Deformation mechanism and dynamic precipitation in a Mg-7Al-2Sn alloy processed by surface mechanical attrition treatment

1. Introduction

Surface hardening is usually required in metal industry since most failures occur on the surface of metal parts, such as behavior of fatigue, fretting fatigue, corrosion and wear [1,2]. Surface mechanical attrition treatment (SMAT) is one of the surface nanocrystallization methods, which has been proven to be an effective way to strengthen the surface of metal alloys due to refined microstructure [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]]. In order to manipulate the properties of SMATed materials, the understanding of grain refinement mechanism during a SMAT process becomes important. According to previous investigations [1,7,[9], [10], [11],13,16], the governing factors of grain refinement during SMAT process were generally considered to be stacking fault energy (SFE) and number of slip systems. For magnesium (Mg) alloys with a typical hexagonal close packed (hcp) structure and medium SFEs (60-78 mJ/m²) [13], twinning is expected to play a significant role in grain refinement since the slip systems are quite limited, as observed in Mg alloy AZ91D by Sun et al. [13]. However, Wei et al. [14] found that grain refinement of Mg alloy AZ91D during SMAT process is attributed to the dislocation activity without twinning behavior. These two contradictory experimental results indicate that there must be some other factors affecting the refinement mechanism, such as second phase for precipitable Mg alloys. Therefore, as-cast, solution-treated (T4), aged (T6) Mg-7Al-2Sn (AT72) alloys were designed to investigate the effect of second phases on deformation mechanism during SMAT in this paper.

For a high content precipitable Mg alloy, dynamic precipitation usually occurs during hot or warm working [[17], [18], [19], [20], [21]]. However, rare research work reported dynamic precipitation of Mg alloys during SMAT process. Since dynamic precipitates on the surface of precipitable alloys can provide additional strengthening effects along with grain refinement strengthening, the effect of different precipitates on strengthening designed AT72 SMATed alloys was also explored in this paper.

2. Experimental methods

As-cast Mg alloy AT72 (Al 6.91 wt.%, Sn 2.65 wt.% and bal. Mg) was prepared by super vacuum die-casting (SVDC) [22]. Samples were subjected to solution treatment in an air furnace at 410 °C for 16 h, followed by water quenching and then aged in an oil bath at 175 °C for 100 h.

An SMAT process was performed at both sides of the plate alloys at room temperature for 10 min with a vibrating frequency of 20 kHz. Details of the SMAT process can be found in the literature [5,23]. Both sides of the plate alloys were mechanical ground before SMAT to eliminate oxides and severe segregation near the surface of casts.

Microstructures of SMAT AT72 alloys were observed using optical microscopy (OM, Zeiss, Axio Observer A1), transmission electron microscopy (TEM, JEOL, JEM-2100 F) operating at 200 kV and high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM, JEOL, aberration corrected JEM-ARM200 F) with energy dispersive spectroscopy (EDS, JEOL, EX-230**BU_37001-2). Thin foils for TEM observations were prepared with a plasma ion polisher (Gatan, Model 691).

3. Results

According to our previous research [22], a dendritic microstructure of as-cast AT72 alloys consists of α-Mg, Mg₂Sn, and Mg₁₇Al₁₂ phases which were mostly aggregated at grain boundaries. After T4 treatment, most second phases were dissolved into the Mg matrix except for a few ignorable tiny spherical Mg₂Sn particles. In T6 alloy, dispersedly distributed lath-like Mg₂Sn and Mg₁₇Al₁₂ phases with width less than 200 nm were observed. Average size of equiaxed grains in T4 and T6 alloys was $\widetilde{4}$0 and $\widetilde{6}$2 μm, respectively.

3.1. Deformation modes observation in AT72 alloys with and without second phases

Fig. 1(a) and 1(b) shows cross-sectional OM images of as-cast AT72 alloy before and after SMAT, while T4 and T6 alloys after SMAT are shown in Fig. 1(c) and (d), respectively. The bright contrast areas, which represent α-Mg phase, were reduced dramatically after SMAT process, especially near the surface layer. It indicates that SMAT process can divide large dendritic α-Mg structure into small grains and improve precipitation of second phases, as shown in the dark contrast areas. In addition, no twins were observed in AT72 alloy after SMAT process in OM and TEM images.

