Mohammad Shahin,Cuie Wen,Khurram Munir,Yuncang Li

School of Engineering,RMIT University,Melbourne,Victoria 3001,Australia

Abstract Magnesium (Mg)-based biomaterials have gained acceptability in fracture fixation due to their ability to naturally degrade in the body after fulfillin the desired functions.However,pure Mg not only degrades rapidly in the physiological environment,but also evolves hydrogen gas during degradation.In this study,Mg0.5Zr and Mg0.5ZrxZn (x=1-5 wt.%) matrix nanocomposites (MNCs) reinforced with different contents (0.1-0.5 wt.%) of graphene nanoplatelets (GNP) were manufactured via a powder metallurgy technique and their mechanical and corrosion properties were evaluated.The increase in GNP concentration from 0.2 wt.% to 0.5 wt.% added to Mg0.5Zr matrices resulted in decreases in the compressive yield strength and corrosion resistance in Hanks’ Balanced Salt Solution (HBSS).On the other hand,a higher concentration (4-5 wt.%) of Zn added to Mg0.5Zr0.1GNP resulted in an increase in ductility but a decrease in compressive yield strength.Overall,an addition of 0.1 wt.% GNPs to Mg0.5Zr3Zn matrices gave excellent ultimate compressive strength (387 MPa) and compressive yield strength (219 MPa).Mg0.5Zr1Zn0.1GNP and Mg0.5Zr3Zn0.1GNP nanocomposites exhibited 29% and 34% higher experimental yield strength,respectively,as compared to the theoretical yield strength of Mg0.5Zr0.1GNP calculated by synergistic strengthening mechanisms including the difference in thermal expansion,elastic modulus,and geometry of the particles,grain refinement load transfer,and precipitation of GNPs in the Mg matrices.The corrosion rates of Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP measured using potentiodynamic polarization were 7.5 mm/y,4.1 mm/y,6.1 mm/y,and 8.0 mm/y,respectively.Similarly,hydrogen gas evolution tests also demonstrated that Mg0.5Zr3Zn0.1GNP exhibited a lower corrosion rate (1.5 mm/y) than those of Mg0.5Zr1Zn0.1GNP(3.8 mm/y),Mg0.5Zr4Zn0.1GNP (1.9 mm/y),and Mg0.5Zr5Zn0.1GNP (2.2 mm/y).This study demonstrates the potential of GNPs as effective nano-reinforcement particulates for improving the mechanical and corrosion properties of Mg-Zr-Zn matrices.

Keywords: Biodegradation;Graphene nanoplatelet;Magnesium metal matrix composite;Mg-Zr-Zn alloy;Strengthening mechanism.

1.Introduction

Bone injury and degeneration caused by accidents,sports injuries,or the natural process of aging often require biomaterial implants to restore function [1].The ability of modern medicine to replace failed bone tissue has greatly improved the quality of life for millions of people worldwide.Commonly,titanium (Ti),stainless steel (SS),and cobalt-chromium (Co-Cr)-based alloys are used for the manufacturing of bone screws,bone plates,and bone pins in orthopedic fracture-fixation surgery.However,these metallic materials exhibit significantly higher elastic moduli (110-250 GPa) as compared to the elastic modulus of human cortical bone (10-30 GPa).The stiffer implants thus carry the physiological load and cause bone resorption in the surrounding bone tissues due to the stress-shielding effect,which inhibits bone remodeling and causes osteoporosis and subsequent implant loosening [2-4].In addition to these challenges,conventional permanent metallic implants also trigger foreign body reactions in the body which can cause allergies and also require complex revision surgery to remove the implants upon the restoration of function.In this context,biodegradable metallic implants are recommended in order to eliminate these limitations of such conventional metallic implants.

In recent years,magnesium(Mg)and its alloys have gained great attention because of their excellent biocompatibility and biodegradability properties providing closer elastic moduli(40-45 GPa) to that of cortical bone (10-30 GPa).This can promote bone remodeling by reducing the stress-shielding effect in the surrounding and host bone tissues [5-7].Mg is also an essential trace element in bone that provides necessary structural support to bone tissues [8,9].However,despite its excellent biocompatibility,the inadequate mechanical properties,such as yield strength and ductility,and higher corrosion rate (CR) of pure Mg and its alloys impede their utilization as biodegradable metallic implants [10-12].

To overcome these challenges,pure Mg is generally alloyed with different elements including aluminum(Al),zinc (Zn),calcium (Ca),zirconium (Zr),and various rare earth elements (REE).For example,commercial Mg alloys including WE43 (Mg3.5Y2.3Nd0.5Zr),WE54(Mg4.9Y1.6Nd0.3Zr0.2Er0.1Yb0.1Gd),and AZ series containing Al and Zn such as AZ91 (Mg9Al1Zn),AZ61(Mg6Al1Zn),and AZ31 (Mg3Al1Zn) exhibit adequate mechanical properties.However,low corrosion resistance with high H2gas evolution,severe galvanic and uneven pitting corrosion,surface damage,toxicity,and allergic effects on the surrounding cells by the release of Al ions from these Mg alloys were reported in previous studies [13-20].The addition of Zn to the Mg matrices can enhance not only the yield strength and ductility of the Mg alloys,but also the corrosion resistance,inhibiting the release of H2gas in artificia body fluid [21-25].Zn is very important in the functioning numerous enzymes,supporting the immune system,DNA,and protein adhesion,and promoting bone growth with nontoxic reactions [26-28].However,it was reported that high content(＞6 wt.%) of Zn in Mg alloys can cause embrittlement and formation of micro anodic and cathodic sites due to formation of a large number of intermetallic phases in the Mg matrices,which results in decreased yield strength and corrosion resistance [29-32].Low content (＜2.7 wt.%) of Zr can provide grain refinement (GR),preventing unalloyed Zr phases in the Mg matrices [33,34].Moreover,Zr exhibits good corrosion resistance,biocompatibility,and osteocompatibility,with excellent cell adhesion and proliferation,providing low ionic cytotoxicity [35,36].

Recently,Mg matrix nanocomposites (MNCs) reinforced with graphene nanoplatelets (GNPs) have been paid attention for biomedical applications due to their superior mechanical strength (130 GPa).Two-dimensional (2D) GNP sheets containing a honeycomb lattice structure with a unique sp2-carbon (C) array can provide strong interfacial bonding between the Mg matrix and GNPs [37,38].From the perspectives of mechanical properties,it was reported that the addition of GNP to a pure Mg matrix provided higher hardness and strength via synergistic strengthening mechanisms such as the difference in thermal expansion (DTE),difference in elastic modulus (DEM),and difference in geometry (DG) between the nanoparticles and Mg matrix,GR of the composites due to the reinforcement of particles via the Hall-Petch effect,precipitation strengthening (PS) via the Orowan looping effect,and load transfer(LT)from the Mg matrix to the GNPs[39-45].In our recent studies,a low concentration (0.1 wt.%)of GNPs in Mg0.5Zr matrices gave not only higherσUCSandσCYSat 219 MPa and 162 MPa,respectively,as compared to pure Mg (142 MPa and 85 MPa) and Mg0.5Zr (181 MPa and 145 MPa),but also higher wear resistance,providing a lower coefficient of friction [34,46].However,a higher concentration (＞1 wt.%) of GNPs in the Mg matrix may adversely affect the mechanical properties due to the strong van der Waals (vdW) interactions within the GNP layers promoting the agglomeration or stacking of the GNP layers [38,47,48].It was reported that effective dispersion preventing the agglomeration of nanoparticles in metal matrices can be obtained via the high-energy ball-milling (HEBM) technique [3,49].The uniform dispersion of GNPs in pure Mg may also depend on the size of the GNPs.GNPs with 15 μm particle size exhibited better dispersion,reducing the stacking of GNP layers in the Mg matrices,compared to 5 μm GNPs [47].

