Background: Accurate measurement of local critical dimension uniformity (LCDU) and local pattern placement errors (LPPE) caused by stochastics is a critical task in metrology for advanced technology nodes, especially for EUV lithography. Aim: Optimizing the settings of a metrology tool to either reduce biases in LCDU and LPPE measurement, or to reliably measure and remove those biases, is critical to stochastics metrology accuracy. Approach: Four different scan modes and two different scan orientations are compared for the measurement of staggered arrays of contact holes on a scanning electron microscope (SEM), and both biased and unbiased LCDU and LPPE are compared. Results: The Hitachi SEM scan modes designed for the measurement of 2D patterns have the minimum sensitivity to scan orientation and differences in X and Y metrology results. Further, measuring and subtracting SEM biases increases the accuracy of the best scan mode results, but also reduces the differences between scan modes by 3X. Conclusions: SEM scan modes designed for 2D patterns produce lower bias in the measurement of contact hole LCDU and LPPE. Measuring and removing biases produce the maximum accuracy for all scan modes.
|