In this paper, we present a novel lateral double-diffused metal-oxide-semiconductor (LDMOS) transistor for high-temperature and high breakdown voltage applications. The key idea in our study is replacing a 4H-SiC layer in a part of the buried oxide region (BOX) to reduce temperature effects. Moreover, the top of the 4H-SiC layer has multiple trenches to increase the breakdown voltage. These multiple trenches have been filled with an N-type silicon material. So, we call the proposed structures as multiple trenches 4H-SiC LDMOS (MTSiC-LDMOS). The proposed device is simulated by a two-dimensional ATLAS simulator, and we have shown that the maximum lattice temperature decreases and the breakdown voltage improves by optimization of multiple trenches in the 4H-SiC region. Also, the results show that the current flow and specific on-resistance have improved. Therefore, the MTSiC-LDMOS structure is more reliable than a conventional LDMOS (C-LDMOS) for high-temperature and high breakdown voltage applications.

Power transistors due to their high performance are essential for intelligent power applications. Today, among different power electronic devices, lateral double-diffused metal-oxide-semiconductor field effect transistors (LDMOSFET) have been widely used in RF applications, monolithic integrated power electronic systems, and smart-power ICs.1–6 On the other hand, silicon carbide (SiC), after silicon (Si) and gallium arsenide (GaAs) materials, is the third-generation semiconductor material. The SiC material has several advantages, such as high thermal conductivity, wide bandgap, high maximum operating temperature, and high breakdown voltage. Also, for high-frequency and high-power applications among its poly-types, 4H-SiC has gained much attraction.7–10 The LDMOS transistors have several advantages over other power devices such as compatibility with advanced VLSI technologies, multiple output capabilities on the same chip, and also a reduction in the number of fabrication steps.11–13 However, the major problems in LDMOS transistors are low breakdown voltage (VBR) and high specific on-resistance ( R O n . s p ). In order to design a suitable LDMOS transistor, we must consider VBR and R O n . s p, but obtaining both low R O n . s p and high VBR is not possible. So, we have a trade-off between the R O n . s p and the VBR. In the last few years, to overcome this effect, various structures have been accomplished.14–17 

Also, silicon-on-insulator (SOI) LDMOS transistors have several important advantages over traditional LDMOSFETs, such as lower parasitic capacitances, reduced latch-up, low leakage current, and high speed, and with other devices, they can be easily integrated.18–25 However, the maximum lattice temperature of the SOI-LDMOS transistor increases due to low thermal conductivity of the buried insulator material. So, the SOI-LDMOS structure has a self-heating effect problem. In this case, to solve this problem, we can use partial SOI technology (PSOI) that has better thermal conductivity, instead of SOI technology.26–29 

In this paper, by using a 4H-SiC layer, a novel LDMOS transistor is presented to improve the self-heating effect (SHE) and the breakdown voltage that are necessary in high-temperature and high-voltage applications. In our work, we use three key ideas: the first is inserting the p-type 4H-SiC layer in the buried oxide (BOX) to reduce the SHE; the second is locating the multiple trenches at the top of the 4H-SiC region to improve the breakdown voltage; and the third is reducing the specific on-resistance by using a Si layer with higher density than its drift region at the bottom of the drift region. So, we called the new structure as multiple trenches 4H-SiC LDMOS (MTSiC-LDMOS). The MTSiC-LDMOS is simulated by a two-dimensional ATLAS simulator.30 We have shown that by optimization of multiple trenches, the maximum lattice temperature decreases and the breakdown voltage improves by flatting the electric field. The proper breakdown voltage and the self-heating effect can be mentioned as the main approaches to the MTSiC-LDMOS structure. Therefore, for high-temperature applications, the MTSiC-LDMOS device is more reliable than a conventional LDMOS (C-LDMOS).

