Behavior of cyclic fatigue cracks in monolithic silicon nitride

Cyclic fatigue-crack propagation behavior in monolithic silicon nitride is characterized in light of current fatigue-crack growth models for ceramics toughened by grain-bridging mechanisms, with specific emphasis on the role of load ratio. Such models are based on diminished crack-tip shielding in the crack wave under cyclic loads due to frictional-wear degradation of the grain-bridging zone. The notion of cyclic crack growth promoted by diminished shielding is seen to be consistent with measured (long-crack) growth rates, fractography, in situ crack-profile analyses, and measurements of back-face strain compliance. Growth rates are found to display a much larger dependence on the maximum applied stress intensity, Kmax, than on the applied stress-intensity range, ΔK, with behavior described by the relationship da/dN ∝ Kmax29 ΔK. Fatigue thresholds similarly exhibit a marked dependence on the load ratio, R = Kmin/Kmax; such effects are shown to be inconsistent with traditional models of fatigue-crack closure. In particular, when characterized in terms of Kmax, growth rates below ∼ 10-9 m/cycle exhibit an inverse dependence on load ratio, an observation which is consistent with the grain-bridging phenomenon; specifically, with increasing R, the sliding distance between the grain bridges is decreased, leading to less frictional wear, and hence less degradation in shielding, per loading cycle. The microstructural origins of such behavior are discussed