Microstructure Evolution in Mg_98.6Y₁Zn_0.4 Alloys and the Development by Hot Deformation Examined by Synchrotron Radiation Small- and Wide-Angle Scattering

Abstract

Synchrotron radiation small- and wide-angle scattering measurements have been performed for Mg_98.6–Y₁–Zn_0.4 alloys. In the early stage of phase transformation from supersaturated solid solutions, isotropic scattering suggesting segregation at grain boundaries was observed. It grew with temperature during heating the sample at a constant rate of 0.133 K/s. Above 600 K where introduction of stacking faults is expected, needle-like scattering became visible, which represents platelet shape segregation of cluster layers called cluster arranged layer (CAL). The layer eventually developed to form multiple layers, cluster arranged nano plates (CANaP) at higher temperatures. Microstructure change by hot rolling after the heat treatments has been examined from a viewpoint of kink-deformed microstructures.

Fig. 6 SAXS patterns from the samples hot-rolled after heat treatment, compared with the undeformed one (a). SAXS patterns with the incident beam parallel to the transverse direction (b), (d) and rolling direction (c), (e) are shown. Thickening of the CALs was observed after hot rolling. Triangles at q ∼5 nm⁻¹ are the peak positions of diffuse scattering by in-plane ordering of L1₂ clusters.

1. Introduction

Structure and formation process of long-range stacking ordered (LPSO) alloys have been intensively investigated during recent decades.^1–11) Mg–Y–Zn alloy system is a representative alloy system, which was found in the early stage of research,^1–3) and still the most important model alloy system to examine basic properties of LPSO alloys. Their typical structural features of the crystallographic nature have been investigated for MgAlGd system that gives nearly perfectly ordered LPSO structures.¹¹⁾ The clusters in the completely ordered LPSO structures are reported to have a specific relaxed atomic position, with TM₆RE₈ and an interstitial atom, where TM = Al and RE = Gd for the MgAlGd alloys⁹⁾ and TM = Zn and RE = Y for MgYZn alloys.^10,11) The cluster is hereafter referred as an L1₂ cluster. However, a representative LPSO alloy system of MgYZn was found to be atypical in a sense that the L1₂ clusters in the LPSO structure behaves as a molecular in the stacking faults,^5,7,9,10,11) and distance between the clusters change rather continuously upon annealing.^12,13) Such stacking fault layer with segregation of solute atoms in a form of L1₂ cluster may be regarded as a metastable two dimensional system.^14,15) We will hereafter call this layer ‘segregation layer’. Later examination on MgYZn for wide range of composition and on MgGdZn alloys turned out that there are some group of LPSO alloys with relatively loose cluster-cluster distance in the segregation layer. This may results from the cluster-cluster interaction energy.^14,16) Therefore, it is interesting from a viewpoint of microstructure control of two-dimensional cluster system. As a candidate for light-weight structure materials, Kawamura et al.^17,18) found that dilute MgYZn alloy with rapidly quenched and heat-treated alloys may exhibit microstructures that induce kink strengthening upon hot rolling. Hagihara et al.⁸⁾ proved that such strengthening may apply even for single crystals.

The microstructures of several segregation layers dispersed in Mg matrix, as demonstrated by Kawamura et al.^17,18) and Hagiwara et al.,⁸⁾ are nonequilibrium ones appearing as transient microstructures during heat treatment and deformation processes. Since these microstructures are transient structures rather than a metastable phase, it is important to understand the kinetics and stability of such structures and how such microstructures changes after processing to induce kink deformation. In the present work, we examine the evolution of nanostructures in dilute MgYZn alloys upon annealing and deformation by using small- and wide angle scattering measurements.

2. Experimental Procedure

Small- and wide-angle X-ray scattering (SWAXS) measurements have been made at Beam-line 40B2 of SPring8, and Beam-line 10C and 15A2 of Photon Factory, Japan. Rapidly quenched and hot-pressed Mg_98.6Y₁Zn_0.4 samples were solution treated at 813 K for 72 hours in vacuum before in-situ or ex-situ annealing. To cover appropriate range of the scattering vector, two settings were used for static SWAXS measurements, i.e., the wavelength of 0.092 nm with the camera length of 1 m and the wavelength of 0.15 nm with the camera length of 3 m. In-situ measurements were made with wavelength of 0.1 nm and the camera length of 1 m. Samples were heated in an evacuated chamber at a heating rate of 0.133 K/s from room temperature up to the solution treatment temperature. SAXS measurement with Pilatus 2M and WAXD measurements with Eiger 500 K were performed during heating. For ex-situ measurements, WAXD measurements were made with Pilatus 200 K. Sample preparations used in the present work are shown in Fig. 1. A sample block with the composition of Mg_98.6Y₁Zn_0.4 was prepared by hot-pressing of rapidly quenched ribbons and homogenized at 813 K for 260 ks before quench (route A) and used for in-situ heating measurements. For hot-rolled samples (route B), the quenched samples were annealed at 623 K for 43.2 ks to form CAL/CANaP structures to induce kink deformed microstructure, and then rolled down with accumulated strain of ε = 0.5 or ε = 1.0 at 573 K.

