Mechanics of Materials
Effects of Cerium on Crystal Orientation Dependence of Fatigue Fracture Behavior of Magnesium Single Crystals
2022 Volume 63 Issue 1 Pages 27-32
Details
2022 Volume 63 Issue 1 Pages 27-32
Fatigue tests of single crystalline Mg–0.026 at%Ce were carried out to investigate the effects of cerium on crystal orientation dependence of fatigue lives. Three crystal orientations were investigated: B specimen with $[1\bar{1}00](0001)$, D specimen with $[11\bar{2}0](1\bar{1}00)$ and E specimen with $[0001](11\bar{2}0)$. Fatigue lives increased by the cerium addition in D and E crystal orientations, but decreased in B crystal orientation. The cerium additions decreased critical resolved shear stress (CRSS) for second order pyramidal slip, resulting in ridge and stripe (RS) patterns which were observed on fracture surfaces in B specimen. A fatigue crack propagation mechanism was proposed to account for the RS patterns.
Fig. 9 SEM images of (a) fatigue fracture surface of B specimen of Mg–Ce single crystals at a stress amplitude of 120 MPa, (b) ridge and stripe (RS) patterns and (c) coarse grooves in the center of B specimen.
Magnesium alloys have been widely studied for use in transportation industries mainly due to their high specific strength. Fatigue properties of magnesium alloys are also an important barometer for such practical uses. Pure magnesium has a hexagonal close-packed (hcp) crystal structure with low symmetry. It shows a strong crystal orientation dependence of mechanical properties as reported by Ando et al.1–5) using pure magnesium single crystalline specimens. Crystal orientation dependence of fatigue crack propagation and fatigue lives were studied using compact tension (CT) tests1–3) and plane bending fatigue tests,4,5) respectively.
Recently, additions of rare earth elements to magnesium has been studied to improve mechanical properties of magnesium alloys; they showed high fatigue strength compared to commercial magnesium alloys.6) Chino et al.7,8) reported that the 0.03 mol% addition of cerium significantly improved cold workability of magnesium alloys. Inokuchi et al.9) performed fatigue tests of Mg–0.028 mol%Ce single crystals and reported that the addition of cerium increases fatigue lives of magnesium. However, the investigation of fatigue lives was limited to one crystal orientation only. Mg–Ce alloys with other crystal orientations should be investigated regarding their effects on fatigue properties since pure magnesium single crystals show crystal orientation dependence of fatigue properties5) as described above. In this study, Mg–Ce single crystals with three different crystal orientations were applied to fatigue tests in order to investigate effects of cerium additions on crystal orientation dependence of fatigue properties.
Mg–Ce single crystals were grown using a split graphite mold by the modified Bridgeman method. The one-side mold carved with grooves 0.6 mm in depth and 3 mm in width were used for the production of thin sheet specimens of single crystals. Single crystals were crystallographically analyzed by the X-ray back reflection Laue method, and were then cut with a non-distortion cutting machine using a stainless wire and nitric acid. Figure 1 shows schematic illustration of the dimensions and crystal orientations of Mg–Ce single crystalline specimens for fatigue tests; the crystal orientations of the longitudinal direction and surface plane are different: B specimen with $[1\bar{1}00](0001)$, D specimen with $[11\bar{2}0](1\bar{1}00)$ and E specimens with $[0001](11\bar{2}0)$. Dimensions were 3 × 0.3 × 16 mm3. B, D and E specimens were determined to be Mg–0.027 mol%Ce, 0.024 mol%Ce and 0.026 mol%Ce by ICP emission spectroscopy, and the mean cerium quantity was 0.026 mol%. As a crack initiation site, a hole with a diameter of 0.5 mm was introduced into each specimen by electric discharge machining to function. 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 3.6 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. Annealed specimens were chemically polished to remove oxide layers using a chemical polishing solution (HNO3:H2O2:C2H5OH = 5:7:20). A holder and a metal weight (0.146 g) were attached at each specimen using an instant glue. The weight was required to increase the amplitude of specimens in fatigue tests.
Schematic illustration of dimensions and crystal orientations of Mg–Ce single crystalline specimens for fatigue tests: B, D and E specimens.
