Simulation technique for pattern inspection using a projection electron microscope

Abstract

The nature of image formation captured by a projection electron microscope (PEM), and defect detection capability of a PEM inspection system for extreme ultraviolet mask, was investigated by a newly developed Monte Carlo simulator while taking into account an imaging electron optics. The detailed analyses of the electron trajectories clarify the mechanism of the PEM image formation without white band peaks, or undershot dips near the pattern edges. The edge slope of the pattern is affected by the material of the under-layer (capping layer). And its incident beam energy dependence can be derived from the experimentally obtained secondary electron emission coefficients (SEECs) by taking into account the damping effect on the signal intensity at the edge. The incident beam energy dependence on the defect detection capability is in good agreement with the logarithmic plots of the experimentally obtained SEEC difference curves. It was found that the gray level difference (caused by SEEC difference) has much more impact on the defect detectability than the gradient of the edge slope of a pattern has; and thus, the optimal inspection condition can be predicted by measuring the SEEC curves of a sample. These results indicate that high-precision calculation can be carried out using the developed PEM simulator.