Fig. 1

Sample preparation of Mg_98.6Y₁Zn_0.4 alloy used in the present work. The basic phase transformation kinetics and thermal stability were examined by in-situ measurements. Microstructure change by deformation after heat treatment to form segregation layers was examined by the process B.

3. Results and Discussions

3.1 Typical SWAXS patterns in annealed samples

Figure 2 shows SWAXS patterns of the samples annealed ex-situ, which demonstrates the microstructure change before deformation in the process route B in Fig. 1. SAXS of an as-solution-treated sample shows an isotropic scattering pattern as shown in Fig. 2(a), whose intensity monotonically decrease with the magnitude of the scattering vector, q = 4π sin(θ)/λ, where 2θ is the scattering angle and λ is the wavelength of incident X-rays. This scattering component is apparently different either from the one observed for supersaturated Mg₈₅Y₉Zn₆ alloys in hcp structure whose SAXS pattern shows well-defined diffuse cluster-cluster correlation peak,⁵⁾ or the one after formation of periodic or aperiodic LPSO structures.^6,19,20) In the previous works, in-situ SWAXS measurements during heating amorphous Mg₈₅Y₉Zn₆ samples showed that a strong SAXS component monotonically decreasing with q appeared after crystallization, which implies promotion of segregation at the nanocrystalline hcp grains by grain growth during heating that coexisted with intragrain clusters.⁶⁾ In the present work, however, grain growth during heating to the temperatures below 600 K was reported to be negligible.¹⁷⁾

Fig. 2

Two-dimensional SWAXS patterns of heat treated samples after hot-press and solution treatment. Profiles for as solution-treated for (a) SAXS and (d) WAXD, annealed at 523 K for 3.6 ks for (b) SAXS and (e) WAXD, and 43.2 ks for (c) SAXS and (f) WAXD. Number density of streaks increased by longer annealing time in the SAXS pattern. q_x and q_y are the scattering vector in SAXS region, q in (d), (e), (f) is the scattering vector in the radial direction for WAXD. WAXD patterns give a part of Debye-Scherrer ring.

Figure 2(b) and (c) show SAXS patterns after annealing at 623 K. The SAXS profiles are characterized by many streaks that extend in the radial directions, and a diffuse ring. The peak position of the diffuse ring agreed with the one known as the in-plane cluster-cluster distance in the LPSO structures in more concentrated cast MgYZn alloys,^5,6,19,20) which implies the formation of two-dimension cluster layers similar to those found in three dimensional LPSO phases. The peak position for the cluster correlation, q_m. between 4.85 nm⁻¹ and 5.1 nm⁻¹ for the present case in Fig. 2, however, is smaller than the one observed in the more concentrated alloys, for example, 5.3 nm⁻¹ to 6 nm⁻¹ for Mg₉₇Y₂Zn₁ cast alloys¹⁹⁾ or from 5.6 nm⁻¹ to 6.1 nm⁻¹ for Mg₈₅Y₉Zn₆ alloys.^12,20) Since the average distance between the nearest L1₂ clusters on the segregation layers, L, is given by   

\begin{equation*} L = 2\pi/q_{\text{m}}, \end{equation*}

smaller q_m implies that the in-plane cluster density is lower in the present dilute alloy. In contrast, the streak component is not characteristic for LPSO structures. The positions that a streak appears and disappears vary for each streak, meaning that the positions of the observed streaks do not directly reflect the nanostructure of the sample. The sample should show well-defined peaks reflecting periodicity of the stacking of cluster layers if the microstructure maintains the form of three-dimensional LPSO structure, which appears at 3.4 nm⁻¹ for 14H, 4 nm⁻¹ for 18R, or 4.8 nm⁻¹ for 10H stacking sequence, respectively. Since Figs. 2(b) and (c) show streaks without such peaks, microstructure in the sample is not a three-dimensional LPSO structure, but rather consists of very thin platelets with single or a couple of layers, as defined by Kawamura et al.^17,18) as Cluster Arranged Nano Plates (CANaP).

