Effects of Cerium Addition on Fatigue Fracture Behavior of Magnesium Single Crystals
2019 Volume 60 Issue 8 Pages 1601-1604
Details
2019 Volume 60 Issue 8 Pages 1601-1604
Single crystalline thin sheets of pure Mg and two alloys containing cerium — 0.016 mol% Ce and 0.028 mol% Ce — were subjected to plane bending fatigue tests so as to clarify effects of cerium addition on fatigue fracture behavior of magnesium. Both fatigue limit and fatigue life were found to improve by cerium addition. With increasing cerium concentration, fatigue crack initiation lives shorten, while crack propagation rates decrease. Critical resolved shear stress for second order pyramidal slip decreased by cerium addition, resulting in short fatigue crack initiation lives.
This Paper was Originally Published in J. Japan Inst. Light Metals 69 (2019) 128–130.
For energy efficiency purposes, magnesium alloys application to transport components are anticipated due to low density (thus light weight) relative to their high specific strength. To this end, fatigue fracture behavior of magnesium alloys have recently under study. Such studies have shown that magnesium alloyed with rare-earth elements1) show higher fatigue strength than conventional magnesium alloys.
Magnesium is known to show crystal orientation dependence of deformation behavior because of its hexagonal closed-pack structure with low symmetry. Ando et al.2–4) have reported that fatigue crack propagation mechanism depends on crystal orientation in pure magnesium single crystals using compact tension (CT) tests. Tsushida et al.5) have reported crystal orientation dependence of fatigue life in pure magnesium single crystals using bending fatigue tests.
In magnesium alloys with 0.03 mol% cerium addition,6,7) non-basal slips are easily activated, resulting in great improvement of cold workability. However, fatigue fracture behavior has yet to be elucidated. In this study, magnesium-cerium alloy single crystals were subjected to fatigue tests so as to investigate effects of cerium addition on fatigue fracture behavior.
Magnesium of 99.99% purity and two alloy types, magnesium-0.016 mol% cerium (hereafter, 16Ce) and magnesium-0.028 mol% cerium (hereafter, 28Ce) single crystals were produced using a Bridgman furnace. Figure 1 is a schematic illustration of thin sheet specimens for fatigue tests. Dimensions were 3 × 0.3 × 16 mm3 and the longitude, width and normal directions were parallel to $[11\bar{2}0]$, $[1\bar{1}00]$ and (0001), respectively. A hole with a diameter of 0.5 mm was introduced into each thin sheet specimen using an electric discharge machine to function as a crack initiation point. The thin sheet specimens were then annealed in a thermal cyclic to remove strain induced from cutting and polishing. For each cycle, the polished specimens were annealed at two temperatures, 673 K and 723 K; they were held for 3.5 ks at both temperatures with the rate of change between temperatures 6.9 × 10−3 K/s. A total of 8 annealing cycles were applied. After annealing, the specimens were soaked in a 20 ml C2H5OH + 7 ml H2O2 + 5 ml HNO3 solution for several seconds to remove oxide layers. A metal washer with 0.146 g as a weight was attached to each specimen to increase the amplitude. Plane bending fatigue tests for thin sheet specimens5) were used in this study. In the test, thin sheet specimens were fixed to an acoustic speaker to provide resonance which resulted in cyclic bending stress being loaded at the hole in the specimen. Stress amplitude, σa, was calculated using the amplitude applied to thin sheet specimens.5) Plane bending fatigue tests were carried out at a frequency of 192–406 Hz at room temperature with a stress ratio of R = −1. The fatigue tests were stopped at 108 cycles, prior to specimen failure. Crack lengths on normal planes of thin sheet specimens in the fatigue tests were measured with an optical microscope.
Schematic illustration of thin sheet fatigue specimen used in this study.
