1996 Annual Report
Technical Activities: POLYMER COMPOSITES PROGRAM
The polymer composites program seeks to facilitate the introduction of light-weight, corrosion-resistant composite materials into commercial applications by expanding the essential science base and generating test methods, reference data, and standard materials. The outstanding properties of composites mean they can be used to make products that are superior and competitive in international markets. Industries as diverse as transportation, construction, marine, off-shore oil, medical devices, and sporting goods have recognized the benefits and are beginning to make significant use of these materials. For this to continue, however, two significant barriers must be addressed: the lack of rapid, reliable, cost-effective fabrication methods, and the poor understanding of and predictive capability for long term performance. These barriers were identified in a series of industry workshops, exchange visits, and consultations. In response to these challenges, the composites program initiated two tasks: one on processing science and the other on interfacial microstructure. The degradation of the interface over time is primarily responsible for the loss of mechanical properties. The automotive industry strongly influences the composites program since many of the processing and durability issues span many automotive applications, and solutions developed at NIST are expected to rapidly propagate throughout the industry. Additionally, the group interacts with companies interested in off-shore oil platforms, infrastructure, aerospace, and a variety of other applications.
The goal of the Processing Science Task is to develop the technology required to monitor, model, and control the events that occur during composite fabrication. The program focuses on liquid composite molding (LCM) since this fabrication method is of great interest to all industry sectors and is the consensus choice of the automotive industry as the method with the most promise for making structural automotive parts. The approach in this task involves three steps. First, measurement tools are developed and used to characterize the material properties that control processing, for example, permeability. Second, sophisticated process simulation models are formulated to analyze the effects of processing parameters rapidly and inexpensively so they can be optimized. Finally, process monitoring sensors are developed and used to provide feedback for verification and improvement of the simulation models and to help develop the technology for on-line process control. The current activities in this Task involve five projects, including a major industry-university-government program sponsored by the Defense Advanced Research Projects Agency.
The work in the Microstructure Task focuses on developing test methods for assessing the resin/fiber interfacial adhesion, and the subsequent degradation of adhesion resulting from fluid attack, particularly moisture. The long term goals are to first develop effective test methods, then to use those tests to identify the chemical and physical mechanisms of degradation, and finally to formulate reliable predictive models. The program focuses on glass fiber materials since they are the primary candidates for automotive applications. In addition, the work is beginning to look at graphite reinforced composites since these systems are important for marine and infrastructure applications. Microscale tests such as the single fiber fragmentation test are currently being analyzed to determine if they can provide realistic estimates of the performance of the resin/fiber interface in composite systems. A variety of interfacial physical and chemical structures are generated during preparation of microscale test specimens by varying the coating chemistry on the fiber, the resin processing speed, and the moisture content of the material. Full scale composite specimens are also produced and tested with identical fiber coatings and processing conditions for comparison with the microscale tests and to provide, in conjunction with the microscale tests, realistic structure-performance relationships. There are currently four specific projects in this Task, including a collaboration with the Automotive Composites Consortium to determine the effects of processing conditions on the interface of polyurethane matrix composites.
FY-96 Significant Accomplishments
A micromechanical testing method developed for polymer composites was shown to be an effective tool for measuring the interfacial shear properties of dental materials. For example, using the microbond interfacial strength test, a hydrophobic silane coupling agent was demonstrated to produce more durable adhesion than the dental industry standard formula.
A traditional Lattice Boltzmann formulation was modified to model flow in heterogeneous media, where the momentum transport is expressed with a combination of the Stokes and Brinkman equations. Lattice Boltzmann techniques enable efficient computation in real materials and the inclusion of important multiphase flow physics.
A sensor capable of detecting rapid polymerization reactions was developed in cooperation with industry and demonstrated in the fabrication of epoxy, polyurethane, and polyester composite plaques. Ford is transferring the technology to their laboratory where it will be implemented on prototype production equipment as part of a NIST/Ford/GE ATP program.
The role of resin viscoelasticity in determining the strength and durability of the interphase of single-fiber-composite test specimens was demonstrated for a model glass-epoxy system. Previous micromechanical testing has largely ignored resin rheology, and that is now thought responsible for much of the scatter and unreliability of typical test methods.
An international collaboration through the Versailles Advanced Materials and Standards Program (VAMAS) was initiated to develop standard techniques for the preparation and testing of single-fiber-composite test specimens. To date, 17 laboratories representing 6 different countries have agreed to participate in the program.
An industry/NIST workshop considered different types of sensors in the context of composite manufacturing challenges and plausible process control scenarios. Only inexpensive, nonintrusive sensor systems hold the interest of manufacturers and systems developers, but very sophisticated control schemes that can make use of minimal sensor information are in demand.
A new set of technical challenges revolving around quality control were identified in a workshop co-sponsored by the Polymers Division and Ohio State University on Liquid Composite Molding. The focus on quality control results from the emergence of a nascent liquid molding industry since the first workshop on Liquid Molding was held at NIST three years ago.
