Biodegradable WE43 Magnesium Alloy Produced by Selective Laser Melting: Mechanical Properties, Corrosion Behavior, and In ‐ Vitro Cytotoxicity

: In this work, selective laser melting (SLM) technology was used to prepare Mg ‐ 4Y ‐ 3Nd ‐ Zr (WE43) alloy. This alloy and production method are promising for the design of biodegradable implants. The aim of this study was to investigate the chemical composition, microstructure, me ‐ chanical properties, corrosion behavior in simulated body fluid (SBF), and cytotoxicity of the alloy produced by SLM method and to compare it with conventionally gravity cast reference alloy. Anal ‐ ysis of the surface of the revealed an oxygen content of 7 wt.%. Undesirable unmelted and only partially adhered spherical particles of the starting powder were also found. The microstructure of the material was very fine and consisted of α‐ Mg dendritic matrix, β‐ Mg 41 (Nd, Y) 5 intermetallic phase, Y 2 O 3 inclusions, and 0.6 vol.% of residual porosity. The Vickers hardness, compressive yield strength, compressive strength, and maximum compressive strain were 88 HV0.1, 201 MPa, 394 MPa, and 14%, respectively, which are close to the reference values in as ‐ cast. The in vitro corrosion rates determined by immersion and potentiodynamic tests were 2.6 mm/year and 1.3 mm/year, re ‐ spectively. Cytotoxicity tests indicated good biocompatibility of the 3D ‐ printed alloy.


Introduction
Magnesium and its alloys are interesting technical materials.Thanks to its excellent physical and chemical properties such as low density, high specific strength, good damping performance, it is widely used in the automotive and aerospace industries [1].Magnesium has also been widely studied for applications in medicine.In the medical field, good biocompatibility of Mg is important, but also good mechanical properties and Young's modulus, which is very similar to that of human bone [2][3][4][5].Magnesium is one of the most important elements in the human body, so it is non-toxic to the human body and degrades spontaneously in the body, and it can be used as a biodegradable material for implants, especially for bone fixation, so that reoperation after a fracture is not necessary.It can be also used as a cardiovascular stent [2,[6][7][8][9].
One of the commonly used commercial magnesium alloys is an alloy containing rare earth (REE) elements, such as Nd, Y, Dy, Gd, and Zr, referred to as Mg-4Y-3RE-Zr or WE43 alloy.This alloy shows a good combination of strength and plasticity and increased corrosion resistance [10,11].The corrosion resistance is due to the addition of rare earth elements, especially yttrium, which forms a protective layer of Y2O3 slowing down the diffusion and transport element between the material and the corrosive environment [6,12,13].However, this protection only works if sufficient Y is present in the alloy, otherwise localized forms of corrosion can occur [12].In the case of magnesium and magnesium alloys it is very necessary to avoid impurities, especially Fe, Ni, Co, Cu, which significantly worsen the corrosion resistance due to the formation of galvanic microcells [4,11,14].Another problem with magnesium and its alloys is the excessive amount of hydrogen produced during the corrosion process.Hydrogen gas can cause health complications, slow down the healing process, and reduce mechanical contact between the implant and the tissue [4,15,16].WE43 alloy is one of the most promising candidates for biodegradable medical implants and is therefore the subject of our study [10,13,17,18].
SLM technology for various industrial materials and components has been well established in the past.However, to the best of our knowledge, the number of reports on SLM produced Mg-based alloys is still limited due to difficulties in handling and processing highly reactive Mg-based powders [3,18,34].Therefore, a comprehensive study of SLM-prepared WE43 alloy is reported in this work.The study includes structural and chemical characterization, mechanical and corrosion analysis, as well as cytotoxicity investigation.

