Image analysis development for optimization of 4H–SiC trench recovery treatment by combination of gas etching and sacrificial oxidation approaches
M. Barcellona,P. Badalà,M. Boscaglia,M. Cantiano,D. Mello,E. Ferlito,M.D. Pirnaci,D. Tenaglia,M.E. Fragalà
DOI: https://doi.org/10.1016/j.mssp.2024.108216
IF: 4.1
2024-02-11
Materials Science in Semiconductor Processing
Abstract:Silicon Carbide (SiC) is one of the most promising new-generation semiconductor material for high-power and high-efficiency electronic devices overcoming in some respects silicon mature performances. SiC substrates are more and more widely used in many semiconductor-based applications, like DC/AC converters. SiC-based MOSFET technology is nowadays applied in electric/hybrid vehicle power conversion systems. To further enhance the performance of SiC HV devices, trench-gate MOSFET design is preferred to planar-gate MOSFET structure. The trenches frame is defined by Reactive Ion Etching (RIE) by which high aspect ratio structures can be attained. Despite the fine-tuning of the etching process, some sidewall striations' undesired effects occur, degrading the performances of the devices. Accordingly, H 2 gas etching and sacrificial oxidation methods were adopted to recover the wall integrity, trying to restore the atomic roughness that characterizes the 4H–SiC. Since this problem interests vertically oriented trench sidewalls, a quantitative estimation of the roughness is difficult as Atomic Force Microscopy (AFM) measurements are always time-dispending and hard to manage: therefore, only a qualitative evaluation of the sidewall roughness can be accomplished by FIB/SEM micrographs. ImageJ software is herein used for a quantitative analysis starting from SEM images allowing a fast and repeatable quality assessment of trench striations before and after the recovery processes. The image analysis can produce numerical data regarding the status of the SiC trenches. In the present work, we will refer to the image analysis output value as Relative Roughness (RR). It represents the numerical data helpful to fulfill the study with a Design of Experiments (DOE) and Responsive Surface Method (RSM) aimed to optimize the post-etch recovery procedures.
engineering, electrical & electronic,materials science, multidisciplinary,physics, condensed matter, applied