Acta Biomaterialia 9 (2013) 8561–8573

Contents lists available at SciVerse ScienceDirect

Acta Biomaterialia

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

In vitro and in vivo studies on biodegradable CaMgZnSrYb high-entropybulk metallic glass q

1742-7061/$ - see front matter � 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.actbio.2013.01.029

q Part of the Biodegradable Metals Conference 2012 Special Issue, edited byProfessor Frank Witte and Professor Diego Mantovani.⇑ Corresponding authors. Tel./fax: +86 10 62767411 (Y.F. Zheng).

E-mail addresses: yfzheng@pku.edu.cn (Y.F. Zheng), whw@aphy.iphy.ac.cn (W.H.Wang), lingqin@cuhk.edu.hk (L. Qin).

1 This author contributed equally to this study.

H.F. Li a, X.H. Xie b,c,1, K. Zhao d, Y.B. Wang a, Y.F. Zheng a,⇑, W.H. Wang d,⇑, L. Qin b,e,⇑a State Key Laboratory for Turbulence and Complex System and Department of Materials Science and Engineering, College of Engineering, Peking University, Beijing 100871, Chinab Department of Orthopaedics & Traumatology, The Chinese University of Hong Kong, Shatin, Hong Kong, Chinac The Department of Orthopedics, The First Affiliated Hospital of Soochow University, Suzhou 215006, Chinad Institute of Physics, Chinese Academy of Sciences, Beijing 100190, Chinae Translational Medicine Research & Development Center, Institute of Biomedical and Health Engineering, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences,Shenzhen 518055, China

a r t i c l e i n f o

Article history:Available online 4 February 2013

Keywords:Ca20Mg20Zn20Sr20Yb20 alloyHigh-entropy alloysBulk metallic glassesBiodegradable metalsBone

a b s t r a c t

In order to enhance the corrosion resistance of the Ca65Mg15Zn20 bulk metallic glass, which has too fast adegradation rate for biomedical applications, we fabricated the Ca20Mg20Zn20Sr20Yb20 high-entropy bulkmetallic glass because of the unique properties of high-entropy alloys. Our results showed that themechanical properties and corrosion behavior were enhanced. The in vitro tests showed that theCa20Mg20Zn20Sr20Yb20 high-entropy bulk metallic glass could stimulate the proliferation and differentia-tion of cultured osteoblasts. The in vivo animal tests showed that the Ca20Mg20Zn20Sr20Yb20 high-entropybulk metallic glass did not show any obvious degradation after 4 weeks of implantation, and they canpromote osteogenesis and new bone formation after 2 weeks of implantation. The improved mechanicalproperties and corrosion behavior can be attributed to the different chemical composition as well as theformation of a unique high-entropy atomic structure with a maximum degree of disorder.

� 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Bulk metallic glasses (BMGs) are a neoteric class of alloys with afully amorphous microstructure, unlike conventional crystalline al-loys. Due to their unique atomic structure, BMGs have superiorstrength, low Young’s modulus, high elastic strain limit, enhancedwear resistance, excellent corrosion resistance, relatively flexiblecomposition, and broader solubility for alloy elements [1]. Bulkmetallic glasses have attracted much attention in recent years ascandidates for biomaterials. Ti-based [2,3], Zr-based [4–6] andFe-based BMGs [7,8] have all been investigated as new kinds ofnon-degradable materials, whilst Ca-based [9,10], Mg-based[11,12], Zn-based [13,14] and Sr-based BMGs [15] have been devel-oped as the new generation of biodegradable metals.

In our previous study [9], Ca65Mg15Zn20 BMG was evaluated byboth in vitro tests on corrosion behavior, ion release and biocom-patibility and in vivo implantation, aiming at exploring its feasibil-ity for potential skeletal applications. Our results imply that

Ca65Mg15Zn20 BMG is nontoxic and could actually stimulate bonetissue healing. However, Ca65Mg15Zn20 BMG degraded completelyafter 4 h in vitro and within 4 weeks in vivo, which is too rapid toallow sufficient time for healing, as it is desirable to have the im-planted fixture present for at least 12 weeks. The rapid degradationrate of Ca65Mg15Zn20 BMG is one of the greatest limitations for itsuse in orthopedic applications. In an effort to improve the corrosionresistance of metal materials, two different strategies are usuallyapplied: surface treatment and alloying. Although surface coatingscould enhance the corrosion resistance of CaMgZn ternary BMGaccording to our previous study [16], there is great concern aboutthe loosening, flaking or peeling-off of coatings, especially whenimmersed in an aggressive physiological medium for a long period.

In recent years, high-entropy alloys (HEAs), an advanced alloysystem that contains multiple principal elements in equimolar ra-tios instead of a single major element, have been developed.According to the regular solution approach, with an increasingnumber of principal elements in the system, the configurational en-tropy of mixing increases, reaching a maximum when the concen-trations of all the elements are equal. This feature forms the coreconcept of HEAs. HEAs are been defined as alloys composed of fiveor more principal elements in equimolar ratios. HEAs have a stablesimple solid-solution structure, and can easily form nanoprecipi-tates and an amorphous phase due to the high mixing entropy[17]. HEAs have many unique mechanical and electrochemical

8562 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

properties that conventional alloys lack. For instance, HEAs havehigh strength, high hardness, good oxidation resistance, excellentcorrosion resistance and excellent abrasion resistance. Conse-quently, HEAs have great potential for practical applications [18].

Considering that BMGs and HEAs both have unique propertiesthat conventional metals and alloys are unlikely to match andthe fact that an amorphous phase can be easily formed in HEAs,we believe that the research and development of new types of al-loys that combine BMG and HEA concepts together would be ofgreat importance for future novel material studies. In the presentstudy, the system and composition of Ca20Mg20Zn20Sr20Yb20 alloywas determined from the prototype ternary Ca65Mg15Zn20 BMGin accordance with the following strategic alloy designs: (i) HEAsdefined by an equiatomic alloy with five or more elements; (ii)the exchangeability of the constituent elements with a similarchemical nature in the periodic table, significant atomic size ratiosabove 12% and negative heats of mixing to satisfy the empiricalrules of BMG formation which were first put forward by Inoue[19]; and (iii) to guarantee the biosafety of biodegradable materi-als, the constitutional elements should be non-toxic, since all theelements will enter into the human body. Therefore, strontiumand ytterbium were selected as elements for addition to CaMgZnalloy in consideration of both their atomic radii matching andphysiological roles.

Strontium, calcium and magnesium are in the same group ofperiodic table. The distribution of Sr is similar to Ca, with 99%of the element being stored in bone [20]. Under normal condi-tions, the bone strontium/calcium ratio varies between 1:1000and 1:2000, and the highest concentration is present in newlyformed bone [21]. Strontium can inhibit bone resorption andstimulate bone formation in both rodents [22] and osteoporoticpatients [23]. Sr-based drug treatments for osteoporosis, such asProtelos, can reduce the risk of fracture in patients after 1 yearof treatment. In vitro studies have shown that Protelos inhibitsosteoclast activity [24] and stimulates osteoblast proliferation[25].

