A 1.6 kV Ga2O3 Schottky Barrier Diode with a Low Reverse Current of 1.2 × 10−5 A/cm2 Enabled by Field Plates and N Ion-Implantation Edge Termination

In this work, by employing field plate (FP) and N ion-implantation edge termination (NIET) structure, the electrical performance of the β-Ga2O3 Schottky barrier diode (SBD) was greatly improved. Ten samples of vertical SBDs were fabricated to investigate the influence of the relative positions of field plates (FPs) and ion implantation on the device performance. The device with the FP of 15 μm and the ion implantation at the edge of the Schottky electrode exhibited a breakdown voltage (Vbr) of 1616 V, a specific on-resistance (Ron,sp) of 5.11 mΩ·cm2, a power figure of merit (PFOM) of 0.511 GW/cm2, and a reverse current density of 1.2 × 10−5 A/cm2 @ −1000 V. Compared to the control device, although the Ron,sp increased by 1 mΩ·cm2, the Vbr of the device increased by 183% and the PFOM increased by 546.8%. Moreover, the reverse leakage current of the device with the FP and NIET structure decreased by three orders of magnitude. The TCAD simulation revealed that the peak electric field at the interface decreased from 7 MV/cm @ −500 V to 4.18 MV/cm @ −1000 V. These results demonstrate the great potential for the β-Ga2O3 SBD with a FP and NIET structure in power electronic applications.


Introduction
β-Ga 2 O 3 has garnered significant attention owing to its ultra-wide bandgap of 4.9 eV and high breakdown electric field of 8 MV/cm, making it a promising candidate for next-generation power electronics [1,2].The development of the large-scale melt-growth technique for the β-Ga 2 O 3 wafer has enabled cost-effective manufacturing approaches [3].Moreover, thanks to the outstanding chemical and thermal stability of β-Ga 2 O 3 [4,5], Schottky barrier diodes (SBDs) [6][7][8][9] and metal-oxide-semiconductor field-effect transistors (MOSFETs) [10][11][12][13] fabricated on β-Ga 2 O 3 wafers have demonstrated strong competitiveness.However, there remains a significant disparity in the breakdown electric field between the material's theoretical prediction and the fabricated β-Ga 2 O 3 SBDs.This discrepancy is primarily owing to the material defects and the electric field concentration at the Schottky electrode edge.Various approaches, such as field plates (FPs) [14,15], the trench structure [16][17][18], thermally oxidized termination [19], and heterogeneous termination [20,21], have been proposed to suppress the electric field concentration.Meanwhile, the ion implantation technique [22][23][24][25], which creates a high-resistance region in β-Ga 2 O 3 , is considered highly effective for increasing the breakdown voltage (V br ) while reducing the reverse leakage current.Zhou et al. [22] reported a β-Ga 2 O 3 SBD with Mg-implanted edge termination, achieving a V br of 1500 V.He and Mg ion implantation were utilized to fabricate β-Ga 2 O 3 SBDs, resulting in a reduced reverse leakage current and enhanced V br [26].Yet, the influence of the ion implantation position and the relative position between the FP and ion implantation on the device performance has not been studied, which could have a significant impact on the electrical characteristics.
Here, β-Ga 2 O 3 SBDs with FP and N ion-implantation edge termination (NIET) have been demonstrated.The device with the FP of 15 µm and the ion implantation at the edge of the Schottky electrode improved the V br from 572 V to 1616 V and achieved the power figure of merit (PFOM) of 0.511 GW/cm 2 .TCAD simulations were employed to analyze the electric field distribution in β-Ga 2 O 3 SBDs, revealing that the FP and NIET structure significantly enhances the device performance.

