betaFeedback

WorksPhysics and Astronomy

Reliable operation of Cr2_2O3_3:Mg/ ββ-Ga2_2O3_3 p-n heterojunction diodes at 600∘^\circC

Abstract

ββ-Ga2_2O3_3-based semiconductor heterojunctions have recently demonstrated improved performance at high voltages and elevated temperatures and are thus promising for applications in power electronic devices and harsh-environment sensors. However, the long-term reliability of these ultra-wide band gap (UWBG) semiconductor devices remains barely addressed and may be strongly influenced by chemical reactions at the p-n heterojunction interface. Here, we experimentally demonstrate operation and evaluate the reliability of Cr2_2O3_3:Mg/ ββ-Ga2_2O3_3 p-n heterojunction diodes at during extended operation at 600∘^\circC, as well as after 30 repeated cycles between 25-550∘^\circC. The calculated pO2-temperature phase stability diagram of the Ga-Cr-O material system predicts that Ga2_2O3_3 and Cr2_2O3_3 should remain thermodynamically stable in contact with each other over a wide range of oxygen pressures and operating temperatures. The fabricated Cr2_2O3_3:Mg / ββ-Ga2_2O3_3 p-n heterojunction diodes show room-temperature on/off ratios >104^4 at ±\pm5V and a breakdown voltage (VBr_{Br}) of -390V. The leakage current increases with increasing temperature up to 600∘^\circC, which is attributed to Poole-Frenkel emission with a trap barrier height of 0.19 eV. Over the course of a 140-hour thermal soak at 600∘^\circC, both the device turn-on voltage and on-state resistance increase from 1.08V and 5.34 mΩΩ-cm2^2 to 1.59V and 7.1 mΩΩ-cm2^2 respectively. This increase is attributed to the accumulation of Mg and MgO at the Cr2_2O3_3/Ga2_2O3_3 interface as observed from TOF-SIMS analysis. These findings inform future design strategies of UWBG semiconductor devices for harsh environment operation and underscore the need for further reliability assessments for ββ-Ga2_2O3_3 based devices.

Beta gallium oxide (β\beta-Ga2_2O3_3) is a prime candidate for next-generation sensing and power electronic devices, especially for high-temperature applications. Due to its large bandgap and high theoretical breakdown field, β\beta-Ga2_2O3_3 shows promise for high-voltage operation and high-frequency switching. Additionally, β\beta-Ga2_2O3_3 can be doped n-type to relatively high levels, most commonly with Si and Sn. Single-crystal β\beta-Ga2_2O3_3 can be grown via multiple bulk methods, including Czochralski (CZ) and edge-defined film-fed growth (EFG). Techno-economic analysis and projective cost modeling suggests that further development of Ga2_2O3_3-based technologies will improve their commercial viability in competition with current SiC and GaN technologies[1, 2, 3, 4, 5, 6].

Over the last few years significant progress has been made in expanding the capabilities of β\beta-Ga2_2O3_3 devices towards high-performance operation, with several groups making efforts to reach the predicted theoretical high-power limits via maximization of breakdown voltage and simultaneous minimization of on-state resistance. This has been achieved through a shift from relatively simple Schottky/heterojunction device structures towards more complex stacks that incorporate field plates, guard rings, and dielectric layers[7]. However, there are few operational demonstrations of these devices at high temperature, specifically above 300-400 °C, and even fewer that study performance stability[8]. Previously, we found that a combination of repeated cycling between low and high temperatures (“cycling”), as well as long-term exposure to very high temperatures (“soaking”) provides insight into both mechanical and thermodynamic stressors that can degrade device electrical performance[9, 10].

A variety of p-type materials have been explored for use in β\beta-Ga2_2O3_3 device fabrication, including SiC, GaN, NiO, diamond, SnO, and Cu2_2O[11, 12, 13, 14, 15, 16]. Nickel oxide (NiO) is the most widely explored due to its wide bandgap and favorable band alignment with β\beta-Ga2_2O3_3, achieving turn-on voltages in the range of 0.9-1.8V and reliably low leakage currents during room-temperature operation. Furthermore, very high breakdown voltages and excellent corresponding Baliga figures of merit have been reported for NiO/β\beta-Ga2_2O3_3 heterojunctions[17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27]. Of the candidate p-type metal oxides for use in β\beta-Ga2_2O3_3 device fabrication, several (e.g., Cu2_2O, NiO) are predicted to form ternary compounds when in contact with Ga2_2O3_3 based on computational thermodynamics[28]. Here we report β\beta-Ga2_2O3_3 heterojunction diodes using the wide bandgap p-type chromium oxide (Cr2_2O3_3), for which there is virtually no literature. However, recent XPS work on Ga2_2O3_3/Cr2_2O3_3 suggest favorable band alignment (ΔEC\Delta E_\mathrm{C} = 1.68 eV; ΔEV\Delta E_\mathrm{V} = 3.38 eV) for heterojunction device fabrication[29, 30].

In this work, epitaxial Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 vertical pn-diodes are fabricated, and their electrical performance and thermal reliability are evaluated. We demonstrate the first example of high-performance Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 heterojunction with a room-temperature rectification ratio >104>10^{4} at ±5\pm5V, and breakdown voltage of −390-390V. We then subject several Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 vertical pn-diodes to repeated thermal cycling between 2525–550 ∘C550\,^{\circ}\mathrm{C}, as well as separately to a 140-hour thermal soak at 600 ∘C600\,^{\circ}\mathrm{C}. We find that the devices show exceptionally stable reverse bias leakage current during prolonged high-temperature testing, but experience forward-bias changes in the form of increased on-state resistance, turn-on voltage, and forward current drop. The reverse bias leakage current can be described by Poole-Frenkel emission with a trap barrier height of 0.190.19 eV. We attribute the observed evolution in the forward bias electrical properties to the accumulation of Mg and MgO at the Cr2_2O3_3/Ga2_2O3_3 interface after prolonged operation at 600 ∘C600\,^{\circ}\mathrm{C} due to the temperature induced diffusion of the Mg dopants in Cr2_2O3_3. This study provides providing new insights and design considerations for oxide-based UWBG semiconductors for p-n heterojunctions applicable in harsh operating environments.

