Acta Biomaterialia 64 (2017) 421–436

Contents lists available at ScienceDirect

Acta Biomaterialia

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

Full length article

Development of magnesium-based biodegradable metals with dietarytrace element germanium as orthopaedic implant applications

https://doi.org/10.1016/j.actbio.2017.10.0041742-7061/� 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

⇑ Corresponding author.⇑⇑ Co-corresponding author.

E-mail addresses: jiangbaoguo@vip.sina.com (B. Jiang), yfzheng@pku.edu.cn(Y. Zheng).

Dong Bian a, Weirui Zhou a, Jiuxu Deng b, Yang Liu a, Wenting Li a, Xiao Chu c, Peng Xiu d, Hong Cai d,Yuhui Kou b, Baoguo Jiang b,⇑⇑, Yufeng Zheng a,⇑aDepartment of Materials Science and Engineering, College of Engineering, Peking University, Beijing 100871, ChinabDepartment of Trauma and Orthopedics, People’s Hospital Peking University, Beijing 100044, ChinacDepartment of Orthopedics, Guangdong Key Lab of Orthopaedic Technology and Implant Materials, Guangzhou General Hospital of Guangzhou Military Command,Guangzhou 510010, ChinadDepartment of Orthopedics, Peking University Third Hospital, Beijing 100191, China

a r t i c l e i n f o

Article history:Received 19 June 2017Received in revised form 13 September2017Accepted 3 October 2017Available online 4 October 2017

Keywords:Mg-Ge alloyBiodegradable metalsDegradationIn vivoOsseointegration

a b s t r a c t

From the perspective of element biosafety and dietetics, the ideal alloying elements for magnesiumshould be those which are essential to or naturally presented in human body. Element germanium is aunique metalloid in the carbon group, chemically similar to its group neighbors, Si and Sn. It is a dietarytrace element that naturally presents in human body. Physiological role of Ge is still unanswered, but itmight be necessary to ensure normal functioning of the body. In present study, novel magnesium alloyswith dietary trace element Ge were developed. Feasibility of those alloys to be used as orthopaedicimplant applications was systematically evaluated. Mg-Ge alloys consisted of a-Mg matrix and eutecticphases (a-Mg + Mg2Ge). Mechanical properties of Mg-Ge alloys were comparable to current Mg-Ca,Mg-Zn and Mg-Sr biodegradable metals. As-rolled Mg-3Ge alloy exhibited outstanding corrosion resis-tance in vitro (0.02 mm/y, electrochemical) with decent corrosion rate in vivo (0.6 mm/y, in rabbit tibia).New bone could directly lay down onto the implant and grew along its surface. After 3 months, bone andimplant were closely integrated, indicating well osseointegration being obtained. Generally, this is apioneering study on the in vitro and in vivo performances of novel Mg-Ge based biodegradable metals,and will benefit the future development of this alloy system.

Statement of Significance

The ideal alloying elements for magnesium-based biodegradable metals should be those which are essen-tial to or naturally presented in human body. Element germanium is a unique metalloid in the carbongroup. It is a dietary trace element that naturally presents in human body. In present study, feasibilityof Mg-Ge alloys to be utilized as orthopedic applications was systematically investigated, mainly focusingon the microstructure, mechanical property, corrosion behavior and biocompatibility. Our findingsshowed that Mg-3Ge alloy exhibited superior corrosion resistance to current Mg-Ca, Mg-Zn and Mg-Sralloys with favorable biocompatibility. This is a pioneering study on the in vitro & in vivo performancesof Mg-Ge biodegradable metals, and will benefit the future development of this alloy system.

� 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Crystalline magnesium (Mg) and magnesium-based alloysspecifically designed for biodegradable implant applications have

attracted great attention since the early 2000 s [1–4]. The idealMg-based biodegradable metals are expected to corrode gradu-ally in vivo, with an appropriate host response, then dissolvecompletely upon fulfilling the mission to assist with the tissuehealing [5]. During degradation, elements in the implant matrixwill release in the form of metal ions and then they participatein tissue healing positively, neutrally or negatively, depending onthe dose and releasing rate. Therefore, besides magnesium, themajor components of Mg-based biodegradable metals should

422 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

be preferably essential elements or trace elements in humanbody.

Many kinds of essential/possibly essential elements found inhuman body could be used as alloying elements to improvein vitro and in vivo performances of Mg-based biomaterials, includ-ing corrosion control, mechanical improvement, biological regula-tion, etc. Till now, many kinds of biomedical Mg-based alloys withessential/possibly essential elements addition have been devel-oped, including binary Mg-Ca [6–8], Mg-Zn [9,10], Mg-Mn [11],Mg-Fe [12], Mg-Cu [13,14], Mg-Sr [15–18], Mg-Si [19], Mg-Sn[20–23], Mg-Li [24] and multicomponent Mg alloys based on thesebinary alloys, Mg-Zn-Ca [25] and Mg-Si-Ca-Zn [26], etc.

Germanium (Ge) is a lustrous, hard, grayish-white metalloid inthe carbon group, chemically similar to its group neighbors, Si andSn. It is usually used as a semiconductor in transistors and variousother electronic devices. Ge presents in human body and it is adietary trace element which might be necessary to ensure normalfunctioning of the body [27]. For the general population, the majorsource of Ge is from food. Ge presents in practically all food andonly in minute amounts [28]. The estimated average dietary intakeof Ge in humans is in the range of 0.4–3.4 mg per day. Ge normal-izes many physiological functions such as lowering of high bloodpressure in humans, restoring deviant blood characteristics to theirnormal range including pH, glucose, the minerals sodium, potas-sium, etc. [29]. Many pharmacological effects of some organic ger-manium compounds have been investigated. They includeantitumor, anticancer, anti-inflammation, antioxidation, radiopro-tection and pro-immune functions [27]. Organic germaniumappears to be remarkably safe while relatively high doses of ger-manium dioxide and other inorganic germanium compounds cancause severe poisoning including fatal cases [28,30]. Animal stud-ies have already showed promising results, despite sufficient clin-ical trials are still needed [28].

From the viewpoint of material science, limited informationabout Mg-Ge alloys are related to thermodynamic characters[31,32], and the effects of Ge on the microstructure and mechan-ical property [33,34]. To the authors’ knowledge, no report onthe corrosion behavior, in vitro or in vivo biocompatibility ofMg-Ge alloys can be found. Recently, Liu et al. [33] reportedtheir experimental results on the binary Mg-Ge alloys. Theyfound that microstructures of Mg-Ge binary alloys consisted oftwo components: a-Mg dendrites and a-Mg + Mg2Ge eutecticphase. Grain size of a-Mg dendrites was effectively refined withGe addition and tensile strengths of Mg-Ge binary alloys wereenhanced while elongations were reduced. Benefits of Ge onthermal stability and mechanical property of Mg have also beenverified in another work [34]. Even though no referable literatureof Ge on Mg corrosion can be found, effects of this intrinsic met-alloid/semimetallic element on Mg degradation is really interest-ing. The continuously net-like second phase (a-Mg + Mg2Ge) inMg-Ge alloys, as reported in previous work [33,34], might actas a barrier during corrosion [35].

In the present study, binary Mg-xGe (x = 1.5, 2.5 and 3 wt%)alloys were designed and fabricated through traditional casting.The Ge content range (1.5–3 wt%) was determined by referring toRefs. [33] and [34], as the strength was not further improved whenGe content exceeded 3.4 wt% while significantly impaired plastic-ity occurred. For biosafety purpose, Ge content should also be lim-ited in a proper low range. To further improve mechanicalproperty, these binary alloys were hot rolled into thin plates. Theirfeasibility to be utilized as orthopedic applications was systemati-cally investigated, mainly focusing on the microstructure, mechan-ical property, corrosion behavior, in vitro cytocompatibility andin vivo biocompatibility.

2. Experimental details

2.1. Material preparation

Mg-Ge alloys were melted and casted by using commercialmagnesium (99.95 wt%) and commercial pure Ge powder(99.9 wt%) under the protection of a mixed gas atmosphere ofSF6 (1 vol%) and CO2 (99 vol%). Briefly, Mg and Ge powders werewell blended in a mixer, then melted at 700–800 �C for 30 minbefore casting. The ingots were machined into 5 mm thick platesand homogenized at 400 �C for 3 h before rolling. Afterwards, theywere hot rolled into thin plates with a final thickness of 2.5 mm ata rolling reduction of 0.2 mm each pass. The plate was annealed at400 �C for 10 min at each rolling interval. The actual Ge contents inthe Mg-xGe alloys were 1.38 ± 0.30, 2.53 ± 0.21 and 3.26 ± 0.20 wt%, respectively. Specimens for microstructural characterization,corrosion test and in vitro biocompatibility evaluation were cutinto 10 � 10 � 2 mm3 disks. Cylindrical pins with a diameter of2.2 mm and length of 10 mm were machined parallel to the rollingdirection for in vivo test. Samples were mechanically groundedwith SiC sandpaper up to 2000 grit, ultrasonically cleaned in ace-tone and absolute ethanol, then dried in hot air. For cell assays,samples were sterilized under ultraviolet radiation for 2 h beforeuse. Pins for animal test were separately encapsulated and steril-ized under Co60 c ray radiation at 25 KGy.

2.2. Microstructural characterization

Specimens for microstructural observation were polished intomirror-like surface with 5 lm diamond polishing paste. After-wards, they were etched in 4% HNO3/alcohol solution and observedunder an optical microscope (Olympus BX51M, Japan). Averagegrain size and phase proportion were calculated in ImageJ software(ImageJ 1.43 u, USA) by analyzing microstructural images (n = 5).An X-ray diffractometer (XRD, Rigaku DMAX 2400, Japan) wasemployed to identify the constituent phases by using Cu Ka radia-tion at a scan rate of 4�/min operated at 40 kV and 100 mA.

