Yasuzaemon Matsunaga was an executive in the young Japanese electric power industry where his role in defining the sector's future saw him nicknamed the "demon of electricity". He fought and lost a desperate battle against the state control of power supplies leading up to the Second World War, a battle that ended in his resignation.
However, after the war, Matsunaga returned as chairman of the advisory council for the industry's reorganization. From 1949 to 1951, he oversaw the creation of the present system of nine regional private power companies, which replaced the state-run regime. Also in 1951 he established a non-profit public benefit corporation called the Central Research Institute of the Electric Power Industry (CRIEPI).
Today, as then, CRIEPI's mission is to carry out "research for the benefit of society as a whole". Compound semiconductors have a key part to play in this effort, most particularly SiC power electronics for use as an "optimal energy utilization technology".
"We believe that SiC inverter technology with low-loss power conversion can enhance the use of electric power in households, industry, transportation and power distribution," explained Hidekazu Tsuchida, leader of CRIEPI's SiC research team.
The research model used by CRIEPI allows it to undertake studies either independently or collaboratively. It mainly teams up with Japanese regional power firms, who provided the majority of CRIEPI's ¥33.71 billion ($311 million) budget for 2008. The institute's intellectual property can then be transferred for external partners to use in manufacturing or operating electrical power systems.
One such research collaboration saw Tsuchida's team partner with AIST (the Japanese National Institute of Advanced Industrial Science and Technology), and chemical giant and LED maker Showa Denko, to improve SiC wafer growth processes. The lessons learned by this collaboration have since been transferred to SiC epiwafer manufacturing start-up ESiCat. The partners set up ESiCat in 2005, and the company is now selling 2 and 3 inch SiC epiwafers.
Tsuchida's team also collaborates with the Kansai Electric Power Company, investigating SiC defects formed during epitaxial growth and learning how to diminish their effects when fabricating power devices. As is common for CRIEPI's SiC work, this focuses on epilayer growth on 4H-SiC substrates, which it says are particularly suitable for the manufacture of SiC power devices.
The main target is high-voltage bipolar SiC devices, which often suffer from performance degradation when a current is applied. The reason has been narrowed down to imperfections called single Shockley faults that form in the devices. These are nucleated by dislocations on the basal planes of SiC epitaxial layers, known as basal plane dislocations (BPDs).
"Reduction of the dislocation density in SiC substrates and epilayers is crucial to ensure reliability, especially for MOSFETs and bipolar devices," explained Tsuchida. "Understanding the formation of the dislocations can give hints for new methods to control or reduce them."
The Kansai-CRIEPI collaboration has found success in eliminating certain types of BPD and preventing their propagation from substrate to epilayer, particularly by finely tuning which SiC crystal face the team uses as a substrate. It has also produced devices demonstrating the capability of SiC. These include Zener diodes, which are important in regulating constant-voltage circuits, with a pulse operation capability of 8 kW, compared with an approximate maximum of 3 kW for silicon rivals.
Perhaps CRIEPI's most notable recent achievement is in high-speed SiC epitaxy, leaping to a growth rate of 250 µm/h – more than 20 times as great as the rate seen in multiwafer reactors. Induction heating warms the sides of the vertical hot-wall reactor that the group uses (figure 1) and the susceptor on which the substrate sits is heated by the resulting thermal radiation. Substrates are loaded from the bottom and the susceptor can be raised close to gas inlets at the reactor top, preventing the unwanted deposition of reagents on the reactor sides.
Rotating the susceptor at 10 rpm during epitaxy ensures even doping levels and SiC thickness across the entire wafer. Low system pressures and high gas flow rates into the reactor lead to rapid film growth and help to suppress the unwanted material nucleation that leads to defects.
"Our high-speed epitaxy is aimed at making very thick epilayers to produce very-high-voltage devices for electric power conversion in the power transmission/distribution system," Tsuchida explained.
At the October 2007 International Conference on Silicon Carbide and Related Materials in Otsu, Japan, CRIEPI shared the podium with many big-name Japanese companies. With Panasonic, Mitsubishi and Toyota, to name but three, performing their own research independently from CRIEPI, Japan is clearly a major force in SiC. According to Tsuchida, it is the leading force.
"In my opinion, the number of companies in Japan involved in researching SiC materials, processing and devices is now the largest in the world," he said. "Producing low-loss power semiconductor devices is widely accepted as a key issue for energy saving and CO2 emission reduction in this country."
With such a well educated outlook, perhaps we will see the fulfillment of SiC's promise in low-loss electrical networks in Japan. What's more, thanks to the intimate relationship with Japanese electricity companies that Matsunaga specified when it was founded, it looks like CRIEPI is destined to be the research institute at the center of the rise of SiC.
M Ito et al. 2008 Appl. Phys. Express 1 015001.
L Storasta et al. 2008 J. Appl. Phys. 103 013705.
View pdf of article