Figure 1(a) shows the calculated oxygen partial pressure vs. temperature (pO2p\mathrm{O}_2-TT) phase diagrams for the Ga-Cr-O system. The “Experimental Range” box includes the temperature and pO2p\mathrm{O}_2 conditions used in this report, which are limited by instrumentation. Cr2O3\mathrm{Cr}_2\mathrm{O}_3 and Ga2O3\mathrm{Ga}_2\mathrm{O}_3 are predicted to coexist for effectively all desirable temperatures and oxygen partial pressures. This pO2p\mathrm{O}_2-TT phase diagram for the Ga-Cr-O system was calculated using density functional theory (DFT) datasets from NREL MatDB[31, 32] along with correction schemes[33, 34] to achieve chemical accuracy (∼50\sim50 meV/atom). Specifically, we estimate the Gibbs formation energy of a compound, ΔGf(T)\Delta G_f(T), in units of eV/atom by,

ΔGf(T)=ΔHf(298.15K)+GSISSOδ(T)−∑iNαiGi,exp(T)(1)\Delta G_f(T) = \Delta H_f(298.15\mathrm{K}) + G^\delta_{\mathrm{SISSO}}(T) - \sum_{i}^{N} \alpha_i G_{i,\mathrm{exp}}(T) \tag*{(1)}

where ΔHf(298.15K)\Delta H_f(298.15\mathrm{K}) is the standard formation enthalpy of the compound and is calculated using DFT total energies from the NREL MatDB and fitted elemental-phase reference energies. Here, αi\alpha_i is the stoichiometric weight for element ii in the compound using experimental values from FactSage[35] for absolute Gibbs free energy of element ii. GSISSOδ(T)G^\delta_{\mathrm{SISSO}}(T) (eV/atom) captures temperature dependency of the compound due to phonons[34]. Lastly, we collect ΔGf(T)\Delta G_f(T) for all competing phases and use convex hull analysis to calculate the grand potential phase diagram. The ideal gas law, e.g., ΔμO=12kBTln⁡(pO2)\Delta\mu_{\mathrm{O}} = \frac{1}{2} k_B T \ln(p\mathrm{O}_2), is used to connect oxygen chemical potential (μO\mu_{\mathrm{O}}) to oxygen partial pressure (pO2p\mathrm{O}_2) in units of atm, where ΔμO\Delta\mu_{\mathrm{O}} is the deviation from the standard oxygen chemical potential (μO∘\mu_{\mathrm{O}}^\circ) with μO=μO∘+ΔμO\mu_{\mathrm{O}} = \mu_{\mathrm{O}}^\circ+ \Delta\mu_{\mathrm{O}}[36]. A full description of these calculations is available online[37].

Cr2O3\mathrm{Cr}_2\mathrm{O}_3:Mg was deposited via pulsed laser deposition (PLD). During PLD growth, a ceramic target of 8at%Mg-doped Cr2O3\mathrm{Cr}_2\mathrm{O}_3 was ablated using a pulsed KrF excimer laser at a frequency of 10Hz and energy of 300mJ. Growth was performed at a substrate temperature of 600°C and O2\mathrm{O}_2 partial pressure of 3×10−43 \times10^{-4} Torr. To fabricate Ohmic contacts (Fig S1), Cr2O3\mathrm{Cr}_2\mathrm{O}_3:Mg was grown on a c-sapphire substrate, followed by a bi-metal deposition of 5 nm Ti/ 100 nm Au. E-beam evaporation of metals was performed with a Temescal FC2000 evaporation system under high vacuum conditions (≤3×10−6\leq3 \times10^{-6} Torr) without venting between layers. Room temperature Hall measurements were performed with a Lakeshore FastHall system. The Cr2O3\mathrm{Cr}_2\mathrm{O}_3:Mg films were determined to be p-type with a carrier concentration of 2.4×10172.4 \times10^{17} cm−3^{-3}; a resistivity of 31.5 Ω\Omega cm, which is in good agreement with reports on Mg-doped Cr2O3\mathrm{Cr}_2\mathrm{O}_3[38, 39]; and a mobility of 0.820.82 cm2^2 V−1^{-1}s−1^{-1}, which can be significantly improved with choice of dopant[40]. High-temperature Hall measurements were also attempted using a custom instrument[41], but the results were inconclusive. To electrically test the Ohmic contacts to Cr2O3\mathrm{Cr}_2\mathrm{O}_3:Mg, two-probe measurements (source I, measure V) were performed via a Keithley 236 SMU in a custom McAllister probe station in ambient air. IV curves were collected every 15 minutes. To initiate the long-term thermal soak, the probe station was ramped to 600°C at a rate of 100°C per hour. Over the course of the one week testing period at 600°C, the resistance remains effectively constant, and the IV curves remain linear. Electrical measurements reveal that Cr2O3:Mg\mathrm{Cr_2O_3:Mg} is a highly resistive semiconductor, especially at lower temperatures. Full results and figures for the Ohmic contact testing can be found in the Supplementary Material Fig. S1.

