In vitro corrosion degradation behaviour of Mg–Ca alloy in the presence of albumin

C.L. Liu a,b,*, Y.J. Wang c, R.C. Zeng a, X.M. Zhang a, W.J. Huang a, P.K. Chu b,**

a School of Materials Science and Engineering, Chongqing University of Technology, Chongqing 400050, Chinab Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, Chinac School of Applied Science and Technology, Chongqing Institute of Technology, Chongqing 400050, China

a r t i c l e i n f o

Article history:Received 25 March 2010Accepted 4 June 2010Available online 15 June 2010

Keywords:A. MagnesiumB. EISC. CorrosionC. Passive films

a b s t r a c t

The in vitro degradation behaviour of Mg–Ca alloy was studied in the presence of albumin by in-situobservation, hydrogen evolution method and various electrochemical techniques. The corrosion andhydrogen evolution rates decreased and the formation of filiform corrosion induced by Cl� was also sig-nificantly inhibited due to the adsorption of albumin molecule. Moreover, the higher the concentration ofalbumin, the higher is the inhibitive effect. The EIS results further show that addition of albuminenhanced the charge transfer resistance and film resistance at the open-circuit potential. The synergisticeffect of the negatively charged adsorbed albumin molecule and OH� has been discussed.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Magnesium alloys produced with a minimal amount of impuri-ties by new casting techniques have better properties than manycurrently approved and commonly used inert metallic biomateri-als. As an essential element in the human body, magnesium de-grades spontaneously in body fluids via corrosion to remove thestress shielding effect and avoid repeated surgeries [1–3]. In orderto apply magnesium alloys with moderate degradation character-istics to biomedical implants, it is important to evaluate theirdegradability and to understand the degradation mechanisms. Inthis respect, Mg–Al [4], Mg–RE [5], Mg–Ca [6], Mg–Mn–Zn [7],Mg–Zn [8] and pure Mg [9,10] have been investigated and in vivoand in vitro corrosion tests have been performed.

In vivo studies have been performed in small animals such asrats (subcutaneously), guinea pigs, and rabbits [11]. In vitro corro-sion and/or electrochemical tests have been predominantly carriedout in simulated body fluids such as 0.9 wt.% NaCl aqueous solu-tion, phosphate buffered solution (PBS, pH 7.4), Hank’s solution,simulated body fluids (SBF) [12–14]. Witte et al. have exploredthe in vivo degradation of AZ91D, and LAE42 rods in the femur ofguinea pigs. Their results show that the in vivo corrosion rates ofAZ91D and LAE442 after 18 weeks of implantation obtained areapproximately 3.516 � 10�4 mm/yr and 1.205 � 10�4 mm/yr

through the SRlCT analysis, respectively, which are about four or-ders of magnitude smaller than those achieved form in vitro tests insimulated fluids [3]. Similar results have also been obtained fromMg–Ca and Mg–Mn–Zn alloys [6,7]. It has been mentioned thatthe absence of organic ingredients such as proteins and aminoacids may be one reason for the variation in the corrosion rates be-tween in vitro and in vivo corrosion tests [5,6,15].

The influence of albumin or serum proteins on the corrosionrates of magnesium alloys has been studied under different exper-imental conditions. Yamamoto and Hiromoto [12] have investi-gated the corrosion rate of pure magnesium in NaCl, E-MEM andE-MEM with fetal bovine serum (FBS), and found that the averagedegradation rate in NaCl aqueous solutions after 8–14 days expo-sure was about 100 times larger than that in E-MEM + FBS, whichwas also found by Kirkland [15]. This result is very similar to thatin human blood plasma. The decreased corrosion rate is attributedto protein adsorption and precipitation of insoluble salts. Liu [16]has found that addition of 1 g/l albumin in the simulated body solu-tion shifts the open-circuit potentials of AZ91 towards a more posi-tive value. The corrosion resistance under the open-circuitconditions in SBF + 1 g/l albumin (Ab) is approximately twice thatin SBF. Rettig [17,18] have investigated the time-dependent electro-chemical behaviour of WE43 magnesium alloy in simulated bodyfluids (m-SBF) with and without 40 g/l albumin. They have ob-served that albumin may form a blocking layer on the surface dur-ing the initial hours of exposure, and the formed corrosion layersconsisting of amorphous apatite have only a low protective ability.Although the above results show the inhibitive effect of protein, theinterface reaction between magnesium alloy and protein has notbeen elucidated in detail at the early stage of immersion.

0010-938X/$ - see front matter � 2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.corsci.2010.06.003

* Correspondence to: C.L. Liu, Department of Physics and Materials Science, CityUniversity of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China.** Corresponding author. Tel.: +852 27887724; fax: +852 27889549.

