Click here for the enhanced PDF version of this article

The requirements of modern high-performance power electronic systems, in particular for the military applications as shown in Figure 1, are surpassing the power density, efficiency and reliability limitations set by the inherent properties of widely employed silicon-based devices. To overcome this, new device technologies are being explored. The past decade has seen an intense and steady increase in resources being funneled into the research and development of widebandgap devices, such as silicon carbide (SiC). SiC power devices hold the promise of vastly exceeding previously constraining restrictions imposed by silicon-based devices. This new technology is just now beginning to find its way into the commercial marketplace. While the present focus in the market is on developing power devices such as high-voltage, high-current MOSFETs, the truly revolutionizing potential of SiC has yet to be tapped.

The potential applications of SiC are widespread and all encompassing in the area of power electronics. The ability to greatly increase the power density of current power systems makes the technology attractive for every branch of the military. The Army's Future Combat Systems (FCS) program will require lighter, more compact power supplies to easily deploy the new communications and computers systems, networked logistics systems, and intelligence, reconnaissance and surveillance systems. Moreover, FCS will require a range of high-efficiency power supplies for the infantry and new ground vehicles, such as the armed robotic vehicle (ARV), small unmanned ground vehicle (SUGV), multifunctional utility/logistics and equipment (MULE) and the infantry carrier vehicle (ICV). The Air Force, through its More Electric Aircraft (MEA) program, aims to minimize and replace hydraulic control systems with light, low-maintenance electric actuators and motor drives. Last, the Navy's next-generation destroyer DD(X) will require high-voltage and high-power-density systems to implement the envisioned compartmentalized power-distribution architecture.

The ability of SiC-based systems to operate in harsh environments or at high temperatures (up to 600 °C) also opens the door for new systems impacting applications such as space exploration vehicles and landers, nuclear power reactors, and petroleum and geological exploration instrumentation. Ultimately, any system that would see improvement from high-density or high-efficiency power electronics would benefit from SiC.

This includes commercial electronics, automobiles (onboard sensors, electric options), electric vehicles, household appliances, industrial manufacturing equipment, the national electrical grid, and mass transportation systems (trains). Therefore, the future of power electronics will be greatly influenced by the commercialization of SiC semiconductor devices.

During the late 1960s, the electronics industry was revolutionized by the development of silicon integrated circuit (IC) technology, resulting in microelectronics applications shrinking by orders of magnitude in comparison with their discrete component counterparts and, ultimately, leading to vast cost reductions in electronics markets. A similar revolution will occur in the power electronics industry, driven by the SiC power switch. SiC has one-tenth the switching losses of silicon, 10 times the blocking voltage, four times the thermal conductivity, and 10 times the switching speeds. SiC technology also provides a junction temperature threshold in excess of 600 °C. All of these physical advantages that SiC has over current silicon technology will greatly enable increased power density, which is the chief limiting factor of today's power electronic systems. It will also significantly enhance energy efficiency, and shrink the size of power electronics systems by an order of magnitude. All of these factors will also result in cost savings. Whereas the IC drove the computer revolution that shrank mainframes to the size of wall cabinets to fit on a desktop, so too will SiC technology be the prime mover behind shrinking wall-sized power electronics systems to the size of a suitcase.

A powerful argument for using SiC power electronics is the size and weight reductions that can be achieved with high-temperature operation. For example, a silicon-based power module with a 3 kg heatsink can achieve a maximum power of 5 kW assuming a junction temperature of 150 °C; while a SiC-based power module with a 0.3 kg heatsink can achieve a maximum power of 7.5 kW, assuming a junction temperature of 600 °C. This implies that the use of SiC technology allows for a 50% increase in power and a 90% reduction in weight and volume. To take full advantage of the high-power density capabilities offered by SiC electronics, the development of high-temperature electronics as well as high-temperature packaging technologies and design methodologies are required. In particular, the integration of high-temperature power devices and high-temperature control electronics into a single module greatly minimizes parasitics, allowing for very high frequencies of operation.

Arkansas Power Electronics International has developed several SiC-based power converters to demonstrate the true potential of the SiC technology. Figure 2 illustrates a complete SiC-based multichip power module (MCPM) that operates at temperatures in excess of 250 °C ambient. This highly compact 4 kW three-phase MCPM inverter integrates SiC power transistors with high-temperature silicon-on-insulator (HTSOI) control electronics. The high-temperature operation of the MCPM allows for increased power density by an order of magnitude when compared to equivalent silicon-based systems. This opens new possibilities in the design of many power systems. An example of this is shown in Figure 3, where a 10 hp induction motor and drive are integrated into a single unit. The power electronics integrated with the motor is attached to a small passive cooling heatsink in order to minimize the increase to the motor's thermal load. A motor design that allows for increased thermal loading could even eliminate the heatsink altogether.

