The influence of ultrasonic vibrations on material removal in the silicon wafer polishing using DDCAMRF: Experimental investigations and process optimization

Abstract

The present experimental investigation attempts to understand and address the effect of ultrasonic vibrations on material removal in the polishing of silicon wafers (1 0 0). The requisite finishing experimentations were performed on an indigenously developed experimental arrangement of double-disc chemical assisted magnetorheological finishing (DDCAMRF) process with longitudinal vibrations. The MR fluid used in the experiments consists of a water-based suspension prepared by mixing suitable amounts of carbonyl iron particles (CIPs), abrasive particles, and additives or stabilizers. The prepared MR fluid uses both mechanics and chemistry to finish the silicon surface. Mechanics is mainly responsible for micro-scratching of silicon surface, which gets “softened” by hydration utilizing DI water in the MR fluid. In this study, the ‘response surface methodology (RSM)’ was chosen for designing the experiments to evaluate the significance of different process factors, namely polishing speed, abrasive concentration, and ultrasonic power on the material removal rate (MRR) in DDCAMRF process. The material removed from the wafer surface was measured using the precision digital weighing balance. It was observed that the MRR was found to increase with the increase in various process factors used. Further, analysis of variance (i.e., ANOVA) technique with a 95% confidence interval was performed to analyze the significant contribution of different process factors on MRR. The validation of developed model was done by performing experiments on random and optimized set of process factors. From, the statistical investigation it was discovered that ultrasonic power has highest contribution of 57.9% on MRR, followed by the polishing speed (13.3%), and abrasive concentration (12.5%). Furthermore, a genetic algorithm optimization tool was utilized to obtain optimum set of process parameters to maximize MRR.