Material Preparation
In this study, WE43 alloy prepared by SLM method was investigated.The input material was WE43 powder atomized with an inert gas (the analysis of this powder is presented in [35,36].The bulk material was produced from powder by SLM process on an SLM 280HL machine (SLM Solutions Group AG, Lübeck, Germany) equipped with YLR-Laser (IPG Photonics, Oxford, MA, USA) with a wavelength of 1060 nm.The laser beam had a Gaussian profile with a focus of 82 μm.The printing process was carried out under argon atmosphere flow to keep the oxygen value below 0.2%.Cube samples of 8 × 8 × 8 mm in size (Figure 1) were printed with a laser power of 175 W, a scanning speed of 800 mm/s, a hatching distance of 90 μm, and a melted layer thickness of 50 μm.The samples were cut into smaller pieces that were used for microstructural characterization, mechanical, corrosion, and biological testing.As-cast ingot purchased from an industrial supplier was used as a reference material.

Microstructure Characterization
A standard metallographic procedure according to EN ISO 10993-18 standard was used to prepare samples for microstructural characterization.The samples were ground on SiC grinding papers of grain size P800-P2500, polished on D2 diamond paste and finally polished on Etosil E silica suspension.
The microstructure was observed using a TESCAN VEGA-3 LMU scanning electron microscope equipped with an EDS analyzer (Aztec-Oxford Instruments, Brno, Czech Republic).Porosity was evaluated on several cross-sections by image analysis in ImageJ software.
More detailed microstructural observation were carried out using a Jeol 2200FS transmission electron microscope.Samples were cut with a diamond spark (perpendicular to the built-up direction) from the sample of material in thin slices of about 0.5 mm thickness, which were then ground to a thickness of 50 μm.Circular samples of 3 mm diameter were cut from the thinned slices.The samples thus prepared were further thinned in the center to a final thickness of 10 μm using a Gatan Dimple Grinder Model 656.These samples were then ion polished using a Gatan PIPS model 691, which allowed areas of the material to be observed by TEM.

Mechanical Testing
The hardness was measured by the Vickers method according to EN ISO 6507-1 standard at a load of 100 g (HV 0.1) and a loading time of 10-15 s.Twenty measurements were taken for each sample and the mean values and standard deviations were calculated.The mechanical properties in compression were measured on samples with dimensions 3.5 × 3.5 × 6.5 mm according to EN ISO 604 using the universal LabTest 5.250SP1_VM machine.A constant strain rate of 0.001 s −1 was used for the compression tests.

Corrosion Properties
Two corrosion tests were performed on the samples-potentiodynamic and immersion test.Both measurements were carried out in SBF (simulated body fluid) prepared according to Müller-Table 1.An SBF volume of 100 mL per 1 cm 2 of sample surface was used for the immersion test.The samples were fixed in plastic holders and immersed in SBF for 3 days at 37 °C.The corrosion rate was evaluated based on the weight changes after the removal of corrosion products according to standard EN ISO 10993-15.Four measurements were performed for each material.Potentiodynamic measurements were performed in a corrosion cell using a conventional three-electrode circuit, with the sample used as a working electrode.A platinum wire was used as the counter electrode and a saturated calomel electrode (SCE) was used as the reference electrode.The measurements were carried out at room temperature on a Gamry PC3 instrument.The following electrochemical tests were performed: measurement of the free corrosion potential (Ekorr) for 60 min, measurement of the polarization resistance and measurement of the potentiodynamic curve.The polarization resistance was measured in the range ±20 mV vs. Ekorr at a scan rate of 0.1 mV s −1 .Polarization resistance and corrosion rate were calculated according to Equations ( 1)-( 3), where ba and bc represent the anodic and cathodic Tafel slope, respectively [37].
The cathodic part of the curve was measured at the conditions: 0.2 V vs. Ekorr to −2.5 V vs. Ereff.The anodic part of the curve was measured at the conditions: −0.2 V vs. Ekorr to −0.6 V vs. Ereff.The scan rate for both parts of the potentiodynamic curves was 1 mV s −1 .Subsequently, the corrosion products after immersion testing were analyzed by SEM/EDS.

