Crack growth measurements in InP crystals were con-Cracks which resulted from this procedure were typically 1/4 to 1/3 the total length of the specimen, and extended through the ducted using an applied-moment, double cantilever beam entire specimen thickness. Precracked specimens were glued to (DCB) geometry. Controlled, smooth crack growth was aluminum arms with a fast-acting commercial glue (Duro observed in water; however, the crack growth velocities Super Glue, Loctite Corp., Cleveland, OH). were not constant for constant applied stress intensity (K I ).The specimens were used in applied-moment DCB 2 tests to More importantly, there was no clear functional depeninvestigate the crack growth behavior of InP in water. Specidence of the crack growth rate (v) on K I . mens were dead-weight loaded and, in the applied-moment DCB geometry, experienced a constant stress intensity for a I. Introductiongiven load, independent of crack length. Crack lengths were monitored through a traveling microscope at a resolution of E NVIRONMENTALLY enhanced crack growth can limit the life-Ϸ6 m. Crack velocities were determined by measuring time of brittle materials. It is important, therefore, to deterchange in crack length over measured time intervals. Only one mine the relationship between crack velocity, in a particular environment, distilled water, was used for the DCB tests. The environment, and the stress intensity at the crack tip. Typically, cracks were immersed in a glass container filled with water quantitative determination of a material's susceptibility to enviduring the measurements. The sides of the container were optironmentally enhanced crack growth is obtained by evaluating cally polished to permit crack length measurements. the crack growth parameter, N, which is defined in the empirical relation between crack growth rate (v) and stress intensity factor (K I ):
III. Results and DiscussionThe InP specimens showed very unusual crack growth behavior. At times, crack velocities were constant over the where A is an adjustable parameter. 1 The larger N is, the less time interval measured; at other times, crack velocities were susceptible the system is to environmentally enhanced fracture. observed either to increase or to decrease at constant K I . For More generally, traditional views of crack growth assume a most loads, however, no crack growth was observed. Figure 2 steady-state process in which v is described by a single parameshows data from a typical specimen; the open circles denote ter, K I , regardless of the specific functional dependence. In this actual crack velocity measurements at different stress intensity communication, we report results of double cantilever beam (DCB) measurements to determine the relationship between crack tip velocity and stress intensity for single-crystal InP in liquid water. InP is a III-V semiconductor of technological interest due to its infrared optical properties. InP has the same structure (zinc blende) and similar bonding to GaAs. It was initially expected, therefore, that th...