High-Temperature Plastic-Deformation Behavior of Mg–(Y/Zn) Supersaturated Solid-Solution Alloys and the Resulting Dislocation Structures

Abstract

High-temperature plastic deformation behavior and the resulting microstructures of magnesium (Mg)-based supersaturated solid-solution alloys containing zinc (Zn) and/or yttrium (Y) have been thoroughly examined in a comparative study by means of various electron microscopy combined with microanalytical techniques. According to the results of compression tests measured for the alloys at constant testing temperatures ranging from room temperature (RT) to 300°C, it is admittedly found in common to the respective alloys that twinning of $\{ 10\bar{1}2\} $-tensile type dominates the deformation at lower temperatures but this gives way to dislocation-slip with a rise in temperature. Above all, Mg–Y–Zn ternary solid-solution alloys yield remarkably higher levels of flow stresses capable of withstanding high temperatures than binary counterparts. The solid-solution alloy of Mg–0.6Y–0.3Zn (at%) subjected to compression at 300°C, in fact, has many deformation-induced stacking-faults on the (0001) basal planes significantly decorated by Y/Zn-solute segregation, providing the definite evidence that Suzuki effect actually contributes to a substantial enhancement of the flow stresses at elevated temperatures. This study demonstrates that the Suzuki effect is measurably activated in Mg-based solid-solution alloys, especially when those containing an adequate amount of combined solutes of Y and Zn, e.g. 0.6∼1 at%Y and 0.3∼0.5 at%Zn, are plastically-deformed at a temperature of 300°C and at a strain rate of 1.0 × 10⁻³ s⁻¹.

Fig. 12 Atomic resolution HAADF-STEM images showing dislocation structure typically found in the 300°C-compressed alloy of Mg–0.6Y–0.3Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence.

1. Introduction

The interaction between dislocations and solute atoms provides a basis for understanding a broad range of phenomena in mechanical metallurgy. Cottrell and Bilby, in their classic work of 1949, first proposed that the idea of the segregation of carbon atoms in alloy steels, having a propensity to form atmospheres around dislocations (Cottrell atmosphere), could be used to account for the yielding and strain aging behavior experimentally observed.¹⁾ Soon after, Suzuki argued that dissociated dislocations could interact with solute atoms or impurities by means of adsorption on the stacking fault (SF) ribbon.^2,3) When solute atoms accumulate at the SF, the SF energy in equilibrium decreases with increasing solute content (adsorption). This effect manifests itself as a change in the width of an extended dislocation accompanied by adsorption (Suzuki-effect or Suzuki-segregation). Experimental verification of Cottrell-atmosphere as well as Suzuki-segregation, however, has long been a challenging task. It had, in fact, taken several decades after the theories were first proposed until such advanced microscopy and/or microanalytical techniques as energy dispersive X-ray spectroscopy (EDS) combined with a field-emission gun, high-angle annular detector darkfield scanning transmission electron microscopy (HAADF-STEM) and three-dimensional atom probe (3DAP) contributed successfully to provide the data, which were taken as direct evidence for the effects.^4–9)

In the research field of the development of lightweight structural materials, it was a sensational discovery that a dual-phase alloy with a nominal composition of Mg₉₇Y₂Zn₁, which Kawamura et al.¹⁰⁾ produced by means of rapidly solidified powder metallurgy processing, had exhibited a record high of tensile yield strength of over 600 MPa (with an elongation of 5%) at room temperature above those of the rest of Mg-based alloys existing before. Subsequent studies attributed such excellent mechanical properties as possessed by the Mg₉₇Y₂Zn₁ alloy to the dual presence of extremely fine grains of the α-Mg matrix phase and an intermetallic precipitate phase of long-period stacking (LPS) structure synchronized by segregation of Y/Zn-solutes.¹¹⁾ Much attention has, thus far, been directed especially at elucidating real crystal structures of LPS phases occurring with different polytypes^12–15) and assessing their effects on mechanical properties.^16–21) It is remarkable that the mechanical strength of such an Mg₉₇Y₂Zn₁ dual-phase alloy is not so much degraded with an increase of testing temperature, and in fact, the typical strength measured at 300°C for the alloy is allegedly slightly below that recorded at room temperature.²¹⁾ This sharply contrasts with the case of the other metals or alloys where a substantial increase of softening takes place at elevated temperatures. It is undoubted that LPS phases play an important role in strengthening such dual-phase alloys. However, the LPS phases alone are not an essential microstructural component responsible for the superior strength at elevated temperatures, since they occupy only minor parts of the whole materials. As of now, there is increasing evidence to indicate that the α-Mg matrix phase itself, which is coexistent with LPS phases, contributes much to strengthening of Mg–Y–Zn alloys especially at a certain range of temperatures.