E-mail addresses: liuchenglong@cqit.edu.cn (C.L. Liu), paul.chu@cityu.edu.hk(P.K. Chu).

Corrosion Science 52 (2010) 3341–3347

Contents lists available at ScienceDirect

Corrosion Science

journal homepage: www.elsevier .com/ locate /corsc i

In the present study, three kinds of solution (NaCl, NaCl + albu-min and distilled water + albumin) were adopted to measure thecorrosion behaviour of Mg–Ca alloy by continuously monitoringthe interface reaction at the early stage of immersion. Li [6] has ex-plored the feasibility of Mg–Ca alloys for use as biodegradablematerials. However, the reasons for the retarded corrosion rateare not clear. Hence, this work aims to investigate the influenceof albumin on the corrosion behaviour of binary Mg–Ca alloy dur-ing the early stage of immersion, especially in the first hour. Inaddition, the synergistic effects of negatively charged albumin,Cl�, and OH� ions during the exposure period are evaluated anda mechanism explaining the reduced corrosion rate is proposed.

2. Materials and methods

2.1. Materials and solutions

The extruded Mg–xCa alloy with nominal 1.5 wt.% calcium wasmachined into 10 mm � 10 mm � 4 mm disks for microstructurecharacterization, corrosion experiments, and electrochemical tests.The samples used in the corrosion and degradation tests were pre-pared by the standard technique of grinding with SiC abrasive pa-per up to 1200 grit, followed by ultrasonic cleaning in acetone,absolute ethanol, and distilled water sequentially.

In order to reduce the influence of other aggressive ions, wechose four different solutions which were distilled water(DW) + bovine serum albumin (Ab), 0.9 wt.% NaCl and 0.9 wt.% Na-Cl + Ab. The albumin-containing solution had albumin concentra-tions of 1 and 10 g/l.

2.2. Immersion tests

The volume of emitted hydrogen during immersion is related tothe dissolution of magnesium. This technique has the advantagethat variations in the degradation rates can be monitored by thehydrogen evolution rates, thereby allowing the study of the degra-dation rate variation versus exposure time. The samples weresoaked in 250 ml solutions for different periods of time at an ambi-ent temperature of 37 ± 1 �C. It has been verified that the dissolvedoxygen does not influence the measurement results and so ourexperiments were carried without aeration. The hydrogen evolu-tion volumes were measured as a function of immersion time [19].

The immersion test was carried out at 37 ± 1 �C. The sampleswere ground and cleaned prior to the tests. The corrosion morphol-ogy of each alloy immersed in the electrolytes in a special electro-lytic cell was observed and recorded at 20� magnification on aremote measurement system. The surface morphology of the cor-roded alloy sample was recorded and studied simultaneously. In-situ observation was found to be very important because as soonas the samples were taken out from the solution and exposed toair, the corrosion products turned white and the morphology sub-sequently changed. During the experiment, the sample was verti-cally immersed in the solution in the cell with the workingsurface of the sample parallel to the quartz window on the cell sidewall. The digital camera and microscopic head were perpendicularto the outside surface of the window [20]. The morphology was ob-served and recorded at time points of 3, 10, 30 and 60 min.

2.3. Electrochemical tests

The electrochemical corrosion behaviour was investigated byopen-circuit potential evolution and potentiodynamic polarizationusing an EG&G Instrument Verstat™. A three-electrode cell withthe sample as the working electrode, saturated calomel electrodeas the reference electrode, and platinum electrode as the counter

electrode was employed. All the pentiodynamic polarization testswere carried out using a scan rate of 0.1 mV/s. The experimentswere conducted at 37 ± 1 �C in triplicate without aeration withfresh solutions and newly polished electrodes. The electrochemicalimpedance spectra were measured at 37 ± 1 �C from a solution vol-ume of 250 ml. A frequency range from 10 mHz to 10 kHz was se-lected. Before the impedance measurement, the sample was keptfloating at the open-circuit potential (OCP) and for each impedancespectrum, the potential was set to the actual value of the OCP. Thedisturbing voltage amplitude chosen was 10 mV. The period of onemeasurement series was 5, 30 and 60 min. The EIS spectra were fit-ted with the ZSimpWin software.

2.4. Fourier transfer infrared (FTIR) spectroscopy

The structure and compositions of the sample surfaces weredetermined by Fourier transform infrared spectroscopy (FTIR) afterimmersion in 10 g/l Ab containing 0.9 wt.% NaCl for different dura-tions of time.