The electrical design of the MCPM stems from a demonstration three-phase motor control developed for high-temperature operation. The high-temperature motor controller operated in an environment of 250 °C^[1-2]. Figure 4 illustrates the circuit block diagram of a three-phase motor drive MCPM design. The control electronics were developed using HTSOI SiC components, which are guaranteed to operate at 225 °C for five years (and one year at 300 °C).

The digital core control block (block 1) contains a microcontroller, latch, SRAM and proprietary software to generate the control signals required for a three-phase motor drive. The heart of this digital core control block is the monolithic 8-bit microcon-troller that uses the standard MCS-51 instruction set. Key features include the programmable counter array, watchdog timer, enhanced serial port for multiprocessor communication and a hierarchical interrupt structure.

In addition to the digital core control block, the system contains start-up circuitry to deliver power from the dc bus to the low-voltage control logic (block 2). Another feature included is feedback of critical conditions such as overvoltage, overcurrent and over-temperature (block 3). Block 4 takes the low-voltage digital signals from the microcontroller and amplifies the voltage and current to drive the isolation transformers in block 5. Block 6 can be customized to drive different types of SiC power switches. In this case, the design drives the gate of a SiC JFET from -40 V (fully off) to 0 V (fully on). The SiC JFET gate drive circuitry also ensures that there is adequate dead time between the high- and low-side switching.

Figure 5 illustrates the cross-section of the MCPM design approach. The MCPM has two main stages, the control and power stage. The control stage requires a low-power, high-density substrate in order to house all of the HTSOI control electronics. The power stage requires a thermally conductive substrate, with high current-carrying capabilities and high-voltage isolation, to house the SiC power devices. Both of these substrates must withstand high temperatures. Other key components of the mechanical design of a high-temperature MCPM are the heat spreaders or base plates, wire bonds and die attaches.

The selection of packaging materials for a high-temperature (>250 °C) MCPMs is a great challenge because many common packaging materials used in today's electronic systems cannot be used at these temperatures. Therefore, new packaging materials must be selected or developed. An important critical step in choosing packaging materials is matching the coefficient of thermal expansion (CTE) for adjoining parts. Since SiC transistors will be used in the power stage, it is important to closely match the CTE of both the power substrate and heat spreader to that of SiC (between 4.6 ppm/°C and 5.1 ppm/°C) to reduce thermal stress failures. Table 1 compares the characteristics of common substrates used in power electronics to that of SiC. In the design of Figure 5, the power substrate is an aluminum-nitride (AlN) or alumina ceramic substrate, bonded on either side with 10-mil to 12-mil copper (direct bond copper [DBC]), which allows for excellent thermal and electrical conductivity. The copper is nickel-plated or gold-plated in order to enhance surface solderability and long-term resistance to thermal oxidation. The bare-die SiC components are attached using high-temperature solders, and then wire bonded, completing the electrical circuits. These wire bonds are normally large diameter (10 mils) aluminum (Al) bonding wire; but can be gold, depending on the surface metallization. APEI has performed thermal cycling experiments for the optimization of Al wire bonds showing reliable operation at temperatures beyond 250 °C^[3-4].

Since most of the SiC power devices are vertical devices, the connections between the die and the substrate are not only important for heat transfer, but also for electrical or current transfer as well. A metallurgical process such as soldering can accomplish this attachment. Since the operating temperature of the MCPM will be approximately 250 °C, solders above 300 °C solidus are used. Last, the power substrate is attached to a heat spreader. The selection of a heat spreader is important to the package as it provides mechanical strength and is in the direct line of heat transfer from the power substrate to the heatsink or heat exchanger. Heat spreaders are often made out of metals such as copper, aluminum or metal matrix composites (MMCs). The heat spreaders used by APEI are often an AlSiC MMC. The selection of the heat spreader is made from three main criteria. First, the CTE of an MMC can be adjusted to match the CTE characteristics of the rest of the package to reduce the stresses of thermal expansion within the MCPM. Second, the thermal conductivity of AlSiC is excellent for heat transfer. Third, AlSiC is available commercially. Table 2 compares the characteristics of different heat spreader materials.

The control substrate can be implemented in a variety of ways, depending upon the maximum temperature of operation. For temperatures less than 225 °C, Isola P96 (Tg > 260 °C), Rogers 4000 (Tg = 280 °C) or Arlon 527 (Tg = 350 °C) could be used reliably. For higher temperatures of operation, APEI has developed a new approach that allows for high circuit density as well as continuous operation at temperatures as high as 400 °C. The control substrate is normally gold-plated (Au-plated) since many of the control components, such as HTSOIs and passives, have Au-plated pads. Some of these components do not have a backside electrical connection; therefore, non-conductive (or highly resistive) epoxies may be used as well as high-temperature solder to attach the components to the control substrate. Finally, the control components are wire bonded to the control substrate using small diameter (0.75 mil, 1.0 mil or 3.0 mil) Au wire bonds. APEI has performed experiments for the optimization of these wire bonds showing reliable operation at temperatures beyond 500 °C^[3-4].

