Abstract Analysis of the short-circuit characteristics of SiC metal-oxide-semiconductor field-effect transistors (MOSFETs) is very important for their practical application. This paper studies the SiC MOSFET short-circuit characteristics with an improved test setup under different conditions. A high-current Si insulated gate bipolar transistor is used as a circuit breaker in the test circuit rather than the usual short-circuit test conducted without a circuit breaker. The test platform with a circuit breaker does not influence the calculation results regarding the short-circuit withstand time and energy, but the SiC MOSFET will switch off after failure in a very short time. In addition, the degree of failure will be limited and confined to a small area, such that the damage to the chip will be clearly observable, which is significant for short-circuit failure analysis.

  • Using an improved test setup, the short-circuit characteristics of SiC MOSFETs manufactured by BASiC Semiconductor Ltd. are tested under different voltages.

  • Si IGBTs are used in the test circuit in series with the SiC MOSFET to limit the damage to the MOSFET after failure to a small area.

  • The short-circuit failure waveforms of different test platforms are compared in detail.

Silicon carbide (SiC) power devices are being widely applied in high-voltage, high-frequency applications (e.g., motor driver, solar inverter, switch mode power supply), which increases the necessity of their safe and reliable performance.1,2 Compared with silicon (Si) insulated gate bipolar transistors (IGBTs), SiC metal-oxide-semiconductor field-effect transistors (MOSFETs) have a smaller chip size and higher current density.3,4 Thus, SiC MOSFETs have a shorter short-circuit withstand time than Si IGBTs, which requires a shorter response time to ensure safety.

In recent years, many investigations have been conducted on SiC MOSFETs, including theoretical, modeling, and experimental studies. A physically based model of the SiC MOSFET has been used to simulate a short-circuit safe operating area. The electro-thermal effect was also applied in a junction temperature simulation. Good agreement was observed for drain current waveforms and failure times under varying conditions.5 In addition, the reliability of the short-circuit characteristics of commercial SiC MOSFETs has been tested.6 During the short-circuit test, the measured gate leakage current and threshold voltage values indicated the occurrence of damage to the gate oxide. However, overall, research has focused mainly on waveforms and few ultimate short-circuit tests have been reported.

In this paper, we present the results of a series of short-circuit tests under different conditions and SiC MOSFETs. Unlike the SiC MOSFET, the Si IGBT has a more mature fabrication process and a longer short-circuit withstand time with excellent reliability. During application, IGBTs in a series with higher gate resistance turn off slowly, which can decrease the effects of the IGBT current tail. If more failure information from the chip is desired, we suggest the use of Si IGBTs as a circuit breaker in a series with the device under test (DUT) in the test circuit. By doing so, the damage sustained by the failed devices can be restricted to a small area for observation by optical microscopy or hotspot analysis. The resulting samples are thus more meaningful for further research regarding the failure mechanism at the gate area.7,8

In this work, two kinds of typical devices packaged in TO-247-3 from BASiC Semiconductor Ltd. are measured in a short-circuit test. Devices with different on-resistance (RDS(on)) values and current ratings because of their different chip sizes are chosen for comparison with the typical short-circuit characteristics under the same conditions. Table I lists the key parameters of the devices and Figs. 1(a) and 1(b) show the output characteristics of the DUTs at different gate voltages (VGS). In addition, Si IGBTs manufactured by Infineon, which have a higher current rating and longer short-circuit withstand time, are applied in the circuit as a circuit breaker, the output characteristics of which are shown in Fig. 1(c). Figure 2 shows the threshold voltages at specified drain-source current (IDS) of MOSFET and collector-emitter current (ICE) of IGBT.

Table 1.

Key parameters of devices in the short-circuit test.

