Feature Stories - Cool new materials for hot electronics
Moore's Law, the observation by Intel co-founder Gordon E. Moore that processing power doubles approximately every two years, has been accurate for more than half a century. We now carry more processing power in the mobile phones in our pockets than could fit into a house-sized computer just a few decades ago. But there are signs that Moore's Law may not hold true for much longer, not because we can't produce more processing power on a smaller scale (new chip designs are set to solve that issue), but because having so many components in such a confined space produces too much heat. Unless a solution is found to the overheating problem, some analysts predict that the rate of increase of power density defined by Moore's Law will start to slow by 2020, severely limiting advances in mobile phones, computers and a range of other electronic devices. 'Thermal issues are the biggest challenge facing the trend toward smaller and more powerful devices, and new techniques to increase power density, such as embedded chips and 3D chip packaging, will only produce more heat,' explains Afshin Ziaei, a manager of research at Thales Research & Technology in France. Mr Ziaei coordinated the project 'Nano packaging technology for interconnect and heat dissipation' (Nanopack), which received more than EUR 7 million in funding from the European Commission. The project management team at Thales Research and Technology have worked with a team from 13 other companies, universities and research institutes to greatly advance the state of the art in thermal management and electrical interconnects. There are several approaches to tackling thermal problems, including building more efficient and effective cooling systems, for example, or developing components that produce less heat in the first place. But probably the most effective and practical solution, where the biggest gains are to be had, is at the thermal interfaces - the places where a chip connects and distributes heat to its packaging and from the packaging to the cooling system. 'These two thermal interfaces, between the core of the chip and the packaging and between the packaging and the cooling system, represent about 40 to 50 % of the thermal resistance. If we can decrease the resistance we can proportionally increase the effectiveness of the cooling system,' Mr Ziaei explains. That means chips can either run hotter, hence enabling more processing power, or they can run at the same temperature but be more reliable. The key to reducing thermal resistance at the interfaces lies in more conductive 'Thermal interface materials' (TIMs) coupled with designs that allow chips and their packaging to dissipate heat faster. Thanks to work carried out in Europe, new TIMs and processes could soon be in use commercially. Using micro- and nanotechnology, the project team has developed a new range of thermally conductive greases, adhesives, materials, structures and processes to improve the thermal interfacing of chips with their packaging and with cooling systems. Close to market, and promising for the future 'Some of them are very mature and could be in use commercially very soon, others are still in the research and development phase but look promising in the long term,' Mr Ziaei says. Among the more mature materials developed by the project partners are advanced versions of traditional thermal grease and adhesive - similar to the substances used to connect a computer processor to the heat sink in a standard PC. The Nanopack grease and adhesive solutions, enhanced by special micro-fillers, each have thermal conductivity of around 10 watts per square metre per Kelvin (W/mK) - representing the rate of transfer of heat in watts through one square metre of a structure divided by the difference in temperature across the structure. In the case of the grease, developed by Austrian project partner Electrovac from metallic micro-spheres and graphitised carbon nano-fibres in a silicone matrix, the conductivity is in line with the state of the art. The adhesive, developed by researchers at Chalmers University of Technology in Sweden, goes well beyond it, however. 'Most adhesives have a heat transfer of around 4 W/mK. At 10 W/mK, this adhesive is a major improvement,' Mr Ziaei notes. 'It is made by incorporating silver flakes and micro-silver spheres in a heat-resistant bi-epoxy matrix.' Chalmers has set up a spin-off company in Sweden, Smart High Tech (SHT), to commercialise the adhesive along with another material developed in Nanopack, a polymer fibre network infiltrated with a metal alloy. The unique material, which resembles a very fine sheet of aluminium foil and has been named SmarTIM, shows extremely efficient thermal performance of between 18 W/mK and 24 W/mK depending on the alloy used. 'The polymer fibre network creates a robust structure, while the alloy ensures efficient conductivity,' Mr Ziaei says. Nanopack partner IBM, meanwhile, improved upon an existing technology known as 'Hierarchical nested channel' (HNC), which uses microstructures on the surfaces that connect with the thermal interfaces to improve conductivity and reduce the thickness of the thermal layer. Other technologies developed in the project are further from commercial use but, once they are mature enough, they could have a major impact on thermal management. One, developed by Fraunhofer IZM, consists of a gold nano-sponge in which the cavities of the sponge are just a few tens of nanometres across. Another, developed by Thales Research and Technology, uses carbon nanotubes - cylindrical structures made from carbon allotropes with a diameter of around one nanometre, approximately 100,000 times smaller than a human hair. These are oriented vertically in a solution so heat is transferred upwards through the centre of the tube. 'These technologies are all very promising for the future. Carbon nanotubes, for instance, have excellent thermal properties. The conductivity of a single tube is close to 1,000 W/mK, and the dream is to produce materials that harness that feature and can conduct around 100 W/mk. However, if we can achieve 50 W/mK it would still be a real breakthrough,' Mr Ziaei says. In addition to the work on materials and processes, the Nanopack team also developed cutting-edge characterisation tools to measure and test the thermal performance of the materials. Some of the partners will now participate in the EU-funded 'Smart Power' project, which will broaden the scope of their work to different chip and packaging designs, and build upon the materials and processes research carried out in Nanopack. Their efforts promise to keep Moore's Law accurate for at least a few more decades. Nanopack received funding under the EU's Seventh Framework Programme for research (FP7), sub-programme 'Next-generation nanoelectronics components and electronics integration'. Useful links: - 'Nano packaging technology for interconnect and heat dissipation' project - Nanopack project data record on CORDIS Related articles: - Better thermal management promises cheaper, greener, cooler electronics - Designer optoelectronics – quantum mechanics for new materials - Nanospheres stretch limits of hard disk storage - Pushing the limits of chip miniaturisation - EUROPRACTICE offers boon to nanotechnology researchers