Liquid Composite Molding: Development of Permeability Measurement Techniques and Data
R.S. Parnas
Objective
The objective is to establish a data base of permeability values for use in the design tools used by the composites industry for process and mold design.
Technical Description
Permeability measurements conducted over several years at NIST have been documented, collected, and placed into a Clipper based database. External sources of reliable data have been identified and attempts to document the adequacy of their measurements for inclusion in the database continue. Permeability measurements are continuing with an emphasis on the permeability of fabrics deformed around curves as would be found in molds of complex shape.
External Collaborations
NIST: Standard Reference Data Division
Industrial: Textile Research Institute
Academic: University of Montreal, University of Nottingham
Planned Outcomes
Version 1 of the database is planned for release in early 1997 and it is expected that molders will use it to help design their processes and parts. Version 2 is planned for release in 1998 with an expanded dataset and enhanced graphics display.
Accomplishments
Over 100 experiments conducted at NIST with a variety of glass reinforcements have been entered into the database. The database was demonstrated at the Second Workshop on Liquid Molding, and a number of requests for the database were made at that time. Collaborations with two other sources of data have been successfully started, as measured by approximately 20 additional data sets that have been imported into the database.
Impacts
Over the past several years engineers from companies including Ford, Boeing, and Northrup/Grumman, as well as several engineering students, have learned how to make accurate permeability measurements through participation in the NIST permeability measurement project.
Outputs
Publications
R.S. Parnas, Chapter 8. Preform Permeability, in RTM for Aerospace Applications, Chapman & Hall, 1997.
H.L. Friedman, R.A. Johnson, B. Miller, D.R. Salem and R.S. Parnas, In-Plane Movement of Liquids Through Curved Fabric Structures, Proc. ASME Symp. on Processing, Design and Performance of Composite Materials, Nov.1995.
Liquid Composite Molding: Development and Verification of Process Simulation Models
F. R. Phelan Jr.
Objectives
The objectives are to develop and apply models that can simulate the events which occur during the LCM process by including the effects of preform deformation and heat transfer. The model will be developed specifically to simulate injection compression molding for the automotive industry and their suppliers.
Technical Description
Process optimization has been done with time-consuming and expensive trial and error methods on full scale equipment. Simulation models can greatly reduce the cost and increase the speed of this task. The simulation models developed in this project are based on a finite element / control volume numerical solution procedure to the governing transport equations. For example, the momentum transport equation is expressed by Darcy's law. In previous work, a Darcy's law simulation, called CRIMSON, for modeling the mold filling phase of LCM was developed. CRIMSON enables modeling of resin injection for either constant flow rate or constant pressure injection conditions, in geometries ranging from 2-D to fully 3-D.
In the next phase of this project, CRIMSON is extended to a second generation LCM process, called Injection/Compression Liquid Composite Molding (I/CLCM). This process has been selected by the automotive companies as the most promising method for fabrication of large structural parts. I/CLCM differs from conventional "injection-only" LCM in that subsequent to preform placement, the tool is only partially closed. An initial charge of resin is then injected, followed by full mold closure. The final closing action of the mold compresses the preform to the desired net shape and volume fraction while distributing the initial shot of resin throughout the part. There are two main I/CLCM process variants. In "closed mold" I/CLCM, shown in Figure 1a, the tool is closed enough to partially compress the preform, and thus, hold the preform in place during the injection phase of the process. In the "ideal" open mold process shown in Figure 1b, a gap exists between the preform and the upper tool surface. The strategy in this case is to try and fill the gap region with fluid first, and then use the compression step to drive the fluid into the preform in the thickness direction. This scenario is termed "ideal" because during an actual open mold injection, there is some penetration of resin into the preform during the injection phase, so that during the compression step there is also some in-plane flow.
External Collaborations
Industrial: Structural Dynamics Research Corp., Automotive Composites Consortium, Northrup/Grumman
Academic: University of Illinois
Planned Outcomes
Accomplishments
In previous work, CRIMSON was modified to model the case of closed mold I/CLCM. This year, work has begun in cooperation with the University of Illinois at Urbana-Champaign to develop a numerical simulation of the general open mold I/C process. This work is not yet complete, however, a comparison of "ideal" open mold I/C and closed mold I/C is illustrative of the differences between the two processes. Table 1 compares the maximum pressure obtained when filling of a 30.48 x 30.48 x 1 cm flat plaque using the two processes. Since the ideal case is a simple 1-D flow, these calculations were done analytically while the closed mold calculations were performed using the computer simulation. The processing parameters were the same for both cases. The result shows that quite a substantial reduction in the maximum pressure is obtained for the ideal case as compared to the closed mold flow. This difference can be attributed mainly to the very short penetration distance needed in the ideal case, as compared to the long flow path required in the in-plane flow closed mold process.
