While pure silicon power transistors have achieved enviable performance improvements, they seem to have reached their limits for demanding power switching and control applications. Silicon carbide (SiC), as a novel semiconductor material, offers potential advantages: smaller size, higher efficiency, complete elimination of switching losses, low drain current, higher switching frequencies than standard semiconductors (pure silicon semiconductors), and the ability to operate above the standard junction temperature of 125°C. Miniaturization and high operating temperature tolerance make these devices more versatile, even allowing them to be placed directly within the housing of motors. Like any new technology, SiC undergoes a development and maturation process. Standard power switches, such as insulated-gate bipolar transistors (IGBTs), have a large product base and optimized manufacturing technologies. SiC, however, requires significant investment in research and development to address material challenges and refine semiconductor manufacturing techniques. Nevertheless, the ability of these power switching devices to rapidly switch between high forward current and reverse cutoff voltages is a worthwhile prospect. SiC's initial successful application and primary use are in light-emitting diodes (LEDs) for automotive headlights, dashboards, and other lighting applications. Other markets include switching power supplies and Schottky barrier diodes. Future applications include hybrid vehicles, power converters (for reducing the size of active pre-filters), and AC/DC motor control. These more demanding applications are not yet commercialized because they require high-quality materials and large-scale production to reduce costs. Globally, significant research funding is being invested in companies, laboratories, and government facilities to make SiC technology more feasible. Some experts predict that the commercialization, industrialization, and even military applications of SiC technology will become a reality within 2 to 5 years or more. [align=center] Figure 1: This fully functional SiC-based 3kW three-phase converter prototype from APEI Inc. can operate at temperatures above 250°C. [/align] Motor control manufacturers are particularly interested in the development of SiC, with some even collaborating with researchers and semiconductor manufacturers to promote its development. However, most remain silent about such collaborations. Dr. Gary Skibinski, a consultant engineer in the Standard Drives division at Rockwell Automation , a pioneer in SiC technology, said, “Rockwell Automation sees the potential benefits of this new technology and considers itself a promoter of SiC. Rockwell has also identified how SiC technology will be integrated into its future business plans. For a leading company, understanding and embracing emerging technologies is crucial.” Development is underway. Skibinski cited the example of adding a SiC power diode to each standard IGBT in a drive module, acting as a polarity-changing flywheel diode, as a logical first step in improving productivity; this change will then be applied to power switches. He said, “Pure SiC drives are still in the R&D and prototyping stage.” Compared to the progress of the pure SiC model (Si IGBT + anti-parallel diode switch), Rockwell has recently made breakthroughs in reducing energy loss and increasing carrier frequency in its research on the Si-SiC hybrid power model (Si IGBT + commercial SiC diode). The total power loss of this model is Eon + Err + Eoff (see Figure 2). For Si or SiC diodes, the Eoff value remains constant regardless of the Rgate value. However, when using SiC diodes, the other two power loss components change with the Rgate value. For any Rgate value, the diode reverse recovery loss Err is virtually reduced to 0 (94%). When Rgate = 25Ω, the Eon of the GBT is reduced by 37%, and when Rgate = 8Ω, the Eon of the GBT is reduced by 85%. [align=center] Figure 2: Recent research by Rockwell Automation shows that Si-SiC hybrid modules can potentially reduce power losses Eon and Err compared to all-silicon modules. For ease of comparison, the IGBT power loss En of all-silicon modules is normalized to 3.3mJ per unit. [/align] The results demonstrate the possibility of higher switching frequencies, which have previously been limited by the reverse recovery loss of pure silicon diodes. Err has limited further development in reducing turn-on losses. Skibinski explained, “Silicon module suppliers recommend using a gate resistor Rgate (e.g., 25 Ω) to balance the IGBT’s turn-on energy loss (Eon) and turn-off energy loss (Eoff).” However, for SiC diodes, the gate resistor Rgate can be omitted. He said, “SiC diodes can reduce total power loss (Eon + Err + Eoff), a characteristic that has potential advantages in driver applications.” First, under the same cooling system conditions, it can achieve a switching frequency of 4 times, allowing for better performance, smaller size, and lower price for the pre-filter. Alternatively, you can retain the current switching frequency and cooling system, resulting in higher efficiency and stability, lower losses, and higher rated output. Reduced total power loss can potentially reduce cooling costs. Yaskawa