Today’s commercially available PAN-based carbon fibers cover a wide range of properties. The tensile strength ranges from approximately 2.5 to 6.5 GPa (362 to 943 kpsi), while tensile modulus ranges from approximately 220 to nearly 500 GPa (32 to 73 Mpsi). With this broad a product portfolio available, most of the research in the industry is pursuing gains other than improvements in properties. Suppliers of the small-tow carbon fiber (3k to 12k filaments) are working to reduce their costs while maintaining fiber properties. Manufacturers of large-tow fibers (more than 12k and often as much as 320k filaments) are trying to match the performance of the smaller-tow fibers and reduce product variability. Manufacturers of both small- and large-tow carbon fibers are working to improve process control so as to increase process efficiency and improve product uniformity. New monitoring techniques are becoming available that could improve the quality and consistency of both small- and large-tow products. These include high-resolution optical probes for monitoring gels and flaws, portable on-line Raman instruments for monitoring structure, and laser instruments for monitoring fiber size. These and other advanced monitoring techniques are being evaluated on pilot-scale fiber lines in academia and industry.
For PAN-based carbon fiber properties, two important research areas have arisen. First, does a wider distribution of properties such as strength, for example, affect the realizable properties of the final part and, if it does, can this effect be predicted reliably and accommodated in the final composite design? This question is important to both the high-performance aerospace user and the much larger-volume applications for infrastructure, transportation, and energy. The effect of a given proportion of low-strength individual filaments in a tow on the strength of a composite made by impregnating the tow with resin is largely unknown. Further, the effect of the initial filament strength distribution on the strength reliability and lifetime of the impregnated tow and, in turn, on the reliability and lifetime of the final part made from many tows, is also not known. Current lifetime models do not consider the statistics of the failure process. Although a model has been developed to accomplish such a prediction,2 no databases exist to verify it. Furthermore, such verification would require an integrated study among fiber producers, prepreg and towpreg companies, and composite part manufacturers. In addition, in situ monitoring of composite part integrity and health could be used to validate such models. The successful outcome of such an effort would result in a lower factor of safety needed for a given reliability, as well as the ability to certify very large parts with minimal or no full-scale testing. Both of these advantages would result in very large cost savings.
A second research thrust would be to improve the uniformity of the carbon fiber. For example, reducing the variability in fiber weight per unit length will reduce the fiber areal weight variability in prepreg. Because processing technology is generally tightly held by the companies, some mechanism such as a consortium of universities, government, and companies to produce results useful to all might be a mechanism for achieving this goal.
Long-term opportunities to improve the compressive properties of fibers are more difficult to define. Many designs are limited by the compressive strength of the composite. Research has shown that both intra- and intercrystalline disorder must be increased if the compressive properties in both PAN- and pitch-based carbon fibers are to be improved.
Although pitch-based carbon fibers have been around for some time, they have managed to penetrate only some small niche markets. The primary reason is that many applications for carbon fibers to date have been strength-driven. This has given PAN-based carbon fibers, with their less flaw-sensitive nongraphitic structures, a natural advantage. However, because of their graphitic structure, the thermal conductivities of pitch-based carbon fibers are as much as three times that of copper and orders of magnitude higher than those of PAN-based carbon fibers, and applications are now emerging in which heat transfer is critical. For example, management of excess heat has become a limiting factor in the design of many military aircraft, satellite structures, and electronic packages. During high-speed