Development of bonding methods and energy absorption of sandwich panels for thermoplastic advanced composites

Given their high strength-to-weight and stiffness-to-weight ratios, sandwich composites continue to be considered for automotive applications. Thermoplastic materials, while difficult to bond, have an increased ease of manufacture and may be reprocessed, making them an attractive alternate to thermoset composites. This investigation focused on the evaluation of adhesives and surface treatments for both nylon and polypropylene thermoplastic composite adherends made from Towflex® pre-impregnated composite fabric. A manufacturing method was established for thermoplastic plates, which produced an acceptable surface finish without contaminating the bonding surface. Adhesives and surface treatments were evaluated using lap shear (ASTM D 3163) and cleavage (ASTM D 3433) test methods. The most promising adhesive/surface treatment combinations were selected for bonding of sandwich composites with two different core materials: balsa wood and polyurethane foam. Initial sandwich configuration testing consisted of flatwise tensile (ASTM C 297) and core shear (ASTM C 273) test methods. These tests provided insights into the sandwich properties and revealed any incompatibilities between the adhesive and the core. Follow on sandwich configuration evaluation consisted of edgewise compression testing, both statically (ASTM C 364) and dynamically. These tests determined the strength and ability of these sandwiches to absorb energy under two different types of loading.

DEVELOPMENT OF BONDING METHODS AND ENERGY ABSORPTION OF SANDWICH PANELS FOR THERMOPLASTIC ADVANCED COMPOSITES by Erik Bravant Haslam A thesis submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Master of Science Department of Mechanical Engineering The University of Utah May 2012 Copyright © Erik Bravant Haslam 2012 All Rights Reserved vi iii The Unive rs i ty of Utah Graduat e School STATEMENT OF THESIS APPROVAL The thesis of Erik Bravant Haslam has been approved by the following supervisory committee members: Daniel O. Adams , Chair 1/27/2012 Date Approved K. L. Devries , Member 1/27/2012 Date Approved Kenneth L. Monson , Member 1/27/2012 Date Approved and by Timothy Ameel , Chair of the Department of Mechanical Engineering and by Charles A. Wight, Dean of The Graduate School. vi iv ABSTRACT Given their high strength-to-weight and stiffness-to-weight ratios, sandwich composites continue to be considered for automotive applications. Thermoplastic materials, while difficult to bond, have an increased ease of manufacture and may be reprocessed, making them an attractive alternate to thermoset composites. This investigation focused on the evaluation of adhesives and surface treatments for both nylon and polypropylene thermoplastic composite adherends made from Towflex® pre-impregnated composite fabric. A manufacturing method was established for thermoplastic plates, which produced an acceptable surface finish without contaminating the bonding surface. Adhesives and surface treatments were evaluated using lap shear (ASTM D 3163) and cleavage (ASTM D 3433) test methods. The most promising adhesive/surface treatment combinations were selected for bonding of sandwich composites with two different core materials: balsa wood and polyurethane foam. Initial sandwich configuration testing consisted of flatwise tensile (ASTM C 297) and core shear (ASTM C 273) test methods. These tests provided insights into the sandwich properties and revealed any incompatibilities between the adhesive and the core. Follow on sandwich configuration evaluation consisted of edgewise compression testing, both statically (ASTM C 364) and dynamically. These tests determined the strength and ability of these sandwiches to absorb energy under two different types of loading. vii v TABLE OF CONTENTS ABSTRACT....................................................................................................................... iv LIST OF TABLES........................................................................................................... viii ACKNOWLEDGMENTS ................................................................................................. ix 1 INTRODUCTION .........................................................................................................1 2 LITERATURE REVIEW..............................................................................................5 3 THERMOPLASTIC FACE SHEETS AND MANUFACTURING............................18 3.1 Fiber Reinforced Thermoplastic Material..................................................18 3.1.1 Polypropylene ................................................................................18 3.1.2 Nylon..............................................................................................19 3.2 Manufacturing............................................................................................20 3.2.1 Thermoplastic Plate Consolidation................................................21 3.2.2 Contamination Issues .....................................................................30 4 ADHESIVES AND SURFACE TREATMENTS .......................................................33 4.1 Introduction................................................................................................33 4.2 Overview of Adhesives..............................................................................34 4.3 Surface Treatments ....................................................................................35 4.3.1 Acid Washing.................................................................................35 4.3.2 Openair® Plasma ...........................................................................36 4.3.3 ATmaP® Flame.............................................................................37 4.3.4 Surface Activation (Sulfonation) ...................................................39 5 LAP SHEAR TESTING..............................................................................................40 5.1 Introduction................................................................................................40 5.2 Background................................................................................................40 5.3 Methods......................................................................................................41 5.4 Results and Discussion .............................................................................43 5.5 Summary....................................................................................................47 6 FRACTURE TOUGHNESS TESTING......................................................................48 vii vi 6.1 Introduction................................................................................................48 6.2 Background................................................................................................48 6.3 Methods......................................................................................................50 6.4 Results and Discussion ..............................................................................51 6.5 Summary....................................................................................................55 7 ENVIRONMENTAL TESTING.................................................................................56 7.1 Intoduction .................................................................................................56 7.2 Background................................................................................................56 7.3 Methods......................................................................................................56 7.4 Results and Discussion ..............................................................................58 7.4.1 Acid Washed Nylon Lord 320/322................................................58 7.4.2 Openair® Plasma Polypropylene Dow LESA with Betamate™ 5404A Primer.........................................................................59 8 FIBER REINFORCEMENT TESTING......................................................................61 8.1 Introduction................................................................................................61 8.2 Background................................................................................................61 8.3 Methods......................................................................................................61 8.4 Results and Discussion ..............................................................................62 8.4.1 Carbon to Glass Reinforcement Comparison for Nylon Thermoplastic .................................................................................62 8.4.2 Carbon to Glass Reinforcement Comparison for Polypropylene Thermoplastic ....................................................................64 9 ROUND ONE CONCLUSIONS.................................................................................66 10 FLATWISE TENSION................................................................................................68 10.1 Introduction................................................................................................68 10.2 Background................................................................................................68 10.3 Methods......................................................................................................69 10.4 Results and Discussion ..............................................................................70 10.5 Summary....................................................................................................71 11 CORE SHEAR.............................................................................................................74 11.1 Introduction................................................................................................74 11.2 Background................................................................................................74 11.3 Methods......................................................................................................75 11.4 Results and Discussion ..............................................................................76 11.5 Summary....................................................................................................79 12 STATIC EDGEWISE COMPRESSION TESTING ...................................................80 vi vii vii 12.1 Introduction................................................................................................80 12.2 Background................................................................................................80 12.3 Methods......................................................................................................81 12.4 Results and Discussion ..............................................................................82 12.4.1 Static Average Edgewise Compression Strength...........................82 12.4.2 Static Energy Absorption...............................................................86 12.4.3 Static Average Weight-Normalized Edgewise Compression Strength ......................................................................................................90 12.4.4 Static Weight-Normalized Energy Absorption..............................94 12.5 Summary....................................................................................................98 13 DYNAMIC EDGEWISE COMPRESSION................................................................99 13.1 Introduction................................................................................................99 13.2 Methods......................................................................................................99 13.3 Derivation of Dynamic Energy Absorption Method ...............................101 13.4 Results and Discussion ............................................................................104 13.4.1 Dynamic Average Edgewise Compression Strength ...................104 13.4.2 Dynamic Energy Absorption .......................................................108 13.4.3 Dynamic Average Weight-Normalized Edgewise Compression Strength..............................................................................108 13.4.4 Dynamic Weight-Normalized Energy Absorption ......................115 13.5 Summary..................................................................................................115 14 CONCLUSIONS AND RECOMMENDATIONS ....................................................119 REFERENCES ................................................................................................................124 vii LIST OF TABLES 3-1: Data from initial XPS testing. ................................................................................... 31 3-2: XPS results utilizing new tooling and zinc stearate mold release. ............................ 32 4-1: List of adhesives, adhesive type and comments regarding application of those used in this investigation. ...................................................................................................... 35 5-1: Nylon data collected and calculated from lap shear testing. ..................................... 44 5-2: Polypropylene data collected and calculated from lap shear testing......................... 45 6-1: Nylon fracture toughness data. .................................................................................. 52 6-2: Polypropylene fracture toughness data...................................................................... 53 9-1: Selection matrix for nylon candidates in bold for continuation of testing ................ 67 9-2: Selection matrix for polypropylene candidates in bold for continuation of testing .. 67 10-1: Flatwise tension results........................................................................................... 72 11-1: Core shear results ................................................................................................... 77 12-1: Static nylon edgewise compression strengths ......................................................... 83 12-2: Static polypropylene edgewise compression strengths .......................................... 84 12-3: Static nylon edgewise comprssion absorbed energy ............................................... 87 12-4: Static polypropylene edgewise compression absorbed energy ............................... 88 12-5: Static nylon weight-normalized edgewise compression strengths .......................... 91 12-6: Static polypropylene weight-normalized edgewise compression strengths ........... 92 12-7: Static nylon weight-normalized edgewise compression absorbed energy .............. 95 12-8: Static polypropylene weight-normalized edgewise compression absorbed energy 96 13-1: Dynamic nylon edgewise compression strengths.................................................. 105 v viii 13-2: Dynamic polypropylene edgewise compression strengths.................................... 106 13-3: Dynamic nylon edgewise absorbed energy ........................................................... 109 13-4: Dynamic polypropylene edgewise absorbed energy ............................................. 110 13-5: Dynamic nylon weight-normalized edgewise compression strengths................... 112 13-6: Dynamic polypropylene weight-normalized edgewise compression strengths..... 113 13-7: Dynamic nylon weight-normalized edgewise absorbed energy............................ 116 13-8: Dynamic polypropylene weight-normalized edgewise absorbed energy.............. 