Two trends in composites.

1. RTM Goes High-Tech

2. Thermoplastics Inch Ahead

Perhaps one of the brightest spots in a generally slow composites market has been the continually evolving technology of resin transfer molding, or RTM. Extremely attractive to major manufacturers from aircraft to automotive, this technology has been the focus lately of technical development aimed at "maturing" the process.

Evaporating aerospace and defense funds have led to recent speculation that another young technology--thermoplastic advanced composites--might never emerge from the cradle. But, based on recent development, such fears seem to be exaggerated. Technology here is still progressing, though somewhat slowly, and there is growing evidence these materials have a strong future in commercial aviation.

Recent technical conferences and one planned for later this spring afford instructive glimpses at major trends in both these composite sectors.

Toward Higher-Tech RTM

There have been plenty of signs lately that RTM is getting ready for the big time. Not only have major aircraft manufacturers expressed considerable interest in the process, but most large automotive SMC molders have been gearing up for the RTM applications everybody sees coming.

They may not have to wait long. The RTM application receiving the most attention these days is the Viper sports car Chrysler Motors Corp. plans to debut later this year. Around 170 lb of RTM are used for a structural underbody pan and several of the exterior body panels. (see PT, Sept. '91, p. 40). Though it's a low-volume specialty car (5000 units per year), it could spell the beginning of larger scale manufacturing programs than RTM has ever seen before.

Faster molding cycles, improved process controls and developing a better understanding of the process to improve quality and speed up the production of new products are keys to broadening RTM's markets. In an effort to satisfy the need for rapid molding cycles, there have been a number of recent improvements in automation, with many of them concentrating on pre-forming. The quest for better controls has been addressed by new, sophisticated, hydraulic-powered pumping systems with full SPC capabilities (see PT, May '91, p. 19 and fresh late-breaking news next month). And computerized process simulation is shedding light on how RTM works, providing users with more control over the process.

RTM PROCESS MODELING

There has been considerable activity in recent years in RTM/SRIM process modeling at raw-material suppliers, universities and some large processors (see PT, Jan. '90, p. 17; May '91, p. 29). Although this work is either proprietary (in the case of molders) or in R&D stages, one encouraging sign is that three commercial vendors of mold-filling and heat-transfer analysis software for injection molding are actively developing versions for RTM. One is Plastics & Computer Inc., which co-authored a paper at last fall's Detroit meeting of ASM International and the Engineering Society of Detroit on joint RTM modeling work being done with Fiat's research center in Italy. Flow simulation was performed on an automobile rear floor panel to be molded of glass mat, urethane foam core, and ICI Acrylics' Modar thermosetting acrylic filled with 50% calcium carbonate. A 3-D finite-element flow analysis yielded predictions of filling patterns; filling time and pressure; temperature; and shear stress throughout the part. The data are displayed in a color-coded map and actual mold trials produced fill times within 10% of the prediction.

Another commercial software firm is Advanced CAE Technology, which is working with Ohio State University on 3-D modeling of mold filling and curing in RTM/SRIM. OSU has presented progress reports on its R&D project at two previous SPI Composites Institute conferences. Their latest results were presented at the fall ASM/ESD meeting and last month's SPI conference in Cincinnati (see PT, Jan. '92, p. 39). The program predicts such data as pressures, temperature profiles and degree of cure through the thickness of the part. The two papers dealt with modeling the effects of heat transfer between the resin and reinforcement and also between resin and mold wall. The latter paper examined the differences between epoxy and metal tooling of differing thermal conductivities.

Techanalysis is yet another software firm that has adapted its work in injection and compression molding analysis to RTM applications. The company is working with one RTM processor who is using its software. What may be a distinctive feature of the Tech-analysis version is that it monitors inlet and runner conditions.

A materials supplier that has done independent work on injection mold-filling simulation is now also looking into RTM. GE Corporate Research and Development in Schenectady, N.Y., reported at the ASM/ESD meeting on its FEMAP-RTM software, an adaptation of the company's FEMAP finite-element injection molding software. This program tracks fluid flow fronts, pressure, temperature, and viscosity. Additional functions were added for RTM applications, including flow through porous media with antistrophic permeability, and time-dependent viscosity for fast-curing materials. GE is using the program to help it design aerodynamic fan blades for aircraft engines.

In its paper, GE considered multiple gating points and how to use the program to locate vents so converging flow fronts would not result in air entrapment and voids. GE also considered the perhaps unsuspected gravity effects on flow that most other programs reportedly ignore.

