Extrusion-based 3D printed biodegradable porous iron

(1)

Zadpoor, A. A.

DOI

10.1016/j.actbio.2020.11.022

Publication date

2021

Document Version

Final published version

Published in

Acta Biomaterialia

Citation (APA)

Putra, N. E., Leeflang, M. A., Minneboo, M., Taheri, P., Fratila-Apachitei, L. E., Mol, J. M. C., Zhou, J., &

Zadpoor, A. A. (2021). Extrusion-based 3D printed biodegradable porous iron. Acta Biomaterialia, 121,

741-756. https://doi.org/10.1016/j.actbio.2020.11.022

Important note

To cite this publication, please use the final published version (if applicable).

Please check the document version above.

Copyright

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy

Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

(2)

'You share, we take care!' - Taverne project

https://www.openaccess.nl/en/you-share-we-take-care

Otherwise as indicated in the copyright section: the publisher

is the copyright holder of this work and the author uses the

Dutch legislation to make this work public.

(3)

Extrusion-based

3D

printed

biodegradable

porous

iron

N.E. Putra

a,∗

_{, M.A. Leeﬂang}

a

_{, M. Minneboo}

a

_{, P. Taheri}

b

_{, L.E. Fratila-Apachitei}

a

_{, J.M.C. Mol}

b

_,

J. Zhou

a

_{, A}

_.A

_{. Zadpoor}

a

a Department of Biomechanical Engineering, Faculty of Mechanical, Maritime, and Materials Engineering, Delft University of Technology, Mekelweg 2, 2628

CD Delft, the Netherlands

b Department of Materials Science and Engineering, Faculty of Mechanical, Maritime, and Materials Engineering, Delft University of Technology, Mekelweg 2,

2628 CD Delft, the Netherlands

a

r

t

i

c

l

e

i

n

f

o

Article history: Received 31 August 2020 Revised 10 November 2020 Accepted 16 November 2020 Available online 20 November 2020 Keywords: 3D printing Material extrusion Biodegradable Iron Scaffold Bone substitution

a

b

s

t

r

a

c

t

Extrusion-based3Dprintingfollowedbydebindingandsinteringisapowerfulapproachthatallowsfor the fabricationofporous scaffolds frommaterials (or materialcombinations) that areotherwisevery challengingtoprocessusingotheradditivemanufacturingtechniques.Ironisoneofthematerialsthat havebeen recentlyshowntobeamenable toprocessingusingthisapproach. Indeed,afully intercon-nectedporousdesign hasthe potentialofresolving thefundamentalissueregardingbulkiron,namely averylowrateofbiodegradation.However,noextensiveevaluationofthebiodegradationbehaviorand propertiesofporousironscaffoldsmadebyextrusion-based3Dprintinghasbeenreported.Therefore,the

in vitro biodegradationbehavior,electrochemicalresponse,evolutionofmechanicalpropertiesalongwith biodegradation,andresponsesofanosteoblasticcelllinetothe3Dprintedironscaffoldswerestudied. Aninkformulation,aswellasmatching3Dprinting,debindingandsinteringconditions,wasdeveloped tocreateironscaffoldswithaporosityof67%,aporeinterconnectivityof96%,andastrutdensityof89% aftersintering.X-raydiffracometryconﬁrmedthepresenceoftheα-ironphaseinthescaffoldswithout anyresidualsfromtherestoftheink.Owingtothepresenceofgeometricallydesignedmacroporesand randommicroporesinthestruts,the in vitro corrosionrateofthescaffoldswasmuchimprovedas com-paredtothebulkcounterpart,with7%masslossafter28days.Themechanicalpropertiesofthescaffolds remainedintherangeofthoseoftrabecularbonedespite28daysof in vitro biodegradation.Thedirect cultureofMC3T3-E1preosteoblastsonthescaffoldsled toasubstantial reductioninlivingcellcount, causedbyahighconcentrationofironions,as revealedbytheindirectassays.Ontheotherhand,the abilityofthecells tospreadandformﬁlopodiaindicatedthe cytocompatibilityofthecorrosion prod-ucts.Takentogether,thisstudyshowsthegreatpotentialofextrusion-based3Dprintedporousironto befurtherdevelopedasabiodegradablebonesubstitutingbiomaterial.

Statementofsigniﬁcance

3Dprintingtechniqueshavebeenusedtofabricateporousironscaffoldsforbonesubstitution. However, no extensive performance evaluation of porous iron made by extrusion-based 3D printingforbonesubstitutionhaseverbeenreported.Therefore,wecomprehensivelystudied theinvitrobiodegradation behavior,electrochemicalcharacteristics,time-dependent mechan-ical properties,and cytocompatibility ofporous iron scaffoldsmade by means of extrusion-based3D printing.Ourresultsshowedthatextrusion-based3D printingcoulddeliverporous iron scaffoldswithenhanced biodegradabilityandbone-mimickingmechanicalpropertiesfor potentialapplicationasbiodegradablebonesubstitutes.

∗_{Corresponding author.}

E-mail address: n.e.putra@tudelft.nl (N.E. Putra).

1. Introduction

Biodegradable metals and their alloys, including those based on magnesium, iron, and zinc, have been extensively studied to explore their potential as temporary bone substitutes [1,2]. Such https://doi.org/10.1016/j.actbio.2020.11.022

(4)

biodegradable biomaterials could alleviate the risk of prolonged inﬂammation associatedwithnon-biodegradable metals that per-manently remain at the implantation site[3]. Among biodegrad-ablemetallicmaterials,thosebasedonironenjoytheadvantageof possessingsubstantiallyhighermechanicalpropertiesascompared to the magnesium- andzinc-based alternatives [4].Those higher mechanical properties open up unique opportunities for adapt-ing the mechanical properties of iron-based materials to those of the local bony tissue by introducing porous structures. The porousstructurealsohastheaddedvalue ofsigniﬁcantly enhanc-ing the internal surface area of the implants to accelerate the biodegradation process of iron [5],which is otherwise extremely low.

Owing toits abundance innature,ease ofmanufacturing, and highmechanicalperformance,iron-basedmaterials havebeen ex-tensively used asstructural materials andpotentially to be used asbiodegradablematerials forbonesubstitution[6].Ironisan es-sentialelementinthehumanmetabolismfortransportingoxygen, mediatingelectrontransfers,andacceleratingenzymereactions.It plays aroleinimmunefunction[7].Moreover,thehomeostasisof iron is essential foroptimalbone regeneration [8]. Iron actsasa catalystfortheformationofreactiveoxygenspecies(ROS)andan appropriate ROSlevelhasbeenreportedto regulateapathwayin osteoblastdifferentiation[9].Adequateiron uptakeinvivocan as-sistin thedevelopmentofosteoblasts [10,11]andinduce platelet activation [12],whichis importantfortheinitial healingstage of bone recovery.On theone hand,there isevidence that ironions exhibit cytotoxicitywhentherateofionreleaseexceedsacertain threshold [13]. On the other hand, long-term in vivo studies on iron-based bone substitutes haveshown thatthe corrosion prod-uctsarebiocompatible[14,15].Itisimportanttonote thatthe re-portedinvivocorrosionratesareextremelylowwithnosigniﬁcant changesinthemassofiron-basedmaterialsimplantedinthebulk (after 52weeks)[14]andfoam(after6weeks) forms[15],which mayleadtothelonger-than-expectedlongevityofsuchiron-based implants(withinafewyears[16]).Itisstillanopenresearch ques-tionwhetherthebiocompatibilityofiron-basedmaterialsremains favorableenoughwhentheirbiodegradationratesareenhancedto matchtherateofbonetissuehealing.Forimplantapplications,the ferromagnetic nature of iron, causing complications during mag-netic resonanceimaging (MRI), is anotherissue, which hasbeen addressed by alloyingwith28 wt% ormoremanganese to retain theaustenitephaseandchangethealloytobe non-ferromagnetic

[17].

Toaddress theﬁrst issue,i.e.,tooslowbiodegradation ofiron, recent research has been focused on developing new types of iron-basedalloys[18,19],modifyingthesurfacepropertiesthrough sandblasting [20] and bio-functionalizing the surface using poly-mers [21-23]orbioceramics [24,25] toaccelerate biodegradation. In addition, porous structures have been fabricated using vari-oustechniques,includingbothtraditionaltechniques,suchas elec-trodeposition [26],spaceholder method[27-31],powdermolding onto porous template [32], and 3D printing techniques, such as binder/ink-jetting[33,34],materialextrusion[35-39],andselective lasermelting[40-43].

3D printing has recently emerged as the most promising ap-proach tothe fabrication ofporous biodegradable metals[44,45]. Using this approach, tailor-made designs of biodegradable scaf-folds can be realized to meet the speciﬁc requirements of vari-ous treatment conditions.Examples includetailor-made solutions to the treatment of critical bony defects and osteosynthesis as well as site-speciﬁc drug delivery for cancer patients [45]. 3D printed implantswitha geometricallyordered porenetwork pro-vide tissue-mimicking mechanical properties,facilitate osseointe-gration, and greatly increase the surface area to volume ratio, thereby increasing therateofbiodegradation. Extrusion-based3D

printingprovidesastraightforwardapproachtotheexsitu fabrica-tionofporousstructuresusingmetallicmaterials(ormaterial mix-tures),whichwouldotherwisebechallengingforother3Dprinting techniques[46],suchaspowderbedfusion3Dprintingprocesses [47,48].

Inthe last few years, 3D printingtechniques, including mate-rialextrusionandselectivelasermelting(SLM),havebeenusedto fabricate porous pure iron for bone substitution. Extrusion-based 3D-printed topologically ordered porous iron scaffolds with cu-bic unit cell (31% porosity) and hexagonal unit cell (60% poros-ity) have been evaluated in terms of their mechanical behavior

[37-39].Thesurfaceoflay-downpatternedironscaffoldshasbeen biofunctionalizedbyapplyinghydroxyapatitecoatingforenhanced cell-material interaction [35]. Afew in-depthstudies on the ma-terialpropertiesandbiocompatibilityofSLMporousironhave ap-pearedintheliteraturetoo[40,41].However,noextensive evalua-tionofporousironmadebyextrusion-based3DpriFntingis avail-able yet.We, therefore,studied the early-stage in vitro biodegra-dation behavior, electrochemical response, time-dependent me-chanical properties,and cytocompatibility of lay-down patterned porous iron scaffolds made by means of extrusion-based 3D printing.

