ELECTR-BEAM PRCESSABLE PLYIMIDES FR HIGH- PERFRMACE CMPSITE APPLICATIS Andrea E. Hoyt Adherent Technologies, Inc. Development Laboratories 11208 Cochiti SE Albuquerque, M 87123 ABSTRACT Electron beam (e-beam) curing of polymer matrix composites (PMCs) is highly desirable from a manufacturing standpoint in that significant cost savings may be realized from the reduced cycle times, the ability to cure large and irregularly shaped parts, and the ability to use inexpensive tooling. However, for many applications in the aerospace and transportation industries, e-beam processable resins currently available do not meet the performance specifications (e.g., excellent thermal stability, high strength, and high toughness). Adherent Technologies, Inc. has recently been developing linear, low molecular weight polyimides with functionalities amenable to crosslinking under electron beam irradiation. A soluble, fully-imidized polyimide backbone was used as the basis for all materials prepared. ligomers of several different molecular weights were prepared and endcapped with functionalities amenable to electron beam-induced crosslinking reactions. A carbon fiber-based tape impregnated with the most promising resin was prepared and exposed to electron beam irradiation at a total dose of approximately 150kGy. Thermomechanical analysis of unexposed tapes showed considerable compression of the sample at temperatures as low as 150 C. In contrast, the e-beam exposed samples showed less resin softening and less compression during analysis. Glass transition temperatures in the exposed samples were also higher (240 C to 250 C). Current efforts focus on improving polyimide processability for automated tape placement (ATP) and on increasing functionality in the system to afford a better cure. KEY WRDS: E-Beam Curing, Polyimide
ITRDUCTI Polyimides currently in use for PMC applications suffer from several processability problems. High temperatures and often high pressures are required in order to achieve reasonable quality materials. These difficulties arise from the nature of the materials themselves. Condensationtype polyimides are typically processed from solutions of their polyamic acid precursors. The solvents used are generally high-boiling point polar aprotic solvents that are extremely difficult to remove from the system during processing. The result can be an unusually high void content in the matrix [1]. Voids can also arise from the polyimide cure step itself; water vapor is a byproduct of the cyclodehydration required to convert polyamic acid to polyimide. Additiontype polyimides or the polymerization of monomeric reactants (PMR) process can alleviate some of the processing difficulties outlined above. However, high temperatures and pressures may still be required, thus limiting the size of parts that can be manufactured and leading to high equipment and energy costs. E-beam curing has the potential for significantly reducing the production cost for PMC structures; cycle times could be dramatically reduced, leading to a significant energy savings, and inexpensive tooling could be used. Recent developments in tape placement technology may further reduce costs. Additionally, there is virtually no limit to the size of part that can be manufactured with e-beam technology. Adherent Technologies, Inc. has investigated processable polyimide oligomers containing functional groups that may be activated for polymerization by e-beam irradiation. Results of these investigations are discussed in the following sections. Synthetic Methods Synthesis of Reactive Homopolyimides EXPERIMETAL Commercially available monomers were used without purification. m-cresol was distilled before use. Stoichiometric amounts of monomers and endcapper employed depend upon the desired molecular weight for the oligomer. To a solution of the desired amount of aromatic diamine in m-cresol was added the calculated amount of endcapping agent. This mixture was stirred, with slight heating, until all solids had disappeared. The appropriate amount of dianhydride was then added and the reaction flask equipped with condenser and argon purge. The temperature was then raised to approximately 200 C and maintained there for a minimum of 7 h. After cooling to room temperature, the reaction solution was poured into methanol and processed in a blender to precipitate the product polyimide or oligomer. The material was collected by suction filtration, washed with additional methanol and dried. Typical yields are on the order of 99-100%.