Cytotoxicity Test
The in vitro cytotoxicity test (elution test) was performed accordance to ISO 10993-5.Samples (6 × 7.5 × 7.5 mm) were sterilized in 70% ethanol (2 h) and under UV light (2 h).Two alloy samples were transferred in MEM cultivation medium (Minimal Essential Medium, Sigma, M0446) with 5% fetal bovine serum (FBS, Sigma F7524) and agitated (130 RPM) at 37 °C in closed vessels for 24 h.The reduced concentration of FBS was used as recommended in the standard (FBS can mask the toxicity).The surface-to-volume ratio was 87.5 mm 2 per ml because for degradable alloys, it is advisable [38] to use higher S/V than recommended in the standard.
Murine fibroblasts L929 (ATCC ® CCL-1™) were maintained in MEM (Sigma, M0446) medium with 10% FBS at standard conditions.Prior to the exposition, murine fibroblasts L929 were trypsinized and resuspended in MEM + 10% FBS to create a suspension with a concentration of 1 × 10 5 cells per ml.Thereafter, 100 μL of the cell suspension was seeded into a 96-well plate, which means the seeding density of 1 × 10 4 cells per well.
After one day, the medium was replaced by the extracts prepared as described above.Extracts were used either in an undiluted form (100% extracts) or diluted with the medium (50 and 25% extracts).Hexaplicates were used for each sample.Sole MEM with 5% FBS served as a control (untreated cells).After one day of incubation with the extracts, cell metabolic activity was evaluated using resazurin assay [39].Extracts were removed, wells were washed with PBS (phosphate buffered saline), and resazurin solution (final concentration 25 μg/mL) in MEM + 10% FBS without phenol red was added.After 2 h of incubation, fluorescence at 560/590 nm (excitation/emission) was measured.Cytotoxicity of the extracts was depicted as a percentage of metabolic activity of the control.Extracts causing the decrease below 70% of the activity of the control (untreated cells) are considered cytotoxic, as described in the standard ISO 10993-5.
The concentration of released ions was measured by ICP-MS.Ultrapure HNO3 was added to the extracts (10 μL/mL) prior to the measurement.

