In vitro and in vivo corrosion, cytocompatibility and mechanical properties of biodegradable Mg–Y–Ca–Zr alloys as implant materials

Acta Biomaterialia 9 (2013) 8518–8533

Contents lists available at SciVerse ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

In vitro and in vivo corrosion, cytocompatibility and mechanical properties of biodegradable Mg–Y–Ca–Zr alloys as implant materials q Da-Tren Chou a, Daeho Hong a, Partha Saha a, Jordan Ferrero a, Boeun Lee a, Zongqing Tan b, Zhongyun Dong b, Prashant N. Kumta a,c,d,e,f,g,⇑ a

Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA 15213, USA Department of Internal Medicine, University of Cincinnati, Cincinnati, OH 45221, USA Department of Mechanical Engineering and Materials Science, University of Pittsburgh, Pittsburgh, PA 15213, USA d Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh, PA 15213, USA e Center for Craniofacial Regeneration, University of Pittsburgh, Pittsburgh, PA 15261, USA f Center for Complex Engineered Multifunctional Materials, University of Pittsburgh, Pittsburgh, PA 15261, USA g McGowan Institute of Regenerative Medicine, University of Pittsburgh, Pittsburgh, PA 15261, USA b c

a r t i c l e

i n f o

Article history: Received 8 November 2012 Received in revised form 14 June 2013 Accepted 18 June 2013 Available online 27 June 2013 Keywords: Magnesium Biodegradable implant Mg–Y–Ca–Zr alloy Bone replacement material Biocompatibility

a b s t r a c t This study introduces a class of biodegradable Mg–Y–Ca–Zr alloys novel to biological applications and presents evaluations for orthopedic and craniofacial implant applications. Mg–Y–Ca–Zr alloys were processed using conventional melting and casting techniques. The effects of increasing Y content from 1 to 4 wt.% as well as the effects of T4 solution treatment were assessed. Basic material phase characterization was conducted using X-ray diffraction, optical microscopy and scanning electron microscopy. Compressive and tensile tests allowed for the comparison of mechanical properties of the as-cast and T4-treated Mg–Y–Ca–Zr alloys to pure Mg and as-drawn AZ31. Potentiodynamic polarization tests and mass loss immersion tests were used to evaluate the corrosion behavior of the alloys. In vitro cytocompatibility tests on MC3T3-E1 pre-osteoblast cells were also conducted. Finally, alloy pellets were implanted into murine subcutaneous tissue to observe in vivo corrosion as well as local host response through H&E staining. SEM/ EDS analysis showed that secondary phase intermetallics rich in yttrium were observed along the grain boundaries, with the T4 solution treatment diffusing the secondary phases into the matrix while increasing the grain size. The alloys demonstrated marked improvement in mechanical properties over pure Mg. Increasing the Y content contributed to improved corrosion resistance, while solution-treated alloys resulted in lower strength and compressive strain compared to as-cast alloys. The Mg–Y–Ca–Zr alloys demonstrated excellent in vitro cytocompatibility and normal in vivo host response. The mechanical, corrosion and biological evaluations performed in this study demonstrated that Mg–Y–Ca–Zr alloys, especially with the 4 wt.% Y content, would perform well as orthopedic and craniofacial implant biomaterials. Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Currently, biomaterials used for orthopedic and craniofacial applications are primarily chosen based on their ability to withstand cyclic load-bearing [1]. Metallic biomaterials, such as stainless steels, Ti and Co–Cr-based alloys, possess stiffness, rigidity and strength far exceeding those of natural bone. Their elastic moduli differ significantly from bone, causing stress-shielding effects that may lead to reduced loading of bone and decreased implant stability [2]. Current metallic biomaterials also suffer from the risk

q Part of the Biodegradable Metals Conference 2012 Special Issue, edited by Professor Frank Witte and Professor Diego Mantovani. ⇑ Corresponding author at: Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA 15213, USA. Tel.: +1 412 648 0223; fax: +1 412 624 3699. E-mail address: [email protected] (P.N. Kumta).

of releasing toxic metallic ions and particles through corrosion or wear [3–7], causing implant site immune response. They may also lead to hypersensitivity [8], growth restriction (most significantly for pediatric implants) [9], implant migration and imaging interferences [10]. Due to these complications, it is estimated that 10% of patients will require a second operation for the removal of permanent metallic plates and screws [11], exposing patients to additional risks, and increasing surgical time and resources. In order to avoid complications associated with permanent bone fixation implants, degradable biomaterials have recently been developed. However, resorbable polymer fixation plates and screws are relatively weak and less rigid compared to metals [12], and have demonstrated local inflammatory reactions [13]. Magnesium alloys have recently emerged as a new class of biodegradable materials for orthopedic applications with more comparable properties to natural bone [14]. They retain a density

1742-7061/$ - see front matter Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2013.06.025

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

similar to cortical bone, and much less than other implant metals. The elastic modulus of magnesium (41–45 GPa) [14] is much closer to natural bone (6–24 GPa) [15] compared to other commonly used metallic implants, thus reducing the risk of stress shielding. Most importantly, Mg degrades to produce a soluble, non-toxic corrosion hydroxide product which is harmlessly excreted through urine. Unfortunately, accelerated corrosion of Mg alloys may lead to accumulation of gas pockets around the implant as well as insufficient mechanical performance and implant stability throughout the degradation and tissue healing process [16,17]. Much research has therefore been focused on controlling the corrosion rate and improving mechanical properties of magnesium alloys through the introduction of alloying elements and use of different processing conditions. Magnesium corrosion and mechanical properties are strongly affected by alloying elements present in the solid solution [16,18]. In this study, the elements yttrium (Y), calcium (Ca) and zirconium (Zr) were alloyed with Mg to create a class of Mg alloys previously unexplored for biological applications. Mg– Y–Ca–Zr alloys have been pursued as ignition-proof alloys to improve oxidation resistance [19,20]. Y contributes to grain boundary strengthening of magnesium alloys [21] and also improves corrosion resistance when alloyed with Mg above 3 wt.% [22,23]. Ca is a major mineralized component in bone [24], and is known to improve corrosion resistance and mechanical properties of pure Mg up to 1 wt.% addition [25,26]. Density functional theory has shown that alloying with Ca and Y helps to form a stable and chemically less reactive hydroxide layer to impart greater corrosion resistance [27]. Zr serves as an effective grain-refining agent [28–31], imparting grain boundary strengthening [32] and corrosion resistance [33]. Addition of Ca to Zr-containing alloys promotes solutal undercooling, which in turn facilitates a greater number of suitable size Zr nuclei to become effective nucleation sites, thus further refining grain size [19,28–31]. The new alloys reported in this study, Mg–1Y–0.6Ca–0.4Zr (wt.%), denoted henceforth as WX11 (nomenclature according to ASTM B27505) [34], and Mg–4Y–0.6Ca–0.4Zr (wt.%), denoted henceforth as WX41, were compared in their as-cast and T4 solution-treated conditions based on their cytocompatibility, corrosion behavior and mechanical properties with the objective of using in orthopedic medical implants. Cytocompatibility was determined in vitro using direct and indirect cell viability tests and in vivo by implanting the alloys into mouse subcutaneous tissue and analyzing the local host response. Corrosion behavior was also evaluated by electrochemical and mass loss tests in vitro and in vivo. Mechanical properties were measured under both compression and tension.

