In Situ Synchrotron X-ray Study of the Mechanical Properties of Pure Mg Produced by Powder Metallurgy

: In this study, in situ synchrotron X-ray experiments with wide-angle X-ray scattering (WAXS) and small-angle X-ray scattering (SAXS) detectors were performed on two pure magnesium materials produced by powder metallurgy. According to SAXS analysis, each of the two materials has a porosity of less than 0.5%. Line broadening analysis was performed on di ﬀ raction patterns collected by WAXS to analyze the dislocation evolution during material deformation. In both materials, < a > -type dislocation activities dominate the tensile deformation. The inﬂuence of grain size and texture on the di ﬀ erent tensile behaviors of these two materials is also discussed.


Introduction
Yield strength and ductility are two key mechanical properties for metallic structural materials. However, these two properties are usually exclusive of each other. Recently, a deformed and partitioned steel with very high ductility [1] was developed using a "dislocation engineering" concept. Therein, high dislocation density contributes to the high yield strength, and a large fraction of mobile dislocations is responsible for achieving high ductility.
The same alloy design strategy can be applied to magnesium alloys, the lightest metallic structural materials. To increase the mobile dislocation density, it is necessary to activate <c + a> dislocations, which proves to be critical for enhancing the ductility of Mg [2][3][4]. While the type of dislocations in deformed Mg can be examined by transmission electron microscopy (TEM) [4], a more statistically robust method is more desirable.
High-energy X-ray experiments have the potential to address this issue. X-ray line profile analysis (XLPA) has been used to quantitatively analyze the different types of dislocations in hexagonal metals [5][6][7] by analyzing diffraction profiles obtained from wide-angle X-ray scattering (WAXS). The extension Convolutional Multiple Whole Profile (e-CMWP) software and the Hexburger software are established tools for data analysis [8][9][10].
In this work, we use in situ synchrotron X-rays to analyze the dislocation density evolution in two pure Mg materials produced by powder metallurgy (PM). The porosity is first evaluated in these In this work, we use in situ synchrotron X-rays to analyze the dislocation density evolution in two pure Mg materials produced by powder metallurgy (PM). The porosity is first evaluated in these two materials by using the small-angle X-ray scattering (SAXS) profile [11,12]. Texture evolution and dislocation evolution during material deformation were analyzed using the WAXS data [13][14][15][16][17][18][19][20][21].

Materials and Experimental Procedures
Gas-atomized pure Mg powders (99.7 wt %), from a direct current arc plasma evaporation apparatus, with spherical particles ranging from 2.2 to 10 μm and a lognormal distribution (measured by Brunauer-Emmet-Teller (BET)) were cold pressed by a 2000 KN hydraulic press machine. The consolidated powders were milled out (DH) or heat-treated at 500 °C for 1 h and then milled out (FPS). Powders sheathed in cans were compacted at 200 °C and then stripped off to obtain extrusion feedstock. The feedstock was kept at 350 °C for 0.5 h; afterwards, it was extruded at the extrusion temperature of 300 °C, the extrusion ratio of 25:1, and the extrusion speed of 11 mm/s. The processing procedures for the two materials are shown in Figure 1. From the extruded bar, flat dog-bone tensile specimens with 5.0 mm (L) × 1.2 mm (W) × 1.0 mm (T) gauge dimensions were fabricated by electrical discharge machining (EDM). The tensile axis was parallel to the extrusion direction. The uniaxial tensile experiment was performed at a strain rate of From the extruded bar, flat dog-bone tensile specimens with 5.0 mm (L) × 1.2 mm (W) × 1.0 mm (T) gauge dimensions were fabricated by electrical discharge machining (EDM). The tensile axis was parallel to the extrusion direction. The uniaxial tensile experiment was performed at a strain rate of 2 × 10 −4 s −1 at 1-ID beamline of the Advanced Photon Source (APS). WAXS and SAXS signals were simultaneously collected during specimen loading, as illustrated in Figure 2. Instrument parameters (beam center, detector tilts, and sample-to-detector distance) were calibrated using a CeO 2 standard sample (NIST-SRM 674b). More details can be found in [29].
2 × 10 −4 s −1 at 1-ID beamline of the Advanced Photon Source (APS). WAXS and SAXS signals were simultaneously collected during specimen loading, as illustrated in Figure 2. Instrument parameters (beam center, detector tilts, and sample-to-detector distance) were calibrated using a CeO2 standard sample (NIST-SRM 674b). More details can be found in [29]. To more completely characterize the microstructure, optical metallography (OM) and transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan) were employed. All of the microscopy samples were cut from the sections perpendicular to the extrusion directions. OM samples were etched using a mixture of 1 g citric acid, 40 mL ethanol, and 20 mL distilled water. TEM samples were prepared by electro-polishing with a solution containing 5% perchloric acid, 34% butanol, and 61% distilled water.

