Error modeling and analysis in dynamic wafer handling

Wafer handling robots are used to transfer wafers in semiconductor manufacturing. A pick-measure-place method is typically used to transfer wafers accurately between stations. The measurement step is performed by an aligner, which is time-consuming. To improve the wafer transfer efficiency, it is necessary to speed up the measurement process or place it in parallel with other operations. Thus optic sensors are installed at each station to estimate the wafer eccentricity on-the-fly. The estimation process is consist of two stages: sensor calibration and wafer eccentricity estimation. Theoretical analysis and numerical optimization methods are used to accomplish these tasks. Based on the wafer handling robot kinematics model, robot kinematics error, sensor calibration error and eccentricity identification error are analyzed in this paper. The proposed error model based method is verified by both theoritical analysis and real experiment performed on a wafer handling robot system. Experiment results demonstrate that the error modeling methods can greatly reduce the wafer eccentricity estimation error on-the-fly. Hence the developed methods can be used to improve the wafer handling accuracy and reduce the wafer handling cycle time in semiconductor manufacturing.