The rare earth element ytterbium is chosen mainly in consid-eration of its BMG forming ability due to its atomic radius andmixing entropy. Ytterbium has unlimited solubility in calciumand has been found to be effective in improving corrosion resis-tance, glass-forming ability and the mechanical properties ofmetallic glass [13]. Fluorides of ytterbium have low toxicity andaccumulation, inducing no local irritation of skin and eyes andnot causing intoxication if administered via the stomach [26].Rare-earth-doped b-NaYF4:Yb, Er up-conversion nanoparticlescan be used in the in vivo imaging, detection and diagnosis ofcancers [27]. 175Yb-labeled hydroxyapatite (HA) particle can beused as an agent for radiation synovectomy of small-sized joints[28].

The aim of this study is to introduce the concept of high entropyinto the field of BMGs in the hope of obtaining high-entropy BMGs(HE-BMGs) with excellent properties. The feasibility of the newlydeveloped HE-BMGs as biomaterials for orthopedic applicationswill be investigated by both in vitro and in vivo evaluations.

2. Materials and methods

2.1. Preparation of materials

CaMgZnSrYb HE-BMGs (CMZSY HE-BMGs), with nominal com-positions (at.%) of Ca20Mg20Zn20Sr20Yb20, were prepared by theinduction-melting method. The base elements Ca (99%), Mg(99.9%), Zn (99.9%), Sr (99%) and Yb (99.5%) (Beijing CuibolinNon-Ferrous Technology Developing Co., Ltd.) were melted in aquartz tube under a high vacuum (better than 3.0 � 10�3 Pa) at

900 �C. The melt was then cast into a liquid-nitrogen-cooled cop-per mould with a cavity of 50 � 5 � 2 mm. The CMZSY HE-BMGsamples were further cut into 5 � 5 � 2 mm plates by a low-speedprecision diamond saw (SYJ-150, MTI) for corrosion and biocom-patibility tests, with the surface being polished up to 2000# grit.All the samples were then ultrasonically cleaned in acetone, abso-lute ethanol and deionized water for 15 min respectively. For thebiocompatibility tests, the samples were sterilized by ultravioletradiation for at least 2 h on one side before being turned over foranother 2 h of sterilization.

2.2. Microstructural characterization and composition analysis

An X-ray diffraction (XRD) with a Rigaku DMAX 2400 diffrac-tometer using Cu Ka radiation at a scan rate of 4� min�1, operatedat 40 kV and 100 mA at room temperature, was employed for theidentification of the amorphous structure of samples. Differentialscanning calorimetry (DSC; Mettler Toledo DSC822e) was per-formed under a purified argon atmosphere with a constant heatingrate of 20 K min�1 to identify the thermodynamic parameters ofthe CMZSY HE-BMG alloys.

2.3. Mechanical tests and density measurement

Uniaxial compression testing was conducted with an Instron8562 testing machine at a constant nominal strain rate of1 � 10�4 s�1 at room temperature. Test samples, 2 mm in diameterand 4 mm in length, were prepared according to ASTM E9-89a(2000) standards [29], which suggested a length-to-diameter ratioof 2 for cylindrical specimens.

The density q was measured by Archimedes’ principle in abso-lute alcohol (P99.9%). The travel time of the ultrasonic wavespropagating through the sample was measured using a MATEC6600 ultrasonic system with a measuring sensitive of 0.5 ns anda carrying frequency of 10 MHz. The elastic constants (includingthe Young’s modulus E, the shear modulus G and the bulk modulusK) were derived from the acoustic data and density.

2.4. Immersion test

The immersion test was carried out in Hanks’s solution [30](NaCl 8.0 g, CaCl2 0.14 g, KCl 0.4 g, NaHCO3 0.35 g, glucose 1.0 g,MgCl2�6H2O 0.1 g, Na2HPO4�2H2O 0.06 g, KH2PO4 0.06 g, MgSO4-

�7H2O 0.06 g, dissolved in 1 l of deionized water) according toASTM-G31-72 [31]. After different immersion times, the sampleswere removed from the Hanks’s solution, gently rinsed with dis-tilled water and dried at room temperature. The weight was mea-sured on an electronic balance (Mettler Toledo AL204) with ameasuring sensitivity of 0.1 mg. The pH value of the solution wasalso recorded during the immersion tests, using a Lei-ci PHS-3CpH meter. The surface and cross-sectional morphologies afterimmersion were observed by scanning electron microscopy(SEM; Hitachi S-4800), followed by energy-disperse spectrometer(EDS) analysis. XRD using a diffractometer with Cu Ka radiationat a scan rate of 2� min�1, operated at 40 kV and 100 mA at roomtemperature, was used to identify the phase composition of thecorrosion product. The amount of hydrogen generated by theCMZSY HE-BMG was measured in accordance to Ref. [32]; for this,the sample is placed in a beaker containing the immersion solutionand a measuring cylinder filled with the liquid is placed over it tocollect the hydrogen formed during corrosion of the sample. Induc-tively coupled plasma atomic emission spectrometry (Leeman,Profile ICP-AES) was employed to measure the concentrations ofthe alloying element ions which had dissolved from the alloyplates. The prototype ternary Ca65Mg15Zn20 BMG was used as a

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8563

control during the immersion test. An average of three measure-ments were taken for each group.

2.5. Electrochemical tests

A three-electrode cell was used for electrochemical measure-ments. The auxiliary electrode was platinum and the referenceelectrode was a saturated calomel electrode. The electrochemicaltests were conducted with an electrochemical workstation (CHI604D) at a temperature of 37 �C in Hanks’s solution. The open cir-cuit potentials (OCPs) were monitored for 4000 s. Potentiodynamicpolarization tests were carried out at a scanning rate of 1 mV s�1

and the initial potential was about 300 mV below the corrosion po-tential (Ecorr) after OCP measurements.