Figure 1 presents the schematic cross-sectional view of the MTSiC-LDMOS structure. It can be seen that a 4H-SiC layer has replaced a part of the oxide layer in the BOX as a layer that benefits from good temperature reduction effects. Moreover, in the new device, the multiple trenches are located at the top of the 4H-SiC layer. As the figure shows, the L and the H are the length and the height of the trenches in the 4H-SiC layer, respectively. The p-type 4H-SiC layer that has better thermal conductivity than silicon dioxide can decrease the maximum lattice temperature in the proposed MTSiC-LDMOS. Also, these multiple trenches have been filled with an N-type silicon material. It is worth noting that the doping density of the N-type silicon is more than the drift region’s doping density. Therefore, the depletion region expands in the drift region and additional peaks are created in the electric field distribution. So, the proposed structure can effectively enhance the electric field in the drift region. So, a significant improvement in the breakdown voltage is obtained in the MTSiC-LDMOS structure.

FIG. 1.

Schematic cross section of the MTSiC-LDMOS structure.

FIG. 1.

Schematic cross section of the MTSiC-LDMOS structure.

Close modal

In Table I, a list of optimum parameters for the MTSiC-LDMOS and C-LDMOS devices that are simulated in our work is presented.14,31 Also, the fundamental parameters of 4H-SiC and silicon32–36 are given and compared in Table II.

TABLE I.

List of optimum parameters for the proposed MTSiC-LDMOS device.

ParametersMTSiC-LDMOSC-LDMOS
Drift region length (μm) 10 10 
Channel length (μm) 
Buried oxide thickness (tBOX) (μm) 0.3 0.3 
Source/drain doping density (cm−31 × 1019 1 × 1019 
Silicon thickness (tSi) (μm) 0.5 0.5 
Drift region doping density (cm−32 × 1016 2 × 1016 
Length of windows (L) 0.5 μNot defined 
Channel doping density (cm−35 × 1017 5 × 1017 
Substrate doping density (cm−31 × 1014 1 × 1014 
Height of windows (H) 0.1 μNot defined 
N-windows doping density 1 × 1017 cm−3 Not defined 
Gate oxide thickness (tox) (nm) 50 50 
Doping of the p-type 4H-SiC layer 1 × 1015 cm−3 Not defined 
Length of 4H-SiC layer (LS8.5 μNot defined 
Distance of 4H-SiC layer from the left side of the device 3 μNot defined 
Distance of 4H-SiC layer from the right side of the device 1.5 μNot defined 
ParametersMTSiC-LDMOSC-LDMOS
Drift region length (μm) 10 10 
Channel length (μm) 
Buried oxide thickness (tBOX) (μm) 0.3 0.3 
Source/drain doping density (cm−31 × 1019 1 × 1019 
Silicon thickness (tSi) (μm) 0.5 0.5 
Drift region doping density (cm−32 × 1016 2 × 1016 
Length of windows (L) 0.5 μNot defined 
Channel doping density (cm−35 × 1017 5 × 1017 
Substrate doping density (cm−31 × 1014 1 × 1014 
Height of windows (H) 0.1 μNot defined 
N-windows doping density 1 × 1017 cm−3 Not defined 
Gate oxide thickness (tox) (nm) 50 50 
Doping of the p-type 4H-SiC layer 1 × 1015 cm−3 Not defined 
Length of 4H-SiC layer (LS8.5 μNot defined 
Distance of 4H-SiC layer from the left side of the device 3 μNot defined 
Distance of 4H-SiC layer from the right side of the device 1.5 μNot defined 
TABLE II.

Comparison of the fundamental parameters of 4H-SiC and silicon.