From the perspective of biodegradation,GNPs with hydrophobic properties indicatinga ＞90° water contact angle(WCA)can reduce water deposition on the surface and inhibit chemical reaction [50-52].It was reported that the dispersion of low content (＜0.5 wt.%) of GNPs in Mg can provide a hydrophobic composite surface [38].In addition,it has been established that sp2-C containing single-layer GNP is like an impermeable sheet providing a thin atomic-scale barrier due to high electron density and can resist the passing of any gas molecules through the GNPs [53-55].It was reported that a low concentration (0.1-0.5 wt.%) of GNPs can reduce the corrosion current density,reducing the reaction rate and ionic transfer in saltwater such as KCl,NaCl,Na2SO4,and HBSS[34,56-58].Addition of a higher concentration (＞1 wt.%)of GNPs to pure Mg can cause large agglomerations due to stacking of GNP layers,causing hydrophilic surfaces (WCA＜90°),which results in galvanic and crevice attacks in saltwater [38,47,59].

From the biological perspective,π-πstacking or covalent crosslinking via strong vdW on the large surface areas of GNPs can enhance the adsorption of biomolecules onto the surface [60,61].Hydrophilic surfaces can promote better biological responses compared to the super-hydrophobic surfaces[62-64].It was reported that an increase in GNP content in Mg matrices can increase hydrophilicity,providing high surface energy [38].Addition of 0.1 wt.% GNPs exhibited better cell viability in osteoblast-like SaOS2 cells compared to pure Mg and 0.2 and 0.3 wt.% GNP-containing Mg MNCs [47].However,it was reported that 0.5-1 wt.% GNP gave better cell viability without any toxic reaction in MG63 cell culture[38].Based on these perspectives,the optimum concentration of GNP to control the mechanical/corrosion properties and biological response of Mg MNCs is still unclear.

Considering the limitations of existing Mg alloys and based on the above perspectives,Mg MNCs reinforced with GNPs containing Zn and Zr might be alternative biomaterials to overcome these challenges.In this context,this research manufactured novel GNP-reinforced Mg MNCs containing Zn and Zr via the powder metallurgy technique.A low concentration(0.5 wt.%) of Zr was used in the manufacturing of the composites.In this study,Mg0.5ZrxGNP (x=0.2-0.5 wt.%) and Mg0.5ZrxZn0.1GNP (x=1-5 wt.%) nanocomposites were manufactured to investigate their microstructural changes and evaluate their mechanical and corrosion properties.The synergistic strengthening and corrosion mechanisms of the fabricated Mg-Zr and Mg-Zr-Zn MNCs are also discussed intensively.

2.Experimental methods

2.1.Starting materials

Mg powder (99.9% purity,particle size:65 μm),Zr powder (99.5% purity,particle size:40 μm),Zn powder (99.5%purity,particle size:5 μm),and GNPs (particle size×thickness:15 μm×5 nm) (Sigma-Aldrich,Australia) were selected as the raw materials in this study.Stearic acid (SA)(≥99.9% purity,C18H36O2) (Sigma-Aldrich,Australia) was used as a lubricant during HEBM of the powders.

2.2.Fabrication of GNP-reinforced Mg-Zr and Mg-Zr-Zn MNCs

As the firs step,pure Mg powders with 0.5 wt.%SA in SS vials were milled in the presence of SS balls under argon(Ar)gas using a planetary ball mill (QM 3SP2,China) for 4 h to reduce their particle size to the range 30-40 μm.SA was used to establish a balance between the cold-welding and fracturing processes that occur during the milling process.SS balls of different diameters (15,10,and 7 mm) were employed to promote fracturing of the Mg particles by enhancing the collision energy between the SS balls and powder particles [65].The ball-to-powder ratio (BPR) was kept at 20:1 during the ball-milling processes.An interval of 30 min was maintained after every 30 min of ball-milling to avoid overheating of the charged powders [66].This ball-milled Mg powder was used for further milling with the Zr,Zn,and GNP powders.In the second step,a low concentration (0.5 wt.%) Zr powder was added to the ball-milled Mg powder and further milled for 2 h under similar milling conditions.The powder mixture of Mg0.5Zr was used for further milling with the Zn and GNP powders.

To prepare the batch-1 composite powders,0.2,0.3,0.4,and 0.5 wt.% GNP were added to the SS vials containing the Mg0.5Zr powder mixtures and further ball milled for 2 h under similar milling conditions.The Mg0.5ZrxGNP (x=0.2,0.3,0.4,and 0.5 wt.%) powder mixtures were then used to synthesize the firs batch of Mg-Zr MNCs.To prepare the batch-2 nanocomposite powders,0.1 wt.% GNPs were added to the Mg0.5Zr powder mixtures and ball-milled for 2 h under similar milling conditions.As the final step,1,3,4,and 5 wt.% Zn powder was added to the ball-milled Mg0.5Zr0.1GNP powder mixtures and milled for another 2 h.To disperse the Zn particles in the Mg0.5Zr0.1GNP powdermixture,zirconia ceramic balls were selected to prevent structural changes and reduce the dynamic energy maintaining the thermoelectric properties of Zn [67,68].The chemical compositions of batch-1 and batch-2 composites are provided in Table 1.The powder mixtures were consolidated into green compacts using SS dies (1.6 cm in diameter and 1.8 cm in height) via hydraulic cold-pressing at 760 MPa maintained for 30 min.The compact samples were stepwise sintered in a high-purity Ar gas environment in a furnace with 92%heating efficiency.The stepwise sintering temperatures were 400 °C for 1 h to remove the SA and then 610 °C for 2 h.

Table 1Chemical composition of batch-1 Mg-Zr and batch-2 Mg-Zr-Zn MNCs.

Table 2Characteristic intensity ratios of ID/IG and IG/I2D in Mg0.5Zr and Mg0.5ZrxZn (x=1-5 wt.%) MNCs containing 0.1-0.5 wt.% GNPs during dispersion and sintering stages of nanocomposites measured from corresponding Raman spectra (average values are reported).

