Comparative analysis of microstructure, mechanical, and corrosion properties of biodegradable Mg-3Y alloy prepared by selective laser melting and spark plasma sintering

Available online at www.sciencedirect.com Journal of Magnesium and Alloys 12 (2024) 1496–1510 www.elsevier.com/locate/jma Full Length Article Comparative analysis of microstructure, mechanical, and corrosion properties of biodegradable Mg-3Y alloy prepared by selective laser melting and spark plasma sintering P. Minárik a , b , ∗,M. Zemkováa ,S. Šašek a ,J. Dittrich a ,M. Knapek a ,F. Lukáˇc a , c ,D. Koutnýd , J. Jarošd ,R. Král a a Faculty of Mathematics and Physics, Charles University, Ke Karlovu 5, Praha 121 16, Czech Republic b Research Centre, University of Žilina, Univerzitná 8215/1, Žilina 010 26, Slovakia c Institute of Plasma Physics of the Czech Academy of Sciences, Za Slovankou 1782/3, 182 00 Praha 8, Czech Republic d Institute of Machine and Industrial Design, Brno University of Technology, Technická 2896/2, Brno 616 69, Czech Republic Received 29 January 2024; received in revised form 28 February 2024; accepted 7 April 2024 Available online 27 April 2024 Abstract This work explored possibilities of biodegradable magnesium alloy Mg-3Y preparation by two modern powder metallurgy techniques – spark plasma sintering (SPS) and selective laser melting (SLM). The powder material was consolidated by both methods utilising optimised parameters, which led to very low porosity ( ∼0.3%) in the SLM material and unmeasurably low porosity in the SPS material. The main aim of the study was the thorough microstructure characterisation and interrelation between the microstructure and the functional properties, such as mechanical strength, deformability, and corrosion resistance. Both materials showed comparable strength of ∼110 MPa in tension and compression and relatively good deformability of ∼9% and ∼21% for the SLM and SPS materials, respectively. The corrosion resistance of the SPS material in 0.1 M NaCl solution was superior to the SLM one and comparable to the conventional extruded material. The digital image correlation during loading and the cross-section analysis of the corrosion layers revealed that the residual porosity and large strained grains have the dominant negative effect on the functional properties of the SLM material. On the other hand, one of the primary outcomes of this study is that the SPS consolidation method is very effective in the preparation of the W3 biodegradable alloy, resulting in material with convenient mechanical and degradation properties that might find practical applications. © 2024 Chongqing University. Publishing services provided by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ) Peer review under responsibility of Chongqing University Keywords: Magnesium; Yttrium; Powder metallurgy; Microstructure; Mechanical strength; Corrosion resistance. 1. Introduction Magnesium (Mg) and its alloys have garnered considerable interest in recent years as promising materials for various engineering applications due to their exceptional combination of low density, high specific strength, and excellent machinability. Magnesium alloys also combine superior bio- ∗Corresponding author at: Faculty of Mathematics and Physics, Charles University, Ke Karlovu 5, Praha 121 16, Czech Republic. E-mail addresses: peter[email protected] , peter[email protected] (P. Minárik). compatibility, mechanical properties close to human bones, and natural biodegradability. Biodegradable magnesium devices and implants are therefore believed to become a costeffective option for medical implants in the future [1] . Among the myriad of magnesium alloys, those incorporating yttrium (Y) as an alloying element have been subject to extensive investigation, as Y is biocompatible [ 2 , 3 ] and can significantly enhance the mechanical and corrosion properties of Mg-based materials [ 4 , 5 ]. Yttrium was found to be an excellent element for improving mechanical properties both at ambient and elevated temperatures due to the solid solution strengthening [6] . Moreover, yttrium was reported to play a crucial role in https://doi.org/10.1016/j.jma.2024.04.018 2213-9567/© 2024 Chongqing University. Publishing services provided by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ) Peer review under responsibility of Chongqing University P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 1497 moderating corrosion resistance [7] . There are two significant effects on corrosion resistance resulting from yttrium incorporation: firstly, yttrium shows