Paper EM-MoA4
Effects of Surface Cleaning and Different Metals as Schottky Contacts to Bulk and Epitaxial β-Ga₂O₃

Monday, November 7, 2016, 2:40 pm, Room 102A

Beta-gallium oxide (β-Ga₂O₃) has emerged over the past few years as a promising next-generation wide bandgap semiconductor. It has a bandgap of ~4.8 eV and a breakdown electric field of ~8 MV/cm, giving it a superior figure-of-merit compared to traditional wide bandgap semiconductors like SiC and GaN. Moreover, it can be produced from the melt, and single-crystal (2-in diameter) substrates have recently become commercially available. Devices based on β-Ga₂O₃ that have so far been demonstrated include Schottky diodes, metal-semiconductor field effect transistors (MESFETs), metal-oxide-semiconductor field-effect transistors (MOSFETs), and ultra-violet (UV) photodiodes. However, since research on β-Ga₂O₃ as a wide bandgap semiconductor is in its very early stages, there is little understanding on how to control device-relevant interfaces to this material. In this work, we have investigated Schottky diodes fabricated on Sn-doped (5x10¹⁸ cm^-3) single-crystal Ga₂O₃ (-201) substrates and lightly doped (~10¹⁷ cm^-3) Ga₂O₃(010) homoepilayers. A surface study was first performed to evaluate the effect of different surface cleaning techniques on contact performance. The surface cleaning methods consisted of (1) an organic solvent clean only (acetone and isopropanol), and an organic clean followed with a (2) HCl, (3) BOE, (4) HCl and H₂O₂ or (5) BOE and H₂O₂. The corresponding Schottky barrier heights (SBHs) were calculated from the I-V and C-V behaviour of Ni Schottky diodes fabricated on bulk Ga₂O₃ (-201). SBHs were lowest for the organically cleaned sample, and highest for the sample treated in HCl and H₂O₂. The latter also had the lowest leakage current in reverse bias and showed the most stable performance even after a period of several weeks after deposition. We have therefore established that organic clean followed by HCl and H₂O₂ treatment is the most effective of the cleaning methods tested. We have also investigated Schottky diodes fabricated using different Schottky metals. On the bulk Ga₂O₃ (-201) substrates, we calculated SBHs from the I-V behavior of Ir, Ni, Au and Sn to vary from ~1.0–0.7 eV in approximate correspondence with the metal workfunctions. On the lightly doped β-Ga₂O₃ (010) epilayer, preliminary measurements indicate a SBH > 1.0 eV for Ni. Electrical behavior of other metals on the (010) epilayer will also be investigated and reported in the presentation.