Preparation of Polyimide-Siloxane ligomers The calculated amounts of endcap monoanhydride and dianhydride were dissolved in a 50/50 mixture of tetrahydrofuran (THF) and,-dimethylformamide (DMF) at room temperature. To this mixture was then added the calculated amount of an α,ω-amine-terminated polydimethylsiloxane oligomer. This mixture was stirred for several hours to allow reaction of the diamine and anhydride components. The aromatic diamine was then added and the mixture stirred for an additional few hours. THF was then distilled off and the reaction refluxed for a minimum of 7 h. After cooling to room temperature, the reaction mixture (frequently a gel) was precipitated into methanol using a blender. The material was collected by suction filtration, washed with additional methanol and dried. Typical yields are on the order of 70% for the molecular weights desired here. Characterization Methods Spectroscopy Infrared spectra were collected using a Mattson Cygnus 100 Fourier transform infrared (FTIR) spectrometer. Samples were in the form of freestanding films, films on KBr windows, or KBr pellets. Resolution was 4 cm -1 unless otherwise noted. Thermal Analysis Differential scanning calorimetry (DSC) was conducted using an indium-calibrated Perkin-Elmer DSC-7 with nitrogen purge at a heating rate of 10 C/min. unless otherwise noted. Samples were heated over the range 50 C to 350 C. Dynamic thermogravimetric analysis (TGA) was conducted using a Perkin-Elmer TGA-7 at a heating rate of 10 C/min over the range 200 C to 800 C. Samples were heated at 200 C prior to analysis to ensure complete vaporization of residual solvents from the materials. Samples were analyzed in both flowing nitrogen and in static air. Temperatures at which 5% and 10% weight losses occurred are reported. Thermomechanical analysis (TMA) was conducted using a Perkin-Elmer TMA-7 at heating rates of 10 or 20 C/min. Both expansion and penetration experiments were conducted to evaluate glass transition temperatures in the materials. Size Exclusion Chromatography (SEC) Size exclusion chromatography (SEC) was conducted using a Waters 150C operating with Shimadzu Class VP data acquisition and analysis software.,-dimethylacetamide (DMAc) containing 0.1M lithium chloride was used as the solvent. All analyses were conducted at 60 C and a flow rate of 1.0 ml/min.
Tape Preparation Carbon-polyimide tape was prepared by applying a chloroform solution of a 50/50 blend of reactive homopolyimide and reactive polyimide-siloxane containing 5% (w/w) of a free radical photoinitiator to carbon tow (AS4-12k, unsized) on a drum winder. The resulting tape was dried at approximately 70 C in circulating air for 4 h to remove residual chloroform. RESULTS The number average degree of polymerization, X n, in a step-growth polymer is given by: X n = ( 1+ 1 ) r 1+ r [ ( 1 p) + ( 1 rp) )]/ 2 1+ r 2rp A A B where A is the number of molecules of monomer A (diamine, in our case), B is the number of molecules of monomer B (dianhydride or monofunctional anhydride endcapper), r is the stoichiometric imbalance (r = A / B < 1), and p is the extent of reaction. For any given value of X n, assuming p = 1 allows the calculation of the stoichiometric imbalance, r, from the following relationship: X n / 2 ( r) ( r) = 1 1+ The calculated value of r then allows the determination the amount of monofunctional endcapper required to achieve the desired molecular weight. Initially, we chose to investigate oligomers of a known soluble polyimide with a reactive endcap. Reaction compositions for the desired molecular weights were calculated using the stoichiometric methods described by dian [2]. Molecular weights of 2000, 5000, and 10000 were targeted initially. The synthetic scheme for these polyimide materials is shown in Figure 1, below. =
Ar R + + m-cresol H 2 Ar H 2 reflux R Ar Ar Ar n R Figure 1. Reaction scheme for the preparation of soluble polyimide oligomers endcapped with either maleic or phenylmaleic anhydride. % Weight 90 80 70 60 50 40 Unexposed Exposed, 10MR, no init. 300 400 500 600 700 800 900 Temperature ( C) Figure 2. Thermogravimetric analysis of unexposed and 10 MR exposed e-beam reactive polyimide films. Weight loss below 300 C (not shown) was due to residual solvent in the films. The initial electron beam exposure of these materials involved irradiating films of phenylmaleimide-endcapped materials at doses of approximately 5, 10, and 15 megarad (MR) at room temperature. Thermogravimetric analyses of the exposed films showed a 20-30 C increase in the onset of decomposition relative to the starting polyimide material. TGA data for the 10 MR exposure is shown in Figure 2. We believe that the relatively high glass transition temperature (T g ) of these reactive oligomers (~250-280 C, as determined using TMA) and the low concentration of reactive functionalities are inhibiting full cure in these systems at room temperature. Incorporation of siloxane units at a level of 10 mole percent was used to lower the T g of the polyimide while retaining a high level of thermal stability. Reactive oligomers of imide-siloxane block copolymers were prepared according to the scheme presented in Figure 3, below. Again, several different molecular weights were evaluated.