2.3.Characterization of as-received and ball-milled powders and sintered composites

The morphology of the as-received powders,ball-milled powders (BMP),elemental dispersion of BMP mixtures,and sintered samples was identified using scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectrometry (EDX) (FEI Nova NanoSEM).The grain structure,residual pores,and secondary phases of the sintered samples were analyzed via optical microscopy (OM) (Leica DM2500M with 3.1MP CCD).An X-ray diffractometer(XRD,BrukerAXS D4 Endeavor)was used to identify the different phases with their miller indices and structural changes in the raw powders,BMP mixtures,sintered samples,and corroded Mg-Zr and Mg-Zr-Zn MNCs.For XRD,samples were scanned in the range 2θ=10-90° at a 0.02°/secscanning rate using Cu-Kαradiation (λ=0.154 nm).A Raman spectroscope (Raman/PL HORIBA LabRAM HR) was used to characterize the quality of GNPs and defect concentrations in the sp2graphitic structure in the as-received GNP powders,BMP mixtures,and sintered Mg-Zr and Mg-Zr-Zn MNCs.

For grain size analysis,disk samples (1 cm in diameter and 0.3 cm in thickness) were machined from the sintered samples using electrical-discharge machining (EDM).A progressive grinding process with those disk samples was carried out using 800,1200,2400,and 4000 grit silicon carbide(SiC) grinding papers.Mechanical polishing of the ground samples was performed using MD-Chem polishing cloths lubricated with 0.05 μm colloidal silica-based oxide polishing suspension (OPS) and ethanol in a volumetric ratio of 50:50.The polished samples were then ultrasonically cleaned in an ethanol solvent.After polishing,the samples were etched using a picral reagent containing 50 ml ethanol,3 ml picric acid,5 ml acetic acid,and 10 ml distilled water.The grain size of the sintered samples was measured using the line intercept method according to ASTM E112-12.The relative density of the sintered samples was determined by Archimedes’ principle according to ASTM B962-15.

2.4.Mechanical property testing of fabricated samples

The mechanical properties of the batch-1 and batch-2 nanocomposite samples were evaluated by compression tests using a uni-axial 50 kN Instron servo-hydraulic testing machine at a 0.5 mm/min displacement rate.Five samples from each composition were tested to report an average value.According to ASTM E9-09,cylindrical samples with 5 mm diameter and 10 mm length were machined using EDM to perform the compression tests.The nanohardness and modulus of elasticity of the samples were determined by nanoindentation tests using a Hysitron TI-950 TriboIndenter with a Berkovich diamond tip.The modulus of elasticity of the sintered samples was determined using the Oliver-Pharr method [69].During nanoindentation,a maximum applied load of 10,000 μN was applied and maintained for a dwelling time of 10 s.The indentation was performed in the pattern of a 5×5 array at three different positions on each sample.Each indent was separated by 20 μm from its adjacent indent to eliminate the generation of residual stress from other indents.The Vickers hardness (HV) of the composite samples was determined via microindentation tests under a 0.3 kgf applied load using a micro-hardness tester (Buehler Omnimet MHT 5104).To obtain average the HV value,at least ten microindents were performed,maintaining a 250 μm inter-microindentation distance to prevent any residual stresses caused by neighboring indents.

2.5.Evaluation of corrosion properties of Mg-Zr and Mg-Zr-Zn MNCs

To determine the corrosion properties of the fabricated samples,electrochemical tests and H2evolution tests were performed in HBSS at 37±1 °C.The disk samples (1 cm diameter and 0.3 cm thickness) were connected to a copper wire using a conductive silver paste and then capsulized with cold epoxy resin for electrochemical tests while the opposite surface was exposed.The exposed surface area (?0.8 cm2)was ground using up to 1200 grit SiC grinding papers followed by degreasing with ethanol and acetone.The ground samples were then ultrasonically cleaned in an ethanol solvent and dried in a stream of hot air.The cold-mounted samples were immersed in HBSS.In this study,a saturated calomel electrode (SCE) as the reference electrode,a platinum electrode (1.5×1.5 cm2) as the counter electrode,and the cold-mounted samples as the working electrode were used in a three-electrode cell system.The potentiodynamic polarization curves were obtained via an open circuit potential at a scan rate of 1.5 mVs-1using a multichannel electrochemical workstation(VSP-300 multi,BioLogic science instrument,France).The potential was recorded for every 10 mV ranging from -2 V to 1 V.The corrosion current density (Icorr)(μA cm-2) and the corresponding CR (mmy-1) were determined by the Tafel extrapolation method using EC-Lab software based on ASTM G-102.To measure the average value,at least two samples were tested for each chemical test according to ISO Standard 10993-12.The morphologies of batch-1 and batch-2 corroded samples were also investigated by SEM to identify the corrosion effects on the surfaces.

For the H2evolution tests,the samples were mounted in epoxy with one surface exposed which was then ground using SiC grinding papers.The mounted samples were kept in HBSS with their exposed surfaces facing upwards.During the degradation tests,the ratio of surface area (cm2) to volume (ml) of HBSS was maintained at 1:300.A funnel and burette apparatus was used to cover the samples and collect the volume of released H2gas (ml) as per the methods described in the literature [70].The ion concentration tests were performed in HBSS after 24 h degradation of fabricated Mg-Zr and Mg-Zr-Zn MNCs using a Microwave Plasma-Atomic Emission Spectrometer (MP-AES) (Agilent 4200) to identify the concentrations (mg/L) of Mg2+and Zn2+ions released in the HBSS.

The static initial WCA tests of the composite surfaces were carried out using an optical Tensiometer (model TL100) to investigate the surface energy of the surfaces via hydrophobicity and hydrophilicity effects.At least five positions were selected on each composite surface to measure an average WCA.The gravitational force (g=9.81 ms-2) and the pull-off force of water droplets were prevented via microbalancing [52].

3.Results and discussion

3.1.Morphology of as-received powders and BMP mixtures

Fig.1 shows SEM micrographs with corresponding EDX spectra of the as-received powders of Mg,Zn,Zr,and GNP.The as-received pure Mg (Fig.1a) and pure Zn (Fig.1b)exhibited spherical particles with average particle sizes of 65 μm and 5 μm,respectively,whereas the Zr powder(?20 μm)possessed an irregular morphology(Fig.1c)and the as-received GNPs (Fig.1d) exhibited fla e-shaped platelets with an average particle size of 15 μm.The highlighted regions in the SEM micrographs indicate the area selected for the EDX analysis.

The morphologies of the batch-1 BMP mixtures containing 0.2-0.5 wt.% GNPs and their corresponding EDX elemental maps are shown in Fig.2.The dispersion of Zr,O,and C representing GNPs in the EDX maps are observed in the SEM morphologies of 0.2,0.3,0.4,and 0.5 wt.%GNP-containing BMP mixtures,as shown in Fig.2a,2b,2c,and 2d,respectively.It was observed that the increase in GNP concentration from 0.2 to 0.5 wt.% provided inhomogeneous dispersion in the batch-1 BMP mixtures,whereas 0.1 wt.%GNPs exhibited homogeneous dispersion in Mg matrices in previous studies [34,47].The reason for the inhomogeneous dispersion of GNPs may be attributable to the stacking and agglomeration of GNPs in the Mg matrices during the milling process [47].The distribution of particle size was measured using the Feret’s diameters (dF) [71] of more than 300 particles selected from SEM images to calculate the mean particle size of the batch-1 BMP mixtures.The mean particle sizes were 33 μm,36 μm,37 μm,and 37 μm for the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP BMP mixtures,respectively.

