The influence of growth rate and temperature on high cycle fatigue of directionally solidified AI-AI(sub 3)Ni eutectic alloy

Abstract

Crack initiation and propagation mechanisms were established by metallographic and fractographic examination. Crack initiation occurred at intense slip offsets which formed on the matrix surface and were supplemented by local fiber failures. Crack propagation occurred primarily in a zone of matrix material adjacent to fibers. Crack initiation and propagation in rapidly solidified Al-Al3Ni occurred along grain and colony boundaries.

Dislocation substructure-fiber interactions were studied by transmission electron microscopy. The dislocation cell size and the amount of dislocation substructure generated during high cycle fatigue was found to be a function of interfiber spacing. In 11 cm/hr material, the fiber spacing was smaller than the mutual trapping distance of the matrix dislocations and very little dislocation substructure was formed during fatigue. Extensive cellular substructure was formed in material grown at 0.5 and 3.5 cm/hr, apparently because the fiber spacing was greater than the mutual trapping distance of the dislocation. Differences in the substructure generated were used to interpret the hysteresis responses of the individual growth rates.

High-cycle fatigue tests have been conducted on specimens of the Al-Al3Ni eutectic alloy unidirectionally solidified at selected rates from 0.5 cm/hr to 1200 cm/hr. Tests were conducted in air at 298°K, 458°K, and 683°K. Room temperature fatigue lives were independent of growth rate at low solidification rates (0.5 - 30 cm/hr), but markedly improved in samples grown at 1200 cm/hr. Materials grown at 30 cm/hr exhibited fatigue lives similar to those of lower growth rate materials, despite gross misalignment due to cellular growth. At 0.5 T(sum) (458°K) and 0.75 T(sub m) (683°K), the fatigue lives of mm material grown at low solidification rates were dependent on growth rate. The dependence of fatigue life on growth rate at elevated temperatures appears to result primarily from differences in cyclic creep rates produced by varying interfiber spacings.