15 MeV proton damage in NiO/β-Ga2O3 vertical rectifiers

15 MeV proton irradiation of vertical geometry NiO/β-Ga2O3 heterojunction rectifiers produced reductions in reverse breakdown voltage from 4.3 kV to 3.7 kV for a fluence of 1013ions·cm−2 and 1.93 kV for 1014 ions·cm−2. The forward current density was also decreased by 1–2 orders of magnitude under these conditions, with associated increase in on-state resistance R ON. These changes are due to a reduction in carrier density and mobility in the drift region. The reverse leakage current increased by a factor of ∼2 for the higher fluence. Subsequent annealing up to 400 °C further increased reverse leakage due to deterioration of the contacts, but the initial carrier density of 2.2 × 1016 cm−3 was almost fully restored by this annealing in the lower fluence samples and by more than 50% in the 1014 cm−2 irradiated devices. Carrier removal rates in the Ga2O3 were in the range 190–1200 for the fluence range employed, similar to Schottky rectifiers without the NiO.


Introduction
β-Ga 2 O 3 is attracting significant recent attention for power switching devices [1][2][3] and solar-blind UV photodetectors [4]. In particular, significant advancements have been achieved in the development of NiO/β-Ga 2 O 3 power rectifiers, surpassing the performance limitations observed in GaN structures, particularly in one-dimensional configurations [5][6][7][8]. Notably, the latest experimental demonstrations have shown maximum breakdown voltages exceeding 8 kV, corresponding to critical electric fields over 8 MV·cm −1 . The determination of the precise critical electric field of β-Ga 2 O 3 remains challenging due to its dependence on various factors such as doping in the drift region of the rectifiers, operating temperature, and geometric configuration. These factors contribute to electric field crowding, thereby influencing the critical field and necessitating further investigation.
The use of p-type NiO to form heterojunctions with n-type Ga 2 O 3 has mitigated the absence of practical p-type doping for the latter. This has led to recent demonstrations of vertical rectifiers with breakdown voltages more than 8 kV with excellent high temperature operation [9]. While the device performance is promising in terms of dc and switching applications [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26], little is known about the effects of radiation on these heterojunctions. While the Ga 2 O 3 is known to be relatively resistant to total dose damage [27,28], large reversible changes in current-voltage characteristics of the heterojunctions have been observed after Co-60 gamma ray exposure which appears to be due to conductivity changes in the NiO [29]. Other low dose ion irradiated oxides have also shown enhanced conductivity which in some cases has been linked to irradiation/illumination assisted desorption of oxygen containing species from the oxide surface [30][31][32][33]. There have also been recent demonstrations of single event burnout (SEB) in Ga 2 O 3 rectifiers [34], while simulations show the SEB threshold voltage of conventional Ga 2 O 3 MOSFETs is lower than that of state-of-the-art AlGaN/GaN HEMTs [35,36]. Field management approaches can provide some mitigation of single event effects in Ga 2 O 3 [37]. There is clearly scope for additional studies of radiation effects in Ga 2 O 3 -based device structures.
In this paper we report the effects of 15 MeV protons on the electrical performance of NiO/β-Ga 2 O 3 vertical rectifiers. The displacement damage from the protons reduces the carrier concentration in the drift region of the rectifiers and reduces the breakdown voltage. Partial recovery after annealing at 400 • C is observed but is limited by degradation of the contacts.

Experimental
The vertical rectifiers have been described in detail previously [23][24][25], but in brief consist of 10 µm drift region of lightly n-type Ga 2 O 3 grown on a conducting n + Ga 2 O 3 . NiO with a total thickness of 20 nm is deposited on the top surface by sputtering and contacts made to both sides by e-beam evaporation of Ti/Au to the rear surface and Ni/Au to the NiO. A schematic of the device structure is shown in figure 1.
Proton irradiations were performed at the Korean Institute of Radiological and Medical Sciences using an MC 50 (Scanditronix) cyclotron. The proton energy was held constant at 15 MeV. The irradiated fluence was either 10 13 or 10 14 cm −2 at a constant beam current of 10 nA. The irradiation times were 102 s for the low fluence and 1018 s for the high fluence. The Stopping and Range of Ions in Matter (SRIM) simulation code [38] was used to estimate the projected range of ∼600 µm (figure 2, top), which means the protons end up in the conducting substrate but create damage throughout the NiO and the epitaxial Ga 2 O 3 layer. We also used to the SR-NIEL simulator [39] to estimate the vacancy density distribution as a function of proton irradiation energy. The non-ionizing energy loss at this energy is 5.7 × 10 −3 MeV·cm 2 g −1 , as shown in figure 2 (bottom) [39]. There is still a larger energy loss due to ionization, but this is dissipated as heat and does not create lattice damage. The device DC characteristics were measured with an HP 4156 parameter analyzer. Capacitance-voltage values were taken with an Agilent 4284A Precision LCR Meter. Post-irradiation annealing was carried out in a N 2 ambient for 60 s in a Rapid Thermal Annealing furnace.