Ar H 2 Ar H 2 and + and THF/DMF, R.T., overnight reflux, DMF, 7h H 2 (CH 2 ) 3 Si Si n Si (C H 2 ) 3 H 2 Ar Ar (CH 2 ) 3 Si Si Si (C H 2 ) 3 n x Ar Ar y Figure 3. Synthesis of polyimide-siloxane block copolymer endcapped with maleic anhydride. There are significant solubility differences between the amine-terminated siloxane oligomers and the other monomers used in this synthesis. Such problems have previously been reported by McGrath [3,4]. Based on the work of McGrath's group, we employed a cosolvent (THF) to improve the solubility of the amine-terminated siloxane oligomer in the reaction mixture. f the imide-siloxane materials prepared, that with a calculated molecular weight of 7500 appeared to be the most promising. FTIR analysis of the material from this synthesis confirmed the presence of the siloxane segment in the polymer (C H stretch corresponding to the methyl groups at ~ 2960 cm -1 ). These imide-siloxane block copolymers showed a significant decrease (~30 C) in T g, affording greater potential for successful processing. Thermal analysis of this material indicated a glass transition at approximately 228 C as determined using TMA. Thermal stability was also excellent, the onset of decomposition (in 2 ) was 518 C and the weight loss at 800 C was approximately 50%. The decomposition temperatures and weight loss were only very slightly lower than those for the homopolyimide initially investigated. Casting of this material was quite difficult. Chloroform solutions of the copolyimide did not wet substrates well, even if the substrates had been treated to improve wetting. When dry, the films flaked off the substrates, indicating that they were probably extremely low molecular weight. Therefore, a 50/50 (w/w) blend of this material with a 3500 molecular weight homopolymer was
prepared for further experiments. Films of the blend still showed some cracking, but could be handled without significant sample loss or damage. Final Materials Evaluation A chloroform solution of the blend prepared above, containing 5% (w/w) of a free radical photoinitiator was applied to 12k unsized AS4 fiber to make a tape approximately one-half inch wide. A tape transport system was devised such that the tape was heated to near the T g of the imidesiloxane block copolymer immediately before progressing through the electron beam. Immediately after exiting the beam area, the tape was passed through a springloaded roller system to consolidate the material. The processed tape was then respooled on the other side. A schematic of the tape handling system is shown in Figure 4. Figure 4. Tape handling system for e-beam irradiation. Photographs of the tape passing through the rollers and of the completed tape are given in Figures 5 and 6.
Figure 5. Carbon/polyimide tape being transported through electron beam and consolidation rollers. Figure 6. E-beam processed polyimide tape. Qualitatively, the resin was much more coherent after electron beam processing than before. Thermogravimetric analysis of the tape after cure (shown in Figure 7) indicated an onset of decomposition (T d ) at approximately 570 C; this is 20 C higher than the T d of the polyimide component and 40 C higher than the T d of the imide-siloxane component. Five percent weight loss occurred at approximately 575 C and ten percent weight loss occurred at approximately 630 C, 20-80 C and % Weight 110 100 90 80 70 60 50 40 200 300 400 500 600 700 800 Temperature ( C) Component 1 Cured Sample Component 2 Figure 7. Comparison of thermogravimetric behavior of resin components vs. cured materials. 50-100 C higher than the individual components before cure. Additionally, the total resin weight loss at 800 C was decreased from approximately 50% to 30% after e-beam cure.
