A Survey of Run-to-run Control for Batch Processes.

Run-to-run (R2R) control is widely used in semiconductor manufacturing systems to minimize the process drift, shift and variability. The R2R controller adjusts control actions or recipes in a supervisory manner after each batch. This paper provides a comprehensive literature review of R2R control methods for the batch process. First, the principles of major R2R controllers are introduced and analyzed, such as exponentially weighted moving average (EWMA), double exponentially weighted moving average (d-EWMA), model predictive control (MPC), optimizing adaptive quality controller (OAQC), artificial neural network (ANN). Besides, simulation examples with different R2R controllers are made to compare the robustness and adaptability. Then, several case studies concerning a chemical mechanical planarization (CMP) process, a multi-input and multi-output (MIMO) system of the furnace process control and the management of blood glucose (BG) are presented. Finally, the paper concludes with some recommendations and directions for the future research.