Device Structure and Fabrication Process
The β-Ga 2 O 3 SBDs were fabricated on a β-Ga 2 O 3 wafer acquired from Novel Crystal Technology (NCT, Tokyo, Japan).The wafer consists of a 587 µm thick β-Ga 2 O 3 (001) substrate with a Sn doping concentration of ~5.1 × 10 18 cm −3 and a 11 µm~3.2× 10 16 cm −3  Si doping epitaxial layer by halide vapor phase epitaxy (HVPE) [27,28].To investigate the effect of the FP and NIET structure on the device performance, ten samples, shown in Figure 1, were prepared and categorized into three groups based on the FP dimensions.Group 1, Group 2, and Group 3 had FP lengths of 0, 15, and 5 µm, respectively.Sample 1 served as a control without any special treatment.
Nanomaterials 2024, 14, x FOR PEER REVIEW 2 of 10 influence of the ion implantation position and the relative position between the FP and ion implantation on the device performance has not been studied, which could have a significant impact on the electrical characteristics.
Here, β-Ga2O3 SBDs with FP and N ion-implantation edge termination (NIET) have been demonstrated.The device with the FP of 15 μm and the ion implantation at the edge of the Schottky electrode improved the Vbr from 572 V to 1616 V and achieved the power figure of merit (PFOM) of 0.511 GW/cm 2 .TCAD simulations were employed to analyze the electric field distribution in β-Ga2O3 SBDs, revealing that the FP and NIET structure significantly enhances the device performance.

Device Structure and Fabrication Process
The β-Ga2O3 SBDs were fabricated on a β-Ga2O3 wafer acquired from Novel Crystal Technology (NCT, Tokyo, Japan).The wafer consists of a 587 μm thick β-Ga2O3 (001) substrate with a Sn doping concentration of ~5.1 × 10 18 cm −3 and a 11 μm ~3.2 × 10 16 cm −3 Si doping epitaxial layer by halide vapor phase epitaxy (HVPE) [27,28].To investigate the effect of the FP and NIET structure on the device performance, ten samples, shown in Figure 1, were prepared and categorized into three groups based on the FP dimensions.Group 1, Group 2, and Group 3 had FP lengths of 0, 15, and 5 μm, respectively.Sample 1 served as a control without any special treatment.The device fabrication process is illustrated in Figure 2. Initially, the β-Ga2O3 wafer underwent sequential cleaning with acetone, isopropanol, and deionized (DI) water for 10 min each to eliminate surface contaminants.N ion implantation was performed three times on the wafer to form a 500 nm deep box profile with a simulated doping concentration of around 2~3 × 10 19 cm −3 .The implantation was carried out at energies/concentrations of 50 keV/5 × 10 12 cm −2 , 150 keV/1 × 10 13 cm −2 , and 230 keV/2 × 10 13 cm −2 in sequence, as plotted in Figure 3.Then, thermal annealing was executed at 1100 °C for 30 min in N2 to recover the implantation damage and activate the implanted N atoms.The back-Ohmic contact was formed by the electron-beam evaporation (EBE) of Ti/Au (30 nm/100 nm), followed by rapid thermal annealing (RTA) for 60 s at 470 °C in N2.Additionally, 100 nm of Al2O3 was deposited onto the β-Ga2O3 surface via EBE as the FP oxide.The Al2O3 layer was amorphous.During the process, the vacuum was maintained below 5 × 10 −5 Pa, and