Modeling of thermally induced error for high-precision step imprint tool

Thermally induced error during the exposure process is a major contribution to the repeatable alignment accuracy of the step imprint tool in imprint lithography system. A modeling based on the multiple regression analysis was proposed for the real-time prediction of thermally induced errors with multiple temperature measurements, and the measurement points were reduced from 15 to 6. The experimental results show that the method accurately predicts the time-variant thermally induced error under different condition, the error predicting veracity reaches to 89% in x direction and 93% in z direction. As a highly accurate predictor, this method solves the difficulties of the variable coupling.