POLYMER COMPOSITES PROGRAM
The Polymer Composites Program seeks to facilitate the introduction of lightweight, 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 lead to products that are superior and competitive in international markets. Industries as diverse as transportation, construction, marine, offshore oil, medical devices, and sporting goods have recognized those 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, two tasks were initiated, one on processing science and the other on microstructure, which constitute the composites program. A knowledge of microstructure and how it changes with time is a key factor for understanding long term performance. The automotive industry strongly influenced the composites program initially since the processing and durability issues span many automotive applications. Currently, the group interacts with the automotive, offshore oil, infrastructure, and aerospace industries.
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. Liquid composite molding (LCM) simulations were developed since this fabrication method is of 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 work has since broadened to involve the development of generic measurement tools to characterize material properties that control processing, for example, permeability and fiber architecture. General multiphase flow models are being formulated that can be applied to a variety of composites processes including those involving viscoelastic flow of thermoplastics. 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 six projects, including the support of a database and a SBIR collaboration to develop a commercial sensor system.
The work in the Microstructure Task is currently developing methods for characterizing and interpreting test data on hybrid reinforced composites. Hybrid reinforcements consisting of glass/carbon mixtures are currently of intense interest in the off-shore oil, infrastructure, and automotive industries due to the desirable mix of properties and economics that are obtainable. The particular microstructural features of interest include resin/fiber interfacial adhesion, especially under moisture attack, fiber architecture, and imperfections in composites such as voids, cracks and delaminations. The long term plan is to first develop effective test methods, and then to use the tests to identify the failure mechanisms in hybrids, particularly when synergistic reinforcement effects are present, and finally to formulate reliable predictive models useful for design. Microscale tests such as the single fiber fragmentation test have been put on a firm theoretical foundation, and are now being modified for multiple fiber tests designed to provide realistic estimates of the interface performance in hybrid composite systems. Imaging technology based on both classical optical and SEM, as well as nondestructive tomographic techniques are being applied to measure the fiber architecture and the imperfections in hybrids. Such microstructural data obtained before, during and after testing carefully produced specimens is expected to provide the information necessary to develop the structure-property relationships necessary for model generation. There are currently six specific projects in this Task, including an international round-robin on interface testing and a collaboration with the Automotive Composites Consortium to determine the effects of processing conditions on the interface of polymer composites.
The Composites Program promotes knowledge dissemination by actively organizing and participating in meetings worldwide. For example, the 1999 Gordon Research Conference on Composites, scheduled for January 10-15, 1999, was organized by Richard Parnas, and includes the world?s leading authorities on a number of topics including processing, interfaces, compression, nanocomposites, and ceramic composites. In addition, Composites Program staff helped the National Academy of Engineering organize materials oriented sessions at its Frontiers of Engineering meetings for the past three years, organized all the processing sessions at the 1998 American Society of Composites meeting, and serve as international advisory members for several European meetings.
Significant Accomplishments
Developed facility and test procedures for assessing the effects of processing on interface performance, and demonstrated its use by showing a large reduction of interface quality in a rapidly cured e-glass / vinyl ester system, relative to the slowly cured material.
Co-sponsored a workshop with the Department of Energy, Advanced Light Source (UC Berkeley), North Carolina State University, Carnegie Institution of Washington, Dow Chemical, and General Electric, entitled X-ray Microscopy of Synthetic and Natural Polymers: Towards Chemical Speciation at 10 nm Spatial Resolution. Assessment of Research Opportunities and the Need for X-ray Microscopy Facilities, which was held at NIST on May 11-12, 1998. The workshop resulted in a successful DOE proposal for $500,000 to develop an X-ray imaging facility at the Advanced Light Source devoted to the study of polymers.
Demonstrated the potential of the sandwich beam specimen for characterizing the mechanical properties of adhesives. The data highlighted the shortcomings of currently available analyses, but the simple testing geometry provides incentive for improving the analysis.
Liquid Composite Molding: Development of Permeability Database
R. Parnas, K. Flynn, M. DalFavero1, and H. Friedman2
1NIST, Standard Reference Data Program
2Textile Research Institute, Princeton, NJ
Objective
The objective is to establish a data base of permeability values for use as a design tool by the composites industry.
Technical Description
Permeability measurements conducted over several years at NIST are documented, collected, and placed into a Clipper based database. External sources of reliable data are identified and their measurements are being documented for inclusion in the database.
External Collaborations
Henry Friedman, Textile Research Institute - perform comparative measurements.
Raymond Gauvin, University of Montreal - provide data.
Christopher Rudd, University of Nottingham - provide data.
Staffan Lundström, Chalmers University, Sweden - perform comparative measurements.
Planned Outcomes
The database released in 1998 is expected to be used by molders to help design their processes and parts. Version 2 of the database is planned for release in 1999 with an expanded data set and enhanced graphics display.
Accomplishments
Version 1.0 of the Permeability Database, consisting of work done at NIST by NIST personnel, was released in collaboration with the NIST Standard Reference Data Division, and has resulted in the sale of 11 copies to date. Version 2.0 of the database is being developed by documenting data submitted by external collaborators, and the size of the database has been doubled.
Impacts
NIST permeability data are being used in flow simulations at several companies, including ATP recipients, in the automotive and aerospace industries. 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, Preform Permeability, in RTM for Aerospace Applications, Chapman & Hall, NY, 1998, Ch. 8.
R. S. Parnas, K. M. Flynn and M. DalFavero, A Permeability Database for Composites Manufacturing by Liquid Molding, Polymer Composites 18(5), 623-33, 1997.
Liquid Composite Molding: Development of Permeability Measurement Techniques
R. S. Parnas, F. R. Phelan, R. C. Peterson, and H. Friedman1
1Textile Research Institute, Princeton, NJ
Objective
To develop new permeability measurement techniques for assessing flow behavior in deformed and model materials.
Technical Description
Permeability measurements are continuing at the Textile Research Institute with an emphasis on the permeability of fabrics deformed around curves, as would be found in molds of complex shape. Measurements conducted on a 3-D woven fabric were completed, and in the flat sections of the material the measurements agreed well with previous measurements conducted at NIST on the same material. Although difficult to quantitate precisely, the measurements in the curved regions of the 3-D fabric clearly indicated a substantial increase in flow resistance. Quantitation of the permeability measurements in the curved regions of the fabric was difficult due to nonuniform flows and an entrance effect. Partly to overcome these difficulties and partly to contrast the behavior with another material, experiments were conducted more recently with a continuous strand random mat (CSRM). The entrance effect is not present in the CSRM, as indicated by previous work at NIST, and the flow moves through the fabric more uniformly. We therefore expect the data analysis of the CSRM to be more straightforward than with the 3-D woven fabric, allowing quantitation of the permeability reduction in the curved region of the material.
Measurements are conducted with the 3-D woven fabric and with the CSRM at 3 different radii of curvature. A measurement system has been built to provide constant flows to L-shaped molds. The molds are designed to maintain constant channel thickness around the curves, and use 5.08 cm (2") thick plexiglas external sections to provide for flow visualization. A video capture system, controlled by a LabView software application, works in conjunction with the flow apparatus to permit detailed analysis of the progression of the flow front throughout the experiment.
A model porous medium measurement system was constructed at NIST to verify the microflow predictions of the Lattice-Boltzmann flow model. A specialized mold was designed and built to permit the anchoring of model media components in a variety of arrangements within the mold. Model media are then constructed by press-forming CSRM to desired fiber volume fractions, and stamping out circular or elliptical shapes of the pressformed material. These shapes are arranged and anchored in the mold in square or hexagonal arrays to model the arrangement of tows in a preform. Experiments are conducted by flowing aqueous corn syrup solutions through the array under constant injection pressures, recording the filling behavior on video tape and measuring steady state flows and pressures after the mold filled.
External Collaboration
Henry Friedman, Textile Research Institute - performing flow studies in deformed fabrics.
Planned Outcomes
Develop knowledge-based rules to guide process designers in controlling resin flows in complex molds. Verify a microflow simulation tool for predicting permeability and unsaturated flow behavior in porous media.