3. Results

Fig. 1 shows the metallographic microstructure of the extrudedMg1.5Ca alloy. The typical a(Mg) primary grains with the Mg2Caprecipitating phase [6] can be observed and the Mg2Ca phase existsalong the grain boundaries or in the a(Mg) primary grains. Fig. 2indicates the open-circuit potential (OCP) measured from Mg1.5Caalloy in the solutions. The variations of the OCP values were similarduring 1 h immersion in the different solutions. The OCP value wasvery low at the beginning of all experiments. It afterwards shiftedrapidly to the positive direction after approximate 500 s of expo-sure, finally becoming nearly constant indicating establishmentof the equilibrium. However, the final OCP values measured fromall the solutions were different. In the NaCl aqueous solution, theOCP rose fast to a value of about �1.38 V(SCE) and it was morenegative than that in NaCl + 1 g/l Ab or DW + 10 g/l Ab. Comparingthe electrolytes NaCl solution with NaCl + 10 g/l Ab solution, theOCP value in the latter was significantly higher than that in theNaCl solution, which was about �1.34 V(SCE) after 1 h immersion.Another interesting phenomenon can be observed from the OCPcurves for 0.9 wt.% NaCl and DW + 10 g/l Ab solutions. The increasein the OCP values in both electrolytes was similar during immer-sion, but it was followed by a slight increase in DW + 10 g/l Ab.In addition, the similar variation and slight increase in the OCP va-lue in NaCl solution including 1 g/l Ab suggest that the lower con-centration of albumin in NaCl may play a limited role in inhibitingthe corrosion attack of Mg–Ca alloy.

Fig. 1. Optical micrograph of the extruded Mg1.5Ca alloy (500�).

3342 C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347

In order to characterize the early corrosion stages of Mg–Ca al-loy in all used electrolytes, the corrosion morphology and hydro-gen evolution process were monitored in-situ by an opticalmicroscope with a digital camera. Fig. 3 provides selected illustra-tions from the in-situ observations of the details of the corrosionmorphologies for Mg1.5Ca in all used electrolytes. Comparing theelectrolytes distilled water and DW + 10 g/l Ab, two main differ-ences can be observed from the corrosion morphologies. One isthat addition of albumin resulted in an increase of the amountand density of hydrogen bubbles compared with distilled water

at the same exposure time. The other is that corrosion became vis-ible on the surface in the form of pitting corrosion in both solu-tions. Also visible were a number of vertical streams composedof a larger number of small hydrogen bubbles and subsequentlythe formation of corrosion pits. It is concluded that addition ofalbumin in distilled water enhances the corrosion of Mg–Ca alloy.In every case illustrated in Fig. 3c, the diameter of hydrogen bub-bles are larger, meaning that Mg1.5Ca sample suffers aggressivecorrosion attack in 0.9 wt.% NaCl. It appears that corrosion pitsagglomerated into an irregular array from which there was a dra-matic hydrogen evolution. In addition, filiform corrosion also ap-peared. After 1 h immersion, filiform corrosion which developedwith hydrogen evolution was concentrated exclusively at thegrowing tips. At the filiform corrosion area, the evolution of hydro-gen may be inhibited by the corrosion products [20]. Comparingthe electrolytes 0.9 wt.% NaCl and 0.9 wt.% NaCl + 10 g/l Ab, fili-form corrosion may be inhibited by the adsorption of albumin. Alsoobserved was the decrease in the amount and diameter of hydro-gen bubbles in the same exposure time. Moreover, the size of thehydrogen bubbles increased very slowly in the same exposuretime. Furthermore, when albumin was added to the NaCl solution,hydrogen evolution took place uniformly over the sample surfaces.Otherwise, hydrogen evolution was concentrated at the corrodingsites. The results further suggest that albumin affects the corrosionprocess of Mg–Ca alloy significantly.

The potentiodynamic polarization curves of Mg1.5Ca obtainedafter open-circuit exposure to different test electrolytes for about10 min are exhibited in Fig. 4. Generally, the cathodic polarizationcurves are assumed to represent the cathodic hydrogen evolutionthrough water reduction, while the anodic polarization curves rep-resent the dissolution of magnesium [6,21]. It can be seen that the

Fig. 2. Variations of open-circuit potential (OCP in 0.9 wt.% NaCl, 0.9 wt.%NaCl + 1 g/l Ab, 0.9 wt.% NaCl + 10 g/l Ab, and DW + 10 g/l Ab) versus immersiontime at 37 ± 1 �C.

Fig. 3. Typical in-situ corrosion morphology for Mg1.5Ca in different electrolytes after different duration of time (3 min, 10 min, 30 min, and 60 min): (a) Distilled water; (b)Distilled water + 10 g/l Ab; (c) 0.9 wt.% NaCl; (d) 0.9 wt.% NaCl + 10 g/l Ab.