Figure 1(b) gives a 2θ2\theta-ω\omega scan of a 120 nm Cr2O3:Mg\mathrm{Cr_2O_3:Mg} layer grown on (001) Si-doped HVPE Ga2O3\mathrm{Ga_2O_3} substrate, performed using a Bruker 8 Discover diffractometer. In addition to the (001) Ga2O3\mathrm{Ga_2O_3} planes, there are also peaks that indicate the presence of a (-201) Ga2O3\mathrm{Ga_2O_3} sublattice within the HVPE layer. In the region from 36 to 42 degrees, the (006) plane of Cr2O3\mathrm{Cr_2O_3} can be observed and its proximity to the (-402) plane of Ga2O3\mathrm{Ga_2O_3} suggests preferential growth on this sublattice. This relationship is further supported by Figure S4, which shows a similar 2θ2\theta-ω\omega scan of Cr2O3:Mg\mathrm{Cr_2O_3:Mg} layer preferentially growing in the (001) orientation on a reference (-201) β\beta-Ga2O3\mathrm{Ga_2O_3} substrate. Figure 1(c) and Figure 1(d) give a top-down and profile visualization of the stacked (001) planes for both materials in the Cr2O3\mathrm{Cr_2O_3}-Ga2O3\mathrm{Ga_2O_3} relationship, respectively.

Phase stability diagram, X-ray diffraction scan, and structural views of the Cr2O3–Ga2O3 interface

Figure 1. (a) Calculation showing pO2_2-temperature phase stability diagram for the Ga-Cr-O system. The “Experimental Range” box highlights the expected operating conditions for devices made from these materials, including the conditions used in this report. (b) 2θ2\theta-ω\omega scan of a 120 nm Cr2_2O3_3:Mg layer grown on (001) HVPE Ga2_2O3_3 substrate on a log scale. The region 36–42 deg shows both the (2ˉ\bar{2}01) sublattice of the HVPE layer, as well as the preferential growth of (006) Cr2_2O3_3 on that sublattice. (c) Top-down and (d) profile views of the (001) Cr2_2O3_3 – (001) Ga2_2O3_3 interface.

Cr2O3:Mg/β\mathrm{Cr_2O_3:Mg}/\beta-Ga2O3\mathrm{Ga_2O_3} p-n vertical heterojunction devices were fabricated using 1 μm1\ \mu\mathrm{m} thick β\beta-Ga2O3\mathrm{Ga_2O_3} drift layers (n≅4×1016 cm−3\mathrm{n} \cong4 \times10^{16}\ \mathrm{cm}^{-3}) grown via HVPE on (001) Ga2O3:Sn\mathrm{Ga_2O_3:Sn} from Novel Crystal Technologies. Photoresist was removed from the as-delivered substrates via an organic wash followed by a sulfuric acid/peroxide rinse. Next, large-area back Ohmic contacts (5 nm Ti / 100 nm Au) and the p-Cr2O3:Mg\mathrm{Cr_2O_3:Mg} layer were deposited via e-beam and PLD, respectively, via the methods previously described. Two different geometries were used for the top Ohmic contacts to Cr2O3:Mg\mathrm{Cr_2O_3:Mg}. For the thermal cycling devices, shadow masks were used to define an array of 1 mm diameter pads that were sufficiently large to ensure reliable probe contact despite repeated thermal expansion and contraction. To fabricate devices for the long-term thermal soak, contact aligner lithography was used to make circular pads of increasing size, from 50-300 μm\mu\mathrm{m} in diameter. In both cases, the 5 nm Ti/100 nm Au contacts were deposited as before. Following contact deposition, both Ohmic and heterojunction samples were annealed a single time via rapid thermal processing (ULVAC-RIKO MILA-3000 Rapid Thermal Processing Unit) at 550°C for 90s in flowing N2\mathrm{N_2}.

Figure 2 gives the room-temperature electrical characteristics of a 300 μ\mum diameter Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 p-n diode under 50 sccm flowing Ar. Figure 2(a) shows the entire JVJV characteristic on a log scale, measured from +5 to −10 V. A rectification ratio greater than 10410^4 is achieved at ±5 V. The diode reaches a turn-on current density of 1 A/cm2^2 at approximately 1.65 V and fitting of the “On” portion of the curve on a linear scale (Fig. S2) yields an RonR_{\mathrm{on}} of 12.96 mΩ cm2^2. The inset shows the device p-n heterojunction architecture for devices studied in this report. Details about doping levels and layer thicknesses can be found in Fig. S2. Figure 2(b) gives a reverse bias breakdown voltage of 0.39 kV. Room-temperature breakdown voltage was measured using a Keysight B1505A Power Device Analyzer with a reverse current limit of 1 mA. This device is among the first demonstrations of a Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 p-n diode; significant further improvements can be expected based on architectural enhancements to control the electric field and optimize the growth conditions.

Room-temperature JV characteristics and reverse bias breakdown performance of a Cr$_2$O$_3$:Mg/$\beta$-Ga$_2$O$_3$ p-n heterojunction diode

Figure 2. (a) Room temperature IV characteristics for 300 μ\mum Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 p-n heterojunction diode, shown in forward-bias on a linear scale. Inset shows the device architecture of the p-n heterojunctions used in this report. (b) Room-temperature reverse bias breakdown performance.

Both thermal cycling and soaking of the heterojunction diodes were performed with a Keithley 2635A SMU in an Instec HCP621G-PMH temperature-controlled probe station. Thermal cycling consists of repeated ramping from 25–550°C and back down in steps of 75°C. The instrument was allowed to equilibrate for 15–20 minutes at each temperature before electrical measurements were collected. Thermal soaking was performed in a similar fashion; the instrument was manually ramped from 25°C to 600°C in steps of 75–100°C. The instrument was allowed to equilibrate for 20–30 minutes at each temperature, after which electrical measurements were collected. Upon reaching 600°C, electrical measurements were performed every 30 minutes for the full duration of the soaking experiment. For thermal cycling, IVIV curves were collected from 0 → +5 V, followed by a reverse bias sweep from 0 → −20 V.