Figure 2 shows S-N plots indicating the relationship between stress amplitude and fatigue life. In this study, for specimens which had not failed by 108 cycles, the stress amplitude, σa, was defined as the fatigue limit. Fatigue limit of pure magnesium was approximately 70 MPa. On the other hand, 28Ce had fatigue limits between 75 and 90 MPa, meaning that cerium addition increased fatigue limit by approximately 10 MPa. When the stress amplitude was 100 MPa, fatigue life of 16Ce was slightly longer than that of pure magnesium and 28Ce showed the longest fatigue life, indicating that fatigue life lengthens with increasing cerium addition.
S-N plots of Mg–Ce and pure magnesium single crystals.
Figure 3 shows optical micrographs of crack profiles. As shown in Fig. 3(a), many lines consisting of micro-cracks (hereafter, micro-crack rows) were observed around the hole just after loaded at a strain amplitude of 120 MPa in 28Ce. Most of them were along a direction inclined 30 degrees from $[1\bar{1}00]$. A crack initiated at a micro-crack row around the hole; however, the initial crack crossed micro-crack rows and then macroscopically propagated along $[1\bar{1}00]$. At a stress amplitude of 90 MPa, the crack propagated along $[1\bar{1}00]$ although the number of micro-crack rows decreased, as shown in Fig. 3(b). Micro-crack rows were also observed in pure magnesium when the stress amplitude was 118 MPa (Fig. 3(c)). In addition, cracks propagated along $[1\bar{1}00]$ just after the initiation in pure magnesium, similar to 28Ce, while some cracks propagating along micro-crack rows were also observed. Conversely, when the stress amplitude was 87 MPa, most cracks propagated along micro-crack rows, as shown in Fig. 3(d). These results indicate that cracks tend to propagate along $[1\bar{1}00]$ by cerium addition.
Optical micrographs of crack profiles of Mg–0.028 mol%Ce ((a) and (b)) and pure magnesium single crystals ((c) and (d)).
Figure 4 shows SEM images of fatigue fracture surfaces of 28Ce and pure magnesium single crystals. Despite the stress amplitude, striation-like patterns parallel to $[1\bar{1}00]$, which is the crack propagation direction, were observed around crack initiation sites on fracture surfaces in both 16Ce and 28Ce (Fig. 4(a) and (b)), and pure magnesium (Fig. 4(c) and (d)) single crystals. The striation-like patterns were widely observed in both 16Ce and 28Ce. On the other hand, striation-like patterns were seldom observed in pure magnesium single crystals under low stress amplitude.
SEM images of fatigue fracture surfaces of Mg–0.028 mol%Ce ((a) and (b)) and pure magnesium single crystals ((c) and (d)).
Figure 5 shows fatigue crack lengths, a, of (a) pure magnesium and 16Ce, and (b) 28Ce, as a function of the number of fatigue cycles, N. Fatigue crack growth rates, da/dN, were estimated from the slopes: 0.019 µm/cycle in pure magnesium (σa = 87 MPa); 0.014 µm/cycle in 16Ce (σa = 100 MPa); and 0.0036 µm/cycle in 28Ce (σa = 100 MPa). da/dN were found to decrease with increasing cerium addition. Here, two specimens were used for 16Ce: open circles (○) and open squares (□). In the specimen drawn as open circles, the da/dN between the first and second plots differ from that between the second and third plots, meaning the crack rapidly propagated between the second and third plots. The plot thus indicates the mean values of the slope between the first and second plots drawn by open circles and the slope between five plots drawn by open squares in Fig. 5(a) were used as da/dN in 16Ce.
Surface crack lengths of (a) pure magnesium and Mg–0.016 mol%Ce single crystals, and (b) Mg–0.028 mol%Ce single crystals, as a function of fatigue cycle.
Figure 6 shows surface crack lengths, a, as a function of cycle ratio (N/Nf). Here, N/Nf refers to the number of cycles to failure divided by cycle number. Surface crack lengths of each sample show a linear relationship with cycle ratio but no constancy with each stress amplitude. Also, the slope of the lines decreased with decreasing cerium addition. Here, contact points of the slope and cycle ratio (i.e., horizontal axis) indicate the absence of surface cracks and would be defined as fatigue crack initiation ratio against fatigue life. Fatigue crack life is expected to shorten due to cerium addition. Therefore, cerium addition facilitates crack initiation and but slows the da/dN, resulting in that Mg–Ce alloys show long fatigue life.