2.6. Biocompatibility evaluation

A human osteoblast-like cell line (MG63) was adopted to evalu-ate the biocompatibility of CMZSY HE-BMG. MG63 cells were cul-tured in minimum essential medium (MEM), 10% fetal bovineserum (FBS), 100 U ml�1 penicillin and 100 mg ml�1 streptomycinat 37 �C in a humidified atmosphere of 5% CO2. The biocompatibilitytests were carried out by indirect contact. Extracts were preparedusing serum-free MEM as the extraction medium with the surfacearea of extraction medium ratio 1.25 ml cm�2 in a humidifiedatmosphere with 5% CO2 at 37 �C for 72 h. The supernatant fluidwas withdrawn and centrifuged to prepare the extraction medium.The extraction medium was serially diluted from 100% to 50, 25, 10and 5% concentration, and refrigerated at 4 �C before the biocom-patibility test. The control groups involved the use of MEM with10% FBS as negative controls and 10% dimethylsulfoxide MEM(10% FBS) as positive controls. Cells were incubated in medium in96-well cell culture plates at 5 � 104 cells ml�1 and incubated for24 h to allow attachment. The medium was then replaced with100 ll of extracts (10% FBS). After incubating the cells in a humid-ified atmosphere with 5% CO2 at 37 �C for 1, 3 and 5 days, the 96-well cell culture plates were observed under an inverted phase con-trast microscope. After that, 10 ll of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma) was added to eachwell. The samples were incubated with the MTT for 4 h at 37 �C,then 100 ll of formazan solubilization solution (10% sodium dode-cylsulfate in 0.01 M HCl) was added to each well for further incuba-tion overnight in a humidified atmosphere. The pH value and theconcentrations of alloying element ions in the extraction mediumwere measured. Optical density (OD) measurements of the sampleswere conducted using a microplate reader (Bio-RAD680) at 570 nmwith a reference wavelength of 630 nm. The cell relative growthrate (RGR) was calculated according to the following formula:

RGR ¼ ODtest=ODnegative � 100% ð1Þ

2.7. Fluorescent dye

Besides the MTT method, fluorescent dye (LIVE/DEAD Viability/Cytotoxicity Kit [L-3224], Molecular Probes�, USA) was also ap-plied to further identify the biocompatibility of the HE-BMG.Briefly, MG63 cells cultured with and without HE-BMG extractswere allowed to grow in the 96 wells for 3 days until acceptablecell densities were obtained. The cells were then washed gentlywith Dulbecco’s phosphate-buffered saline (D-PBS) three timesprior to the assay in order to remove any esterase activity, whichcan be present in the serum-supplemented growth medium andcould cause an increase in extracellular fluorescence due to thehydrolysis of calcein AM. After the last wash, 100 ll of D-PBSwas added to each well. After that, an additional 100 ll of fluores-cent dye working solution containing 2 lM calcein AM was added,

yielding 200 ll of solution per well containing 1 lM calcein AM.The cultured cells were stained by incubation for 45 min at roomtemperature, protected from light. Living cells stained with calce-in-AM show as a green color under fluorescence microscopy.

2.8. Alkaline phosphatase (ALP) activities

To assay ALP activity, MG63 cells were cultured in 5, 10, 25, 50and 100% CMZSY HE-BMG extracts respectively for 7 days at an ini-tial seeding density of 4 � 104 cells per well. ALP activity was eval-uated as the amount of p-nitrophenol (pNP) produced onhydrolysis of p-nitrophenyl phosphate (pNPP) in the presence ofALP as a catalyst. Briefly, 100 ll of 1-step pNPP (Sigma) was addedto each well containing a 100 ll aliquot of supernatant. After30 min incubation at 37 �C the reaction was terminated by adding50 ll of 2 M NaOH. The absorbance was measured at 405 nm (Bio-RAD680). ALP activity results are presented as percentages of thenegative control group.

2.9. Animal tests

Three-month old C57BL/6 mice were used to examine thein vivo degradation properties of CMZSY HE-BMG rods. In brief,the mice were anesthetized using ketamine (75 mg kg�1) and xyla-zine (10 mg kg�1). A tunnel, 0.7 mm in diameter and 5 mm inlength, was created into the medullary cavity along the axis ofthe femoral shaft from the distal femur. Sterilized CMZSYHE-BMG rods, 0.7 mm in diameter and 5 mm in length, wereimplanted into the bone tunnels. Created bone tunnels withoutimplants worked as control groups. After the operation, the micewere housed in an environmentally controlled animal care labora-tory. The animal experimental protocol was approved by the Ani-mal Ethics Committee of the Chinese University of Hong Kong(No. 10/049/MIS). The in vivo analysis included: (i) X-ray observa-tion: sequential radiographs of the distal femur were taken (30 kV,3 s) every week post-operation for general inspection; (ii) in vivomicrocomputerized tomography (micro-CT) (viva CT40, ScancoMedical AG, Brüttisellen, Switzerland) with a voxel size of 20 lmwas used to scan the distal femurs of mice at weeks 0, 1, 2, 3and 4 after operation. The change in the CMZSY HE-BMG in thebone tunnel was measured using our published protocol [33].Two-dimensional (2-D) images were acquired directly from thescanning, while the 3-D structure was reconstructed from the vol-ume of interest, where an optimized threshold was used to isolatethe bone and materials from the background. The density changesof the CMZSY HE-BMG and the changes of the cortical bone aroundthe CMZSY HE-BMG were measured on the digitally extracted tis-sue; and (iii) histological examination: 4 weeks after implantation,the femora with the CMZSY HE-BMG rods were harvested and pre-pared for decalcification using 10% buffered formalin acid beforebeing embedded in paraffin, sectioned longitudinally at a thicknessof 7 lm and stained with hematoxylin and eosin (H&E).

2.10. Statistical analysis

Statistical analysis was conducted with SPSS 17.0. Differencesbetween groups were analyzed using one-way analysis of variancefollowed by Tukey’s test.

3. Results

3.1. Microstructural characterization

Fig. 1(a) shows the XRD patterns of CMZSY HE-BMG. The typicalamorphous features of the CMZSY HE-BMG samples can be

Fig. 1. (a) XRD pattern and (b) DSC measurement curve of CMZSY HE-BMG.

8564 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

confirmed by the absence of detectable crystalline diffractionpeaks, together with the broad scattering signals around 30 and52�, respectively.

Fig. 1(b) shows the corresponding DSC curve of CMZSY HE-BMG. The glass transition temperature (Tg), the first crystallizationtemperature (Tx1) and the supercooled liquid region (DTx), definedby DTx = Tx1 � Tg, of CMZSY HE-BMG are determined from the DSCcurve and are found to be 357, 390 and 33 K.

3.2. Mechanical properties

Table 1 presents the mechanical properties of CMZSY HE-BMG.The fracture strength rf for CMZSY HE-BMG is 370.7 ± 25.4 MPa,which is much higher than that of Ca65Mg15Zn20 BMG (300 MPa)and cortical bone (130–180 MPa). The Young’s modulus of CMZSYHE-BMG (19.4 ± 3.4 GPa) is quite close to that of cortical bone (3–20 GPa).

3.3. Corrosion behavior

Fig. 2 shows the weight loss and pH value changes of CMZSYHE-BMG compared with its prototype ternary Ca65Mg15Zn20 BMG(CMZ BMG for short) as a function of immersion time. From the

Table 1Mechanical properties of CMZSY HE-BMG, Ca65Mg15Zn20 BMG and bone.