117 ix v ix ACKNOWLEDGMENTS I would like to take the opportunity to thank some of the key players in my graduate education. First, I would like to thank Dr. Dan Adams, my advisor and committee chairman for his ideas, guidance, experience, friendship and the doors he has opened for me. I am also appreciative of the time given by the remaining members of my supervisory committee: Dr. Larry DeVries and Dr. Kenneth Monson. I would like to thank the members of the Automotive Composites Consortium (ACC) for their financial support and guidance. I am especially grateful for the support from Kedzie Fernholz and Ann Straccia of Ford RIC. I would also like to thank everyone at the composites lab at the University of Utah. Specifically, I would like to thank Nathan Guymon for his hours of help in making this investigation a success. In addition, I would like to thank all of the other research assistants who assisted in testing, given me advice, and helped in any other way. 1 1 INTRODUCTION The use of advanced composites materials in automobiles is on the rise. For example, the Chevrolet Corvette has been focusing on using lighter materials to increase vehicle performance with the use of sheet molding compound (SMC) reinforced with random chopped fiberglass. Although this is still the primary use on the Corvette, monolithic and sandwich composites utilizing continuous fibers are increasingly being incorporated. The benefit of incorporating these structural materials is to reduce weight thereby improving efficiency and performance. Such applications have focused primarily on thermoset composites. However, thermoplastic composites offer reduced cycle-times, improved ease-of-manufacturing and are recyclable. The greatest difficulty, however, is bonding them to other structures. Previous research at the University of Utah, focused on characterizing candidate sandwich composites for use in automotive floor applications, indicated that thermoplastic composite facesheets were difficult to bond to core materials [1]. Based on this previous research, this current study focused on developing and evaluating joining methods between thermoplastic composite facesheets and candidate core materials in sandwich composites. Sandwich composites are comprised of a core material that is sandwiched between two laminated facesheets. The core material is much thicker than the facesheets, is generally low density and is intended to carry shear stresses from transverse loading. The facesheets are thin, resin/fiber reinforced sheets with high specific stiffness and strength 2 to resist tensile and compressive loadings in bending. Overall, sandwich structures offer increased strength-to-weight and stiffness-to-weight ratios. When bonding thermoplastics, the removal of surface contamination due to low molecular weight molecules can improve bonding and is commonly achieved by mechanical abrasion and solvent cleaning. Mechanical abrasion includes sanding, grinding, and grit blasting all of which mechanically alter the surface to increase wet-out and surface energy. Solvent cleaning relies on the fact that low molecular weight molecules go into solution more readily than the larger molecules. Since there are no known low-temperature solvents to remove contaminants from nylon or polypropylene, solvent cleaning is not a viable option for removing contamination as it is with thermoset epoxies. However, using an acidic solution for surface etching can improve adhesion of thermoplastics. Increasing the surface energy of unmelted thermoplastics to improve wetting and increase bond strength can be achieved by a number of methods. Even though it requires handling of hazardous materials, oxidation by strong chemicals can improve bondability. The use of flame, sulfonation and plasma treatments not only can clean the surface, but also can introduce polar chemical groups though oxidation. These procedures are currently practiced in industry and can result in good adhesion with common adhesives. A downside is that the benefits of these surface treatments are prone to fade over time or under elevated temperature, requiring them to be a process step just prior to bonding. The thermoplastic facesheets used in this study featured both polypropylene and nylon matrices reinforced with continuous carbon fiber. The investigation has focused on improving upon the promising bonding methodologies for joining thermoplastic 3 substrates developed in a previous study at the University of Utah [2] and characterizing the bond performance. Additionally, these improved bonding methodologies were applied to the manufacture of thermoplastic sandwich composites. The core materials investigated were balsa wood, structural foam, and honey-combs. The thermoplastic sandwich composites were evaluated for their structural and energy absorption properties. A consolidation method was developed yielding contaminant-free plates using a heated press with aluminum tooling surfaces and zinc stearate mold release for nylon reinforced carbon fiber and Kapton film for polypropylene reinforced carbon fiber. The manufacturing process was also adapted to improve surface finish and ensure plates were manufactured flat and free of warping. This process is comprehensively outlined and a summary of the removal of contaminants is given. This investigation focused on improving the most promising adhesive, surface preparation methods, surface treatments, and procedures for polypropylene reinforced carbon fiber and applying the most promising adhesive, surface preparation methods, surface treatments, and procedures for nylon reinforced carbon fiber determined in a previous investigation at the University of Utah [2]. Multiple rounds of experimental evaluation were performed to investigate and characterize the bonding methodologies of the two different types of adherends, and the mechanical properties of the sandwich composites. First, a round of lap shear (ASTM D 3163) testing was used on the polypropylene reinforced carbon fiber to identify bond strength. Next, a round of fracture toughness testing (ASTM D 3433) on polypropylene and nylon reinforced carbon fiber to investigate fracture toughness. 4 The two most promising candidates from each thermoplastic group identified in lap shear and fracture toughness testing were selected for follow-on testing under different environmental conditions. The two most promising candidates from each thermoplastic group identified in lap shear and fracture toughness testing were selected for follow-on testing to evaluate their strength using glass fiber reinforced composites. Eight of the most promising candidates were selected for sandwich panel manufacture and testing: four from nylon thermoplastic composite facesheets and four from polypropylene thermoplastic composite facesheets. Flatwise tension (ASTM C 297) and core shear (ASTM C 273) were used to determine the strength of the core/adherend interface, as well as the strength of the overall sandwich composite. These tests were also an excellent way to determine any adhesive/core issues. After identifying any adhesive/core issues, a down-selection of the initial sandwich panel candidates occurred where both nylon and polypropylene were selected for energy absorption testing. Static edgewise compression testing (ASTM C 364) and a dynamic edgewise compression test [3] were used to determine the ultimate strength of the thermoplastic sandwiches as well as the energy absorbed. 5 2 LITERATURE REVIEW The use of composite materials in the automotive industry continues to grow, although affordability, energy absorption, surface finish quality, and recyclability are difficult issues. Although a majority of the focus to date has been on monolithic composite laminates, composite sandwich structures are being considered for selected automotive applications to provide even greater strength-to-weight and stiffness-to-weight ratios. Composite sandwich structures consist of an inner, low-density core encased by and bonded to a pair of thin, outer composite facesheets. Currently, the primary interest for these sandwich composites appears to be in roof and floor applications, where the improved bending properties of sandwich structures are a major consideration [4]. Unlike thermosetting materials, thermoplastic materials can be repeatedly heated and melted at elevated temperatures. Combining this characteristic with the decreased cycle time compared to thermosets, thermoplastics become quite attractive for manufacture. However, the end use must be considered, because as the glass transition temperature of the thermoplastic is approached, the Young's modulus decreases [5]. Thus, the use of thermoplastics in higher temperature environments such as in engines could cause softening and loss of structural integrity. Details on the specific thermoplastics used in this investigation are provided later in this report. Thermoplastics are typically insoluble in each other so that the adhesive and the adherent must be the same polymer for melt bonding. Since they do not mix chemically, 6 polyethylene will not adhere to polypropylene when melted together. Welding of adherends of like material has become an accepted method of joining; however this can be difficult with complex shaped parts [6]. Adhesives tend to have difficult wetting out or spreading on thermoplastics because of a low surface energy and thus require elevated temperatures to develop strong bonds. Due to a smooth, impermeable surface and low molecular weight molecules interfering with the bond, mechanical interlocking directly onto the surface can be difficult [7]. Often the surface is contaminated due to low molecular weight molecules migrating to the surface during consolidation. Removal can improve bonding and is commonly achieved with mechanical abrasion or solvent cleaning. Mechanical abrasion includes sanding, grinding, grit blasting or any other process mechanically changing the surface. Mechanical abrasion (sanding or grit blasting) of the thermoplastic surface was found to improve bond strength in the previous University of Utah research [1]. Solvent cleaning relies on the fact that low molecular weight molecules go into solution more readily than the larger molecules. While solvent cleaning has been shown to be successful in removing such material from thermoset epoxies, it is much more difficult for polypropylene and nylon thermoplastics, since there are few if any reasonable low-temperature solvents for these materials [7, 8]. Increasing the surface energy of unmelted thermoplastics to improve wetting and increase bond strength can be achieved by a number of methods. Even though it requires handling of hazardous materials, chemical oxidation can greatly improve bondability. The use of flame, sulfonation and plasma treatments not only cleans the surface, but also introduces polar chemical groups though oxidation. These procedures are practiced in 7 industry and can be successful when combined with the proper adhesive materials. These surface treatments do tend to fade under elevated temperature conditions or over time and thus such processing should be done shortly prior to bonding [7, 8]. Primers are often used to enhance the bonding of an adhesive to a substrate. In some cases where the adhesive will not bond strongly directly to a substrate, a substance may be used as an "intermediary" between the adherend and the adhesive. If a very thin layer of this substance is applied to the substrate, it may form a primer to which the adhesive will develop a strong bond [7, 8]. The aerospace industry has provided much of the research of joining thermoplastic composites to date because joining is critical to the manufacture of aerospace structures. Much of this research took place in the late 1980s and early 1990s in an attempt to move the use of thermoplastics into mainstream aerospace use. The joining methods commonly tested can be grouped into two categories: standard thermoset joining methods adapted for thermoplastics and methods specific for thermoplastics. The standard thermoset joining methods are co-consolidation, adhesives and fasteners. However, co-consolidation of thermoplastics is actually a melt fusion process as opposed to co-curing for thermosets. Joining methods specific for thermoplastics include resistance welding, induction bonding, ultrasonic welding and microwave bonding [9]. Previous research performed at the University of Utah focused on the characterization of candidate sandwich composites for the use of automotive floor applications [1]. Six sandwich configurations utilizing thermoplastic materials- carbon/nylon, carbon/polypropylene and glass/polypropylene with polyurethane and balsa wood cores-were tested. Candidate adhesives were identified and flatwise tensile 8 testing was used to identify the best-suited adhesives to bond the facesheets to the core: Lord 320/322 toughened epoxy and Loctite 907 epoxy. Preliminary results also showed that surface preparation of the composite-specifically abrasion-increased the bond strength. Further flatwise tensile and core shear testing identified that these thermoplastic sandwiches predominantly failed between the adhesive and the facesheet as shown in Figure 2-1. As noted in these previous studies, the bond strength between the facesheet and the adhesive was lower than the strength of the core itself which is undesirable (Figure 2-2). Based on the poor bond strength of the parameter groups used, only thermoset facesheets were used for future sandwich composite evaluation. However, this previous testing [1] identified the need for further research to develop and evaluate joining methods between thermoplastic composite facesheets and candidate core materials as is undertaken in this study. Figure 2-1: Previous testing at the University of Utah involving bonding of thermoplastic composites resulted in adhesive or interface failure [1]. 9 Figure 2-2: Three-point flexure testing results from previous University of Utah research [1] with lower bond strength thermoplastic composite facesheets indicated. 