Compliance was set to 0.1A0.1\mathrm{A} for all measurements. For thermal soaking electrical measurements, IV curves were collected from 0→+5V0 \to+5\mathrm{V} with a 0.2A0.2\mathrm{A} current limit, followed by a reverse bias sweep from 0→−10V0 \to-10\mathrm{V} with a 0.1A0.1\mathrm{A} current limit.

Figure 3 shows the high-temperature reliability of several Cr₂O₃:Mg/β\beta-Ga₂O₃ pn diodes under both thermal soaking and cycling conditions. Figure 3(a) shows results of a >100-hour thermal soak at 600∘C600^{\circ}\mathrm{C} for a 300 μm300\,\mu\mathrm{m} Cr₂O₃:Mg/β\beta-Ga₂O₃ pn diode under 50 sccm flowing Ar. JV curves were measured every 30 minutes, with traces plotted every 10 hours to demonstrate the time evolution. A line was fit to the “On” portion of Figure 3(a) from 100 to 250 A/cm2^{2} and extrapolated to determine the turn-on voltage and specific on-state resistance. After a break-in period of 10 hours, VonV_{\mathrm{on}} steadily increases from 1.08V to 1.59V after 140 hours, at a rate of approximately 4 mV per hour. Correspondingly, the value of RonR_{\mathrm{on}} also increases from 5.34 mΩ\Omega cm2^{2} to 7.1 mΩ\Omega cm2^{2} over the same period but appears to grow with the square root of time. The evolution of reverse bias leakage currents was also examined and found to stabilize at all voltages after a break-in period of ∼\sim30 hours. Plots of RonR_{\mathrm{on}}, VonV_{\mathrm{on}}, and JrevJ_{\mathrm{rev}} vs. time can be found in Fig. S2(c-e). Fig S2(f) shows the forward-bias JVJV curves for the same 300 μ\mum Cr2_{2}O3_{3}:Mg/β\beta-Ga2_{2}O3_{3} pad taken before (“Pre”, solid lines) and after (“Post”, dashed lines) the 140-hour soak at 600 °C, measured out to +5V. Significant changes in performance can be observed, especially at lower temperatures. Room-temperature VonV_{\mathrm{on}} and RonR_{\mathrm{on}} increase from 1.65V and 12.96 mΩ\Omega cm2^{2} to 3.84V and 312 mΩ\Omega cm2^{2}, respectively.

Four-panel plots showing high-temperature reliability measurements of Cr₂O₃:Mg/$\beta$-Ga₂O₃ p-n diodes

Figure 3. (a) Linear JV curves of a 300 μm300\,\mu\mathrm{m} diameter Cr₂O₃:Mg/β\beta-Ga₂O₃ p-n diode operating at 600∘C600^{\circ}\mathrm{C} for 140 hours, with traces plotted every 10 hours. The current limits were set to 200 mA in forward bias and 100 mA in reverse bias to prevent damage during operation. (b) normalized Mg signal from TOF-SIMS for the thermally soaked diode and an as-deposited reference device. The Ga₂O₃/Cr₂O₃ interface is indicated by the blue rectangle. The signals have been normalized relative to their baseline signal taken from between 20–80 nm in the Cr₂O₃ layers. (c) Forward bias JV curves of a 1 mm diameter Cr₂O₃:Mg/β\beta-Ga₂O₃ pn diode for three select temperatures before and after repeated thermal cycling. (d) Evolution of temperature dependent leakage current with repeated cycling, measured at -20V.

To understand the cause of the increase in forward drop, both the thermally soaked (“test”) sample and an as-deposited (“reference”) sample were studied using time-of-flight secondary ion mass spectrometry (TOF-SIMS) analysis. Results are shown in Figure 3(b) and Fig S5. The test sample reveals a significant amount of Mg dopant diffusing to the Cr2_{2}O3_{3}/Ga2_{2}O3_{3} interface after thermal soaking at 600°C. Beyond this, there appears to be no interdiffusion of any other species between the p-type and n-type layers – the interface remains abrupt, which supports our theoretical calculations. The normalized Mg and MgO signals at the interface from the test and reference sample are also compared. Both signals at the Cr2_{2}O3_{3}/Ga2_{2}O3_{3} interface in the test sample are significantly higher than in the reference sample. While there is evidence of a small amount of Mg accumulation in the reference sample, very little of it appears to have oxidized. This diffusion can be attributed to the short-term high-temperature growth conditions experienced during PLD. MgO is a well-known wide bandgap material that is frequently used for its dielectric strength46. Given the band alignment of Ga2_{2}O3_{3} with MgO and Cr2_{2}O3_{3}[29] [47], this observed reaction amounts to inserting a thin dielectric layer between the p-type and n-type layers. Assuming the thickness of this MgO layer is increasing with time at temperature, this potentially explains the gradual increase in both forward bias resistance and turn-on voltage observed in the electrical measurements. The reverse bias leakage current stabilizes after 20-40 hours, which suggests that the mechanism for current transport has a relatively low energy barrier and isn’t strongly dependent on the thickness of the Mg/MgO layer. While the final recorded turn-on voltage at 600°C approaches the band offset of Cr2_{2}O3_{3} and Ga2_{2}O3_{3} measured with XPS[29, 30], it’s more likely that this attributable to the MgO layer. Temperature-dependent XPS measurements are needed to better understand this relationship.

A separate device utilizing 1 mm diameter pads was subject to repeated thermal cycling between 25-550°C, with current compliance set to 100mA. Forward bias curves were measured out to 5V; reverse bias was measured to -20V. Figure 3(c) shows the forward-bias JV-T curves for cycle 5 and cycle 30 for three different temperatures, showing slight increases in VonV_{on} with negligible changes in RonR_{on}. Figure 3(d) shows the evolution of leakage current with increasing cycle, as measured at -20V. With a few exceptions at lower temperature, the leakage current appears to stabilize after ~20 cycles.

