Homojunction, graded bandgap normalGaAlAs:normalSi LED's have been demonstrated to be highly reliable. During accelerated aging (30 mA forward bias, 250°C), these LED's generally degrade slowly without the formation of dark lines or dark spots. It has recently been found, however, that some LED's show a rapid reduction of light output even when aged without current bias at 200°C. <110> oriented dark lines appeared in the electroluminescence image of heavily degraded devices. To determine the source of the DLD's, the material quality of degraded LED's and unaged LED's from the same wafer was assessed using etch pitting and transmission cathodoluminescence. Examination of the devices revealed the presence of small pyramids, composed mainly of silicon, on the p‐surface of some of the LED's; no unusual features were found on the remaining LED's. A dislocation network surrounding each pyramid was found to be initially present. Due to bonding and thermal stresses applied to the pyramids, these networks enlarged with aging along the {111} planes and appeared as <110> DLD's in the electroluminescence image upon reaching the p‐n junction. In degraded LED's from pyramid‐free regions of the wafer with pyramids, <110> DLD's initiate at dicing damage and regions of the p‐surface subject to high bonding stress, i.e. edges and corners. The growth of these DLD's is identical to those initiating at pyramids. A similar study of LED's from a wafer that is totally free of pyramids shows that DLD's are not formed in devices that degrade slowly.

An automated IR ellipsometer was used to measure the ellipsometric parameters, ψ and Δ, on multilayer mirrors at 3.80 μm. Spatial variations in differential phase shift, Δ, are modeled by thickness variations in the multilayer coating and variations in ψ are modeled by an absorption change.

The University of Dayton Research Institute (UDRI) has designed and is building three infrared ellipsometers in support of the Air Force high-energy laser program. One of these instruments is a conventional null type elliposmeter for operation in the 3.39 to 4.00 μm wavelength range. Ellipsometric parameters are determined with a precision of 0.01° at any angle of incidence from 20° to 85°. Repeatability and absolute accuracy also approach 0.01° which yields optical constants accurate to 0.1 percent in some cases. All components are mounted on removable carriers for easy interchange, although operation is normally in the PCSA (Polarizer, Compensator, Sample, Analyzer) configuration. The system software has special provisions for the second instrument which is also a null type ellipsometer dedicated to monitoring vacuum deposited coatings in real time at 3.39 μm. The software can model non-ideal components in the optical path using Jones or Mueller matrices and provide completely corrected data with a single zone measurement after the instrument is characterized. The third instrument is a rotating polarizer, automated ellipsometer tunable from 1 to 12 μm. A dedicated computer operates the instrument and reduces data.

The University of Dayton has designed and built an Infrared Ellipsometer/ Reflectometer (E/R) to characterize curved surfaces in support of the ALPHA Laser Program. The instrument is tunable in wavelength and angle of incidence, is remotely controlled by a dedicated microcomputer, is positioned and scanned by a six-degree-of-freedom translation stage, and makes simultaneous measurements of Rp, Rs, Ψ, and Δ. A return path configuration was chosen to provide the versatility necessary for the variety of curved surfaces to be used in the ALPHA laser. This E/R is capable of measuring Rp and Rs to 0.1% accuracy and Delta to better than 0.1° accuracy on curved surfaces to 1.0 centimeter radius of curvature. The optical design and theory of operation as well as automatic alignment and focusing features are discussed. Data taken with this instrument are presented.