9 10 In a study focusing on newly developed adhesives in the late 1980s, Powers and Trzaskos [10] investigated fastening techniques for the use of advanced thermoplastic composites for structural applications. Lap shear testing was performed with two structural epoxy film adhesives and two structural curing epoxy adhesives. Grit blasting, as well as grit blasting followed by a solvent wipe and plasma treatments, were used as surface treatments. The authors concluded that each of these treatments was effective with these adhesives and sufficient strengths were achieved in all cases. It is also interesting to note that a correlation was performed between the strength of the bonded joint and the number of plies and ply orientation used. Shear strength was shown to increase as laminate thickness increased to a nominal thickness of 0.25 cm (0.10 in). In addition, laminates with surface plies having fibers oriented parallel to the shear load had at least 15% improved shear strength. Prior to the above studies, Kodokian and Kinloch [11, 12] determined that the use of light abrasion/solvent wipe treatment with structural epoxy adhesives was inadequate. The resulting joints were relatively weak and the data had significant scatter. Adhesive or interfacial failure was observed. They determined a more comprehensive study was needed, which led to these later studies of more specific areas. Two composite materials were used-carbon fiber reinforced PEEK thermoplastic and epoxy thermoset. Two toughened epoxies and PEEK hot-melt were used in conjunction with simple abrasion, molding next to clean aluminum foil followed by an acid wash, acid etch and corona discharge. A double cantilever beam test was used with a crack initiated at one end. All four treatments were successful with the use of the epoxies in bonding the thermosets, but the simple abrasion was not successful. In addition, the use of the 11 aluminum foil and acid etch did not increase bond strength significantly, indicating that surface contamination from mold release is not the essence of the problem. This finding was confirmed with the use of XPS to determine that silicone and fluorine levels were very low. Parker [13] similarly determined that initial bond strength is related to the presence of contaminants on the adherend surface. Thus, confirmation of a clean surface after adherend manufacture is necessary. Kodokain and Kinloch [11, 12] also found that the use of the hot-melt did increase bond strength with the PEEK samples, but had little adhesion with the thermoset adherends. Corona discharge did improve fracture energy (Gc) values and cohesive failures were noted with both materials. Another factor that can influence bond strength is the bondline thickness. McKnight et al. [14] investigated the effects of adhesive cure cycle and final bondline thickness on bond strength. Using adherends of S-2 glass/polyphenylene sulfide, thermoplastic composites with a co-molded polysulfone surface layer were bonded using a high performance epoxy film adhesive and an epoxy paste adhesive. Results indicated that bondline thickness reaches an optimum level before resulting in decreasing strength values. Silverman [15] chose a slightly different approach using six different joining concepts to achieve strengths approaching 93.0 MPa (13.5 ksi). The concepts included adhesive bonding using a commercially available epoxy system, mechanical fasteners plus adhesive bonding, resistance heating, focused infrared heating, ultrasonic welding, and a novel technique employing an amorphous thermoplastic film. Advanced thermoplastic composites specimens were made using PEEK reinforced with AS-4 12 graphite pre-preg tapes. Tensile lap shear coupons were used for screening the alternative joining methods and measuring the bond strengths. Surface preparation methods examined included grit blasting, acid etching, and cold gas plasma etch treatment. The fusion bonding methods of resistance heating, focused infrared heating, and ultrasonic welding used a thin film of the neat thermoplastic PEEK in the joint prior to fusion. Even though none of the methods achieved the intended strength, epoxy with plasma treatment had the highest joint strength at 41.6 MPa (6.04 ksi) with cohesive and adhesive failure occurring equally. The strengths of the other methods dropped significantly. More recently, Loven [16] investigated the use of adhesives for structural bonding for the automotive and marine industry. Even though this study does not investigate scientifically beyond a case study, it does point out that thermoplastics and composites can be structurally bonded in application. Little research has been performed to date using nylon thermoplastic composites. Wade et al. [17] did investigate the surface modification and adhesive bonding of a nylon 6,6 reinforced with unidirectional glass fiber. Wettability studies of plasma-treated specimens showed a significant reduction of contact angles in water, relative to untreated material. The most effective treatment used oxygen plasma. The increases in wettability observed were determined to be the result of two effects. First, the treatment resulted in an increase in the concentration of oxygen and nitrogen containing functional groups on the surface of the polymer. Second, the plasma treatment removed fluorine contamination, the source of which was identified as the PTFE mold release agent. The surface modification resulted in significantly improved adhesion between the composite 13 and an applied toughened epoxy adhesive. Thus, a considerable increase in the Mode II critical strain energy release rate, GIIc, was observed following plasma treatment. Specimens treated in oxygen plasma showed the greatest improvement in GIIc and resulted in a cohesive failure. Without plasma treatment, the specimens failed in an adhesive mode at very low values of GIIc. Adhesion was further optimized by molding the nylon-6,6 plates directly against steel plates instead of PTFE thereby removing any chance of fluorine contamination. As noted, previous significant research has focused on determining methods to successfully join thermoplastic composites. These investigations have largely used different surface treatment methods that include abrasion, acid etching and plasma treatment. However, other surface treatments exist such as flame, corona and sulfur exposure. Development of these treatments and advances in the others result in the need for further investigation and testing of a wider variety of surface treatments as is taken on in this study. Details of the surface treatments investigated and used are found below. After suitable surface treatment and adhesive combinations have been found, they must then be applied to the making of sandwich panels. This investigation utilizes two ASTM standard tests, flatwise tension and core shear, and two modified tests, static and dynamic edgewise compression. These particular test methods were selected based on their ability to readily identify performance problems and to use in assessing performance for automotive applications. Flatwise tensile testing (ASTM C 297) is used to evaluate the strength of the sandwich under loading in a direction perpendicular to the facesheets [18]. This test is often conducted first for sandwich composites to assess the as-manufactured performance 14 of the facesheets, the core material and the bond between the constituents. The primary functions of this test are to identify material compatibility problems associated with the facesheet, core, and adhesive as well as manufacturing related problems such as core crushing, improper facesheet consolidation, etc. The flatwise tensile strength and the failure location/failure mode are the most important information from these tests. A schematic of the test setup is illustrated in Figure 2-3. In order to produce flatwise stress in the sandwich construction, steel loading blocks are adhesively bonded to each facesheet. A cylindrical pin is seated in each loading block for a means of load application. The pins are used to distribute the applied load and reduce bending moments. Core shear testing (ASTM C 273) is used to evaluate shear properties of the core material [19]. Strength and stiffness properties are determined from shear deformation in planes parallel to the facesheets. This test does not produce pure shear. However specimen lengths are sufficient so that secondary stresses are minimal. Figure 2-3: Flatwise tension test setup Facesheets Core 15 The resulting properties from these tests are fundamental to the use and design of sandwich composites. From the load versus crosshead deflection curve, the following determinations can be made: the core shear stress at any point (including failure), the effective core shear modulus, and the energy absorbed in the core under shear deformation. Figure 2-4 illustrates the core shear testing setup. A metal plate is adhesively bonded to each facing of the sandwich construction. Load is applied at cylindrical pins seated in the metal plates such that the line of action passes through opposite corners in the core. Edgewise compression testing (ASTM C 364) is used to determine the compressive properties of a sandwich panel [20]. Additionally, the area under the load versus deflection curve can be calculated to determine the energy absorption of the sandwich under edgewise loading. Figure 2-4: Core shear test setup 16 Previous work by Van Otten [21] investigated different edgewise crush initiators. These initiators encouraged the correct failure sequence of the sandwich to get the maximum possible energy absorption. It showed that there was a dependence between the stiffness of the facesheet and the strength of the core. If the facesheet was too stiff, and the core insufficiently strong, the facesheet would fail in buckling and separate from the core. This finding led to the use of a modified static edgewise testing fixture, which had been used in a previous investigation at the University of Utah [3]. The fixure does not provide end clamping over 25.4 mm of the specimen at the top fixture instead it has two V-blocks as shown in Figure 2-5. These V-blocks realign the specimen during a buckling failure. This realignment allows the specimen to absorb more energy in the direction of compression. Figure 2-5: Modified edgewise compression fixture 17 There is currently no standard for dynamic edgewise compression testing. Thus, ASTM C 364 for edgewise compressive strength of sandwich constructions was used as a rough guideline for a previous investigation at the University of Utah [3]. The fixture used is the same geometry as that used in the modified static edgewise compression testing. However, the fixture is modified for use in a drop tower. 18 3 THERMOPLASTIC FACE SHEETS AND MANUFACTURING 3.1 Fiber Reinforced Thermoplastic Material Two thermoplastic matrix materials reinforced with carbon fiber were used in all testing: polypropylene and nylon. The material, TowFlex® woven fabric reinforced thermoplastic, was produced by Hexcel Composites [22, 23]. TowFlex® laminates are manufactured using a compression molding process, where the material is heated to a specified melt temperature, pressure is applied for consolidation, and the resulting part is cooled while maintaining pressure. Details of this manufacturing process are outlined below. The two thermoplastic materials used are very common and their use is widespread in the plastics industry. Use in the composites industry has been hindered by not only the ability to bond to other structures, but also the ability to adhere the thermoplastic matrix to the fiber reinforcement. Hexcel has successfully accomplished this bonding by spreading and powder coating continuous carbon and glass fibers with thermoplastic particles. The thermoplastic particles are subsequently melt-fused to the reinforcement fibers, producing a flexible material. Similar to a preimpregnated material, TowFlex® is then ready for consolidation off of the roll [22, 23]. 3.1.1 Polypropylene Polypropylene has a wide variety of uses from reusable storage to clothing. It is highly resistant to chemicals and has higher softening and melting temperatures than 19 other common plastics. The Young's (elastic) modulus is midrange, allowing for use where stiffness is needed, but has some resistance to brittle failure [5]. Polypropylene is an addition polymer, meaning it is synthesized from propylene. The propylene monomers are added together without the loss of any atoms or molecules to create the polypropylene chain as seen in Figure 3-1. 3.1.2 Nylon Nylon is a family of synthetic polymers created by Dow Chemical in the 1930s. It was originally used in women's stockings, parachutes and more recently has been used in ropes, guitar strings and mechanical parts. Like polypropylene, the Young's modulus is favorable to provide stiffness without brittle failure. Unlike polypropylene the monomers used donate carbon to the polymer to create chains of six- sided carbon groups as seen in Figure 3-2. This process defines the two most common types of nylon: nylon 6 and nylon 6,6. Nylon 6,6 has an additional 6 carbon molecules repeated in the polymer chain that give extra strength and the resilience compared to nylon 6 [5]. The TowFlex® material in this study utilizes nylon 6. Figure 3-1: A general polypropylene chain. 20 Figure 3-2: A general nylon 6 chain. 3.2 Manufacturing The thermoplastic adherends used were fabricated at the University of Utah using a heated press to consolidate the TowFlex® woven fabric reinforced thermoplastic. Aluminum tooling surfaces were used to enhance surface finish. The consolidation cycle for the nylon-based TowFlex® was 1.0 MPa (~150 psi) at approximately 260ºC (500ºF) for approximately 15 minutes. The consolidation cycle used for the polypropylene-based TowFlex® was 1.0 MPa (~150 psi) at approximately 190ºC (375ºF) also for approximately 15 minutes. The tool was shuttled to a room temperature press and placed under 1.0 MPa (~150 psi) to increase the cooling rate. Details of the manufacturing process for the thermoplastic plates are outlined below. Images, where useful, follow the specific process step. This detail includes modifications made for improved surface finished and consolidation while minimizing warping. Following the process is a summary of the contamination issues faced in using mold release and film barriers, justifying the contamination-free method that has been used. 21 3.2.1 Thermoplastic Plate Consolidation Below is a detailed process instruction developed and used for the manufacture of the thermoplastic plates used in this investigation. 1. Put on sterile gloves to prevent contamination during all steps of the process (Figure 3-3). 2. Cutting Material a. Wipe table and tools with acetone and paper towel to prevent contamination. b. Using two squares and a straight edge for alignment, a 25.4 cm (10 in.) piece of Towflex® is cut off at the width of the roll (Figure 3-4). c. Noting the orientation, this piece is cut into 3 pieces 25.4 cm (10 in.) square, ensuring the selvage edge of the roll is not used (Figure 3-5). d. Extra material is stored for potential later use. Figure 3-3: Sterile gloves are worn throughout the process to reduce contamination. 22 Figure 3-4: 10" strips of material are cut from the roll of thermoplastic material. Figure 3-5: The 10" strips are then cut into pieces making 10" by 10" squares. 23 3. Tool Preparation - Nylon a. Aluminum plate 30 cm by 30 cm by 6.5 mm (12in x 12in x 1/4in) is used as a tooling surface. b. Aluminum foil is placed over the surface and folded around the edges. This provides a smooth repeatable surface with an adequate release so no initial or additional polishing is required with the aluminum sheet stock (Figure 3-6). c. Aluminum plate with foil is taken away from all thermoplastic materials and zinc stearate is applied with a side-to-side motion as a mold release. This is to prevent contamination from the spraying of the mold release in the laboratory. d. The released aluminum plate is brought back into the laboratory. 4. Tool Preparation - Polypropylene e. Aluminum plate 30 cm by 30 cm by 6.5 mm (12in x 12in x 1/4in) is used as a tooling surface. 5. Kapton film 30 cm by 30 cm (12in x 12in) is places on the lower aluminum plate (Figure 3-7). 6. Material Layup a. Lap Shear Specimens 1. Four plies of the Towflex® material are used. 2. Material is oriented in the 0/90 direction only. 24 Figure 3-6: Aluminum foil is applied to the mold surface. Figure 3-7: Kapton film placed on aluminum tool. 25 3. The curl of the material is placed onto the prepared tooling surface in the same direction so the edges curl up. This was found to give the smallest amount of warping of a finished plate (Figure 3-8). b. GIC Specimens 1. Eight plies of the Towflex® material are used. 2. Material is oriented in the 0/90 direction only. 3. The curl of the material is placed onto the prepared tooling surface in the same direction so the edges curl up. This was found to give the smallest amount of warping of a finished plate (Figure 3-8). c. Flatwise Tension, Core Shear, Static Edgewise Compression, and Dynamic Edgewise Compression Specimens 1. Four plies of the Towflex® material are used. 2. Material is oriented at [0/90,±45]T for a quasi-isotropic layup. 3. The curl of the material is placed onto the prepared tooling surface in the same direction so the edges curl up. This was found to give the smallest amount of warping of a finished plate (Figure 3-8). 