The development of I/CLCM stems from the need to mold high fiber volume fraction components in applications with fast cycle times. In pure-injection LCM, such process constraints can result in excessively high injection pressures that induce undesirable fluid-structure interactions involving preform, foam core, or tool deformation.
In an effort to link the CRIMSON software with the user community, Structural Dynamics Research Corp. (SDRC) has developed a graphical user interface enabling the interfacing of the NIST flow modeling software CRIMSON with their I-DEAS Master Series mechanical design software. The interface allows the user from within I-DEAS to design a part, specify preform and fluid properties such as permeability and viscosity, enter boundary conditions, and then run the CRIMSON program. Results are automatically read back into I-DEAS when the simulation is finished.
Impacts
The simulation software has been transfered to the ACC, Budd, and Northrup/Grumman, and is regularly used as a design tool.
Outputs
Publications
Phelan Jr., F.R., Analysis of Injection/Compression Liquid Composite Molding Process Variants, Proceedings of the ASME96 Symposium on Multi-Disciplinary Issues in Manufacturing of Composites, Non-metals and Metals, (1996), in press.
Phelan Jr., F.R., Simulation of Injection/Compression Liquid Composite Molding, Proceedings of the First Joint Topical Conference on Processing Structure and Properties of Polymeric Materials, to appear, (1996).
Presentations
Phelan Jr., F.R., Numerical simulation of injection/compression liquid composite molding, presented at the 11th Annual ASM/ESD Advanced Composites Conference and Exposition (ACCE95), ASM International, (November 1995).
Phelan Jr., F.R., Simulation of injection/compression liquid composite molding, presented at The Second Workshop on Liquid Composite Molding, Ramada University Hotel, The Ohio State University, Columbus, OH, (June 14,1996).
Liquid Composite Molding: Development and Verification of Permeability Prediction Models
F.R. Phelan Jr. and Michael A. A. Spaid
Objective
The objective is to develop theoretical tools to aid industrial designers for predicting the permeability tensor of the fiber reinforcement materials used in Liquid Composite Molding (LCM) from a knowledge of their microstructures.
Technical Description
During the past year, a novel approach for computing the flow behavior in real reinforcement materials was developed. Rather than solving the fluid flow problem with standard numerical techniques such as finite elements or finite differences, a lattice Boltzmann (LB) method has been adopted. The primary advantage of the LB method is in its ability to model robustly the two-phase nature (resin/air) of the flow processes.
LB methods involve the solution of the discrete Boltzmann equation for the particle distribution function f(x,v,t), given by
where v is the discrete representation of the velocity space, and c, is the collision operator which couples the velocity states. For suitably symmetric lattices, it is possible to prove that the Navier-Stokes equations are recovered from the LB formulation. Traditional fluid flow quantities such as velocity and density may be recovered by taking moments of the particle distribution function as follows
Most LB formulations employ a linear form of the collision operator c,, under the assumption that the particle distribution may be expanded about its equilibrium value:
In (3), feq is the equilibrium distribution function, and is a parameter which controls the rate of relaxation to equilibrium. For a 2-D hexagonal lattice, the velocity space is comprised of six vectors of equal magnitude v0 which point along the lattice links, and the zero vector to incorporate rest particles. The particle equilibrium distribution function for the hexagonal LB model is given by
where d0 is the fraction of rest particles.
Accomplishments
In this study, a traditional LB formulation was modified to enable the modeling of flow in heterogeneous media, where the momentum transport is expressed with a combination of the Stokes and Brinkman equations.
To validate the LB formulation of the Stokes and Brinkman equations, simulations were performed for flow over porous cylinders with elliptical cross section. The effective permeability predicted by the simulations was then compared to a semi-analytical lubrication model previously developed in this project, and known to agree with rigorous finite element results. Excellent agreement between the LB calculations and the semi-analytical model were obtained for both cylinders and ellipses with relatively small lattice sizes.
The LB model developed is a viable alternative to directly solving the Stokes-Brinkman equations using standard numerical techniques. One of the advantages of the LB method is the ability to model the dynamics of the flow, which is of particular interest in an inherently unsteady flow problem such as the infiltration of RTM preforms. In addition, modifications of the LB model to include important physics such as wicking forces are relatively simple to implement. Future work will focus on the implementation of a multiple phase LB code in order to study the mechanisms of void formation as a function of the relative importance of viscous and surface tension forces, as measured by the dimensionless capillary number.
Outputs
Publications
Spaid, M.A.A., and F.R. Phelan Jr., Lattice Boltzmann Methods for Modeling Microscale Flow in Fibrous Porous Media, Proceedings of the Third International Conference on Composites Engineering (ICCE/3), pp. 785-786, (1996).
Spaid, M.A. and F.R. Phelan, Jr., Lattice Boltzmann methods for modeling microscale flow in heterogeneous porous media, Physics of Fluids, (submitted).
