分子束外延
兴奋剂
材料科学
氧化物
外延
光电子学
纳米技术
冶金
图层(电子)
作者
A. Ardenghi,Oliver Bierwagen,A. Falkenstein,G. Hoffmann,Jonas Lähnemann,M. Martin,Piero Mazzolini
标识
DOI:10.48550/arxiv.2202.05762
摘要
The oxidation-related issues in controlling Si doping from the Si source material in oxide molecular beam epitaxy (MBE) is addressed by using solid SiO as an alternative source material in a conventional effusion cell. Line-of-sight quadrupole mass spectrometry of the direct SiO-flux ($Φ_{SiO}$) from the source at different temperatures ($T_{SiO}$) confirmed SiO molecules to sublime with an activation energy of 3.3eV. The $T_{SiO}$-dependent $Φ_{SiO}$ was measured in vacuum before and after subjecting the source material to an O$_{2}$-background of $10^{-5}$ mbar (typical oxide MBE regime). The absence of a significant $Φ_{SiO}$ difference indicates negligible source oxidation in molecular O$_{2}$. Mounted in an oxygen plasma-assisted MBE, Si-doped $β$-Ga2O3 layers were grown using this source. The $Φ_{SiO}$ at the substrate was evaluated [from 2.9x10$^{9}$ cm$^{-2}$s$^{-1}$ ($T_{SiO}$=700°C) to 5.5x10$^{13}$ cm$^{-2}$s$^{-1}$ (T$_{SiO}$=1000°C)] and Si-concentration in the $β$-Ga2O3 layers measured by secondary ion mass spectrometry highlighting unprecedented control of continuous Si-doping for oxide MBE, i.e., $N_{Si}$ from 4x10$^{17}$ cm$^{-3}$ ($T_{SiO}$=700°C) up to 1.7x10$^{20}$ cm$^{-3}$ ($T_{SiO}$=900°C). For a homoepitaxial $β$-Ga2O3 layer an Hall charge carrier concentration of 3x10$^{19}$ cm$^{-3}$ in line with the provided $Φ_{SiO}$ ($T_{SiO}$=800°C) is demonstrated. No SiO-incorporation difference was found between $β$-Ga2O3(010) layers homoepitaxially grown at 750°C and $β$-Ga2O3(-201) layers heteroepitaxially grown at 550°C. The presence of activated oxygen (plasma) resulted in partial source oxidation and related decrease of doping concentration (particularly at $T_{SiO}$<800°C) which has been tentatively explained with a simple model. Degassing the source at 1100°C reverted the oxidation.
科研通智能强力驱动
Strongly Powered by AbleSci AI