In addition, GE modeled RTM with its developmental cyclic thermoplastic oligomers. These novel materials have low molecular weight and low viscosity, but chemically react in the mold to build a high-molecular-weight thermoplastic polymer (see PT, Oct. '90, p. 131). FEMAP-RTM models the effects on pressure and flow of the viscosity build-up that occurs while the mold is filling. (This is not normally a concern with relatively slow-curing thermosets.)

According to researchers at the University of Nottingham, England, a comprehensive modeling technique is essential if RTM is ever going to be used for high-volume production of composite automotive components. These researchers presented their modeling technique at the ASM/ESD conference. Their finite-difference model reportedly can provide computational aids for designing molds and complex shapes as well as assist in determining manufacturing economics. They used the example of an automotive undershield to show how their technique can model successive flow fronts advancing from the injection gate, as well as pressure and temperature distributions.

Also at the ASM/ESD meeting, reports on modeling fluid flow through fiber reinforcements were given by researchers from the University of Michigan in Ann Arbor, Michigan State University in Lansing, and the National Institute for Standards and Technology (NIST) in Gaithersburg, Md. NIST is working in cooperation with the Automotive Composites Consortium, Troy, Mich.

The Center for Applied Polymer Research of the Ecole Polytechnique de Montreal in Quebec is also developing a finite-element simulation program for RTM. The program will be discussed in detail in Orlando, Fla., this May at a conference sponsored by the American Composites Manufacturing Learning Center (see PT, Feb. '92, p. 12). This program also computes and displays flow fronts inside the mold and pressure at each step of the molding process. Their technique, the researchers say, allows easier simulation of complex shapes, as well as of parts containing openings and inserts or multiple injection points.

Computer modeling of a special variant of RTM for "advanced fiber architectures" has been developed by Virginia Polytechnic Institute and State University, Blacksburg. It models the hot-pressing of a lay-up of B-staged resin film and dry fiber preform. The program simulates the infiltration of hot-melt resin into the preform and cure of the resin-saturated preform. For a specified temperature cure cycle, the researchers say, the model was able to accurately predict the compaction pressure, initial resin mass and time required to infiltrate a fabric preform to the desired fiber volume fraction. The Virginia Polytechnic team says future work with the modeling technique will focus on selecting optimal cure cycles by using electromagnetic sensing to monitor the position of the resin infiltration front and the state of resin curing during fabrication of the composite.

AUTOMATING RTM

First discussed at the ASM/ESD show in Detroit was an innovative, automated RTM system developed by Owens-Corning Fiberglass Corp. at its Technical Support Laboratory in Belgium (see PT, Nov. '91, p. 11). The system is designed to produce large RTM or SRIM structural components in high volume. It consists of a robotized directed-fiber preforming system and a resin injection system; plus automatic preform transfer between the two systems. The concept is adaptable to a fully automated production-line set-up. Owens-Corning is working with General Motors and the Automotive Composites Consortium to commercialize this technology.

The jointed-arm preforming robot has the flexibility to deposit random and oriented, chopped and continuous fibers, powdered resin binder and even particulate fillers. Thus, the robot can build up a precisely engineered preform employing different kinds of reinforcements.

Several feed rollers controlled by stepper motors feed rovings from several creels into the robot arm. Fibers are chopped, if desired, at a point on the "shoulder" of the robot arm, and transported inside the arm to a spray deposition nozzle. A special device is added if the chopped fibers are to be oriented. Surfacing veil is applied by chopping low-tex roving and separating the fibers with a breaker bar. No mat handling is required.

The preform is sprayed into a mold half, which shuttles into a press containing a matching tool. The tools are metal screens fed with hot and cold air through two parallel ducting systems.

Owens-Corning has designed, but not built, a mechanism for automatically transferring the preform from the female spray-up mold half to a male injection mold half. As shown in the diagram, the female mold half is flipped over 180|degrees~ to deposit the preform like a pancake onto the male mold. Stripping the preform from the female mold is facilitated by a combination of a movable center screen section and fixed outer screen area.

The injection unit uses a hydraulically driven, one-stroke pump (from Applicator AB of Sweden) with three separate delivery cylinders for resin, initiator and accelerator. It must be sized to fill the mold in one continuous stroke. This has the important value of eliminating the pulsating or hesitating flow caused by other pumps, which results in air entrapment and voids.

Also, this pump eliminates variations in initiator content. The main resin cylinder is linked to the other two smaller cylinders by cam followers to ensure a consistent ratio. Furthermore, the cams can be engineered to add more accelerator to the latter part of the shot, helping it gel and cure at the same rate as the first material into the mold and leading to shorter overall cure times.

