Study of Chemical-Physical Cleaning Technologies for Better Particle Removal Trapped on the Crystal Defects of 4H-SiC

Abstract

Listen

Translate article

SiC, compared to conventional Si, has gained attention as a power semiconductor material due to its outstanding physical characteristics such as wide bandgap, high breakdown voltage, fast electron saturation velocity, excellent thermal conductivity, and superior mechanical strength. The fabrication process of SiC wafers follows a similar procedure to most processes except for Si single crystal growth and it requires management of various contaminants such as particles and metal impurities for high-quality SiC wafer production. The cleaning process of SiC wafers is crucial for achieving high-quality power semiconductor fabrication and high yield. The cleaning mechanism of contaminated particle for Si wafers is well known, typically utilizing RCA cleaning based on the principles of oxidation and etching with an increase in electrostatic repulsion. However, according to our previous studies, SiC shows characteristics where the chemical mixtures used in RCA cleaning do not induce oxidation and etching, it is necessary to find alternative cleaning methods for better contaminant removal distinct from Si cleaning mechanisms. Particularly, for removing particle contaminants trapped on crystal defects like triangular surface defects occurring during SiC fabrication, chemical-based cleaning methods are not effective. In this study, we evaluate various chemical-mechanical cleaning techniques for effectively removing particle contaminants on SiC wafers, specifically researching particle contamination removal within crystal defects.Given the superior mechanical strength of SiC compared to Si, this study assesses chemical-mechanical cleaning techniques, including PVA brush scrubbing technique and megasonic cleaning technique employing hydrogen dissolved functional DIW. Since most particle contamination occurs during the CMP process, 4H-SiC wafers were polished using CMP equipment to replicate particle contamination for cleaning evaluation. Furthermore, to compare and evaluate particle removal performance according to particle contamination methods, particles were contaminated using the dipping method. Particle removal efficiency was assessed by calculating the particle removal rate on SiC using optical microscopy, and the removal of particles trapped on the crystal defects was evaluated using FESEM. Both PVA brush scrubbing and megasonic cleaning methods showed excellent performance in surface particle removal. However, PVA brush scrubbing showed lower efficiency in removing particles trapped on the crystal defects, whereas megasonic cleaning method demonstrated higher efficiency, particularly when using hydrogen dissolved DIW, showed a better particle removal performance. This is attributed to the higher bubble cavitation effect observed in megasonic cleaning with hydrogen dissolved DIW, resulting in greater energy imparted upon collapse of cavitation bubbles, thus more effectively cleaning particles on the substrate surface and within crystal defects. For particle cleaning of SiC wafers, a chemical-mechanical cleaning process is excellent, and particularly, megasonic with hydrogen dissolved DIW was confirmed to be effective in removing particle contaminants trapped on crystal defects.