Materials Compressivestrength, rf

(MPa)

Young’smodulus,E (GPa)

Shearmodulus,G (GPa)

Bulkmodulus,K (GPa)

Ca20Mg20Zn20Sr20Yb20 370.7 ± 25.4 19.4 ± 3.4 7.6 ± 1.3 15.0 ± 2.6Ca65Mg15Zn20 300 [34] 22.3 [34] 10.1 [35] 22.6 [35]Cortical bone [36] 130–180 3–20 _ _

Fig. 2. (a) Weight loss measurement and (b) pH value change in CMZSY HE-BMG a

figure, it can be seen that the CMZSY HE-BMG has a far lowerdegradation rate, and thus a lower pH value, compared with theprototype CMZ BMG. The average degradation rate calculated fromthe weight loss is 0.17 mg cm�2 h�1 for CMZSY HE-BMG and45 mg cm�2 h�1 for CMZ BMG.

The hydrogen evolutions of CMZSY HE-BMG and CMZ BMG as afunction of immersion time are shown in Fig. 3. Similar to theweight loss test, the H2 evolution is significantly decreased withthe addition of the elements Sr and Yb to form the high-entrophyglassy alloy, the results showing the average hydrogen evolutionrates of CMZSY HE-BMG and CMZ BMG to be 0.02 and 48 ml cm�2

h�1, respectively.Fig. 4 shows the SEM micrographs and EDS analysis of the sam-

ple surface after the immersion test in Hanks’s solution for 72 h.The surface of the CMZSY HE-BMG after the immersion test re-mains flat and is still relatively whole, though some corrosionproducts have precipitated onto it. Further EDS analysis showedthe presence of oxygen, calcium, phosphorus, carbon, magnesiumand strontium on the surface. The EDS analysis results demonstratethat the ratio of Ca/P is 1.40 whilst that of (Ca + Sr)/P is 1.64.

In order to confirm the phase composition of the corrosionproducts, XRD structure and phase analysis was employed; theXRD pattern is shown in Fig. 5. From the XRD analysis results,we can see that the corrosion products of CMZSY HE-BMG consistof HA, strontium apatite (Sr-HA), calcium strontium carbonate[(Sr,Ca)CO3] and calcium strontium magnesium phosphate[(Ca,Sr)2Mg(PO4)2�H2O].

Fig. 6 shows the cross-sectional SEM image and the correspond-ing EDX maps of the post-immersion CMZSY HE-BMG. Three dis-tinct layers can be observed in the cross-sectional SEM image.The outermost layer consists mainly of Ca, Sr, C and P, representingthe corrosion products deposited on the CMZSY HE-BMG. The

nd CMZ BMG samples as a function of the immersion time in Hanks’s solution.

Fig. 3. Hydrogen evolution volumes of CMZSY HE-BMG and CMZ BMG samples as afunction of the immersion time in Hanks’s solution. Fig. 5. XRD patterns corresponding to the corrosion products of HE-BMG.

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8565

transition layer between the HE-BMG substrate and the precipi-tated layer shows a significant drop in Sr content and some ofthe Ca has dissolved. The inner layer is the uncorroded HE-BMGsubstrate. In contrast, the nobler elements Mg, Zn and Yb remainin the transition layer, together with a significant amount of oxy-gen, indicating that the compositions of the transition layer mayconsist of magnesium oxide, zinc oxide and ytterbium oxide.

The open circuit potentials of CMZSY HE-BMG in Hanks’s solu-tion were recorded for 3000 s. Their evolution with time is shownin Fig. 7(a). During the initial 800 s of immersion time, the opencircuit potential (Eocp) of the alloy increases with time, indicatingthe precipitation of a protective surface film [37,38]. After 800 sof immersion, the Eocp of the alloy achieves a relatively stable va-lue. The electrochemical polarization curve of CMZSY HE-BMG inHanks’s solution is shown in Fig. 7(b). The corrosion current den-sity calculated from the Tafel plot was 9.16 lA cm�2. Due to thefast degradation rate of CMZ BMG, we could not obtain a stableOCP curve and hence did not conduct a Tafel plot.

3.4. Ion release

Fig. 8 shows the ion concentrations in Hanks’s solution after theimmersion period of 72 h. The pre-existing Mg and Ca ion levels inthe Hanks’s solution were subtracted from the plotted ICP data. Itcan be seen in Fig. 8 that the ions released into solution are mainlyCa, Mg and Sr, while the Zn and Yb ion levels are quite low.

3.5. Biocompatibility test

Fig. 9(a) and (b) shows the ion concentrations and pH value of theextraction medium used for biocompatibility test. The pre-existingMg and Ca ion concentrations in the MEM solution were subtracted

Fig. 4. (a) Surface morphology and (b) EDS spectrum of the corrosion products of

beforehand in Fig. 9(a). The extraction medium has high concentra-tions of Ca, Mg and Sr but low concentrations of Zn and Yb, similar tothe ion concentrations released into the Hanks’s solution. The pH va-lue increased with the extraction concentration and the pH value ofthe 100% CMZSY HE-BMG extract was about 8.5.

The MG63 cell viability and ALP activity cultured in 100, 50, 25,10 and 5% CMZSY HE-BMG extraction medium for 1, 3 and 5 daysare shown in Fig. 9(c) and (d). It can be seen that cells cultured inthe CMZSY HE-BMG extract showed significantly higher cell viabil-ity and ALP activity than the control group, indicating that it couldpromote the proliferation and differentiation of osteoblasts.

Fig. 10 presents images of fluorescently dyed MG63 cells cul-tured in series of CMZSY HE-BMG extracts and negative controlgroups. Healthy morphologies of MG63 cells, whether polygon orspindle, were the well spread and extended MG63 cells showedpolygonal or spindle shapes and had filopodia while pooly spreadcells showed rounded shape without filopodia. At lower concentra-tions (5%, 10% and 25%), the pH values are similar to the negativecontrol group (P > 0.05), thus the higher cell proliferation wascaused by the positive effect of the metal ions (Ca, Mg, Sr, Zn).However, the 100% HE-BMG extracts did not exhibit much highercell proliferations than the negative control group and showedobvious decreased cell proliferation compared with the 10% and25% extracts, which may be attributed to the higher pH value com-pared with the lower concentration extracts (10% and 25%) (P <0.05) and the negative control group (P < 0.05).

3.6. Animal tests

3.6.1. X-ray observationFig. 11 shows the radiographs of distal femora with and without

implantation of CMZSY HE-BMG at 0, 1, 2, 3 and 4 weeks. The

CMZSY HE-BMG after being immersed in Hanks’s solution at 37 �C for 72 h.

Fig. 6. SEM micrograph (a) and EDX maps (b) from the cross-section of HE-BMG after immersion in Hanks’s solution for 3 days.