The glass transition temperature of the cured resin in the tape was difficult to measure. A single layer of tape was placed in a thermomechanical analyzer (TMA) and analyzed in expansion mode. Compression of the samples was noted, most likely due to softening of uncured resin in the samples. Representative TMA data are shown in Figure 8. The unexposed tape showed considerable compression of the sample at temperatures as low as 150 C, near the T g of the siloxane component in the resin formulation; softening continued throughout the temperature scan. Ultimately, the sample was compressed by approximately 40% during analysis. In contrast, the e-beam exposed samples showed less resin softening and less compression during analysis. Glass transition temperatures in the exposed samples were also higher (onsets ~240 C and 250 C). However, it should be % Height 100 90 80 70 60 100 150 200 250 300 350 Figure 8. Thermomechanical analysis of unexposed and exposed carbon/polyimide tapes noted that the cure was not uniform throughout the tape. Exposed sample #2 even showed a high T g point of ~300 C (small step) and only about 10% compression during analysis. We believe that the cure nonuniformity is due in part to the relatively high T g s of the two components in the resin formulation; due in part to the relatively low concentration of functional groups in the system, which may inhibit cure at temperatures below the system T g ; and due in part to inadequacies in the tape handling system designed for these experiments. CCLUSIS unexposed tape Temperature ( C) exposed tape, sample #2 exposed tape, sample #1 The TGA and TMA data presented above indicate that system cure is indeed occurring upon electron beam exposure of these reactive polyimide systems. The main barriers to full success are the relatively high glass transition temperatures of even low molecular weight polyimides and the low concentration of e-beam reactive functional groups in the system. Materials development is currently underway at Adherent Technologies, Inc. to address these processability issues. The objectives of this development effort are lowering the uncured resin T g into the range accessible by current automated tape placement (ATP) technology (T max = 180 C) and increasing the functionality of the systems without sacrificing thermal stability.
REFERECES 1. Geldermans, P., Goldsmit, C., and Bedetti, F., In Polyimides: Synthesis, Characterization, and Applications, Vol. 2, Mittal, K.L., Ed., Plenum, ew York, pp. 695-711, 1984. 2. dian, G. Principles of Polymerization, 2nd Ed., Wiley: ew York, 1981, pp.83-84. 3 C. A. Arnold, J. D. Summers, and J. E. McGrath, Polym. Eng. Sci., 1989, 29, 1413. 4 C. A. Arnold, J. D. Summers, Y. P. Chen, R. H. Bott, D. Chen, and J. E. McGrath, Polymer, 1989, 30, 986. ACKWLEDGEMETS Many thanks to Mr. Dana Finley and Mr. Jim Allred of Adherent Technologies for their expertise and assistance with the tape handling apparatus used in this work. The use of the electron beam facilities at orth Star Research, Inc. of Albuquerque, M is also gratefully acknowledged. This work was funded by ASA Glenn Research Center under the Small Business Innovation Research (SBIR) program. Continuing support from Dr. Michael Meador of ASA Glenn is also greatly appreciated. BIGRAPHIES Dr. Andrea E. Hoyt received her Ph.D. in Polymer Science from the University of Connecticut Institute of Materials Science in 1993. She has a B.A. in Chemistry from the University of Colorado and an M.S. in Polymer Science from the University of Connecticut. Dr. Hoyt is the Manager of Polymer Projects at Adherent Technologies. In this position, she has led projects in the development of electron beam curing resin systems, coatings for chemical sensors, liquid crystalline thermoset adhesives, and new moisture resistant coupling agents for glass fiberreinforced composites. Her current research focuses on the development of novel inorganic polymers for scintillator applications and on UV-curable resin systems. Before joining Adherent Technologies, she was engaged as a postdoctoral research associate in the Microsensors Research and Development Department at Sandia ational Laboratories. She has also done research in liquid crystalline polymers and liquid crystalline thermosets at Los Alamos ational Laboratory and in polyimide synthesis and characterization at the University of Connecticut. Dr. Hoyt has over 10 publications in technical journals and conference proceedings and holds 4 patents in the areas of liquid crystalline thermoset materials and chemical sensors.