2. Materialsandmethods

2.1. Scaffolddesign,3Dprinting,andpost-processing

In this research, pure iron was chosen to understand the in vitrobiodegradation behavior, electrochemicalresponse, evolution ofmechanicalproperties,andcytocompatibilityof3Dprinted scaf-foldsso as to build up a solid base,on which further endeavors could be madetoaddresss the other issuesrelatedto iron-based materials for biomedical applications, forexample, by adding al-loyingelementsorfunctionalagents.Ironpowderwith99.88wt% purity,minorimpurities(Cu,Ni,Mo,Al,Mn,Si,andCr),and spher-ical particle morphology (Fig. 1a)produced through nitrogen gas atomization(MaterialTechnologyInnovationsCo.,Ltd.,China)was sieved to reach a particle size distribution of D10 = 25.85 μm, D50 = 39.93μm andD90 = 53.73 μm.A printable ink with iron powder loadingwasprepared bymanually mixing theiron pow-der with a 5 wt% hydroxypropyl methylcellulose (hypromellose) polymer (Mw ~86 kDa, Sigma Aldrich, Germany) aqueous solu-tion [35] at a mass ratio of 7:1 (corresponding to a volume ra-tioof49:50),basedonthepreliminaryexperimentswithdifferent powder-to-binder(mass)ratios inorderto choosea 3D printable ink. Then, the shear-thinning properties of the chosen ink were studiedusinganMCR302rheometer(AntonPaarGmbH,Germany). Inaddition,thermogravimetricanalysis(TGA)ofhypromellosewas performed using an SDT Q600 v20.9 thermogravimetric analyzer (TAInstruments,USA).

Porousironscaffolds(10.5mminheightand10mmin diame-ter)weredesignedwiththeGeSiMRoboticssoftware(GeSiM Bio-instruments, Germany) to have a strut width of 410 μm, a strut spacingof400μm,alayerthicknessof328μm,adesigned poros-ityof50%,andaninitialdesignsurfaceareaof40.4cm2 ₍_Fig.₁_b). The iron ink wasextrudedinto 3D porousiron scaffolds through a 410

μ

m tapered nozzle tip using 3D BioScaffolder 3.2 (GeSiM Bio-instruments, Germany) with angles of 0° and 90° that inter-changed every layer (Fig.1c). The printing pressure andprinting speed were set at 200 kPa and 5 mm/s, respectively. After 3D printing, the as-printed iron scaffoldswere allowed to dryforat least30minin adesiccator, beforebeingloaded intoatube fur-naceSTF16/180(CarboliteGeroLtd.,UK)underahighlypureargon atmosphere(purity: 99.9999%; inlet pressure: 1 bar) andheld at 350°Cfor1h fordebindingandat1200 °Cfor6h forsintering. Finally, the as-sintered iron scaffolds were ultrasonically cleaned

(5)

Fig. 1. The starting material, extrusion-based 3D printing, and scaffold design: (a) iron powder particle morphology, (b) an illustration of extrusion-based 3D printing, and (c) the scaffold with the 0 ° and 90 ° lay-down pattern design.

in isopropyl alcohol for 15 min prior to characterization and investigation.

2.2. Characterizationofmacrostructureandmicrostructure

Theheightanddiameteroftheporousironscaffoldswere mea-sured before andafter sintering to determine the shrinkage.The micro-architectureoftheironscaffoldswasobservedusinga scan-ningelectronmicroscope(SEM,JEOLJSM-IT100,Japan).FromSEM images, thestrut width, strut spacing, andlayer thicknessofthe scaffoldsweremeasured.Thecrosssectionsoftheas-sinterediron scaffolds, after polishing up to 1 μm, were observed using SEM. Theregions ofinterestonthecrosssectionofthestrutswere de-ﬁned and analyzed withcolor thresholdin ImageJ (NIH, USA)to select the porous area. The solid fractions of the struts(X) were calculatedbasedonthefollowingequation:

X=

1− Porearea

Totalareao f ROI

× 100% (1)

Furthermore, theabsoluteporosities ofthe as-printed and as-sinteredironscaffoldswerecalculated,usingtheweighingmethod andEqs.(2and3):

ϕ

p=

1−mp/

ρ

ink Vbulk

× 100% (2)

ϕ

s=

1−ms/

ρ

iron Vbulk

× 100% (3)

where

ϕ

p and

ϕ

s are, respectively, theabsolute porosities ofthe as-printed andtheas-sinterediron scaffolds[%],misthemassof theas-printedoras-sinteredironscaffolds[g],Vbulkisthebulk vol-umeofthescaffold[cm3_],

_ρink

_is _the_density_of_the _iron_ink ₍_i.e._, 4.41 g/cm3_), _and

ρiron

_is_the _theoretical _density_of_pure _iron ₍_i.e._, 7.874g/cm3_).

Inaddition,theinterconnectedporosity oftheas-sintered iron scaffolds was assessed using the Archimedes’ principle as de-scribed intheASTMstandardB963–13[49].Thevalue was calcu-latedusingEq.(4):

ϕ

i=

ρ

e

ρ

o× mao− ma mao− meo

× 100% (4)

where

ϕi

is the interconnected porosity of the as-sintered iron scaffolds [%],

ρ

e is the density of ethanol (i.e., 0.789 g/cm3),

ρ

o isthedensityofoil(i.e.,0.919g/cm3_), _m

aoisthemassofthe oil-impregnatedironscaffoldweighedinair[g],maisthemassofthe iron scaffold weighed in air[g], and meo is the mass of the oil-impregnatedironscaffoldweighedinethanol[g].

2.3. Phaseidentiﬁcation

Thephasecompositionoftheas-sinteredironscaffoldswas de-terminedusinganX-raydiffractometer(XRD,D8Advance,Bruker, USA). XRD in the Bragg-Brentano geometry was equipped with

a graphitemonochromatoranda Vantecposition-sensitive detec-tor that was set to work at 45 kV and 35 mA. A step size of 0.020° withacountingtimeof10sperstepusingCoK

α

radiation wasemployed. The XRD patternwas evaluated using the Diffrac Suite.EVAv5.0software(Bruker,USA)andtheInternationalCentre forDiffractionDataPDF-4database.

2.4. Staticinvitroimmersiontests

Staticin vitroimmersiontests(upto 28days,triplicates)were performedintherevisedsimulatedbodyfluid(r-SBF)[50]usinga cellcultureincubator,underthefollowingconditions: 5%CO2,2% O2, relative humidity = 95%, temperature= 37 °C± 0.5 °C. The solutionvolume-to-surfacearea ratiowas6.7mL/cm2 _[51]_._Before thetests, the sampleswere sterilized andthe r-SBFsolutionwas filteredusinga0.22μmfilter(MerckMillipore,Germany).ThepH values were monitored during the immersion period using a pH electrode (InLab Expert Pro-ISM, METTLER TOLEDO, Switzerland). Theconcentrationsofsolublecalcium,phosphateandironionsin the r-SBF solution were quantified using an inductively coupled plasma optical emission spectroscope (ICP-OES, iCAP 6500 Duo, Thermo Scientific, USA) after 1,2, 7, 14, and 28 days of in vitro biodegradation.Todeterminethemassloss,the invitrocorrosion products were removed by immersing theas-corroded iron scaf-folds atthe abovementioned time pointsin a 50 vol% HCl solu-tion(withaspecificgravityof1.19,SigmaAldrich,Germany) con-taining3.5g/Lhexamethylenetetramine (SigmaAldrich,Germany) for 10 min, followed by ultrasonic cleaning in isopropyl alcohol for15 min [52]. Subsequently, the samples were dried overnight ina desiccatorandweighed usinga balance withan accuracyof 0.1 mg. The cycle was repeated and the mass loss was plotted against the cleaning cycle toobtain the most accurate value, ac-cordingtotheASTMstandardG1–03[52].Fromthemassloss val-ues,theaveragecorrosionratewasdeterminedbasedontheASTM standardG31–72[53]usingEq.(5):

CRimmersion [mm/year]=8.76 × 104 ×

m

A ×

ρ

× t (5) wheremisthemassloss[g],Aisthesurfaceareaofthescaffolds [cm2_]_calculated_based_on_the_initial_scaffold_design_value,

_ρ

_is_the theoreticaldensityofpureiron(i.e.,7.874g/cm3_),_and_t_is_the du-rationofinvitroimmersion[h].

2.5. Characterizationofinvitrocorrosionproducts

Thephasesoftheinvitrocorrosionproductsofthescaffolds af-ter28daysofimmersionwereidentiﬁedusingXRD(D8Advance, Bruker,USA).Inaddition,themorphologiesoftheinvitrocorrosion productsontheperipheryoftheironscaffoldsafter7,14,and28 daysofcontinuousimmersionwereobservedusingSEMandtheir compositionswereanalyzedwithanX-rayenergydispersive spec-troscope(EDS,JEOLJSM-IT100,Japan).Inadditiontotheperiphery,

(6)

thescaffoldsafter7,14,and28daysofcontinuousimmersionwere groundwithSiC#2000,andthecorrosionproductsinthecenterof thestructurewerecharacterizedusingSEM-EDS.Also,thefraction oftheremainingbasematerial(pureiron)afterinvitroimmersion attheselectedtimepointswascalculatedusingImageJ(Eq.(1)). 2.6. Electrochemicalmeasurements

To study the electrochemical corrosion behavior, iron scaffold specimens were carefully prepared by partial mounting them in thermoplasticacrylic resin andexposingthem tor-SBF(pH=7.40, temperature= 37± 0.5 °C).Theexposed surfaceareawas calcu-lated,basedonthedesignvalueofthescaffold.Beforethe experi-ments,themountedspecimenswereultrasonicallycleanedin iso-propylalcohol andthendried thoroughly. Athree-electrode elec-trochemical system was conﬁgured, in which a graphite rod, an Ag/AgClelectrode,andtheironspecimen,respectively,actedasthe counter electrode,the referenceelectrode, andthe working elec-trode.AllthetestswerecarriedoutintriplicateusingaBio-Logic SP-200potentiostat(Bio-LogicScienceInstruments,France).