In recent years, Suzuki et al.^22–25) demonstrated that high temperature creep strength of Mg–Y alloys is considerably improved by the addition of Zn, giving evidence pointing to the possible interpretation: Mg solid-solution alloys containing both Y and Zn, when subjected to creep test at certain elevated temperatures, had more planar faults and extended dislocations generated on the (0001) basal planes than in the case without Zn. They claimed further that a combined addition of Y and Zn to Mg induced a progressive decrease of the SF energy and hereby made dislocation motions confined in the basal planes by the influence of the Suzuki-segregation, therefore accounting for the improvement of creep strength observed. Meanwhile, Hiraga et al.⁸⁾ made thorough investigations on deformation microstructure of a Mg₉₇Y₂Zn₁ dual-phase alloy, which had been subjected to hot extrusion at 350°C, providing the first direct evidence for the Suzuki-effect observable for both the α-Mg matrix phase and the LPS phase by means of HAADF-STEM: the α-Mg matrix phase, after the extrusion, had many basal a-dislocations introduced and dissociated, also making the resultant SF-ribbons decorated by Y/Zn-segregation. Whereas, the LPS phase after the extrusion had many Y/Zn-rich SF-regions locally dissociated and hereby changed back to the regular HCP-lattice, conversely making the localized solutes expelled from the corresponding local regions. Either case, whether it corresponds to adsorption or desorption of Y/Zn taking place on the SF-ribbon, is typical of the Suzuki effect.^2,3) Besides, Yang et al.⁹⁾ applied aberration-corrected STEM technique capable of performing the ultimate atomic-scale imaging to examine crystal defects in an Mg₉₇Y₂Zn₁ dual-phase alloy plastically-deformed at 300°C, and they documented in the most detail that the deformed α-Mg matrix had three types of perfect dislocations, such as 60° a-, screw a-, and a + c-dislocations, dissociated into two partials and additionally made the resultant SF-ribbons segregated with Y/Zn-solutes. These observations allowed them to estimate the SF energy of their Mg–Y–Zn alloy to be in the range of 4.0–10.3 mJ m⁻².

The Suzuki-effect occurs for certain alloy systems, but not all. In case of Mg-alloys, it has become widely recognized that Mg–Y–Zn is the system capable of having the Suzuki-effect activated during plastic deformation made at certain elevated temperatures. However, a controversy still exists as to what metallurgical and/or physical parameters are a key factor to optimize the performance of the Suzuki effect in Mg–Y–Zn alloys, and further, how much this effect alone, excluding any effect by LPS phases, contributes to enhance the strength capable of withstanding high temperatures. Thorough investigations targeted at single-phase alloys consisting only of Mg–(Y/Zn) supersaturated solid-solution, by means of advanced electron microscopy, are therefore an essential prerequisite for resolving the problems addressed above. The present study focuses on the description of both high temperature compression behavior and the resulting deformation microstructures of nearly single-phase alloys as addressed above, in an attempt to quantitatively measure the real effect of Suzuki-segregation on the flow stresses varied with elevated temperatures.

2. Experimental Procedures

According to Lee et al.,²⁶⁾ a master alloy with a nominal composition of Mg–1.4Y–0.7Zn (at%) had a certain dual-phase microstructure developed after solution treatment set at 520°C, consisting of both a major part of α-Mg supersaturated solid-solution with an average composition of Mg–1.1Y–0.6Zn and a minor part of (Y, Zn)-rich second phase. In the present study, we prepared master alloys with even less amounts of solutes and made them solution-treated at a lower temperature of 500°C than the conditions selected by Lee et al., in an attempt to satisfy both effects of the supersaturated solid-solution and the suppression of grain coarsening. Four different master ingots of dilute Mg-based alloys having nominal compositions of Mg_99.4Zn_0.6 (at%), Mg_99.4Y_0.6, Mg_99.1Y_0.6Zn_0.3 and Mg_98.5Y₁Zn_0.5 were each prepared from a mixture of high-purity metals of Mg (99.99%), Zn (99.99%) and master alloys of Mg–15 mass%Y by induction heating under an Ar gas in a graphite crucible (as-solidified). The as-solidified alloys were subsequently subjected to a fixed pattern of thermomechanical treatments as illustrated in Fig. 1, in order to make them developed to single-phase alloys of the α-Mg supersaturated solid-solution with the respective composition. The 1st step in Fig. 1 is hot rolling operated at a working temperature of 400°C and with a thickness reduction ratio of 30%, where the reduction per one pass is approximately 0.25 mm (as-rolled). The as-rolled alloys were hereby supplied as plates with a thickness of 3 mm. The following annealing process is two-step: the 1st annealing set to 400°C for 1 hour is intended for the effect of recovery and recrystallization but not allowing annealing-twins (as-annealed), while the 2nd annealing set to 500°C for 1 hour followed by water-quenching is for the effective solution treatment while suppressing grain coarsening (as-solution). Each of the as-solution alloys was machined to a rectangular parallelepiped with dimensions of 3 × 3 × 5 mm³, so that the longest side was parallel to the direction of hot-rolling (RD) subjected. The surfaces of the parallelepipeds were polished by an emery paper. The rectangular parallelepipeds so prepared were supplied for the following compression tests.

Fig. 1

Specimen processing method for all the as-solidified alloys.

The compression tests were executed by using a load testing machine equipped with a thermostatic chamber capable of controlling ambient temperatures (INSTRON 5985), where the loading axis was set in parallel to RD of the specimen direction and the testing temperature was set to one of the following: room temperature (RT), 100°C, 200°C and 300°C. After holding one specimen in the chamber for 10 min at a prescribed temperature, the specimen was compressed at a strain rate of 1.0 × 10⁻³ s⁻¹ until the specimen had approximately 7% nominal strain imposed, and afterwards it was unloaded and moved out of the chamber immediately (as-compressed) to minimize thermo-driven microstructural evolution after deformation. Such alloy specimens as described above, having different compositions and different processing histories, are hereafter denoted as the ‘300°C-compressed alloy of Mg–0.6Y–0.3Zn’ and so on.

