Composites have replaced metals in a wide range of applications, realizing considerable weight savings as well as associated cost and performance improvements. One factor that can limit the use of composites is their thermal or electrical conductivity properties. In particular, composites based on unidirectional tapes or 2-D laminates are often not suitable in areas where heat is generated due to their limited thermal or electrical conductivity perpendicular to the direction of the fibers.
The use of high-conductivity fibers can provide high in-plane thermal conductivity of laminated or unidirectional composites. However, in the through-thickness direction, heat must traverse polymer-rich inter-laminar regions with relatively low thermal conductivity, as well as pass through several fiber-matrix interfaces, each with an attendant loss in thermal conductivity.
In the direction of the fibers, the parallel circuit analogy is well accepted in composite literature and leads to a Rule-of-Mixtures (ROM) relation for the in-plane thermal conductivity. In the direction perpendicular to the fibers, the through-thickness direction, many models have been proposed and examined over a number of years, each predicting and many experimentally proving thermal conductivities much lower than ROM would suggest.
Thus, current composite technologies do not provide structural composites with high through-thickness thermal conductivity. As an example, a uniaxial composite of K1100 pitch carbon fibers (950-1100 W/m·K) in an epoxy matrix (0.2 W/m·K) can provide 595 W/m·K in the direction of the fibers while only measuring 1.0 W/m·K perpendicular to them. In 2-D laminated composites, parts based on a plain weave prepreg of YS-90 pitch carbon fibers (500 W/m·K) in epoxy measured 145 W/m·K along the warp and fill directions, but only 1.0 W/m·K through the thickness.
Loading the matrix resin with particulate can provide some increase in the through-thickness thermal conductivity, but the overall influence is limited and it comes at the cost of both lowering the mechanical properties of the resin and increasing the fabrication complexity of the composite. In high-end applications, high through-thickness thermal conductivity is achieved by making a carbon-carbon (C-C) composite and partially graphitizing the carbon matrix, providing through-thickness thermal conductivities in the range of 20-50 W/m·K. Again, this processing is expensive and the resultant C-C composites are not suitable as structural members.
3-D woven preforms offer a low-cost method of manufacturing composites with high through-thickness thermal conductivity without an attendant loss of strength or stiffness. In research by 3TEX, a survey of available pitch carbon yarns and a set of tests to determine the ability to 3-D weave them led to the manufacture and through-thickness thermal conductivity testing of composites based on orthogonal 3-D woven preforms with pitch carbon tows and with copper yarns in the through-thickness (z) direction.
Results of this research indicate that composites based on 3-D fiber architectures with high thermal conductivity fiber in the z direction have the potential to greatly increase the through-thickness thermal conductivity of polymer-based composites. The most attractive high thermal conductivity fiber type in terms of performance and weight are pitch carbon fibers, though they create difficulty in preforming. 3TEX conducted tests on a variety of pitch carbon fibers and found that YS-80 (320 W/m·K), CN-80 (320 W/m·K), and YS-95 (600 W/m·K) could be woven in the Z direction.
Tests on composites based on 3-D fiber architectures with YS-80 and with copper yarns in the Z direction measured through-thickness thermal conductivities between 4 and 7.2 W/m·K, a substantial increase over that reported in other literature. The results, however, did not follow a parallel circuit thermal flow model that would predict a ROM relationship between the conductive fibers in the through-thickness direction and the remaining “matrix” of non-aligned fibers and polymer. Rather, the measured thermal conductivities were around 25 to 35% lower than ROM would suggest.
The lower-than-expected conductivities were explained in terms of heat flow paths. How much heat flows through the “matrix” or through the conductive fibers depends on the relative thermal conductivities of the contact material, the “matrix,” and the conductive through-thickness fibers, as well as the thickness of the part and the distance between the through-thickness fibers. This ratio in turn determines the effective thermal conductivity of the composite material.