X-ray Data Analysis
The WAXS diffraction patterns were transformed to diffraction line profiles using the GSAS-II software (GSAS-II, Argonne National Laboratory, United States) [30]. The azimuthal range of [−25°, 25°] around the tensile direction was used (tensile direction is marked with a white arrow marked in Figure 2) [19,29]. The XLPA procedure was further performed to estimate the fraction of different dislocation slip systems during material deformation using the CMWP (http://csendes.elte.hu/cmwp/) and Hexburger software (1.0, HexBurger, Prague, Czechia). Assuming that the two materials have a fiber texture, an open-source software, MTEX (5.3.1, Mtex, Chemnitz, Germany) [18], was utilized to calculate inverse pole figures from the WAXS diffraction patterns. The SAXS data were analyzed using the Irena package in Igor Pro (1.4.1, Irena, Argonne National Laboratory, DuPage County, IL, United States) [31] to estimate the porosity in the materials. Figure 3 shows the OM images (a,c) and the measured grain size distributions (b,d) of the two materials. The grain size was measured using the linear intercept method with a software named Nano-measurement. For both DH and FPS, at least three images were used to measure the grain size. The average grain size is 3.1 μm in the DH material and 11.6 μm in the FPS material. For the DH sample, the grain size is close to the size of the original powders, which have a lognormal grain size To more completely characterize the microstructure, optical metallography (OM) and transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan) were employed. All of the microscopy samples were cut from the sections perpendicular to the extrusion directions. OM samples were etched using a mixture of 1 g citric acid, 40 mL ethanol, and 20 mL distilled water. TEM samples were prepared by electro-polishing with a solution containing 5% perchloric acid, 34% butanol, and 61% distilled water.

X-ray Data Analysis
The WAXS diffraction patterns were transformed to diffraction line profiles using the GSAS-II software (GSAS-II, Argonne National Laboratory, United States) [30]. The azimuthal range of [−25 • , 25 • ] around the tensile direction was used (tensile direction is marked with a white arrow marked in Figure 2) [19,29]. The XLPA procedure was further performed to estimate the fraction of different dislocation slip systems during material deformation using the CMWP (http://csendes.elte.hu/cmwp/) and Hexburger software (1.0, HexBurger, Prague, Czechia). Assuming that the two materials have a fiber texture, an open-source software, MTEX (5.3.1, Mtex, Chemnitz, Germany) [18], was utilized to calculate inverse pole figures from the WAXS diffraction patterns. The SAXS data were analyzed using the Irena package in Igor Pro (1.4.1, Irena, Argonne National Laboratory, DuPage County, IL, United States) [31] to estimate the porosity in the materials. Figure 3 shows the OM images (a,c) and the measured grain size distributions (b,d) of the two materials. The grain size was measured using the linear intercept method with a software named Nano-measurement. For both DH and FPS, at least three images were used to measure the grain size. The average grain size is 3.1 µm in the DH material and 11.6 µm in the FPS material. For the DH sample, the grain size is close to the size of the original powders, which have a lognormal grain size distribution. The porosity was less than 0.5% in each of these two materials according to both the Archimedes measurement and the SAXS measurement (see Figure 4). distribution. The porosity was less than 0.5% in each of these two materials according to both the Archimedes measurement and the SAXS measurement (see Figure 4).   distribution. The porosity was less than 0.5% in each of these two materials according to both the Archimedes measurement and the SAXS measurement (see Figure 4).