C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347 3343

various solutions give rise to different cathodic polarization cur-rents. The cathodic polarization currents in 0.9 wt.% NaCl solution,associated with the hydrogen evolution reaction, was approxi-mately two orders of magnitude higher than that in 0.9 wt.%NaCl + 10 g/l Ab solution. In comparison with the solution includ-ing 10 g/l Ab, addition of 1 g/l albumin may have little effect onthe hydrogen evolution reaction. Comparing the electrolyteDW + 10 g/l Ab and 0.9 wt.% NaCl, the cathodic polarization currentin the former was much lower than that in the latter, indicatingthat the cathodic reaction is kinetically easier in 0.9 wt.% NaCl thanin DW + 10 g/l Ab. In addition, addition of albumin shifted the ano-dic polarization curves to more positive potentials.

Although the results of in-situ observation drew a conclusionthat the existence of albumin inhibits the corrosion process ofMg1.5Ca alloy, the difference is qualitative. Hence, to investigatethe effect of albumin on the degradability of the Mg–Ca alloy fur-ther, the hydrogen evolution of Mg–Ca alloy was measured (seeFig. 5) from the same total area of 0.9 cm2. The hydrogen evolutioncurves show that the Mg1.5Ca alloy suffered different corrosion at-tack in the decreasing order of 0.9 wt.% NaCl > 0.9 wt.%NaCl + 1 g/lAb > 0.9 wt.% NaCl + 10 g/l Ab > DW + 10 g/l Ab. Among the fourtest electrolytes, NaCl or albumin enhanced corrosion of theMg1.5Ca alloy. But when NaCl and albumin coexisted, corrosion

of this alloy might be inhibited. The average rate of hydrogen evo-lution from the Mg1.5Ca is about 15 ml/cm2/day in 0.9 wt.% NaCland in contrast, that in 0.9 wt.% NaCl + 10 g/l Ab is approximately4.5 ml/cm2/day. Based on the open-circuit potential, in-situ obser-vation, and hydrogen evolution tests, it is concluded that albuminis a corrosion inhibitor for Mg–Ca alloy in NaCl + albumin solution,and with increasing the concentration of albumin, the inhibitive ef-fect becomes higher.

Fig. 6 shows the EIS of Mg1.5Ca alloy immersed in differentelectrolytes for different time. The plots show three time constantsin 0.9 wt.% NaCl with and without albumin, including one high fre-quency capacitance loop, one medium frequency capacitance loop,and one low frequency inductance loop. The diameter of the highfrequency capacitance loop increases in the albumin solution.The influence on the degradation process of Mg1.5Ca in the solu-tion with and without albumin is analyzed based on the EIS results.Three time constants correspond to the characteristics of the elec-tric double layer, surface film, and existence of the metastableMg2+, respectively [21–23]. The plots can be explained by theequivalent circuit shown in Fig. 7. Herein, Rs is the solution resis-tance, Rct refers to the charge transfer resistance, Cdl representsthe electric double layer capacity at the interface of the Mg–Ca al-loy substrate and electrolyte, and Rf is the film (consisting of thecorrosion product and albumin adsorption layer) resistance. Thefilm capacity is described by a constant phase element (CPEf).The CPEf is defined by two values, Y0 and n. If n is equal to 1, CPEf

is identical to a capacitor. Often a constant phase element is usedin a model in place of a capacitor to compensate for the nonhomo-geneity in the system [22,23]. RL and L indicate the existence ofmetastable Mg2+ during dissolution of the Mg–Ca alloy. The dataare fitted using the ZSimpWin 3.20 software and the errors are lessthan 10%. The fitted results are shown in Table 1. Comparing theEIS in 0.9% NaCl with and without albumin, the existence of albu-min caused a gradual decrease in the diameter of the medium fre-quency capacitance loop and finally it was replaced by aninductance loop. The EIS implies that the film layer formed onthe Mg1.5Ca alloy surface were different in 0.9% NaCl with andwithout albumin, thereby suggesting different influence on thecorrosion process of Mg1.5Ca alloy.

4. Discussion

Mg–Ca alloy has been shown to be suitable candidate as ortho-pedic biodegradable materials [6]. However, the in vitro and in vivocorrosion rates show great differences, and Gu et al. [3,6] have sug-gested that protein may play a significant role in retarding the cor-rosion. Our results here showed that addition of albumin into aNaCl solution retarded the corrosion of Mg–Ca alloy, which is sim-ilar to Kirkland’s results [15].