a high solubility in the Mg matrix, preventing undesirable galvanic corrosion between the Mg matrix and intermetallic phases [8] . Secondly, during corrosion, a dense protective film of Y2 O3 is formed, effectively slowing down or suspending the corrosion attack [5] . The Mg-3Y alloy was investigated in our previous study and it was shown that it is a good candidate for the medical applications. The alloy exhibited a very good combination of strength, ductility, and in vitro degradation properties, especially in the ultrafine-grain state prepared by equal channel angular pressing [9] . In this study, we turn our focus on two innovative powder metallurgical techniques – spark plasma sintering (SPS) and selective laser melting (SLM). Both of these techniques provide a relatively easy and time-effective method for preparing near-net-shape implants; therefore, they are of high interest to modern medicine [ 10 , 11 ]. The SPS process involves the application of high temperature and pressure in a fast and energy-efficient manner, resulting in rapid powder consolidation and excellent material properties [12] . The electrical current passes through the conductive die and the powder, generating localised heating at the powder particle contacts. Before the actual sintering begins, pressure is applied to the powder inside the die, which helps to eliminate any voids or porosities within the powder bed and ensures uniform densification during the sintering process. Unlike traditional sintering methods, SPS heats the material internally, which enables quicker heating rates and reduces the overall processing time [13] . The SLM process belongs to the family of additive manufacturing techniques, which are commonly addressed as 3D printing [10] . This process involves joining metal powder to build a 3D object layer by layer, based on a digital 3D model. SLM utilises a high-powered laser beam to melt and fuse the metal powder. After completing each layer, the printer adds a new layer (powder bed), continuously building the 3D object. The advantages of SLM are design flexibility, material efficiency, and the ability to create complex geometries suitable for medical applications [14] . During the last decade, several magnesium alloys have been prepared by SPS and/or SLM techniques with different results, commonly depending on the alloy composition [ 10 , 12 , 15–18 ]. The key factor, which generally has a strong impact on the physical properties of the compacts, is the residual porosity. For both manufacturing methods, there is some general knowledge regarding the effects of the individual processing parameters on the resulting microstructure, but they need to be tailored for each alloy separately to prepare the compacts with favourable properties. This paper presents a systematic comparative investigation of the microstructure, mechanical properties, and corrosion resistance of the binary Mg-3Y magnesium alloy processed using three different techniques: Selective Laser Melting, Spark Plasma Sintering, and Extrusion. Extrusion is a common manufacturing process used to create objects with a fixed crosssectional profile by forcing a material through a specially designed die or a set of dies. This technique is well-known and presents a benchmark for the two powder metallurgy techniques. It is important to note that each of these processes has a distinct effect on the microstructure, which subsequently affects the mechanical and corrosion properties of the alloy [9] . A comprehensive comparison of these three processing techniques concerning their influence on the mechanical and corrosion properties of the biodegradable Mg-3Y alloy provides important insights on its utilisation, especially in medical applications. 2. Materials and methods 2.1. Initial material In this work, Mg-3Y alloy (wt.%, W3) was cast from pure elements and subsequently gas-atomised at Nanoval GmbH & Co. KG, Berlin. The obtained powder was sifted and the resulting D50 value was 41 μm. The powder was prepared under an inert atmosphere and all powder handling was carried out in a low-oxygen environment to prevent oxidation. The SEM micrographs showing the microstructure of the powder particles are presented in Fig. 1 . The powder particles exhibit a round shape with occasional deformations ( Fig. 1 a). An inner structure of the powder particles exhibited a typical dendritic structure ( Fig. 1 b) resulting from the rapid cooling [19] . The brighter areas correspond to a higher concentration of heavier yttrium atoms due to the compositional contrast, as shown in Fig. 1 c (top) and confirmed using energy-dispersive X-ray spectroscopy (EDS) (bottom). An electron back-scatter diffraction (EBSD) analysis of the powder microstructure ( Fig. 1 d) revealed that the grain distribution strongly depends on the size of the individual particles. Smaller particles contained only one or a few grains, while larger particles consisted of many grains. 