2.3. Tensile test

Tensile specimens were machined parallel to the rolling direc-tion according to ASTM-E8-04 [36]. Tensile tests were performedat a crosshead speed of 1 mm/min on a universal material testingmachine (Instron 5969, USA) at room temperature. An average ofat least three parallel samples was taken for each group.

2.4. Electrochemical test

A traditional three-electrode cell system was used for electro-chemical test with a platinum foil as counter electrode (CE), a sat-urated calomel electrode (SCE) as reference electrode.Experimental samples with an exposed area of 1 cm2 acted asworking electrode. Measurements were carried out in Hank’s solu-tion [37], on an electrochemical workstation (CHI660C, China).Open circuit potential (OCP) was continuously monitored for3600 s and then electrochemical impedance spectroscopy (EIS)was measured from 100 kHz to 10 mHz using a 5 mV amplitudeperturbation. Potentiodynamic polarization tests were performedat a scanning rate of 1 mV/s and the initial potential was 300 mVbelow the corrosion potential (Ecorr). Electrochemical parametersand corrosion rates were calculated according to ASTM-G102-89[38]. Three duplicate samples were taken for each material for sta-tistical analysis.

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 423

2.5. Immersion test

Immersion tests were carried out in Hank’s solution at37 ± 1 �C according to ASTM-G31-72 [39], with an exposure ratioof 20 mL/cm2. The volume of emerged hydrogen was recorded byusing a calibrated burette in accordance with Ref. [35]. The corre-sponding corrosion rate was calculated based on the followingequation [40]: pH = 2.279VH (PH: corrosion rate, mm/y; VH: hydro-gen evolution rate, mL/cm2/d). After immersion for 3 and 20 days,samples were removed out of the solution, gently rinsed withdistilled water, and dried in open air. Surface morphology as wellas cross section morphology of the corroded samples was observedunder an environmental scanning electron microscope (ESEM, FEIQuanta-200F, the Netherlands), operating in backscatteringelectron mode. Surface chemistry of the corroded samples wasdetermined by using an imaging X-ray photoelectron spectrometer(XPS, Axis Ultra, Kratos Analytical Ltd.) with Al Ka radiation. Highresolution narrow scanning was performed to determine thebinding states of Mg 2p, O 1s, Ca 2p, P 2p, C 1s, and Na 1s. Fouriertransform infrared spectroscopy (FTIR, Nicolet iS 50, ThermoScientific) was utilized to identify the functional groups in thecorrosion products. The spectrum was recorded from 4000 to500 cm�1. Constituent phases in the corrosion product layer weredetermined by using an X-ray diffractometer (XRD, X’Pert ProMPD). The test was performed by using a Mono-Capillary PreFixModule operated at 45 kV and 40 mA.

2.6. Cytotoxicity

Cytotoxicity of the experimental materials was evaluated byindirect cell assay, in which human osteoblast-like MG63 cell linewas used. Cells were cultured in minimum essential medium(MEM) supplied with 10% fetal bovine serum (FBS), 100 U/mL peni-cillin and 100 lg/mg streptomycin at 37 �C in a humidified atmo-sphere with 5% CO2. Alloy extracts were prepared by using aserum-free MEM with an extraction ratio of 1 cm2/mL under stan-dard cell culture condition. After incubation for 72 h, the super-natant fluid was withdrawn, centrifuged and stored at 4 �Cbefore use. Ion concentrations and pH values of the extracts weremeasured by inductively coupled plasma atomic emission spec-troscopy (ICP-AES, Leeman) and a pH meter (PB-10, sartorius),respectively.

Methylthiazol tetrazolium (MTT) assay was adopted to evaluatethe cytotoxicity of the alloy extracts according to ISO 10993-5:2009(E) [41]. In brief, cells were seeded onto 96 well cultureplate at a density of 3 � 103 cells per 100 lL. After incubation for24 h to allow attachment, the medium was replaced by alloyextracts with normal MEM as control. After 1, 3 and 5 days’ incuba-tion, 10 lL MTT was added to each well and another 4 h was kept.Afterwards, 100 lL formazan solubilizing solution (10% SDS in0.01 M HCl) was added into each well and left overnight in theincubator. Then, spectrophotometric absorbance of each well wasmeasured by using a microplate reader (Bio-Rad 680) at 570 nmwith a reference wavelength of 630 nm.

2.7. Alkaline phosphate activity measurement

Alkaline phosphate activity (ALP) of MG63 cells was determinedby hydrolysis of p-nitrophenyl phosphate (p-Npp) using a test kit(Nanjing Jiancheng Bioengineering Institute, China). Briefly, after5-day cultivation with extracts, cells were rinsed with a phosphatebuffer solution (PBS, pH 7.4) for three times and then lysed in 0.1%Triton X-100 by freezing and thawing for 60 min at 37 �C. 30 lL ofthe resulting cell lysates was then transferred to each well of a new96 well plate, cultured with 50 lL carbonated buffer solution (pH= 10) and 50 lL substrate solution (4-amino-antipyrine) for 15 min

at 37 �C. Afterwards, 150 lL of chromogenic agent (potassium fer-ricyanide) was added into the mixed solution. The spectrophoto-metric absorbance was measured at 520 nm by a microplatereader (Bio-RAD680). For normalization, total protein contentwas determined using a Bicinchoninic Acid (BCA) protein assaykit (Beijing Biosea Biotechnology, China). The ALP activity wasexpressed and normalized by the total protein content (U/gprot).

2.8. Animal test

2.8.1. Animal model and surgeryBased on the in vitro performances of Mg-xGe alloys, as-rolled

Mg-3Ge with the best mechanical properties and superior corro-sion resistance was chosen to perform in vivo test. Twelve femaleNew Zealand rabbits (2–2.5 kg) were provided by Laboratory Ani-mal Center of Peking University People’s Hospital (animal use per-mit No.: SYXK (jing) 2011-0010). Animals were kept in separatedcages and health condition was verified by preoperative observa-tion 1 week before operation. Three animals were set as normalcontrol and they were sacrificed at 3 months. The remaining nineanimals all received implantation. Briefly, animals were anaes-thetized with 30 mg/kg pentobarbital sodium by ear vein injection.Operation sites beneath the bilateral knee-joint were shaved andsterilized followed by decortication. A pre-drilled hole, 2.2 mm indiameter, was made on lower edge of the lateral tibial plateau byusing an orthopaedic drill. An as-rolled Mg-3Ge pin was insertedinto the predrilled hole. Then, 4 � 104 U penicillin was droppedin the incision site before suturing. 40 � 104 U penicillin wasintramuscular injected in case of postoperative infection and thisinjection lasted for 3 days. The anesthetic, surgical andpost-operative care protocols were examined by and fulfilled therequirements of the Ethics Committee of Peking UniversityPeople’s Hospital.

2.8.2. Postoperative observation, micro-CT, and histological evaluationAll animals were clinically examined for general condition and

in particular for significant signs of lameness, infection, subcuta-neous emphysema formation and loss of appetite. Three animalswere sacrificed by over-dosage anesthesia at each time pointpost-operation (1, 2 and 3 months), respectively. Rabbit tibias wereretrieved and the implantation section (�1.5 cm) was truncated formicro-CT analysis and histological analysis. Bones were scanned ina self-built micro-CT device at a spatial resolution of 35 lm. Three-dimensional (3D) reconstruction was performed in Amira software(Amira 5.4.1, Visage Imaging). Retrieved bones were fixed in 10%formalin and dehydrated in gradient ethanol/distilled water mix-tures. Then, they were embedded in methacrylate and sectionedinto 150 lm slices on a EXAKT system (EXAKT Apparatebau,Norderstedt, Germany). The hard tissue sections were furthergrounded to 50–80 lm before staining with hematoxylin-eosin(H&E) or toluidine blue (Guge Biological Technology Co., Ltd.).Details about the hard tissue slicing process could be found inour previous work [42]. Animal liver and kidney were alsoretrieved for pathological examination by conventional H&Estaining.

2.9. Statistical analysis

Data in this work were expressed as means ± standarddeviation. Statistical analysis was performed with SPSS 18.0software (SPSS Inc., Chicago, USA). Differences between groupswere analyzed using one-way analysis of variance (ANOVA)followed by Tukey test. The statistical significance was defined asd p < 0.05.

424 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

3. Results

3.1. Microstructure

Fig. 1 shows the optical microstructures, average grain size andcorresponding XRD patterns of experimental materials. All of thealloys were composed of two phases, a-Mg and intermetallic com-pound Mg2Ge with face centered cubic crystal system, as shown inFig. 1(c). As-cast Mg-Ge alloys consisted of dendrites of a-Mgmatrix (bright white dendrites) and eutectic phases (a-Mg + Mg2Ge, brown rod-like/network shaped) at the interdendritic regions,as shown in Fig. 1(a). Area fraction of the eutectic phases increasedfrom 17.0 ± 1.3% in Mg-1.5Ge to 24.3 ± 2.9% in Mg-2.5Ge and thento 33.9 ± 2.5% of Mg-3Ge. Meanwhile, average size of the eutecticphase continuously increased with increasing Ge addition. Thea-Mg dendrites in Mg-2.5Ge and Mg-3Ge were significantly

Fig. 1. (a) Microstructures of as-cast and as-rolled Mg-Ge alloys; (b) average size of thexperimental alloys.

refined compared to Mg-1.5Ge, as shown in Fig. 1(b). Microstruc-tures of the as-rolled Mg-Ge alloys were much more uniform com-pared to their as-cast counterparts. Area fractions of the eutecticphases in the as-rolled Mg-Ge alloys (16.2 ± 1.1%, 25.1 ± 2.3% and35.9 ± 4.3%) were similar to their as-cast counterparts. Besides,Eutectic phases were significantly refined after rolling.