Beforetheelectrochemicaltests,thesetupwasallowedtoreach astableopencircuitpotential(OCP)for1h.Thelinearpolarization resistance(LPR)testsoftheironspecimensatdifferenttimepoints up to 28dayswere carriedout from−25to +25mV versusOCP at a scan rate of 0.167 mV/s. Consecutively, the electrochemical impedancespectroscopy(EIS)testsoftheironspecimensat differ-enttimepointsupto28dayswereconductedusingasine ampli-tudeof10mVversusOCPatafrequencyscanbetween100kHzto 10 mHz.Moreover, potentiodynamicpolarization(PDP)testswere performed on the specimens after1 day and28 days of immer-sion, withpolarizationbetween−300 to +500 mVversus OCP at a scanrateof0.5mV/s.Fromthe PDPresults,the corrosionrates werecalculatedaccordingtotheASTMstandardG102–89[54]and usingEq.(6):

CRelectrochemical [mm/year]=3.27 × 10−3 × EW ×

icorr

ρ

(6)

whereEW istheequivalentweight ofiron (valence2),icorristhe currentdensity[μA/cm2_],_and

_ρ

_is_the_theoretical _density_of_pure iron[g/cm3_].

2.7. Uniaxialcompressiontests

An Instronuniversaltestingmachine(ElectroPulsE10000, Ger-many) witha10 kNload cellwasusedto evaluatethe compres-sivemechanicalpropertiesoftheas-sinteredironscaffoldsaswell as the specimens retrieved after 1, 2, 7, 14 and 28 days of in vitro immersion. The tests were performedat a crosshead speed of3 mm/min.The mechanicalpropertiesoftheporous iron scaf-folds,includingthequasielasticgradient(referredaselastic modu-lus)andyieldstrength,were obtainedfollowingtheISOstandard 13,314[55].Theslopeoftheﬁrstlinearregioninthestress-strain graphwasdeﬁnedastheelasticmodulus.Aparallellinetothe ini-tiallinearelastic region,offsetby0.2%strain, wasdrawn andthe stressvalueattheintersectionwiththestress-straincurvewas de-termined asthe yieldstrength.The testswereperformedin trip-licate. The average values with standard deviations are reported hereafter.

2.8. Cytocompatibilityevaluation

2.8.1. Pre-cultureofMC3T3-E1cellsandthepreparationofiron extractmedia

Preosteoblasts MC3T3-E1 (Sigma Aldrich, Germany) were pre-cultured for 7 days in

α

-minimum essential medium (

α

-MEM,

Thermo Fisher Scientific, USA) withoutascorbic acid and supple-mentedwith10%fetalbovineserum(FBS,ThermoFisherScientific, USA)and1%penicillin/streptomycin(p/s,ThermoFisher Scientific, USA).

α

-MEMwithoutascorbicacidwasusedtomaintainthecells inthe preosteoblastic state. The cellswere incubatedat37 °C in ahumidiﬁedatmospherewith5%CO2 and2%O2(relative humid-ity₌95%).Theculturemediumwasrefreshedevery2to3days.

Theironextractculturemediawerepreparedbyimmersingthe sterilizedporousironscaffolds(10.25mminheightand9.75mm indiameter)in

α

-MEM(without ascorbicacid, butwith10%FBS, 1%p/s)for72hat37°Cina5%CO2 and2%O2atmospherewith 95% relativehumidity [56].The specimen-to-medium ratiowas5 cm2_/mL,_in_which_the_surface_area _of_the_scaffolds_was_calculated based on the design value. Thereafter, the supernatant was col-lected,filtered,anddiluted into75%, 50%,25%,and10% fromthe original concentration.Furthermore,theiron ionconcentration in the original (100%) iron extract media wasquantified using ICP-OES(iCAP6500Duo,ThermoScientific,USA).Allmediawerekept at4°Cpriortocelltests.

2.8.2. Indirectcytotoxicitytests

To evaluate the indirect cytocompatibility of the iron scaf-folds,thePrestoBlueassay(ThermoFisherScientiﬁc,USA)was per-formed. The MC3T3-E1 preosteoblasts (1 × 104 _cells) _were cul-turedin200μLof100%,75%,50%, 25%and10% ironextract me-diausing 48-well plates.The same numberofcellswascultured inthe original

α

-MEM (without ascorbic acid, butwith10% FBS, 1% p/s) asthenegativecontrol. After 1,3, and7days ofculture, the iron extract media were replaced with200 μL freshpure

α

-MEM(withoutascorbicacid,butwith10%FBS, 1%p/s)toprevent theinterference ofiron ions withtheassay. Consecutively, 20μL ofPrestoBluereagents(ThermoFisherScientiﬁc,USA)wereadded, andthespecimenswereincubatedat37°Cfor1h.Afterwards,the absorbancevalues were measured witha Victor X4 Wallac plate reader(PerkinElmer, USA)at awavelength of530- 590nm. The testswere performedintriplicate. Theaverage metabolic activity ofthecellsisreportedasapercentageofthenegativecontrol, cal-culatedusingEq.(7):

Metabolicacti

v

ity[%]= Absorbance

(

specimen

)

Absorbance

(

negati

v

econtrol

)

× 100

(7) To observe the morphology of the preosteoblasts grown in the iron extract media, the cytoskeleton and nucleus of the cellswerestainedusingrhodaminephalloidinand4 ,6-diamidino-2-phenylindole (DAPI) dyes, respectively. The MC3T3-E1 pre-osteoblasts (1 _{× 10}4 _cells) _were _cultured _for ₃ _days _on _48-well glassdisks in200 μLofiron extractmedia. The samenumberof cellswascultured in the original

α

-MEM(without ascorbic acid, but with10% FBS, 1% p/s) asthe negativecontrol. After culture, the specimens were washed with PBS (Thermo Fisher Scientific, USA), fixed using 4% formaldehyde (Sigma Aldrich, Germany)for 15min atroom temperature, then washedwithPBS,and perme-abilizedwith0.5%triton/PBS(SigmaAldrich,Germany)at4°Cfor 5 min.Then, 1% bovine serum albumin/PBS (BSA, SigmaAldrich, Germany) was added per well and followed by incubation for 5min.Consecutively,1:1000rhodaminephalloidin(ThermoFisher Scientific, USA) in 1% BSA/PBS was added per well, followed by 1hincubationat37°C.Afterwards,thespecimenswererinsedin 0.5% tween/PBS (SigmaAldrich, Germany) andwashed with PBS. Finally,thespecimensweremountedonaglassslidewithProlong gold(LifeTechnologies,USA),containingthe DAPIdye. Thereafter, themorphologyofthecellsculturedindifferentironextractswas observed using a fluorescence microscope (ZOE cell imager, Bio-Rad,USA).From stainingimages,thenumberofcellsandthecell

(7)

Fig. 2. Morphology and phase composition of the porous iron scaffolds: SEM images of (a, b, c) the as-printed iron scaffolds and (d, e, f) the as-sintered iron scaffolds at different magniﬁcations, (g) the cross-section of the polished struts, and (h) the XRD pattern of the scaffolds after sintering. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

spreading area were determined by counting the nucleus of the cellsusingImageJ(NIH,USA).

2.8.3. Directcytotoxicitytests

To evaluate the direct cytocompatibility of the iron scaffolds, the Trypan blue cell counting assay was performed. First, the porous iron scaffolds (10.25 mm in height and 9.75 mm in di-ameter)werepre-immersedin

α

-MEM(withoutascorbicacid,but with10%FBS,1%p/s)for5minandplacedinthe6-wellplate. Af-terthat,theMC3T3-E1preosteoblasts(1× 106_cells_per_specimen) were pipettedintotheporous ironscaffolds.Subsequently,10mL of

α

-MEMwasaddedperwell toimmersethescaffold.The same numberofcellswasculturedforall experimentalgroups. For ev-ery time point, the testswere performedin triplicate.After 1, 3, and7daysofculture,thecellsweretrypsinizedfromthescaffolds and the well plates andthen suspended. 10 μL ofeach cell sus-pensionwasmixedwith10 μLoftrypanbluedye(Bio-Rad,USA) andsubsequentlypipettedintoadual-chambercellcountingslide. The numberof livingcells wascounted using an automated cell counter (TC20,Bio-Rad, USA). The averagenumber ofliving cells withstandarddeviationsarereported.

In addition, to observe the distribution and morphology of MC3T3-E1 cells seeded on the porous iron scaffolds, a live/dead

staining assay was performed, followed by SEM imaging. The MC3T3-E1preosteoblasts(2× 105_cells_per_specimen)_were_seeded on theiron scaffolds (2.05 mm inheight and9.75 mm in diam-eter) and cultured in 2 mL

α

-MEM (without ascorbic acid, but with10%FBS,1%p/s).After3daysofculture,thespecimenswere washedwithPBSandimmersedinPBScontaining2μL/mLof cal-cein (Thermo Fisher Scientific, USA) and 1.5 μL/mL of ethidium homodimer-1 (Thermo Fisher Scientific, USA) for 30 min in the darkat room temperature. Thereafter, the live anddead cellson theiron scaffoldswere observed usingafluorescence microscope (ZOEcellimager, Bio-Rad,USA).Furthermore,SEMimagingofcell morphologies on the iron scaffolds was performed. Briefly, after 3days ofculture,the specimenswere washed inPBS, fixedwith 4% formaldehyde (Sigma Aldrich, Germany) for20 min, followed bydehydrationstagesin30,50,70, and100%ethanolfor10min each,andtheywerefurtherpreservedusinghexamethyldisilazane (Sigma Aldrich,Germany) for30 min. The specimenswere dried for2hpriortoSEMimaging.