View original graphic| Download| PPT slide

Fig. 1.   (a) and (b) are cross-sectional optical micrographs of as-cast AT72 alloys before and after SMAT deformation, respectively; (c) and (d) are cross-sectional optical micrographs of T4 and T6 AT72 alloys after SMAT deformation, respectively. “S” in each image represents the surface of the SMAT sample.

With respect to T4 alloy subjected to SMAT (Fig. 1(c)), a high density of twins near the surface and the twin density decreases from surface to the center of plate alloys due to gradient stress, strain and strain rate. Similar to SMATed AT72 alloy, no twins were observed in SMATed T6 alloy, as shown in Fig. 1(d). In order to further verify the absence of twins in T6 alloy, several typical depths from the surface of each sample were observed through TEM, i.e. 0-30 μm, 200-300 μm and 800-900 μm. No twins were found in T6 alloy after SMAT, however, some interactions between precipitates and dislocations, as shown in Fig. 2, were observed from layer 200-300 μm which bore quite high plastic strain during SMAT process but formed fewer boundaries than the topmost layer.

View original graphic| Download| PPT slide

Fig. 2.   (a) Bright-field TEM images of the layer that is 200-300 μm below the treated surface in T6 AT72 alloy; (b) SAED patterns with electron beam parallel to [1-210] zone axis of Mg matrix; (c) HRTEM image of phase boundary between α-Mg and precipitate; (d), (f), (h) are bright-field images under two-beam conditions; (e), (g), (i) are dark-field images under two-beam conditions.

Fig. 2 (a) shows a high density of lath-like precipitates lies on the basal plane of grains with electron beam parallel to [1$\bar{2}$10] of Mg matrix as shown in Fig. 2(b). Many Morie fringes can be readily observed in both Mg matrix and precipitates, which indicates that there is a high density of dislocations in the alloy due to SMAT deformation. These dislocations can be considered to be derived from grain boundaries and phase boundaries (Fig. 2 (c)), slipped in Mg matrix and then were mostly blocked by the large quantity of phase boundaries, leading to strengthening of T6 AT72 alloy. Similar with grain boundaries, phase boundary can hinder and absorb dislocations during deformation, as well as emit dislocations as sources due to the lattice misfit between the two phases. Orientation relationships between lath-like precipitates (Mg₁₇Al₁₂ and Mg₂Sn) and Mg matrix are reported as 111β/(0001)α-Mg and <110>β/<11$\bar{2}$0>α-Mg [[24], [25], [26], [27], [28], [29], [30], [31]]. The phase boundary shown in Fig. 2(c) is parallel to the basal plane of Mg matrix and thus < a> component dislocations can be formed from this phase boundary during SMAT deformation.

Dislocation types can be confirmed under two-beam condition, as shown in Fig. 2(d) to 2(i). According to the criterion of dislocation extinction g·b × u = 0 [32,33], the < c> dislocation component can be shown in Fig. 2(d) and (e) while < a > dislocation component can be observed in Fig. 2(f) and (g). If dislocations containing both < a> and < c> components but disappear when $\vec{g}$=01$\bar{1}$$\bar{1}$, as shown in the right top corner of Fig. 2(i), then dislocations belong to < c+a > type with $\vec{b}$=1/3[$\bar{2}$11$\bar{3}$] or $\vec{b}$=1/3[11$\bar{2}$3] as Burgers vector.

3.2. Dynamic precipitation phenomenon in AT72 alloys during SMAT

Oversaturated α-Mg structures can be obtained in as-cast AT72 alloy because of the high cooling rate during SVDC process. Therefore, precipitates will appear at an appropriate temperature or under some extent of strain. SMAT was conducted at room temperature without heating source. However, the highest strain rate and strain in the topmost surface layer were calculated as $\widetilde{1}$0⁴ s^-1 and $\widetilde{1}$ × 10^-1, which leads to the precipitation as shown in Fig. 3.

