Brief Overview of Silicon Carbide

Power/Voltage Range of Semiconductor Devices

Figure 1 shows where each of the Silicon Carbide (SiC), Insulated-Gate Bipolar Transistors (IGBTs) (plain silicon (Si) Metal Oxide Semiconductor Field-effect Transistors (MOSFETs)) and Gallium Nitride (GaN) are with respect to the operating voltage and the frequency range.

• MOSFETS are very good at low voltage and they can operate at very high frequencies at low voltage.
• GaN is similar. GaN has devices in the market that are between 600V and 700V and operate well at high frequencies.
• IGBT devices are very good at high voltage and high power but their construction and operation limit how fast you can switch (turn on and off).
• SiC also operates at high voltages, but the technology allows them to operate at higher switching frequencies. As we saw before, this means smaller systems and higher efficiency.

If you can switch correctly, the benefits of SiC during operation are higher voltage and lower switching losses. For the power stage, we could have smaller passives. This leads to fewer sections or components and then, at higher frequencies and higher efficiency, you reduce the cooling requirements.
This means that for the system, you get an improvement in size, cost and weight.

Wide Bandgap

Refer to the accompanying Figure 2.

• Bandgap is the energy required to move an electron from the valence band to the conduction band
  • In semiconductor physics, the bandgap is the energy difference between the valence band (where electrons are bound to atoms) and the conduction band (where electrons are free to move and conduct electricity). For an electron to jump from the valence band to the conduction band, it must gain energy equal to the bandgap. This concept is fundamental in determining the electrical properties of a material.

• Wider bandgap devices can support a higher electric field
  • Materials with a wider bandgap can withstand higher electric fields before breaking down. This makes them suitable for high-power applications because they can operate at higher voltages and temperatures compared to traditional semiconductors like silicon. Examples of wide bandgap materials include SiC and GaN.

• Lower R_DSon per die area
  • R_DSon refers to the on-resistance of a transistor when it is in the conducting state. Wide bandgap materials typically have lower R_DSon values for a given die area compared to silicon. This means that devices made from wide bandgap materials can be more efficient, as they exhibit lower power losses during operation.

• Graphic explanation
  • The "bandgap" or "forbidden band" is the energy range in a solid, where no electron states can exist. The valence band is below this gap, and the conduction band is above it. In wide bandgap materials, this gap is larger, which is visually represented by a wider region between the valence and conduction bands.

In summary, bandgap materials require more energy to move electrons from the valence band to the conduction band, can handle higher electric fields, and offer lower on-resistance, making them advantageous for high-power and high-efficiency applications.

Back to Top

High Voltage and High Power Comparison (SiC vs Si)

Let's look a bit more into why SiC performs better than Si at higher voltages.  Figure 3 compares Sic to Si for five different characteristics.

• Breakdown field
  • One is the ability to support an electric field. The higher the voltage, the higher the electric field. SiC can support a much higher electric field for a given thickness of the device which means lower ON resistance and higher efficiency.
• Saturation velocity
  • Saturation velocity is the maximum speed at which we can conduct electricity through a device or through a medium, and SiC has a higher saturation velocity than Si, which leads to faster speeds and faster switching.
• Band gap energy
  • Band gap energy is the amount of energy it takes to push an electron from its steady state over to its electric band. 3X higher means that, because it has more energy, you can operate at a higher junction temperature without the higher leakage currents.
• Thermal conductivity
  • Thermal conductivity is higher for SiC and allows heat to be removed from the device more easily. This property contributes to higher power density, relaxed cooling requirements, and the ability to conduct higher currents.
• Positive temperature coefficient
  • While Si and SiC both have a positive temperature coefficient, which means you could parallel these devices, SiC has a lower temperature coefficient. This means you can still parallel devices, but when you operate at a higher temperature, you have lower conduction loss.

Back to Top

Wide Band Gap Device Comparison

When comparing SiC MOSFETs to GaN MOSFETs, several key differences emerge.  Refer to Figure 4.