ParametersSi4H-SiC
ε S 11.8 9.66 
Eg (300 K)(eV) 1.08 3.23 
ni (300 K)(cm−31.45 × 1010 1.73 × 10−8 
V S A T ( cm / s ) 1.03 × 107 2.2 × 107 
E c r i t ( 300 K ) ( V / cm ) 0.25 × 106 2.2 × 106 
μ n m a x ( c m 2 / V S ) 1430 970 
μ p m a x ( c m 2 / V S ) 450 124 
ParametersSi4H-SiC
ε S 11.8 9.66 
Eg (300 K)(eV) 1.08 3.23 
ni (300 K)(cm−31.45 × 1010 1.73 × 10−8 
V S A T ( cm / s ) 1.03 × 107 2.2 × 107 
E c r i t ( 300 K ) ( V / cm ) 0.25 × 106 2.2 × 106 
μ n m a x ( c m 2 / V S ) 1430 970 
μ p m a x ( c m 2 / V S ) 450 124 

The 2D device simulator SILVACO is performed to have a good comparison between the MTSiC-LDMOS and the C-LDMOS transistors. In this simulation, proper models, including the analytic model, field-dependent mobility (FLDMOB) model, CCSMOB model, SHI model, LAT.Temp model, Auger recombination model, impact Selb model, and Shockley–Read–Hall (SRH) recombination, are activated. Also, to achieve more acceptable results, the interface between 4H-SiC, Si, and SiO2 is considered at 3 × 1010 cm−2. It is worth noting that the ATLAS two-dimensional (2D) simulator is calibrated to the experimental data.37 The output characteristics of an SOI-LDMOS at VD = 0.1 V are shown in Fig. 2. It can be concluded from the figure that the simulation results agree well with the experimental data.

FIG. 2.

Output characteristic of an SOI-LDMOS that is extracted from experimental data and compared with simulations data (Ref. 37) at VD = 0.1 V.

FIG. 2.

Output characteristic of an SOI-LDMOS that is extracted from experimental data and compared with simulations data (Ref. 37) at VD = 0.1 V.

Close modal

We know that the applied voltage magnitude is the area underneath the electric field curve. Therefore, at a certain voltage, the heights of common peaks near the drain and source junctions are reduced by producing additional peaks in the electric field. In other words, these additional peaks in the drift region of the structure modulate surface electric field distribution such that the electric field distribution becomes more uniform and the breakdown voltage becomes more enhanced.12 In the proposed structure, these additional peaks, which are created due to the multiple trenches that are located at the bottom of the drift region, lead to a flatter electric field. Figures 3 and 4 present the distribution of the electric field of the MTSiC-LDMOS and C-LDMOS devices in their breakdown voltages along cutline AA′ and cutline BB′, respectively. The maximum peak of the electric field happens near the junction of the drift and drain regions in the C-LDMOS transistor. However, additional peaks are created in the drift region in the MTSiC-LDMOS transistor. These additional peaks lead to an electric field distribution for the MTSiC-LDMOS that is flatter than the C-LDMOS device. As a result, a higher BV for the MTSiC-LDMOS occurs. It is important to note that to have a good comparison, we need to examine the proposed structure and the C-LDMOS device at the same drain voltage (VDS). Therefore, in Fig. 5, the MTSiC-LDMOS and C-LDMOS devices are compared while VDS is equal for both devices.

FIG. 3.

Electric field distribution of the MTSiC-LDMOS structure along cutline AA′.

FIG. 3.

Electric field distribution of the MTSiC-LDMOS structure along cutline AA′.

Close modal
FIG. 4.

Electric field distribution of the MTSiC-LDMOS structure along cutline BB′.

FIG. 4.

Electric field distribution of the MTSiC-LDMOS structure along cutline BB′.

Close modal
FIG. 5.

Electric field distribution of the MTSiC-LDMOS structure along cutline BB′ at the same VDS.

FIG. 5.

Electric field distribution of the MTSiC-LDMOS structure along cutline BB′ at the same VDS.

Close modal

Also, the vertical potential distribution along the CC′ cutline is illustrated in Fig. 6. It is clear that the BV of the MTSiC-LDMOS and C-LDMOS devices is 65 and 36 V, respectively.

FIG. 6.

Vertical potential distribution of the MTSiC-LDMOS structure along the CC′ cutline.

FIG. 6.