the growth rate was kept at ~0.1 nm/s.Finally, Schottky anode electrodes were fabricated The device fabrication process is illustrated in Figure 2. Initially, the β-Ga 2 O 3 wafer underwent sequential cleaning with acetone, isopropanol, and deionized (DI) water for 10 min each to eliminate surface contaminants.N ion implantation was performed three times on the wafer to form a 500 nm deep box profile with a simulated doping concentration of around 2~3 × 10 19 cm −3 .The implantation was carried out at energies/concentrations of 50 keV/5 × 10 12 cm −2 , 150 keV/1 × 10 13 cm −2 , and 230 keV/2 × 10 13 cm −2 in sequence, as plotted in Figure 3.Then, thermal annealing was executed at 1100 • C for 30 min in N 2 to recover the implantation damage and activate the implanted N atoms.The back-Ohmic contact was formed by the electron-beam evaporation (EBE) of Ti/Au (30 nm/100 nm), followed by rapid thermal annealing (RTA) for 60 s at 470 • C in N 2 .Additionally, 100 nm of Al 2 O 3 was deposited onto the β-Ga 2 O 3 surface via EBE as the FP oxide.The Al 2 O 3 layer was amorphous.During the process, the vacuum was maintained below 5 × 10 −5 Pa, and the growth rate was kept at ~0.1 nm/s.Finally, Schottky anode electrodes were fabricated by evaporating Ni/Au (50 nm/100 nm).The forward current-voltage (I-V) and V br characteristics were measured using the semiconductor device analyzers Keysight B1500A and B1505A (Santa Rosa, CA, USA), respectively, at room temperature.To evaluate the impact of the NIET on the breakdown performance of the β-Ga 2 O 3 SBD, a physicsbased technology computer-aided design (TCAD) simulation in Sentaurus was performed, where the parameters of the β-Ga 2 O 3 in this simulation are shown in Table 1.The Okuto impact ionization model, field-dependent mobility, Shockley Read Hall recombination, and bandgap-narrowing effect were also taken into account.
by evaporating Ni/Au (50 nm/100 nm).The forward current-voltage (I-V) and Vbr characteristics were measured using the semiconductor device analyzers Keysight B1500A and B1505A (Santa Rosa, California, USA), respectively, at room temperature.To evaluate the impact of the NIET on the breakdown performance of the β-Ga2O3 SBD, a physics-based technology computer-aided design (TCAD) simulation in Sentaurus was performed, where the parameters of the β-Ga2O3 in this simulation are shown in Table 1.The Okuto impact ionization model, field-dependent mobility, Shockley Read Hall recombination, and bandgap-narrowing effect were also taken into account.by evaporating Ni/Au (50 nm/100 nm).The forward current-voltage (I-V) and Vbr characteristics were measured using the semiconductor device analyzers Keysight B1500A and B1505A (Santa Rosa, California, USA), respectively, at room temperature.To evaluate the impact of the NIET on the breakdown performance of the β-Ga2O3 SBD, a physics-based technology computer-aided design (TCAD) simulation in Sentaurus was performed, where the parameters of the β-Ga2O3 in this simulation are shown in Table 1.The Okuto impact ionization model, field-dependent mobility, Shockley Read Hall recombination, and bandgap-narrowing effect were also taken into account.