The OCP results in Fig. 2 showed similar variations in the differ-ent electrolytes, where the initial increase was followed by a rela-tively steady state. In the presence of 10 g/l albumin in thesolution, the OCP value increased gradually with the prolongedimmersion, which is different from that in 0.9 wt.% NaCl. The dif-ference can probably be ascribed to the surface being accessibleto the formation of corrosion product and adsorption of albumin.Gu et al. have pointed out that when the fresh Mg–Ca alloy surfacewas exposed to the aqueous solution, the internal second phase(Mg2Ca) acted as an efficient cathode for hydrogen evolution, andanodic dissolution of magnesium occurs. In this process, a magne-sium hydroxide film formed on the surface of the Mg–Ca alloy dueto significant alkalization near the surface [6]. However, in thepresence of Cl� ions, the Cl� ions are small enough to displacewater molecules in the hydrogen sheath and then Cl� preferen-tially combines with Mg2+ to transform magnesium hydroxide into

Fig. 4. Potentiodynamic polarization curves obtained from Mg1.5Ca in differentelectrolytes after 10 min immersion.

Fig. 5. Hydrogen evolution volumes of Mg1.5Ca alloy as a function of immersiontime in different electrolytes.

3344 C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347

soluble MgCl2. Therefore, the variations of the OCP values shouldbe attributed to the dissolution–precipitation process of Mg(OH)2

film layer in NaCl solution [6,24,25]. In the presence of albumin,at the initial stage of immersion, the OCP values showed little dif-ference before about 600 s immersion, which was probably as-cribed to the quick formation of magnesium hydroxide. However,with longer exposure period, the OCP values showed obvious dif-ference in different electrolytes. Addition of 10 g/l albumin intothe NaCl solution resulted in a positive shift in the potential. Thissuggests that albumin might interact with the surface impeding

further dissolution–precipitation process of Mg1.5Ca controlledby Cl�. Similarly, the in-situ observation results at 10, 30, and60 min in Fig. 3c also indicate that aggressive corrosion took placein 0.9 wt.% NaCl. In comparison, the Mg1.5Ca alloy just suffersslight corrosion attack and no filiform corrosion in 0.9 wt.%NaCl + 10 g/l Ab, as shown in Fig. 3d. The potentiodynamic curvesin Fig. 4 after 10 min open-circuit immersion also show that thecorrosion potential in 0.9 wt.% NaCl + 10 g/l Ab is around�1.35 V(SCE), which is about 140 mV(SCE) higher than that in0.9 wt.% NaCl. The above results indicate the synergistic effect ofNaCl and bovine serum albumin could play a key role in the corro-sion process of Mg1.5Ca alloy.

It has been reported that the bovine serum albumin moleculemainly consists of amino acid joined by peptide bonds. The basicstructure of an amino acid consists of four parts: an amino group(ANH2), a carboxyl group (ACOOH), a central carbon (C) and a sidechain. At physiological pH of 7.4, albumin undergoes a neutral–acidic transition and becomes negatively charged (isoelectric pH4.7–4.9). Divalent cations such as Ca2+ or Mg2+ can serve as bridg-ing agents to enhance adsorption of albumin molecules throughelectrostatic interaction [16,26]. In order to examine the adsorp-tion of Ab on the Mg1.5Ca alloy surface, FTIR spectra were takenbefore and after immersion in 0.9 wt.% NaCl + 10 g/l Ab for differ-ent time durations and the results are displayed in Fig. 8. Thestrong amide I (1650 cm�1), amide II (1545 cm�1) and a weak

Fig. 6. EIS of Mg1.5Ca immersed in different electrolytes for different time: (a) 0.9 wt.% NaCl and (b) 0.9 wt.% NaCl + 10 g/l Ab.

Fig. 7. Equivalent circuit of Mg1.5Ca immersed in 0.9 wt.% NaCl with and without10 g/l Ab for different period.

Table 1Fitted results from Mg1.5Ca immersed in different electrolytes for different time.

Electrolytes Time (min) Rs (X cm2) Cdl (lF cm�2) Rct (X cm2) Y0 (X�1 cm�2 s�n) n Rf (X cm2) L (mH cm�2) RL (X cm2)

0.9 wt.% NaCl 5 27.6 131.5 246.1 7.9 � 10�4 0.68 847.2 19.2 136330 27.8 142.7 253.5 6.3 � 10�4 0.66 700.6 43.7 143560 27.2 161.4 280.1 4.7 � 10�4 0.66 554.1 78.3 1813

0.9 wt.% NaCl + 10 g/l Ab 5 27.4 168.2 296.3 4.6 � 10�4 0.69 784 23.8 102930 27.7 181.5 319.5 3.9 � 10�4 0.69 931 35.4 108360 27.6 184.7 330.2 3.6 � 10�4 0.69 1057 60.6 1656

C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347 3345

amide III 1248 cm�1) signals of Ab can be observed from thesespectra [27,28].

The corrosion process of Mg–Ca alloy in the aqueous solutionincludes the following chemical reactions.

