Characterization of diffraction gratings in a rigorous domain with opticalscatterometry: hierarchical neural-network model

Characterization of microstructures with features from submicrometers to hu ndreds of micrometers requires versatile methods. Profilometry and optical microscopy cannot cope with submicrometer features, and atomic-force micros copy, scanning-electron microscopy, and near-field microscopy are inherentl y slow, off-line methods. In optical scatterometry, the laser light scatter ed from a sample is measured and the sample profile is subsequently charact erized. We propose the use of a two-stage model based on neural networks: r ough categorization followed by refinement, thus reducing the need for prio r information on the sample. We simulate the method for a submicrometer dif fraction grating characterized by five parameters. It is shown that intensi ty measurements of few diffraction orders by use only of one wavelength are enough to yield rms errors of less than 2 nm for the parameters (approxima tely 2-3% of the optimal values of the parameters). (C) 1999 Optical Societ y of America.