ENVIRONMENT-ASSISTED CRACKING OF HIGH-STRENGTH MAGNESIUM ALLOYS WE43-T6 A. Ahmad and T. J. Marrow* School of Materials, University of Manchester, M1 7HS, UK. *Corresponding author: james.marrow@manchester.ac.uk ABSTRACT Notched specimens of WE43-T6 alloy (4.2 wt% Y, 2.3 wt% Nd, 0.7 wt% Zr, 0.8 wt% HRE, bal. Mg) were static- fatigue tested in ambient air. The statistical distributions of stable cracks initiated at the notch root below the failure stress and the clusters of intergranular intermetallic phase have been compared. The strain at the notch root, as a function of applied stress, was measured using Electronic Speckle Pattern Interferomery (ESPI). The crack density increases with notch root strain, and at high stress, the crack population distribution changes due to crack coalescence. The critical event in failure of the static fatigue specimen, approaching the ultimate tensile strength, is deduced to be nucleation of a single large unstable crack nucleus, formed by coalescence of micro- cracks within clusters of intermetallic phase. Key Words: magnesium, static-fatigue, crack coalescence, ESPI, statistics. Introduction Some magnesium alloys experience environment assisted cracking (EAC) in ambient atmospheric air, and hydrogen embrittlement is the dominant mechanism [1-3]. The highest susceptibility has been found in aluminum-containing alloys (6% max) [4-5]. EAC has also been observed in pure magnesium [3]. The EAC sensitivity of aluminum-free rare earth-containing alloys, such as WE43 has not been previously investigated in detail. Previous work has shown that plastic strain nucleates EAC in WE43-T6 by cracking of the intergranular intermetallic [6]. This paper reports a quantitative study of the crack nucleation mechanism in WE43-T6. This is part of a project that aims to predict the probability of nucleation and propagation of cracks at notches and micro- shrinkage porosity in cast engineering materials. Experimental The aim of the experiments was to observe the effect of microstructure on crack initiation and propagation under static fatigue conditions. WE43 alloy was supplied by Magnesium Elektron Ltd. as cast plates, measuring 200 x 200 x 25 mm and peak aged to the T6 condition. The test specimens were machined from heat-treated blanks. The alloy composition is given in Table 1. The average grain size of the material, measured by the intercept method [7] was 93 ± 12 µm (standard deviation). Static fatigue tests (i.e. constant load) were performed using cylindrical notched tensile specimens (FIGURE 1) to determine the time to failure and the failure probability as a function of nominal applied stress for 100 hours run-out. The majority of specimens (67%) failed in less than 1 minute, indicating that once initiated at high stress, approaching the ultimate tensile strength, EAC crack propagation was rapid. Stable crack nuclei were observed in specimens that did not fail within the run-out period of 100 hours. Six such specimens were analysed to determine the statistical distribution of surface crack nuclei length at the notch root. It had previously been shown that the cracks in the run-out specimens were stable, and did not increase in length on re-loading the sample [6]. The number and lengths of the surface cracks at the notch root for each specimen were measured using scanning electron microscopy. The cracks tended to occur in aligned clusters. Metallographic observations have indicated that the observed surface cracks are formed by coalescence of individual microcracks, initiated by fracture of the intermetallic phase (FIGURE 3) [6]. The distribution of the distance between adjacent surface cracks is given in FIGURE 3c. The mean of the distribution is 4 µm. Those with tips that were separated by less than 4 µm were considered to form a single crack, and their total end-to-end length was measured. The crack lengths were found to follow lognormal distributions (FIGURE 4) for all the nominal stress levels. The crack length distributions were found by ANOVA and Kruskal-Wallis method hypothesis testing of the data (Table 2) to originate from the same population, with a certainty of 95%, for applied stress below 279 MPa. The cracks observed at 279 MPa were concluded not to come from the same population as those observed at lower stress.