7. Shim Placement a. Shim thickness is 1.5 mm (0.060 in). b. Four shims are placed just inside the tool edge to ensure thermoplastic plate thickness. 26 Figure 3-8: Material is placed onto the tooling surface. Figure 3-9: Shims are placed onto the tooling surface clear of interaction with thermoplastic material. 27 c. The shims must be placed in such a fashion so that when the curl of the material is flattened they do not interfere with the layup (Figure 3-9). 8. Pressing a. Press is preheated to 260ºC (500ºF) for the nylon and 190ºC (375ºF) for the polypropylene (Figure 3-10). b. The tool with the layup is placed into the press making sure that none of the shims shift inside. c. Press the tool with the hydraulic ram to achieve 1.0 MPa (~150 psi). d. Allow the press to come back up to 260ºC (500ºF) for the nylon and 190ºC (375ºF) for polypropylene, and allow to consolidate for 15 minutes at these temperatures. e. Ensure pressure remains constant at 1.0 MPa (~150 psi) for the entire consolidation cycle. f. After the 15 minute consolidation time, shuttle the mold to a room temperature press, and apply 1.0 MPa (~150 psi) as the mold cools. g. After the mold has cooled to below 120ºF, it can be removed from the room temperature press. 28 Figure 3-10: The mold is pressed to consolidate the thermoplastic material. 9. Removal of the Panel from the Tool a. Use a scraper or large screw driver to pry the two halves of the tool apart (Figure 3-11). b. Aluminum foil that is still attached to the plate can be removed by hand or with a razor blade (Figure 3-12). c. In one corner of the panel write: Date, Type of Material (PP or N6), and Number. d. Place the panel in the storage bag. e. Record the process in lab book and all deviations from the procedure. 29 Figure 3-11: The mold is opened. Figure 3-12: Excess aluminum foil is removed. 30 3.2.2 Contamination Issues Imperative to bonding plates successfully is ensuring the plates are free of contaminants. Contamination of the surface can result in a barrier between the adhesive and the adherend, resulting in decreased bond strength. Contaminants of great concern are the same materials that are often used in mold releases and include fluorine and silicone based materials. These particular materials reduce friction and surface energy when they are included on the surface of the adherends, which is of benefit when de-molding and a hindrance when bonding. The manufacture of clean, contaminant-free adherends requires measures be taken to ensure a clean environment around the material and that only certain mold releases be used. Initially, a tool was used that was well seasoned with PTFE mold release. This led to contamination of the bonding surface as was confirmed by poor bonding and X-ray Photoelectron Spectroscopy (XPS) performed at Ford Research Center. Based on these results, additional samples were made using a new aluminum plate as a tooling surface. A TFE film was used as a barrier and as a release layer. However, XPS analysis confirmed that high levels of fluorine (up to 7.9%) were still found on the plate surface as shown in Table 3-1. It was assumed that transfer of fluorine was occurring from the TFE film to the thermoplastic adherend. Given the level of fluorine on the surface, a series of plates were made utilizing new aluminum tooling surfaces and zinc stearate as a mold release. Plates were manufactured with no mold release; however the thermoplastic bonded to the aluminum plates necessitating some level of mold release. XPS analysis was performed in the University of Utah Nanofabrication Laboratory. Results indicated, the use of new aluminum tooling 31 Table 3-1: Data from initial XPS testing. surface and zinc stearate eliminated the fluorine contaminant, and improved surface finish as shown in Table 3-2. After an initial round of lap shear testing, the comparative bond strengths obtained for polypropylene in the initial qualitative testing were unsatisfactory. It was believed that the zinc stearate mold release was contaminating the surface of the polypropylene, resulting in the low comparative bond strengths. A Kapton film (polyolefin) was found to have the desirable characteristics of suitable release and low contamination for this application. Kapton films have been used as releases for higher processing temperature thermoplastics such as PEEK. It was found the Kapton would release from the polypropylene adequately but would not release from the nylon. Since the nylon showed excellent bond strength with the zinc stearate on aluminum foil, it was deemed not necessary to revise the release for the nylon. The polypropylene was inspected by XPS for and a reduction in zinc was found. Sample C O N F S Si K Na Cl Al PP-2 89.3 5.9 - 3 0.79 - - 1 - - PP-3 87.6 9.2 1.2 1.7 0.13 - - 0.27 - - PP-4 74.8 17.7 - - - 5.1 - 0.4 0.47 1.5 PP-Bulk 97.9 2.1 - - - - - - - - N6-2 69.1 13.2 7.3 7.9 0.74 0.31 0.66 0.88 - - Elemental Composition--Atomic % 32 Table 3-2: XPS results utilizing new tooling and zinc stearate mold release. Peak Position FWHM Raw Area RSF Atomic Atomic Mass BE (eV) (eV) (CPS) Mass Conc % Conc % Zn 2p 1022 3.982 39896.2 5.589 65.387 0.45 2.27 F 1s 689 1.822 17.5 1 18.998 0 0 O 1s 532 3.648 144893.8 0.78 15.999 10.64 13.07 Ca 2p 348 3.667 19175 1.833 40.078 0.62 1.92 C 1s 285 3.197 394655.9 0.278 12.011 86.18 79.51 Si 2s 153 3.206 4605 0.324 28.086 0.89 1.91 N 1s 401 2.82 9843.3 0.477 14.007 1.22 1.31 33 4 ADHESIVES AND SURFACE TREATMENTS 4.1 Introduction In a bonded joint, the adhesive joins two surfaces often called adherends or substrates. An adhesive is generally considered a material that bonds or unites two adherends together such that they act as a unit and load is transferred through the joint. Most structural adhesives are thermosetting, meaning that they cross-link and thus increase the number of primary bonds when curing [7, 8]. Epoxies have historically been the most commonly used adhesive with composites. However, new chemistries have improved acrylics, cyanocrylates and urethanes for use in this industry. Using the proper adhesive is only part of the solution when bonding composites and particularly thermoplastic composites. Surface treatments to increase surface tension can increase wet-out of adhesives increasing bond strength. In addition, surface treatments can remove contaminants that can result in a barrier between the adherend and adhesive causing low bond strength. There are four common theories used to explain how adherends and adhesives are bonded together. Mechanical interlocking is the simplest and suggests that the adhesive fills into voids in the surface of the adherend and cures. Once it is cured, it cannot be pulled out of the voids unless the adhesive or the adherend fails. Surface treatments can remove contamination and open up these voids for this bonding to take place. In addition, treatments that mechanically change the surface also increase the surface area of interaction between the adherend and adhesive [7, 8]. 34 Covalent bonding is a second theory that suggests that primary bonds occur between the adhesive and adherend during curing. For this to occur, chemical groups must cause a reaction at the interface of the materials. Electrostatic attraction is a third theory and suggests that the materials bonded together have an ionic attraction. Finally, acid-base interactions result in an interaction where one material gives up a hydrogen atom and the other accepts it [7, 8]. It is important to note that these theories are not completely established and much debate continues to surround them. No attempt is made to understand which type of bonding has occurred in this investigation. Instead, an understanding of failure type is deemed more important. Thus, failures are judged to be adhesive or cohesive. Adhesive failures occur between the adhesive/adherend interfaces and are considered unfavorable as the interface is the weakest point of the joint. Cohesive failures occur within the adhesive itself. In addition, failure within the adherend can occur. These are considered favorable because the bond strength is greater than the adhesive strength [5]. Below are a summary of the adhesives and surface treatments investigated and an outline of the decision process used in the selection process. 4.2 Overview of Adhesives In a previous study conducted by the University of Utah [2] only thermosetting structural adhesives were considered. This resulted in the identification of four types of adhesives: epoxies, acrylics, cyanoacrylates, and urethanes. However during testing it was found that only three adhesives were practical for use with this investigation: epoxies, acrylics, and urethanes. The adhesives used in this investigation are shown in Table 4-1. 35 Table 4-1: List of adhesives, adhesive type and comments regarding application of those used in this investigation. Supplier/Adhesive Adhesive Type Comments Lord 320/322 [24] toughened epoxy Moderate difficulty to use and spread. High viscosity Lord 7542 [24] urethane Easy to use and spread. Slight noxious odor. 3M 8239 [25] urethane Difficult to use due to short work time (~1 min). Slight noxious odor Dow LESA [26] acrylic Easy to use and spread. Highly noxious odor Dow LESA w/ BetaprimeTM 5404A [26] primer acrylic Easy to use and spread. Highly noxious odor 4.3 Surface Treatments Surface treatments were divided into two groups: in-house and specialty treatments. The in-house treatments are those that do not require highly-specialized processing equipment. The only in-house treatment investigated was acid washing. The specialty treatments require highly-specialized equipment which in each case investigated requires significant capital investment. Specialty treatments investigated include ATmaP® flame, Openair® plasma, and Sulfonation treatments. Each surface treatment method is summarized below. 4.3.1 Acid Washing The acid wash treatment was a simple cleaning with a 3% phosphoric acid solution to potentially remove contaminants. This is a commonly used method for removing small amounts of zinc stearate left on the surface after manufacture. Adherends were 36 completely submerged (Figure 4-1) in the 3% phosphoric acid solution for approximately 5 minutes before being rinsed clean with distilled water. Adherends were bonded within 24 hours of treatment. 4.3.2 Openair® Plasma A plasma is a higher energy state of matter composed of charged particles. It is what comprises the sun, stars and other objects seen outside of our solar system. Here on Earth, plasma can be found in neon lights, flames and electric discharges. While these sources are often associated with heat, the temperature of plasma is often below the temperature of the same material in a gaseous phase. Thus, a plasma treatment can be performed at a lower temperature thereby reducing potentially negative effects of heat. Openair® is a trademarked name owned by Plasmatreat North American Inc. Previously, plasma treatment had existed but it had to be contained within a chamber, Figure 4-1: Acid washing comprised of submersion in a 3% phosphoric acid solution for approximately 5 minutes. 37 making treatment difficult for large or complexly shaped parts. The Openair® method uses electricity to excite molecules in the air, thereby generating plasma. This plasma removes contaminants, static electricity and dust by oxidizing these compounds. The exposed polymer chains of the thermoplastic are also affected and the ends are broken off and replaced with highly reactive -OH and -NH groups. This increases the wettability and surface energy which potentially increases the bond strength of the treated adherend. The Openair® system utilizes a robot arm to move the nozzle that projects the plasma (Figure 4-2), allowing for varying speed and hence time of exposure of the surface being treated. In addition, the amount of plasma, or intensity, can be varied [27]. Adherends were treated courtesy of Ford RIC of Dearborn, MI. Five different parameter sets were used where speed and intensity of the plasma beam were varied. In all cases, adherends were bonded within 24 hours of treatment to remove fade as an unwanted potential parameter. 4.3.3 ATmaP® Flame Flame treatment utilizes combustion to burn contaminants, increase surface energy and improve wettability. Historically, flames of most all types have been used though significant advances have been made in the past decade. The ATmaP® flame Technology is a trademarked process by FTS Technologies of Flint, MI. The ATmaP® delivers a highly controlled flame that burns with an oxygen content of 0.2-1.2% oxygen. A water-borne diimine solution is evaporated into the flame and interacts with the surface to replace lower energy molecules that are removed from the plasma contained in the flame. Nitrogen is used to encompass the flame which results in the surface becoming 38 Figure 4-2: The Openair® plasma machine used by Ford RIC to treat adherends. polarized. The system allows for each of these parameters to be controlled in addition to flame dwell time through the use of a computer-controlled robot arm. Adherends were treated courtesy of FTS Technologies. FTS performed a series of tests to determine the ideal treatment parameters based on contact angle measurements. Contact angle measures the angle of various fluids dropped onto a surface. As the angle between the droplet and surface decreases, the surface is considered to have higher wettability and therefore have higher bond strength. Even though contact angle is not a 39 direct correlation to bond strength, it is assumed that smaller angles result in better bond strength. Even though previous research by FTS has shown that fade does not occur, the treated adherends were bonded within 24 hours of treatment to remove fade as an unwanted potential parameter [28]. 4.3.4 Surface Activation (Sulfonation) Surface Activation is a modification of traditional sulfonation. It is the exposure to specific concentrations of sulfur trioxide gas, thus attaching sulfur and oxygen atoms to the surface of the adherend. Next, the sulfonate is made chemically inert by treating the surface with a neutralizing agent. The result is a chemically modified polar surface layer that is up to 25 microns thick. The exposure takes place inside of the chamber filled with sulfur trioxide where most of the sulfur trioxide is recaptured, filtered and reused. The Surface Activation was performed by Surface Activation Technologies of Troy, MI. Exposure time was varied from 3 to over 10 minutes. Given the depth of the treatment, fade was not considered to be an issue [29]. However, adherends were bonded within 24 hours of treatment. 40 5 LAP SHEAR TESTING 5.1 Introduction The single lap shear test was used to identify the adhesive and surface treatment combinations with the best adhesion. It was selected due to its relative quickness, expense and ease of manufacture. In this investigation, lap joints were tested in tension at room temperature following procedure ASTM D 3163 [30]. Load versus displacement curves were generated for each test and compared to determine consistency of the bond strength between replicate specimens. The peak load was determined and the bonded area was measured for each parameter group to be used to calculate the comparative bond strength. 