Liquid Composite Molding: Bulk Resin Measurements for Process Monitoring and Control
J. P. Dunkers, R. S. Parnas, K. M. Flynn, R. E. Neff, D. L. Woerdeman, and D. D. Sourlas
Objective
The objective is to develop optical fiber sensors, spectroscopic measurement methods, and control structures for monitoring and controlling chemical and physical processes during composites manufacturing.
Technical Description
The need to reduce the variation in composite quality has been recognized for many years. Variation in cure between parts and within a part is a major contributor to composite non-uniformity. The cure monitoring work focuses on fluorescence and near infrared spectroscopies, using an optical fiber drawn from a NIST standard glass as the sensing element. The information from the cure sensor is transferred to a process control computer that makes adjustments to the mold temperature to bring the cure into agreement with the desired cure trajectory, thus reducing part non-uniformity. For slowly curing materials, a model-assisted feedback controller uses the cure data provided by the sensor and a kinetic model of the resin to bring the actual cure trajectory into agreement with the desired cure profile. For rapidly curing materials, receding horizon control strategies are under development.
External Collaborations
Industrial: Automotive Composites Consortium, Ford, ICI Polyurethanes, Northrup/Grumman (DARPA)
Academic: University of Missouri, Florida State University
Accomplishments
Fluorescence Sensor:
Advances were made this past year in data analysis, interpretation, and instrumentation. A number of molding experiments were done with varying processing parameters and resin systems to evaluate the accuracy and range of operation of this monitoring technique. The sensor was found to work well in epoxy, polyurethane (see figure), and polyester systems. Furthermore, the fluorescence
measurements distinguished resin quality degradation in epoxy/amine systems. From statistical treatment of the data, it was found that errors in the measurements were largest early in the reaction and more pronounced at higher temperatures. Time resolution was concluded to be the major source of error at early times, with the measurements approaching the instrument error in the later stages of cure.
Development of a high speed fluorescence system continued during the past year to solve the problems associated with poor time resolution. A charged coupled device (CCD) camera detector and on-line spectral analysis were implemented, providing real-time data for rapidly curing systems. An example of the system capability is illustrated by the cure data obtained in a glass reinforced polyurethane composite. Data were obtained every 10 seconds, and the gel point is clearly indicated by the intensity peak at approximately 3 minutes. This speed and gel point detection is a requirement of the automotive industry for cure sensing. The sensor was made rugged and portable through the use of fiber optic components and a portable Ar+ laser. The sensor system is implemented in an industrial liquid molding facility, where the ability of the sensor to predict the final molded part quality will be assessed.
Near Infrared Sensor:
The fiber optic sensor is coupled with a Fourier transform infrared spectrometer to monitor the cure of the resin in the middle of a resin transfer molded part using evanescent wave spectroscopy. Near infrared spectroscopy provides a direct measure of chemical information and the ability to follow multiple reactions simultaneously. A mini-bundle fiber optic sensor was constructed of three, 130 mm diameter fibers with a refractive index of 1.62. A mini-bundle configuration was chosen to optimize the signal to noise ratio by increasing energy throughput. The spectrometer, optics, and detector are constructed in a 0o configuration where light is launched into one end of the fiber mini-bundle and detected at the other end.
A number of molding and sample cell cure experiments were successfully performed using an epoxy/amine resin system up to 150oC. From these spectra, information about the oxirane and amine consumption or the hydroxyl group evolution can be analyzed in real time to provide cure data to the controller. The isothermal sample cell spectra are used to generate a kinetic model for the control algorithm. Using rapid scanning software, spectra of a reacting polyurethane resin system were also obtained, demonstrating that the near IR system can also be used with rapidly curing systems.
Control:
A model-assisted feedback cure control algorithm, previously developed in this project, was tested with cure simulations and experiments. Simulations helped optimize the controller parameters and experiments demonstrated the controller effectiveness with epoxy/amine resins in glass reinforced composites. The control algorithm was also analyzed for robustness to measurement noise. Both simulations and experiments demonstrated that the presence of noise did not alter the optimum control parameters, and that acceptable control was maintained up to approximately 5% measurement noise.
The previously developed algorithm was useful for controlling the cure to relatively simple pathways, and was generalized to permit arbitrarily complex cure pathways as functional setpoints. Analytical approximations to nonisothermal cure models were developed to permit the algorithm to invert measured cure data to find history dependent temperature setpoints.
Impact
Sensor and control technology implemented at Grumman for the manufacture of the greenbody of a ceramic matrix composite part of a jet engine. Sensor technology implemented at Ford shortens the time to assess part quality from 1 day to 1 minute.
Outputs
Publications
R.A. Neff, D.L.Woerdeman and R.S. Parnas, Use of a Charged Coupled Device (CCD) Camera for Evanescent Wave Optical Fiber Cure Monitoring of Liquid Composite Molding Resins, Polymer Composites (submitted).
D. L. Woerdeman, J. K. Spoerre, K. M. Flynn and R. S. Parnas, Cure Monitoring of the Liquid Composite Molding Process Using Fiber Optic Sensors, Polymer Composites (in press).