Figures 2(d)–(f) show two-dimensional wide angle X-ray diffraction (WAXD) patterns that correspond to Figs. 2(a)–(c). A noticeable point is that sharp streaks are observed across specific diffraction spots for the heat treatment conditions where streaks are observed in SAXS patterns. The direction of the streak is almost the same for the spots with the same diffraction index. This clearly indicates that thin layers with scattering contrast and stacking contrast appears on the specific crystallographic planes. Such streaks are not observed for the cast samples with three-dimensional LPSO structures.

3.2 Interpretation of the streaks in the SWAXS patterns

The streaks observed in Fig. 2 are apparently different from the streaks observed for the solution-treated Mg₉₇Gd₂Zn₁ alloys which are cast and then solution-treated and quenched prior to annealing.²¹⁾ The streaks observed for the MgGdZn samples extended from the origin (q = 0) and show a correlation peak corresponding to the periodicity of the segregation layers. For qualitative explanation of the streak components in Figs. 2(b) and (c), model calculations using kinematical scheme were made. Figure 3 schematically explains how a plate-like scattering object are detected in the two-dimensional SAXS detector. Ellipsoidal pattern in the figure show the needle-like form factor of a CANaP drawn with the Ewald sphere. When the plate is parallel to the incoming X-rays, the needle is tangential to the Ewald sphere at origin, giving a needle-like scattering profile symmetric around q = 0 on the sphere. When the plate is not parallel to the incoming X-rays, the intersection of the needle-like form factor with the Ewald sphere gives a spot shown as the case A, or a partial line extending in the radial direction that is similar to a fog box pattern as shown as the case B. This accounts for the streaks appearing and disappearing at the q irrelevant to the characteristic structure of the sample. For more detailed examination, kinematical calculations with a multilayered cluster layers using simple Laue functions are presented. Figure 4(a) and (b) show the SAXS intensity along the streak on the detector, i.e., on the Ewald sphere with a fixed direction of incident beam. The calculation uses the experimental condition of λ = 0.092 nm and the distance between the segregation layer as that of 14H structure. The results suggest that when the incident beam is set to be perpendicular to the plane-normal direction of CANaPs, the correlation peak between the segregation layers can be detectable even for bilayer of the CALs as interference peaks, although the peak height decreased due to the distance from the exact streak position running along q_z axis. The results imply that when the streaks are aligned to be tangential to the Ewald sphere, the multilayered segregation plate should be detectable. Figure 4(c) and (d) show the asymmetric case, where the angle between the direction of the streak in the reciprocal lattice and that for the tangential direction of the Ewald sphere in the scattering plane is defined as ϕ in Fig. 3. The figures show that asymmetric peak position does not gives the interference peak originated from the nanostructure of the sample, and the peak may appear even when the segregation layer consists of a monolayer of clusters. Therefore, it is concluded that the effect of interference from the multilayers can be analyzed only when the streak symmetric with respect to the origin is obtained, which is not the case for Fig. 2, where hot-pressed samples show powder pattern with too dense random streaks. SAXS pattern for a coarse grain sample is necessary for the detailed analysis of streak components, where a set of symmetric streak around q = 0 can be picked up as is the case for MgGdZn,²¹⁾ and left for future work.

Fig. 3

Schematic explanation of observed streaks that appear at the q irrelevant to the layer periodicity and extend in the radial direction. For the streak A, a spot appears at the intersection with the Ewald sphere, while the profile show slight asymmetry for the streak B.

Fig. 4

Streak patterns simulated for several cases in the SAXS region to explain how the curvature of Ewald sphere and the tilt angle ϕ in Fig. 3 affect the measured profiles. The domain size of the CALs are 20 nm for (a) and 100 nm for (b), with the thickness of the plate shown as the number of layers in the figure, with the distance between the layers are that for the 14H LPSO structure, giving the peak position of 3.4 nm⁻¹. It should be noted that the interference peaks are clearly visible for bilayers as shown by the arrows even with the effect of Ewald sphere curvature. Figure 4(c) and (d) show the effect of the tile angle ϕ on the asymmetric profiles of the streak for the bilayer and the monolayer.

