Title:
Boosting Lithography optimization with AI for thin film device fabrication

Abstract:
The development of thin-film sensing solutions traditionally involves time-consuming and expensive optimization cycles to deliver devices whose performance meets the requirements of a specific application. More efficient process optimization is not only a technical challenge but also a fundamental technological requirement for the next generation of sensor technologies for industrial applications.

Lithography plays a crucial role in thin-film device fabrication, as it defines one of the critical design parameters: the sensing element geometry. However, its optimization still largely relies on a trial-and-error approach.

In this presentation, an alternative approach to lithography optimization based on machine learning and inverse analysis is presented. The central idea is to move beyond the conventional question, “What pattern will these lithography parameters produce?” and instead ask, “What lithography parameters should be used to obtain the pattern we want?”

This talk showcases the potential of AI to move lithography from trial-and-error toward a more data-driven methodology, enabling the development of specification-driven process optimization strategies as a means of shortening the gap between R&D and prototyping, while also highlighting the challenges that still need to be addressed.