β-Ga₂O₃ orientation dependence of band offsets with SiO₂ and Al₂O₃

Ga₂O₃ is currently attracting interest for high power electronic devices with breakdown performance beyond the one-dimensional limits of both SiC and GaN.^1–12 The relatively low cost of manufacture makes these devices an option for kV class rectifiers for power converters in electric vehicles and their charging infrastructure.^1–5 Dielectrics are needed on β-Ga₂O₃ as gate insulators on transistors, edge termination structures on rectifiers, and surface passivation layers for encapsulation.^3–5 In most of these applications, there is a need to know the band alignment between the dielectric and the semiconductor to understand carrier accumulation and electric field profiles at the heterointerfaces. Two of the most common dielectrics for β-Ga₂O₃ are SiO₂ and Al₂O₃, which can be deposited by a range of techniques, including low damage atomic layer deposition, and selectively patterned with dry and wet etchants.^1,3,4 The band offset between SiO₂ and Al₂O₃ insulators in heterojunction conduction in Ga₂O₃ devices demonstrates its crucial significance, influencing the electronic transport characteristics and interfacial charge transfer dynamics at the junction interface. While several studies have reported the band alignments of these materials on β-Ga₂O₃ for specific crystallographic orientations,^13–18 there has not been a self-consistent examination of the band offsets for the three main commercially available substrate orientations.^17,19 Since the properties of β-Ga₂O₃ are anisotropic,^19–22 it is of value to establish the orientation dependence of band alignments. The use of the same cleaning procedure, deposition method, and thermal budget reduces variations in the measured valence band offsets.²³ 

X-ray photoelectron spectroscopy was used to measure the valence band offsets of SiO₂ and Al₂O₃ on different β-Ga₂O₃ orientations of (001), (010), and $( 2 ¯01)$ using the standard Kraut method.^24–26 The valence band offsets for SiO₂ on these three orientations were 1.4, 1.4, and 1.1 eV, respectively, while for Al₂O₃ the corresponding values were 0.0, 0.1, and 0.2 eV, respectively. Care was taken to minimize sample charging and other issues.²⁷ Conduction band offsets were derived from the directly measured valence band offsets and bandgaps of dielectrics and Ga₂O₃.²⁸ 

Three β-Ga₂O₃ bulk substrates (10 × 15 mm) with different orientations, (001), (010), and $( 2 ¯01)$⁠, were purchased from NCT, Japan. All were nominally undoped, 0.65 mm thick, with n-type carrier concentrations of <9 × 10¹⁷ cm⁻³. To indicate the crystal quality, the x-ray diffraction full width at half maximum in [100] and [010] azimuth directions was 50 arc sec in both cases. The root-mean-square surface roughness was <2 nm in all cases, as measured by atomic force microscopy of 10 × 10 μm² areas. The front surface was chemically mechanically polished while the rear faces had ground finishes. The band alignments of atomic layer deposited (ALD) SiO₂ or Al₂O₃ on these substrates were measured after deposition conditions that have been described previously.^17,25,26 The layers for both dielectrics were deposited at 200 °C using tris (diethylamino) silane or trimethylaluminum as precursors, respectively, for SiO₂ or Al₂O₃. The temperature was chosen based on this being the optimum for insulator quality as judged by morphology and crystal quality. The layers are polycrystalline under these conditions.¹⁷ For substrate cleaning predeposition, the following rinse sequence was employed: acetone, IPA, N₂ dry, and finally ozone exposure for 15 min. The ALD layers were deposited at 200 °C in a Cambridge Nano Fiji 200 using a remote inductively coupled plasma at 300 W to generate atomic oxygen. The bandgaps of the dielectrics were 8.7 eV for SiO₂ and 6.9 eV for Al₂O₃ as measured previously by reflection electron energy loss spectroscopy^29–31 and by the O1s peak from 200 nm dielectric reference layers.^25,26 1.5 nm layers of SiO₂ or Al₂O₃ were deposited on the three Ga₂O₃ substrates in the same ALD deposition for measuring their core levels. This thickness is chosen so that the core levels in the underlying substrate can be measured through the insulator. The thicknesses were obtained from crystal balance calibrated measurements. The RMS roughness measured over 10 × 10 μm areas by atomic force microscopy was 1.8–2.1 nm for the 200 nm layers.

Valence band offsets were obtained from the Kraut method using XPS with a Physical Instruments ULVAC PHI system.²⁴ This uses an Al X-ray source with an energy of 1486.6 eV, a source power of 300 W, and an analysis area of 20 μm diameter. The standard take-off angle of 50° with an acceptance angle of ±7° was used for all spectra. For high resolution scans, the electron pass energy was 23.5 eV. The total energy resolution of the system is 0.5 eV, with an accuracy of 0.03 eV for the binding energies measured.

The procedure for determining the valence band offsets entails the determination of core level and valence band energy positions using three distinct Ga₂O₃ substrates along with 200 nm thick layers of dielectrics.^17,25–27 Subsequently, the subsequent phase involves the quantification of core level displacements within the SiO₂/Ga₂O₃ and Al₂O₃/Ga₂O₃ heterostructures. These displacements can be translated into their corresponding valence band offsets. Finally, the disparities in conduction band positions are inferred by leveraging the measured energy bandgaps and the valence band offsets.