Accomplishments
The reduction in permeability due to fabric curvature was demonstrated for flow around corners by building specialized flow molds with 90° curves of several radii of curvature. Material deformation in curves and at the corners of molds is responsible for modeling difficulties, and characterization data for such deformed material is very sparse.
Void formation and depletion dynamics were successfully imaged in model porous media with enough resolution to permit direct comparison with model calculations.
Outputs
Publications
Henry L. Friedman, Alexander V. Neimark, David R. Salem, and Richard S. Parnas, Visualization and Quantification of Forced In-Plane Flow Through Deformable Porous Media, Proc. American Society of Composites Conference, Baltimore, MD, Sept. 21, 1998, in press.
R. C. Peterson, and F. R. Phelan Jr., An Investigation of Void Formation Dynamics in Fiber Preform Materials, Proc. American Society of Composites Conference, Baltimore, MD, Sept. 21, 1998, in press.
H. L. Friedman, R. A. Johnson, B. Miller, D. R. Salem, and R. S. Parnas, Forced In-Plane Flow Through Complex Deformable Structures: Influence of an Imposed Curve, Polymer Composites 18(5), 663-71, 1997.
H. L. Friedman, R. A. Johnson, V. Gusev, A. V. Neimark, D. Buvel, D. R. Salem, and R. S. Parnas, Visualization and Quantification of Forced In-Plane Flow through Deformable Porous Media, Polymer Composites, in press.
Presentations
Henry L. Friedman, Alexander V. Neimark, David R. Salem, and Richard S. Parnas, Visualization and Quantification of Forced In-Plane Flow Through Deformable Porous Media, American Society of Composites Conference, Baltimore, MD, Sept. 21, 1998.
Liquid Composite Molding: Development and Verification of Process Simulation Models
F. R. Phelan Jr., K.M. Pillai1, and C. L. Tucker III1
1University of Illinois, Urbana, IL
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 perform deformation and heat transfer. The model will be developed specifically to simulate injection/compression liquid composite molding (I/CLCM) for the automotive industry and their suppliers.
Technical Description
LCM process optimization has typically been done with time-consuming and expensive trial and error methods on full scale equipment. Simulation models have the potential for greatly reducing the cost and increasing the speed of this task. In previous work, a simulation called CRIMSON which models the mold filling phase of LCM was developed. CRIMSON is based on a finite element/control volume numerical solution to the governing transport equations in which the momentum equation is expressed by Darcy?s law. 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 current phase of this project, CRIMSON has been 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 satisfying their need to fabricate large structural parts in fast cycle times. Research has revealed that meeting the required process constraints with pure-injection LCM, results in excessively high injection pressures that induce undesirable fluid-structure interactions involving perform, foam core, or tool deformation. I/CLCM differs from conventional "injection-only" LCM in that subsequent to perform 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 perform to the desired net shape and volume fraction while distributing the initial shot of resin throughout the part. During the initial injection stage of ?open mold? I/CLCM there is a gap between the perform and the upper tool surface. The strategy is to try and fill the gap region with fluid first, and then use the compression step to drive the fluid into the perform in the thickness direction. However, because there is some penetration of resin into the perform during the injection phase, during the compression step there is a combination of in-plane and through-thickness flow. Thus, a fully 3-D flow model is needed.
External Collaborations
Dave Pinella, Structural Dynamics Research Corp. - I-DEAS/CRIMSON interface development.
Doug Denton, Automotive Composites Consortium - I/CLCM flow simulation.
Gilbert Carpenter, Northrup/Grumman - RTM flow simulation.
Chuck Stuart, The Budd Co. - RTM flow simulation.
Chihdar Yang, Specialty Plastics, Inc. - RTM flow simulation.
Kurt Schultz, Louisiana State University - RTM flow simulation.
Chuck Tucker and Krishna Pillai, University of Illinois - perform deformation modeling.
Planned Outcome
Provide the automotive industry and other interested organizations with simulation tools for design and optimization of the liquid molding process.
Accomplishments
This year, work was completed in cooperation with the University of Illinois at Urbana-Champaign (UIUC) to develop a numerical simulation of the open mold I/CLCM process. There are three main elements to the flow modeling algorithm, the preprocessor module, the perform compression module, and the flow module. The preprocessor and compression modules were developed in collaboration with UIUC. In the pre-processor, a 3-D flow mesh is constructed as input for the CRIMSON program. Two inputs are needed in the pre-processing step: a 2.5-D mesh of the final molded geometry and all information on the perform "layup" in this geometry. This information is normally known by a part designer. From this information, a 3-D mesh of the final preform geometry and "spines" are constructed. "Spines" are vectors that run through groups of co-linear nodes from the lower surface to the upper surface of the FE mesh. The mechanical deformation module models the non-linear deformation of the fibrous preform during compression in three-dimensional geometries. Local 1-D mechanics with no shear deformation are assumed in the model for the sake of robustness. In response to compression, the preform deforms along the "spines" constructed in the preprocessing step as the spines rotate in response to the imposed deformation. Linking this module with CRIMSON enables CRIMSON to compute the volume changes that occurs in the elements during compression of the preform which is necessary in order to track the movement of the flow front.
In an effort to link the CRIMSON software with the user community, Structural Dynamics Research Corp. (SDRC) has developed a second generation graphical user interface to interface 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. Phase II of this effort includes Injection/Compression parameters.
Impact
The simulation software has been transferred to the ACC, Budd, GM, Specialty Plastics and Northrup/Grumman, and is regularly used as a design tool.
Outputs
Publications
F. R. Phelan Jr., Analysis of In-Plane Injection/Compression Liquid Composite Molding, Composites A, in press.
K. M. Pillai, C. L. Tucker III and F. R. Phelan Jr., Numerical Simulation of Injection/Compression Liquid Composite Molding. Part 1: Mesh Generation, Composites A, submitted.
K. M. Pillai, C. L. Tucker III and F. R. Phelan Jr., Numerical Simulation of Injection/Compression Liquid Composite Molding. Part 2: Preform Compression, Composites A, submitted.
K. M. Pillai, C. L. Tucker III, and F. R. Phelan Jr., Injection/Compression Liquid Composite Molding 3-D Flow Analysis, Proc. American Society of Composites Conference, Baltimore, MD, September, 1998, in press.
Presentations
F. R. Phelan Jr., K. M. Pillai, and C. L. Tucker III, Injection/Compression Liquid Composite Molding 3-D Flow Analysis, AIChE Annual Meeting, Los Angeles, CA, November, 1997.
Liquid Composite Molding: Development of Processing Models for Viscoelastic Flows
A. M. Reiff, and F. R. Phelan Jr.
Objectives
The objectives are to develop and apply numerical models to describe the flow of viscoelastic fluids in composites processing operations such as liquid molding, pultrusion and injection molding.
Technical Description
Due to the nonlinear rheological behavior of curing thermosets and of polymer melts, and the multi-phase nature of injection flows, the processing of composite materials is little understood and quite difficult to analyze. Lattice-Boltzmann (LB) methods have been shown to be quite powerful for modeling multi-phase flows, but have been limited to Newtonian fluids. In this project, a generalized Lattice-Boltzmann scheme is developed to simulate two dimensional viscoelastic flow of fluids. We solve the Navier-Stokes equations, representing the conservation of mass and momentum. Along with the conservation of mass and momentum, we use the upper Maxwell family of constitutive laws for the extra stress to describe the viscoelastic rheology of the fluid.
The basic approach in LB is to solve the discrete Boltzmann equation for the particle velocity distribution function on a lattice. Each particle velocity distribution function represents a packet of particles moving with a fixed velocity along the lattice. The evolution of these distribution functions is governed by the discrete Boltzmann equation which is approximated by a single relaxation time collision operator called the BGK approximation. This collision term forces particle distributions toward their local equilibrium value. The Boltzmann equation is discretized by permitting velocities of magnitude zero or one and directions determined by the spatial lattice geometry. For example, with a two dimensional hexagonal lattice, seven velocities are permitted including six of magnitude one connecting a given lattice point to its six nearest neighbors, and one velocity of magnitude zero.
