Motor power signal analysis for end-point detection of chemical mechanical planarization

In the integrated circuit (IC) manufacturing, in-situ end-point detection (EPD) is an important issue in the chemical mechanical planarization (CMP) process. In the paper, we chose the motor power signal of the polishing platen as the monitoring object. We then used the moving average method, which was appropriate for in-situ calculation process and made it easy to code for software development, to smooth the signal curve, and then studied the signal variation during the actual CMP process. The results demonstrated that the motor power signal contained the end-point feature of the metal layer removal, and the processed signal curve facilitated the feature extraction and it was relatively steady before and after the layer transition stage. In addition, the motor power signal variation of the polishing head was explored and further analysis of time delay was performed.