2. Materials and methods 2.1. Materials preparation and characterization Elemental ingots of Mg (US Magnesium Inc., Salt Lake City, UT, 99.97%), Ca (Alfa-Aesar, Ward Hill, MA, 99.5%) and Y (Alfa-Aesar, 99.9%) were weighed according to the nominal composition and melted together in a graphite crucible inside an induction furnace (MTI Corporation, Richmond, CA) purged with ultrahigh-purity (UHP) Ar and vacuumed to avoid oxidation of the pure elements. The initial alloy produced by induction melting was cleaned thoroughly to remove residue and oxide scale and re-melted in a mild steel crucible using an electrical resistance furnace (Wenesco Inc., Chicago, IL) under the protection of Ar +1.5% SF6 cover gas. The melting and pouring temperature was 780 °C, and once the temperature was reached, an equivalent amount of zirconium was added using ZirÒ max (Mg–33.3% Zr) master alloy (Magnesium Elektron Ltd, Manchester, UK). After zirconium addition, the melt was stirred for 10 s at intervals of 1 and 5 min to dissolve and disperse the zirconium par-

8519

ticles uniformly into the melt. The melt was held for 30 min and poured into a cylindrical mild steel mold preheated to 500 °C with dimensions of 44.5 mm diameter  82.5 mm length. Appropriate stirring and holding times was essential to release Zr particles uniformly from ZirmaxÒ master alloy in order to achieve high solubility of Zr in the melt and optimal grain refinement [35]. The as-cast samples were solution-treated (T4) at 525 °C for 6 h inside a tubular furnace covered under continuous UHP Ar flow and quenched in water. The alloy nominal compositions, determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, iCAP duo 6500 Thermo Fisher, Waltham, MA), are listed in Table 1. The WX11 and WX41 alloys were compared to as-cast Mg (US Magnesium Inc.) and asdrawn AZ31 (Goodfellow Corp., Coraopolis, PA) in reported tests. In order to determine the phase formation, X-ray diffraction (XRD) was conducted using a Philips X’Pert PRO diffractometer employing Cu Ka (k = 1.54056 Å) radiation with a Si detector (X’celerator). The X-ray generator operated at 45 kV and 40 mA at a 2h range of 10–80°. 2.2. Microstructure characterization Square plate samples (10 mm  10 mm  1 mm) of the WX11 and WX41alloys were mounted in epoxy, mechanically polished (Tegramin-20, Struers, Ballerup, Denmark), and chemically etched in a solution of 5 ml acetic acid, 6 g picric acid, 10 ml water and 100 ml ethanol. The microstructure was observed using optical microscopy (Axiovert 40 MAT, Carl Zeiss, Jena, Germany) and scanning electron microscopy (SEM; JEOL JSM-6610, JEOL Ltd, Tokyo, Japan) with energy-dispersive X-ray (EDX; EDAX Genesis, Mahwah, NJ) to conduct elemental analysis. Average grain size was measured according to ASTM E112 [36] following the Abrams three-circle procedure. 2.3. Mechanical characterization Samples were machined along the long axis of the Mg alloy ingots in dimensions in accordance with ASTM-E8-04 [37] for tensile testing and ASTM-E9-09 [38] for compressive testing. Tensile bar samples with a gage area of 3  3 mm were machined for tensile samples. Cylindrical samples of 10 mm diameter  20 mm length were machined for compressive tests. Tensile and compressive stress–strain curves were obtained for each as-cast and T4 solution-treated alloy, and compared to as-cast pure Mg and as-drawn AZ31. The tensile and compressive tests were conducted by Ortho Kinetic Testing Technologies, LLC (Southport, NC) at room temperature using an MTS11 – 50 kN electro-mechanical load frame (MTS, Eden Prairie, MN) with laser extensometer. Tensile tests were carried out at a cross-head speed of 1.3 mm min–1, while compression tests were carried out at a speed of 2 mm min–1. Tensile and compressive yield strength, ultimate compressive and tensile strength, Young’s modulus (E), percentage elongation and compression (%) were determined from the stress–strain curves. An average and standard deviation of at least three measurements was taken for each group, with the exception of T4-treated WX11, for which only one sample was available for testing. 2.4. Electrochemical corrosion test To test corrosion of the WX alloys, the potentiodynamic polarization technique was used. Samples were connected to a copper wire using silver paste and mounted in epoxy resin. The mounted samples of dimensions 10 mm  10 mm  1 mm were mechanically polished, sonicated in isopropyl alcohol and dried in air. The potentiodynamic polarization was carried out with an electrochemical workstation (CH-604A, CH Instruments, Inc., Austin, TX) at a scanning rate of 1 mV s–1 and potential window of 500 mV

8520

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Table 1 Chemical composition of Mg–Y–Ca–Zr alloys as measured by ICP-OES. Alloy

Mg–1Y–0.6Ca–0.4Zr Mg–4Y–0.6Ca–0.4Zr

Chemical composition (wt.%) Y

Ca

Zr

Cu

Fe

Mn

Ni

Si

Mg

0.66 ± 0.03 3.18 ± 0.10

0.52 ± .01 0.63 ± 0.41

0.13 ± 0.004 0.074 ± 0.013

0.016 0.015

0.003 0.009

0.008 0.005

0.008 0.003

0.006 0.007

Balance Balance

above and below the open circuit potential. A three-electrode cell was employed with platinum as the counter electrode, Ag/AgCl as the reference electrode and the sample as the working electrode. The test was performed in Dulbecco’s modified Eagle medium (DMEM, with 4.5 g l–1 glucose, L-glutamine and sodium pyruvate, Cellgro, Manassas, VA) supplemented with 10% fetal bovine serum (FBS), 100 U ml–1 penicillin and 100 lg ml–1 streptomycin at pH 7.2 ± 0.2 and held at 37.4 °C. Before each measurement, the sample was immersed in the corrosion medium to provide stability. The cathodic and anodic portions of the generated Tafel plots were fitted linearly to allow calculation of corrosion potential, Ecorr, and corrosion current density, icorr. Samples were cleaned by immersion in 200 g l–1 of chromic acid and 10 g l–1 of AgNO3 for 10 min to remove corrosion products and corrosion morphology was characterized using SEM and EDX. 2.5. Immersion corrosion test Corrosion of the Mg alloys was also measured using mass loss in a solution of media-volume –to-surface-area ratio of 50 ml cm2, a ratio where mass loss was observed to stabilize [39]. Immersion tests of cylindrical samples of 5 mm diameter  2 mm thickness polished to 1200 grit were carried out in the same media as the electrochemical corrosion test at 37 °C in a humidified atmosphere with 5% CO2. Samples were removed after 1, 2 and 3 weeks of immersion and dried at room temperature. The sample masses were measured after immersion in 200 g l–1 of chromic acid and 10 g l–1 of AgNO3 for 10 min to remove corrosion products. The corrosion rate was calculated according to ASTM G31-72 [40] with the following equation:

C ¼ ðK  WÞ=ðA  T  DÞ

10% extract solutions. Extracts were sterile filtered through 0.2 lm syringe filters before being added to cells. The murine osteoblastic cell line (MC3T3-E1, American Type Culture Collection, Rockville, MD) was used for in vitro cell cytotoxicity experiments and cultured in aMEM, 10% FBS, 100 U ml–1 penicillin and 100 lg ml–1 streptomycin at 37 °C in a humidified atmosphere with 5% CO2. The cells were seeded in 96-well cell culture plates at 6  103 cells/200 ll medium in each well and incubated for 24 h. Culture medium without extract served as the negative control and 10% DMSO culture medium as the positive control. After 24 h incubation, medium was replaced with 200 ll of extraction medium with varying concentrations and incubated for 1 and 3 days. The cytotoxicity of the corrosion extracts was tested using the MTT assay. Media and extracts were replaced with fresh cell culture medium to prevent interference of the magnesium in the extract with the tetrazolium salt [42]. The MTT assay was performed according to the Vybrant MTT Cell Proliferation Kit (Invitrogen Corp., Karlsruhe, Germany) by first adding 10 ll of 12mM3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dissolved in phosphate buffer solution (PBS) to each well. The samples were incubated with MTT for 4 h, after which 100 ll formazan solubilization SDS-HCl solution was added to each well and incubated for 12 h. The absorbance of the samples was measured using the Synergy 2 Multi-Mode Microplate Reader (BioTek Instruments, Winooski, VT) at a wavelength of 570 nm. The absorbance of the samples was divided by the absorbance of the mean positive control subtracted from the mean negative control

ð1Þ 1

where C is the corrosion rate (mm year , mmpy), the constant K is 8.76  104, W is the mass loss (g), A is the sample area exposed to solution (cm2), T is the time of exposure (h) and D is the density of the material (g cm–3). An average and standard deviation of three measurements was taken for each group. Corrosion morphology was characterized using SEM (JEOL and Philips-XL30FEG, Philips, Amsterdam, the Netherlands) and EDX before and after the removal of corrosion products. Magnesium concentration in the immersion media was measured using ICP-OES after each time point. 2.6. Indirect cytotoxicity tests Mg–Y–Ca–Zr alloy samples and as-cast pure magnesium sectioned into cylindrical samples of 5 mm diameter  5 mm length were polished up to 1200 grit, ultrasonically cleaned in isopropyl alcohol, air-dried and sterilized by ultraviolet radiation for 1 h. The specimens were incubated in modified Eagle’s medium alpha (aMEM; Life Technologies, Carlsbad, CA) supplemented with 10% FBS, 100 U ml–1 penicillin and 100 lg ml–1 streptomycin at 37 °C in a humidified atmosphere with 5% CO2 for 72 h. The sampleweight-to-extraction-medium ratio was 0.2 g ml–1 in accordance with the EN ISO standard 10933:12 [41]. This starting extraction ratio was designated as 100% extract, with less concentrated extracts prepared by diluting the 100% extract into 50%, 25% and

Fig. 1. XRD patterns of as-cast pure Mg, WX11 and WX41.

