New materials & processes at SPI Composites meeting.

Novel processes and new resins will be featured among the technical papers at the annual SPI Composites Institute Conference and Expo in Cincinnati, Feb. 8-11. Also evident will be the growing emphasis on computerized process simulation for RTM/SRIM, SMC, pultrusion and even corrosion-resistant laminating. Composite recycling will be a less prominent focus this year, limited to a panel discussion on composite recycling programs around the world.

A HYBRID OF RTM AND SMC

Among the many machinery and process innovations will be a discussion of a unique low-pressure molding technique that derives from both RTM and SMC. The VPM (Vacuum Press Molding) process was developed by Lotus Engineering of Norfold, England, to produce body parts for its Elan sports car (production of which was just recently canceled). The process is said to offer several advantages over Lotus's current vacuum-assisted RTM (VARI) method. For one thing, VPM reportedly can mold net-size components in only 5 min, and parts need no post-trimming. No liquid resin or catalyst is mixed or dispensed, and the tooling can be either plastic or nickel-faced. VPM uses a press only for mold handling, not pressure generation, because only vacuum is required to distribute material throughout the mold. Since no glass preform is required and neither pumping equipment nor a high-tonnage press is needed, capital investment is far below that for SMC or standard RTM.

The material system is a pasty molding compound, containing chopped fiber, resin, catalyst, and release agent. The resin is a blend of 75% standard liquid polyester and 25% Crystic Impreg, a special crystalline polyester made by Scott Bader Co. Ltd. of England and available from Ashland Chemical Canada Ltd., Mississauga, Ont. That material is a soft solid at room temperature, but melts at around 194 F to a low-viscosity liquid. When molten Crystic Impreg is blended with liquid polyester and then cooled, the mix solidifies. In a hot mold, it becomes highly fluid again. (This approach is used by Total Composites, Inc. of South Bend, Ind., to make an SMC that needs no chemical thickener and molds at low pressure. See PT, April '90, p. 36).

With VPM, a pre-weighed charge of compound is placed in the gel-coated lower mold half, and the mold is closed. The material melts virtually instantly in the hot mold and flows out to completely fill the mold as vacuum is applied to withdraw air from the cavity. (Vacuum also provides the only force needed to hold the tool closed.) The resin gels in 70 sec, giving a 3-5 min total cycle. Each tool can produce 12 parts/hr. Parts need only a light manual deflashing.

Parts can be made without gel coating, but some porosity remains (Scott Bader is working to eliminate this). Leftover trimmings are stable at room temperature and can be compression or injection molded like BMC or SMC. Lotus also notes that physical properties of moldings can be enhanced by sandwiching layers of reinforcing fabric between charges of resin compound.

Taking the latter approach even further, Lotus is experimenting with an alternative VPM technique employing polyester resin with heat-activated catalyst in a film form that is laid in the mold on top of a conventional glass preform. The catalyzed resin film, developed for Lotus by Ashland Chemical Inc., Columbus, Ohio, melts in the hot mold and impregnates the preform, curing in 5 min. Again, no liquid resin or catalyst handling is involved.

VACUUM FORMING SMC

Another novel vacuum-assisted "dry" processing technique will be the topic of a paper from Honda Engineering Co., Ltd., Honda R&D Co. Ltd., and Nippon Shokubai Co., Ltd. of Japan. The process consists essentially of vacuum forming a prepreg sheet consisting of a thickened polyester SMC-type compound between gel-coat layers, all of which is sandwiched between two stretchable films. The gel coat is blade coated onto the film before deposition of the SMC. The 40%-glass SMC is created by impregnating resin, catalyst and thickener into chopped-glass mat.

The process involves preheating the prepreg for 20 min at 140 F and then forming with vacuum into a temperature-controlled female mold. Gel time was about 5 min and cycle time 20 min. Key problems faced by the developers were tailoring the viscosity of SMC and gel-coat materials to provide uniform thickness throughout the part and selection of an appropriate carrier-film material. The developers consider the process to be suitable for medium- and large-scale manufacturing of medium- to large-size FRP moldings such as lighting domes, kitchen countertops, bathtubs and artificial marble.

RTM DEVELOPMENTS

Also scheduled for the conference is an update on the automated RTM system developed by Owens-Corning Fiberglas Corp., Toledo, Ohio, at its Technical Support Laboratory in Battice, Belgium (see PT, Nov. '91, p. 11 and March '92, p. 42). Owens-Corning calls it the "P-4" system, standing for Programmable Powdered Preform Process. It's designed to produce as many as 150,000 large RTM or SRIM structural components per year. It consists of a robotized directed-fiber preforming system (using a powder binder) and a resin injection system, with automatic preform transfer between the two systems. It's adaptable to a fully automated production-line setup. This month, the company is expected to announce the first commercial use of the system.

