Advances in sizings and surface treatments for carbon fibers
After developing a new low-temperature-cure carbon/epoxy prepreg for a customer recently, Lewcott Corp. (Millbury, Mass., U.S.A.) produced small runs of the prepreg for testing, using 3K and 12K fabric. The customer reported that the cured 12K prepreg performed as expected, but the 3K product would not cure at the desired temperature, says Nick DiMeo, Lewcott director of sales and marketing. A subsequent investigation turned up the culprit: the only difference between the fibers in the two fabrics was the sizing, which in the case of the 3K fabric apparently interfered with the matrix resin's crosslinking. Not surprisingly, Lewcott's weaver could not tell DiMeo exactly how the sizings differed. Surface treatment and sizing, the final two steps in the carbon fiber manufacturing process, are the subject of notable trade secrecy. "These are the subtleties that make one fiber different than another fiber," explains Chris Levan, technical service manager for Cytec Carbon Fibers LLC (Piedmont, S.C., U.S.A.). In fact, sizings and surface treatments play vital roles in composite performance: Surface treatment produces additional bonding sites on the fiber surface, while sizing enhances fiber processability with a protective coating on the fiber surface and can provide a coupling agent for the fiber/resin bond. The relationship between surface treatment and sizing is complex, and the effectiveness of each depends on its compatibility with both the fiber and the resin.
Until recently, there was little demand for information about or changes in sizing and surface treatments. The vast majority of carbon fiber is still incorporated into standard, qualified epoxy resin systems via well-known fabricating processes, for which sizing and surface treatment "recipes" are so well established and reliable that choice of fiber is typically based on economics. Lewcott, for example, routinely shops carbon fiber from a dozen different weavers. "They never ask you what sizing you want on it," says DiMeo. "Our purchasing guy is looking for the best price."
However, in DiMeo's customer's case, outlined above - and despite the fact that the fiber manufacturers had optimized each of the sizings for the prevalent epoxy resins - the accelerators and other modifiers used to reduce the cure temperature effectively created a "new" epoxy resin. This moved the prepreg into uncharted territory. "As we were pushing the temperature envelope lower and lower," says DiMeo, "we found a sizing/resin incompatibility that we didn't know existed."
Historically, such fiber/resin incompatibilities have been rare. But in a significant and growing number of new applications the risk is increasing, as "off-the-shelf" carbon reinforcements are introduced to applications for which the matrix will be highly modified epoxy resins or non-epoxies such as vinyl ester, polyimide or PEEK. Carbon in chopped form also is being paired with thermoplastic molding compounds. And carbon is beginning to be used in atypical molding processes, such as resin infusion. As molders step into these new territories and confront the inevitable challenges, those who surface treat and size carbon fiber have begun to draw aside the veil.
Adhering to treatments tried-and-true
While fiber manufacturers are not forthcoming about their approach to surface treatment, it is widely believed that all of the manufacturers use a time-honored electrolytic process to alter the fiber surface. Without mentioning which surface treatment Cytec uses, Levan agrees that for commercial fibers, surface treatment technology is relatively static. "The dials on [the process] can be changed to enhance a particular characteristic," he says, "but fundamental changes are typically not required. Most carbon fiber manufacturers have optimized their surface treatments, and our robust processes ensure repeatability." Consequently, fiber manufacturers have had little incentive to experiment with other treatment methods, such as plasma treatment and gaseous or wet chemical processes. "There haven't been many applications calling for better performance from the surface than we get now," Levan maintains. "However, if industry momentum builds such that new chemistries and applications call for different surface treatments, Cytec Carbon Fibers is well-positioned to meet that challenge."