The pump feeds an inlet valve on the mold that is closed at the end of injection to hold the mold's contents under pressure until cured. The valve also provides for flushing the mixers and pipes.

Molds have a special sealing system consisting of a silicone seal fitted into the upper half and a steel edge in the lower half. The reinforcement is compressed between the seal and the steel edge and allow air to bleed out all around the mold. This means no vent holes to clean and no displacement of the reinforcement because it is fixed along all edges.

To achieve faster flow of resin into the mold, Owens-Corning has devised a method that introduces the resin via a channel along one entire edge of the mold instead of from a single point. The resin then flows along the full width of the part in a parallel flow front and at a constant speed. This method can deliver resin to the mold 10 times faster than through a central inlet, the company says.

NEW APPROACHES

High-performance RTM for manufacturing secondary structural parts for aircraft was discussed by French supplier Brochier SA, sub. of Ciba-Geigy Corp., at January's Society of Manufacturing Engineers (SME) Composites in Manufacturing '92 conference in Anaheim, Calif. Brochier sees RTM as 20-40% less expensive than traditional manual prepreg-tape layup and autoclave curing, making it feasible to use composites in more applications.

Brochier's Injectex RTM process injects resin under low pressure (7.25 to 21.75 psi) into an airtight mold under high vacuum. Such low injection pressures, and fiber contents higher than in normal RTM, are reportedly made possible by a special family of Injectex reinforcing fabrics and by special resins (developed with Ciba-Geigy) that have viscosities below 100 cps at temperatures as high as 212 F. Injectex fabrics range from standard weaves to highly drapeable, heat-formable fabrics and even woven socks. Resins include four epoxies and a phenolic; a flame-retardant epoxy and a bismaleimide are in development.

Brochier stresses the importance of degassing the resin prior to injection to ensure low porosity in the parts. Injection must be done under very low pressure into an airtight mold under high vacuum.

A novel approach called "flexible" RTM (FRTM) was presented in two papers at last October's SAMPE conference in Kiamesha Lake, N.Y., from MIT's Laboratory for Manufacturing Productivity in Cambridge, Mass., and from the nearby Charles Stark Draper Lab, Inc. FRTM clamps dry reinforcing fabrics between two elastomeric diaphragms. Resin is then impregnated into the reinforcement with the aid of vacuum. Finally, the wet fabric sandwiched between the diaphragms is placed in a press with matched tooling or perhaps a single tool and a pressure box to apply compressed air to deform the diaphragms and prepreg. The advantage of this process is that it eliminates the need either to make a preform or for tedious hand lay-up of fabric in a tool. Impregnation of the fabric in a flat sheet may also be somewhat easier than in a formed shape. Also, mold halves reportedly can be relatively lightweight compared with conventional RTM tooling, since they do not need to withstand the pressure of resin injection.

Thermoplastics Inch Ahead

Despite the slow commercial progress of thermoplastic advanced composites, which has caused some materials suppliers to scale back their development efforts (see PT, Dec. '91, p. 55), the subject is far from dead. In fact, substantial technical effort in this field continues, judging from recent conference papers.

NEW HIGH-HEAT RESINS

Composites that can withstand high temperatures are an ongoing subject of aerospace research. Among new thermoplastic composite materials unveiled recently are two high-temperature hexafluorinated polybenzoxazoles (PBO) developed by Northrop Corp.'s Aircraft Div., Hawthorne, Calif., and the Polymer Branch of the U.S. Air Force at Wright-Patterson AFB in Dayton, Ohio. As reported at the SAMPE conference, two different PBO resins were found to have physical and mechanical properties similar to PEEK, polyetherimide (GE's Ultem), and polyarylene sulfide (Phillips 66 Co.), together with glass-transition temperatures of 550-580 F and thermal stability to 950-970 F, based on TGA tests in nitrogen atmosphere. The resins have very high melt viscosity and very low solubility. Initial efforts at prepregging into carbon-fiber tape with the BASF powder method were only partly successful. Further work on impregnating technique and fiber/resin coupling is needed.