2.2. Processing The specimens from gas-atomised W3 powder were fabricated by the Laser Powder Bed Fusion process. A 3D printer SLM 280HL (SLM-Solutions AG, Germany) was used for this purpose. The machine featured a 400 W YLR-Laser (IPG Photonic) with a Gaussian profile and a spot size diameter of approx. 82 μm. To prevent oxidation or burning of the powder during the application of the laser beam, the building chamber was filled with an inert argon atmosphere with an oxygen level below 0.2%. The optimal processing parameters were established by a series of experiments, starting from the fabrication and evaluation of single weld tracks, followed by a series of volumetric specimens in the form of 10 ×10 ×10 mm3 cuboids. For all experiments, 50 μm layer thickness was used. Using single weld tracks, 90 combinations of processing parameters with variations of Laser Power (LP) and Scanning Speed (SS) were evaluated. The parameters were set within the range of LP = 60–240 W and SS = 200–1500 mm/s. Single weld tracks were evaluated with regard to weld track continuity 1498 P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 Fig. 1. Gas-atomised powder: (a) SEM micrograph, (b) SEM cross-section, (c) detail of dendritic microstructure with Y elemental map of the same area, and (d) EBSD orientation map. using a light microscope Keyence VHX 6000, followed by a more detailed evaluation of weld track cross-section by metallography, considering only continual tracks of good quality. Based on the weld track width, depth, and the presence of defects, 10 of the best combinations of processing parameters were chosen for volumetric testing with cuboid specimens. The cuboid experiments consisted of several iterations evaluated by metallography in the XY cross-section (parallel to the baseplate, Fig. 2 ), while the variation in hatching distance (HD) with a resulting overlap of 30–50% between individual weld tracks was used. The best combination was found as LP = 180 W, SS = 600 mm/s, HD = 0.09 mm. Cuboids fabricated with such parameter combinations showed a relative density of 99.8% in XY and XZ sections. Based on small cuboid samples, rectangular specimens with dimensions of 20 ×10 ×6 mm3 were produced (schematically drawn in Fig. 2 ). Spark Plasma Sintering (SPS) was selected as a second way of consolidating the W3 powder. The consolidation was performed in a Thermal Technology LLC device type 10–4. 10 g of the powder was used for one sample. Sintering was conducted in the cylindrical graphite tool at 550 °C for 10 min under a pressure of 100 MPa. These processing parameters for SPS were selected based on our previous results, for details see Refs. [ 12 , 20 ]. Utilisation of these parameters resulted in unmeasurably low residual porosity. The temperature was controlled by a thermocouple placed inside the graphite tool. A protective argon atmosphere was used to prevent oxidation. The resulting sample had a diameter of 30 mm and a height of 8 mm (schematically drawn in Fig. 2 ). Fig. 2. Schematic illustration of selective laser melting (SLM), spark plasma sintering (SPS) and extrusion process (EX), together with major sample directions. SLM orientation is denoted through the text according to the coordinate system, SPS according to the sintering loading direction (SLD), and EX according to the extrusion direction (ED). P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 1499 Finally, the third W3 material was prepared by means of extrusion. Before extrusion (EX), the cast billet was homogenised at 400 °C for 16 h. Subsequently, the billet was extruded at 350 °C with an extrusion ratio (ER) of 30 and an extrusion rate (ram speed) of 3 mm/s, as schematically drawn in Fig. 2 . The diameter of the extruded rod was 17 mm. 2.3. Microstructure The microstructure of the gas-atomised powder and the prepared materials were analysed using the scanning electron microscope (SEM) ZEISS Auriga Compact equipped with the EDAX EDS and the EDAX EBSD camera. The relative porosity of the samples prepared via powder metallurgy was also determined from high-resolution SEM images as the surface fraction of the polished specimens. At least 10 images with suitable