By contrast, R&D specialists like Advanced Fiber Technologies Inc. (Wilmington, Del., U.S.A.) are actively seeking the outer limits of interfacial performance from traditional electrolytic treatments and alternative systems as well, especially in high-performance niches, which demand the absolute optimum interfacial properties. In space applications, for example, the need for penultimate strength-to weight ratios can justify the investment. Company president Warren Schimpf explains that electrolytic treatment forms a number of different functional groups for bonding, including hydroxyls, carboxyls and carbonyls. Surface treatments of commercial carbon fiber lines, optimized for epoxies, generally seek to minimize the carbonyl groups, which do not bond as well with epoxies as the other groups. But other resin systems may tolerate or even prefer carbonyls. Advanced Fiber Technologies' goal is to find the optimum ratio of these groups for particular applications. "We may vary parameters such as current, voltage, electrolyte type or concentration to preferentially develop the functional groups," says Schimpf.
His team also is exploring alternate technologies such as electropolymerization, which (as the name suggests) polymerizes different species onto the fiber surface. This process could potentially serve the dual role of surface treatment and sizing; or it could be performed on a conventionally surface-treated fiber, directly replacing sizing. "In particular areas, there's no real reason why it couldn't be cost-competitive with other sizings," Schimpf speculates. This process, however, is in a very early developmental phase. And, he points out, "it has to be better than existing sizings to be worth it."
Similar realities confront R&D companies that are experimenting with other alternatives such as plasma treatment. To date, none have changed the performance of the finished composite significantly enough to warrant large scale development. "We've proven that we can generate the kind of functional groups we're trying to generate," says Schimpf, "but we're not always successful making the composite better than we can with the electrolytic process." The small improvements that have come to fruition would benefit only a small set of applications. For most others, he concludes, "[Electrolysis] is an efficient enough process to make the composites perform as well as they need to."
Rethinking the sizing job description
The story is somewhat different, however, with sizings. In the carbon fiber market, conventional wisdom views surface treatments as the bonding agents and sizings as the handling agents that, at best, have no effect on interfacial bonding, and at worst, must be removed to achieve good bonding. One cannot overestimate sizing's contribution to ease of handling. Weavers have worked with fiber manufacturers for decades to develop sizings that, when fibers are passed through a liquid bath, form a film or coating on the fiber surface, which effectively "lubricates" the fiber to optimize production speed and facilitate the weaving process in ways that produce more cost-efficient, higher quality carbon fiber fabrics. "In very rare instances, we recommend making fabrics unsized for highly specialized applications," says Mary Shafer, general manager of Fabric Development Inc. (Quakertown, Pa., U.S.A.) by way of illustration, "but we get one-tenth the production rate." Outside the relatively narrow field of carbon fiber/epoxy applications, however, the assumption that sizing is only a handling aid may not hold up.
According to Dr. Andy Brink, vice president of sizing developer/manufacturer Hydrosize Technologies Inc. (Raleigh, N.C., U.S.A.), the notion that "'unsized carbon fiber is best' may be true for epoxy composites, but that isn't necessarily true with all matrix resins. The typical carbon fiber surface has several functional groups and can even be considered reactive to epoxy and other condensation reactions. However, it has no reactive groups capable of reacting with vinyl ester during cure (a free radical process)." For carbon to bond well with vinyl esters and other nonepoxies, the sizing must form a compatible bond to both the carbon surface and the resin matrix.
The solution, however, is nowhere near as simple as adapting the sizings used on glass fibers with vinyl esters. Brink notes that the film formers that work so well with glass are designed to "dissolve quickly into the resin, thus freeing up the silane coupling agents for reaction with the resin and providing good interfacial shear strength." But in the case of carbon, he continues, a quickly dissolving sizing "frees up the carbon fiber surface that is nonreactive to the vinyl ester, and a good interface is not formed. It needs to have a 'controlled solubility' - so that it partly dissolves in the resin, but does not completely leave the fiber interface," Brink concludes.
