Degradable magnesium-hydroxyapatite interpenetrating phase composites processed by current assisted metal infiltratio in additive-manufactured porous preforms

Mrino Css-Lun,Er B.Montur,Norbrt Hort,Sbstin Díz--l-Torr,JoséCluio Ménz-Grcí,Luci Vi?tjnová,Am Brínk,Al?Dˇnhl,Krl Dvoˇrk,Joz Kisr,LislvˇClko

a Central European Institute of Technology,Brno University of Technology,Purkyˇnova 123,Brno 612 00,Czech Republic

b Institute of Metallic Biomaterials,Helmholtz-Zentrum Hereon,Max-Planck-Stra?e 1,Geesthacht 21502,Germany

c Institute of Product and Process Innovation,Leuphana University Lüneburg,Universit?tsallee 1,Lüneburg 21335,Germany

d Centro de Investigación e Innovación Tecnológica,Instituto Politécnico Nacional,Cerrada de Cecati s/n,Mexico City 02250,Mexico

e Biomedical Center,Faculty of Medicine in Pilsen,Charles University,Alej Svobody 1655/76,Plzeˇn 323 00,Czech Republic

fInstitute of Biophysics of the Czech Academy of Sciences,Královopolská135,Brno 612 65,Czech Republic

g Faculty of Civil Engineering,Brno University of Technology,Purkyˇnova 139,Brno 612 00,Czech Republic

Abstract This work explores ceramic additive manufacturing in combination with liquid metal infiltratio for the production of degradable interpenetrating phase magnesium/hydroxyapatite(Mg/HA)composites.Material extrusion additive manufacturing was used to produce stoichiometric,and calcium deficien HA preforms with a well-controlled open pore network,allowing the customization of the topological relationship of the composite.Pure Mg and two different Mg alloys were used to infiltrat the preforms by means of an advanced liquid infiltratio method inspired by spark plasma sintering,using a novel die design to avoid the structural collapse of the preform.Complete infiltratio was achieved in 8 min,including the time for the Mg melting.The short processing time enabled to restrict the decomposition of HA due to the reducing capacity of liquid Mg.The pure Mg-base composites showed compressive yield strength above pure Mg in cast state.Mg alloy-based composites did not show higher strength than the bare alloys due to grain coarsening,but showed similar mechanical properties than other Mg/HA composites that have significantl higher fraction of metallic phase.The composites showed faster degradation rate under simulated body conditions than the bare metallic component due to the formation of galvanic pairs at microstructural level.Mg dissolved preferentially over HA leaving behind a scaffold after a prolonged degradation period.In turn,the fast production of soluble degradation products caused cell metabolic changes after 24 h of culture with not-diluted material extracts.The topological optimization and reduction of the degradation rate are the topics for future research.? 2022 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:Interpenetrating phase composite;Biodegradable metal;Topological relationship;Direct ink writing;Metal infiltration Computed aided design.

1.Introduction

A wide variety of materials have been used for bone repair.Metallic devices are usually implanted for load-bearing bone fracture fixatio[1,2],whereas calcium phosphate(CaP)ceramics are mostly used as osteoconductive and degradable synthetic bone grafts to guide bone regeneration in non-load-bearing applications[3].Composites consisting of a biodegradable metallic phase and CaP ceramics represent a promising solution to combine in one material the osteoconductive,mechanical and degradation properties valuable for bone repair.Magnesium(Mg)is attractive as metallic phase because it is a biodegradable metal well tolerated in the human body,it has osteogenic effect and mechanical properties similar to the human bone[4-7].Different methods have been developed to produce Mg/CaP composites;most of them are based on traditional powder metallurgy routes to obtain particle-reinforced composites,with either Mg or CaP as matrix having discrete,dispersed,and isolated particles homogenously embedded[8-12].Another approach explored recently is the fabrication of interpenetrating Mg/CaP composites by liquid Mg infiltratio in a CaP foam[13-19].

Interpenetrating phase composites are high-performance materials,in which both the matrix and the reinforcement form a continuous three-dimensional(3D)percolating network through the structure[20].The two phases provide structural stability to the material,while one of them the required multifunctionality,in the case of CaPs represented by the osteoconductivity and promotion of new bone formation.The mechanical properties of the interpenetrating composites depend on multiple parameters including the mechanical behaviour,fraction,geometry and distribution of the continuous components,as well as the interfacial bonding strength and processing conditions[21-24].Thereby,the macroscopic properties can be tuned by selecting the appropriate components and designing their 3D topological distribution inside the manufactured structure.Such way that the ability of fabricating on-demand 3D topologies raises new horizons for the development of novel load-bearing,osteoconductive and biodegradable interpenetrating Mg/CaP composites.

Ceramic additive manufacturing provides the means for the computer-aided design and fabrication of CaP preforms with complex porous structures that satisfy the topological requirements,whereas liquid Mg infiltratio is the most straight forward method for the fabrication of the interpenetrating composite[25-27].Although multimaterial additive manufacturing has facilitated the fabrication of interpenetrating polymer-polymer composites[23,28-32],only few works have explored the combination of additive manufacturing with infiltratio for the processing of interpenetrating metallic/metallic and ceramic/metallic composites[33-36],none of them for biomedical applications neither degradable.Therefore,this novel processing approach represents a significan step forward in the development of interpenetrating Mg/CaP composites for biomedical applications,beyond the infiltratio of CaP foams with fi ed pore network obtained by replication of polymeric templates.

As may not be obvious,the infiltratio of liquid Mg in CaP preforms represents a technical challenge,due to the high reactivity of Mg,resulting in Mg oxidation and CaP degradation[12,37,38].Moreover,the liquid metal infiltratio is not a spontaneous process,requiring the degassing of the ceramic preform or the application of an external pressure to overcome the surface tension for the liquid metal intrusion in small pores.Therefore,the mechanical strength of the preform is a limitation factor for infiltratio[26].In order to surpass these two drawbacks,this work uses a new fast and low-pressure liquid metal infiltratio method inspired by the electric fiel assisted sintering technology,also known as spark plasma sintering(SPS).In general,the SPS technique is practiced under high vacuum conditions,where a pulsed electric current is supplied directly on a compacted powder to heat it up to the sintering temperature in a few minutes.Within the current assisted metal infiltratio(CAMI)method,the high vacuum state protects Mg from oxidation and degasses the preform for easier infiltration First,Mg is melted at a fast-heating rate,using the principle of SPS and rapidly solidifie when the current is cut off.Both rapid heating and cooling rates,along with a short dwelling time,minimize the time for chemical reaction between liquid Mg and CaP preform.The infiltratio of Mg in the liquid state is driven by the effect of gravitational force.Furthermore,CAMI may apply a gentle mechanical force through the SPS electrode to overcome the surface tension for the infiltration without applying a compaction force directly on the ceramic preform.Therefore,CAMI represents an innovation respect to a previous work,where mechanically tough Ti2AlC foams were sandwiched between the SPS electrodes for aluminium alloy infiltratio[39].

