Modelling and experimental study of roughness in silicon wafer self-rotating grinding

Abstract

Self-rotating grinding is the most widely used technology to thin silicon wafer. The roughness is an important indicator of thinning quality and processing accuracy. To get a better grinding quality, rigid control of roughness is required. Although the models of roughness in metal and ceramic machining were extensively studied, mechanism of roughness formation in silicon wafer self-rotating grinding was not well understood. In this article, starting from the mechanism of grinding grooves formation, Rayleigh probability density function was used to characterize the depth of grinding grooves. By establishing a relationship between the roughness and the depth of grooves, a theoretical model of roughness was developed. The overlapping effect of abrasive grains and wheel-workpiece deflection were also considered to improve the accuracy of the model. The model could identify the effects of processing parameters, abrasive grain size, material properties and grinding mark geometry on roughness quantitatively. Verification experiments under seven grinding conditions were performed to validate the theoretical model. The experimental values agreed with the predictive values with less than 20% deviation. Effects of wheel rotation speed, wafer rotation speed, feed rate and wafer radial distance on roughness were discussed in detail.