Fig.1.SEM morphologies of as-received powders:(a) pure Mg;(b) Zn;(c) Zr;(d) GNPs (red circles indicate EDX mapping areas).

Fig.3 shows the morphologies and corresponding EDX elemental maps of the batch-2 BMP mixtures containing 1-5 wt.% Zn.The EDX point analysis of the BMP mixtures also confirmed the dispersion of Zr,Zn,and GNP powders in the Mg powders,showing the effectiveness of the HEBM process.The mean particle sizes were measured as 31 μm,30 μm,35 μm,and 38 μm for the Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP BMP mixtures,respectively.

3.2.XRD and Raman analysis of powder mixtures and sintered samples

The XRD patterns of the starting powders are shown in Fig.4a showing characteristic peaks of Mg,Zn,Zr,and GNP.The XRD pattern of the GNPs revealed diffraction peaks at 26.5° and 54.5° which correspond to the 002 and 004 planes of GNPs.Fig.4b shows the XRD patterns of the batch-1 BMP mixtures containing 0.2-0.5 wt.% GNPs and their Mg-Zr MNC samples.A less intense GNP peak was observed in the batch-1 BMP mixtures containing 0.2-0.5 wt.% GNPs.However,interestingly,XRD patterns collected from the sintered Mg-Zr MNCs reinforced with GNP samples did not reveal any peaks associated with GNPs.The reason for the absence of C peaks in the XRD pattern of the Mg0.5ZrxGNP (x=0.2-0.5 wt.%) nanocomposite samples may be attributable to the much lower concentration of GNPs in the Mg matrices,which made GNP undetectable in the sintered samples.It may also be attributable to the significant difference in mass attenuation coefficient between Mg(4.5×102cm2g-1) and C (1.3×103cm2g-1) for the Kαradiation of X-rays [34].

Fig.4c shows the XRD patterns of the batch-2 BMP mixtures containing 0.1 wt.% GNPs and different concentrations of Zn particles (1,3,4,and 5 wt.%),and their sintered samples.The peaks at 2θ=43.17° in the XRD patterns of the BMP powder mixtures were associated with Zn.No carbide peaks were observed in the XRD patterns collected from the sintered Mg-Zr-Zn MNCs,confirming the chemical stability of GNPs at a high sintering temperature (610 °C).However,a less intense peak of an MgZn intermetallic phase(γ-MgZn) at 2θ=38.98° was also observed in the sintered Mg0.5ZrxZn0.1GNP (x=3-5 wt.%) nanocomposites,showing the reactivity of Mg with Zn during the composite processing [26,72].A weak magnesium oxide (MgO)peak was observed in batch-1 Mg-Zr MNCs as shown in Fig.4b,whereas batch-2 Mg-Zr-Zn MNCs did not reveal any MgO peak (Fig.4c).The MgO peaks in the XRD spectra of batch-1 samples may be attributed to the presence of oxygen(O2) atoms in the internal grain boundaries of the composites which reacted with Mg matrices at ?410°C[72-74].However,batch-2 composites containing Zn as an alloying element revealed an additionalγ-MgZn intermetallic phase at the grain boundaries which resulted in a lower porosity of this batch of composites as compared to batch-1 composites.Another reason may be attributed to the high affinit of GNPs to attach O2in the sp2carbon networks as GNPs have a large surface area (?2630 m2/g) providing more sites for the reaction with oxygen [47].

Fig.2.SEM images,and corresponding elemental maps of the batch-1 BMP mixtures:(a) Mg0.5Zr0.2GNP;(b) Mg0.5Zr0.3GNP;(c) Mg0.5Zr0.4GNP;and(d) Mg0.5Zr0.5GNP.(Red arrows indicate the dispersion of GNPs in the Mg particles).

The structural changes in the characteristic sp2graphitic structure of GNPs during ball-milling and sintering were characterized by Raman spectroscopy and the results are shown in Fig.4d.Raman spectra obtained from the as-received GNPs revealed three characteristic graphitic peaks at 1345 cm-1,1577 cm-1,and 2700 cm-1which were associated with the D,G,and 2D bands of GNP,respectively,where the characteristic G band represents first-orde Raman scattering and the D and 2D bands were generated from the second-order Raman scattering process.In the Raman spectra of GNPs,the D band revealed non-sp2defects,whereas the G band exhibited the in-plane stretching mode of C-C bonds in GNPs which is attributed to the crystallinity or graphitization of GNPs [75].

Fig.3.SEM images of batch-2 BMP mixtures,corresponding EDX elemental maps and spectra:(a) Mg0.5Zr1Zn0.1GNP;(b) Mg0.5Zr3Zn0.1GNP;(c) Mg0.5Zr4Zn0.1GNP;and (d) Mg0.5Zr5Zn0.1GNP (insets show highmagnification elemental maps of powder particles highlighted in red circles showing distribution of various elements in batch-2 BMP mixtures).

Fig.4.XRD patterns and Raman spectra obtained from the received powders,BMP mixtures,and sintered Mg0.5ZrxGNP (x=0.2-0.5 wt.%) and Mg0.5ZrxZn0.1GNP (x=1-5 wt.%) nanocomposites:(a) XRD patterns of as-received powders;(b) XRD patterns of batch-1 Mg0.5ZrxGNP (x=0.2-0.5 wt.%) BMP mixtures and their sintered samples;(c) XRD patterns of batch-2 Mg0.5ZrxZn0.1GNP (x=1-5 wt.%) BMP mixtures and their sintered samples;(d) Raman spectra of as-received GNP,BMP mixtures,and sintered nanocomposites containing different contents of GNPs (0.1-0.5 wt.%).

The characteristic intensity ratios quantified from the D,G,and 2D bands of GNPs in the as-received powders,BMP mixtures,and sintered samples are summarized in Table 2.The intensity ratio of the D band to G band (ID/IG) is indicative of the accumulation of defects or induced disorders in GNPs and changes in the structural integrity of dispersed GNPs in the Mg matrices,whereas the characteristic intensity ratio of the G band to 2D band (IG/I2D) is indicative of the changes in the number of layers in GNPs during the composite processing[38,47].Disordering of the graphitic lattice as shown by the existence of sp3or non-sp2C can be found in vacancies or basal edges of GNPs that may emerge during the composite processing stages [34,65].These defects may occur due to impact energy developed in the individual GNPs by the high friction within agglomerated GNPs during the milling process[47].In the Raman spectra,the shape of the characteristic 2D peak and its full-width half maxima (FWHM) can quantify the number of GNP layers.A symmetrical 2D peak associated with a singleπelectron valance band excited via only one Raman scattering cycle represents a single-layer GNP(SLG),whereas a multi-layer GNP (MLG) causes multiple pairs ofπ-πelectron bands resulting from multiple Raman scattering cycles;therefore,higher resonant phonons with different frequencies contribute to the wider FWHM in the 2D band compared to SLGs [76].The ID/IG,IG/I2D,and FWHM in the 2D band of the as-received GNPs were measured as 0.57,4.8,and 81 cm-1,respectively.It can be seen that the characteristic ID/IGratio increased with increasing content of GNPs in the Mg matrices,which is attributed to their agglomeration and stacking in the Mg matrices [47].The value of IG/I2Dof the BMP mixtures and sintered nanocomposites decreased with increasing GNPs content in the Mg-Zr and Mg-Zr-Zn matrices.From the characteristic 2D peak of the Raman spectra of the BMP mixtures containing GNPs and their nanocomposites,the FWHM was measured and varied between 90 and 95 cm-1.

