it . 3:2... 3 v I: 51.; :v r .3 a. .q‘ “dis: 1 . -1... 4‘ r: _ t\; A 12:...1... 21.41:.“ . . L... 1...». .'II iii THE ‘C.\“‘\\\\‘\\ \Wil [i W Twin 131293 LIBRARY Michigan State University l This is to certify that the dissertation entitled ADHESION MECHANISMS OF POLYURETHANES T0 GLASS SURFACES presented by RAJ K. Agrawal has been accepted towards fulfillment of the requirements for Ph.D. degreein Chemical Engineering afimzw Major professor Date WWgé /?97/ MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE ll RETURN BOXtorunavombehockouflommm TO AVOID FINES return onetbdorodntoduo. DATE DUE DATE DUE DATE DUE MAR Q 1 2333 U 0 ‘0 2, ’ MSU leAn Mun-five NONE“ Oppomnly Institution W1 ADHESION MECHANISMS or POLYURETHANES TO GLASS SURFACES By Raj K. Agrawal A DISSERTATION Submitted to Michigan State University in partial fiilfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemical Engineering 1994 ABSTRACT ADHESION MECHANISMS or POLYURETHANES T0 GLASS SURFACES By RniKAsmvnl Adhesion mechanisms of segmented polyurethanes to glass surfaces were investigated in this study. Polyurethanes were prepared fi'om caprolactone based polyols and toluene diisocyanatewith 1,4butanediol asthechainextender. Amodelbased onhard and soft segment miscibility was developed to predict phase separation in these polyurethanes. The model was compared with experimentally determined phase separation data from near- infrared, fourier transform infrared, difl‘erential scanning calorimetry, dynamic mechanical analysis, and mechanical characterizations. Adhesion of various polyurethanes to glass surfaces was evaluated using ”block-shear” adhesion test method. It was found that phase separation in polyurethanes significantly afi‘ects their adhesion. Polyurethanes with higher modulus showed better adhesion, but in polyurethanes with the same modulus, phase separated samples showed better adhesion than the phase mixed samples. Scanning electron microscopy and x-ray photoelectron spectroscopy of the fi'actured surfaces revealed the formation of a 20A - 100A thick interphase region. Composition and thickness of this interphase region was found to be dependent on the matrix phase separation. It was determined that phase separation in the matrix could cause preferential segregation of butanediol to the interphase region. Thepowbflityofdranicdbondmgbetwwnthepolymuhanesmdtheglaumrfacuwu orploredbymaldngtheglaumrficechafiuflymatfluoughmethyluimahoxydlmemd trimethylchlorosilaneu'eatments. Chemicalbondingwasfoundnottobeanimportant factor in glass/polyurethane adhesion. The efl‘ect of increased modulus of the BDO rich interphaseregion, duetoBDOarfidunamwasdetectedbtrtwasconcludedtobeaminor fictor in adhesion of the phase-separated polyurethanes to glass. The surface flee energies of the various polyurethanes were evaluated using the molar parachors and contact angle measurements. The work of adhesion was calculated from the surface energetics and compared to the experimentally determined adhesion data A linear correlation was found between the polar surface fi'ee energy (7’) of the polyurethanes and their adhesion values; adhesion being higher for higher 1’ polyurethanes. An adhesion experiment with butanediol coated glass plaques and the various polyurethanes demonstrated that adhesion of these polyurethanes can be significantly improved by creating an interphase region rich in butanediol type species. Hydrogen bonding between the constituents ofthe interphase region and the glass surface was concluded to be the primary mechanism of adhesion. A model of the interphase region is proposed with various possible hydrogen bonding mechanisms. To my parents iv ACKNOWLEDGEMENTS I would like to express my sincere appreciation to Dr. Lawrence T. Drzal for his guidance, support, encouragement and mentoring throughout this project. It has been a real pleasure to work with him and I have certainly learned a great deal during my past nine years of association with him. I also wish to thank my fellow engineering students at the composite center past and present - Shri, Tesh, Javad, Brent, Ed, Sanjay, Murry, Venkat, Rik and many others for their assistance and friendship. In addition, I wish to thank Dan Hook, Henjen Ho, Brian Rook and Mike Rich for their help in equipment operation, experimental analysis as well as their friendship. I am thankfirl to the other members of my committee - Dr. J. Jackson, Dr. K. Jayaraman, Dr. R. Narayan, Dr. W. Mungall and Dr. A. Scranton for the time they dedicated to this dissertation. I am also thankful to Niall Lynam, Anoop Agrawal, Desaraju Varaprasad and Hamid Habibi, all of Donnelly Corporation, for their help, encouragement and thoughtful discussion throughout this project. I thank the Donnelly Corporation, the office of Naval Research and the Automotive Composites Consortium for supporting this work. I thank my parents and family for their emotional and financial support and their never ending encouragement through out my educational career. Finally I thank my wife Sunita for her love, constant support, encouragement and help, without which I would not have been able to complete this work. LIST OF TABLES LIST OF FIGURES NOMENCLATURE CHAPTER 1 INTRODUCTION AND BACKGROUND 1.1 BACKGROUND ON POLYURETHANES 1.1.1 Reaction Injection Molding (RIM) 1.1.2 RIM Urethane Chemistry 1.1.3 Modular Windows 1.2 BACKGROUND ON ADHESION CHAPTER 2 PROJECT DESCRIPTION 2.1 PROBLEM DEFINITION CHAPTER 3 STRUCTURE-PROPERTY RELATIONSHIPS IN POLYURETHANES AND THEIR EFFECTS ON ADHESION TO GLASS 3.] ABSTRACT 3.2 INTRODUCTION 3.3 SCOPE OF THE PRESENT WORK l4 I8 22 30 37 37 38 39 3.4 3.5 3.6 EXPERIMENTAL 3.4.1 Materials 3.4.2 Sample Preparation 3.4.3 Tests and Characterization RESULTS AND DISCUSSION 3.5.1 Thermal Characterization 3.5.2 Mechanical Characterization 3.5.3 Adhesion to Glass CONCLUSIONS CHAPTER 4 PHASE SEPARATION IN POLYURETHANES AND 4.1 4.2 4.3 4.4 ITS EFFECTS ON ADHESION TO GLASS ABSTRACT INTRODUCTION EXPERIMENTAL 4.3.1 Materials 4.3.2 Wide Angle X-ray Difi’action (WAXD) 4.3.3 Near-Infrared Spectroscopy (NIR) 4.3.4 Fourier Transform Infi'ared Spectroscopy 4.3.5 X-ray Photoelectron Spectroscopy (XPS) THEORETICAL CONSIDERATION OF PHASE SEPARATION 40 40 41 43 49 49 59 68 80 82 82 83 84 84 87 87 87 89 4.5 4.6 RESULTS AND DISCUSSION 4.5.1 Bulk Morphology 4.5.2 Interphase Composition 4.5.3 Adhesion Results CONCLUSIONS CHAPTER 5 INVESTIGATION OF POSSIBLE PHYSICO-CHEMICAL 5.1 5.2 5.3 5.4 5.5 INTERACTIONS AT THE INTERPHASE ABSTRACT INTRODUCTION EXPERIMENTAL 5.3.1 Materials 5.3.2 Surface Energy Measurements 5.3.3 Surface Tension Measurements 5.3.4 Dynamic Mechanical Analysis 5.3.5 Glass Pretreatment 5.3.6 X-ray Photoelectron Spectroscopy (XPS) THEORETICAL CONSIDERATION OF SURFACE FREE ENERGY RESULTS AND DISCUSSION 5.5.1 Physical Interactions 92 92 102 112 114 117 117 118 120 120 121 124 124 124 125 126 129 129 5.5.2 Chemical Interactions 5.5.3 Adhesion Mechanisms 5.6 CONCLUSIONS CHAPTER 6 CONCLUSIONS AND RECOMMENDATTONS 6.1 CONCLUSIONS 6.2 RECOMMENDATIONS FOR FUTURE WORK APPENDICES APPENDIX A Group Contn'butions to Solubility Parameter (6“) of the Hard Segment Based on Toluene Diisocyanate and Butanediol APPENDIX B Surface Free Energy Estimation of the Hard Segment by Molar Parachors and the Cohesive Energy Densities BIBLIOGRAPHY 150 157 163 165 165 I67 169 171 173 TABLE 2.1 TABLE 2.2 TABLE 3.1 TABLE 3.2 TABLE 3.3 TABLE 3.4 TABLE 3.5 TABLE 4.1 TABLE 4.2 TABLE 4.3 TABLE 4.4 TABLE 4.5 TABLE 4.6 TABLE 4.7 TABLE 4.8 TABLE 5.1 TABLE 5.2 TABLE 5.3 LIST OF TABLES RelativeReactivityofPhenyl IsocyanatewithVariousActive Hydrogen Compounds Typical Compositions of Soda-Lime Glass (wt. %) Urethane Formulations at Isocyanate Index of 1.0 Efl‘ects of Hard Segment and Polyol Molecular Weight on Thermal Transitions in Cross-Linked Segmented Polyurethanes Mechanical Properties of Various Polyurethane Systems Adhesion of Various Polyurethanes to Glass Surface Atomic Concentration Ratio on Failed Glass Surface fi'om Xray Photoelectron Spectroscopy Urethane Formulations at Isocyanate Index of 1.0 'N,‘ Values for Various Polyols 'Nfl' Values for Different Hard Segment Lengths 'Na + N,‘ Values for Various Urethane Formulations Calculation of Critical Interaction Parameter xc C 1s Peak Binding Energies and Relative Peak Areas of Carbon Chemical States on Various Failed Glass Samples 0 1s, N Is, C 1s, and Si 2p Atomic Percent Concentrations at Difl‘erent Photoelectron Take-OE Angles Adhesion of Various Polyurethanes to Glass Surfaces (PSi) Urethane Formulations at Isocyanate Index of 1.0 Surface Free Energies of Liquids Used for Contact Angle Measurements Calculated Surface Free Energies of the Polyurethanes X 29 32 42 53 65 73 79 86 93 93 94 95 1 O9 110 113 122 123 129 TABLE 5.4 TABLE 5.5 TABLE 5.6 TABLE 5.7 TABLE 5.8 TABLE 5.9 TABLE 5.10 TABLE 5.11 Observed Surface Free Energies of the Polyurethanes Based on Contact Angle Measurements Thermodynamic Work of Adhesion Between the Polyurethanes and the Glass Surface Adhesion Values (Psi) of Various Polyurethanes to Bare Glass Surface and to 1,4 Butanediol Coated Glass Surface C 1sPeakBindingEnergies andRelativePeakAreasof Carbon Chemical States on Failed Glass Samples Elastic Storage Modulus of Various Polyurethanes at Difi‘erent Temperatures Adhesion Values (Psi) of Various Polyurethanes to Silane Treated Glass Surfaces 0 Is, N Is, C 1s, and Si 2p Atomic Percent Concentrations at Difi‘erent Photoelectron Take-OE Angles for the Adhesion Failed 10D Glass Sample Coated with Methyltrimethoxysilane Adhesion Values (Psi) of 1D Polyurethane to Various Treated Glass Surfaces 132 140 143 I46 148 154 155 159 FIGURE 1.1 FIGURE 1.2 FICHJRE 1.3 FIGURE 1.4 FICRJRE 1.5 FIGURE 1.6 FIGURE 1.7 FIGURE 1.8 FIGURE 2.1 FIGURE 2.2 FIGURE 2.3 FIGURE 2.4 FIGURE 3.1 FIGURE 3.2 FIGURE 3.3 LIST OF FIGURES SchematicofaRIMMachine TheEightUnitOperationsforRM Schematic Representation of Phase Separation in Polyurethane Segmented Block Copolymers Flernml Modulus at 25°C vs. Soft Segment Molecular Weight and Hard Segment Content in Polyurethanes Dynamic Modulus vs. Temperature for Three Hard Segment Types All at 50% Hard Segment Content Difl‘erential Scanning Calorimetric Curves of Urethane Polymers Containing Various Glycols Load vs. Strain for Polyurethanes at Various % Hard Segment Content An Encapsulated Modular Window A Liquid Drop Resting at Equilibrium on a Solid Surface Schematic Representation of Interphase Region in a Polymeric Composite A Transmission Electron Micrograph of a Segmented Polyurethane Chemical Structure of Polyurethane Constituents Glass - Urethane Adhesion Sample Glass-Urethane Adhesion Testing Fixture a) Front View of Fixture with Specimen Clamped in Place b) RearViewofFixturewithSpecimenClampedinPlace Model of Molecular Arrangements in Urethanes 10 11 11 13 13 15 24 25 27 34 47 50 FIGURE 3.4 DSC Thermograms of Tone 0310 Based Polyurethanes with Varying Hard Segment Content (1st Run) FIGURE 3.5 DSC Thermograrns of Tone 0310 Based Polyurethanes with Varying Hard Segment Content (2nd Run) FIGURE 3.6 DSC Thermograms of Tone 0305 Based Polyurethanes with Varying Hard Segment Content (1st Run) FIGURE 3.7 DSC Thermograrns of Tone 0305 Based Polyurethanes with Varying Hard Segment Content (2nd Run) FIGURE 3.8 DSC Thennograms of Tone 0301 Based Polyurethanes with Varying Hard Segment Content (1st Run) FIGURE 3.9 DSC Thermograrns of Tone 0301 Based Polyurethanes with Varying Hard Segment Content (2nd Run) FIGURE 3.10 Thermal Transition Temperature Versus Hard Segment Content in Difl‘erent Molecular Weight Triol Based Urethane Systems (1st DSC Run) FIGURE 3.11 Thermal Transition Temperature Versus Hard Segment Content in Different Molecular Weight Triol Based Urethane Systems (2nd DSC Run) FIGURE 3.12 DMA Scan of Tone 0310 and Tone 0301 Polyols Based urethanes with Constant Hard Segment Content FIGURE 3.13 Load/Displacement Response of Tone 0310 Based Polyurethanes at Difi'erent Hard Segment Content FIGURE 3.14 2% Secant Tensile Modulus Versus Hard Segment Content in Various Urethane Systems FIGURE 3.15 Tensile Strength Versus Hard segment Content in Various Urethane Systems FIGURE 3.16 Fractured Urethane Sample After Iosipescu Shear Testing 51 54 55 56 57 58 61 62 63 66 67 69 FIGURE 3.17 Stress/Strain Response of Difi‘erent Urethanes in Iosipescu Shear Testing FIGURE 3.18 2% Secant Shear Modulus Versus Hard Segment Content in Various Urethane Systems FIGURE 3.19 2% Secant Shear Modulus Versus Molecular Weight per Cross-Link (M9 in Various UrethaneSystems FIGURE 3.20 Adhesion to Glass Versus Hard Segment Content in Various Urethane Systems FIGURE 3.21 Adhesion to Glass Versus Shear Modulus of Various Urethane Systems FIGURE 3.22 SEM Micrographs of Glass Surfaces After Adhesion Testing of Tone 0310 Based Polyurethanes FIGURE 4.1 NIR Spectra of Tone 0310 Based Polyurethanes with Varying Hard Segment Content FIGURE 4.2 NIR Spectra of Tone 0305 Based Polyurethanes with Varying Hard Segment Content FIGURE 4.3 NR Spectra of Tone 0301 Based Polyurethanes with Varying Hard Segment Content FIGURE 4.4 FTIR Spectrum of Sample 10B FIGURE 4.5 FTIR Spectra of N-H Band in Tone 0310 Based Polyurethanes with Varying Hard Segment Content FIGURE 4.6 FTIR Spectra of Free (1736 cm") and Hydrogen Bonded (1712 cm") Carbonyl Bands in Tone 0310 Based Polyurethanes with Varying Hard Segment Content FIGURE 4.7 XPS Survey Scans (a) Bare Soda-Lime Glass, (b) Failed Glass Sample 10B FIGURE 4.8 Curve Fit C Is Spectra of Failed Glass Samples at 90° Photoelectron Take-Ofl‘ Angle xiv 70 71 72 75 76 97 98 101 103 104 106 107 FIGURE 5.1 FIGURE 5.2 FIGURE 5.3 FIGURE 5.4 FIGURE 5.5 FIGURE 5.6 FIGURE 5.7 FIGURE 5.8 FIGURE 5.9 Zisman Plots of (a) 1C Polyurethane Sample with 46.7 wt. % Hard Segment and (b) HS Polyurethane Sample with 100% HardSegrnent p 96 1 C 0 1% vs. Y“ + as ) plotfor r; 20?)" Tone 0310 Based Polyurethanes with Difi'erent Hard Segment Contents. Theoretical (X) and Measured (0) Surface Free Energy Values for Tone 0310 Based Polyurethanes at Difl‘erent Hard Segment Contents Polar Surface Free Energy Component of Various Polyurethanes at Difl'erent Hard Segment Contents Adhesion to Glass Versus Polar Component of the Surface Free Energy for Various Polyurethane Systems Curve Fitted C Is Spectra of Failed Glass Surfaces Taken at 15° Photoelectron Take-OE Angle (a) Sample ID (b) Sample ID with BDO Coated Glass Idealized Monolayer of Condensed (a) Methyltrimethoxysilane (b) Trimethylchlorosilane on Glass Surface Curve Fitted C 1s Spectrum of Methyltrimethoxysilane Treated Glass Surface at 45° Photoelectron Take-OE Angle Curve Fitted C 1s Spectra of Adhesion Failed 10D Glass Samples Precoated with Methyltrimethoxysilane Taken at 15°, 45°, and 90° Photoelectron Take-OE Angles FIGURE 5.10 Curve Fitted C 1s Spectra of (a) Control Glass Sample (b)GlassTreatedwithBDOandRinsed FIGURE 5.11 Schematic of Polyurethane/Glass Interphase Region with Probable Hydrogen Bonding Mechanisms 131 134 136 138 141 144 151 153 157 161 162 ADXPS BDO DABCO DBTDL DETDA DMA DSC PVC NOMENCLATURE Angular dependent x-ray photoelectron spectroscopy Butanediol 1,4-disabicyclo[l.1.1]octane Dibutyl tin dilaurate Diethyl toluene diamine Dynamic mechanical analyzer Difl‘erential scanning calorimetry Elastic storage modulus Loss modulus Cohesive energy Cohesive energy density Ethylene glycol Electron spectroscopy for chemical analysis Molar attraction constant Fourier transform infi’ared spectroscopy Injection molding In-mold coating Molecular weight Molecular weight per cross-link 4,4'-diphenylmethane diisocyanate Number of monomer units in the hard segment Number of monomer units in the soft segment Near-infi'ared spectroscopy Molar parachor Polypropylene oxide Polyvinyl chloride Universal gas constant Reaction injection molding RRIM Reinforced reaction injection molding SEM Scanning electron microscope SIMS Secondary ion mass spectroscopy SRIM Structural reaction injection molding T Temperature T. Glass transition temperature TDI Toluene diisocyanate V Molar volume V,l Reference volume Vw Van der Waals volume WAXD Wide angle x-ray difi'action W... Work of adhesion W" Polar work of adhesion ‘ W”... Dispersive work of adhesion XPS X-ray photoelectron spectroscopy GREEK SYMBOLS 1 Surface fi'ee energy 7’ Polar surface fi'ee energy 7” Dispersive surface free energy 70 Surface fi’ee energy of glass 7,, Surface flee energy of hard segment 7,, Surface free energy of liquid 1, Surface fi'ee energy of soft segment 70 Surface the energy of polyurethane Solubility parameter , Solubility parameter of soft segment 6,, Solubility parameter of hard segment Contact angle Specific snvity Interaction parameter Critical interaction parameter Hard and soft segmem interaction parameter Interaction parameter due to entropy change CHAPTERI W Glass/polymeradhesioningeneralhasbecomeveryimportam duetothepopularityof glass fiber reinforced composite materials. Good adhesion between the polymer matrix anddwglmrdnforcunenthshowntobemhnpomfiaorindetamirungtheoverafl mechanical properties of these composite materials (Dawson et al, 1982; Dau et al, 1989; Yang and Lee, 1987; Schwarz, 1979). Among the wide variety of polymers for composites manufacturing, polyurethanes are becoming increasingly important due to their fast molding cycles. Reinforced reaction injection molding (RRIM) and structural reaction injection molding (SRIM) of polyurethanes are widely used in the automotive industry to make body panels, fascias, bumper beams, etc. In addition to these, glass/polyurethane combinations can be found in a variety of applications such as reaction injection molded modular windows for automobiles, laminated Windshields, and fiber optic cables. Also the use of polyurethane coating on the inside surface of windshields is being investigated for making these antilacerative. In all of these applications, good adhesion between the glass and the polyurethane is imperative for their viability and efi‘ectiveness. W Polyurethanes are commonly processed by liquid molding process such as reaction injection molding (RIM). The RIM process is quite difi‘erent fiom the injection molding process commonly used for thermoplastics, and can have a significant efi‘ect on urethane 2 toglass adhesion. T'heRMproeesaRIlt/lurethanechemistryand encapsulated modular windows are discussed below. 111 B .1. . 1111' [Bill] RIM is a process for rapid production of complex polyurethane plastic parts from the combination and rapid reaction of low viscosity monomers and oligomers. These liquids are combined by impingement mixing just as they enter the mold. Mold pressures are very low. The solid polymer forms by cross-linking or phase separation and parts can ofien be dernolded in less than one minute. Figure 1.1 shows a schematic of a RIM machine. Two or more chemical streams flow at high pressures (around 2,000 psi) into a mixing chamber. In the mixhead the streams impinge at high velocity, mix, and begin to polymerize as they flow out into the mold cavity. Because the mixture is initially at a low viscosity, a lower pressure of around 50 psi is needed to fill the mold cavity. The RIM process consists of eight unit operations which are illustrated in Figure 1.2. Supply tanks are used to store and blend components. They maintain the level in the conditioning tank in the machine. The conditioning tanks control temperature and degree of dispersion of the reactants by low pressure recirculation. An important step in RIM is highpressuremeteringofthereactantstothemixheadatsuficientflowrateforgood mixing and at the proper ratio for complete polymerization. From the impingement chamber, the reacting mixture flows into the mold, filling it in typically less than five High Pressure Ratio Control A? '2 $4 :4 s § Mixhead 2|? Reactant A / Reactanl B \\ j ' \\ // FIGURE 1.1 Schematic of a RIM Machine SUEP'Y——> Condition (2) a Blend ‘— Componenl(s) H Rec cle Re cle \ Supply Pressurize. Meter (3) Mix (4) Fill (5) -—> Cure (6) —> Demold (7)—> Finish (8) poslcure. painl FIGURE 1.2 The Eight Unit Operations for RIM 4 seconds. There it polymerizes and solidifies suficiently to take the stresses of demolding. The final operation consists of various finishing steps including trimming of flash and cleanings. RMhuthreepmcesdngeharadaisficswhichmakehespedaflyuuacfiveforhigh- volumeproductionoflargeparts. Theseuethelowpresmresrequiredthelow processing temperatures involved, and the use of reactive liquid intermediates. TheRMprocessrequiresmaterialtobemeteredataround 2,000 psi. Thispressure requirement is much lower than that involved in other high-volume plastic fabrication A processes. For example, injection molding machines generate barrel pressures flom 8,000— 40,000 psi to force high viscosity resin into the mold. The low viscosity of the injected fluids in RIM make it possible to fill molds completely at pressures below 50 psi. This leads to much smaller and less expensive mold clamps for large parts. Low viscosity and low pressure during filling also translates into lighter weight and lower cost molds. Complex shape parts with multiple inserts can readily be fabricated with the RIM process. The low viscosity involved in the process also opens many options for reinforcements. One is to place long fiber mats into the mold and then inject reactive monomers into them in a second step (Gonzalez and Macosko, 1983; Eckles and Wilkinson, 1986; Carleton et al, 1986). Thus RIM can be used for high-speed resin transfer molding. This process is generally called structural RIM (S-RIM). 5 RIMurethanesystemsarealso filled/reinforcedwithglass for achieving highstifiress, low thermalexpansion coeficientandbetter dimensional stability. Inoneapproacltglass fillers(choppedglass,milledglass,andflakeglass)areaddedtothereactantsofthe urethanefommlaflonmdflreprocessisrefaredureinforcedRIMR-RIM). Temperatures utilized in RIM processing are also low when compared with thermoplastic injection molding process. In RIM urethane processing, the reactive liquid streams are maintained at temperatures between 75° and HOT with specific temperature depending on the chemical stream being processed. The mold is kept at a temperature between 130° to 170°F. Since the urethane polymerization is highly exothermic, minimal heat input is required to maintain tooling temperature during production. RIM‘s use of liquid intermediates has additional benefits beyond the low pressures and temperatures involved. A tremendous amount of design flexibility is possible with RIM. Since the mold is filled with a low viscosity liquid, very large parts with complex designs can be produced. These process advantages have allowed the designer to take full advantage of the remarkable versatility that is possible with urethane chemistry. By selecting flom a wide range of intermediates, the forrnulator can develop a polymer which will meet a specified set of physical property requirements (Gillis et al, 1983). The material can range flom very flexible elastomers to very stifl' plastic. It can be modified with a variety of 6 fillers/reinforcements. Since the intermediates are liquid, this compounding can be done economically in relatively small batches. RIM has a relatively short history of about 20 years since their first major commercial use for producing automotive bumpers and fascias in flexible polyurethane in 1974. About 95% of all the RIM production is in polyurethanes or urethanes (Macosko, 1989). Several other chemical systems suitable for RIM process, for example, Nylon 6 (Hedrick et al, 1985) and dicyclopentadiene (Geer, 1983) are also currently in use. LLLRIMllmhansflzemim In the RIM urethane process, two chemical streams, one of diisocyanate and the other of polyol, are impingement mixed. Isocyanate reacts with polyol to form a urethane linkage. R—N=C=O + HO—R' —> R—NH—C—O—R' Several diisocyanates can be used to form the polyurethane elastomer, but the majority of RIM formulations are built on derivations of 4,4'-diphenylmethane diisocyanate (MDI). Popular derivatives include use of polymeric MIDI and uretonimines. ocri—.