3.3 Evolution of SAXS intensity during heating solution-treated samples

To understand the relationship between the as-quenched microstructure and the thin plate nanostructure in Fig. 2, in-situ measurements of solution-treated samples were performed during heating at a constant rate of 0.133 K/s. Figure 5 shows the evolution of radially averaged SAXS profile. At temperatures below 600 K, the profiles are characterized by an isotropic scattering, which decreased monotonically with q, as the profile shown in Fig. 2(a). With the shape of the scattering profile unchanged, the intensity increased and reached maximum at 500 K, and then decreased up to 600 K. Above 600 K, the profiles show development of two peaks as pointed by the arrows, corresponding to the periodicity of 14H LPSO at 3.4 nm⁻¹ and in-plane L1₂ cluster distance at ∼5.5 nm⁻¹. Therefore, LPSO-like segregation structures appear above 600 K. Since the SAXS patterns below 550 K do not give streaks nor cluster correlations, the microstructure related to the LPSO structure appears only at the higher temperatures. During heating, SAXS intensity corresponding to the LPSO microstructures vanished at the solution treatment temperature of 813 K used for the isothermal annealing for Fig. 2.

Fig. 5

In-situ SAXS profiles during heating solution-treated samples. Two arrows show the peak positions for 14H periodicity at q = 3.4 nm⁻¹ and the in-plane cluster correlation. Several spikes from streak components appear randomly in the profiles at higher temperatures.

3.4 SAXS pattern after strong deformation

Figure 6 shows SAXS profiles after rolling the sample down to the accumulated strain of ε = 0.5 and ε = 1.0 at 573 K, respectively. The streaks in the SAXS patterns represent the scattering from single or several layers of segregation layers with stacking faults, i.e., CANaPs as discussed for Fig. 2. Figure 6(a) shows the SAXS profile from the sample without deformation. The streaks extend randomly in all the radial directions, reflecting that the samples maintain microstructures of randomly oriented fine grains with several tens of micrometers in size.¹⁷⁾ Figure 6(b) and (c) show the SAXS patterns for the sample hot rolled with ε = 0.5, taken with the incident beam parallel to the transverse direction (TD) and rolling direction (RD) respectively. For further deformation up to ε = 1.0, the SAXS profile shown in Fig. 6(d) and (e), the streaks are more strongly accumulated into two major directions, suggesting development of kink microstructures. Maximum in the intensity at 3.4 nm⁻¹ appears along the streaks in deformed samples. This suggests that the thickness of the segregation layers increased during hot rolling. The diffuse peak corresponding to the in-plane cluster distance observed as a diffraction ring in Fig. 6(a) are observed as a pair of diffuse regions in the direction normal to the streaks in the deformed samples. From the diffuse regions, the size of the clusters was evaluated by Guinier approximations as Rg = 0.34 nm for ε = 0.5 and 0.35 nm for ε = 1.0 respectively. Therefore, the clusters in the hot rolled sample are in the form of the L1₂ clusters identical as those observed in concentrated MgYZn alloys. The orientation of the CANaPs are strongly rearranged by larger accumulated deformation of ε = 1.0. Orientation distribution of the streak for the deformed sample in Fig. 6 is shown in Fig. 7. Strong rolling with ε = 1.0 suggests that strong and well-aligned kink nanostructures with multiple L1₂ cluster layers (CANaP) are prepared by the process.

Fig. 6

SAXS patterns from the samples hot-rolled after heat treatment, compared with the undeformed one (a). SAXS patterns with the incident beam parallel to the transverse direction (b), (d) and rolling direction (c), (e) are shown. Thickening of the CALs was observed after hot rolling. Triangles at q ∼5 nm⁻¹ are the peak positions of diffuse scattering by in-plane ordering of L1₂ clusters.

Fig. 7

Angular distribution of SAXS intensity for several positions in the scattering vector. For the larger deformation, the plane-normal direction of the CANaPs tend to accumulate to a couple of orientations as shown in (b) and (d).

4. Conclusion

Small- and wide-angle scattering measurements have been made for Mg_98.6Y₁Zn_0.4 alloys hot-pressed and solution treated as the initial condition. Heat treatments above 550 K and below 623 K lead to segregation layers composed of L1₂ clusters with stacking faults, whose thickness is mostly monolayer and dispersed in α-Mg matrix. From in-situ heating measurements, further annealing was found to result in multiple layers of segregation layers leading to 14H LPSO structures. After forming such cluster layers in Mg matrix, hot rolling with accumulated strain of ε = 1.0 results in well-oriented streaks suggesting controlled kink nanostructures with increased thickness of cluster layers (CANaP). Accumulation in angle distribution of CANaPs were examined in the present work. Still, the distribution in CANaP and kink spacing requires better low angle resolution, which is now in progress as ultra small-angle X-ray scattering (USAXS) research.

Acknowledgments

Present work has been supported by grant-in-aid for scientific research under proposal numbers 18H05476 and 22K18886. SWAXS and EXAFS measurements have been performed under permission of PAC under proposal numbers 2019G685, 2021G553 at Photon Factory and 2019A1235, 2019B1298 and 2022A1327 at SPring8.