A plane bending fatigue test for thin sheet specimen developed by Tsushida et al.4) was used in this study. In the fatigue test, a thin sheet specimen with the holder is attached at the center of a voice coil of a loudspeaker and then oscillates by the resonance, resulting in that cyclic stress loads at the edge of the hole of the specimen. Stress amplitude, σa was calculated using the amplitude of the bottom of specimens.4) Fatigue tests were carried out until approximately 5 × 107 cycles at a stress ratio R = −1 at room temperature in ambient atmosphere. Frequency of the cyclic stress was between 45 and 246 Hz. Crack lengths on specimen surfaces were measured using an optical microscope during fatigue tests. Fracture surfaces were observed by scanning electron microscopy (SEM) after fatigue tests.
Figure 2(a) shows S-N plots of B, D and E specimens. D specimen had the longest fatigue life; the fatigue life of Mg–Ce single crystals were strongly affected by the crystal orientation. While fatigue limit of D specimen was approximately 90 MPa, that of B and E specimens was lower and approximately 80 MPa. For comparison, S-N plots of pure magnesium single crystals with the same crystal orientations as B, D, and E specimens are shown in Fig. 2(b)–(d). While fatigue lives of D and E specimens increased by cerium addition as shown in Fig. 2(b) and (c), that of B specimen tended to decrease as shown in Fig. 2(d). In three crystal orientations, E specimen of pure magnesium and B specimen of Mg–Ce showed the shortest fatigue life. Cerium addition was found to change the crystal orientation dependence of the fatigue life.
(a) Fatigue S-N plots of Mg–0.026 at%Ce single crystals. Fatigue S-N plots of Mg–Ce and pure Mg single crystals are compared in three different orientations: (b) D specimen, (c) E specimen and (d) B specimen.
Figure 3 shows optical micrographs of fatigue cracks around the hole in D specimen of Mg–Ce single crystals at stress amplitudes of (a) 100 MPa and (b) 140 MPa. Fatigue cracks indicated by white arrows initiated at the edge of the hole with basal slips and macroscopically propagated along [0001] with zigzag patterns. Several lines inclined approximately 47° from [0001] were observed around the hole. Similar lines were observed in the previous study.9) Each line consisted of micro cracks (hereafter, a micro crack row). The micro crack rows resulted from $\{ 10\bar{1}2\} $ twins which occurred by the loading of the compression stress in fatigue tests. However, the lines seemed to not affect fatigue crack propagation behavior. Figure 4 shows SEM image of the fatigue fracture surface of D specimen of Mg–Ce single crystals at a stress amplitude of 140 MPa. Coarse stripe patterns perpendicular to [0001] were observed in the entire fatigue fracture surface. Directions of the stripe patterns and slip lines on specimen surfaces were parallel to basal planes, implying that fatigue cracks must be propagated with basal slips. Similar stripe patterns were observed in the same fatigue tests5) and CT tests3) of pure magnesium single crystals.
Optical micrographs of fatigue cracks around holes in D specimen at stress amplitudes of (a) 100 MPa and (b) 140 MPa.
SEM image showing the fatigue fracture surface of D specimen at a stress amplitude of 140 MPa.
Figure 5 shows the relationship between crack length, a, and the number of cycles, N, of D specimen. When the data of Mg–Ce and pure magnesium were compared at a stress amplitude of 120 MPa, the cerium additions were found to delay fatigue crack initiation. Also, fatigue crack growth rates were compared using the slope of the plot, and the fatigue crack growth rate of Mg–Ce was slightly slower than that of pure magnesium. The slower rate results in long fatigue lives of D specimen.
Surface crack length of D specimens as a function of number of cycles.
Figure 6 shows optical micrographs of fatigue cracks around the holes in E specimen at stress amplitudes of (a) 100 MPa and (b) 80 MPa. Many micro crack rows inclined approximately 43° and 23° from $[1\bar{1}00]$ occurred around the holes just after the onset of fatigue tests. The micro crack rows resulted from $\{ 10\bar{1}2\} $ twins occurred by the loading of the tensile stress in fatigue tests. When the stress amplitude was more than 100 MPa, fatigue cracks initiated between micro crack rows with different inclinations; however, propagated independent to micro crack rows but parallel to basal planes, as shown in Fig. 6(a). When the stress amplitude was 80 MPa, fatigue cracks propagated parallel to basal planes simultaneously with the formation of micro crack rows. Also, basal slip lines were observed around fatigue cracks as shown in Fig. 6(b). Figure 7 shows SEM image of fatigue fracture surface of E specimen at a stress amplitude of 80 MPa. Flat fracture surfaces parallel to basal planes were observed near specimen surfaces and seemed to not correspond to micro crack rows observed on the normal plane. Therefore, the formation of micro crack rows barely affects crack propagation behavior inside the specimen. On the other hand, complex fracture morphology was observed at the middle part of the specimen; the formation mechanism is unclear.