Electric, another driver manufacturer using SiC technology, applies SiC technology to radar screens. Yaskawa Electric summarizes the basic advantages of SiC as: high operating temperature, high switching speed, and lower losses in both on and off modes, making the driver system more efficient. Tsuneo J., a special member of IEEE at the research laboratory of Yaskawa Electric, Kogura, Japan. Dr. Kume stated in Control Engineering, "This low-loss characteristic, coupled with a high operating junction temperature, allows for smaller sizes of silicon carbide devices and cooling systems, making it possible to drive systems with higher power density. Furthermore, high-frequency switching performance significantly improves the response and bandwidth of control systems." Yaskawa is working closely with leading semiconductor manufacturers, such as Mitsubishi Semiconductors, and will launch advanced SiC devices once the technology matures. According to Kume, this technology is undergoing further testing for practical applications and quality; drive products using this new technology have not yet begun development. Agile and Innovative Small, innovative companies often contribute to the advancement of technologies. In the SiC field, one such example is Arkansas Power Electronics International Inc. (APEI). APEI specializes in developing high-performance power electronic systems using SiC devices as a core technology. Dr. Alexander B. Lostetter, President of APEI, said, "APEI focuses particularly on technologies for applications in extreme environments (temperatures above 500°C) and/or with very high power density." Figure 3: TranSiC's BitSiC1206 bipolar silicon carbide power transistor is a 6-A prototype that has already been shown to users. A 30-A prototype is planned for release by the end of 2007. APEI has developed, manufactured, and tested SiC-based DC and AC motor drives, single-phase and three-phase converters (rated at 3kW and 5kW), and DC-DC converters. Lostetter introduced other research advancements by the company, including: high-temperature packaging technology that allows single devices to operate at 500°C or higher; and SiC-based analog/digital low-voltage circuit control, which allows circuitry to be integrated into power control systems operating above 300°C. Also under development is an operational amplifier based on discrete SiC junction field-effect transistors (JFETs) that can operate at 500°C. Lostetter said, "High junction temperatures reduce the size of thermal management systems in electronics and enable them to operate at high power densities." In 2005, SiC expert Swedish founded another small company active in SiC power transistor development—TranSiC AB, spun out from the Royal Institute of Technology (KTH) in Stockholm. Recently, TranSiC AB successfully completed the prototype demonstration of its bipolar transistor in a standard TO247 package. The first model, BitSiC1206, is a 1200V, 6A device. TranSiC CEO Bo Hammarlund mentioned that the chip packaging was very successful, and the switching performance was excellent compared to similar products. The company sources SiC wafers and external materials from a variety of sources, but the critical chip processing is all done in the KTH laboratory. Hammarlund explained that BitSiC's industrial packaging is done by an experienced outsourcing company, but when customers are pilots, TranSiC can provide rapid packaging with short lead times because, in such applications, packaging price and development speed are related. TranSiC hopes to mature and cost-effectively make its BitSiC products within two years. Currently, each of these chips is very expensive. Hammarlund explained, "We will implement every cost reduction measure, and we hope to reduce costs by 30% every six months over the next two years." The next goal is to have a large customer base for BitSiC by the end of 2006 and to provide instruction manuals on TranSiC's website. Hammarlund told Control Engineering, "Our goal is to complete a 30A device prototype, plus a package prototype that can withstand a junction temperature of 225°C, by the end of 2007." This is part of the company's long-term development roadmap for higher currents. Different Perspectives Not everyone agrees on the prospects of silicon carbide power control. ABB, an expert in high-power semiconductors, terminated its SiC development project in 2002 at its joint development center in Sweden. One explanation from the company's semiconductor R&D department for this move was the bipolar conduction attenuation effect caused by Basel planar faults. This reflects a prospect solely based on high-voltage/high-power devices and applications. Munaf, chief engineer of ABB Switzerland Semiconductors' R&D department... Dr. Rahimo stated, “Silicon carbide is suitable for low-voltage unipolar diodes in the short term, and it also has potential for low-power bipolar transistors and junction field-effect transistors in discrete high-frequency applications. However, due to the higher PN junction gate height between SiC and Si materials, bipolar diodes are only somewhat useful for reducing conduction losses in devices with rated voltages above 4.5kV. On the other hand, SiC bipolar transistors are not hindered by