Researchers from the Center for Applied Polymer Research at the Ecole Polytechnique in Montreal will discuss the state of development of computerized process simulation for RTM (see PT, March '92, p. 41). They will illustrate the use of RTMFLOT software to model the filling of a complex 3-D part. The software consists of an integrated series of program modules for calculating the resin permeability of the preform, producing a finite-element mesh of the mold, computing and displaying the successive positions of the resin flow front, and calculating the heat transfer from the mold to determine resin viscosity during the process.

In a separate paper, Ecole Polytechnique researchers will discuss the development of an empirical model to predict permeability of commercial continuous-strand mats. Mat permeability prediction will also be discussed by a group from the Departments of Mechanical and Chemical Engineering at Ohio State University, Columbus, which has been developing RTM computer models for several years.

PET FIBER TOUGHENS RTM, SMC

Hoechst Celanese Corp., Technical Fibers Group, Charlotte, N.C., will present two papers on toughness enhancement of composites with Trevira-brand continuous-strand mat of polyester (PET) fiber. This material has been used to enhance the toughness and reduce the density of marine laminates through partial replacement of glass. In RTM, PET mat also improved resin flow, but it did decrease tensile and flexural properties, compared with all-glass laminates. In SMC experiments with Premix Inc. of North Kingsville, Ohio, Hoechst Celanese found similar enhancements in toughness, balanced by reductions in tensile, flexural and compressive properties, though adequate mechanical properties could still be obtained. Some development work was necessary to overcome the lofting tendency of the PET mat. Also, since the PET mat does not flow, the SMC charge must be cut to the full size of the part; and only relatively simple shapes can be molded with PET mat reinforcement.

SMC/BMC MACHINERY AND RESINS

In equipment developments for SMC, John T. Hepburn, Ltd., Mississauga, Ont., will discuss the Paradyne Multimold system on which a patent has been applied for. For simultaneous molding of multiple parts, this system uses four pancake-type pushback cylinders on a bolster plate of the press bed. All cylinders are connected to a common pressure source--typically an accumulator. Analogous to the use of such pushback cylinders for parallelism control, they are said to ensure uniform pressure in each quadrant of the press, regardless of thickness variation between the different molds.

Researchers from the Dept. of Chemical Engineering at Ohio State University will discuss progress in computer simulation of SMC material flow in the mold, heat transfer to and from the mold, and material curing. Now that the process can successfully be modeled, the authors say, the next phase of development is to create software that will automatically optimize the process to meet goals and constraints set by the user.

Another step beyond SMC flow simulation is predicting the residual stresses molded into the part so as to calculate shrinkage and warpage. Researchers from the Polymer Processing Research Group of the Dept. of Mechanical Engineering at the University of Wisconsin-Madison will present a finite-element/finite-difference program for SMC shrink/warp prediction taking into account the effects of fiber content, part thickness, asymmetric curing and flow-induced fiber orientation.

Ashland Chemical Inc. will report on a three-component resin system it recently developed for structural SMC. It consists of a new thickenable vinyl ester (Arotech 2000, which will be the subject of a separate paper), a thickenable isophthalic resin (Arotech 1900) and a patent-pending, low-shrink additive concentrate (Arotech 2100). The proportions can be tailored for various levels of cost-performance in high-temperature, zero-shrink applications. Its max. glass-transition temperature is 358 F, almost 50|degrees~ higher than previous Ashland vinyl esters. Reportedly, the system can be tailored to balance economics with properties such as resistance to high temperatures, corrosion and fatigue. Regardless of the blend ratios, the HDT, |T.sub.g~ and flexural modulus do not change, Ashland says. Tensile strength, flexural strength, elongation and shrinkage control increase slightly as the vinyl ester content of the blend increases, the company reports, while tensile modulus increases with a rise in isophthalic content. The novel low-shrink additive, a polyether polyester that's soluble in vinyl ester, is designed to give zero shrink and good physical properties at a range of blend ratios of the other components.

Ashland Chemical will also discuss making SMC more flexible for use in auto panels such as liftgates, doors and side panels of vans. Ashland says that by combining a tough matrix resin with a novel surface-enhancing flexible resin and additives, plus lowering the amounts of glass and filler, SMC can be made with excellent surface quality and reduced flexural modulus. The new "Arophlex" resin system processes just like conventional SMC, Ashland says, but has greater toughness and flexibility.

A new type of low-profile unsaturated polyester resin from Japan, designed specifically for making artificial marble by a BMC process, will be discussed by Hitachi Chemical Co., Ltd., Ibaraki, Japan. The polyester and low-profile additive in the system can produce compression molded composites that Hitachi says are transparent and have excellent boiling water resistance (300 hr with no change in appearance). The new polyester is a bisphenol-A isophthalic type. The low-profile additive is an unspecified thermoplastic with special functional groups that react with the polyester; and the rest of the system consists of glass powder filler, chopped-strand glass fiber, and magnesium oxide thickener.

