Brussels, 09 Mar 2004
Increasing demands on electronic devices are driving the search for new semi-conductor materials. JESICA confirmed silicon carbide offers higher performance and energy savings – but a European source is needed.
Silicon is the bedrock on which 50 years of spectacular growth in microelectronics has been founded. However, as demands on performance increase inexorably, silicon is reaching its physical limits. Increasingly, electronics are required to operate reliably under extreme environments not possible with silicon. Industry is therefore looking for next-generation 'hard' electronic materials that can operate at higher temperatures, higher powers, higher frequencies, higher efficiencies and shorter wavelengths. Such requirements arise from space, aerospace and other transport sectors, to manufacturing and energy conversion.
Silicon carbide (SiC) offers several attractive properties that make it a key candidate for such high-performance devices. In particular, it has a wide-bandgap – meaning there is a large energy gap between bonding electrons and the conduction band, so the transistor can switch more energy, and more energy is released when an electron falls between bands.
In addition, SiC:
Can support high electrical fields without breakdown – eight times more than silicon;
Offers a higher thermal conductivity at room temperature than any metal or other semiconductor, diamond excepted; and
Has a high saturated electron drift velocity that favours high frequency applications.
Taken together, these properties enable SiC-based devices to switch higher power densities, emit light at shorter wavelengths as a substrate for Group III nitride compounds such as gallium nitride, allow tighter packing densities and operate more efficiently at radio and microwave frequencies. In addition, such devices are more resistant to radiation and heat – allowing operation in extreme environments such as under car bonnets, in space or in electrical power switching.
For these reasons, SiC can have strategic applications in next-generation control devices. For example, national electricity distribution grids have high power losses as silicon-based devices operate at their limits – switching one MW can cost up to 200 kW in losses.
Estimates based on using SiC switches suggest that European greenhouse gas emissions could be reduced by 20 million tonnes annually, with a cost saving of €1 billion.
However, to capitalise on such a strategic material requires a reliable supply of raw material: the SiC wafers that will allow European device manufacturers to develop world-beating applications.
The three-year JESICA project that ran from 1999 to 2002 established the know-how and innovative manufacturing techniques for SiC wafers. It formed the basis for a European SiC source on which device manufacturers and systems houses could develop future applications.
"We aimed to start with pure SiC material, grow wafers and characterise them, and then produce working devices," explains project coordinator Christian Brylinski of Thales Research and Technology in France. "Recognising this would be a strongly iterative process that needed commitment, we formed a consortium which covered all of these aspects – from A to Z."
French members included SiC powder manufacturer Saint-Gobain, together with the French Atomic Energy Commission (CEA) and the Institut national polytechnique de Grenoble (INPG), both experts in sublimation growth – the classical method for making SiC. Swedish partners Okmetic and the University of Linköping (LiU) developed an alternative high temperature, chemical vapour deposition (HTCVD) process that gives very pure SiC crystals. French start-up Novasic offered the wafer-polishing expertise critical for a superhard material such as SiC. Finally, together with the Irish National Microelectronics Research Centre (NMRC), Thales undertook material characterisation, device fabrication and testing.
Great strides in technology
JESICA made great strides in SiC growth technology. At the start, the Swedish innovative HTCVD process grew SiC films typically 10 microns thick but, by project end, deposition rates reached more than 100 microns per hour producing 6-mm thick plates. Today, it is possible to produce plates that are several centimetres thick.
After processing into wafers, Thales manufactured device prototypes that were characterised with NMRC. "We made and tested microwave power devices – SiC-based metal-semiconductor field-effect transistors (MESFETs) – better than ever shown before on commercial SiC material," says Brylinski. "The high quality SiC from our partners ensured our devices had stable DC characteristics. This is the first time this has been clearly demonstrated and a great advance for next-generation devices."
"It is now vital for Europe not to miss opportunities with SiC," he insists. "Japan has a large national programme on SiC, driven by the potential energy savings, and in the USA there are already a few microwave device samples commercially available.
"It is not only the European component manufacturers who will benefit. The key users will be European power systems manufacturers installing switching power supplies, mobile phone networks, military and civilian air traffic control radar equipment, electricity distribution systems and TV transmitters worldwide. We cannot afford to allow a technology gap to open with our competitors."
Next generation devices
Two years after JESICA ended, the partners are going forward with SiC. Okmetic is raising €20 million for a SiC wafer plant in Sweden, while the CEA is also seeking to create a start-up company to commercialise substrates. Novasic is now a world leader in SiC processing technology – a service it is providing to customers worldwide. Thales and LiU are co-operating on advanced device development, using substrates from Okmeti, polished by Novasic.
"We have given a European SiC industry a hard push in the back, plans for SiC foundries are advanced, material is available to EU device developers, and there are now discussions on a European MESFET foundry," says Brylinski. "We expect to see these devices in professional and military equipment within five years, with consumer applications following five years later. From lagging behind Japanese and US efforts, we now have the building blocks for a European SiC industry."