I. Preface
With the development of electric vehicles, automotive power device chips are also looking for components that can effectively handle higher operating voltages and temperatures. At this point, silicon carbide MOSFETs are the technology of choice for EV building blocks such as traction inverters. Silicon carbide-based inverters enable electrical systems up to 800V to significantly extend EV range and halve charging time. According to industry research firm IHS Markit, up to 45% of global vehicle production will be electrified by 2025, with around 46 million electric vehicles sold each year. It is estimated that by 2030, these numbers will rise to 57%, with around 62 million electric vehicles sold annually. Power devices are evolving from silicon-based IGBTs to silicon carbide MOSFETs.
The material properties of silicon carbide have been significantly improved compared with silicon: the critical breakdown field strength of silicon carbide material is nearly 10 times that of silicon, the bulk mobility is close to that of silicon, the band gap width is 3 times that of silicon, the electron saturation drift speed is 2 times that of silicon, and the thermal conductivity is also 3 times that of silicon. Compared with silicon devices of the same voltage range, the thickness of silicon carbide devices is about 1/10 of that of silicon devices, and theoretically the on-state voltage drop can be greatly reduced, and the advantages in switching rate and switching loss are more obvious.
△ Figure 1 Comparison of the characteristics of SiC, GaN and Si materials
So far, SiC MOSFETs have not shown the expected advantage over silicon-based IGBTs in terms of on-state voltage drop. There is still plenty of room for improvement in the performance of silicon carbide MOSFETs in terms of substrate and epitaxial layer material mobility, and SiC/SiO2 interface surface mobility. Figure 2 shows the conduction losses of silicon carbide cool MOS compared to silicon-based IGBTs.
△ Figure 2 Comparison of SiC MOSFET and Si IGBT characteristics
2. Optimization of the specific on-resistance of silicon carbide MOSFET devices
Reducing the specific on-resistance of silicon carbide affects the specific on-resistance due to the following factors:
2.1 Mobility of silicon carbide MOSFETs
Depending on the position in the MOSFET, it can be divided into channel mobility and bulk mobility: when the MOSFET is on, the mobility in the channel below the gate is called the channel mobility, and in the area away from the gate and the surface of the material, it is called the bulk mobility.
Channel mobility can be obtained by measuring the output characteristics of the MOSFET: here it can be divided into (1) effective mobility, (2) field-effect mobility, and 3) saturation mobility.
The channel mobility is limited by a number of defects at the SiC/SiO2 interface, which results in the device's field-effect mobility being two orders of magnitude lower than its Hall mobility. It is generally believed that the selective oxidation of SiC leads to carbon precipitation and the formation of carbon clusters (C-clusters) in SiO2 equilibrium) confines the high carbon chemical potential, which leads to a low formation energy of the interfacial C-cluster defects, which explains the high interfacial density of states, and the defect energy level is located close to the bottom of the conduction band of SiC, thus reducing the mobility of the carriers.
The main scattering mechanisms of SiC MOSFET inverted channel electrons include: Coulomb scattering of interfacial charges, bulk lattice scattering, Coulomb scattering of ionized impurities, surface rough scattering, and surface phonon scattering. The following diagram shows the electron scattering at the SiC MOSFET channel:
△ Fig.3 Influencing factors of electron mobility at the channel
At the 4H-SiC interface, the performance of the device is mainly affected by the class acceptor defects, and the coulombic mobility is determined by the scattering of channel electrons by the interface trap. In terms of reducing the specific on-resistance, the resistance of the JFET, the bulk resistance of the MOSFET, and the resistance of the substrate with a high concentration of the planar SiC MOSFET provide a considerable contribution to the total specific on-resistance.
2.2. The device structure design is optimized to reduce the specific on-resistance
Where channel resistance:
thereinto
Accumulation Resistance:
JFET Area Resistance:
The drift zone resistance is:
Taking a SiC MOSFET with a pitch of 5.0μm as an example, the thickness of the gate oxide layer is about 450A, Vth~3.0V, LCH=0.5μm, change the gate width to adjust the value of a, under the conditions of a=1μm, 1.5μm, 2μm, with the average concentration of JFETs, the thickness of the drift zone is about 10μm:
△ Table 1 Composition of specific on-resistance when substrate 1 is used
△ Table 2 Composition of specific on-resistance when substrate 2 is used
△ Table 3 Effect of drift zone concentration on Ronsp and Vth
△ Fig.4 SiC MOSFET current expansion diagram and doping distribution in JFET region
When the thickness of the epitaxial layer and the doping concentration are constant, reducing the resistance in the JFET region and optimizing the concentration in the current expansion region contribute greatly to reducing the total specific on-resistance, and the RDsp has a high correlation with the substrate state of the device, so it is necessary to achieve a compromise design between the epitaxial doping concentration and thickness. Reduced substrate thickness: When the thinning thickness is reduced from 175 μm to 110 μm, the on-resistance is expected to be reduced by about 3 mohm for the common 1200V/80 mohm SiC Planar MOSFET, which translates to about 0.13 mohm.cm2 reduction in the specific on-resistance.
2.3. Trench MOSFET取代Planar MOSFET降低比导通电阻
At present, SiC MOSFETs are also ushering in the trend of developing from planar MOSFET to trench MOS, and the resistance in the JFET region of the planar MOSFET accounts for a large proportion of the total on-resistance, while the JFET region in the trench MOS theoretically does not exist.
△ 图5 ROHM Trench MOSFET结构
△ Figure 6 Infineon's trench MOSFET structure
The double-trench structure represented by ROHM and the half-pack structure represented by infineon represent two structures that have been developed independently of SiC trench MOSFETs. The figure below shows the performance of ROHM's Ron as it moves from a flat structure to a trench.
△ Fig.7 Comparison of ROHM plane and trench MOSFET structures, i.e., comparison of specific on-resistance
3. Performance comparison of mainstream products
The following figure shows the performance comparison of MOSFETs of major manufacturers on the market:
△ Figure 8 Comparison of the specific on-resistance of SiC MOSFETs by major manufacturers
IV. Conclusions
Silicon carbide MOSFETs have various advantages such as low loss, high blocking, high temperature operation, and fast switching speed, but in terms of technology: although SiC single crystal materials have made great progress in recent years in reducing and eliminating the fatal defects that lead to the degradation of the performance and reliability of SiC power semiconductors, the impact of other defects such as dislocation defects on the characteristics of components has not been solved. At present, there are two major technical difficulties in silicon carbide MOSFET devices: low inverted layer channel mobility and gate oxygen reliability at high temperature and high electric field. After silicon carbide MOSFETs overcome the above problems, they will usher in explosive growth.
END
Source: Power Semiconductor Ecosystem
*Disclaimer: This article was originally written by the author. The content of the article is the author's personal point of view, and the wide bandgap semiconductor technology innovation alliance is reprinted only to convey a different point of view, and does not mean that the alliance agrees with or supports the view, if you have any objections, please contact us.