The macroscopic flow quantities, density and velocity, are recovered by summing over the particle distribution function solution of the discrete Boltzmann equation.
Planned Outcome
To model multi-phase viscoelastic flow and provide the developed simulation tools to manufacturers for optimizing the production of composite materials.
Accomplishments
Within the LB framework, a second order accurate numerical scheme was derived for viscoelastic constitutive laws in the upper Maxwell family. The Lagrangian coordinate system used in the LB method proved to be convenient for discretizing the convected stress terms in the constitutive law. Assuming isotropic conditions along with the condition that the macroscopic stress matches the nodal stress, we derived the equilibrium distribution functions for the seven velocity vectors on the 2-D hexagonal lattice.
Using the LB method described, the conserved quantities mass and momentum can be found. Observing that the LB update is explicit, the new values for mass and velocity are utilized in the approximation of the stress tensor. From the dimensionless form of the upper Maxwell constitutive laws, we derive the second order approximation to the extra stress required in the LB calculation of the equilibrium distribution function.
At this time, the numerical model simulates flow in a square cavity with convex impermeable tows. We assume symmetric conditions at the walls parallel to the flow direction; at the walls perpendicular to the flow direction, we assume periodic velocity along with a constant pressure (or density ) gradient. Along the boundary of the tow, we assume a no-slip condition, i.e. U = 0. The next phase of this project will be to alter the no-slip boundary conditions to allow for permeable tows.
Liquid Composite Molding: Process Monitoring and Control
J. P. Dunkers, R. S. Parnas, K. M. Flynn, C. Zimba, S. Keuh1, and K. Murphy2
1University of Delaware, Newark, DE
2F&S Inc., Blacksburg, VA
Objectives
To develop measurement methods based on optical fibers for monitoring the flow and reaction processes that occur in composites manufacturing, and to begin commercialization efforts for such technology.
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, and flow inconsistencies are major contributors to composite non-uniformity. The cure monitoring work focuses on fluorescence and near infrared spectroscopies, using an optical fiber drawn from commercially available high refractive index glass as the sensing element. Optical systems are designed and built to provide high speed spectral acquisition and chemometric methods are explored to provide equally high speed spectral analysis. The interfacial sensitivity of the evanescent wave sensors is determined with optical theory that quantifies the coupling between the excitation radiation and the fluorescence radiation.
The same optical fiber sensors used for cure monitoring are also being explored for flow monitoring by taking advantage of the change in signal strength that occurs as a fiber is covered by fluid when operated in evanescent wave mode. The effects of fiber bending and local environment around the fiber are being explored. In this work, flow simulation is being used to optimize the optical fiber trajectory in the mold to provide flow data most indicative of proper mold filling.
External Collaborations
Kent Murphy, F&S Inc. - developing a commercial sensor system.
Rob Bannerjee, EDX - developing quality control methods for sandwich panel products.
Dennis Sourlas, University of Missouri - developing process control algorithms.
Suresh Advani and Sylvia Kueh, University of Delaware -applying sensors to flow monitoring.
Planned Outcome
Demonstrate usefulness of process monitoring to the composites industry and indicate a potential low cost route with optical fiber sensors. Assist industry to develop a commercial sensor system.
Accomplishments
An SBIR subtopic was announced and a proposal from F&S selected for phase I feasibility study. In this study the market for an optical fiber sensor system is being assessed in the composites industry. The technical design of such a system is being explored by designing fibers and electronics to multiplex multiple sensors on a single fiber.
The potential to use a well developed fiber optic cure sensor system for flow monitoring during mold filling has been demonstrated with simulations. Using a reasonable sensor response function, the ability of a single fiber sensor system to detect flow anomalies such as edge effects was demonstrated in simple edge filled molds as well as in complex molds containing inserts. This result leads to the intriguing possibility of a simple and fast system suitable for manufacturing that can provide real time information for controlling mold filling.
Outputs
Publications
Joy Dunkers, Kathleen Flynn, Mitchell Huang, and Walter McDonough, Fourier Transform Near-Infrared Monitoring of Reacting Resins Using and Evanescent Wave High-Index Fiber-Optic Sensor, Applied Spectroscopy 52(4), 552 (1998).
Richard Parnas, Joy Dunkers, and Raymond Neff, Monitoring Composites with Optical Fiber Sensor Systems, Proc. SPE ANTEC 1998, Atlanta, GA, May, 1998.
Richard S. Parnas, Process Monitoring and Control for High Speed Composites Manufacturing, Proc. SPE Thermoset Regional Technical Conference 1998, Chicago, IL, March, 1998.
Dennis Sourlas, Susmito Naha, Gary Patterson, and Richard Parnas, Study on the Estimation and Control of a Liquid Composite Molding Process, Proc. Automatic Control Conference, Philadelphia, PA, June 24-26, 1998.
Sylvia R. M. Kueh, Suresh G. Advani, and Richard S. Parnas, Virtual Sensor Simulation Study of Flow During Resin Transfer Molding Process, Proc. American Society of Composites Meeting, Baltimore, MD, Sept. 22-25, 1998, in press.
Presentations
Richard S. Parnas, Optical Fiber Sensors for Monitoring and Control of Composite Processing, Second International Conference on Composite Materials for Offshore Operations, Houston, TX, October 29, 1997.
Richard S. Parnas, Optical Fiber Monitoring of Fast Systems, American Institute of Chemical Engineers 1997 Annual Meeting, Los Angeles, CA, November 19, 1997.
Richard S. Parnas, Experiments with Evanescent Wave Optical Fiber Sensors, 1997 International Mechanical Engineering Congress & Exposition, Dallas, TX, November 21, 1997.
Richard S. Parnas, Mode Coupling Theory for Evanescent Wave Optical Fiber Sensors, poster at Gordon Research Conference on Composites, Ventura, CA, January 6, 1998.
Richard Parnas, Joy Dunkers, and Raymond Neff, Monitoring Composites with Optical Fiber Sensor Systems, SPE ANTEC ?98, Atlanta, GA, May, 1998. Best Paper Award.
Richard S. Parnas, Process Monitoring and Control for High Speed Composites Manufacturing, SPE Thermoset Regional Technical Conference 1998, Chicago, IL, March, 1998.
Dennis Sourlas, Susmito Naha, Gary Patterson, and Richard Parnas, Study on the Estimation and Control of a Liquid Composite Molding Process, Automatic Control Conference, Philadelphia, PA, June 24-26, 1998.
Sylvia R. M. Kueh, Suresh G. Advani, and Richard S. Parnas, Virtual Sensor Simulation Study of Flow During Resin Transfer Molding Process, American Society of Composites Meeting, Baltimore, MD, Sept. 22-25, 1998.
Liquid Composite Molding: Interface Sensitive Optical Fiber Sensor
J.P. Dunkers, R.S. Parnas, J. Lenhart1, and J. VanZanten1
1The Johns Hopkins University, Baltimore, MD
Objective
Develop a method of measuring the chemical and physical behavior of the polymer matrix in the first 100 Å around glass fibers.
Technical Description
Fluorescence monitoring becomes more practical by grafting the fluorophore onto the glass fiber instead of dissolving the fluorophore into the resin. The fluorophore responds to its environment by undergoing a fluorescence shift to shorter wavelengths as the viscosity and polarity of the environment increases. A fluorescing silane coupling molecule, Robello siloxane (RBS), is being synthesized and grafted on model flat glass and silicon surfaces. The RBS is grafted onto the surfaces strongly diluted by other typical co-silane coupling agents to minimize fluorescence self quenching and to mimic the typical silane chemistry used in composites as closely as possible. The silane hydrolysis and deposition conditions are systematically varied with respect to hydrolysis and deposition time, solvent, and co-silane. The grafted layer is assessed using fluorescence and contact angle measurements, atomic force and scanning electron microscopy, and infrared spectroscopy.
External Collaborators
William Birch, Corning, France - silane deposition procedures.