Electrical Properties and Simulations of Devices in Group 1
Figure 4a shows the linear forward current density-voltage (J-V) characteristics of Sample 1 and Sample 2. The specific on-resistances (R on,sp ) of Sample 1 and Sample 2 are 4.16 mΩ•cm 2 and 4.68 mΩ•cm 2 , respectively.The increase in the R on,sp post-ion implantation is attributed to the N-implanted Ga 2 O 3 acting as a current-blocking layer [29,30], thereby reducing the effective area of the Schottky contact.Figure 4b depicts the semilogarithmic J-V curves of Sample 1 and Sample 2, where the turn-on voltage (V on ) is extracted at the current density of 1 A/cm 2 .The V on also increases from 0.84 V to 0.89 V by ion implantation.This could be mainly due to the lateral depletion problem of edge termination [26].

Electrical Properties and Simulations of Devices in Group 1
Figure 4a shows the linear forward current density-voltage (J-V) characteristics of Sample 1 and Sample 2. The specific on-resistances (Ron,sp) of Sample 1 and Sample 2 are 4.16 mΩ•cm 2 and 4.68 mΩ•cm 2 , respectively.The increase in the Ron,sp post-ion implantation is attributed to the N-implanted Ga2O3 acting as a current-blocking layer [29,30], thereby reducing the effective area of the Schottky contact.Figure 4b depicts the semi-logarithmic J-V curves of Sample 1 and Sample 2, where the turn-on voltage (Von) is extracted at the current density of 1 A/cm 2 .The Von also increases from 0.84 V to 0.89 V by ion implantation.This could be mainly due to the lateral depletion problem of edge termination [26].Figure 5a presents the reverse breakdown characteristics, indicating that Sample 2 has a higher Vbr and significantly lower leakage current compared to Sample 1.The leakage current of Sample 2 is less than 1/100th of that of Sample 1 at −500 V. Figure 5b shows the surface electric field distributions of Sample 1 and Sample 2 simulated by the TCAD.The implanted region with a depth of 500 nm is set as the semi-insulating layer because the crystal lattice is damaged by the high-energy N ion implant.The peak electric fields of both samples are located at 30 μm, the edge of the Schottky electrode.Additionally, the vertical electric field distributions along Line Y, which is shown in Figure 2, are plotted in the inset of Figure 5b.The right half of the device is set as the simulation cell, with the center of the device as the origin.The positive x-axis extends to the right, parallel to the material surface, and the positive y-axis extends downward, perpendicular to the material surface.The peak electric field of Sample 1 exceeds 7 MV/cm at −500 V, while, for Sample 2, it is around 6.5 MV/cm.These results show that the ion-implantation ET structure can reduce the peak electric field at the interface and prevent premature device breakdown.Figure 5c,d display the statistical distributions of the reverse breakdown and the leakage current @ −500 V for Sample 1 and Sample 2. The average breakdown voltages of Sample 1 and Sample 2 are 572 V and 920 V, respectively, illustrating a 60.8% increase after ion implantation, which is mainly owing to the electric field concentration