Optical micrographs of fatigue cracks around holes in E specimen at stress amplitudes of (a) 100 MPa and (b) 80 MPa.
SEM image of fatigue fracture surface of E specimen of Mg–Ce single crystals at a stress amplitude of 80 MPa.
Figure 8 shows optical micrograph of the fractography of B specimen at a stress amplitude of 100 MPa. Many micro crack rows parallel to $[11\bar{2}0]$ occurred around the hole right after the onset of fatigue tests in B specimen, and then fatigue cracks propagated along micro crack rows. The same propagation was also observed under other stress amplitude conditions. Figure 9 shows SEM images of fracture surfaces of B specimen at a stress amplitude of 120 MPa. Fracture morphology near specimen surfaces differed from that inside the specimen, as shown in Fig. 9(a). Fracture surface areas indicated by white arrows inclined approximately 43° from [0001]. In the areas, ridges in white perpendicular to the crack propagation direction and tilted stripe patterns between ridges were observed, as shown in Fig. 9(b). Similar fracture surfaces were observed after uniaxial fatigue tests of pure magnesium single crystals at low stress amplitudes,10) and it was termed the ridge and stripe (RS) pattern. The RS pattern was observed in the area ranging from specimen surfaces to approximately 100 µm. On the other hand, grooves parallel to $[11\bar{2}0]$ were observed in the middle parts of the specimen, as shown in Fig. 9(c). Fatigue crack propagation mechanism was found to be different between the near specimen surface and the inside of the specimen.
Optical micrograph of B specimen of Mg–Ce single crystals at a stress amplitude of 100 MPa.
SEM images of (a) fatigue fracture surface of B specimen of Mg–Ce single crystals at a stress amplitude of 120 MPa, (b) ridge and stripe (RS) patterns and (c) coarse grooves in the center of B specimen.
Mg–Ce single crystals were found to show a strong relationship between the crystal orientation dependence and fatigue fracture behavior, similar to pure magnesium single crystals. However, effects of cerium addition were different between specimens, and the cause will be discussed.
Fatigue lives of D specimen increased by cerium addition as shown in Fig. 2(b). In D specimen, fatigue cracks propagated with basal slips (Fig. 3) and stripes perpendicular to the crack propagation direction (Fig. 4) were observed in the whole of the fracture surface. Similar cracks were observed after CT tests of pure magnesium single crystals. Therefore, fatigue cracks propagate in D specimen of Mg–Ce based on the same fatigue crack propagation mechanism.3) The schematic illustration of the fatigue crack propagation mechanism is shown in Fig. 10. Basal slips are activated by cyclic loading, resulting in the crack initiation. Shear stress loads along basal planes at the crack tip, and then other basal slips are activated. Many basal dislocations occurred from the crack tip moves some distance, piles up and stops. As a result, cracks perpendicular to basal planes initiate by stress concentration at the crack tip and propagate some distance. Basal slips are activated at the crack tip again. The repeating forms the fracture surface shown in Fig. 4. That is, fatigue crack propagation depends on basal slip activity in D specimen. Therefore, to investigate critical resolved shear stress (CRSS) for basal slip, Mg–0.049 at%Ce single crystals were also prepared and the tensile tests were performed using the same specimen shape as the fatigue specimen. As a result, the CRSS for basal slip was 0.8–1.1 MPa and was higher than approximately 0.5 MPa11) reported in pure magnesium single crystals. Cerium addition was found to increase CRSS for basal slip. Therefore, cerium addition decreases the basal slip activity and delays crack initiations. Furthermore, fatigue crack propagation rates become slow. As a result, D specimen shows the longest fatigue life.
Schematic illustration of the fatigue crack propagation mechanism of D specimen of Mg–Ce single crystals.