For polyimides, PEEK and other high-temperature resins, carbon poses a similar challenge. The functional groups provided by the traditional epoxy-compatible sizings, "do not chemically react with polyimides," says Hydrosize president Dr. Heather Brink. "Consequently weak interfacial shear strengths result." Worse, for carbon parts molded with polyimides or PEEK, high processing temperatures during manufacture and continuous use in high-temperature end-use environments can degrade the epoxy sizing and, as a result, weaken the fiber/matrix interface, producing voids and delaminations.
Optimizing a sizing for a particular resin system involves a degree of scientific knowledge, but much of the developmental work also is based on practical trials. Suggesting that the theoretical knowledge and the empirical data are sometimes hard to separate, Andy Brink nevertheless admits that fiber manufacturers often do rely on the empirical. "There's some truth to that," agrees Cytec's Levan. "However, fiber scientists have progressed past the empirical stage and they know how to formulate and experiment in order to optimize the characteristics." Among other parameters, Cytec can change the amount (wt%), type, application conditions or drying conditions of the sizing to meet its customers' needs for new products. A similar, knowledge-based process occurs at Hydrosize, "If a customer wants a film former with a certain modulus, tensile strength, Tg, and so on," Heather Brink reports, "we can usually obtain those properties in one or two attempts."
Matching sizing to the process
The optimal sizing also differs for different fabricating methods, though little information is available because developmental work is in its early stages. Levan reports that Cytec has the expertise to vary its sizing for fibers destined for infusion versus prepreg processes, for example, but beyond that, he could reveal very little. Heather Brink explains that Hydrosize alters the ratios of the monomers to change the chemical composition of the polymer, yielding different properties. But again, methods for altering these ratios are proprietary.
In the case of chopped versus continuous fibers, sizings will differ to provide different spreadability characteristics. "In chopped fiber," says Andy Brink, "you need to have extremely good bundle integrity or the fiber is difficult to handle and feed into the compounding machine." Thus, continuous-fiber applications might use a sizing that enables greater filament spreading for easier wetout, but this spreading must be prevented in chopped-fiber applications until the resin is introduced (i.e. during compounding).
Is the improbable possible?
While technical solutions are available, those pioneering carbon fiber use in innovative fiber/resin/process combinations face some obstacles to easy reformulation of sizing to suit their unique applications. While suppliers of glass fiber may offer 30 to 40 fiber products, where the primary difference between these is sizing adapted to various resins, molding processes and end uses, the carbon/epoxy market, by nature, is resistant to change. In the aerospace arena, qualification processes dictate what fiber, sizing, and matrix resin system are used - both in military and in many high-volume commercial applications. "T-300 today is essentially the same as it was 20 years ago," Levan points out. "If we've qualified the fibers, we do not change unless otherwise stipulated by the customer." Any reformulation of sizing in this market would have to warrant the cost and time involved in requalification.
Further, many nonepoxies are "compatible enough" with the epoxy-based sizing that it has not been necessary to push the envelope. In some cases, Fabric Development's Shafer notes, "We haven't realized the full theoretical strength of the composite." However, people using polyesters, vinyl esters, or even infusion processes, she points out, are concerned about factors other than achieving optimal performance out of the sizing, such as cost.
Production volume also is an issue. Currently, most carbon-reinforced composites that are incompatible with epoxy-based sizings correspond to small product runs for niche markets. While the sizing material itself is a relatively small expense, entailing only 1 to 2 wt% of the fiber, the cost of switching over a commercial manufacturer's sizing line to a specialty sizing is very high for these small markets. Sizings can be applied after fiber manufacture, but Shafer estimates a 50 percent increase in material costs to produce a fabric with a specialty sizing, compared to using off-the-shelf fibers; and Lewcott's DiMeo notes similar cost increases for his company's prepreg. Such expense is absorbed only in unusual circumstances.