MgðsÞ !Mg2þðaqÞ þ 2e�ðanodic reactionÞ ð1Þ

2H2Oþ 2e� ! H2 þ 2OH� ðaqÞðcathodic reactionÞ ð2Þ

Mg2þðaqÞ þ 2OH�ðaqÞ !MgðOHÞ2ðproduct formationÞ ð3Þ

The Mg2+ cations may bridge the negatively charged albuminmolecule and Mg–Ca alloy surface. During initial immersion, thecollective effect of H2O and Cl� creates a large amount of Mg2+

[6], which may result in rapid adsorption of albumin. Accordingto hydrodynamic studies, the highly helical molecule of albuminis described as three domains within the shell of a prolate ellipsoid(6–12 nm2) or three spheres (diameters of 3.8, 5.3, and 3.8 nm)positioned adjacent to each other. The different shapes arise fromthe flexibility of the helical molecule allowing rapid expansionand contraction as ligands bind with ions. On the surface, the trian-gular albumin molecule with equilateral 8 nm sides and a 2.9 nmaverage thickness transforms to a flattened triangular shape with13 nm sides and a 1.2 nm average thickness. The albumin withits marked size can easily inhibit the transfer of ions from the solu-tion to the metal surface [28,29]. Therefore, the albumin adsorp-tion layer may inhibit the dissolution of Mg(OH)2 under theeffect of Cl�, leading to the gradual positive shift in the OCP valuein 0.9 wt.% NaCl + 10 g/l Ab.

In order to further explain the in-situ observation results, hydro-gen evolution measurement was performed. It is known that thedegradation rate of magnesium can be measured by the evolvedhydrogen volume [19]. For all the electrolytes, the hydrogen evolu-tion volume increased monotonously as a function of immersiontime, but if the immersion time approaches infinity, the hydrogenevolution volume becomes a constant value. Fig. 9 shows the vari-ations in the hydrogen evolution volume per square centimeterevery hour during 12 h immersion. Comparing the solutions0.9 wt.% NaCl without and with 10 g/l Ab, the hydrogen evolutionrate in 0.9 wt.% NaCl was higher than that in the latter, which indi-cates some inhibitive factors during the corrosion dissolution ofMg1.5Ca. Based on the chemical reaction (3), the formation ofMg(OH)2 film is beneficial to the retardation of corrosion. However,the aggressive Cl� results in further dissolution of the Mg(OH)2

film [24]. Therefore, the competitive effect between Cl� and OH�

could lead to the dissolution–precipitation process on the Mg1.5Casurface. When albumin and NaCl coexist, negatively charged albu-min molecule and OH� can form a mixed protective layer com-posed of an albumin adsorption layer and Mg(OH)2. On the onehand, the albumin layer can easily inhibit the transfer of Cl� fromthe solution to the sample surface [24]. On the other hand, the neg-atively charged albumin molecule and hydroxyl may inhibit thepenetration of Cl� from the solution through electrostatic repulsiveeffects. Therefore, the protective effect from the combination of thealbumin adsorption layer and Mg(OH)2 film may restrain theaggressive effect from Cl�.

If the albumin molecules absorb onto the sample surface, theinterface between the sample and test electrolytes could be chan-ged, which must result in the variation of some electrochemicalcorrosion parameters [17,18]. Based on the fitted results in Table 1from the EIS test, the competitive effect between Cl� and the pro-tective layer composed of the albumin adsorption layer andMg(OH)2 can be investigated better. In 0.9 wt.% NaCl, the Rct valueshows smaller increase with immersion. The change was from246.1 to 280.1 X cm2, indicating that the transfer of charge at theinterface of the Mg1.5Ca/solution was retarded slightly. It is inter-esting to note that higher Rct values were measured from 0.9 wt.%NaCl + 10 g/l Ab than 0.9 wt.% NaCl. It may result from the nega-tively charged albumin molecule, and the adsorption of albumincould increase the integrity of the film layer covering the Mg1.5Casurface, changing the interfacial behaviour of charge transfer inhigh frequency range and hinders the diffusion of ions [10,23].It’s interesting to note that the difference of hydrogen evolutionrate was bigger than that of Rct values in both solutions. The reasonmay be that the Rct values referred to the charge transfer resistanceat the interface of sample and test solution at a special potentialduring immersion time [23], however, the volume of hydrogenevolution is related to the average dissolution rate of magnesiumalloy during immersion [19], which is not synchronous with the re-sult of EIS. The increase of Rct value indicates the hard charge trans-fer process. In addition, the Rf values in 0.9 wt.% NaCl showedgreater changes than those in 0.9 wt.% NaCl + 10 g/l Ab after thesame immersion period. In 0.9 wt.% NaCl solution, there was grad-ual decrease in the Rf values and it could be ascribed to the gradualcorrosion attack of the original MgO and Mg(OH)2 by Cl�. The in-situ observation results indicate that the corrosion attack on theMg1.5Ca alloy becomes greater with increasing immersion time.In contrast, the Rf values of the Mg1.5Ca alloy increased from784.3 to 1057.4 X cm2 and the associated CPEf values decreasedonly slightly with immersion in 0.9 wt.% NaCl + 10 g/l Ab. This

Fig. 8. FTIR spectra of Mg1.5Ca alloy after exposure to 0.9 wt.% NaCl containing10 g/l Ab for different times: 0 min, 10 min, 30 min, and 60 min.