Silicon Carbide (SiC) MOSFETs

• Thermal conductivity
  • SiC offers excellent thermal conductivity.
• High voltage operation
  • SiC is capable of switching and operating at very high voltages.
• Stable RDSon temperature coefficient
  • SiC has a very stable RDSon temperature coefficient, with only a small increase at maximum operating temperature.
• Avalanche capability
  • SiC can handle overvoltage and absorb energy during fault conditions, enhancing reliability.
• Design flexibility
  • A wide range of vendors, voltage ratings, and package options are available, simplifying the design process.

Gallium Nitride (GaN) MOSFETs

• Electron mobility
  • GaN has superior electron mobility, allowing for faster switching and lower switching losses.
• Voltage limitation
  • Commercially available GaN devices are typically limited to operation below 1000 volts.
• Temperature coefficient
  • GaN has a higher RDSon temperature coefficient, which can be as high as 2:1 or more across the temperature range, compared to a 25-40 percent increase for SiC.
• Lack of avalanche rating
  • GaN devices do not have a Unclamped Inductive Switching (UIS) or avalanche rating, meaning they can be permanently damaged if the voltage exceeds the breakdown threshold.

In summary, while GaN MOSFETs offer faster switching and lower switching losses, SiC MOSFETs provide better reliability, higher voltage operation, and more robust performance, making them a preferred choice in many applications.

Back to Top

SiC Performs Better at High Voltage and High Power

SiC MOSFETs perform better at high voltage for several reasons, particularly related to ON resistance. Refer to Figure 5.

• Depletion region (drift layer)
  • In a Si planar MOSFET, the depletion region or drift layer is responsible for most of the voltage blocking. Si cannot support as high an electric field as SiC, necessitating a much thicker drift layer in Si devices. Specifically, the drift layer in Si needs to be about 10 times thicker than that in SiC for the same blocking voltage.
• Higher doping concentration
  • SiC can support a higher doping concentration compared to Si. This higher doping concentration results in lower resistance within the device, contributing to better performance at high voltages.

In summary, the ability of SiC to support a higher electric field and higher doping concentration leads to a thinner drift layer and lower ON resistance, making SiC MOSFETs more efficient and effective in high-voltage applications.

Back to Top

Conduction Loss

When comparing SiC MOSFETs to Si IGBTs in terms of conduction loss, a couple key points emerge as shown in Figure 6.

• Conduction loss mechanism
  • SiC MOSFETs behave essentially as resistive devices when turned on, resulting in a fairly linear relationship between current and voltage.
  • IGBTs have a different conduction mechanism, characterized by a lower bulk resistance.
• Performance at different current levels
  • At lower currents, SiC MOSFETs exhibit much lower conducting voltage (or ON voltage) compared to IGBTs, due to their linear resistance behavior.
  • At higher currents, IGBTs have a lower voltage drop than SiC MOSFETs, owing to their lower bulk resistance.

In summary, SiC MOSFETs offer lower conduction losses at lower currents, while IGBTs perform better at higher currents due to their lower bulk resistance.

Back to Top

Conduction Loss vs Temperature

SiC generally exhibits the smallest increase in voltage across the device while it is ON, compared to other technologies. Figure 7 presents a graph normalized to 25°C. As the device's temperature rises during operation, potentially reaching 100°C, 125°C, or higher for junction temperatures, the performance characteristics change.

For instance, Si FETs, particularly high-voltage ones, experience a significant increase in ON resistance at elevated temperatures. Similarly, IGBTs show an increase in their Vce voltage under higher temperature conditions. In contrast, SiC maintains the lowest voltage increase across the operating temperature range and also benefits from lower conduction losses.

Back to Top

Switching Loss

The other significant contributor to power dissipation in a MOSFET or any switching device is the switching loss, which occurs during the device's turn-on and turn-off phases. In the accompanying diagram shown in Figure 8, we observe the turn-on process; as the device turns on, the current begins to rise while the voltage across the device starts to fall. During this transition period, substantial power is dissipated. The area under the green power curve represents the energy for that specific transition. By multiplying this energy by the switching frequency, you can determine the average power dissipated during turn-on. The same calculation applies to turn-off losses.

During the turn-on phase, in addition to the current flowing through the device, there are typically two other sources of current—and consequently, two other sources of loss. These are the discharge of the output capacitance and the reverse recovery loss, the latter of which will be discussed in more detail later.