magnification were analysed for this purpose. The samples for SEM investigation were mechanically polished with the decreasing abrasive grain size down to 50 nm. Samples for EBSD were subsequently ion-polished using a Leica RES102 device. The EBSD analysis was performed on the area of 2 ×2 mm2 with a step size of 2 μm to acquire sufficient datasets for all statistical calculations in the TSL OIM 8 software. The measured data were partially cleaned by one step of confidence index (CI) standardisation and one step of grain dilatation. Only points with CI > 0.1 were used for the analysis. A transmission electron microscope (TEM) JEOL F2000FS equipped with Bruker EDS was used for precise identification of the secondary phase particles. As samples for TEM, discs of 3 mm in diameter were cut from the investigated materials, mechanically thinned down to 150 μm, and finally ion-polished to electron-transparent thickness. 2.4. Mechanical properties The mechanical properties of the manufactured materials were studied through Vickers microhardness tests (HV), compressive, and tensile deformation tests. The QNESS Q50 device was used for Vickers microhardness tests, calculating the average microhardness from at least 150 indents. The deformation tests were performed using the Instron 5882 universal testing machine at room temperature with a constant crosshead speed, providing an initial strain rate of 10−3 s−1 . The tensile samples had a flat dog-bone shape with gauge dimensions of 8 ×4 ×1.5 mm3 , while the dimensions of the compression samples were 5 ×3 ×3 mm3 . At least three tests were performed for each condition. For the digital image correlation (DIC) measurements, the tensile samples were ground using #1200 SiC emery paper followed by the application of fine speckle pattern by means of the airbrush technique utilising water-based paints. The sample images were captured during loading at a rate of 1 fps using the Nikon D3100 camera. The DIC data were evaluated using the Ncorr package implemented in Matlab [21] . The same DIC parameters were used for all samples: subset radius of 40 px, subset spacing of 5 px and strain radius of 4 points. 2.5. Corrosion resistance A linear polarisation method was used to study the initial corrosion resistance of the studied materials. The tests were conducted with the potentiostat Metrohm AUTOLAB128N using the three-electrode setup. The polarisation curves were measured after 10 min of stabilisation in 0.1 M NaCl aqueous solution at room temperature. The measurement was performed from −150 mV to 200 mV vs. the open circuit potential (OCP) with the scan rate of 1 mV/s. A rotating disc electrode (sample) was used for better homogeneity of the measurement, using 300 rpm. The samples were cut in the YZ plane (X-direction) and, prior to each measurement, the exposed surface was ground with #1200 SiC emery paper and rinsed with ethanol. The values of corrosion potential ( Ecorr ), corrosion current density ( icorr ) and polarisation resistance ( Rp ) were determined from the measured data by the Tafel analysis. The long-term corrosion performance of the studied materials was evaluated through weight loss measurements. Specimens with dimensions of 1 ×6 ×6 mm3 (XYZ) were cut from each material, and their initial weight m0 was recorded. Subsequently, these samples were submerged in a 0.1 M NaCl aqueous solution for a duration of 7 days. Following this immersion period, the specimens were carefully removed from the solution, and the corrosion products were removed according to the ISO 8407:2009 standard. Upon cleaning, the weight of each sample m was remeasured, and the corrosion rate CR was calculated using the formula: CR = (m0 - m) / (t ·S) , where t represents the immersion time and S denotes the total surface area of the sample. At least three samples were examined for each type of material. 3. Results 3.1. Microstructure Fig. 3 shows an overview of the microstructure of the investigated materials in terms of SEM micrographs. Both materials prepared via powder metallurgy exhibited compact structure, however, the inner structure, especially with an emphasis on the distribution of the secondary phases, was significantly different. The bright secondary phase particles are distributed relatively uniformly in the magnesium matrix of the SLM sample ( Fig. 3 a). On the other hand, the SPS sample clearly exhibits distinguishable former powder particles surrounded by a continuous bright layer, as documented in Fig. 3 b. In addition, small discontinuous precipitates inside the former powder