3.3.Microstructure of sintered samples

The morphologies of the batch-1 sintered Mg-Zr MNCs containing 0.2-0.5 wt.% GNPs are shown in Fig.5.The average grain sizes of the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP were measured as 25 μm,35 μm,36 μm,and 34 μm,respectively.Inhomogeneous dispersion of GNPs and Zr particles in the Mg matrices adversely affected the GR in the Mg0.5ZrxGNP(x=0.2-0.5 wt.%),which complements the results published in previous studies [34,47,77,78].Figs.5(e-h)show the SEM micrographs and corresponding EDX elemental maps of Mg-Zr MNCs containing 0.2-0.5 wt.%GNPs.Micron-sized pores were observed in the microstructures of Mg0.5Zr0.2GNP(Fig.5e),Mg0.5Zr0.3GNP(Fig.5f),Mg0.5Zr0.4GNP (Fig.5g),and Mg0.5Zr0.5GNP (Fig.5h)which adversely affected the relative densities of these composites.The average relative densities of the Mg-Zr MNCs were measured as 91%,88%,90%,and 89% for the composites containing 0.2,0.3,0.4,and 0.5 wt.% GNP,respectively,indicating a decrease in the densification of the nanocomposites with increasing GNP content in the Mg-Zr matrices.This may be attributable to the stacking and re-agglomeration of GNPs in the deeper zones of the Mg particles during the high sintering temperature [47].EDX elemental maps reveal the aggregated Zr particles in the Mg0.5Zr0.5GNP nanocomposite,as shown in Fig.5h,which resulted in insufficient GR in the composite.In addition,most of the elemental dispersion was observed on the grain boundaries of the composites.

Fig.6 shows the microstructures of the batch-2 Mg-Zr-Zn MNCs reinforced with 0.1 wt.% GNPs.A low content of GNPs (0.1 wt.%) was added to this batch of nanocomposites to investigate the effect of Zn addition.Also,the addition of 0.1 wt.% GNPs exhibited uniform dispersion in Mg-Zr matrices in a previous study resulting in higher yield strength of the Mg0.5Zr0.1GNP [34].Addition of Zn to the Mg0.5Zr0.1GNP samples resulted in the formation ofγ-MgZn phases during the sintering stages,as indicated by the white arrows in Fig.6a,6b,6c,and 6d for the Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP,respectively.This complements the XRD results for this batch of nanocomposites as shown in Fig.4c.The formation ofγ-MgZn phases may be attributed to the slow cooling rate (-5 °C/min) from the peak sintering temperature of 610 °C.The overall mechanism of the formation of theseγ-MgZn phases during the sintering process in an Ar atmosphere can be described by the following equations [72]:

Heating:

where the subscripts‘s’and‘l’indicate solid and liquid states,respectively,of the reactants and the products.

Due to the low melting point of Zn (419.5 °C),it took on a liquid state during the high sintering temperature of the Mg-Zr-Zn MNCs (Eqn.(1)) and,therefore,interacted with the surrounding Mg matrices at the temperature of 530 °C to form moltenγ-MgZn,as shown by the reaction in Eqn.(2).During the cooling stages,the moltenγ-MgZn compounds tended to aggregate along the grain boundaries (Eqn.(3)).

Fig.5.OM and SEM micrographs with corresponding EDX elemental maps showing the distributions of various elements in batch-1 sintered Mg-Zr MNCs:(a and b) Mg0.5Zr0.2GNP;(c and d) Mg0.5Zr0.3GNP;(e and f) Mg0.5Zr0.4GNP;(g and h) Mg0.5Zr0.5GNP (insets show high-magnification images of the pores,and dispersion of GNPs indicated by black arrows in Mg-Zr MNCs).

Compared to the Mg0.5ZrxZn0.1GNP (x=4,and 5 wt.%)nanocomposites,the addition of 1 and 3 wt.% Zn to the Mg0.5Zr0.1GNP resulted in better GR of these batches of nanocomposites,as shown in Fig.6a and 6b,respectively.The reason for GR may be attributable to the dissolved Zr particles as a grain-growth restricting element [33] and a pinning effect due to intermetallic compounds at the grain boundaries,which inhibited the grain growth of the Mg[79].However,the unalloyed Zr particles in the Mg matrices and inhomogeneous precipitation of the intermetallic compounds at the grain boundaries can adversely affect GR [34,79].The average grain sizes of the Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP nanocomposites were measured as 18 μm,14 μm,21 μm,and 32 μm,respectively.

Figs.7a-d show the morphologies and corresponding EDX elemental maps of the batch-2 Mg-Zr-Zn MNCs.A few aggregated Zr particles were observed in the Mg0.5Zr1Zn0.1GNP and Mg0.5Zr5Zn0.1GNP,as shown in Figs.7a and 7d,respectively,showing the poor dispersion of Zr particles in these composites.The unalloyed aggregated Zr particles become stress-concentration regions in the composites during the applied loads and act as nucleation sites to initiate cracks in the composites [80,81].In the batch-2 Mg-Zr-Zn MNCs,elemental mapping of these composites revealed the precipitation ofγ-MgZn phases at grain boundaries (Fig.7b) and oxygen (O) rich regions as shown in Fig.7c.GNPs have a high affinit to attaching O to their sp2carbon networks and their large surface areas (?2630 m2/g) provide more sites for their reaction with oxygen [47].Moreover,uniform distribution ofγ-MgZn phases was observed at the grain boundaries of the Mg0.5Zr3Zn0.1GNP compared to the Mg0.5Zr4Zn0.1GNP and Mg0.5Zr5Zn0.1GNP.From the SEM images of the batch-2 Mg-Zr-Zn MNCs,the average size of the intermetallic particles was measured as 900 nm.The average relative densities were also measured as 94%,98%,96%,and 96% for the Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP,respectively.It is noted that the addition of Zn to batch-2 nanocomposites positively impacted the densification of these composites.The relative densities of this batch of composites were measured approximately 8% higher than those of batch-1 nanocomposites.This may be attributed to the melting of Zn at elevated temperature sintering which led to the formation ofγ-MgZn phases at the grain boundaries of composites therefore reduced the porosities in this batch of composites [82].

Fig.6.OM micrographs of batch-2 Mg-Zr-Zn MNCs:(a)Mg0.5Zr1Zn0.1GNP;(b) Mg0.5Zr3Zn0.1GNP;(c) Mg0.5Zr4Zn0.1GNP;and (d) Mg0.5Zr5Zn0.1GNP (white arrows indicate the γ-MgZn secondary phases).

