Investigation of polishing parameters and slurry composition on germanium chemical mechanical planarisation using response surface methodology

Abstract

Chemical mechanical polishing/planarisation (CMP) stays a widely used process for complete planarisation in semiconductor fabrication. CMP process provides surface uniformity, high selectivity, low defects with the desired material removal rate. The removal rate is influenced by various independent parameters namely turntable speed, down pressure, slurry pH, and H2O2 concentration. Modelling the CMP process from fundamental principles is limited. Hence, in this study, the design of the experimental methodology has been adopted to design the CMP process. Different models such as linear, quadratic, two-factor interaction, and cubic mathematical were developed and statistically analysed in identifying the suitable model by Box-Behnken design. The consequence of each parameter and their interactions on Ge removal rate is analysed. A quadratic model is proposed from the outcome. The predicted values achieved using model equations exhibited appropriate fit by experimental values (R2 value for rutile and anatase as 0.943 and 0.942, respectively). The present work verified that response surface methodology and Box-Behnken design can be expeditiously functional for modelling of chemical mechanical planarisation.