View original graphic| Download| PPT slide

Fig. 3.   (a) Bright-field TEM image of the topmost layer with the thickness of 30 μm in the as-cast AT72 alloy after SMAT deformation; (b) is the SAED patterns of the particle circled in (a) with electron beam close to [011] of FCC crystal.

As shown in Fig. 3(a), grain size of the surface layer in SMATed as-cast AT72 alloy is around 400 -600 nm and a large quantity of dark contrast particles with size of around 50-300 nm can be easily observed at the grain boundaries and inside grains occasionally. The electron diffraction patterns of the particle circled in Fig. 3(a) is shown in Fig. 3(b). This particle is confirmed as Mg₂Sn phase after the lattice parameter measurement and comparison to the standard PDF card. It is worth noting that these large quantities of homogeneous distributed particles were not observed in cast, solution-treated and aged AT72 alloys before SMAT, which is also demonstrated in our previous research [22].

In order to further testify the chemistry and phases of all particles, EDS mapping was captured at the HAADF-STEM mode, as shown in Fig. 4. Different from the relatively uniform distribution of Mg, Al was not enriched in the particles, where Sn was obviously segregated. Therefore, those particles appeared after SMAT were verified as Mg₂Sn phase and no Mg₁₇Al₁₂ dynamic precipitates were observed. In addition, Mg and Al were separated from each other to some extent, especially in Fig. 4(b), which is a new phenomenon probably related to the decrease in system energy when SMAT provides energy to cross energy barrier.

View original graphic| Download| PPT slide

Fig. 4.   HAADF-STEM images and EDS mapping results of the topmost layer with the thickness of 30 μm in SMATed cast AT72 alloy (Red, green and blue represent Mg, Al and Sn, respectively, and the overlapping images are at last.) (a) and (b) are recorded from two different specimens in order to verify this phenomenon. Rectangular marked area in (a) corresponds to EDS scanning area.

In order to testify the dynamic precipitation phenomenon in T4 and T6 AT72 alloys, 30 μm thick topmost layers in SMATed T4 and T6 alloys were characterized by TEM, respectively (Fig. 5(a) and (b)). Grain size of T4 alloy in the surface layer is around 500 nm and the size of particles is about 200-300 nm, which are smalller than those of SMATed as-cast AT72 alloy. This is probably caused by the more homogeneous distribution of elements in the solution treated alloy and a lower density of dislocations in the alloy as nucleation sites of precipitation.

View original graphic| Download| PPT slide

Fig. 5.   Bright-field TEM image of the topmost layer with the thickness of 30 μm in the (a) T4 and (b) T6 AT72 alloys after SMAT deformation.

Grain size of the surface layer in T6 alloy is about 600-700 nm and there are also dark particles with size of around 100-250 nm. The quantity of these precipitates is less than that in SMATed cast alloy, which is probably due to homogeneous distribution of Sn after solution treatment and the consumption of Sn during ageing treatment. Comparing to the microstructure of 200-300 μm below the surface in Fig. 2 without precipitated particles, the dynamic precipitates in the topmost layer is attributed to the highest strain rate of $\widetilde{1}$0⁴ s^-1.

4. Discussion

4.1. The inhibiting effect of second phases on twinning behavior of AT72 alloy during SMAT

As aforementioned, the formation of twins was suppressed by second phases in SMATed and T6 AT72 alloys due to the increase in microstructure diversity. Moreover, the large quantity of phase boundaries with mismatched lattice can absorb dislocations as barrier and emit dislocations as source to accommodate plastic strain during SMAT deformation. Many < c+a > dislocations with a high critical resolved shear stress (CRSS) were observed in SMATed T6 AT72 alloy. On the contrary, T4 AT72 alloy, with a uniform microstructure without second phase, did not exhibit diverse dislocation slips during SMAT deformation. Thus, twinning is required to coordinate strain and consume energy.