Vertical potential distribution of the MTSiC-LDMOS structure along the CC′ cutline.

Close modal

As already discussed in the Introduction, the lattice temperature of the C-LDMOS transistor increases because the BOX material is silicon dioxide. Due to the low thermal conductivity of the buried SiO2 material, the lattice temperature of the C-LDMOS transistor increases with drain voltage. However, for the proposed MTSiC-LDMOS device to reduce the self-heating effect, we use the 4H-SiC layer that has better thermal conductivity than silicon dioxide. By using the 4H-SiC layer at the BOX of the MTSiC-LDMOS structure, a lower lattice temperature compared with the C-LDMOS device is achieved. The maximum lattice temperature against the drain voltage at VD = 20 V and VG = 8 V for both devices is depicted in Fig. 7. When the drain voltage bias increases, the lattice temperature and, then, the self-heating effect for both devices increase, but a lower lattice temperature in the MTSiC-LDMOS will achieve this. The maximum lattice temperatures in the MTSiC-LDMOS and C-LDMOS transistors at VG = 8 V and VD = 20 V are 363 and 310 K, respectively.

FIG. 7.

Maximum lattice temperature against the drain voltage of the MTSiC-LDMOS structure.

FIG. 7.

Maximum lattice temperature against the drain voltage of the MTSiC-LDMOS structure.

Close modal

Other main parameters in power transistors are the drain current and the R O n . s p, which are mentioned earlier. Figure 8 shows the drain current at VG = 8 V and VD = 20 V for the MTSiC-LDMOS and C-LDMOS structures. The proposed MTSiC-LDMOS device has a higher drain current than the C-LDMOS structure. Also, the specific on-resistances as a function of N-drift length for the MTSiC-LDMOS and C-LDMOS devices are shown in Fig. 9. It is clear from Figs. 8 and 9 that the MTSiC-LDMOS has a smaller R O n . s p. This is because of the multiple trenches in the proposed MTSiC-LDMOS device that have higher doping density than its drift region.

FIG. 8.

Drain currents at VG = 8 V and VD = 20 V for the MTSiC-LDMOS and C-LDMOS structures.

FIG. 8.

Drain currents at VG = 8 V and VD = 20 V for the MTSiC-LDMOS and C-LDMOS structures.

Close modal
FIG. 9.

Specific on-resistances as a function of N-drift length for the MTSiC-LDMOS and C-LDMOS devices.

FIG. 9.

Specific on-resistances as a function of N-drift length for the MTSiC-LDMOS and C-LDMOS devices.

Close modal

Now, we present design considerations of the proposed MTSiC-LDMOS device to improve the specific on-resistances, the breakdown voltage, and the lattice temperature. As the doping density of the multiple trenches increases, the R O n . s p decreases. However, on the other hand, the breakdown voltage will decrease. So, we need to find an optimal amount of doping density for the multiple trenches, where the BV will be the highest and the R O n . s p will be the lowest. To find the best optimum value of the trench doping density, Fig. 10 is drawn where the BV and R O n . s p are plotted in this figure. The gate voltage is considered 0 V to obtain the BV. Also, the R O n . s p is obtained in the drain voltage of 1 V and the gate voltage of 10 V. The best value of the doping density as the figure shows is 1017 cm−3.

FIG. 10.

Best optimum value of the trench doping density.

FIG. 10.

Best optimum value of the trench doping density.

Close modal

As mentioned earlier, our proposed structure is suitable for applications where the temperature is of great importance. On the other hand, in the proposed MTSiC-LDMOS structure, we have to optimize the multiple trenches to obtain suitable results. For this purpose, the maximum lattice temperature for the MTSiC-LDMOS structure for different lengths of multiple trenches is drawn in Fig. 11. As can be seen from the figure, the maximum lattice temperature decreases as the length of multiple trenches increases while the breakdown voltage increases. So, we seek the optimum value where the breakdown voltage is the highest and the lattice temperature is the lowest. Therefore, according to Fig. 11, the best length for the multiple trenches is 0.5 μm, in which the maximum lattice temperature and the breakdown voltage are 310 K and 65 V, respectively.

