed Polymer Layers," Macromolecules, Vol. 18, pp 1882, 1985 Theodorou, D.N. , " lattice Models for Bulk Polymers at Interfaces," Macromole- cules, Vol. 21, pp 1391, 1988 Scheutjens, J.M. , "Mean Field Lattice Models for Polymers at Interfaces," in Physics of Polymer Surfaces and Interfaces (edited by LC. Sanchez), Butterworth- Heinemann, 1992 Madden, W.G., "Monte Carlo Studies of Melt Vacuum Interfaces of a Lattice Polymer, " J. Chem. Phys, Vol. 87, pp 1405, 1987 Mansfield, K. F. and Theodorou, D.N., " Interfacial Structure and Dynamics of Macromolecular Liquids: A monte Carlo Simulation Approach, Macromolecules, Vol. 22. pp 3143, 1989 Chakrabarti, A. , and Toral, R., Density Profiles of Terminally Anchored Polymer Chains: A Monte Carlo Study, Macromolecules, Vol. 23, pp 2016, 1989 Pangelinan, A.B. , "Surface Induced Molecular Weight Segregation in Thermo- plastic Composites," Ph.D dissertation, University of Delaware, 1991 Theodorou, D.N. , " Molecular modeling of Polymer Surfaces and Polymer/Solid Interfaces," in, Physics of Polymer Surfaces and Interfaces (edited by LC. Sanchez), Butterworth-Heinemann, 1992 McCullogh, R.L. , " Deformation Mechanism of Constrained Polymer Chains. I. Geometrical Induced Correlations in Deformation Response Mechanisms of Tie Molecules," J.Polym. Sci., Polym. Phys. Ed. Vol. 15, pp 1805-1835, 1977 Fitzgibbon, DR. and McCullough, R.L., " Influence of Neutral Surfaces on Polymer Molecules in the Vicinity of the Surfaces," J. Polym. Sci., Vol. 27, pp 655-671, 1989 Brady, R.L. and Porter, R.S., "Interfacial Crystallization and Adsorption in Polycarbonate/Carbon Fiber Composites," J. Thermoplastic Comp. Mates, Vol. 2. Pp 164-171, 1989 ‘ ASTM D 638M, "Standard Test Method for Tensile Properties Of Plastics (metric), " ASTM Standards and Literature References for Composite Materials, 1991. PP. 172-180 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 100 Golovoy, A., Zinbo, M., "Water Sorption and Hydrolytic Stability of Polycarbonates," Polym. Eng. and Sci., Vol. 29(24), pp.- 1733-1738, 1989 Munjal, S. and Kao, C.I. , "Mathematical Model for Experimental Investigation of Polycarbonate Pellet Drying," Polym. Engr. and Sci., Vol. 30(21), 1352, 1990 Granick, 8., "Dynamics of Adsorption and Desorption at Polymer/Soild Interfaces" in, Physics of Polymer Surfaces and Interfaces, (edited by LC. Sanchez), Butterworth-Heinemann, 1992 Drzal, L.T., Rich, M.J., Camping, J.D., and Park, W.J., "Interfacial Shear Strength and Failure Mechanisms in Graphite Fiber Composites," Paper 20-C, 35th Annual Technical Conference, Reinforced Plastics/Composites Institute, The Soceity of the Plastics Industry, 1980 Kelly, A., and Tyson, W.R., "Tensile Properties of Fiber-Reinforced Metals: Copper/Tungsten and Copper/Molybdenum, J. Mech. Phys. Solids, Vol. 13, pp 329-350, 1965 Tse, M.K. , US. Patent 4,662,228, (assigned to Dow Chemicals Company Inc.), 1987 Mandell, J.F., Grande, D.H., Tsiang, TH, and McGarry, P.J., " A Modified Microdebonding Test for Direct In-Situ Fiber/Matrix Bond Strength Determina- tion in Fiber Composite, " Research report R84-3, Massachusetts Institute of Technology, December 1984 Drzal, LT. and Herrera-Franco, P.J., "Composite Fiber-Matrix Bond Tests," Engineered Materials handbook, Vol. 3 : Adhesives and Sealants, ASM International, pp 391-405, 1991 Aharoni, S.M. , "On Entanglements of Flexible and Rodlike Polymers, Macromol- ecules, Vol. 16, pp 1722-1728, 1983 Myers. D.. " Surfaces, Interfaces and Colloids: Principles and Applications," VCH Publishers, 1991 Gutowski, W.R. , " Effect of Fibre-Matrix Adhesion on Mechanical Properties of Composites," in, Controlled Interphases in Composite Materials (edited by H. Ishida), Elsevier, New York, pp505-520, 1990 Adamson, A.W., "physical Chemistry of Surfaces," Wiley-Interscience, 1990 89. 90. 91. 92. 93. 101 Carafagna, C., Netti, P.A., Nicolais, L., Dibenedetto, A.T., "In-Situ Compos- ites: Evaluation of the adhesion Between the Thermoplastic Matrix and the Fibers of Liquid Crystalline Polymer, " Polym. Comp. Vol. 13(3), pp 169-173, 1992 Stamm, M. , " Reflection of Neutrons for the Investigation of Polymer Interdiffusion at Interfaces," in, Physics of Polymer Surfaces and Interfaces, (edited by LC. Sanchez), Butterworth—Heinemann, 1992 Fowkes, F.M., "Attractive Forces at Interfaces," Ind. and Engr. Chem, Vol. 56, pp 40-62, 1964 Kaelble, D.H. , "Dispersion-Polar Surface Tension Properties of Organic Solids, " J. Adhesion, Vol. 2, pp 50-72, 1970 Ferry, J.D., "Viscoelastic Properties of Polymers," Wiley-Interscience, 1980