Overall, Figure 3 demonstrates the different effects of thermal cycling and thermal soaking on the performance of electronic devices. While both impart a significant thermal load, the 140-hour soak at 600°C appears to have greater impact on device recovery, especially at lower temperatures. In contrast, the changes in electrical behavior between cycles 5 and 30 in the temperature cycling test are significantly less severe. In both experiments, the reverse bias leakage current stabilizes after a somewhat significant break-in period (~30 hours or 20 cycles, respectively); the forward bias current appears to be more affected in both cases. Another consideration is the size of the pad being measured. Given the poor thermal conductivity of β\beta-Ga2_2O3_3, it is possible that device performance degradation can be attributed to substantial Joule heating. Since there is a greater than 10-fold increase in the area of the thermally cycled devices when compared to those that were soaked, the generated heat can be more evenly distributed across the contact area.

To understand the corresponding leakage current mechanism, we examined reverse bias J-V-T measurements from cycle 30 (Figure 4(a)). The results show that reverse bias current has a strong dependence on both temperature and voltage. To analyze these results, we use the Poole-Frenkel emission (PFE) model, which describes conduction via thermal excitation of electrons. from trap states and is typically present in insulating layers. PFE can expressed by the following equation:

Reverse bias J-V curves and Poole-Frenkel plots for cycle 30

Figure 4. (a) Reverse bias JV curves from cycle 30, showing dependences on both temperature and voltage. (b) Poole-Frenkel plots of leakage current for cycle 30, with linear fits to each temperature trace shown in dotted black lines.

J0∝Eexp⁡(−q(ϕF−qEπεsε0)kBT)(2)J_0 \propto E \exp\left(-\frac{q\left(\phi_F-\sqrt{\frac{qE}{\pi\varepsilon_s\varepsilon_0}}\right)}{k_BT}\right) \tag*{(2)}

Here, EE is the electric field across the depletion region, ϕF\phi_F is the barrier height of the trap state, εs\varepsilon_s is the relative dielectric constant, and kBk_B is Boltzmann’s constant. To extract the barrier height of these trap states, graphs of ln⁡(JE)\ln\left(\frac{J}{E}\right) vs. E\sqrt{E} are first plotted for each temperature, showing excellent linear fits (Figure 4(b)). Subsequently, the intercepts of these linear fits are plotted against q/kBTq/k_BT (Fig S3), yielding trap barrier heights of 0.191 eV[42, 43, 44]. This relatively low barrier height is smaller than what has been reported for NiO/β\beta-Ga2_2O3_3[45], which results in the leakage current of the Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 diodes that is several orders of magnitude higher than NiO/Ga2_2O3_3 heterojunctions.

The J-V-T results and analysis presented in Figure 4 indicate that the cycling and soaking reliability trends presented in Figure 3 may be related to non-ideal interfaces in the overall Au/Ti/Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3/Ti/Au multi-layer device stack (Figure 1a, inset). The contribution of Au/Ti/Cr2_2O3_3:Mg or β\beta-Ga2_2O3_3/Ti/Au Ohmic contacts to this device degradation can be ruled out. This is because the independent Au/Ti/Cr2_2O3_3:Mg/Al2_2O3_3 Ohmic contacts stability study (see Fig. S1) showed that they remain stable for a longer time than the device as a whole. Our earlier published results on β\beta-Ga2_2O3_3 showed stable Ohmic behavior after >500 hours of continuous operation at 600°C[9]. Another potential interface where degradation can occur is Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3. As shown in Figure 1(b), there is a quasi-ordered relationship between the Cr2_2O3_3 layer and the (-201) character of the (001)-oriented HVPE Ga2_2O3_3 substrate. The lack of direct epitaxy can result in a lower interface quality, which can lead to a high density of interfacial trap states and thereby facilitate reverse-bias conduction. Furthermore, this might be exacerbated by the accumulation of Mg and MgO at this interface. However, this is unlikely because similar lack of ideal epitaxy exists at the NiO/β\beta-Ga2_2O3_3 interface which shows higher barrier extracted from the J-V-T measurement using PFE model[45]. It is worth noting that based on the minimal changes to forward bias conduction and the shorter cumulative exposure time to high temperatures, we do not anticipate as much Mg/MgO migration in the thermally cycled samples as in the thermally soaked samples.

In this work, we demonstrate reliable operation of Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 p-n heterojunction diodes at 600°C. This experimental work is motivated by theoretical calculations that predict thermodynamically stable phase coexistence of n-Ga2_2O3_3 and p-Cr2_2O3_3 for an extremely wide range of temperatures and pO2_2. We fabricate some of the first reported Cr2_2O3_3:Mg/β\beta-Ga2_2O3_3 vertical p-n heterojunction diodes with VonV_{on}, RonR_{on}, and VbrV_{br} of 1.65 V, 12.96 mΩ\Omega cm2^2, and -390 V, respectively at room-temperature. We subject several diodes to repeated thermal cycling between 25-550 °C and long-term continuous operation at 600 °C for 140 hours. Analysis of the reverse bias current shows that the leakage mechanism is very well described by Poole-Frenkel emission. We find that prolonged thermal soaking has more deleterious effects on electrical performance than repeated cycling, leading to increased turn-on voltage and on-state resistance. Since we observed stability of the ultrathin Ti/Au Ohmic contact to Cr2O3\mathrm{Cr_2O_3}:Mg/cc-Al2O3\mathrm{Al_2O_3} as well as Ga2O3\mathrm{Ga_2O_3}/Ti/Au contacts under the same high-temperature operating conditions, we attribute this performance evolution to the migration of Mg to the Cr2O3\mathrm{Cr_2O_3}/Ga2O3\mathrm{Ga_2O_3} interface, where SIMS reveals the presence of a thin MgO layer. Future work should focus on understanding these degradation phenomena and their temperature dependence, as well as comparisons to other material systems under similar conditions.