Joe Lenhart and John VanZanten, The Johns Hopkins University - silane layer characterization.
Steve Pollack, Howard University - RBS synthesis.
Planned Outcome
To use the interface sensitive fiber optic sensor in conjunction with a commercial sensor system for quality control of industrial composites manufacturing.
Accomplishments
A room temperature deposition procedure of the RBS and co-silane upon glass was developed. The layer was shown to be strongly bonded to the glass by performing contact angle measurements before and after washing with solvent. The potential for the grafted RBS dye to be used as an interface sensitive cure monitoring sensor was demonstrated. Both fluorescence intensity changes and spectral shifts from the RBS can be followed when an epoxy resin cures over the mixed coupling agent layers. It was also shown that layer thickness had an effect on the emission shift upon cure. For the RBS/ isocyanopropyltrimethoxysilane system, a 1 µm layer underwent a blue shift of 17 nm whereas a 10 µm layer only underwent a 4 nm shift upon cure. A correlation was found between the polarity of the co-silane and the emission wavelength of the RBS. With the more polar co-silanes the emission is around 640 nm and with non-polar co-silanes the emission is near 540 nm. In addition, the polarity of the co-silane also influenced the emission intensity, with the polar co-silanes exhibiting an intrinsically lower fluorescence intensity than the non-polar silanes. Using atomic force microscopy, the roughness of the RBS/co-silane film on the glass surface was measured and was found to vary drastically with co-silane. Using these results, it is now possible to design a sensor that has the optimized properties for the interface sensor and the cure sensor.
Output
Publications
Joseph L. Lenhart, John VanZanten, Joy P. Dunkers, Carl G. Zimba, Steven K. Pollack, and Richard S. Parnas, A Fiber Optic Sensor for Composite Cure Monitoring, Proc. American Society of Composites Meeting, Baltimore, MD, Sept. 21-23, 1998, in press.
Presentations
J. L. Lenhart, J. VanZanten, J. P. Dunkers, C. G. Zimba, S. K. Pollack, and R. S. Parnas, A Fiber Optic Sensor for Composite Cure Monitoring, American Society of Composites Meeting, Baltimore, MD, Sept. 22, 1998.
Microstructure Studies: Chemical Imaging
C. G. Zimba, and R. S. Parnas
Objective
Apply chemical imaging techniques in the infrared and x-ray to measure the spatial distribution of individual chemical species in a variety of polymeric materials, including composites, blends, and other multi-component polymer systems.
Technical Description
In most cases, the polymer systems of interest are not homogeneous or single-components, but are blends, composites, or copolymers. To understand the nature of the molecular and physical processes in these heterogeneous systems, it is important to understand both the spatial distribution and the chemical behavior of the various components within the polymer matrix. However, there is typically a trade-off between spatial resolution and chemical specificity. Techniques, such as transmission or scanning electron microscopies (TEM and SEM), have high spatial resolution but limited chemical information, while infrared microscopy has low spatial resolution and high chemical content.
While many chemical imaging tools exist, those that use photon probes offer the best combination of chemical identification with spatial resolution. Two such techniques have been the focus of this effort: infrared microspectroscopy and x-ray microscopy. The infrared technique allows spectra of areas as small as 10-20 microns to be acquired with conventional IR sources and microscopes. While this is still a relatively large size scale, the infrared spectrum affords one of the best probes of molecular structure and morphology of polymers. X-ray microscopy is a useful adjunct to infrared analysis, providing somewhat less chemical information but with a spatial resolution of 50 nm. Using Near Edge X-ray Absorption Fine Structure (NEXAFS) in the soft x-ray region as the spectroscopic probe, x-ray microscopy is sensitive to the hybridization and molecular bonding of carbon. Thus, different moieties of carbon, i.e., carbonyl, phenyl, nitrile, can easily be distinguished. When combined with an imaging capability, this spectroscopic sensitivity makes it possible to ascertain the spatial distribution of chemically different species. Using both infrared and x-ray microscopy, many systems, including laminates, blends, composites, and block copolymers, have been characterized.
External Collaborations
Nicolet Instrument Corporation and Spectra-Tech, Inc. - infrared array detectors.
Rina Dukor, Vysis, Inc. - infrared microspectroscopy of single cells in breast cancer biopsies.
Sanjeeva Murthy, Allied Signal - NEXAFS microscopy of nanocomposites.
E. L. Thomas (MIT) C. Ober (Cornell) - NEXAFS microscopy for block copolymers.
C. Jacobsen, SUNY Stoneybrook, - improved spatial resolution in NEXAFS microspectroscopy.
Internal Collaborations:
E. Heilweil (Physics) - infrared microspectroscopy using array detectors
N. Eidelman (ADAHF) - infrared and NEXAFS microscopy of calcified tissue
Planned Outcomes
Dedicated X-ray microscope for polymers at the Advanced Light Source, and improved infrared imaging equipment available commercially.
Accomplishments
The most noteworthy examples of the use of infrared and x-ray microscopy include the characterization of calcified tissue and the observation of molecular orientation in individual lamellae of a block-copolymer.
The calcified tissue sample was a cryo-microtomed cross-section taken from a larger piece of bovine pericardium which was pretreated with gluteraldehyde, implanted subdermally in a laboratory rat for 28 days, retrieved, washed with distilled water and lypholized to remove excess water. In these sections, it was observed by x-ray microscopy (Beamline X1A, NSLS) that the calcified growths were irregularly shaped nodules of 3 to 6 microns in size and that they were preferentially deposited on the side of the implanted pericardium away from the muscle tissue of the rat. Infrared analysis of the tissue, using the infrared microscope at Beamline U4IR at the NSLS with a Cu:Ge detector, showed that the calcified deposits were composed of biological apetite. This is an excellent example of the combined power of infrared and x-ray microscopy.
The block copolymer, p(S-b-HIC), having a styrene (S) block of 9K and a hexylisocyanate (HIC) block of 245 K molecular weight, respectively, spontaneously organizes into a lamellar structure. Previous analysis by TEM has shown that the styrene blocks form arrowhead like structures that alternate in orientation in adjacent lamellae. Electron diffraction from an area comprised of 10-20 lamellae shows that the HIC blocks form highly-crystalline regions which alternate in orientation between adjacent lamellae. X-ray microscopy of these thin cast films also shows the arrowhead morphology of the styrene blocks as seen with TEM. Moreover, the orientation of individual lamellae has been observed using dichroic imaging. This is the first time that orientation in block copolymers has been directly observed on a size scale significantly smaller that than the lamellar spacing.
Improvements in instrumentation and methodology have also been achieved. In x-ray microscopy, methods have been developed and used to obtain NEXAFS spectra of smaller sample areas, approaching the spatial resolution of the optics. Previously, it has not been possible to achieve this due to uncertainties in the position of the zone plate focus as the x-ray energy was changed. By obtaining images at finely spaced energies and then aligning them to account for any positional drift, spectra from single pixels can now be obtained. This has proven to be quite useful for the characterization of features that are close in size to the spatial resolution of the microscope. In infrared spectroscopy, an existing infrared microscope at NIST has been modified to accept a two-dimensional infrared array detector. By using an array for a detector, it is possible to simultaneously obtain infrared spectra from the sample area corresponding to each pixel, greatly reducing the time required to characterize large sample areas. While this experiment is still under development, the imaging and spectroscopic capability has been demonstrated for both transmission and reflection samples. Future work will emphasize time- dependent phenomena.
The Division of Materials Sciences in the Office of Basic Energy Science of the Department of Energy awarded a $450K grant to a multi-partner team, including NIST, to build a new x-ray microscope that will be dedicated to polymer science. This new facility, which will be built at the Advanced Light Source in Berkeley, CA, will greatly augment the amount of beam time available for the investigation of polymers and make possible more extensive and detailed studies which thus far have been impossible due to limited beam time. While this new x-ray microscope will use proven technology for the most part, improvements that have a high benefit/cost ratio, such as improved detectors, will also be pursued.
