You are here

Filled polymers with high thermal conductivity*

News International-French

9 Mar 2011

Adding a conductive filler in a polymer is an effective way to improve the thermal conductivity of plastics. Extensive thermal and composition measurements on injection-moulded polybutylene terephthalate plates show that the best filler consists of short aluminium fibres, and that thermal conductivity can be increased by a factor of 2, 6 and 10 using 20, 42 and 44 % vol. aluminium, respectively.


(Published on July-August 2007 - JEC Magazine #34)


The challenge of thermally conductive polymers

Thermally conductive filled polymers are heterogeneous materials containing at least two solid phases: a continuous insulating phase and a disperse conductive phase. The latter is introduced as a filler during the compound processing stage; therefore, the filler particles are distributed irregularly in the end product.


Applications relate to heat transfer in the car industry (radiators, hoods, battery cases for electric vehicles...), electronics or electrical engineering (components, cases...). The use of injection-moulded filled polymers would make it possible to produce geometrically complex parts while saving weight compared to the metal version.


In many applications, the suitable thermal conductivity in the transversal direction of the plate is 2 to 4 W.m-1.K-1 instead of the lower conductivity of unfilled technical polymers (from 0.15 to 0.3 W.m-1.K-1). The in-plane conductivity is less relevant for the end-user, although high values can help prevent hot points.



The practical difficulty when producing thermally conductive filled polymers is to retain the technical characteristics of the bare polymer. As the presence of filler leads to poor mechanical properties (inducing brittleness, for example), great care must be taken to achieve not only good conductivity, but also stability, innocuousness, mechanical quality and, last but not least, low price. Intrinsically conductive polymers are an alternative to filled polymers, but their high cost and lack of stability are major drawbacks.


The ease of production of thermally conductive filled polymers depends on the type of polymeric material, decreasing in the following order: inks, adhesives/mastics, rubbers, thermosets, and thermoplastics. The first two classes are already technically mature, and rubbers are often made conductive by blending with carbon black; but thermosets are seldom made conductive by filling. Finally, the commercial availability of conductive thermoplastics, though repeatedly announced, has never become effective.


Materials, processing and measuring methods

Tests were performed using polybutylene terephthalate (PBT) and various types of fillers such as carbon black, ceramics, boron nitride, and aluminium. Two sets of aluminium fibres (ALF1 and ALF2) with the same average length (1.1 mm) and with different average diameters (respectively 90 μm and 160 μm) were considered. The polymer and filler were mixed using a twin screw extruder. Flat plates with dimensions 100 x 100 x 2 or 4 mm (Figure 1) were moulded using a semi-industrial injection machine.


Small 15 x 15 x 2 or 4 mm samples were cut from the flat injection-moulded plates to analyse the thermal and electrical conductivities and the filler contents. Thermal conductivity was measured using a hot guarded plate apparatus designed for small samples analysis. Electrical conductivity was measured after deposition of a 100 nm silver electrode on both large faces of the samples.


Main results

The best results were obtained with aluminium fibres.


Effect of filler content

As shown in Figure 2, electrical conductivity increases sharply for a filler content around 30% by volume, which corresponds to the percolation threshold of the compound. Higher filler contents do not affect electrical conductivity. Thermal conductivity shows a different behaviour (Figure 2): it always increases as the filler content increases, the sharp increase being obtained at higher volumetric filler contents (44%). At this metal concentration, the composite exhibits a thermal conductivity 10 times higher than the pure polymer.



Effect of particle size

Particle size (ALF1 or ALF2) does not have any noticeable effect, as seen in Figures 2 and 3. The differences occurred when studying the uniformity of the two studied properties. The electrical and thermal conductivities and filler content were measured on 15 to 20 samples distributed in each injectionmoulded part (see Figure 1). Strong non-uniformity of both properties was found for thick fibres (ALF1); this was mainly due to the non-uniform distribution of the filler particles (Figure 4). Reducing the particle size (ALF2 fibres) significantly improved the uniformity of the filler distribution (Figure 5) and the dispersion of the measured properties around their mean values in the whole part was reduced by a factor of 2 [3].



Effect of particle orientation Optical observations showed that the particles were preferentially oriented along the main flow direction (x-axis). This explained the anisotropy of the injection-moulded parts, the in-plane thermal conductivity being 2.5 (y axis) to 4.5 (x axis) higher than the transversal one.


Further directions for thermal conductivity improvement

One could not wholly benefit from the fibre shape due to their in-plane orientation during the injection process. The use of particles with 3D structure (foam-like particles for example) could result in higher thermal conductivities. Further research should also focus on the ways to reduce thermal contact resistance between the particle and matrix, since their value and effect on the effective thermal conductivity is quite high as the particles become closer, as shown in [4] and [5].




  1. D. M. Bigg, Polym. Compos., 7(3), (1986) 125
  2. F. Danes, B. Garnier, T. Dupuis, Int. J. Thermophys. 24, (2003) 771
  3. F. Danes, B. Garnier, T. Dupuis, Compos. Sci. and Techn. 65, (2005) 945
  4. B. Garnier, T. Dupuis, J. Gilles, J.P. Bardon, F. Danes, Proc. 12th Int. Heat Transf. Conf., publ. Elsevier, Paris, 4, (2002) 9
  5. C. Filip, B. Garnier, F. Danes, to be published J. Heat Transfer, Dec. 2007