3.2. Mechanical property

Fig. 2 displays the tensile mechanical behaviors of as-cast andas-rolled Mg-Ge alloys. Tensile yield strengths (YS) and ultimatetensile strengths (UTS) of the as-cast Mg-Ge alloys were in therange of 36.6–51.4 MPa and 116.4–143.2 MPa, respectively.As-cast Mg-2.5Ge and Mg-3Ge exhibited higher strengths andelongations than as-cast Mg-1.5Ge. After rolling, YSs and UTSs ofMg-Ge alloys were significantly improved, however, elongations

e a-Mg matrix and eutectic phase, d p < 0.05 and * p < 0.01; (c) XRD patterns of

Fig. 2. (a) Tensile stress-strain curves of experimental Mg-Ge alloys; (b) tensile yield strengths, ultimate tensile strengths and elongations of Mg-Ge alloys.

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 425

of Mg-1.5Ge and Mg-2.5Ge were impaired. Significantly improvedYS, UTS and elongation were found in as-rolled Mg-3Ge as theywere improved to 175.1 ± 5.1 MPa, 236.4 ± 0.7 MPa and 17.7 ± 2.5%, respectively.

3.3. Corrosion behavior

OCP, EIS and Potentiodynamic polarization results were dis-played in Fig. 3(a)–(c) and corresponding electrochemical parame-ters were listed in Table 1. OCP values of all the experimentalmaterials could generally stabilize after immersion for 1000 s.OCP curves of as-cast Mg-1.5Ge and as-rolled Mg-1.5Ge fluctuatedfrequently compared to the remaining ones, implying that it wasdifficult to reach a steady state for the formation and dissolutionof the corrosion product film at the electrode/solution interface.

Fig. 3. (a) OCP curves of Mg-Ge alloys in Hank’s solution; (b) potentiodynamic polarizativolume of evolved hydrogen during immersion; (e) magnesium ion released into Hank’

For potentiodynamic polarization curves, the cathodic polarizationcurrent reflected the severity of hydrogen evolution reaction onplatinum electrode while the anodic branch represented dissolu-tion of magnesium. As-cast Mg-2.5Ge, as-cast Mg-3Ge and as-rolled Mg-3Ge exhibited lower cathodic current density comparedto the remaining alloys. A little potential increase above the Ecorrwould cause a soaring increase in anodic current of Mg-1.5Ge.The increase on the current of the anodic branches was reducedfor the samples with higher Ge content (2.5 and 3 wt%) whenapplying the same voltage disturbance above the Ecorr, implyinginhibited anodic reaction. The highest corrosion current densitywas found in as-cast Mg-1.5Ge and it was more than 2 orders ofmagnitude of the lowest one, as-rolled Mg-3Ge. Among as-castMg-Ge alloys, corrosion current density significantly decreasedwith Ge content reaching 2.5 wt%. After hot rolling, reduced

on curves of Mg-Ge alloys in Hank’s solution; (c) Nyquist plots of Mg-Ge alloys; (d)s solution; (f) Ge release after immersion.

Table 1Corrosion rates of Mg-Ge alloys in Hank’s solution.

Alloy Ecorr (V) icorr (lA/cm2) vcorr (mm/y) PH (mm/y)

as-cast Mg-1.5Ge �1.456 188.2 4.32 5.06as-cast Mg-2.5Ge �1.603 1.2 0.03 0.16as-cast Mg-3.0Ge �1.532 0.9 0.02 0.22as-rolled Mg-1.5Ge �1.532 35.6 0.82 1.25as-rolled Mg-2.5Ge �1.542 13.6 0.31 0.58as-rolled Mg-3.0Ge �1.536 0.7 0.02 0.13

Ecorr: corrosion potential; icorr: corrosion current density; vcorr: corrosion rate calculated from polarization plots by using Tafel region extrapolation; PH: corrosion ratecalculated from hydrogen evolution.

Fig. 4. Surface morphologies (a) and cross section morphologies (b) of Mg-Ge alloys after immersion in Hank’s solution for 3 and 20 days. CP in (b) means corrosion products.

426 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

corrosion current densities were revealed for Mg-1.5Ge, whileintensified corrosion was happened in Mg-2.5Ge. As-cast Mg-3Geand as-rolled Mg-3Ge owned the largest loop in the Nyquist plots,demonstrating their superior corrosion resistance among thealloys. Meanwhile, as-cast Mg-1.5Ge exhibited the smallest loop,indicating its poor corrosion resistance. This finding was in agree-ment with the potentiodynamic polarization results.

Evolved hydrogen during immersion test was revealed in Fig. 3(d). At the initial stage of immersion (0–50 h), no much differencewas found among the materials, all showing a low hydrogen evo-lution rate. Thereafter, as-cast Mg-1.5Ge kept being corroded at arelatively high corrosion rate level throughout the whole immer-sion period. However, other alloys maintained a low hydrogen evo-lution trend and hydrogen volume gradually stabilized. Except foras-cast and as-rolled Mg-1.5Ge, volume of generated hydrogen wasall well below 10 mL/cm2 after 20 days. As-cast Mg-2.5Ge, as-castMg-3Ge and as-rolled Mg-3Ge exhibited the slowest hydrogenevolution rate, suggesting their superior corrosion resistance.

Fig. 3(e) and (f) present released ion concentrations of Mg andGe in Hank’s solution at different immersion durations. Continuousdissolution of Mg and alloying element Ge into Hank’s solution wasfound as their concentrations increased persistently with pro-longed immersion time. Mg and Ge concentrations in Hank’s solu-tion decreased with increased Ge content in the alloys, suggestingthe positive effect of Ge on corrosion resistance. Hot rolling couldfurther improve corrosion resistance as as-rolled Mg-Ge alloys

exhibited less dissolved Mg and Ge ions than their as-cast counter-parts at day 10 and 20. Results of ion release were well consistentwith the hydrogen evolution behaviors and all the corrosion testsdemonstrated that as-rolled Mg-3Ge exhibited the best corrosionresistance.

Fig. 4(a) shows typical corrosion morphologies of Mg-Ge alloysafter immersion in Hank’s solution for different durations. A corro-sion product layer was formed on sample surface with some whiteclusters/particles deposition. Some microcracks emerged on thesample surface as a result of the shrinkage of corrosion productlayer during dehydration. Severely localized corrosion wasobserved on as-cast Mg-1.5Ge, which was consistent with itshighest corrosion rate among all the alloys. Except for as-castMg-1.5Ge, corrosion product layers on the remaining alloys wererelatively uniform and compact. Judging from the cross-sectionmorphology in Fig. 4(b), differences in the corrosion behaviorscould be easy to tell. As-cast Mg-1.5Ge and as-rolled Mg-1.5Geboth suffered severely localized corrosion. A thick corrosion pro-duct layer with holes and cracks was observed above those deepcorrosion sites. With increasing Ge addition, localized corrosionwas suppressed as the boundary of the Mg-Gematrix becamemoreregular and smoother. Thickness of the corrosion product layerdecreased with the increase of added Ge, suggesting improvementin the corrosion resistance. The thin corrosion product layer onas-cast and as-rolled Mg-3Ge could be hardly distinguished fromthe Mg matrix and surrounding resin. This layer was closely

Fig. 5. (a) FTIR spectra of as-rolled Mg-Ge alloys after immersion in Hank’s solution for 20 days; (b) XRD patterns after 20-d-immersion; (c) XPS analysis of the corrosionproducts; (d) typical high resolution narrow scan results of Mg 2p, O 1s, Ca 2p, P 2p, C 1s and Na 1s among the corroded Mg-Ge alloys.

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 427

adherent to the Mg-3Ge matrix and could possibly provide protec-tion to the matrix.

Fig. 5 shows the XPS, FTIR and XRD results of the corrosionproducts on Mg-Ge alloys. Since the results of as-cast Mg-Ge alloyswere similar to their as-rolled counterparts, so only the as-rolledresults were listed here. XPS results revealed that the corrosionproducts were mainly composed of Mg, O, Ca, P, C and limitedNa, as shown in Fig. 5(c). Ge was not detected mainly because ofits minor content in this layer. FTIR indicated the presence ofH2O, OH�, CO3

2�, PO43� and possible HPO4

2�, as depicted in Fig. 5(a)[43,44]. Further XRD analysis confirmed the existence of Mg(OH)2 in the corrosion products. Apparent diffraction peaks ofCaCO3 were only detected on the Mg-1.5Ge sample. Due to a thickcorrosion product layer on Mg-1.5Ge, strong peaks from Mg(OH)2could be detected with minor signals from the a-Mg matrix. Highresolution narrow scan of XPS also confirmed the presence ofCaCO3 on Mg-2.5Ge and Mg-3Ge, since one Ca 2p3/2 peak assignedto 347.4 eV, accompanied by C 1s at 289.7 eV and O 1s at 531.6 eV,was detected [45]. No obvious CaCO3 signals on the XRD patternsof Mg-2.5Ge and Mg-3Ge should be ascribed to their limitedamounts. Single Mg 2p peak at 50.4 eV along with O 1s at 532.9eV and C 1s at 289.7, as well as the presence of CO3

2� proved byFTIR, indicated possible presence of MgCO3 or hydrated MgCO3

[46]. P showed a single peak at 133.4 eV. This peak was relatedto PO4

3� or HPO42�, as also proved in FTIR. Generally, corrosion prod-

ucts of Mg-Ge alloys were mainly composed of Mg(OH)2 accompa-nied with small amount of CaCO3. Besides, a small quantity ofcarbonates, phosphates or their complex compounds might alsoexist.

3.4. Cytotoxicity and ALP activity

Fig. 6(a) and (b) present pH values and ion concentrations of theextract media. Due to the buffer agent in MEM, no much differencewas found in the pH values among all the extracts. Distinctive ionreleasing behaviors observed in Hank’s solution were not observedin MEM. Magnesium concentrations in various extracts were nolonger apparently different. On the contrary, they seemed to beclose to each other. Magnesium concentrations fluctuated between132 lg/mL and 160 lg/mL, while Ge concentrations were in therange of 1.8–3.8 lg/mL. Ge concentrations in as-rolled Mg-2.5Geand as-rolled Mg-3Ge extracts were apparently higher than theremaining extracts.

Fig. 6(c) and (d) show the MG63 cell viability cultured in Mg-Gealloy extracts and ALP activity after culturing for 5 days. Accordingto ISO 10993-5, cytotoxicity of biomaterials should be Grade 0 or

Fig. 6. (a) pH values and (b) ion concentrations of the extract media; (c) MG63 cell viability and (d) ALP activity to Mg-Ge extracts.

428 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

Grade 1, which means cell viability should exceed 80% [41]. Exceptfor as-rolled Mg-3Ge, the remaining alloys all satisfied this crite-rion, exhibiting good cytocompatibility. Cell viability in as-rolledMg-3Ge extract was significantly lower than other extracts duringthe whole culture period. After 3 days, cell viability of as-rolledMg-3Ge was already lower than 80%, and further impaired viability(47.9 ± 3.6%) was found at day 5. Cell viability of as-rolled Mg-Gealloys was lower than the as-cast ones, mainly due to the higherMg and Ge concentrations in their extracts. Compared to control,increased ALP activity was observed for MG63 cells treated withMg-Ge extracts. As-cast Mg-1.5Ge owned the highest ALP concen-tration among all the experimental materials. Among the Mg-Gealloys in as-rolled condition, as-rolled Mg-3Ge exhibited the high-est ALP concentration.

3.5. In vivo performance

Animals received implantation all survived and no obvioussigns of lameness and loss of appetite were observed. No infectionwas found on wounds or peri-implant tissues through autopsy andmicro-CT examination.

3.5.1. Micro-CT analysisIn vivo degradation of Mg-Ge implants and osseointegration

between implants and surrounding bones were characterized

through micro-CT assessment, as depicted in Fig. 7. Generally,implantation of Mg-Ge pin did not cause observable bone damageand normal bone morphology was observed, as revealed from the3D reconstruction (Fig. 7(a) and (b)).

Continuous degradation of the implant could be observed andno obvious signs of localized severe corrosion was found. In thefirst two months, a few gas bubbles in quite limited size could bedetected in medullary cavity of rabbit tibia, as marked in greencolor in Fig. 7(b). Up to 3 months, the previously existed gas bub-bles around the implant could be no longer observed. During thewhole implantation period, Mg-Ge implant maintained intact, asshown in Fig. 8 (a). Surface morphology of the Mg-Ge implantbecame rougher as time went on. Some mild and shallow corrosionpits could be observed on implant surface after 3 months, implyingthe appearance of minorly localized corrosion. Residual implantvolume calculated from 3D reconstruction is presented in Fig. 8(b). At the initial stage after implantation (�1 mon), the implantsuffered a fast degradation as implant volume significantlyreduced with a volume loss of 15.4 ± 5.9%. Mitigated degradationwas observed in the subsequent 2 months and the implant couldmaintain 75.6 ± 3.2% of its original volume after 3 months.

Bone conductive property of Mg-Ge implant was also proved inthe micro-CT assessment as newly formed bones could directlyattach to the material surface and stretched along implant surface,as clearly revealed in Fig. 7(c). Qualitatively, continuous

Fig. 7. Micro-CT analysis of implant degradation and bone-implant interfaces. (a) 3D reconstruction of bone and implant; (b) 3D reconstruction with different parts renderedwith different colors, semitransparent gray for bone, orange for Mg-Ge implant, green for gas bubble; (c) 2D slice.

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 429

osseointegration and strengthened bonding between bone andimplant could be observed from the 2D and 3D reconstructions.At 1 month, both ends of the implant were embed in the corticalbone, showing fast osseointegration. However, weak bondingcould be recognized as some voids could be found betweenimplant and bone (not completely integrated). In the followingmonth (2 months), two end faces of the implant were graduallycovered by newly formed bones, but still not directly contacted.After 3 months, bone and implant were well integrated as theimplant was tightly sealed by cortical bones. In addition, corticalthickness surrounded the implant was thicker than that far fromthe implant, as revealed in Fig. 7(c). Bone-implant contact ratio(BIC) can quantitatively reflect the osseointegration status. In thisstudy, BIC was calculated as the proportion of bone-implantdirectly contacted area to the total implant surface in Amira soft-ware. Continuously increased BIC value suggested enhancedosseointegration and bone-implant bonding, as shown in Fig. 8(c).

3.5.2. Histological analysisFig. 9 shows tissue responses adjacent to the Mg-Ge implant

after 1, 2 and 3 months with normal tissue as control. Clearly, con-tinuous degradation of the Mg-Ge implant could be observed sincethe implant surface became much rougher with time. Generally,the implant went through a macroscopic uniform corrosion modewithout severely localized corrosion. No abnormality was foundwith bone tissues surrounding the implant, showing good histo-compatibility. No microscopic differences were noted in the tissuestructures of liver and kidney compared to normal controls, Sup-plementary Fig. 1.

New bone was laid down directly onto the implant surface andgrew along the surface without interposed soft tissue layer, imply-ing direct bonding of bone and implant. No scar tissue or fibrous

tissue was presented between bone and implant surface. Enhancedosseointegration could be observed as directly contacted areas ofbone and implant increased with prolonged implantation time,as shown in Fig. 9(a) and (c). Bone-implant interfaces were closelyrelated to the osseointegration and bonding strength. Typical fea-tures at bone-implant interfaces were presented in Fig. 9(b). Inarea A, implant was closely integrated with surrounding bones,showing well osseointegration. In area B, a transitional zoneexisted between the implant and new bone. Signs of new bonepenetration to the implant surface could be observed in this zone,marked with white arrows. In area C, localized corrosion happened.Some surface fragmentations of the implant materials could befound. A white gap without tissue filling presented betweenimplant and bone tissue, showing delayed osseointegration.

4. Discussion

4.1. Microstructure, mechanical property and corrosion behavior

According to the Mg-Ge binary phase diagram, there is no solu-bility of Ge in Mg and Mg2Ge phase must be precipitated, as shownin Fig. 10(a) [34]. Phase diagram also shows a high stability of thisMg2Ge intermediate phase (melting temperature 1117.4 �C),substantially higher than Mg17Al12 phase ((melting temperature455 �C)) in Mg-Al alloys [34]. In our Mg-xGe alloys, Ge contentwas below 3.36 wt% (eutectic point). So, they all belonged to thehypoeutectic group. Mg2Ge phase presented in the form ofa-Mg + Mg2Ge hypoeutectic, surrounding the dendrites of primarya-Mg phase. Area/volume ratio of those hypoeutectic phasesincreased with the increasing addition of Ge. Besides, net-like mor-phology was formed when Ge content exceeded 2.5 wt%, as shownin Fig. 1(a) and in Fig. 10(b). The improved strengths of Mg-Ge

Fig. 8. (a) 3D reconstruction of the implant; (b) quantitative analysis of implant volume; (c) bone-implant contact ratio.

430 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

alloys were mainly derived from the second phase strengthening ofMg2Ge. Effects of solid solution strengthening could be neglectedsince there is no solubility of Ge in Mg. Morphology of the eutecticstructures might also have impact on alloy strength as the net-likesecond phases could act as a skeleton in the transmission and dis-tribution of stress, possibly showing similar roles of concrete ironand stones in reinforced concrete structures in architecture.

The matrix a phase in magnesium alloys is normally anodic tothe second phases and is usually preferentially corroded. Eutecticphase in our Mg-Ge alloys played two opposite roles during corro-sion: (1) acted as cathode to the a-Mg matrix which acceleratedthe corrosion via galvanic corrosion; (2) acted as a barrier whichprovided protection to the Mg matrix and thus slowed down thecorrosion. Those two aspects interacted with each other and theywere closely related to the volume fraction and morphology ofthe eutectic phase [35]. Overall corrosion could be enhanced orsuppressed depending on which process was the dominated one.When the eutectic phase particles were in a low area/volume frac-tion, they could not form a net-like morphology and could not pro-vide protection of the matrix, as shown in the cross-sectionmicrostructure of corroded Mg-1.5Ge in Fig. 10(b). On the contrary,micro-galvanic corrosion between the a phase and eutectic phasewould promote the evolution of severely localized corrosion [47].However, for a high-volume fraction, the eutectic phase was nearlycontinuous like a net over the a matrix. Corrosion of the a phasewas then quite easily obstructed by the continuous eutectic phase,and so the corrosion was greatly retarded, as demonstrated in thecross-section microstructure of corroded Mg-3Ge in Fig. 10(b).

Since Ge has no solubility in Mg and Mg-Ge alloys owned the sameconstituent phases of a-Mg and Mg2Ge, differences among theircorrosion behaviors should mainly depend on the volume fractionand morphology of the Mg2Ge phase. The high-volume fractionand continuously net-like eutectic phase containing Mg2Ge shouldbe responsible for the high corrosion resistance of Mg-2.5Ge andMg-3Ge alloys. At the same time, low volume fraction and inter-mittent distribution of the eutectic phase (Fig. 10(b)) was believedto be closely related to the poor corrosion resistance of Mg-1.5Ge.

4.2. Toxicity of Ge and Mg-Ge alloys

The use of germanium compounds in dietary supplements topromote health or cure diseases has become increasingly popularin 1970s, especially in Japan. Until 1982, no systemic toxicity ofGe in man was reported [30]. Thereafter, numerous reports showedthat such supplements present a potential hazard to humans athigh dose [28]. However, till now, Ge-containing health careproducts, commonly available in organic forms and in capsulesor tablets, as well as in powdered form, are still available in healthfood stores and over the Internet. Cases of Ge induced toxicity inhuman were closely related to the high dosage (>100 times higherthan the estimated daily intake) and long-term supplement. Con-tamination of Ge in the inorganic forms, such as GeO2, was alsosuspected to be related to those clinical cases [28].

Ge compounds are relatively less toxic compared to other met-alloids and metals. However relatively high dose of germaniumdioxide and other inorganic Ge compounds can cause severe

Fig. 9. Typical histological staining of hard tissue sections. (a) H&E staining; (b) local magnifications of specific areas in (a); (c) toluidine blue staining.

Fig. 10. (a) Phase equilibrium diagram of binary Mg-Ge system; (b) typical microstructures (upper) of as-cast Mg-1.5Ge and Mg-3Ge alloys, corresponding eutectic phasemorphologies (middle) and corresponding cross-section microstructures (under) after immersion in Hank’s solution for 20 days.

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 431

poisoning including death. The median lethal dose (LD50) of GeO2

for mice ranges from 2020 mg/kg to 6300 mg/kg, with correspond-ing values for rats of 1620 mg/kg and 3700 mg/kg, respectively[30]. It has been reported that daily intake of Ge from food inhuman ranges from 0.4-3.4 mg, of which �96% is absorbed. At thisconcentration level, Ge would not cause toxicity [48]. Absorbed Geis preferentially excreted in the urine in both humans and animals[48,49]. Based on human case reports, the lowest observed adverseeffect levels of inorganic Ge, GeO2, organic Ge-132 ((GeCH2CH2

COOH)2O3, a novel organogermanium compound originally

synthesized in Japan in 1967) and Ge-lac-cit (germanium–lactate–citrate) are in the range of 0.7–23 mg/kg/d, equivalent to42–1380 mg/d in an adult weighing 60 kg [28], much higher thanthe daily intake.

Except for as-rolled Mg-3Ge, the other experimental Mg-Gealloys exhibited no toxicity to MG63 cells in this study. From theperspective of ion toxicity, the coexistence of 132–160 lg/mL Mgand 1.8–2.4 lg/mL Ge would not induce toxicity to MG63 cells.According to the results of Feyerabend et al. [50], 15 mM (360lg/mL) Mg2+ would not cause cytotoxicity to MG63 cells. So,

Table 2Some basic properties of typical essential/possibly essential elements in human body which can be used as alloying elements in magnesium.

Element Essential/tracein humana

Contentin human

Content in bone Bloodserumlevel

Recommendeddaily allowance

Standardelectrodepotential(V) b

Physiological roles Perspective of material science Reference

Mg Major,essential

21–28 g 60–65% in bones and teeth 0.75–1.2mM

320–400 mg �2.37 Activator of many enzymes; maintain muscleand nerve function; co-regulator of proteinsynthesis and muscle contraction; stabilizerof DNA and RNA

Poor mechanical properties of pureMg hinder its use in load bearingapplications; magnesium alloys withsufficient mechanical properties areneeded

[63–65]

Ca Major,essential

1000 g 99% in bones and teeth 2.1–2.6mM

1000–1300 mg �2.87 Second messenger in signal transductionpathways; calcium ions as a cofactor in manyenzymes; maintain proper bone formation

Improves strength of Mg (solidsolution strengthening, precipitationhardening, grain refinement) andcreep resistance; reduces castability

[66,67]

Zn Trace, essential 2–4 g 29% in bones 12–16mM

12–15 mg �0.76 Zn has ubiquitous roles; participation inmetabolism of RNA and DNA, signaltransduction, and gene expression, alsoregulation in apoptosis; appearance in allenzyme classes

Improves strength, yet reducesductility in high concentrations (solidsolution hardening, precipitation);improves castability

[68,69]

Mn Ultra-trace,essential

10–12 mg 43% in bones 10.9–78.3nM

1.8–5 mg �1.18 Necessary for development, metabolism, andthe antioxidant system; regulation of cellularenergy, bone and connective tissue growth,and blood clotting

Improves strength and ductility(grain refinement); improves creepresistance; improves corrosionresistance in combination withaluminium (precipitation that picksup iron)

[70,71]

Fe Trace, essential 3–5 g 20 mg is transferred daily tothe bone marrow forerythropoiesis and isefficiently recycled bymacrophages

6.3–26.9mM

0.9 mg �0.44 The most important transition metal in allliving organisms; iron-containing proteins arecrucial for the transport and storage of oxygen

Fe is the most common impurity inMg; it is detrimental to the corrosionresistance due to the formation of Fe-containing particles

[71–73]

Cu Ultra-trace,essential

84–126mg (1.4–2.1 mgper kg ofbodyweight)

0.53–5.69 ppm in bones 11–24mM

0.9–1.2 mg 0.34 Helps with the formation of collagen,increases the absorption of iron and plays arole in energy production; helps keep bloodvessels, bones, nerves, and the immunesystem healthy

Cu is a common impurity in Mg; asmall amount of Cu has a beneficialeffect on the creep strength ofmagnesium die castings but stronglyaccelerates corrosion

[74–76]

Sr Trace, possiblyessential

0.3 g 99.1% in bones 0.11–2.48mM

2mg �2.89 No harmful effects of stable strontium inhumans; substitution for calcium in a smallextent and deposition in bones; showing dosedependent metabolic effect on bone; lowdoses stimulated new bone formation

Improves strength and ductility(grain refinement); improves creepresistance and high temperaturestrength

[5,15,77,78]

Si Trace, possiblyessential

1–2 g No certain value. The highestlevels are found in bone andother connective tissuessuch as, skin, nail, hair,trachea, tendons and aorta

7.7–11.1mM

/ �0.86 Involves in bone mineralization andprevention of osteoporosis, collagensynthesis, and prevention of the aging of skin,overall condition of hair and nails

Reduces ductility; improves creepresistance and high temperaturestrength (precipitation); reducescorrosion resistance and castability

[71,79,80]

Sn Ultra-trace,possiblyessential

30 mg / 0.067–0.34 mM(in wholeblood)

/ �0.14 Tin-deficient diets in rat studies resulted inpoor growth, reduced feeding efficiency,hearing loss, and bilateral (male) hair loss

Improves strength, hardness andcreep strength, however too much Snimpairs strength and elongation; lowconcentration positively affects thecorrosion resistance

[5,21,81,82]

432D.Bian

etal./A

ctaBiom

aterialia64

(2017)421–

436

Table2(con

tinu

ed)

Elem

ent

Essential/trace

inhuman

aCon

tent

inhuman

Con

tentin

bone

Blood

seru

mleve

l

Recom

men

ded

dailyallowan

ceStan

dard

electrod

epo

tential

(V)b

Physiologicalroles

Perspe

ctiveof

materialscience

Referen

ce

LiUltra-trace,

possibly

essential

21–7

63pp

b10

0ng/gin

bone

<1mM

/�3

.04

Lithium

isusedin

thetrea

tmen

tof

bipo

lar

diso

rder

Red

ucesstrengthye

tim

prov

esdu

ctility/form

ability(chan

geto

bcc

latticestru

cture);

redu

cesco

rros

ion

resistan

cean

dde

nsity

[5,83,84

]

Ge

Ultra-trace,

dietarytrace

elem

ent

(pos

sibly

essential?)

//

0.07

mM(inplasma)

0.4–

3.4mg/d

(estim

ated

dietaryintake

ofGe)

0.24

Organ

icGehas

been

usedas

adietary

supp

lemen

twiththerap

euticbe

nefi

tsincludingan

ticancer,an

ti-inflam

mation,

antiox

idation,rad

ioprotection

andpro-

immunefunctions

Improv

esroom

tempe

rature

and

hightempe

rature

strength(grain

refiningan

dseco

ndph

ase

strengthen

ing)

[27,33

,85,86

]

aElem

ents

listed

below

as‘‘E

ssen

tial

inhuman

s”arethos

elisted

bythe(U

S)Fo

odan

dDru

gAdm

inistrationas

essential

nutrients.U

ltra-trace

mineralsarethos

eelem

ents

withestablished

,estim

ated

,orsu

spectedrequ

irem

ents

below

1mgpe

rda

y.bStan

dard

electrod

epo

tential

was

excerptedfrom

Electron

icScience

Tutor,G

RDelpierre

andBTSe

well,19

92–2

012.

FMscCfRfrt

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 433

higher Ge concentration (3.8 lg/mL) accompanied with higherMg concentration (148 lg/mL) in as-rolled Mg-3Ge extractshould be responsible for its low cell viability. Even though as-rolled Mg-3Ge induced cytotoxicity to MG63 cells in vitro, thein vivo tissue response and bone-implant osseointegration werequalified. New bone could directly lay down onto the implant sur-face and grew along the surface, implying decent biocompatibilityof this alloy. Owing to the capability of in vivo circulation system,local ion concentration surrounding the implant could be diluted.In order to minimize the huge gap between in vitro and in vivobiocompatibility tests, 6–10 times dilution of extracts forin vitro cytotoxicity test was recommended for biodegradablemagnesium-based materials [51]. In present study, local Mg andGe concentrations in vivo should be much lower than thatin vitro and that’s why as-rolled Mg-3Ge showed good biocom-patibility in bone while exhibited cytotoxicity in vitro. Duringthe 3 months’ implantation, the total released Ge was only 0.58mg, corresponding to 0.0064 mg per day under the assumptionof a constant degradation rate. Daily Ge dissolution from theimplant was far below its daily intake from normal diet and bio-safety of this alloy could be guaranteed.

