SiC power device advantages enhance power conversion systems

Compared to silicon, SiC has ten times the dielectric breakdown field strength, three times the bandgap and three times the thermal conductivity.

BY TAKU HAMAGUCHI,ROHM Semiconductor, Santa Clara, CA.

Three major factors are influencing the evolution and implementation of next-generation power semiconductor devices: regulatory requirements for ongoing improvements to efficiency in power conversion systems; market demands for lighter, smaller, more cost-effective systems with more integrated features and emerging new applica- tions such as electric vehicles (EVs) and solid state transformers (SSTs). Up until recently, silicon has been the primary material used in power electronics, and although silicon technology continues to improve, it does have certain limitations that must be taken into consideration when designing for the growing list of essential power system requirements.

Device manufacturers have proven during the last ten years that wide bandgap (WBG) materials such as silicon carbide (SiC) and gallium nitride (GaN) provide multiple advantages in the development of next-generation power semiconductor devices. WBG-based devices offer dramatic improvements in performance, operating temperature, power handling efficiency and the ability to deliver new capabilities, which are not possible with silicon-based devices. For this reason, WBG power devices are now considered the future of power semiconductor devices.

The growing popularity of SiC devices can be traced to the availability of all components needed to build complete power systems, namely SiC diodes, switches and modules. This increased availability is the result of an expanded supply chain with a growing number of suppliers that can offer more economically viable pricing. GaN power devices have been available commercially for a much shorter period of time. Because of this varying state of maturity, SiC and GaN devices have evolved to support separate but complementary functions in different market segments.

SiC is being used for power systems because it has been proven to be more efficient than silicon. The primary advantage of SiC MOSFETs is their very low switching losses, which increase efficiency and enable higher frequency operation.

Because the wide bandgap discussion can be a lengthy topic, this article will focus primarily on the advantages of SiC technology in power conversion systems.

FIGURE 1. Comparison of specific on-resistance of Si- MOSFET and SiC-MOSFET.

FIGURE 1. Comparison of specific on-resistance of Si- MOSFET and SiC-MOSFET.

SiC material advantages

The wide bandgap material properties shown in TABLE 1 explain why SiC-based power devices can outperform silicon. SiC’s breakdown field strength is ten times higher than that of silicon, plus SiC devices can be constructed to withstand the same breakdown with a much smaller drift region. In theory, SiC can reduce the resistance per unit area of the drift layer to 1/300 compared to silicon at the same silicon breakdown voltage.

TABLE 1. Physical characteristics of major wide bandgap materials.

TABLE 1. Physical characteristics of major wide bandgap materials.

Compared to silicon, SiC has ten times the dielectric breakdown field strength, three times the bandgap and three times the thermal conduc- tivity. Both p-type and n-type regions, which are necessary to fashion device structures in semicon- ductor materials, can be formed in SiC. These devices can be produced with a much thinner drift layer and have very high breakdown voltage (600V and up), but provide very low resistance relative to silicon devices. Resistance of high-voltage devices is predominantly determined by the width of the drift region. Compared to silicon, the resistance per unit area of the drift layer can be reduced up to 1/300 at the same breakdown voltage with SiC materials. These properties make SiC an optimal power device material that can far exceed the performance of their silicon counterparts.

The first commercial SiC Schottky Barrier Diodes (SBDs) were introduced more than ten years ago and have been designed into many power systems, most notably into power factor correction (PFC) circuits of switch mode power supplies. Technology maturity, performance and dramatic cost reduction due to increasing volume and competition are the main reasons SiC MOSFETs have been adopted in more and more applications. SiC SBDs are currently available with breakdown voltage ratings of 600V-1700V and 1A-60A current ratings. Thus, SiC devices tend to compete with silicon MOSFETs in the 600V-900V range and with IGBTs in the 1kV+ range.

SiC MOSFETs are now experiencing greater demand with power designers for its normally-off operation and voltage controlled device advantages. Plus, SiC MOSFETs offer gate drive simplicity versus that of junction gate field-effect transistors (JFETs) and bipolar junction transistors (BJTs).

High temperature advantages

The high temperature capabilities of SiC power devices have not been fully exploited because of limitations in packaging technology and the associated lower operating temperatures of other components in systems.

Currently available products are rated only at 150 ̊C to 175 ̊C, and SiC power modules that use special die bonding technology can operate at 250 ̊C. R&D tests on SiC have shown operation up to 650 ̊C is possible, whereas the upper limit of silicon semiconductors is 300 ̊C.

Additionally, SiC’s thermal conductivity is three times higher than that of silicon. These properties contribute to lower cooling needs, making it simpler to cool SiC components. This results in supporting thermal systems that can be smaller, lighter and lower cost.

Enablers of improved power switches

An ideal power switch is able to carry large current with zero voltage drop in the on-state, blocks high voltage with zero leakage in the off-state and incurs zero energy loss when switching from off- to on-state and vice versa. In silicon-based devices, it is difficult to combine these desirable but diametrically opposed characteristics, especially at high voltage and current. To address this problem, many designs have employed Insulated Gate Bipolar Transistor (IGBT) devices. With IGBTs, low resistance at high breakdown voltage is achieved at the cost of switching performance using minority carriers injected into the drift region to reduce conduction (on-) resistance. Therefore, when the transistor is turned off, it takes time for these carriers to recombine and “dissipate” from the base region, thus increasing switching loss and time.