4.2. Dynamic precipitation of Mg₂Sn in AT72 alloys during SMAT

A considerable amount of nano Mg₂Sn articles was observed but no Mg₁₇Al₁₂ was found in the surface layer of as-cast, T4 and T6 AT72 alloys after SMAT process. The content of Al and Sn in α-Mg phase has little effect on precedence of Mg₂Sn precipitation since a similar phenomenon was found in adequate SMATed samples in three AT72 alloys with different original phase configurations. The involved reasons for which Mg₂Sn is more easily to dynamicaly precipitate during SMAT than Mg₁₇Al₁₂ are prudently speculated as following. First, Sn maintains a positive enthalpy of mixing (+ΔH) with Al so that Sn atoms tend to be rejected by Al [26,30,31,34]. Therefore, the dissolution of Al in α-Mg phase promotes the precipitation of Sn in Mg-7Al-2Sn alloy. Second, the formation of intermetallics was related to the electronegative difference between different elements. Electronegative value of Mg, Al, and Sn is 1.31, 1.61, and 1.96, respectively. Therefore, Mg₂Sn is preferred to form because of the highest electronegative difference between Mg and Sn elements in a Mg-Al-Sn system [35,36].

Although dynamic precipitates are same, size and distribution of Mg₂Sn show a little difference in the topmost layer in the three AT72 alloys under the same SMAT process, which can be attributed to solute concentration and microstructure difference. The microstructure of as-cast alloy is composed of α-Mg, island-like second phases Mg₁₇Al₁₂ and Mg₂Sn with size of several microns and even larger. The solute distribution inside α-Mg phase is not homogeneous due to the high solidification speed during high vacuum die casting. However, after fully solution heat treatment, T4 alloy with only α-Mg phase shows the most homogeneous of the three alloys in both solute distribution and microstructure. T6 alloy was obtained through age treatment based on T4 alloy composing of α-Mg phase and fine Mg₁₇Al₁₂ and Mg₂Sn precipitates with size of nanometers. Therefore, comparing to T4 alloy, solute distribution and microstructure in T6 is quite balanced and homogeneous except for the precipitates. The pre-existed second phase of both eutectic phase and precipitate would bring in large quantity of boundaries which could block and accumulate dislocations during plastic deformation of SMAT. Since solute segregation is severe in as-cast AT72 alloy and large second phases can accumulate many dislocations to provide nucleation sites, the quantity of Mg₂Sn particles is larger than T4 and T6 alloys after SMAT process. In addition, the quantity of Mg₂Sn particles in SMATed T4 alloy is the smallest due to homogeneous distribution and most uniform microstructure.

Dynamic precipitation in AT72 alloys during SMAT process is considered to be driven by the high concentration of solutes, the high strain rate in the surface layer and unavoidable temperature increment. In addition, the high strain rate in this case can provide the extra energy for precipitation at ambient temperature, known as strain induced dynamic precipitation, which is different from the usual dynamic precipitation driven by heat in Mg alloys [[17], [18], [19], [20], [21]].

5. Conclusions

The effect of second phases on deformation mechanism during SMAT process has been investigated in as-cast, T4 and T6 AT72 alloys. Conclusions are listed as follows:

(1)The second phases in as-cast and T6 AT72 alloys, which can provide phase boundaries for activating dislocations, were observed to suppress twinning during SMAT deformation.

(2)The dynamic precipitation of Mg₂Sn particles observed on the surface of the as-cast, T4 and T6 AT72 alloys after SMAT process is driven by the high content of alloying element, some local heat as well as the high strain rate. The high strain rate is considered to be the key point of inducing precipitation.

Acknowledgements

The authors gratefully acknowledge the financial supports of the National Natural Science Foundation of China (No. 51474149, No. 51301107 and No. 51671101), and Natural Science foundation of Jiangxi Province (No. 20172BCB22002). The authors would also like to express their gratitude to Professor Alan A. Luo from The Ohio State University for helpful discussions.

The authors have declared that no competing interests exist.