To verify the MCPM design for high-temperature operation, a thermal model was generated in FLOTHERM for detailed 3-D thermal analysis. The goal of the thermal simulations was to closely estimate and simulate the actual conditions experienced during operation. In these simulations, a worst-case thermal load of 500 W was modeled, which represents a 10 hp motor drive with an overall efficiency of 93%. The results of these simulations in Figure 6 show excellent thermal spreading and an effective removal of generated heat through the heatsink. Figure 6 also shows that the maximum die junction temperature is 240 °C (or a 190 °C rise over the ambient of 50 °C), and the control electronics' maximum temperature is close to 200 °C.

The penetration and widespread use of SiC devices and systems will be heavily dependent on the cost and availability of different types of devices. Since the first SiC power device was released to the commercial market (a SiC Schottky diode in 2001), there has been an extensive effort to commercially develop other SiC power devices^[5]. These devices include the metal-oxide-semiconductor field-effect transistor (MOSFET), junction field-effect transistor (JFET), static induction transistor (SIT), gate-turn-off (GTO) thyristor and bipolar junction transistor (BJT). A large selection of SiC devices will allow designers to include them in a larger number of applications. In turn, the cost of these devices will be reduced as production levels are ramped up and confidence is gained in the new technology. Finally, other markets such as the rapidly growing SiC LED market, will help the commercialization of SiC power devices, since technologies developed for this market (such as the growing of SiC wafers) can be directly transferred to the SiC power device manufacturing process. When the cost of SiC begins to drop, the technology will have the potential to become ubiquitous in the power electronics industry.

1. A. Lostetter, J. Hornberger, S. Magan Lal, K. Olejniczak, A. Mantooth, and Aicha Elshabini, “Development of Silicon-Carbide (SiC) Static-Induction-Transistor (SIT) Based Half-Bridge Power Converters,” Proceedings of the 2003 IMAPS Conference, Boston, MA, October 2003.

2. J. Hornberger, A. Lostetter, T. McNutt, S. Magan Lal, and A. Mantooth, “The Application of Silicon Carbide (SiC) Semiconductor Power Electronics to Extreme High-Temperature Extraterrestrial Environments,” Proceedings of the 2004 IEEE Aerospace Conference, MT, March 2004.

3. J. Hornberger, A. B. Lostetter, K. J. Olejniczak, S. Magan Lal, and A. Mantooth, “A Novel Three-Phase Motor Drive Utilizing Silicon on Insulator (SOI) and Silicon Carbide (SiC) Semiconductor Power Electronics for Extreme High-Temperature Environments,” IMAPS 37^th International Symposium on Microelectronics, Long Beach, CA, November 2004.

4. H.A. Mustain, A.B. Lostetter, W.D. Brown, “Evaluation of Gold and Aluminum Bond Performance for High Temperature (500 °C) Silicon Carbide (SiC) Power Modules,” the 55th Electronic Components and Technology Conference, Lake Buena Vista, FL, 2005.

5. Phlippen and Burger, “A New High Voltage Schottky Diode Based on Silicon Carbide (SiC),” 2001 EPE, Graz, Austria.

Marcelo Schupbach is the senior engineering manager of APEI. He received his B.S.E.E. at the Universidad Nacionalde la Plata (Argentina) in 1998, and his M.S.E.E. and Ph.D. at the University of Arkansas in 2000 and 2004, respectively. Schupbach's expertise lies in the design and development of state-of-the-art extreme environment SiC electronic systems, in which he has more than two dozen internationally chaired conference and journal publications. Schupbach is the technical lead of internal projects and commercial contracts for the development of high-power density SiC-based power electronic systems. He has also led work in creating SiC-based dc-dc converters for extreme environment applications involving the Army's Future Combat Systems program. Schupbach heads the development of several commercial and military SiC programs at APEI. He can be reached at mschupb@APEI.NET.

Alexander Lostetter is the president of APEI. He received his B.S.E.E. and M.S.E.E. degrees from Virginia Polytechnic Institute and State University in 1996 and 1998, respectively, and his Ph.D. in Microelectronics from the University of Arkansas in 2003. He was employed as a reliability and failure analysis engineer in the Semiconductor Technology Center of the Space Electronics Division for Lockheed Martin. Lostetter joined APEI in 1999, where he led the extreme environment power electronics research initiative. He was promoted to chief operations officer in August 2002 and promoted to president in October 2003. He has published more than 40 articles and journal papers in the area of power electronics systems, design, miniaturization and packaging, including a chapter in the IEEE released Advanced Packaging, 2^nd edition textbook with the chapter entitled “Power Electronics Packaging.” He can be reached at alostet@APEI.NET.

RSS    Save to Del.icio.us  Digg This