Microstructure and Chemical Composition
The microstructure of the WE43 alloy in the as-cast condition is illustrated in Figure 2. It can be seen that this material has a relatively coarse microstructure composed of α-Mg dendrites (dark) and eutectic mixture (light) in the interdendritic regions.According to EDS and XRD analysis (not shown), the eutectic is mainly enriched in Nd and contains α-Mg + β-Mg41Nd5 phases [36,40].In the β intermetallic phase, Nd is partially substituted by Y and other RE elements and can thus be written as β-Mg41(Nd, Y)5.The coarse structure results from the slow cooling of the melt during gravity casting.The porosity of the alloy in the as-cast state was not detected.Figure 3 shows the surface of the 3D-printed samples studied.It can be seen from the figure that the surface contains unmelted and only partially attached powder particles.Such particles are often found on the surface of alloys produced by SLM [24,[41][42][43][44].In these areas, the powder material was only partially melted by the laser beam and the connection with the previous layer was insufficient.This is the so-called balling effect, which results from (i) discontinuous weld mark, (ii) insufficient heating of the powder at the end of the SLM building process, and (iii) oxide coatings on the powder particles that prevent their melting.The powder particles are prone to release from the surface of the material, which can have detrimental effects on the surrounding tissues.In addition, they can also form stress concentrators and negatively affect the mechanical (especially fatigue) properties of the material.For these reasons, they should always be removed from the implant surface by a mechanical means or chemical etching.The element distribution maps in Figure 4 and the EDS analysis of the surface show an increased oxygen concentration (light area) over almost the entire surface of the 3D printed material.Residual oxygen coming from the atmosphere in the 3D printing chamber reacts with both Mg and rare earth metals to form surface oxide skins.EDS analysis showed 7 wt.%O concentration on the surface, but the oxides were not detected by XRD due to their small thickness.Figure 5 shows SEM images of a cross-section of the 3D-printed material at different magnifications.The microstructure is very fine and consists of α-Mg matrix (grey), oxide particles (light), and β-Mg41(Nd, Y)5 intermetallic phases (light).In Figure 5, at the highest magnification, the very fine dendritic morphology is visible.The microstructural refinement of the 3D-printed material is a direct consequence of the very high melt cooling rates (up to 106-108 K/s) during the SLM process [32].
The macro view in Figure 5 also shows the presence of residual porosity within the 3D-printed WE43 alloy.The volume fraction of porosity was measured by imaging analysis on several cross-sections of the material and was determined to be 0.6 vol.%.Residual porosity is an intrinsic property of materials produced by SLM and can result from several phenomena such as (i) gas entrapped in the powder during SLM; (ii) local superheating and evaporation of the alloy due to high laser power density; (iii) local imperfect melting and bonding of powder particles due to insufficient laser power density and other factors [10,36,[45][46][47][48][49][50].In our case, the volume of residual porosity is very low, indicating that the laser beam parameters in the SLM process were selected and controlled properly.A more detailed view and maps of the distribution of elements within the reference as-cast material, where we can see the structure without oxide (Figure 6) and the 3Dprinted material (Figure 7) show that there are separate oxide inclusions (light) enriched mainly in Y, which is due to a high affinity of Y for O.The oxides inside the matrix come from (i) the oxidized surface of the starting powder particles and (ii) the reaction of the powder and melt with residual oxygen in the working chamber of the SLM machine.For this reason, the oxide particles have elongated shapes, as will be shown in the detailed TEM view below.The SEM image in Figure 7 also clearly shows the dendritic boundaries with intermetallic phases β-Mg41(Nd, Y)5.In the SEM mode, these boundaries appear lighter than the surrounding Mg matrix due to enrichment of Nd and other RE elements.In addition, very fine light particles of intermetallic phases are observed inside the dendrites as light spots.The most detailed microstructural observation of the 3D-printed WE43 alloy was performed by TEM + EDS (Figure 8). Figure 8a shows a detail of the microstructure in the vicinity of dendrite boundaries.According to the X-ray elemental maps included in this figure, three structural components are identified: (i) α-Mg matrix dendrites (light); (ii) β-Mg41(Nd, Y)5 particles inside the dendrites and at the dendrite boundaries (dark); and (iii) a very fine mixture of α-Mg + β-Mg41(Nd, Y)5 along the dendrite boundaries.The β-Mg41(Nd, Y)5 intermetallic phases are extremely fine (well below 1 μm).The microstructure of the WE43 alloy in Figure 8a is a direct result of the specific temperature regime during the SLM process.It is conducted in the layer-by-layer manner so that the laser beam melts only a small volume of powder each time.As a result, rapid cooling the melt at rates up to 103-106 K/s occurs during SLM [32,[51][52][53].The rapidly solidified microstructure is therefore extremely fine (Figures 5, 7 and 8) and contains dendrites of supersaturated solid solution α-Mg [11,36,54,55] and interdendritic particles β-Mg41(Nd, Y)5.Afterward, very fine intermetallic phases β-Mg41(Nd, Y)5 precipitate in the solid state inside the dendrites and along their boundaries due to the decomposition of the solid solution of α-Mg supersaturated with RE elements.Precipitation is accelerated by the elevated temperature in the SLM working chamber.The finest β-Mg41(Nd, Y)5 phases are located mainly along the dendrite boundaries where α-Mg supersaturation is highest.Figure 8 (a), (b) shows details of the oxide inclusions in the 3D-printed material together with EDS maps of the Mg, O, Y, and Nd elements.The elemental maps show that yttrium oxide dominates in the inclusions.Nd is mainly concentrated in the fine intermetallic phases β-Mg41(Nd, Y)5.In cross-section, the yttrium oxide inclusions have a typical elongated "river-like" shape, suggesting that their origin is in the oxide shells on the particles of the input powder.

Mechanical Properties
Table 2 summarizes the results of the hardness tests.For both materials, 20 measurements were carried out.It can be seen that the hardness of the 3D-printed WE43 alloy is slightly lower compared to the as-cast reference material, despite its refined microstructure (Figures 5, 7, and 8).This difference can be attributed mainly to the presence of oxide particles, which weaken the compactness of the 3D-printed material (Figures 7 and 8b).In addition, the residual porosity observed in the 3D-printed alloy may also play a role (Figure 5).