Fig. 9. Variations of hydrogen evolution volume every hour during 12 immersion inthe electrolytes.

3346 C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347

A

may be evidence of the improvement in the integrity for the albu-min adsorption and Mg(OH)2 mixed layer covering the Mg1.5Casurface during the exposure period [23,27], and a competition be-tween albumin adsorption molecule, OH�, and Cl� for sites on thesurface could inhibit the in-leakage of Cl� and its corrosion attackon Mg1.5Ca substrate under the layer.

These results demonstrate that albumin can act as an inhibitorin the presence of aggressive Cl� ion. Its adsorption on the Mg1.5Casurface impedes further dissolution of this alloy and increases theresistance by the cooperative effect of the adsorbed albumin layerand Mg(OH)2. However, until now we have no verification of thereaction process of a protein layer on the sample surface and in fu-ture studies, the reaction process must be analyzed in more details.

5. Conclusion

The corrosion degradation behaviour of Mg–Ca alloy was inves-tigated by electrochemical techniques, hydrogen evolution tests,and in-situ observation. Three test electrolytes simulating body flu-ids: NaCl, DW + Ab and NaCl + Ab were used in this study. The re-sults indicate that the solution chemistry modified the corrosionbehaviour on Mg1.5Ca at the early stage of immersion. Accordingto the in-situ observation and hydrogen evolution results, Mg1.5Caalloy suffered greater corrosion attack in 0.9 wt.% NaCl than in0.9 wt.% NaCl + Ab. Albumin acted as a corrosion inhibitor onMg1.5Ca and gave rise to diminished hydrogen evolution ratesand cathodic corrosion current densities. Moreover, the higherthe concentration of albumin, the greater is the inhibitive effect.The EIS plots show that addition of albumin enhanced the chargetransfer resistance and the resistance of the surface film layer atthe open-circuit potential. The synergistic effect of the negativelycharged adsorbed albumin molecule and OH� appeared to retardfurther corrosive effect from the aggressive Cl� and to decreasethe corrosion rate of Mg1.5Ca in solutions including Cl� to someextent.

Acknowledgements

This work was financially supported by Natural Science Foun-dation Project of Chongqing CSTC (2008BB4062; 2008BB0063;KJ080615) and Hong Kong Research Grants Council (RGC) GeneralResearch Fund (GRF) No. CityU 112307. The authors would like tothank Prof. Chen RS for his assistance in MgCa alloy.

References

[1] R.C. Zeng, W. Dietzel, F. Witte, N. Hort, C. Blawert, Progress and challenge formagnesium alloys as biomaterials, Advanced Biomaterials 10 (2008) B3–B4.

[2] A.C. Hanzi, P. Cunde, M. Schinhammer, On the biodegradation performance ofan Mg-Y-RE alloy with various surface conditions in simulated body fluid, ActaBiomater. 5 (2009) 162–171.

[3] F. Witte, J. Fischer, J. Nellesen, In vitro and in vivo corrosion measurements ofmagnesium alloys, Biomaterials 27 (2006) 1013–1018.

[4] W. Zhou, T. Shen, N.N. Aung, Effect of heat treatment on corrosion behavior ofmagnesium alloy AZ91D in simulated body fluid, Corr. Sci. 52 (2010) 1035–1041.

[5] F. Witte, V. Kaese, H. Haferkamp, E. Switzer, In vivo corrosion of fourmagnesium alloys and the associated bone response, Biomaterials 26 (2005)3557–3563.

[6] Z.J. Li, X.N. Gu, S.Q. Lou, Y.F. Zheng, The development of binary Mg–Ca alloys foruse as biodegradable materials within bone, Biomaterials 29 (2008) 1329–1344.

[7] L.P. Xu, G.N. Yu, E.L. Zhang, In vivo corrosion behaviour of Mg–Mn–Zn alloy forbone implant application, J. Biomed. Mater. Res. 83A (2007) 703–711.

[8] S.X. Zhang, X.N. Zhang, C.L. Zhao, Research Mg–Zn alloy as degradablebiomaterial, Acta Biomaterialia (2009), doi:10.1016/j.actbio.2009.06.028.

[9] M.P. Staiger, A.M. Pietak, J. Huadmai, Magnesium and its alloys as orthopedicbiomaterials: a review, Biomaterials 27 (2006) 1728–1734.