Supplementary Material Description

The supplementary material contains:

  • S1) Device architecture for Ohmic contact devices. Temperature-dependent performance of Ohmic contact.

  • S2) Room-temperature and high-temperature JV curves of Cr2O3\mathrm{Cr}_{2}\mathrm{O}_{3}:Mg/β\beta-Ga2O3\mathrm{Ga}_{2}\mathrm{O}_{3} p-n devices, as well as extracted parameters from a 140-hour soak at 600 ∘C600\,^\circ\mathrm{C}.

  • S3) Poole-Frenkel barrier height extraction.

  • S4) XRD of (001) Cr2O3\mathrm{Cr}_{2}\mathrm{O}_{3}:Mg / (-201) β\beta-Ga2O3\mathrm{Ga}_{2}\mathrm{O}_{3}.

  • S5) TOF-SIMS results for thermally soaked and reference Cr2O3\mathrm{Cr}_{2}\mathrm{O}_{3}:Mg / β\beta-Ga2O3\mathrm{Ga}_{2}\mathrm{O}_{3} samples.

Author Declarations

The authors have no conflicts to declare.

Data Availability

The data supporting the findings of this study are available within the paper and its Supplementary Material. Any additional data connected to the study are available from the corresponding author upon reasonable request.

Acknowledgements

This work was authored by the National Renewable Energy Laboratory (NREL), operated by Alliance for Sustainable Energy, LLC, for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. Funding is provided by the Office of Energy Efficiency and Renewable Energy (EERE) Advanced Manufacturing Office. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. This material makes use of the TOF-SIMS system at the Colorado School of Mines, which was supported by the National Science Foundation under Grant No.1726898. We would especially like to thank V.S. for his contributions to the computational methods used in this report; C.E.P for her assistance in high-temperature transport measurements; and M.A.W for his analysis with the TOF-SIMS system.

Author Contributions

William A. Callahan – Conceptualization (equal), Data Curation (lead), Formal Analysis (lead), Methodology (lead), Software (equal), Validation (lead), Visualization (equal), Writing/Original Draft Preparation (lead).

Kingsley Egbo– Conceptualization (equal), Resources (equal), Writing – review & editing (equal)

Cheng-Wei Lee– Software (equal), Visualization (equal), Writing – review & editing (equal)

Ryan O’Hayre: Funding acquisition (equal), Writing – review & editing (equal), Supervision (equal)

David Ginley: Funding acquisition (equal), Writing – review & editing (equal), Supervision (equal)

Andriy Zakutayev: Project administration (lead), Supervision (equal), Funding acquisition (equal), Writing – review & editing (equal)