FIG. 11.

Maximum lattice temperature for the MTSiC-LDMOS structure in different lengths of multiple trenches.

FIG. 11.

Maximum lattice temperature for the MTSiC-LDMOS structure in different lengths of multiple trenches.

Close modal

After optimizing the value of the length of multiple trenches, we must also check the value of the height of multiple trenches. So, the breakdown voltage and the maximum lattice temperature for the MTSiC-LDMOS structure are plotted for different heights of multiple trenches in Fig. 12. As indicated in the figure, for a specific height value of the multiple trenches, there is a maximum lattice temperature and a breakdown voltage for the proposed MTSiC-LDMOS structure. As can be seen from the figure, the best height value to have a simultaneous highest breakdown voltage and the lowest temperature is 0.1 μm. It is worth noting that when the height of trenches is 0.1 μm, the electric field distribution for the MTSiC-LDMOS is flatter than the other cases. As a result, a maximum BV for the MTSiC-LDMOS for H = 0.1 μm occurs. Also, we set the gate voltage to 0 V to obtain the breakdown voltage in Figs. 11 and 12 and to obtain the maximum lattice temperature, we set the gate and the drain voltages to 8 and 20 V, respectively.

FIG. 12.

Best optimum value of the trench’s height.

FIG. 12.

Best optimum value of the trench’s height.

Close modal
To compute the trade-off between BV and R O n . s p, the figure-of-merit (FOM) that is measured as Eq. (1),38 
FOM = ( BV ) 2 ( R O n . s p ) ( W c m 2 ) ,
(1)
is used. Also, we should optimize the multiple trenches to get the best result. Therefore, we present the FOM for the proposed MTSiC-LDMOS against the number of trenches in Fig. 13. It is obvious that the best number of trenches is equal to 7, while H, L, and LS are 0.1, 0.5, and 8.5 μm, respectively. On the other hand, to find an optimal amount of the length of the 4H-SiC layer (LS), the BV and R O n . s p are plotted versus the length of the 4H-SiC layer of the MTSiC-LDMOS structure in Fig. 14. As can be seen from the figure, the best value of the length of the 4H-SiC layer is 8.5 μm. The gate voltage is considered 0 V to obtain the BV. Also, the R O n . s p is obtained in the drain voltage of 1 V and the gate voltage of 10 V.
FIG. 13.

FOM for the proposed MTSiC-LDMOS against the number of trenches.

FIG. 13.

FOM for the proposed MTSiC-LDMOS against the number of trenches.

Close modal
FIG. 14.

Best optimum value of the length of the 4H-SiC layer.

FIG. 14.

Best optimum value of the length of the 4H-SiC layer.

Close modal

In conclusion, a comparison between the proposed MTSiC-LDMOS, C-LDMOS, the proposed MTSiC-LDMOS with 6H-SiC material instead of 4H-SiC material (6H-MTSiC-LDMOS), and other devices is mentioned in Table III. It is worth noting that in addition to the advantages of the proposed structure, the use of 4H-SiC material in the proposed MTSiC-LDMOS also has some problems such as making the part more complicated compared to the C-LDMOS.

TABLE III.

Comparison between proposed MTSIC-LDMOS and LDMOS transistors.