Results and discussion
The forward current density-voltage (J-I) characteristics from devices after irradiation with 10 13 cm −2 or 10 14 cm −2 are shown in figure 3. The associated on-state resistance, R ON , calculated from the slope of these are also shown. The effect of the irradiation is to reduce the forward current density by approximately one order of magnitude for the lower fluence and 2 orders of magnitude for the higher fluence. The relation between charge and electric field E (Poisson's equation) and transport (drift/diffusion) equations means that the current density depends on both carrier mobility and density, e.g.
where n/p are the electron/hole densities, e is the electronic charge and µ is the mobility. Thus the creation of traps that remove carriers from the drift region and degrade the carrier mobility in that region are the causes of the change in forward current. Radiation creates traps that remove carriers from the conduction process and degrade mobility, i.e. n, µ are reduced [26]. The R ON values show only a slight degradation for low fluence and two orders of magnitude for the high fluence. Under forward bias, the drift region conducts with an on-resistance R on = 1/neµ. Therefore, a decrease in carrier density increases on-resistance. Subsequent annealing brought a partial restoration of forward current for both fluences, but degradation of the front contact was apparent above 300 • C and prevented further improvement in device performance. This is clearly a result of the Ni/Au interacting with the NiO, since previous studies have shown the rear Ti/Au/Ga 2 O 3 contact is stable under these conditions [40,41]. Control samples that were not irradiated but annealed at the same temperatures confirm the origin of the degradation at 400 • C. The trends in forward density with fluence and annealing temperature are made clearer in the linear plot of figure 4, which shows that annealing alone degrades the forward current for 400 • C anneals. Note that companion Schottky rectifiers fabricated on the same wafer without the NiO layers also showed similar changes in current, indicating that changes in the resistance of the NiO are not responsible for the current changes. The NiO is also very highly doped (>10 19 cm −3 ) and would not be affected at the fluences used here.
The trends in reverse current in the low bias voltage region as a function of fluence are shown in figure 5. The current increases by ∼50% for the low fluence and ∼100% for the higher fluence, but is further degraded by annealing, independent of whether the sample had also been irradiated. Thus, two different mechanisms are present-the introduction of generation-recombination centers by the proton damage and the contact degradation at 400 • C. The increase in current due to the contact degradation is caused by a lowering of the effective barrier height as the contact metal reacts with the Ga 2 O 3 and creates a non-uniform interface [42,43].     The breakdown voltages were extracted from the high bias reverse current characteristics and defined as the bias for a reverse current reaching 0.1 A·cm 2 . These measurements were performed in Fluorinert atmosphere. As shown in figure 6, the V B values were 4.3 kV for the unirradiated rectifiers and decreased to 3.7 kV for the low fluence and 1.93 kV for the high fluence. Annealing at 300 • C brought little change for either of the fluences but at 400 • C there was an increase to 3.7 kV for the higher fluence sample showing at this condition that the reduction in G-R centers outweighed the effect of contact degradation.
The change in drift region carrier density was obtained from the C 2 -V characteristics for figure 7. The initial carrier density of 2.2 × 10 16 cm −3 was reduced to 9.6 × 10 15 cm −3 for the low fluence and 3.1 × 10 15 for the high fluence. These reductions in carrier density do not account for the magnitude of the reduction in forward current and increase in on-state resistance, showing that carrier mobility reductions are equally important in the change of device electrical performance. Subsequent annealing at 400 • C brought a partial restoration in carrier densities, to 2.1 × 10 16 and 1.7 × 10 16 cm −3 , respectively, for the two fluences.
The carrier removal rate per ion was then obtained from the change in carrier density divided by the fluence and is included in the compilation plot of figure 8, which shows the published values for different polymorphs of Ga 2 O 3 irradiated with different forms of radiation [26,27]. Since the NIEL is approximately constant over the drift layer thickness, this provides a reasonable estimate the carrier removal rate. The values obtained here for the NiO/Ga 2 O 3 , 1200 for 10 13 cm −2 and 190 cm −1 for 10 14 cm −2 are consistent with values reported for proton irradiation at similar energies/fluences but in Schottky diode structures that did not include the NiO [26,27]. This indicates that in contrast to gamma irradiation [29], the heterojunction rectifiers do not show any significant differences with the presence of the NiO for proton damage. The value determined at the higher fluence may already be in the saturation region where most of the carriers are already trapped at damage sites and therefore the value at low fluence is likely to be closer to the true number.
The fact that proton damage does not produce the hysteresis in I-V characteristics seen for gamma irradiation in these structures indicates that the reversible resistivity changes seen in other irradiated oxides is not occurring. For example, Borgersen et al [30,32] reported that irradiation of the related oxide, In 2 O 3 , with low doses of Si ions or and long UV exposures result in similar resistivity drops, interpreted as irradiation/illumination assisted desorption of oxygen containing species from the surface. This was consistent with the effect of post-irradiation exposure of the samples to an oxygen atmosphere partially  restoring the resistivity. Swallow et al [33] reported that Ga 2 O 3 surfaces are terminated by O-H groups, producing downward band bending and electron accumulation. Annealing to remove these hydroxyl groups converts the surface to depletion with upward band bending of 0.26 eV. In our case, the fluences of protons used here does not produce changes in the interfacial properties that affect the electrical performance of the heterojunction rectifiers.

Summary and conclusions
In summary, the main effects of proton irradiation on vertical geometry NiO/β-Ga 2 O 3 heterojunction rectifiers were found to be a reduction in reverse breakdown voltage 4.3-3.7 kV for a fluence of 10 13 cm −2 and to 1.93 kV for a fluence of 10 14 cm −2 . The forward current density also decreased significantly, resulting in an increase in the on-state resistance (R ON ) of the rectifiers. These were due to a reduction in carrier density and mobility in the drift layer. Additionally, the reverse leakage current increased by approximately a factor of 2 for the higher fluence. Subsequent annealing at temperatures up to 400 • C further increased the reverse leakage current due to deterioration of the contacts. However, the initial carrier density was almost fully restored by this annealing process in the lower fluence samples and by more than 50% in the 10 14 cm −2 irradiated devices. The carrier removal rates in the Ga 2 O 3 layer were found to be in the range of 190-1200 for the employed fluence range, which is similar to Schottky rectifiers without the NiO layer.

Data availability statements
All data that support the findings of this study are included within the article (and any supplementary files).