this drawback, and in the long run, they are still more noteworthy than other types of switches in high-voltage applications.” Regarding the fast switching capability of SiC technology, ABB Semiconductors confirmed that this high-frequency operating capability is only applicable in environments with low stray inductance, such as low-power/low-voltage systems. Rahimo said, “In high-power systems, stray inductance is large, requiring the semiconductor to perform switching actions slowly. For SiC devices, this means allowing them to switch slowly to accommodate the requirements of the buffer, which reintroduces losses that we were originally trying to eliminate with expensive SiC devices.” Furthermore, Rahimo stated that the price of the underlying material is currently 100 times that of ordinary materials (for 3-inch SiC wafers), but may drop to 10 times in the future. While the quality of SiC wafers has improved, allowing device manufacturers to use smaller molds (5mm² wafers) while maintaining production volume, the yield of larger molds (e.g., 25mm² wafers for 50A diodes) is very low. Compared to monolithic integrated circuit diodes with 6-inch and 12-inch wafer diameters, the quality of 4-inch SiC wafers remains very poor. He added, "The defect rate for Si wafer production is very low, leading SiC power devices by 5 to 10 years." The timeline for high-power SiC devices may be even longer. Other developers are also aware of the shortcomings of SiC but continue development. Expected Breakthrough For APEI APEI Inc.'s application is aimed at stimulating breakthroughs in SiC technology, focusing on devices and packaging. Lostetter stated, "At the device level, wafer yield is the primary concern; low yield means a slow pace of device commercialization." APEI is reportedly collaborating with numerous international manufacturers to achieve high-yield devices, and is currently in the R&D phase. However, if devices cannot be commercialized, systems using these devices cannot be commercialized either. Other key points mentioned by the company include the need for voltage-controlled normally closed devices that do not reduce power density performance, and semiconductor gold plating processes that can withstand long-term reliability and high-temperature environments. Many current devices are normally open or current-controlled. Lostetter explained that developing power systems using normally open devices requires extreme caution, particularly regarding protection in the event of a catastrophic failure, such as preventing all power devices from turning on simultaneously by directly grounding the power supply. [align=center] Figure 4: A DC motor drive made entirely of SiC material, developed by APEI Inc. for NASA's Venus landing robot, can be used in temperatures exceeding 500°C. [/align] At the materials level, APEI sees a demand for high-performance mold materials. These materials must be long-term reliable, diffusion-resistant, and corrosion-resistant at the power layer, and mechanically reliable enough to withstand extreme temperature fluctuations without damage from thermal expansion and stress cracking. Some prototypes have emerged, while some SiC drive projects still have a long way to go. APEI mentioned exciting developments in its collaboration with US government clients. High-power-density three-phase motor drives are being used in the US military's Future Combat Systems (FCS) program. This system connects fully electronic or hybrid electronic combat vehicles, with a target completion date of 2020. Lostetter said, "When APEI demonstrated 3-5kW SiC motor drives, the actual demand was likely already at 100-1000kW. We believe these goals will be demonstrated within 2-3 years, and commercialization within the next 3-5 years." APEI's more ambitious project is developing an extreme-environment DC motor drive for NASA's Jet Propulsion Laboratory's Venus Explorer (VISE) lander. This project, still in the design phase, aims to deploy the robot to the surface of Venus. VISE is similar to successful Mars exploration robots, but Venus faces a much harsher environment: surface temperatures exceed 485°C, pressures exceed 90 atmospheres, and the atmosphere has a high concentration of sulfuric acid. NASA's timeline, according to Lostetter, states, "APEI Inc.'s SiC-based motor drive would allow the robot to perform traction, load-bearing, and other functions using gear transmission, eliminating the need for expensive and heavy electronic protection and heat treatment systems, which were previously essential. Furthermore, silicon-based systems would fail shortly after coolant depletion." APEI estimates the all-SiC motor drive will be completed by 2010, aligning with NASA's timeline to launch the VISE project in 2013. Meanwhile, TransSiC's Hammarlund is confident in overcoming the challenges of SiC materials, summarizing, "At that time, our production volume will be very large, and we can provide 50A chips." The current state of power-controlled silicon carbide is similar to a "Catch 22" situation. When reliable SiC devices are widely available and inexpensive, users and corresponding applications will naturally emerge. On the other hand, SiC can only be commercialized when there is sufficient user demand.