-—c«[-‘H2 ‘]-:H2—.'NCO POLYMERIC MDI 1.1—NCO OCN-R—NjirN—R—NCO R = ““2 O MONOMERIC URETONIMINE Prepolymers obtained by the reaction of low molecular weight diols with excess MDI are also used in some urethane formulations. The second chemical stream is of a hydroxyl terminated flexible chain oligomer. Most popular oligomers are based on polypropylene oxide (Speckhard and Cooper, 1986). Ethylene oxide capped, polypropylene oxide diols and triols with molecular weights between 3,000 and 7,000 are the main oligomers used today in RIM urethane formulations. s THz-Ofcsmof'fczfirmflfil‘l CH2—0'6'C3II50')‘:”“(’C2H40§3C2H40H ETHYLENE OXIDE CAPPED POLYPROPYLENE OXIDE BASED POLYOL The polyol stream also contains chain extenders, catalysts, and other additives such as fillers, pigments, surfactants, and internal mold release agents. Chain extenders are added to form a segmented block copolymer (hard segment) when it reacts with diisocyanates. The most common chain extenders in use today are ethylene glycol and isomeric mixture of 2,4 and 2,6 diamine isomers of 3,5 diethyltoluene diamine (DETDA) (Pannone and Macosko, 198 8). HO—CHz—CHz—OH ETHYLENE GLYCOL H3 H3 I NH; HzN . N112 €sz €sz C2H5 C2H5 NH; 0 80% 20/9 DIETHYLTOLUENE DIAMINE Efl‘ective catalysts for isocyanate-hydroxyl RIM formulations are tertiary amine (Wongkamolsesh and Kresta, 1985) and tin catalysts (Camargo et al, 1985), and are used 9 in low concentrations (<1%) in the polyol stream. The two most commonly used tertiary amine and tin catalysts are 1,4-diazabicyclo-[l,1,1] octane (DABCO) and dibutyl tin dilaurate (DBTDL). NearlyallRMurethanesystemsbuild structurebyphase separationratherthanbycross linking (Gillis, 1982). Diisocyanate combines with the chain extender to form a hard block in the segmented copolymer which phase separates to build structure (Figure 1.3). The oligomer portion of a RIM formulation is called the soft segment because it typically has a low glass transition temperature. There is extensive data (Saunders and Frisch, 1983) reported in the literature on final properties of segmented polyurethanes as a flmction of various formulation parameters. The formulation parameters which can affect phase separation of segmented polyurethanes include the chemical nature of the hard and soft segment, individual segment length and segment length distn’bution, intra- and interdomain hydrogen bonding, hard segment content, overall molecular weight, and molecular weight distn’bution, as well as the nature of the domain interface and the mixing of hard segments in the soft phase. With hard segment content being the same, longer sofl segment sequence length improves the degree of phase separation, hence increases the flexural modulus as shown in Figure 1.4 (Macosko, 1989). The reinforcing efi‘ect of the hard segment is also clearly demonstrated as the flexural modulus increases with hard segment content at a given sofl segment molecular weight. Polyurethanes based on polyether polyol exhibit better low 10 HARD SEGMENT [] ousocvauare .7 ‘.' 2‘ ' ....... .-.«_. . . EXTENDER _9_ POLYO L SOFT SEGMENT HARD SEGMENT FIGURE 1.3 Schematic Representation of Phase Separation in Polyurethane Segmented Block Copolymers Ambient Temperature Flexural Modulus (MPa) 6 a ll s S -O. ..ra.. s 3 1,000 .5. s s 3.000 "I. 5 5 0,000 0“ I ~ 'H—-O—————O|otus . 00‘ S F ,. l/o - p- l’A—- " : -/ 0 “é » t—“OPO‘MIS )- O 1 '0 20 50 40 M, at Soft Segment (x 10”) FIGURE 1.4 Flexural Modulus at 25°C vs. Soft Segment Molecular Weight and Hard Segment Content in Polyurethanes [. “0 1L 1 fl I I T I 1 .1 q 1 W molznsrmmzooo. us.sos '5 / 1 \rotmtroamrm h‘ “‘ ‘ . . MOI/IDAIDZOOO.HSISO§ 1 <2 / i . 1 c9 ‘1 ‘ N. 7 a x . .0 . 1 .. am ° = - a" 4‘ 1 0 ." 0.0.4 1 ‘0 ~ . d e 9 . as . O 0’ °, . , "it o I O 0 ¢ . : : - . . O 0 v T ' 1 r r r r 1 r r W' 1* Temp.deg. (°C) FIGURE 1.5 Dynamic Modulus vs. Temperature for Three Hard Segment Types All at 50% Hard Segment Content 12 temperature flexibility than the counterparts based on polyester polyol. This is attributed tolesshydrogenbondingbetweenthehardsegmentandtheethergroupsinpolyethersofl segment in contrast to the ester groups. Highmoduhrsandlowheatsagpropertiesareoflenpromotedbyhardaegmem crystallinity and rigid bulky groups. Crystallization provides additional driving force for phase separation besides the thermodynamic incompatibility. In glycol extended polyurethanes, crystallinity is defined as an important mechanism for property development. Polymers with amorphous hard segments made of 2,4'-MDI and butanediol (BDO) shows extensive phase mixing and poor mechanical properties. Figure 1.5 (Macosko, 1989) shows the dynamic modulus vs. temperature data for typical glycol extended systems. The ethylene glycol extended system exhibits higher softening temperature and flatter rubbery plateau modulus than the BDO extended system. This is probably due to its higher hard segment transition temperature. Figure 1.6 (Cornell et al, 1984) compares difi‘erential scanning calorimetry (DSC) curves for RIM systems containing difl'erent glycol chain extenders. The increasing hard segment melting point with decreasing glycol molecular weight is readily apparent. Diamine chain extenders, even without the benefit of crystallinity, show improved properties over the glycol extended polyurethanes as demonstrated in Figure 1.5. Higher polarity difi'erence between hard and soft segments and three dimensional hydrogen 13 FIGURE 1.6 Differential Scanning Calorimetric Curves of Urethane Polymers Containing Various Glycols rs 100°C CURE 7 w W 90% vr o v so w so III 40 Load (kg) ‘r 29 min ‘ ado robo ‘ ’ Strain (96) FIGURE 1.7 Load vs. Strain for Polyurethanes at Various % Hard Segment Content 14 bonding, madepossiblebytheextraNI-Ig'oupintheurealinkage, havebeenproposedto account for the property improvement (Schwarz et al, 1979). Figure 1.7 (Chang et al, 1982) shows how composition changes the stress-strain behavior of polyurethanes based on MDI/BDO/PPO—EO. With increasing hard segnent content, the segnented polyurethanes exlu‘bit a wide range of behavior, flom a soft rubber at low hardsegnuncomenttoalughmodulushardplasficathighhardsegnemcomau. LLLMMMSIM Modular assemblies are widely used in the automotive industry. This concept involves supplying of modules or integated subsystems to automobile manufacturers. A gowing portion of rear quarter windows, Windshields, and backlites are supplied as modules which can be directly attached to body sheet metal of an automobile. Figure 1.8 shows one such module. The conventional gasketing technique of gluing or otherwise mechanically attaching an extruded elastomeric gasket to glass can no longer meet design and performance needs of today's automobiles with improved aerodynamics and better styling. In the modular approach, all the componmts required for fit and function of a window are supplied to automobile manufacturers as an integated preassernbly suitable for direct installation at the automobile production line. These components typically include glass, elastomeric gasketing material, tracks for sliding the window, and associated trims such as bezels, stud mounts, etc. The modular glass part is produced by placing the glass in a molding press FIGURE 1.8 An Encapsulated Modular Window 16 andtheninjectingapolymericgasketingmaterialintothemold cavity. Themoldcavity maycontainclips, studs, andguidesthatareusedtoultirnatelyfastentheglasswindowto afl'ameofan automobile. Theseclipsand studsgetmoldedwiththerestofthemodule and thus form a one-piece molded modular assembly. Modular glass is amenable to complex and curved glass shapes, variable gasket cross-sections, and incorporating trim materials. Someoftheadvantagesinusingmodularglassare: - Single source responsibility for quality and (1ng changes - Elimination of stack up of tolerances associated with various window components - Reduced handling damages - Less installation labor - Less storage space - Suitability for just-in-time inventory controls - Robotic installation - Drag and noise reduction in vehicles The two primary modular window gasketing materials are plasticized polyvinylchloride (PVC) processed by injection molding and methane processed by reaction injection molding (RIM) (Agawal et al, 1991). PVC is the preferred material for fixed small side windows while RIM urethane is used for larger and more complex parts like windshields and backlites. The high melt viscosity of PVC during injection molding makes it dificult to an larger window cavities. High molding pressure and temperature associated with PVC are known to cause excessive glass breakage and laminating material degradation in lower strength laminated windshields. 17 RIM urethane's low viscosity (~ 1 Poise) and low process temperature (~75°C) make it auitableforgasketingmaterial inalltypesofmodularwindows. Urethanesforwindow encapsulation fall into two categories. Those based on aliphatic isocyanates are referred toasaliphaticurethaneswhereasthosebasedon aromaticisocyanatesarereferredtoas aromatic urethanes. Aliphatic urethane systems are inherently resistant to discoloration fiom solar weathering whereas aromatic urethanes are not. Due to cost and toxicity concerns, most RIM encapsulated modular windows use aromatic urethanes based on 4,4'-diphenylmethane diisocyanate (MDI) and are required to be painted to maintain weatherability. Painting of urethane surfaces can be done either in an in-mold coating (IMC) process (Agrawal, Fox and Lynam, 1991) or by painting alter the molding operation. In the IMC process, a coating material is applied onto Class "A" surfaces of a mold. Upon injection of urethane into the mold, the coating bonds to the fieshly formed urethane surface, detaches fiom the mold and becomes an integral part of the encapsulant. Adhesion between the polymeric gasket and the glass panel is crucial for a modular window to firnction. Adhesion between the gasket and the glass ensures glass retention in an automobile, avoids water leakage and maintains structural integity of the window in the long-term use. Neither PVC nor RIM urethane adhere well to the glass panel by itself. Adhesion promoters are applied to the glass perimeter prior to encapsulation for promoting adhesion of gasketing materials. 18 W Ihaeareseveralflleoflesofadhesionbetweentwodissimilarmbsuates. Accordingto Kinloch (Kinloch, 1937), these theories can be categorized as follows: (a) Mechanical Interlocking (b) Difliuion “wow . (c) Electronic Theory (4) Adsorption Theory The mechanical interlocking, as the name implies, pr0poses that mechanical keying, or interlocking, of one substrate into the irregularities of the other substrate surface is the major source of intrinsic adhesion. The difl‘usion theory states that the intrinsic adhesion ofpolymerstothernselvesandtoeachotherisduetomutual diflirsion ofpolymer molecules across the interface. The electronic theory of adhesion suggests that electrostatic forces arising flom contact of two dissimilar substrates (due to electron transfer) may contribute sigiificantly to the intrinsic adhesion. The adsorption theory proposes that, provided nrflicicntly intimate molecular contact is achieved at the interface, the materials will adhere because of the interactomic and intermolecular forces which are established between the atoms and molecules in the surfaces of the substrates. Interactions that are reversible, such as Van der Waals forces and hydrogen bonding, or irreversible, such as ionic, covalent, and metallic bond formation may contribute to the overall adhesion. 19 The adsorption theory is the most widely applicable and accepted among all the theories. Forglass/polymer systems, adsorptiontheoryhasbeen successfirllyusedto explainthe adhesion phenomena. Polyurethanes in general exhibit good adhesion with glass due to severalfactorsincludingtheirpolarnature, surfacewettingpropertiesand chemical reactivity with a variety of functional groups (Hepburn, 1991). It is generally accepted thatthesurfaceofglassiscoveredbysilanolg'oups(Si-OH)asaresultofinteractions with the atmosphere (Mohai et al, 1990; Markus et al, 1981; Vaughan et al, 1974). It is quite conceivable that reaction of these silanol goups with the isocyanates present in the urethane formulation could account for the adhesion levels achieved between the glass/polyurethane systems. Adhesion promoters such as coupling agents are commonly used in glass/polymer systems to improve adhesion and increase environmental stability of the bond. Silane based coupling agents are the most widely used adhesion promoters for glass/polymer systems. The structure of such silanes may be represented by the general structure Y (CH,),SiX,, where n = 0 to 3, Y is an organofunctional group usually selected for reactivity with a given matrix and X is a hydrolyzable goup on silicon (Plueddernann, 1991). Several researchers have studied bonding mechanisms of Silane coupling agents to glass surfaces. Garbassi et al and Vaughan et al have used surface analysis techniques such a x-ray photoelectron spectroscopy to study the bonding mechanisms (Garbassi et al, 1987; Vaughan et al, 1974). It is generally accepted that the mechanism of adhesion is through the formation of covalent bond across the interface. Koenig et a1 (Koenig et al, 1971) using Laser Raman spectroscopy and Chiang et a1 (Chiang et al, 1980) using fourier 20 transform infiared spectroscopy have clearly established the presence of Si-O-Si bonding across glass/silane interface. Shown below is an idealized monolayer of bonded silane coupling agent on glass surface. i i i i 1: Ho—Si—o—Si—o—S'—o— i—o—si—o \ 0H 0 1) (I) H 9 | //%.E/// For the coupling agent/polymer matrix interface, the reaction of organoflrnctional goup 'Y’ with the matrix is believed to be a key mechanism of adhesion. It has been suggested, however, that at this interface, some form of interpenetrating network structure might well be formed instead of a simple covalent bond (Kinloch, 1987). In glass/polyurethane systems, the most commonly used silane coupling agent is aminosilane due to the possible reaction of the amine functionality with the isocyanates present in the urethane formulation. A common aminosilane is aminopropyl triethoxysilane H,NCH,CH,CH,Si (OC,H,),. f! 21 Aminosilanes have been used in fiber sizing compositions to improve adhesion ofgass fibers to urethane matrices (McWilliarns et al,1974). Drown et al (Drown et al, 1991) have audied the role of various glass fiber sizings on fiber/matrix adhesions. Aminosilane based sin'ngsinRRIMmethaneshavebeensmdiedbyseveralresearchers(Galli, 1982; Otaigbe,1992)andbeenfoundtobeveryimportantfortensileand othermechanical properties ofthe resultant composites. Darnani and Lee (Damani and Lee, 1990) have studied glass fiber/polyurethane interphase using single fiber fiagnentation test (Broutrnan, 1969). They have found the fiber sizing to be an efi‘ective tool for improving chemical interaction during the interphase formation and for improving the overall adhesion of the matrix to the fibers. CHAPTERZ W Whmmethanewnsfimuusuebrougluinconuawithglassmrfaceahsadhedonadflbe influenced by many factors, physical and chemical in nature. The physical interactions are causedbyVandaWaalsforceswhichcanbeatm'butedtodifi‘ereruefi‘ects: (a) dispersion forces arising fi’om internal electron motions which are independent of dipole moments and (b) polar forces arising flom the orientation of permanent electric dipoles and the induction efi‘ect of permanent dipoles on polarizable molecules. The dispersion forces are usually weaker than the polar forces but they are universal and all materials exhibit them. Another type of force that may operate is the hydrogen bond, formed as a result of the attraction between a hydrogen atom and a second, small and strongly electronegative atom such as oxygen, nitrogen, fluorine or chlorine. Thermodynamics is a very useful way of descn’bing and quantifying some of these physical interactions. The surface-free energy (7) or surface tension is a fundamental parameter which can describe the interactions of liquids with other phases. In the case of a liquid droprestingonasolidsurface, ywisaforceperunitlengthactingonthemrfaceofa solid in equilibrium with the liquid vapor. The molecules at the surface of a liquid are subject to very difi‘erent forces than molecules in the bulk liquid. For the liquid vapor interface, molecules in the bulk have the same environment in each direction. Long 22 23 dimceatuacfiveforcesmdshondiumcerepulfiveforcescwsethemoleadesto maintain an intermediate spacing. Molemles on the surface have weaker interactions with the vapor phase but strong interactions with other liquid molecules. These strong forces pullthesurfacemoleculestowardtheliquid phase, sotheyoppose spreading. Figure2.1shomtheveaonwluchrepmuumeforcuacfinguthebmndarybetweala solid (S), liquid (L), and vapor (V). The equilibrium interactions in terms of surface tensions are represented by the Young equation: 7”, - 1,; - 73L COS 6 (2.1) The Young equation allows determination of information about the solid surface flom knowledge of the liquid-vapor surface tension and the measured contact angle. The chemical interactions can lead to chemical bond formation across the interface. In glass/polyurethane system, such chemical bonds may include covalent and ionic bonds. These physical and chemical interactions along with glass surface roughness and porosity have been shown to create an interphase region in polymeric composites (Drzal, 1983; Verpoest et al, 1988) as illustrated schematically in Figure 2.2. The interphase region is definedastheregionthatisformedasaresultofthebondingbetweenthefiberand matrix; it has a siglificantly distinct morphology or chemical composition as comwed to the bulk fiber or the bulk matrix. The interphase may be a difi‘usion zone, a nucleation zone, a chemical reaction zone, and so forth, or any combination of the above (Swain et al, 1990). 24 Vapour Lkufid ///s;57///////‘ FIGURE 2.1 A Liquid Drop Resting at Equilibrium on a Solid Surface 25 0 : \Zit‘. ,1 «”irgi‘fiig‘f FX’:.\ .' 1. . “a \\\\\\\ o ‘\ " -. N\ ° W W/ // J " Ill-IIIIII' a A a. The bulk fiber b. A fiber surface layer possessing a difi'erent microstructure or chemical composition compared to the bulk fiber. c. All outer fiber layer altered by fiber surface treatments d. A layer in which the fibers bond to the sizing A sizing or coupling agent layer A layer, in which the matrix microstructure or chemical composition or both gadually changes flom that of the sizing to that of the bulk matrix g. The bulk matrix can FIGURE 2.2 Schematic Representation of Interphase Region in a Polymeric Composite 26 Thechanaaisficsofthismtaphaseregonhawbemshowntodepwdonwvadfaaom includingthe composition ofthematrix, surfacechernistryofglass and processing vafiablesindudingauefimetemperamreandpresmre. Thethiclmessofthisregioncan extendfl‘omafewtoafewhundrednanometers. Amulticomponentinterphasehasa complex microstructure or chemical composition or both and can have siglificant efi‘ects on the physical and thermomechanical properties of the composite. In comparison to other polymers used in composites such as epoxies, which have single phase morphology, polyurethanes in general are phase separated systems with multiphase morphology. The hard and soft segnents in polyurethanes phase separate into hard and soft domains. Figure 2.3 is a transmission electron microgaph of a polyurethane sample showing hard and soft domains (Oertel, 1993). Polyurethanes build their structure florn phase separation. The hard segnent domains act as internal reinforcements in the matrix. A majority of the mechanical properties in segnented polyurethanes depend on the phase separation. Thephasesepamfionhlthebulkofthepolyurethmemauixcanhaveasiguficant influence on its interphase region with glass surfaces. A number of studies have shown that the surface of polyether based urethanes can exhibit an enrichment of polyether compared to the bulk (Y oon and Rattler, 1988; Heam et al, 1988). Heam et al have studied a cast urethane surface using secondary ion mass spectroscopy (SIMS) and x-ray photoelectron spectroscopy (XPS). They have reported that the surface of segnented 27 FIGURE 2.3 A Transmission Electron Micrograph of a Segmented Polyurethane 28 polyurethane was enriched in polyether (soft segnent). The surface layer of polyether was not pure but was interdispersed (in the upper IDA-15A) with small quantities ofhard segments. Recently, Deng and Schreiber (Deng and Schreiber, 1991) have discussed orientation phenomena at polyurethane surfaces when brought into contact with difi‘erent media. YoonandRatner (YoonandRatner, 1986) have related thephase separation to its surface composition and found that where siglificant phase separation takes place, little or no hard segnent could be found in the outermost few molecular layers of the polymers. In addition to the phase separation, the reaction kinetics between isocyanate and polyol, and especiallyinaRIMurethane system, isveryfastand complex. Therelativereactivity of phenyl isocyanate with various active hydrogen compounds is shown in Table 2.1 (Macosko, 1989). In addition to the isocyanate-hydroxyl reaction, isoyanates can react with themselves to undergo dirnerization or trimerization, or can react with urethane and urea linkages to form allophanate and biuret linkages respectively. The isocyanates in polyurethanes can also react with the active hydrogen species present on the glass surface. The glass surfaces are usually found to be hydrated due to adsorption of water vapors. It has also been suggested (Pantano, 1981) that NaHCO, type species may be present on the surface due to adsorption of carbonaceous species flom the atmosphere. The surface ofsoda-lime glass has also been shown to be rich in Na+ ions relative to the bulk. Many possible reactions of urethanes with these surface active species could lead to covalent and ionic bond formations. 