2.9. Statisticalanalysis

Allvaluesareexpressedasmean± standarddeviation.The sta-tisticalanalysisofthePrestoBlueresultswasperformedwith

(8)

two-Fig. 3. In vitro corrosion characteristics of the porous iron scaffolds: (a) visual observation of the scaffolds before and after 28 days of immersion, (b) XRD pattern of the corrosion products on the iron scaffolds at day 28, (c) pH values of the r-SBF solution, (d) the Ca, PO 4 , and Fe ion concentrations in the r-SBF throughout the immersion

period, (e) the mass loss percentages, and (f) the corrosion rate of the scaffolds.

Table 1

Characteristics of the porous iron scaffolds made by extrusion-based 3D printing.

Sample group Strut width (μm) Strut spacing (μm) Layer thickness (μm) Absolute porosity (%) Interconnected porosity (%)

Design 410 400 320 50 –

As-printed iron 420.3 ± 5.5 390.6 ± 5.6 327.4 ± 4.4 47 ± 4 –

As-sintered iron 411.2 ± 5.9 398.9 ± 5.7 320.2 ± 3.7 67 ± 2 65 ± 1

way ANOVA followed by the Tukey post hoc test. For the direct cytocompatibility cell count results, statistical analysis was con-ductedwithone-wayANOVA,followedbytheTukeyposthoctest. 3. Results

3.1. Structuralcharacteristicsoftheporousironscaffolds

Theporousironscaffoldsexhibitedafree-standing characteris-tic,wherethestrutsbridgedabovetheunderlyinglayers(Fig.2a). Suﬃcientpowderloadingintheink(Fig.2b) andstrongbond be-tween powderparticles andbinder(Fig.2c) allowedthescaffolds withahighaspectratiotobebuiltandtostayintactwithout dis-tortions orshrinkage after3D printing. The as-printedspecimens hadanabsoluteporosity of47 ± 4%,withanaveragestrut width of420.3± 5.5μm,andastrutspacingof390.6± 5.6μm(Table1). After sintering (Fig. 2d-e), the dimensions of the specimens were slightlyreduced by2.3 ± 0.2%(height) and2.5± 0.2% (di-ameter). The strut width changed to 411.2 ± 5.9 μm while the

strutspacingbecame398.9_{± 5.7}μm(Table1).Attheperipheryof thescaffolds,thefusionofiron particleswithevidentopen pores between the necks of iron powder particles could be observed (Fig. 2f). On thecross section, a partially sintered microstructure withpores rangingbetween26and135μminsidethestrutswas observed(Fig.2g).Despitethepresenceofmicropores,theaverage solid fractionof thestruts washigh(i.e., 89_{± 4%).} Theabsolute porosity ofthe as-sintered specimenswas67± 2% andthe total interconnectedporositywas65± 1%, correspondingtoapore in-terconnectivityof96%.BasedontheXRDanalysis,theas-sintered iron scaffolds only contained the

α

-iron phase (Fig. 2h) without anydetectablebinderresidues.

3.2. Invitrocorrosionbehaviorandthecorrosionproducts

After28daysofstaticinvitroimmersion,athick brown corro-sionlayeralmostentirelycoveredtheperipheryofthespecimens (Fig.3a).Thecorrosionproductsweremostlycomposedofiron ox-ide hydroxide(

γ

-FeOOH), magnetite (Fe3O4), andiron phosphate

(9)

Fig. 4. Morphology and chemical compositions of the in vitro corrosion products at the periphery of the porous iron scaffolds: SEM and EDS point analysis of the corrosion products after (a, d, g) 7 days, (b, e, h) 14 days, and (c, f, i) 28 days of immersion. The arrow and number indicate where the EDS point analysis was performed and the corresponding elemental composition, respectively.

(FePO4)(Fig.3b).Theimmersiontestsperformedwith5%CO2and 2%O2slightlyincreasedthepHofther-SBFsolutionto7.62atday 28(Fig.3c).

Throughouttheinvitroimmersiontests,thereleaseofironions to the r-SBF solution remained _< 1 ppm (Fig. 3d). The concen-tration ofiron ions increasedfrom0.15 ppm to0.64 ppm inthe ﬁrst14daysofimmersionandthendecreasedto0.08ppm(atday 28). On theother hand, theconcentrations ofcalcium and phos-phate ionsinther-SBFsolutioncontinuously decreasedovertime (Fig.3d).Amuchhigherrateofreductionintheconcentration of thephosphateioninthesolutionwasobservedintheﬁrst14days of immersion (Fig. 3d), while a slightly higher precipitation rate of calcium ions on the specimens wasnoticed betweendays 14 and 28. The mass of the iron scaffolds was reduced by 7 ± 1% after 28 days of immersion (Fig. 3e). The average in vitro corro-sion rateafter1 dayofimmersion was0.28± 0.05 mg/cm2_/day, but declined to 0.11 ± 0.01 mg/cm2_/day _after ₂₈ _days. Accord-ing to the ASTM standard G31–72 [53], the average corrosion rates after 1 day and 28 days of immersion could be con-verted into 0.13 ± 0.03 and 0.05 ± 0.01 mm/year, respectively (Fig.3f).

After7daysofimmersion,loosecorrosionproductssurrounded the strutsofthescaffolds(Fig.4a).Finegranules ofthe corrosion products, which were rich in iron, oxygen, and carbon and con-tainedsodium,calcium,phosphorus,andchlorine,wereidentiﬁed (Fig.4d andg). After 14 daysof immersion,the thicknessof the corrosion layerincreasedandalmostﬁlled themacroporesofthe

ironscaffolds(Fig.4b).Themorphologiesofthecorrosionproducts consistedofamixtureoffineandcoarse granules(Fig.4e)anda flake-likestructure(Fig.4h).Thecompositionofthefinergranules wassimilartotheoneseenafter7daysofimmersion(Fig.4e,EDS 1),whereasthecoarsergranulesmainlycontainedcalcium,oxygen, andcarbon withtrace amounts ofiron, sodium, andmagnesium (Fig.4e,EDS2).Intheflake-likecorrosionproducts,ironand oxy-genwere dominant,butcarbonandsodiumexisted too(Fig. 4h). After28daysofimmersion,thecorrosionproductsdevelopedinto a more compact structure (Fig. 4c, f). The dense corrosion layer wascomposed ofiron,carbon,andoxygen (Fig.4i,EDS1)oriron andoxygen(Fig.4i,EDS2).

Inaddition tothe periphery,the corrosionproducts were also formed in the interconnected pore network of the struts of the scaffolds(Fig.5a-c).Thecorrosionproductsinthemicroporeswere predominantly composed of oxygen with traces of iron, calcium, and carbon (Fig. 5d-e). Moreover, two other types of corrosion products were identiﬁed in the center of the scaffolds after 28 daysofimmersion(Fig.5f-g).Onetype, onthestrut surface,had a ﬂake-like morphology and contained iron, oxygen, and carbon (Fig. 5g, EDS1). The other type hada crystal-like spherical mor-phologyandcontainediron,oxygen, carbon,andcalcium(Fig.5g, EDS 2). As the biodegradation progressed, the pure iron present inthescaffoldswasgraduallyconsumed.Thepureironfractionin thestrutsreducedfrom89_{± 4%}beforethestartoftheimmersion testto87± 2%,79± 2%,and79± 2%after7,14,and28daysof immersion,respectively(Fig.5h).

(10)

Fig. 5. Morphology and chemical compositions of the in vitro corrosion products in the center of the porous iron scaffolds: the cross section of the scaffolds after (a) 7, (b) 14, and (c) 28 days of immersion, (d) a magniﬁed view of the corrosion products at day 28 in the pore network and (e) the elemental mapping, (f) corrosion products on the struts after 28 days of immersion and (g) the enlarged view of the corrosion products with EDS point analysis, and (h) reductions in the pure iron fraction in the struts during the immersion period. The arrow and number indicate where the EDS point analysis was performed and the corresponding elemental composition, respectively. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

3.3. Electrochemicalresponse

Accordingto thePDPcurves(FigureS2a),thecorrosion poten-tial of theiron scaffolds shifted from -781_{± 22} mVafter 1 day ofimmersionto-676± 17mVafter28daysofimmersion.Atthe same time points,the corrosioncurrentdensities decreased from 8.0 ± 0.3μA/cm2 _to_3.5_{± 0.6} _μA/cm2 _(Figure _S2a)._According_to theASTMstandardG102–89[54],thecorrosionrate,basedonthe currentdensities,werecalculatedtobe0.09and0.04mm/year af-ter1dayand28daysofimmersion,respectively.Theresultswere quiteclosetothevaluesobtainedfromtheinvitroimmersiontests inthecellculture incubator.Moreover,thepolarizationresistance oftheiron scaffoldswas3.3± 0.6 k

.cm2 _after₁_day_of immer-sion. It increased to 8.4 _{± 1.4} k

.cm2 _at _day₇ _and_reduced _to 4.9 ± 0.7k

.cm2 _at _day₉ _of_immersion _(Figure_S2b). _From _day 9, the value continuously increasedwith an average rateof 0.35 k

.cm2 _per_day_until_day_21,_became_9.1_{± 2.1}_k

_.cm2_._At_day₂₈ ofimmersion,thepolarizationresistancewas26.7± 6.5k

.cm2_.

The Bode plots of the impedance against frequency (Fig. 6a, S2c)generallyshowedincreasesintheimpedancemodulusforlow

(e.g., 0.01 Hz) and medium (e.g., 10 Hz) frequencies for the en-tire duration of the immersion tests. The Bode plots impedance at0.01 Hzafter1day and28daysofimmersion were 2.8± 0.6 k

.cm2 _and_33.1_{± 2.4}_k

_.cm2_,_{respectively.}_At_these_time_points, theBodeimpedanceat10Hzwere0.3± 0.1k

.cm2_and_11.4_{± 1.4} k

.cm2_,_{respectively.}_In _the_Bode _plot_of_phase_angle_against fre-quency(Fig.6b), thephase anglevaluesathighfrequencies (e.g., 10kHz)decreasedfrom−1.3° after1dayto −53.9° after28days of immersion.In the case oflow frequencies (e.g.,0.1 Hz), how-ever,thephaseanglevaluesmoved towardsmorepositivevalues, showingashiftinthepeakoftheplotfromahigherfrequencyto alowerfrequencythroughouttheimmersiontime(Fig.6b).

The Nyquist plots displayed three different types of spectra throughouttheimmersiontest.TheNyquistplotafter1dayof im-mersion showed a singlecapacitive arc (Fig. 6c). After 7days of immersion, the Nyquist plot showed a larger curve than that of the 1-dayimmersion plot (Fig. 6c). From 9 until 24 days of im-mersion, theNyquistplots exhibited atwo-semicircle continuous arcthatbeganwithasmallersemicircleatahigherfrequencyand continuedwithaninclinedslopeformingasecondlargercurveat

(11)

Fig. 6. Electrochemical impedance spectroscopy of the porous iron scaffolds: the Bode plots of (a) impedance modulus and (b) phase angle against frequency for up to 28 days of immersion, Nyquist spectra (c) after 1 and 7 days, (d) 9, 14, 21, and 24 days, and (e) 26 to 28 days of immersion. The arrows in Bode plots indicate the change in the impedance and phase angle values with time for a speciﬁc frequency.

a lower frequency(Fig.6d). Inaddition tothe increasein the di-ameter of thesemicircle arc overtime, the angle ofthe inclined slopedecreasedascorrosionprogressed(Fig.6d).After 28daysof immersion,thesemicircleatthehigh-frequencyregionwaslarger andextendedwithamoderatelyinclinedstraightlineforthelower frequencies(Fig.6e).fail

3.4. Mechanicalproperties

Theuniaxialcompressivestress-straincurvesoftheporousiron scaffolds(Fig.7a)beganwithalinearelasticregion,followedbya trend resemblingstrain-hardeningduring theplastic deformation stage until specimenfailure. The as-sintered ironscaffolds exhib-ited a yield strengthof 7.2± 0.4 MPaandan elastic modulus of 0.6_{± 0.1}GPa.Over theinvitroimmersionperiodof28days,the yield strength of the scaffolds increasedto 10.6 ± 0.2 MPa after 2 days of immersion and slightly reduced to 9.4 ± 0.9 MPa af-ter 28daysof immersion(Fig.7b). Likewise, the elasticmodulus increased to 0.6± 0.1 GPa after 2 daysof immersion and stabi-lizedatanaveragevalueof0.7± 0.2GPaafter28daysof immer-sion(Fig.7b).Ontheotherhand,theultimatestrengthremarkably dropped from96± 6 MPato32± 2MPaafter7daysof immer-sion.Then,itfurtherdecreasedto19± 7MPaafter28daysof im-mersion(Fig.7c).Moreover, theiron scaffoldsbecamelessductile aftertheimmersiontestsandanotabledeclineinstrain atfailure from53± 3%to31± 3%wasobservedafter7daysofimmersion.

Thestrainatfailurefurtherdecreasedto13± 3%after28daysof immersion(Fig.7c).

3.5. InvitrocytotoxicityagainstMC3T3-E1preosteoblasts

The 100%iron extract medium contained120.4 ± 3.7ppm of ironions.Theexposuretothe100%ironextractmedium resulted ininstantcytotoxicity,with25.6_{± 4.8%}metabolicactivity remain-ingafter24handalmostnometabolicallyactivecellsafter7days ofculture(Fig.8a).Aftertheironextractwasdilutedto75%ofits original,ametabolicactivityof73.3± 9.3%wasdetectedafter24h culture.However,thetrendofdecreasedactivitywithculturetime (i.e.,33.3± 5.9%after7days)remainedunchanged.Whentheiron extract medium wasfurther diluted to≤ 50%,the growthof the preosteoblastswasonlyslightlyinhibited,asindicatedbythehigh metabolicactivity(>80%)evenafter7daysofculture(Fig.8a).

Fromthestainedsamples,thecellcountsfrom100%,75%,and 50%ironextracts were133_{± 16}(Fig.8c),190_{± 63}(Fig.8d), and 753 ± 59 (Fig. 8e),respectively. More cells were observed when theywere culturedinalower-concentrationironextract medium. Furthermore,theindividual stressﬁbersandﬁlopodiaofeach cell cultured in the 100% and 75% iron extract remained recogniz-able(Fig.8c-d).The preosteoblastsinthe 75%iron extract exhib-iteda polygonalshapeandweremoreoutspread (2580.5± 247.5 μm2 _per _cell) _than _the _cells _grown _in _the _100% _iron _extract (1736.4± 209.9μm2_per_cell).

(12)

Fig. 7. Mechanical properties of the porous iron scaffolds: (a) compressive stress-strain curves, (b) the yield strength and elastic modulus, and (c) the ultimate strength and strain at failure of the scaffolds before and after in vitro immersion for up to 28 days.

The direct seeding of the preosteoblasts on the porous iron scaffoldsresultedinareductioninthelivingcellcountafter24h of culture (i.e.,2.8 ± 0.3(× 105₎ _cells) ₍_Fig. ₈_b)._The _viable _cells continuouslydeclinedto5.6± 1.5(× 104₎_cells,_after₃_days_of cul-ture(p< 0.001).Interestingly,atday7ofculture, theviablecell countslightlyincreasedto6.9± 2.0(× 104₎_cells_{(statistically}_not significant).Although itwaschallengingtoquantifythe live/dead staining results, a number of viable (green) preosteoblasts were detectable onthe iron scaffolds(Fig.8f-h). Moreimportantly,the stainedpreosteoblastsshowedahomogenousdistributionoverthe strutsoftheporousironscaffolds(Fig.8f-h),indicatingagoodcell seeding efficiency in the direct assays. Furthermore, on the SEM images,thesurvivingcellswereidentifiedwithextendedfilopodia intheporous structureofthescaffolds(Fig.8i) aswell ason the corrosionproducts(Fig.8j-k).

4. Discussion

In thisstudy,anextrusion-based 3D printingprocessfollowed bydebindingandsinteringwasdevelopedtofabricateporousiron scaffolds withalay-downpatternforbonesubstitution.Sincethe biodegradation rate of bulk iron is generally too low [14,15], a geometrically ordered porous design was employed to speed up therateofbiodegradationofthespecimenswhilemaintainingthe other importantpropertiesofporous iron,includingits structural integrity and bone-mimicking mechanical properties. Our results showed that the lay-down pattern design, indeed, enhanced the corrosionrateoftheironscaffolds,whilethebone-mimicking me-chanicalpropertieswereretained.Asforthecytocompatibility,the results obtained here were comparable to those reported in the otherstudiesofporousiron[32,40,41].

4.1. Extrusion-based3Dprintingandpost-processing

The developed fabrication process delivered porous pure iron scaffolds with a strut size and a strut spacing close to the de-signvalues(Table1).Theuseofanoptimizedinkformulationand ink synthesis process wasofgreat importance forachievingsuch

promisingresults.Thechoiceofhypromelloseasthebinderis be-causeofitsstraightforwardpreparation,biocompatibility[57],and its suitable rheologicalproperties forextrusion. Theiron ink was preparedwith49vol%powderloadingtostrikeabalancebetween the printability(Figure S1a-c) andthe self-holding characteristics ofthestruts. Theoptimizedpowder-to-binderratiodemonstrated shear-thinning ﬂow behavior (Table S1) and ensured sustainable printingwithout clogging in the nozzletip and enabledthe fab-ricationoftheporousironscaffoldswithahighaspectratio.

Following the extrusion-based 3D printing, heat treatment of theporousironscaffoldswasperformedaspost-processing,which startedwithadebindingstage wherethetemperaturewassetat apoint wherethermaldecompositionofhypromellosetook place (FigureS1d).Afterwards,theheattreatmentcontinuedtoa solid-state sinteringstage ata highertemperature, where iron powder particlesbondedtogetherandformednecks.Theparametersofthe sinteringprocess(i.e.,temperatureandholdingtime)werechosen basedontheobtainedsolidfractionsofthestrutsofthescaffolds undervarious conditions (Figure S1e).The chosen sintering tem-perature(1200°C)andtime(6h)resultedintheporousiron scaf-foldswithan89± 4%solidfractionofthestrutsandanabsolute porosity of 67_{± 2%.} It should be mentioned that a higher solid fractioncanbeachievedandwillleadtoahigherstrengthofthe scaffoldsasaresultofenhanceddensiﬁcation[58].However,inthe caseof thescaffolds intended for useas biodegradable implants, partialsinteringmayofferanadvantage,becausethescaffoldswill have moreexposed powder particle boundary area in thestruts. Under the present sintering condition, the porous iron scaffolds possessedaporeinterconnectivityof96%(Fig.2).Itiswellknown thatporousbone substituteswithinterconnectedmacro- and mi-croporesofferimprovedbiofunctionalitybyfavoringtheadhesion, growth,anddifferentiationofcellsduringboneregeneration [59]. Duetoapartialsolid-statesinteringprocess,theshrinkageofthe ironscaffoldswassmall(i.e.,<2.6%).Furthermore,underthe cho-sensinteringcondition, theironscaffolds wereonly composedof the

α

-iron single phase (Fig. 2h), conﬁrmingthe absence ofany carbonresiduesfromthedecomposedpolymerthatcouldhave dif-fusedintothebasematerial.

(13)

Fig. 8. In vitro biological evaluation of the iron scaffolds towards preosteoblasts MC3T3-E1: (a) indirect metabolic activity of preosteoblasts cultured in iron extracts for 1, 3, and 7 days, (b) the number of living preosteoblasts after 1, 3, and 7 days of direct culture on the porous iron scaffolds, (c-e) rhodamine phalloidin (red) and DAPI (blue) stained preosteoblasts after 3 days of culture in (c) 100%, (d) 75%, and (e) 50% iron extracts, (f-h) calcein acetoxymethyl (green, indicating living cells) and ethidium homodimer-1 (red, indicating dead cells) stained preosteoblasts on the scaffolds, and (i-k) the morphology of the cells after 3 days of direct cell culture on the iron scaffolds.

