To support the ongoing size reduction in integrated circuits, the need for accurate depth measurements of on-chip structures becomes increasingly important. Unfortunately, present metrology tools do not offer a practical solution. In the semiconductor industry, critical dimension scanning electron microscopes (CD-SEMs) are predominantly used for 2D imaging at a local scale. The main objective of this work is to investigate whether sufficient 3D information is present in a single SEM image for accurate surface reconstruction of the device topology. In this work, we present a method that is able to produce depth maps from synthetic and experimental SEM images. We demonstrate that the proposed neural network architecture, together with a tailored training procedure, leads to accurate depth predictions. The training procedure includes a weakly supervised domain adaptation step, which is further referred to as pixel-wise fine-tuning. This step employs scatterometry data to address the ground-truth scarcity problem. We have tested this method first on a synthetic contact hole dataset, where a mean relative error smaller than 6.2% is achieved at realistic noise levels. Additionally, it is shown that this method is well suited for other important semiconductor metrics, such as top critical dimension (CD), bottom CD and sidewall angle. To the extent of our knowledge, we are the first to achieve accurate depth estimation results on real experimental data, by combining data from SEM and scatterometry measurements. An experiment on a dense line space dataset yields a mean relative error smaller than 1%.

No abstract

No abstract

There is a growing need for accurate depth measurements of on-chip structures. Since Scanning Electron Microscopes (SEMs) are already regularly being used for fast and local 2D imaging, it is attractive to explore the 3D capabilities of SEMs. This paper presents a comprehensive study of depth estimation performance when single- or multi-angle data is available. The research starts with an analytical line-scan model to show the major contributors of the signal change with increasing height and angle. We also analyze Monte-Carlo scattering simulations for height sensitivity on similar structures. Next, we validate the depth estimation performance with a supervised machine learning model and show correlation with the previous studies. As predicted by the sensitivity studies, we show that the height prediction greatly improves with increasing tilt angle. Even with a small angle of 3 degrees, a threefold average performance improvement is obtained (RMSE of 16.06 nm to 5.28 nm). Finally, we discuss a preliminary proof-of-concept of a self-supervised algorithm, where no ground-truth data is needed anymore for height retrieval. With this work we show that a data-driven tilted-beam approach is a leap forward in accurate height prediction for the semiconductor industry.