29 TABLE 2.1 Relative Reactivity of Phenyl Isocyanate with Various Active Hydrogen Compounds Time in s to 25% conversion butylcu'banilate CH,(CI-1,),-O-CO-1~lHoCJ-I, 3 x 10’ (films allophanate) diphalyl urea CJ-irNH-CO-NH-Cfi, 1800 forms biuret) ‘ wata’ 11,0 450 2-butmol CH,CH 0H CH, CH, 300 l-butanol HO(Cl-1,),C1-l, 92 l-butanol 25 + 0.1 mol% dr‘butyltin dilaurate (DBTDL) 1-butanol+ 2 mol% DBTDL 6.5 l-butanol 56 + 2 mol% l,4-diazabicyclo (2.2.2) octane (DABCO) l-butanol ~10 + 0.2 DABCO + 0.1 DBTDL o—toluidine i-l,N-CJL(CH,) l9 o-toluidine + 2 mol% DABCO 7.5 aliphatic amine 1-1,N[Ci~l(CH,)C1-l, - ~10" OLCILCH, 30 The fast and complex reaction kinetics in polyurethane systems, the phase separation kinetics, the multiphase morphology, and the possibility of mrmerous physical and chemical irneractionswiththeglassnrrfacemaketbeglass/polyurethaneadhesion studya complicated and challenging phenomenon. W There is extensive data reported in literature related to the adhesion of polymeric materials to glass. But the majority of the work reported is limited to homopolymers (e.g., polyolefins) or single phase morphology polymer systems (e.g., epoxies, polyesters, etc.) The small amount of work done related to polyurethane/glass adhesion (see Chapter 1) is either empirical in nature or is an extrapolation of the findings based on homopolymers or single phase materials. As explained earlier in this chapter, polyurethanes, in the majority of instances, are multiphase copolymers commonly processed through reaction injection molding. In contrast to homopolymers or single phase materials, polyurethanes undergo extremely fast and complex reaction kinetics coupled with phase separation and morphology development. All of these factors can significantly influence structure! morphology, physical, mechanical and chemical properties of a glass! polyurethane interphase region. This study is an attempt to use the existing literatures on structure/property relationships in polyurethanes and the surface chemistry of glass, and couple that with the understanding of adhesion of homopolymers or single phase materials, to develop a thorough understanding of polyurethane to glass adhesion. The objective of this research 31 istoidarfitysnduudyimfldslphenomararelatedtoadhesionmechsnimof polyurethanetoglsss. Thefomsofflfisreswchhuoundthehnaphasebetweenapolymethancmauksnda glass substrate. The characterization for structure/morphology, physical, mechanical and chemical properties ofthe interphase region are the subject ofthis work The factors controlling the interphase region such as matrix chemistry, its properties, phase separation and glass chemistry are studied in detail in order to develop theoretical models supported with experimental results to further the understanding of adhesion mechanisms. Various adhesion theories are explored to determine the important adhesion mechanisms acting at the glass/polyurethane interphase. The experimental anddreoreticalworkinflristhesisisorganizcdmainlyinflueechapters, 3, 4, and 5. Each of these chapters is self-contained in that there is an introduction section to overview the pertinent literature, an experimental section, a results and discussion section, and finally, a conclusions section. The glass surface chosen for this study is soda-lime float glass due to its wide use and availability in plate form. The main constituents in a soda-lime float glass are Na20:Ca0:GSi0,. There may also be small amounts of A130,, Mg0, Ba0, 3,0,, and 110,. Typical soda-lime float glass compositions are shown in Table 2.2 (Uhlmann and Kreidl, 1983). There are two sides to a soda-lime float glass, the tin side and the air side. The air side is used in this study for adhesion testing. 32 TABLE 2.2 Typical Compositions of Soda-Lime Glass (Wt. %) 1972 1977 VARIATION 810, 72.19 72.15 66.2 - 74.7 111,0, 1.81 2.13 1.25 - 2.5 1.1-,0, 0.12 0.11 0.07 - 0.18 CaO 9.55 10.66 9.16 - 13.40 Mg0 1.51 0.91 0.55 - 1.91 380 0.17 0.08 0.0 - 0.47 191,0 13.96 13.83 12.88 - 17.30 11,0 0.59 0.57 0.40 - 0.85 so, 0.16 0.14 0.08 - 0.22 33 Theurethsnematrixdevelopedforthisstudyistransparent. Thepolyolusedisa caprolactone based triol available fi'om Union Carbide under the trade name "Tone”. The dfiwcyanueusedismMenedfiwcymnesdectedforhsuymmeUymdthechsinmenda usedisbunnediol.1hechanicalsuucmresofdltheseconsfima1tsueshownm Figure 2.4. The third chapter is focused on developing model urethane matrix systems, developing adhesion testing methodology, building structure-property relationships in these polyurethanes and finally correlating the findings to their adhesion behavior to glass surfaces. The model urethane matrices cover a wide range of properties fiom being very soft and elastomeric to very rigid and of high modulus. Difl‘erent polyol molecular weights are used with difi'erent amounts of hard segments to tailor make polyurethanes with varying mechanical properties but the same chemistry. This type of experimental design allows for the study of the efl‘ects of polyol molecular weights and the hard segment contents on polyurethanes structure, property, and adhesion characteristics. In Chapter 4, a theoretical model is developed to predict phase separation in polyurethanes. The model utilizes solubility parameters of the hard and the soft segments to determine a miscibility interaction parameter between the two phases. A phase diagram for copolymers derived fiom scattering studies, correlating block lengths with their weight fiaction compositions, is used to calculate onset of phase separation. By comparing the onset of phase separation with the miscibility interaction parameter, the model is capable of predicting extent of phase separation in polyurethane formulations with variables such 34 IO 0 H ioMjD‘io/Ro’W21H Caprolactone based polyol H3 H3 . NCO OC ' NCO NCO Toluene diisocyanate “iii-E3“ 1,4 Butanediol FIGURE 2.4 Chemical Structure of Polyurethane Constituents 35 aspolyolmolecularweights, lengthofchainonendermdisocyanateaandthehardand sofi segment contents. Experimental techniques such as near-infiared (NIR) and fourier transforminfi'areda-‘I'IIO spectroscopyareusedtoverifythetheoretical predictions. AnglihrdcpendentXPS (ADXPS)isutilizedtoanalyzethefailed glasssurfacefi’om adhesion testings. Composition and thickness ofthe interphase region are also analyzed using ADXPS. Matrix phase separation is correlated to interphase composition and its adhesion. Based on the findings, a “beneficial” interphase region is created to improve adhesion of the poorly bonded polyurethane matrices. It is firrther demonstrated that by coating glass surfaces with a thin layer of butanediol (chain extender), adhesion of poorly phase separated polyurethanes can be significantly improved. The fifth chapter is focused on investigating possible physical and chemical interactions in the interphase region. A theoretical model is developed to predict surface-flee energies of the various polyurethanes. Contact angle measurements are done to experimentally determine polar and dispersive components of the surface-free energies. The surface-free energy data is used to calculate the work of adhesion and is correlated with the experimental adhesion values. The preferential segregation of butanediol to the interphase region is correlated to polar surface fiee energy. Silane coupling agents are used to treat the glass surface and with the use of the XPS technique, chemical interactions between urethane and glass surfaces are explored. Mechanical properties of the interphase region with excess butanediol are also evaluated to study the efi‘ects of the preferential segregation of butanediol on adhesion. 36 Finally, Chapter 6 presents the conclusions in a manner to coherently tie together the observations and findings of all the work done in this project. Also, Chapter 6 provides gtfiddinuforwmefimueworkwhaethefindingsofthisworkanbemdedmfimha dudduethegludurethmeadhedonmwhanimhrdifi'aanmethmemauicuandother glasssurfaces. Theworkpresentedmflfischaptuhubeenacceptedforpubficafionintheloumd of Adhesion (1994). 3.1m Polyurethanes were prepared fiom toluene diisocyanate (TDI), 1-4-butane diol (BDO) and polycaprolactone based triols with varying molecular weights. Among each molecular weight triol based urethane, hard segment content was varied from 20% to 70%. Difl‘erential scanning calorimetry, tensile testing, and Iosipescu shear testing were done on all the various urethanes prepared. Thermal characterization data revealed the dependence of phase separation on hard segment content as well as on the trio] molecular weight. Tensile data and Iosipescu shear data firrther confirmed the observations made fi'om the DSC data. The data fiirther indicated that phase separation can greatly improve modulus of cross-linked segmented urethanes. Adhesion of these urethanes to glass surface was evaluated using soda-lime float glass plate. Urethane samples were cast on the air side of the glass plates and adhesion was measured in shear mode. Adhesion data indicated that in addition to hard segment content, modulus, cross-link density, and molecular weight of the triols; phase separation seems to be a major factor in controlling adhesion. Surfaces of the failed adhesion samples were also analyzed and the failure mode was found to be cohesive with varying degree with the difl‘erent urethane systems. 37 38 W Reinforced reaction injection molding (RRIM) and structural reaction injection molding (SRIM) ofurethanesarewidelyused intheautomotiveindustryto makebodypanels, fascias, bumper beams, etc. In these applications, good adhesion between the urethane mauixsndthereinforcunun(umaflyglau)isshownmbeanimponamfactorin determining overall mechanical properties of these composites (Dawson and Shortall, 1982; Kau ete al, 1989; Yang and Lee, 1987; Schwarz et al, 1979). Another important and emerging application of RIM-urethane is integral molding of gaskets onto glass panels to produce modular window assemblies (e. g., windshields) for automobiles (Reilly and Sanok, 1988; Fielder and Csrsell, 1990; Agrawal et al, 1991). In these modular windows, good adhesion between urethane gasket and glass panels is essential for the structural integrity of these assemblies in the automobiles. In the above mentioned applications of RIM-urethane polymers, urethane matrix properties vary from one extreme to the other, fiom being very soft and elastometric in modular window application to very rigid and of high modulus in RRIM and SRIM applications. There is extensive data reported in literature (Sanders and Frisch, 1983; Macosko, 1989) relating final properties of urethanes to various formulation parameters. These parameters control cross-linking density and phase separation in segmented polyurethanes, thus determining final matrix properties. It has been shown (Zdrahala et al, 1979; Chen et al, 1987; cm et al, 1982; Camargo et .1, 1985) that phase separation depends on individual segment length segment length distribution, intra- and interdomain hydrogen bonding, and several other factors. Recent work by Rao and Drzal (Rao and 39 Drzal, l991)hasshownthatforthesamemrfacechemisfiyandmatrixchemisuy, adhesionvariesdirectlywiththematrixmodulusinglassycross-linked epoxies. This smdyismauanptwmthisrdafionshipmmhasystammdcondatethesuucnue property relationship ofsegrnented polyurethanetoitsadhesion characteristicstoglass substrates. Model urethane matrix compositions have been developed that produce a transparent matrix which can be easily prepared and studied in laboratory. Thermal and mechanical characterization of the matrices have been done to establish matrix properties. Adhesion characteristics of the various matrices to soda-lime float glass plates have been evaluated and correlated to their compositions, structure, and properties. W In this study, we have not used any catalyst in the urethane formulations, thus increasing handling time for sample preparation. Also, the polyol used is a triol, allowing us to prepare and study a wide range of mechanical property urethanes without having to change its chemistry. Urethane formulations have ranged from the lowest possible hard segment to 100% hard segment. In this paper, the hard segment content is defined as the percentbyweight oftheisocyanateandthechain extenderinthepolymerat afixed stoichiometry or isocyanate index. Thus, in the lowest possible hard segment formulation, thereisnochainextenderwhereasinthe100%hardsegmentformulation,thereisno polyol. 40 To study the role of polyol molecular weights, three different polyol molecular weights were used to prepare urethane formulations with the same hard segment contents. W W The polyurethanes used in this study were caprolactone-based trifirnctional polyols available from Union Carbide under the trade name ”Tone.” Characteristics of these polyols are listed below: Polyols Supplier Molecular Weight Hydroxyl Number Tone 0310 Union Carbide 900 187 Tone 0305 Union Carbide 540 312 Tone 0301 Union Carbide 300 S60 Hard segments were made from a 80°/o-20% mixture of toluene 2,4-diisocyanate and toluene 2,6-diisocyanate ('I'DI, Aldrich Chemical Co.), and 1,4-butanediol (BDO) as chain extender (Aldrich Chemical Co.). H3 H3 I NCO CC I NCO NCO Toluene diisocyanate 7??? H ' —0H 0 ‘r‘ ‘f ‘f’? H H H H 1,4 Butanediol 41 The various urethane formulations studied in this work are shown in Table 3.1. Their nufacefieearergiegbuedoncomactmglemamremanambetweenWJandflz dyneslcm (AgrawalandDml, 1994). AlsoshowninTable3.1aretheweightpercent hard segrnentcontentandthemolecular weightpercross-link (Me). Mcistheunitweight ofthepolymerdividedbythemrmberofcross-linkjunctionsintheunitweightofthe polymer. The glass substrates used for adhesion testing were annealed 2" x 5" x 1/4" soda-lime float glass plaques. Adhesion testing of urethanes were carried out on the air-side of the glass plaques. A one step urethane preparation approach was used to prepare all the samples. Trio] and BDO chain extender were mixed and degassed for 2-4 hours at 60°C. Silicone molds for casting tensile dogbone specimens and Iosipescu shear specimens were also subjected to degassing at the same time. The stoichiometric amount of TDI was then added to the triol-chain extender mixture and homogenized with a magnetic stirrer for about a minute. The resultant mixture was then quickly cast into the degassed silicone molds. Silicone molds were then heated for 24 hours at 90°C in a convection oven. After curing, samples were taken out of the moldsiand were sanded and polished to achieve uniform thickness. 42 TABLE 3.1 Urethane Formulations at Isocyanate Index of 1.0 43 E 5 I] . I . ‘ Theairsideof2'x5'x1/4" annealed soda-limefloatglassplaqueswascleanedwith methyl-ethyl-ketone solvent and dried. Silicone molds with a 1/4' x 1/4' x 1/4' cavity wereplacedontheairsideoftheglassplaquesandweresecuredtotheglassplaques usingclamps. Thedegassedandhomogenizedmixhrreofthepolyokthechainernender, andtheisocyanatewasthenpouredintothe 1/4'x1/4'x1/4" cavitiesformedbytheglass plaques and silicone molds. The glass plaque-silicone mold assemblies were then cured for 24 hours at 90°C in a convection oven. Upon cooling, silicone molds were separated fiom the glass plaques and the samples were stored for adhesion testing. Figure 3.1 shows the drawing of an adhesion sample with two urethane blocks cast on a soda-lime glass plate. 3 I 3 I I Q] . . D'fli 'lS . C . [DSC] DSC scans on all the cured urethane samples were done on a Shimadzu TASO thermal analysis system. Sample weights used for the scans were approximately 20-25 mg. The DSC cell, with the sample inside the cell, was cooled down to ~80°C by liquid nitrogen and scans were done at 10°C/min to up to 280°C. For the second DSC scan on the same sample, the DSC cell was allowed to cool down to ambient temperature and then liquid nitrogen was used to further cool it down to -80°C. E 0 .m mus _ .0 .. mam 28595 u< so: no. am 29: \ V 320 Q assesses: o: 45 D 'lll°l!l°mllil Rectangularbars(30mmx4mrnx1.25mm)ofvarioustnethanesampleswereusedfor dynamicmechanicalanalysisonaSeikoInstrument‘sDMS-m system. ValuesofE' and tanb uvafioustunperanueswereobtainedmadampedthreepointbmdingosdnmon mode of deformationat 1 Hzfixed frequency. Theternperaturewasvariedfiom-70°C to 280°C at 10°Clmin. I . SI I . All urethane samples for the Iosipescu testing were sanded and polished to a uniform thickness of 2.5 mm Strain gage rosettes (fi'om Micro Measurements Inc.) were attached to the front of each specimen. Testing was done on a servohydraulic material testing machine MTS 900 using a modified Wyoming fixture (Ho et al, 1993) at .OS'Iminute crosshead speed. Strain gage on each sample was connected to a wheatstone bridge with a half-bridge configuration. The wheatstone bridge was connected to a signal- conditioning amplifier, and the amplified analog signal was converted to digital signal through a circuit completion box which was connected to a microcomputer controlled data acquisition system. At least three samples were tested for each urethane formulation. For softer urethane formulations such as 10A and 108, reinforcing tabs were glued to the sample ends to facilitate testing. I '1 I . Dogbone-shaped urethane samples were tested on a tabletop material testing machine, Instron 4201, using pneumatically actuated grips. Grips were separated at 2"lminute and 46 theamplestress—suainarrvetoitsfailurewasrecordedonachanrecorda. Atleastfour samples were tested for each urethane forrmrlation Ell . I . Adhesionofmethmemanbrmglassmrfacewuevduatedusingwda-fimeflouglass platesuflrembstrate.Plateglasswuchosa1rathaflunfibmutheglass-msuix mta'facd'unaphasemplaeglmwuldfimhabemdyndwithrdafiwaseudngviwal, microscopic, spectroscopic, and chemical means. The lap-shear configuration for adhesion testing with glass plates could not be used successfully due to the brittleness of the glass plates. In the lap-shear configuration trials, glass plates broke during sample loading or during sample testing due to the slight misalignment or bending. To overcome this, ASTM test method D4501 for measuring shear strength of adhesive bonds between rigid substrates by the block-shear method was modified for this study. Figure 3.2 shows the front and the rear view of the test fixture with an adhesion sample clamped in place. Instron 4201,tabletopmaterialtestingmachine,wasutilizedfortheadhesion testing. The testfixturewasmountedonthelnstron Anadhesionsamplewasloadedinthetest fixturecarefullysuchthatthecasturethaneblockofthesamplewouldengagewiththe shearing bar ofthe test fixture. Upon the sample loading, thejaws ofthe Instron machine weremovedapartatO.2"/mimrte. Inthis fashiortthecasturethane blockwasshear loaded in a plane parallel to the glass plaque. The maximum load required for the 47 82m 5 woman—U 5.50on 5:5 235m “0 33> Eon 3 82m 5 comp—£0 5.50on 5:» 235m Co 33> :55 Au 235m @389. 5.505? 05595-820 Nd 5505 48 detachment oftheurethaneblockfi'omtheglasssurfacewasrecorded. Atleastfive sarnplesweretestedforeachmethaneforrmrlation. Afierfaflureglusandmethane samplesweresaved for failuremodeanalysis. WWW Glasssurfacesofthefailed adhesion sarnpleswereexsminedby SEM. Thesurfaceswere gold coated by a Denton Vacuum DESK II coater. A total thickness of approximately 100A gold film was deposited on the sample mrfaces. An 181-88130 scanning electron microscope was used to examine the samples. A 50X magnification was utilized in the SEM examination. MW Subsequent to adhesion testing, failed glass surfaces were analyzed using Perkin-Elmer P1115400 x-ray photoelectron spectrometer. Approximately 1/4" x 1/4" square area was sectioned fi'om the failed glass surface and was placed inside the XPS chamber. The XPS spectra were obtained at a base pressure of approximately 10" Torr. The standard Mg Kx source was used for all samples analysis and was operated at 300W (15 kV, 20 mA). A continuously variable angle sample stage was used and was set to 45° (photoelectron take- ofi‘ angle). The portion of the sample analyzed by the spectrometer is set through an initial lens system andwassetfora2.0mmdiametercircle. Datawascollected inthefixed analyzer transmission mode utilizing a position sensitive detector and a 180° hemispherical analyzer. Pass energies were set at 89.45 eV for the survey scans (0-1000 eV) and at 49 35.75 eV for the narrow scans of the elemental regions. Data collection and manipulation was perfonned with an Apollo 3500 workstation running PHI ESCA software. W Figure 3.3 shows a simplified two-dimensional schematic representation of possible molecular arrangements in some of the urethane formulations studied in this work Figure 3.3A shows urethane formulation 10A which has no chain extender. From the schematic, it is clear that sample 10A is a cross-linked single-phase urethane system. Figures 3.33, C, and D represent urethane formulations 1013, SE, and 1E respectively, all with the same hard segment content (~67°/o). among these three samples, sample 10E has the highest amount of the chain extender. This can lead to longer hard segment chain lengths resulting in better phase separation (Macosko, 1989; Zdrahala et al, 1979). In Figure 3.3D, the soft segment chain length is almost equal to the chain extender and thus very little phase separation is expected. 3 E I II 1 Cl . . Phase separation phenomena in segmented urethanes can be related to their thermal transition behavior. DSC analysis was used to study thermal transitions in all the synthesized urethane samples. DSC thermograms for Tone 0310 are shown in Figure 3.4. Therecanbeseveralthermaluansifionsinasegmentedurethanerelatedto soft segment, hard segment, hydrogen bonding between domains, crystallization, phase melting, etc. (Bymc et al, 1992; Chen et al, 1992; Hepburn, 1992). In Figure 3.4, 8011 segment transition, which is below 0°C, is dificult to detect consistently and reliably, and thus is 8550.5 :_ mucoEomSEa‘ 3.300.023 .302 nd g0?— m m . 29:8 O n I can u . . 28.3 o .9. J\ 38 2.3. fwfl 33 2.8 a .r: \r\, o w 28 2.8 5 m n 37.55am. (n < 2 . seen..." 51 End a C 28:00 2.2:me .5: waves; 5.3 855238 83m 8 8 28h a6 «52882; 08 a...” menu—a .0...th 8.8... 8.8. 86 <0 — .88. mo F .86. 00 P 1 a . F 8. .86 m: 18.Ns .3... umo 52 notreportedhere. ThethermaltransitionaboveO°thichisthemostprominentinthese scmscanberdatedtodrehardsegmau.1hishudsegmauthamdmsifiondmis shown in Table 3.2 for all the samples based on Tone 0310 (the highest molecular weight trial) based polyurethanes. The glass transition temperature for all the Tone polyols is approximately -60°C, whereas for the formulation with 100% hard segment (Sample HS, Table 3.1) is 98.4°C. From Figure 3.4 it is clear that in Tone 0310 system, as the hard segrnerrtconterrtirrcreaseathen'ansitionternperamreincreases. Withhigherhard segment content, hard segment chain length increases and longer chain lengths improve phase separation (Macosko, 1989; Zdrahala, 1979). Thus, the increase of thermal transition temperature with hard segment suggests that degree of phase separation improves with hard segment content. Figure 3.5 shows second DSC thermograrns on the same samples afler annealing and quenching. The thermal transition data fiom the second DSC run are, in general, higher than the corresponding thermal transition temperature fi'om the first DSC run. This suggests that sample annealing-quenching improves phase separation in segmented polyurethanes (Hepburn, 1992). The first and second DSC run thermograrns of Tone 0305 and Tone 0301 polyol-based urethanes are shown in Figures 3.6, 3.7, 3.8, and 3.9 respectively, and the thermal transition temperature data is shown in Table 3.2. In the first and second DSC runs of Tone 0305 system, the thermal transition temperature increases with increasing hard segment content. But sample annealing-quenching does not seem to have much efl‘ect on the transition temperatures as evident by comparing the corresponding first and second DSC transition temperatures. In Tone 0301 based polyurethanes, transition temperature decreases with increasing hard segment content in the first DSC run and remains the same or increases somewhat in the second DSC run. 53 TABLE 3.2 Efl'ects of Hard Segment and Polyol Molecular Weight on Thermal Transitions in Cross-Linked Segmented Polyurethanes Sample . . lst Run . . 