∗∗∗_{, p < 0.001;}∗∗∗∗_{, p < 0.0001. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)}

4.2. Invitrocorrosioncharacteristicsandmechanisms

The static in vitro corrosion rate of porous iron scaffolds (0.11± 0.01mg/cm2_/day_at_day₂₈₎₍_Fig.₃_f),_is_much_higher_than thatofcastpureiron(i.e.,0.04– 0.06mg/cm2_/day)_[60]_and_that_of rapidlyprototypedporousiron (i.e.,0.04– 0.08mg/cm2_/day)_[32]_, andelectrodeposited iron foam(i.e., 0.023 mg/cm2_/day) _[61]_._The highercorrosion rateofthescaffoldswasattributedto thestruts that containedsinteredpowderparticleboundarieswitha

micro-porenetwork.Therandommicroporousinterconnectivityprovides manyintricate sitesthat are favorableforautocatalytic corrosion

[62–64],therebyresultingintheenhanceddissolutionofiron. Fur-thermore,corrosionoccurrednotonlyattheperiphery(Fig.4)but alsointhestrutsandthecenterofthescaffolds(Fig.5).The pen-etrationofthe r-SBFsolutioninto thecenterof thescaffolds un-der static immersion tests can be attributed to the capillary ac-tionintheporenetwork[65],therebyenhancingthecorrosionof thescaffoldasa whole.Nevertheless,thestaticin vitrocorrosion

(14)

rate(0.05± 0.01mm/yearatday28)isstill lower thanthepure iron withareﬁned grainstructure, manufacturedby electroform-ing(i.e.,0.40mm/year)[66],cross-rolling(i.e.,0.11– 0.14mm/year)

[67], and selective laser melting (i.e., 0.085 mm/year) [68], and thusrequiresanimprovementtoachievean expected biodegrada-tionrateofanidealbonesubstitute(i.e.,0.2to0.5mm/year)[69]. Whenironisexposedtoacorrosiveﬂuid,thecorrosionis gov-ernedbyananodicreaction,followedbythereductionofthe dis-solvedoxygenandwaterintohydroxideions.Asthecorrosion pro-ceeds, the pH ofthe medium increases, andiron (II, III) hydrox-ides form via thereaction ofiron and hydroxide ions.Since iron (III) hydroxide is lessstable, it will transforminto a more stable compound,i.e.,ironoxidehydroxide.Moreover,duetothehypoxia condition in the incubator, the iron (II) hydroxide undergoes an anaerobicSchikorrreactiontoformmagnetite.Also,thereleaseof iron ionsincreasethesurroundingpHlevelandpromotethe pre-cipitationofphosphateandbicarbonateionsfromther-SBF solu-tion.Theabovementionedcorrosionreactions[68]explainthe for-mation of iron oxide (hydroxide) and iron phosphate at the pe-riphery(Figs.3b and4), aswell asthe presenceofiron, calcium, carbon,andoxygeninthecenterofthescaffolds(Fig.5)thatwere identiﬁedafter28daysofinvitroimmersion.

Ironionsinther-SBFwerehardlydetectedbytheICP-OES anal-ysis (Fig. 3d), dueto the participationof iron ions inthe forma-tion ofiron-basedcorrosionproducts thatare largelyinsoluble in a physiological condition[40].The releasetrendof iron ionswas upwardintheﬁrst14daysofimmersion,correspondingtothe el-evationof thesolutionpH valueto 7.57.Duringthe sameperiod, a faster reduction of phosphate ions in the r-SBF was observed, as compared to the reduction of calcium ions. Besides the reac-tion with iron ions, the early period of calcium and phosphate ion reduction could also be related to the precipitation of cal-cium phosphate-based corrosion products (Fig. 4). Between days 14 and28, a declining trend of iron ion release was noted. The slight increase in the pH level after 14 days of immersion may have acceleratedthe precipitationof iron phosphateand carbon-atecompounds,whichexplainsthedecreasingtrendsofironions andmoreconsumptionofcalciumionsinthemediumthaninthe ﬁrst 14days(Fig.3d). Theprecipitationofiron/calciumcarbonate releaseshydrogenionsthatbalancesthealkalinityofthesolution. Inaddition,withtheinvitrocorrosionoccurringina5%CO2 envi-ronment,thepHofther-SBFwasalsoeffectivelymaintained.The pH value slightly increasedto 7.62 at the end ofthe immersion tests.

Fromtheelectrochemicalperspective,thecorrosionmechanism and kinetics ofthe porous iron scaffolds evolvedduring the im-mersionperiod(Fig.6).TheBodeimpedancemodulusvaluesatthe lowfrequency(i.e.,0.01Hz)indicatesthechargetransferresistance of the material during biodegradation and at medium frequency (i.e.,10Hz) is relatedtothe corrosionproduct formation[70,71]. Theincreasingtrendofimpedancemodulusvaluesatthelowand mediumfrequencies(FigureS2c)suggeststhecontinuousbuild-up ofcorrosionproductsontheironscaffoldsthroughoutthe immer-sion tests[20,72]. In addition, the Bode plot of the phase angle values isindicative ofthe corrosionsusceptibilityof thematerial aswell. Aphase anglevalue atahighfrequency(e.g.,at10 kHz) close to 0°, which was observed in the specimen after 1 day of biodegradation (Fig.6b),indicates that thematerialis susceptible to corrosion [73,74]. Onthe contrary, thephase angle shifts toa more negative value when the material exhibits more corrosion resistance,asobserved forthe specimenssubjectedto28 daysof biodegradation[73,74].

BasedontheNyquistplotafter1dayofimmersion(Fig.6c),a singlearcwasobserved,implyingtheintialactivecorrosionofiron withacapacitivebehavior[64].Thediameterofthecapacitivearc indicates thecorrosionrateofthematerial.Thegreaterthe

diam-etersize,theslowerthecorrosionrate[75].Asthe corrosiontest progressed to day 7,the diameter ofthe arc increased, showing thatevenattheinitialstage, thecorrosionproductshinderedthe progressofthecorrosionprocess.Thereafter, betweendays7and 9, the Nyquistspectra transformed into a two-semicircle pattern (Fig. 6d). The pattern remaineduntil 24 days of immersion. The two-semicirculararcspresentintheNyquistspectrahavebeen re-portedas an indicationof a corrosionmechanism that combines active anddiffusive processes[76]. Fromthe observations after7 and14 daysofimmersion,corrosionproductsaccumulatedatthe periphery(Fig.4),inthecenter(Fig.5),andinthemicropore net-workofthespecimens.Theactivecorrosionprocessmostprobably occurredonanyavailableironsurface,whilethediffusionactionis expectedtohavetakenplacethroughthecorrosionproduct layer that wasstill relativelyloose anddispersed.As thecorrosion ad-vanced fromday 9 to day 24(Fig. 6d), the diameterof the arcs grewlargeralongwithatendencytowardsamorediffusive corro-sionmechanism, assuggestedby an increasein thearcdiameter anda decreasedincline slope forthelower frequencies [77]. Ap-proaching day28,when nearly all of the corrosionproducts had beenturned intoadense layer,thecorrosionmechanismevolved into a solely diffusion-controlled mechanism (Fig. 6e), which is hallmarked by a straight linein the Nyquist spectra in the low-frequencyregion[64].

Fromtheinvitroimmersiontestsandtheelectrochemical anal-ysis of the porous iron scaffolds, it is clear that the formation ofthe corrosion products hinderedthe direct dissolutionof iron, hencereducingthecorrosionrateovertime. Theevolutionofthe corrosionproductsfromalooselayertoadenselystructuredlayer duringthein vitroimmersionperiodis observed.Thismakes the long-termcorrosion mechanism of porous iron scaffolds strongly dependent on thediffusion process [64].An improvementin the corrosionrateofporousironscaffoldscanbeachievedby perform-ingdynamicinvitroimmersiontests[41].Althoughwiththe con-tinuousliquidﬂowduringinvitroimmersion,substantialamounts ofcorrosionproductsmaystilladheretothescaffoldsafter28days

[41].In an in vivostudyon iron-based stents,a slowercorrosion ratethantheoneestimatedfromtheinvitrotestswasfound,due to the formationof insoluble iron-based corrosion products [78]. Therefore,thestrategiestomodifythecorrosionproducts (e.g.,to makethemlessadhering) areofgreatinterestforthefurther de-velopmentofporousironscaffoldsasbonesubstitutes.

4.3. Bone-mimickingmechanicalproperties

Forad interimmetallic bone scaffolds, themechanical proper-tiesmust be maintainedfor a certain perioduntil thenew bone tissuecan take overthe mechanicalroleof theimplant andbear physiologicalloading.Therefore,thecontrolovertheinvivo degra-dationrateofthematerialisofgreatimportancetobetterestimate therateofbiodegradationandavoidasuddenlossofthestructural integrity ofthescaffold.The mechanicalpropertiesofthe porous iron scaffolds made by extrusion-based 3D printing remained in the range of those of trabecular bone (i.e., E = 0.5–20 GPa and

σy

= 0.2 – 80 MPa) despite 28 days of in vitro biodegradation (Fig.7)[79,80].