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

8521

Fig. 2. Optical micrographs of Mg–Y–Ca–Zr alloys after polishing and etching: (a) WX11 as-cast; (b) WX11 T4 heat treated; (c) WX41 as-cast; (d) WX41 T4 heat treated; (e) pure Mg as-cast; (f) AZ31 as-drawn.

to determine percentage viability of cells compared to the controls, where cells cultured with fresh media constituted 100% cell viability. An average and standard deviation of five samples was taken for each group.

2.7. Direct cell viability and adhesion test MC3T3-E1 cells were cultured directly on Mg–Y–Ca–Zr alloys and as-cast pure magnesium. Cell culture conditions and media were the same as in the indirect cytotoxity test. Samples had dimensions of 10 mm  10 mm  1 mm and were polished up to 1200 grit, ultrasonically cleaned in isopropyl alcohol, air-dried, and UV sterilized for 1 h. The alloy samples were incubated in cell culture media for 10 min, after which cells were seeded on the samples at a density of 4  104 cells ml–1. Cell viability was evaluated at 1 and 3 days using the LIVE/DEAD Viability/Cytotoxicity Kit (Invitrogen) following the manufacturer’s protocol. This kit determines cell viability by differentiating between live and dead cells with fluorescence microscopy of two different wavelengths. Briefly, the Mg–Y–Ca–Zr samples with attached MC3T3-E1 cells were washed with PBS and stained for 30 min at room temperature with 2 lmol l–1 ethidium homodimer-1 and 4 lmol l–1 calcein AM in PBS. After incubation in the LIVE/DEAD solution for 30 min in

room temperature, live cell and dead cell images were captured using fluorescence microscopy.

2.8. In vivo murine subcutaneous study All experimental procedures for the murine subcutaneous study were approved by the Animal Care and Use Committee (IACUC) at the University of Cincinnati. Healthy nude mice were housed under controlled conditions and maintained with a standard pellet diet and water. Mice were anesthetized using isoflurane through a nosecone. A skin incision was made to create a subcutaneous pocket on the back of the mouse. Pure Mg, AZ31 and WX11 and WX41 as-cast alloys of dimensions 5 mm diameter  1.4 mm thickness were inserted into the pocket and the incisions were closed by surgical staples. After 7, 40 and 70 days, the mice were sacrificed in a CO2 chamber followed by cervical dislocation. The Mg/Mg alloy implants with surrounding tissue were recovered, carefully separated from the tissue, cleaned, air-dried and the final mass was measured to determine change in mass due to corrosion to calculate corrosion rate according to Eq. (1). The tissue was fixed in 10% formalin in PBS, paraffin-embedded and sectioned (4 lm per section) for hematoxylin–eosin (H&E) staining. 70 day explanted samples were also imaged using SEM and analyzed using EDX before and after removal of corrosion products.

8522

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Fig. 3. SEM images of Mg–Y–Ca–Zr alloys after polishing and etching: (a) WX11 as-cast; (b) WX11 T4 heat treated; (c) WX41 as-cast; (d) WX41 T4 heat treated with corresponding EDX of grain boundaries and precipitates as noted by arrows.

2.9. Statistical analysis Statistical analysis was conducted with SPSS Statistics 17.0 (SPSS Inc., Chicago, IL). Differences between groups were analyzed using one-way ANOVA followed by Tukey’s test when group sizes were equal or Gabriel’s pairwise test when group sizes were unequal. P < 0.05 was accepted as a statistically significant difference between means and is denoted in figures. Error bars within figures represent standard deviation. 3. Results 3.1. Phase and microstructure characterization of Mg–Y–Ca–Zr alloys Table 1 shows that the final composition of each alloy determined by ICP analysis was close to nominal composition, with reduction in alloying element content likely due to the re-melting process. The reduction in Zr was primarily due to settling of large zirconium particles and clusters in the liquid melt [43]. Importantly, the impurity levels of each alloy compositions was low, reducing the risk of rapid corrosion due to impurities being below their tolerance limit [44]. The phase formation of the WX11 and WX41 alloys were characterized by XRD as shown in Fig. 1. The XRD patterns show that all the alloys were composed of hexagonal close packed (hcp) a-Mg, without X-ray detecting unalloyed Y, Ca, Zr or intermetallic phases. Fig. 2 shows the optical microstructure of the as-cast and solution-treated WX11 and WX41 alloys containing grains of a-Mg with secondary phases located at the grain boundaries and as precipitates within the matrix. The average grain sizes of WX11 as-cast, WX11 T4-treated, WX41 as-cast and WX41 T4-treated alloys were 79, 98, 98 and 200 lm, respectively, with the presence of uniform equiaxed a-Mg grains throughout the microstructure. Addition of the alloying elements greatly reduced the grain size compared to high-purity Mg (Fig. 2e), which lacked the presence of secondary phases. AZ31 purchased from Goodfellow Corp., having undergone grain refinement through drawing, exhibited much finer grain size (Fig. 2f) compared to the Mg–Y–Ca–Zr alloys and

pure Mg. SEM and EDX (Fig. 3) revealed that the WX11 and WX41 as-cast alloys consisted of grain boundaries containing higher amounts of Y and Ca (Fig. 3a and c), indicating the presence of secondary phases due to precipitate segregation during solidification, a common phenomenon prevalent in casting. Dissolution of these secondary phases at the grain boundaries into the bulk a-Mg was observed in T4 solution-treated alloys (525 °C for 6 h followed by water quenching), leading to homogenization of the as-cast alloy. Secondary phase of Y-rich intermetallic particles were also observed, which were still present after T4 solution treatment (Fig. 3b and d). More second phase particles were also present in the WX41 alloys due to higher Y content. The increase in grain size after T4 solution treatment was likely due to coalescence of smaller grains along the triple point grain boundary regions and the formation of supersaturated a-Mg solid solution after the precipitates dissolved into the matrix. 3.2. Mechanical properties of Mg–Y–Ca–Zr alloys Compression and tensile mechanical properties of the reported alloys are presented in Fig. 4. Ultimate compressive strains (Fig. 4a) of as-cast WX11 and WX41 alloys were significantly larger than those of as-cast pure Mg and as-drawn AZ31. Ultimate compressive strength (Fig. 4a) for the as-cast experimental alloys was also greater than for as-cast pure Mg, while AZ31 exhibited significantly greater ultimate compressive strength over all the other materials tested due to the work hardening and grain refinement imparted on it. T4 solution treatment resulted in a significant reduction in compressive strength and strain, while an increase in compressive yield strength (Fig. 4b) was observed with an increased Y content. As-cast WX41 demonstrated significantly higher ultimate tensile strength (Fig. 4c) than as-cast WX11. Similar to compression test results, applying the T4 solution treatment to the Mg–Y–Ca–Zr alloys resulted in a reduction in the ultimate tensile strength and tensile yield strength compared to the as-cast alloys, while no significant effect on tensile strain could be observed after heat treatment. AZ31 displayed much higher tensile strength and

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

8523

Fig. 4. Mechanical properties of Mg–Y–Ca–Zr alloys, pure Mg, and AZ31. (a) Ultimate compressive strength and ultimate compressive strain; Significant difference (p < 0.05) between ⁄ and all other groups;  , à and other groups. (b) Compressive yield strength; significant difference (p < 0.05) between ⁄ and all other groups;  , à and other groups. (c) Ultimate tensile strength and ultimate tensile strain; significant difference (p < 0.05) between ⁄, § with all other groups;  , à with other groups. (d) Tensile yield strength and Young’s modulus; significant difference (p < 0.05) between ⁄; § with all other groups;  , à with other groups.