2nd Run Transrtron Termerature (°C) Transrtron Temperature (°C) 10A 6.1 5.3 I 10B 19.2 26.5 1 10C 35.5 36.3 10D 46.1 56.7 I 10E 54.3 68.2 I 5B 44.4 49.4 I SO 59.4 59.9 I SD 64.0 64.8 I 5E 73.0 - 72.2 I 1C 100.6 102.2 1D 98.7 101.3 I 1E 86.7 106.9 HS 98.4 109.8 54 8.8a 00.8w ASE 05V 0:00:00 2.25% Bum wcmba> firs 85583.0m 008m 9 no 0:8. .«0 mEeBoscoxh 0mm On 9503 .03th 098' - 0— we— 9.. m0— 0* (OF >>E own $5 €3— 8: 82:00 Eon—mom Ea: marte> 5.3 852.236.. .88: 88 2.8 :6 “signage 0m: 3 950:. BEES. DodoN 8.8. 8.0 d a a I q _ a a u q - mI own 56 A53 0:8 0:00:00 Eon—mow :33 9.3.5, 5.3 8:28.an 83m 88 25... no 8885.0. 0m: 2“ 550.: 8.094. 11.... '4 q q q .0...th 8.8— cod d d 1 q d - mI .omd. 69". 100.0. 1.6.No >>E own S7 23— 35 28:00 Eon—mom Bum m:_b:> . 5.3 8:28.28: 88: 88 25: :6 mesaoéofi 0m: .3 850.... .03th 8.8a 8.8. 8.: d‘ 4 d 4 11 .- d d d. dr - 58 OOdON Acsm ES 3850 EoEwom Ea: w:_ba> 5.3 85.52.53: 88: 88 9.8. 0o 28865.85 0m: 3 $50.... .065... 8.00— 8.0 - m: mI ! a J u d d d d g ago‘a >>E omo 59 WethinkthelowertransitiontemperatureoflEsampleinthefirstDSCruncouldbethe resultofextrernelyshortgeltime. ThismayexplainwhythetransitiontemperanrreoflE increases to 106.9°C in the second DSC run fiom 867°C in the first DSC run. The DSC transition temperature of Tone 0310, 0305, and 0301 based polyurethanes fiom thefirstrunaregraphedinFigure3.10andfromthesecondmnaregnphedin Figure 3.11. Fromboth ofthesegraphsitis dearthattherateofincrease oftransition temperature with the hard segment content is the highest for Tone 0310 followed by Tone 0305 and Tone 0301. This suggests that degree of phase separation increases with hard segment content at a greater rate for higher polyol molecular weight polyurethanes. These observations are also supported by the dynamic mechanical analysis. Figure 3.12 shows log (E’) vs. temperature for samples 10B and 1E which have the same hard segment content. In the figure, we see that the modulus plateau in the rubbery region for the sample 10B is flatter than that for the sample 15 even though the cross-link density in 105 is lower than 15. This suggests that the degree ofphase separation in 105 is greater than in IE (Camargo et al, 1985). 352 II I . 1C] . . All the urethane samples were tested in the tensile mode at 2"lminute cross-head separation speed. Load versus displacement response of Tone 0310 based polyurethanes is shown in Figure 3.13 and is found to be highly nonlinear. Samples 10A and 10B showed typical nonlinear and high elongation characteristics of soft rubbery material. This 5...: 0m: 8.: 852% 05583 008m .01... 2E0? 3.80.02 Eobhfi :. 58:00 EoEwom Bum m300> 0:880:80... 5.8.88... HEB—F o. .m .9505 AN. .cmEomm 0.0: 0.8 .. mwm . owm .WE 5.8 . m...m . vmm . 5...... . mm». . New . Ewe m0. 4 - m. .m . #8 Wm... l 0 .mo ”.20.. 4 88 029 - 0.5m .. mew r 0.0m 1 How 1 a. a. 58sz . New , Qwoe (D) UO!1!SUD11 loweui 61 5...: 8: 5.0 588% 9552: .88: .85 55.83 558.2... .885: c. 58:00 Eon—mom .08: «80> 8380:th 5.8.895. .558... 26. 950—...— .5 .8898. So: 0.009 mNm 0.4m 5mm odw m..© inn 5&8 mRm Ndm 8mm d — q d u A u d a — q q d — d — q u o 3.0 mZOH O o 88 020. 4 .m. Fomo mzop his I‘ll! rm. o m. 5N mm we Om mm em mm mo. Om. (O)Uomsuc>11|owreqi 62 28:00 EoEmom .5: 2.8200 .23 8558: 33m flex—om 88 2.9—. 28 2 no ocohuo 50m <2Q N. .m 950E on mLBOLmQEmH com no 0mm mm: om? mm? om mm ON n_l on! d _ . ‘ q _ ‘ _ . q . _ . _ 4 _ . d 4 _ . v .U 7 n I 1 7 1 r 4 f . ”T m 1 n w H: mm: H T 1 w u ,. 1+ 1 4 I L T l 1 J WL _ . _ . _ _ . _ . _ P _ . r r _ _ n oh: no? 08») ‘3 22:00 EoEmom v.8: 2.20.05 8 mocafiosbom 38m 2 no 25,—. mo 822.8% «5.583%:qu 2 .n 950$ Ac; EmEmuoEmé 059 mm; mm; 09; m0; 000 00.0 5.0 31.0 mfio 00.0 q a fi 4 . _ a — q 4 q _ q a q — d a « m I. >’ D 1 t I ’h" 1141‘“! [11“ It?! ' i {‘1‘ ‘ wt? 1.1 m3 <3 0 T 63 (Squ) p001 64 could be due to strain induced crystallintion in these samples (Macosko, 1989; Hepburn l992)eventhoughsamplesrernainedtranslucentsthighstrain Higherhardsegment content samples 10C, 10D, ICE and all other samples based on Tone 0305 and Tone 0301 polyols showed yield behavior. Strain at yield for these samples are shown in Table 3.3. Figure 3.14 shows a semilogarithmic plot of2% secant tensile modulus versus hard segment content of all the urethane samples. The dependence of modulus on hard segment is nonlinear for Tone 0310 based urethanes whereas is almost linear on this semilogarithrnic plot for Tone 0301 and Tone 0305 based urethanes. The abrupt change in the slope of Tone 0310 system between sample 108 and 10C suggests that modulus buildup is taking place due to phase separation (Macosko, 1989; Camargo et al, 1985). This further supports the observation made earlier with DSC data that higher polyol molecular weight increases the degree of phase separation. Tensile strength data for all the urethanes tested is shown in Table 3.3 and is graphed in Figure 3.15. It is clear that for a given hard segment content, lower molecular weight polyol system has a higher tensile strength. This could be due to the higher cross-linking density associated with lower molecular weight triols. Because the adhesive or matrix shear properties have been shown to be a key predictor and scaling parameter, Iosipescu shear testing of all the samples was conducted. Stress- straindatawasrecorded onlyupto 8% strainduetothestraingage limitation. Onlythe shear modulus was determined. Shear strength of most of the samples could not be determined as the samples could not be strained to failure due to Iosipescu testing fixture's Mechanical Properties of Various Polyurethane Systems TABLE 3.3 Strain at Yield (°/.) 67,036 150,653 10E 178,971 250,000 8.7 12,300 85 SB 128,506 1 16,000 8.0 8,000 80 SC 149,699 277,000 8.0 1 1,600 84 5D 165,404 287,000 8.5 12,600 85 I SE 171,787 306,000 9.0 13,600 85 1C 176,060 152,000 1 1.0 15,500 87 l ID 184,962 310,000 11.0 16,500 88 I 188 733 9.5 15,300 mEszm 0:285 30.5; 5 23:00 Ens—mom Bum msmco> 3.35: 0:83. Emoom .XR E .n 550—..— ANV EmEomm Ea: 0N0 0N0 v.00 0.00 wdv 0.30 v.0e 0.00 AV. 3.. 0.0m wNm . _ . _ . _ . _ . _ a _ . _ . _ . _ . Woe Q So 029 en: P000 0200. |P .. m (!Sd) bunsel elgsuei ‘snlnpow ruooeg ZZ 67 “"8830 23523 m:0t~> E 83:00 Eon—mom Pam «38> 50550 0:33. 2 .m B505 ANV EmE0mm Ugo: vN0 0N0 0.00 0.00 0.00 0.3V «.00 0.00 040 0.0m .QNN 1a q a _ a q a _ a — 4 fi « — . _ . _ 000020» “ 0000 020.? .000 020% - J 0 000— 0000 0000 000m. 0000 0000? 000m? 0000? 00NOP 0000? 0%!) mbuens ensuer 68 limitation Thesamplesthatdidfiaenneshowedafailurepattemcharacteristicofpure planar shear loading (Figure 3.16). Reproducible data for soil low modulus samples 10A and 103 could not be obtained and arenotreportedhere. Stress-strainresponseforalltheothersamplesareshownin Figure 3.17. A27. seeantmoduluswas calwlated and shown in Table 3.3. Good agreementwas foundbetween Iosipescu testing andtensiletesting. Figure 3.18 shows shearmodulusversushard segmentcontentsforallthe difi‘erenttypes ofurethanes. From the graph we see that for the same hard segment content, lower molecular weight polyol based urethanes have higher modulus. This is due to the higher cross-linking density for the same hard segment content in lower molecular weight trial-based systems. Figure 3.19 shows the efi‘ect of molecular weight per cross-link (M9 on shear modulus. As M, increases, cross-link density decreases (see Figure 3.3) and shear modulus increases. For the same M, lower molecular weight polyol systems exhibit higher modulus due to higher hard segment content. Also, in Figure 3.18, the slope of Tone 0310 based urethane system is higher than that of Tone 0305 and Tone 0301 systernswhichisconsistentwithtensiletestingand suggeststhathighermolecularweight polyol undergoes phase separation more readily. 3.1mm Adhesion samplesweretestedintheshearfixuueandthepeakloadvaluesrecordedare shown in Table 3.4. Samples with Tone 0310 and Tone 0305 showed very good reproducibility whereas Tone 0301 based samples had a larger amount of variation and Sumo], 18qu nosodrsol row aldunzs suaqrorn pormoerg 91's gunou 69 70 NR 0.0 0.0 0.0 0.0 0.0 Ease 8% 383m£ 5 85505 £20.05 .3 8:88,; emubmufihm S .n ”.5503 A5 28% 0? m0 OF \\ Um. \ OD \ w e 0.0 0N 0.0 N0? 0. : (15%) 888.118 71 0x. 383% 0550.5 32.5, =_ 88:00 205000 Eu: mamao> «2:002 32—0 “580 gm 3 .m ”550$ A5 EmE0m0 EDI 00 N0 00 v0 00 00 we 00 v0 00 1* 0 so 029 H .. .H a mono 0on U a 0 L g . 72 258% 2358: 300:5 5 35 0.05.805 .80 2303 0£000.02 320> 00—0002 30:0 2300 .\.N 2 .m 959..“ A030 xc__lmmo..0 En. E062, 00300.02 0000 00mm 00.00 008 000? 0009 0004 000 000 000 0 W A a u q u 4 — 1 0 ’00 wzoh 0000 020.? P000 020% (lsd) 8nlnpOl/V .10qu wages ZZ 73 TABLE 3.4 Adhesion of Various Polyurethanes to Glass Surface Shear Adhesion (PS3 706 :1: 4 1590 :k 30 2690 :h 60 4640 :t 80 5370 :1: 300 4160 :1: 110 4280 :1: 90 5240 :1: 220 5080 :1: 360 825 :1: 340 2120 :1: 530 3020 :1: 1360 74 especially with sample 115. Thus, a large number of 1B samples were tested to get reliable mean and standard deviation. Figure3.20 shows adhesionvaluestoglassversushard segrnentcontentforallthethree typesofurethane systems. Within eschfamilyofurethane systems, adhesionvalues increasewithincreasinghard segrnentcontent. Forthesarnehard segmentcontent (for example, 10C, 5C and 1C or 10D, 5D, and 1D) Tone 0305 based urethane shows better adhesion than Tone 0310, and Tone 0310 based urethane shows better adhesion than Tone 0301 based urethane. Rao et al (Rao and Dml, 1991, 1992) have shown that adhesion ofgraphite fibers to epoxy resin is dependent on the shear modulus of the matrix. A plot is made between the shear modulus of the various urethanes and their respective adhesion values to glass and is shown in Figure 3.21. From this figure we see that adhesion does increase with modulus in all the urethane systems and difl‘erent urethane systems show different depencies. Higher modulus urethanes based on Tone 0301 show poor adhesion to glass. From the least square fit lines, we see that same level of adhesion could be obtained fiom lower modulus sample based on Tone 0310 polyol. Tone 0310 polyol has the highest molecular weight among all the polyols studied and based on thermal and mechanical characterization of these urethanes, we have seen that higher molecular weight polyol enhances phase separation. This suggests that along with modulus and hard segment content, phase separation in a segmented urethane system can have significant efl‘ect on its adhesion to glass. 75 «8830 0.8523 msotu> 5 22:00 EoEmom 03$ namao> mum—O 0. 0000.030 00.0 EOE ANV 0300:0000 000: 00 00 00 00 00 00 00 00 0m 0N — a _ a _ q _ a _ . fl . _ a _ q _ 90 1 osmsmo z/Ovo I 00. 0 F000 020H 0 F00 MZOH .. . o U 0% .. “ mono 0on 000 000— 0009 000m 000m 0000 0000 0000 0000 ooom comm. (!Sd) 880:9 01 UO!Saupv 76 mp56>m 0550.5 2.00.5 00 3.0002 300m mamao> nae—O 2 0002:}. _N.m EOE Afiavm33002000£00c0000NN oooo_m oooome 0000m_ oooom_ oooon_ cocoa, oooom oooon oooom _ . _ . _ . _ . _ 0F00020H 000 F 000m 0000 0000 I 0000 (!Sd) 88015) or uorseupv 77 Afiaadhesiontesfing,glauwrfacesoftheumpleswereobsavedforfaflmemode. Someofthe sampleshadclumksofurethanelefiontheglusmrfacewhfleothersamples showed brittle fitilure and did not leave any visibly noticeable urethane. SEM micrographs ofTone 0310 based urethane samples are shown in Figure 3.22. As seen under optical microscope, SEMconfinnstheadhesivefiiluremodeinsamples 10A, 103, and 10Cand cohesivefailuremodeinsarnples 10Dsnd 10E. Thecircularpatternsinmicrographsof 100 and 10B could be due to the microvoids present in the sample. To further analyze failure modes, x-ray photoelectron spectroscopy was used for all the samples. A control sample ofglasswasalsoruntoobtainbaselinedata Availabilityofnitrogenonthe surface was then used as the indicator ofurethane presence. Surface atomic concentration ratio of nitrogen per 100 carbon atoms was calculated fiorn the narrow scans of the elemental regions for all the samples and are shown in Table 3.5. The control sample has atomic ratio of 0.84. All the urethane samples showed higher N/ 100C ratio than the control glass. This suggests that all the samples had some degree of cohesive failure in urethanes. Table 3.5 also shows theoretically calculated values of N/ 100C for all the urethane samples. These values are higher than the experimental values obtained fi'om the glasssurfaces.1-Iearn, eta1(I-1earnetal, 1988) havereportedthataircuredurethane surfaces in segmented polyurethanes tend to be richer in soft segments. This observation would help explain why surface concentration of nitrogen at the interface could be less than the bulk nitrogen concentration. 802005300 003m 2 8 2.8 0o 0:080 502.3. .22 ”80:6 3.0 0o 2020202 2% «an 050$ w 2 - 29:3,. 0 2 . anum < 2 . 20.5w 79 TABLE 3.5 Atomic Concentration Ratio on Failed Glass Surface from X-Rsy Photoelectron Spectroscopy Sarnjles NI 100C Theoretical Value 10A 2.8 6.8 103 3.5 9.0 ' 10C 5.9 10.4 10D 6.2 12.2 10E 9.0 13.3 SD 3.7 10.1 5C 6.2 11.4 I 5D 8.6 12.9 I SE 7.8 13.9 I 1C 4.2 13.0 1D 5.7 14.1 115 8.2 14.7 Glass 0.8 80 W Tlueedifi‘erMurethmesystamwereprepuedfiomcapmlsctonebaseduiohwith different molecular weights and the same chain extender and the same diisocyanate. Withineachurethane system, hard segmentcontentwasvariedbysdding difl'erent amountsofthechainextenderandtheisocyanate. Thistypeofexperimental design allowed us to study efi‘ects ofhard segment and triol molewlarweight on thermal, mechanical, and adhesion characteristics of cross-linked segmented polyurethanes. DSC results showed that the transition temperature related to hard segment increased with increasing amount of hard segment content. This could be due to the increased hard segment chain length at higher hard segment content, promoting phase separation in urethanes. Thus, in a urethane system, phase separation phenomena is favored by increasing hard segment content. Also, the rate of increase of thermal transition temperature in difl‘erent molecular weight triol-based urethanes indicated that phase separation phenomena was also favored by the higher triol molecular weight. Both tensile and Iosipescu testing of the urethane samples showed nonlinear stress-strain behavior. The tensile and shear modulus data indicated that even though with increasing hard segment content within a urethane system, cross-linking density decreases, its modulus increases. This could be due to several factors including increased amount of aromatic isocyanate content, increased hydrogen bonding, increased phase separation, and many other factors. The modulus of the same hard segment content urethanes was found to be higher for lower molecular weight triols. Thus, cross-linking density seems to be the determining factor for modulus in the constant hard segment content urethanes. In the 81 case ofhigher molecular weight triol-based urethanes, the ttte ofincrease ofmodulus with hard segment content is higher than that for the lower molecular weight triol-based polyurethanes. This indicates that phase separation is favored by larger polyol molecules. Adhesion to glass for these cross-linked polyurethanes seems to be a coupled phenomena controlled by several factors, including hard segment content, modulus, molecular weight of triols, cross-linking density, and phase separation. In urethanes based on same molecular weight triols, higher modulus and higher hard segment content proved to improve adhesion to glass. In urethanes made with difl‘erent molecular weight polyols, same level of adhesion was obtained with higher molecular weight polyol-based urethanes having lower modulus and lower hard segment content as compared to urethanes made with lower molecular weight polyols having higher modulus and higher hard segment content. This suggests that among many other factors, phase separation in cross-linked segmented urethanes can be a key factor in controlling adhesion to glass surfaces. CHAPTER4 W Theworkpresanedinfluschaptahasbeenmbnfiuedforpubficafioninthelmnnd of Adhesion Science and Technology (1994). W Polyurethanes were prepared fi'om difl'erent molecular weight polycaprolactone based polyols with varying amount of hard segment content made fiom toluene diisocyanate and butanediol. A theoretical model based on hard and soft segment miscibility was used to predict phase separation in these polyurethanes. Wide angle x-ray difl‘raction, near- infinred spectroscopy, and fourier transform infi’ared spectroscopy were used to experimentally determine phase separation in these polyurethanes. Good agreement was found between the experimental results and the theoretical predictions. The glass surfaces of the previously tested glass/urethane adhesion samples were analyzed using angular dependent x-ray photoelectron spectroscopy (ADXPS). ADXPS data revealed that an interphase region of approximately 20-100A thickness was present between the urethane matrix and the glass substrate in each sample. The data also showed that the composition ofthe interphase region was influenced not only by the matrix composition but also by the phase separationinthematrix Thecurvefitted C-lsspectraoftheinterphase region showed the presence of C-0 type linkages which could be due to the presence of C-OH from the polyols and/or the butanediols and C-O.Si type bonds. These observations were 82 83 supportedbytheadhesionresultsofthevariousurethanestoglass substratescoatedwith a 2% solution of butanediol in acetone. W Polyurethanesandespecially segmented polyurethanesarewidelyused aselastomersand as engineering plastics. Polyurethanes are available as thermoplastics as well as thermosets. Thermoplastic polyurethanes are typically processed by injection molding and thermoset (cross-linked) polyurethanes are typically processed by reaction injection molding. In both types of urethanes, chopped glass fibers are commonly used as fillers to improve urethanes mechanical propaties. Good adhesion between the glass reinforcement and the urethane matrix is desired to achieve the maximum benefits of the reinforcement. In other structural applications, good adhesion between urethane and glass surfaces may also be desired. Such applications include laminated windshields and modular windows. In modular windows for automobiles, for example, a glass panel is insert molded and encapsulated by a urethane gasket (Reilly and Sanok, 1988; Fielder and Carswell, 1990; Agrawal et al, 1991). Good adhesion between the glass panel and the urethane gasket is necessary to prevent water leaks and also to maintain the structural integrity of these glass modules. In our previous study (Agrawal and Drzal, 1994), we investigated the relationship between the urethane structure and properties to its adhesion to glass surfaces. We found that, along with modulus and cross-link density, phase separation in the urethane also is a factor in adhesion to glass surfaces. The surface composition of the failed glass surfaces 84 afier adhesion testing, as determined by the x-ray photoelectron spectroscopy (XPS), was quite difi‘erent than the stoichiometric compositions of the urethane matrices indicating that an interphase had formed. Severalreportshavebeenpublishedinrecentyearsdiscussingphase separationin polyurethanes. Numerous experimental techniques including fourier transform infiared spectroscopy (Y oon and Ratner, 1988; Carnargo et al, 1982), near-infrared spectroscopy (Miller et al, Part I and 11,1990), difi‘erential scanning calorimetry (Yoon andRatner, 1988; Carnberlin and Pascault, 1983; Byrne et al, 1992; Chem et al, 1992), dynamic mechanical analysis (Camargo et al, 1985; Estes and Cooper, 1970), transmission electron microscopy (Schneider et al, 197 5; Chang, 1984), x-ray photoelectron spectroscopy (Yoon and Ratner, 1988; Heam et al, land 2, 1988; Vargo et al, 1991), and many others havebeenusedbyresearcherstoanalyzephase separationinpolyurethanes. Thehardand sofi segments have generally been found to phase separate into discrete phases with a domain size of 10-20 nm (Heam et al, 1, 1988). Interfacial studies have shown that the surface of polyether based urethanes are enriched with the polyether when compared to the bulk (Yoon and Ratner, 1988; Heam et al, 1, 1988). Recently, Deng and Schreiber (Deng and Schreiber, 1991) have discussed orientation phenomena at polyurethane surfaces in contact with difl‘erent media. Yoon and Ratner (Yoon and Ratner, 1986) have related the urethane phase separation to its surface composition and found that where significant phase separation takes place, little or no hard segment could be found in the outermost few molecular layers of the polymers. 85 Tlussmdyistamsionofourpreviousworkandisdirectedatducidafingthe rdafionshipbetweenpolymethmestrucnnemdadhesiontotheglasswrface. Inthis nudy,wehaveuiedwconeluephasesepuafioninpolymethanestotheiradhedon behaviortoglasssurfaces. Theoreticaland ertperimerualtechniqueshavebeenusedto explore phase separation in model urethane compounds. A theoretical model based on phasemiscibilityhasbeenusedto predict phase separation and is coupledwithbynea- infi'ared and fourier transform infiared spectroscopy to experimentally determine the phase separation. Adhesion experiments have been performed to provide evidence regarding the role of phase separation in polyurethanes and its role in promoting adhesion to glass surfaces. X-ray photoelectron spectroscopy has been utilized to study the failed glass stu'facesafteradhesiontestingto quantifytheinterphase structure andtherebyunderstand the role of phase separation in urethanes and its role in adhesion. W W The polyurethanes used in this study were based on caprolactone polyols available from Union Carbide under the trade name “Tone." Hard segments were made fiom a 80°/e-20°/e mixture of toluene 2,4-diisocyanate and toluene 2,6—diisocyanate (TDI, Aldrich Chemical Co.) and 1,4-butanediol (BDO, Aldrich Chemical Co.) as the chain extender. The various urethane formulations studied in this work are shown in Table 4.1. For adhesion testing, annealed soda-lime float glass plaques were used. W x Vt" x Vt" blocks of urethanes were cast on the air side of the glass plaques. The details of urethane Urethane Formulations at Isocyanate Index of 1.0 86 TABLE 4.1 Diem“? T wimp; BDO “’1 sgét 32m gnatron 0:3: 0:3; 03:18 (Mole%) (Mole%) (wt %) Cr ass-Link (Me) 10A - - x 0 6o 22 1160 1013 - - x 22 S6 37 1424 100 - - x 31 54 47 1692 101) - - x 34 53 60 2222 1015 - - x 41 52 67 2752 513 - x - 7 ss 37 854 so - x - 20 56 47 1013 so - x - 31 54 59 1324 SE - x - 36 53 67 1646 1c x - - 0 6o 47 563 11) x - - 18 57 59 739 1r: x - - 26 55 67 912 as - - - so so 100 - 87 mixing and adhesion sample preparation can be found in our previous publication (Agrawal and Drzal, 1994). ”2!!" ! 121- DE . 