Microstructural investigations of the as-solution alloys as well as the as-compressed alloys were made for cross sections perpendicular to RD of the specimen direction by means of various electron microscopy combined with microanalytical techniques. Scanning electron microscopy (SEM; JEOL JSM-7800F), electron back-scattered diffraction analysis (EBSD; HITACHI SU-70 equipped with OXFORD INSTRUMENTS HKL-CHANNEL5), transmission electron microscopy having a scan image observation function (TEM/STEM; JEOL JEM-2100F) were applied to examine the deformation microstructures. Specimens for SEM/EBSD and TEM/STEM were cut from the alloys of interest and thinned by mechanical polishing and completed by ion-milling. All the TEM/STEM images as well as selected area diffraction (SAD) patterns presented in this paper were recorded along the zone axis of $[2\bar{1}\bar{1}0]$ direction of the α-Mg HCP-structure.

3. Results and Discussions

Figure 2 shows SEM back-scattered electron images obtained from three different as-solution alloys of (a) Mg–0.6Y (at%), (b) Mg–0.6Y–0.3Zn and (c) Mg–1Y–0.5Zn. RD, ND and TD indicated here are respectively the rolling-, the normal- and the transverse-direction of the specimens defined when the corresponding as-solidified alloys were subjected to hot rolling. These alloys have been identified as nearly single-phase alloys: the α-Mg supersaturated solid-solution phase with the respective composition occupies most of the material, which is each recognized as dark-grey contrast in Fig. 2, and additionally no more than trace of (Y, Zn)-rich precipitate in bright contrast is present. It should be noted here that the microstructural features of the other alloys not shown here are all qualitatively similar.

Fig. 2

SEM back-scattered electron images of the as-solution alloys which were prepared by solution treatment set at 500°C for 1 hour followed by water quenching: (a) Mg–0.6Y; (b) Mg–0.6Y–0.3Zn; (c) Mg–1Y–0.5Zn. RD, ND and TD are the rolling-, the normal- and the transverse-direction of the specimens which are defined when the respective as-solidified alloys were subjected to hot rolling.

Figure 3 illustrates inverse pole figure (IPF) maps which were created for four different as-solution alloys as well as a pure-Mg, i.e., (a) pure-Mg; (b) Mg–0.6Zn; (c) Mg–0.6Y; (d) Mg–0.6Y–0.3Zn; (e) Mg–1Y–0.5Zn, where the colour is composed of image quality and a colour mode for the crystal direction parallel to RD of the specimen direction. Note that the specimen of pure-Mg was also prepared by the processing method shown in Fig. 1. Morphological features such as grain size and grain shape, grain boundaries, twin boundaries, crystal grain orientations and so on, can be deduced from the maps. A clear difference has been found in the average grain sizes calculated among all the specimens: (a) 126 µm for pure-Mg, (b) 138 µm for Mg–0.6Zn, (c) 103 µm for Mg–0.6Y, (d) 67 µm for Mg–0.6Y–0.3Zn and (e) 76 µm for Mg–1Y–0.5Zn. These calculation values are not a total surprise, since a fixed pattern of thermomechanical treatments prescribed made the specimens with different compositions recrystallized differently and therefore the resultant microstructures showed up with differences of the average grain size. It is notable that the ternary alloy specimens particularly have small crystal grains of about a half as large as those of the pure-Mg or the binary ones. Obviously, combined addition of Y and Zn has an appreciable effect on the refinement of crystal grains of Mg solid-solution alloys. Otherwise all the specimens have rather similar microstructural and textural features: they are almost a single-phase alloy of the α-Mg supersaturated solid-solution with the respective composition, allowing very little amounts of the second phase and annealing-twins to coexist. In regard to the textural feature, every map displays a roughly similar combination of colours such as yellow-green, blue or light-blue, indicating in common that the basal planes of their crystal grains mostly lie in nearly parallel to RD of the respective specimen, i.e., roughly a basal texture. The specimens having such microstructural and textural features as described above were supplied for the following compression tests.

Fig. 3

IPF-maps which are created for four different as-solution alloys as well as a pure-Mg, where the colour is composed of image quality and a colour mode for the crystal direction parallel to RD of the specimen direction: (a) pure-Mg; (b) Mg–0.6Zn; (c) Mg–0.6Y; (d) Mg–0.6Y–0.3Zn; (e) Mg–1Y–0.5Zn. RD, ND and TD are the rolling-, the normal- and the transverse-direction of the specimens which were defined when the respective as-solidified alloys were subjected to hot rolling.