3.4.Mechanical properties of sintered samples

Fig.8a presents the compressive stress-strain curves of the Mg0.5ZrxGNP(x=0.2-0.5 wt.%)nanocomposites which were interpreted to evaluate the compressive behavior of each nanocomposite and the results are summarized in Table 3.It can be seen that the ultimate compressive strength (σUCS),compressive yield strength(σCYS),and compressive strain(εc)decreased with increasing GNP content in the Mg-Zr matrices.Higher content of GNPs led to their agglomeration in the Mg matrices,which adversely affected the mechanical properties of the Mg-Zr MNCs [45].Another reason could be the formation of stress-concentration regions in the aggregated GNPs which may have resulted in the generation of microcracks that caused premature failure of the nanocomposites during load transfer from the Mg matrix to the stiffer GNPs[38,47,83],as shown in Fig.9a.TheσUCSandσCYSof the nanocomposites were measured as 187 and 142,182 and 132,158 and 117,and 161 and 134 MPa,respectively,for the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP.The compressive stress-strain curves for the batch-2 Mg-Zr-Zn MNCs are shown in Fig.8b.It can be seen that the addition of 3 wt.%Zn to the Mg0.5Zr0.1GNP resulted in a higherσUCSas compared to batch-2 Mg0.5ZrxZn0.1GNP(x=1,4,and 5 wt.%) nanocomposites.It should be noted that maximum solubility of Zn in Mg is 1.6 wt.% at room temperature in the equilibrium state.Zn can partially dissolve into primary Mg matrix and provides solid solution strengthening [84,85].However,further increase of Zn content (4 wt.% and 5 wt.%) in the Mg matrices led to a decreased compressive strength of the composite.This may be attributed to the increase of dissolved Zn in the Mg particles,which promotes the formation of secondary phases with dendritic segregation along grain boundaries,which adversely affect the strength of composites [29].The uniform distribution ofγ-MgZn intermetallic phases at the grain boundaries may cause precipitation strengthening and therefore result in increased strength while decreasing the ductility of these composites [29].

Fig.8c shows the load vs displacement curves obtained from nanoindentation tests of the batch-1 Mg-Zr MNCs reinforced with 0.2-0.5 wt.% GNPs and the results are summarized in Table 3.The indentation depths for the batch-1 Mg-Zr MNCs were measured as 540 nm for the Mg0.5Zr0.2GNP composite as compared to 580 nm,584 nm,and 674 nm for the Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP,respectively,under similar applied loads,whereas the nanohardness values for this batch of nanocomposites were measured as 1.0,0.8,0.6,and 0.8 GPa for the Mg-Zr MNCs containing 0.2,0.3,0.4,and 0.5 wt.% GNP,respectively.It can be seen that the hardness values of the Mg-Zr MNCs deteriorated with increasing GNP content.This can be attributed to the inhomogeneous dispersion of higher contents of GNPs(＞0.2 wt.%) in the Mg matrices,which is in agreement with the results published in a previous study [45].Fig.8d shows the load-displacement curves of the batch-2 Mg-Zr-Zn MNCs reinforced with 0.1 wt.% GNPs and the results are summarized in Table 3.It can be seen that higher Zn (5 wt.%) content also reduced the hardness of the composites.This complements the compressive behavior of this batch of nanocomposites,as higher Zn content (5 wt.%) significantly enhanced ductility (εc=23%).

Table 3Mechanical properties of batch-1 and batch-2 sintered samples measured by compression and nanoindentation tests.

Fig.10 shows the assessments of strength of the batch-1 and batch-2 nanocomposites considering the potential strengthening mechanisms including DTE,DEM,and DG of the particles,and GR,LT,and PS mechanisms,which were discussed with their governing equations in previous studies [3,34,47].Fig.10a shows a comparison between the experimentalσCYSandσpredictedof the batch-1 Mg-Zr MNCs containing 0.2,0.3,0.4,and 0.5 wt.% GNPs.The experimentalσCYSof the Mg0.5Zr0.2GNP nanocomposite exhibited good agreement with the predicted value of yield strength.It was observed that the deviation betweenσCYSandσpredictedvalues increased with increasing GNP content in the Mg-Zr matrices,indicating the insufficient functioning of the strengthening mechanisms to increase the yield strength of the batch-1 Mg-Zr MNCs.This may be attributable to the agglomeration of GNPs or stacking of GNP layers leading to the inhomogeneous dispersion of GNPs in the Mg matrix.Theσpredictedwas measured as 169,181,192,and 203 MPa for the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP,respectively,whereas Mg0.5Zr0.1GNP made via a similar manufacturing technique revealed very close agreement between the experimental and predicted yield strength in a previous study [34] that was attributed to the uniform dispersion of GNPs in the Mg-Zr matrices.Fig.10b shows the effect of the addition of Zn on the yield strength of the batch-2 Mg-Zr-Zn MNCs.The experimentalσCYSof the Mg-Zr-Zn MNCs was compared to theσpredictedof an Mg0.5Zr0.1GNP composite (straight line) made via a similar manufacturing technique in a previous study [34].It was observed that the Mg0.5Zr1Zn0.1GNP and Mg0.5Zr3Zn0.1GNP exhibited 29% and 34% higherσCYS,respectively,compared to theσpredictedof the Mg0.5Zr0.1GNP nanocomposite (164 MPa) [34],whereas the Mg0.5Zr4Zn0.1GNP and Mg0.5Zr5Zn0.1GNP nanocomposites revealed 17% and 46%lowerσCYS,respectively,compared to the Mg0.5Zr0.1GNP nanocomposite.This may be attributable to the inhomogeneous distribution ofγ-MgZn intermetallic phases at the Mg grain boundaries which resulted in the formation of coarse grains due to an insufficient pinning effect leading to poor GR [26,32].

Fig.7.SEM images and corresponding EDX elemental maps showing the distributions of different elements in Mg-Zr-Zn MNCs:(a) Mg0.5Zr1Zn0.1GNP;(b) Mg0.5Zr3Zn0.1GNP;(c) Mg0.5Zr4Zn0.1GNP;and (d) Mg0.5Zr5Zn0.1GNP (Insets show high-magnification image of the dispersion of GNPs in the Mg-Zr-Zn matrices indicated by white arrows).

Fig.8.Compressive and nanoindentation behaviors of Mg-Zr and Mg-Zr-Zn MNCs containing different concentrations of GNPs:(a) compressive stressstrain curves of batch-1 Mg-Zr MNCs;(b) compressive stress-strain curves of batch-2 Mg-Zr-Zn MNCs during nanoindentation;(c) nanoindentation loaddisplacement curves of batch-1 Mg-Zr MNCs;and (d) nanoindentation load-displacement curves of batch-2 Mg-Zr-Zn MNCs.

Fig.9.Schematic illustration of load transfer and fracturing of Mg MNCs containing:(a) agglomerated GNPs and (b) uniformly dispersed GNPs.