0!”sz Widemglex-nydifi'aefionpattamofthevafiousurethaneumpleswererecordedona Rigaku RU2003 difi'actometer using a graphite monochromated CuKa radiation operatedatSOkVand lOOrnA Theurethanesamplesizewas6mmx6mmx2mmandthe reflected intensity was recorded as a firnction of 20 angle at a rate of 0.25°/minute. W A Lambda-9' spectrophotometer made by Perkin-Elmer was used in transmission mode for near-infi'ared analysis of the urethane samples. Urethanes were cast directly onto the air-side of soda-lime glass plaques (similar to the plaques used for adhesion testing) and werecuredfor24 hours@90°C. Thecasturethanefilrnthieknessonglasswas approximately Va". Although the total spectral range was 1100-2500 am, only the region 1600-2100 nm was used for phase separation analysis. A slit width of2 nm was used along with a 240 urn/minute scanning speed. The urethanes of this study were cross-linked urethanes. Thus, the conventional sample preparation methods of solution casting thin urethane films or salt plate methods could not be used. Also, liquid nitrogen grinding of urethanes for difi'use reflectance was not used due to the possible disruption of the urethane structure. To overcome this, urethane was 88 directlycuredontoasiliconewafer. Inthismethod,asmallsmountofurethanemi1mrre wasukenfiomflrembdngvidandwuquicklyuneuedonasificonwafa'usingacouon swab. Thesihconwaferwasflrenmbjectedtotheumearfingcydeutheglass—urethme adhesion samples (24 hours @ 90’C). Fourier transform infi'ared spectra were acquired on a Bio-rad FTS-40 spectrometer in transmission mode. Asiliconwaferwasusedinthebackground scan. All sarnpleswere scanned at room temperature using 4 cm" resolution and sixty-four scans were averaged for each sample. 12W Subsequent to adhesion testing, failed glass surfaces were analyzed using a Perkin-Elmer P1115400 x-ray photoelectron spectrometer. An approximate W x W square area was sectioned from the failed glass surface and was placed inside the XPS chamber. The angular dependent XPS spectra were obtained at a base pressure ofapproximately 10” Torr. The standard Mg Kc source was used for all sample analysis and was operated at 300W (15 kV, 20 mA). A continuously variable angle sample stage was used and was programmed for 15°, 45°, and 90° angles (photoelectron take-ofl‘ angle). A 2.0 mm diarnetercircleofthe samplewasanalyzedbythespectrometer. Datawascollectedinthe fixed analyzer transmission mode utilizing a position sensitive detector and a 180° hemispherical analyzer. Pass energies were set at 89.45 eV for the survey scans (0- 1000 eV) and at 35.75 eV for the narrow scans of the elemental regions. Data collection and manipulation was performed with an Apollo 3500 workstation running PHI ESCA 89 software. ThecurvefittingwascaniedoutusingamodifiedGauss-Newtonnonlinear leastsquaresopfimizafionprocedmethatispmofflieinsnumanalsofiware. TheC-ls bindingemrgyofthegraphiticpeakwassetto285.0erorcalibrationpurposes. WW Thehudmdwfisegmmtsinpolymethanemoleaducanphasesepuateintohudmd soft domainsinthebulk. One ofthe important drivingforces forthisphase separationis the difl‘erence in solubility between the hard and the soft blocks (Macosko, 1989). Several researchers (Macosko, 1989; Camberiin and Pascault, 1984; Rayn et al, 1988) have used the Flory-Huggins interaction parameter 1 to successfully estimate soft and hard segment miscibility in segmented polyurethanes. 4.1 1m ' lam + £- (68.55): ( ) Here xm is interaction parameter between hard and soft segments, 6,, and 6, are solubility parameters of hand and soft segments and V, is a reference volume. has) is an entropy term and can be neglected for polymer systems so that equation (4.1) becomes 4.2 Km ' % (53 " as): ( ) At T - 298K with RT in calories and the reference volume V,l taken as 100 cm’lmol (Carnberlin and Pascault, 1984), the equation (4.2) reduces to 4.3 1m " 'é' (511-591 ( ) 90 The soft segment in the urethanes of this study is based on the polycaprolactone polyol and its solubility parameter is 6 s - 9.1 (cal/cm’)” (Brandrup and Immergut, 1989). The hard segment is based on toluene-diisocyanate and butanediol and its structure is shown below: ‘—-$—J Hard Segment Unit Weight = 264 CH3 Since the solubility parameter of this hard segment could not be found in the literature 6 H was estimated fi'om the group contribution method (Van Krevelen, 1990) using cohesive energy data reported by various authors; and Small's approach. Very good agreement was found between the 6 a values obtained by the different approaches. An average of the various calculated 6H value was taken and was found to be 12.44 (cal/cm’)”. The details of 6H calculations are listed in Appendix A. 91 Thtgthehardandsofisegrnentmiscibilitycanbecharacterizedby x” = % (12.44-9.1)2 - 1.86 cal/cm’ (4,4) Theonset ofphase separation, foramixtureoftwopolymers, AandB, canbepredicted using the Flory-Huggins relationship (Macosko, 1989) to estimate the critical interaction parameter. .1 x" 2 1 1 * I (4.5) «an: whereNAandN. arethenumbersofrepeatunitsinpolymersAandB. In segmented polyurethanes, the above equation is not applicable, as polyurethanes are not blends of two polymers but are block copolymers. Benoit and Hadziioannou (Benoit and Hadziioannou, 1988) have developed phase diagrams as a function of repeat sequence in a multibloclr copolymers having various architectures. The phase diagrams provide xc(N,, + N.) forvarioushard segment contentswhere xcisthecr‘itical interactionparameterforthe onset of phase separation. Na and N, are numbers of monomer units in hard and sofi chain segments. These monomer units are not taken in the classical sense of polymer chemistry. They are subunits having the same volume. The molar volume of caprolactone is ~ 105 cm’lmol (calculated by group contribution method, Van Krevelen, 1990) and its molecular weight is 114. These values were 92 assignedtoarepeatunitandusedtocalculatethevaluesof N,andN..,. Listedin Table 4.2 aretheN. values forthevarious polyols and in Table 4.3 arethe Nfl values for thedifiaanhudwgmanlmgths.1hewaagelargthsofhudblocksuecdaduedfor stoichiometric formulations of the various urethanes and are shown in Table 4.4. Table 4.5 shows the calculated xcvalues based on the phase diagram developed by Benoit and Hadziioannou (Benoit and Hadziioannou, 1988). Phase separation will occur when 1391c (x3, - 1.86). The phase separation predictions for the various urethane formulations are also shown in Table 4.5. The relative degree of phase separation can also be judged by the difi'erence between the x", and xc values. W W WAXD curves for the various urethane samples were obtained. Due to the small sample size, the scans were noisy but adequate for qualitative assessment of hard segment crystallinity. The curves showed no sharp crystalline reflection and were characteristic of amorphous systems. This indicated the absence of three dimensional order in the hard segment domains of these samples. This finding is as expected due to the molecular asymmetry in TDI and also with the observations reported by other researchers (Legge et al, 1987). Irrfiued spectroscopy was used to evaluate phase separation in the samples. Hydrogen bonding has been considered as an essential part in the stabilization of hard-segment 93 TABLE 4.2 'N.’ Values for Various Polyols Sofi Segment Molecular Weight N, Tone 0301 300 ~ 3 Tone 0305 540 ~ 5 Tone 0310 900 ~ 8 TABLE 4.3 'Na' Values for Difi'erent Hard Segment Lengths Calculated Molar Volume H“ 5‘9““ cm’lMol (Ref. 25) N" 0 132.4 ~l O-O 347.6 ~3 O-O—O 562.8 ~5 O-O-O-O 778.0 ~7 Designation: Toluene diisocyanate - ‘0' Butanediol - '-' 94 TABLE 4.4 'Nfl + N,‘ Values for Various Urethane Forrmrlations F Urethane WWW Length Avg. Na N, fi'om Na + N, orrnulatrons Table 2 0 0-0 O-O-O O-O-O-O 10A 100 l 8 9 10B 33 66 2.3 8 ~10 10C 66 33 3.6 8 ~12 10D 28 72 4.4 8 ~12 1013 100 7 8 ~15 5B 86 14 1.3 S ~6 SC 46 54 2.1 5 ~7 5D 68 32 3.6 5 «9 5E 86 14 5.3 5 ~10 1C 100 l 3 4 1D 55 45 1.9 3 ~S 1E 100 3 3 6 Designation: Toluene diisocyanate - '0' Butanediol - '-' 95 TABLE 4.5 Calculation of Critical Interaciton Parameter xc Urethane Wt. % Hard Phase Separation Formulations Segment Xe (x3, >xc) 10A 22 5.6 No 1013 37 1.4 Yes 10C 47 0.97 Yes 10D 60 0.89 Yes 10B 67 0.76 Yes 5B 37 2.33 No SC 47 1.66 Yes 5D 59 1.19 Yes 5E 67 1.14 Yes 1C 47 2.9 No 1D 59 2.14 No 1E 67 1.9 No 96 sancturemdhubearukmhnoaccoumindifl‘auuhardsegmaumoddsmomnetd, 1974; Blackwell and Gardner, 1979). The extent of interurethane hydrogen bonding can beestirnstedbyinfiaredspectroscopy. Inurethanes,alargefi’actionoftheN-ngoupsof (-N-C-O-) the urethane linkage is hydrogen bonded. ; v. A recent review article (Miller, 1991) discusses use of near-inflated spectroscopy (NIR) in polymeranalysis. Also,Milleretal(Millertetal,IandII,1990)haveshownthe usefulness of NIR in phase separation studies in polyurethanes. Since soda-lime float glass istransparent inNIRregion and sample preparationwasveryeasyNIRwas selected for phase separation evaluation in this study. Figures 4.1, 4.2, and 4.3 show the NIR transmission spectra (in the region 1600-2100 nm) of the various urethane samples. This region of NIR spectra is dominated by carbonyl stretching bands (1915 nm), C-H stretching first overtone region (1600-1800 nm) and N-H combination band (around 2045 nm). Peak assignments are taken fiom the published literature (Miller et al, I and II, 1990; Miller, 1991). The absorption peaks in general are very broad and the carbonyl peak at 1915 nm is very weak. Thus, the N-H peak alone was used to study hydrogen bonding. Each sample shown in Figure (4.1, 4.2, and 4.3) has a NIR spectrum similar to the 100% hard segment (HS) composition which is independent of the polyol type and quantity. The N-I-I band in the HS sample appears at 2065 nm, which has a high degree of hydrogen bonding. In Figure 4.1, for urethanes based on the high molecular weight polyol Tone 0310, the N-H band for sample 10A appears at 2045 nm which can be assigned to 97 28:00 2.25% .5: 35> 55 8552a?— Bsm 2 8 Soto 2.8% 62 2. $59..— AECV £96653 GORP 00m: 00 Pm ooom com: com: OHQ m: (spun KJoJrquo) uogssgwsumi z 98 28:00 Eofiwom Bum miss 5? $552.53 83m 88 2.8. ea 88% 52 S. gem AECV £92925 ooom com: com: oonr com: 00 rm - —\ . (Sigun KJDquqJD) uorssgwsumi Z 2.2.80 .58me .5: 9.2.95 .23 ”055220.— 83 88 25 co 2.8% E2 m... 9.25 A85 £92325 00 Pm Doom com: com P GORP com F d u d 4 meow IIZ mI (spun KJDJrquD) uogssruusumi Z 100 unbondedN-H(Milleretal, II,1990). Weseethatforthehigherhard segrnentcontent samples 10B, 10C, 10D, and 10B, the NJ! bands incremently shift toward the HS sample bandat2065mnwithlargestshifiinthesample10E. Smallspectralshifiscanbedueto severalflfingsindudingaystaflhutyandhydmgmbondingudiswssedbyGuton (Gaston, 1992). Since the hard segments in these polyurethane samples are amorphous, thesespectral shifiscouldbeduetothehrcreasedhydmgenbondingwithincreasedhard segment content in these samples. This suggests that phase separation in the urethanes based on Tone 0310 increases with increasing hard segment content and samples 10D and 10E have some degree ofphase separation Sirnilartrendsbutto alesser degreecan also be seen in urethanes based on the medium molecular weight polyol Tone 03 05 (Figure 4.2). On the other hand, in the urethanes based on the lowest molecular weight polyol Tone 0301 (Figure 4.3), the NJ! band occurs around 2045 nm for all the samples with difl‘erent hard segment content (1C, 1D, and 1E) and no spectral shift is observed. This suggests that the urethane samples based on Tone 0301 are not phase separated. FTIR spectroscopy was used to verify the observations fi’om the NIR spectroscopy. Figure 4.4 is an FTIR spectra ofthe cured urethane sample 10B. The isocyanate asymmetric vibration band usually occurs between 2260-2280 cm“. In Figure 4.4, there is nofi’eeisocyanatepeakpresent. Thus,itcanbeconcludedthattheurethanereactionsin sample 10B must have been completed. The peak at ~333S cm" is assigned to hydrogen bonded amide hydrogen and the region around 1700 cm" is assigned to a carbonyl band (Yoon and Ratner, 1988; Camargo et al, 1982). These regions are analyzed here in detail for the various urethane formulations. 101 me 295m ao 538% SE 3. mecca ooo_ coca coon aces . o :rz w S O I. 0. D» u 0 u m. a ONO r u 102 Figure 4.5 shows the N-H band for Tone 0310 based urethane samples 10A, 10B, 10D, and 10Eintheincreasingorderofthehardsegmentcontent. Thepesklocationsare 3343 cm“ for sample 10A, 3337 cm" for sample 103, 3323 cm" for sample 10D, and 3323 cm"forsample10E. Thisspectral shifiisinagreementwiththeNIRdataand mggensthatthedegreeofphasesepuafionhtcrusuwhhinausinghudsegmmt content. Figure 4.6 shows the carbonyl region of the FTIR spectra for the samples 10A through 10E. The band at 1736 cm" is associated with C=O stretching absorption in a non- hydrogen bonded state while the lower wave number peak at 1712 cm" represents the absorption of C-0 hydrogen bonded with the N-H groups. The 1736 cm" peak is a result of the nonbonded carbonyl peaks of the urethane linkages in the samples and the ester linkages in the polyol portion of the samples. Thus we do not expect the 1736 cm" band to ever disappearbut the 1712 cm" peak should increase inintensitywith respect tothe 1736 cm" peak as phase separation increases. It is clear horn the graph that the 1712 cm" peak intensity increases with increasing hard segment content. This supports the earlier made observation fi'om the N-H peak regarding phase separation. Thus, the order of phase separation in the samples studied is 10910D>10C>10B>10A I E 2 I l C . . The phase separation discussed applies to the bulk polyurethane samples. The bulk phase separation afi‘ects mechanical properties such as modulus, but in our previous study (Agrawal and Drzal, I, 1994) we found that modulus alone could not explain the adhesion 103 coon 28:60 Eon—mow Ea: wfib~> .23 mocafioazbom 38m 2 no 626,—. E 23m 3-2.3 383m 5.: mé 950E comm ooem <0“ mean mmmm mmmm :1 aoueqaosqv 104 28:60 EoEwom as: 9.35, .23 85.238 82m 2 8 2.8. 5 $55 3898 £5 «.5 Beam Swain as £8 85 8i no 3.8% at 3. 950E :IEo. mama—55:28; 00%: Gem; coma p 1. 3 5v 1 » Ar 1 dh ls . 1 A g: T i : 4 3g + ‘ g L 1 fi ‘i U 107 ',."S‘32 1.3259! FIGURE 4.8 Curve Fit C ls Spectra of Failed Glass Samples at 90° Photoelectron Take- Ofl‘ Angle 108 chargecorrectedwiththeC-Cpeakreferencedto285.0eV. Acursorylookatthespectra reveakthatthaeuesevudbindinguuuofcubonpresanontheumplqbothu higherandatlowerbirndingenergies ofC-Cbonds(285.0 eV). Thelowerbindingenergy peakcanbeassociatedtocarbonattachedtopositivelycharged sodium(Na‘)andcalcium (Ca“)ionspresentinthesoda—limeglasssubstrate. Thehigherbindingenergycomponent (~286.5 eV) canbeduetotheenergyshificausedbyC—Osinglebond environment. This may include C-O-C ether type linkage, unreacted free hydroxyl: C-OH fiom polyols and chain extenders and C-O-Si type linkages with the glass surface. Since the polyols used in these urethanes are polyester type, the presence of the ether linkage is unlikely. The even higher binding ernergy component at ~289.5 eV can be due to the urethane linkage (Heam et al, 2, 1988; Vargo et al, 1991) present on the samples. The C-1s peak binding ernergies and the relative peak areas for the samples are shown in Table 4.6. As seen fiom the table, the lower binding energy componernt (~283.5 eV) is much smaller for the sample 10D than the other samples and is very high for the sample 10B. This might be due to the tlnickness variation of the residual film on the gass surfaces, with the film being tlnickest in the sample 10D and thinnest irn the sample 10B. The Na” and the Ca++ ions might have migated intotheresidualfilmandarewithinthesamplingdepths oftheXPS in 10B, 1D, and 5thilenotsointhecaseofthe sample 10D duetotlnegeaterresidualfilm thickness. The atomic percernt concentrations of 0, N, C, and Si obtained from the high resolution XPS spectra at difi‘erent photoelectron take-ofl‘ angles are shown in Table 4.7. It is clear fiom the data that as the sampling depth increases, the Si concentration increases with ClsPeskBindingEnergiesandRelativePeskAress 109 TABLE 4.6 of Carbon Chemical States on Various Failed Glass Samples PeakI PeakII PeakIII PeskIV Sm?“ 1315. Area B.E. Area 13.13. Area 13.13. Area (eV) ('4) (eV) 1%) (eV) ('4) (eV) ('4) 10D 283.71 3.24 284.96 76.94 286.38 13.33 288.86 6.50 5D 283.23 8.26 284.89 74.17 286.85 10.52 289.45 7.05 1D 283.86 23.92 285.06 62.67 287.15 9.15 289.49 4.27 1013 283.98 33.38 285.32 54.62 287.09 8.99 289.51 3.00 110 ands «ed 8.." 3.: 3.3 3.. 3.3 3.3 2.3 .a... 3.3. min. 2.6m «a... 3.3. m... 3.3 a... arrow 8.3 9.3 2.. 8.3 and. as...” 3... 3.2. and. 2.3. on... shun a. 2.2. 36 Exam 2.3 on.v~ 3.. we.“ 36. Sen 8... no.3. and. amen as. «.6... an 3.2. «c.» 8.3 3.3 3.2.. .n. .3... 3.2 3.9. vufi 3.2. $6. 3.3 .Nd 3.8 Dc. 0 z O .m U z 0 .m U z O .m U 2 0 8.95m §.3m ace one on. Ev Named. 8&5. .883 =§8_8§E 6896 a 239888 285.. 6.562 a“ a as .2 o .2 z .2 o lll beinghighestatthe90°anglesndisclosetothatofthebareglass. TheSiatomic concenuufimcandsobeusedwcompuetherdafiwflucbressesofthereddudfilmson theglasssm'faces. TheSiconcentrationinlODisthelowestfollowedbySDandlDin the15°anglescans. TheSiconcentrationinlDanleBarealmostthesame. Theme tendcanalsobeseenirrthe45°andthe90°XPSscans Thismggeststhattheresidual filmisthicker inthesarnple 10D, followedbythesamples SDand 1D. Theresidualfilms inthe samples 1D and 10B are almost ofthe sarnethickness. The C-ls concentrations in , the various samples also indicate the same residual film thickness trend. In Table 4.7, the stoichiometric atomic concentrations of the various urethane samples are also listed in the far right column. On comparing the XPS derived atomic concentrations with the stoichiometric values, we see that the carbon and nitrogen atomic concentrations are much less than that of the bulk matrix. Based on the atomic compositions of the residual films in all the samples and their variation with the sampling depths, we can state that the residual films on the glass substrates have a general thickness range lO-IOOA and have both inorganic and organic constituents. These residual films are an interphase region which has a composition intermediatetothatoftheglass mbstrateandtheurethanematrices. The urethane matrix composition directly afi‘ects the proposed interphase region between the matrix and the glass substrate. The XPS data also suggests that the phase separation in the matrix also afi‘ects the composition and thickness of the interphase region. The high 112 Oh atomic concentrationandthe greaterthickness of the interphase region inthe sample 10D could be the result of phase separation phenomena, causing the preferential segregation of the polyol and/orthelowmolewlar weight chain extenderbutanedioltothe surface. Whereasinthepoorlyphaseseparatedsample 1D, preferential segregation ofthe polyoland/orthechaineaenderdidntukeplacetomyappredablementremlfingina thinnerinterphaseregionnotsorichintheC-Otypelinkages. Thecurvefit C-lsspectraalsorevealsthattheamount ofthehighenergycomponent (~289.5 eV) carbon is in good agreement with the N-ls atomic concentration This suggeststhat the 286.5 eV component in the interphase region is likely dueto the presence of the butanediol rather than the polyols since the polyol should contribute to the ~289.5 eV peak. MW An adhesion experiment was designed to check the validity of the above made observations. The air side of clean soda-lime glass plaques were coated with a thin layer of a 2% (by weight) solution of 1,4 butanediol in acetone in order to facilitate phase separation in the interphase. Upon drying ofthe film, various urethane blocks were cast ontotheglass surface(detailed procedurecanbefoundinRef. Agrawal andDml, 1994). Upon curing, the samples were allowed to age for a week prior to the adhesion testing. The adhesion of the urethanes to the glass substrate was evaluated using a block-shear method (Agrawal and Drzal, l, 1994) and the results are shown in Table 4.8. Also 113 TABLE 4.8 Adhesion of Various Polyurethanes to Glass Surfaces (Psi) Adhesion to 233;?!“ (:11? Adhesion to Glass Substrate Samples M “’T ‘3 BareGlass Substrate Coatedwithar/o 9““ ““3 (Psi)(Ref. 4) BDO Solution ‘1‘“) (Psi) 10D 150,553 4640: so 4550.1: so so 155,404 5240:220 456M270 ID 184,962 2120: 530 4980* 350 10 176,060 325 s 340 4120 s 130 1013 1,4004 1590130 1510* so i'Tensile Modulus, sample was too soft for Iosipescu testing 114 showninTable4.8aretheadhesiondatafortheurethaneswithoutanycoatingonthe glass substrates. The data reveal that the adhesion values ofthe urethane samples 10D, 5D, and 103 are notsignificantlyafl‘ectedbyther/oBDOcoatingontheglasssurface. Ontheotherhand, the adhesion values ofthe samples 1D and 10 have significantly improved overthe previouslyreportedvaluesobtainedonbareglass. This experimentindicatesthatby coating the glass substrate with the 2% BDO solution, we have modified the interphase region in the samples 1D and 1C to be similar to the interphase regions in the other urethane samples. Interphase modification of the samples 10D and 5D with the BDO coating showed little efi‘ect on the adhesion in these samples because the adhesion values obtained with or without the 2% BDO coating are all close to the cohesive strength of the glass substrate. The adhesion data for the sample 103 remain unchanged because it is an elastomeric material with very low modulus as compared to the other urethane samples (Table 4.8). This suggests that both the interphase region and the matrix are contributing factors in adhesion of urethanes to bare glass and either one may be a limiting factor. W A theoretical approach based on miscibility of hard and soft segments was used to predict phase separation in the various urethane formulations consisting of caprolactone polyols of difi‘erent molecular weights and hard aegrnents made fiom toluene diisocyanate and 1,4 butanediol. The Flory-Huggins interaction parameter (1) between the hard and the soft segment was calculated from their solubility parameters and compared with the onset of 115 phase separation (1c) estimated fiom a phase diagram for the various stoichiometric hard and sofi blockchainlengths. Thecomparison oftheestirnated xc valuestothe Flory- Huggins xvalues suggestedthathigher molecular weight polyolspromotephase aepuafionandinaphueseparafingurethmsyamphasesepmfionmaeaseswhh increasingamomtofhardsegmentcontent. Wide angle x-ray difi‘action of the urethane samples confirmed the expected amorphous '- nature of the hard segments. Both near-inflated and fourier transform infrared spectroscopy were used to experimentally evaluate the phase separation in these samples. In both the infiared techniques, hydrogen bonding analysis was used as the basis for determining the phase separation. The spectral shift in NIR N-H band wavelength due to possible hydrogen bonding indicated phase separation in higher molecular weight polyols (Tone 0310 and 03 05) with hard segment content above 40% and was also supported by the spectral shift ofboth the N-H band and the carbonyl band in FTIR spectroscopy. Both theFl'IRandNIRdatawereingood agreementwiththetheoreticallypredicted phase separation results. Angle resolved x-ray photoelectron spectroscopic analysis was conducted on selected failed glass samples after adhesion testing. The XPS derived atomic concentrations indicated the presence of a urethane type polymeric layer with constituents of glass such as silicon present in it. The various sampling depths analyzed by the angle resolved XPS indicated gradients of atomic concentration present in all the samples. The Polymeric layer thickness was thinnest in the low molecular weight and thickest in the high molecular 116 weight samples. The C-ls and N-ls atomic concentrations in all the residual films were foundtobelowerthantheirrespectivestoichiometricvalues. Itwasproposedthatthe reddudfilmswaeanhaphaseregionwithcomposifiomhamediatetothuofthe respective matrices. Based on the variations in compositions and thicknesses of these interphaseregionsinthevariousurethanesamples studied, itwasalsoconcludedthat matrix composition and phase separation in the matrix afi‘ected the interphase region significantly. ThemrvefittedspectraoftheC—lsregionindicatedthattheseinterphase a regions might be rich in 00 type linkages which could be fiom C-O-Si and/or C-OH type species. To support the above mentioned hypotheses and create a 'beneficial" interphase, adhesion of the various urethanes to glass substrates coated with a 2% solution of butanediol was measured. The data indicated that modification of the interphase region of a poorly phase separated urethane sample dramatically improved the adhesion. 