An effective system for parameter optimization in photolithography process of a LGP stamper

In the current thin-film transistor liquid crystal display industry, the light guide plate (LGP) of the backlight module has become thinner and smaller, and the backlight module needs to be illuminated uniformly and effectively. The parameter setting for the photolithography process of a LGP stamper often relies on the engineers' experiences by means of trial-and-error or design of experiment to obtain a suitable and more reliable process parameter setting, which requires a large amount of time, manpower, and cost. This research proposes a novel two-stage optimization system for photolithography process integrating the Taguchi method, back-propagation neural networks, genetic algorithms, particle swarm optimization, and related technologies to effectively generate optimal process parameter settings. The first stage is to reduce the process variance. The second stage is to find the final optimal process parameter settings for the best quality specification. Experimental results show that the proposed system can create the best process parameters which not only meet the quality specification for the micro-dots on the photoresist, but also effectively enhance the overall process stability.