The effects of alloying elements and deformation temperature on flow stresses in compression were investigated using pure-Mg and its diluted solid-solution alloys containing Y and/or Zn. Figure 4 shows typical stress-strain curves measured for the respective alloys presented in Fig. 3. In general, a level of flow stresses of solid-solution alloys increases with a rise in solute concentration and decreases with a rise in testing temperature, and these common features are actually evident from the present results. When simply comparing the levels of flow stresses recorded at a fixed temperature for all the specimens, they are ranked according to their magnitudes in the following order: pure-Mg < Mg–0.6Zn < Mg–0.6Y < Mg–0.6Y–0.3Zn < Mg–1Y–0.5Zn. Such a direct comparison, however, may not be an accurate measure to evaluate the effects of alloying elements on the solid-solution hardening, since there is also another important factor to be taken into account, that is, the effect of grain sizes/boundaries which typically manifests itself as a change in the yield strength according to the Hall-Petch relation. It was addressed above that there are rather large differences in the average grain sizes among all the as-solution alloys, which are ranged between 67 µm and 138 µm, amounting maximally to a nearly two-fold difference. Such differences in grain size must have been reflected to some extent to the yielding behaviors. According to a number of relevant previous studies,^27–30) however, the differences in the levels of flow stresses presently observed seem to be well beyond the range of the differences explainable by grain size effects. In other words, the effects of grain size on the stress-strain curve become rather small especially after yielding and the plastic-flow parts in the curves are considered to be reliable enough to compare the effects of alloying elements on Mg solid-solution hardening. Therefore, it can be deduced from Fig. 4 that the addition of Y to Mg have a larger solid-solution hardening effect than that of Zn (compare Figs. 4(b) and 4(c)) and the effect is more pronounced when a combined addition of Y and Zn is made. Comparison of Figs. 4(c) and 4(d) allows us to estimate that the addition of 0.3 at%Zn to Mg–0.6 at%Y makes the level of flow stresses at RT increased by 30% approximately, and further, when the testing temperature is increased to be as high as 300°C, the level reaches more than 50%-increase. These observations also agree well with the recent reports by Suzuki^22–25) that creep strength of Mg is substantially improved by alloying with Y and Zn. The solid-solution alloy of Mg–1Y–0.5Zn, indeed, gave the highest performance of all in flow stresses capable of withstanding high temperatures. It is also remarkable that the flow curve measured for Mg–1Y–0.5Zn at 300°C is accompanied with continuously serrated plastic-flow, which is known as the Portevin-Le Chatelier effect. This characteristic behavior may also be a manifestation of Suzuki-segregation as will be discussed later.

Fig. 4

Stress-strain curves obtained from compression tests executed at various temperatures ranging from RT to 300°C for the pure-Mg and the four different as-solution alloys: (a) Pure-Mg; (b) Mg–0.6Zn; (c) Mg–0.6Y; (d) Mg–0.6Y–0.3Zn; (e) Mg–1Y–0.5Zn.

The IPF-maps created for the five different specimens after the compression tests made at RT as well as 300°C are presented in Fig. 5 and 6, respectively. It should be remembered that the as-solution alloys, regardless of alloy composition, allowed most of their crystal grains to be displayed in rather ‘cool-tones’ of colour such as blue or yellowish green (see Fig. 3). After the compression test made at RT, a dramatic change has appeared in the colour-tones: the RT-compressed alloys have their many grains as well as some interior regions of the grains displayed in rather hot-tones of colour such as red, orange and pink (see Fig. 5). Such colour changes recognized in the IPF-maps, as typically seen from blue (or yellowish green) to red, indicate changes of crystal orientation taking place in many grains during compression by about 90°. It is most apparent in Figs. 5(a) and 5(b) that there are many strip-regions displayed in red which are sandwiched by the surroundings in blue or yellowish green, showing a typical microstructural feature of deformation twinning, most of which are identified as $\{ 10\bar{1}2\} $-tensile type. These observations admittedly confirm that, at a testing temperature of RT, twinning contributes much to the deformation in compression. The classical theory of crystal deformation³⁾ explains that dislocation-slip and twinning, sometimes grain boundary sliding to a lesser extent,³¹⁾ can contribute to the plastic deformation of HCP-structure materials. At lower temperatures, deformation twinning is essential to satisfy the Von-Misses criterion established for homogeneous deformation of HCP-polycrystals. The role of twinning on the plastic deformation behavior has abundantly been investigated and reported during the past several decades. Among all microstructural parameters conceivable, grain size is allegedly the critical factor for deformation twinning: the occurrence frequency of deformation twinning increases with an increasing grain size.²⁹⁾ As is evident from Fig. 5, the RT-compressed specimens such as those of pure-Mg, Mg–0.6Zn and Mg–0.6Y, which had had larger grains with over 100 µm on average before compression, have allowed deformation-twinning to abundantly occur within the grains, while the ternary alloys, whose grains had been smaller by an approximate factor of 0.5 than the others, have had less twins. Twin boundaries in crystals, whether they are annealing- or deformation-twins, can function as barriers to dislocation movement, and so the increase of twins is basically equivalent to the decrease of grain size. Thus, all the specimens, after subjected to yielding in compression at RT, could have had the grain size/boundary effects almost comparable with each other, not making measurable differences in contribution to the flow stresses.

Fig. 5

IPF-maps created for five different RT-compressed specimens, where the colour is composed of image quality and a colour mode for the crystal direction parallel to RD of the specimen direction: (a) Pure Mg; (b) Mg–0.6Zn; (c) Mg–0.6Y; (d) Mg–0.6Y–0.3Zn; (e) Mg–1Y–0.5Zn. RD, ND and TD are the rolling-, the normal- and the transverse-direction defined when the respective as-solidified alloys were subjected to hot rolling.

Fig. 6

IPF-maps created for five different 300°C-compressed specimens, where the colour is composed of image quality and a colour mode for the crystal direction parallel to RD of the specimen direction: (a) Pure-Mg; (b) Mg–0.6Zn; (c) Mg–0.6Y; (d) Mg–0.6Y–0.3Zn; (e) Mg–1Y–0.5Zn. A diagram indicated by (f) shows misorientation angle plotted against distance along the linear path indicated by a white arrow present in the upper left part of (e), where several twin boundaries lie perpendicular to the path.