3.5.In vitro corrosion behavior of fabricated Mg-Zr and Mg-Zr-Zn MNCs

Fig.11 shows the corrosion behaviors of the batch-1 and batch-2 nanocomposites assessed by electrochemical and H2evolution tests,and the results are summarized in Table 4.Fig.11a shows the typical potentiodynamic polarization curves of Mg-Zr MNCs reinforced with 0.2-0.5 wt.%GNP in HBSS.It can be seen that the corrosion current density (Icorr)increased with increasing GNP content in the Mg-Zr matrices.This is mainly attributed to agglomerated GNPs in the Mg-Zr MNCs which became cathodic sites and resulted in galvanic corrosion of the nanocomposites.The logarithmic current densities(log I)were measured as-3.1 mA/cm2,-2.75 mA/cm2,-2.41 mA/cm2,and -1.92 mA/cm2for the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP,respectively.The CRs of these nanocomposites were quantified using the Tafel extrapolation method and validated using Eqn.(4),and the corresponding polarization resistance (Rp)was calculated by Eqn.(5) [34,86]:

Table 4Corrosion properties of Mg-Zr and Mg-Zr-Zn MNCs containing 0.1-0.5 wt.% GNPs evaluated via electrochemical and H2 evolution tests.

where Icorris the corrosion current density (mA/cm2) measured by the Tafel extrapolation fi intersecting the linear portion of the anodic and cathodic curves,andβaandβcare the Tafel constants (mV) for the anodic and cathodic parts,respectively.

Fig.10.Validation of predicted yield strength (MPa),considering potential strengthening mechanisms,with experimental values:(a) batch-1 Mg-Zr MNCs(black dot and red dot indicate predicted and experimental values,respectively);(b) batch-2 Mg-Zr-Zn MNCs (straight line indicates predicted value of yield strength of Mg0.5Zr0.1GNP and black dots indicate the experimental σCYS of Mg0.5ZrxZn0.1GNP) (x=1-5 wt.%).

Fig.11.Corrosion behavior of batch-1 Mg-Zr MNCs containing 0.2-0.5 wt.% GNPs and batch-2 Mg-Zr-Zn MNCs containing 0.1 wt.% GNPs evaluated via electrochemical and H2 evolution tests in HBSS at 37 °C:(a and b) potentiodynamic polarization curves of batch-1 and batch-2 nanocomposites,respectively;(c and d) H2 evolution during immersion time of batch-1 and batch-2 nanocomposites,respectively;(e and f) CR (mm/y) measured from electrochemical and H2 evolution tests for batch-1 and batch-2 nanocomposites,respectively.

The inhomogeneous dispersion of GNPs in Mg matrices may decrease the diffusivity of O ions,which adversely affects the formation and stability of passivation layers on composites[47,87].During chemical reaction in HBSS,Mg metal hits H2ions out of solution due to its higher reactivity than elemental H2,as expressed in Eqns.(6) and 8.The rest of the Mg ions and hydroxide ions result in corrosion products Mg(OH)2(Eqn.(9)).The overall corrosion mechanism and formation of subsequent corrosion products of the batch-1 composites are governed by the reactions below [47]:

Anodic reaction:

Fig.11b shows the potentiodynamic polarization curves of the batch-2 Mg-Zr-Zn MNCs containing 0.1 wt.% GNPs in HBSS.It can be seen that the Icorrof the Mg0.5Zr4Zn0.1GNP and Mg0.5Zr5Zn0.1GNP increased during polarization,which is in agreement with previous studies [32,88],whereas the Mg0.5Zr3Zn0.1GNP exhibited lower Icorrproviding higher Rpcompared to the other Mg-Zr-Zn MNCs.This is attributed to the better GR which stimulated the formation of passive film during immersion in HBSS by breaking the intermetallic particles along the grain boundaries [89,90].Also,the batch-2 nanocomposites exhibited higher corrosion protection with lower Icorrin HBSS compared to the batch-1 nanocomposites.The mechanisms of the higher corrosion resistance of the batch-2 Mg-Zr-Zn MNCs reinforced with 0.1 wt.% may be attributable to the following reasons:(i)the hydrophobicity of the nanocomposites reducing the release of electrons (2e-)during the anodic reaction [51,52];(ii) the impermeability of uniformly dispersed GNPs in the Mg matrices providing an atomic-scale barrier to inhibit H2gas production during the cathodic reaction [53];and (iii) the lower reactivity of Zn compared to Mg,which resulted in slower H2evolution.

During chemical reaction,Zn can displace H2ions from HBSS due to the higher reactivity of Zn compared to H2,as in Eqn.(10).However,the evolution of H2gas by Zn occurs at a slower rate compared to Mg,as the reactivity order of those elements is Mg＞Zn＞H.In addition,Mg can prevent Zn2+ions transferring electrons(2e-)into the solution,which results in the production of further Zn leading to the reverse reaction (Eqn.(12)).The additional chemical reactions due to the addition of Zn to the batch-2 Mg-Zr-Zn MNCs are governed by the equations below [91]:

Compared to the Mg0.5Zr0.4GNP and Mg0.5Zr0.5GNP,the Mg0.5Zr0.2GNP,and Mg0.5Zr0.3GNP exhibited a slower H2evolution process,as shown in Fig.11c.In the inset view of Fig.11c,no significant increase in H2evolution of the 0.2 wt.% and 0.3 wt.% GNP-reinforced Mg-Zr MNCs was observed during the initial 3 h immersion time,whereas the 0.4 wt.% and 0.5 wt.% GNP-reinforced nanocomposites exhibited faster H2evolution after 1 h immersion.This may be attributable to the increased affinit to adhesion interaction between water and the GNP-free surfaces of the composites due to the inhomogeneous dispersion of GNPs.On the other hand,the 0.1 wt.%GNP-reinforced batch-2 Mg-Zr-Zn MNCs exhibited a slower H2evolution process in HBSS as compared to the batch-1 Mg-Zr MNCs,as shown in Fig.11d.It can be seen that the evolution of H2gas decreased with increasing Zn content in the Mg0.5Zr0.1GNP.The CR using the H2evolution (ml) values of the nanocomposites was quantified by Eqn.(13) [34]:

where ΔvH2is the release of H2gas (ml) during immersion time t(day)and A(cm2)is the exposed surface area in HBSS.

Fig.12.(a) Schematic of effects of hydrophobic and hydrophilic surfaces on WCA;(b-e) WCA of batch-1 nanocomposites:(b) Mg0.5Zr0.2GNP;(c)Mg0.5Zr0.3GNP;(d) Mg0.5Zr0.4GNP;(e) Mg0.5Zr0.5GNP;and (f-i) batch-2 nanocomposites:(f) Mg0.5Zr1Zn0.1GNP;(g) Mg0.5Zr3Zn0.1GNP;(h)Mg0.5Zr4Zn0.1GNP;and (i) Mg0.5Zr5Zn0.1GNP.

The CR (mm/y) measured from electrochemical and H2evolution test results is plotted in Fig.11e and 11f for the batch-1 and batch-2 nanocomposites,respectively.The corresponding values from the corrosion tests are listed in Table 4.

Agglomeration of hydrophobic GNPs in the Mg matrices also exposed more sites of the hydrophilic Mg matrices to the corrosive medium,which enhanced the corrosion attack,as shown in Fig.12a [38].The initial WCA of the batch-1 Mg-Zr MNCs’ surface decreased with increasing GNPs in the Mg-Zr matrices leading to increased hydrophilicity,as shown in Fig.12 (b-e),which is in agreement with a previous study [38].However,the batch-2 Mg-Zr-Zn MNCs revealed hydrophobic surfaces during the WCA test,as shown in Fig.12 (f-i).The WCA of the batch-2 nanocomposite surfaces decreased due to the decrease in surface tension over time.

