Heterojunctions as the Weakest Link: A Fundamental Investigation of Damage Evolution in Electronic Devices

项目摘要提案题目：异质结作为最薄弱的环节：电子器件损伤演变的基础研究非技术摘要微电子器件（如晶体管）的可靠性对其应用至关重要。通常，这是通过加载晶体管到失效，然后分析数据的电气特性来研究的。死后显微镜也可以观察到装置内部的损伤。但是，必须小心地重新创建故障事件。在这项研究中，引入了一个新的概念，在显微镜中测试晶体管，允许近原子分辨率的缺陷演变映射以及局部热量和电流传输的作用。将特别考虑晶体管内部各层之间的接口，其对器件可靠性的重要作用很难用传统方法研究。这种对机械、电气和热领域的高分辨率访问将消除现有的挑战，即识别晶体管中的局部弱点以及它们对全局或器件级故障影响背后的基本机制。这种独特的方法将应用于下一代全电动交通、能量存储和射频通信技术中使用的高功率晶体管的研究。这项研究的成功将导致晶体管在这些应用中减少相关设备的尺寸和重量，同时提高功率和可靠性。除了在高功率晶体管可靠性的基础科学方面取得进展外，该项目还将确保培养具有前沿和多学科科学技术的研究生和本科生。为了吸引作为下一代劳动力的K-12学生，将开展外展活动。技术摘要本研究的目的是探讨异质结在功率晶体管整体器件可靠性中的作用。这项研究的动机是高功率和高频设备的预测可靠性与实际可靠性之间的差距，这些设备被吹捧为下一代全电动运输、能量存储和射频通信技术。将研究基于氮化镓的高电子迁移率晶体管（HEMT），以回答两个基本问题：(a)就缺陷成核而言，晶体管中最弱的组件是什么；(b)是缺陷成核还是原子扩散在器件降解中更具病毒性。本研究假设跨界面和沿界面的扩散可能是答案，并提出了一种独特的基于操作中的显微镜的验证方法。透射电子和热反射显微镜将用于实时绘制和监测设备在降解开始时的结构和传输特性。无论是在空间上还是在时间上，高分辨率都将当前在设备数据中查找故障签名的范例从设备级转移到原子接口级。对缺陷和界面的原子结构和化学的实时研究，以及对它们对电子-声子相互作用的影响的理解，将消除准确识别主要机制和量化其影响的障碍。本研究在分离了各种机械应力（残余应力、热弹性应力和逆压电应力）后，精心设计了非热实验。这项研究的结果将为GaN HEMT故障提供基本见解，并为性能和可靠性改进提供途径。它还将在课堂和实验室环境中培养研究生和本科生。该项目还将接触到下一代劳动力，即K-12学生，以吸引他们走向科学和技术。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

PROJECT ABSTRACTProposal Title:Heterojunctions as the Weakest Link: A Fundamental Investigation of Damage Evolution in Electronic DevicesNon-technical AbstractReliability of microelectronic devices, such as transistors, is critical to applications. Typically, this is studied by loading the transistors to failure and then analyzing the electrical characteristics of the data. Post-mortem microscopy is also performed to visualize the damage inside the device. However, care must be taken to recreate the failure events. In this research, a novel concept is introduced where the transistor is tested inside microscopes that allow near-atomic resolution mapping of defect evolution and the role of local heat and current transport. Special consideration will be given to the interfaces between various layers inside a transistor, whose important role in device reliability has been difficult to study using conventional approaches. Such high-resolution access to mechanical, electrical and thermal domains will remove the existing challenge of identifying the local weak spots in the transistor and the fundamental mechanisms behind their impact on the global or device level failure. This unique approach will be applied to the study of high-power transistors that will be used in next generation all-electric transportation, energy storage and radio-frequency communication technologies. Success of this research will lead to transistors that will reduce the size and weight of relevant equipment in these applications while increasing power and reliability. In addition to the advancements in the fundamental science of high-power transistor reliability, the project will ensure training of the graduate and undergraduate students with cutting edge and multi-disciplinary science and technology. Outreach activities will be performed to attract K-12 students, who are the workforce of the next generation.Technical AbstractThe objective of this research is to investigate the role of heterojunctions in overall device reliability for power transistors. The research is motivated by the gap between the predicted and actual reliability of high power and high frequency devices touted to enable next generation all-electric transportation, energy storage and RF communication technologies. Gallium nitride based high electron mobility transistors (HEMT) will be studied to answer two fundamental questions: (a) what is the weakest component in a transistor in terms of defect nucleation and (b) is it defect nucleation or atomic diffusion that is more viral in device degradation. This research hypothesizes that diffusion across and along interfaces could be the answers and lays out a unique in-operando microscopy based validation approach. Transmission electron and thermo-reflectance microscopy will be used in real time to map and monitor the structural and transport characteristics of the devices during the onset of degradation. The high resolution, both spatially and temporally, shifts the current paradigm of looking for failure signatures in device data from the device-level to the atomic interface level. Realtime investigation of the atomic structure and chemistry of defects and interfaces, along with an understanding of their influence on electron-phonon interactions, will eliminate the roadblock of accurately identifying the dominant mechanism and quantifying its impact. The proposed research elaborately designs non-thermal experiments after isolating the various components of mechanical stress (residual, thermo-elastic and inverse piezo-electric). The outcome of this research will provide fundamental insights on GaN HEMT failure and suggest paths towards performance and reliability improvements. It will also train graduate and undergraduate students in both class and laboratory settings. The project will also reach out to the next generation workforce, the K-12 students, to attract them toward science and technology.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.