High-Q-Power-GaN - Development of transistors for high-power electronics on low-defect free-standing gallium nitride substrates

本研究计划的目的是在低缺陷的自支撑氨化GaN衬底上开发用于功率转换的AlGaN/GaN基大功率晶体管。使用具有非常低的位错密度的氨生长GaN衬底使得能够实现具有出色的电特性的AlGaN/GaN高功率晶体管，以利用GaN材料的特定物理特性。该方案的一个重要方面是确定AlGaN/GaN晶体管的退化机制，特别是GaN缓冲结构中的击穿电压。击穿电压降低的原因是GaN中的杂质或结构缺陷，其潜在地充当电荷载流子陷阱并在带隙中产生局部能级。许多产生载流子陷阱的机制非常复杂，至今尚未完全理解。这是因为GaN外延异质结构中的缺陷（例如位错和点缺陷）的多样性和密度由于非原生衬底上的沉积工艺而非常高。原则上，该电子陷阱可以位于AlGaN势垒层中的表面处、AlGaN/GaN界面处靠近二维电子气的位置处、或者GaN缓冲层中。在本项目中，将研究缺陷对AlGaN/GaN功率晶体管的漏电流、阈值电压和载流能力等参数的影响。在这种情况下，一个特定的调查对象将是AlGaN/GaN晶体管的退化模式上的碳掺杂的影响。碳在GaN中充当受体，其用于补偿GaN缓冲结构的本征n型导电性。最近的证据表明，AlGaN/GaN晶体管的GaN缓冲层中的某些退化与碳和高位错密度有关。然而，异质外延AlGaN/GaN HEMT中的上述陷阱相关效应总是被高缺陷密度引起的效应所重叠。到目前为止，还不可能明确地澄清例如碳掺杂在某些条件下是否会导致这些器件的击穿电压。使用氨氧化物GaN衬底将使得能够实现具有碳补偿GaN缓冲层的低缺陷密度AlGaN/GaN功率晶体管。通过这种方法，有一种方法可以通过AlGaN/GaN基器件中的位错相关效应来分离碳相关。为了这个目的，处理的功率晶体管的缺陷丰富的AlGaN/GaN结构生长在非原生衬底（SiC和Si）的功率晶体管，这是同质外延生长在低缺陷ammonohydrated GaN衬底进行了系统的比较。

The aim of this research project is to develop AlGaN/GaN based high-power transistors for power conversion on low defect free-standing ammonothermal GaN substrates. The use of ammonothermal grown GaN substrates having a very low dislocation density enables the realization of AlGaN/GaN high power transistors with outstanding electrical properties to exploit the specific physical properties of the GaN material. An important aspect of this proposal is the determination of degradation mechanism of AlGaN/GaN transistors, particularly the breakdown voltages in the GaN buffer structure. Reasons for the reduced breakdown voltages are impurities or structural defects in the GaN, which potentially act as charge carrier traps and generate localized levels in the band gap. Many of the mechanisms for the generation of charge carrier traps are very complex, and as yet incompletely understood. This is because the diversity and density of defects, such as dislocations and point defects, in GaN epitaxial heterostructures is very high due to the deposition process on non-native substrates. In principle, this electron traps can be located at the surface, in the AlGaN barrier layer, at the AlGaN/GaN interface close to the two-dimensional electron gas, or in the GaN buffer layer. In this project, the effects of defects on parameters such as leakage current, threshold voltage, and current-carrying capacity in AlGaN/GaN power transistors will be investigated. A particular object of investigation in this context will be the influence of the carbon doping on the degradation pattern of AlGaN/GaN transistors. Carbon acts as an acceptor in GaN, which is used to compensate the intrinsic n-type conductivity of GaN buffer structures. Recent evidence suggests that certain degradations in the GaN buffer layers of AlGaN/GaN transistors are related to carbon and a high dislocation density. However, the above traps-related effects in heteroepitaxial AlGaN/GaN HEMT are always overlapped by the effects arising from the high defect densities. An unambiguous clarification of whether e.g. a carbon doping may under certain conditions lead to the breakdown voltage of these devices was not possible until now. Using ammonothermal GaN substrates will enable the realization of low defect density AlGaN/GaN power transistors with carbon-compensated GaN buffer layers. With this approach, there is a way to separate the carbonrelated by the dislocation-related effects in AlGaN/GaN based devices. For this purpose, processed power transistors of defect-rich AlGaN/GaN structures grown on non-native substrates (SiC and Si) are systematically compared with power transistors, which were grown homoepitaxially on low-defect ammonothermal GaN substrates.