Devices Device type RDS(on) (mΩ) Threshold voltage (V) Voltage rating (V) Current rating (A) Drive voltage (V)
SiC MOSFET  160  2.9
@IDS = 2.5 mA 
1200  10  20 
SiC MOSFET  80  3
@IDS = 5 mA 
1200  20  20 
Si IGBT  5.8
@ICE = 1.5 mA 
1200  40  15 
Devices Device type RDS(on) (mΩ) Threshold voltage (V) Voltage rating (V) Current rating (A) Drive voltage (V)
SiC MOSFET  160  2.9
@IDS = 2.5 mA 
1200  10  20 
SiC MOSFET  80  3
@IDS = 5 mA 
1200  20  20 
Si IGBT  5.8
@ICE = 1.5 mA 
1200  40  15 
Fig. 1.

Output characteristics of (a) Device A, (b) Device B, and (c) Device C at room temperature.

Fig. 1.

Output characteristics of (a) Device A, (b) Device B, and (c) Device C at room temperature.

Close modal
Fig. 2.

Threshold voltages of Device A, Device B, and Device C at room temperature.

Fig. 2.

Threshold voltages of Device A, Device B, and Device C at room temperature.

Close modal

The short-circuit characteristics of discrete SiC MOSFETs are evaluated on the platform shown in Fig. 3. The gate voltage of the DUTs and IGBTs are controlled by the gate driver controllers 1EDI20H12AH from Infineon and SID1182K from Power Integrations, respectively. The short-circuit waveforms are measured at various drain voltages (VDS) and short-circuit pulse widths but the same gate voltage (VGS) and gate resistance (RG). High-voltage differential probes are used to test the drain voltage and high-frequency current probes are used to measure the short-circuit current.

Fig. 3.

Schematic (a) and photo (b) of the test bench used in the short-circuit test.

Fig. 3.

Schematic (a) and photo (b) of the test bench used in the short-circuit test.

Close modal

In this short-circuit test, we used Si IGBTs with better short-circuit current capability, which can prevent uncontrollable surges in the current. Figure 4(a) shows the normal short-circuit test waveforms of the SiC MOSFET in Device A.9Figure 4(b) shows the waveforms of the short-circuit test with a circuit breaker for the DUT and IGBTs in Device B. This test can be divided into five periods.

Fig. 4.

Short-circuit waveforms after failure of the DUTs. (a) Short-circuit test results without IGBTs as circuit breaker; (b) short-circuit test results with IGBTs as circuit breaker.

Fig. 4.

Short-circuit waveforms after failure of the DUTs. (a) Short-circuit test results without IGBTs as circuit breaker; (b) short-circuit test results with IGBTs as circuit breaker.

Close modal

Period A: Both the DUT and IGBTs are turned off, with the DC bus voltage on at 800 V. Due to the different drift-layer resistances of the devices, the drain voltage of the DUT (VDS.MOSFET) is smaller than the DC bus voltage.

Period B: The turn-on signal for IGBTs is 1 μs earlier than that of the DUT. When the IGBTs are switched on, the collector voltage of the IGBT (VCE.IGBT) falls to minimum and the VDS.MOSFET increases to the DC bus voltage under the same starting conditions of the typical test shown in Fig. 4(a).

Period C: The DUT is switched on and maintains the short-circuit status. The drain current of the DUT (IDS) increases to its peak value as it is in a high-voltage saturation area, then decreases later due to the increase in the junction temperature. Both tests show the same process of change in the current for the different device types with different short-circuit test platforms. Thus, the calculation of the short-circuit energy is not affected by the novel test platform.

Period D: When the DUT is switched off but fails with increasing current and voltage oscillations, the increasing VCE.IGBT is detected and the gate driver controller turns the IGBTs off within a short period of time. The increasing power of the DUT can be maintained within the required limits quickly to prevent continuous destruction. However, in the test circuit without a circuit breaker, the current soars immediately, which is the scenario that occurs prior to explosion of the device by thermal runaway.

Period E: Lastly, the DUT turns off safely as the VDS.MOSFET and VGS.MOSFET have been limited to their minimum values. The IGBT is turned on for a short time by the gate driver controller. However, this is a much shorter time than the short circuit withstand time of Si IGBT. There is no dangerous for the circuit breaker. And the whole test is accomplished safely with complete test waveforms obtained.

