Heat conducting polymers

Materials World magazine
1 Jun 2010

Stretched nanofibres of polyethylene can have a thermal conductivity larger than that of bulk material, meaning polymers can be used as efficient heat conducting materials, says Professor Gang Chen and his team at the Massachusetts Institute of Technology, USA.

Bulk polymers have low thermal conductivity, on the order of 0.1Wm-1K-1, however, they 
are much cheaper to manufacture than metals and over the last 20 years there has been an interest in trying to improve their thermal conductivity. Most attempts have focused on synthesising composites by adding metal nanoparticles or carbon nanotubes, which have higher thermal conductivity. This approach is expected to yield order of magnitude improvements based on an effective medium rationale.

However, the composite approach has not shown any drastic increase in thermal conductivity because of the additional interfacial resistance incurred at the boundaries between the additives and the matrix. The approach taken by researchers at the Massachusetts Institute of Technology, USA, focuses on harnessing the intrinsically good heat conduction properties of the individual polymer molecules, rather than adding other high thermal conductivity materials.

Computer simulations of atomic motion have predicted that polyethylene (PE) could have a thermal conductivity as high as 180±65Wm-1K-1. The challenge has been finding ways to align the molecules in the same direction. The subsequent experimental work has resulted in three major developments:

• A new technique has been developed for fabricating crystalline PE nanofibres – highly aligned molecules forming crystals with diameters ranging from 50-500nm.
• A technique for stretching and measuring the fibre thermal conductivity, using a bi-metallic AFM cantilever.
• Demonstrations of thermal conductivities for individual fibres as high as 104Wm-1K-1 have been reported. These are not only the highest values recorded for any polymer, but they are also higher than 60% of elemental metals.

The fabrication procedure used at MIT is a hybrid process based on gel spinning and drawing. First, a PE nanofibre 100-200µm long is drawn from a PE gel, composed of 0.8wt% ultra high molecular weight PE (Molecular weight: ~6,000,000) dissolved in decalin solvent at 145ºC using a drawing stylus. The drawing stylus consisted of a tungsten tip (tip diameter <1µm) and an atomic force microscope bi-material cantilever (tip diameter <4µm), where the cantilever tip is only needed for subsequent thermal conductivity measurements. The gel is initially heated to 120ºC. Once the fibre is drawn, it is heated again to 90ºC, and the fibre is mechanically stretched to achieve an even higher draw ratio. The resulting PE nanofibres had diameters ranging from 50-500nm and lengths up to tens of millimetres. Thermal characterisation of these fibres resulted in thermal conductivities ranging from 53-104Wm-1K-1 with the higher conductivity values corresponding to the nanofibres with the larger draw ratios.

The new synthesis technique needs to be scaled up to form fibre bundles, or sheets as bulk material. If this goal can be reached, high thermal conductivity polymers could serve as a cheaper alternative to metals, especially in applications where a nominal downgrade in performance is acceptable in exchange for a decrease in cost.

The fibres are lightweight compared to metals. This would make them an attractive alternative in automotive and aerospace applications, where decreasing weight also reduces costs.

An additional a feature of thermally conductive PE is that it remains electrically insulating. This means that high thermal conductivity polymers could be used in applications where high thermal conductivity and low electrical conductivity is needed. For example, uses could be found as sheathing for superconducting wires, where the aim is to electrically insulate the high current carrying wire from their surroundings, while presenting minimal resistance to heat transfer from the cryogenic fluid. Polyethylene nanofibres are chemically inert, meaning they may be useful in chemical plants where non-corrosive materials are required, or applications such as solar hot-water collectors and heat exchanger fins.

Origin of the phenomenon

Understanding the underlying physics that give rise to high thermal conductivity in nanofibres is not only beneficial for further developments, but could aid in identifying more applications where they would be useful.

The high thermal conductivity comes from the ability of the molecular chains to act as efficient conduits for vibrational energy in the atoms. This is fundamentally different than thermal conduction in metals, where electrons are the primary heat carriers. As a result, heat conduction in metals is intimately interconnected with the electrical conductivity.

The key with polymers is to recognise that individual polymer chains behave as 1D conductors, resulting from the one dimension of symmetry in the atomic arrangement. This effect has important consequences that can be traced back to the pioneering work of Fermi, Pasta and Ulam, where they discovered that a nonlinear chain of oscillators could theoretically have infinite thermal conductivity. With this in mind, the important criterion for high thermal conductivity in polymers is to restructure their molecular arrangement so they form an aligned array of chains. This is challenging because the chains have a strong tendency to entangle, forming a disordered network in the bulk material. This causes the 
bulk thermal conductivity to be as low as ~0.35 Wm-1K-1.

The nanostructuring effect

Previous experiments have shown that when bulk PE is stretched, thermal conductivity can be increased by as high as 42Wm-1K-1 along the stretch direction, which is around 120 times that of the bulk. The ceiling in these experiments, however, appears to be dictated by the fact that, at some point, further stretching results in fracture.

Fractures are typically initiated at defect sites in the forming polymer chain lattices, which are often due to voids or amorphous regions. Comparing MIT’s nanofibres to bulk PE, the defect density would be expected to be comparable. However, the key benefit of nanostructuring PE, is that the total number of defects is reduced. Therefore, as the fibre is stretched, there is less likelihood of fracture, since only a single crack initiated by a defect is needed for failure.

Consequently, stretching a nanofibre brings it closer to an idealised single crystal before it fractures, as opposed to stretching bulk samples, which are more likely to fracture earlier in the process.

Scaling up

A number of issues and challenges remain with scaling up the nanofibre fabrication process so the high thermal conductivity is preserved. One way this could be accomplished is by manufacturing large batches of fibres that are woven or collected into bundles or sheets. Similar rope-like structures have been formed from carbon nanotube bundles, which then form fibres and can be manufactured into a macroscopic material that preserves many of the properties exhibited by individual fibres.

The other major constraint in scaling up the manufacturing process is cost. Using a high thermal conductivity polymer as opposed to a metal reduces cost. If the manufacturing processes required to create macroscopic materials that maintain these properties are expensive, it can underwrite the overall objective. The MIT team believes, however, that these challenges can be overcome by new processing techniques.

Further information: Dr Gang Chen