The present work uses CAMI for the fabrication of degradable interpenetrating Mg/hydroxyapatite composites,in combination with the robocasting technique for the topological design and fabrication of two types of hydroxyapatite(HA)porous preforms.Robocasting is a cost-effective material extrusion additive manufacturing technology in which 3D structures are built layer by layer with high fidelit by the automatic deposition of filament made of high solid content ceramic pastes[40,41].A key step in this technique is the conversion of the deposited powder into a dense ceramic,HA normally requires sintering at high temperature[42],although the cementitious reaction at room temperature is also possible[43].The effects of the microstructural features of HA after sintering and cement setting on the subsequent infiltra tion performance were explored.Two monophasic Mg alloys with low content of either calcium(Ca;0.2 wt.%)or zinc(Zn;1 wt.%)were developed in order to decrease the degradation rate in comparison to pure Mg.Both the chemical and structural characterizations of the interpenetrating composites were performed,and their degradation rate and cytotoxicity were studiedin vitroas an initial attempt to introduce these novel composites as possible degradable biomaterials for orthopaedic applications.Interfacial reaction,mechanical strengthening and degradation mechanisms were identifie as fundamental bases for future improvements.

2.Materials and methods

2.1.Casting of magnesium alloys

Ultra-high purity Mg was used for the production of two Mg alloys containing low amount of Ca or Zn by permanent mould direct chill casting[44],i.e.,0.2 wt.% Ca and 1 wt.%Zn.Firstly,blocks of pure Mg were melted at 670 °C using an electric resistance furnace under a protective atmosphere of Ar with 2 vol.% of SF6.After melting,pre-weighted pellets of Ca(99%,Alfa Aesar)or Zn(99.9%,Alfa Aesar)were added when temperature reached the 750 °C and after 5 min of manual mixing the 1.0044-steel crucible was taken out from the furnace for cooling by immersion in water at roomtemperature.The obtained ingots were machined into cylinders of 100 mm in dimeter and 150 mm in length in order to extrude them at 300 °C in one step to obtain bars of 10 mm in diameter using a ram speed of 0.6 mm·s-1.The chemical composition of the extruded alloys was determined by X-ray fluorescenc(Spectrolab M9,Spectro Analytical GmbH,Germany)and in the case of the calcium content,it was measured by flam atomic absorption spectroscopy(AAS,240 FS,Agilent,USA).

2.2.Synthesis and robocasting of hydroxyapatite preforms

Two different powders were used to fabricate the ceramic preforms for metal infiltration First,stoichiometric HA preforms were fabricated using a HA powder synthesized by chemical precipitation between phosphoric acid(H3PO485%,Lachner,Czechia)and calcium hydroxide(Ca(OH)299%,Merck,Germany)at room temperature[3].Second,calcium deficien HA(CDHA)preforms were obtained by the hydrolysis of preforms made of alpha tricalcium phosphate(α-TCP)powder.Theα-TCP was synthetized at 1400 °C by solid-state reaction from 2:1 molar mixture of CaCO3(≥99%,Sigma Aldrich,Germany)and anhydrous CaHPO4(＞98.0%,Sigma Aldrich,USA)followed by air quenching.Both,HA andα-TCP were milled for 15 min at 450 rpm in a planetary ball mill(Fritsch Pulverisette 6)using agate jar and balls(25 balls of 20 mm in diameter)and later sieved below 45μm particle size for the fabrication of the preforms by robocasting.

The pastes for additive manufacturing by robocasting technique were prepared by mixing the powder with an aqueous solution containing 40 wt.% of Pluronic F-127(Sigma Aldrich,Germany)at a liquid-to-powder ratio of 0.6 g·g-1.A dual asymmetric centrifugal mixer(DAC 150,Speedmixer,USA)was used for homogenization during 60 s at 3500 rpm.Afterwards,the pastes were placed into a 3 mL cartridge set of a robotic deposition device(Pastecaster,FundacióCIM,Spain)and the preforms were robocast in air using 410μm tapered dispensing tips(SmoothFlow Tapered Tips,Nordson EFD,USA)at a velocity of 8 mm·s-1.Cylindrical preforms were built up following an orthogonal grid pattern of parallel filament with an in-layer separation between them of 500μm and an overlapping between layers of 5%.The HA preforms,having 10 mm in diameter and 11 filament per layer,were left to dry overnight in air at room temperature and later sintered at 1200°C for 5 h in a furnace(LHS 08/15,LAC,Czech Republic)with a heating ramp of 2.5 °C·min-1.In contrast,theα-TCP preforms,having 8 mm in diameter and 9 filament per layer,were stored at 37 °C in 100%humidity atmosphere for 24 h and afterwards immersed in distilled water at 37 °C for 6 days to complete the hydrolysis ofα-TCP into CDHA according to the following chemical reaction[43].

Scheme 1.Design of graphite-die and arrangement for the current-assisted metal infiltratio technique.

The diameter of the preforms after sintering or hydrolysis was measured with a calliper to calculate the linear shrinkage.The characterization of the preforms was conducted according to Sections 2.4 and 2.5.

2.3.Current assisted metal infilt ation

The interpenetrating phase Mg/CaP composites were obtained by the new CAMI method,conducted inside a spark plasma sintering apparatus(SPS,Dr.Sinter 1050,Japan).The method used a graphite die with two concentric sections(Scheme 1).The upper section had an internal diameter of 10 mm and height of 18 mm,whereas the bottom section had an inner diameter of 9 mm and height of 22 mm.The connection between the two sections had a 90° diameter reduction.The preform was placed in the bottom section meanwhile the metal(a disc of 10 mm in diameter and 10 mm in height)was in the upper section.The bottom of the die was closed with a T-shape graphite lid of 9 mm in diameter.In contrast,the upper part was closed with a mobile graphite piston that gently pushed the molten metal into the bottom chamber containing the porous preform.A zirconium oxide disc(10 mm diameter and 2 mm thick)and graphite paper were placed between the metal and the piston to promote the heating of the metal and prevent the backfl w of the molten metal towards the piston.

The assembled die was set inside the SPS apparatus,and the air pressure was reduced dawn to 8 Pa.The CAMI process consisted in increasing the temperature at a heating ramp of 100 °C·min-1until 670 °C,followed by a dwell time of 2 min.Heating was achieved by applying a direct electrical current in on-off cycles of 12 and 2 ms,respectively.The temperature was controlled automatically by the SPS system with a type-K thermocouple with the hot joint placed in the centre of the wall of the die,half the height of the preform.A constant load of 1 kN(3.2 MPa)was applied during the entire process to close the electric circuit and gently promote the liquid metal infiltration After the infiltration the samples were cooled down inside the SPS apparatus below 75 °C before increasing the air pressure.Finally,the samples were machined to cylinders with a diameter of 8 mm and length of 12 mm,which exposed the edges of the preform embedded in the metal.All the samples were cleaned before characterization in an ultrasonic bath in acetone and isopropanol series for 5 min each and dried with nitrogen at room temperature.

