Enhancing projection based iterative learning control: A set-membership approach.

Area of Science:

• Control Engineering
• Mechatronics
• Semiconductor Manufacturing

Background:

• Lithographic machine tools demand high precision and flexibility for semiconductor fabrication.
• Projection-based iterative learning control (P-ILC) is a promising approach for wafer stages.
• Existing P-ILC methods face challenges with model uncertainties and error convergence.

Purpose of the Study:

• To enhance the performance of Projection-based iterative learning control (P-ILC) for lithographic machine tools.
• To improve both nonparametric and parametric feedforward control modes within the P-ILC framework.
• To address limitations related to repetitive error compensation, non-repetitive error accumulation, and convergence speed under model uncertainties.

Main Methods:

• Incorporation of a set-membership based frequency-domain iterative learning control (SM-F-ILC) algorithm.
• Application of SM-F-ILC within the nonparametric feedforward control mode of P-ILC.
• Leveraging the benefits of SM-F-ILC to also enhance the parametric feedforward control mode.
• Validation through experimental results on wafer stages.

Main Results:

• The enhanced P-ILC scheme demonstrated improved performance in compensating for repetitive errors.
• Effective attenuation of non-repetitive error accumulation was observed.
• Fast convergence speed was achieved, even in the presence of model uncertainties.
• Experimental validation confirmed the efficacy of the proposed control strategy.

Conclusions:

• The integration of SM-F-ILC significantly enhances the capabilities of P-ILC for lithographic applications.
• The improved P-ILC scheme offers superior error compensation and faster convergence, crucial for wafer stage precision.
• This approach provides a robust solution for achieving high motion accuracy and flexibility in semiconductor manufacturing processes.