Monitoring of Photoresist Pattern Thickness Using Top-down SEM Image

In semiconductor manufacturing, CD-SEM (Critical Dimension-Scanning Electron Microscopy) is widely used as a process monitoring tool. As design rules become more fine and dense, needs for controlling three-dimensional shape in addition to two-dimensional size of the pattern, such as a CD value, have risen for the process monitoring. In this paper, we propose a method for estimating a thickness of photoresist pattern only from a single top-down SEM image. In this method, the relationship between the image features calculated from the top-down SEM image and photoresist thickness is learned beforehand, and the thickness is estimated using the relationship. High throughput can be expected owing to the small number of necessary images, compared to stereovision techniques that require two or more tilt images with different viewpoints. In addition, we propose a method for estimating lithography process parameters that can cause thickness decrease. It is confirmed that the thickness of photoresist pattern (designed thickness 270nm) could be estimated with accuracy of 19.9nm (3σ), which is so small that the method can be applied to photoresist thickness monitoring.