The comparison of the proposed test platform with the common short-circuit test reveals that use of the Si IGBT as a circuit breaker does not influence the test conditions or the short-circuit test waveforms. The SiC MOSFET after short-circuit failure maintains its usefulness for further research by the protection of the circuit breaker. Therefore, we suggest the use of Si IGBTs in the circuit as a protection measure for both the DUTs and the circuit.10,11

Figure 5 shows the short-circuit current waveforms for Devices A and B under different VDS values with the improved test circuit. Both devices are tested at VGS = 20 V and RG = 2.2 Ω. With increases in the VDS, the saturation current increases higher and faster. The peak currents are 120 A and 220 A, which are 12 and 11 times the normal continuous current ratings, respectively. The short-circuit protection system is a critical factor when working with higher voltages. Device B with the higher current is used in the next experiment.

Fig. 5.

Short-circuit current waveforms under different VDS of (a) Device A and (b) Device B.

Fig. 5.

Short-circuit current waveforms under different VDS of (a) Device A and (b) Device B.

Close modal

Table II shows the short-circuit characteristics of three B-type devices, which are measured on the novel test bench at various DC bus voltages, and Fig. 6 presents the current waveforms. With increases in the DC bus voltage, both the short-circuit withstand time and critical energy decrease when the junction temperature rapidly increases.

Table 2.

Summary of the short-circuit characteristics of B-type devices.

Devices Drain voltage (V) Short-circuit time (μs) Short-circuit energy (J) Damage degree
B1  400  21  1.4  Hotspot 
B2  600  Hotspot 
B3  800  0.8  Burned 
Devices Drain voltage (V) Short-circuit time (μs) Short-circuit energy (J) Damage degree
B1  400  21  1.4  Hotspot 
B2  600  Hotspot 
B3  800  0.8  Burned 
Fig. 6.

Short-circuit current waveforms of Devices B1, B2, and B3.

Fig. 6.

Short-circuit current waveforms of Devices B1, B2, and B3.

Close modal

Optical microscopy and hotspot analysis of the devices were performed next. Regarding the results obtained for Devices B1 and B2 in an ultimate test with VDS = 400 V and VDS = 600 V, respectively, the breakdown voltages are observed to decrease to 200 V and the hotspot test indicates the presence of damage, as shown in Figs. 7(a) and 7(b).12 For the ultimate test of Device B3, the VDS increased to 800 V and the chip burned, as shown in Fig. 7(c), which reveals a more catastrophic failure than those of Devices B1 and B2.

Fig. 7.

Optical microscopic images from the hotspot tests of (a) Device B1, (b) Device B2, and (c) Device B3 after the ultimate test.

Fig. 7.

Optical microscopic images from the hotspot tests of (a) Device B1, (b) Device B2, and (c) Device B3 after the ultimate test.

Close modal

Based on the results of the above experiments, the DUTs are determined to fail because of the high temperature generated by the high short-circuit current and high voltage. The metal melted at temperatures higher than 933 K, but the degree of failure was limited to a low grade. In the short-circuit test under lower voltage, the degree of damage was limited to a single hotspot. Thus, further investigation of the failure mechanism of the gate oxide area can be made using the single hotspot sample.13 ,14

In this paper, we presented the results of short-circuit tests conducted under different conditions and the failure performances of devices manufactured by BASiC Semiconductor Ltd. The short-circuit peak current of SiC MOSFETs was determined to be approximately 12 times the nominal continuous current rating and the short-circuit withstand time is longer than 4 μs. As the VDS is gradually increased, the short-circuit withstand time and critical energy decrease due to the rapid increase in power. The damage regions of the failed devices tested with a circuit breaker are restricted to a single hotspot or a small area. Therefore, when further experimental analysis is necessary regarding the short-circuit capacity and device failure, the use of Si IGBTs is suggested as a circuit breaker.

The authors declare no conflict of interest.

This work was supported by the Shenzhen Science and Technology Program [Grant No. KQTD2017033016491218].