5.2 Background The lap shear joint test is a commonly used adhesive test. Specimens are simple to build, test, and readily resemble the geometry of many practical joints. Even though different configurations can be used, the most common for rigid plastic adherends is the single lap test outlined in the ASTM designation D 3163 shown in Figure 5-1. This test method is titled "Determining the Strength of Adhesively Bonded Rigid Plastic Lap Shear Joints in Shear by Tension Loading" [30]. It must be noted that although the test is called lap "shear" the failure is usually related to the tensile stresses rather than strictly the shear stresses. Typically, the apparent adhesive strength or lap shear strength is reported as the ratio of the load at 41 Figure 5-1: The single lap shear test configuration outlined in ASTM D 3163. failure to the area of overlap. This value is often reported in adhesive handbooks and on data sheets supplied by adhesive manufacturers even though the value is often different from the maximum stress. This has led to the use of overly simplified design rules such as average stress criteria that can lead to inaccurate assumed bond strength. Often left out are effects resulting from adherend thickness, adhesive bond thickness and overlap length, which Powers and Trzaskos determined affected bond strength significantly [10]. This test can be useful for comparison of the parameter groups in this investigation, but the results must be interpreted with caution. Thus, the term "comparative" bond strength is used to imply that this value should be used only for comparison within this investigation. For more information regarding the known issues associated with the single lap shear joint test, the reader is referred to ASTM D 4896 [31]. 5.3 Methods For lap shear testing, the test specimen geometry was selected in accordance with ASTM D 3163. The manufacturing procedure of the thermoplastic plates outlined in 42 Chapter 3 was followed. An acid wash was performed prior to all surface treatments, in an attempt to remove additional contaminants. The specimens were 4 plies thick and 25.4 mm (1.0 in) wide. Each specimen was approximately 101.6 mm (4.0 in) long and a bond length, L, was chosen to be 12.3 mm (0.5 in). Bond thickness was controlled by placing 0.51 mm (0.020 in) thick wires in the bond area. Fixtures were machined to hold the specimen plates at the proper places with respect to each other and pins were used to ensure proper placement. Plates at least 127 mm (5 in) wide bonded for each parameter group and all samples tested were cut from the same larger plate. The plates were held in the fixtures for at least 12 hours which at least doubled the hardening time for each adhesive. After removal from the fixtures, the specimens were cured at room temperature for at least 48 hours before they were cut, ensuring that a complete and consistent cure was achieved. A minimum of three specimens were tested for each test parameter group resulting in over 60 tests. In several cases, additional specimens were tested if one of the other specimens was damaged when handled or loaded into the test frame. Specimens were loaded into an Instron 4303 tensile testing frame with a 5 kip load cell. A constant displacement rate of 1.27 mm/min (0.050 in/min) was used. Load versus displacement curves, bond area, peak load, and failure type were noted for each specimen tested. The bond area and peak load were used to calculate comparative bond strength for each specimen and average stress for each parameter group. These bond strengths were then used to compare the different parameter groups and assess which parameter group achieved highest bond strength. The calculated strengths are not intended for use in 43 design, but only as a reference to compare the different parameter groups within this investigation. 5.4 Results and Discussion The different parameters were tested and the data are summarized in Tables 5-1 and 5-2. The data from Tables 5-1 and 5-2 are also presented graphically to compare average comparative bond strength for nylon and polypropylene, respectively. As seen in Figure 5-2 the parameter groups involving nylon achieved the highest comparative bond strength. The average strengths of the polypropylene parameter groups were considerably lower than the nylon with the exception of Openair® plasma treated Dow LESA with BetaprimeTM primer. It should also be noted that in general higher strength parameter groups had a smaller relative variation. Nylon not only outperformed polypropylene, it has the five highest comparative bond strength parameters. However, differences between these parameter groups are not statistically significant as shown in Figure 5-2. The Lord 320/322 epoxy had four of the five highest overall comparative bond strengths. It is also interesting to note that the two highest comparative bond strengths were achieved with the acid wash surface treatment. The comparative bond strength of nearly all the polypropylene candidates increased with the removal of the zinc stearate as the mold release as shown in Figure 5-2 with the exception of the Openair® plasma 3M8239. It is evident from Figure 5-2 that the polypropylene comparative bond strength parameters can be divided into five distinct levels of strength. Openair® plasma LESA with Betaprime™ 5404A primer is the only member that occupies the highest strength 44 Table 5-1: Nylon data collected and calculated from lap shear testing. Combination Bonded Area (m2) Peak Force (N) Bond Strength (MPa) Ave. Bond Strength (MPa) Standard Deviation (MPa) Failure Type Acid Wash Nylon/Lord 320/322 2.68E-04 3407 12.7 2.77E-04 3428 12.4 12.2 0.6 cohesive 2.78E-04 3181 11.5 Acid Wash Nylon/3M 8239 2.68E-04 3181 11.9 11.6 0.2 cohesive w/ 2.83E-04 3280 11.6 fiber pull 2.78E-04 3199 11.5 ATmaP® Nylon/Lord 320/322 3.04E-04 3433 11.3 2.84E-04 3342 11.8 11.6 0.3 cohesive 2.82E-04 3342 11.9 Openair® Nylon/Lord 320/322 2.53E-04 2877 11.4 2.46E-04 3128 12.7 11.3 1.4 cohesive 2.57E-04 2559 9.9 Surface Act Nylon/Lord 320/322 2.87E-04 3117 10.9 10.2 0.7 cohesive w/ 2.86E-04 2963 10.4 fiber pull 2.88E-04 2732 9.5 Openair® Nylon/3M 8239 2.79E-04 613 2.2 6.3 3.0 mixed 2.81E-04 2382 8.5 2.76E-04 2404 8.7 2.82E-04 1604 5.7 Surface Act Polypro/Lord 7542 w/ Zinc 3.04E-04 700 2.3 2.8 0.7 cohesive w/ 2.99E-04 706 2.4 fiber pull 3.03E-04 1092 3.6 Surface Act Polypro/Lord 320/322 w/ Zinc 2.61E-04 791 3.0 2.5 0.8 cohesive w/ 2.68E-04 513 1.9 fiber pull 2.78E-04 n/a n/a Openair® Nylon/Bondmaster M1315 2.84E-04 428 1.5 2.82E-04 620 2.2 2.4 1.1 adhesive 2.80E-04 1002 3.6 Openair® Polypro/3M 8239 w/ Zinc 2.93E-04 407 1.4 2.2 0.7 adhesive 3.11E-04 910 2.9 3.00E-04 793 2.6 3.18E-04 613 1.9 45 Table 5-2: Polypropylene data collected and calculated from lap shear testing. Combination Bond Area (m2) Peak Load (N) Stress (MPa) Average Stress (MPa) Failure Type Openair® Polypro Dow LESA w/Primer 2.87E-04 3262 11.4 10.0 cohesive w/ fiber pull 2.93E-04 2886 9.9 2.91E-04 2872 9.9 3.00E-04 2668 8.9 Surface Act 481 Polypro Dow LESA 2.94E-04 2210 7.5 2.95E-04 1570 5.3 7.4 adhesive 2.87E-04 2342 8.2 2.88E-04 2531 8.8 Surface Act 482 Polypro Dow LESA w/Primer 2.95E-04 1802 6.1 6.7 adhesive w/ fiber pull 2.79E-04 2175 7.8 2.82E-04 1898 6.7 2.89E-04 1784 6.2 Surface Act 482 Polypro Lord 320/322 2.89E-04 1690 5.8 6.5 cohesive w/ fiber pull 2.82E-04 1985 7.0 2.87E-04 1953 6.8 2.91E-04 1879 6.4 Surface Act 482 Polypro Lord 7542 2.86E-04 1714 6.0 2.82E-04 1647 5.8 6.5 adhesive 2.91E-04 2054 7.0 2.97E-04 2141 7.2 Surface Act 483 Polypro Dow LESA 2.89E-04 2378 8.2 2.78E-04 1982 7.1 6.2 fiber pull 2.80E-04 1463 5.2 2.64E-04 1112 4.2 Surface Act 481 Polypro Dow LESA w/Primer 2.83E-04 1528 5.4 2.87E-04 1751 6.1 6.1 fiber pull 2.84E-04 1919 6.8 Surface Act 482 Polypro LESA 2.87E-04 1496 5.2 5.1 adhesive w/ fiber pull 2.83E-04 1502 5.3 2.83E-04 1338 4.7 2.86E-04 1455 5.1 ATmaP® Polypro Dow LESA w/Primer 2.83E-04 1473 5.2 4.7 adhesive w/ fiber pull 2.87E-04 974 3.4 2.87E-04 1563 5.5 2.97E-04 n/a n/a Openair® Polypro Dow LESA 2.97E-04 1148 3.9 4.3 adhesive w/ fiber pull 3.07E-04 1416 4.6 3.17E-04 1319 4.2 3.28E-04 1513 4.6 Surface Act 483 Polypro Dow LESA w/Primer 3.01E-04 854 2.8 2.91E-04 894 3.1 2.6 adhesive 2.93E-04 510 1.7 Openair® Polypro 3M8239 22..8879EE--0044 225705 01..09 0.9 adhesive 46 Figure 5-2: Comparative bond strength of each specimen tested using the lap shear test. 0 2 4 6 8 10 12 14 Acid Wash Nylon Lord320/322 Acid Wash Nylon 3M8239 ATmaP Flame Nylon Lord320/322 Plasma Nylon Lord320/322 SAT Nylon Lord320/322 Plasma Nylon 3M8239 Plasma Nylon Bondmaster M1315 Openair® Plasma PP LESA w/ Primer SAT-481 PP LESA SAT-482 PP LESA w/ Primer SAT-482 PP Lord 320/322 SAT-482 PP Lord 7542 SAT-483 PP LESA SAT-482 PP LESA ATmaP® Flame PP LESA w/ Primer SAT-481 PP LESA w/ Primer Openair® Plasma PP LESA SAT PP Lord7542 w/ Zinc Stearate SAT-483 PP LESA w/ Primer SAT PP Lord320/322 w/ Zinc Stearate Openair® Plasma PP 3M8239 w/Zinc Stearate Openair® Plasma PP 3M8239 Comparative Bond Strength (MPa) 50 46 47 level. There is then a step down in strength to the next level, which has five members from Surface Activated Technologies 481 LESA to Surface Activated Technologies 483 LESA. The next lower strength level has four members from Surface Activate Technologies 482 LESA to Openair® plasma LESA. The fourth strength level has four members and contains all three previous parameters which were manufactured with the zinc stearate mold release. The fifth and lowest strength level has only one member, Openair® plasma with 3M 8239. 5.5 Summary The highest scoring parameter groups from the previous investigation through spot adhesion testing [2] were tested quantitatively using the lap shear test outlined in ASTM D 3163. Specimens were fabricated for each of these high scoring parameter groups. Peak load was established and average bond strength-referred to as comparative bond strength-was calculated to be used as a comparative measure within this investigation. From the lap shear test results, the previous conclusion that the nylon thermoplastic composite is capable of higher bond strengths than the polypropylene thermoplastic composites was again confirmed. In addition, the epoxy adhesive achieved higher strengths and more consistent results than the other adhesives investigated. A method for polypropylene manufacture for adequate comparative bond strength was achieved. 48 6 FRACTURE TOUGHNESS TESTING 6.1 Introduction Peel and cleavage tests can be useful for characterizing different adhesive and surface treatment properties. Peel testing was not included due to the more compliant adherends to be used. Cleavage testing was performed, due to its relative quickness, expense and ease of manufacture. In this investigation, cleavage specimens are tested according to ASTM D 3433 [32]. Load versus displacement curves were generated for each test and from that the fracture toughness was calculated. 6.2 Background Cleavage tests are designed with an intentional, nonuniform distribution of stress. These tests differ from peel tests in that both adherends are relatively rigid, resulting in an approximately 0º peel angle. While ASTM has standardized several cleavage tests for adhesives, the one most relevant to this proposed study is that described in ASTM D 3433, "Standard Test Method for Fracture Strength in Cleavage of Adhesives in Bonded Metal Joints." This test method uses a longer cleavage specimen composed of two long slender rectangular strips bonded together over part of their length as shown in Figure 6- 1. The adherends require a certain stiffness and strength, so they can bend through a certain angle without failing. Near the end of the unbonded section of the adherends, a 49 Figure 6-1: Flat adherend cleavage specimen for determining fracture toughness flexible wire is attached. These wires are clamped in the grips of the tensile testing machine. As the specimen is subjected to tensile loading, the wires separate the unbonded section, subjecting the bondline to tension and bending. This result of peak force per specimen width is of little use in the design of joints that differ in detail from the specimen and loading configuration. However, this joint may be analyzed with a fracture mechanics perspective. The fracture mechanics approach considers the adhesive joint as a system in which failure or growth of the crack requires that the stresses at the crack tip be sufficient to break bonds, and its analysis involves an energy balance. This means that a crack can grow only if sufficient energy is released from the stress field that will balance with the amount of energy required to create the new crack surface as the fractured region enlarges. The critical value of this energy release rate is referred to by various names, 25.4 mm 356 mm Adhesive a 50 including the adhesive fracture energy, adhesive fracture toughness, and work of adhesion. Here G1c will be used to represent this critical energy release rate from ASTM D 3433 [32]. ( )( ) 2 3 2 2 2 1 4 max 3 EB h G L a h C + = (6.1) where, C G1 is the fracture toughness (J/m2) L(max)= load to start the crack (N) E = tensile modulus of adherend (MPa) B = specimen width (mm) a = crack length (mm) h = thickness of adherend, normal to plane of bonding (mm) The word "adhesion" is dropped from the comparable term when considering cohesive failure. 6.3 Methods For cleavage testing, the test specimen geometry was selected in accordance with ASTM D 3433. The manufacturing procedure of the thermoplastic plates outlined in Chapter 3 was followed. An acid wash was performed prior to all surface treatments, in an attempt to remove additional contaminants. The specimens were 8 plies thick and 25.4 mm (1.0 in) wide. Each specimen was approximately 127 mm (5.0 in) long and the initial crack length, a, was chosen to be 12.3 mm (0.5 in). Bond thickness was controlled by placing 0.51 mm (0.020 in) thick wires in the bond area. Fixtures were machined to hold the specimen plates at the proper places with respect to each other and pins were 51 used to ensure proper placement. Plates at least 127 mm (5 in) wide were bonded for each parameter group and all samples tested were cut from the same larger plate. The plates were held in the fixtures for at least 24 hours which at least doubled the hardening time for each adhesive. After removal from the fixtures, the specimens were cured at room temperature for at least 48 hours before they were cut, ensuring that a complete and consistent cure was achieved. A minimum of three specimens were tested for each test parameter group. In several cases, additional specimens were tested if one of the other specimens was damaged when handled or loaded into the test frame. Specimens were loaded into an Instron 4303 tensile testing frame with a 5 kip load cell. A constant displacement rate of 12.7 mm/min (0.50 in/min) was used. As the crack propagated along the length of the specimen, crack length was noted and associated with the load versus displacement curves to calculate G1C according to ASTM D 3433 using Equation 6.1. 