DeviceVBR
(V)
ROn.sp
(mΩ mm2)
FOM
(MW/cm2)
T (K) (VG = 8 V, VD = 20 V)
C-LDMOS 36 50.0 2.59 363 
MTSiC-LDMOS 65 35.08 12.04 310 
6H-MTSiC-LDMOS 38 35.21 4.10 310 
Ref. 39  87 63.6 11.90 — 
Ref. 40  64 48.82 8.39 — 
DeviceVBR
(V)
ROn.sp
(mΩ mm2)
FOM
(MW/cm2)
T (K) (VG = 8 V, VD = 20 V)
C-LDMOS 36 50.0 2.59 363 
MTSiC-LDMOS 65 35.08 12.04 310 
6H-MTSiC-LDMOS 38 35.21 4.10 310 
Ref. 39  87 63.6 11.90 — 
Ref. 40  64 48.82 8.39 — 

We presented a novel LDMOS transistor for improving the self-heating effect and the breakdown voltage in this paper. A 4H-SiC layer was replaced in a part of the buried oxide region (BOX). Moreover, the top of the 4H-SiC layer has multiple trenches that have been filled with an N-type silicon material. The proposed device was called multiple trenches 4H-SiC LDMOS (MTSiC-LDMOS). By using the multiple trenches 4H-SiC layer at the proposed device, we saw an enhancement in the breakdown voltage and a reduction in the maximum lattice temperature. Also, the results of the MTSiC-LDMOS structure that is extracted from the two-dimensional ATLAS simulator showed an improvement in the current flow and the specific on-resistance. Therefore, for high breakdown voltage and high temperature, the MTSiC-LDMOS device is more reliable than a conventional LDMOS (C-LDMOS) device.

The authors received no financial support for the research, authorship, and/or publication of this article.

The authors have no conflicts to disclose.

Amir Sohrabi-Movahed: Conceptualization (lead); Software (equal); Writing – original draft (lead). Ali Asghar Orouji: Methodology (equal); Supervision (lead); Writing – review & editing (equal).