particles and clearly depleted zones along the earlier powder particle boundaries were formed, marked by red lines (detail in Fig. 3 b). As mentioned above, the processing method also had a strong impact on the residual porosity. A low residual porosity of ∼0.3% was measured in the SLM sample by SEM image analysis of both investigated planes. On the other hand, no residual porosity was observed in the material prepared by SPS. These results correspond well with the aforementioned investigation performed 1500 P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 Fig. 3. SEM micrographs of the microstructure of (a) SLM sample, (b) SPS sample, and (c) EX sample. Fig. 4. SEM micrographs and corresponding elemental distribution maps for Y and O measured in SLM and SPS samples. by light microscopy during the optimisation of SLM parameters. Fig. 3 c presents the microstructure of the EX condition. Because of the highly limited presence of any secondary phase particles, it is clear that almost all yttrium is dissolved in the magnesium matrix. The occasional small bright particles were identified in our previous investigations as the stable Mg24 Y5 phase [22] . The secondary phase characterisation in SLM and SPS samples was performed via a combination of EDS chemical mapping ( Fig. 4 ) and selected area diffraction (SAED) pattern analysis in TEM ( Fig. 5 ). The EDS maps showed that the secondary phase particles in both materials contain yttrium, and some of them also oxygen. The dominant secondary phase particles in the SLM sample have a fine “curtain-like” character and contain both yttrium and oxygen ( Fig. 4 ). According to the subsequent TEM analysis, these particles are the nanocrystalline Y2 O3 phase ( Fig. 5 a). Their shape and distribution strongly suggest that these oxides formed on the outer shell of the former powder particles during gas-atomisation, which were abruptly disrupted during the SLM process. TEM investigation also revealed the presence of the Mg24 Y5 phase, however, the number of these particles was very low ( Fig. 5 a). The Y2 O3 phase in the SPS sample forms a continuous 3D net along the former powder particle boundaries, as disclosed in Figs. 4 and 5 b. Its form is nanocrystalline, similar to the SLM sample, proving that this oxide was formed already on the initial powder particles during gas-atomisation. Nevertheless, the concentration of yttrium along these former powder particle boundaries was much higher than the Y/O ratio corresponding to the Y2 O3 phase. Therefore, the depletion of yttrium from the magnesium matrix along these boundaries ( Fig. 3 b) is caused by its segregation into these boundaries. The fine precipitates inside the former powder particles do not contain oxygen ( Fig. 4 ) and were identified as the YH2 phase ( Fig. 5 c). Occasional presence of larger clusters of crystalline (200 nm) Y2 O3 phase particles was found in TEM samples of both materials ( Fig. 5 c). Based on their type and shape, they are considered to be residuals from the casting process prior to atomisation. The grain structure of all investigated materials was thoroughly examined using EBSD. The orientation maps of the SLM and SPS samples were measured in two different planes, the top plane and the side plane, to investigate a possible anisotropy of the grain structure. All orientation maps are P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 1501 Fig. 5. TEM micrographs of (a) SLM sample and (b, c) SPS sample with corresponding SAED patterns. Fig. 6. EBSD orientation maps of the microstructure of (a) SLM top section (Z-direction), (b) SLM side section (X-direction), (c) SPS top section (SLD), (d) SPS side section, (e) EX samples (ED) and (f) distribution of grain boundaries type. (a-e) overlays: black –HAGBs, red –LAGBs, green –twin GBs. shown, together with the EX sample (only perpendicular to ED), in Fig. 6 . According to the micrographs, there is little difference between the two sections of the individual samples. Nevertheless, there is a significant difference in the grain structure resulting from the different processing methods. The SLM sample is formed by relatively large grains containing a high number of twins ( Fig. 6 a and 6 b). The grains are prolonged to several hundreds of microns, have curly grain boundaries, and the colour variation within the grains indicates a relatively high residual strain. The character and 1502 P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 Fig. 7. EBSD pole figures calculated for the investigated samples. distribution of the twin boundaries and