References

  1. [1]S.J. Pearton, J. Yang, P.H. Cary, F. Ren, J. Kim, M.J. Tadjer, and M.A. Mastro, “A review of Ga2O3 materials, processing, and devices,” Applied Physics Reviews 5(1), 011301 (2018).DOI
  2. [2]Z. Galazka, “β-Ga2O3 for wide-bandgap electronics and optoelectronics,” Semicond. Sci. Technol. 33(11), 113001 (2018).
  3. [3]C. Wang, J. Zhang, S. Xu, C. Zhang, Q. Feng, Y. Zhang, J. Ning, S. Zhao, H. Zhou, and Y. Hao, “Progress in state-of-the-art technologies of Ga2O3 devices,” J. Phys. D: Appl. Phys. 54(24), 243001 (2021).DOI
  4. [4]S. Reese, and A. Zakutayev, in Oxide-Based Materials and Devices XI, edited by F.H. Teherani, D.C. Look, and D.J. Rogers (SPIE, San Francisco, United States, 2020), p. 87.
  5. [5]K.N. Heinselman, D. Haven, A. Zakutayev, and S.B. Reese, “Projected Cost of Gallium Oxide Wafers from Edge-Defined Film-Fed Crystal Growth,” Crystal Growth & Design 22(8), 4854–4863 (2022).DOI
  6. [6]E.M. Garrity, C.-W. Lee, P. Gorai, M.B. Tellekamp, A. Zakutayev, and V. Stevanović, “Computational Identification of Ternary Wide-Band-Gap Oxides for High-Power Electronics,” PRX Energy 1(3), 033006 (2022).DOI
  7. [7]C. Venkata Prasad, and Y.S. Rim, “Review on interface engineering of low leakage current and on-resistance for high-efficiency Ga2O3-based power devices,” Materials Today Physics 27, 100777 (2022).DOI
  8. [8]A.E. Islam, N.P. Sepelak, K.J. Liddy, R. Kahler, D.M. Dryden, J. Williams, H. Lee, K. Gann, A. Popp, K.D. Leedy, N.S. Hendricks, Jeff.L. Brown, E.R. Heller, W. Wang, W. Zhu, M.O. Thompson, K.D. Chabak, and A.J. Green, “500 °C operation of β-Ga2O3 field-effect transistors,” Applied Physics Letters 121(24), 243501 (2022).DOI
  9. [9]W.A. Callahan, E. Supple, D. Ginley, M. Sanders, B.P. Gorman, R. O’Hayre, and A. Zakutayev, “Ultrathin stable Ohmic contacts for high-temperature operation of β-Ga2O3 devices,” Journal of Vacuum Science & Technology A 41(4), 043211 (2023).DOI
  10. [10]K. Heinselman, P. Walker, A. Norman, P. Parilla, D. Ginley, and A. Zakutayev, “Performance and reliability of β-Ga2O3 Schottky barrier diodes at high temperature,” Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 39(4), 040402 (2021).DOI
  11. [11]T. Watahiki, Y. Yuda, A. Furukawa, M. Yamamuka, Y. Takiguchi, and S. Miyajima, “Heterojunction p-Cu2O/n-Ga2O3 diode with high breakdown voltage,” Applied Physics Letters 111(22), 222104 (2017).DOI
  12. [12]Y. Zhang, Y. Li, Z. Wang, R. Guo, S. Xu, C. Liu, S. Zhao, J. Zhang, and Y. Hao, “Investigation of β-Ga2O3 films and β-Ga2O3/GaN heterostructures grown by metal organic chemical vapor deposition,” Sci. China Phys. Mech. Astron. 63(11), 117311 (2020).DOI
  13. [13]P. Sittimart, S. Ohmagari, T. Matsumae, H. Umezawa, and T. Yoshitake, “Diamond/β-Ga2O3 pn heterojunction diodes fabricated by low-temperature direct-bonding,” AIP Advances 11(10), 105114 (2021).DOI
  14. [14]S. Nakagomi, K. Hiratsuka, Y. Kakuda, and K. Yoshihiro, “Beta-Gallium Oxide/SiC Heterojunction Diodes with High Rectification Ratios,” ECS J. Solid State Sci. Technol. 6(2), Q3030–Q3035 (2017).DOI
  15. [15]¹⁵ T. Minami, Y. Nishi, and T. Miyata, “High-Efficiency Cu₂O-Based Heterojunction Solar Cells Fabricated Using a Ga₂O₃ Thin Film as N-Type Layer,” Appl. Phys. Express 6(4), 044101 (2013).
  16. [16]¹⁶ M. Budde, D. Splith, P. Mazzolini, A. Tahraoui, J. Feldl, M. Ramsteiner, H. Von Wenckstern, M. Grundmann, and O. Bierwagen, “SnO/β-Ga₂O₃ vertical pn heterojunction diodes,” Applied Physics Letters 117(25), 252106 (2020).DOI
  17. [17]¹⁷ J. Zhang, S. Han, M. Cui, X. Xu, W. Li, H. Xu, C. Jin, M. Gu, L. Chen, and K.H.L. Zhang, “Fabrication and Interfacial Electronic Structure of Wide Bandgap NiO and Ga₂O₃ p–n Heterojunction,” ACS Appl. Electron. Mater. 2(2), 456–463 (2020).DOI
  18. [18]¹⁸ X. Xia, J.-S. Li, C.-C. Chiang, T.J. Yoo, F. Ren, H. Kim, and S.J. Pearton, “Annealing temperature dependence of band alignment of NiO/β-Ga₂O₃,” J. Phys. D: Appl. Phys. 55(38), 385105 (2022).DOI
  19. [19]¹⁹ Y. Wang, H. Gong, Y. Lv, X. Fu, S. Dun, T. Han, H. Liu, X. Zhou, S. Liang, J. Ye, R. Zhang, A. Bu, S. Cai, and Z. Feng, “2.41 kV Vertical p-NiO/n-Ga₂O₃ Heterojunction Diodes With a Record Baliga’s Figure-of-Merit of 5.18 GW/cm²,” IEEE Trans. Power Electron. 37(4), 3743–3746 (2022).DOI
  20. [20]²⁰ C. Wang, H. Gong, W. Lei, Y. Cai, Z. Hu, S. Xu, Z. Liu, Q. Feng, H. Zhou, J. Ye, J. Zhang, R. Zhang, and Y. Hao, “Demonstration of the p-NiOₓ/n-Ga₂O₃ Heterojunction Gate FETs and Diodes With BV²/Rₒₙ,sp Figures of Merit of 0.39 GW/cm² and 1.38 GW/cm²,” IEEE Electron Device Lett. 42(4), 485–488 (2021).DOI
  21. [21]²¹ H. Luo, X. Zhou, Z. Chen, Y. Pei, X. Lu, and G. Wang, “Fabrication and Characterization of High-Voltage NiO/β-Ga₂O₃ Heterojunction Power Diodes,” IEEE Trans. Electron Devices 68(8), 3991–3996 (2021).DOI
  22. [22]²² X. Lu, X. Zhou, H. Jiang, K.W. Ng, Z. Chen, Y. Pei, K.M. Lau, and G. Wang, “1-kV Sputtered p-NiO/n-Ga₂O₃ Heterojunction Diodes With an Ultra-Low Leakage Current Below $1~\mu$ A/cm²,” IEEE Electron Device Lett. 41(3), 449–452 (2020).DOI
  23. [23]²³ X. Lu, Y. Deng, Y. Pei, Z. Chen, and G. Wang, “Recent advances in NiO/Ga₂O₃ heterojunctions for power electronics,” J. Semicond. 44(6), 061802 (2023).DOI