As for the deformation behavior of the 300°C-compressed specimens, the resulting IPF-maps, as presented in Fig. 6, have their most grains to remain unchanged in blue or yellowish green as they did before compression (compare Figs. 3 and 6). A close inspection, however, allows us to recognize a crucial difference between two groups of the as-compressed specimens: the RT-compressed specimens have deformation-twins to occur widespread in the crystals, while the 300°C-compressed specimens also allow them to occur but in much smaller numbers and scale. Figure 6(f) is a diagram of misorientation angle plotted against distance along the linear path indicated by a white arrow in the upper left part of Fig. 6(e), where several twin boundaries lie perpendicular to the path. Then, the misorientation angles calculated at the respective twin boundaries are mostly found to be around 86°, confirming that the twin boundaries formed by the compression are of $\{ 10\bar{1}2\} $-tensile type. This is agreeable with what was deduced from Fig. 5. In this way, the EBSD analyses presently conducted have confirmed that twinning of $\{ 10\bar{1}2\} $-tensile type dominates the deformation during the compression made at RT, but at a testing temperature of 300°C, the twinning gives way to dislocation-slip. Grain boundary sliding is also another cause for the deformation but its contribution is probably negligibly small, since crystal grains comprising of the present specimens are too coarse for grain boundary sliding to become active.³¹⁾ We applied TEM/STEM for the further investigation to identify a unique deformation microstructure which could account for the higher levels of flow stresses experimentally determined for the ternary solid-solution alloys.

Deformation microstructures of the as-compressed alloys of Mg–0.6Y–0.3Zn, Mg–0.6Zn and Mg–0.6Y were examined on the atomic scale by means of TEM/STEM. The observations focused on local regions including SF-ribbons formed between two partial dislocations. TEM-Brightfield (BF)- as well as BF-STEM imaging can make a sharp contrast on the faulted region, responding sensitively to strain fields accumulated at the fault, while HAADF-STEM imaging can also make a similar contrast for the fault but by responding far more sensitively to atomic numbers (Z) of constituent elements of the region. Figure 7 is a set of TEM/STEM results obtained from the RT-compressed alloy of Mg–0.6Y–0.3Zn, which were recorded in the $[2\bar{1}\bar{1}0]$ incidence for a local region including deformation-induced dislocations: a TEM-BF image together with the corresponding SAD pattern for (a), a BF-STEM image for (b) and a HAADF-STEM image for (c). Dark contrast features appearing similarly in Figs. 7(a) and 7(b), which exhibit as line segments running perpendicular to the c-axis of the α-Mg HCP-structure as well as curved segments, are assumed to be due to basal perfect a-, non-basal perfect a- or their dissociated partial-dislocations. On the other hand, the corresponding HAADF-STEM image shown in Fig. 7(c) has the defect feature only dimly and/or incompletely exposed with an inverted contrast of light and dark, which are apparently due to diffraction contrast rather than Z-contrast. The 200°C-compressed alloy of Mg–0.6Y–0.3Zn is found to have a qualitatively similar contrast as well. Figure 8 presents an equivalent set of TEM/STEM images showing the resulting dislocation microstructure of the 200°C-compressed alloy of Mg–0.6Y–0.3Zn, where linear-contrast features recognized are likely due to deformation-induced SFs. Here again, characteristic Z-contrast can be hardly recognized at the locations of the SFs in Fig. 8(c). Figure 9 shows high-resolution transmission electron microscopy (HRTEM) images of SFs found in two different as-compressed alloys, which were taken respectively from the RT- for (a) and the 200°C-compressed alloy for (b) in the $[2\bar{1}\bar{1}0]$ incidence. In each micrograph, one of the basal planes of the α-Mg structure along which the deformation-induced SF is present is indicated by a big arrow. The closure failures of the SF/RH Burgers circuits,³⁾ each of which encloses one of the partial dislocations, indicate that they are both associated with the dissociation of either a 60° a- or a screw a-dislocation.⁹⁾ The stacking sequences across the SFs in both images have the same order ⋯ABABCACA⋯. The middle B and C layers have a local face-centered cubic (FCC) environment, which is characteristic of SF of intrinsic 2 (I2)-type. These observations make us claimed that the as-compressed alloys at a testing temperature of 200°C or lower have many basal a-dislocations dissociated but without accompanying Y/Z-segregation on the faulted regions. Hence it follows that a solid-solution alloy of Mg–0.6Y–0.3Zn does not have Suzuki-effect activated during the compression at temperatures of 200°C and lower.

Fig. 7

TEM and STEM images showing dislocation microstructure typically found in the RT-compressed alloy of Mg–0.6Y–0.3Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) TEM/SAD; (b) BF-STEM; (c) HAADF-STEM.

Fig. 8

TEM and STEM images showing dislocation microstructure typically found in the 200°C-compressed alloy of Mg–0.6Y–0.3Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) TEM/SAD; (b) BF-STEM; (c) HAADF-STEM.

Fig. 9

HRTEM images showing extended dislocation structures typically found in the as-compressed alloys of Mg–0.6Y–0.3Zn, which were taken respectively from the RT- for (a) and the 200°C-compressed alloy for (b) in the $[2\bar{1}\bar{1}0]$ incidence.