8566 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

figure indicates that CMZSY HE-BMG rods do not degrade muchduring the whole experiment period, as evidenced by there beingno obvious reduction in the diameter of the rods. If a material de-grades rapidly in vivo, the large amount of hydrogen gas generated

in a short period will not be able to be absorbed or diffused com-pletely, so will show up in the X-ray radiographs as a gas shadow[39,40]. In the present study, due to the relatively slow degradationrate in vivo, no gas shadow was observed around the rods. With

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8567

increasing time after implantation, the CMZSY HE-BMG rods re-tained their original shape under X-ray observation. The radio-graphs show that all the CMZSY HE-BMG rods were wellpositioned (meaning that the HE-BMG rods can be seen clearly, indi-cating their excellent radiopacity) in the distal femora and no translu-cence was found around the metal alloy. There was no obviousinflammation around the CMZSY HE-BMG rods and no mice diedpost-operation. Interestingly, although there was no evidence ofobvious degradation during the whole experiment period, a perios-teal reaction was observed around the CMZSY HE-BMG rods at 2, 3and 4 weeks, showing as circumferential osteogenesis, i.e. the cor-tical bone around the rods became thicker than that in the controlbone groups (yellow arrows in Fig. 11), indicating the formation ofnew bone.

Fig. 8. Ion concentrations released into Hanks’s solution. n = 3; the error barrepresents the standard deviation.

3.6.2. Micro-CT observationFig. 12 shows 2-D and 3-D images of a representative mouse fe-

mur with an intramedullary CMZSY HE-BMG implant scannedin vivo at different post-implantation time points. In the 3-D mi-cro-CT reconstruction of the mouse distal femur and the 2-D tomo-graphs of cross-sections, less degradation of the CMZSY HE-BMGcan be seen, no degraded debris was found around the metal alloyrods and new bone formation was observed during the 4 weekimplantation period. The peri-implant cortical bone thickness ofthe CMZSY HE-BMG implantation group was significantly higherthan that of the non-implanted control from 1 week post-surgery(⁄p < 0.05 and ⁄⁄p < 0.01) and the bone thickness increased withincreasing implantation period (yellow arrows in Fig. 12).

From the 2-D and 3-D images, we can calculate the bone thick-ness. The results are shown in Fig. 13. This demonstrates that theaverage thickness of the peri-implant cortical bone is 563 ± 70 lmat week 4, which is significantly higher than that of the non-im-planted control group (249 ± 24 lm).

3.6.3. Histological analysisFig. 14 shows the histology of cross-sections of the distal femora

and diaphyseal region of the mice. No inflammation or osteonecro-sis was found at the bone area around the implants. The histolog-ical observations at 4 weeks after implantation show that the bonethickness around the implanted CMZSY HE-BMG is greater thanthat of normal bone (yellow arrows in Fig. 14), which is consistentwith the radiographs and micro-CT results. The new bone at the in-ner edge of the cortical bone can be clearly observed (black arrowsin Fig. 14). There is no new bone formation around the wound areawhere no implants were embedded in the control groups.

Fig. 7. (a) OCP curve and (b) potentiodynamic polarization cu

4. Discussion

HE-BMGs, relative newcomers to the world of metallic glasses,with a maximum degree of disorder, have unique properties thatare unlike those of conventional alloys. According to the formulafor free energy (G = H – TS, G = free energy; H = enthalpy; T = abso-lute temperature (K); S = entropy) and Boltzmann’s hypothesis be-tween the entropy and the complexity of the system, high mixingentropy can significantly lower the free energy, thus lowering thetendency to order and segregate. On the other hand, high mixingentropy relaxes the constraints set by the Hume–Rothery rulebetween two elements, thus promoting mutual effects amongmulti-principal elements and making the alloys more stable [41].HE-BMGs are not only of fundamental interest but can be used inpractical applications, their mechanical properties and corrosionbehavior being of particular interest for biomedical applications.

From the above results, it can be seen that the CMZSY HE-BMGshowed better mechanical behavior and corrosion resistance thanits prototype ternary Ca65Mg15Zn20 BMG. The reason for this canbe ascribed to the cocktail effects of HEAs. The cocktail effects referto the ideal mixing and inevitably greater number of interactionsbetween the multi-principal elements, thus enhancing the bondingbetween the alloying elements [42]. The cocktail effects result inan atomic-scale composite effect on properties wherein the inter-actions between the different elements themselves play an impor-tant role [43]. Besides the cocktail effects of HEAs, the alloyingelements Sr and Yb also enhance the mechanical behavior and cor-rosion resistance of CMZSY HE-BMG. Previous studies demon-strated that the Sr and Yb additions could obviously increase thestrength and enhance the resistance to corrosion of Mg-based al-loys [44,45] and Al-based alloys [46,47].

rve of electrochemical measurement in Hanks’s solution.

Fig. 9. (a) Ion concentrations of CMZSY HE-BMG released into MEM solution, (b) pH value change for the different concentration gradients of the CMZSY HE-BMG extracts, (c)MG63 cell viability cultured in the different concentration gradients of the CMZSY HE-BMG extracts and (d) ALP activity of MG63 cells cultured in the different concentrationgradients of the CMZSY HE-BMG extracts. The error bar represents the standard deviation. n = 5; ⁄p < 0.05, ⁄⁄p < 0.01.

Fig. 10. Cell staining at 3 days in the negative control group and in cells cultured in series of HE-BMG extracts. (a) Negative control group; (b) 5% HE-BMG extract; (c) 10% HE-BMG extract; (d) 25% HE-BMG extract; (e) 50% HE-BMG extract; (f) 100% HE-BMG extract. 200�; the scale bar represents 100 lm.

8568 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

In the design of BMGs, the ‘‘confusion principle’’ is widely ac-cepted: the more the elements involved, the greater the chaos,the greater the chance of glass formation and the greater the like-lihood of special properties. The reason why so few attempts havebeen made to exploit this idea and analyze the thermodynamicsand kinetics of the most disordered HE-BMG alloys fully is thatthe more the elements, the greater the disorder and the more dif-ficult the analysis [48].

4.1. Comparison of the mechanical properties of CMZSY HE-BMG withas-reported biodegradable BMGs

Mechanical properties are important aspects of biomaterials.Generally speaking, high strength and low elastic modulus arerequired for biomedical metals used in orthopedic applications.The high strength may improve the capacity of the material forload-bearing applications; the low elastic modulus can avoid

Fig. 11. Radiographs of mice distal femora with and without CMZSY HE-BMG rod implants for 0, 1, 2, 3 and 4 weeks of implantation.

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8569

problems of stress shielding, which could lead to the loosening andfailure of a bone implant [49].

Fig. 15(a) presents a comparison of the mechanical properties ofpreviously reported biodegradable BMGs and cortical bone withthe present CMZSY HE-BMG. The CMZSY HE-BMG shows a highercompression strength than that of Ca65Mg15Zn20 BMG and corticalbone but a lower one than that of Mg–Zn–Ca BMGs and Zn38Ca32-

Mg12Yb18 BMG. The different mechanical properties of the CMZSYHE-BMG and the other biodegradable BMGs can be attributed tothe differences in their chemical composition and the intrinsic lo-cal atomic scale structure.

4.2. Comparison of the corrosion behavior of CMZSY HE-BMG with as-reported biodegradable BMGs

Because of the large mixing entropy, CMZSY HE-BMG tends tobe more stable than its prototype CMZ BMG [18]. The cross-sectionanalysis, combined with XRD phase analysis, indicates the forma-tion of a protective zinc oxide and ytterbium oxide intermediatelayer, with HA, Sr-HA and other carbonates and phosphates precip-itated on the outside layer. The formation of Sr-HA was found todecrease the corrosion rate of Mg–0.5Sr [50] and to improve thebioactivity and biocorrosion resistance of titanium alloys [51].However, a previous study of Ca–Mg–Zn BMG demonstrated that

the corrosion layer was porous corroded calcium oxide, withoutHA or other phosphate being deposited [10].

As can be seen from Fig. 15(b), the CMZSY HE-BMG shows a re-duced degradation rate compared to Ca65Mg15Zn20 and Sr40Mg20-

Zn15Yb20Cu5 BMGs, but a slightly higher rate than Mg–Zn–Ca andZn38Ca32Mg12Yb18 BMGs. There are two reasons for this: first, itcan be attributed to the chemical composition; and secondly, theformation of a unique high-entropy atomic structure also playsan essential role in the corrosion behavior.

However, due to the complex composition and structure of HE-BMG, the precise mechanisms of how structure and compositionaffect the mechanical and corrosion behavior at the fundamentallevel are not well understood yet and need to be further explored.

4.3. In vitro and in vivo biocompatibility of CMZSY HE-BMG

In the in vitro cellular test, we found increased cell proliferationand differentiation of MG63 osteoblasts cultured in CMZSY HE-BMG extracts. In the in vivo tests, we observed a significantly high-er cortical bone thickness as well as newly formed endosteal bonearound the CMZSY HE-BMG compared with the control group. Theexcellent biocompatibility can be attributed to two main factors.First, the physiological roles play by the alloying elements andtheir combined effects.

Fig. 12. (a) 2-D and (b) 3-D images of a representative mouse femur with an intramedullary CMZSY HE-BMG implant scanned in vivo at 0, 1, 2, 3 and 4 weeks of implantation.Scale bar = 1.0 mm.

Fig. 13. Cortical bone thickness in the as-rolled CMZSY HE-BMG group and theempty tunnel control group at different implantation intervals. The error bar

8570 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

(i) Calcium is the most abundant cation in humans. The totalbody content averages about 25,000 mmol, or 1 kg, in a70 kg man, virtually all of it occurring in bone [52].

(ii) Magnesium is essential to human metabolism and is natu-rally found in bone tissue. It is the fourth most abundant cat-ion in the human body, with an estimated 1000 mmol ofmagnesium stored in the body of a normal 70 kg adult, withapproximately half of it present in bone tissue. Magnesiumis a co-factor for many enzymes, and stabilizes the struc-tures of DNA and RNA [36]. The correct quantity of magne-sium is essential for maintaining the strength of bone.Intake of magnesium increases the bone density, and isimportant for keeping and remodeling the strength andstructure of the bone; it also may prevent a decrease in bonequantity in post-menopausal women and in old people ingeneral [53].

(iii) Zinc is one of the most abundant nutritionally essential ele-ments in the human body. It has been found in all body tis-sues, with 85% of the total in muscle and bone, 11% in theskin and the liver, and the remainder in all the other tissues.In multicellular organisms, virtually all zinc is intracellular,with 30–40% located in the nucleus, 50% in the cytoplasm,organelles and specialized vesicles (for digestive enzymesor hormone storage), and the remainder in the cell mem-brane [54]. Zinc ions are a constituent of more than 200enzymes. Zinc deficiency causes a marked decline in immu-nological competence. Zinc ion affects T-cell activation, act-ing as a mitogen [55].

(iv) Strontium has an affinity for bone and is incorporated into itby surface exchange and ionic substitution. Strontium saltshave been found to stimulate bone formation and inhibitbone resorption both in vitro and in vivo [56]. Local deliveryof strontium with HA in cement has been shown to inducenew bone formation and is effective in reducing the risk offracture in osteoporosis [57]. The drug strontium ranelate

has been shown to reduce the incidence of fractures in oste-oporotic patients [58,59]. Strontium is present in the min-eral phase of bone, especially in the regions of highmetabolic turnover [60]. In vitro, it increases the numberof osteoblasts and reduces the number and activity of osteo-clasts [61]; in vivo, it inhibits bone resorption and improvesbone formation [62].

(v) Nanoparticles labeled with Yb are widely used in the in vivoimaging, detection, diagnosis and radiotherapy of cancers[27,28].

On the other hand, the formation of HA and Sr-HA is a keyadvantage for the CMZSY HE-BMG as biodegradable implant.

HA has the ability to bond directly to bone tissue without anintervening fibrous layer based on the chemical resemblance of

represents the standard deviation. n = 6; ⁄p < 0.05, ⁄⁄p < 0.01.

Fig. 14. Histological observation of cortical bone by light microscopy with H&E staining. (a, c) Normal bone and the bone marrow cavity; (b, d) CMZSY HE-BMG rod in thefemur of a mouse 4 weeks after implantation.

Fig. 15. Comparison of (a) the mechanical properties and (b) the corrosion behavior between CMZSY HE-BMG and the as-reported biodegradable BMGs [12,14,15,34].

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8571

HA to bone minerals and it plays an important role in bone forma-tion and the normal functions of bone tissue [63].

It is well known that Sr-HA has better thermal stability, biocom-patibility and surface activity compared with pure HA [64]. Previ-ous in vitro and in vivo studies have shown that Sr-HA promotesosteoblast response and stimulates new bone formation [65].

To sum up, CMZSY HE-BMG has excellent mechanical propertiesand could stimulate bone formation. It therefore shows greatpotential for the prevention and cure of osteoporosis and fracture.This new kind of BMG alloy, with a large amount of disorder, mayherald a new era of BMGs in biomedical applications.

As the research of HE-BMGs is still in its infancy, many new HE-BMGs are still unexploited and their mechanisms need to be ex-plored and studied much more deeply in future research work.

6. Conclusions

In this study, CMZSY HE-BMG was prepared and investigatedin vitro and in vivo as a biodegradable metallic material. The fol-lowing conclusions can be drawn.

(1) The CMZSY HE-BMG showed higher compressionstrength than the previously reported prototype ternaryCMZ BMG and has a similar Young’s modulus to corticalbone.

(2) By adding the alloying elements Sr and Yb to the prototypeternary CMZ BMG, a new kind of BMG, i.e. a high-entropyBMG (CMZSY HE-BMG), has been created which has muchbetter corrosion resistance.

8572 H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573

(3) In vitro tests showed that CMZSY HE-BMG samples couldstimulate the proliferation and differentiation of culturedosteoblasts.

(4) In vivo tests showed that, although the CMZSY HE-BMGsamples did not show any obvious degradation after 4 weeksof implantation, they can promote osteogenesis and newbone formation rapidly, after just 2 weeks of implantation.

