Typical polymer heat conductivites are in the order of 0,1-0.5 W/mK (typically around 0.2). To increase the heat conductivity to 1-3 W/mK a considerable amount of high heat conductivity additive is necessary. Several years ago I was involved in the development of heat conductive silicone rubber on LSR (liquid silicone rubber) and on HTV (high temperature vulcanizin silicone rubber) basis. We used alumium nitride and boron nitride powders, but could hardly go above 2 W/mK, and even then the mechanical properties and the processability went down. Professor Ishida published a polybenzoxazine based compound if I remember well with 70 W/mK, that's a feat! Nitrides, carbides and diamond exhibit heat conductivites ranging form a few hunderd to some thousand W/mK, depending on the perfection of the crystal. If you use metals, you are in a sometwhat easirer situation, but the composite will be electrically conductive. The key problem is the scattering of phonons at the interface of the dissimilar phases, the imperfection of particle/particle contacts. Usually better result are achieved, if micro- and nano-fillers are combined (there are lots of articles on that), where the inter-particle polymer layers may be bridged over by nano-particles.

I am not an expert, but perhaps you could consider using carbon nanotubes.

You might want to see our recent Chemistry Letters paper I E where we measured templated MWCNT coatings with thermal conductivities in the low to mid diamond range, 0.5-3 kW/m/K [“Advanced nanocarbon materials facilitated by novel stimuli-responsive stabilizers;” J. Texter, D. Ager, V. Arjunan Vasantha, R. Crombez, D. England, X. Ma, R. Maniglia, and N. Tambe, Chem. Lett. 41, 1377-1379 (2012); doi:10.1246/cl.120632]. I believe a key factor is alignment, and our high values were obtained in the plane of the template; much smaller values were obtained perpendicular to the films. I think a strong factor is the exotic interaction of the imidazolium group with graphenic surfaces.

Typical polymer heat conductivites are in the order of 0,1-0.5 W/mK (typically around 0.2). To increase the heat conductivity to 1-3 W/mK a considerable amount of high heat conductivity additive is necessary. Several years ago I was involved in the development of heat conductive silicone rubber on LSR (liquid silicone rubber) and on HTV (high temperature vulcanizin silicone rubber) basis. We used alumium nitride and boron nitride powders, but could hardly go above 2 W/mK, and even then the mechanical properties and the processability went down. Professor Ishida published a polybenzoxazine based compound if I remember well with 70 W/mK, that's a feat! Nitrides, carbides and diamond exhibit heat conductivites ranging form a few hunderd to some thousand W/mK, depending on the perfection of the crystal. If you use metals, you are in a sometwhat easirer situation, but the composite will be electrically conductive. The key problem is the scattering of phonons at the interface of the dissimilar phases, the imperfection of particle/particle contacts. Usually better result are achieved, if micro- and nano-fillers are combined (there are lots of articles on that), where the inter-particle polymer layers may be bridged over by nano-particles.

Well percolation is the basic prerequisite for order of magnitude imporvement. Below that you will have maginal imporvement (almost linear with volume fraction).

Tough one! How about a composite with single crystalline diamond microparticles? You get to buy these particles from commercial sources and are of decent quality. Obviously you will not reach (and not even approach) the thermal properties of diamond but a 20 - 30 wt% composite will have good thermal conductivities. This way you will not increase electrical conductivity -- although I am not sure if that is a potential problem. And if electrical conductivity is not an issue then I would suggest (following John) carbon nanotubes -- and possibly align them along the required direction of heat flow. Another option is some form of silk -- (not the polymer that dow corning markets) -- spider silk is supposed to have very high thermal conductivity, the chinese and indian silkworm silk is not bad either. But silk is not easy to process or mold. Cheers!

Dear Palash, do you have heat conductivity data on silk? It is a polymer, so what can be the reason for extra heat conductivity???

Dear Gyorgy, here are two of the most recent articles on this: DOI: 10.1002/adma.201104668 on spider silk, supposedly with thermal conductivity as much as ~350 - 400 W/(mK) and DOI: 10.1039/C2SM26146D on relatively less conductive silkworm silk ~5 W/(mK). The high thermal conductivity and mechanical strength in spider silk is attributed to the highly oriented antiparallel β-pleated sheets in the fibrils, apparently unique to spider silk. First paper actually goes through the internal structure in detail and it could be interesting if someone could follow up on the design paradigms to synthesize new polymer. I am interested in barrier properties of silk films, not so much on thermal properties though and have not had opportunities to check on those numbers. But I am pretty sure those will strongly depend on how the fibers were harvested. Palash

Where I can find these silk fibers or films?

Thank you for the interesting references

Dear Celeste, I am not aware of any commercial source for silk protein, not yet. Mostly via bio research groups who actually specializes in silkworm / silk can provide you with processable protein -- naturally supply is -- as of now -- very limited. But please do let me know if you find any commercial source. Thanks -- Palash

After all, graphene is supposed to be the best choice. Exotic interaction results caused by the interface interaction only or by other effects?