An auto-focus algorithm using multiple band-pass filters for a scanning electron microscope (SEM) is proposed. To acquire sharp images of various kinds of defects by SEM defect observation in semiconductor manufacturing, the auto-focus process must be robust. A method for designing a band-pass filter for calculating the ‘focus measure’ (a key parameter of the auto-focus process) is proposed. To achieve an optimal specific frequency response for various images, multiple band-pass filters are introduced. As for the proposed method, two series of focus measures are calculated by using multiple band-pass filters independently, and it is selected according to reliability of the series of focus measures. The signal-to-noise ratio of an image for acceptable auto-focus precision is determined by simulation using pseudo images. In an experiment using the proposed method with real images, the success rate of auto focus is improved from 79.4% to 95.6%.

A scanning electron microscope transition edge sensor has been developed to analyze the minor or trace constituents contained in a bulk sample and small particles on the sample under a low accelerating voltage (typically <3 keV). The low accelerating voltage enables to improve the spatial analysis resolution because the primary electron diffusion length is limited around the sample surface. The characteristic points of our transition edge sensor are 1) high-energy resolution at 7.2 eV@Al-Kα, 2) continuous operation by using a cryogen-free dilution refrigerator and 3) improvement of transmission efficiency at B-Kα by using thin X-ray film windows between the sample and detector (about 30 times better than our previous system). Our system could achieve a stabilization of the peak shift at Nd-Mα (978 eV) within 1 eV during an operation time of 27 000 s. The detection limits with B-Kα for detection times 600 and 27 000 s were 0.27 and 0.038 wt%, respectively. We investigated the peak separation ability by measuring the peak intensity ratio between the major constitute (silicon) and the minor constitute (tungsten) because the Si-Kα line differs from the W-Mα line by only 35 eV and a small W-Mα peak superimposed on the tail of the large Si-Kα peak. The peak intensity ratio (I(W-Mα)/I(Si-Kα)) was adjusted by the W particle area ratio compared with the Si substrate area. The transition edge sensor could clearly separate the Si-Kα and W-Mα lines even under a peak intensity ratio of 0.01.

No abstract

No abstract

Abstract222222222222222A highly sensitive inspection algorithm is proposed that extracts defects in multidimensional vector spaces from multiple images. The proposed algorithm projects subtraction vectors calculated from test and reference images to control the noise by reducing the dimensionality of vector spaces. The linear projection vectors are optimized using a physical defect model, and the noise distribution is calculated from the images. Because the noise distribution varies with the intensity or texture of the pixels, the target image is divided into small regions and the noise distribution of the subtraction images are calculated for each divided region. The bidirectional local perturbation pattern matching (BD‐LPPM), which is an enhanced version of the LPPM, is proposed to increase the sensitivity when calculating the subtraction vectors, especially when the reference image contains more high‐frequency components than the test image. The proposed algorithm is evaluated using defect samples for three different scanning electron microscopy images. The results reveal that the proposed algorithm increases the signal‐to‐noise ratio by a factor of 1.32 relative to that obtained using the Mahalanobis distance algorithm. © 2012 Wiley Periodicals, Inc. Electron Comm Jpn, 95(9): 44–53, 2012; Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ecj.11390