  24. [24]²⁴ C. Liao, X. Lu, T. Xu, P. Fang, Y. Deng, H. Luo, Z. Wu, Z. Chen, J. Liang, Y. Pei, and G. Wang, “Optimization of NiO/β-Ga₂O₃ Heterojunction Diodes for High-Power Application,” IEEE Trans. Electron Devices 69(10), 5722–5727 (2022).DOI
  25. [25]²⁵ Y. Kokubun, S. Kubo, and S. Nakagomi, “All-oxide p–n heterojunction diodes comprising p-type NiO and n-type β-Ga₂O₃,” Appl. Phys. Express 9(9), 091101 (2016).DOI
  26. [26]²⁶ H. Gong, X. Chen, Y. Xu, Y. Chen, F. Ren, B. Liu, S. Gu, R. Zhang, and J. Ye, “Band Alignment and Interface Recombination in NiO/β-Ga₂O₃ Type-II p-n Heterojunctions,” IEEE Trans. Electron Devices 67(8), 3341–3347 (2020).DOI
  27. [27]²⁷ Y. Deng, Z. Yang, T. Xu, H. Jiang, K.W. Ng, C. Liao, D. Su, Y. Pei, Z. Chen, G. Wang, and X. Lu, “Band alignment and electrical properties of NiO/β-Ga₂O₃ heterojunctions with different β-Ga₂O₃ orientations,” Applied Surface Science 622, 156917 (2023).DOI
  28. [28]²⁸ A. Jain, S.P. Ong, G. Hautier, W. Chen, W.D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder, and K.A. Persson, “Commentary: The Materials Project: A materials genome approach to accelerating materials innovation,” APL Materials 1(1), 011002 (2013).DOI
  29. [29]²⁹ S. Ghosh, M. Baral, R. Kamparath, R.J. Choudhary, D.M. Phase, S.D. Singh, and T. Ganguli, “Epitaxial growth and interface band alignment studies of all oxide α-Cr₂O₃/β-Ga₂O₃ p-n heterojunction,” Applied Physics Letters 115(6), 061602 (2019).DOI
  30. [30]S. Ghosh, M. Baral, J. Bhattacharjee, R. Kamparath, S.D. Singh, and T. Ganguli, “Evaluation of valence band offset and its non-commutativity at all oxide α-Cr2O3/β-Ga2O3 heterojunction from photoelectron spectroscopy,” Journal of Applied Physics 130(17), 175303 (2021).DOI
  31. [31]S. Lany, “Band-structure calculations for the 3 d transition metal oxides in G W,” Phys. Rev. B 87(8), 085112 (2013).DOI
  32. [32]S. Lany, “Semiconducting transition metal oxides,” J. Phys.: Condens. Matter 27(28), 283203 (2015).DOI
  33. [33]V. Stevanović, S. Lany, X. Zhang, and A. Zunger, “Correcting density functional theory for accurate predictions of compound enthalpies of formation: Fitted elemental-phase reference energies,” Phys. Rev. B 85(11), 115104 (2012).
  34. [34]C.J. Bartel, S.L. Millican, A.M. Deml, J.R. Rumptz, W. Tumas, A.W. Weimer, S. Lany, V. Stevanović, C.B. Musgrave, and A.M. Holder, “Physical descriptor for the Gibbs energy of inorganic crystalline solids and temperature-dependent materials chemistry,” Nat Commun 9(1), 4168 (2018).
  35. [35]C.W. Bale, E. Bélisle, P. Chartrand, S.A. Decterov, G. Eriksson, A.E. Gheribi, K. Hack, I.-H. Jung, Y.-B. Kang, J. Melançon, A.D. Pelton, S. Petersen, C. Robelin, J. Sangster, P. Spencer, and M.-A. Van Ende, “FactSage thermochemical software and databases, 2010–2016,” Calphad 54, 35–53 (2016).DOI
  36. [36]J. Osorio-Guillén, S. Lany, S.V. Barabash, and A. Zunger, “Magnetism without Magnetic Ions: Percolation, Exchange, and Formation Energies of Magnetism-Promoting Intrinsic Defects in CaO,” Phys. Rev. Lett. 96(10), 107203 (2006).
  37. [37]C.-W. Lee, A. Zakutayev, and V. Stevanović, “Computational insights into phase equilibria between wide-gap semiconductors and contact materials,” (2023).arxiv.org/abs/2312.11749
  38. [38]E. Arca, K. Fleischer, S.A. Krasnikov, and I. Shvets, “Effect of Chemical Precursors On the Optical and Electrical Properties of p-Type Transparent Conducting Cr2O3:(Mg,N),” J. Phys. Chem. C 117(42), 21901–21907 (2013).DOI
  39. [39]K. Fleischer, D. Caffrey, L. Farrell, E. Norton, D. Mullarkey, E. Arca, and I.V. Shvets, “Raman spectra of p-type transparent semiconducting Cr2O3:Mg,” Thin Solid Films 594, 245–249 (2015).DOI
  40. [40]E. Arca, A.B. Kehoe, T.D. Veal, A. Shmeliov, D.O. Scanlon, C. Downing, D. Daly, D. Mullarkey, I.V. Shvets, V. Nicolosi, and G.W. Watson, “Valence band modification of Cr2O3 by Ni-doping: creating a high figure of merit p-type TCO,” J. Mater. Chem. C 5(47), 12610–12618 (2017).DOI
  41. [41]K.A. Borup, E.S. Toberer, L.D. Zoltan, G. Nakatsukasa, M. Errico, J.-P. Fleurial, B.B. Iversen, and G.J. Snyder, “Measurement of the electrical resistivity and Hall coefficient at high temperatures,” Review of Scientific Instruments 83(12), 123902 (2012).DOI
  42. [42]W. Hao, Q. He, X. Zhou, X. Zhao, G. Xu, and S. Long, in 2022 IEEE 34th International Symposium on Power Semiconductor Devices and ICs (ISPSD) (IEEE, Vancouver, BC, Canada, 2022), pp. 105–108.DOI
  43. [43]L. Zhou, X. Lu, L. Chen, X. Ouyang, B. Liu, J. Xu, and H. Tang, “Leakage Current by Poole–Frenkel Emission in Pt Schottky Contacts on ( )-β-Ga2O3 Grown by Edge-Defined Film-Fed Growth,” ECS J. Solid State Sci. Technol. 8(7), Q3054–Q3057 (2019).DOI
  44. [44]S.M. Sze, Y. Li, and K.K. Ng, Physics of Semiconductor Devices (John wiley & sons, 2021).
  45. [45]Z. Wang, H.-H. Gong, X.-X. Yu, X. Ji, F.-F. Ren, Y. Yang, S. Gu, Y. Zheng, R. Zhang, and J. Ye, “Trap-mediated bipolar charge transport in NiO/Ga2O3 p+-n heterojunction power diodes,” Sci. China Mater. 66(3), 1157–1164 (2023).DOI

Paper details

vlnc:W_bpv7rxaz
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents