Mihl Knpk,Mári Zmková,Adm Gr?,Ev Jblonská,Frnti?k Lukáˇ,d,Robrt Král,Jn Bohln,Ptr Minárik

a Charles University,Faculty of Mathematics and Physics,Ke Karlovu 5,12116 Prague,Czech Republic

bNuclear Physics Institute of CAS,ˇRe?130,25068ˇRe?,Czech Republic

c University of Chemistry and Technology,Department of Biochemistry and Microbiology,Technická5,Prague,16628 Prague,Czech Republic

d Institute of Plasma Physics of CAS,Za Slovankou 3,18200 Prague,Czech Republic

e Helmholtz-Zentrum Geesthacht,Magnesium Innovation Centre(MagIC),D21502 Geesthacht,Germany

Abstract Alloying of Mg with rare-earth(RE)elements proved to be beneficial for their in-vitro and in-vivo performance.In this work,a novel WN43(Mg-4wt%Y-3wt%Nd)alloy with a well-defined composition was prepared,where,unlike in the commercial WE43 alloy,the possibly harmful RE mischmetal was substituted by less toxic Nd.A modern spark plasma sintering(SPS)technique was used to effectively produce WN43 samples from atomized powders.Sintering temperatures of 400°C–550°C and holding times of 3 or 10min were used and wellcompacted final materials were successfully prepared.It was shown that a superior combination of corrosion and mechanical properties was attained in the samples sintered at 500°C and 550°C,while the effect of sintering time was rather negligible.The performance of this material was exceptional within the group of Mg alloys prepared by powder metallurgy and comparable with conventionally prepared alloys.Moreover,it was shown that a great variety of mechanical and corrosion characteristics can be obtained by altering the SPS parameters so as to fulfill case-specific requirements typical of biomedical materials.Consequently,the novel WN43 alloy prepared by SPS seems to be a particularly suitable material for biomedical use.? 2021 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;Biocompatibility;Powder;Sintering;Corrosion;Rare earth element.

1.Introduction

Magnesium alloys exhibit a number of distinctive properties,such as high strength-to-weight ratio,high damping,low elastic modulus,impact resistance,dimensional stability,non-toxicity,and biodegradability[1,2].The combination of these characteristics has brought new possibilities for practical,particularly biomedical applications and the studies of magnesium alloys have intensified in the last decade[3–5].Research in the field of biological implants has been growing substantially and new types of materials are being persistently developed and clinically tested[6].The primary motivation for the medical use of magnesium is that the human body is able to effectively metabolize and dispose even excessive amounts through kidneys[3,7].Unlike permanent implants made of e.g.stainless steel,titanium or cobalt-chromium alloys,the employment of magnesium alloys allows for new prospects in orthopedics and surgery because natural degradation of magnesium in biological media is vastly advantageous for designing temporary implants.In this way,secondary surgeries are avoided(as the implant naturally dissolves in the body)and patients benefit from shortened recovery periods and decreased costs.Furthermore,conventional permanent implants impose the risk of heavy metals release due to undesired long-term degradation,whereas magnesium is naturally present and harmless to the human body[8],provided pertinent alloying elements are carefully selected[1,9–11].Moreover,the mechanical properties of magnesium alloys are reasonably similar to those of cortical bone and,therefore,can be used as load-bearing implants introducing rather low stress shielding[10,12].

Yet,biodegradable implants must satisfy various requirements.One of the essential concerns in magnesium-based materials is the achievement of appropriate corrosion rates.The material must not decompose too fast and its mechanical properties must be retained for the required period.It was shown that pure magnesium exhibits a rather rapidinvitroandin-vivodegradation resulting in dubious mechanical performance,excess hydrogen gas production,and local pH increase,which can counteract proper healing[13].On the other hand,additions of alloying elements to pure magnesium can be essentially beneficial for biomedical applications[14].Certain alloying elements,including RE,are known to affect the mechanical properties especially through the formation of secondary phases,precipitation hardening,and/or solid solution strengthening[15–17].Moreover,particular elements,such as aluminum[18]or rare-earth(RE)metals[19]were shown to be capable of slowing down the corrosion rate.Magnesium alloys representative of this group are,for example,the commercially available WE43,AE42,and LAE442 alloys[20–25].The RE elements are typically added in relatively small amounts in the form of mischmetals,where one or two RE elements dominate.It was shown that the yttrium can effectively decrease corrosion rates,especially if homogeneously distributed in the magnesium matrix[19,26].Yttrium,although such classification is not impeccably accurate,is often regarded as RE element due to similar chemical and toxicological characteristics[27].The mechanism underlying the corrosion resistance improvement by RE elements is the stabilization of protective oxide or hydroxide films[28,29].The Mg-RE alloys were documented to exhibit favorablein-vitroandin-vivoperformance confirmed also in the pre-clinical tests[25,30,31]and first implants have become commercially available[32].

Furthermore,reduction in the grain size was found to be another important factor for the improvement of the corrosion resistance and the mechanical properties of magnesium alloys[33–35].Fine-grained microstructures can be produced using various methods of severe plastic deformation(SPD)[18,34].The alternative and often more effective way to produce materials featuring fine grains is the sintering of atomized material particles.Powder particles can be consolidated using various techniques,such as hot isostatic pressing or extrusion.However,these techniques involve long exposure times at high temperatures,leading to detrimental coarsening of the initial fine microstructures[36–38].These drawbacks can be eliminated using the modern spark plasma sintering(SPS)method,which operates at considerably lower sintering times and temperatures.Modern SPS facilities accurate control of the sintering parameters and,therefore,offer precise tailoring of the final microstructures.This method also allows for the production of desired material shapes unlike the conventional wrought magnesium alloys,which exhibit very restricted forming capabilities associated with low ductility.Yet,there are only a limited number of works focusing on magnesium-based materials produced by SPS[39–41].The present investigation was focused on the corrosion performance and compressive behavior of a novel WN43(Mg-4wt%Y-3wt%Nd)alloy produced by the SPS technique using different sintering parameters.In this alloy,neodymium was used instead of the RE mischmetal present e.g.in the commercial WE43 and WE54 alloys since the potentially detrimental effect of RE mischmetal elements in these alloys and possible related health issues have been recently discussed[42,43].It should be noted that the WE43 magnesium alloy was investigated in a number ofin-vivostudies as a promising candidate for the implant material,see e.g.[25,31].In addition,the only commercially available magnesium implant is the MAGNEZIX screw with the composition of Mg-Y-REZr which does not considerably differ from the commercial WE43 alloy[32].Yet,among the RE elements,neodymium exhibits relatively low toxicity and is presumably harmless at the involved concentrations[27,44,45].Hence,it was used in this study as a replacement for the unspecified mischmetal in order to obtain a“purified”version of WE-based alloys.Moreover,additions of neodymium were shown to have a positive impact on the corrosion resistance of magnesium alloys[11,46].In addition,this novel alloy is of a well-defined and relatively simple composition,which allowed us to effectively study and understand the evolution of microstructure as well as related corrosion and mechanical performance as a function of sintering parameters.Based on the experimental findings,the potential employment of this material for the production of biomedical implants was assessed.

2.Materials and methods

The novel WN43(Mg-4Y-3Nd)magnesium alloy was cast using a modified gravity casting procedure.The exact composition was measured by spark emission spectroscopy to be Mg-3.6wt%Y-3.4wt%Nd-0.02wt%Fe-0.003wt%Cu-0.001wt%Ni.The cast billets were machined,remelted and gas-atomized in an inert atmosphere at the Clausthal University of Technology,Germany.

The microstructures of the initial powder(and,later,of the sintered materials)were analyzed using a scanning electron microscope(SEM)FEI Quanta 200F operating in the backscattered electron(BSE)imaging mode,further equipped with an EDAX electron backscatter diffraction(EBSD)camera.The samples for SEM were mechanically polished down to 0.05μm using ethanol-based suspensions to prevent corrosion.The samples for EBSD were additionally ion-polished using a Leica EM RES102 ion beam milling system.The SEM microstructure observations revealed that the as-received gas-atomized powder particles were of a characteristic globular shape(Fig.1).The examination of the particle crosssections showed a very fine cell-like dendritic structure of the secondary phases typical of gas-atomized powders as a result of rapid cooling[39,40].These phases were identified by X-ray diffraction in the previous study[47]to be of theβ1type,having Mg3(Nd,Y)composition,as Y is partially dissolvable in the common Mg3Nd phase[35].The cell-like structure was evidently finer in smaller particles.Such a variance in the phase structure was most likely associated with differences in the cooling rate,which strongly depends on the size of powder particles.Moreover,the Y2O3phase formed as a shell around the individual particles due to the natural oxidation process.The volume fractions of the Mg3(Nd,Y)and Y2O3phases were both determined by XRD to be～1.5%,and the residual Y was supposedly dissolved in the Mg matrix[47].The EBSD investigation in that study revealed that the grain structure depended considerably on the size of individual particles.Small particles featured only a few grains or were single-crystalline,whereas larger particles(>～50μm)typically consisted of a number of smaller grains.The particle size distribution investigated by light microscopy showed a relatively broad character with the particle sizes up to～170μm while the calculated D50 value was～49μm[47].

Fig.1.The microstructure of the as-received WE43 powder observed by SEM(BSE).

SPS was carried out using the device type 10–4 from Thermal Technology LLC.The sintering was performed at four temperatures:400,450,500,or 550°C and holding times of 3 or 10min.The temperature was controlled using a thermocouple located in the graphite tooling in the vicinity of the sample.During heating,the external load of 100MPa was applied and held for the sintering period.After the sintering,the load was disengaged and the compacted material was cooled down(the apparatus itself serves as an effective heat sink).The phase composition was investigated by X-ray diffraction(XRD)using the D8 Discover diffractometer(Bruker)featuring a CuKαsource and a NiKβradiation filter.

The compression tests were carried out using the universal testing machine Instron 5882 at room temperature with an initial strain rate of 10?3?s?1.The samples with dimensions of 5×3×3mm3were cut from each material from the central part of the compacted pellet so that the loading direction(i.e.along the 5mm edges)was perpendicular to the direction of applied load during the production of the pellets.At least two samples were tested for each material condition.

The initial corrosion resistance was determined by the linear potentiodynamic polarization(PDP)and electrochemical impedance spectroscopy(EIS)methods using the potentiostat AUTOLAB128N utilizing a standard three-electrode setup and a rotating disk electrode(300rpm).The exposed surface was ground with SiC 1200(15μm)grit paper prior to each measurement.The tests were performed in the 0.1mol NaCl aqueous solution at room temperature after 10min stabilization for both the PDF and EIS measurements.The PDP measurement was performed in the potential range from?150mV to 200mV with respect to the open circuit potential(OCP)at a constant scanning rate of 1 mV?s?1.The EIS analyses were performed in the frequency range of 0.1–20MHz and 10mV amplitude with respect to OCP was used.At least three tests were carried out for each sample type.The corrosion rates of the materials prepared at different sintering conditions were determined using weight loss measurements.Samples with dimensions of 1.3×8×8mm3were cut from each material.The initial weight(m0)of each sample was documented and the sample was subsequently immersed into a 0.1mol NaCl aqueous solution for 7 days.After this period,the specimens were taken out and cleaned in order to remove the corrosion products according to the ISO 8407:2009 standard[48].Consequently,the weight of the samples(m)was measured again and the corrosion rate,Rc,was calculated as:Rc=(m0–m)/(t?A),wheretis the immersion time andAis the total surface area of the sample.At least three samples were studied for each material type.

The corrosion performance was also investigated in biological medium.Minimal essential medium(MEM,Sigma M0446)supplemented with 10% fetal bovine serum(FBS)was used as a medium frequently used for cultivation of cells during cytotoxicity testing.The corrosion test was performed in falcon tubes with vented caps(Corning)in a cultivation incubator in CO2atmosphere(so-called physiological corrosion).Samples(6×6×1mm3)were ground(P1200),washed with ethanol,weighted and sterilized in 70% ethanol for 2h and under UV for additional 2h.Samples were immersed into 40ml(surface to volume ratio wasS/V=～2.4 mm2/ml,i.e.V/S=～42ml/cm2)of media and incubated in 5% CO2atmosphere at 37°C on an orbital shaker for 14 days.Three replicates were used for corrosion rate determination from mass loss.

The in vitro cytotoxicity test(the elution test with extracts)was performed according to ISO 10,993–5.Samples(6×6×1mm3)were ground(P1200,P2000,and P4000),washed with ethanol and sterilized in 70% ethanol for 2h followed by irradiation with UV for additional 2h.The samples(in four replicates)were transferred to MEM cultivation medium(Sigma M0446)with 5% fetal bovine serum(FBS,Sigma F7524)and agitated(130rpm)at 37°C in closed vessels for 24 h.S/Vwas 17 mm2?ml?1.Murine fibroblasts L929(ATCCR○CCL-1TM)were maintained in MEM(Sigma,M0446)medium with 10% FBS at standard conditions of 37°C,5% CO2and 100% relative humidity.Cells were used from 3rdpassage after thawing and only up to 20thpassage.Prior the exposition,murine fibroblasts L929 were trypsinized and resuspended in MEM+10% FBS to create a suspension with a concentration of 1?105cells per ml.Thereafter,100μl of the cell suspension was seeded into a 96-well plate,giving the seeding density of 1?104cells per well.After one day,the medium was replaced by the extracts prepared as described above.The extracts were used in undiluted form(100% extracts)as well as diluted with the cultivation medium(50%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[49].Extracts were removed and resazurin solution(final concentration 25μg/ml)in MEM+10%FBS without phenol red was added.After 2 h of incubation,fluorescence at 560/590nm(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.

Fig.2.The SEM micrographs of the WN43 samples prepared by SPS at(a)400°C for 3min,(b)400°C for 10min,(c)550°C for 3min,and(d)550°C for 10min.

3.Results

3.1.Microstructure

The SEM(BSE)microstructure observations of the selected samples sintered at 400 and 550°C for 3 and 10min are shown in Fig 2.The micrographs revealed that the effect of sintering temperature on the resulting microstructure was significant while the effect of holding time was practically negligible.It can be seen that a relative density close to 100% was achieved in all the compacts.Nevertheless,the residual porosity(black areas in the micrographs,some of them marked by yellow arrows)decreased with increasing sintering temperature from～0.5%(at 400°C)to<0.1%(at 550°C)of volume fraction for both sintering times[47].After the sintering at 400°C for 3 and 10min,the distribution of the secondary phase particles did not change significantly and only a minor disruption of the original cell-like distribution(cf.Fig.1)accompanied by a coarsening of initial dendrites was witnessed(Fig.2a,b).On the other hand,sintering at 550°C led to a depletion of the Mg matrix of the secondary phases.Subsequently,significant segregation of Y and Nd took place primarily at the particle boundaries and some amount of newly formed fine precipitates appeared also in the central part of former particles,as disclosed in Fig.2c,d,and in detail in Ref.[47].Because of short sintering time and fast cooling,equilibrium was not reached and precipitates corresponding to the entire Mg-Nd-Y precipitation series[50]were identified in the sintered samples[47].Fig.3a-d shows the EBSD orientation maps in combination with image quality maps.Incorporation of the image quality maps was chosen as they are particularly sensitive to matrix defects,such as particle or grain boundaries and secondary phases.

Consequently,the configuration of individual grains,with respect to the secondary phase particles and former powder boundaries,is well visible in Fig.3.These micrographs showed that the grains were,in each case,oriented rather randomly(i.e.only insignificant crystallographic texture was observed,as also shown in[47]).On the other hand,the EBSD micrographs gave evidence for changes in the size distribution of the grains.With increasing sintering temperature,the grain size distribution narrowed as grains became more homogeneous in shape and size.Small grains in the samples sintered at 400°C were observed predominantly along the former powder particles as a result of strain localization due to the application of external load during the SPS processing.The temperature of 400°C was evidently not sufficient to promote extensive grain growth.On the other hand,the highest sintering temperature of 550°C caused distinctive recrystallization and grain growth,leading to the formation of new grain boundaries.Nevertheless,Fig.3c,d revealed that the grain growth was significantly limited by the former powder particles even during sintering at 550°C.It should be also noted that the average grain size did not change and the value of～18μm was documented across all the sample sets independently of the sintering conditions,suggesting that even though the grain size distribution became narrower,the effects of recrystallization and grain growth canceled out in terms of the average grain size[47].A similar effect was also observed in the AE42 magnesium alloy produced by SPS[39].For more details on the microstructure and phase evolution in the studied material see our previous work[47].

Fig.3.The EBSD orientation maps overlaid over the EBSD image quality maps of the WN43 samples prepared by SPS at(a)400°C for 3min,(b)400°C for 10min,(c)550°C for 3min,and(d)550°C for 10min.

3.2.Corrosion properties

Fig.4.The anodic potentiodynamic polarization tests performed on the WN43 samples sintered at 400 or 550°C with holding time of 3 or 10min.

The PDP corrosion tests were performed on all the studied materials and the selected PDP curves(Evans diagrams)for the samples sintered at 400 and 550°C are shown in Fig.4.The tests were reproducible as only negligible variations in the PDP curve shapes and positions were obtained for the samples cut from the same pellet.Therefore,only one representative curve is shown for each sample type.It can be observed that with the increasing SPS temperature the corrosion potential values,Ecorr,were gradually shifted to the less noble region.This effect was the most significant upon sintering at 500 and 550°C(see Table 1 where data for all the samples are presented).The values of polarization resistance,Rp,PDP,were evaluated directly from the Evans diagrams using Stern’s analysis.In addition,theRp,EISvalues were also determined from the EIS data by means of Nyquist plot fitting(Fig.5)using the equivalent circuit shown in the figure inset.The Nyquist plots clearly exhibited two capacitive loops,one in the high-frequency range and the other in the medium-frequency range.The partial polarization resistances,Rp1,EISandRp2,EIS,corresponding to the high-frequency and medium-frequency loops,respectively,were evaluated first.Consequently,the total polarization resistance was calculated asRp,EIS=Rp1,EIS+Rp2,EIS.It can be observed in Fig.5 and Table 1 that,irrespective of the holding time,higher sintering temperature resulted in a considerable increase in the“diameter”of both,the mid-frequency(predominantly)and the highfrequency capacitive loops.The values,Rp,PDPandRp,EIS,of all the samples calculated using both methods are presented in Table 1 and imply that the corrosion resistance was improved gradually up to the highest sintering temperature of 550°C.The values from both,PDP and EIS analyses,followed analogous qualitative trends and,quantitatively,were also rather comparable,albeit theRp,EISvalues were somewhat higher(by～10–30%)than theRp,PDPvalues mostly after sintering at higher temperatures.

Table 1The corrosion parameters determined from the PDP and EIS experiments for all the investigated samples.

Fig.5.The electrochemical impedance spectra of the WN43 samples sintered at 400 or 550°C with holding time of 3 or 10min;and the equivalent electrical circuit used for the EIS data fitting(inset).

Fig.6.The corrosion rates of the WN43 samples sintered at 400,450,500,or 550°C with holding time of 3 or 10min tested in the 0.1mol NaCl solution;and the corrosion rate of the WN43 sample sintered at 550°C for 3min tested in the biological medium(MEM+10% FBS).

The corrosion behavior of the sintered WN43 samples was further examined by means of the immersion tests and the calculation of corresponding corrosion rates.The samples were immersed into the aqueous solution of 0.1mol NaCl for 7 days.After this period,the corrosion products were removed according to the 8407:2009 ISO standard[48].Subsequently,the weight of the tested samples was measured and the respective weight losses and corrosion rates were calculated.The corrosion rates for the studied materials are presented in Fig 6.It is obvious that the corrosion rate was relatively high upon sintering at the lowest temperature of 400°C,for both holding times.Sintering at 450°C resulted in a minor increase in the corrosion rate,whereas there was a rather significant decrease in this quantity for the samples sintered at higher temperatures,500 and 550°C,again for both holding times.Even though the improvement in the corrosion rate versus SPS temperature was not utterly gradual,the immersion tests revealed superior corrosion performance upon sintering at higher temperatures,500 and 550°C,in accordance with the PDP and EIS data.

The sample with the lowest degradation rate measured in the 0.1mol NaCl solution(i.e.550°C,3min)was subsequently also tested using simulated body fluid as a medium.The test was performed in MEM+10%FBS at 37°C and controlled pH,and the resulting value of 0.030(2)mg·cm-2·h-1,which is lower by～40% than that of the sample tested in the 0.1mol NaCl solution,was obtained(it is also shown in Fig.6 for the sake of comparison).The investigation of degradation rate of the sample sintered at 550°C for 3min was also extended by an evaluation of the cytotoxicity according to the ISO 10993-5 standard.The results showed that the relative metabolic activity of the L929 murine fibroblasts incubated with 50% extracts and 100% extracts was 106(5)%and 105(4)%,respectively,and was comparable to that of the untreated cells.Note that these values are well above the normative limit of toxicity,which is 70%of the relative metabolic activity.

Fig.7.The true plastic stress-strain compression curves of the WN43 samples sintered at 400,450,500,or 550°C with holding time of(a)3min and(b)10min.

Table 2The mechanical parameters(compression yield strength-CYS,ultimate compression strength–UCS,elongation at UCS–εUCS,and elongation to fracture–εmax)determined from the compression curves and Vickers microhardness–HV0.5 of all the investigated samples.

3.3.Mechanical properties

In order to assess the mechanical performance of the studied materials,standard compression tests were performed.The true plastic stress-strain curves presented in Fig.7 reflect a substantial effect of different SPS parameters on the mechanical properties.Only one sample is shown for each condition,as the mechanical response was rather similar for the samples of each set.The sigmoidal“s-shaped”deformation curves observed for practically all the samples indicate that mechanical twinning took place during compression so as to compensate for the limited number of independent active slip systems(i.e.to fulfill the von Mises criterion)[51,52].The mechanical parameters of all the tested samples are presented in Table 2.It is evident that the average strain at ultimate compressive strength(εUCS)and strain to fracture(εmax)of the samples sintered at the lowest temperature of 400°C were relatively poor.However,increase in the sintering temperature resulted in a significant enhancement of maximum deformability,especially between the samples sintered at 450°C and 500°C asεUCSincreased from 0.05(1)to 0.16(1)or from 0.08(1)to 0.16(1)for the samples sintered for 3min and 10min,respectively andεmaxincreased from 0.06(1)to 0.23(1)or from 0.09 to 0.18 for the samples sintered for 3min and 10min,respectively.A further rise in the SPS temperature to 550°C led to a small decrease in both these quantities for both sintering times,3 or 10min.Evaluation of theεmaxparameter in addition to the standardεUCSdetermination was employed in order to better characterize the formability of the samples as certain weak softening/plateau in the flow stress was observed for the samples sintered for 3min.Similar trends as for the formability were also observed in the case of the ultimate compressive strength(UCS),which increased steadily from 202(8)or 190(7)MPa(3 or 10min at 400°C,respectively)to 331(2)or 343(2)MPa(3min or 10min at 500°C,respectively),whereas sintering at 550°C resulted in a drop in the UCS values to 317(1)or 308(8)MPa(3 or 10min,respectively).The compressive yield strength(CYS)values effectively decreased with increasing SPS temperature from 450°C to 550°C(from 189(2)to 161(2)MPa for 3min and from 180(2)to 149(1)MPa for 10min of SPS processing).The samples sintered at 400°C exhibited similar CYS values as the ones sintered at 450°C in the case of 3min SPS time or lower CYS values(by～30MPa)for 10min SPS time.

In addition,the results of Vickers microhardness testing(HV0.5)were included in Table 2.Rather constant HV0.5values of～70 were documented across all the sample sets sintered at 400,450,or 500°C,whereas the samples sintered at 550°C exhibited slightly lower HV0.5of 66(2)(holding time of 3min)or 62(3)(holding time of 10min).

4.Discussion

Apart from the chemical composition,the corrosion behavior and the severity of attack the material undergoes are controlled,in the first place,by its microstructural features such as material homogeneity,grain size[53],morphology and distribution of the secondary phases[18],or residual strain[54].The PDP tests presented in Fig.4 revealed that the SPS processing of WN43 powders progressively shifted the PDP curves to the less noble regions(i.e.theEcorrvalues decreased)as a function of increasing sintering temperature.On the other hand,the corrosion resistance,Rp,determined using both,the PDP and EIS techniques,was gradually enhanced with rising SPS temperature(Table 1).Similarly,the corrosion rates improved(decreased)with the rise in the SPS temperature,particularly between 450 and 500°C.It is worth recalling that the effect of sintering time on the corrosion parameters was rather negligible.A similar evolution of the above-mentioned corrosion parameters was also observed e.g.in the cast Mg-5Y-1.5Nd[55]and Mg-5Y-2Nd-3Sm-0.5Zr[56]alloys after homogenization heat treatment.

The occurrence of less nobleEcorrvalues with increasing SPS temperature can be likely explained in terms of the dissolution of the secondary phases and their segregation mostly at the particle boundaries(Fig.2a-d).The boundaries enriched with Nd and Y can be considered as more noble compared to the Mg matrix,thus leading to an intensification of the corrosion potential between the boundaries and the Mg matrix[57](for the exact phase composition after sintering using different SPS conditions see also[47]).On the other hand,the decrease in the corrosion kinetics with rising SPS temperature,represented by theRp,PDPandRp,EISparameters and the corrosion rate,requires a more complex clarification.First,it was shown in our recent work[47]that the magnitude of residual strain due to SPS processing of this material decreases significantly as a consequence of increasing sintering temperature due to recovery and recrystallization.Additionally,higher SPS temperature also brought about changes in the meso–and microstructure of the material.The grain size became more uniform and the dissolution,segregation at particle boundaries,and certain re-precipitation of finer secondary phase particles took place,resulting in the reduction of the amount of primary defects acting as corrosion nucleation sites(see Fig.2 and Ref.[47]).Homogeneously distributed solute RE elements or fine secondary phases rich in RE were shown to be an effective corrosion retardant in Mg[49,58].The combination of these effects consequently contributed to the overall improvement in the corrosion resistance.Furthermore,a closer examination of the EIS data presented in Fig.5 revealed that the two capacitive loops followed somewhat different trends.It was reported that the high frequency capacitive loop and respectiveRp1,EISvalues are related to the film effect of corrosion products and the charge transfer process associated primarily with the presence of cathodic secondary phase particles.On the other hand,the second(mid-frequency)capacitive loop and theRp2,EISvalues are indicative of a decrease in the diffusion rate of Mg2+ions(i.e.a reduction in the mass transport)through the growing protective passivation layer on the sample surface,thus making it more compact[59,60].In this context,a more pronounced increase in theRp2,EISvalues compared to theRp1,EISvalues as a function of increasing sintering temperatures(Fig.5 and Table 1)implies that the dissolution of initial secondary phases,consequent microstructure homogenization and fine particle precipitation,which enhanced surface passivation with rising SPS temperature,play a dominant role in the improvement of the overall corrosion resistanceRp,EISand,similarly,Rp,PDP.It should be also noted that an additional low frequency inductive loop arising due to the relaxation processes of adsorbed species on the electrode surface[60]was not observed in this study.

The longer-term corrosion behavior of the compacted WN43 samples was also examined by means of corrosion rate measurements.The samples were immersed into the 0.1mol NaCl aqueous solution for 7 days and the corrosion rates were evaluated(Fig.6).The general trend of reduced severity of the corrosion attack with increased SPS temperature was observed also in the immersion tests data,yet the experimental errors were more pronounced due to the more stochastic nature of such tests.The greatest two-to three-fold drop in the corrosion rates was observed when the sintering temperature was increased from 450 to 500°C for both holding times,3 and 10min.Considerable weight loss and related high corrosion rate observed in the samples sintered at 400 and 450°C can be,again,ascribed to a number of factors such as higher porosity(Fig.2a-d),heterogeneous distribution of secondary phases,and non-recrystallized microstructure[47],which effectively hindered the formation of the stable passivation layer.To validate this explanation,Fig.8a-d show the SEM micrographs of the transversal cross-section of the sample sintered at 400°C and 550°C for 10min after 7 days of immersion.Note that the magnification in Fig.8a,c is lower(800×)than in Fig.8b,d(1600×).The darkest areas correspond to the epoxy resin used for the sample fixation during its preparation for SEM,the dark gray layers account for the corrosion layers on the sample surface or its near interior,and the light gray zones correspond to the sample.It is evident from these micrographs that the severity of the corrosion attack was greater for the samples sintered at 400°C(Fig.8a,b),compared to sintering at 550°C(Fig.8c,d).It is worth recalling that sintering at lower temperatures only led to a partial dissolution of secondary phases and,consequently,the depleted magnesium matrix was in close contact with a relatively large amount of big particles(mostly Mg41Nd5[47]).Therefore,the galvanic corrosion at these spots was active also inside the former powder particles.In addition,compactness of the samples sintered at lower temperatures was lower(i.e.the porosity was higher),leading to a larger specific surface susceptible to corrosion.Fig.8a,b clearly show the propagation of corrosion attack along the porous grain boundaries,thus forming an unstable passivation layer and,in turn,effectively penetrating and deteriorating the sample.

Conversely,the zoomed-in cross-section of the corrosion layer of the sample sintered at 550°C(Fig.8d)indicates preferred corrosion in the vicinity of the former particle boundaries,i.e.in the regions which were substantially depleted of alloying elements and are in close contact with significantly enriched former boundaries(particularly by Nd,Y,and O,as was observed using the energy dispersive X-ray(EDX)analysis in our recent study[47]).This effect can be explained by the fact that the RE-rich boundaries are nobler and strongly cathodic relative to the grain interiors and,especially,the zones depleted of RE elements.At the same time,it is also evident that the well-compacted particle boundaries themselves effectively hindered the propagation of the corrosion fronts.Similarly,dissolved Nd and Y and partial reprecipitation of fine secondary phases in the interior parts of the former powder particles resulted in higher corrosion resistance within these areas and,therefore,decreased the rate of propagation of the corrosion attack.On the other hand,these local observations seem to flatten out macroscopically,leading to relatively uniform corrosion over the sample surface(Fig.8c).The other beneficial factor of a more uniform distribution(especially of Y)is its ability to increase the protectiveness of the porous Mg(OH)2corrosion layer.Yttrium oxides tend to incorporate into the pores and make the layer stiffer and more resistant to charge and mass transfer[61].This effect is especially visible in the results acquired by EIS,where both theRp1andRp2values grew with increasing sintering temperature(Table 1).Naturally,the particle boundaries rich in Nd,Y and O were significantly less prone to degradation than the Mg-matrix,as already mentioned above.According to EDX and XRD analyses in our previous work[47],the as-received WN43 powder particles already contained the Y2O3layer on the surface and subsequent exposure of the powder to elevated temperatures during SPS processing resulted in further oxidation and the content of the Y2O3phase increased in all the samples roughly from 1.5 to 2.0%[47].Similarly,the corrosion dynamics in the WE43 magnesium alloy prepared by SPS was shown to be related to the phase segregation and formation of the stable layer of Y2O3at the particle boundaries[62].As a result,enriched and more compact particle boundaries formed an effective barrier against the corrosion attack and were responsible for the additional reduction in the corrosion rate of the material.

Fig.8.The SEM micrographs of the transversal cross-section(two different magnifications)of the samples sintered for 10min after 7 days of immersion in the 0.1mol NaCl aqueous solution at 400°C:(a)800×magnification,(b)1600×magnification,and at 550°C:(c)800×magnification,(d)1600×magnification.

Taken together,the overall decrease in the corrosion rate with increasing SPS temperature is an exceptionally complex process featuring a great number of variables.The corrosion resistance of RE-containing Mg alloys is controlled by two competing effects,namely the cathodic activity of the intermetallic particles containing RE and the formation of protective surface films[61].As confirmed above,the segregation of the secondary phase particles causes a localized corrosion attack due to the micro-galvanic effect.At the same time,the Yand Nd-rich boundaries effectively hindered the propagation of corrosion fronts and the more uniform distribution of Y enhanced the protectiveness of the corrosion layer.Moreover,it was shown that enhanced thermodynamical stability(e.g.reduction in the residual strain,segregation of the secondary phases at the boundaries,and re-precipitation of fine particles in the matrix)with rising SPS temperature have a beneficial effect on the corrosion resistance[63].Consequently,it was shown that the positive effects related to the presence of RE elements in the matrix counterbalanced the detrimental effect of galvanic coupling in the studied material sintered at higher temperatures.Furthermore,the degradation rate measured in the MEM+10% FBS solution was significantly lower than the one measured in the NaCl solution.It was repeatedly shown that the main reason is the corrosion inhibiting effect of calcium phosphates forming on magnesium hydroxide and the presence of FBS,regardless of higher testing temperature[34,64].The parallel measurement of the material cytotoxicity showed that the samples were clearly cytocompatible(i.e.the metabolic activity of the cells was way above the toxicity limit)in the elution tests with extracts.

The mechanical properties of the sintered WN43 samples assessed by means of conventional compression tests at room temperature showed a significant dependence on the sintering parameters(primarily on the sintering temperature).A gradual drop in the CYS values with increasing SPS temperature was observed also in other related studies on SPS processing of AZ91[63,65],AE42[39],and WZ21[40]magnesium alloys.The samples sintered at 400°C in the present work,however,did not follow this trend and exhibited similar(for 3min holding time)or lower(for 10min holding time)CYS than the ones sintered at 450°C.At this condition,the samples were still rather brittle and less homogeneous,as also indicated by very limited deformability and scatter in the experimental data(see Table 2).Such behavior is most likely associated with residual porosity in these samples(around 0.5%[47])which demonstrates imperfect compaction,as also seen in Fig.2a,b.

Likewise,this fact is reflected in the UCS values,which are the lowest for the samples processed at 400°C,for both holding times.Diffusion-assisted compaction which was more intensive at higher temperatures,450 and 500°C,led to a progressive improvement in UCS,whereas a decrease in UCS was evidenced for the samples sintered at 550°C for both,3 and 10min.A comparable non-monotonous concave dependence of UCS on the sintering temperature in Mg alloys processed by SPS was documented also by other authors[40,65].The flow stress and UCS of a material are governed by various effects,such as material compactness,grain size(through the Hall-Patch relation),solid solution strengthening,precipitation strengthening,and Taylor strengthening related to the dislocation density(associated with the residual strain).Firstly,the increase in these quantities up to an SPS temperature of 500°C in the studied material is assumed to be primarily related to decreasing porosity and enhanced rigidity of the samples.Furthermore,after sintering at these temperatures,the eutectic precipitates and solute alloying elements identified in the initial powders partially re-precipitated in the Mg matrix[47]and,therefore,contributed to the material strength.On the other hand,as the highest sintering temperature of 550°C exceeded the solvus temperature of secondary phases,the former precipitates were effectively dissolved,the Mg matrix was purified,and segregation of the alloying elements and reprecipitation occurred along the particle boundaries(Fig.2c,d).Furthermore,extensive recovery and recrystallization(i.e.formation of new strain-free grains)took place at this temperature(Fig.2c,d and Fig.3c,d),which both alleviated the overall material dislocation density.The contribution of these effects clearly overwhelmed the effect of better compaction,which was most likely saturated already after sintering at 500°C,irrespective of the sintering time.Moreover,this explanation also elucidates the observation of a drop in the(otherwise rather constant)microhardness values after sintering at the highest sintering temperature.It should be mentioned that the effect of grain size is believed to be negligible as it was shown to be rather constant across the samples studied in this work[47].Little variations in the grain size of Mg alloys as a function of SPS parameters were also reported elsewhere[39,66].

Inferior compressive performance of the samples sintered at temperatures below 500°C is also demonstrated in Fig.9 showingpostmortemsurfaces of the samples sintered for 10min at 400°C(Fig.9a)and 550°C(Fig.9b).Apart from one larger crack,the sample sintered at 400°C underwent significant crushing manifested by small disintegrated fragments of the material.On the other hand,the fracture of the sample sintered at 550°C took place by means of a single well-defined crack and the two resultant fragments remained fully compact.

Relatively high deformability of the samples sintered at higher temperatures observed in this work can be primarily attributed to a very low porosity of the studied samples when compared with other works on Mg-based materials prepared by SPS[39,67].It should be noted that the strain to fracture values(εmax)were evaluated along with the standard strain at UCS values(εUCS).This approach was employed to characterize the deformability of the samples in more detail since the shape of the deformation curves of some of the samples was somewhat atypical.Especially,the samples sintered at 500 and 550°C for 3min reached UCS relatively early(ε=～15%)and they deformed steadily for several more percent,exhibiting certain strain softening(i.e.decrease in the flow stress).This behavior is typically observed in magnesium alloys only at elevated temperatures[68]and its occurrence in the WN43 samples prepared by SPS is not clear at the moment.Theoretical modeling based on the Considère criterion modified for compression(e.g.[69])or digital image correlation technique could be used in the future to elucidate this observation.Such analyses are,however,out of the scope of the present work.

Fig.9.The deformed surfaces after the failure of the compressed WN43 samples sintered for 10min at(a)400°C and(b)550°C.Loading direction was parallel with the horizontal axis of the images.

As there is a lack of studies regarding Mg-Y-Nd alloys,the obtained mechanical results cannot be critically assessed.On the other hand,there are certain works focusing on the compositionally similar Mg-Y-RE(WE43)alloy processed by SPS.Apart from the observed trends,the quantitative comparison of the mechanical parameters of the best-performance samples sintered at 500°C(listed in Table 1)revealed very similar results as for the WE43 alloy[66,70].However,such a comparison must be taken with caution as the effect of different SPS device and slightly different conditions cannot be completely quantified.Nevertheless,these results argue for a robust comparability of characteristics of the materials prepared using SPS and also confirm that the substitution of RE with pure Nd in these alloys is entirely viable.Furthermore,the obtained results are also in the range or even exceed the mechanical performance of the WE43 alloy processed by conventional methods(CYS of 150–200MPa,UCS of 300–400MPa)[71–73].Likewise,the microhardness(HV0.5)values of～70(Table 2)are in the upper range of values reported in the literature for conventional(HV of 40–95)[2,71,74]or SPS-ed(HV of 35–85)[66,70]WE43 samples.

The supreme combined performance in terms of corrosion as well as mechanical properties was attained in the samples sintered at 500 and 550°C(whereas preferred holding time,3 or 10min,could not be determined downrightly).Especially,the strain to fracture of～20% is seldom reported for the Mgbased materials prepared via powder metallurgy[65].On the other hand,biomedical applications are typically case-specific and,for this reason,suitable materials need to be capable of fine-tuning.It should be noted that besides sintering time and temperature,additional means of microstructure modification of the SPS products have been also elaborated.Among them,the greatest attention has been paid to the milling of the powder particles before the SPS processing.In this way,superior mechanical performance can be achieved due to microstructure refinement and disruption of oxide films on the particle surfaces.Similarly,an improved corrosion performance with decreasing grain size can be also expected in the REcontaining Mg alloys[34].Secondly,the effect of sintering pressure was examined and a higher pressure during SPS proved to be beneficial for the Mg-6Al powder compaction and subsequent mechanical performance[75].A great variety of parameters resulting in different corrosion and mechanical parameters,therefore,render the SPS method a feasible processing technique for Mg-based bioimplants and the novel WN43 alloy investigated in this work appears to be one of the prominent candidates.

5.Conclusions

In this work,a novel WN43 alloy was prepared by the SPS technique,featuring a well-defined composition in comparison with the commercially available WE43 alloys(due to the presence of RE mischmetal).Moreover,Nd used in this study instead of the RE mischmetal is assumed to exhibit better biocompatibility.The atomized WN43 powders were compacted by SPS at temperatures of 400°C–550°C and holding times of 3min or 10min and a relative density very close to 100% was achieved.The SPS temperature significantly affected the corrosion and mechanical properties of final materials whereas the effect of holding time was only negligible.These properties and related implications can be summarized as follows:

?The corrosion potential,Ecorr,gradually decreased with increasing SPS temperature due to the dissolution of the secondary phases and the segregation of Nd and Y at the particle boundaries,leading to the intensification of the corrosion potential between the boundaries and the Mg matrix.

?The polarization resistance,Rp,determined using both,the PDP and EIS techniques,in the 0.1mol NaCl aqueous solution exhibited an increasing tendency with rising SPS temperature.Similarly,the corrosion rates(determined by the immersion tests for 7 days)decreased with higher SPS temperature.These observations were related to the combined effect of decreasing residual strain(due to recovery and recrystallization)and microstructure homogenization with increasing SPS temperature.

?The degradation tests in the biological medium(MEM+10% FBS)yielded even lower corrosion rates than the ones performed in the 0.1mol NaCl solution.Moreover,the cytotoxicity tests showed that the relative metabolic activity of cells was well above the toxicity limit,thus confirming good biocompatibility of the studied material.

?The overall mechanical performance was the most favorable upon SPS processing at 500°C as a result of sufficient compaction and only partially recovered/recrystallized microstructure.Specifically,the maximum deformability(up to～20%)and UCS(～340MPa)were achieved with the SPS temperature of 500°C.The CYS values exhibited their maxima(180MPa–190MPa)after sintering at 450°C and slightly decreased with rising SPS temperature.

?In general,the mechanical performance of the prepared materials was exceptional within the group of Mg alloys prepared by powder metallurgy(particularly due to outstanding deformability)and similar or better than in the case of Mg-RE alloys processed by conventional techniques.Additionally,this confirms that the substitution of RE with pure Nd is fully feasible.

?Accordingly,it was shown that a superior combination of corrosion and mechanical properties was attained after sintering at higher temperatures of 500 or 550°C.As mentioned,the effect of sintering time was insignificant and sufficient compaction was achieved already with 3min holding time,which can prospectively lead to a reduction in the production costs.

?Manifestly,a wide range of corrosion and mechanical characteristics can be obtained by altering the SPS parameters in order to achieve case-specific requirements typical of biomedical materials.Biocompatibility and favorable properties of the novel WN43 alloy together with the ability to fine-tune its performance by varying the SPS parameters,therefore,make it a particularly promising material for biomedical use.

Acknowledgments

This work was financially supported by the Czech Science Foundation under the project GA18–19213Y.Partial financial support by ERDF,project No.CZ.02.1.01/0.0/0.0/15 003/0000485 is also acknowledged.M.Z.gratefully acknowledges additional financial support by the Charles University Grant Agency under the grant 410119.M.K.acknowledges partial financial support from OP RDE,MEYS,grant No.CZ.02.1.01/0.0/0.0/16 013/0001794.