What you need to know about silicon carbide in power applications

Silicon carbide (SiC), a semiconductor compound consisting of silicon (Si) and carbon (C), belongs to the wide bandgap (WBG) family of materials. Its physical bond is very strong, giving the semiconductor a high mechanical, chemical and thermal stability. The wide band gap and high thermal stability allow SiC devices to be used at junction temperatures higher than those of silicon, even over 200°C. The main advantage offered by silicon carbide in power applications is its low drift region resistance, which is a key factor for high-voltage power devices. [Here “10 things to know About GaN”]

SiC-based power devices are driving a radical transformation of power electronics, thanks to a combination of excellent physical and electronic properties. Although the material has been known for a long time, its use as a semiconductor is relatively recent, in great measure due to the availability of large and high-quality wafers. In recent decades, efforts have focused on developing specific and unique high-temperature crystal growth processes. Though SiC is available with different polymorphic crystalline structures (also known as polytypes), the 4H-SiC polytype hexagonal crystal structure is the most suitable for high power applications. A six-inch SiC wafer is shown in Fig. 1.

Fig. 1: 6-inch SiC wafer (Image: STMicroelectronics)

1. What are the main properties of SiC?

The combination of silicon with carbon provides this material with excellent mechanical, chemical and thermal properties, including:

• High thermal conductivity
• Low thermal expansion and excellent thermal shock resistance
• Low power and switching losses
• High energy efficiency
• High operating frequency and temperature (operating up to 200°C junction)
• Small die size (with the same breakdown voltage)
• Intrinsic body diode (MOSFET device)
• Excellent thermal management which reduces cooling requirements
• Long lifetime

2. Which are the applications of SiC in electronics?

Silicon carbide is a semiconductor that is perfectly suited to power applications, thanks above all to its ability to withstand high voltages, up to ten times higher than those usable with silicon. Semiconductors based on silicon carbide offer higher thermal conductivity, higher electron mobility, and lower power losses. SiC diodes and transistors can also operate at higher frequencies and temperatures without compromising reliability. The main applications of SiC devices, such as Schottky diodes and FET/MOSFET transistors, include converters, inverters, power supplies, battery chargers and motor control systems.

3. Why SiC overcomes Si in power applications?

Despite being the most widely used semiconductor in electronics, silicon is beginning to show some limitations, especially in high-power applications. A relevant factor in these applications is the bandgap, or energy gap, offered by the semiconductor. When the bandgap is high, the electronics it uses can be smaller, run faster, and more reliably. It can also operate at higher temperatures, voltages, and frequencies than other semiconductors. While silicon has a bandgap of around 1.12eV, silicon carbide has a nearly three times greater value of around 3.26eV.

4. Why can SiC handle so high voltages?

Power devices, especially MOSFETs, must be able to handle extremely high voltages. Thanks to a dielectric breakdown intensity of the electric field about ten times higher than that of silicon, SiC can reach a very high breakdown voltage, from 600V to a few thousand volts. SiC can use higher doping concentrations than silicon, and the drift layers can be made very thin. The thinner the drift layer, the lower its resistance. In theory, given a high voltage, the resistance of the drift layer per unit area can be reduced to 1/300 of that of silicon.

5. Why SiC can outperform IGBT at high frequencies?

In high-power applications, IGBTs and bipolar transistors have mostly been used in the past, with the aim of reducing the turn-on resistance that occurs at high breakdown voltages. These devices, however, offer significant switching losses, resulting in heat generation issues that limit their use at high frequencies. Using SiC, it is possible to make devices, such as Schottky barrier diodes and MOSFETs, which achieve high voltages, low turn-on resistance and fast operation.

6. Which impurities are used to dope SiC material?

In its pure form, silicon carbide behaves like an electrical insulator. With the controlled addition of impurities or dopants, SiC can behave like a semiconductor. A P-type semiconductor can be obtained by doping it with aluminum, boron, or gallium, while impurities of nitrogen and phosphorus give rise to a N-type semiconductor. Silicon carbide has the ability to conduct electricity under some conditions but not in others, based on factors such as the voltage or intensity of infrared radiation, visible light, and ultraviolet rays. Unlike other materials, silicon carbide is capable of controlling the P-type and N-type regions required for device fabrication over wide ranges. For these reasons, SiC is a material suitable for power devices and able to overcome the limitations offered by silicon.

7. How can SiC achieve better thermal management than silicon?

Another important parameter is the thermal conductivity, which is an index of how the semiconductor is able to dissipate the heat it generates. If a semiconductor is not able to dissipate heat effectively, a limitation is introduced on the maximum operating voltage and temperature that the device can withstand. This is another area where silicon carbide outperforms silicon: the thermal conductivity of silicon carbide is 1490 W/m-K, compared to the 150 W/m-K offered by silicon.

8. How is SiC reverse recover time compared to Si-MOSFET?

SiC MOSFETs, like their silicon counterparts, have an internal body diode. One of the main limitations offered by the body diode is the undesired reverse recovery behavior, which occurs when the diode switches off while carrying a positive forward current. The reverse recovery time (trr) thus becomes an important index to define the characteristics of a MOSFET. Fig. 2 shows a comparison between the trr of a 1000V Si-based MOSFET and a SiC-based MOSFET. As can be seen, the body diode of the SiC MOSFET is extremely fast: the values of trr and Irr are so small as to be negligible, and the energy loss Err is considerably reduced.

Fig. 2: Reverse recovery time comparison (Image: Rohm)

9. Why is soft turnoff important for short circuit protection?

Another important parameter for a SiC MOSFET is the short circuit withstand time (SCWT). Since SiC MOSFETs occupy a very small area of the chip and have a high current density, their ability to withstand short circuits that can cause thermal breaks tends to be lower than that of silicon-based devices. In the case, for example, of a 1.2kV MOSFET with TO247 package, the short-circuit withstand time at Vdd=700V and Vgs=18V is about 8-10 μs. As Vgs decreases, the saturation current decreases and the withstand time increases. As Vdd decreases, less heat is generated and the withstand time is longer. Since the time required to turn off a SiC MOSFET is extremely short, when the turnoff rate Vgs is high, a high dI/dt can cause severe voltage spikes. A soft turnoff should therefore be used to gradually lower the gate voltage, avoiding overvoltage peaks.

10. Why is isolated gate driver a better choice?

Many electronic devices are both low and high voltage circuits, interconnected to each other to perform control and power functions. A traction inverter, for example, typically includes a low voltage primary side (power, communication, and control circuits) and a secondary side (high voltage circuits, motor, power stage and auxiliary circuits). The controller located on the primary side normally uses feedback signals from the high voltage side and is susceptible to possible damage if no isolation barrier is present. An isolation barrier electrically isolates the circuits from the primary to the secondary side forming separate ground references, implementing the so-called galvanic isolation. This prevents unwanted AC or DC signals from being transferred from one side to the other, resulting in damage to the power components.

The article originally published at sister publication Power Electronics News.