Fromthestress-straincurves,theas-sinteredporousiron scaf-folds (Fig. 7a) exhibited the typical mechanical characteristics of a ductile material.Unlike mostgeometrically ordered 3D printed porous metallic scaffolds [81,82], the stress-strain curves of the iron scaffolds did not exhibit a plateau stage. The stress-strain curves of the iron scaffolds with a 0° to 90° lay-down pattern were found to be quite similar to those of the 3D printed scaf-folds designed using the solid Schwartz p-unit cells, which also have90° interconnectionsbetweentheirstruts[83,84].Inaddition tothepattern,theporenetworkinthestrutsofthescaffoldsacted

(15)

tweenironanditscorrosionproductsmightaffectthemechanical properties too [87]. Since the elastic modulus andyield strength were measured over asmall rangeofstrainsduringinitial defor-mation,whereloadingcouldstillbetransferredacrossironpowder particle necks andthe interfacesbetweeniron andthecorrosion products, theformationofthecorrosionproductsled toincreases in theyield strength andelastic modulus after2days of immer-sion. Afterwards,thevaluesfluctuated throughoutthe28days of immersionbutremainedhigherthanthoseoftheinitialscaffolds. Theslightincreaseintheyieldstrengthandelasticmoduluscould be explainedby considering thecorrosionproducts asa reinforc-ing phase in theporous iron matrixthat provided a strengthen-ing effect.The addition of iron oxide (i.e., 2 and5 wt%)has, in-deed, beenreportedtoimprove theelasticmechanical properties oftheiron-ironoxidecomposite[88].Thefluctuationsintheyield strengthandelasticmodulusovertheimmersiontimemaybedue to the evolution ofthe corrosion product morphology that influ-encedtheir interfacialbondingwiththe specimens.Another rele-vantfactorcouldbethedecreasingfractionofironinthescaffolds as corrosion progressed. The latter would significantly affect the ductilityofthescaffolds(Fig.7c).

4.4. Cytocompatibilitycharacteristics

Differentlevelsofpreosteoblastviability(Fig.8a)wereobserved depending on the dilutionlevel ofthe iron extract. According to the ISO 10993-5 [89], the 100% iron extract was severely toxic (level4)withnearlynometabolicallyactivecells.The75%iron ex-tractwascategorizedasmoderatelytoxic(level3)withmorethan 50%growthinhibition,whiletheironextracts≤ 50%wereslightly toxic (level 1) with more than 80% metabolic activityeven after 7 days of culture. Moreover, the rhodamine-phalloidin and DAPI staining(Fig.8c-e)clearlyshowedtheunfavorableeffectsof undi-lutedextractsonthegrowthofthepreosteoblasts.Thenumberof cells increased in the diluted extracts contributingto the higher metabolicactivitydetected.Thecytotoxicityofironisoftenrelated totheuncontrolledformationofROS,includingsuperoxideanions and hydroxyl radicals via the Fenton reactions [90]. Although a properROSlevelplaysanimportantpartinsomebiologicalevents, such astheactivationofsignalingpathwaysandgeneexpression, excessiveROSmayleadtooxidativestresses,whichwillharmthe cells[90].

Ofnote,theiron extractswerepreparedwithanextraction ra-tio of5 cm2_/mL,_which _is_higher _than _the_most_reported extrac-tion ratiosinthein vitrostudies ofiron-basedmaterials (e.g.,2.5 cm2_/mL _[32] _and_1.25_cm2_/mL _[₄₃_,₆₈_,₉₁_])._In _addition_to _the rel-atively high solution volume-to-surface area ratio,the 72 h iron extractioninthecellculturemediumresultedinahigherironion concentration ascomparedtotheextractfromther-SBFmedium, whichismostlikelyduetothepresenceofserum[92,93].Ahigher extractionratiocombinedwithahighironionreleaseofthe scaf-folds createdinthisstudyexplainsthe immediatecytotoxicity of the 100% iron extract that was observed. A safe iron ion con-centration for bone marrow stem cells has been reported to be < 75 mg/L [13], which is similar to the results obtained in our

froman implantistoosmalltocausesystemictoxicity[97].Local tissuetoxicityanalysisis,therefore,morerelevantforsuchporous iron.

The presentdirectcell seeding tests demonstrateda cytotoxic effectofthe porousiron scaffoldswitha 17-foldreduction of vi-able cellswithin the first 3 days ofculture, and afterwards only growthinhibition. Reductionsin cell viabilitywere also observed inthedirectcultureofmousebonemarrowstemcellson60vol% porous Fe-30Mn scaffolds [31], human osteosarcoma cells on 40 vol%porous Fe-30Mn6Si1Pdscaffolds[29],3T3fibroblasts[32]as well as rabbitbone marrow stem cells[35] on porous iron scaf-folds. Interestingly, despite significant reductions in the number of viable cells, the preosteoblasts were still found stretching in the pores (Fig. 8i)and adheringto the corrosionproducts ofthe porousironscaffolds(Fig.8j-k).The spreadingmorphologyofthe preosteoblasts may be due to the combined effect of the sur-face morphology and the presence of calcium- and phosphate-based corrosion products that are known to be osteoconductive

[14].Sinceinvitrocytocompatibilityassessmentcannotcompletely mimictheinvivoconditions,thesepreliminaryresultswarrantin vivo studies on 3D printed porous iron scaffolds, particularly to study the local cyto- and histo-compatibility of the porous iron scaffolds.

5. Conclusions

Inthisstudy, theiron scaffoldswitha lay-downpattern were successfully fabricated by using extrusion-based 3D printing and subsequentsintering,whichallowedforcomprehensive character-ization of the material for application as a biodegradable bone substitute. The structure of the fabricated porous iron scaffolds washighly interconnected,owing to the presence of macropores fromthepatterndesignandrandommicroporesinthestruts. Af-ter 28 days of staticin vitro immersion,the mass of the porous iron reduced by 7%. The in vitro corrosion rate decreased from 0.28 down to 0.11 mg/cm2_/day, _with _different _corrosion mecha-nisms operatingover the 28days of static immersion. The yield strength andelasticmodulus oftheporous iron scaffoldsslightly increasedduetotheformationofcorrosionproductsinthestruts during the 28 days of in vitro corrosion. These values remained withinthe rangeofthe mechanicalpropertiesoftrabecular bone. The direct culture of preosteoblasts on the iron scaffolds re-vealed cytotoxicity, due to the high concentration of iron ions, asexplainedbytheresultsfromtheextraction-basedassays. Fur-ther in vitrocytotoxicity experiments(e.g., co-cultureof multiple types ofcells andin vivo studies)should be performed[84] un-der dynamic conditions. Taken together, our results showedthat extrusion-based 3D printing could deliver porous iron scaffolds with enhanced biodegradability and bone-mimicking mechanical propertiesfor potential applicationasbone substitutes. Introduc-ingbioactivecomponentsintheformofcoatings[15]orinsidethe scaffoldbodytoformcomposites [36,69]maybeexplored to fur-therimprovethebiofunctionalitiesofsuch3Dprintedporousiron scaffolds.

(16)

DeclarationofCompetingInterest

Theauthorsdeclarethattheyhavenoknowncompeting ﬁnan-cialinterestsorpersonalrelationshipsthatcouldhaveappearedto inﬂuencetheworkreportedinthispaper.

Acknowledgments

This work is partof the3DMed project that hasreceived the fundingfromtheInterreg2Seasprogram2014– 2020,co-funded by the European Regional Development Fund undersubsidy con-tractno. 2S04-014.Mr. RuudHendrikx attheDepartmentof Ma-terials Science andEngineering,Delft University ofTechnology,is acknowledged for theXRD analysis.Mr. Michel vandenBrink at theDepartmentofProcessandEnergy,DelftUniversityof Technol-ogy, isacknowledged fortheICP-OES analysis.The authors thank Ms. Agnieszka Kooijman at the Department of Materials Science andEngineering,Delft UniversityofTechnology fortheassistance intheelectrochemicalexperiments.

Supplementarymaterials

Supplementary material associated with this article can be found,intheonlineversion,atdoi:10.1016/j.actbio.2020.11.022. References

[1] H. Hermawan, Updates on the research and development of absorbable metals for biomedical applications, Prog. Biomater. 7 (2018) 93–110, doi: 10.1007/ s40204- 018- 0091- 4 .

[2] H.S. Han, S. Loffredo, I. Jun, J. Edwards, Y.C. Kim, H.K. Seok, F. Witte, D. Man- tovani, S. Glyn-Jones, Current status and outlook on the clinical translation of biodegradable metals, Mater. Today. 23 (2019) 57–71, doi: 10.1016/j.mattod. 2018.05.018 .

[3] Y.F. Zheng, X.N. Gu, F. Witte, Biodegradable metals, Mater. Sci. Eng. R. 77 (2014) 1–34, doi: 10.1007/978- 1- 4614- 3942- 4 _ 5 .

[4] M. Schinhammer, A.C. Hänzi, J.F. Löﬄer, P.J. Uggowitzer, Design strategy for biodegradable Fe-based alloys for medical applications, Acta Biomater. 6 (2010) 1705–1713, doi: 10.1016/j.actbio.2009.07.039 .

[5] J. He, F.L. He, D.W. Li, Y.L. Liu, Y.Y. Liu, Y.J. Ye, D.C. Yin, Advances in Fe- based biodegradable metallic materials, RSC Adv. 6 (2016) 112819–112838, doi: 10.1039/C6RA20594A .

[6] R. Gorejová, L. Haverová, R. Ori ˇnaková, A. Ori ˇnak, M. Ori ˇnak, Recent advance- ments in Fe-based biodegradable materials for bone repair, J. Mater. Sci. 54 (2019) 1913–1947, doi: 10.1007/s10853- 018- 3011- z .

[7] N. Abbaspour , R. Hurrell , R. Kelishadi , Review on iron and its importance for human health, J. Res. Med. Sci. (2014) .

[8] E. Balogh, G. Paragh, V. Jeney, Inﬂuence of iron on bone homeostasis, Pharma- ceuticals 11 (2018) 1–18, doi: 10.3390/ph11040107 .

[9] F. Atashi, A. Modarressi, M.S. Pepper, The role of reactive oxygen species in mesenchymal stem cell adipogenic and osteogenic differentiation: a review, Stem Cells Dev. 24 (2015) 1150–1163, doi: 10.1089/scd.2014.0484 .

[10] M. Parelman, B. Stoecker, A. Baker, D. Medeiros, Iron restriction negatively af- fects bone in female rats and mineralization of hFOB osteoblast cells, Exp. Biol. Med. 231 (2006) 378–386, doi: 10.1177/153537020623100403 .