3D Printed Material
Reference As-Cast Material HV 0.1 88 ± 7 96 ± 12 Figure 9 and Table 3 summarize the results of compression tests.It can be seen that the compressive strength of tested materials shows a similar trend to the hardness, i.e., the 3D-printed material has a slightly lower strength than the reference material as-cast product.From the point of view of compressive plasticity, both materials behave similarly.The decreased compressive strength of the 3D-printed WE43 alloy reflects its microstructural properties, mainly the presence of oxide particles and slight residual porosity (Figures 5, 7, and 8).It is important to note that neither pores nor oxides have a significant effect on the plasticity in compression, as demonstrated by the comparable plasticity of the 3D-printed material to that of the WE43 reference as-cast materials.The 3D-printing SLM technology thus offers a promising alternative to conventional metallurgy in the production of Mg alloy components.3D printing of Mg alloys has been described in [3,13,45,56].The authors prepared WE43 alloy by 3D printing and the results of the compression test were comparable to the values reported in this paper.The variation in values depended on the porosity of the material and whether the measurements were taken before or after exposure to the corrosive environment.

Corrosion Properties
The results of the potentiodynamic measurements are summarized in Figure 10 and Table 4.It can be seen that the corrosion potential of the 3D-printed material is slightly shifted to more positive values compared to the reference as-cast material.This is probably due to the presence of oxides with nobler behavior in the 3D-printed alloy (Figures 5  and 7).The corrosion rates in the SBF solution determined by Tafel extrapolation of the potentiodynamic curves are close to the value 1 mm/year for both tested materials.The corrosion test was also carried out by immersion testing in SBF and the results are summarized in Table 5.For the reference as-cast material, the corrosion rate is similar to that determined by Tafel extrapolation.In contrast, the 3D-printed alloy shows twice the corrosion rate in the immersion test.The enhanced corrosion rate of the 3D printed alloy is related to the existence of a specific microstructure of this alloy.As shown previously (Figures 5, 7 and 8), the alloy contains yttrium oxide inclusions and residual porosity.Both factors facilitate corrosion attack by creating paths for the corrosion environment to penetrate the material more rapidly.

3D-Printed Material
Reference As-Cast Material Corrosion rate (mm/year) 2.6 ± 1.9 1.0 ± 0.5 The SEM view and EDS analysis of the corrosion products on the 3D-printed material after SBF exposure are presented in Figure 11 and Table 6.The corrosion products are relatively uniformly distributed over the entire surface of the material.Due to the inorganic nature of corrosion products and the drying process before the SEM examination, they contain numerous cracks.The main components of corrosion products are Mg, O, P, and Ca.This means that they consist mainly of mixed calcium-magnesium phosphates and carbonates.Similar findings have been reported by other authors [36,[57][58][59][60].The mechanism of alloy corrosion in SBF includes several chemical reactions, which can be expressed by the following simplified equations [4,15,16,61]: Mg 2+ (Ca 2+ ) + OH − + HCO3 -→ (Mg,Ca)CO3 + H2O (formation of insoluble product) Table 6.EDS analysis of the corrosion products on the 3D-printed alloy after exposure in SBF EDS analysis of the corrosion products on the 3D-printed alloy after exposure in SBF.In terms of biocompatibility, the formation of phosphates during corrosion facilitates bone tissue ingrowth and good osseointegration of the 3D-printed implant with bone because of the similarity between the corrosion products and the inorganic part of the human bone [3,5,57,58,62,63].Also, a corrosion rate of 2.6 mm/year may be acceptable for medical implants such as bone fixation screws.If we consider, for example, a screw with a thickness of 5 mm, its degradation will occur approximately one year after implantation.Therefore, the screw could have a load-bearing capacity for a sufficiently long period of bone tissue healing.The minimum degradation time of a biodegradable implant should not be less than 6-12 months [64].The biodegradable metallic material should provide sufficient strength to the healing bone for at least 12 weeks, provided, of course, that it is non-toxic and does not have a negative effect on the body [65].In the case of magnesium alloy implants, the rate of implant degradation should not exceed the rate of healing of the tissue itself, which in an adult is at least 12-18 weeks, to maintain the mechanical integrity of the implant [66].