INDUCTION BONDING FLIES

An interesting case study demonstrating the economic advantages of thermoplastic composites over high-performance thermosets was provided at the SAMPE meeting by Grumman Aircraft Systems, Bethpage, N.Y. The project also demonstrated the advantages of electromagnetic induction bonding over autoclaving for joining TP composite sections. Grumman fabricated an all-thermoplastic horizontal stabilizer out of prepregs of carbon fiber and PAS-2 polyarylene sulfide resin from Phillips 66 Co. (known commercially as Ryton S). Grumman compared the labor, materials, tooling and facilities costs for fabricating 100 parts from thermoplastic and from BMI prepreg (the latter has already been flight tested). The thermoplastic was compared in two variations--one with conventional autoclave bonding of the outer skins to the I-beam stiffeners, and one with induction bonding of the two using 3M's AF-191 amorphous thermoplastic film adhesive, and a vacuum bag to apply pressure. Induction heating acted directly on the carbon-fiber reinforcements, eliminating the need for any metal susceptor material.

Tooling for both TP approaches required about 40 man-hr vs. 60 man-hr for thermoset. Despite the fact that the thermoplastic part required a total of 14 separate manufacturing steps, versus one-step autoclave co-curing of the thermoset part, fabricating labor costs were only 56 hr for induction-bonded TP, 78 hr for autoclave-bonded TP and 89 hr for thermoset. Materials costs were $958 for induction-bonded TP, $3260 for autoclave-bonded TP (presumably higher because of the vacuum-bagging materials) and $2378 for thermoset BMI. Overall manufacturing costs were virtually identical for the TP and BMI autoclaved parts, and both were 73% higher than the induction-bonded TP part. In addition, Grumman was quite encouraged that the induction-bonded TP component showed slightly greater bond strength and 10% greater stiffness than its autoclave-bonded counterpart.

SPECIAL FOR THERMOFORMING

Since Phillips 66 Co., Bartlesville, Okla., is no longer actively pursuing TP composites, the company presented what may have been its last paper on the subject at the October SAMPE meeting. Phillips discussed thermo-forming of TP prepregs. Specialized equipment included a platen press with two moving platens and a floating third platen that actually applies the forming pressure via inflatable air bags. The floating platen allows for some platen misalignment and reportedly lowers the cost of the press because the lateral force on the tool acts upon the "bayonets" or tie rods rather than the moving platens.

The press incorporates a clamping frame that allows the prepreg sheet to slip into the mold during forming while under tension so as to resist wrinkling. Also, the clamp frame shuttles the prepreg quickly from the preheat oven into the press.

To reduce the cost of matched tooling, one mold half is metal and the other has a cast RTV silicone rubber surface. The latter, a layer only 0.06-0.25 in. thick, compresses during forming to apply uniform hydrostatic pressure on the laminate. Also, the pliable silicone permits varying the laminate thickness to some degree without having to build a new tool--helpful in prototyping.

TP PULTRUSION ADVANCES

Interest in thermoplastic pultrusion continues to grow, according to Jeff Martin, president of equipment maker Pultrusion Technology Inc. His firm is embarking on a nylon pultrusion development project; and Martin says a major extruder of PVC profiles has expressed interest in pultruding PVC with glass fibers. He also expects to introduce fairly soon a machinery system specialized for pultruding thermoplastic prepregs.

One approach to solving the problem of how to impregnate rovings and mats with a high-viscosity thermoplastic melt is to start with low-viscosity precursors that chemically react in-situ to yield a high-molecular-weight polymer composite. This "RIM-pultrusion" or "reaction injection pultrusion (RIP)" method has been pursued by some university researchers as a means of pultruding nylon 6/glass-fiber composites (see PT, Dec. '87, p. 29; Jan. '90, p. 19). This approach is said to be easier than handling "boardy" TP prepregs and much less expensive than softer powder-impregnated rovings.

An effort to commercialize this process was reported last month at the SPI meeting in Cincinnati. American Composite Technology (ACT), working with the Plastics Engineering Dept. of the University of Lowell, Mass., has developed proprietary technology for RIM-pultruding nylon 6 from caprolactam. ACT developed its own formulation, based on a proprietary catalyst and accelerator, in order to get faster reaction rates than were available with any commercial system.

The two-component system, which has a viscosity in the range of 1-10 cps, was preheated to drive off moisture and pumped through a static mixer and injected directly into the 42-in.-long die. E-glass rovings were fed into the die to produce 3/8-in. rod. At a die temperature of 320 F, pulling speeds of 15-20 in./min were achieved with monomer conversion rates of 96-97%.

ACT v.p. of engineering Stephen C. Nolet says the material system will probably be marketed commercially before the end of this year by an unnamed company that provided funding for the project. Total material cost for RIP with this system is expected to be less than $2/lb.