Effect of twinning, slip, and inclusions on the fatigue anisotropy of extrusion-textured AZ61 magnesium alloy

Revision as of 16:34, 1 August 2014 by Caleb (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search



In this study, experiments were conducted to quantify structure-property relations with respect to fatigue of an extruded AZ61 magnesium alloy using a MultiStage Fatigue (MSF) model. Experiments were conducted in the extruded and transverse directions under low and high cycle strain control fatigue conditions. The cyclic behavior of this alloy displayed varying degrees of twinning and slip depending on the strain amplitude as observed in the hysteresis loops of both directions. Under low cyclic conditions, asymmetrical stress strain response was observed for both orientations. However, systematic stabilization of the hysteresis occurred by half-life due to subsequent twinning and detwinning mechanisms. In addition, under high cycle fatigue, pseudo-elasticity was observed at the first and at half-life cycles. Structure-property relations were quantified by examining the fracture surfaces of the fatigued specimens using a scanning electron microscope. In terms of crack incubation, fatigue cracks were found to initiate from intermetallic particles (inclusions) that were typically larger than the mean size. Quantified sources of fatigue crack incubation, microstructurally small cracks, and cyclic stress–strain behavior were correlated to the MSF model. Based on the specific material parameters, the MSF model was able to predict the difference in the strain-life results of the AZ61 magnesium alloy in the extruded and extruded transverse directions including the scatter of the experimental results. Finally, the MSF model revealed that the inclusion size was more important in determining the fatigue life than the anisotropic effects from the texture, yield, and work hardening.


Fig. 4. Tensile true-stress versus true-strain behavior for the extruded magnesium AZ61 alloy in the extrusion direction (ED) and (b) extrusion transverse direction (ETD).

As a lightweight metal, magnesium is being intensively sought to be integrated in large mechanical structures for more energy savings and green house emission reductions. This industrial renaissance is reviving wide attention to the complex fundamentals of the effect of materials processing on the fatigue resistance of magnesium alloys. In fact, wrought magnesium alloys present generally higher mechanical strength and fatigue resistance than cast magnesium alloys due to their ability to have smaller inclusion sizes and essentially no porosity, and their precipitation hardening after the severe deformation during the processing conditions. Wrought magnesium alloys, however, present sharper textures over cast magnesium alloys and thus are more anisotropic by nature. Texture has indeed a profound effect on the mechanical response due to the low symmetry of the hexagonal close-packed (HCP) crystal structure.

In this research we characterize the fatigue anisotropy of an extruded AZ61 magnesium alloy and develop a structure-property modeling framework for the fatigue life using the Multistage Fatigue (MSF) modeling approach. In general, characterization of fatigue of magnesium alloys is still in the infancy stage and cur- rent fatigue modeling approaches for wrought magnesium alloys have thus far been purely empirical exercises. Thus, a microstructurally sensitive modeling approach for wrought magnesium alloys is warranted


Fig. 9. Fracture surface of magnesium AZ61 alloy in the extrusion direction (ED). Note the twinning on the surface and the intermetallic particle that initiated the fatigue crack. This specimen was fatigue tested at 0.3% strain amplitude.

Monotonic tensile tests were performed using an MTS servo- hydraulic load frame with a capacity of 25 kN. Fig. 4 displays the stress–strain behaviors under tension along both the ED and the ETD illustrating the anisotropic mechanical behavior. The two orientations exhibited different yield strengths, hardening rates, and ultimate strengths. Specifically, the yield under the ED was 180 MPa and was slightly different compared to the yield strengths of this same alloy reported in other recent work [26,27][1]. These differences could be attributed to the texture intensity and grain size variations in the extrusion processes. The yield stress under the ETD was approximately 85 MPa, which is approximately 50% lower than that under the ED. The difference in the yield stress between the ED and the ETD results from the difference in the critical resolved shear stress (CRSS) of deformation modes activated in either directions. In fact, under the ETD, the Schmid Factor (SF) for twinning was the highest, so twinning was profuse, while under the ED, prismatic <a> and basal <a> elaborate the highest SF with a predominance for the prismatic <a> slip.


Extensive fractography analyses were performed on the fatigue fractured surfaces to measure and chemically identify inclusions that initiated fatigue cracks. Fractured intermetallic particles near the free surface of the specimen were found to be the preferential sites of crack initiation on most fractured surfaces as typified in Fig. 9. Chemical analyses revealed that these particles corresponded to Al8 Mn5 binary compounds. These particles that initiated the fatigue cracks had an average equivalent size (derived from the square root of the particle area) of 8.4 mm for the ED and a 9.4 mm for ETD. Multiple initiation sites were found in a same fracture surface even for specimens tested under high cycle conditions. These cracks would eventually coalesce into a main dominant crack ultimately leading to failure.

Model correlations

Fig. 16. Upper and lower bounds for the Multistage fatigue model for the (a) extrusion direction (ED) and (b) extrusion transverse direction (ETD). These bounds were determined by only varying the intermetallic particle size by the maximum and minimum size observed during the fractography study.

Figures 16a and 16b show the correlations of the MSF model to experimental data for both ED and ETD. The difference in the two directions was due to the difference in grain size (GS), particle size (PS), yield stress (Sy ), ultimate strength (Sult ) and cyclic hardening behavior (K , n ). All of the other material constants were consistent for both loading orientations. This is of considerable importance, because it illustrates the sensitivity of the MSF model to capture the effects of microstructural influences on fatigue life. In this particular case, the fatigue life was found to be most sensitive to the inclusion size related to incubation and secondarily to the cyclic hardening. In an attempt to capture the stochastic nature of the experimental fatigue data, the maximum and minimum particle sizes were employed to obtain error bands in the ED and the ETD.


  1. [26] Z.B. Sajuri, Y. Miyashita, Y. Hosokai, Y. Mutoh, Int. J. Mech. Sci. 48 (2006) 198. [27] A.N. Chamos, S.G. Pantelakis, G.N. Haidemenopoulos, E. Kamoutsi, Fatigue Fract. Eng. Mater. Struct. 31 (2008) 812.

Citation: Effect of twinning, slip, and inclusions on the fatigue anisotropy of extrusion-textured AZ61 magnesium alloy, J.B. Jordon, J.B. Gibson, M.F. Horstemeyer, H. El Kadiri, J.C. Baird, A.A. Luo, Materials Science and Engineering A, 528 (2011) 6860– 6871

Personal tools

Material Models