Molders of automotive Class A SMC may be interested in a paper from Diffracto Ltd., Windsor, Ont., correlating surface waviness measurements from its D-Sight TPS-2 instrument and the LORIA device from Ashland Chemical. The authors show "a significant correlation between the two instruments over a variety of test plaques, even though the experimental conditions were not rigorous."

PULTRUSION NEWS

Owens-Corning will deliver a paper on pultrusion processing of phenolic resins. According to the company, phenolics are attractive because of their low VOC emissions and good fire and smoke performance. However, Owens-Corning warns, there are a few processing differences between phenolics and polyesters--namely the creation of moisture volatiles during curing and the very low viscosity of phenolics--that may require modification of the pultrusion process. Owens-Corning found that using a die profile with a cool entrance zone, high-temperature middle zone, and an exit zone that cools the part surface to near its glass-transition temperature best fits the rheology and kinetics of phenolics.

Thermoplastics Pultrusion Techniques of Yorktown, Va., and the NASA/Langley Research Center in Hampton, Va., will discuss the potential for low-cost thermoplastic pultrusion of industrial products rather than the high-tech/high-cost aerospace/defense applications that have been explored in the past. According to the authors, PET polyester offers good potential for pultrusion because its combination of shrinkage, viscosity, water absorption, mechanical properties and cost are closest to those of unsaturated polyester, considered to be the ideal pultrusion material. Polypropylene is also a good choice, the researchers say, because of its light weight, low water absorption, and low cost. However, they warn, a pultruded E-glass/PP will have lower tensile, flexural and shear strengths than a thermoset.

For predicting resin pressure and backflow in the tapered inlet of a pultrusion die, researchers at the University of Mississippi, University, Miss., have developed a mathematical model that examines the effect of resin viscosity, preform size, pulling speed and profile of the inlet taper in order to obtain the best die inlet shape and process conditions to achieve maximum possible pressure rise in the die inlet. According to the Mississippi researchers, the model shows that the pressure can be increased significantly (to minimize voids) by changing the preform size without altering the existing die inlet.

CORROSION-RESISTANT RESINS

New corrosion-resistant resin systems will be the topic of several papers. The Dept. of Polymer Science at the University of Southern Mississippi, Hattiesburg, will discuss an all-hydrocarbon thermoset (MRG-XR333) it developed as a matrix material for carbon- and glass-fiber composites. University researchers expect the resin can be made commercially for $2-3/lb. Composites made from the new thermoset reportedly have extremely high thermal resistance and hydrolytic stability, and can be cured by various methods. A carbon-fiber composite with about 30% by weight resin exhibits a flexural modulus of 9.6 million psi, compared with a 6.1 million psi modulus for an epoxy/carbon composite. When exposed to 572 F heat, the composite with the all-hydrocarbon matrix maintains 60% of its modulus, while the composite with the epoxy matrix retains only 10% of its modulus. After bringing the composites back to ambient temperature, the modulus of composites made with the new thermoset increases to 11.75 million psi, while the epoxy composite's modulus rises only to 3.3 million psi. While the flexural moduli for glass composites were about the same for the hydrocarbon and epoxy matrices at room temperature, the hydrocarbon resin retained nearly all of its original stiffness at temperatures from 392 F to 662 F, while the epoxy composite saw its modulus drop by about 50% at 392 F.

Interplastic Corp., Vadnais Heights, Minn., will report on a new terephthalic acid-based polyester resin that reportedly offers better acid resistance than standard PG/isophthalic corrosion-resistant resins, as well as superior resilience and toughness. High HDT, tensile strength and elongation are also claimed.

Reichhold Chemicals Inc., Research Triangle Park, N.C., will report on polymer concrete made with its low-shrink polyester, Polylite Profile 32-490. Reichhold compared the benefits of polyester concrete with an epoxy-based concrete, finding that raw-material cost for the polyester was 25.6|cents~/lb vs. 45.8|cents~/lb for the epoxy; the workability, castability and surface finish of the two concretes were equal;linear shrinkage for the polyester concrete was 0.0007 in./in. after 4-hr cure at 150 F, while the epoxy concrete shrunk 0.0016 in./in.; dimensional stability of the polyester concrete was better; and the polyester concrete had better linear coefficient of thermal expansion than the epoxy. However, compressive strength and modulus of the epoxy concrete was superior, especially at elevated temperatures, Reichhold concedes.

And Ashland Chemical has developed a computer program to model the effect of laminate thickness, ambient temperatures and air cooling on cure response of corrosion-resistant laminates. The program computes heat transfer and resin cure kinetics to predict exotherm temperature and degree of cure. It's useful, Ashland says, in designing processes to minimize residual stresses, cracking and part distortion. It can answer questions about temperature gradients, whether staging is required, and effect of laminate thickness on cure response.