Statistical information enhanced robust design method of optical thin film

Area of Science:

• Optical Engineering
• Materials Science

Background:

• Traditional optical film design methods often struggle to simultaneously optimize multiple performance metrics.
• Sensitivity to parameter errors can limit the practical yield of optical coating systems.

Purpose of the Study:

• To develop a novel, multi-objective optimization strategy for optical film design.
• To improve the simultaneous control over spectral coefficient error and its standard deviation.
• To enhance the production yield and reduce sensitivity to parameter errors in optical film systems.

Main Methods:

• Construction of a new merit function incorporating spectral coefficient average error and standard deviation.
• Application of a multi-objective optimization strategy combining Non-Dominated Sorting Genetic Algorithm (NSGA) and Sequential Quadratic Programming (SQP).
• Comparative analysis of the new algorithm against conventional methods for wideband anti-reflection films, cut-off filters, and infrared dual-band filters.

Main Results:

• The novel merit function effectively optimizes both spectral error expectation and standard deviation envelope.
• The proposed multi-objective optimization strategy demonstrated superior control over sensitivity to film parameter errors.
• Comparative designs showed the new algorithm's effectiveness in improving film system performance and yield.
• The new design method offers time-efficient calculations compared to traditional approaches.

Conclusions:

• The developed multi-objective optimization strategy provides a significant advancement in optical film design.
• This novel approach enhances production yield and offers computational efficiency for optical coating systems.
• The method has practical significance for the manufacturing of advanced optical filters and films.