Figure 1 shows the determination of the bandgaps of the three different orientations of β-Ga₂O₃ from the onset of the plasmon loss feature in the O 1s photoemission spectra. Table I shows the valence band maxima and core level data used to calculate the bandgaps of Ga₂O₃. The bandgaps were determined to be 4.64, 4.71, and 4.59 eV for the $( 2 ¯01)$⁠, (001), and (010) oriented β-Ga₂O₃ substrates, respectively. These correspond closely to the previously reported values of 4.67, 4.72, and 4.57 eV for $( 2 ¯01)$⁠, (001), and (010) oriented β-Ga₂O₃ substrates, respectively, as determined from Tauc plots of the transmittance spectra.¹⁷  Table I also shows the same data for the thick SiO₂ and Al₂O₃ layers, showing respective bandgaps of 8.7 and 6.9 eV. These are also consistent with previous reports.¹⁷ 

Figure 2 shows the delta core level energies for interfaces of thin SiO₂/β-Ga₂O₃ with orientation (a) (001), (b) (010), and (c) $( 2 ¯01)$ from the differences in Ga 2p3 and Si 2p transitions. These are tabulated in the top of Table II for the SiO₂/β-Ga₂O₃ heterojunctions.

These data lead to the band alignment results of Fig. 3 for the SiO₂/β-Ga₂O₃ structures for the three different orientations of Ga₂O₃. All of the band alignments are type I, nested gap. The valence band offsets are, respectively, 1.4, 1.4, and 1.1 eV, for the (001), (010), and $( 2 ¯01)$ orientations. This is consistent with a value of 1 eV reported by Konishi et al.¹⁴ for plasma enhanced chemically vapor deposited SiO₂. The corresponding conduction band offsets are then 2.59, 3.01, and 2.66 eV for the (001), (010), and $( 2 ¯01)$ orientations. Clearly, SiO₂ is an appropriate choice as a dielectric for β-Ga₂O₃ of any orientation in any of the possible applications noted earlier, since both the valence and conduction band offsets are larger than the commonly quoted rule of thumb of >1 eV.²⁸ 

The data for delta core level energies for the Al₂O₃/β-Ga₂O₃ heterostructures are shown in Fig. 4 for (a) (001), (b) (010), and (c) $( 2 ¯01)$ orientations of β-Ga₂O₃ from the differences in Ga 2p3 and Al 2p transitions. The data are also tabulated at the bottom of Table II.

The corresponding band alignment diagrams are shown in Fig. 5 for the Al₂O₃/ β-Ga₂O₃ structures. The valence band offsets are very small, being 0.1, 0.2, and 0.0 eV, respectively, for the (001), (010), and $( 2 ¯01)$ orientations. These would not provide significant confinement of holes in device structures, although Al₂O₃ could still be used as passivation of field plate layers. The result for the $( 2 ¯01)$ orientation is consistent with the previously reported value of 0.07 eV by Carey et al.^15,18 The corresponding conduction band offsets are 2.29, 2.51, and 2.26 eV for the (001), (010), and $( 2 ¯01)$ orientations.

We determined the band alignments for two common dielectrics on Ga₂O₃ for three orientations of this semiconductor. The dielectrics were deposited by low damage ALD, to avoid the type of interfacial disorder that is present during physical vapor depiction methods. The conduction band offsets in both SiO₂/β-Ga₂O₃ and Al₂O₃/β-Ga₂O₃ heterostructures are large and provide excellent electron confinement for all three major crystal orientations of β-Ga₂O₃, but the valence band offsets in the latter system are smaller than desirable for limiting hole transport. The differences in band alignment are relatively small for the different orientations, even though the properties of monoclinic β-Ga₂O₃ are anisotropic because of the low crystal symmetry. SiO₂ is a superior choice because of its larger valence band offsets, although Al₂O₃ could still have utility for surface passivation, field management, and encapsulation purposes.

Work was performed as part of Interaction of Ionizing Radiation with Matter University Research Alliance (IIRM-URA), sponsored by the Department of the Defense, Defense Threat Reduction Agency under Award No. HDTRA1-20-2-0002. The content of the information does not necessarily reflect the position or the policy of the federal government, and no official endorsement should be inferred. The authors thank the staff of the Nanoscale Research Facility at UF, part of the Herbert Wertheim College of Engineering's Research Service Center (RSC), for assistance in device fabrication.

The authors have no conflicts to disclose.

Hsiao-Hsuan Wan: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Software (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Jian-Sian Li: Formal analysis (equal); Investigation (equal); Methodology (equal). Chao-Ching Chiang: Data curation (equal); Formal analysis (equal); Methodology (equal). Xinyi Xia: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal). David C. Hays: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal). Fan Ren: Conceptualization (equal); Funding acquisition (equal); Methodology (equal); Project administration (equal); Resources (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Stephen J. Pearton: Conceptualization (equal); Project administration (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available within the article.