0n the other hand, in the well phase separated samples the interphase modification had little efi'ect on the overall adhesion values which were close to the cohesive strength of the glass substrate. This suggests that phase separation in urethane plays a large role in adhesion to glass substrates by afi‘ecting the composition, thickness, and properties of the interphase region. The introduction of phase separated material at the glass-urethane interphase in elastomeric systems had little efi‘ect on the adhesion. This indicated that along with optimum interphase properties, the modulus of the urethane matrix itself can be a limiting factor in adhesion to the glass substrates. CHAPTERS khl I a I 0 a 0 '1'. '3 _ '9. .‘Qr'au' ;. Imn; a ”I 0a W Theworkpresentedintlfischapterwillbesubnfitted forpublicationinthelournal of Adhesion (1995). W Surface fi'ee energies of polyurethanes made fi'om toluene diisocyanate and 1,4 butanediol based hard segments and caprolactone polyol based soft segments were theoretically calculated using additive functions such as molar parachors and cohesive energy densities. Good agreement was found between the theoretically calculated values and the experimentally determined values based on contact angle measurements. The phase separated polyurethanes were found to have higher polar surface fiee energy component (1'). This was linked to the preferential segregation of butanediol to the polyurethane surfaces due to phase separation. The adhesion values of these polyurethanes to soda-lime glass were correlated with their respective 1' values and a linear relationship was found. It was also shown that the adhesion values of the low .1" polyurethanes improved substantially when the glass surfaces were coated with a thin layer ofbutanediol prior to the bonding. The modulus of the interphase region rich in butanediol was evaluated and was determined not to be a significant factor in controlling adhesion of the butanediol coated glass mbstrates. The chemical interactions at the polyurethane/glass interphase were investigated by pretreating the glass surfaces with methyltrimethoxysilane and 117 118 trimethyl chlorosilane prior to adhesion testing. The adhesion data showed no significant difi‘erencebetweentheuncoatedandthesilanetreatedglass substrates. Basedonthis experimental evidence, the possibility ofany covalent or ionic bonding at the polyurethane/glass interphase was assumed negligible. It was determined that the mechanism ofadhesionbetweenthepolyurethanesmdtheglassmrfacecouldbethrough the formation of an interphase region in which hydrogen bonding between the butanediol rich interphase region and the hydroxylated glass surface plays a key role. W Glass/polyurethane adhesion has become increasingly important in the automotive and other industries in a variety of applications including laminated windshields, reaction injection molded modular windows for automobiles, long and short glass fiber reinforced thermoplastic and thermoset composites, etc. Also the use of polyurethane coatings on the inside glass surface of windshields is being investigated to impart antilacerative property to the windshield to protect occupants in the event of a collision. In all of these applications, good adhesion between the glass and the polyurethane is imperative. Adhesionbetweentwo dissimilarmrfacessuchasglassandpolyurethaneisregardedasa complex phenomenon influenced by many factors including physical and chemical interactions. The reversible physical interactions are caused by Van der Waals forces which may also include hydrogen bonding. The irreversible chemical interactions may include ionic and covalent bond formation across the interface between the two materials. Several researchers have studied polyurethane surfaces to determine the surface 119 compositionsandmrfaceproperties. Vargoetsl(Vargoetal, l991)havefoundthatthe polymethmemrfaceisennchedmthelowmolewluwdghtpolyetbucomponem. Sengupuetd(Sengupuetal,l991)havedetaminedthepohrandd1edispmive componentsofthe surface-fies energies ofpolyurethanesandhavetriedtorelatetheseto thesurficesoftsegmentcontents. The glass surface is also very complex reflecting its composition and history of environmental exposure. Fawkes et al (Fawkes et al, 1990) have used the acid-base interaction approach to study adhesion of glass to various polymers. The chemical composition ofthe surface ofglass has also been studied by several authors using x-ray photoelectron spectroscopy and other surface sensitive techniques (Pantano, 1981; Vaughan and Peek, 1974; Lassas et al, 1993; Markus and Priel, 1981; Mohai et al, 1990). Theresearchindicatesthatglasssurfaces, exposedtoambientatmosphere, sreenrichedin sodium ions relative to the bulk. The glass surfaces are usually found to be hydrated due to the adsorption of water vapors. It has also been suggested that NaHCO3 type species may be present on the surface due to the adsorption of carbon monoxide and dioxide species fi'om the atmosphere. All of these studies have addressed either the surface composition oftheurethanesortheglasssurfacesbuthave donelittleto correlatethe actual chemical and physical nature oftheinterfaceto adhesion ofthese polyurethanesto the glass surfaces. In our previous studies on glass/polyurethane adhesion (Agrawal and Drzal, I and II, 1994) we found that polyurethane to glass adhesion is greatly influenced by the modulus 120 ofthepolymerintheirrterphase. Wealsoformdthatalongwithmodulus,phase separationinpolyurethanealsoinfluencesitsadhesiontoglass. Aninterphaseregionwas fmmdbaweenthepolyurethanemaubtmdtheglmmbstrateofomprevimnuudies whichhadacomposifionhuanwdiatetothatofflremauixmdtheglassmrface. The composifionmdflreflficknessofflfishaphueregionwufoundtoberdatedtothe phase separation in the matrix. In the present study, we have investigated the physico-chernical interactions at the polyurethane/ glass interface. Surface fiee energies of the various urethane formulations were evaluated using theoretical and experimental techniques, and the dependence ofthe surface fi'ee energies on the surface composition and/or the phase separation has also been studied. The thermodynamic work of adhesion to the glass surface has been evaluated for the various phase-mixed and phase-separated polyurethanes and compared with the expaimental adhesion data reported earlier (Agrawal and Drzal, I, 1994). The role of chemical interactions at the polyurethane/glass interface has been explored by coating the glass surfaces with an alkyl silane to make the surface chemically "inert" prior to the adhesion testing. W W The polyurethanes used in this study were based on caprolactone polyols available fiom Union Carbide under the trade name ”Tone.“ Hard segments were made fi'om a 80%-20% mixture of toluene 2,4-diisocyanate and toluene 2,6—diisocyanate (TDI, Aldrich Chemical 121 Co.) and 1,4-butanediol (BDO, Aldrich Chemical Co.) as the chain extender. The various urethaneformulations studied inthiswork and previous are shown in Table 5.1. For adhesion testing, annealed soda-lime floatglassplaqueswereused. '/4"x'/4"x%" blocksofurethaneswerecastonthesirsideoftheglassplaques. Thedetailsofurethane mixing and adhesion sample preparation can be found in our previous study (Agrawal and Drzal, I, 1994). W The various urethane formulations were cast on clean 75 x 50 mm glass slides. The samples were then heated for 24 hours at 90°C in a convection oven. The final polyurethane coating thickness on the glass slides was about 1-2 mm. The glass slides were handled carefully throughout the sample preparation to avoid any surface contamination of the polyurethane surfaces. The surface fi'ee energy of the samples thus prepared were obtained from the experimental determination of contact angles for sessile drops using a Rams-Hart Model 100 goniometer. All experiments were conducted at room temperature (~23°C). The data wereobtained forwettingbyaseries offluidswithvarying surfacetensionswhichare shown in Table 5.2. The liquids used covered wide range of properties from being non- polar such as “obromo-naphthalene to a highly polar liquid such as water. By doing so, a comprehensive wettability profile was developed for each urethane formulation. Liquids 122 TABLE 5.1 Urethane Formulations at Isocyanate Index of 1.0 Sample 1’01”le BDO m1 33;; $315131: Damon Tone Tone Tone (MOW/0) NOVA) (wt %) Cross-Lank 0301 0305 0310 (Me) 10A - - ' x 0 50 22 1150 1013 - - x 22 55 37 1424 100 - - x 31 54 47 1592 10D - - x 34 53 50 2222 10B - - x 41 52 57 2752 513 - x - 7 ss 37 354 so - x - 20 55 47 1013 51) - x - 31 54 59 1324 SE - x - 35 53 57 1545 1C x - - 0 50 47 553 1D x - - 18 57 59 739 IE x - - 25 55 57 912 HS - - - 50 50 100 - 123 TABLE 5.2 Surface Free Energies of Liquids Used for Contact Angle Measurements (Van Krevelen, 1990; Table 8.5) SW Free Energy (We!!!) Liquid 7" 7’ 1 Water 21.8 51.0 72.8 Glycerol 37.0 26.4 63.4 Formamide 39.5 18.7 58.2 Methylene Iodide 48.5 2.3 50.8 “-bromonapthalene 44.6 0.0 44.6 124 withvaylawmrfacetarsionsmchun—dkaneswaenotusedintlfissmdyduetothe dificulty and errors associated with measuring small (<10°) contact angles. W Themrfacetensiansofthevariouscaprolactone—basedpolyolscouldnotbefoundinthe WandflmwaemeaarredusingaCahnDCA-nZdynamiccomaamglemflyzer. Afieshlyflunedglasssfidewuusedforthemeamremuu.1hesfidesadvandngand recedingspeedwassetat22u/mimrte. Thesurfacetensianafwaterwasdeterminedusing this technique and was found to be accurate (72.8 dynes/cm) and reproducible. 531D 'lll'lll' Rectangular bars (30 mmx4 mmx 1.5 mm) ofurethane samples were used for dynamic mechanical analysis on a Polymer Laboratories MK III DMTA system. Elastic storage modulus (E') at various temperatures were obtained in a single cantilever bending oscillation mode of deformation at 1 Hz fixed frequency. The temperature was varied fiom 10°C to 50°C at 5°C/min. W A 2.0 weight percent solution of methyltrimethoxysilane, available from Dow Corning under the trade name Z6070, was prepared in reagent grade methanol. Approximately 10 weight percent (of the amount of 26070) deionized water was added to the solution and the solution was aged for a week prior to use. Another 2.0 weight percent solution of trimethyl chlorasilane, available from Aldrich Chemical Co., was prepared in reagent grade 125 tetrahydrafuran. Cottonswabswereusedtoapplyanevencoatingofthesesolutionsto theairsideafthesada-limeglassplaques. Thecoatedglassplaqueswereallawedtodry for30mimrtesandthenwereusedtopreparetu'ethaneadhesionsunples. MW Subsequenttoadhesiantesting, failedglass surfaceswereanalyzedusingaPerkin-Elmer P1115400 x-ray photoelectron spectrometer. Approximately V4" x V4" square area was sectioned fi'om the failed glass surface and was placed inside the XPS chamber. The angulardependent XPS spectrawereobtained at abase pressure afapproximately 10" Torr. The standard Mg Kx source was used for all samples analysis and was operated at 300W (15 kV, 20 mA). A continuously variable angle sample stage was used and was programmed for 15°, 45°, and 90° angles (photoelectron take-ofl‘ angle). The portion of the sarnpleanalyzedbythe spectrometerissettluaughanirfitial lenssystemandwasset for a 2.0 mm diameter circle. Data was collected in the fixed analyzer transmission mode utilizing a position sensitive detector and a 180° hemispherical analyzer. Pass energies were set at 89.45 eV for the survey scans (0-1000 eV) and at 35.75 eV for the narrow scans of the elemental regions. Data collection and manipulation was performed with an Apollo 3500 workstation running PHI ESCA software. The curve fitting was carried out using a modified Gauss-Newton nonlinear least squares optimintion procedure that is part oftheinstrumental sofiware. TheC-lsbinding energy ofthe graphitic peakwas set to 285.0 eV for calibration purposes. 126 4Is|01=fl a, 0 1P__'sl0 0F 11.5 31.3511 The surface fiee energy of the polyurethanes of this study can be estimated theoretically. The polyurethanes are composed of sofl and hard segments. By knowing the fiee energies of these segments, the overall surface free energies of the polyurethanes can be evaluated by using an approach described by Eberhardt (Eberhardt, 1966). structure is shown below: The hard segment in the urethanes is based on toluene diisocyanate and butanediol and its' 131—H El—E—O‘fCI-lg) H—: H; $0 L(_?__)_‘l Hard Segment Unit Weight - 264 —o—n 55 if" H The surface free energy y" of this hard segment can be estimated by an additive function, Z the molar parachors Ps as proposed by Sugden (Sugden, 1924) and discussed by Van Krevelen (Van Krevelen, 1990). The relationship between the surface fi'ee energy and the molar parachors is as follows: _ ( 3P3]‘ (5.1) 7H ' ‘57 where V is the molar volume contribution of the groups in the glassy (amorphous) state of the polymer. The molar volume contributions can be evaluated fiom the Van der Waals volume (Vw) as follows: 127 v--1.5vw (5.2) ThveandP.valuesforfl1ehardsegmemgroupsaretakenfiamtherefamce Van Krevelen, 1990 (Tables 4.2 and 8.1). The calculated 2V and DP. values are 210.22 cm’lmol and 556.7 (cm’lmol) (erg/cm’)“ respectively. The details of these calculations areshowninAppendixB. Thus, 5.3) 555 7 ‘ ( 8 ' 8 49.2 dynes/cm 7” ( 210.22) The surface tension of the hard segment can also be estimated using an empirical relationship (Van Krevelen, 1990).. (5.4) 711 ~ 0.75 e3 where 1,, is expressed in dynes/cm and the cohesive energy density e... in MJ/m’. The cahesiveenergydensityafthehard segrnentcanalsobeestimatedusinggroup contribution approach and is calculated to be 519.2 MJ/m3 (Appendix B). Thus, 7,, s 0.75 (519.2)”3 - 48.5 dynes/cm. (5.5) Both the 1,, values estimated fiam the parachors and the cohesive energy density contributions are very close to each other and bath values were used here to determine the overall polyurethane surface fi’ee energies. The free energy of the salt segment 1, is dificult to estimate fi'om the group contributions of the molar parachors or the cohesive energy density as the exact structure of the repeating unit is not known. Thus, the 1, values were experimentally determined using dynamic contact angle measurements. The advancing and receding surface tensions were determined for the various polyols and little difl‘erence was found between the values for 128 thedifi‘erentpalyols. Ahysteresisaf-Gdyneslcmwasobservedbetweentheadvancing mdflwncedingmrficetarsiom.Theadvmdnganfacetu1sianvdueisusedhaeuh signifiesthefirstcantactbetweentheglassslidesandthepolyals. Theaveragenvalue wasfaundtabe38.l dynea/crnwithastandarddevistionaf0.4dynes/cm. Themrfacefieeenagyofthewfisegmwtandwbeesfinntedindirecflybywfluafing' cohesive energy density fiom its solubility parameter. The solubility parameter of the polyols is b, - 18.6 (I/cm’)m (Brandrup and Immergut, 1989). vs z 0.75 kw)” (5.5) Ys ” 0-75 (532)” (5-7) vs 2 0.75 (18.6)"3 = 37.0 dynes/cm (5.8) This value is close to the experimentally determined value of38.1. The 1. =- 38.1 is used here for the overall calculation of the surface free energy of the polyurethanes. According to Eberhardt's approach, the surface fiee energy of polyurethanes can be expressed as: 711 = Ns'Ys + NH‘YH (5-9) whereyuisthesurfacefi’eeenergyafthepolyurethanes andtheN,andNn arethemale fiactions of the soft and the hard segments on the surface of the polyurethanes. Assuming that the surface composition is same as the bulk composition, the N, and the Na values can be calculated fi'om the various urethane's stoichiometric formulations and are listed in Table 5.3. Based on the N8, NH, 7,, and 7, values, the n, for the various polyurethanes 129 TABLE 5.3 Calculated Surface Free Energies of the Polyurethanes Mole Fraction Surface Free Energy (dynes/cm) Na Na 70 ~70 3""9'“ 85a Segment Hard Segment (Molar Parachor) (“mum ‘5’) 10A 0.401 0.599 44.7 44.3 1013 0.222 0.778 45.7 45.2 10c 0.154 0.846 47.5 45.9 101) 0.132 0.868 47.7 47.1 1013 0.059 0.931 48.4 47.7 1c 0.398 0.502 44.8 44.3 11) 0.250 0.740 45.3 45.8 1E 0.194 0.806 47 45.4 as 0.0 1.0 49.2 48.5 130 werecalculatedandareshowninTable 5.3. Thecalculated yuvaluesbasedonmolar parachorsandcahesiveenergydenaityareveryclose. Itshauldbenotedthatasthehard segmancamwtinausuinthepolymethmeaflreyuvduesflsoinauseandisthe highestfor100%hard segrnentcontent sample HS. W E E 1 El . l I . The surface fiee energies of the various model polyurethanes were determined from the contact angle measurements of several liquids with difi’erent surface tensions and chemical firnctionalities. Zisrnan (Zisman, 1964) plots of Case versus 1 were developed for all the urethane samples. Figures 5.1a and 5.1b show the sample Zisman plots for polyurethane sample 1c with 46.7 wt. % hard segment and sample HS with 100 wt. % hard segment. The extrapolated critical surface tension yc at Case = l for all the samples are shown in Table 5.4. The 7c signifies the empirical value of the maximum surface tension of liquids able to spread on the given surface. The yc data reveal that the critical surface tension of the various polyurethanes studied are close to each other, considering the experimental errors invalvedinthemeasurernents. Also, no cleartrendinthedatawithincreasinghard segment content can be observed, as can be inferred fiam the theoretically calculated surfacefree energies. Along with the critical surface fiee energy, polar and dispersive components of the surface fies energies of these polyurethanes were also evaluated from the contact angle measurements. According to Schultz et al (Schultz et al, 1977), the surface fiee energy 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Cos (0) 0.0 0 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Cos (9) 0.0 0 I 1 a 1 L 1 4 1 10 20 30 40 50 60 70 80 90 100 Surfoc e Tension (dynes/c m) (a) ' I ' I I 1 n l 4 l a l a l a 1 n 1 . ] 10 20 30 40 50 60 70 80 90 100 Surfoc e Tension (dynes/c m) 0)) FIGURE 5.1 Zisman Plots of (a) 1C Polyurethane Sample with 46.7 wt. % Hard Segment and (b) HS Polyurethane Sample with 100% Hard Segment 132 TABLE 5.4 Observed Surface Free Energies of the Polyurethanes Based on Contact Angle Measurements Surface Free Energy (dynes/cm) Samples 75 7’ 7" Yu' (7"+ 7") 10A 44.4 6.83 33.64 40.47 10B 44.0 6.50 34.20 40.70 10C 44.6 10.55 30.58 41.13 10D 44.3 13.55 30.49 44.04 10E 42.2 14.97 30.25 45.22 1C 41.83 7.72 34.22 41.94 1D 42.46 10.72 32.49 43.21 1E 42.76 11.11 32.49 43.60 HS 43.64 8.73 35.76 44.49 Glass - 37.59 19.18 56.77 133 canberepresentedbythesumoftwo components, namelyadispersion (1°)andapolar consonant (7'). 7 = 7" + 7" (5.10) Byusingthisandtheexpressians derived farinterfacialfi’eeenergybyFowkes (Fawkes, 1957), Owens and Wendt (Owens and Wendt, 1959) and Kaelble and Uy (Kaelble and Uy, l970),theuna'fadalfieeu1agybetwearthepolymethmesmdthefiquidscanbe expressed as: run = Yu + 71 - 2 (70" 71")” - 2 (75" up)“ (5.11) Combining this with the Young's equation yo = 7m + yLCose, we get (5.12) D K a P8 1 + an. a g 205 v?) + 2(‘10 Yr.) (5,13, 71. Yr. 7 (1 +0230) 1' “ or 1' x = (73) + -—f)— (72)” (5.14) 2(YL Yr. ' " y (1 +Cas0) The .Y—f). and L values for the various polyurethanes Yr. 2(be were plotted and (705m and HUD)"z values were obtained fiom the slope and the intercept of the linear regression fit lines through the experimental data points. One such plot for Tone 0310 based polyurethanes is shown in Figure 5.2. From the graph, we see 134 95:80 Eon—mom Ea... EobEQ 5.3 85523.9. woman o. no 25... xGSN w» . 8.. .0... q .m> II N m 950.... 889+.v». a» . It w T... w. v. N. 0.. wd 0.0 4.0 Nd 0.0 2...._....1.I._.Z._..I...272...... m 'TYVVYITWYI I 1 "TijI‘YTV'I fifi I I b F by r * h P P - 1P P — b p b P — F n F F — h F bl P h F - hr hr — h F P b 1? P P L {2031.30 82:. o ..\. u mkm A to thm [are 0. : hm. 135 flututhehudugrnentcomaumaeasesintheseumplesthdrdopesdwinaease. Thishdicuuthnwifliuthehudsegrnentcommmflwpohrcomponanofthe methanesurfacefieeenergyalsoincreases. They'andynvaluesabtainedfiomthese graphsareshowninTable 5.4. ThelastcohrmninTable 5.4 also shawstheaverall urrfacefieeeriergyofthesepolyurethanes(yu)aduchisdrenrmofthey’andynvalues. Figure 5.3 showsthegraphofthemeasured yuvaluesversusthehard segmentcontentin. the Tone 0310 based polyurethanes. Also shown are the graphs of the theoretically predicted 7‘, values for these polyurethanes. Both the theoretical and experimental graphs show the same trend that the surface tension increases with the increasing hard segment content. The theoretically predicted values are in very good agreement with the experimental values. They are higher than the experimental values by only 3-4 WM and the difl‘erence is less for the higher hard segment content samples (10D, 105, etc.) which are phase separated polyurethane systems (Agrawal and Drzal, I and II, 1994). This small amount of difl‘erence could be due to the assumption made in the theoretical prediction regarding the surface composition being the same as the bulk composition. The experimental 7" curve shows an increase in the value after the sample 10C and then maintains the high 7., values for the samples 10D, 10B, and HS. This sharp rise in 70 values could be due to the phase separation in these samples leading to the surface enrichment in high surface energy species. This is further explored by plotting the polar component of the surface fiee energy for these samples. 136 2:350 «casewom .5... «scuba 3 8559539. Beam 0. no one... 8.. 82.5 02m on... seesaw 8. 8.332 e5 cc 3528.: 3 950E A5365 .cmEoo .cmEomm v.6: oo— om cm on 00 om 0.4 0m. ON A . 1 . _ . _ 4 _ . . r _ . . Om, . S n u: 1 mm D 3 8 00. mo. <0. . H Cut—«5002 1 0.? “W m: P 10 .1... a .J or m 1. - e. ) xoeoco ozmoco av 69.809: ,m. fi>fimcm0 I; \l U >1 r 909: x] 0 0902309.. 8 e. e - on V . a mm 137 Figure 5.4 shawsthe y'versushard segrnentcontent for Tone 0301 andTone 0310based palymethuws.AlsoshownisthedflapohuHSforthe100°/ehudsegmaucament polym'ethane(containsonlyBDOandTD1). Thedatashowsthatthey'valuesincrease withincreasinghardsegmentcontentinboththe 1 and 10seriespolyuretbanesamples. They'valuesforthe 1 seriesareingenerallowerthanthatforthe 10 seriesforthesame hardsegrnentcontents. A sharp increaseinthey’valueofsample 10Cisalsoserved and the trend continues with the samples 10D and 10B. This observation can be explained by noting that with increasing hard segment content, phase separation increases in higher molecular weight polyol (Tone 0310) based polyurethanes (Agrawal and Drzal, I and II, 1994). The phase separation in the lower molecular weight polyol (Tone 0301) based polyurethanes is not as significant and thus the 7’ values for these polyurethanes are lower than 10 series polyurethanes. The 1’ value for the 100% hard segment content polyurethane is 8.73 dynes/cm which is lower than the 1’ values for the samples 10C, 10B, and 10B; all of which have lower hard segment content. The higher 7’ values for 10C, 10D, and 10E suggests that the surfaces ofthese polyurethanes are not rich in the hard segment. Further, the higher 7’ values can be explained by the phase separation in these polyurethanes. In our previous studies (Agrawal and Drzal, 1 and II, 1994), we have discussed phase separation in these polyurethanes and have shown that due to phase separation, the surface composition in these polyurethanes tend to be rich in hydroxyl containing species and deficient in nitrogen containing groups when compared with the stoichiometric composition. The two hydroxyl containing species in these polyurethanes are the polyols and the chain extender (BDO). Due to the lower molecular weight of the chain extender, BDO is more likely to come to the surface (Agrawal and Drzal, II, 1994). 