6.4 Results and Discussion The data from the cleavage testing were more involved than the lap shear testing for quantifying the G1c of the particular adherend adhesive combinations. Each sample represents three to five G1c calculations along its length as the crack propagated. Coupling this with the three specimens per combination results in a total of 9 to 15 G1c calculations per combination. The data for nylon are presented in Table 6-1. The data were also compiled into Figure 6-2 for easier graphical reference. Nylon had overall higher G1c values than polypropylene as seen in Figure 6-2. This result was expected due to the higher comparative bond strengths achieved with nylon. Within the nylon adherends, the 3M 52 8239 urethane based adhesive outperformed the Lord 320/322 epoxy based adhesive. While both adhesives achieved excellent comparative bond strengths, the urethane adhesive had a much greater fracture toughness over the epoxy adhesives which are generally more brittle. The data for polypropylene are presented in Table 6-2. Polypropylene was outperformed by nylon due to the reduced bond strength between the adhesive and the adherend. Polypropylene failures were mainly between the adhesive and the adherend and were not cohesive failures within the adhesive itself. The fracture energies measured were mainly between the polypropylene adherend and the adhesive. Table 6-1: Nylon fracture toughness data. Combination Sample G1c (J/m2) Std Deviation (J/m2) Avg G1c (J/m2) Openair® Nylon 3M 8239 1 554.44 446.78 2 1372.79 902.64 1223.40 3 1742.96 682.22 Surface Act. Nylon Lord 320/322 1 912.41 317.25 2 869.57 196.19 878.67 3 854.04 119.60 ATmaP® Nylon Lord 320/322 1 787.69 117.60 2 812.19 46.98 852.45 3 957.46 116.66 Acid Wash Nylon Lord 320/322 1 774.07 69.71 2 834.48 101.67 834.29 3 894.32 141.41 Openair® Nylon Lord 320/322 1 567.40 63.46 2 707.60 101.50 762.92 3 1013.77 244.10 53 Table 6-2: Polypropylene fracture toughness data. Combination Sample G1c (J/m2) Std Deviation (J/m2) Avg G1c (J/m2) Surface Act. 482 Polypro Lord 7542 1 265.66 58.84 2 611.55 172.86 462.04 3 638.67 150.52 4 332.28 132.04 Surface Act. 482 Polypro Lord 320/322 1 340.52 75.95 2 500.99 130.61 434.20 3 544.89 98.84 4 350.40 61.86 Openair® Polypro Dow LESA w/Primer 1 205.10 42.73 2 340.35 122.60 378.74 3 590.76 206.58 Surface Act. 483 Polypro Dow LESA 1 187.94 38.79 2 666.69 166.05 357.79 3 305.60 31.53 4 270.92 74.89 Surface Act. 482 Polypro Dow LESA 1 314.61 76.80 2 324.76 46.73 318.39 3 206.56 82.30 4 427.62 355.83 Surface Act. 482 Polypro Dow LESA w/Primer 1 290.49 66.00 2 361.65 148.39 298.73 3 244.05 95.63 Surface Act. 481 Polypro Dow LESA w/Primer 1 161.53 95.73 2 264.29 94.52 218.94 3 230.99 27.76 Surface Act. 483 Polypro Dow LESA w/Primer 1 204.34 30.60 2 210.04 99.26 218.44 3 240.93 43.03 Openair® Polypro 3M 8239 1 179.13 96.10 2 218.15 7.96 196.99 3 193.67 95.09 Surface Act. 481 Polypro Dow LESA 1 128.40 27.24 2 231.01 95.95 159.57 3 115.15 37.83 4 163.73 15.34 Openair® Polypro Dow LESA 1 57.04 9.31 2 124.72 38.29 93.65 3 80.22 12.47 4 112.62 35.70 57 Figure 6-2 : Fracture toughness energies for all tested parameters 0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 Openair® Plasma Nylon 3M 8239 Surface Activation Technologies Nylon Lord 320/322 ATmaP® Flame Nylon Lord 320/322 Acid Wash Nylon Lord 320/322 Openair® Plasma Nylon Lord 320/322 Surface Activation Technologies 482 Polypro Lord 7542 Surface Activation Technologies 482 Polypro Lord 320/322 Openair® Plasma Polypro Dow LESA w/Betaprime™ 5404A Surface Activation Technologies 483 Polypro Dow LESA Surface Activation Technologies 482 Polypro Dow LESA Surface Activation Technologies 482 Polypro Dow LESA w/Betaprime™ 5404A Surface Activation Technologies 481 Polypro Dow LESA w/Betaprime™ 5404A Surface Activation Technologies 483 Polypro Dow LESA w/Betaprime™ 5404A Openair® Plasma Polypro 3M 8239 Surface Activation Technologies 481 Polypro Dow LESA Openair® Plasma Polypro Dow LESA G1c (J/m2) 54 55 6.5 Summary The highest scoring parameter groups from the previous investigation through spot adhesion testing [2] were tested quantitatively using the cleavage test outlined in ASTM D 3433. Specimens were fabricated for each of these high scoring parameter groups. The fracture toughness of these specimens was calculated to be used as a comparative measure within this investigation. From the cleavage test results, the previous conclusion that the nylon thermoplastic composite is capable of higher bond strengths than the polypropylene thermoplastics composites was again confirmed. In addition, the urethane adhesive had higher fracture toughness than did the epoxy adhesive. 56 7 ENVIRONMENTAL TESTING 7.1 Intoduction In order to determine the effectiveness of different adhesive, surface treatment and aherend combinations under actual working conditions, it was deemed necessary to expose them to various environmental conditions. Input was sought from the ACC, and the acceptable environmental parameters were selected. 7.2 Background The top performing adherend adhesive combinations of acid washed nylon thermoplastic with Lord 320/322 adhesive and Openair® plasma polypropylene with Dow LESA and BetaprimeTM 5404A primer were selected for the additional environmental testing. The three environmental parameters decided upon for the environmental tests were a hot test of 80ºC (176ºF), a cold test of -40ºC (-40ºF), and a hot/wet test of 50ºC (122ºF) and 85% relative humidity. 7.3 Methods For lap shear testing, the test specimen geometry was selected in accordance with ASTM D 3163. The manufacturing procedure of the thermoplastic plates outlined in Chapter 3 was followed. An acid wash was performed prior to all surface treatments, in an attempt to remove additional contaminants. The specimens were 4 plies thick and 25.4 mm (1.0 in) wide. Each specimen was approximately 101.6 mm (4.0 in) long and a bond length, L, was chosen to be 12.3 mm (0.5 in). Bond thickness was controlled by 57 placing 0.51 mm (0.020 in) thick wires in the bond area. Fixtures were machined to hold the specimen plates at the proper places with respect to each other and pins were used to ensure proper placement. Plates at least 127 mm (5.0 in) wide were bonded for each parameter group and all samples tested were cut from the same larger plate. The plates were held in the fixtures for at least 12 hours which at least doubled the hardening time for each adhesive. After removal from the fixtures, the specimens were cured at room temperature for at least 48 hours before they were cut, ensuring that a complete and consistent cure was achieved. The methods for the hot and cold tests are analogous except for the temperatures implemented. All specimens to be tested for the respective test were placed in the base of the test chamber. This was to ensure the specimens would be at the same temperature as the test chamber. The test chamber was turned on and set to the proper temperature. A thermocouple was placed inside the test chamber to ensure the proper temperature was reached. After the proper temperature had been reached, the test chamber was opened and a specimen was loaded into the grips. The test chamber was closed again and a 5 min. dwell time was used to allow the test chamber to return to the proper temperature. The hot/wet test consisted of a exposing the specimens to a 50ºC (122ºF) and 85% relative humidity environment until the specimens stopped gaining weight. Before testing, they were placed in a 23ºC (73ºF) and 50% relative humidity environment for one hour to "dry out" then tested at room temperature. A minimum of three specimens were tested for each test parameter group. In several cases, additional specimens were tested if one of the other specimens was damaged when handled or loaded into the test frame. Specimens were loaded into an Instron 4303 58 tensile testing frame with a 5 kip load cell. A constant displacement rate of 1.27 mm/min (0.050 in/min) was used. Load versus displacement curves, bond area, peak load, and failure type were noted for each specimen tested. The bond area and peak load were used to calculate comparative bond strength for each specimen and average stress for each parameter group. These bond strengths were then used to compare the different parameter groups and assess which parameter group achieved highest bond strength. 7.4 Results and Discussion 7.4.1 Acid Washed Nylon Lord 320/322 A graphical comparison of the different environmental parameters is shown in Figure 7-1. The parameter group of acid washed nylon thermoplastic adherends with Lord 320/322 adhesive at -40ºC (-40ºF) had an average comparative bond strength of 12.0 MPa (1.7 ksi). The failure modes at -40ºC (-40ºF) were primarily cohesive and fiber/matrix failures. At 80ºC (176ºF), the specimen had an average comparative bond strength of 1.8 MPa (0.27 ksi). The failure mode at 80ºC (176ºF) was cohesive. The adhesive at the elevated temperature had been softened and was no longer structurally sound, as evidenced by the cohesive failure with a very low strength. A cohesive failure is representative of achieving the highest bond strength to the adherend. At the 50ºC (122ºF) and 85% relative humidity, the specimen had an average comparative bond strength of 7.7 MPa (1.1 ksi). The reduction in strength for the hot/wet test suggests there is some reduction in strength due to temperature and humidity. 59 Figure 7-1: Comparative bond strength for acid wash nylon with Lord 320/322 adhesive under different environmental conditions. 7.4.2 Openair® Plasma Polypropylene Dow LESA with Betamate™ 5404A Primer A graphical comparison of the different environmental parameters is shown in Figure 7-2. The parameter group of Openair® plasma polypropylene thermoplastic adherends with Dow LESA adhesive with Betamate™ 5404A primer at -40ºC (-40ºF) had an average comparative bond strength of 5.9 MPa (0.86 ksi). The failure modes observed at -40ºC (-40ºF) were adhesive, cohesive and fiber/matrix interface failures. At 80ºC (176ºF), the specimen had an average comparative bond strength of 1.9 MPa (0.27 ksi). The failure modes at 80ºC were adhesive and fiber/matrix interface failures. At the 50ºC (122ºF) and 85% relative humidity, the specimen had an average comparative bond strength of 3.9 MPa (0.57 ksi). 0 2 4 6 8 10 12 14 Acid Wash PA6 Lord320/322 Acid Wash PA6 Lord 320/322 -40⁰C Acid Wash PA6 Lord 320/322 80⁰C Acid Wash PA6 Lord 320/322 50⁰C 85% RH Comparative Bond Strength (MPa) 60 Figure 7-2: Comparative bond strength for Openair® plasma polypropylene with Dow LESA with Betamate 5404A primer under different environmental conditions. 0 2 4 6 8 10 12 Plasma PP LESA w/ Primer Plasma PP LESA w/ Primer - 40⁰C Plasma PP LESA w/ Primer 80⁰C Plasma PP LESA w/ Primer 50⁰C 85% RH Comparative Bond Strength (MPa) 61 8 FIBER REINFORCEMENT TESTING 8.1 Introduction In order to determine the effectiveness of different adhesive, surface treatment and adherend combinations, an investigation of the fiber reinforcement was conducted. Input was sought from the ACC, and the best performing combinations with carbon reinforcement were selected for further testing with glass reinforcement. 8.2 Background The top performing adherend/adhesive combinations for nylon were selected as acid washed with Lord 320/322 and 3M 8239. The top performing adherend adhesive combinations for polypropylene were selected as Openair® plasma with Dow LESA and BetaprimeTM 5404A primer and Surface Activation Technologies 482 with Dow LESA and BetaprimeTM 5404A primer. 8.3 Methods For lap shear testing, the test specimen geometry was selected in accordance with ASTM D 3163. The manufacturing procedure of the thermoplastic plates outlined in Chapter 3 was followed. An acid wash was performed prior to all surface treatments, in an attempt to remove additional contaminants. The specimens were 4 plies thick and 25.4 mm (1.0 in) wide. Each specimen was approximately 101.6 mm (4.0 in) long and a bond length, L, was chosen to be 12.3 mm (0.5 in). Bond thickness was controlled by placing 0.51 mm (0.020 in) thick wires in the bond area. Fixtures were machined to hold 62 the specimen plates at the proper places with respect to each other and pins were used to ensure proper placement. Plates at least 127 mm (5.0 in) wide were bonded for each parameter group and all samples tested were cut from the same larger plate. The plates were held in the fixtures for at least 12 hours which at least doubled the hardening time for each adhesive. After removal from the fixtures, the specimens were cured at room temperature for at least 48 hours before they were cut, ensuring that complete and consistent cure was achieved. A minimum of three specimens were tested for each test parameter group. In several cases, additional specimens were tested if one of the other specimens was damaged when handled or loaded into the test frame. Specimens were loaded into an Instron 4303 tensile testing frame with a 5 kip load cell. A constant displacement rate of 1.27 mm/min (0.050 in/min) was used. Load versus displacement curves, bond area, peak load, and failure type were noted for each specimen tested. The bond area and peak load were used to calculate comparative bond strength for each specimen and average stress for each parameter group. These bond strengths were then used to compare the different parameter groups and assess which parameter group achieved highest bond strength. The calculated strengths are not intended for use in design, but only as a reference to compare the different parameter groups within this investigation. 8.4 Results and Discussion 8.4.1 Carbon to Glass Reinforcement Comparison for Nylon Thermoplastic There was a decrease in strength found in both cases with nylon thermoplastic in the change from carbon reinforcement to glass reinforcement (Figure 8-1). The failure 63 Figure 8-1: Comparative bond strength for nylon glass carbon comparison modes also changed from a cohesive failure to a fiber matrix interface failure. These results suggest that the nylon thermoplastic does not adhere to the glass reinforcement as well as it adheres to the carbon reinforcement. The parameter group of acid washed nylon thermoplastic adherends with Lord 320/322 adhesive had an average comparative bond strength of 8.1 MPa (1.2 ksi). The failure mode was primarily fiber/matrix interface failure. The parameter group of acid washed nylon thermoplastic adherends with 3M8239 adhesive had an average comparative bond strength of 8.9 MPa (1.3 ksi). The failure mode was primarily fiber/matrix interface failure. 0 2 4 6 8 10 12 14 Glass Acid Wash Nylon Lord320/322 Carbon Acid Wash Nylon Lord320/322 Glass Acid Wash Nylon 3M8239 Carbon Acid Wash Nylon 3M8239 Comparative Bond Strength (MPa) 64 8.4.2 Carbon to Glass Reinforcement Comparison for Polypropylene Thermoplastic There was a decrease in strength found in both cases with polypropylene thermoplastic in the change from carbon reinforcement to glass reinforcement (Figure 8- 2). The failure modes also changed from a mixed failure with adhesive, cohesive, and fiber/matrix interface failures to a pure fiber/matrix interface failure. These results suggest that the polypropylene thermoplastic does not adhere to the glass reinforcement as well as it adheres to the carbon reinforcement. The parameter group of Openair plasma polypropylene reinforced glass thermoplastic adherends with Dow LESA adhesive with Betaprime primer had an average comparative bond strength of 6.6 MPa (0.96 ksi). The failure mode was primarily fiber/matrix interface failure. The parameter group of Surface Activation Technologies 482 polypropylene reinforced glass thermoplastic adherends with Dow LESA adhesive with Betaprime primer had an average comparative bond strength of 4.9 MPa (0.71 ksi). The failure mode was primarily fiber/matrix interface failure. 65 Figure 8-2: Comparative bond strength for polypropylene glass carbon comparison 0 2 4 6 8 10 12 Glass Plasma PP LESA w/Primer Carbon Plasma PP LESA w/ Primer Glass SAT-482 PP LESA w/Primer Carbon SAT-482 PP LESA w/ Primer Comparative Bond Strength (MPa) 66 9 ROUND ONE CONCLUSIONS The most promising adhesive, surface treatment, and adherend combinations are shown in Table 9-1 for nylon and Table 9-2 for polypropylene. A greater weighting was given to the lap shear strengths than to the fracture toughness. For nylon, input from the ACC was sought to determine the best selections for follow-on testing. The four combinations shown in bold in Table 9-1 were selected. An adhesive comparison between Lord 320/322 and 3M 8239 could be made across the common surface treatment of acid wash. A surface treatment comparison between acid wash, ATmaP Flame, and Openair Plasma could be made across the common adhesive Lord 320/322. For polypropylene, input was also sought from the ACC to determine the best selections for follow-on testing. The four combinations shown in bold in Table 9-2 were selected. An adhesive comparison between Lord 320/322, 3M 8239 and Dow LESA with Betamate 5404A could be made across the common surface treatment of Surface Activation Technologies 482. A surface treatment comparison between Openair Plasma and Surface Activation Technologies 482 could be made across the common adhesive Dow LESA with Betamate 5404A primer. 