strain over the other tested materials. Tensile properties of the Mg–Y–Ca–Zr alloys were either greatly improved or observed to be comparable to those of pure Mg. Values of Young’s modulus (Fig. 4d) for the tested alloys varied between 34 and 60 GPa, similar to the measured value for AZ31 (42 GPa), with high variance within groups observed. 3.3. Corrosion behavior of Mg–Y–Ca–Zr alloys The corrosion rates of the as-cast and T4 solution-treated Mg– Y–Ca–Zr alloys compared to pure as-cast Mg and as-drawn AZ31 were calculated based on mass loss after immersion as well as by potentiodynamic polarization measurements, and are shown along with corrosion potential and concentration of Mg ions released into immersion solution in Fig. 5 and Table 2. The corrosion potential (Fig. 5b) of the WX41 alloys was higher than the WX11 alloys and the control materials, pure Mg and AZ31, respectively. T4 solution-treated alloys also demonstrated a higher corrosion potential and breakdown potential compared to their as-cast counterparts. Potentiodynamic corrosion rates of the WX41 alloys were lower than the WX11 alloys and were similar to those of AZ31. Solution

treatment increased potentiodynamic corrosion rate of the WX11 alloy, though it did not affect WX41. Fig. 5c shows that the alloy corrosion rate calculated from the mass loss immersion test appeared to start stabilizing after 2 weeks of immersion. After 2 weeks of immersion, the corrosion rate of WX41 as-cast was significantly lower than those of the WX11 alloys. The 2 week corrosion rates of the WX41 alloys were also not significantly different from high-purity Mg. After 3 weeks of immersion, once again the WX41 alloys demonstrated a lower corrosion rate compared to WX11, and similar to high purity Mg. T4 solution treatment did not appear to have a major effect on immersion corrosion rate. The corrosion medium used for the mass loss corrosion experiments was used to determine concentration of Mg released into solution after 1, 2 and 3 weeks (Fig. 5c). The WX41 alloys again showed higher corrosion resistance compared to WX11, releasing a lower concentration of Mg at each timepoint, comparable to Mg and AZ31. Heat treatment did not affect Mg released. Fig. 6 shows the SEM micrographs of the Mg–Y–Ca–Zr alloys surface cleaned with CrO3/AgNO3 solution to remove the corrosion products after the potentiodynamic polarization test. All the samples demonstrated localized pitting corrosion, a commonly-seen

8524

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Fig. 5. Corrosion properties of Mg–Y–Ca–Zr alloys, pure Mg and AZ31 in DMEM with 10% FBS. (a) Corrosion rate as measured using mass loss after 1, 2 and 3 weeks’ immersion and potentiodynamic corrosion. Significant difference (p < 0.05) between ⁄ and ⁄⁄; between   and à, between §. (b) Corrosion potential and breakdown potential. (c) Concentration of Mg released in corrosion media after 1, 2 and 3 weeks’ immersion. Significant difference (p < 0.05) between ⁄ and ⁄⁄;   and à.

phenomena in Mg alloys [18] including Y-containing alloys [45], while the as-cast alloys (Fig. 6a and c) also show corrosion occurring at corrosion-prone grain boundary regions (arrows) due to higher secondary phase localization. Micrographs of the dried samples after 3 weeks’ immersion in DMEM with 10% FBS were captured using SEM to further assess corrosion products (Fig. 7). Localized corrosion was observed with deposits of corrosion product rich in C and O on the corroded surface (EDX shown in Fig. 7a and b). The region immediately surrounding these corrosion deposits was generally rich in Mg (Fig. 7a and e). Finally, most of the surface was covered in a layer rich in O, Ca, P and Mg. Ca and P are essential components in bone, and have been reported as forming a layer of amorphous calcium phosphate on rare-earth (RE)-containing alloys after degradation in vivo [46]. After removal of corrosion products, the SEM shown in Fig. 8 revealed that the 3 week immersion corrosion samples exhibited linked-up corrosion cavities on all materials.

3.4. In vitro cytotoxicity of Mg–Y–Ca–Zr alloys Fig. 9 shows the indirect cytotoxicity results of WX11 samples performed using MC3T3-E1 cells and the MTT assay. For both culture periods, cell viability was expectedly most reduced with 100% extract concentration, and increased as the extract percentage decreased. After 1 day of culture with extracts (Fig. 9a), no cytotoxicity was observed for 25% and 10% extract concentrations as no reduction in cell viability was observed. After 3 days of culture (Fig. 9b), cell viability was reduced to above 70% at 25% and 10% extract concentrations. This is consistent with findings that show high extract concentrations are cytotoxic and lead to osmotic shock, suggesting that a ten-fold extract dilution should be used for as-cast magnesium materials [47]. WX11 as-cast also displayed significantly higher cell viability at 50% extract concentration compared to WX11 T4, WX41 as-cast pure Mg. After 1 day of culture with extract, WX11 and WX41 as-cast and T4-treated alloys

8525

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Table 2 Electrochemical and immersion corrosion measurements conducted in DMEM with 10% FBS at 37 °C of Mg–Y–Ca–Zr alloys, pure Mg, and AZ31 and murine subcutaneous corrosion rate after 40 and 70 days’ implantation. Material

Corrosion potential, Ecorr (V)

Corrosion current density, icorr (lA cm2)

Corrosion rate, mm year–1 (mmpy

WX11 as-cast WX11T4 WX41 as-cast WX41T4 Pure Mg AZ31

1.73 1.58 1.59 1.54 1.68 1.68

14.86 24.96 4.98 5.22 24.50 8.19

0.34 0.57 0.11 0.12 0.56 0.18

Immersion corrosion rate (mmpy)

Subcutaneous corrosion rate (mmpy)

Week 1

Week 2

Week 3

40 days

70 days

0.49 0.56 0.57 0.58 0.69 0.65

0.69 0.79 0.41 0.61 0.46 0.29

0.72 0.64 0.46 0.55 0.46 0.40

0.31

0.73

0.18

0.068

0.060 0.067

0.059 0.024

Fig. 6. SEM images showing surface morphologies of (a) WX11 as-cast; (b) WX11 T4 heat treated; (c) WX41 as-cast; (d) WX41 T4 heat treated; (e) pure Mg; and (f) AZ31 after potentiodynamic polarization in DMEM with 10% FBS at 37 °C and cleaning with CrO3/AgNO3 solution. Arrows denote corrosion at grain boundaries.

showed significantly higher cell viability compared to pure Mg at 25% extract concentration, though no difference between them could be observed after 3 days of culture. Fig. 10 shows pre-osteoblast MC3T3-E1 cells cultured directly on the WX11 and WX41 alloys for 1 and 3 days, and then stained with calcein-AM (green fluorescence in live cells) and ethidium homodimer-1 (red fluorescence in dead cells). After 1 and 3 days of culture, both WX11 and WX41 as-cast and T4 heat treated alloys (Fig. 10a–d and h–k) demonstrated comparable live cell density

compared to pure Mg (Fig. 10e and l) and AZ31 (Fig. 10f and m). Relatively few apoptotic cells compared to live cells were observed on each material, indicating generally good cell viability. Tissue culture plastic displayed higher cell viability compared to the Mg-based materials. No significant differences in cell morphology were observed among groups. Cell density was still high among the Mg–Y–Ca–Zr alloys after 3 days of culture compared to 1 day, indicating that the cells were not significantly affected by prolonged direct exposure times to the corrosion environment.

8526

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Fig. 7. SEM images showing surface morphologies of (a) WX11 as-cast; (b) WX11 T4 heat treated; (c) WX41 as-cast; (d) WX41 T4 heat treated; (e) pure Mg; (f) AZ31 after 3 weeks’ static immersion in DMEM with 10% FBS at 37 °C. EDX was performed at various spots as denoted by arrows.