[10] L.C. Zhao, C.X. Cui, Q.Z. Wang, S.J. Bu, Growth characteristics and corrosionresistance of micro-arc oxidation coating on pure magnesium for biomedicalapplications. Corr. Sci. (2010) doi:10.1016/j.corsci.2010.03.008.

[11] F. Witte, N. Hort, C. Vogt, Degradable biomaterials based on magnesiumcorrosion, Curr. Opinion Solid State Mater. Sci. 12 (2008) 63–72.

[12] A. Yamamoto, S. Hiromoto, Effect of inorganic salts, amino acids and proteinson the degradation of pure magnesium in vitro, Mater. Sci. Eng. C 29 (2009)1559–1568.

[13] Y.C. Xin, C.L. Liu, X.M. Zhang, Corrosion behaviour of biomedical AZ91magnesium alloy in simulated body fluids, J. Mater. Res. 22 (2007) 2004–2012.

[14] G.L. Song, Control of biodegradable of biocompatible magnesium alloys, Corr.Sci. 49 (2007) 1696–1701.

[15] N.T. Kirkland, J. Lespagnol, N. Birbilis, M.P. Staiger, A survey of bio-corrosionrates of magnesium alloys, Corr. Sci. 52 (2010) 287–291.

[16] C.L. Liu, Y.C. Xin, X.B. Tian, P.L. Chu, Degradation susceptibility of surgicalmagnesium alloy in artificial albumin-containing biological fluid, J. Mater. Res.22 (2007) 1806–1814.

[17] R. Rettig, S. Virtanen, Time-dependent electrochemical characterization of thecorrosion of a magnesium rare-earth alloy in simulated body fluids, J. Biomed.Mater. Res. 85 (2008) 167–175.

[18] R. Rettig, S. Virtanen, Composition of corrosion layers on a magnesium rare-earth alloy in simulated body fluids, J. Biomed. Mater. Res. J. Biomed. Mater.Res. 88 (2009) 359–369.

[19] Y.C. Xin, K.F. Huo, T. Hu, P.K. Chu, Influence of aggressive ions on thedegradation behaviour of biomedical magnesium alloy in physiologicalenvironment, Acta Biomater. 4 (2008) 2008–2015.

[20] G.L. Song, A. Atrens, X.L. Wu, B. Zhang, Corrosion behaviour of AZ21, AZ501,and AZ91 in sodium chloride, Corr. Sci. 40 (1998) 1769–1791.

[21] W.C. Jian, W.G. Xing, H.F. Peng, M.P. Li, J.D. Wen, Effect of heat treatment oncorrosion and electrochemical behaviour of Mge3Nde0.2Zne0.4Zr (wt%) alloy,Electrochim. Acta 52 (2007) 3160–3167.

[22] G.L. Song, D. Atrens, X. Wu, J. Nairn, The anodic dissolution of magnesium inchloride and sulphate solutions, Corr. Sci. 39 (1997) 1981–2004.

[23] Y.W. Song, D.Y. Shan, R.S. Chen, F. Zhang, E.H. Han, Biodegradable behaviour ofAZ31 magnesium alloy in simulated body fluid, Mater. Sci. Eng. C 29 (2009)1039–1045.

[24] N. Hara, Y. Kobayahsi, D. Kagaya, N. Akao, Formation and breakdown of surfacefilms on magnesium and its alloys in aqueous solutions, Corr. Sci. 39 (1997)188–194.

[25] J.H. Nordlien, S. Ono, N. Masuko, K. Nisancioclu, A TEM investigation ofnaturally formed oxide films on pure magnesium, Corr. Sci. 39 (1997) 1397–1414.

[26] A. Klinger, D. Steinberg, D. Kohavi, M.N. Sela, Mechanism of adsorption ofhuman albumin to titanium in vitro, J. Biomed. Mater. Res. 36 (1997) 387–392.

[27] Y.H. Shen, Z.L. Yang, J.G. Wu, FT-IR study on the precipitates of bovine serumalbumin reacted with calcium hydroxyapatite, Acta Scientiarum NaturalimUniversitaties Pekinensis 35 (1999) 431–436.

[28] C. Vogt, K. Bechstein, S. Gruhl, M. Lange, F. Witte, in: K.U. Kainer, D.G.M.Weimar (Eds.), Proceeding of the 8th International Conference on MagnesiumAlloys and their Applications, Wiley-VCH, Germany, 2009, pp. 1162–1174.

[29] N. Padilla, A. Bronson, Electrochemical characterization of albumin protein onTi6Al4V alloy immersed in simulated plasma solution, J. Biomed. Mater. Res.81 (2007) 531–543.

C.L. Liu et al. / Corrosion Science 52 (2010) 3341–3347 3347