A technique was developed to obtain the subcritical crack growth velocity in a 4-point bending sample by analyzing the load-displacement curve. This was based on the observation that the compliance of a beam increases as the crack grows. Beam theory was used to analyze the general configuration where two cracks propagated in the opposite directions. A simple equation relating the crack velocity to the load and displacement was established, taking advantage of the fact that the compliance was linearly proportional to the crack lengths; thus the absolute crack length was not important. Two methods of obtaining crack velocity as a function of load were demonstrated. First, by analyzing a load-displacement curve, a corresponding velocity curve was obtained. Second, by changing the displacement rate and measuring the corresponding plateau load, a velocity value was calculated for each plateau load. While the former was capable of obtaining the dependence of crack velocity versus load from a single test, the latter was found to be simpler and more consistent. Applications were made to a CVD SiO2 system. In both cases of crack propagation either inside the SiO2 layer or along its interface with a TiN layer, the crack growth velocity changed with the stress intensity at the crack tip exponentially. As a result, a small crack will grow larger under essentially any tensile stresses typically existing in devices, provided that chemical agents facilitating stress corrosion mechanisms are also present.

Pure bulk AlN substrates were prepared by hot-pressing to eliminate the influence of an aid-sintering substance on the interface reactions. AlN thin films were deposited on Si(111) substrates to decrease the influence of charging on the analysis of metal/AlN interfaces with x-ray photoelectron spectroscopy (XPS). Thin films of titanium were deposited on bulk AlN substrates by e-gun evaporation and ion beam assisted deposition (IBAD) and deposited on AlN films in situ by e-gun evaporation. Solid-state reaction products and reaction mechanism of the Ti/AlN system annealed at various temperatures and under IBAD were investigated by XPS, transmission electron microscopy (TEM), x-ray diffraction (XRD), and Rutherford backscattering spectrometry (RBS). Ti reacted with AlN to form a laminated structure in the temperature range of 600 °C to 800 °C. The TiAl3 phase was formed adjacent to the AlN substrate, TiN, and Ti4N3−x as well as Ti2N were formed above the TiAl3 layer at the interface. Argon ion bombardment during Ti evaporation promoted the interface reactions. No reaction products were detected for the sample as-deposited by evaporation. However, XPS depth profile of the Ti/AlN/Si sample showed that Ti–N binding existed at the interface between the AlN thin films and the Ti thin films.

Sapphire (α–Al2O3) and Al were joined by means of the surface activated bonding (SAB) method in an ultrahigh vacuum at room temperature. Tensile tests have shown that failure occurred not along the interface but inside the Al bulk near the interface. High resolution transmission electron microscopy has revealed the formation of a direct interface between Al and sapphire, indicating the possibility to artificially fabricate an atomically direct interface of dissimilar materials at room temperature. However, an intermediate layer was partially observed, which might be attributed to the effect of fast atom beam irradiation of the sapphire surface.

A frequently advocated design for the Information Superhighway that will provide voice, video, and two-way digital communication for businesses and residences calls for transmission on optical fibers from the source (head end) to neighborhood nodes and coaxial cable for distribution from the neighborhood nodes to individual subscribers. Materials reliability challenges abound from the head end to the customer's premise. The maximum allowed service outage per customer is expected to be less than one hour per year. The fiber portion of the network, including lasers, amplifiers, lenses, and the fiber itself, carries two-way transmission to large neighborhoods. With one fiber serving hundreds of customers, cost/reliability tradeoffs emphasize reliability. At neighborhood nodes, where the network branches, coaxial components carry the signal to the street level, where each subscriber is served by at least one drop cable and one network interface unit. The volume of components is much higher. The challenge is to achieve high reliability at low cost. Indeed, one of the impediments to replacing coaxial distribution in the neighborhood with fiber is the difficulty of achieving low-cost, reliable packaging for lasers and other optical components, while maintaining reliability. Since covering all the many materials aspects of system reliability would fill an entire book, the important design considerations, materials issues, and test methods that bear on the reliability of hybrid fiber coaxial systems will be illustrated by reviewing several aspects of coaxial reliability. The intent is to illustrate the methods used to identify and prevent failure mechanisms and the information that must be assembled for making appropriate decisions on reliability/cost tradeoffs. While coaxial systems have been used for over 20 years to bring cable TV to homes across North America, the requirements for reliability and signal quality with these one-way analog rf systems are less demanding than those for two-way digital data transmission. Many of the reliability issues that could potentially degrade performance characteristics of coaxial cable systems are materials and environment related. This paper discusses these issues to the extent that they are currently understood, presents results from some ongoing research to improve our understanding of the potential materials/environment related causes of signal deterioration, describes some prevention strategies, and discusses some of the many materials challenges that remain.