Our modern information-based society has largely been built on two pillars: the information itself and the technological means to access, store, and process this information. The enormous growth in the stream of information has been made possible by the giant technological progress in the production of integrated circuits (ICs) that nowadays form the technology backbone of our information society. This progress is commonly known as “Moore’s law”, and it embodies five decades of relentless innovation, research, and development in the semiconductor industry.
For successful and economically viable production of ICs, a key and critical ingredient is the ability to accurately measure and monitor the quality of each individual step of the production process and its alignment with respect to previous steps. Optical scatterometry is currently a widely used technique in high-volume manufacturing. For several technical reasons it is the most viable and cost-effective candidate for current and future high-quality high-throughput wafer-metrology. However, this technique critically depends on fast and accurate Maxwell solvers, which numerically solve Maxwell’s equation to predict how light is scattered by structures on a wafer.
We aim at the construction and demonstration of an advanced, accurate, and efficient Maxwell solver for optical metrology. This Maxwell solver needs to include cross-talk with product structures and coupling to the metrology sensor, and should be in line with state-of-the-art fast numerical modeling techniques recently developed at TU/e. This will address the urgent need for a next-generation Maxwell solver for optical scatterometry.