4.3. Mg-Ge alloys compared to other Mg-X (X: essential/possiblyessential elements in human health) alloy systems

Table 2 summarizes the basic properties of essential/possiblyessential elements in human body. As is well-known, elementsCa and Mg are the major minerals in human body and the otherelements belong to the group of ‘‘trace elements” [52]. Ca seemsto be a proper alloying element in magnesium, especially fororthopaedic implants, since it is the most abundant metal inhuman body and it is crucial to the health of bones and teeth.But, poor formability and fast degradation might hinder the appli-cation of Mg-Ca alloys [7]. Strontium (Sr), zinc (Zn) and silicon (Si)commonly found in bone tissues are believed to benefit for bonegrowth and health [53]. Strontium can replace calcium in the

ig. 11. Comparison of the mechanical properties among various biodegradableg-X alloys (X: essential/possibly essential element in human health). Differentymbols represent different alloy systems and symbol color reflects the processondition. Mechanical properties were deduced from the following references: Mg-a alloys from Ref. [61,87,88], Mg-Zn alloys from Ref. [55,62,89,90], Mg-Sr alloysrom Ref. [15,16,18,91], Mg-Sn alloys from Ref. [20–23,82,92], Mg-Si alloys fromef. [92,93], Mg-Mn alloys from Ref. [92], Mg-Li alloys from Ref. [24], Mg-Ge alloysrom Ref. [33] and from present study (purple marked). (For interpretation of theeferences to colour in this figure legend, the reader is referred to the web version ofhis article.)

Fig. 12. In vitro and in vivo corrosion rates of binary Mg-Ge alloys explored in this study compared to previously reported Mg-X alloys (X: essential/possibly essential elementin human health): (a) corrosion rates in Hank’s solution calculated from electrochemical tests; (b) corrosion rates in Hank’s solution calculated from immersion tests; (c)corrosion rates in various bone tissues. In vitro corrosion rates: Mg-Ca alloys include as-cast Mg-1Ca [94]; Mg-Zn alloys include as-cast Mg-(1, 5, 7)Zn [55]; Mg-Sr alloysinclude as-rolled Mg-(1, 2, 3, 4)Sr [15], and as-extruded Mg-(0.5, 1, 1.5, 2.5)Sr [18]; Mg-Sn alloys include as-cast Mg-(1, 3, 5, 7)Sn [20], as-extruded Mg-(1, 3, 5, 7)Sn [22], andas-cast and as-rolled Mg-1Sn [92]; Mg-Si alloys include as-cast and as-rolled Mg-1Si [92]; Mg-Mn alloys include as-cast and as-rolled Mg-1Mn [92]; Mg-Li alloys include as-extruded Mg-3.5Li and Mg-8.5Li [24]; Mg-Cu alloys include as-cast Mg-(0.05, 0.1, 0.25)Cu [13], and Mg-(0.03, 0.19, 0.57)Cu [14]; corrosion data of Mg-Ge alloys were resultsin this study. In vivo corrosion rates of as-cast Mg-1Ca, as-extruded Mg-6Zn and as-rolled Mg-2Sr were derived from Ref. [61,62,15], respectively.

434 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

mineral of growing bones and thus lead to bone growth problems[54]. Thus, whether Mg-Sr alloys are applicable to children needsfurther research. Zn benefits for the strength and ductility simulta-neously. It is a proper alloying addition in both orthopaedic andcardiovascular applications [55,56]. Formability and corrosionresistance of Mg-Si alloys might be restricted by the brittle Mg2Siphase since Si has no solubility in Mg and Mg2Si must be precipi-tated [56,57]. Fe and Cu are normally recognized as impurities inmagnesium since they are inevitably introduced from the rawmaterials or during the production process [58]. Mg-Fe and Mg-Cu based alloys do not seem suitable for load bearing implants asthey possess a fast degradation rate. However, they could be appli-cable in special situations, such as Mg-Cu alloy in the treatment ofosteomyelitis [13]. Among all the elements in Table 2, lithium (Li)is the only one who can change the hexagonal close-packed (hcp)structure of Mg into body-centered cubic (bcc) type, thus improvesductility/formability. However, Li impairs strength. Mg-Li alloys tobe used as orthopaedic implants should be strengthened by otheralloying elements, such as Ca [59]. Generally, a common problemfaced by Mg-Ca, Mg-Sr, Mg-Si and even Mg-Zn alloys is that theyexhibited fast degradation (>0.5 mm/y) either in vitro or in vivo[15,60–62]. Information about Mg-Sn and Mg-Mn alloys is quite

limited till now. Possibility of those alloys to be used as orthopae-dic implants is still uncertain.

Compared to other essential/possibly essential elements, bio-logical information on Ge is quite limited. But, basically, it is a com-mon trace element found in human diet and well controlled intakewill not induce safety issues. Blood level of Ge is 3.99 lM (0.29 mg/L) in human serum, even higher than Li, Sn, Sr and Mn, as shown inTable 2. Dietary intake of Ge is similar to some other nutrient ele-ments. Standard electrode potential of Ge (Ge-Ge2+ + 2e, 0.24 V) islower than Cu and higher than the remaining elements in Table 2,for example, group neighbors Sn (Sn-Sn2+ + 2e, �0.14 V) andSi (Si + 2H2O-SiO2 + 4H++4e, �0.86 V).

An overall comparison of mechanical properties among variousMg-X alloys (X represents for essential/possibly essential elementsin human health) is displayed in Fig. 11. Under the same processingcondition, at as-cast or as-rolled state, mechanical properties ofMg-Ge binary alloys are comparable to other alloy systems.Strengths and elongations of as-cast Mg-Ge alloys in our studyare higher than most of the remaining as-cast binary alloys, asshown in Fig. 11. Among the limited binary alloys in as-rolled con-dition, our as-rolled Mg-3Ge exhibits the highest strength and sig-nificantly improved elongation. Generally, processing history has

D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436 435

remarkable impacts on the mechanical properties and corrosionbehaviors. Performances of Mg-Ge alloys can be adjusted throughvarious heat treatments and by deformation processing. In vitrocorrosion rates of typical Mg-X alloys measured by electrochemicaltest and immersion test are displayed in Fig. 12 for comparison.Corrosion resistance of Mg-Ge alloys is quite outstanding. Amongthe Mg-X alloys, as-cast Mg-2.5Ge, as-cast Mg-3Ge and as-rolledMg-3Ge exhibited the lowest corrosion rates (electrochemical), asshown in Fig. 12(a). Superior corrosion resistance of those alloyswas also verified in the immersion corrosion test, as shown inFig. 12(b). Besides, as-rolled Mg-3Ge exhibited much betterin vivo corrosion resistance compared to typical Mg-Ca, Mg-Znand Mg-Sr alloys of which orthopaedic implantation wasperformed, as shown in Fig. 12(c).

5. Conclusions

In this study, biodegradable Mg-based alloys with dietary traceelement germanium were developed and their feasibility to beused as orthopedic applications was systematically investigatedthrough in vitro and in vivo evaluations. Mechanical properties ofMg-Ge alloys were superior to many other biodegradable alloysstrengthened with essential elements in human body. Besides,they could be further improved by deformation processing, suchas rolling and extrusion. Mg-Ge (>2.5 wt% Ge) alloys exhibitedexcellent corrosion resistance since as-rolled Mg-3Ge exhibited aslow degradation rate in vitro (vcorr = 0.02 mm/y, pH = 0.13 mm/y).A proper degradation rate of 0.6 mm/y in vivo has also been veri-fied. Quite limited H2 generated during implant degradation didnot cause bone destruction and it could be completely absorbedin 3 months. Osseointegration of bone-implant was proved sincenew bone could directly lay down onto the implant and grew alongits surface. In conclusion, Mg-Ge alloys show great potential to beused as orthopaedic implants considering their good combinationsof mechanical property, corrosion resistance, cytocompatibilityand histocompatibility.

Acknowledgements

This work was supported by the National Key Research andDevelopment Program of China (Grant No. 2016YFC1102402),National Natural Science Foundation of China (Grant No.51431002), NSFC/RGC Joint Research Scheme (Grant No.51361165101 and 5161101031) and NSFC-RFBR CooperationProject (Grant No. 51611130054).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at https://doi.org/10.1016/j.actbio.2017.10.004.

References

[1] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Magnesium and its alloys asorthopedic biomaterials: a review, Biomaterials 27 (9) (2006) 1728–1734.

[2] D. Zhao, F. Witte, F. Lu, J. Wang, J. Li, L. Qin, Current status on clinicalapplications of magnesium-based orthopaedic implants: a review from clinicaltranslational perspective, Biomaterials 112 (2017) 287–302.

[3] H. Hermawan, D. Dubé, D. Mantovani, Developments in metallic biodegradablestents, Acta Biomater. 6 (5) (2010) 1693–1697.

[4] A.M. Sammel, D. Chen, N. Jepson, New generation coronary stent technology–isthe future biodegradable?, Heart Lung Circ 22 (7) (2013) 495–506.

[5] Y.F. Zheng, X.N. Gu, F. Witte, Biodegradable metals, Mater. Sci. Eng. R. 77(2014) 1–34.

[6] J.W. Seong, W.J. Kim, Development of biodegradable Mg–Ca alloy sheets withenhanced strength and corrosion properties through the refinement anduniform dispersion of the Mg2Ca phase by high-ratio differential speed rolling,Acta Biomater. 11 (2015) 531–542.

[7] Z. Li, X. Gu, S. Lou, Y. Zheng, The development of binary Mg-Ca alloys for use asbiodegradable materials within bone, Biomaterials 29 (10) (2008) 1329–1344.

[8] M.H. Idris, H. Jafari, S.E. Harandi, M. Mirshahi, S. Koleyni, Characteristics of as-cast and forged biodegradable Mg-Ca binary alloy immersed in Kokubosimulated body fluid, Adv. Mater. Res. 445 (2012) 301–306.

[9] S. Zhang, X. Zhang, C. Zhao, J. Li, S. Yang, C. Xie, H. Tao, Z. Yan, Y. He, J. Yao,Research on an Mg-Zn alloy as a degradable biomaterial, Acta Biomater. 6 (2)(2010) 626–640.

