The resolution of the PEM image is basically determined by the quality of the imaging EO, such as the point spread function, aberrations, and optical magnification, and also by the pixel size of the TDI sensor. In our currently developed PEM for the 16-nm node, the pixel size of the acquired image is set to . In previous work using our developed tool, hp 64- and 44-nm L/S patterns were successfully resolved and 12-nm-sized defects were detected without any false defect.15–16,24 On the other hand, the resolution of the SEM image is basically determined by the spot size of the incident beam, and that is determined by the beam current and beam energy.10 In order to analyze the defect detectability using PEM and SEM, simulated PEM and SEM images with hp 64-nm L/S patterned EUV masks were obtained using CHARIOT Monte Carlo software (Abeam Technologies Inc.).25 The software with 72 cores was installed in an all-in-one server computer, Proliant DL 980 G2 (Hewlett-Packard) with 80 cores. In order to simulate the PEM image obtained by our developed tool, the pixel size and the number of electrons/pixel on the simulated PEM image were set to and , respectively. On the other hand, simulated SEM images were modeled on the multibeam SEM, developed by Malloy et al., and the current of a single beam was set to 672 pA.7 The resolution of the simulated SEM image was determined by the pixel size of the image, which was varied from 1.5 to 6 nm. In order to obtain the same electron dose per unit area as that of the PEM, the dwell time was set to 6.1 ns (in the case of 1.5-nm-sized pixel). According to the ITRS-2014 update, the defect size on the EUV mask is defined as the square root of the defect area on a two-dimensional mask surface. Therefore, square-shaped programmed defects were used in the simulation to simplify the analysis of defect detection. In order to compare the PEM and SEM images with the same SE yields, a fixed landing energy of 1 keV was used to acquire these images. The detailed simulation conditions for PEM have been described in earlier reports.9,21–22 The difference between the simulated image with defects and that without defects is defined as the difference image. In order to enhance the defect signal intensities, different types of image processing operations for PEM and SEM images were applied to the simulated images.9,17 The EUV aerial image and the wafer printability of the mask defects were simulated using a LAIPH™ defect printability simulator (Luminescent Technologies Inc.).26 For a typical defect type, we focused on edge extrusion and edge intrusion, and we evaluated the printability with the printed space-width difference on a wafer. Illumination conditions were defined to match a state-of-the-art EUV pattern exposure system: , to 0.9, and illumination type as a dipole. Defect sizes that generated a 10% critical dimension error for 16-nm node patterns were then calculated. The LER of the L/S patterns was formed with normal random numbers and was described by a 3-sigma deviation of an edge from a straight line (3 sigma).