Outputs
Workshops
Infrared Spectroscopy Using Synchrotron User Facilities, 24th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), Providence, RI, October 29, 1997. This was sponsored by the User Executive Committee of the National Synchrotron Light Source, and Carl Zimba helped organize and teach it.
X-ray Microscopy of Synthetic and Natural Polymers: Towards Chemical Speciation at 10 nm Spatial Resolution. Assessment of Research Opportunities and the Need for X-ray Microscopy Facilities, NIST, Gaithersburg, MD, May 11-12, 1998. This was co-sponsored by NIST, Department of Energy, Advanced Light Source, North Carolina State University, Carnegie Institution of Washington, Dow Chemical, and General Electric.
Publications
C. G. Zimba, P.A. Martoglio, and J. A. Reffner, Infrared Microspectroscopy with Synchrotron Radiation Sources, Applied Spectroscopy, submitted.
C. G. Zimba, Application of Two-Dimensional Correlation Methods to Infrared Microspectroscopy, Applied Spectroscopy, submitted.
Presentations
Carl G. Zimba, Developments in Infrared Microspectroscopy Using Synchrotron Radiation, 24th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), Providence, RI, October 30, 1997.
Carl G. Zimba, Chemical Imaging in the Infrared and X-ray Regions, Syracuse University, Department of Chemistry, Syracuse, NY, March 31, 1998.
Carl G. Zimba, Chemical Imaging in the Infrared and X-ray Regions, Society of Applied Spectroscopy, Chicago, IL, April 14, 1998.
Carl G. Zimba, Chemical Imaging in the Infrared and X-ray Regions, Society of Applied Spectroscopy, Reading, PA, April 15, 1998.
Carl G. Zimba, X-ray Microscopy: Introduction and Applications, Workshop on X-ray microscopy of synthetic and natural polymers, NIST, Gaithersburg, MD, May 11-12, 1998.
Naomi Eidelman, Catheryn Jackson, and Carl G. Zimba, Measuring and Mapping the Calcium in Calcified Bovine Pericardium Implants Retrieved from Rats by Scanning Transmission X-ray Microscopy, poster at Workshop on X-ray Microscopy of Synthetic and Natural Polymers, NIST, Gaithersburg, MD, May 11-12, 1998.
Carl G. Zimba, X-ray Microscopy of Multilayer Polymer Coatings, poster at workshop X-ray Microscopy of Synthetic and Natural Polymers, NIST, Gaithersburg, MD, May 11-12, 1998.
Carl G. Zimba, and Chris Jacobsen, NEXAFS Microspectroscopy Using STACK, poster at workshop on X-ray Microscopy of Synthetic and Natural Polymers, NIST, Gaithersburg, MD, May 11-12, 1998.
Carl G. Zimba, Chemical Imaging: Application to Polymers, Bundesanstalt fur Materialforschung und -prufung (BAM), Berlin, Germany, June 29, 1998.
Carl G. Zimba, Chemical Imaging in the Infrared and X-ray Regions, University of Heidelberg, Department of Chemistry, Heidelberg, Germany, July 1, 1998.
Carl G. Zimba, Microspectroscopy and Imaging of Polymers, Max Planck Institut fur Polymerforschung, Mainz, Germany, July 2, 1998.
Carl G. Zimba, Application of 2D Correlation Techniques to Infrared and X-ray Microspectroscopy Data, poster at International Symposium on Advanced Infrared and Raman Spectroscopy, Vienna, Austria, July 6-9, 1998.
Carl G. Zimba, Chemical Imaging in the Infrared and X-ray Regions, Laboratoire pour l'Utilisation du Rayonnement Electromagnetique (LURE), Orsay, France, July 15, 1998.
Carl G. Zimba, Polymer Applications of X-ray Microspectroscopy and Imaging, poster at Polymer Physics Gordon Research Conference, Newport, RI, August 16-21, 1998.
Microstructure Studies: Improve the Utility of the Single Fiber Fragmentation Test
G. A. Holmes, R. C. Peterson, W. G. McDonough, D. L. Hunston, R. S. Parnas, and J. F. Cheng1
1Massachussetts Institute of Technology, Cambridge, MA
Objective
The objective in this project is to develop the tools required to obtain reliable measurements of the interfacial adhesion in composites having rheologically complex matrices.
Technical Description
Perhaps the single most important problem in the measurement of fiber-matrix interface strength is the lack of standardization. To address this issue, NIST and Michigan State University have organized an international program under the auspices of the Composites Technical Working Group of VAMAS, the Versailles Project on Advanced Materials and Standards. This is an international organization that promotes pre-standards research. The objectives are to establish consensus test procedures, to conduct round robins to verify these procedures, and to provide a forum for identifying critical research issues, exchanging results, and encouraging cooperation among researchers active in the area. The initial focus is the fragmentation test, but the work may expand to other measurement methods if the program is successful. Participants include 15 laboratories in 8 countries.
The interfacial shear strength (or shear stress transfer coefficient) is typically calculated from experimental test data using models that assume the matrix material is linear elastic or elastic-perfectly plastic. The polymer matrix is rarely linear elastic or elastic-perfectly plastic. This project seeks to quantify the impact of non-ideal matrix behavior on the interfacial shear strength determined by the most widely used test method, the single fiber fragmentation test. This will be accomplished by: (1) monitoring the change in load in single fiber fragmentation test (SFFT) specimens with increasing strain, (2) monitoring the evolution of the fiber fragmentation process with increasing strain, (3) investigating the impact of loading rate on the matrix modulus and fragmentation process, (4) development of shear stress transfer models to account for the impact of matrix property changes with increasing strain and strain rate, and (5) comparing results from the new models with values predicted by traditional linear elastic and elastic-plastic based shear stress transfer models.
An important aspect of the work is to assess the durability of adhesion promoting silane coupling agents that are applied to the fiber surface before composite manufacturing. These coupling agents are designed to promote adhesion of the matrix to the embedded fiber and provide a protective barrier against moisture attack. Because moisture absorption alters the properties of the matrix material, traditional data analysis methods that do not account for matrix property changes cannot decouple the effects of moisture on the matrix from the effects of moisture on the interface. Hence, assessing the durability of a fiber coating requires decoupling the changes resulting from differences in matrix properties from the true changes in interfacial shear strength due to degradation of the chemical bonds formed between the silane coupling agent and the fiber. To accomplish this aspect of the project, model systems will be prepared in which the impact of resin stiffness on the calculated interfacial shear strength can be determined independently of fiber-matrix interface degradation and fiber strength degradation. The matrix behavior from these model systems will be compared with resin systems plasticized by moisture to assess the applicability of this approach. Analytical models will be developed to predict the apparent change in interfacial shear strength arising from changes in matrix properties. These values will be compared with values derived from experimental data to determine the durability of E-glass coatings.
External Collaborations
Eric Pohl, OSI is collaborating in the study of durability of industry coatings.
Larry Drzal, Michigan State University is cooperating test method development.
G. Mao, Wayne State University is using atomic force microscopy to study silane coatings.
John Nairn, University of Utah is investigating dynamics of debond region formation.
Leigh Phoenix, Cornell University is correlating NIST experimental data with theoretical models.
Graham Sims, Versailles Advanced Materials and Standards Program (VAMAS) and 14 other participating laboratories are participating in an effort to establish standardized tests for interfacial shear strength.
Planned Outcome
Establish international protocol for single fiber fragmentation test.
Accomplishments
VAMAS Program: During the past year, a consensus test procedure was developed, and a model material system was selected. A single batch of samples is now being prepared with this system, and specimens will be distribution to all participants. The round robin with these samples is scheduled for completion in December 1998.
Research on Fragmentation Test and Viscoelastic Characterization: An empirical nonlinear viscoelastic constitutive equation was developed to model the behavior of the resin during the single fiber fragmentation test. As a result of these modeling efforts, a nonlinear viscoelastic micromechanics model was developed to more accurately determine the interfacial shear stress transfer coefficient (I-STC). Noting that viscoelastic materials are history dependent, a reasonable estimate of the I-STC was obtained by using the stiffness of the matrix at the end of the test. This approach allows comparison of the I-STC from different resin systems.