67 Table 9-1: Selection matrix for nylon candidates in bold for continuation of testing Average of Test Results Lap Shear (MPa) G1C (J/m2) Acid Wash Nylon Lord320/322 12.18 834 Acid Wash Nylon 3M8239 11.65 N/A ATmaP Flame Nylon Lord320/322 11.64 852 Plasma Nylon Lord320/322 11.33 763 SAT Nylon Lord320/322 10.23 879 Plasma Nylon 3M8239 6.26 1223 Plasma Nylon Bondmaster M1315 2.43 N/A Table 9-2: Selection matrix for polypropylene candidates in bold for continuation of testing Average of Test Results Lap Shear (MPa) G1C (in-lbs/ in2) Plasma PP LESA w/ Primer 10.00 2.16 SAT-481 PP LESA 7.44 0.91 SAT-482 PP LESA w/ Primer 6.71 1.71 SAT-482 PP Lord 320/322 6.54 2.48 SAT-482 PP Lord 7542 6.52 2.64 SAT-483 PP LESA 6.20 2.04 SAT-482 PP LESA 5.09 1.82 ATmaP Flame PP LESA w/ Primer 4.68 N/A SAT-481 PP LESA w/ Primer 4.57 1.25 Plasma PP LESA 4.31 0.50 SAT-482 PP Lord7542 w/ Zinc Stearate 2.76 N/A SAT-483 PP LESA w/ Primer 2.55 1.25 SAT-482 PP Lord320/322 w/ Zinc Stearate 2.47 N/A Plasma PP 3M8239 w/ Zinc Stearate 2.22 N/A Plasma PP 3M8239 0.91 1.12 68 10 FLATWISE TENSION 10.1 Introduction Flatwise tensile testing may be used to assess the as-manufactured performance of the bond between the facesheets and core. A square specimen is adhesively bonded to pin-loaded steel blocks. The primary functions of this test are to assess the bond strength under an out-of-plane tensile loading and to identify the weakest interface within the sandwich composite. In this investigation, flatwise tensile specimens were tested according to ASTM C 297 [18]. The flatwise tensile strength, failure location, and the failure mode were determined from these tests. This test is also useful for identifying manufacturing related problems such as core crushing, poor adhesive bonding, and improper facesheet consolidation. 10.2 Background Originally it was envisioned that a larger sandwich panel would be fabricated and the smaller flatwise tension specimens would be cut from this larger sandwich panel. The smaller flatwise tension specimens would then be bonded to the steel fixture blocks. Due to concerns over contamination and fade of the surface treatments on the thermoplastic facesheets, a method for simultaneously bonding the facesheets, core, and fixture blocks was devised, which included a specially manufactured alignment fixture. With both sides of each facesheet having the same surface treatment and the same time amount of time until bonding, the concerns of fade and contamination were minimized. It was also 69 shown in preliminary testing that the facesheet/fixture block bond was just as susceptible to failure as the facesheet/core bond. With both sides of the facesheet having the same surface treatment and bond time, a failure on either side of the facesheet could be considered possible. Preliminary testing also showed that the strength of the balsa core/facesheet bond was stronger than the facesheet/fixture block bond. However, this was not an issue with the foam core because of the reduced strength of foam as compared to balsa. Therefore, the balsa core was reduced from a 51 mm by 51 mm (2 in x 2 in) cross section to a 38 mm by 38 mm (1.5 in x 1.5 in) cross section which resulted in a 43% reduction in cross sectional area while the facesheets remained full size. This change was intended to eliminate failures at the facesheet/fixture block bond, and try to force the specimen to fail in the core or the facesheet/core interface. 10.3 Methods Flatwise tension testing specimen geometry was chosen in accordance with ASTM C297 [18]. The specimen facesheets were 4 plies thick and 51 mm (2.0 in) by 51 mm (2.0 in). The dimensions of the foam cores were 51 mm (2.0 in) by 51 mm (2.0 in) and 12.7 mm (0.50 in) thick. The dimensions of the balsa cores were 38 mm (1.5 in) by 38 mm (1.5 in) and 12.2 mm (0.48 in) thick. The steel fixture blocks were prepared by grit blasting followed by an acetone wipe to remove contaminants. The fixture blocks were masked with duct tape to prevent the adhesive from adhering to the tool. The core was masked with duct tape to prevent the adhesive from adhering to the core and affecting results. Due to the reduction in size for the balsa core, a 1.1 mm (0.04 in) hole was drilled in the center of each facesheet, and an alignment pin was installed. The alignment pin pressed into the reduced balsa core and 70 prevented its motion during the adhesive bonding process. The alignment fixture was used to check the alignment of the facesheets, reduced balsa core, and fixture blocks, before the inside of the facesheets around the outside of the reduced balsa core were masked with duct tape. During the bonding process, each surface received an even coating of adhesive, and two spacing wires were placed in each bond line before the surfaces were pressed together. As each piece of the specimen was completed during the bonding process, it was placed in the fixture and alignment was checked at each step. When the specimen had been bonded to the fixture blocks, a steel plate was placed on the specimen and clamped to a solid flat surface with a C-clamp. A C-clamp was used on both ends of the fixture blocks to squeeze the bond lines together. The C-clamp was incrementally tightened to supply sufficient pressure with care being taken to not crush the core. The specimens were left in the fixtures for 24 hours until adequate adhesive strength had been reached for their removal. The tape mask was removed from the tools and core. The excess adhesive was removed with a dremel or a razor blade. A total of three to six specimens were tested for each test parameter group. Specimens were tested at a constant displacement of 1.3 mm/min (0.05 in/min). The peak load was measured and the flatwise tensile strength was calculated according to ASTM C 297 [18]. Failure mode was also noted for each specimen. 10.4 Results and Discussion The different parameters that were investigated and the resulting test data are summarized in Table 10-1 and can be viewed graphically in Figure 10-1. Nylon adherends with balsa core were the highest performers. The bond strengths with nylon 71 were sufficient to produce cohesive failures in the adhesive and core failures in the balsa. In fact the flatwise tension strengths of nylon with balsa were comparable to previous tests on thermoset sandwich composites as seen in Figure 10-1. The bond strength with polypropylene was not high enough to achieve a failure in the balsa core, but was strong enough to achieve failure in the foam core. The strength of the foam core was not affected by either adherend, nylon or polypropylene, because the bond strengths were strong enough to achieve core failure in every combination. It was discovered that the 3M 8239 adhesive was incompatible with the balsa core. 3M 8239 is a fast cure with a work time of approximately 1 min. This fast cure time prevents the adhesive from wetting out the balsa core and achieving good mechanical interlocking. 10.5 Summary The recommended parameter groups from lap shear and cleavage testing were used to produce sandwich panels with two different core materials, balsa and polyurethane foam. These sandwiches were tested in flatwise tensile according to ASTM C 297 [18]. A special method of manufacture was devised to provide better adhesion to the steel fixture blocks. The nylon adherends with the balsa core were the highest scoring group. Adhesion strength with the polypropylene adherends was not sufficient to cause failure in the balsa core but was sufficient to cause failure in the foam core. The adhesive, 3M 8239, was found to be incompatible with the balsa and foam core. 72 Table 10-1: Flatwise tension results Sandwich Configuration Stress (MPa) Std. Dev. (MPa) Failure Mode Openair® Plasma Nylon Lord 320/322 Balsa 7.39 0.92 Core Cohesive ATmaP® Flame Nylon Lord 320/322 Balsa 5.88 0.50 Core Cohesive Acid Wash Nylon Lord 320/322 Balsa 5.80 0.57 Core Cohesive Surface Act Polypro Lord 320/322 Balsa 2.14 0.15 Mixed Acid Wash Nylon 3M 8239 Balsa 2.11 0.48 Adhesive ATmaP® Flame Nylon Lord 320/322 Polyurethane Foam 1.82 0.13 Core Surface Act Polypro Lord 320/322 Foam 1.61 0.10 Core Openair® Plasma Polypro Dow Lesa w/ Primer Balsa 1.57 0.57 Adhesive Acid Wash Nylon 3M 8239 Polyurethane Foam 1.54 0.14 Core Acid Wash Nylon Lord 320/322 Polyurethane Foam 1.54 0.29 Core Openair® Plasma Nylon Lord 320/322 Polyurethane Foam 1.52 0.12 Core Surface Act Polypro Lord 7542 Foam 1.49 0.10 Core Openair® Plasma Polypro Dow Lesa w/ Primer Foam 1.06 0.23 Core Surface Act Polypro Lord 7542 Balsa 1.02 1.24 Adhesive 73 Figure 10-1 : Flatwise tension test data 0 1 2 3 4 5 6 7 8 9 10 Openair® Plasma Nylon Lord 320/322 Balsa ATmaP® Flame Nylon Lord 320/322 Balsa Acid Wash Nylon Lord 320/322 Balsa Surface Act Polypro Lord 320/322 Balsa Acid Wash Nylon 3M 8239 Balsa ATmaP® Flame Nylon Lord 320/322 Polyurethane Foam Surface Act Polypro Lord 320/322 Foam Openair® Plasma Polypro Dow LESA w/ Primer Balsa Acid Wash Nylon 3M 8239 Polyurethane Foam Acid Wash Nylon Lord 320/322 Polyurethane Foam Openair® Plasma Nylon Lord 320/322 Polyurethane Foam Surface Act Polypro Lord 7542 Foam Openair® Plasma Polypro Dow LESA w/ Primer Foam Surface Act Polypro Lord 7542 Balsa Woven Carbon-Epoxy Balsa P4 Carbon-Epoxy Balsa Woven Carbon-Epoxy Polyurethane Foam P4 Carbon-Epoxy Polyurethane Foam Stress (MPa) 73 74 11 CORE SHEAR 11.1 Introduction Core shear testing may be used to assess the as-manufactured performance of the bond between the facesheets and core. The primary functions of this test are to assess the bond strength under shear loading and to identify the weakest interface within the sandwich composite. In this investigation, core shear specimens were tested according to ASTM C 297 [19]. The core shear strength, failure location, and failure mode were determined. This test is also useful for identifying manufacturing related problems such as core crushing, poor adhesive bonding, and improper facesheet consolidation. 11.2 Background Originally it was envisioned that a larger sandwich panel would be fabricated and the smaller core shear test specimens would be cut from this larger sandwich panel. The smaller core shear specimens would then be bonded to the steel fixture plates. Due to concerns over contamination of the specially surface treated thermoplastic facesheets, a method for simultaneously bonding the facesheets, core, and fixture plates was devised, which included a specially manufactured alignment fixture. With both sides of each facesheet having the same surface treatment and the same amount of time until bonding, the concerns of fade and contamination were minimized. It was also shown in preliminary testing that the facesheet/fixture plate bond was just as susceptible to failure as the facesheet/core bond. With both sides of the facesheet having the same surface 75 treatment and bond time, a failure on either side of the facesheet could be possible. Preliminary testing also showed that the strength of the balsa core/facesheet bond was stronger than the facesheet/fixture plate bond. However, this was not an issue with the foam core because of the reduced strength of foam as compared to balsa. Therefore, the balsa core was reduced in length by 12.2 mm (0.5 in) while the facesheets remained the full size. This change was intended to eliminate failures at the facesheet/fixture plate interface, and to try to force the specimen to fail in the core or the facesheet/core interface. The alignment fixture controlled the facesheets and core alignment, while allowing control of the bond line thickness by placing wires within the bond line. 11.3 Methods Core shear testing specimen geometry was chosen in accordance with ASTM C273 [19]. The manufacturing procedure of the thermoplastic plates outlined in Chapter 3 was followed. An acid wash was performed prior to all surface treatments, in an attempt to remove additional contaminants. The specimen facesheets were 4 plies thick and 51 mm (2.0 in) wide, and 203 mm (8.0 in) long. The dimensions of the foam cores were 203 mm (8.0 in) long, 51 mm (2.0 in) wide, and 12.7 mm (0.50 in) thick. The dimensions of the balsa cores were 191 mm (7.5 in) long, 51 mm (2.0 in) wide, and 12.2 mm (0.48 in) thick. The steel fixture plates were prepared by grit blasting followed by an acetone wipe to remove contaminants. The steel fixture plates were masked with duct tape to prevent the adhesive from adhering. The core was masked with duct tape to prevent the adhesive from adhering to the core and affecting results. Due to the reduction in size for the balsa core, the alignment of the facesheet, reduced balsa core, and fixture plates were checked 76 in the alignment fixture prior to bonding. The inside of the facesheets around the outside of the reduced balsa core was masked with duct tape. During the bonding process, each surface received an even coating of adhesive, and three spacing wires were placed in each bond line before the surfaces were pressed together. As each piece of the specimen was completed during the bonding process, it was placed in the alignment fixture and alignment was checked at each step. When the specimen had been bonded to the fixture plates, a steel plate was placed on the specimen and clamped to a solid flat surface with a C-clamp. A C-clamp was used on the side of the alignment fixture and on the steel fixture plates to squeeze the bond lines together. The C-clamp was incrementally tightened to supply sufficient pressure with care being taken to not crush the core. The specimens were left in the fixtures for 24 hours until adequate strength had been reached for their removal. The tape mask was removed from the fixture plates and core. The excess adhesive was removed with a dremel or a razor blade. A total of three to six specimens were tested for each test parameter group. Specimens were tested at a constant displacement rate of 1.3 mm/min (0.05 in/min). The peak load was measured and the core shear strength was calculated according to ASTM C 273 [19]. The failure mode was also noted for each specimen. 