J. P. Dunkers, K. M. Flynn and R. S. Parnas, A Mid-Infrared Attenuated Total Internal Reflection Cure Sensor for the Control of the Resin Transfer Molding of a Pre-Ceramic Polymer, Composites: Part A, December 1996.
J. P. Dunkers, K. M. Flynn, R. S. Parnas and D. D. Sourlas, The Effect of Noise and Control Parameters on the Efficiency of a Model Assisted Feedback Control Algorithm for Liquid Composite Molding, Proceedings of the 1st Joint Topical Conference on Processing, Structure, and Properties of Polymeric Materials, in press.
D. L. Woerdeman, K. M. Flynn, J. P. Dunkers and R. S. Parnas, The Use of Evanescent Wave Fluorescence Spectroscopy for Control of the Liquid Molding Process, J. of Rein. Plas. & Comp., 15(9), 922 (1996).
Presentations
J. P. Dunkers, K. Flynn and R. S. Parnas, An IR-ATR Cure Sensor for the Control of the Resin Transfer Molding of a Pre-Ceramic Polymer, Gordon Research Conference on Composites, Ventura, CA, January 11, 1996.
R. Neff, Evanescent Wave Optical Fiber Sensors for Monitoring and Control of the Liquid Molding Process, Advanced Composites Conference and Exposition, Dearborn, MI, November 15, 1995.
R. S. Parnas, Attenuated Total Reflection Infra-Red Cure Monitoring of the Liquid Molding Process for Ceramic Matrix Composites, AIChE Meeting, Miami, FL, November 16, 1995.
R. S. Parnas, Evanescent Wave Fluorescence Cure Monitoring for Use with Process Control of the Liquid Molding Process, AIChE Meeting, Miami, FL, November 16, 1995.
R. S. Parnas, Evanescent Wave Optical Fiber Cure Monitoring and Model Assisted Control of the Liquid Molding Process, American Society of Nondestructive Testing Spring Conference, Norfolk, VA, March 19, 1996.
R. S. Parnas, Sensors and Control, Second Workshop on Liquid Composite Molding, Columbus, OH, June 14, 1996.
R. S. Parnas, A Fiber Sensor for Simultaneous Fluorescence and Near Infrared Measurements During Composite Processing, 1996 Fifth World Congress of Chemical Engineering, San Diego, CA, July 17, 1996.
D. L. Woerdeman and R. S. Parnas, In-Mold Sensors to Monitor the Liquid Molding Process and Post Process Performance, Society of Engineering Science, New Orleans, LA, October 31, 1995.
Liquid Composite Molding: Interphase Sensitive Sensors for Process Monitoring
R.S. Parnas
Objective
The objective is to develop sensors with sensitivity to the 100 region near the fiber surface, a critical region of composite materials that determines many of the bulk mechanical properties.
Technical Description
One possible route to such a sensor lies through localizing fluorescent dyes at the surface. The process of localizing fluorescent dyes at the surface entails synthesizing the appropriate dye with a chemical tail containing a surface reactive group, typically a silane coupling group. After completing the synthesis, surface grafting is conducted onto a substrate such as a flat glass slide or an optical fiber. The fluorescence of the surface bound material is measured in the presence of various solvents and resins to assess the ability of the surface bound fluorophore to respond to the environment. Critical technical issues that must be resolved after the surface chemistry is understood concern the structure of the surface layer, and the calibration of the surface bound fluorophore.
A second possible route to an interphase sensitive sensor lies through fiber ultrasonics, and work has begun to build on a previous project based on ultrasonic shear wave reflection. A substrate with known mechanical properties and a low attenuation for the propagation of ultrasonic shear waves is used, and a polymer is placed on its surface. A shear wave is generated in the substrate and reflects from the substrate-polymer interface. The amplitude and phase of the reflected wave depend on the mismatch in mechanical properties. By measuring the reflected wave and combining the results with the properties of the substrate, the viscoelastic behavior of the polymer in the region close to the interface can be calculated. Even thin polymer films can be characterized with this technique.
External Collaborations
Academic: Howard University, The Johns Hopkins University
Accomplishments
A second batch of Robello siloxane (fluorescent dye linked to silane coupling agent) was synthesized. Most importantly, the difficulty of reproducibly synthesizing the dye was elucidated by the different behavior exhibited by the second batch of dye. The solubility of the second batch was much higher than the first batch in organic solvents such as xylene, and the fluorescence behavior of the grafted layers was much worse. Once it was determined that the second batch of material was actually purer than the first batch, experiments were carried out to redevelop the grafting procedure. Controlled hydrolysis experiments demonstrated the necessity of water in the surface reaction to produce a fluorescing surface layer, and this result is consistent with the solubility behavior of the materials. Alternative synthesis routes involving the production of Robello alcohols are being explored in an attempt to better control the structure of the surface bound layer.