suppressed with the ion-implanted ET structure.On the other hand, the average leakage current density @ −500 V of Sample 2 is 5.2 × 10 −7 A/cm 2 , an approximately nearly three orders of magnitude reduction in the reverse leakage current.Figure 5a presents the reverse breakdown characteristics, indicating that Sample 2 has a higher V br and significantly lower leakage current compared to Sample 1.The leakage current of Sample 2 is less than 1/100th of that of Sample 1 at −500 V. Figure 5b shows the surface electric field distributions of Sample 1 and Sample 2 simulated by the TCAD.The implanted region with a depth of 500 nm is set as the semi-insulating layer because the crystal lattice is damaged by the high-energy N ion implant.The peak electric fields of both samples are located at 30 µm, the edge of the Schottky electrode.Additionally, the vertical electric field distributions along Line Y, which is shown in Figure 2, are plotted in the inset of Figure 5b.The right half of the device is set as the simulation cell, with the center of the device as the origin.The positive x-axis extends to the right, parallel to the material surface, and the positive y-axis extends downward, perpendicular to the material surface.The peak electric field of Sample 1 exceeds 7 MV/cm at −500 V, while, for Sample 2, it is around 6.5 MV/cm.These results show that the ion-implantation ET structure can reduce the peak electric field at the interface and prevent premature device breakdown.Figure 5c,d display the statistical distributions of the reverse breakdown and the leakage current @ −500 V for Sample 1 and Sample 2. The average breakdown voltages of Sample 1 and Sample 2 are 572 V and 920 V, respectively, illustrating a 60.8% increase after ion implantation, which is mainly owing to the electric field concentration suppressed with the ion-implanted ET structure.On the other hand, the average leakage current density @ −500 V of Sample 2 is 5.2 × 10 −7 A/cm 2 , an approximately nearly three orders of magnitude reduction in the reverse leakage current.

Electrical Properties and Simulations of Devices in Group 2
Figure 6a depicts the forward J-V characteristics of Sample 3 to Sample 6, where the specific on-resistances of Sample 3, Sample 4, Sample 5, and Sample 6 are 4.20 mΩ•cm 2 , 4.67 mΩ•cm 2 , 5.11 mΩ•cm 2 , and 6.31 mΩ•cm 2 , respectively.Sample 6 shows a 23.4% decrease in the Ron,sp compared to Sample 5, which could be ascribed to the fact that a larger ion-implanted region of Sample 6 leads to more N ion diffusion after annealing.The semilogarithmic J-V characteristics of Sample 3 to Sample 6 are shown in Figure 6b.The turnon vlotages of Sample 3, Sample 4, Sample 5, and Sample 6 are 0.84 V, 0.85 V, 0.90 V, and 0.91 V, respectively, where the turn-on voltages of Sample 5 and Sample 6 show a more significant increase.This implies that the ion implantation close to the edge of the Schottky electrode has a greater impact on the Von of the devices.Figure 7a conveys the reverse breakdown characteristics of Sample 3 to Sample 6.Under the same reverse bias, Sample 3 has a significantly higher leakage current than