Acknowledgements

This work was supported by the National Basic Research Pro-gram of China (973 Program) under Grant No. 2012CB619102, Na-tional Science Fund for Distinguished Young Scholars (Grant No.51225101), the National Natural Science Foundation of China un-der Grant No. 31170909, the Research Fund for the Doctoral Pro-gram of Higher Education under Grant No. 20100001110011, theNatural Science Foundation of Heilongjiang Province (ZD201012)and the Guangdong Province innovation R&D team project underGrant No. 201001C0104669453. The authors thank W. Jiao fromthe Institute of Physics, Chinese Academy of Sciences, for his helpwith the preparation of materials.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 2, 3, 5, 6, 8–15,are difficult to interpret in black and white. The full colour imagescan be found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2013.01.029.

References

[1] Ashby M, Greer A. Metallic glasses as structural materials. Scripta Mater2006;54:321–6.

[2] Qin C, Oak J, Ohtsu N, Asami K, Inoue A. XPS study on the surface films of anewly designed Ni-free Ti-based bulk metallic glass. Acta Mater2007;55:2057–63.

[3] Oak JJ, Louzguine-Luzgin DV, Inoue A. Fabrication of Ni-free Ti-based bulk-metallic glassy alloy having potential for application as biomaterial, andinvestigation of its mechanical properties, corrosion, and crystallizationbehavior. J Mater Res 2007;22:1346–53.

[4] Qiu C, Chen Q, Liu L, Chan K, Zhou J, Chen P, et al. A novel Ni-free Zr-based bulkmetallic glass with enhanced plasticity and good biocompatibility. ScriptaMater 2006;55:605–8.

[5] Liu L, Qiu C, Chen Q, Chan K, Zhang S. Deformation behavior, corrosionresistance, and cytotoxicity of Ni-free Zr-based bulk metallic glasses. J BiomedMater Res Part A 2008;86:160–9.

[6] Liu L, Qiu C, Sun M, Chen Q, Chan K, Pang GKH. Improvements in the plasticityand biocompatibility of Zr–Cu–Ni–Al bulk metallic glass by the microalloyingof Nb. Mater Sci Eng A 2007;449:193–7.

[7] Wang Y, Li H, Cheng Y, Wei S, Zheng Y. Corrosion performances of a nickel-freeFe-based bulk metallic glass in simulated body fluids. Electrochem Commun2009;11:2187–90.

[8] Wang YB, Li HF, Zheng YF, Li M. Corrosion performances in simulated bodyfluids and cytotoxicity evaluation of Fe-based bulk metallic glasses. Mater SciEng C 2012;32:599–606.

[9] Wang Y, Xie X, Li H, Wang X, Zhao M, Zhang E, et al. Biodegradable CaMgZnbulk metallic glass for potential skeletal application. Acta Biomater2011;7:3196–208.

[10] Cao JD, Kirkland NT, Laws KJ, Birbilis N, Ferry M. Ca–Mg–Zn bulk metallicglasses as bioresorbable metals. Acta Biomater 2012;8:2375–83.

[11] Zberg B, Uggowitzer PJ, Löffler JF. MgZnCa glasses without clinically observablehydrogen evolution for biodegradable implants. Nat Mater 2009;8:887–91.

[12] Gu X, Zheng Y, Zhong S, Xi T, Wang J, Wang W. Corrosion of, and cellularresponses to Mg–Zn–Ca bulk metallic glasses. Biomaterials2010;31:1093–103.

[13] Jiao W, Zhao K, Xi XK, Zhao DQ, Pan MX, Wang WH. Zinc-based bulk metallicglasses. J Non-Cryst Solids 2010;356:1867–70.

[14] Jiao W, Li HF, Zhao K, Bai HY, Wang YB, Zheng YF, et al. Development of CaZnbased glassy alloys as potential biodegradable bone graft substitute. J Non-Cryst Solids 2011;357:3830–40.

[15] Li HF, Zhao K, Wang YB, Zheng YF, Wang WH. Study on bio-corrosion andcytotoxicity of a Sr-based bulk metallic glass as potential biodegradable metal.J Biomed Mater Res B Appl Biomater 2012;100B:368–77.

[16] Li HF, Wang YB, Cheng Y, Zheng YF. Surface modification of Ca60Mg15Zn25bulk metallic glass for slowing down its biodegradation rate in water solution.Mater Lett 2010;64:1462–4.

[17] Lin YC, Cho YH. Elucidating the microstructure and wear behavior formulticomponent alloy clad layers by in situ synthesis. Surf Coat Technol2008;202:4666–72.

[18] Yeh JW, Chen SK, Lin SJ, Gan JY, Chin TS, Shun TT, et al. Nanostructured high-entropy alloys with multiple principal elements: novel alloy design conceptsand outcomes. Adv Eng Mater 2004;6:299–303.

[19] Inoue A. Stabilization of metallic supercooled liquid and bulk amorphousalloys. Acta Mater 2000;48:279–306.

[20] Skoryna SC. Metabolic aspects of the pharmacologic use of trace elements inhuman subjects with specific reference to stable strontium. Trace SubstEnviron Health 1984;18:3–23.

[21] Cabrera WE, Schrooten I, De Broe ME, D’Haese PC. Strontium and bone. J BoneMiner Res 1999;14:661–8.

[22] Arlot M, Roux J, Boivin G, Perrat B, Tsouderos Y, Deloffre P, et al. Effects of astrontium salt (S12911) on both tibial metaphysis and epiphysis in normalgrowing rats. J Bone Miner Res 1995;10:M415.

[23] Marie PJ. Strontium ranelate: a dual mode of action rebalancing bone turnoverin favour of bone formation. Curr Opin Rheumatol 2006;18:S11–5.

[24] Su Y, Bonnet J, Deloffre P, Tsouderos Y, Baron R. The strontium salt S12911inhibits the expression of carbonic anhydrase and the vitronectin receptor inchicken bone marrow cultures and bone resorption in mouse calvaria andisolated rat osteoclasts. J Bone Miner Res 1992;7:S306.

[25] Canalis E, Hott M, Deloffre P, Tsouderos Y, Marie P. The divalent strontium saltS12911 enhances bone cell replication and bone formation in vitro. Bone1996;18:517–23.

[26] Neizvestnova E, Grekhova T, Privalova L, Babakova O, Konovalova N, Petelina E.Hygienic regulation of yttrium, terbium, ytterbium and lutetium fluorides inthe air of the workplace. Med Tr Prom Ekol 1994;7:32–5.

[27] Wang M, Mi CC, Wang WX, Liu CH, Wu YF, Xu ZR, et al. Immunolabeling andNIR-excited fluorescent imaging of HeLa cells by using NaYF4:Yb, Erupconversion nanoparticles. ACS Nano 2009;3:1580–6.

[28] Chakraborty S, Das T, Banerjee S, Subramanian S, Sarma HD, Venkatesh M.175Yb-labeled hydroxyapatite: a potential agent for use in radiationsynovectomy of small joints. Nucl Med Biol 2006;33:585–91.

[29] American Society for Testing and Materials. ASTM E9–89a (2000): standardtest methods of compression testing of metallic materials at roomtemperature. In: Annual book of ASTM standards. Philadelphia,PA: American Society for Testing and Materials; 2000.