Previous ultrasonics studies used a manual technique for determining phase shift. It was time consuming, used just a few selected points in the signal, and produced a resolution that was only marginal. A new technique has now been developed that digitizes the signal for the first two reflections and transfers the data to a computer for analysis. The data acquisition and analysis processes are fully automated and use the complete signal for maximum accuracy. To determine the stability of the result, several hundred measurements were made over a two day period with a bare substrate. The data show very little variation in the transit time, two standard deviations were less than 100 picoseconds. The practical limit to the scatter is governed by the temperature control, and this value is very close to that limit.
Environmental Durability Studies: Effect of Fiber Coatings and Interfaces
Gale A. Holmes and Donald L. Hunston
Objective
The objective is to determine the effects of fiber surface treatments and coatings on the durability of composites. The research is focused primarily on determining the effect of moisture attack on the fiber and fiber-matrix interface in glass fiber/epoxy systems.
Technical Description
Model composites containing a single bare E-glass fiber, with various silane coating treatments, embedded in an epoxy resin were prepared. Four types of silane coatings were employed, n-octadecyl trichlorosilane (n-OTCS), n-octadecyl triethoxysilane (n-OTES), g-aminopropyl triethoxysilane (g-APTMS), and g-glycidyloxypropyl trimethoxysilane (g-GOPTMS). In addition, a 35%/65% mole percent mixture of n-OTCS and g-APTMS, respectively, was prepared. To determine the effect these coatings have on the interfacial shear strength and durability of the fiber and fiber-matrix interface when exposed to moisture, the single fiber fragmentation test (SFFT) has been used. Additionally, the coatings on the fibers are characterized by dynamic contact angle measurements for quality control and to determine the hydrophobicity for later correlation with the durability results.
Previous research in this project indicated that simplifying assumptions concerning the resin properties used in the analysis of SFFT data may have a significant impact on the interfacial shear strength and durability results. The micro mechanics models developed for the SFFT use simple linear elastic or elastic-plastic models for the resin mechanics, whereas it is well known that the resins are viscoelastic. Improved micromechanics models of the SFFT are being developed and tested with the data obtained from the samples with various coatings. To support the analysis effort, improved instrumentation and experimental techniques have been implemented to insure that the necessary data are acquired for the improved models.
External Collaborations
Industry: Owens Corning, OSI
Academic: University of Utah, Michigan State University
International: Versailles Advanced Materials and Standards Program (VAMAS)
Planned Outcome
Accomplishments
Previous work used a coating procedure that was vulnerable to pH variation, and consequently caused experimental inconsistencies. Therefore, a new procedure, using 95% ethanol, was devised to simplify the precipitating solution and control the pH. To eliminate the production of HCl, n-octadecyl triethoxysilane (n-OTES) was used instead of n-OTCS. In an effort to avoid excessive protonation of the amine group on g-APTMS and minimize the hydrolysis of the epoxide group on g-GOPTMS all solutions were precipitated under basic conditions. Characterization of the coated fiber surfaces by dynamic contact angle measurements showed that a fiber surface coated by the new procedure was slightly more hydrophobic than when prepared by the old procedure.
Careful attention to the experimental procedures used in the SFFT have revealed the following: (1) detectable viscoelastic relaxation often occurs with each strain increment, (2) the resin matrix is nonlinear elastic over the strain region of interest, (3) fiber fracture often occurs primarily in the nonlinear region of the matrix stress strain curve, and (4) absorption of moisture in the resin matrix reduces the yield stress of the matrix and significantly alters the stress strain curve of the matrix. These observations suggest that the linear elastic micro mechanics models do not accurately reflect the impact of the matrix on the shear-stress transfer process during the SFFT. In addition, significant changes in the matrix properties, e.g., plasticization of the matrix due to moisture absorption, can result in significant changes in the shear-stress transmissibility of the fiber-matrix interface. Changes in the transmissiblity of the fiber matrix interface are also caused by degradation of that interface. Therefore, it is important to ascertain the importance of changes in the matrix viscoelastic properties on the durability analysis and shear-stress transfer process. If these changes are important, the durability of the fiber-matrix interface can only be accurately ascertained by developing a method to decouple transmissibility changes due to changes in the matrix properties from transmissibility changes due to degradation of the fiber matrix interface.
To accomplish these tasks the fragmentation apparatus has been modified to include a load cell and a data acquisition program was written to take load measurements and fragment data during the SFFT. To obtain a better understanding of the fragmentation process, accurate measurements of the fragment lengths are being made after each strain increment. As a result, fragmentation maps are readily generated and the fragmentation results are being compared with updated micro mechanics models. To compare these results with the predictions of stress analysis, the basic viscoelastic properties of the resin are being determined by conducting stress relaxation experiments at a variety of strains using pure epoxy resin. The characterization data will be used for stress analyses conducted at NIST and will also be made available to others in the field through the VAMAS program.