Rate Effects: To assess the impact of strain rate on the single fiber fragmentation process, two sets of model E-glass composites were made from diglycidyl ether of bisphenol-A / m-phenylenediamine (DGEBA/m-PDA) epoxy resin and polyisocyanurate. When the effective strain rate of the test was decreased, the I-STC of the E-glass model composites made from DGEBA/m-PDA increased. This is opposite to what one might expect based solely on viscoelastic behavior. Similar tests on model composites composed of E-glass and polyisocyanurate, however, do not show dramatic changes in the I-STC with decreasing strain rate. These results indicate that the viscoelastic behavior of the matrix must be considered when establishing a single fiber fragmentation testing protocol.
Outputs
Publications
G. A. Holmes, R. C. Peterson, D. L. Hunston, W. G. McDonough, and C. L. Schutte, Non-linear Viscoelasticity in a DGEBA/m-PDA Resin System and Its Impact on E-Glass Interfacial Shear Strength Measurements, ASTM STP 1357, in press.
D. L. Hunston, G. A. Holmes, R. C. Peterson, Viscoelastic Properties of a Resin Commonly used in the Single Fiber Fragmentation Test, Proc. American Society of Composites Meeting, Baltimore, MD, Sept. 21-23, 1998, in press.
G. A. Holmes, R. C. Peterson, and D. L. Hunston, Modeling of Multi-Step Nonlinear Stress Relaxation in DGEBA/m-PDA Epoxy Resins, Proc. of the 2nd Int. Conference on Mechanics of Time Dependent Materials, Pasadena, CA, March, 1998, p. 26.
G. A. Holmes, R. C. Peterson, D. L. Hunston, and W. G. McDonough, The Influence of the Matrix Modulus on the Interfacial Shear Strength Parameter, Proc. of the 21st Annual Meeting of the Adhesion Society, Savannah, GA, February, 1998, p. 175.
Gale A. Holmes, Richard C. Peterson, Donald L. Hunston, Walter G. McDonough, The Applicability of Nonlinear Melt Rheology Concepts to the Nonlinear Solid Rheology of DGEBA/m-PDA Epoxy Resin in the Glassy Region, Proc. of NOBCChE 1998, Dallas, TX, April 13, 1998, in press.
Presentations
D. L. Hunston, G. A. Holmes, R. C. Peterson, Viscoelastic Properties of a Resin Commonly used in the Single Fiber Fragmentation Test, American Society of Composites Meeting, Baltimore, MD, Sept. 21-23, 1998.
D. L. Hunston, G. A. Holmes, W. G. McDonough, Standardization of Test Methods for Fiber-Matrix Interface Strength in Polymer Composites, poster at NIST Assessment Panel Meeting, NIST, Gaithersburg, MD, March 19, 1998.
G. A. Holmes, Debond Region Formation during Fiber Fracture, Imperial College London, UK, June 10, 1998.
G.A. Holmes, Effect of Strain Rate on Fiber Fracture in Single Fiber Fragmentation Test Specimens, UMIST, Manchester, UK, June 11, 1998.
G. A. Holmes, Modeling of Multi-Step Relaxation Behavior of DGEBA/m-PDA using Melt Rhelogy Concepts, University of Leeds, Leeds, UK, June 12, 1998.
G. A. Holmes, The Deformation Mechanics of the Single Fiber Fragmentation Test, CIMTEC 1998 Conference, Florence, Italy, June 14, 1998.
G. A. Holmes, R. C. Peterson, and D. L. Hunston, Modeling of Multi-Step Nonlinear Stress Relaxation in DGEBA/m-PDA Epoxy Resins, The 2nd Int. Conference on Mechanics of Time Dependent Materials, Pasadena, CA, March, 1998.
G. A. Holmes, R. C. Peterson, D. L. Hunston, and W. G. McDonough, The Influence of the Matrix Modulus on the Interfacial Shear Strength Parameter, The 21st Annual Meeting of the Adhesion Society, Savannah, GA, February, 1998.
Gale A. Holmes, Richard C. Peterson, Donald L. Hunston, Walter G. McDonough, The Applicability of Nonlinear Melt Rheology Concepts to the Nonlinear Solid Rheology of DGEBA/m-PDA Epoxy Resin in the Glassy Region, NOBCChE 1998, Dallas, TX, April 13, 1998.
Hybrid Reinforced Polymer Composites
D. L. Hunston, K. M. Flynn, R. S. Peterson, J. P. Dunkers, G. A. Holmes, W. G. McDonough, and S. Zaghi
Objective
The objective is to formulate a scientific basis for testing the mechanical properties of hybrid composite materials by developing an understanding of the way that the microstructures of these materials influence the behavior and complicate the interpretation of test results.
Technical Description
There is a great deal of interest in composites which are reinforced with two or more different types of fibers. By combining the advantageous features of carbon, glass, and Kevlar fibers, the fabricator can achieve significantly better trade-offs between various properties and between cost and performance. Industries currently exploring the use of hybrids include off-shore oil, infrastructure, and alternate fuel vehicles. The challenge, however, is that the design methodology for hybrids is inadequate. For example, studies at the University of Houston have shown that some properties of hybrid tubes cannot be understood with current theories and some properties measured with tubes do not agree with equivalent properties obtained from plate experiments. The explanation seems to be that the failure modes and the interpretation of test results depend on details of the sample microstructure in ways that we do not understand. Moreover, characterizing this microstructure is a challenge since the available tools have not been applied to these types of systems. Because the variety of possible hybrid materials is almost unlimited, the work here will focus initially on the filament wound tubes of interest to the off-shore oil industry and the questions raised by the Houston study. The work will first develop/adapt tools to characterize the microstructure of these materials. A systematic effort will then measure the microstructure and mechanical properties of tubes and plates prepared with a model system. The project is a major collaboration with Drs. Wang and Williams at the University of Houston who will be conducting much of the large scale mechanical testing. Since the project must eventually extend to other types of hybrids as well, some selected experiments will be initiated immediately using the layered hybrids of interest for infrastructure applications and the fabric hybrids being considered for use in alternate fuel vehicles.
External Collaborations
Prof. S. S. Wang and Dr. J. Williams, University of Houston - testing of hybrid tubes.
Prof. Frank Ko, Advanced Product Development, Bristol, PA - providing hybrid fabrics for tests.
Guy Rossi, Solectria, Wilmington, MA - application of hybrids in electric vehicles.
Planned Outcomes
Accomplishments
This project is new, and the first step was the design and purchase of test specimens made with a model material system. Based on the Houston results, both tubes and plates were included. The resin and curing chemistry were chosen to be acceptable to industry and commercial fabricators. Carbon and glass fibers were selected for the mix since this combination gives the best opportunity to balance high specific strength and stiffness with low cost -- the goal for many commercial applications. The mix pattern was selected to combine both fiber types within each ply. This avoids the weak planes that can occur when each ply is a single fiber type. Based on these choices, detailed specifications for the samples and the filament winding process were developed, and specimens were ordered.
The second step in the project was the development of microstructure characterization techniques. A number of methods were examined and proved useful. One of the most interesting was the combination of optical and electron microscopy (see Figure 1). The optical experiments easily distinguished the carbon fibers while the glass fibers stood out clearly in electron micrographs. These characterization methods provide the tools needed for the remainder of the program.
In a complimentary study, experiments were initiated on a hybrid system of interest for alternate fuel vehicles. Several reinforcement geometries are being considered including a hybrid fabric. Because such systems are totally new, there is a need to obtain basic mechanical property data.
Unfortunately, most existing test methods were designed to accommodate simple reinforcement geometries like unidirectional layers or random mat materials. Consequently, these tests must be examined and where necessary adapted for use with hybrids. During the past year, a matrix of desired properties was established and possible measurement methods were identified.