low/high angle grain boundaries (LAGB/HAGB) are depicted in overlays. Note that the lowand high-angle grain boundaries had misorientation limitations set to be 2 °−15 °(LAGB) and > 15 °. The overall number fraction of each grain boundary type is shown for all samples in Fig. 6 f. The SPS resulted in a significantly different grain structure. Fully recrystallised equiaxed grains with insignificant residual strain, a low fraction of LAGBs and only a negligible fraction of twin boundaries were observed in this material ( Fig. 6 c and 6 d). The average grain size was calculated to be ∼20 μm. The grain boundary maps reveal that the final grain structure after SPS is partially related to the initial powder particle shape. Locally, the grain growth through the former powder particle boundaries was hindered due to the presence of an oxide layer. The EX material exhibited a fully recrystallised grain structure with equiaxed grains of ∼32 μm in diameter. No twin boundaries and only a low number fraction of LAGB were observed. The texture analysis results, in the form of EBSD pole figures, are shown in Fig. 7 . Interestingly, the EX and SPS samples showed only a very weak texture ( Fig. 7 ). In the SPS sample, the texture is {0001} basal texture parallel to the sintering loading direction, and in the EX sample, it is {11–20} fibre texture parallel to the ED. The texture strength in these two samples was below 2 times random. A significantly stronger texture was measured in the SLM sample. The orientation of the dominant component is close to {11– 20} < 10–10 > parallel to the Z direction, and its strength is ∼5 times random. 3.2. Mechanical properties The impact of microstructural differences resulting from different processing techniques on the mechanical strength was at first investigated using Vickers microhardness testing. In the case of the SLM sample, the microhardness was measured in two planes (XY, YZ) in order to reveal potenTable 1 Vickers microhardness of all the investigated samples. Material microhardness [HV] SLM (XY plane) 50.0 ±4.7 SLM (YZ plane) 52.0 ±3.5 SPS 49.9 ±1.7 EX 57.0 ±3.0 tial anisotropy with respect to the building direction. The tests revealed that all samples, including the two planes in the SLM sample, exhibited uniform microhardness distribution over the measured areas; the resulting average values are summarised in Table 1 . All three microstructural conditions prepared by different methods resulted in very similar microhardness ( ∼50 HV). Furthermore, no anisotropy was observed in the SLM sample as microhardness values measured in XY and YZ planes of the SLM sample agree within the statistical error. The compressive and tensile deformation tests were performed to reveal the mechanical response of the investigated samples under uniaxial load. The sample orientation was the same for tensile/compressive tests: parallel to X direction for the SLM sample, perpendicular to SLD for the SPS sample and parallel to ED for the EX sample. DIC was also performed to observe strain distribution during tensile tests. The true plastic stress-strain curves for compressive and tensile tests are presented in Fig. 8 . The compressive deformation curves displayed a sigmoidal shape ( Fig. 8 a) for all samples, indicating activity of the {10–12} mechanical twinning during compression [ 23 , 24 ]. In contrast, the tensile tests exhibited a power-law curve typical of deformation mediated solely by dislocation slip ( Fig. 8 b) [25] . Mechanical parameters evaluated from the deformation curves, namely the compressive and tensile yield strength, ultimate tensile strength, and elongation to fracture, are presented in Table 2 . Note that the elongation to fracture was evaluated from the DIC data and represents the maximal elongation in tension before specimen fracture. P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 1503 Fig. 8. The representative true stress-strain (a) compressive and (b) tensile curves of the SLM, SPS and extruded samples. Note different y-axis. Table 2 The mechanical parameters (compressive/tensile yield strength - CYS/TYS, ultimate tensile strength –UTS, and fracture elongation in tension - εmax ). Material SLM SPS EX CYS [MPa] 117 ±7 108 ±3 107 ±3 TYS [MPa] 113 ±3 115 ±2 99 ±1 UTS [MPa] 180 ±4 200 ±3 186 ±1 εmax [%] 9 ±1 21 ±3 30 ±3 Despite the different microstructures, the resulting yield stress values are very similar for all studied conditions and both deformation directions. In compression tests, the major difference in the samples lies in the magnitude of twinning (more pronounced