3.5. In vivo biocompatibility of as-cast Mg–Y–Ca–Zr alloys H&E staining of the local site of implantation of Mg–Y–Ca–Zr ascast alloys, pure Mg, and AZ31 in the subcutis of nude mice is shown in Fig. 11a–l. Minimal toxicity to the surrounding tissue was observed, while the region surrounding the implants appeared to be undergoing normal tissue repair. No significant accumulation of inflammatory cells was observed, whereas a layer of tissue composed of pink-stained collagen fibers and reactive fibroblasts was seen after 7 days. At this time point, a relatively high density of fibroblasts in tissues adjacent to as-cast WX11 and WX41 alloy pellets suggests that their presence did not inhibit the normal healing response in the implantation site. After 40 and 70 days’ implantation, dense collagenous connective tissue was seen surrounding the location of the Mg implants, without the presence of a high density of chronic inflammatory cells. Normal adipocytes could be faintly distinguished past the dermis. In vivo corrosion as determined through mass loss (Table 2 and Fig. 11m) of as-cast WX41 was much lower than as-cast WX11 and similar to pure Mg and AZ31 after 70 days. This result is consistent with the in vitro corrosion behavior of the Mg–Y–Ca–Zr alloys, in which the higher Ycontaining alloy, WX41, corroded more slowly than WX11. The samples explanted after 70 days’ implantation were dried and imaged using SEM to assess corrosion products formed (Fig. 12) in comparison to those formed after static immersion (Fig. 7). After in vivo corrosion, corrosion products and layers rich

in C and O were observed to have formed on the alloy surface, similar to results from the immersion study. The cracked corrosion layer was seen to contain higher Ca and P content compared to agglomerates of corrosion deposits. Removal of corrosion products revealed widespread corrosion, resulting in irregular surface topography on all materials as seen in Fig. 13. 4. Discussions 4.1. Effect of alloying elements and heat treatment on microstructure and mechanical properties Alloying elements strongly affect the microstructure and mechanical properties of magnesium alloys. Higher Y content has been reported to result in grain coarsening [48], which was observed to a small degree in Fig. 2 when comparing the WX11 to the WX41 alloys. The measured Zr content as seen in Table 1 was also reduced in the WX41 alloy, which may also have contributed to the higher grain size compared to WX11. Y-rich intermetallic particles were observed in the Mg–Y–Zr–Ca alloys as has been observed in Mg–Y binary alloys [23], though implementation of T4 solution treatment resulted in dissolution of second phase precipitates from the grain boundaries as well as grain coarsening. Strength of Mg alloys is known to be sensitive to changes in grain size according to the Hall–Petch relation, where finer grains result in higher grain boundary strengthening [49,50]. Appropriately, the

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

8527

Fig. 8. SEM images showing surface morphologies of (a) WX11 as-cast; (b) WX11 T4 heat treated; (c) WX41 as-cast; (d) WX41 T4 heat treated; (e) pure Mg; (f) AZ31 after 3 weeks’ static immersion in DMEM with 10% FBS at 37 °C and cleaning with CrO3/AgNO3 solution.

Fig. 9. Cytotoxicity of MC3T3 cells cultured for (a) 1 day and (b) 3 days in extract media from as-cast and T4 heat treated WX11 and WX41 alloys and as-cast pure Mg. Significant difference (p < 0.05) between ⁄ and other groups or as connected.

as-cast Mg–Y–Ca–Zr alloys maintained higher strength in compression and tension and higher compressive strain compared to the alloys after grain coarsening occurred. The higher presence of secondary phases in the as-cast Mg–Y–Ca–Zr alloys may have contributed to precipitation strengthening by acting as impediments for dislocation movement during plastic deformation. This phenomenon as well as solid solution strengthening has been

utilized in RE containing alloys [32], and may explain the marginally higher tensile strengths for as-cast alloys containing a higher Y content, also observed in other studies comparing Mg–Y–Ca–Zr alloys [19]. As-drawn AZ31 possesses a fine grain size to meet mechanical property requirements for industrial applications [51], while the high purity Mg devoid of any alloying elements exhibited comparatively low mechanical properties, suffering from

8528

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Fig. 10. Fluorescent images of live (green) and dead (red) MC3T3-E1 cells attached after 1 and 3 days culture on (a, h) WX11 as-cast; (b, i) WX11T4 heat treated; (c, j) WX41 as-cast; (d, k) WX41T4 heat treated; (e, l) pure Mg; (f, m) AZ31 as-drawn; (g, n) tissue culture plastic.

a lack of solid solution strengthening or precipitation strengthening. The Young’s modulus of the tested alloys was similar to that of AZ31 and seems to be appropriate for orthopedic applications, being close to the modulus of natural bone of 6-24 GPa [15]. Fan et al. [20] measured mechanical properties of Mg–3Y–0.5Ca– 0.1Zr alloy, which yielded a tensile strength of 140 MPa and an elongation of 4%, very similar to those of the WX41 as-cast alloy. It is essential to note that the Mg–Y–Ca–Zr alloys presented here were not subject to any forming processes to impart further grain refinement. Due to their as-cast nature, the tested alloys likely contained casting defects such as inclusions, oxide scale, etc., which would require a forming process to remove and improve upon the mechanical properties. 4.2. Effect of alloying elements and heat treatment on corrosion behavior Corrosion resistance has been a major point of interest in magnesium alloy research due to severe pitting corrosion observed in Mg alloys when exposed to the physiological chloride environment beyond 150 mmol l–1 [52] as well as at sites of low hydrogen overpotential [18]. Microstructure and grain size again are critical in controlling corrosion, with reports of a reduction in grain size resulting in improved corrosion resistance. Aung and Zhou [53] demonstrated that increasing grain size through heat treatments at various temperatures of AZ31B-H24 resulted in an increase in corrosion rate as explained by the grain boundaries acting as a physical corrosion barrier. Fine grained alloy microstructures processed by friction stir processing of Mg–Y–RE [54] and equal channel angular pressing (ECAP) of AZ31 [55] also displayed better corrosion behavior than coarse grained alloys. However, conflicting studies have shown that a reduction in grain size leads to a

decrease in corrosion resistance, through the addition of a grainrefining agent [56] or ECAP [57]. In this study, T4 solution treatment was attempted to reduce the volume fraction of secondary phases by diffusion into the a-Mg matrix, thus limiting the extent of microgalvanic corrosion where the anodic a-Mg matrix has a lower corrosion potential than any secondary phase, corroding preferentially as a galvanic couple [58]. This was observed in comparing the corrosion surface of as-cast to solution heat treated alloys in Fig. 6, where the as-cast alloys with higher secondary phase concentration at the grain boundaries encountered visible corrosion at these grain boundaries, which was not seen in the solution-treated alloys. Differing the Y content in Mg alloys may affect the corrosion behavior as reported by Liu and colleagues [23]. By increasing the Y content, the volume fraction of Y-containing intermetallics increases, thereby enhancing the microgalvanic corrosion. On the other hand, Liu et al. [23] determined that increasing Y above 3% may also result in a more protective passivation layer, which was also determined theoretically [27]. This stronger passivation of the WX41 alloys compared to the WX11 alloys was evident in the higher corrosion potentials (Fig. 5b) of as-cast and T4-treated WX41 compared to the as-cast and T4-treated WX11. T4 solution treatment was also observed to shift the corrosion potential of the Mg–Y–Ca–Zr alloys to more noble potentials due to the diffusion of intermetallic particles precipitated along the grain boundaries into the solid solution, possibly allowing for the protective oxide layer to cover a larger region of the alloy surface, as suggested by Neil et al. [56] after heat treating Mg–Zn–RE–Zr alloys. After solution treatment, more positive breakdown potentials of the Mg–Y–Ca–Zr alloys were also observed, indicating a higher resistance to the breakdown of the passive film and onset of pitting corrosion [56]. The electrochemical corrosion rates (Fig. 5a) as calculated using corrosion current

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

8529

Fig. 11. (a–l) Histology images (H&E staining) of the skin above the implants of (a, e, i) WX11 and (b, f, j) WX41 as-cast alloys, (c, g, k) pure Mg, and (d, h, l) AZ31 after 7 days (a–d), 40 days (e–h), and 70 days (i–l) in the subcutis of nude mice. (m) Corrosion rate as calculated by mass loss of pellet samples before and after murine subcutaneous implantation at timepoints of 40 and 70 days.