Inokuchi et al.9) evaluated fatigue properties of A specimen with $[11\bar{2}0](0001)$ of Mg–0.028 at%Ce single crystals using the same plane bending fatigue test methodology and reported that the fatigue life of A specimen increases by cerium addition. Stripe patterns parallel to [1100] which is the crack propagation direction were observed in A specimen. Similar patterns were observed after crack propagation with $\{ 11\bar{2}2\} \langle \bar{1}\bar{1}23\rangle $ second order pyramidal slip (SPCS) mechanism in CT tests of pure Mg single crystals. Therefore, they concluded that fatigue crack propagates in A specimen based on SPCS mechanism. Figure 11(a) shows schematic illustration of fatigue crack mechanism of A specimen. Two SPCSs are activated alternately, and then fatigue cracks propagate with forming stripe patterns along $[1\bar{1}00]$, the line of intersection of two SPCSs. CRSS for SPCS decreased by cerium addition,12) leading that the activation of SPCS becomes more easily at the crack tip. As a result, the crack closure phenomenon easily occurs and crack propagation rates decreases by decreasing stress intensity factor. Therefore, fatigue lives increased in A specimen.9)
Schematic illustrations of fatigue crack propagation mechanism of (a) A specimen with $[11\bar{2}0](0001)$ of pure magnesium single crystals and (b) B specimen with $[1\bar{1}00](0001)$ of Mg–Ce single crystals.
B specimen had the normal plane of (0001) and fatigue crack propagation parallel to basal planes, similar to A specimen. However, cerium addition decreased fatigue lives in B specimen. While RS patterns formed near specimen surfaces on fracture surfaces in B specimen as shown in Fig. 9, similar fracture surfaces were not reported in A specimen of Mg–0.028 at%Ce.9) On the other hand, stripe patterns parallel to the crack propagation direction on fracture surfaces of B specimen in pure magnesium,5) similar to A specimen9) in Mg–Ce. Figure 11(b) shows schematic illustration of fatigue crack mechanism of B specimen in Mg–Ce. Stripe patterns at the middle part of specimens are caused by SPCS mechanism, similar to A specimen shown in Fig. 11(a). On the other hand, RS patterns were observed near specimen surfaces. We proposed the fatigue crack propagation mechanism10) through uniaxial fatigue tests of pure magnesium single crystals, as described below. $\{ 10\bar{1}2\} $ twins and SPCS are activated in specimens by cyclic stress. Here, SPCS dislocations pile up at twin boundaries, and then void arrays form along twin boundaries. Void arrays grow and connect each other by cyclic stress, resulting in the formation of RS patterns on fracture surfaces. Resolved shear stress (RSS) on $\{ 10\bar{1}2\} $ twins in B specimen was higher than that in A specimen since the loading direction was parallel to $[1\bar{1}00]$ in B specimen. Therefore, $\{ 10\bar{1}2\} $ twins are easily activated in B specimen. Also, Schmid factor for SPCS with respect to the loading direction in B specimen is lower than that in A specimen; RSS decreases in B specimen. However, cerium additions decrease CRSS for SPCS, and fatigue cracks easily propagate with RS patterns near specimen surfaces. As a result, fatigue lives decreases in pure magnesium. Effects of cerium additions on $\{ 10\bar{1}2\} $ twins are under investigation in separate research.
In E specimen, fatigue cracks propagated along [0001]. Brittle and flat fracture surfaces were observed near specimen surfaces, but not regular morphology corresponding to repeating fatigue stress. Also, gaps inclined approximately 30° from $[1\bar{1}00]$ were observed on the fracture surface, but it was not regular. In [0001] uniaxial fatigue tests of pure magnesium single crystals,10) regular stripes and RS patterns were observed at higher and lower stress amplitude, respectively. The loading direction of E specimen in this study was also [0001]. However, the fracture surface significantly differs from that in the uniaxial fatigue tests. Therefore, we found that the fatigue crack propagation mechanism depends on the loading stress in fatigue tests.
Three different Ce% (0.024, 0.026 and 0.027 mol%) of Mg–0.026 at%Ce single crystals with different crystal orientations were applied to fatigue tests so as to investigate effects of cerium additions on crystal orientation dependence of fatigue fracture behavior. Cerium additions increased fatigue lives in D specimen with $[11\bar{2}0](1\bar{1}00)$ and E specimen with $[0001](11\bar{2}0)$. On the other hand, cerium additions decreased fatigue lives in B specimen with $[1\bar{1}00](0001)$. Crystal orientation dependence of fatigue lives in Mg–Ce was found to differ from that in pure magnesium. Cerium additions increased CRSS for basal slip system, resulting in slower crack propagation rates in D specimen. Moreover, cerium additions decreased CRSS for second order pyramidal slip. As a result, B specimen showed ridge and stripe patterns after fatigue tests.
A portion of the present study was financially supported by “The Light Metal Educational Foundation, Inc.” The authors are very grateful for the support.