Surface crack lengths as a function of fatigue life ratio (N/Nf).
Based on the above results, effects of cerium addition on fatigue fracture of magnesium are discussed. Pure magnesium and 28Ce fractured under stress amplitudes of 70 MPa and 90 MPa, as shown in Fig. 2. Due to second order pyramidal slips, Mg–0.052 mol%Ce and pure magnesium single crystals yielded at approximately 50 MPa8,9) and 70 MPa10) in $\langle 11\bar{2}0\rangle $ tensile tests whose loading direction was identical to that in the present fatigue tests. Since cerium contents of Mg–Ce alloys used in this study are approximately half of those in the Refs. 8–10), the yield stress of the present Mg–Ce alloys would be lower than that of pure magnesium. Therefore, stresses higher than that required for activation of second order pyramidal slips were loaded in the present fatigue tests, resulting in Mg–Ce alloys fracturing due to second order pyramidal slip.
Striation-like patterns parallel to $[1\bar{1}00]$ were observed on fracture surfaces in both 16Ce and 28Ce, as shown in Fig. 4(a) and (b). Ando et al.3) have reported that similar fracture surfaces were observed in pure magnesium single crystals after CT tests in which crack propagation direction was parallel to $[1\bar{1}00]$ and have proposed slip-off mechanism of $\{ 11\bar{2}2\} \langle \bar{1}\bar{1}23\rangle $ second order pyramidal slip for crack propagation. Observed results of crack propagation direction and fracture surface in the present fatigue tests indicate that an identical mechanism likely occurs. In addition, Hayashi et al.8,9) have performed $[11\bar{2}0]$ tensile tests of Mg–0.052 mol%Ce alloy single crystals and discussed the cause of the decreasing yielding stress by cerium addition as described below; cerium addition decreases stacking fault energy of second order pyramidal planes, decreasing the frequency of the immobilization of (c+a) dislocations. They found that critical resolved shear stress for second order pyramidal slip decreases with the addition of cerium. It was, therefore, assumed that fatigue cracks probably initiate due to second order pyramidal slip, similar to the fatigue crack propagation mechanism in pure magnesium single crystals. Second order pyramidal slip would be easily activated as CRSS decreased by cerium addition, which results in a decrease in fatigue crack initiation lives of Mg–Ce alloy single crystals compared with those of pure magnesium. However, fatigue lives of Mg–Ce alloy single crystals were longer than those of pure magnesium. The decreasing fatigue life would result from the occurrence of crack closure phenomenon10) since plastic deformation at crack tips is facilitated due to decreasing CRSS for second order pyramidal slip by cerium addition. That is, the occurrence of crack closure is suggested to degrade effective stress intensity factor range at crack tips, decreasing the da/dN significantly. The effects of cerium addition on crack initiation mechanism and crack propagation mechanism will be discussed in the near future.
Two types of magnesium-cerium (Mg–Ce) alloy single crystals, Mg–0.016 mol%Ce and Mg–0.028 mol%Ce, and pure magnesium single crystals were subjected to fatigue tests to clarify the effects of cerium addition on fatigue fracture behavior of magnesium. Cerium addition was found to increase both fatigue limit and fatigue life of magnesium. In pure magnesium single crystals, cracks propagated parallel to $[1\bar{1}00]$ at lower stress amplitude and along micro-crack rows inclined 30 degrees from $[1\bar{1}00]$ at higher stress amplitude. Conversely, cracks propagated parallel to $[1\bar{1}00]$ in Mg–0.028 mol%Ce despite the stress amplitude. Addition of cerium to pure magnesium was found to shorten fatigue crack initiation lives and to slow crack propagation rates.