While weavers like Fabric Development do run their own sizing lines for developmental work or small production runs, Shafer point out, "We don't want to do a 6,000 lb sizing run." In any case, applying the sizing during fiber manufacture is much preferred - to avoid fiber damage during spooling and the costs of a secondary process, but also because the fiber manufacturers are best equipped to produce sizing runs of commercial quantity. Yet a "critical mass" of sorts must be reached before a specialty sizing run becomes economically feasible.
In addition, most sizing providers are large resin or coating manufacturers for whom the relatively small market of specialty sizings is not a priority. For carbon fiber, these larger companies "do not put a lot of resources into sizing development," says Heather Brink. "It often falls below their radar." Thus the rapid rise of a company like Hydrosize Technologies, which has grown specifically because it is responsive and focuses entirely on this market, she notes, adding, "Our prices are very cost-competitive because we maintain a very low overhead."
Team effort a must
Market forces are not the only factor; both surmountable and inherent technological limitations also come into play. Carbon behaves differently from glass, both physically and chemically. As a result, those working on new fiber/matrix combinations must consider not only the fiber's surface treatments and sizings, which afford a certain degree of versatility, but also the matrix resin, to which additional modifications might be made to improve interfacial characteristics. A team approach involving the resin and fiber suppliers and all other stakeholders is warranted.
This kind of cooperation is challenging because of the varying needs of each team member, DiMeo suggests. For example, his company is developing some phenolic applications with carbon reinforcements uniquely sized to yield desired properties. "We're making sure it works for the weaver and that the fiber supplier can apply it."
Carbon sized up for SMC/BMC
DiMeo believes that as carbon prices drop and more low-cost tooling options become available, carbon composites will continue to move into high-volume applications that will cost-justify such collaborative efforts. A case in point is compounder Quantum Composites Inc. (Bay City, Mich., U.S.A.), which manufactures glass- and carbon-reinforced sheet molding compounds, using vinyl ester, polyimide and other resins as the matrices. To make its carbon-reinforced products, the company chops tow into a matrix paste that has been laid down on a carrier film. "We want [the fiber] to chop and fall into nice discrete bundles," says account manager Mark Imbrogno. "But once it gets into the paste, we want it to be fully infiltrated with matrix into the bundle." The fibers are sized with these handling characteristics and the interfacial bonding challenges of the specific resin system in mind. As a result, Quantum produces carbon-fiber SMC products with optimum interfacial properties. These yield the desired level of part performance from a smaller volume fraction of fibers, reducing material cost. Further, a sizing-enabled combination of chop bundle control and spreadability during wetout keeps Quantum production lines moving at more cost-efficient speeds.
Quantum's markets promise plenty of potential for volume production. "The automotive and offshore oil industries have focused on carbon fiber for a long time," Imbrogno points out. Quantum is, in fact, actively selling engineered structural composite molding compounds for high-temperature applications (such as bushings and bearings) using polyimides, as well as in automotive applications with hybrid vinyl esters, and carbon fiber/epoxy in military and ordinance applications. While price continues to be an issue in these markets, interest in carbon fiber is growing. "Quantum feels it has a feasible pathway to more widespread use of carbon fiber." As they develop these markets, Imbrogno emphasizes, "The sizing is a key focus. There's no secret there."
Unlocking the future
As carbon fiber enters new arenas - in automotive and industrial markets, this has only begun - today's "best kept secrets" are the keys to doorways that lead to greater composite performance and more efficient manufacturing methods. Now more than ever, those who make those keys will be key players in carbon fiber's future. Ultimately, success of such ventures will depend on the effectiveness with which carbon fiber manufacturers and sizing/surface treatment providers, like their fiberglass industry counterparts, understand and meet the processing and bonding requirements of weavers, fabricators and product end users.
Focused on optimizing traditional hand layup, nacelle and thrust reverser manufacturers cast an eye on future use of automation and closed molding.
Fiber-reinforced plastic (FRP) replacing coated steel in more reinforced-concrete applications.
Oven-cured, vacuum-bagged prepregs show promise in production primary structures.