density of the Mg–Y–Ca–Zr alloys decreased with an increase in Y content, which occurs through suppression of the cathodic reaction in the corrosion process [57]. Immersion corrosion (Fig. 5a) and Mg concentration measurements (Fig. 5c) confirmed this reduction in corrosion rate from WX11 to WX41 alloys. Solution treatment caused a large increase in the electrochemical corrosion rate of WX11, while only a slight increase in corrosion rate was observed for WX41. However, these trends were not reflected in the immersion corrosion and Mg concentration measurements, where T4 treatment did not result in a marked change in corrosion rate. This may have been due to a combination of T4 treatment dissolving the intermetallic phases and reducing microgalvanic galvanic corrosion, while also increasing grain size, which may retard passivation

kinetics [54]. The linked-up corrosion cavities seen on the surface of the alloys are commonly observed on corroded Mg alloys [59]. Overall, the WX41 alloys demonstrated favorable corrosion behavior relative to WX11 and pure Mg and AZ31, confirming the utility in Y addition for corrosion resistance. Corrosion rate calculated from potentiodynamic polarization curves were lower than those calculated from mass loss during immersion, possibly because the corrosion mechanism of Mg involves a fraction of the uni-positive Mg ion reacting chemically as well as electrochemically, causing electrochemical measurements to underestimate the corrosion rate compared to weight loss or hydrogen evolution methods [23,59]. Corrosion products developed over a 3 week immersion period in DMEM with 10% FBS

8530

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Fig. 12. SEM images showing surface morphologies of (a) WX11 as-cast; (b) WX41 as-cast; (c) pure Mg; (d) AZ31 after 70 days implantation in murine subcutaneous tissue. EDX was performed at various spots as denoted by arrows.

Fig. 13. SEM images showing surface morphologies of (a) WX11 as-cast; (b) WX41 as-cast; (c) pure Mg; (d) AZ31 after 70 days’ implantation in murine subcutaneous tissue and cleaning with CrO3/AgNO3 solution.

(Fig. 7) revealed large amounts of Ca and P in deposits (EDX in Fig. 7a and b), consistent with observations of aggregations of calcium-containing corrosion products [60]. 4.3. In vitro cytocompatibility The alloying elements used in this study have been shown in the literature to be biocompatible. Y has been shown to be non-toxic in longevity studies [61], non-hepatotoxic [62] and incorporated in alloys which were clinically well-tolerated as an absorbable metal stent [63]. Ca is an essential component in bone, and requires Mg for incorporation into bone [64]. Zr ions have low cytotoxicity

[65] and zirconium coatings have demonstrated improved osseointegration of metal implants in vivo [66]. In vitro cytocompatibility was assessed on the Mg–Y–Ca–Zr alloys by (1) exposing MC3T3 pre-osteoblast cells to media containing degradation products of the materials and analyzing cell viability through the colorimetric MTT assay, and (2) seeding cells directly onto the alloys and observing live and dead cell densities through the LIVE/DEAD assay. Fig. 9 presents the results from the MTT assay, in which at high extract concentrations (100% and 50%), cell viability was reduced; however, this was due to high extract concentrations causing cellular osmotic shock, suggesting a ten-fold extract dilution be

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

used for as-cast magnesium materials [47]. Indeed, at lower concentrations of extract (25% and 10%), high cell viability was observed for the alloys, indicating that higher concentrations of alloy degradation products resulted in lower cytotoxicity. Higher than 100% cell viability was observed from 25% and 10% extracts of WX11 and WX41 as-cast and T4-treated alloys after 1 day of culture, which may be facilitated by enhanced osteoblast activity in the presence of corrosion product magnesium hydroxide [67]. Similarly, a high density of live cells was observed on the surface of the Mg–Y–Ca–Zr alloys in Fig. 10. By direct culturing cells onto Mg-alloy substrates, this assay qualitatively shows the effect of hydrogen gas production, increasing pH and local concentration of corrosion products on cell attachment and viability. The differences observed in corrosion rate did not appear to have any bearing on the in vitro cytocompatibility results, with all Mg materials harboring a similar number of attached active cells, while appearing to have a lower cell density compared to tissue culture plastic. This is consistent with the 3 day MTT results at 25% and 10% extract concentrations in which cell viability was lower than the negative control. The elongated, spread morphology of the adherent cells on the Mg– Y–Ca–Zr alloys after both 1 and 3 days of culture confirmed the cytocompatibility of the alloys, which did not corrode too rapidly to inhibit cell attachment. In this study, cytocompatibility was conducted using MC3T3-E1 cells with samples immersed in aMEM+ 10% FBS, and the presented results may vary with the use of different cell lines and respective cell culture media. Cells cultured in other media such as DMEM may exhibit different cytocompatibility results compared to aMEM, due to the presence of higher L-glutamine content in DMEM, which functions as a chelating agent for magnesium ions [68], potentially causing increased Mg ion release from alloys immersed in DMEM [69]. Thus, the obtained cytocompatibility results presented here should be solely considered under cell culture conditions using aMEM as opposed to other media such as DMEM, which would likely result in a higher corrosion rate of tested alloys, applying a different local environment to cultured cells. 4.4. In vivo biocompatibility Past studies have shown acceptable host response and biocompatibility of RE-containing Mg alloys upon in vivo implantation [70]. Subcutaneous implantation of Mg–1.5% Nd–0.5% Y–0.5% Zr alloys with and without 0.4% Ca demonstrated adequate Mg metabolism, renal function and host–tissue response compared to the Ti–6Al–4 V control implant [71]. The murine subcutaneous study was conducted here in order to compare the in vivo corrosion and local tissue response of the selected materials: WX11 as-cast, WX41 as-cast, pure Mg and AZ31. Histological slides stained using H&E (Fig. 11a–l) showed that the Mg–Y–Ca–Zr alloys did not introduce toxicity or excessive inflammatory response to the surrounding tissue, presenting an acceptable host response with natural wound healing occurring as seen in the presence of a high density of fibroblasts which deposit collagen fibers in order to form new extracellular matrix during the tissue repair process [72]. This dense connective tissue was observed adjacent to the Mg alloy implants at all time points, indicating biocompatibility at varying magnitudes of alloy degradation. The in vivo mass loss corrosion rates (Table 2 and Fig. 11m) of the tested Mg materials were lower than those measured from in vitro immersion. Witte et al. [73] concluded that in vitro corrosion tests could not accurately predict in vivo corrosion rates after observing a four orders of magnitude reduction of corrosion rate in vivo compared to in vitro. Different environmental conditions, dynamic blood flow around the implant and variations in pH were among the speculative reasons provided to explain the discrepancy. The lower corrosion rates measured in vivo vs. in vitro was in agreement with

8531

studies of Mg alloys implanted intramedullary in guinea pig femurs [73] and in a subcutaneous environment in Lewis rats [74]. Explanted samples displayed the formation of similar corrosion products and morphology to samples from immersion testing. After 70 days’ implantation, the extent of in vivo corrosion is clearly greater than after 3 weeks in vitro, though corrosion products rich in C and O with the presence of Ca and P were still identified. The corrosion product morphology of a cracked corrosion layer with corrosion products deposited on top as seen both in vitro and in vivo is in agreement with results found in other studies [74,75]. Nonetheless, in vivo corrosion rate following 40 and 70 days of implantation for as-cast WX41 was lower than as-cast WX11 and comparable to pure Mg and AZ31, which was consistent with the measured in vitro corrosion rates discussed in Section 4.2. This result is in agreement with established knowledge that while in vitro corrosion tests currently cannot accurately predict those in vivo, they may serve as a relatively inexpensive screening method to assess the relative corrosion resistance of Mg alloys.

