Abstract
Trial linewidth standards for the calibration of scanning electron microscopes (SEM's) used in semiconductor production have been manufactured and evaluated. A quantitive investgation of the relationship between peak-to-peak linewidth and cross sectional dimension has been made. The results show that the peak-to-peak values are 7nm smaller than the values obtained from cross sectional measurements. The edge signal peak shift has been simulated using a newly modified diffusion model and this also gives smaller linewidth than the defined edge distance. Linewidths measured optically have been compared with those measured by an SEM. The optical method gives wider linewidths by some dozens of nanometer. The influence of focusing errors and limit of the measurements have also been investigated.