138 25:50 Eon—mom Bum .56....n. .u 8:§$a§.o.. esota> .o .caconEoU 3.3m ace... coatsm 8.0.. v6. 950.... .5506: .coEoo .coEomm 0.6: 00. om om on om on 0.4 on om - — q d a u n d 1 1— ‘1 — u — q u m... .omo mco. 0 .mo 9.8. Lost-No N. .0. o. w. ON (LU o/seu/(p)d,( 139 Thisisalsosupportedbythey’vahresobservedhere. Thelawery’valuesamples 10A, 1C,andlOBhavenoorverysmallBDOcantentandalsopoorphaseseparafianandtlurs theirsurfacesarenotrichinBD0.0ntheother'hand,inthesamples10Dand10E,BDO wgregatumthemrfaceuflrenflumhighay’values.1hemrfacefieea1agyvalues forBDOare(BrandrupandImmergut, 1989): y'-14.6dynes/cm y”=29.6dynes/cm 78 44.2dyneslcm The 1' of polyalisexpectedtobelessthan14.6 duetohigher hydrocarbon content and this firrther supports the conclusion that the surfaces ofthese polyurethanes are rich in BDO. The adhesion values of these polyurethanes to soda-lime glass surface were determined previously (Agrawal and Drzal, I, 1994) and are shown in Table 5.5. We had shown that the polyurethane to the glass adhesion improved with phase separation in the matrix and alsawiththemodulus afthematrix Aplat oftheadhesionvaluesasafirnction ofy'is showninFigure 5.5. Alinearrelationship canbe seenbetweenthe y'andtheadhesion values. The above observations suggest that in the samples exhibiting good adhesion to the glass, thehhaphaungionbetweenthepolyurethmemafiixmdtheglassmrfaceflgmwd and Drzal, II, 1994) consists of a larger concentration of higher polar fiee energy components than the bulk and the components are butanediol type species. 140 TABLE 5.5 Thermodynamic Work of Adhesion Between the Polyurethanes and the Glass Surface Shear Adhesion si SW” Wu- (dy'm’m) (Agrawal and During 1994) 10A 83 706 :1: 4 10B 83 1590 :1: 30 10C 88 2690 :l: 60 10D 92 4640 :1: 80 10E 96 5370 :l: 300 1C 85 825 t 340 1D 90 2120 :1: 530 IE 91 3020 :1: 1360 HS 89 - 141 0.80.25 00050.5...9. 2.0.53 ._0.. ancm 00.... eschew 05 .0 20:09:00 00.0.. 08:5 0.00.6 0. 00.02.34 m.n 950.... m. AEo\m0c.€v an : o. m m . 1 . 4 . a . 4 . 88505300 mmcmm . 85505300 motmm o. ooo . 000m 000... 000.0 0000 0000 000m 000m (gsd) ssolg) 01110159qu 142 The polar forces on the surface arise fiom the orientation of permanent electric dipoles and the induction effect of permanent dipoles on polarizable moleculu. In the case of . ‘N'C'O' «4 -8+8 polyurethanes,thesemoleculeswould 1nclude -8 I II and g 0-(6394-0 at?!” +0 H 0 species and especially the 1?;wa 9:352, due to their preferential segregation to the surface. These dipoles can create hydrogen bonds (permanent dipole-dipole interactions), other dipole-dipole bonds (excluding hydrogen) and dipole-induced dipole interactions with the hydroxyl rich glass surfaces (Kinloch, 1987). To firrther explore the role of the hydrogen bonding and other polar interactions between the polyurethane surface and the glass surface, an additional adhesion experiment was conducted. In this experiment, selected polyurethanes with varying surface 1' values (and thus with varying surface BDO content) were bonded to the bare soda-lime glass plaques and to glass plaques coated with 2% (weight) layer of BDO fiom acetone. The samples were tested for adhesion values in a shear mode (details of adhesion testing are discussed in Agrawal and Drzal, I, 1994) and the data is shown in Table 5.6 The data reveals that the adhesion values of the polyurethanes with law 7’ values can be significantly improved bycoatingtheglass surfacewithBDO, orin otherwords, bymakingtheirsurfacerichin higher 7’ component BDO. The failed glass surfaces after the adhesion testing were analyzed using XPS. Figure 5.6 shows the curve fitted C Is spectra of sample 1D and Samfle 1D with the glass surface coated with BDO, taken at 15° photoelectron take-ofi' angle. These spectra have been charge corrected with the C-C peak referenced to 285.0 eV. A cursory look at the spectra 143 TABLE 5.6 Adhesion Values (Psi) of Various Polyurethanes to Bare Glass Surface and to1,4 Butanediol Coated Glass Surface RelativeBDO 2% Secant Shear Consultation Modulus (Psi) arthe Surface Adhesionta l, 4 IosipescuTcsting 1' (Arbitrary Amundsen-re Bmdiot Coated Samples Ref 9 (dynealcm) Units) Glass Sta-face (Psi) Glass Staface (Psi) 10D 150,653 13.6 -H-t- 4640 :t: 80 4550 :t: 80 ID 184,962 10.7 -H- 2120 :1: 530 4980 :1: 850 1C 176,060 7.7 + 825 :1: 340 4120 :1: 130 10B 1400‘ 6.5 1590 :l: 30 1510 :1: 50 ’Tensile Modulus, sample was too soft for Iosipescu testing 10 10 144 A A l-_ I l j I Y Y I I I 0) FIGURE 5.6 Curve Fitted C 1s Spectra of Failed Glass Surfaces Taken at 15° Photoelectron Take-Off Angle (a) Sample ID (b) Sample ID with BDO Coated Glass 145 revulsthatfliereanseva'dbhrdingmuofarbonpresartonfltesunplexbaflru higherandatlowerbindingenergiesofC—Cbonds(285.0eV). TheClspeakbinding energiesandtherelativepeakareasareshowninTableS.7. Thepeakat~289.5chan bemociatedwithmetlnnefinkage(AgrswalmdDruLfl,l994)mditspresence indicatescohesivemodeoffailureinboththesarnples. Thehigherurethanepeakareain flieBDOcouedglaunufacemdicuesmethmenchaaufacethmthenglmnuface without BDO coating. Thepeak~286.5 chanbeduetotheunreactedfieehydroxylsC— OH fiom the polyol and BDO. The significantly higher area fiaction forthis peak in BDO coated adhesion sample as compared to the uncoated adhesion sample, suggests the presence of partially or completely unreacted BDO on the surface. Thus we can conclude that some BDO molecules fiom the coating react with isocyanates to form urethane linkages and integrate in the matrix while others remain unreacted. It is conceivable that the unreacted BDO molecules present in the interphase region of BDO coated glass adhesion samples can lead to localized change in the mechanical properties of the interphase region. The modulus of the modified interphase region could influence the adhesion of the matrix in addition to the hydrogen bonding mentioned earlier. To understand the mechanical properties of the butanediol rich interphase region, polyurethane rectangular bars were prepared with 5% and 15% (by total weight) excess of 1,4 butanediol. The excess BDO was incorporated into the urethane composition. Excess BDO was added to a homogeneous stoichiometric mixture of polyol, BDO, and isocyanate. It is expected that the urethane produced would closely represent the interphase formation process in the experimental samples used in this study. The samples 146 TABLE 5.7 C lsPeakBindingEnergiesandRelativePeakAreas of Carbon Chemical States on Failed Glass Samples Peak 1 Peak 11 Peak III Peak IV Sample . B.E. Area B.E. Area B.E. Area B.E. Area (0V) (%) (eV) (“/0 (eV) M) (6") (70) 1D 283.76 23.90 285.13 62.81 286.98 9.11 289.46 4.17 1D 283.80 5.07 284.91 61.57 286.3 23.87 289.24 9.48 with BDO Coating on Glass 147 with 15%excessbutanedialdidnatcurewell, andthesampleswereeithertacky and ”putty-like“ or very brittle with poor tensile properties. These samples could not be tested for mechanical properties. The sarrrples with 5% excess butanediol were tested for elastic storage modulus using Dynamic Mechanical Analysis and the data is shown in Table 5.8. Also shown in Table 5.8 is the elastic storage modulus for the corresponding polyurethane with stoichiometric formulation. The data indicates that the elastic storage modulus of polyurethane 1D with 5% BDO is about 10%-12% higher than for the stoichiometric composition 1D. Since the interphase formation in BDO coated glass plaques is expected to be simulated by this method, we can conclude that the localized interphase modulus in BDO coated 1D adhesion sample is higher than that of in the 1D adhesion sample without any coating. Based on the observations fiom our previous study (Agrawal and Drzal, I, 1994), the higher interphase modulus should result in higher adhesion values which is in agreement with the adhesion data obtained with BDO coated glass plaques. This observation suggests that the preferential segregation of BDO type species to the interphase region in polyurethane/glass samples influences its adhesion not only through increased polar interactions and hydrogen bonding but also by increasing the modulus of the interphase region. The thermodynamic work of adhesion due to the polar and the dispersive interactions was also calculated for the various polyurethanes/glass systems using Kaelble's expression. Kaelble (Kaelble, 1971) has modified the expression originally developed by Good (Good, 148 TABLE 5.8 Elastic Storage Modulus of Various Polyurethanes at Difi‘erent Temperatures Elastic Storage Modulus (Psi) Temperature (°C) 1D 1D with 5% Excess BDO 10 390,000 ' 438,000 15 381,000 438,000 20 381,000 428,000 25 381,000 418,000 30 373,000 418,000 35 364,000 399,000 40 356,000 390,000 45 348,000 373,000 50 348,000 356,000 149 1967)andFowkes(Fawkes,1967)andhasrepresenteditasthesumoftheworkdueto the dispersion and the polar components. W“, = Via“...D + Wu,” (5.15) Using the geometric mean relation to predict the interactions, wadh = 2 “Yo” Yo 1n + ('YrrP 7013172] (5-15) The W. values for the polyurethane/glass systems are shown in Table 5.5. A quantitative comparison between the calculated work of adhesion and the measured adhesion values in termsoftheload requiredtoinducebondfailureisnotrelevantbut, aqualitative comparison of the trends in the data can provide valuable information The data shows thattheW‘increases slightlywithincreasinghard segmentcontentbutnottotheextent observed in the experimental adhesion data. The shear adhesion data increases from 706 psi for Sample 10A, with 22 weight percent hard segment content to 5370 psi for ssample 10D, with 67 weight percent hard segment content. The lack of agreement between the calculated W... and the experimental adhesion data may be explained by the recent observations made by Fawkes et al (Fawkes et al, 1990) that the geometric mean relation to predict polar interactions may not include the efl‘ects of hydrogen bonding at the interface which might be an important factor in the polyurethane/glass systems. Another reason for this discrepancy can be the fact that the W...| calculation does not take into consideration the modulus of the polyurethane matrix. In the adhesion experiment, the matrix modulus afl‘ects the energy less going into the deformation of the matrix prior to the band failure. 150 5 E 2 Q] . l I . Glasssurfacesareknowntoberichinisolated, vicinal, andgerninal silanolgroups. In additiantothephysical interactionsbetweenthesurfacesilanolsandtheurethanematrirg there could be various chemical interactions. To investigate the chemical interactions such ucovdahmdionicbondhgflwglaunnfaceswaeprefiutedwithtwodifi‘mdlme coupling agents. A 2% (by weight) solution of prehydrolyzed methyltrimethoxysilane and a 2% (by weight) solution of trimethyl chlorosilane in tetrahydrofuran were used to pretreat glass surfaces prior to the adhesion testing. The purpose of the silane treatment was to make the otherwise hydroxylated glass surface chemically inert towards any subsequent covalent bond formation with polyurethanes. Another firnction of the silane coating was to provide a barrier layer between the glass and the urethane matrix to avoid any possibility of ionic bond formation between the Na’, K‘, and Ca++ ions and the urethane matrix. On the other hand, the silane coatings could also influence wettability of the various polyurethanes by providing lower energy surfaces. However, Plueddernann (Plueddemann, 1982) had shown that little correlation existed between the surface energy of silanes and their efl‘ectiveness as coupling agents between glass and polymeric matrices. The polyurethanes of this study showed adequate wetting to the silane treated glass surfaces. Shown in Figure 5.7a and 5.7b are the idealized monolayers of the condensed methyltrimethoxysilane and trimethyl chlorosilane on glass surfaces respectively. The methyltrimethoxysilane treated glass surface was analyzed using XPS. The atomic concentrations determined fiom the XPS analysis did not show the presence of Na+ ions which indicated that the glass surface was covered completely with the silane layer. 151 if? $“3 $33 is HO—Si-O—ii—O—S'\QO ‘O—SiH—O—ii—O /ci5//// H3 H3 $113 CH3—?1—CH3 CH3—§l-CH3 CH3—?1—CH3 }/AQ/// FIGURE 5.7 Idealized Monolayer of Condensed (a) Methyltrimethoxysilane (b) Trimethylchlorosilane on Glass Surface 152 Figure5.8 showsthecurvefittedC-lsspectrurnofthesilanetreatedglasssurfaceat45o photoelectron take-ofi‘ angle. The curve fit spectra reveals that the major portion of the carbon is Si-CH, type carbon (75.74%) followed by a C-O type linkage (23.23%). The presenceofthisethafinkageindicatesthatthaenughtbesomeunhydmlyzed methoxy groups (Si-O-CH,) present in the silane layer. This type of mahoxy group will not be present on the trimethyl chlorosilane treated glass surfaces. The adhesion data fiom these glass surfaces will not be influenced by the methoxy groups presence as might be the case in the methyltrimethoxysilane treated glass surfaces. Blocks ofseveral polyurethanes were cast on the silane treated glass surfaces and were cured in the same manner as reported previously. These samples were tested for adhesion values and the data is shown in Table 5.9. _ A comparison of this data on glass surfaces completely blocked by the silane with adhesion data for the bare glass surface (Agrawal and Drzal, I, 1994) reveals that the methyltrimethoxysilane coating on the glass surface had little influence on the adhesion values within experimental error. Similar adhesion data is obtained from the trimethyl chlorosilane treated glass surfaces. Both these data suggest that the adhesion mechanism of the polyurethanes to the glass surface is probably not due to covalent or ionic bonding in the interphase region. The failed glass surface of the sample 10D fi'om the methyltrimethoxysilane adhesion test was also analyzed using the XPS to determine the locus of failure. Table 5.10 shows the atomic concentrations of N,O,C and Si at three difl‘erent photoelectron take-off angles 15°, 45°, and 90°. The presence of nitrogen reveals that nitrogen containing species such 153 o_m=< e093 assessoi .2 a custom ”30 use: o§__..§o§eE_§o§o 5.58am 2 0 BE 2:6 an 950$ >o .>cmuzu uzmozmn e.~m~ a.vw~ o.m«~ o.w«~ o.om~ a.~m~ cu- “ P n T!- \1\1 ole mmlnmu ~0- up -- uh- uh- clu- un- ‘ di- un- «1- ea GOUS'JQS'3/(3)N 154 TABLE 5.9 Adhesive Values (Psi) of Various Polyurethanes to SilaneTreated Glass Surfaces Adhesion to Bare Adhesion to Aldhesron to Samples Glass Surface Methyltrimethoxysilane '3'": “M (1m 9) Treated Glass Surface “mm” “m“ ' Glass Surface 10D 4640a: 80 49401 390 4290*310 SD 52403: 220 4790 :1: 450 4740 :1: 560 1D 2120 :l: 530 2480 :l: 440 2380 :l: 550 103 1590 :l: 30 1470 :l: 170 1370 :k 140 155 TABLE 5.10 013,ng C Is, and Si 2p AtonficPacentConcentrationsatDifl‘erentPhotoelectron Take-OE Angles for the Adhesion Failed 10D Glass Sample Coated with Methyltrimethoxysilane Photoelectron Take-OE Angle am“ 15° 45° 90° 0 36.44 40.19 44.30 N 0.93 1.62 2.17 C 45.40 40.01 34.69 Si 17.23 18.19 18.84 156 asTDI and/orurethanelinkageswerepresent onthesurface. Figure 5.9showsthecurve fitted C—ls spectra at 15°, 45°, and 90° photoelectron take-ofl' angles. The peak at ~289.5 thrdicatesthatniuogmisconfingfi’omthepolyurethanefinhgesntherflrm 60111 the TDI. The nitrogen concentration gradient increases with the sampling depth Thisindicatesthatinterdifiirsion might havetakenplacebetweenthe silanelayerandthe urethanematrixresultinginaninterphaseregion. Theloeusoffailureseemstobe between the silane layer and the urethane matrix, through the interphase region. This interphase region observation is consistent with the previously made observations regarding interphase formation in the polyurethane/glass systems (Agrawal and Drzal, II, 1994). E 5 3 ! ll . I l l . Based on our previous studies regarding; the structure-property relationships in polyurethanes and their efi'ects on adhesion; the phase separation in polyurethane and the formation of an interphase region between the polyurethanes and the glass surface (Agrawal and Drzal, I and II, 1994); and the observations made in this study regarding the role of physico-chemical interactions in the interphase region, it can be inferred that the reversible physical interactions and not covalent chemical interactions in the interphase region play a key role in determining adhesion of the polyurethanes of this study to the soda-lime glass surfaces. The permanent dipole—dipole interactions and especially the hydrogen bonding between the urethane surface and glass surface appear to be very important for the overall adhesion. Phase separation in the matrix tends to cause a preferential segregation of BDO type species in the interphase region. Based on the linear 10 ‘ v A LALLLAAAAA LA A v I v vv v v v IKE M. sacs. Sh! u 1r 1L 4 r 1» L 1» J 1. o 1 1» 2 ‘ 1h 1 0 ‘ 1b 1’ o O f 232 10 r (P 1» 1» ’ 4b 1} 45° 1} . 1) 4} 1} 1r 7 1» 11 a 1i ‘1 . ‘ 1» V g 1» 1} . 5 1b it ( 1b 1» A bit ‘ 1a T 1» 1h 1» 1r 1? 1» 2 4» 1) I0 1L 1 1» 1y 4? 1 0 ¢ 232 _L A J A L A u a W T a 1 v j f . *l g 1} db 1’ 90° “ . 1} 1|- 0 45 J» 7 1 a 4? db . ‘ 1b 1» 3 1p 1) s 1» 1» g 1» 1» g 4 4> 1b 1h 11 1 41- 1) f» 1r 2 1» up J} 1p ‘ 1p 1 z" \ l A. _ A A ; . 4 v T V7 fl an an ass-.0 an ass FIGURE 5.9 Curve Fitted C 1s Spectra of Adhesion Failed 10D Glass Samples Precoated with Methyltrimethoxysilane Taken at 15°, 45°, and 90° Photoelectron Take-OE Angles 158 rdafionshipbetwemthepolumrfacefieemgyfifimdflremauhradhedommdthe findingsofmeadhedonarpaimaruvfithBDOcoatedglassphquthanfimherbe mfaredthatthepresmceofatcusBDOindrehuaphaseregioninflumcesmaubr adhedonbyinausingthemodrflusofthehnaphueregionmdbymaeasingthepolu interactionswiththeglasssurface. To explore the roles of the increased interphase modulus and the increased polar interactions (hydrogen bonding) on adhesion, another experiment was conducted. In this experiment, the air sides of soda-lime glass plaques were coated with 2% (by weight) solution of trimethyl chlorosilane in tetrahydrofirran solvent. Alter 30 minutes of air drying, some of the glass plaques were rinsed with tetrahydrot‘uran to remove any unbonded trimethyl chlorosilane fiom the glass surfaces. The glass plaques thus prepared were overcoated with a 2% (by weight) solution of BDO in acetone and were allowed to dry at room temperature. The glass plaques were then evaluated for adhesion with 1D polyurethane and the data is shown in Table 5.11. The purpose of coating glass plaques with trimethyl chlorosilane was to obtain an inert surface with which the possibility of hydrogen bonding with the subsequent coating of BDO can be minimized. In the case where the trimethyl chlorosilane treated glass plaques were further rinsed with tetrahydrofuran, the possibility of any potential polar interactions between the unbonded trimethyl chlorosilane and the BDO was eliminated. In this fashion, the efi‘ect of increased interphase modulus due to excess BDO on the overall adhesion of the matrix can be studied. A comparison of the adhesion data between the 159 TABLE 5.11 Adhesion Values (Psi) of 1D Polyurethane to Various Treated Glass Surfaces Glass Treatments Adhesion (Psi) Bare Glass 2120 :1: 530 Glass Coated with BDO 4980 :1: 850 Glass Coated with Trimethyl Chlorosilane 2380 :1: 550 Glass Coated with Trimethyl Chlorosilane andR' 1 28701530 Glass Coated with Trimethyl Chlorosilane and Overcoated with BDO 1570*“ Glass Coated With Tnmethyl Chlorosilane 1780 :1: 600 and Rinsed and Overcoated with BDO 160 BDO coated glass plaques and the trimethyl chlorosilane/BDO coated glass plaques revealsthsttheadhesionvsluewentdownfi'om4980psito1570psiwhentheBDOlayer wasdevoidofpotentialhydrogenbondingswiththeglassmrface. Similarremltswere obuinedfiomtheglmphquesuwedwithuimethylchlomdlmefimedmdwercoated withBDO.Thisexpahnannrggeusthathydmga1bondingbetweentheexcess bumnediolmthehuuphasemgionmdtheglassmrfaceisthepmminaumechufimof adhesion for phase separated polyurethanes. The suggested hydrogen bonding between BDO and the glass surface was firrther investigated using XPS. The air side of a soda-lime glass plaque was coated with a 2% (by weight) solution of BDO and was subjected to the polyurethane curing cycle (90°C for 24 hours). Afierwards, the glass plaque was rinsed several times with acetone to remove any unadsorbed BDO hour the glass surface. A control sample was also subjected to the same procedure except for the BDO coating step. The curve fitted Cls spectra ofthe control sample and the BDO coated sample are shown in Figures 5.10a and 5.10b respectively. The bands at ~286.5 eV, associated with C-OH type species, indicate higher hydroxyl content (9.83%) on BDO treated sample as compared to the control sample (6.63%). This suggests that BDO can be adsorbed on the glass surfaces. A schematic of the interphase region is shown in Figure 5.11 with some possible hydrogen bonding mechanisms. It is apparent that the majority of the hydrogen bonding is between the butanediol species which segregate to the polyurethane surfaces and the hydroxyl groups present on the glass surface. The hydrogen bonding in the interphase region can "(E NE. SN. $1109 "(D/£51112 $1109 NwAurostsoa Os-o Nwmwmwasm'é' Os-o 161 A l I Y I l I fiT I A l V T Y 1 T 1b 1 .4 4» J 232 2% 288 288 284 282 l A A I V l l A l l A 1 1 1 A A A 1 I s s I I v 00 FIGURE 5.10 Curve Fitted C 1s Spectra of (a) Control Glass Sample (b) Glass Treated with BDO and Rinsed 162 POLYURETHANE X X X X [XX X X X N:O:/:3\0/"‘{m\ POLYOL 3:950 (”Ct/13130111)rr~1(:\(P FIGURE 5.11 Schematic of Polyurethane/Glass Interphase Region with Probable Hydrogen Bonding Mechanisms 163 takeplaceinmsmerouspossiblewsys. Oneendofthebutanediolscanbehydrogen -y- bondedtothe 1': or E. linkagesoftheurethanegroupsandtheotherendcanbe hydrogenbondedwiththeglasssurfacehydroxyls. Inanotherscenario, partiallyreacted butanediolswillhavefi'eehydroxyl groupstohydrogenbondwiththeglasssurface. The polyester fianctionalities of the polyols and also the unreacted fi’ee hydroxyl ends of the polyols can potentially hydrogen bond with the glass surface. Yet another mechanism can be the direct hydrogen bonding of the urethane groups to the glass surface either through the -3- or through the j- linkages of the urethane groups. The preferential segregation of BDO at the interphase region observed in this study is not inconsistent with the previously reported studies. Heam et a1 (Heam et al, I, 1983) and Vargo et al (Vargo et al, 1991) have reported the enrichment of air/polyurethane interface in low molecular weight polyether based polyol components. The glass/polyurethane interphase of this study is expected to be different than the air/polyurethane interface studied by the above mentioned researchers. The high surface energy (56.77 dynes/cm) and especially the polar nature of the glass surface (7' = 37.59 dynes/cm) will