Electrical Properties and Simulations of Devices in Group 2
Figure 6a depicts the forward J-V characteristics of Sample 3 to Sample 6, where the specific on-resistances of Sample 3, Sample 4, Sample 5, and Sample 6 are 4.20 mΩ•cm 2 , 4.67 mΩ•cm 2 , 5.11 mΩ•cm 2 , and 6.31 mΩ•cm 2 , respectively.Sample 6 shows a 23.4% decrease in the R on,sp compared to Sample 5, which could be ascribed to the fact that a larger ion-implanted region of Sample 6 leads to more N ion diffusion after annealing.The semi-logarithmic J-V characteristics of Sample 3 to Sample 6 are shown in Figure 6b.The turn-on vlotages of Sample 3, Sample 4, Sample 5, and Sample 6 are 0.84 V, 0.85 V, 0.90 V, and 0.91 V, respectively, where the turn-on voltages of Sample 5 and Sample 6 show a more significant increase.This implies that the ion implantation close to the edge of the Schottky electrode has a greater impact on the V on of the devices.

Electrical Properties and Simulations of Devices in Group 2
Figure 6a depicts the forward J-V characteristics of Sample 3 to Sample 6, where the specific on-resistances of Sample 3, Sample 4, Sample 5, and Sample 6 are 4.20 mΩ•cm 2 , 4.67 mΩ•cm 2 , 5.11 mΩ•cm 2 , and 6.31 mΩ•cm 2 , respectively.Sample 6 shows a 23.4% decrease in the Ron,sp compared to Sample 5, which could be ascribed to the fact that a larger ion-implanted region of Sample 6 leads to more N ion diffusion after annealing.The semilogarithmic J-V characteristics of Sample 3 to Sample 6 are shown in Figure 6b.The turnon vlotages of Sample 3, Sample 4, Sample 5, and Sample 6 are 0.84 V, 0.85 V, 0.90 V, and 0.91 V, respectively, where the turn-on voltages of Sample 5 and Sample 6 show a more significant increase.This implies that the ion implantation close to the edge of the Schottky electrode has a greater impact on the Von of the devices.Figure 7a conveys the reverse breakdown characteristics of Sample 3 to Sample 6.Under the same reverse bias, Sample 3 has a significantly higher leakage current than Figure 7a conveys the reverse breakdown characteristics of Sample 3 to Sample 6.Under the same reverse bias, Sample 3 has a significantly higher leakage current than other samples, confirming that the combined FP and NIET structure results in lower reverse leakage current compared to the conventional FP structure.Figure 7b shows the surface electric field distributions of Sample 3 to Sample 6, with two peak positions at 30 µm and 45 µm, corresponding to the edge of the Schottky electrode and the edge of the FP, respectively.The higher electric field at 30 µm can trigger premature breakdown.The inset displays the vertical electric field along Line Y, revealing that, at 30 µm, the peak electric fields of Sample 5 and Sample 6 are nearly equivalent and significantly lower than those of Sample 3 and Sample 4. This demonstrates that ion implantation at the edge of the Schottky electrode in Sample 5 and Sample 6 effectively suppresses the peak electric field.Meanwhile, the peak electric field of Sample 4 is slightly lower than that of Sample 3, indicating that ion implantation at the edge of the FP has little effect on the peak electric field.Figure 7c,d plot the statistical distributions of V br and the reverse leakage current density @ −1000 V for Sample 3 to Sample 6.The average breakdown voltages for Sample 3, Sample 4, Sample 5, and Sample 6 are 1276 V, 1422 V, 1616 V, and 1610 V, respectively, while the average leakage current densities are 2.6 × 10 −2 A/cm 2 , 4.4 × 10 −4 A/cm 2 , 1.2 × 10 −5 A/cm 2 , and 4.1 × 10 −6 A/cm 2 , respectively.These results suggest that the FP and NIET structure significantly improves the reverse characteristics, with ion implantation at the Schottky electrode edge offering greater enhancement compared to implantation at the FP edge.
Nanomaterials 2024, 14, x FOR PEER REVIEW 6 of 10 other samples, confirming that the combined FP and NIET structure results in lower reverse leakage current compared to the conventional FP structure.Figure 7b shows the surface electric field distributions of Sample 3 to Sample 6, with two peak positions at 30 μm and 45 μm, corresponding to the edge of the Schottky electrode and the edge of the FP, respectively.The higher electric field at 30 μm can trigger premature breakdown.The inset displays the vertical electric field along Line Y, revealing that, at 30 μm, the peak electric fields of Sample 5 and Sample 6 are nearly equivalent and significantly lower than those of Sample 3 and Sample 4. This demonstrates that ion implantation at the edge of the Schottky electrode in Sample 5 and Sample 6 effectively suppresses the peak electric field.Meanwhile, the peak electric field of Sample 4 is slightly lower than that of Sample 3, indicating that ion implantation at the edge of the FP has little effect on the peak electric field.Figure 7c,d plot the statistical distributions of Vbr and the reverse leakage current density @ −1000 V for Sample 3 to Sample 6.The average breakdown voltages for Sample 3, Sample 4, Sample 5, and Sample 6 are 1276 V, 1422 V, 1616 V, and 1610 V, respectively, while the average leakage current densities are 2.6 × 10 −2 A/cm 2 , 4.4 × 10 −4 A/cm 2 , 1.2 × 10 −5 A/cm 2 , and 4.1 × 10 −6 A/cm 2 , respectively.These results suggest that the FP and NIET structure significantly improves the reverse characteristics, with ion implantation at the Schottky electrode edge offering greater enhancement compared to implantation at the FP edge. -