In a further effort to examine fiber-matrix test methods, an international research effort has been organized with the assistance of Prof. Larry Drzal at Michigan State University. The initial focus will be on the single fiber fragmentation test (SFFT). Previous round robins on the SFFT have been discouraging in that the data show a large scatter. Close examination of the results, however, suggests that the variations within each laboratory were significantly less, and where laboratories had close collaboration, the agreement between them was better. Consequently, it was suggested that much of the scatter may be attributable to differences in sample preparation technique and testing procedure among the various groups involved. The proposed program will examine this hypothesis. To date, 17 laboratories representing 6 different countries have agreed to participate in the program. It will be conduced under the auspices of the Composites Working Group of VAMAS, the Versailles Advanced Materials and Standards Program.
Outputs
Publications
K. S. Macturk, C. R. Schultheisz, D. L. Hunston and C. L. Schutte, The Effect of Coupling Agent on Composite Durability, p. 403 in Proc. Adhesion Soc. (Adhesion Society, Blacksburg, 1996).
D. L. Hunston, K. S. Macturk, C. R. Schultheisz, G. Holmes, W. G. McDonough and C. L. Schutte, The Role of Silane Surface Treatments in Strength and Durability of Fiber-Matrix Bonding in Composites, Vol. 2, pp-427-432 in Proc. EurAd'96, European Adhesion Conference (The Institute of Materials, London, 1996).
Environmental Durability Studies: Comparison of Test Results for Laminated Composite Samples and Single-Fiber Composite Specimens
C.R. Schultheisz and D.L. Hunston
Objective
The objective is to determine if micromechanical measurements on model systems can predict the long-term durability of composite materials.
Technical Description
The single-fiber fragmentation test (SFFT) has been shown to provide a qualitative ranking of resin/fiber interface strengths. This test also provides information about the degradation of the interface when samples are subjected to long term exposure to heat, water, and other environmental effects. To determine the accuracy of SFFT tests for predicting composite material behavior, the results have been compared with mechanical measurements on composites.
An E-glass/epoxy model system was chosen as the initial focus in the overall task, because glass-fiber composites involve lower raw- material costs and a potentially larger-volume market as structural composites when compared to carbon-fiber composites. The program involves testing single-fiber samples in a dry state and after varying times of immersion in 25 C and 75 C distilled water, and evaluating the degradation of the fiber and fiber/matrix interface. In addition, tests on macroscopic samples in the dry state and at two levels of moisture uptake at 25 C and 75 C were performed for comparison. Four types of mechanical tests have been used: tension tests to reflect the degradation of the fiber strength, while compression tests and interlaminar fracture tests in Mode I and Mode II reflect the degradation of the interfacial strength and changes in the matrix material.
External Collaborations
Industrial: Automotive Composites Consortium, Owens-Corning, Dow Chemical, Morrison Molded Fiber Glass, Textile Research Institute.
Academic: University of West Virginia, Northwestern University
Accomplishments
Fabrication, immersion (in 25 C and 75 C distilled water for up to 5000 hours), and testing of the single-fiber fragmentation samples and the macroscopic composites have been completed. Although there is considerable scatter in the data, degradation of both the glass fibers and the fiber/matrix interface is apparent. Tension and Mode I fracture tests on the resin alone have also been completed. The tests on the macroscopic samples show good correlation with the results from tests on the micro-composites up to a point. The tensile test results do reflect the changes in the fiber strength determined using the single-fiber test quite well. However, the compression samples were found to be too thick, leading to failure by crushing in the grips; the transverse failure mode may also reflect changes in the interfacial strength, but the test is not designed to measure that property. The results from the interlaminar fracture tests are also somewhat ambiguous because of the large increase in fracture toughness in the matrix material alone associated with moisture uptake: the interlaminar fracture results reflect a competition between decreasing interfacial strength and increasing matrix toughness.
A model for the degradation of the glass fibers has been developed. The model includes the effects of stress corrosion (attack by moisture of stressed fibers) and zero-stress aging (attack by moisture of unstressed fibers). From the rate of degradation (initially rapid, then slowing) it appears that zero-stress aging is the dominant effect. This point is important, as it implies that the presence of water at the surface of the glass fiber is sufficient to cause significant degradation, even without the influence of additional tensile stresses (which were quite large in the single-fiber fragmentation samples, caused by swelling of the matrix). It appears that the zero-stress aging behavior can be described by an activation energy, in which case elevated temperature can be used as a means of accelerated testing of the strength of the fibers. If the interfacial degradation and matrix changes could also be described with activation energies, elevated temperatures would be a means of accelerating the changes in these properties as well, with different activation energies associated wth different rates of change with temperature.
Additional experiments have been performed in collaboration with the Automotive Composites Consortium to investigate the effects of water and other fluids (notably windshield washer and brake fluids) on candidate E-glass/polyisocyanurate materials of interest to the automotive industry; the work at NIST has focused on the micro- mechanical tests for comparison with tests on macroscopic composites performed at Oak Ridge National Laboratory.