Microstructure Studies: Performance Relationships
D. L. Hunston, W. G. McDonough, G. A. Holmes, C. R. Schultheisz, Z. Miyagi1, D. Raghavan2, J. He2, and S. Zaghi
1National Research Laboratory for Metrology, Tsukuba, Japan
2Howard University, Department of Chemistry, Washington, DC
Objectives
The objectives are to develop and apply measurement tools to establish relationships between the microstructural features generated during processing and the performance properties of polymer composite systems. The work focuses on the role of the polymer-polymer, polymer-metal, and the polymer-fiber interfaces in performance of structural adhesives and composites.
Technical Description
The work in this project involves four areas: bulk resins, coatings, composites, and adhesives. The common thread is structure-property relationships. With bulk resins, toughened thermosets were studied. In these two phase material systems, the fracture resistance or toughness depends on the morphology of second phase (the toughener). Since the morphology in most such systems develops by phase separation during cure, the microstructure is difficult or impossible to control and vary systematically. Te work here uses materials made from suspensions of preformed toughener particles to minimize morphological change during cure. Consequently, features like particle size, size distribution, and concentration can be systematically varied by changing the starting suspension and the dilution prior to cure. The work on coatings is developing an ultrasonic technique to characterize non-destructively polymer layers of 1 mm or less bonded to a flat surface. Once developed, the technique will be used to monitor adhesives and coatings during polymer cure or attack of the cured polymer by water. The work on composites is examining the fatigue behavior of pultruded samples to develop test methods for durability and to generate data that explore degradation mechanisms. The final area of work is developing a new technique for characterizing the viscoelastic properties of adhesives. The specimen involves two strips of metal bonded together with the adhesive. This sandwich is subjected to three point bending perpendicular to the adhesive plane. The specimen is simple to make, disposable, and simulates the constrains and processing environment of an adhesive joint. In addition, there is a published elastic analysis which would permit the calculation of the shear modulus for the adhesive from the bending stiffness of the sandwich.
External Collaborations
Dwight Hoffman, Dow Chemical Co. - information on morphology of rubber-toughened epoxy.
David Dwight, Owens Corning - the relationship between fragmentation test and full composites.
Glen Barefoot and Daniel Witcher, Strongwell - materials for durability of pultruded composites.
Prof. Catherine Brinson, Northwestern University - durability for pultruded composites.
Prof. D. Raghavan, Howard University - structure-property relationships in toughened epoxies.
Prof. K. Liao, Nanyang Technological University, Singapore - fatigue of pultruded composites.
Planned Outcomes
The work on toughening will generate structure-property relationships that provide generic guidelines to help industry develop more fracture resistant materials.
The ultrasonic studies will produce a fully automated instrument for characterizing and monitoring changes in the shear mechanical properties of thin films.
The adhesive and composite studies will provide guidelines and methodologies for testing the properties and durability of structural adhesives and composites.
Accomplishments
In the areas of toughening, ultrasonics, and environmental fatigue, the focus during the past year was publication of the results. In addition, an important question was addressed for the model toughened thermoset. The material used here is designed to produce samples where the size and size distribution of toughener particles remain constant while the concentration of toughener is varied over a wide range. Although all the previous data support with this assertion, the evidence was indirect. During the past year, transmission electron micrographs were taken of cryogenically sliced thin sections stained to reveal the acrylic rubber. Efforts to quantitatively analyze the micrographs is now underway, but simple visual inspection of the photographs confirm that the rubber remains phase separated as particle over the complete range of concentrations studied. Moreover, there is no indication of a major change in the size or size distribution of the particles. This is important because high concentration of toughener in other systems often produces phase inversion.
The behavior of adhesives in the new sandwich specimen is being explored by testing a model thermoset with a Tg of 30 °C over a range of temperatures from 5 °C to 55 °C, using both stress relaxation and dynamic mechanical experiments. The results clearly show the contribution of the adhesive can easily be measured so the technique has good sensitivity to the properties of the adhesive. Surprisingly, when the curves for different temperatures were shifted, they seemed to superimpose to form a master curve even though they represent stiffness of a structure and not a simple modulus. To test the elastic analysis which is available for the sandwich specimen, results at 5 °C and 55 °C were used since the adhesive behavior is approximately elastic at these temperatures. Unfortunately, the calculated values for the adhesive's shear modulus are completely unrealistic. Examination of the theory suggests that some of the assumptions made in the analysis may not be valid unless the adhesive is a very thin. Work on a new, viscoelastic analysis is now underway.
Outputs
Publications
J. He, D. Raghavan, D. Hoffman, and D. Hunston, The Influence of Elastomer Concentration on Toughness in Dispersions Containing Preformed Acrylic Elastomeric Particles in an Epoxy Matrix, Polymer, in press.
D. L. Hunston, J. He, R. Raghavan, and D. Hoffman, Limits on toughening in Structural Adhesives, Proc. Adhesion Society Meeting, Adhesion Society, Blacksburg, 1998, pp. 200-202.
D. L. Hunston, Automated Monitoring of Thin Film Properties via Ultrasonics, Proc. Adhesion Society Meeting, Adhesion Society, Blacksburg, 1998, pp. 440-442.
K. Liao, C. R. Schultheisz, D. L. Hunston, and L. C. Brinson, The Effect of Water on the Fatigue Behavior for a Pultruded Glass-Reinforced Composite, Proc. 56th Annual Technical Conference of Society of Plastics Engineers, ANTEC 1998, SPE, Brookfield, 1998, pp. 2245-2250.
D. L. Hunston, Performance Property Relationships, in Micro-Mechanics Measurement Technologies for Fiber-Polymer Interfaces, ed. by McDonough, Parnas, Holmes, and Hunston, NISTIR-6102, NIST, Washington, DC, 1997, pp. 169-191.
D. Raghavan, D. L. Hunston, J. He, and D. Hoffman, The Influence of Morphology and Concentration on Toughness in Dispersions Containing Preformed Acrylic Elastomer Particles in an Epoxy Matrix, Polymeric Materials Science and Engineering Preprints, American Chemical Society, in press.
Presentations
D. L. Hunston, "Thermosets", Adhesion Short Course, Adhesion Society, Savannah, Feb. 21, 1998.
D. L. Hunston, J. He, R. Raghavan, D. Hoffman, "The Effect of Elastomer Concentration on Toughening in Rubber-Modified Thermosets", 49th ACS Southeast Regional meeting , Roanoke, October 20, 1997.
D. L. Hunston, J. He, R. Raghavan, and D. Hoffman, "Limits on Toughening in Structural Adhesives," Adhesion Society Meeting, Savannah, Feb. 21-24, 1998.
K. Liao, C. R. Schultheisz, D. L. Hunston, and L. C. Brinson, "The Effect of Water on the Fatigue Behavior for a Pultruded Glass-Reinforced Composite," 56th Annual Technical Conference of Society of Plastics Engineers, ANTEC 1998, Atlanta, April 16-May 1, 1998.
D. Raghavan, D. L. Hunston, J. He, and D. Hoffman, "Toughening of Epoxies - Dispersion of Preformed Acrylic Elastomeric Particles in an Epoxy Matrix," poster at poster at Gordon Conference on Composites, Ventura, Jan. 4-10, 1998.
D. L. Hunston, "Automated Monitoring of Thin Film Properties via Ultrasonics," poster at Adhesion Society Meeting, Savannah, Feb. 21-24, 1998.
D. Raghavan, D. L. Hunston, J. He, and D. Hoffman, "The Influence of Morphology and Concentration on Toughness in Dispersions Containing Preformed Acrylic Elastomer Particles in an Epoxy Matrix," poster at Symposium on Toughening of Plastics, Polymeric Materials Science and Engineering Division, American Chemical Society, Boston, August 23-27, 1998.
Z. Miyagi, M. Koike, and D. L. Hunston, "Development of Testing Method for in-situ Shear Modulus of Composites Material using Sandwich Beam," Symposium by National Research Laboratory of Metrology, Tsukuba, March 10, 1998.