Electrical Properties and Simulations of Devices in Group 3
Figure 8a depicts the linear forward J-V characteristics of Sample 7 to Sample 10, with specific on-resistances of 4.05 mΩ•cm 2 , 4.29 mΩ•cm 2 , 5.16 mΩ•cm 2 , and 5.34 mΩ•cm 2 , respectively.The Ron,sp of Sample 4 increased by 0.47 mΩ•cm² compared to Sample 3, while the Ron,sp of Sample 8 increased by 0.24 mΩ•cm² compared to Sample 7. The smaller increase in Ron,sp for Sample 8 compared to Sample 7 could be mainly caused by the current crowding in the 5 μm FP and NIET composite structure, which results in a lower Ron,sp for the device.Figure 8b presents the semi-logarithmic J-V curves of Sample 7 to Sample 10,

Electrical Properties and Simulations of Devices in Group 3
Figure 8a depicts the linear forward J-V characteristics of Sample 7 to Sample 10, with specific on-resistances of 4.05 mΩ•cm 2 , 4.29 mΩ•cm 2 , 5.16 mΩ•cm 2 , and 5.34 mΩ•cm 2 , respectively.The R on,sp of Sample 4 increased by 0.47 mΩ•cm² compared to Sample 3, while the R on,sp of Sample 8 increased by 0.24 mΩ•cm² compared to Sample 7. The smaller increase in R on,sp for Sample 8 compared to Sample 7 could be mainly caused by the current crowding in the 5 µm FP and NIET composite structure, which results in a lower R on,sp for the device.Figure 8b presents the semi-logarithmic J-V curves of Sample 7 to Sample 10, with turn-on voltages of 0.89 V, 0.90 V, 0.95 V, and 0.95 V, respectively.The R on,sp and V on of the samples in Group 3 exhibit analogous tendencies to those in Group 2, indicative of a positive correlation between V on and the distance from the ion implantation position to the Schottky contact area, as well as the ion implantation area.
with turn-on voltages of 0.89 V, 0.90 V, 0.95 V, and 0.95 V, respectively.The Ron,sp and Von of the samples in Group 3 exhibit analogous tendencies to those in Group 2, indicative of a positive correlation between Von and the distance from the ion implantation position to the Schottky contact area, as well as the ion implantation area.Figure 9a presents the reverse breakdown characteristics of Sample 7 to Sample 10, showing that Sample 7 has a significantly higher leakage current than the other samples, consistent with the results of the Group 2. Figure 9b displays the surface electric field distributions of Sample 7 to Sample 10, with the inset portraying the vertical electric field along Line Y.The electric field at 30 μm remains higher than that at 45 μm. Figure 9c,d depict the Vbr and the reverse leakage current density @ −1000 V statistics for Sample 7 to Sample 10, with average breakdown voltages of 1223 V, 1307 V, 1454 V, and 1469 V, respectively, and average leakage current densities of 4.1 × 10 −2 A/cm 2 , 3.8 × 10 −5 A/cm 2 , 6.4 × 10 −6 A/cm 2 , and 5.5 × 10 −6 A/cm 2 , respectively.The tendencies of the Vbr and reverse leakage current densities of the samples in Group 3 are consistent with those in Group 2. Additionally, the breakdown voltages of Sample 5 and Sample 6 are higher than those of Sample 9 and Sample 10.This reveals that ion implantation at the edge of the Schottky electrode enhances Vbr more effectively when the FP length is 15 μm. Figure 9a presents the reverse breakdown characteristics of Sample 7 to Sample 10, showing that Sample 7 has a significantly higher leakage current than the other samples, consistent with the results of the Group 2. Figure 9b displays the surface electric field distributions of Sample 7 to Sample 10, with the inset portraying the vertical electric field along Line Y.The electric field at 30 µm remains higher than that at 45 µm.Figure 9c,d depict the V br and the reverse leakage current density @ −1000 V statistics for Sample 7 to Sample 10, with average breakdown voltages of 1223 V, 1307 V, 1454 V, and 1469 V, respectively, and average leakage current densities of 4.1 × 10 −2 A/cm 2 , 3.8 × 10 −5 A/cm 2 , 6.4 × 10 −6 A/cm 2 , and 5.5 × 10 −6 A/cm 2 , respectively.The tendencies of the V br and reverse leakage current densities of the samples in Group 3 are consistent with those in Group 2. Additionally, the breakdown voltages of Sample 5 and Sample 6 are higher than those of Sample 9 and Sample 10.This reveals that ion implantation at the edge of the Schottky electrode enhances V br more effectively when the FP length is 15 µm.

Comprehensive Comparisons and Analysis
Table 2 summarizes the V br , R on,sp , V on , and reverse leakage current densities (J rev ) of Samples 1-10. Figure 10a-c show the benchmark plots of R on,sp versus V br , V on versus PFOM, and J rev versus PFOM, respectively.Sample 5 achieves the highest PFOM, with a very low reverse leakage current of 1.2 × 10 −5 A/cm 2 @ −1000 V, although it has a relatively high V on .On the other hand, Sample 4 shows little increase in V on compared to the conventional device; however, its reverse characteristics are inferior to Sample 5. Figure 10d sums up the simulated peak electric fields of Group 2 and Group 3, where the former is generally lower than the latter, suggesting that the devices in Group 2 exhibit superior breakdown tolerance compared to the devices in Group 3. Ion implantation with an FP length of 15 µm significantly improves device electrical performance.Figure 10e,f compare the electrical characteristics of Sample 1 and Sample 5, showing that the β-Ga 2 O 3 SBD with the FP and NIET structure shows a 183% increase in V br and a reduction in the reverse leakage current by three orders of magnitude relative to the conventional device.Although the R on,sp increases by 1 mΩ•cm 2 , the PFOM of the device increases by 546.8%.