In-Vitro Cytotoxicity According to ISO 10993-5
In vitro cytotoxicity of the 3D-printed WE43 alloy was tested according to ISO 10993-5 standard.Results are expressed in Figure 12 as relative metabolic activity of L929 cells after 1 day incubation with 1 day extracts of the alloy.Figure 12 shows that metabolic activity of the cells incubated with 100% extracts of both samples was above the cytotoxicity limit of 70% and, moreover, none of the diluted extracts show any cytotoxicity.Table 7 gives concentrations of released ions in extracts of both samples and pure medium.It is observed that magnesium matrix of the alloy dissolves preferentially in the medium.Concentrations of potentially toxic RE elements is lower than Mg by three orders of magnitude.Magnesium is generally considered as relatively non-toxic element [18,38,60,67].Cytotoxicity experiments indicate a good biocompatibility of the 3D-printed WE43 alloy.

Conclusions
The aim of this work was to provide a comprehensive study of 3D-printed biodegradable alloy WE43, including chemical, microstructural, mechanical, corrosion, and cytotoxicity tests.The main conclusions can be summarized as follows: 1. Great attention should be paid to the surface state of the 3D-printed alloy.Unmelted and only partially adherent spherical particles of the initial powder alloy were found on the surface.These particles are undesirable because they are detrimental to the surrounding tissue and negatively affect the mechanical (especially fatigue) properties.2. The microstructure of the material is very fine and contains α-Mg dendritic matrix, β-Mg41(Nd, Y)5 intermetallic phase, Y2O3 inclusions, and 0.6 vol.% of residual porosity.3. Thanks to the fine microstructure, the hardness, compressive yield strength, ultimate compressive strength, and maximum compressive strain showed values similar to those of a conventional casting.4.An in vitro corrosion rate of 2.6 mm/year was obtained in the SBF immersion test.
This value may be acceptable for some implant types.Despite the satisfactory results of the corrosion experiments, it seems that further improvement of the corrosion properties of the material is necessary for future applications.This improvement can be achieved, for example, by appropriate control of the SLM process, reduction of the oxidation rate, surface treatment, etc. 5. Standard in vitro cytotoxicity tests indicated good biocompatibility of the 3D-printed WE43 alloy studied in this work.

Figure 2 .
Figure 2. Microstructure of the reference as-cast WE43 alloy at different magnifications.

Figure 4 .
Figure 4. SEM image and EDS maps of elements distribution on the surface of 3D-printed material.

Figure 5 .
Figure 5. Microstructure of the 3D-printed material at different magnifications.

Figure 6 .
Figure 6.SEM image and EDS maps of elements distribution inside the reference as-cast material.

Figure 7 .
Figure 7. SEM image and EDS maps of elements distribution inside the 3D-printed material.

Figure 8 .
Figure 8. TEM micrograph and elements distribution maps of the 3D-printed WE43: (a) detail in the vicinity of dendrite boundaries, (b) detail of oxide inclusions.

Figure 11 .
Figure 11.Surface of the 3D-printed alloy corroded in SBF solution (SEM) and element distribution maps (EDS).

Figure 12 .
Figure 12.Relative metabolic activity of the L929 cells (resazurin assay) after 1 day incubation with extracts of the samples.Sole extraction medium served as a control (untreated cells, for which metabolic activity was taken as 100%) (sample 1 and sample 2 are 3D-printed materials).The line indicates the normative limit of 70% of the metabolic activity of the control.Error bars stand for the standard error deviation of six replicates.

Table 1 .
Composition of SBF used in corrosion tests.

Table 3 .
Mechanical compressive properties of the 3D-printed material and as-cast reference material.

Table 4 .
Results of potentiodynamic test of 3D-printed and reference material (SBF solution).

Table 5 .
Results of immersion tests in SBF solution at 37 °C.

Table 7 .
Concentration of elements in extracts and pure medium (ICP-MS).