1.
Tobias
Erlbacher
,
Lateral Power Transistors in Integrated Circuits
(
Springer
,
Berlin
,
2014
).
2.
P.
Antognetti
,
Power Integrated Circuits: Physics, Design and Applications
(
McGraw-Hill
,
New York
,
1986
).
3.
G.
Ma
,
W.
Burger
, and
M.
Shelds
, “High efficiency 0.4 /spl mu/m gate LDMOS power FET for low voltage wireless communications,” in
1999 IEEE MTT-S International Microwave Symposium Digest
(Cat. No.99CH36282), Anaheim, CA (IEEE, 1999), Vol. 3, pp.
1195
1198
.
4.
Z.
Dong
,
B.
Duan
,
C.
Fu
,
H.
Guo
,
Z.
Cao
, and
Y.
Yang
,
IEEE Electron Device Lett.
39
,
1358
(
2018
).
5.
Y.
Wang
,
B.
Duan
,
L.
Sun
,
X.
Yang
,
Y.
Huang
, and
Y.
Yang
,
Superlatt. Microstruct.
151
,
106810
(
2021
).
6.
M. S.
Adhikari
,
R.
Patel
,
S. L.
Tripathi
, and
Y.
Singh
,
Indian J. Pure Appl. Phys.
58
,
678
(
2020
).
8.
N.
Sghaier
,
J. M.
Bluet
,
A.
Souifi
,
G.
Guillot
,
E.
Morvan
, and
C.
Brylinski
,
IEEE Trans. Electron Devices
50
,
297
(
2003
).
9.
H.
Hjelmgren
,
F.
Allerstam
,
K.
Andersson
,
P. Å.
Nilsson
, and
N.
Rorsman
,
IEEE Trans. Electron Devices
57
,
729
(
2010
).
10.
Z.
Ramezani
,
A. A.
Orouji
, and
P.
Keshavarzi
,
Physica E
59
,
202
(
2014
).
11.
12.
A. A.
Orouji
,
S.
Sharbati
, and
M.
Fathipour
,
IEEE Trans. Device Mater. Reliab.
9
,
449
(
2009
).
13.
M.
Saremi
,
B.
Ebrahimi
,
A.
Afzali-Kusha
, and
S.
Mohammadi
,
Microelectron. Reliab.
51
,
2069
(
2011
).
14.
A.
Gavoshani
,
M.
Dehghan
, and
A. A.
Orouji
,
Silicon
14
,
5801
(
2022
).
15.
16.
M.
Saremi
,
M.
Saremi
,
H.
Niazi
,
M.
Saremi
, and
A. Y.
Goharrizi
,
J. Electron. Mater.
46
,
5570
(
2017
).
17.
M.
Mehrad
,
M.
Zareiee
, and
A. A.
Orouji
,
IEEE Trans. Electron Devices
64
,
4213
(
2017
).
18.
A.
Gavoshani
,
A. A.
Orouji
, and
A.
Abbasi
,
Silicon
14
,
597
(
2022
).
19.
M. S.
Adhikari
and
Y.
Singh
,
Indian J. Phys.
91
,
1211
(
2017
).
20.
M. K.
Anvarifard
,
Z.
Ramezani
,
I. S.
Amiri
, and
A. M.
Nejad
,
Mater. Sci. Semicond. Process.
107
,
104849
(
2020
).
21.
Y.
Wang
,
Z.
Wang
,
T.
Bai
, and
J. B.
Kuo
,
IEEE Trans. Electron Devices
65
,
1056
(
2018
).
22.
M. K.
Anvarifard
,
Superlatt. Microstruct.
98
,
492
(
2016
).
23.
M. S.
Adhikari
and
Y.
Singh
,
IETE Tech. Rev.
36
,
234
(
2019
).
24.
A.
Gavoshani
and
A. A.
Orouji
,
Silicon
15
,
109
(
2023
).
25.
D.
Madadi
,
IEEE Access
10
,
130293
(
2022
).
26.
S. E.
Jamali Mahabadi
,
S.
Rajabi
, and
J.
Loiacono
,
Superlatt. Microstruct.
85
,
872
(
2015
).
27.
A.
Gavoshani
and
A. A.
Orouji
,
J. Comput. Electron.
20
,
1513
(
2021
).
28.
B. A.
Tahne
,
A.
Naderi
, and
F.
Heirani
,
Silicon
12
,
975
(
2020
).
29.
A.
Gavoshani
and
A. A.
Orouji
,
IET Circuits Devices Syst.
16
,
272
(
2022
).
30.
Device Simulator ATLAS
, Silvaco, International, 2012, see www.silvaco.com.
31.
A.
Sohrabi-Movahed
and
Ali A.
Orouji
,
J. Electron. Mater.
52
,
1366
(
2023
).
32.
B. J.
Baliga
,
Modern Power Devices
, 1st ed. (Wiley-Interscience,
1987
).
33.
M.
Ruff
,
H.
Mitlehner
, and
R.
Helbig
,
IEEE Trans. Electron Devices
41
,
1040
(
1994
).
34.
B. J.
Baliga
, “
Silicon carbide power devices
,” in
Springer Handbook of Semiconductor Devices
(
Springer International Publishing
,
Cham
,
2022
).
35.
C. E.
Weitze
,
J. W.
Palmour
,
C. H.
Carter
,
K.
Moore
,
K. J.
Nordquist
, and
S.
Allen
,
IEEE Trans. Electron Devices
43
,
1732
(
1996
).
36.
J.
St Keszewski
,
A.
Jakubowski
, and
M. L.
Korwin-Pawlowski
,
J. Telecommun. Inf. Technol.
3
,
93
(
2007
), see http://www.itl.waw.pl/czasopisma/JTIT/2007/3/93.pdf.
37.
X.
Cheng
,
Z.
Song
,
Y.
Dong
,
Y.
Yu
, and
D.
Shen
,
Microelectron. Eng.
81
,
150
(
2005
).
38.
Y.
Singh
and
R. S.
Rawat
,
Eng. Sci. Technol. Int. J.
18
,
141
(
2015
).
39.
H. A.
Mansoori
,
A. A.
Orouji
, and
A.
Dideban
,
J. Comput. Electron.
16
,
666
(
2017
).
40.
X.
Luo
et al,
IEEE Electron Device Lett.
37
, 1185 (
2016
).