Finally, an additional collaboration to study the use of composites in infrastructure applications (such as bridges and roads) has begun with Northwestern University's Basic Industrial Research Laboratory, the University of Kentucky, Morrison Molded Fiber Glass and a number of State Departments of Transportation. The corrosion resistance and light weight of composite materials offer many advantages for civil engineering uses. An initial literature survey on this topic is nearly complete, and experiments have been performed to investigate the durability of pultruded composite materials that can be used to retrofit existing structures or for new construction.
Outputs
Publications
C. Schultheisz, C. Schutte, W. McDonough, K. Macturk, M. McAuliffe and S. Kondagunta, The Durability of Glass-Fiber/Epoxy Composites Evaluated from Single-Fiber Fragmentation Tests and Full-Scale Composites Immersed in Water, Proceedings of the Society for Experimental Mechanics, VIII International Congress on Experimental Mechanics, pp. 106-107, June 10-13, 1996, Nashville, Tennessee.
C. Schultheisz, W. McDonough, S. Kondagunta, C. Schutte, K. Macturk, M. McAuliffe and D. Hunston, Effect of Moisture on E- glass/Epoxy Interfacial and Fiber Strengths, American Society for Testing and Materials, 13th Symposium on Composite Materials: Testing and Design, ASTM STP 1242, Orlando, Florida, May 20, 1996.
K. Liao, R. Altkorn, S. Milkovich, J. Gomez, C. Schultheisz, L. Brinson and J. Fildes, Long-Term Durability of Composites in Secondary Infrastructure Applications, 28th International SAMPE Technical Conference, November 4-11, 1996, Seattle, Washington.
Presentations
C.R. Schultheisz, W.G. McDonough and G.A. Holmes, The Effects of Automotive Fluids on the Degradation of Glass/Isocyanurate Composites, American Institute of Chemical Engineers 1995 Annual Meeting, Miami Beach, Florida, November 16, 1995.
W.G. McDonough, C.R. Schultheisz and G.A. Holmes, Durability Issues in Glass/Polyisocyanurate Composites when Exposed to Automotive Fluids, Materials Research Society 1995 Fall Meeting, Boston, Massachusetts, December 1, 1995.
C. Schultheisz, Durability of Composites, The Materials Science Club of New York, Hoboken, New Jersey, April 4, 1996. Also, Exxon Research and Engineering Company, Annandale, New Jersey, April 5, 1996.
C. Schultheisz and G.A. Holmes, Durability of E-glass/Epoxy Composites Immersed in Water, Regional Technical Conference, Society of Plastics Engineers, October 28-29, 1996.
C. Schultheisz, Durability of Composites, University of Delaware, November 8, 1996.
Environmental Durability Studies: Development of Processing Methods to Fabricate Urethane Samples
W.G. McDonough and R.S. Parnas
Objective
The objective is to develop new processing procedures that will enable the preparation of urethane test specimens that can be used in the microstructure program and which are equivalent to materials made in industry by structural reaction molding (SRIM).
Technical Description
Single fiber specimens for the SFFT are typically prepared by pouring premixed resin into an open rubber mold, and then curing the resin in an autoclave. That method cannot work with rapidly curing resins of interest to the auto industry. Consequently, an injection molding procedure is being developed that will closely mimic the processing speed, temperature, and pressure observed in the SRIM process used with the resins of interest. The dogbone samples thus prepared will be tested by SFFT to determine if the interface strength is degraded under rapid processing conditions.
External Collaborations
Industrial: Automotive Composites Consortium (ACC), The Dow Chemical Co., and Bayer Corporation
Accomplishments
A mold from the Liquid Composite Molding Projects was modified by adding an insert that contains several dog bone shaped cavities. When the resin is injected, the cavities produce multiple fragmentation samples. A specially designed injection system was made to simulate the SRIM process. The isocyanate resin is put into one chamber and the polyol mixture is put into another chamber, and during processing, the liquids are combined together in a static mixer and injected into the mold. Preliminary trials with the Dow system were very encouraging, but a shift in requirements at the ACC has necessitated the shift to the Bayer system.
Flow visualization experiments have been carried out with nonreacting fluids to simulate the hydrodynamic loads experienced by the single fiber in each mold cavity. The fibers survived high speed injections, indicating that dogbone samples can be prepared in a rapid injection and cure process.
Impact
Outputs
Publications
G. Holmes, W. McDonough and C. Schultheisz, Comparison of Polyisocyanurate Networks Using Swelling Tests, Proc. of 11th Annual Advanced Composites Conference and Exhibition 11/95.
Presentations
W. G. McDonough, Durability of composites: processing and testing issues in model single fiber composites, Tokyo Institute of Technology, Tokyo, Japan, September 2, 1996.
W. G. McDonough, Evaluating a polyurethane/glass fiber composite for interfacial strength studies, National Fisheries University of Pusan, Pusan, South Korea, September 12, 1996.