Mechanical Properties Determined by In Situ Tensile Tests at Room Temperature
Engineering stress-strain curves of the DH specimen (grain size = 3.1 µm) and the FPS specimen (grain size = 11.6 µm) were obtained from in situ testing and are shown in Figure 7. Key tensile properties extracted from the curves are listed in the built-in table. Elongation of the FPS specimen is almost twice the elongation of the DH specimen. The yield strength and UTS of the two specimens are close to each other.

Mechanical Properties Determined by In Situ Tensile Tests at Room Temperature
Engineering stress-strain curves of the DH specimen (grain size = 3.1 μm) and the FPS specimen (grain size = 11.6 μm) were obtained from in situ testing and are shown in Figure 7. Key tensile properties extracted from the curves are listed in the built-in table. Elongation of the FPS specimen is almost twice the elongation of the DH specimen. The yield strength and UTS of the two specimens are close to each other.  Figure 8a,c shows the initial texture of the two specimens based on their WAXS diffraction patterns calculated by MTEX. The two specimens show similar texture, with the tensile axis close to the {1-100} pole. The only difference between these two materials is that the powder was heat-treated at 500 °C for 1 h before extrusion for the FPS material. This step increases the grain size but is unlikely to change the texture significantly. Figure 8b,d shows the texture of the specimens before their fracture. Because the two specimens failed at relatively small strain, the texture did not change much for both specimens.  Figure 8a,c shows the initial texture of the two specimens based on their WAXS diffraction patterns calculated by MTEX. The two specimens show similar texture, with the tensile axis close to the {1-100} pole. The only difference between these two materials is that the powder was heat-treated at 500 • C for 1 h before extrusion for the FPS material. This step increases the grain size but is unlikely to change the texture significantly. Figure 8b,d shows the texture of the specimens before their fracture. Because the two specimens failed at relatively small strain, the texture did not change much for both specimens.

Texture Evolution
Metals 2020, 10, x FOR PEER REVIEW 6 of 11

Mechanical Properties Determined by In Situ Tensile Tests at Room Temperature
Engineering stress-strain curves of the DH specimen (grain size = 3.1 μm) and the FPS specimen (grain size = 11.6 μm) were obtained from in situ testing and are shown in Figure 7. Key tensile properties extracted from the curves are listed in the built-in table. Elongation of the FPS specimen is almost twice the elongation of the DH specimen. The yield strength and UTS of the two specimens are close to each other.  Figure 8a,c shows the initial texture of the two specimens based on their WAXS diffraction patterns calculated by MTEX. The two specimens show similar texture, with the tensile axis close to the {1-100} pole. The only difference between these two materials is that the powder was heat-treated at 500 °C for 1 h before extrusion for the FPS material. This step increases the grain size but is unlikely to change the texture significantly. Figure 8b,d shows the texture of the specimens before their fracture. Because the two specimens failed at relatively small strain, the texture did not change much for both specimens.

Peak Broadening during the Tensile Test
During the in situ tensile test, shifting and broadening of diffraction peaks were observed, as shown in the inset of Figure 9. We are more interested in peak broadening, which reflects the dislocation activity in the specimen. Figure 9 shows the evolution of the full width at half maximum (FWHM) of the {1010}, {1011}, and {1120} peaks as a function of strain. Those peaks all show a gradual increase of FWHM values.
Metals 2020, 10, x FOR PEER REVIEW 7 of 11 Figure 8. Inverse pole figure maps of (a) the DH specimen before deformation, (b) the DH specimen before fracture, (c) the DH specimen before deformation, and (d) the FPS specimen before fracture.

Peak Broadening during the Tensile Test
During the in situ tensile test, shifting and broadening of diffraction peaks were observed, as shown in the inset of Figure 9. We are more interested in peak broadening, which reflects the dislocation activity in the specimen. Figure 9 shows the evolution of the full width at half maximum (FWHM) of the {101 � 0}, {101 � 1}, and {112 � 0} peaks as a function of strain. Those peaks all show a gradual increase of FWHM values.