Comprehensive Comparisons and Analysis
Table 2 summarizes the Vbr, Ron,sp, Von, and reverse leakage current densities (Jrev) of Samples 1-10. Figure 10a-c show the benchmark plots of Ron,sp versus Vbr, Von versus PFOM, and Jrev versus PFOM, respectively.Sample 5 achieves the highest PFOM, with a very low reverse leakage current of 1.2 × 10 −5 A/cm 2 @ −1000 V, although it has a relatively high Von.On the other hand, Sample 4 shows little increase in Von compared to the conventional device; however, its reverse characteristics are inferior to Sample 5. Figure 10d sums up the simulated peak electric fields of Group 2 and Group 3, where the former is generally lower than the latter, suggesting that the devices in Group 2 exhibit superior breakdown tolerance compared to the devices in Group 3. Ion implantation with an FP length of 15 μm significantly improves device electrical performance.Figure 10e,f compare the electrical characteristics of Sample 1 and Sample 5, showing that the β-Ga2O3 SBD with the FP and NIET structure shows a 183% increase in Vbr and a reduction in the reverse leakage current by three orders of magnitude relative to the conventional device.Although the Ron,sp increases by 1 mΩ•cm 2 , the PFOM of the device increases by 546.8%.

Figure 2 .
Figure 2. Process flow of the fabricated β-Ga2O3 SBDs with the FP and NIET structure.

Figure 3 .
Figure 3. N-implanted total concentration after three N ion implantations.

Figure 2 .Figure 3 .
Figure 2. Process flow of the fabricated β-Ga2O3 SBDs with the FP and NIET structure.

Figure 4 .
Figure 4. (a) Forward J-V characteristics and Ron,sp; (b) semi-log scale J-V characteristics and Von of Sample 1 and Sample 2.

Figure 4 .
Figure 4. (a) Forward J-V characteristics and R on,sp ; (b) semi-log scale J-V characteristics and V on of Sample 1 and Sample 2.

Figure 5 .
Figure 5. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of Vbr; (d) statistical plots of the reverse leakage current densities of Sample 1 and Sample 2.

Figure 6 .
Figure 6.(a) Forward J-V characteristics and Ron,sp; (b) semi-log scale J-V characteristics and Von of Sample 3 to Sample 6.

Figure 5 .
Figure 5. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of V br ; (d) statistical plots of the reverse leakage current densities of Sample 1 and Sample 2.

Figure 5 .
Figure 5. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of Vbr; (d) statistical plots of the reverse leakage current densities of Sample 1 and Sample 2.

Figure 6 .
Figure 6.(a) Forward J-V characteristics and Ron,sp; (b) semi-log scale J-V characteristics and Von of Sample 3 to Sample 6.

Figure 6 .
Figure 6.(a) Forward J-V characteristics and R on,sp ; (b) semi-log scale J-V characteristics and V on of Sample 3 to Sample 6.

Figure 7 .
Figure 7. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of Vbr; (d) statistical plots of reverse leakage current densities of Sample 3 to Sample 6.

Figure 7 .
Figure 7. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of V br ; (d) statistical plots of reverse leakage current densities of Sample 3 to Sample 6.

Figure 8 .
Figure 8.(a) Forward J-V characteristics and Ron,sp; (b) semi-log scale J-V characteristics and Von of Sample 7 to Sample 10.

Figure 9 .
Figure 9. (a) Reverse characteristics; (b) simulations of the surface electric field distributions; (c) statistical plots of Vbr; (d) statistical plots of reverse leakage current densities of Sample 7 to Sample 10.

Table 1 .
Material parameters used in this simulation.

Table 1 .
Material parameters used in this simulation.
3 Figure 3. N-implanted total concentration after three N ion implantations.

Table 1 .
Material parameters used in this simulation.

Table 2 .
V br , R on,sp , V on , and reverse leakage current densities of Samples 1-10.