For the 300°C-compressed alloy of Mg–0.6Y–0.3Zn, HAADF-STEM imaging gave a different contrast feature for its faulted regions. Figure 10 shows an equivalent set of TEM/STEM images showing dislocation microstructure typically found in the 300°C-compressed alloy of Mg–0.6Y–0.3Zn. In Figs. 10(a) and 10(b), there are many deformation-induced SF-ribbons mostly associated with the dissociation of basal a-dislocations. As is evident from Fig. 10(c), the HAADF-STEM image has the SF-ribbons distinguished by making bright Z-contrast accompanied. This demonstrates that the SF-ribbons induced during the compression at 300°C are accompanied by the segregation of heavier solutes of Y and Zn, which is typical of Suzuki-segregation. A number of separate experiments have confirmed that the width of the SF-ribbons, i.e., the distance between corresponding two dissociated partials are ranged from a few tens nm to one hundred nm. There is, in fact, another possible cause to be considered for the formation of SFs accompanying Y/Zn-segregation: since the materials were statically heat-treated at 300°C for a few tens of minutes before compression test, the SFs could then form during the static-aging process. We have actually confirmed by a number of separate experiments that SFs can form during the static-heat treatment but they are apparently different and can be distinguished from the deformation-induced SFs. Figure 11 illustrates typical (a) BF- and (b) HAADF-STEM images taken in the $[2\bar{1}\bar{1}0]$ incidence from the static-aged alloy of Mg–0.6Y–0.3Zn at 300°C for 30 min (without being subjected to compression), showing the resulting SFs in the alloy, where one can find the presence of a type of SFs extending over several hundred nanometers only. It should be emphasized here that the dislocation-induced SFs resulting from the compression test mostly have lengths in the order of a few tens nanometers, while those formed during the static-aging are over several hundred nanometers long. It is also worthwhile pointing out that many dislocations found here have linear shapes and seldom deviate from the basal plane to exhibit curved shapes. These observations imply that as the Suzuki-segregation develops, the dislocations on the spots may become increasingly confined to the basal plane in mobility and hereby discourage themselves from making cross slip or climbing motion, therefore resulting in the high level of flow stresses observed. It is noted here that the morphological features of dislocations described above correspond much to the Suzuki’ observations made for an as-crept alloy of Mg–0.9Y–0.04Zn at 377°C.^24–26) It should also be emphasized that the contrast features of the 300°C-compressed alloy of Mg–1Y–0.5Zn, which are not shown here, is qualitatively similar to those of Mg–0.6Y–0.3Zn, as shown in Fig. 10.

Fig. 10

TEM and STEM images showing dislocation microstructure typically found in the 300°C-compressed alloy of Mg–0.6Y–0.3Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) TEM/SAD; (b) BF-STEM; (c) HAADF-STEM. Note that the dislocation-induced SFs present here can be recognized with bright Z-contrast in the HAADF-STEM imaging.

Fig. 11

STEM images showing SFs formed in the static-aged alloy of Mg–0.6Y–0.3Zn at 300°C for 30 min, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) BF-STEM and (b) HAADF-STEM. Note that the presence of SFs extending over a long distance exceeding several hundred nanometers only, which are indicated with small arrows, can be recognized in the alloy.

Figure 12 shows atomic-resolution HAADF-STEM images recorded at three local regions with the bright contrast present somewhere in Fig. 10(c). The SFs revealed here have exactly the same stacking sequence in common as those revealed in Fig. 9. Besides, the two middle layers including the SF are found to be brighter due to a higher concentration of Y/Zn than the outer layers. Furthermore, there are certain contrast fluctuations irregularly occurring on the atomic scale along the deformation-induced SF-ribbons, which are associated with two-dimensional random distribution of Y/Zn. Such an irregular and/or volatile contrast feature arising from the deformation-induced SFs is clearly different from the contrast from the ‘growth SFs’ which is typified by the words of regular and/or uniform.⁹⁾ It has, thus, been demonstrated that the supersaturated solid-solution alloys of Mg–Y–Zn make Suzuki-effect measurably activated when compressed at 300°C, and then have the SF-ribbons significantly decorated by solutes of Y and Zn. These observations make it highly probable that, in case of solid-solution alloys of Mg–Y–Zn, the Suzuki-segregation is a responsible cause for the high levels of flow stresses capable of withstanding a high temperature of 300°C.

Fig. 12

Atomic resolution HAADF-STEM images showing dislocation structure typically found in the 300°C-compressed alloy of Mg–0.6Y–0.3Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence.