Dislocation Evolution Analysis
The observed peak broadening in the two samples were caused by dislocation activity. To quantitatively understand the activity of different dislocation slip systems, the Hexburger software was used. Table 1 shows the 11 possible slip systems that can be activated [5,8]. Following the instructions of the Hexburger software [5,8,19], we computed the relative fraction of each slip system. We grouped the slip systems into three categories according to their Burgers vectors: <a>, <c>, and <c + a>. Figure 10 shows the evolution of the relative fraction of <a>, <c>, and <c + a> dislocations in the DH and the FPS specimens during deformation. In both specimens, the <a> dislocations dominate the deformation. The DH specimen shows an increase of <c + a> dislocation fraction during plastic deformation. This specimen has smaller grain size. The increased <c + a> activity could be induced by the stress concentration at grain boundaries. The DH specimen fractured earlier than the FPS

Dislocation Evolution Analysis
The observed peak broadening in the two samples were caused by dislocation activity. To quantitatively understand the activity of different dislocation slip systems, the Hexburger software was used. Table 1 shows the 11 possible slip systems that can be activated [5,8]. Following the instructions of the Hexburger software [5,8,19], we computed the relative fraction of each slip system. We grouped the slip systems into three categories according to their Burgers vectors: <a>, <c>, and <c + a>. Figure 10 shows the evolution of the relative fraction of <a>, <c>, and <c + a> dislocations in the DH and the FPS specimens during deformation. In both specimens, the <a> dislocations dominate the deformation. The DH specimen shows an increase of <c + a> dislocation fraction during plastic deformation. This specimen has smaller grain size. The increased <c + a> activity could be induced by the stress concentration at grain boundaries. The DH specimen fractured earlier than the FPS specimen, which may also be explained by the sudden increase of <c + a> dislocations, which are usually less mobile [3]. specimen, which may also be explained by the sudden increase of <c + a> dislocations, which are usually less mobile [3].

Discussion
The two materials (DH and FPS) were both obtained by extrusion of pressed Mg powders. Indeed, both materials have lower yield strength than that of conventional pure Mg extrusion bars. We conducted some ex situ tensile and compression tests, and similar findings were obtained. This decrease in yield strength could be due to the texture effect: both DH and FPS materials were prepared by extruding powders instead of cast ingots. This difference causes a weak texture, as demonstrated in Figure 7. Under a weak texture, the easy basal slip can be activated in more grains, which reduces the yield strength. The DH material with a grain size of ~3 μm shows a yield strength similar to that of the FPS material with a grain size of ~10 μm. A possible explanation is that the heat treatment at 500 °C not only increased the grain size but also eliminated some porosity in the FPS material. The latter effect may offset the Hall-Petch effect, so that the two materials show similar strength.

Discussion
The two materials (DH and FPS) were both obtained by extrusion of pressed Mg powders. Indeed, both materials have lower yield strength than that of conventional pure Mg extrusion bars. We conducted some ex situ tensile and compression tests, and similar findings were obtained. This decrease in yield strength could be due to the texture effect: both DH and FPS materials were prepared by extruding powders instead of cast ingots. This difference causes a weak texture, as demonstrated in Figure 8. Under a weak texture, the easy basal slip can be activated in more grains, which reduces the yield strength. The DH material with a grain size of~3 µm shows a yield strength similar to that of the FPS material with a grain size of~10 µm. A possible explanation is that the heat treatment at 500 • C not only increased the grain size but also eliminated some porosity in the FPS material. The latter effect may offset the Hall-Petch effect, so that the two materials show similar strength.
From the X-ray line profile analysis, we were able to estimate the relative fraction of <a>, <c>, and <c + a> dislocations during material deformation. The <a> dislocations dominated the deformation of both materials. This observation is consistent with other researchers' findings in conventional pure Mg [22]. The <c + a> dislocations became more active with strain. Nucleation of <c + a> dislocations can accommodate the stress concentration and strain incompatibility across grain boundaries. On the other hand, the <c + a> dislocations have low mobility [3], so both materials fractured at a relatively small strain.
The DH material shows a higher fraction of <c + a> dislocations during its deformation. The smaller grain size causes more dislocation pile-up at grain boundaries, which tends to induce <c + a> dislocations. Similar results were reported in [28].