[30] Chang E, Lee T. Effect of surface chemistries and characteristics of Ti6Al4V onthe Ca and P adsorption and ion dissolution in Hank’s ethylene diamine tetra-acetic acid solution. Biomaterials 2002;23:2917–25.

[31] ASTM-G31-72: standard practice for laboratory immersion corrosion testing ofmetals. In: Annual book of ASTM standards. Philadelphia, PA: AmericanSociety for Testing and Materials; 2004.

[32] Song G, Atrens A. Understanding magnesium corrosion—a framework forimproved alloy performance. Adv Eng Mater 2004;5:837–58.

[33] Siu W, Qin L, Cheung WH, Leung K. A study of trabecular bones inovariectomized goats with micro-computed tomography and peripheralquantitative computed tomography. Bone 2004;35:21–6.

[34] Wang G, Liaw PK, Senkov ON, Miracle DB, Morrison ML. Mechanical andfatigue behavior of Ca65Mg15Zn20 bulk-metallic glass. Adv Eng Mater2009;11:27–34.

[35] Senkov O, Miracle D, Keppens V, Liaw P. Development and characterization oflow-density Ca-based bulk metallic glasses: an overview. Metall Mater Trans A2008;39:1888–900.

[36] Staiger MP, Pietak AM, Huadmai J, Dias G. Magnesium and its alloys asorthopedic biomaterials: a review. Biomaterials 2006;27:1728–34.

[37] Xin Y, Huo K, Tao H, Tang G, Chu PK. Influence of aggressive ions on thedegradation behavior of biomedical magnesium alloy in physiologicalenvironment. Acta Biomater 2008;4.

[38] Zheng Y, Gu X, Xi Y, Chai D. In vitro degradation and cytotoxicity of Mg/Cacomposites produced by powder metallurgy. Acta Biomater 2010;6:1783–91.

[39] 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.

[40] Li ZJ, Gu XN, Lou SQ, Zheng YF. The development of binary Mg–Ca alloysfor use as biodegradable materials within bone. Biomaterials 2008;29:1329–44.

[41] Zhou YJ, Zhang Y, Wang YL, Chen GL. Microstructure and compressiveproperties of multicomponent Alx(TiVCrMnFeCoNiCu)100�x high-entropyalloys. Mater Sci Eng A 2007;454–455:260–5.

[42] Du Y, Lu Y, Li T, Wang T, Zhang G. Effect of aluminium content ofAlxCrFe1�5Ni0�5 multiprincipal alloys on microstructure and alloy hardness.Mater Res Innovations 2011;15:107–10.

[43] Tang WY, Yeh JW. Effect of aluminum content on plasma-nitridedAlxCoCrCuFeNi high-entropy alloys. Metall Mater Trans A 2009;40:1479–86.

[44] Gunde P, Hanzi AC, Sologubenko AS, Uggowitzer PJ. High-strength magnesiumalloys for degradable implant applications. Mater Sci Eng A 2011;528:1047–54.

[45] Yu WB, Liu ZY, He H, Cheng NP, Li XL. Microstructure and mechanicalproperties of ZK60–Yb magnesium alloys. Mater Sci Eng A 2008;478:101–7.

H.F. Li et al. / Acta Biomaterialia 9 (2013) 8561–8573 8573

[46] Fang HC, Chen KH, Zhang Z, Zhu CJ. Effect of Yb additions on microstructuresand properties of 7A60 aluminum alloy. Trans Nonferrous Met Soc China2008;18:28–32.

[47] Liao HC, Sun Y, Sun GX. Correlation between mechanical properties andamount of dendritic a-Al phase in as-cast near-eutectic Al-11.6% Si alloysmodified with strontium. Mater Sci Eng A 2002;335:62–6.

[48] Greer AL. Confusion by design. Nature 1993;366:303–4.[49] Navarro M, Michiardi A, Castaño O, Planell JA. Biomaterials in orthopaedics. J R

Soc Interface 2008;5:1137–58.[50] Bornapour M, Muja N, Shum-Tim D, Cerruti M, Pekguleryuz M.

Biocompatibility and biodegradability of Mg–Sr alloys: the formation of Sr-substituted hydroxyapatite. Acta Biomater 2013;9:5319–30.

[51] Kung K-C, Lee T-M, Lui T-S. Bioactivity and corrosion properties of novelcoatings containing strontium by micro-arc oxidation. J Alloys Compd2010;508:384–90.

[52] Bleich HL, Moore MJ, Lemann Jr J, Adams ND, Gray RW. Urinary calciumexcretion in human beings. N Engl J Med 1979;301:535–41.

[53] Morii H. Magnesium and bone. JJPEN 2001;23:511–3.[54] Tapiero H, Tew KD. Trace elements in human physiology and pathology: zinc

and metallothioneins. Biomed Pharmacother 2003;57:399–411.[55] Ishii K, Usui S, Sugimura Y, Yoshida S, Hioki T, Tatematsu M, et al.

Aminopeptidase N regulated by zinc in human prostate participates intumor cell invasion. Int J Cancer 2001;92:49–54.

[56] Pors Nielsen S. The biological role of strontium. Bone 2004;35:583–8.

[57] Cheung KMC, Lu WW, Luk KDK, Wong CT, Chan D, Shen JX, et al.Vertebroplasty by use of a strontium-containing bioactive bone cement.Spine 2005;30:S84–91.

[58] Ammann P. Strontium ranelate: a novel mode of action leading to renewedbone quality. Osteoporos Int 2005;16:11–5.

[59] Marie PJ. Strontium ranelate: new insights into its dual mode of action. Bone2007;40:S5–8.

[60] Blake G, Zivanovic M, McEwan A, Ackery D. Sr-89 therapy: strontium kineticsin disseminated carcinoma of the prostate. Eur J Nucl Med Mol Imaging1986;12:447–54.

[61] Chang W, Tu C, Chen TH, Komuves L, Oda Y, Pratt SA, et al. Expression andsignal transduction of calcium-sensing receptors in cartilage and bone.Endocrinology 1999;140:5883–93.

[62] Grynpas M, Hamilton E, Cheung R, Tsouderos Y, Deloffre P, Hott M, et al.Strontium increases vertebral bone volume in rats at a low dose that does notinduce detectable mineralization defect. Bone 1996;18:253–9.

[63] Aubin JE, Liu F, Bilezikian JP, Raisz LG, Rodan GA, editors. Principles of bonebiology. San Diego, CA: Academic Press; 1996. p. 51.

[64] Suganthi RV, Elayaraja K, Joshy MIA, Chandra VS, Girija EK, Kalkura SN. Fibrousgrowth of strontium substituted hydroxyapatite and its drug release. Mater SciEng C Mater Biol Appl 2011;31:593–9.

[65] Habibovic P, Barralet JE. Bioinorganics and biomaterials: bone repair. ActaBiomater 2011;7:3013–26.