[10] J. Li, Y. Song, S. Zhang, C. Zhao, F. Zhang, X. Zhang, L. Cao, Q. Fan, T. Tang, In vitroresponses of human bone marrow stromal cells to a fluoridatedhydroxyapatite coated biodegradable Mg–Zn alloy, Biomaterials 31 (22)(2010) 5782–5788.

[11] E.M. Salleh, S. Ramakrishan, Z. Hussain, Characterization of binary Mg-Mnalloy synthesized through mechanical alloying: effects of milling speed, Adv.Mater. Res. 1087 (2015) 479–483.

[12] G. Xie, H. Takada, H. Kanetaka, Development of high performance MgFe alloyas potential biodegradable materials, Mater. Sci. Eng. A 671 (2016) 48–53.

[13] Y. Li, L. Liu, P. Wan, Z. Zhai, Z. Mao, Z. Ouyang, D. Yu, Q. Sun, L. Tan, L. Ren,Biodegradable Mg-Cu alloy implants with antibacterial activity for thetreatment of osteomyelitis: In vitro and in vivo evaluations, Biomaterials106 (2016) 250–263.

[14] L. Chen, X. Fu, H. Pan, W. Peng, W. Lei, L. Tan, K. Wang, Z. Ying, Y. Ke, P.K. Chu,Biodegradable Mg-Cu alloys with enhanced osteogenesis, angiogenesis, andlong-lasting antibacterial effects, Sci. Rep. 6 (2016) 27374, https://doi.org/10.1038/srep27374.

[15] X.N. Gu, X.H. Xie, N. Li, Y.F. Zheng, L. Qin, In vitro and in vivo studies on a Mg–Sr binary alloy system developed as a new kind of biodegradable metal, ActaBiomater. 8 (6) (2012) 2360–2374.

[16] M. Bornapour, M. Celikin, M. Pekguleryuz, Thermal exposure effects on thein vitro degradation and mechanical properties of Mg-Sr and Mg-Ca-Srbiodegradable implant alloys and the role of the microstructure, Mater. Sci.Eng. C 46 (2015) 16–24.

[17] H.S. Brar, J.Wong,M.V.Manuel, Investigationof themechanical anddegradationproperties of Mg–Sr and Mg–Zn–Sr alloys for use as potential biodegradableimplant materials, J. Mech. Behav. Biomed. Mater. 7 (2012) 87–95.

[18] C. Zhao, F. Pan, L. Zhang, H. Pan, K. Song, A. Tang, Microstructure, mechanicalproperties, bio-corrosion properties and cytotoxicity of as-extruded Mg-Sralloys, Mater. Sci. Eng. C 70 (Pt 2) (2017) 1081–1088.

[19] A. Gil-Santos, I. Marco, N. Moelans, N. Hort, O.V.D. Biest, Microstructure anddegradation performance of biodegradable Mg-Si-Sr implant alloys, Mater. Sci.Eng. C 71 (2017) 25–34.

[20] C. Zhao, F. Pan, S. Zhao, H. Pan, K. Song, A. Tang, Microstructure, corrosionbehavior and cytotoxicity of biodegradable Mg–Sn implant alloys prepared bysub-rapid solidification, Mater. Sci. Eng. C 54 (2015) 245–251.

[21] J. Kubásek, D. Vojtech, J. Lipov, T. Ruml, Structure, mechanical properties,corrosion behavior and cytotoxicity of biodegradable Mg–X (X=Sn, Ga, In)alloys, Mater. Sci. Eng. C 33 (4) (2013) 2421–2432.

[22] C. Zhao, F. Pan, S. Zhao, H. Pan, K. Song, A. Tang, Preparation andcharacterization of as-extruded Mg–Sn alloys for orthopedic applications,Mater. Des. 70 (2015) 60–67.

[23] J. Kubásek, D. Vojtech, D. Dvorsky, Structural and mechanical study on Mg–xLM (x = 0–5 wt.%, LM = Sn, Ga) alloys, Int. J. Mater. Res. 107 (2016) 459–471.

[24] W.R. Zhou, Y.F. Zheng, M.A. Leeflang, J. Zhou, Mechanical property,biocorrosion and in vitro biocompatibility evaluations of Mg–Li–(Al)–(RE)alloys for future cardiovascular stent application, Acta Biomater. 9 (10) (2013)8488–8498.

[25] A.F. Cipriano, A. Sallee, R.-G. Guan, Z.-Y. Zhao, M. Tayoba, J. Sanchez, H. Liu,Investigation of magnesium–zinc–calcium alloys and bone marrow derivedmesenchymal stem cell response in direct culture, Acta Biomater. 12 (2015)298–321.

[26] E. Zhang, L. Yang, J. Xu, H. Chen, Microstructure, mechanical properties andbio-corrosion properties of Mg-Si(-Ca, Zn) alloy for biomedical application,Acta Biomater. 6 (5) (2010) 1756–1762.

[27] L.G. Menchikov, M.A. Ignatenko, Biological activity of organogermaniumcompounds (A review), Pharm. Chem. J. 46 (11) (2013) 635–638.

[28] S.H. Tao, P.M. Bolger, Hazard assessment of germanium supplements, Regul.Toxicol. Pharm. 25 (3) (1997) 211–219.

[29] K. Yokoi, Germanium, Toxicity, Encyclopedia of Metalloproteins, SpringerScience+Business Media, New York, 2013.

[30] A.G. Schauss, Nephrotoxicity and neurotoxicity in humans fromorganogermanium compounds and germanium dioxide, Biol. Trace Elem.Res. 29 (3) (1991) 267–280.

[31] G.R. Belton, Y.K. Rao, Thermodynamic properties of Mg-Ge alloys, Metall.Mater. Trans. B 2 (8) (1971).

[32] F. Islam, A.K. Thykadavil, M. Medraj, A computational thermodynamic modelof the Mg–Al–Ge system, J. Alloys Compd. 425 (1–2) (2006) 129–139.

[33] H.B. Liu, G.H. Qi, Y.T. Ma, X.G. Zhang, Microstructure evolution and mechanicalproperties of Mg–Ge binary magnesium alloys, Mater. Res. Innov. 14 (2) (2013)154–159.

[34] J. Šerák, T. Kovalcík, D. Vojtech, P. Novák, The influence of Ge on the propertiesof Mg alloys, Key. Eng. Mater. 647 (Suppl. 2) (2015) 72–78.

[35] G. Song, A. Atrens, Understanding magnesium corrosion—a framework forimproved alloy performance, Adv. Eng. Mater. 5 (12) (2003) 837–858.

[36] ASTM-E8-04, Standard test methods for tension testing of metallic materials,annual book of ASTM standards, American Society for Testing and Materials,Philadelphia, PA, 2004.

436 D. Bian et al. / Acta Biomaterialia 64 (2017) 421–436

[37] H. Kuwahara, Y. Alabdullat, N. Mazaki, S. Tsutsumi, T. Aizawa, Precipitation ofmagnesium apatite on pure magnesium surface during immersing in Hank’ssolution, Mater. Trans. 42 (7) (2001) 1317–1321.

[38] ASTM-G102-89, Standard practice for calculation of corrosion rates andrelated information from electrochemical measurements, Annual book ofASTM standards Philadephia, American Society for Testing and Materials,Pennsylvania USA, 1999.

[39] ASTM G31-72, Standard practice for laboratory immersion corrosion testing ofmetals, ASTM International, Philadelphia West Conshohocken, PA, 2004.

[40] A.A. Ghoneim, A.M. Fekry, M.A. Ameer, Electrochemical behavior ofmagnesium alloys as biodegradable materials in Hank’s solution,Electrochim. Acta 55 (20) (2010) 6028–6035.

[41] ISO 10993-5:2009(E), Biological evaluation of medical devices—Part 5: Testsfor in vitro cytotoxicity, International Organization for Standardization, 2009.

[42] P. Xiu, Z. Jia, J. Lv, C. Yin, Y. Cheng, K. Zhang, C. Song, H. Leng, Y. Zheng, H. Cai,Tailored surface treatment of 3D printed porous Ti6Al4V by micro-arcoxidation for enhanced osseointegration via optimized bone in-growthpatterns and interlocked bone/implant interface, ACS Appl. Mater. Int. 8 (28)(2016) 17964–17975.

[43] P.K. Bowen, J. Drelich, J. Goldman, Magnesium in the murine artery: Probingthe products of corrosion, Acta Biomater. 10 (3) (2014) 1475–1483.

[44] A.F. Anvari-Yazdi, K. Tahermanesh, S.M.M. Hadavi, T. Talaei-Khozani, M.Razmkhah, S.M. Abed, M.S. Mohtasebi, Cytotoxicity assessment of adipose-derived mesenchymal stem cells on synthesized biodegradable Mg-Zn-Caalloys, Mater. Sci. Eng. C 69 (2016) 584–597.

[45] C.S. Gopinath, S.G. Hegde, A.V. Ramaswamy, S. Mahapatra, Photoemissionstudies of polymorphic CaCO3 materials, Mater. Res. Bull. 37 (7) (2002) 1323–1332.

[46] S. Ardizzone, C.L. Bianchi, M. Fadoni, B. Vercelli, Magnesium salts and oxide: anXPS overview, Appl. Surf. Sci. 119 (3) (1997) 253–259.

[47] J. Wang, Y. Jang, G. Wan, V. Giridharan, G.L. Song, Z. Xu, Y. Koo, P. Qi, J. Sankar,N. Huang, Flow-induced corrosion of absorbable magnesium alloy: in-situ andreal-time electrochemical study, Corros. Sci. 104 (2016) 277–289.

[48] T.J. Chen, C.H. Lin, Germanium: environmental pollution and health effects,Encyclopedia Environ. Health 16 (918) (2011) 927–933.

[49] K. Yokoi, T. Kawaai, A. Konomi, Y. Uchida, Dermal absorption of inorganicgermanium in rats, Regul. Toxicol. Pharmacol. 52 (2) (2008) 169–173.