Fig.13a shows the XRD spectra of the corroded batch-1 and batch-2 nanocomposites to identify the corrosion products formed on the surfaces of these nanocomposites.Compared to the batch-1 Mg-Zr MNCs,the XRD pattern of the batch-2 Mg-Zr-Zn MNCs revealed higher intensity of Mg peaks at 34.5°,36.7°,47.9°,and 63.3°,and lower intensity of Mg(OH)2peaks at 38.1°,50.8°,and 58.7°,indicating the corrosion-inhibition ability of the nanocomposites containing Zn.This is attributed to the lesser H2gas evolved during degradation in HBSS [34].In addition,the Mg-Zr-Zn MNCs exhibited additional peaks associated with Zn(OH)2which were formed on the substrates as a corrosion product.The XRD spectra of the batch-2 nanocomposites also revealed a tiny MgO peak,indicating the formation of a thin layer of MgO on the Mg substrates.

Fig.13.(a)XRD patterns of corroded batch-1 and batch-2 nanocomposites;(b)Mg2+ and Zn2+ ion concentrations in HBSS after 24 h degradation of nanocomposites;(c) SEM images of corroded batch-1 Mg-Zr MNCs:(ci) Mg0.5Zr0.2GNP;(cii) Mg0.5Zr0.3GNP;(ciii) Mg0.5Zr0.4GNP;and (civ) Mg0.5Zr0.5GNP;and (d) SEM images of corroded batch-2 Mg-Zr-Zn MNCs:(di) Mg0.5Zr1Zn0.1GNP;(dii) Mg0.5Zr3Zn0.1GNP;(diii) Mg0.5Zr4Zn0.1GNP;and (div)Mg0.5Zr5Zn0.1GNP (Red circles indicate the positions in the low magnification for the high magnification SEM images).

During degradation,the batch-1 Mg-Zr MNCs released more Mg2+ions in HBSS,indicating lower corrosion resistance as compared to the batch-2 Mg-Zr-Zn MNCs,as shown in Fig.13b.Also,the Mg2+ion concentration increased with increasing GNP content in the Mg-Zr matrices of the batch-1 nanocomposites,whereas no significant difference in Mg2+was observed in the Zn-containing batch-2 nanocomposites.The measured Mg2+ion concentrations were 7.3,8.0,8.4,and 9.2 mg/L for the batch-1 Mg0.5ZrxGNP wherex=0.2,0.3,0.4,and 0.5,respectively,and 6.6,5.7,5.6,and 6.0 mg/L for the batch-2 Mg0.5ZrxZn0.1GNP wherex=1,3,4,and 5,respectively.The Zn2+ion concentration was significantly lower compared to the Mg2+ion concentration in HBSS for the batch-2 nanocomposites.This may be attributable to the addition of low concentrations (1-5 wt.%) of Zn to Mg and the formation of Zn from Zn2+ions by receiving 2e-from Mg in further reactions during degradation [91].However,there was no significant difference in Zn2+ion concentration (mg/L) observed in HBSS after 24 h degradation of the batch-2 nanocomposites.The Zn2+ions were measured as 0.3,0.31,0.36,and 0.35 for the batch-2 Mg0.5ZrxZn0.1GNP wherex=1,3,4,and 5,respectively.Figs.13ci-civ,and 13di-div show the SEM morphologies of the corroded surfaces of the batch-1 and batch-2 nanocomposites,respectively,after 24 h degradation.The batch-1 nanocomposites exhibited micron-sized pillars surrounded by large holes indicating high levels of galvanic attack in HBSS,as shown in Figs.13ci,cii,ciii,and civ for the Mg0.5Zr0.2GNP,Mg0.5Zr0.3GNP,Mg0.5Zr0.4GNP,and Mg0.5Zr0.5GNP,respectively,whereas the corroded surfaces of the batch-2 nanocomposites exhibited uniform pitting corrosion with some cracks on the surface as shown in Figs.13di,dii,diii,and div for the Mg0.5Zr1Zn0.1GNP,Mg0.5Zr3Zn0.1GNP,Mg0.5Zr4Zn0.1GNP,and Mg0.5Zr5Zn0.1GNP,respectively.The formation of cracks on the corroded surfaces may be attributed to the development of internal stresses during degradation [21].

Conclusions

In this study,Mg0.5Zr and Mg0.5ZrxZn (x=1-5 wt.%)MNCs reinforced with different contents of GNPs (0.1-0.5 wt.%) were fabricated via a powder metallurgy process.The mechanical and corrosion behaviors of these nanocomposites were evaluated in conjunction with the underlying strengthening mechanisms.Based on the experimental results along with theoretical analysis,the study can be concluded as follows:

1 The ultimate compressive strength,yield strength,ductility,and hardness of the Mg-Zr matrices were adversely affected by increasing the GNP content.Addition of 0.2 wt.% GNPs to the Mg-Zr matrices gave higher nano-hardness and elastic modulus as compared to the 0.3-0.5 wt.% GNP-reinforced nanocomposites.

2 Addition of 1 and 3 wt.% Zn to the Mg0.5Zr0.1GNP resulted in higher compressive yield strength and elastic modulus as compared to the nanocomposites containing higher 4 and 5 wt.% Zn content.The Mg0.5Zr5Zn0.1GNP showed significantly higher ductility (?23%) as compared to the other batch-2 nanocomposites containing lower Zn content (＜5 wt.%).

3 The deviation between experimentalσCYSandσpredictedincreased with increasing GNP contents in the Mg-Zr matrices,whereas the Mg0.5Zr1Zn0.1GNP and Mg0.5Zr3Zn0.1GNP nanocomposites showed 29% and 34% higher compressive yield strength,respectively,as compared to the predicted compressive yield strength of the Mg0.5Zr0.1GNP nanocomposite (164 MPa).

4 The batch-2 Mg-Zr-Zn MNCs reinforced with 0.1 wt.%GNPs had significantly higher Rpresulting in lower Icorrand corrosion rate in HBSS as compared to the batch-1 Mg-Zr MNCs containing 0.2-0.5 wt.% GNPs.The polarization resistance decreased with increasing GNP content in the Mg-Zr matrices.

5 The Mg0.5Zr3Zn0.1GNP revealed lower Icorrthus showing higher corrosion resistance as compared to the nanocomposites containing higher Zn content (4-5 wt.%).

6 The batch-2 Mg-Zr-Zn MNCs released fewer Mg2+ions during their degradation in HBSS as compared to the batch-1 Mg-Zr MNCs.

7 Based on the mechanical and corrosion properties,the novel Mg0.5Zr3Zn0.1GNP nanocomposite can be considered as a promising candidate material for biomedical applications.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could appear to have influence the work reported in this paper.

Acknowledgment

The authors acknowledge the financial support for this research by the Australian Research Council(ARC)through the Future Fellowship (FT160100252) and the Discovery Project(DP170102557).The authors also acknowledge the scientific and technical assistance of RMIT University’s Microscopy and Microanalysis Facility (RMMF),a linked laboratory of the Australian Microscopy &Microanalysis Research Facility.