As the power of laser produced plasma sources has increased, EUV lens heating has become a major component of process variation. Differential lens heating can cause thermal aberrations which affect system drift during operation, therefore pupil plane characterization will play a critical role in process optimization for EUV lithography (EUVL). In-situ full pupil characterization, which depicts the pupil in its in-use state, is essential for these tools. To this end we have developed Quick Inverse Pupil (QUIP)—a software suite developed for rapid characterization of pupil plane behavior based on images formed by that system. This algorithm is based on statistical modeling, which correlates image-space variation with known aberrations. Previously we have presented variations on this algorithm which can only measure third-order aberrations and requires aerial image data. In this paper, we will present an approach to measure high order aberrations from images formed in resist. An inverse pupil solution will be obtained from CD-SEM image analysis. We will show that the additional degrees of freedom required to measure high-order aberrations can be achieved through using different pitches of the metrology targets. We will demonstrate that this technique can accurately determine third- and fifth-order aberrations with a retrieval error below 0.5 miliwaves in under one second. A combination of synthetic and experimental data will be presented.
|