[50] F. Feyerabend, J. Fischer, J. Holtz, F. Witte, R. Willumeit, H. Drücker, C. Vogt, N.Hort, Evaluation of short-term effects of rare earth and other elements used inmagnesium alloys on primary cells and cell lines, Acta Biomater. 6 (5) (2010)1834–1842.

[51] J. Wang, F. Witte, T. Xi, Y. Zheng, K. Yang, Y. Yang, D. Zhao, J. Meng, Y. Li, W. Li,K. Chan, L. Qin, Recommendation for modifying current cytotoxicity testingstandards for biodegradable magnesium-based materials, Acta Biomater. 21(2015) 237–249.

[52] C.D. Berdanier, D. Johanna, H. David, Handbook of Nutrition and Food, thirded., CRC Press, 2013.

[53] H.F. Li, X.H. Xie, Y.F. Zheng, Y. Cong, F.Y. Zhou, K.J. Qiu, X. Wang, S.H. Chen, L.Huang, L. Tian, L. Qin, Development of biodegradable Zn-1X binary alloys withnutrient alloying elements Mg, Ca and Sr, Sci. Rep. 5 (2015) 10719, https://doi.org/10.1038/srep10719.

[54] Public Health Statement for Strontium, Agency for Toxic Substances & DiseaseRegistry.

[55] S. Cai, T. Lei, N. Li, F. Feng, Effects of Zn on microstructure, mechanicalproperties and corrosion behavior of Mg–Zn alloys, Mater. Sci. Eng. C 32 (8)(2012) 2570–2577.

[56] Y. Chen, Z. Xu, C. Smith, J. Sankar, Recent advances on the development ofmagnesium alloys for biodegradable implants, Acta Biomater. 10 (11) (2014)4561–4573.

[57] M.E. Moussa, M.A. Waly, A.M. El-Sheikh, Effect of Ca addition on modificationof primary Mg2Si, hardness and wear behavior in Mg–Si hypereutectic alloys,J. Magn. Alloys 2 (3) (2014) 230–238.

[58] D. Persaud, A. McGoran, Biodegradable magnesium alloys: a review of materialdevelopment and applications, J. Biomim. Biomater. Tissue Eng. 12 (1) (2011)25–39.

[59] Y. Estrin, S.S. Nene, B.P. Kashyap, N. Prabhu, T. Al-Samman, New hot rolled Mg-4Li-1Ca alloy: a potential candidate for automotive and biodegradable implantapplications, Mater. Lett. 173 (2016) 252–256.

[60] M. Erinc, W. Sillekens, R. Mannens, R. Werkhoven, Applicability of existingmagnesium alloys as biomedical implant materials, Magnesium Technology,The Materials Society Annual Meeting, 2009, pp. 209–214.

[61] Z. Li, X. Gu, S. Lou, Y. Zheng, The development of binary Mg–Ca alloys for use asbiodegradable materials within bone, Biomaterials 29 (10) (2008) 1329–1344.

[62] S. Zhang, X. Zhang, C. Zhao, J. Li, Y. Song, C. Xie, H. Tao, Y. Zhang, Y. He, Y. Jiang,Research on an Mg–Zn alloy as a degradable biomaterial, Acta Biomater. 6 (2)(2010) 626–640.

[63] N.-E.L. Saris, E. Mervaala, H. Karppanen, J.A. Khawaja, A. Lewenstam,Magnesium: an update on physiological, clinical and analytical aspects, Clin.Chim. Acta 294 (1) (2000) 1–26.

[64] R.R. Watson, V.R. Preedy, S. Zibadi, Magnesium in Human Health and Disease,Humana Press, 2013.

[65] Y. Zheng, Magnesium Alloys as Degradable Biomaterials, CRC Press, 2015.[66] M. Brini, D. Ottolini, T. Calì, E. Carafoli, Calcium in health and disease, Met. Ions

Life Sci. 13 (4) (2013) 81–137.[67] P. Govind, Role of calcium in human and animal health, Front. Cell. Infect.

Microbiol. Jigyasa 3 (2014) 39–45.[68] G.P. Rink, L, Zinc and the immune system Zinc: immune system, Proc. Nutr.

Soc. 2000 (2000) 541–552.[69] H. Tapiero, K.D. Tew, Trace elements in human physiology and pathology: zinc

and metallothioneins, Biomed. Pharmacother. 57 (9) (2003) 399–411.[70] A.B. Santamaria, Manganese exposure, essentiality & toxicity, Indian J. Med.

Res. 128 (4) (2008) 484–500.[71] A. Sigel, H. Sigel, R.K.O. Sigel, Interrelations between Essential Metal Ions and

Human Diseases, Metal Ions in Life Sciences, Springer, Dordrecht, 2013.[72] R. Crichton, Iron Metabolism: From Molecular Mechanisms to Clinical

Consequences, third ed., John Wiley & Sons, 2009.[73] J.C. Dale, M.F. Burritt, A.R. Zinsmeister, Diurnal variation of serum iron, iron-

binding capacity, transferrin saturation, and ferritin levels, Am. J. Clin. Pathol.117 (5) (2002) 802–808.

[74] S.S. Rajeswari, Role of copper in health and diseases, Int. J. Curr. Sci. 10 (2014)E94–E107.

[75] P. Trumbo, A.A. Yates, S. Schlicker, M. Poos, Dietary reference intakes: vitaminA, vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese,molybdenum, nickel, silicon, vanadium, and zinc, J. Am. Diet. Assoc. 101 (3)(2001) 294–301.

[76] K. Jaouen, E. Herrscher, V. Balter, Copper and zinc isotope ratios in human boneand enamel, Am. J. Phys. Anthropol. 162 (3) (2017) 491–500.

[77] S. Pors Nielsen, The biological role of strontium, Bone 35 (3) (2004) 583–588.[78] W.E. Cabrera, I. Schrooten, M.E. De Broe, P.C. D’Haese, Strontium and bone, J.

Bone Miner. Res. 14 (5) (1999) 661–668.[79] E. Bissé, T. Epting, A. Beil, G. Lindinger, H. Lang, H. Wieland, Reference values

for serum silicon in adults, Anal. Biochem. 337 (1) (2005) 130–135.[80] R. Jugdaohsingh, Silicon and bone health, J. Nutr. Health Aging 11 (2) (2007)

99–110.[81] J. Chen, W.U. Shi-Hua, G.H. Guo, Y.M. Liu, Application of calcium nitrate in the

determination of stannum in blood by graphite furnace atomic absorptionspectrometry, Chin. J. Health Lab. Technol. 26 (15) (2016) 2150–2153.

[82] H. Liu, Y. Chen, Y. Tang, S. Wei, G. Niu, The microstructure, tensile properties,and creep behavior of as-cast Mg–(1–10)%Sn alloys, J. Alloys Compd. 440 (1–2)(2007) 122–126.

[83] A.M. Pérez-Granados, M.P. Vaquero, Silicon, aluminium, arsenic and lithium:essentiality and human health implications, J. Nutr. Health Aging 6 (2) (2002)154–162.

[84] B. Sampson, Determination of low concentrations of lithium in biologicalsamples using electrothermal atomic absorption spectrometry, J. Anal. At.Spectrom. 6 (6) (1991) 115–118.

[85] A. Shinohara, M. Chiba, Y. Inaba, Determination of germanium in humanspecimens: comparative study of atomic absorption spectrometry andmicrowave-induced plasma mass spectrometry, J. Anal. Toxicol. 23 (7)(1999) 625–631.

[86] B.J. Kaplan, W.W. Parish, G.M. Andrus, J.S. Simpson, C.J. Field, Germane factsabout germanium sesquioxide: I. Chemistry and anticancer properties, J.Altern. Complement. Med. 10 (2) (2004) 337–344.

[87] H. Du, Z. Wei, X. Liu, E. Zhang, Effects of Zn on the microstructure, mechanicalproperty and bio-corrosion property of Mg–3Ca alloys for biomedicalapplication, Mater. Chem. Phys. 125 (3) (2011) 568–575.

[88] H.R. Bakhsheshi-Rad, M.H. Idris, M.R. Abdul-Kadir, A. Ourdjini, M. Medraj, M.Daroonparvar, E. Hamzah, Mechanical and bio-corrosion properties ofquaternary Mg–Ca–Mn–Zn alloys compared with binary Mg–Ca alloys,Mater. Design 53 (2014) 283–292.

[89] J. Kubásek, D. Vojtech, Structural characteristics and corrosion behavior ofbiodegradable Mg–Zn, Mg–Zn–Gd alloys, J. Mater. Sci. Mater. Med. 24 (2013)1615–1626.

[90] Q. Peng, X. Li, N. Ma, R. Liu, H. Zhang, Effects of backward extrusion onmechanical and degradation properties of Mg–Zn biomaterial, J. Mech. Behav.Biomed. Mater. 10 (2012) 128–137.

[91] M. Bornapour, N. Muja, D. Shum-Tim, M. Cerruti, M. Pekguleryuz,Biocompatibility and biodegradability of Mg–Sr alloys: the formation of Sr-substituted hydroxyapatite, Acta Biomater. 9 (2) (2013) 5319–5330.

[92] X. Gu, Y. Zheng, Y. Cheng, S. Zhong, T. Xi, In vitro corrosion andbiocompatibility of binary magnesium alloys, Biomaterials 30 (4) (2009)484–498.

[93] E. Zhang, L. Yang, J. Xu, H. Chen, Microstructure, mechanical properties andbio-corrosion properties of Mg–Si(–Ca, Zn) alloy for biomedical application,Acta Biomater. 6 (5) (2010) 1756–1762.

[94] B. Zhang, Y. Hou, X. Wang, Y. Wang, L. Geng, Mechanical properties,degradation performance and cytotoxicity of Mg–Zn–Ca biomedical alloyswith different compositions, Mater. Sci. Eng. C 31 (8) (2011) 1667–1673.