have a significant influence on the interphase composition. This reasoning is further supported by the studies reported by Deng and Schreiber (Deng and Schreiber, 1991) discussing orientation phenomena at polyurethane surfaces in contact with difl‘erent media W In our previous studies (Agrawal and Drzal, I and II, 1994) we concluded that phase separation in polyurethanes significantly afl‘ects its adhesion to glass surfaces. It was 164 daunfimdthnflwmauixphueseparafionmflumcedthecompodfionandthethickneu ofthehuaphueregionbyuudngaprefamfiflwgregafionofhighpolumrficeenagy componenttotheinterphaseregion Inthepresentstudy,theactualmechanismof adhesionbetweenthe interphase constituents sndtheglasssurfacewasexplored. The contribution of chunical bonding in the form of covalent and ionic bonding on the overall glass/polyurethane adhesion was found to be not important. It was concluded that physical interactions are the most important factors in controlling glass/polyurethane adhesion. The polar surface free energy of the various polyurethanes correlated well with the XPS results regarding the BDO enrichment ofthe interphase regions in the phase separated polyurethanes. The work ofadhesion calculated fi’om the surface free energy components of the polyurethanes and the glass surface was found to be a poor predictor of the actual adhesion behavior. However, a linear relationship between the polar surface flee energy and the observed adhesion values emphasized the role of polar interactions on the adhesion. The modulus of the interphase region was found to be higher than that of the matrix due to the preferential segregation of butanediol at the interphase. This increase in modulus was determined not to be the factor for the superior adhesion performance demonstrated by the phase separated polyurethanes. It was concluded that the most important mechanism of adhesion between the polyurethanes and the glass surface of this study is through the formation of an interphase region in which hydrogen bonding between the butanediol rich interphase region and the hydroxylated glass surface plays a key role. CHAPTER6 QLCQHCLHSIQHS Inthiswork, sdhesionmechanisrnsofsegmented polyurethanesto soda-limefloatglass wereinvestigated. Thefocusoftheresearchwasarotmdtheglasslpolyurethane interphase region with emphasis on its structure/morphology, physical, mechanical, and chemical characterization to understand the underlying adhesion mechanisms. The model polyurethanes used in this study have allowed for the evaluation of phase separation, wide range of mechanical properties, difl‘erent structure/morphology and varying surface energetics on their adhesion to glass surfaces, while keeping the urethane chemistry constant. This approach has simplified the experimentation and the interpretation of the results. A model based on hard and soft segment miscibility was developed to predict phase separation in these polyurethanes. The model utilizes urethane chemistry/composition, polyol molecular weight and solubility parameter of the hard and soft segments to predict phase separation The model predictions agreed well with the phase separation data obtained fi'om the spectroscopic (NIR, FTIR), thermal (DSC, DMA), and mechanical (Iosipescu shear test, tensile test) characterizations. It was found that phase separation in polyurethanes significantly afl‘ects adhesion to glass surfaces. Polyurethanes with higher modulus showed better adhesion, but among the 165 166 polymethanesvfithflwumemodrfluaphasesepuuedumplesslnwedbettaadhedon thanthephasemixedsarnples. SEMsndXPS analysesofthefi'actured surfacesfromtheglass/polyurethanessdhesion testsrevealedtheformation ofaninterphaseregion, 20A-100Ainthickness, anda composition intermediatetothatofthematrixandtheglass. Itwasdeterminedthat matrix phase separation had significant influence on the composition, the thickness, and the mechanical properties of the interphase region. It was further determined that phase separation in the matrix along with the highly polar nature of the glass surface could cause preferential segregation of high polar surface energy component to the interphase region. The modulus of the BDO rich interphase was found to be slightly higher than the corresponding stoichiometric composition polyurethane. However, this increased modulus interphase was found not to be contributing to the superior adhesion levels of the phase-separated polyurethanes. Through the use of unreactive silanes such as methyltrimethoxy silane and trimethyl chlorosilane on the glass surface, the role of covalent and ionic bonding between the polyurethane and the glass surface was investigated and was also found to be not important for the adhesion. It was concluded that polar interactions are the most important factors in controlling glass/polyurethane adhesion. The polar surface fi'ee energy of the various polyurethanes correlated well with the XPS determined results regarding the BDO enrichment of the interphase regions. The work of 167 sdhedonuladuedfiomflremrfacefieeenagycomponansofflwpolymethamanddw glasswfacewasfoundtobeapoorpredictoroftheacmaladhesionbehavior. However, afinurrdafionshipbaweenthepolumrfacefieearergyandtbeobsavedadhedon values firrther emphasized the importance of polar interactions for adhesion AnsdhesionatpahnunnithBDOcoatedglassplaquesdanonsuatedthnadhedon of polyurethanes can be significantly improved by creating a "Beneficial" interphase rich in high polar surface fiee energy components. It was concluded that the most important mechanism ofadhesionbetweenthepolyurethanesandtheglasssurfaceofthis studyis through the formation of an interphase region in which hydrogen bonding between the butanediol rich interphase region and the hydroxylated glass surface plays a key role. 9W While the main objectives of this project have been met, still more work is needed to completely understand the adhesion mechanisms of polyurethanes in an actual reaction injection molding process. This work did not examine catalysts, which may play a very important role in the overall adhesion scheme by controlling reaction and phase separation kinetics. The experimental work of this project can be extended to include other glass substrates such as glass fibers and glass mats. The majority of commercial glass fibers have some sort of sizing on them. The efi‘ects of sizing on the glass/polyurethane interphase region and on the overall adhesion add more complexity to the system and should be explored. 168 Onlythethermosetpolyurethanematrixwasstudiedinthiswork. Thecxperimentaland theoretical work can be extended to include thermoplastic polyurethanes which generally have better phase separation than thermoset polyurethanes. Also, some of the polyurethane constituents can be tagged to study their preferential segregation to the surfaceortotheinterphaseregion Finally,thedurabilityoftheglass-polyurethanebond wasnotaddressedinthisstudy. Thedurabilityofthesebondsinhotandlmmid environmenuhuspedficdyuficanceespedaflymflreunomofivemdumywhaeflrese fins-polyurethane bondsmust survive overaperiod ofat least 10 years. APPENDICES Bunn, (Van Krevelen, (Van Krevelen, 6109112129. 1229.) -CH, Phenyl (trisubstituted) HO l I .N-c-o- -CH,- Fedor, Molar Attraction E...“ (J/mol) E... (Jlmol) Constant, F Molor Volume Van Krevelen, (cm’lmol) (Van Krevelen, (Van Krevelen, According to the group contribution approach, 2:5“, " a” ' 13V 169 12%)) 12%) 1229.1 7120 4710 420 33.5 16340 31940 1377 33.4 36620 26370 1483 18.5 2850 4940 280 16.1 zau- 108100 as“: 109150 213- 5883 2v= 168.3 K 5 a: 5% H V 170 a,“ - 25 .34(J/cm,)* - 1239(Ca1/an ’)" 65M - 25.40(J/cm3)“ I 12.45(Cal/cnr 3)“ According to Small's approach (26), 2F 53,. ‘ %‘ ,whereM-molecularweight p - specific gravity For TDI/BDO based hard segment, M - 264, p - 1.1451 (based on stoichiometric calulation) Thus, 53?. 8 25.52(J/cm3)" :- 12.47(CaI/cm s)1s Thus, 5 + 6 + 5 01]“ = [ H... :15» Hr.) g 1244(CaI/cm 3)“ APPENDIX B 1115.1 91 3:111:11. Nusl'a" Del 9619 .‘iumali lol P, V, (cm’lmol) Em (J/mol) Sugden, Van da Waals Volume, Fedor, 9121191 110. W W122i?) W122!» -CH, 1 56.1 13 67 4710 Phenyl 1 155.8 40.80 31940 (trisubstituted) '3 f 2 94.4 18.0 26370 arc-o- -CH,- 4 39.0 10.23 4940 SP, - 556.7 2V“, = 131.39 22E... = 109150 2v = 1.62v,, = 210.22 cm’lmol According to the molar parachor approach, 2 PS ‘ 7” BV 5567 ‘ - ° .. 49.2 dynes/ 7” ( 210.22) a" 171 172 Accordingmmanpificalrdafionshipbetweenfltecohesiveenagydmsitymdthe surfacefi’eemergy, y, a 0.75 “M“ 01’ as ” ~0.75 —— 33 . ( ,MV) ., 109150 210.22 as y” a 0.75 ( ) - 48.5 dynes/an BIBLIOGRAPHY BIBLIOGRAPHY Agrawal, RIC, Agrawal, A., and Thomas, 1., “Developments in Modular Windows for Automotive Applications”, SAE Int. Cong. and Exp, Detroit, MI, Paper 910759, (1991). Agrswal, RIC and Drzal, L.T., “Adhesion Mechanisms of Polyurethanes to Glass Surfaces 1. Structure - Property Relationship in Polyurethanes and Their Efl‘ects on Adhesion to Glass”, accepted for publication in J. Adhesion, (1994). Agrawal, RK. and Drzal, L.T., ”Adhesion Mechanisms of Polyurethanes to Glass Surfaces II. Phase Separation in Polyurethanes and Its Efects on Adhesion to Glass”, submitted for publication in the J. Adhesion Science and Technology, (1994). Agrawal, RIC, Fox, K.L., and Lynarn, N.R, “Characterization of In-Mold Coatings on RIM Encapsulated Modular Windows", SAE Int. Cong. and Exp., Detroit, MI, (1991). Benoit, H. and Hadziioannou, G., “Scattering Theory and Properties of Block Copolymers with Various Architectures in the Homogeneous Bulk State”, Macromolecules, 21, 1449- 1464, (1988). Blackwell, I. and Gardner, K.H., “Structure of the Hard Segments in Polyurethane Elastomers", Polymer, 20, 13, (1979). Bonart, R, Morbitzer, L. and Muller, E.H., 'X-ray Inestigations Concerning the Physical Structure of Crosslinking in Urethane Elastomers. 111. Common Structure Principles for Extensions with Aliphatic Diamines and Diols", J. Macromol. Sci, Phys, B9, 447, (1974). Born, L., and Hespe, H., ”Prediction of Polyurethane Hard Segment Structures Based on Diphenylmethane-Urethane Model Compounds", Colloid and Polmer Sci, 263, 335, (1985). Brandrup, J. and Immergut, E.H., “Polymer Handbook", John Wiley & Sons, (1989). Broutrnan, L.J., 'Measurement of the Fiber-Polymer Matrix Interfacial Strength", Interfacaes in Composites, ASTM Special Technical Publication 452, American Society for Testing and Materials (1969). Byme, CA, Mead, J.L., and Desper, C.R, “Structure-Property Relationships of Aliphatic Polyurethane Elastomers Prepared fi'om CHDI", Advances in Urethane Science and Tech, . Vol. 11, 68-109, Technomic Publication, Lancaster, PA, (1992). 173 174 Camargo, RB., Macosko, C.W., Tirrell, M, and Wellinghofl', S.T., “Experimental Studies of Phase Separation in Reaction Injection Molded (RIM) Polyurethanes”, Polymer Eng'neering and Science, Vol. 22, No. 11, (1982). Camargo, RB., Macosko, C.W., Tirrell, M, and Wellinghofi‘, S.T., "Phase Separation Studies in RIM Polyurethanes: Catalyst and Hard Segment Crystallinity Efi‘ects", Polymer, 26, 1 145-1 154, (1985). Camberiin, Y. and Pascault, JP, “Phase Segregation Kinetics in Segmented Linear Polyurethanes: Relations between Equilibrium Time and Chain Mobility and between Equilibrium Degree of Segregation and Interaction Parameter”, J. Polymer Science: Polymer Physics Edition, Vol. 22, 1835-1844, (1984). Camberlin, Y. and Pascault, J.P., “Quantitative DSC Evaluation of Phase Segregation Rate in Linear Segmented Polyurethanes and Polyurethaneureas", J. Polymer Science: Polymer Chemistry Edition, Vol. 21, 415-423, (1983). Carleton, C.S., Waszeciak, D.P., and Alberino, L.M., ”A RIM Process for Reinforced Plastics”, Session 9-D, Proc. of Reinforced Plastics/Composites Inst, Soc. Plast. Ind., Atlanta, (1986). Chang, AL, "Processing-Marphology-Property Relationships of Polyurethanes", Ph.D. Thesis, Univ. of Minnesota, (1984). Chang, AL, Briber, RM., Thomas, E.L., Zdrahala, RJ., and Critchfield, F.E., ”Morphology Study of the Structure Developed During the Polymerization of a Series of Segmented Polyurethanes”, Polyrn, 23, 1060-1065, 1982). Chen, J.H. and Rukenstein, E., “Solvent-Stimulated Surface Rearrangement of Polyurethanes”, J. Colloid Interface Sci. 135, 496, (1990). Chen, W.P., Frisch, KC, and Wong, S.W., "The Efl’ect of Soft Segment on the Morphology of Polyurethane Elastomers", Advances in Urethane Science and Tech, Vol. 11, 110-136, Technomic Publication, Lancaster, PA, (1992). Chen, Z.S., Yang, W.P., and Macosko, C.W., "Polyurea Synthesis and Properties as a Function of Hard Segment Content”, Rubber Chem. and Tech, 61, 86-99, (1987). Chiang, C.I-I, Ishida, H., and Koenig, J.L., “The Structure of Gamma.- Aminopropyltriethoxysilane on Glass Surfaces”, Journal of Colloid and Interface Science, Vol. 74, 396 (1980). Cornell, M.C., Turner, RB., and Morgan, RB., ”Amine Modified Polyol for RIM Systems”, Proceed. SP126th Ann, (1984). 175 Darnani, 8.6. and Lee, L.J., "The Resin-Fiber Interface in Polyurethane and Polyurethane- Unsaturated Polyester Hybrid”, Polymer Composites, Vol. 11, No. 3, 174-183 (1990). Dawson, LR. and Shortall, J.B., 'The Impact Behaviour of RRIM Polyurethane Elastomers", Cellular Polymers 1, 41-51, (1982). Deng, Z. and Schreiber, H.P., ”Orientation Phenomena at Polyurethane Surfaces”, J. Adhesion, Vol. 36, pp. 71-82, (1991). Drown, ER, Moussawi, HA, and Drzal, L.T., “Glass Fiber Sin'ngs and Their Role in Fiber-Matrix Adhesion", J. Adhesion Sci. Technol., Vol. 5, No. 10, 865-881 (1991). Eberhardt, J.G., 166ml ofPhysical Chemistry, 70, 1183, (1966). Eckles, J.H. and Wilkinson, T.C., “Composite Manufacturing Using the Reaction Injection Molding Process", Session 9-B, Proc. of Reinforced Plastics/Composites Inst., Soc. Plast. Ind., Atlanta, (1986). Estes, GM. and Cooper, S.L., ”Block Copolymers and Related Heterophase Elastomers", J. Macromol. Sci-Revs. Macromol. Chem, C4 (2), 313-366, (1970). Fielder, 1K. and Carswell, R, ”Polyurethane RIM Systems for Automotive Modular Wmdows', SAE Int. Cong. and Exp., Detroit, MI, Paper 900515, (1990). Fowkes, F.M., "Treatise on Adhesion and Cohesion", Vol. 1, Patrick, RL., Ed., Marcel Dekker, p. 352, (1967). Fowkes, F.M., Dwight, D.W., and Cole, D.A., ”Acid-Base Properties of Glass Surfaces”, J. Non-Crystalline Solids 120, 47-60, (1990). Galla, EA, Mascioli, RL., and Bechara, I.S., “RIM Catalysis and Coating Interaction”, Proceed. SP126th Ann. Tech. Conf., 24-31, (1981). Galli, E., ”Surface Modification for RRIM Urethanes”, Plastic Compounding, January, 21- 28 (1982). Garbassi, F., Occhiello, E., Bastioli, C. and Romano, G., ”A Qualitative and Quantitative Assessment of the Bonding of 3-Methacryloxypropyltrimethoxysilane to Filler Surfaces Using XPS and SSIMS (FABMS) Techniques", Journal of Colloid and Interface Science, Vol. 117, No. 1 (1987). Garton, A., “Infrared Spectroscopy of Polymer Blends, Composites and Surfaces”, Hanser Publishers, (1992). 176 Geer, R.P., 'Poly (dicyclopentadiene): A New RIM Thermoset", Proceed, Soc. Plast. Eng., National Techn Conf., Detroit, MI, (1983). Gilli, Ed, “Surface Modifications of RRIM Urethanes”, Plastics Compounding, 21-28, (1982). Gillis, HR, Hurst, AT., Ferrarini, L.J., and Watts, A, ”Liquid MDI Isocyanates: Efl‘ects of Structure on Properties of Derived Polyurethane Elastomers", SPI Polyurethane Div., 6th Intern TechMarketing Conf., 375-380, (1983). Gonzales, VM and Macosko, C.W., 'Properties of Mat Reinforced Reaction Injection Molded Materials“, Polyrn Comp. 4, 190-195, (1983). Good, RJ., 'Treatise on Adhesion and Cohesion“, Vol. 1, Chapter 2, Patrick, RL., Ed., . Marcel Dekker, (1967). Heam, M.J., Ratner, B.D., and Briggs, D., “SIMS and XPS Studies of Polyurethane Surfaces 1. Preliminary Studies", Macromolecules, 21, 2950-2959, (1988). Heam, M.J., Briggs, D., Yoon, S.C., and Ratner, B.D., “SIMS and XPS Studies of Polyurethane Surfaces, 2. Polyurethanes with Fluorinated Chain Extenders", Macromolecules, 21, 2950-2959, (1988). Hedrick, RM, Gabbert, J.D., and Wohl, M.H., "Nylon 6 RIM”, in Reaction Injectin Molding, Editor Kresta, J .E., Am Chem. Soc. Symp. Series 270, Washington, DC, Chapter 10, (1985). Hepburn, C., "Polyurethane Elastomers", Elsevier Science Publication, New York, NY, (1992). Ho, H., Tsai, M., Morton, J., and Farley, G.L., “An Experimental Procedure for the Iosipescu Composite Specimen Tested in the Modified Wyoming Fixture", J. Composites Technology and Research, 15, 1, 52-58, (1993). Kaelble, DH, 'Physical Chemistry of Adhesion”, John Wiley & Sons, (1971). Kaelble, DH. and Uy, K.C., 'Reinterpretation of Organic Uquid-Polytetrafluoroethylene Surface Interactions", J. Adhesion, 2, 50, (1970). Kau, C., Hiltner, A, Beer, 12., and Huber, L., ”Damage processes in Reinforced Reaction Injection Molded Polyurethanes", J. of Reinforced Plastics and Composites, Vol. 8, 18-39, (1989). Kinloch, AJ., ”Adhesion and Adhesives”, Chapman and Hall, (1987). 177 Koenig, JL. and Shih, P.T.K, 'Raman Studies of the Glass Fiber-SilaneResin Interface", Journal of Colloid and Interface Science, Vol. 36, 247 (1971). Lsssas, A, Fagerholrn, LE, Stenlund, B.G. and Nasrnan, J.H., “Modification of E Glass Fiber Surfaces with Polyacrylic Acid Polymers From an Aqueous Solution”, Polymer Composites, Vol. 14, No. 1, (1993). Legge, N.R, Holden, G. and Schroeder, HE, ”Thermoplastic Elastomas", Hanser Publishers, pp. 22, (1987). Macosko, C.W., “Fundamentals of Reaction Injection Molding", Hanser Publishers, New York, NY, (1988). Marlms, S. and Priel, 2., “Chemical Bonding Between Aromatic Molecules and Glass Surfaces", Journal of Colloid and Interface Science, Vol. 82, No. 1 (1981). Markus, S. and Priel, 2., "Chemical Bonding Between Molecules and Glass Surfaces”, J. of Colloid and Interface Science, Vol. 82, No. l, (1981). Mchy, F.J. and Fyiwara, M., Modern Plastics, 45, 143, (1968). McWilliams, D.E., Killmeyer, C.W., and Eilerrnan, G.E., ”Glass Fiber Size of Curable Blocked Polyurethane Emulsion with Amino Silane”, U.S. Patent 3,803,069 (1974). Miller, C.E., 'Near-Infiared Spectroscopy of Synthetic Polymers”, Applied Spectroscopy Reviews, 26 (4), 277-339, (1991). Miller, C.E., Edelman, P.G., Ratner, B.D. and Eichinger, B.E., "Near-Infrared Spectroscopic Analyses of Poly (ether urethane urea) Block Copolymers. Part 1. Bulk Composition”, Applied Spectroscopy, Vol. 44, No. 4, (1990). Miller, C.E., Edelman, P.G., Ratner, B.D. and Eichinger, B.E., "Near-Infrared Spectroscopic Analyses of Poly (ether urethane urea) Block Copolymers. Part II. Phase Separation”, Applied Spectroscopy, Vol. 44, No. 4, (1990). Mohai, M., Bertoti, I. and Revesz, M., 'XPS Study of the State of Oxygen on a Chemically Treated Glass Surface”, Surface and Interface Analysis, Vol. 15, pp. 364-368, (1990). Oertel, G., “Polyurethane Handbook”, Hanser Publishers, (1993). Otaigbe, J.U., "Effect of Coupling Agent and Absorbed Moisture on the Tensile Properties of a Thermoplastic RRIM Composite”, Journal of Applied Polymer Science, Vol.45, 1213-1221 (1992). 178 Owens, DR and Wendt, RC., ”Estimation of the Surface Free Energy of Polymers”, Journal of Applied Polymer Science, 13, 1740, (1969). Pannone, MC. and Macosko, ”Reaction Kinetics of a Polyurea Reaction Injection Molding System”, Polym. Eng. and Sci., 28, 660-669, (1988). Parrtano, C.G., ”Surface and In-Depth Analysis of Glass and Ceramics”, Ceramic Bulletin, Vol. 60, No. 11, (1981). Plueddemann, E.P., ”Silane Coupling Agents”, Plenum Press, (1982). Rso, V. andDml, L.T., ”TheDependenceofInterfacial Shear StrengthonMatrixand Interphase Properties”, Polymer Composites, Vol. 12, No. 1, 48-56, (1991). Rao, V. and Drzal, L.T., ”The Temperature Dependence of Interfacial Shear Strength for Various Polymeric Matrices Reinforced with Carbon Fibers”, J. Adhesion, Vol. 37, 83-95, (1992). Reilly, Albert F. and Sanok, John 1., ”Method and Apparatus for Making Molded Window Gasket”, us. Patent 4,755,339, (1988). Ryan, AJ., Stanford, J.L., and Still, RH., ”Calculation of Solubility Parameters of Typical Structural Units Present in Segmented Polyurethanes, Poly (urethane ureas) and Polyureas Formed by Reaction Injection Moulding”, Polymer Communications, Vol. 29, (1988). Saunders, J.H. and Frisch, K.C., ”Polyurethanes Chemistry and Technology, Part 1”, John Wiley and Sons Inc, (1983). Schneider, N.S., Desper, C.R, Illinger, J.L., King, AO., and Barr, D., ”Structural Studies of Crystalline MDI-Based Polyurethanes”, J. Macromol. Sci-Phys, B11 (4), 527-552, (1975). Schultz, J., Tsutmmi, K and Donnet, J .B., ”Surface Properties of High-Energy Solids. II. Determination of the Nondispersive Component of the Surface Free Energy of Mica and Its Energy of Adhesion to Polar Liquids”, J. Colloidal Interfacial Science, 59, 277, (1977). Schwarz, B.G., Critchfield, F.E., Tackett, LR, and Tar-in, P.M., ”Silane Effects and Machine Processing in Reinforced High Modulus RIM Urethanes Composites”, Journal of Elastomers and Plastics, Vol 11, p. 280, (1979). Sengupta, a and Schreiber, H.P., ”Surface Characteristics of polyurethane Adhesive Formulations”, J. Adhesion Science and Technology, Vol. 5, No. 11, pp. 947-957, (1991). 179 Speckhard, TA and Cooper, S.L., ”Ultimate Tensile Properties of Segmented Polyurethane Elastomers: Factors Leading to Reduced Properties for Polyurethantes Based on Nonpolsr Soft Segments”, Rubber Chem. and Tech., 59, 405-431, (1986). Sugden, S., J. Chem. Soc., 125, 1177, (1924). Trivosono, NM, Lee, LR, and Schriner, SM, Ind. Eng. Chem. 50, 912 (1958). Uhlmann, DR and Kreidl, N.J., ”Glass: Science and Technology”, Vol. 1, Academic Press, (1983). Van Krevelen, D.W., ”Properties of Polymers”, Elsevier Science Publishers, 3rd Ed., (1990). Vargo, T.G., Hook, D.J., Gardella Jr., JA, Eberhardt, M.A., Meyer, AE., and Baler, RE., ”A Surface Spectroscopic and Wettability Study of a Segmented Block Copolymer Poly (ether urethane)”, Applied Spectroscopy, Vol. 45, No. 3, (1991). Vaughan, DJ. and Peek Jr., R.C., ”The Use of ESCA in the Study of the Surface Chemistry of Commercial Glass Fibers”, The Society of Plastics Industry, Inc., 29th Annual Tech. Conf., The Society of Plastics Industry, Sec. 13-D, 1-4 (1974). Westley, SA, ”Some Recent Developments in Bonding RIM Urethane Automotive Modular Glass”, SAE Int. Cong. and Exp., Detroit, MI, (1991). Wongkamolsesh, K. and Kresta J .E., ”Organotin Catalysis in Urethane Systems”, Reaction Injection Molding - Poly. Chem. and Eng, Editor Kresta, J.E., 111-133, (1985). Yang, RC. and Lee, W.M., ”Fracture Mechanism Study of Flake Glass Filled RIM Urethane Systems”, J. of Elastomers and Plastics, Vol. 19, 120-146, (1987). Yoon, SC. and Ratner, B.D., ”Surface Structure of Segmented Poly (Ether Urethanes) and Poly (Ether Urethane Ureas) with Various Perfluoro Chain Extenders. An X-ray Photoelectron Spectroscopic Investigation”, Macromol., 19, 1068, (1986). Yoon, SC. and Ratner, B.D., ”Surface and Bulk Structure of Segmented Poly (ether urethanes) with Perfluoro Chain Extenders. 2. FTIR DSC, and X-ray Photoelectron Spectroscopic Studies”, Macromol., 21, 2392-2400, (1988). Zdrahala, RJ., Gerkin, RM., Hager, SL, and Critchfield, F.E., ”Polyether-Based Thermoplastic Polyurethanes. I. Effect of the Hard-Segment Content”, J. Applied Polymer Science, 24, 2041-2050 (1979). 180 Zisman, WA, ”Relation of the Equilibrium Contact Angle to Liquid and Solid Constitution”, Contact Angle, Wettability and Adhesion, Advances in Chemistry Symposium Series 43, Gould, RF., Ed., ACS, (1964). ‘0111000“