For the past decade, Sandia National Laboratories has been at the forefront of research in vertical gallium nitride power semiconductor devices. Prime examples of this are the demonstration of a world record > 6 kV GaN pn diode, as well as the first demonstration of a > 1.2 kV vertical GaN trench MOSFET with a HfO2 gate dielectric. In the past six years, Sandia has worked on a program funded by the Department of Energy’s Vehicle Technologies Office to develop vertical GaN power semiconductors for next-generation vehicle electrification. This talk focuses on the challenges and cycles of learning from that program as they relate to the characterization and development of wide bandgap (WBG) power semiconductors.Many of the conventional characterization techniques used for WBG semiconductors were developed for Si-based electronics in the latter half of the 20th century and were later adapted to WBG semiconductors. As the understanding of WBG semiconductors evolves over time, it becomes necessary to reevaluate these techniques for sources of error and to provide modifications and alternate methods which adapt to the nuances of the material being evaluated. One such example of this is the challenge in extracting the density of interface traps (DIT ) in a WBG metal oxide semiconductor (MOS) capacitor. We will explore some of the nuances in extracting DIT for a WBG semiconductor compared to silicon. In addition, we propose a method of visual qualification of the shape of the DIT vs. [EC-ET] curve which can be used to assess the accuracy of the extraction.Reducing channel resistance for MOSFETs is a key factor in improving performance and has an especially large impact for GaN or SiC vertical devices rated at 1700 V or below. A large contributing element to high channel resistance comes from low channel mobilities which are often reported as low as 15-30 cm2/V·s for SiC, but have been reported as high as 200 cm2/V·s for GaN. A common and basic technique for extracting mobility in SiC planar devices is to use fatFET structures which are long channel devices where the channel resistance becomes the dominant contributor to total device resistance, and hence provides an accurate method for extracting channel mobility. For trench MOSFETs, making a long channel device is not a viable option due to the channel orientation, hence it becomes challenging to find a good approach for extracting channel mobility. One method sometimes used to extract mobility is the transconductance method. We will discuss this method and the inaccuracy of this technique for extracting channel mobility in trench MOSFETs. In our study, the transconductance technique underestimates channel mobility by as much as 33% and the error is associated with the additional series resistance. We will discuss this issue and as well as alternate approaches to extracting mobility.Another challenge we report on concerns the extraction of contact resistivity and the limitations of various techniques. Using the circular transmission line model (CTLM) for extracting contact resistivity has several challenges related to the thickness of the semiconductor film below the contact, the correction factor approximation, and the error for extremely low resistivity contacts. For thick films, alternative methods like the Cox and Strack [1] method are significantly more area efficient. Understanding the limitations of these techniques helps set boundaries on our interpretation of data and is key in providing accurate information on contact resistivity. References R. H. Cox and H. Strack, Solid-State Electron. 10 (12), 1213 (1967). This work was supported by the Electric Drivetrain Consortium managed by Susan Rogers of DOE’s Vehicle Technologies Office. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government.
Read full abstract