11.4 Results and Discussion The different parameters that were investigated and the resulting data are summarized in Table 11-1 and can be viewed graphically in Figure 11-1. There was no apparent difference between the two different thermoplastic adherends, nylon and polypropylene, in shear as shown in Figure 11-1. The thermoset sandwich composites 77 Table 11-1: Core shear results Sandwich Configuration Average Strength (MPa) Std. Dev. (MPa) Failure Mode Surface Act Polypro Lord 7542 Balsa 1.68 0.07 Core Adhesive ATmaP® Flame Nylon Lord 320/322 Balsa 1.68 0.29 Cohesive Openair® Plasma Nylon Lord 320/322 Balsa 1.60 0.14 Core Cohesive Surface Act Polypro Lord 320/322 Balsa 1.51 0.10 Core Cohesive Acid Wash Nylon Lord 320/322 Balsa 1.50 0.17 Core Cohesive Openair® Plasma Polypro Dow Lesa w/ Primer Balsa 1.30 0.22 Mixed Acid Wash Nylon 3M 8239 Balsa 1.15 0.18 Adhesive Openair® Plasma Polypro Dow Lesa w/ Primer Foam 0.84 0.08 Core Surface Act Polypro Lord 7542 Foam 0.79 0.02 Core Acid Wash Nylon Lord 320/322 Polyurethane Foam 0.76 0.04 Core ATmaP® Flame Nylon Lord 320/322 Poluyurethane Foam 0.76 0.06 Core Acid Wash Nylon 3M 8239 Polyurethane Foam 0.73 0.10 Core Surface Act Polypro Lord 320/322 Foam 0.72 0.07 Core Openair® Plasma Nylon Lord 320/322 Polyurethane Foam 0.68 0.05 Core 78 Figure 11-1 : Overall core shear results 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Surface Act Polypro Lord 7542 Balsa ATmaP® Flame Nylon Lord 320/322 Balsa Openair® Plasma Nylon Lord 320/322 Balsa Surface Act Polypro Lord 320/322 Balsa Acid Wash Nylon Lord 320/322 Balsa Openair® Plasma Polypro Dow LESA w/ Primer Balsa Acid Wash Nylon 3M 8239 Balsa Openair® Plasma Polypro Dow LESA w/ Primer Foam Surface Act Polypro Lord 7542 Foam Acid Wash Nylon Lord 320/322 Polyurethane Foam ATmaP® Flame Nylon Lord 320/322 Poluyurethane Foam Acid Wash Nylon 3M 8239 Polyurethane Foam Surface Act Polypro Lord 320/322 Foam Openair® Plasma Nylon Lord 320/322 Polyurethane Foam Woven Carbon-Epoxy Balsa P4 Carbon-Epoxy Balsa Woven Carbon-Epoxy Polyurethane Foam P4 Carbon-Epoxy Polyurethane Foam Stress (MPa) 78 79 made with the balsa core did not perform in a statistically different manner from each other with the exception of the 3M 8239 adhesive. As reported in the flatwise tensile testing, there was an incompatibility between the balsa and the 3M 8239 adhesive. This incompatibility seemed to be less of an issue in shear, but its effects were still apparent. The thermoplastic sandwich composites with the balsa core did not score as high as the previous thermoset sandwich composites. The thermoplastic sandwich composites with the foam core did not produce statistically significant differences in core shear strengths from each other. The thermoplastic sandwich composites with the foam core scored comparatively with the previous thermoset sandwich composites. 11.5 Summary The recommended parameter groups from lap shear and cleavage testing were used to produce composite sandwich panels with two different core materials, balsa and polyurethane foam. These composite sandwich configurations were tested using a core shear test configuration according to ASTM C 273 [19]. A special method of manufacture was devised to provide better adhesion to the steel fixture plates. The core shear strength was found to be more dependent on the core than the adherend. The balsa core outperformed the foam core regardless of the adherend, nylon or polypropylene. 80 12 STATIC EDGEWISE COMPRESSION TESTING 12.1 Introduction Edgewise compression testing is commonly used to assess the performance of sandwich composites under in-plane compression loading. Although not a direct test of the adhesion of the facesheet to the core, it does provide a good indication of adhesion as well as evaluating failure produced in an end impact. The primary outputs of this test are the facing compression stress at failure and the failure modes. The energy absorption may be obtained from integrating the area under the load versus displacement curve. In this investigation, edgewise compression specimens were tested according to a modified ASTM C 364 [20]. 12.2 Background According to ASTM C 364 [20], the edgewise specimen should be supported over 13 mm (0.51 in.) of length adjacent to the specimen ends. In this study, the edgewise compression fixture was modified so as to only provide clamping on the bottom 6 mm (0.24 in.) and no clamping on the top of the specimen as shown in Figure 12-1. The top and bottom of the fixture had side clamps with a 45° angled surface. These angled side clamps provided realignment of the specimen in the fixture during crushing. Specimen alignment was maintained using alignment rods and linear bearings in the top portion of 81 Figure 12-1: Specially designed static edgewise compression testing fixture. the fixture. Hardened steel contact plates were pressed into the top and bottom specimen contact areas of the fixture to prevent damage due to localized loading. Lateral restraint bolts were also fitted on both the top and bottom clamps to prevent them from spreading during loading. 12.3 Methods For static edgewise compression testing, the specimen geometry was chosen in accordance with ASTM C 364 [20]. The manufacturing procedure of the thermoplastic plates outlined in Chapter 3 was again followed. An acid wash was performed prior to all surface treatments in an attempt to remove additional contaminants. Composite sandwich manufacturing went as follows. Facesheets four plies thick and approximately 140 mm by 254 mm (5.5 in by 10 in) were treated with their respective surface treatment. Core material, balsa or foam, was cut to size on a band saw. Bond line thickness wires were bent and pressed into the core material on each end. During the Lateral Restraint Bolts Linear Guide Rod Specimen Clamp 82 bonding process, each surface received an even coating of adhesive and alignment was checked before the surfaces were pressed together. The composite sandwich was placed in a die set for 24 hours with additional weights to ensure the proper bondline was achieved. After removal from the die set, the sandwich composite panels were allowed to cure at room temperature for a minimum of 72 hours to ensure a complete and consistent cure before being cut. Each specimen was machined to ensure the loading surfaces were parallel. The final specimen geometry was 76 mm by 127 mm (3.0 in x 5.0 in). A total of three to six specimens were tested for each test parameter group. The fixture shown in Figure 12-1 was placed under the crosshead. A specimen was placed between the V-block clamps, and the lateral restraint bolts were tightened to hold the specimen in place. Specimens were tested at a constant displacement rate of 7.6 mm/min (0.3 in/min) for a distance of 50.8 mm (2.0 in). The peak load was measured as well as the load versus displacement recorded. The peak load was used to calculate the edgewise compression strength of the panel, and the load versus displacement curve was integrated to find the total energy absorbed. 12.4 Results and Discussion 12.4.1 Static Average Edgewise Compression Strength The different composite sandwich panels were tested and the average edgewise compression strength was computed and summarized in Tables 12-1 and 12-2 and can be viewed graphically in Figure 12-2. The nylon facesheet sandwich composites had a higher average edgewise compression strength over the polypropylene facesheet sandwich composites. This result 83 Table 12-1: Static nylon edgewise compression strengths Sandwich Configuration Max Load (kN) Ult. Strength (MPa) Initial Failure Mode Average Ult. Strength (MPa) Std. Deviation (MPa) Openair Plasma Nylon Lord 320/322 Balsa 52 266 Facesheet Buckling 265 11 49 249 End Crushing 54 275 Facesheet Buckling 53 271 End Crushing 54 276 End Crushing 50 255 Facesheet Buckling Acid Wash Nylon Lord 320/322 Balsa 48 251 End Crushing 255 21 49 251 End Crushing 56 288 Facesheet Buckling 44 224 End Crushing 50 259 Facesheet Buckling 50 256 Facesheet Buckling ATmaP Flame Nylon Lord 320/322 Foam 46 236 Facesheet Buckling 235 8 44 227 Facesheet Buckling 43 223 Facesheet Buckling 48 247 Facesheet Buckling 46 237 Facesheet Buckling 46 237 Facesheet Buckling Openair Plasma Nylon Lord 320/322 Foam 43 224 Facesheet Buckling 219 8 43 220 Facesheet Buckling 41 210 Facesheet Buckling 45 231 Facesheet Buckling 42 217 Facesheet Buckling 41 212 Facesheet Buckling Acid Wash Nylon Lord 320/322 Foam 40 208 Facesheet Buckling 204 14 40 204 Facesheet Buckling 39 201 Facesheet Buckling 42 215 Facesheet Buckling 42 217 Facesheet Buckling 35 178 Facesheet Buckling 84 Table 12-2: Static polypropylene edgewise compression strengths Sandwich Configuration Max Load (kN) Ult. Strength (MPa) Initial Failure Mode Average Ult. Strength (MPa) Std. Deviation (MPa) Openair Plasma Polypro Dow LESA w/Primer Foam 25 128 End Crushing 118 24 17 89 General Buckling 24 123 General Buckling 29 151 End Crushing 24 125 End Crushing 18 92 End Crushing Openair Plasma Polypro LESA w/Primer Balsa 24 122 Facesheet Buckling 129 12 27 136 End Crushing 26 137 General Buckling 23 120 End Crushing 28 145 Facesheet Buckling 22 114 End Crushing Surface Activation Polypro Lord 320/322 Foam 29 147 End Crushing 27 140 End Crushing 148 9 31 157 General Buckling Surface Activation Polypro Lord 320/322 Balsa 26 134 Facesheet Buckling 29 148 Facesheet Buckling 150 17 33 169 Facesheet Buckling Surface Activation Polypro Lord 7542 Foam 25 126 End Crushing 17 90 General Buckling 114 21 24 125 General Buckling Surface Activation Polypro Lord 7542 Balsa 23 118 End Crushing 22 114 End Crushing 116 2 23 117 Facesheet Buckling 85 Figure 12-2: Static edgewise compression strengths 0 50 100 150 200 250 300 Openair Plasma Nylon Lord 320/322 Balsa Acid Wash Nylon Lord 320/322 Balsa ATmaP Flame Nylon Lord 320/322 Foam Openair Plasma Nylon Lord 320/322 Foam Acid Wash Nylon Lord 320/322 Foam Surface Activation Polypro Lord 320/322 Balsa Surface Activation Polypro Lord 320/322 Foam Openair Plasma Polypro LESA w/Primer Balsa Openair Plasma Polypro Dow LESA w/Primer Foam Surface Activation Polypro Lord 7542 Balsa Surface Activation Polypro Lord 7542 Foam Average Ult. Strength (MPa) 85 86 was expected, since nylon produced a stronger, stiffer facesheet and the edgewise compression strength of the sandwich composite is produced mainly by the facesheets. This result is further evidenced by the fact that nylon facesheets with foam outperformed the polypropylene with the balsa core. Although the edgewise compression strength of the sandwich is mainly produced by the facesheets, the bond between the core and the facesheets also contribute to the overall strength of the sandwich. This is evidenced by the fact that nylon facesheets with the balsa core outperformed the nylon facesheets with the foam core. The polypropylene facesheet sandwich composites had a lower average edgewise compression strength than the nylon facesheet sandwich composites. This result was believed to be a result of the polypropylene facesheet being weaker and less stiff than the nylon facesheets. There was no difference in edgewise compression strength between the two cores with polypropylene facesheets. This result can be attributed to a weaker bond strength between the core and the facesheet. The foam core with the polypropylene facesheets mainly had failures in the foam core where the balsa core with the polypropylene facesheets mainly had adhesive failures. 12.4.2 Static Energy Absorption The average energy absorption was computed and summarized in Tables 12-3 and 12-4 and can be viewed graphically in Figure 12-3. Nylon facesheets with the balsa core had the highest energy absorption of all the candidates. The strength of the bond that can be achieved with the nylon facesheet and the crack arrest properties of the balsa core allowed for more reloading and a higher absorption of energy over the other candidate sandwich configurations. Conversely, nylon facesheets with the foam core had the 87 Table 12-3: Static nylon edgewise comprssion absorbed energy Sandwich Configuration Energy (N m) Average Energy (N m) Std Dev (N m) Openair Plasma Nylon Lord 320/322 Balsa 297.45 546.01 223.95 510.19 520.95 932.00 640.31 375.16 Acid Wash Nylon Lord 320/322 Balsa 624.90 419.42 140.21 510.85 301.14 469.57 355.21 254.84 ATmaP Flame Nylon Lord 320/322 Foam 139.75 122.81 28.81 131.22 167.79 96.78 108.04 93.25 Openair Plasma Nylon Lord 320/322 Foam 168.71 157.60 14.04 141.02 140.68 172.85 156.53 165.81 Acid Wash Nylon Lord 320/322 Foam 97.16 149.11 33.95 185.67 153.33 127.10 147.30 184.09 88 Table 12-4: Static polypropylene edgewise compression absorbed energy Sandwich Configuration Energy (N m) Average Energy (N m) Std Dev (N m) Openair Plasma Polypro Dow LESA w/Primer Foam 386.52 291.68 91.52 160.62 240.75 272.64 404.40 285.14 Openair Plasma Polypro LESA w/Primer Balsa 270.43 335.42 67.21 446.73 271.30 363.75 307.77 352.57 Surface Activation Polypro Lord 320/322 Foam 246.95 190.03 212.26 30.44 199.79 Surface Activation Polypro Lord 320/322 Balsa 327.01 569.22 394.27 152.86 286.58 Surface Activation Polypro Lord 7542 Foam 262.16 147.99 182.89 68.81 138.53 Surface Activation Polypro Lord 7542 Balsa 288.15 463.14 341.39 105.71 272.89 89 Figure 12-3 : Static edgewise compression absorbed energy 0 100 200 300 400 500 600 700 800 900 Openair Plasma Nylon Lord 320/322 Balsa Acid Wash Nylon Lord 320/322 Balsa ATmaP Flame Nylon Lord 320/322 Foam Openair Plasma Nylon Lord 320/322 Foam Acid Wash Nylon Lord 320/322 Foam Surface Activation Polypro Lord 320/322 Balsa Surface Activation Polypro Lord 320/322 Foam Openair Plasma Polypro LESA w/Primer Balsa Openair Plasma Polypro Dow LESA w/Primer Foam Surface Activation Polypro Lord 7542 Balsa Surface Activation Polypro Lord 7542 Foam Average Energy (N m) 89 90 lowest energy absorption of all the configurations. The higher stiffness and strength of the nylon facesheets led to a catastrophic failure of the sandwich composite. The facesheets completely detached from the core, no longer allowing the core/facesheet interaction. After the facesheets were detached from the core, energy could only be absorbed by the buckling of the nylon facesheets. Polypropylene facesheets with the balsa core were the second highest absorber of energy. The crack resistance of the balsa core and the polypropylene facesheets contributed to the higher level of energy absorption. Polypropylene facesheets with the foam core outperformed nylon facesheets with the foam core. The compliance of the polypropylene facesheets allowed for more crushing while remaining attached to the foam core, resulting in higher energy absorption. 12.4.3 Static Average Weight-Normalized Edgewise Compression Strength The average ultimate strength was normalized by weight and summarized in Tables 12-5 and 12-6 and can be viewed graphically in Figure 12-4. When the static average edgewise compression strength data are normalized by weight, the same general trend can be seen, as with the non-normalized data. However, the differences between the balsa core and the foam core sandwiches were greater due to the greater mass per volume of the foam core as compared to the balsa core. The nylon facesheet sandwich composites had a higher average weight- normalized edgewise compression strength than the polypropylene facesheet sandwich composites. The nylon facesheets with the foam core outperformed the polypropylene with the balsa core even while normalized for weight. Although the load applied to the sandwich 91 Table 12-5: Static nylon weight-normalized edgewise compression strengths Sandwich Configuration Weight (g) Length (mm) Width (mm) Ult Strength per Areal Weight (MPa/kg/m²) Average Ult Strength per Areal Weight (MPa/kg/m²) Std. Deviation per Areal Weight (MPa/kg/m²) Openair Plasma Nylon Lord 320/322 Balsa 70.07 127.15 76.50 36.97 35.27 1.53 72.62 126.92 76.71 33.38 72.40 126.85 76.68 36.93 75.28 126.87 76.91 35.09 75.88 127.05 76.94 35.56 73.77 126.92 76.81 33.71 Acid Wash Nylon Lord 320/322 Balsa 76.76 126.37 75.69 31.30 31.58 2.10 78.61 126.75 76.38 30.92 80.02 126.87 76.48 34.98 76.74 126.75 77.06 28.46 79.30 126.87 76.73 31.76 77.52 126.80 76.71 32.06 ATmaP Flame Nylon Lord 320/322 Foam 80.30 126.92 76.48 28.52 29.18 1.45 78.87 126.95 76.53 28.02 79.27 126.75 76.30 27.22 78.34 127.18 76.45 30.70 75.45 127.08 76.35 30.50 76.20 126.97 76.40 30.11 Openair Plasma Nylon Lord 320/322 Foam 81.35 126.92 76.28 26.64 26.79 1.11 80.19 126.95 76.48 26.68 80.90 126.75 76.43 25.14 78.67 127.18 76.38 28.57 77.75 127.08 76.71 27.21 77.55 126.97 76.43 26.52 Acid Wash Nylon Lord 320/322 Foam 81.17 127.03 76.58 24.91 24.28 1.69 80.23 126.64 76.43 24.61 82.22 126.57 76.50 23.62 80.79 126.59 76.73 25.83 82.60 126.47 76.61 25.51 81.17 126.26 76.63 21.21 92 Table 12-6: Static polypropylene weight-normalized edgewise compression strengths Sandwich Configuration Weight (g) Length (mm) Width (mm) Ult Strength per Areal Weight (MPa/kg/m²) Average Ult Strength per Areal Weight (MPa/