Table 1Mg-based biodegradable materials tested in this study.

Different interpenetrating phase composites were produced by the CAMI under the same processing conditions using different combinations of pure Mg,Mg-0.2 wt.% Ca alloy and Mg-1 wt.% Zn alloy with either HA or CDHA preforms.In addition to those composites,the bare metals were characterized according to Sections 2.4 and 2.5.Table 1 summarizes the different materials studied and provides their nomenclature.

2.4.Microstructural characterization

The cross section of the composites and metals was prepared for metallographic observation using the standard methodology for Mg and its alloys.A 0.25μm particle size diamond suspension was used for the last polishing step.Xray diffraction(XRD,Rigaku SmartlLab 3 kW,Japan)was used for the crystalline phase composition identificatio using the HighScore+software(Panalytical,The Netherlands)and ICDD PDF 2 and ICSD 2012 databases.Scans were performed in a Bragg-Brentano geometry between 10° and 90° with a scan speed of 4° min-1with a Cu-Kαradiation(λ=0.154 nm)and a current of 30 mA and 40 kV.Microstructures were revealed with picral solution(1.5 g of picric acid,25 mL of ethanol,5 mL of acetic acid and 10 mL of distilled water)and observed using an optical microscope(Olympus 8500)and scanning electron microscope(SEM;TESCAN Lyra3,Czech Republic)equipped with an energy dispersive X-ray(EDX)spectroscope(INCA,Oxford Instruments,UK).Representative images were obtained using an electron beam voltage of 5 kV.EDX elemental analysis was performed in the different phases identifie to determine the distribution of the elements and track possible atomic diffusion or reactions between phases.The preforms and the composites were in addition coated with a carbon nanometric layer to prevent charging during the SEM observations.SEM-image analysis was used to measure the diameter and separation of the filament in the preforms.In addition,the open-pore size distribution of the preforms was determined in the range from 9 to 150μm by mercury intrusion porosimetry(MIP;Thermofinni an 140/240),and the specifi surface area was determined by N2adsorption according to the Brunauer-Emmett-Teller method(NOVA 3200e,Quantachrome Instruments).

The 3D internal structure of the interpenetrating phase composites was analyzed by X-ray micro computed tomography(μCT;GE Phoenix v|tome|x L240 system).The scans were performed with a Nano focus X-ray tube of 180 kV/15W and a high contrast fla panel DXR250 with 2048×2048 pixel2.The exposure time was 334 ms in every position using an acceleration voltage of 80 kV and an X-ray current of 170μA.A 2 mm aluminium plate filtere the beam.The isotropic linear voxel size(voxel resolution)of obtained volume was 25μm.The tomographic reconstruction was performed using the GE phoenix datos|x 2.0 software(GE Sensing & Inspection Technologies GmbH,Germany)to correct the sample drift and the beam hardening.The visualization of the samples and structural volume calculations were executed in the VG Studio MAX 3.1 software(Volume Graphics GmbH,Germany).Two samples of each interpenetrating composite were analyzed for comparison.Firstly,a region of interest fittin the edges of the sample was segmented based on porosity(air)by the multi-thresholding Otsu method[45].After that,the random walker algorithm was used to segment the metallic and ceramic parts labelling a pixel as an object or background[46].The segmented binary images were saved individually and processed in the VG Studio MAX software to calculate the volume of each phase in the composites,perform porosity analysis and create 3D images.The CAMI efficien y was analyzed by the remnant porosity in the composites and expressed as infiltratio percentage following Eq.(1).

where,VMetis the volume of metal infiltrate in the composite and VAiris the volume of pores in the composite after the infiltration

2.5.Mechanical characterization

Unconfine uniaxial compression test was performed at room temperature using a universal testing machine(Zwick/Roell Z010,Germany)at a strain rate of 0.5 mm·min-1.The specimens were machined to have a cylindrical geometry with 8 mm in diameter and 12 mm in length to meet requirements in the E9/89a ASTM standard for ductile materials[47].The engineering stress was calculated by dividing the load supported by the sample per the crosssection area.Compressive yield strength was determined at 0.2% of deformation,compressive strength was stablished as the maximum stress supported by the samples before failure.At least four samples were used for each condition and the average values are reported with standard deviations.The morphology of the fracture surfaces was examined by means of SEM.

2.6.Degradation test under simulated physiological conditions

The corrosion rate of magnesium was monitored by hydrogen(H2)evolution at a constant temperature of 37 °C.A set of three discs(8 mm in diameter and 3 mm in height with surfaces ground up to a 1200-grit SiC paper)per material were immersed individually in 500 mL of simulated body flui(SBF)solution,conventional SBF in[48].The volume of produced H2by the anodic dissolution of Mg was monitored for 14 days and the average corrosion rate was calculated following Eq.(2)[49].

where,CRHis the instantaneous corrosion rate from the hydrogen evolution test,expressed in mm·y-1,and VHis the volume of hydrogen(mL)produced per unit of surface area(cm2)exposed to SBF solution per time of immersion(days).

Potentiodynamic polarization test was conducted in SBF at room temperature.A three-electrode configuratio was used in which the sample was the working electrode,Ag/AgCl/3M KCl electrode was the reference electrode,and a glassy carbon rod was the counter electrode(both Metrohm-Autolab).The exposed sample surface area was 0.2 cm2and the volume of SBF solution was 5 mL.The polarization scan performed with a Potentiostat/Galvanostat(PGSTAT128N,Autolab)started immediately after fillin the electrode chamber with SBF from-2.0 to-1.0 V vs.Ag/AgCl/3M KCl at a scan rate of 5 mV·s-1.The corrosion current density(jcorr)was estimated applying the Tafel slope analysis according to G102-89 ASTM standard[50].A set of at least 4 measurements were performed per sample composition for calculation of mean values and standard deviation.

2.7.Cytotoxicity-cell culture test

Samples(8 mm in diameter and 3 mm in height)were desinfected by immersion during 30 min in ethanol(70%in water)and rinsed three times with phosphate buffer solution(PBS).Immediately,each sample was immersed in 1 mL of complete cell culture medium at 37 °C for 5 days and the extract(supernatant containing the soluble degradation products)was collected to treat the cells.The SaOS-2 osteogenic cell line(ATCC,Manassas,USA)was used for this study.The cells were maintained in Dulbecco’s Modifie Eagle’s medium(DMEM)supplemented with 10% fetal bovine serum,2 mM L-glutamine,50 U·mL-1penicillin and 50μg·mL-1streptomycin(all from Merck,Germany).The cells were seeded in 96-well cell culture plates at a density of 8×103cells per well.The cells were left to attach during 24 h in 5% CO2humid atmosphere at 37 °C.Afterwards,the cells were rinsed twice with PBS to remove the non-attached cells and 150μL of material extract were added per well(8 wells per group and experimental time for statistical analysis).Note that the extracts were not diluted with fresh cell culture medium and were not renewed during the experiment.Cells treated with fresh medium were used as control group.After 1,3 and 6 days of culture the cells were rinsed twice with PBS and the cell metabolic activity was evaluated using a Resazurin assay kit alamarBlueTMCell Viability Reagent(ThermoFisher Scientific Waltham,USA).Briefl,150μL of cell culture medium containing 10% of alamarBlueTMCell Viability Reagent were added to the cells and incubated for 2 h at 37 °C in 5% CO2humidifie atmosphere.Afterwards,100μL of culture medium were transferred to a black 96-well plate and Resorufi fluorescenc(excitation at 530 nm,emission at 590 nm)was measured using a microplate reader(BioTek,USA).After removing the alamarBlueTMCell Viability Reagent,the cells were rinsed twice with PBS and incubated for 10 min at 37 °C with 150μL of culture medium containing 1000x diluted Hoechst 34222(ThermoFisher Scientific Waltham,USA).The cells were rinsed twice with medium and imaged with a fluorescen microscope IX83 using DAPI filte(Olympus,Japan).The number of cells(nuclei)was count using FIJI-ImageJ software(version 1.53i,NIH,USA).The results of cell metabolic activity and cell number were expressed as relative fold change compared to the control group at the same time point.

2.8.Statistical analysis

The phase fraction volume,mechanical and corrosion properties were analyzed using one-way ANOVA and t-test analyses of variance to determine statistical differences between groups.Significanc was set atp＜0.05.All data are expressed as mean±standard deviation(SD).

Data of metabolic activity and cell proliferation were analyzed for normality using Shapiro-Wilk test.The difference from 100% of untreated control was tested by Student T-test and the difference between the groups of samples was tested by Mann-Whitney U test.Statistica software(TIBCO Software Inc.,Palo Alto,California,USA)was used.

3.Results

3.1.Chemical and microstructural characterization of CaP preforms

Fig.1a and c show the orthogonal pattern of the CaP preforms,where the interfilamen space formed an open network of pores.The diameter of the preforms(7.9±0.3 mm),the diameter of the filament(422±22μm)and the inlayer separation between them(483±36μm)were in accordance with the digital design of the CDHA preform,showing minimal shrinkage duringα-TCP hydrolysis.In contrast,sintering resulted in 23% of linear shrinkage of the HA preform,which was oversized to deliver an equivalent diameter(7.7±0.4 mm)than the CDHA preform.Shrinkage was also evident in the filamen diameter(300±6μm)and in-layer separation of the filament(281±37μm).The two consolidation mechanisms also resulted in preforms having different microstructure and pore size distribution.The microstructure of the CDHA preform consisted of an entangled network of plate-like nanocrystals(Fig.1d),with pore size between plates below 100 nm(Fig.1f).No pores bigger than 100 nm were observed in the CDHA preform,meaning that the size of the interfilamen porosity was above the detection limit of MIP.In contrast,the microstructure of the HA preforms corresponded to equiaxied polyhedral grains with some remaining intergranular porosity(Fig.1b)with size around 1μm(Fig.1f).In addition,the HA preform showed pores with size above 10μm and up to 150μm,representing the interfilamen porosity between layers.In the two cases the crystalline composition of the preforms corresponded to hydroxyapatite(ICSD:087670)(Fig.1e).However,the CDHA preform presented unreactedα-TCP and wider diffraction peaks compared to the sintered HA preform,which was attributed to the nanometric size of the crystalline domains.

Fig.1.Microstructure and crystalline composition of calcium phosphate(CaP)preforms.(a)Top view and(b)microstructure of HA preform.(c)Top view and(d)microstructure of CDHA preform.(e)X-ray diffraction patterns of consolidated preforms compared to diffraction peaks of HA(ICSD:087670).(f)Entrance pore size distribution of consolidated preforms determined by MIP.

3.2.Chemical and microstructural characterization of Mg and Mg alloys

The diffraction patterns of the Mg alloys showed the same diffraction peaks as pure Mg(ICSD:181728)(Fig.2a).However,extrusion led to the preferential orientation of the grains,showing high intensities of specifi crystallographic planes.The chemical analysis corroborated the purity of Mg,registering very low content of impurities,which are inherent to the extraction process(Table 2).The microstructural analysis confirme the results of XRD,i.e.the low content of Ca or Zn did not promote the formation of secondary phases or precipitates.Both,Ca and Zn produced a grain size refin ing effect,where 0.2 wt.% of Ca resulted in smaller grains(～10μm)(Fig.2c)than 1 wt.% of Zn(～20μm)(Fig.2b),in comparison to the coarser grains of the initial Mg ingot(＞500μm)(Fig.2d).

3.3.Chemical and microstructural characterization of interpenetrating Mg/CaP composites

CAMI allowed the efficien and reproducible infiltratio of Mg into the CaP preforms without their structural collapse(Fig.3).In general,there were no statistically significan differences(p=0.1147)in the infiltratio percentage(expressed by Eq.(1))between HA and CDHA preforms,being in average 98.9±0.9% for all the composites.TheμCT analysis showed that the 3D interfilamen porous network of the preforms was completely fille with the molten Mg,producing after solidificatio an interpenetrating biphasic structure,where each phase was self-connected and run continuously through the material(Fig.3).Furthermore,the additive manufacturing of the preform allowed the exact control of the topological relationship between the two phases.No differences in the phase volume fraction were observed based on the Mg composition,although there were certain differences between the two studied preforms(Fig.3).According toμCT analysis,the metallic fraction in the HA composites was around 55%,whereas in the CDHA composites,it was around 45%.Such difference was attributed to the different number of filament per layer in the two types of preforms and the dimensional change of the HA preform during sintering.

The XRD patterns showed in general a large-scale biphasic nature of the composites,however having traces of MgO(Fig.4b).The diffraction peaks of Mg evidenced preferential crystallographic orientation,arisen from fast solidificatio during CAMI,and low intensity of the diffraction peaks of HA in the CDHA composites.High resolution microstructural analysis by SEM showed that Mg did not penetrate the intergranular porosity of the preforms(Fig.4a),presumably due to the small pore size(Fig.1f).However,the molten Mg fille the cracks observed within the ceramic filament(Fig.4a).Moreover,the microstructural analysis revealed a coarse-grain microstructure of the metallic phases after CAMI,which in general had larger grains than after extrusion,and the emergence of a new phase on the grain boundary(Fig.4a).EDX analysis revealed that the new intergranular phase was rich in Mg and Ca,with the elemental ratio corresponding to the Mg2Ca intermetallic(close-up in Fig.4a).

Fig.2.Microstructure and crystalline composition of Mg and Mg alloys.(a)X-ray diffraction patterns of initial cast Mg and extruded Mg alloys.Optical microscope images of the microstructure of(b)extruded Mg-1 wt.% Zn,(c)extruded Mg-0.2 wt.% Ca,and(d)cast Mg.

Table 2Elemental chemical analysis of Mg and Mg alloys.

SEM analysis revealed a continuous layer formed at the ceramic-metal interface in all composites(Fig.5).Lineal elemental analysis by EDX identifie the chemical composition of the layer as MgO(Fig.5a).The thickness of the interfacial layer was about 200 nm in Mg/HA(Fig.5b)and Mg-0.2%Ca/HA(Fig.5d)composites,whereas in the Mg/CDHA(Fig.5c)and Mg-0.2%Ca/CDHA(Fig.5e)composites it was about 10μm.In the case of the Mg-1%Zn/CDHA composite,the thickness of the interfacial layer was close to 1μm.In all cases,the MgO layer was found to lay adjacent to both constituent phases.

3.4.Mechanical properties and fracture mode of interpenetrating Mg/CaP composites

Fig.6 shows the stress-strain curves for all the studied materials in compression,while Table 3 summarizes the mechanical properties extracted from the curves.The CaP preforms showed compressive strength two orders of magnitude below the Mg alloys and a brittle behaviour with less than 1% of strain to failure(Fig.6d).The incorporation of low amount of alloying elements increased more than fi e times the yield strength and more than two times the compressive strength of pure Mg.Ca provided a major strengthening effect than Zn(Table 3).Mg and the Mg alloys disclosed important permanent deformation after the elastic region,with clear work hardening before the maximum stress,followed by sample fracture(Fig.6a-c).The alloys deformed less before fracture than pure Mg.The infiltratio with Mg significantl increased the mechanical strength and strain to failure of the CaP preforms(Table 3 and Fig.6).Despite the CDHA preforms developed 2.5 times higher compressive strength than the HA preforms,the HA-based composites revealed almost three times higher yield and compressive strengths than the CDHA-based composites(Fig.6a,b).This result was attributed to the larger fraction of metallic phase(Fig.3)and also to the thinner MgO layer present at the ceramic-metal interphase(Fig.5),in the HA-based composites.With the exception of the composites made of pure Mg,which had a higher yield strength as compared to Mg alone,all interpenetrating composites reached elastic limit and compressive strength below those of the corresponding metallic phase(Table 3).

Fig.3.Three-dimensional(3D)virtual reconstructions of the interpenetrating phase Mg/CaP composites,together with the infiltratio percentage and volumetric phase content calculated byμCT.

Fig.4.Microstructure and crystalline composition of the Mg/CaP composites.(a)Cross-section SEM micrograph of Mg-0.2%Ca/HA composite,highlighting the presence of Mg2Ca intermetallic on the Mg grain boundary.(b)XRD patterns of interpenetrating phase Mg/CaP composites.

The fracture surface of the composites exhibited ductile failure of the metallic phase,with important plastic stretching of the metal,perpendicular to the direction of the compression force(Fig.6e,f).Moreover,multiple cracks formed within the ceramic phase,whose progression paths were restricted and deflecte along the ceramic-metal interphase,producing the interfacial debonding.The fla surfaces in both phases suggested a weak bonding.Only the Mg/HA composite suffered crack invasion and limited propagation through the metallic phase,where the crack progress was arrested(Fig.6e).The interpenetrating structure physically constrained the ceramic phase,causing its progressive breakdown into several pieces as the plastic deformation of the metallic phase continued,rather than catastrophic failure.The broken ceramic fragments detached after the fracture of the CDHA composite,whereas they remained fasten in the HA composite.

Fig.5.Magnesium-hydroxyapatite interface in the composites.(a)Representative image of the interface and elemental analysis by EDX showing the MgO layer formed between Mg and HA.The proposed mechanism for the formation of the MgO layer is shown in the coloured part of the image,for the description see the discussion.Representative interface region for(b)Mg/HA,(c)Mg/CDHA,(d)Mg-0.2%Ca/HA,and(e)Mg-0.2%Ca/CDHA composites.

Fig.6.Compression test results.Stress-engineering vs.strain curves for(a)pure Mg-based materials,(b)Mg-0.2%Ca-based materials,(c)Mg-1%Zn-based materials and(d)CaP preforms.Representative fracture surfaces for(e)Mg/HA and(f)Mg/CDHA interpenetrating phase composites.Solid arrows indicate some points of the plastic deformation of Mg before failure bridging the main fracture.Arrow heads show some cracks in the preforms,deflecte at the ceramic-metal interface.Dash arrows show the invasion of some cracks in the Mg phase.

3.5.Corrosion behaviour in simulated body flui

The polarization curves revealed that the two Mg alloys displayed similar reaction rates in SBF(Fig.7c).Moreover,the Mg alloys exhibited more positive Ecorrand higher jcorrthan pure Mg(Fig.7c and Table 4),indicating higher corrosion resistance,but faster instantaneous corrosion process.In general,the composites exhibited values of Ecorrbetween pure Mg and the Mg alloys,meaning that they own higher corrosion resistance than pure Mg,but lower than Mg alloys.The composites also had larger values of jcorrthan the bare metals.All the surfaces exhibited pitting corrosion after polarization,which was preferentially localized along the ceramic-metal interface(crevice corrosion)in the composites(Fig.7d).Mg alloys revealed smaller pits compared with pure Mg and Mg/CaP composites.The Mg-0.2%Ca alloy presented the largest density of small pits.

Table 3Mechanical properties under compression of CaP preforms,bare Mg-based metals and Mg/CaP interpenetrating phase composites.

Table 4Corrosion rates of interpenetrating Mg/CaP composites and Mg alloys by H2 evolution assessment and their corrosion potential(Ecorr)and corrosion current density(jcorr)from potentiodynamic polarization test.

Fig.7.In vitro corrosion test results.Corrosion rate profile calculated from the H2 evolution test of:(a)bare metallic materials,and(b)interpenetrating Mg/CaP composites.(c)Characteristic polarization curves in SBF.(d)Macroscopic aspect of representative interpenetrating Mg/CaP composites after polarization test,showing pitting corrosion(top),and after 2 weeks of degradation in SBF showing the galvanic dissolution of the metallic phase(bottom).

The corrosion-rate profile obtained from the H2evolution showed very high and decreasing corrosion rate during the firs 24 h of immersion in SBF(Fig.7a,b).The corrosion rate decreased due to the formation of a protective Mg(OH)2layer,reaching the equilibrium after the 5th day of immersion,without showing changes in the corrosion rate during the second week of measurement.In general,the composites exhibited higher corrosion rate than their bare-metallic-phase counterparts.The Mg-0.2%Ca/HA composite reached the highest corrosion rate after passivation(11.1±2.3 mm·year-1),whereas,among the composites,the Mg/HA showed the lowest corrosion rate(3.7±1.7 mm·year-1).The other composites had a corrosion rate between 5 and 6 mm·year-1.In the case of the metallic materials the corrosion rate at equilibrium revealed the following trend:pure Mg(4.2±0.3 mm·year-1)＞Mg-0.2%Ca(2.9±0.2 mm·year-1)＞Mg-1%Zn(1.9±0.7 mm·year-1).Note that this ranking does not mirror the trends of the corrosion current density because short-term corrosion tests by polarization curves measured immediately upon specimen immersion differ from long-term behaviour related to steady state corrosion.The macroscopic analysis of the composites at the end of the experiment demonstrated the galvanic dissolution of the metallic phase,leaving basically intact the ceramic preform(Fig.7d).

3.6.Materials’cytotoxicity

The materials induced time-and composition-dependent cell metabolic changes(Fig.8).They caused a moderate cytotoxicity after 1 day of incubation as the cell number and the cell metabolic activity were close or above the 70% of the control.Afterwards,the cell number and the cell metabolic activity decreased with the incubation time below 70% of the control.Therefore,the extracts of the materials were cytotoxic from day 3 according to the ISO standard 10993-5[51].In general,the composites caused a major reduction of the cell number and the cell metabolic activity than their bare metallic phase,without a clear trend between HA and CDHA composites.The extracts of the Mg-1%Zn alloy induced the largest reductions between the metallic samples and from shorter incubation time.

4.Discussion

The high infiltratio efficien y achieved by the CAMI(above～98%)puts forward this processing technology for the production of interpenetrating ceramic/metal composites with near net shape and controlled topological relationship of the components through the additive manufacturing of ceramic preforms.Furthermore,the low mechanical strength of the ceramic preforms was not a limitation for the infiltra tion,as the 3D architecture and continuity of the preforms were maintained after CAMI due to the special design of the die that avoids the direct application of load on the preform.No significan differences in the infiltratio efficiencie were observed among the different interpenetrating Mg/CaP composites.One of the major advantages of the CAMI processing was to limit the reaction between liquid Mg and the CaP preform.The reduction of CaPs and the oxidation of Mg have been observed even during solid state sintering of Mg/β-TCP composites by spark plasma sintering[37,38].The reaction was recorded as an exothermic event around 530 °C that involves an abrupt mass loss due to the decomposition of the phosphate group[37].However,the mechanism through which Mg reduces CaPs has not been described extensively;the chemical reaction assumes the formation of gaseous phosphorous compounds such as phosphine(PH3)and the formation of MgO[37,38,52].It is well accepted that the reducing capacity of Mg increases in liquid state.Nonetheless,the fast heating and cooling rates,together with the short dwell time,limited the formation of MgO during the CAMI to a continuous layer localized at the Mg-CaP interface.Therefore,CAMI appears as an emerging processing route in cases where there is a high chemical reactivity between the components.

The proposed mechanism for the formation of the MgO layer is shown in Fig.5a and involves the following steps.First,liquid Mg enters in direct contact with HA or diffuses through the MgO layer to be in contact with HA.Second,Mg reduces HA according to the following general reaction:

It is probable that Mg reduces firs the phosphate group producing CaO and subsequently generates Ca[53].It is also possible the formation of other phosphorous allotropes or compounds,such as calcium phosphide or phosphine[52].The progression of the reaction increases the thickness of the MgO layer shifting the reaction front towards the new surface of the HA preform.After the CaP reduction,the elemental Ca diffuses through the MgO layer and it is dissolved in the liquid Mg.The solubilized Ca is segregated during the metal solidificatio and precipitates on the grain boundaries as Mg2Ca intermetallic,such way that the presence of the intermetallic has a higher frequency close to the Mg-MgO interface.The volatile species generated are removed due to the continuous degassing of the SPS chamber.

The stoichiometry and microstructure of the preforms had a significan effect on the formation of the MgO layer.CDHA preforms had a Ca/P ratio of 1.5,plate-like crystal microstructure,and high specifi surface area(25.2 m2·g-1),which promote the reaction with Mg and consequently generating almost 50 times thicker MgO layer than HA preforms,which were stoichiometric(Ca/P ratio of 1.67)and had a smooth surface with low specifi surface area(0.6 m2·g-1).

HA-based composites showed around 2.7 times higher mechanical strength than the CDHA-based composites,but the difference being mainly attributed to the higher metal fraction in the HA-based composites(55%)than in the CDHA-based composites(45%),as the strength of the interpenetrating composites increases with higher metal content.However,an important effect of the MgO layer and its thickness may not be ignored,and further research is required to conclude on the mechanical stability of the layer and its interfacial bonding with Mg and HA,as to ensure an effective load transfer.Besides,it has to be considered that MgO transforms into Mg(OH)2in contact with water,generating a signifi cant increment of volume,which may cause stresses,microcracking and interfacial debonding[54,55].Therefore,in principle,the formation of MgO should be avoided or extremely minimized.

Fig.8.Results of the cytotoxicity assessment in material extracts.Cell number(a-c)and cell metabolic activity(d-f)expressed as % of untreated control(dash lines at 70 and 100%),for pure Mg based materials(a and d),Mg-0.2%Ca based materials(b and e),and Mg-1%Zn based materials(c and f).Data are expressed as the mean±S.D.The significan difference of a particular group from 100% of untreated control is marked by #(Student T-test,p＜0.05;n=16 per each group).The significan differences between similar sets of samples at specifi time-points(Mann-Whitney U test;p＜0.05;n=16 per each group)are marked by corresponding group legend pattern.

One of the most relevant microstructural improvements of the single-phase extruded Mg alloys was the grain refinement The grain refinemen with homogenous grain size significantl increased the yield and compressive strengths of the studied alloys respect to the as-cast pure Mg.In agreement with the Hall-Petch relationship[56],the Mg-0.2%Ca alloy achieved higher grain-boundary strengthening and lower ductility than the Mg-1%Zn alloy.

The minor dependence of the mechanical properties of the composites on the composition of the Mg phase was attributed to grain coarsening after the CAMI,as the fina grain size of the metallic phase was similar for all the Mg/CaP composites because of the melting and solidificatio during the liquid infiltratio process.Thus,grain coarsening after CAMI removed the grain-boundary strengthening leading to Mg-alloy composites with yield and compressive strengths far below the values of the corresponding extruded alloy.In contrast,the Mg grain size of the pure Mg materials did not show significan differences before and after CAMI,resulting in Mg/HA and Mg/CDHA composites with yield strength above pure cast Mg,the improvement being clearer in the case of the Mg/HA composites.The major resistance to plastic deformation was attributed to structural strengthening mechanisms particular for interpenetrating phase composites.The phase interlocking at nearly 50 to 50% along the material increased the stress transfer(relieving stresses from Mg)and the damage tolerance(deflecting bridging and arresting cracks formed in hydroxyapatite at the interfaces).In addition,the constriction limited the Mg expansion produced by the Poisson effect(promoting the structural integrity).Such mechanisms act in synergy or independently with the interfacial bonding,which is a key factor for the performance of particle and fibr reinforced composites[57].The reinforcement effect is lost at the yield stress due to the accumulation of damage in the CaP preform that is no longer able to constrain the local deformation of Mg.Notably,the composites showed 10 to 20 times higher compressive strength,and were signifi cantly less brittle,than the porous CaP preforms,putting forward the relevance of this type of materials for load-bearing applications.

The correlation of the mechanical performance of the interpenetrating composites with particle-reinforced composites is difficult because such composites had normally no more than 20% of CaP to prevent the agglomeration of particles,and the mechanical behaviour highly depends on the Mg alloy used as matrix,especially on the particular microstructure achieved after processing[58].However,an exception that allows the comparison is the particle reinforced Mg composite fabricated by powder metallurgy with 30% of HA particles,which developed 72 MPa and 92 MPa of yield and compressive strength,respectively[59].These values are around or below of those obtained for the interpenetrating Mg/HA composites with smaller fraction of Mg,55%.

The fracture mechanism of the interpenetrating composites is a complex process that deserves more detailed analysis in future to be completely understand and thus acquire the fundamental bases for the topological optimization of the interpenetrating 3D structure.Despite of a non-strict topological optimization of the Mg/CaP composites,an increase of grain size after Mg solidification and the presence of MgO layer at the Mg-CaP interface,the produced interpenetrating Mg-0.2%Ca/HA and Mg/HA composites exhibited an average compressive strength(110 MPavs.70-150 MPa)and strain to failure(～14%vs.8-18%)within the range of other interpenetrating Mg/CaP composites manufactured by suction or squeeze casting,which in general had significantl higher content of Mg(55% in this workvs.around 90% of Mg)[13-18].Moreover,the specifi yield strength of the strongest interpenetrating Mg/CaP composites achieved in this work was within the lower range of the compressive strength of cortical bone(～0.03 MPa/(Kg·m-3))[2,60].

The improved mechanical performance of the interpenetrating Mg/CaP composites introduced here and the possibility to tailor the mechanical behaviour by the topological design are two of the highlights of these materials.It is foreseen that the optimization of the porosity and the solid pattern of the preform may allow to achieve Mg/CaP composites with the required mechanical strength for fixatio of load-bearing bone fractures or fixatio of bone with a soft tissue.In fact,the roles of fraction,morphology,surface area and orientation of the reinforcement on the mechanical performance of fibre-reinforce and interpenetrating phase composites have been demonstrated[23,61-63].In case that the optimal 3D architecture of the ceramic preform was complex to be produced by robocasting,other ceramic additive manufacturing technologies such as ceramic vat photopolymerization can be explored to produce the preforms.

The dissimilarity between the degradation mechanisms and degradation rates between Mg and HA makes challenging the study of the bio-degradation of Mg/HA composites.On the one hand,Mg and its alloys are degraded by galvanic corrosion,generating hydrogen gas and Mg ions that under physiological condition form Mg(OH)2as solid corrosion product,thus increasing the alkalinity of the environment.On the other hand,HA is not soluble at physiological pH nor at the pH generated by the degradation of Mg.Therefore,the resorption of HA requires the active participation of cells,normally osteoclast which are the bone cells specialized in breaking down bone tissue through locally creating an acidic microenvironment by pumping hydrogen ions[64].It is hard to recreate the conditions for the cell mediated resorption of HAin vitro,therefore,as an alternative,the degradation of HA is studied in acidic solutions that mimic the microenvironment generated by osteoclast[65].However,Mg is highly prone to degradation in acidic pH,thus not reflectin thein vivosituation.Furthermore,the methods to quantify the degradation of Mg and HA are based in different principles.The degradation of Mg is evaluated by electrochemical and hydrogen evolution tests,whereas the degradation of HA is mainly evaluatedin vitroby mass loss.Despite mass loss can be applied with Mg,the method is not straightforward in the case of the interpenetrating composites because the removal of the adherent Mg(OH)2fil can erode the ceramic phase,introducing errors in the measurement.Considering low electrochemical corrosion potential and relative high corrosion rate of Mg,together with the slowin vivoresorption of HA(depending in the implant size it can take years to be fully resorbed by osteoclasts),it is reasonable to assume that Mg galvanic dissolution will be the main degradation mechanisms of the composites following the implantation.Therefore,in this work the study of the degradation of the interpenetrating Mg/CaP composites was focused on the degradation of Mg under simulated physiological conditions.This approach also allowed to compare the results with Mg matrix composites reported in the literature.

The refine grains in the extruded Mg-Ca and Mg-Zn alloys decreased the sensibility to galvanic corrosion by shifting towards more positive values the corrosion potential,and decreased the corrosion rate after one week of immersion in SBF respect to the cast Mg.However,the short-term corrosion process(before passivation)was faster due to localized corrosion mode.The effect of grain size on the corrosion of Mg and its alloys has been discussed previously[66].In general,the degradation rates of the interpenetrating Mg/CaP composites were faster than the ones reported for interpenetrating composites obtained by an infiltratio of Mg-1%Ca alloy in HA and TCP foams,with an average corrosion rate of about 3.4 mm·year-1[15,58].However,they exhibited higher degradation resistance than the conventional particle reinforced Mg matrix composites produced by powder metallurgy,which had degradation rates from 1.5 to 720 mm·year-1,depending on homogeneity in the distribution of the CaP particles and the densificatio of the fina composite[67-69].The interpenetrating Mg/CaP composites exhibited much faster degradation rate than their corresponding bare metallic phases due to the following factors:(i)the coarse grain size developed after the CAMI,(ii)the formation of the Mg2Ca intermetallic on the grain boundaries of Mg,(iii)the open porosity inside the fil aments of the CaP preform that allows the absorption of the corrosion medium deep into the composite,and(iv)the possible transformation of the MgO layer into Mg(OH)2producing interfacial debonding that further promotes the absorption of the corrosion medium.

The analysis of the corroded surfaces after the potentiodynamic polarization test revealed that the mechanism that governs the degradation of Mg in the interpenetrating composites was a mixture of pitting and crevice corrosion,preferentially localized along the metal-ceramic interface,where the amount of the Mg2Ca intermetallic was greater.While pitting was originated due to the heterogeneous Mg grain size[66,70,71]and the presence of Mg2Ca intermetallic[72-74]that form galvanic pairs at microstructural level,crevice was the result of the absorption of the corrosion medium in the ceramic fil aments,reaching the internal Mg surfaces on the Mg-CaP interfaces and promoted by the higher density of the intermetallic near the interface.The localized corrosion at the Mg-CaP interface has been reported in other interpenetrating Mg/CaP composites,assuming but not proving the formation of some galvanic couple at the interface[13,14,19].Microporous fil aments are one of the microstructural features of the CaP ceramics produced by robocasting,thereby producing Mg interpenetrating composites more prone to crevice corrosion,because the filament were exposed to the corrosion medium.The preferential degradation of Mg and the chemical stability of HA left behind the preform after a prolonged immersion period.This is in fact the concept of these degradable interpenetrating composites.Mg should provide mechanical stability during bone fracture consolidation,and then the degradation of Mg should give space for bone ingrowth into the porous preform.Furthermore,HA should promote bone ingrowth because it is a bioactive and osteoconductive material.Finally,the HA preform should be resorbed and substituted with new bone after bone remodelling.However,the degradation rate of Mg in the composites should be significantl reduced to match the bone regenerative capacity and minimize the rate of generation of corrosion products in order to guarantee the load-bearing capacity and prevent adverse host response reactions.According to the corrosion mechanisms identified a possible approach to reduce the degradation rate of the composite at short-term is to supress crevice corrosion by avoiding corrosion medium absorption in the ceramic fil aments through enabling the formation of a Mg shell during the infiltratio or by the deposition of a corrosion protective coating.

The extracts of the interpenetrating Mg/CaP composites exhibited mild to moderate cytotoxicity on osteoblast-like cells,without a clear relationship with the corrosion rate beyond that composites were less cytocompatible and had a faster degradation rate than the corresponding metallic phase.The response of the cells depends not only on the amount but also on the type of soluble corrosion products generated,Ca2+ions appear to be better tolerated by the cells than Zn2+ones.The lowin vitrobiological tolerance is not exclusive for the interpenetrating composites,but in general to Mg based materials,due to the fast accumulation of soluble degradation products that significantl increases the pH and osmolality of the cell culture medium,causing an osmotic shock on the cells[5,15,74-76].Therefore,it is not surprising that the biological effects of Mg-based materials strongly depend on the cell culture procedure and type of cells used[8,15],with nearly all Mg-based materials classifie as cytotoxic when different cells are cultured in static conditions and without the dilution of the soluble degradation products.A false outcome that does not correlate properly with thein vivodata[76,77].In contrast,the Mg-based materials produce slight cytotoxicity(grade I)when at least two-fold dilution of the soluble corrosion products is used for the cell culture[15].Furthermore,thein vitroresults of the Mg toxicity with diluted extracts agreed better with the biocompatibility observed,both,locally in direct contact with bone tissue and systemically in visceral organs of the host[11,75,77-79].In the case of the insoluble corrosion products,Mg(OH)2is well tolerated and provides corrosion resistance through the formation of a passivation film The use of non-diluted extracts under the current experimental protocol challenged the cells,which after one day of culture reduced their metabolic activity and cell number significantl.Even that the dilution of the extracts may result in better cell activityin vitro,extensive work remains to decrease the degradation rate of the interpenetrating Mg/HA composites in the biological environment.Only then,this still promising degradable materials will offer unprecedented alternatives for bone repair.

The mechanical performance and the degradation resistance can be further improved through the optimization of the metal infiltratio parameters to avoid the decomposition of HA and the grain coarsening.The use of stoichiometric HA preforms would slow down the chemical reaction,while increasing the heating rate and shortening the soak segment during CAMI would reduce the time available for reaction.In addition,the preform can be coated with a protective layer that avoids the reaction.As for the grain size refinement it might be achieved by increasing the solidificatio rate by the reduction of the dimensions of the graphite die to enable faster cooling or by the introduction of nucleation agents in the metal if needed.

5.Conclusion

CAMI allowed the complete infiltratio of liquid Mg in HA-based preforms without their structural damage.The combination of the CAMI with ceramic additive manufacturing allowed the production of 50 to 50% interpenetrating phase composites with customized topological design.Despite CAMI settings limited the time for chemical reactions,the high reactivity of liquid Mg led to the surface decomposition of the HA-based preforms,forming a MgO layer at the Mg-HA interface.The thickness of the MgO layer significantly depended on the stoichiometry and surface morphology of HA.The decomposition of HA also formed Mg2Ca on the Mg grain boundaries,mostly near the Mg-HA interface.Melting and solidificatio induced grain coarsening of the Mg-alloys,which led to composites with lower compressive strength than the extruded alloys.In contrast,the compressive yield strength of pure Mg composites was above the yield strength of cast Mg due to the similar grain size and structural strengthening mechanisms.A combination of pitting and crevice corrosion produced the degradation of the composites,which was promoted by the absorption of the corrosion medium through the ceramic filament and the presence of Mg2Ca intermetallic in the coarse Mg grain microstructure.Preventing Mg grain growth and completely suppress HA decomposition during the CAMI are requirements to improve both the mechanical performance and degradation resistance of Mg/CaP composites.It may be feasible to produce interpenetrating phase composites with a minimum fraction of Mg that provides load-bearing capacity during bone healing,and afterwards,the degradation of Mg gives space to new bone ingrowth promoted by the osteoconductivity of HA.Thus,offering more alternatives for the fabrication of temporal implants.However,the reduction of the degradation rate under physiological conditions is necessary to produce functional composites.The methodology introduced in this work can be extrapolated to process other reactive metal-ceramic systems different from Mg-HA,with diverse and controlled properties in accordance with their application.

Declaration of Competing Interest

There are no conflict to declare.

CRediT authorship contribution statement

Mariano Casas-Luna:Formal analysis,Investigation,Data curation,Writing-original draft,Visualization.Edgar B.Montufar:Methodology,Formal analysis,Investigation,Funding acquisition,Writing-original draft.Norbert Hort:Methodology,Resources,Writing-review & editing.Sebastian Díaz-de-la-Torre:Methodology,Resources,Writing-review & editing.JoséClaudio Méndez-García:Resources,Writing-review & editing.Lucie Vi?tejnová:Formal analysis,Investigation,Writing-review & editing.Adam Brínek:Formal analysis,Investigation,Writing-review & editing.Ale?Daˇnhel:Resources,Writing-review& editing.Karel Dvoˇrak:Resources,Writing-review &editing.Jozef Kaiser:Resources,Writing-review & editing.LadislavˇCelko:Conceptualization,Methodology,Supervision,Funding acquisition,Writing-review & editing.

Acknowledgments

This work was supported by the Czech Science Foundation(grant 19-22662S).CzechNanoLab project LM2018110 funded by MEYS CR is gratefully acknowledged for the support of the measurements at CEITEC Nano Research Infrastructure.MCL acknowledges to Brno Ph.D.Talent scholarship and to the Brno University of Technology Internal Project:CEITEC VUT-J-19-5915.SDT acknowledges to CONACYTSNI and SIP-IPN(SAPPI 20220438).LV acknowledges to project no.NU20-08-00150(MH,Czechia).Special thanks to A.Pati?o-Pineda from CIITEC-IPN for their technical assistance during CAMI,M.Horynová,P.Gejdo?,P.Skarvada and T.Zikmund from CEITEC-BUT for their technical assistance during sample characterization and to Z.Pavlousková from CEITEC-BUT for her assistance in administrative tasks.