5. Conclusions The current study presents feasibility testing of Mg–Y–Ca–Zr alloys for orthopedic applications, analyzing the effects of increasing Y content from 1 to 4 wt.% and performing T4 solution treatment on the alloy performance criteria of mechanical properties, corrosion and cytocompatibility. The higher Y content present in WX41 contributed to an increase in grain size, higher corrosion resistance in terms of formation of a more stable passivation layer and reduced corrosion rate in in vitro and in vivo conditions, and slight increase in strength. T4 solution treatment resulted in grain coarsening, which may have contributed to lower compressive and tensile strength and compressive strain, thus an additional artificial aging step may be pursued in the future to produce higher strength alloys through precipitation hardening at the expense of ductility. The mechanical properties of the Mg–Y–Ca–Zr alloys showed improvement over pure Mg, and may be further improved through forming processes to refine grain size and reduced casting defects and porosity. Both WX11 and WX41 alloys displayed favorable in vitro cytocompatibility and in vivo host response compared to pure Mg and AZ31. Overall, the encouraging mechanical, biocorrosion and biological properties of the Mg–Y–Ca–Zr alloys, especially the WX41 alloy, are indicative of their potential as viable orthopedic and craniofacial implant biomaterials. Future studies to investigate the effects of hot extrusion and rolling with an in vivo load-bearing bone model will help further the current research described herein and will promote our ultimate aim to optimize Mg alloys for orthopedic medical implants.

Acknowledgements The authors would like to gratefully thank the assistance from Ortho Kinetic Technologies, LLC for conducting the mechanical tests. The authors would like to acknowledge the financial assistance of NSF-ERC Grant #EEC-0812348 for supporting this work. The authors also acknowledge the financial support of the Swanson School of Engineering, University of Pittsburgh, while PNK acknowledges the Edward R. Weidlein Chair Professorship, the Center for Complex Engineered Multifunctional Materials (CCEMM), and NSF–CBET Grant #0933153 for partial support of this research.

8532

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533

Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 10 and 11, is difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2013.06.025. References [1] Ratner BD, Hoffman AS, Schoen FJ, Lemons JE, Biomaterials science: an introduction to materials in medicine. 2nd ed. London: Elsevier Academic Press, 2004. [2] Huiskes R, Weinan H, Van Rietbergen B. The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res. 1992:124–34. [3] Puleo DA, Huh WW. Acute toxicity of metal ions in cultures of osteogenic cells derived from bone marrow stromal cells. J Appl Biomater 1995;6:109–16. [4] Jacobs JJ, Gilbert JL, Urban RM. Corrosion of metal orthopaedic implants. J Bone Joint Surg Ser A 1998;80:268–82. [5] Lhotka C, Szekeres T, Steffan I, Zhuber K, Zweymüller K. Four-year study of cobalt and chromium blood levels in patients managed with two different metal-on-metal total hip replacements. J Orthop Res 2003;21:189–95. [6] Jacobs JJ, Hallab NJ, Skipor AK, Urban RM. Metal degradation products: a cause for concern in metal–metal bearings? Clin Orthop Relat Res 2003:139–47. [7] Jacobs JJ, Skipor AK, Patterson LM, Hallab NJ, Paprosky WG, Black J, et al. Metal release in patients who have had a primary total hip arthroplasty: a prospective, controlled, longitudinal study. J Bone Joint Surg Ser A 1998;80:1447–58. [8] Orringer JS, Barcelona V, Buchman SR. Reasons for removal of rigid internal fixation devices in craniofacial surgery. J Craniofac Surg 1998;9:40–4. [9] Yaremchuk MJ, Fiala TGS, Barker F, Ragland R. The effects of rigid fixation on craniofacial growth of rhesus monkeys. Plast Reconstr Surg 1994;93:1–10. [10] Sullivan PK, Smith JF, Rozzelle AA. Cranio-orbital reconstruction: safety and image quality of metallic implants on CT and MRI scanning. Plast Reconstr Surg 1994;94:589–96. [11] Gilardino MS, Chen E, Bartlett S. Choice of internal rigid fixation materials in the treatment of facial fractures. Craniomaxillofac Trauma Reconstruc 2009;2:49–60. [12] Cordewener FW, Schmitz JP. The future of biodegradable osteosyntheses. Tissue Eng 2000;6:413–24. [13] Bell RB, Kindsfater CS. The use of biodegradable plates and screws to stabilize facial fractures. J Oral Maxillofac Surg 2006;64:31–9. [14] Staiger MP, Pietak AM, Huadmai J, Dias G. Magnesium and its alloys as orthopedic biomaterials: a review. Biomaterials 2006;27:1728–34. [15] Reilly DT, Burstein AH. The mechanical properties of cortical bone. J Bone Joint Surg 1974;56:1001–22. [16] Witte F. The history of biodegradable magnesium implants: a review. Acta Biomater 2010;6:1680–92. [17] Gu XN, Zhou WR, Zheng YF, Cheng Y, Wei SC, Zhong SP, et al. Corrosion fatigue behaviors of two biomedical Mg alloys – AZ91D and WE43 – in simulated body fluid. Acta Biomater 2010;6:4605–13. [18] Song GL, Atrens A. Corrosion mechanisms of magnesium alloys. Adv Eng Mater 1999;1:11–33. [19] Chang S-Y, Tezuka H, Kamio A. Mechanical properties and structure of ignition-proof Mg–Ca–Zr alloys produced by squeeze casting. Mater Trans JIM 1997;38:526–35. [20] Fan J, Chen Z, Yang W, Fang S, Xu B. Effect of yttrium, calcium and zirconium on ignition-proof principle and mechanical properties of magnesium alloys. J Rare Earths 2012;30:74–8. [21] Wu A-r, Xia C-q. Study of the microstructure and mechanical properties of Mgrare earth alloys. Mater Des 2007;28:1963–7. [22] J-c Gao, Wu S, Qiao L-y, Wang Y. Corrosion behavior of Mg and Mg–Zn alloys in simulated body fluid. Trans Nonferrous Met Soc China 2008;18:588–92. [23] Liu M, Schmutz P, Uggowitzer PJ, Song G, Atrens A. The influence of yttrium (Y) on the corrosion of Mg–Y binary alloys. Corros Sci 2010;52:3687–701. [24] Ilich JZ, Kerstetter JE. Nutrition in bone health revisited: a story beyond calcium. J Am Coll Nutr 2000;19:715–37. [25] Drynda A, Hassel T, Hoehn R, Perz A, Bach F-W, Peuster M. Development and biocompatibility of a novel corrodible fluoride-coated magnesium–calcium alloy with improved degradation kinetics and adequate mechanical properties for cardiovascular applications. J Biomed Mater Res Part A 2010;93A:763–75. [26] Li Z, Gu X, Lou S, Zheng Y. The development of binary Mg–Ca alloys for use as biodegradable materials within bone. Biomaterials 2008;29:1329–44. [27] Velikokhatnyi OI, Kumta PN. First-principles studies on alloying and simplified thermodynamic aqueous chemical stability of calcium-, zinc-, aluminum-, yttrium- and iron-doped magnesium alloys. Acta Biomater 2010;6:1698–704. [28] Saha P, Viswanathan S. An analysis of the grain refinement of magnesium by zirconium. Magnesium technology 2011. New York: Wiley; 2011. p. 175–80. [29] Saha P, Viswanathan S. Engineering an efficient zirconium-based grain refiner for magnesium alloys. Int. J. Metalcast. 2010;4:70–1. [30] Saha P, Lolies K, Viswanathan S, Gokhale A, Batson R, A Systematic Study of the Grain Refinement of Magnesium by Zirconium. Warrendale, PA: Minerals, Metals and Materials Society/AIME; 2010.