[11] J.G. Messer, P.T. Cooney, D.E. Kipp, Iron chelator deferoxamine alters iron- regulatory genes and proteins and suppresses osteoblast phenotype in fetal rat calvaria cells, Bone 46 (2010) 1408–1415, doi: 10.1016/j.bone.2010.01.376 . [12] D. Praticò, M. Pasin, O.P. Barry, A. Ghiselli, G. Sabatino, L. Iuliano, G.A. Fitzger-

ald, F. Violi, Iron-dependent human platelet activation and hydroxyl radical formation: involvement of protein kinase C, Circulation (1999), doi: 10.1161/01.CIR. 99.24.3118 .

[13] E. Zhang, H. Chen, F. Shen, Biocorrosion properties and blood and cell compatibility of pure iron as a biodegradable biomaterial, J. Mater. Sci. Mater. Med. 21 (2010) 2151–2163, doi: 10.1007/s10856-010-4070-0 .

[14] T. Kraus, F. Moszner, S. Fischerauer, M. Fiedler, E. Martinelli, J. Eichler, F. Witte, E. Willbold, M. Schinhammer, M. Meischel, P.J. Uggowitzer, J.F. Löﬄer, A. Wein- berg, Biodegradable Fe-based alloys for use in osteosynthesis: outcome of an in vivo study after 52 weeks, Acta Biomater 10 (2014) 3346–3353, doi: 10.1016/ j.actbio.2014.04.007 .

[15] S. Ray, U. Thormann, M. Eichelroth, M. Budak, C. Biehl, M. Rupp, U. Som- mer, T. El Khassawna, F.I. Alagboso, M. Kampschulte, M. Rohnke, A. Henß, K. Peppler, V. Linke, P. Quadbeck, A. Voigt, F. Stenger, D. Karl, R. Schnettler, C. Heiss, K.S. Lips, V. Alt, Strontium and bisphosphonate coated iron foam scaffolds for osteoporotic fracture defect healing, Biomaterials 157 (2018) 1–16, doi: 10.1016/j.biomaterials.2017.11.049 .

[16] M.H. Emily Walker, Magnesium, iron and zinc alloys, the trifecta of bioresorbable orthopaedic and vascular implantation - a review, J. Biotechnol. Bio- mater. (2015), doi: 10.4172/2155-952x.10 0 0178 .

[17] Y. Ishikawa, Y. Endoh, Antiferromagnetism of γ-FeMn alloys, J. Appl. Phys. 39 (1968) 1318–1319, doi: 10.1063/1.1656274 .

[18] Y.P. Feng, A. Blanquer, J. Fornell, H. Zhang, P. Solsona, M.D. Baró, S. Suriñach, E. Ibáñez, E. García-Lecina, X. Wei, R. Li, L. Barrios, E. Pellicer, C. Nogues, J. Sort, Novel Fe-Mn-Si-Pd alloys: insights into mechanical, magnetic, corrosion resistance and biocompatibility performances, R, Soc. Chem. 4 (2016) 6402–6412, doi: 10.1039/c6tb01951j .

[19] J. Hufenbach, H. Wendrock, F. Kochta, U. Kühn, A. Gebert, Novel biodegradable Fe-Mn-C-S alloy with superior mechanical and corrosion properties, Mater. Lett. 186 (2017) 330–333, doi: 10.1016/j.matlet.2016.10.037 .

[20] J. Zhou, Y. Yang, M. Alonso Frank, R. Detsch, A.R. Boccaccini, S. Virtanen, Accel- erated degradation behavior and cytocompatibility of pure iron treated with sandblasting, ACS Appl. Mater. Interfaces 8 (2016) 26482–26492, doi: 10.1021/ acsami.6b07068 .

[21] A.H.M. Yusop, N.M. Daud, H. Nur, M.R.A. Kadir, H. Hermawan, Controlling the degradation kinetics of porous iron by poly(lactic-co-glycolic acid) inﬁltration for use as temporary medical implants, Sci. Rep. 5 (2015) 1–17, doi: 10.1038/ srep11194 .

[22] L. Haverová, R. Ori ˇnaková, A. Ori ˇnak, R. Gorejová, M. Baláž, P. Vanýsek, M. Kup- ková, M. Hrubov ˇcáková, P. Mudro ˇn, J. Rado ˇnák, Z.O. Králová, A.M. Turo ˇnová, An in vitro corrosion study of open cell iron structures with PEG coating for bone replacement applications, Metals (Basel) 8 (2018) 1–21, doi: 10.3390/ met8070499 .

[23] R. Gorejová, R. Ori ˇnaková, Z. Orságová Králová, M. Baláž, M. Kupková, M. Hrubov ˇcáková, L. Haverová, M. Džupon , A. Ori ˇnak , F. Ka ľ avský, K. Kova ľ ,In

vitro corrosion behavior of biodegradable iron foams with polymeric coating, Materials (Basel) 13 (2020) 1–17 .

[24] Y. Su, S. Champagne, A. Trenggono, R. Tolouei, D. Mantovani, H. Hermawan, De- velopment and characterization of silver containing calcium phosphate coat- ings on pure iron foam intended for bone scaffold applications, Mater. Des. 148 (2018) 124–134, doi: 10.1016/j.matdes.2018.03.061 .

[25] J. He, H. Ye, Y. Li, J. Fang, Q. Mei, X. Lu, F. Ren, Cancellous-bone-like porous iron scaffold coated with strontium incorporated octacalcium phosphate nanowhiskers for bone regeneration, ACS Biomater. Sci. Eng. 5 (2019) 509–518, doi: 10.1021/acsbiomaterials.8b01188 .

[26] R. Alavi, A. Trenggono, S. Champagne, H. Hermawan, Investigation on mechanical behavior of biodegradable iron foams under different com pression test conditions, Metals (Basel) 7 (2017) 202, doi: 10.3390/met7060202 .

[27] J. ˇCapek, D. Vojt ˇech, Microstructural and mechanical characteristics of porous iron prepared by powder metallurgy, Mater. Sci. Eng. C. 43 (2014) 494–501, doi: 10.1016/j.msec.2014.06.046 .

[28] J. ˇCapek, D. Vojt ˇech, A. Oborná, Microstructural and mechanical properties of biodegradable iron foam prepared by powder metallurgy, Mater. Des. 83 (2015) 46 8–4 82, doi: 10.1016/j.matdes.2015.06.022 .

[29] Y.P. Feng, N. Gaztelumendi, J. Fornell, H.Y. Zhang, P. Solsona, M.D. Baró, S. Suriñach, E. Ibáñez, L. Barrios, E. Pellicer, C. Nogués, J. Sort, Mechani- cal properties, corrosion performance and cell viability studies on newly developed porous Fe-Mn-Si-Pd alloys, J. Alloys Compd. 724 (2017) 1046–1056, doi: 10.1016/j.jallcom.2017.07.112 .

[30] M. Heiden, E. Nauman, L. Stanciu, Bioresorbable Fe–Mn and Fe–Mn–HA m for orthopedic implantation: enhancing degradation through porosity control, Adv. Healthc. Mater. 6 (2017) 1–12, doi: 10.10 02/adhm.20170 0120 .

[31] S.M. Huang, E.A. Nauman, L.A. Stanciu, Investigation of porosity on mechanical properties, degradation and in-vitro cytotoxicity limit of Fe30Mn using space holder technique, Mater. Sci. Eng. C. 99 (2019) 1048–1057, doi: 10.1016/j.msec. 2019.02.055 .

[32] P. Sharma, K.G. Jain, P.M. Pandey, S. Mohanty, In vitro degradation behaviour, cytocompatibility and hemocompatibility of topologically ordered porous iron scaffold prepared using 3D printing and pressureless microwave sintering, Mater. Sci. Eng. C. 106 (2020) 110247, doi: 10.1016/j.msec.2019.110247 . [33] D.T. Chou, D. Wells, D. Hong, B. Lee, H. Kuhn, P.N. Kumta, Novel processing of

iron-manganese alloy-based biomaterials by inkjet 3-D printing, Acta Biomater. 9 (2013) 8593–8603, doi: 10.1016/j.actbio.2013.04.016 .

[34] D. Hong, D.T. Chou, O.I. Velikokhatnyi, A. Roy, B. Lee, I. Swink, I. Issaev, H.A. Kuhn, P.N. Kumta, Binder-jetting 3D printing and alloy development of new biodegradable Fe-Mn-Ca/Mg alloys, Acta Biomater. 45 (2016) 375–386, doi: 10.1016/j.actbio.2016.08.032 .

[35] C. Yang, Z. Huan, X. Wang, C. Wu, J. Chang, 3D printed Fe scaffolds with HA nanocoating for bone regeneration, ACS Biomater. Sci. Eng. 4 (2018) 608–616, doi: 10.1021/acsbiomaterials.7b00885 .

[36] H. Ma, T. Li, Z. Huan, M. Zhang, Z. Yang, J. Wang, J. Chang, C. Wu, 3D printing of high-strength bioscaffolds for the synergistic treatment of bone cancer, NPG Asia Mater 10 (2018) 31–44, doi: 10.1038/s41427- 018- 0015- 8 .

[37] D.K. Mishra, P.M. Pandey, Experimental investigation into the fabrication of green body developed by micro-extrusion-based 3D printing process, Polym. Compos. 41 (2020) 1986–2002, doi: 10.1002/pc.25514 .

[38] D.K. Mishra, P.M. Pandey, Effect of sintering parameters on the microstructure and compressive mechanical properties of porous Fe scaffold fabricated using 3D printing and pressure less microwave sintering, Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 0 (2020) 1–16, doi: 10.1177/0954406220921416 .

[39] D.K. Mishra, P.M. Pandey, Mechanical behaviour of 3D printed ordered pore topological iron scaffold, Mater. Sci. Eng. A. 783 (2020) 139293, doi: 10.1016/j. msea.2020.139293 .