Focus, edge detection, and CCD camera characterization for development of an optical overlay calibration standard.

摘要

In order to ensure continued growth and development, a consortium of IC manufacturers has produced a "roadmap" of critical technologies immediately needed, and predicted to be needed, by the industry in the near future. Reduction of critical dimensions (the smallest dimensions of an IC, typically the CMOS gate length) necessitate tighter control over the alignment of one mask (i.e., lithographic) level relative to another. Measurement of the relative alignment of two such masks is known as "overlay metrology." Reference standards for calibration of present and planned overlay metrology tools must be developed for the IC industry to meet their anticipated needs.;This work contributes to the development of calibration standards and methods for overlay metrology by consideration and characterization of several aspects of overlay measurement that introduce error into the measurement. Unavoidable variations in the focus response of an overlay tool lead to errors due to coupling of lateral motion of the measuring microscope with its focus motion, and its variation of optical aberrations with focus. We consider various algorithms available for autofocus of an optical microscope. The algorithms have been tested with simulated and real data. We have found that an algorithm's response depends crucially on the material system being investigated. We also determined an optimal algorithm of those tested for use on the NIST optical overlay metrology tool. Detection of feature edges and their positions on the IC are critical to overlay metrology. We investigated various algorithms for edge detection appropriate for the optical overlay metrology tool at the National Institute of Standards and Technology (NIST). Results comparing the performance of the recommended algorithm against various algorithms used in the industry are presented. Length standards are normally calibrated with a scanning photometric stage monitored by laser interferometry. Optical overlay patterns are measured with digital charge coupled device cameras viewing a stationary stage. We investigated the potential errors introduced into the overlay measurement by the use of such cameras. We present methods for mapping and correction of the errors introduced by the cameras and their associated optical systems.