Microstructure Studies: Measurement of Interface Strength after High Speed Processing of Polyurethanes
W. G. McDonough, Y. H. Kim1, K. M. Flynn, and R. S. Parnas
1 Korea Maritime University, Pusan, S. Korea
Objectives
The effects of processing on interfacial shear properties as measured by the single fiber fragmentation test (SFFT) will be determined. The materials to be evaluated are resins and reinforcements used in automotive applications. New processing procedures will also be developed to enable the preparation of fast curing (in the order of minutes) specimens that can be used in the microstructure program and which are equivalent to materials made in industry by structural reaction injection 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. This method cannot work with rapidly curing resins of interest to the auto industry. Consequently, an injection molding procedure has been developed that will closely mimic the processing speed, temperature, and pressure observed in the SRIM process used with the resins of interest. The dog bone samples thus prepared will be tested by SFFT to determine if the interface strength is degraded under rapid processing conditions. Whereas resin transfer molded epoxy resins may take hours to cure, the rapid reaction of polyurethanes can result in gel formation in less than a minute and vinyl ester resins can gel in a few minutes.
External Collaboration
Thomas Dearlove, Automotive Composites Consortium - coordinating results with the automotive industry.
Planned Outcome
Provide the automotive industry methods for estimating the tradeoffs between processing speed and composite properties.
Accomplishments
A multi-cavity mold was designed and built to produce 8 single fiber dogbone specimens from a single injection. A specially designed injection system was made to simulate the SRIM process. Flow visualization experiments were carried out with non-reacting fluids to simulate the hydrodynamic loads experienced by the single fiber in each mold cavity. The fibers survived high speed injections, verifying that SFFT samples can be prepared in an SRIM-like process.
Vinyl ester/E-glass single fiber fragmentation specimens were made at different processing times. Initial results indicate changing the gel time changes the measured interfacial properties. A single formulation of vinyl-ester was either catalyzed with 2% or 10% catalyst to produce either a 45 minute or a 10 minute gelation time. The materials were processed to the same total level of cure. The fiber fragmentation patterns observed upon testing, and subsequent analysis, indicated that the interfacial shear strength was approximately 30% less in the more rapidly cured material.
Outputs
Publications
W. McDonough, R. Parnas, G. Holmes, D. Hunston, Workshop on Micromechanics Measurement Technologies for Fiber-Polymer Interfaces, NISTIR-6102, NIST, Washington, DC, 1997.
W. McDonough, G. Holmes, R. Peterson, Interface Adhesion of E-glass Fibers in Model Polyisocyanurate Networks, Proc. American Society of Composites Baltimore, MD, Sept. 21-23, 1998, in press.
Presentations
W. McDonough, G. Holmes, R. Peterson, Interface Adhesion of E-glass Fibers in Model Polyisocyanurate Networks, American Society of Composites Baltimore, MD, Sept. 21, 1998.
Microstructure Studies: Non-destructive Characterization by Optical Coherence Tomography
J. P. Dunkers, F. R. Phelan, C. G. Zimba, K. M. Flynn, R. Prasankumar1 and J. G. Fujimoto1
1MIT, Boston, MA
Objective
The goal of this work is to image composite microstructure using a new technique, optical coherence tomography (OCT). Knowledge of tow position and orientation are applied to permeability prediction. Characterization of defects such as voids, wrinkles, cracking, delaminations, and crazing is important for optimizing processing parameters and evaluation of composite health.
Technical Description
OCT will be applied to a variety of composites systems to determine its potential for characterizing microstructure and damage. OCT is a non-invasive, non-contact optical imaging technique that allows the visualization of microstructure within scattering media. OCT uses light in a manner analogous to the way ultrasound imaging uses sound and, while typically affording shallower penetration depth, provides significantly higher resolution (5-30 µm). To perform OCT imaging, broad-spectrum laser light (20 - 200 nm bandwidth) is transmitted using a single mode fiber and coupled into a 50/50 fiber optic splitter that illuminates the sample and a linearly translating, constant velocity reference mirror. The fiber optic splitter, fixed sample and constant velocity reference mirror can be thought of as a Michelson interferometer. Light back-reflected from each interferometer arm is recombined at the fiber optic splitter. Interference fringes are registered at the detector only when the optical path length of the reference arm matches that of the sample arm to within the coherence length of the light source. The axial resolution with which this ranging can be performed is therefore determined by the coherence length or inverse spectral width of the source. Therefore, the axial resolution can be as low as 5 µm. Low coherence sources such as mode-locked solid-state pumped lasers or superluminescent diodes are used. Transverse resolution in OCT is determined by the focal spot size of the probing beam, which is usually 10-30 µm. Higher numerical aperture optics provide superior transverse resolution, but at the expense of a diminished depth-of-field.
Planned Outcome
To provide the composites community a versatile and low cost non-destructive technique for the study of composite microstructure.
Accomplishments
Reinforcement permeability was calculated using OCT images and the lattice-Boltzman fluid flow program. Images from OCT that show tow location, size, and shape of a unidirectional epoxy/unidirectional E-glass composite (see Figure 2) were used in these permeability calculations. Before the permeability calculation could be run, the low contrast grayscale image had to be converted into a binary image through an automated and robust image processing program that used various techniques to enhance the contrast between the tows and resin, and then automatically identify the tows. There was good agreement between the axial permeability values calculated using the OCT images and the experimental permeability. In addition, the calculated permeabilities from two different regions in the sample were almost identical. By comparing automatically and manually processed images, it was determined that the tow roughness had a large influence on the calculated permeability.
The versatility of OCT was demonstrated by imaging a variety of defects. A three dimensional tomographic reconstruction of the data enabled visualization of the defect within the composite along any plane. A re-slicing along the fiber length revealed voids that are oriented in the direction of the fiber tows. Analysis of a 0-90° woven reinforcement showed the relative position and orientation of layers to one another and a wrinkle in one of the interior layers.
It was also demonstrated that OCT could image cracks and delaminations in composite. The epoxy/unidirectional E-glass composite was subjected to impact damage. OCT was then performed along a selected surface crack. OCT revealed the crack penetrating through a tow in the top layer. When the crack reached 575 mm, a rectangular delamination zone that was 2.9 mm by 660 mm was revealed. The 575 mm corresponded to the epoxy rich area between the first and second layer of composite.
External Collaborations
R. Prasankumar and J. Fujimoto, MIT - Hardware and expertise in OCT Imaging.
Outputs
Publications
Joy P. Dunkers, Richard S. Parnas, Carl G. Zimba, Richard C. Peterson, Kathleen M. Flynn, James C. Fujimoto, and Brett E. Bouma, Optical Coherence Tomography of Glass Reinforced Polymer Composites, Composites A, in press.
J. P. Dunkers, C. G. Zimba, K. M. Flynn, R. S. Parnas, D. L. Hunston, J. G. Fujimoto, B. Bouma, and R. Prasakumar, Optical Coherence Tomography of Polymer Composites, Proc. of the American Society of Composites Conference, Baltimore, MD, Sept. 21-23, 1998, in press.
Joy P. Dunkers, Carl G. Zimba, Frederick R. Phelan, Richard S. Parnas, and James Fujimoto, Real Structure Determination for Permeability Prediction, Proc. of 8th US-Japan Conference on Composite Materials, Baltimore, MD, Sept. 24-25, 1998, in press.
Presentations
J. P. Dunkers, R. S. Parnas, R. C. Peterson, C. G. Zimba, K. Flynn, B. Bouma and J. G. Fujimoto, Optical Coherence Tomography of Composites, Fall Meeting of the Materials Research Society, Boston, MA, December 4, 1997.
J. P. Dunkers, C. G. Zimba, K. M. Flynn, R. S. Parnas, D. L. Hunston, J. G. Fujimoto, B. Bouma, and R. Prasakumar, Optical Coherence Tomography of Polymer Composites, American Society of Composites Conference, Baltimore, MD, Sept. 21, 1998.
Joy P. Dunkers, Carl G. Zimba, Frederick R. Phelan, Richard S. Parnas, and James Fujimoto, Real Structure Determination for Permeability Prediction, 8th US-Japan Conference on Composite Materials, Baltimore, MD, Sept. 24, 1998.
