Ihiv i\ a pri-piinl of « paper intended for publicalion in I journal or proceedings Since hanni-s m»i be made Wfnre publication, thtt prrprnit i* made atailahlc with the un dtrstandinn that il will not be cite d or reproduced without the pirmKsion of the author DISTRIBUTE!. UF TKiS HUUiiHENT IS UN! IR1ITED DISCLAIMER This document was prepared as an ammnt of work sponsored by an agency of the United Stales Government.

Video monitoring of pulsed visible-wavelength mirror damage led to the previously reported observation of illuminated coating defects as precursors to microscopic damage. Typically, many defects are observed in the laser interaction area, but only a few act as damage initiators. The present study seeks to identify the latter defects via any special characteristic of their laser-excited emission spectra that may serve as a recognizable signature. The ability to make such recognition nondestructively is crucial to producing optical coatings that perform at intrinsic damage levels. The video microscopy apparatus and methodology for measuring emission spectra of individual defects will be described. Results from various dielectric mirrors will be presented, including conclusive evidence that defect illumination is primarily the result of defect scattering. The dependence of defect image intensity on laser wavelength is examined more closely.

A video microscopy system has demonstrated the ability to detect and locate defects nondestructively in multilayer optical coatings. The system uses laser-excited scattering to illuminate defects responsible for laser-induced damage in multilayer dielectric mirrors. Scatter intensity maps and contours can be generated by digitizing the video image of scattering from individual defects. These can reveal characteristic defect scatter features that can be related to a damage probability. Due to the relatively low resolution of the stored video image, the scattering from a defect may be contained in only a few pixels. Using digital-enhancement techniques such as cross correlation, noise filtering, and Laplacian edge enhancement combined with fast-Fourier transforms, it is possible to increase the resolution of the characteristic features of a scattering defect for a more detailed analysis. Examples of these techniques for a dielectric mirror will be presented and discussed.

The ability to detect and locate defects nondestructively in multilayer optical coatings has been demonstrated using a video microscopy system. The system uses laser-excited scattering to illuminate defects responsible for laser-induced damage in multi-layer dielectric mirrors. So far, attempts to predict the particular defect that will initiate damage have been unsuccessful. Presently, intensity maps and contours can be generated by digitizing the video images of scattering from individual defects. These can reveal characteristic defect scatter features possibly related to damage probability, and they offer a means to follow the development of these features from incipient to catastrophic damage. Examples of defect scatter maps and intensity contours for a dielectric mirror will be presented and discussed.