Proceedings PaperCorrelation of real-time-monitored process module parameters and wafer results
|Format||Member Price||Non-Member Price|
The business climate of the integrated circuit industry in the 1990's is very different from that of the 1970's. Today's industry demands that manufacturers produce newer more complex designs with shrinking design time windows and higher return on investment for fabrication facilities. This tighter business climate inposes stricter controls on the day to day operation of the fabrication process. Now, not only must new equipment make significant improvements in process capability but every effort must be made to insure that older (already depreciated) equipment is used to its fullest. To this end extensive work has been undertaken to establish "sensor based manufacturing" (SBM) schemes with the goal of improving wafer to wafer as well as within a wafer repeatability and extending equipment useful lifetime. This improvement in repeatability translates into reduced need for expensive, time consuming, in-line metrology. The end result is improved cycle time and reduced overall facility costs. Early work on this topic centered around the generation of a parametrical view of the plasma process module. By viewing the module as the convolution of a plasma based chemical factory and the hardware necessary to establish and sustain it, a scheme for improved control becomes clear. Present control schemes are based only on hardware or machine parameters, yet the wafer results are defined by the plasma parameters. Therefore, shifting the focus of module control from the traditional machine parameters to the less conventional plasma parameters would place the focus of control closer to the wafer and improve process performance. This paper deals with the application of SBM concepts to establish the real time monitoring of fundamental machine and plasma parameters known to have direct correlation to wafer results such as etch rate and uniformity. Proper usage of SBM generated information includes improved end point capabilities, real time SPC and "go-no-go" decision analysis.