Clear, transparent (0.5–5.0 μ), solution-grown, single crystals of negative uniaxial LiIO3 show no optically induced damage and remain stable in the temperature range 20°–256°C. The crystals are of point group 6 and are optically active and pyroelectric. We have measured the linear electrooptic coefficients in the visible and the optical activity in the infrared. For λ1=1.0845 μ phase-matched second harmonic generation (PMSHG) occurred at θm=28.9° and this angle changed by less than 0.3° in the range 20–256°C. PMSHG was employed to determine the single domain quality of the crystals and to measure | d31(LiIO3)/d36(KDP) |=11. At 1.06 μ we have measured the coherence lengths and the ratio of the nonlinear coefficients | d33/d31 |=0.8.

No abstract

We discuss the implementation of a strategy designed to provide laser-lightemitting reliability assurance for 1.3-μνα InGaAsP/lnP lasers of the planar mesa, buried heterostructure type for use in a submarine cable application. The testing regimes include initial characterization (cosmetic and light-current curve inspection), passive aging (elevated temperatures [85 to 175°C] without bias, with and without humidity [<85-percent relative humidity]), overstress active aging (high temperatures [150°C], high currents [250 mAdc]), and longterm rate-monitoring active aging (elevated temperature [60°C] burn-in [3 mW/facet]). Overstress testing is designed to compel a timely (~10 2 -hour) identification of premature failures, due to modes of degradation other than the long-term ultimately controlling wear-out mode, and to stabilize transient modes. To identify premature failures of the wear-out type, survivors of overstressing are subjected to rate monitoring in which wear-out degradation rates, established in a reasonable time (~10 3 hours), may be sorted. The principal results of the important overstress aging were the detection of an initially occurring saturable degradation mode, present to some extent in most lasers, and a regimen to force its rapid stabilization, so that it would not obscure determination of the activation energy of the wear-out mode. With a credibly determined value for the latter, it was deterministically inferred from * Authors are employees of AT&T Bell Laboratories.

Outlines are given for eight alternative black-box (i.e., input-output) meth odologies that are appropriate for estimating, from external characteristics, the reliability of semiconductor lasers or other gradually degrading manufac tured products with lifetimes too long to measure directly over practical time spans. These reliability estimates, which are essential for various components of such systems as submarine communication cables or satellites, are obtained from two classes of data. One class consists of the measured properties of statistically equivalent components, i.e., samples from the manufactured pop ulation, that have been operated to failure or at least to a significant degree of degradation. This degradation is often brought about in a shortened time span through the application of a temperature or other "accelerating stress" that is large compared to the operating temperature or other stress of the intended application. The other class of data is, for each component, comprised of the predeployment properties ofthat very component, including particularly its own predeployment degradation rate (which may also be measured under accelerating stresses). Brief consideration is given in passing to important special cases when only one of these two classes of data is available.

Various types of InP-based semiconductor lasers, Fabry–Perot (FP), and distributed feedback (DFB), in different wavelength regions of 1.3, 1.48, and 1.55 μm have been subjected to human-body-model electrostatic discharge (ESD) testing. The reverse V-I characteristics of these diode lasers were found to be generally most sensitive in detecting ESD damage than the forward characteristics (e.g., threshold current) of the laser. The laser ESD failure voltages were much lower for the reverse than the forward polarity and DFB lasers were found to be more vulnerable to ESD than FP lasers. The failure mechanism was found to be due to localized melting—a thermal effect—in both polarities of ESD testing. We also report the study of the latent ESD effects on the long-term aging rates of semiconductor lasers.