[31] Saha P, Viswanathan S. Grain refinement of magnesium by zirconium: characterization and analysis (11–066). AFS Transactions-American Foundry Society 2011;119:469–80. [32] Witte F, Hort N, Vogt C, Cohen S, Kainer KU, Willumeit R, et al. Degradable biomaterials based on magnesium corrosion. Curr Opin Solid State Mater Sci 2008;12:63–72. [33] Song G, StJohn D. The effect of zirconium grain refinement on the corrosion behaviour of magnesium–rare earth alloy MEZ. J Light Met 2002;2:1–16. [34] ASTM B275–05 Standard practice for codification of certain nonferrous metals and alloys, cast and wrought. Annual book of ASTM standards. Philadelphia, PA: ASTM; 2005. [35] Qian M, StJohn DH, Frost MT. Zirconium alloying and grain refinement of magnesium alloys. Magnes Technol 2003;2003(2003):209–14. [36] ASTM E112 Standard test methods for determining average grain size. Annual book of ASTM standards. Philadelphia, PA: ASTM; 2010. [37] ASTM E8-04 Standard test methods for tension testing of metallic materials. Annual book of ASTM standards. Philadelphia, PA: ASTM; 2004. [38] ASTM E9-09 Standard test methods of compression testing of metallic materials at room temperature. Annual book of ASTM standards. Philadelphia, PA: ASTM; 2009. [39] Kirkland NT, Birbilis N, Staiger MP. Assessing the corrosion of biodegradable magnesium implants: a critical review of current methodologies and their limitations. Acta Biomater 2012;8:925–36. [40] ASTM G31-72 Standard practice for laboratory immersion corrosion testing of metals. Annual book of ASTM standards. Philadelphia, PA: ASTM; 2004. [41] International Organization for Standardization. ISO-10993-12: biological evaluation of medical devices part 12: Sample preparation and reference materials. Arlington, VA: ANSI/AAMI; 2007. [42] Fischer J, Prosenc MH, Wolff M, Hort N, Willumeit R, Feyerabend F. Interference of magnesium corrosion with tetrazolium-based cytotoxicity assays. Acta Biomater 2010;6:1813–23. [43] Qian M, Zheng L, Graham D, Frost MT, StJohn DH. Settling of undissolved zirconium particles in pure magnesium melts. J Light Met 2001;1:157–65. [44] Liu M, Uggowitzer PJ, Nagasekhar AV, Schmutz P, Easton M, Song G-L, et al. Calculated phase diagrams and the corrosion of die-cast Mg–Al alloys. Corros Sci 2009;51:602–19. [45] Krause A, von der Höh N, Bormann D, Krause C, Bach F-W, Windhagen H, et al. Degradation behaviour and mechanical properties of magnesium implants in rabbit tibiae. J Mater Sci 2010;45:624–32. [46] Witte F, Kaese V, Haferkamp H, Switzer E, Meyer-Lindenberg A, Wirth CJ, et al. In vivo corrosion of four magnesium alloys and the associated bone response. Biomaterials 2005;26:3557–63. [47] Fischer J, Pröfrock D, Hort N, Willumeit R, Feyerabend F. Improved cytotoxicity testing of magnesium materials. Mater Sci Eng B 2011;176:830–4. [48] Pan F-s, Chen M-b, Wang J-f, Peng J, Tang A-t. Effects of yttrium addition on microstructure and mechanical properties of as-extruded AZ31 magnesium alloys. Trans Nonferrous Met Soc China 2008;18(Suppl. 1):s1–6. [49] Hall EO. The deformation and ageing of mild steel. III. Discussion of results. Proc Phys Soc Sec B 1951;64:747. [50] Petch NJ. The cleavage strength of polycrystals. J Iron Steel Inst 1953;174:25–8. [51] Ghali E, Dietzel W, Kainer K-U. General and localized corrosion of magnesium alloys: a critical review. J Mater Eng Perform 2004;13:7–23. [52] Xu L, Yu G, Zhang E, Pan F, Yang K. In vivo corrosion behavior of Mg–Mn–Zn alloy for bone implant application. J Biomed Mater Res Part A 2007;83A:703–11. [53] Aung NN, Zhou W. Effect of grain size and twins on corrosion behaviour of AZ31B magnesium alloy. Corros Sci 2010;52:589–94. [54] Argade GR, Panigrahi SK, Mishra RS. Effects of grain size on the corrosion resistance of wrought magnesium alloys containing neodymium. Corros Sci 2012;58:145–51. [55] Alvarez-Lopez M, Pereda MD, del Valle JA, Fernandez-Lorenzo M, GarciaAlonso MC, Ruano OA, et al. Corrosion behaviour of AZ31 magnesium alloy with different grain sizes in simulated biological fluids. Acta Biomater 2010;6:1763–71. [56] Neil WC, Forsyth M, Howlett PC, Hutchinson CR, Hinton BRW. Corrosion of heat treated magnesium alloy ZE41. Corros Sci 2011;53:3299–308. [57] Wang Q, Liu Y, Fang S, Song Y, Zhang D, Zhang L, et al. Evaluating the improvement of corrosion residual strength by adding 1.0 wt.% yttrium into an AZ91D magnesium alloy. Mater Charact 2010;61:674–82. [58] Song G, Atrens A. Understanding magnesium corrosion – a framework for improved alloy performance. Adv Eng Mater 2003;5:837–58. [59] Shi Z, Atrens A. An innovative specimen configuration for the study of Mg corrosion. Corros Sci 2011;53:226–46. [60] Xin Y, Hu T, Chu PK. Influence of test solutions on in vitro studies of biomedical magnesium alloys. J Electrochem Soc 2010;157:C238–43. [61] Schroeder HA, Mitchener M. Scandium, chromium(VI), gallium, yttrium, rhodium, palladium, indium in mice: effects on growth and life span. J Nutr 1971;101:1431–7. [62] Nakamura Y, Tsumura Y, Tonogai Y, Shibata T, Ito Y. Differences in behavior among the chlorides of seven rare earth elements administered intravenously to rats. Toxicol Sci 1997;37:106–16. [63] Zartner P, Cesnjevar R, Singer H, Weyand M. First successful implantation of a biodegradable metal stent into the left pulmonary artery of a preterm baby. Catheter Cardiovasc Interv 2005;66:590–4.

D.-T. Chou et al. / Acta Biomaterialia 9 (2013) 8518–8533 [64] Serre CM, Papillard M, Chavassieux P, Voegel JC, Boivin G. Influence of magnesium substitution on a collagen–apatite biomaterial on the production of a calcifying matrix by human osteoblasts. J Biomed Mater Res 1998;42:626–33. [65] Yamamoto A, Honma R, Sumita M. Cytotoxicity evaluation of 43 metal salts using murine fibroblasts and osteoblastic cells. J Biomed Mater Res 1998;39:331–40. [66] Plecko M, Sievert C, Andermatt D, Frigg R, Kronen P, Klein K, et al. Osseointegration and biocompatibility of different metal implants – a comparative experimental investigation in sheep. BMC Musculoskelet Disord 2012;13:32. [67] Janning C, Willbold E, Vogt C, Nellesen J, Meyer-Lindenberg A, Windhagen H, et al. Magnesium hydroxide temporarily enhancing osteoblast activity and decreasing the osteoclast number in peri-implant bone remodelling. Acta Biomater 2010;6:1861–8. [68] Schmidbaur H, Bach I, Wilkinson DL, Müller G. Metal ion binding by amino acids. Preparation and crystal structures of magnesium, strontium, and barium L-glutamate hydrates. Chem Ber 1989;122:1433–8.

8533

[69] Bobe K, Willbold E, Morgenthal I, Andersen O, Studnitzky T, Nellesen J, et al. In vitro and in vivo evaluation of biodegradable, open-porous scaffolds made of sintered magnesium W4 short fibres. Acta Biomaterialia, 2013; in press. [70] Witte F, Fischer J, Nellesen J, Vogt C, Vogt J, Donath T, et al. In vivo corrosion and corrosion protection of magnesium alloy LAE442. Acta Biomater 2010;6:1792–9. [71] Aghion E, Levy G, Ovadia S. In vivo behavior of biodegradable Mg–Nd–Y–Zr–Ca alloy. J Mater Sci Mater Med 2012;23:805–12. [72] Midwood KS, Williams LV, Schwarzbauer JE. Tissue repair and the dynamics of the extracellular matrix. Int J Biochem Cell Biol 2004;36:1031–7. [73] Witte F, Fischer J, Nellesen J, Crostack H-A, Kaese V, Pisch A, et al. In vitro and in vivo corrosion measurements of magnesium alloys. Biomaterials 2006;27:1013–8. [74] Walker J, Shadanbaz S, Kirkland NT, Stace E, Woodfield T, Staiger MP, et al. Magnesium alloys: predicting in vivo corrosion with in vitro immersion testing. J Biomed Mater Res B Appl Biomater 2012;100B:1134–41. [75] Ullmann B, Reifenrath J, Dziuba D, Seitz J-M, Bormann D, Meyer-Lindenberg A. In Vivo degradation behavior of the magnesium alloy LANd442 in rabbit tibiae. Materials 2011;4:2197–218.