sigmoidal curve in the case of the EX material) and in the deformation to fracture. In the tensile tests, the strengthening rate is similar for all three materials, but the fracture elongation differs significantly. The SLM sample exhibited lower UTS and εmax (180 MPa, 9%) compared to both SPS (200 MPa, 21%) and EX (186 MPa, 30%) materials. Fig. 9 shows the evolution of strain distribution of all the sample types during tension up to 1% before fracture calculated using the DIC technique. The least ductile SLM sample developed considerable strain localisation as early as at 1% global plastic strain in terms of several highly strained deformation bands and spots. Subsequently, some of these sites evolved into the critical crack, leading to failure, as marked with two white ellipses in Fig. 9 . On the other hand, the SPS and EX samples exhibited much more homogeneous local strain evolution up to higher global strains, although they differed in the critical crack initiation. Specifically, in the case of the SPS sample, the strain distribution was homogeneous up to at least 15% of global strain, and a distinct critical crack formed at the edge of the sample only shortly before failure. The most ductile EX sample exhibited homogeneous local strain up to 15%, too, however, the crack formation was preceded by the characteristic necking: some strain localisation in the central part of the sample occurred already at 22% of global strain and was followed by a development of distinctive neck identified before failure. Both the nature of the fracture and the strain localisation were consistent for all measured samples of the given material set. The origin of the strain localisation in the SLM sample and, particularly, in the weak spots 1 and 2 (cf. Fig. 9 ) was revealed by the observation of fracture surfaces by SEM, see Fig. 10 a and b. The fracture morphology and only limited signatures of plasticity (plastic dimples) confirm that the fracture took place in a brittle manner. The fracture surfaces of SPS and EX samples are shown in Fig. 10 c and d, respectively. The fracture surface of the SPS sample contains signatures of intensive plastic deformation and almost no silhouettes of the original powder particles. In the case of the EX sample, the analysis of the fracture surface revealed both the presence of intergranular fracture without visible plastic dimples, as well as severely plastically deformed areas. 3.3. Corrosion resistance The linear polarisation tests were conducted on all studied materials to investigate the initial susceptibility (10 min of stabilisation) of the samples to the corrosion attack in 0.1 M NaCl aqueous solution. The corresponding curves are presented in Fig. 11 . Notably, all polarisation curves showed a similar shape, and the most pronounced difference is the shift of the corrosion potential ( Ecorr ) of both powder metallurgy samples towards less noble values compared to the extruded sample. The corrosion current density ( icorr ) evaluated according to the Stern analysis was the same for the SPS and extruded samples, while the SLM sample showed a little lower corrosion current density, see Table 3 . To further examine the corrosion behaviour of the investigated samples, 7-day immersion tests were conducted, and the corresponding corrosion rates were calculated. The corrosion rates of the studied materials are presented in Table 3 . From the results, it is evident that the extruded and SPS materials exhibited similar corrosion rates within the error. On the 1504 P. Minárik, M. Zemková, S. Šašek et al. / Journal of Magnesium and Alloys 12 (2024) 1496–1510 Fig. 9. The DIC results show the evolution of strain distribution during tensile tests in terms of Lagrangian strains ( εyy ) in the loading, i.e. vertical, direction for all three types of materials. εyy was evaluated at several points of global strain ( ε) depending on the fracture strain εmax of particular material: SLM – 1% and 1% before fracture ( ∼7%), SPS –1%, 8%, 15%, and 1% before fracture ( ∼21%), EX –1%, 8%, 15%, 22%, and 1% before fracture ( ∼29%). Note different colorbar scales throughout the figure. Table 3 The corrosion characteristics determined from Fig. 11 and corrosion rate (CR) evaluated from the immersion test for all investigated samples. Material SLM SPS EX Ecorr [V vs. SCE] −1.620 ±0.004 −1.620 ±0.003 −1.604 ±0.003 icorr [mA cm2 ] 0.34 ±0.03 0.45 ±0.03 0.46 ±0.02 CRicorr [mg cm2 h−1 ] 0.15 ±0.01 0.20 ±0.01 0.21 ±0.01 CR [mg cm2 h−1 ] > 0.603 ∗0.19 ±0.02 0.17 ±0.01 ∗estimated degradation within 5 days of immersion.