Contrary to IGBTs, MOSFETs are majority carrier devices so they have no “tail” current. SiC MOSFETs, therefore, can deliver all three requirements of power switch — high breakdown voltage, low on-resistance and fast switching speed (FIGURE 1). For example, compared with silicon IGBTs and fast recovery diodes (FRDs), ROHM combines a SiC MOSFET and SiC SBD in one package, which provides 88 percent lower turn-off loss and 34 percent lower turn-on loss enabling switching frequency in hundreds of kHz range. The improvement in turn-off is due to absence of tail current in the MOSFET. The improvement in turn-on is due to the much lower recovery loss of the SiC diode.

Power systems designs can gain significant benefits through low switching losses:
• Less heat generated translates into simpler, cheaper, smaller, and/or lighter cooling systems and ultimately higher power density.
• Allows switching frequency to increase to reduce sizes of passive components (capacitors, inductors), reducing system cost, size, and weight.
• Enables lower operating temperatures so compo- nents do not have be derated as much, allowing smaller, less expensive components to be used. At the system level, this means a lower-rated SiC system can replace higher-rated silicon system.

The tests shown in FIGURE 2 and 3 were conducted at Vdd = 400V, Icc = 20A, and 25 ̊C, and diode recovery losses are included.

FIGURE 2. Shows 88% reduction of turn-off loss: Sic-MOSFET + SiC SBD v. Si IGBT + FRD

FIGURE 2. Shows 88% reduction of turn-off loss: Sic-MOSFET + SiC SBD v. Si IGBT + FRD

FIGURE 3. 4% reduction of turn-on loss: Sic-MOSFET + SiC SBD v. Si IGBT + FRD.

FIGURE 3. 4% reduction of turn-on loss: Sic-MOSFET + SiC SBD v. Si IGBT + FRD.

 

FIGURE 4 shows that at 20 kHz switching frequency, a 100-A SiC half bridge module that is forced-air cooled can replace a 200-A IGBT module that is water cooled.

FIGURE 4. Lower switching losses allow 100a SiC module to replace 200a IGBT module.

FIGURE 4. Lower switching losses allow 100a SiC module to replace 200a IGBT module.

SiC MOSFET reliability

Reliability is a one of the most important considerations in power electronics design. Therefore, one of the first questions from power system engineers is: “Is SiC as reliable as silicon?” The three most important aspects related to overall reliability are gate oxide reliability, stability of gate threshold voltage Vt, and the robustness of the body diode with reverse conduction.

Electrical overstressing of the gate oxide is a common failure mode of MOS devices. Gate oxide quality, consequently, directly affects SiC MOSFET’s reliability. The good news is that manufacturers have solved the problem of developing high-quality oxide on SiC substrates to minimize defect density (interface and bulk traps) without compromising device life or electrical characteristics stability.

A standard test that measures the quality of gate oxide MOS is the Constant Current Stress Time- Dependent Dielectric Breakdown (CCS TDDB), shown in FIGURE 5. The accumulated charge QBD is a quality indicator of the gate oxide layer. The value of 15-20°C/ cm2 is equivalent to that of silicon MOSFETs.

FIGURE 5. Constant current - time dependent dielectric Breakdown measurements

FIGURE 5. Constant current – time dependent dielectric Breakdown measurements

FIGURE 6 shows that when a positive voltage is applied to the gate for an extended period of time, crystal defects at the oxide-SiC interface trap electrons and cause Vth to increase.

FIGURE 6. Vth increases due to extended application of positive gate voltage.

FIGURE 6. Vth increases due to extended application of positive gate voltage.

In FIGURE 7, when a negative voltage is applied, trapped holes cause Vth to decrease — the shift in Vth is 0.3V or less. These tests are performed on the ROHM Semiconductor SCT2080KE SiC MOSFET.

FIGURE 7. Vth decreases due to extended application of negative gate voltage.

FIGURE 7. Vth decreases due to extended application of negative gate voltage.

The results are comparable to that of a silicon MOSFET. However, the shift would be much smaller in practical usage since MOSFETs are alternately switched on and off. This allows trapped electrons and holes to “escape” between switching cycle. Thus, the accumulated trapped carriers, which cause shift in Vth, are much less.

A new era of power conversion systems

Even though there have been many significant technology advances in the last decade, and the supply chain continues to expand, the wide bandgap technology industry for SiC devices has a long way to go to reach its full potential. Making great strides is the next generation of SiC power devices, which are well-positioned to enable new era of high-volume power conversion applications such as EVs and solid state transformers. SiC can also be a positive catalyst for future technology development that enhances application capabilities while continuing to stimulate market demand.

POST A COMMENT

Easily post a comment below using your Linkedin, Twitter, Google or Facebook account. Comments won't automatically be posted to your social media accounts unless you select to share.