REFERENCES

• 1)   Y.  Kawamura,  K.  Hayashi,  A.  Inoue and  T.  Masumoto: Mater. Trans. 42 (2001) 1172–1176. doi:10.2320/matertrans.42.1172
• 2)   T.  Itoi,  T.  Semiya,  Y.  Kawamura and  M.  Hirohashi: Scr. Mater. 51 (2004) 107–111. doi:10.1016/j.scriptamat.2004.04.003
• 3)   Y.M.  Zhu,  A.J.  Morton and  J.F.  Nie: Acta Mater. 58 (2010) 2936–2947. doi:10.1016/j.actamat.2010.01.022
• 4)   E.  Abe,  A.  Ono,  T.  Itoi,  M.  Yamasaki and  Y.  Kawamura: Philos. Mag. Lett. 91 (2011) 690–696. doi:10.1080/09500839.2011.609149
• 5)   H.  Okuda,  M.  Yamasaki,  Y.  Kawamura,  M.  Tabuchi and  H.  Kimizuka: Sci. Rep. 5 (2015) 14186. doi:10.1038/srep14186
• 6)   H.  Okuda,  H.  Tanaka,  S.  Shiratake,  M.  Yamasaki and  Y.  Kawamura: Acta Mater. 118 (2016) 95–99. doi:10.1016/j.actamat.2016.06.028
• 7)   S.  Kurokawa,  A.  Yamaguchi and  A.  Sakai: Mater. Trans. 54 (2013) 1073–1076. doi:10.2320/matertrans.M2013123
• 8)   K.  Hagihara,  R.  Ueyama,  M.  Yamasaki and  Y.  Kawamura: Acta Mater. 209 (2021) 116797. doi:10.1016/j.actamat.2021.116797
• 9)   H.  Yokobayashi,  K.  Kishida,  H.  Inui,  M.  Yamasaki and  Y.  Kawamura: Acta Mater. 59 (2011) 7287–7299. doi:10.1016/j.actamat.2011.08.011
• 10)   D.  Egusa and  E.  Abe: Acta Mater 60 (2012) 166–178. doi:10.1016/j.actamat.2011.09.030
• 11)   K.  Kishida,  K.  Nagai,  A.  Matsumoto,  A.  Yasuhara and  H.  Inui: Acta Mater. 99 (2015) 228–239. doi:10.1016/j.actamat.2015.08.004
• 12)   H.  Okuda,  M.  Yamasaki and  Y.  Kawamura: Scr. Mater. 139 (2017) 26–29. doi:10.1016/j.scriptamat.2017.06.013
• 13)   H.  Okuda,  T.  Horiuchi,  T.  Tsukamoto,  S.  Ochiai,  M.  Yamasaki and  Y.  Kawamura: Scr. Mater. 68 (2013) 575–578. doi:10.1016/j.scriptamat.2012.11.041
• 14)   H.  Kimizuka,  M.  Fronzi and  S.  Ogata: Scr. Mater. 69 (2013) 594–597. doi:10.1016/j.scriptamat.2013.07.003
• 15)   H.  Kimizuka,  S.  Kurokawa,  A.  Yamaguchi,  A.  Sakai and  S.  Ogata: Sci. Rep. 4 (2014) 7318. doi:10.1038/srep07318
• 16)   K.  Takeuchi,  K.  Yuge,  S.  Kurokawa and  A.  Sakai: Sci. Rep. 8 (2018) 6841. doi:10.1038/s41598-018-25283-1
• 17)   Y.  Kawamura,  H.  Yamagata,  S.I.  Inoue,  T.  Kiguchi and  K.  Chattopadhyay: J. Alloys Compd. 939 (2023) 168607.
• 18)   Y.  Kawamura,  M.  Hayashida,  M.  Ifuku,  S.I.  Inoue and  K.  Chattopadhyay: in preparation.
• 19)   H.  Okuda,  T.  Horiuchi,  T.  Maruyama,  M.  Yamasaki,  Y.  Kawamura,  K.  Hagihara and  S.  Kohara: Metall. Mater. Trans. A 45 (2014) 147–151. doi:10.1007/s11661-013-1766-z
• 20)   H.  Okuda,  K.  Kinstsu,  S.  Kurokawa,  M.  Tabuchi,  H.  Nitani,  H.  Kimizuka,  S.  Inoue,  M.  Yamasaki and  Y.  Kawamura: submitted to Acta Mater.
• 21)  H. Okuda, M. Yamasaki and Y. Kawamura: Magnesium Technology 2022, ed. by Maier et al., (TMS, 2022) pp. 149–153.