Back to Top

Switching Loss Comparison (SiC vs Si)

Let's compare the switching losses for an IGBT and a SiC MOSFET, both sized for the same application. In this scenario, we are considering devices rated at 600 volts and 30 amps, and we will estimate these losses at a junction temperature of 25°C based on the available data sheets.  Refer to the accompanying bar chart in Figure 9.

The first two bars in Figure 9 illustrate the ON and OFF losses for the SiC MOSFET, while the second two bars show the ON and OFF losses for the IGBT. For ON losses, you can observe an improvement with SiC due to its faster switching speed. During turn-off, there is a significant improvement—almost a four-to-one advantage—primarily because SiC devices have a much faster turn-off time compared to IGBTs.

IGBTs exhibit a phenomenon known as current tailing, where they are slow to turn off. As previously noted from the graph, a prolonged turn-off period results in a longer transition time during which power is dissipated, leading to higher losses. This characteristic makes SiC MOSFETs more efficient in terms of switching losses.

Back to Top

Reverse Recovery

Referring to Figure 10, Si diodes with a PN junction exhibit a mechanism called "reverse recovery" when conducting current. In these diodes, minority carriers, electrons in the P-type material and holes in the N-type material, are essential for current flow. When the current stops and the diode needs to block voltage, these minority carriers must be cleared from the junction. The speed at which these carriers are removed determines how quickly the diode can turn off.

In contrast, Schottky diodes, or metal-semiconductor junction diodes, operate differently. These diodes consist of a metal layer in contact with an N-type or P-type semiconductor, forming a junction barrier. In Schottky diodes, only majority carriers are involved in current flow, meaning there are no minority carriers within the junction. As a result, when the device turns off, it does so more quickly. The only reverse current is due to the charge built up in the capacitance between the metal and the semiconductor, which is typically very low. This makes the reverse recovery energy in Schottky diodes significantly lower than that in PN junction diodes.

Back to Top

Datasheet Dynamic Parameters

Let's examine the diagram in Figure 11 to understand the current flow through the diode. Initially, current flows in a positive direction through the diode. When you want to turn the diode off, the current naturally decreases to zero. However, due to the buildup of minority carriers, reversing the voltage across the diode forces these carriers out, resulting in a negative current.

The amplitude and duration of this negative current depend on several factors. Firstly, it depends on the device itself. Secondly, it is influenced by the amount of forward current that was flowing before the diode was turned off; higher forward current results in more minority carriers. Thirdly, the speed at which the device is turned off plays a role. If the turn-off is slower, more minority carriers recombine within the device, reducing the number available to contribute to the reverse current. Temperature also affects this current and the duration of the reverse recovery period.

The area under the curve, from the point where the current goes negative to where it returns to zero, is known as the reverse recovery charge. This charge, when multiplied by the switching frequency, determines the power loss due to reverse recovery. The energy associated with this recovery is equal to the reverse recovery charge multiplied by the voltage across the device.

Back to Top

Diode Reverse Recovery

Let's take a look at an actual example comparing SiC and Si diodes during reverse recovery events. The orange trace shown in Figure 12 represents the SiC diode, while the two silicon traces represent reverse recovery at 25°C and 125°C.

In general, SiC diodes exhibit minimal temperature dependence at slow to medium reverse recovery turn-off speeds. At very high di/dt (rate of change of current), there might be about a 20 percent increase in reverse recovery charge at 175°C compared to 25°C, but for this example, we will ignore that.

Examining the graph, we can calculate the power loss due to reverse recovery. For the silicon diode at 25°C, the peak reverse recovery current is approximately 25 amps, and the reverse recovery time is about 175 nanoseconds. Calculating the area under this curve gives us around 2200 nanocoulombs. With a typical operating frequency of 65 kHz, which is common for power factor correction and other applications, this results in about 14 watts of power dissipation solely due to reverse recovery.

When the operating temperature of the Si device increases to 125°C, the reverse recovery charge and power dissipation increase significantly. This is one of the challenges in operating high-voltage Si MOSFETs at high frequencies, as the reverse recovery losses become substantial.

In contrast, the SiC body diode has a much lower reverse recovery charge (Qrr), leading to significantly lower power dissipation due to reverse recovery. This characteristic allows SiC devices to operate at much higher frequencies with lower losses, thereby increasing efficiency and reducing the size of passive components.

Back to Top