Whereas, the binary solid-solution alloys, regardless of compression temperature, have hardly had the Suzuki-segregation detected by HAADF-STEM imaging. Two sets of TEM and STEM images showing dislocation microstructures found in the 300°C-compressed alloy of Mg–0.6Zn as well as Mg–0.6Y are presented in Figs. 13 and 14, respectively. In common to both alloys, the TEM and the BF-STEM images reveal the presence of many deformation-induced dislocations: some of the dislocations have straight linear shapes extending along the basal plane, and others have certain curved shapes extending out of the basal plane. These morphological features of dislocations are in sharp contrast to those of the 300°C-compressed alloy of Mg–Y–Zn, where the dislocations formed in curved-shapes are rarely seen (compare with Fig. 10). There have been a growing number of experimental results taken as evidence that certain Mg alloys, especially containing rare-earth metals such as Y, show significantly enhanced ductility, which is due to a high activity of compression twining, secondary twinning and further pyramidal a + c-slip.^32–35) Evidence for increased a + c-slip activity has also been given by texture simulation.^36,37) When a testing temperature for the present binary alloys is increased up to 300°C, two different effects of twinning and pyramidal slip like a + c-dislocation are likely in competition for deformation of the α-Mg matrix phase. The increased number of the dislocations formed in curved-shapes recognized in Fig. 12 as well as Fig. 13 is certainly associated with the mass generation of the a + c-dislocation which is supposed to become active with a rise in temperature. In any case, however, no characteristic Z-contrast appears on the faulted regions in the respective HAADF-STEM images. It is, thus, concluded that Suzuki effect is measurably activated in Mg-based solid-solution alloys, especially when those containing an adequate amount of combined solutes of Y and Zn are plastically-deformed at an elevated temperature of 300°C and at a strain rate of 1.0 × 10⁻³ s⁻¹. These observations allow us to claim that as the Suzuki-segregation develops, the dislocations on the spots are effectively confined to the basal planes in mobility and increasingly locked in the planes, thereby discouraging themselves from making cross slip or climbing motion. This interpretation accounts well for the assumption that the Mg-based solid-solution alloys consisting of not only Mg–Y–Zn but also the other systems capable of allowing LPS-phases to precipitate can yield remarkably high levels of flow stresses even at elevated temperatures.

Fig. 13

TEM and STEM images showing dislocation microstructure typically found in the 300°C-compressed alloy of Mg–0.6Zn, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) TEM/SAD; (b) BF-STEM; (c) HAADF-STEM.

Fig. 14

TEM and STEM images showing dislocation microstructure typically found in the 300°C-compressed alloy of Mg–0.6Y, which were taken in the $[2\bar{1}\bar{1}0]$ incidence; (a) TEM/SAD; (b) BF-STEM; (c) HAADF-STEM.

4. Conclusions

High-temperature deformation behavior and the resulting microstructures of Mg–(Y/Zn) supersaturated solid-solution alloys were thoroughly examined by means of advanced electron microscopy combined with microanalytical techniques. The results are summarized as follows:

1. (1)    Regardless of alloy composition, twinning of $\{ 10\bar{1}2\} $-tensile type dominates the deformation at lower temperatures. Crystal grain size is an important factor for the occurrence frequency of deformation twinning: the alloys consisting of grains ranging from 100 µm to 140 µm in an average diameter allow the deformation-twins to abundantly occur within the grains during compression at RT, while those of smaller grains by an approximate factor of 0.5 have appreciably less twins. With a rise in testing temperature, twinning increasingly gives way to dislocation-slip.
2. (2)    Mg–Y–Zn ternary solid-solution alloys yield remarkably higher levels of flow stresses capable of withstanding high temperatures than the binary counterparts. The alloying element of Y has a comparatively larger effect on Mg solid-solution hardening than that of Zn and also the effect is more pronounced when a small amount of combined solutes of Y and Zn is added. Under the compression temperature of RT, the addition of 0.3 at%Zn to Mg–0.6 at%Y alloy makes the level of flow stresses increased by over 30%, whereas in case of 300°C, the level reaches more than 50% increase.
3. (3)    The solid-solution alloy of Mg–0.6Y–0.3Zn, after the compression at 300°C, has its extended dislocations significantly decorated by Y/Zn-segregation, providing the definite evidence that Suzuki-effect actually contributes increasingly to the strength of the alloy with a rise in temperature. The Suzuki effect is measurably activated in Mg solid-solution alloys, especially when those containing an adequate amount of combined solutes of Y and Zn, e.g. 0.6∼1 at%Y and 0.3∼0.5 at%Zn, are plastically deformed at an elevated temperature of 300°C and at a strain rate of 1.0 × 10⁻³ s⁻¹. Continuously-serrated plastic-flow observed in the stress-strain curve of the Mg–1Y–0.5Zn solid-solution alloy during the compression at 300°C is certainly associated with the Suzuki-effect.

Acknowledgements

This work was supported by JSPS KAKENHI Grant-in-Aid for Scientific Research (C), Grant Number 19K05075 from MEXT, Japan. It was also supported by AMADA FOUNDATION Grant Number AF-2018011.