1.
Kreutzer
O
,
Heckel
T
,
Maerz
M
.
Using SiC MOSFET’s full potential—Switching faster than 200 kV/μs
.
Mater Sci Forum
2016
;
858
:
880
884
.
2.
Dimitrijev
S
,
Han
J
,
Moghadam
HA
, et al.
Power-switching applications beyond silicon: Status and future prospects of SiC and GaN devices
.
MRS Bull
2015
;
40
(
5
):
399
405
.
3.
Yin
S
,
Gu
Y
,
Deng
S
, et al.
Comparative investigation of surge current capabilities of Si IGBT and SiC MOSFET for pulsed power application
.
IEEE Trans Plasma Sci
2018
;
46
(
8
):
2979
2984
.
4.
Romano
G
,
Fayyaz
A
,
Riccio
M
, et al.
A comprehensive study of short-circuit ruggedness of silicon carbide power MOSFETs
.
IEEE J Emerging Sel Top Power Electron
2016
;
4
(
3
):
978
987
.
5.
Duong
TH
,
Ortiz
JM
,
Berning
DW
, et al.
Electro-thermal simulation of 1200 V 4H-SiC MOSFET short-circuit SOA
. in
2015 IEEE 27th International Symposium on Power Semiconductor Devices & IC’s
(
IEEE
,
2015
).
6.
Maxime
B
,
Remy
O
,
Thibault
C
, et al.
Electrical performances and reliability of commercial SiC MOSFETs at high temperature and in SC conditions
. in
European Conference on Power Electronics & Applications
(
IEEE
,
2015
), p.
1
9
.
7.
Boige
F
,
Tremouilles
D
,
Richardeau
F
.
Physical origin of the gate current surge during short-circuit operation of SiC MOSFET
.
IEEE Electron Device Lett
2019
;
40
(
5
):
666
669
.
8.
Boige
F
,
Richardeau
F
.
Gate leakage-current analysis and modelling of planar and trench power SiC MOSFET devices in extreme short-circuit operation
.
Microelectron Reliab
2017
;
76-77
:
532
538
.
9.
Masaki
N
,
Junjie
A
,
Hiroshi
Y
, et al.
Investigation of short-circuit failure mechanisms of SiC MOSFETs by varying DC bus voltage
.
Jpn J Appl Phys
2018
;
57
(
7
).
10.
Shi
Y
,
Xie
R
,
Wang
L
, et al.
Switching characterization and short-circuit protection of 1200 V SiC MOSFET T-type module in PV inverter application
.
IEEE Trans Ind Electron
2017
;
64
(
11
):
9135
9143
.
11.
Sasagawa
K
,
Abe
Y
,
Matsuse
K
.
Study on gate-controlled protection methods for IGBTs under a short-circuit condition
.
IEEJ Trans Ind Appl
2007
;
127
(
5
):
478
484
.
12.
Liu
J
,
Zhang
G
,
Wang
B
, et al.
Gate failure physics of SiC MOSFETs under short-circuit stress
.
IEEE Electron Device Lett
2019
;
41
(
1
):
103
106
.
13.
An
J
,
Namai
M
,
Iwamuro
N
.
Experimental and theoretical analyses of gate oxide and junction reliability for 4H-SiC MOSFET under short-circuit operation
.
Jpn J Appl Phys
2016
;
55
(
12
):
124102
.

Shen Diao, Research and Development Department, Shenzhen BASiC Semiconductor Ltd., China. His research interests include the dynamic characteristics of SiCMOSFET and the Spice modeling of SiC power devices.

Ziwei Zhou, Research and Development Department, Shenzhen BASiC Semiconductor Ltd., China. Her research interests include the material characteristics of 4H-SiC and the failure analysis of power devices.

Jun Sun, Research and Development Department, Shenzhen BASiC Semiconductor Ltd., China. His research interests include chip design and the fabrication of SiC power devices.

Weiwei He, Research and Development Department, Shenzhen BASiC Semiconductor Ltd., China. His research interests include high-power semiconductor devices such as IGBTs and MOSFETs, and their applications.