REFERENCES

• 1)   A.H.  Cottrell and  B.A.  Bilby: Proc. Phys. Soc. London Sect. A 62 (1949) 49–62.
• 2)   H.  Suzuki: Sci. Rep. Res. Inst. Tohoku Univ. A4 (1952) 455–463.
• 3)  J.P. Hirth and J. Lothe: in Theory of Dislocations, Second Edition, (A Wiley-Interscience Publication, Hoboken, 1982) pp. 517–518, pp. 17–24.
• 4)   T.  Kamino,  Y.  Ueki,  H.  Hamajima,  K.  Sasaki,  K.  Kuroda and  H.  Saka: Philos. Mag. Lett. 66 (1992) 27–31.
• 5)   J.  Wilde,  A.  Cerezo and  G.D.W.  Smith: Scr. Mater. 43 (2000) 39–48.
• 6)   G.W.  Han,  I.P.  Jones and  R.E.  Smallman: Acta Mater. 51 (2003) 2731–2742.
• 7)   Y.  Koizumi,  T.  Nukaya,  S.  Suzuki,  S.  Kurosu,  Y.  Li,  H.  Matsumoto,  K.  Sato,  Y.  Tanaka and  A.  Chiba: Acta Mater. 60 (2012) 2901–2915.
• 8)   K.  Hiraga,  A.  Yasuhara and  K.  Saito: Mater. Trans. 53 (2012) 1385–1390.
• 9)   Z.  Yang,  M.F.  Chisholm,  G.  Duscher,  Xi.  Ma and  S.J.  Pennycook: Acta Mater. 61 (2013) 350–359.
• 10)   Y.  Kawamura,  K.  Hayashi,  A.  Inoue and  T.  Masumoto: Mater. Trans. 42 (2001) 1172–1176.
• 11)   E.  Abe,  Y.  Kawamura,  K.  Hayashi and  A.  Inoue: Acta Mater. 50 (2002) 3845–3857.
• 12)   T.  Itoi,  T.  Semiya,  Y.  Kawamura and  M.  Hirohashi: Scr. Mater. 51 (2004) 107–111.
• 13)   M.  Matsuda,  S.  Ii,  Y.  Kawamura,  Y.  Ikuhara and  M.  Nishida: Mater. Sci. Eng. A 393 (2005) 269–274.
• 14)   Y.M.  Zhu,  A.J.  Morton and  J.F.  Nie: Acta Mater. 58 (2010) 2936–2947.
• 15)   E.  Abe,  A.  Ono,  T.  Itoi,  M.  Yamasaki and  Y.  Kawamura: Philos. Mag. Lett. 91 (2011) 690–696.
• 16)   K.  Hagihara,  N.  Yokotani and  Y.  Umakoshi: Intermetallics 18 (2010) 267–276.
• 17)   K.  Hagihara,  A.  Kinoshita,  Y.  Sugino,  M.  Yamasaki,  Y.  Kawamura,  H.Y.  Yasuda and  Y.  Umakoshi: Acta Mater. 58 (2010) 6282–6293.
• 18)   M.  Yamasaki,  K.  Hashimoto,  K.  Hagihara and  Y.  Kawamura: Acta Mater. 59 (2011) 3646–3658.
• 19)   K.  Hagihara,  A.  Kinoshita,  Y.  Fukusumi,  M.  Yamasaki and  Y.  Kawamura: Mater. Sci. Eng. A 560 (2013) 71–79.
• 20)   K.  Hagihara,  T.  Okamoto,  H.  Izuno,  M.  Yamasaki,  M.  Matsushita,  T.  Nakano and  Y.  Kawamura: Acta Mater. 109 (2016) 90–102.
• 21)   X.H.  Shao,  Z.Q.  Yang and  X.L.  Ma: Acta Mater. 58 (2010) 4760–4771.
• 22)   M.  Suzuki,  H.  Sato,  K.  Maruyama and  H.  Oikawa: Mater. Sci. Eng. A 319–321 (2001) 751–755.
• 23)   M.  Suzuki,  T.  Kimura,  J.  Koike and  K.  Maruyama: Scr. Mater. 48 (2003) 997–1002.
• 24)   M.  Suzuki,  T.  Kimura,  J.  Koike and  K.  Maruyama: Mater. Sci. Eng. A 387–389 (2004) 706–709.
• 25)   M.  Suzuki and  K.  Maruyama: Mater. Sci. Forum 638–642 (2010) 1602–1607.
• 26)   J.  Lee,  K.  Sato,  T.J.  Konno and  K.  Hiraga: Mater. Trans. 50 (2009) 222–225.
• 27)   A.  Jain,  O.  Duygulu,  D.W.  Brown,  C.N.  Tome and  S.R.  Agnew: Mater. Sci. Eng. A 486 (2008) 545–555.
• 28)   B.  Raeisinia and  S.R.  Agnew: Scr. Mater. 63 (2010) 731–736.
• 29)   H.  Somekawa and  T.  Mukai: Mater. Sci. Eng. A 561 (2013) 378–385.
• 30)   B.Q.  Shi,  Y.Q.  Cheng,  X.L.  Shang,  H.  Yan,  R.S.  Chen and  W.  Ke: Mater. Sci. Eng. A 743 (2019) 558–566.
• 31)   J.  Koike,  R.  Ohyama,  T.  Kobayashi,  M.  Suzuki and  K.  Maruyama: Mater. Trans. 44 (2003) 445–451.
• 32)   A.  Galiyev,  R.  Kaibyshev and  G.  Gottstein: Acta Mater. 49 (2001) 1199–1207.
• 33)   S.  Sandlöbes,  S.  Zaefferer,  I.  Schestakow,  S.  Yi and  R.  Gonzalez-Martnez: Acta Mater. 59 (2011) 429–439.
• 34)   S.  Sandlöbes,  M.  Friák,  S.  Zaefferer,  A.  Dick,  S.  Yi,  D.  Letzig,  Z.  Pei,  L.-F.  Zhu,  J.  Neugebauer and  D.  Raabe: Acta Mater. 60 (2012) 3011–3021.
• 35)   S.  Sandlöbes,  M.  Friák,  J.  Neugebauer and  D.  Raabe: Mater. Sci. Eng. A 576 (2013) 61–68.
• 36)   S.R.  Agnew,  M.H.  Yoo and  C.N.  Tome: Acta Mater. 49 (2001) 4277–4289.
• 37)   S.R.  Agnew,  L.  Capolungo and  C.A.  Calhoun: Acta Mater. 82 (2015) 255–265.