You are on page 1of 4

Available online at www.sciencedirect.

com

Materials Letters 62 (2008) 2181 – 2184


www.elsevier.com/locate/matlet

Corrosion process of pure magnesium in simulated body fluid


Yong Wang a,b , Mei Wei b,⁎, Jiacheng Gao a , Jinzhu Hu c , Yan Zhang a
a
College of Materials Science & Engineering, Chongqing University, Chongqing, 400045, China
b
The Institute of Materials Science, University of Connecticut, Storrs, CT 06269, USA
c
College of Materials Science and Engineering, Southwest University, Chongqing, 400700, China
Received 30 September 2007; accepted 20 November 2007
Available online 28 November 2007

Abstract

The chemical and physical processes of magnesium in simulated body fluid (SBF) were investigated. The corrosion rate of magnesium was
measured after 3, 5, 7, 14 and 21 days of immersion, respectively. It was found that the corrosion rate decreased with increasing immersion time,
while the pH of SBF changed inversely. Network-like cracks and pits were the main damages resulting from corrosion, and the localized buildup
of chloride ions was the major cause of pit formation.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Magnesium; SBF; Corrosion; Mg(OH)2 layer; Hydroxyapatite

1. Introduction ing phases may deposit on magnesium surfaces after a certain


period of immersion. In addition, the biocarbonate ions in the
Magnesium and its alloys can be used as biodegradable SBF may also favor the delay of magnesium corrosion due to
biomaterials [1–3], but the high corrosion rate of the material the formation of low soluble carbonate products [7]. Much
is a major concern [4]. It is essential to understand the research has been conducted to evaluate the effect of surface
corrosion process of magnesium under physiological condi- modification on the corrosion behavior of Mg and its alloys
tions in order to develop better magnesium-based biomater- in SBF [8,9], but fundamental studies on the corrosion
ials. Witte et al. [5] reported that amorphous calcium processes of uncoated magnesium are lacking. Therefore, the
phosphates can be formed on the surface of the corrosion corrosion behavior of pure magnesium in SBF was studied in
layer of magnesium alloys, and the calcium phosphate-coated the current report to develop a better understanding of
magnesium forms direct contact with the surrounding tissue corrosion mechanisms of pure magnesium in physiological
in vivo. The same research group also indicated that the conditions.
substitute ocean water is not suitable for the prediction of in
vivo corrosion rate of magnesium-based alloys [6]. To mimic 2. Experimental part
the body environment, simulated body fluid (SBF) is
employed in the current study to assess the corrosion Magnesium with a purity of 99.9% was used. The as-
behavior of magnesium. Similar to sodium chloride solution, annealed ingot was cut into rectangular samples of 10 ×
SBF also induces severe corrosion of magnesium at body 28 × 5 mm3. The samples were ground using SiC papers
temperature and pH. However, since SBF is supersaturated ranging from 320 to 800 grits, and ultrasonically rinsed in
with calcium and phosphate, calcium- or phosphate-contain- acetone for 10 min. Their weight and surface area were
recorded. The composition of SBF is listed in Table 1, which is
prepared based on Ref. [10]. The samples were soaked in SBF
⁎ Corresponding author. Tel.: +1 860 486 9253; fax: +1 860 486 4745. at 37 °C for 3, 5, 7, 14, and 21 days. Each sample was soaked
E-mail address: m.wei@ims.uconn.edu (M. Wei). in 500 mL SBF separately. At each time point, a group of
0167-577X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.matlet.2007.11.045
2182 Y. Wang et al. / Materials Letters 62 (2008) 2181–2184

Table 1
Ion concentrations of SBF
Ion K+ Na+ Ca2+ Mg2+ Cl− HCO−3 SO2−
4 HPO2−
4

Concentration (mM) 6.0 109.5 7.5 1.5 110.0 17.5 0 3.0

samples were removed from SBF, rinsed with de-ionized


water, and subjected to different characterizations.
The corrosion rate (CR, mm/yr) of the samples was
calculated via the following equation [11]:

CR ¼ 8:76T104 W =ATD ð1Þ

where W is the mass loss (g), A is the original surface area of


each sample (cm2), T is the immersion time (h) and D is the
sample density (g/cm 3 ). The corrosion topography was
Fig. 2. Cracks and pits on the corroded metal surface, where p is for pits and c is
observed using a JEOL JSM 6335F field emission scanning for cracks.
electron microscope (FESEM). The phase composition of
sample surface was identified using a Bruker D5005 X-ray
diffractometer (XRD) and/or an Amary 1000A energy dis- increasing immersion time. The dissolution of magnesium consumes
persive X-ray spectroscopy (EDX). For cross-section observa- H+ but releases OH−, leading to the pH increase in SBF.
tions, the samples were mounted in epoxy (BUEHLER, USA), The corrosion topography of the samples is shown in Fig. 2.
ground and polished, etched with a solution of 2% nitric acid in Network-like cracks and small pits were the main damages resulting
ethanol, and examined using an optical microscope (OM). from corrosion. As the immersion time increased, the cracks and pits
became wider and larger. The corrosion of magnesium is localized
3. Results and discussion corrosion [12], leading to the formation of cracks and pits. The
localized corrosion of magnesium is different from auto-catalyzing
Fig. 1 shows the variations of the CR and the SBF pH with pitting [12], where the dissolution occurs on the entire surface rather
immersion time. The CR was approximately 4.4 mm/yr at the 3rd day, than concentrates on the pits at an increasing rate.
and it dropped dramatically to about 3.0 mm/yr at the 5th day. After At the 5th day, a porous deposit layer was formed at some regions
that, CR continued to decrease, but the decrease was moderate. An on the sample surface (Fig. 3). This deposit increased with immersion
even slower drop was observed from the 7th day to the 21st day. The time, and it covered the entire sample surface at the 21st day. XRD
SBF pH increased with the immersion time. Corresponding to the examination (Fig. 4) shows that high-intensity Mg(OH)2 peaks were
abrupt drop of CR between the 3rd and the 5th day, the pH increase detected after 21 days of soaking, suggesting a substantial amount of
during this period was rather slow comparing to any other time Mg(OH)2 is formed on sample surfaces. Meanwhile, a MgO film,
intervals studied. which was not detected by XRD, could be formed through the
Magnesium is a highly reactive metal. It corrodes in water via the decomposition of Mg(OH)2 [13,14]. MgO/Mg(OH)2 acts as a
following reaction [12]: protective film for pure Mg, so the corrosion can only take place at

Mg þ Hþ þ H2 O→Mg2þ þ OH− þ H2 ð2Þ


It dissolves in SBF while hydrogen gas bubbles rise from its
surface, which was noted in the current study as soon as the samples
were immersed in SBF. The rate of bubble release decreases with

Fig. 1. CR of magnesium and pH of SBF at different time points. Fig. 3. Porous deposit on the sample surface after 5 days of immersion in SBF.
Y. Wang et al. / Materials Letters 62 (2008) 2181–2184 2183

Table 2
Sample surface and cluster composition at different time points (wt.%)
O Mg P Ca Cl
Surface (7 days) 50.06 7.12 14.20 27.01 1.63
Surface (21 days) 64.58 30.66 0.68 0.62 3.47
Cluster (3 days) 62.09 19.03 1.46 1.33 16.10

layer acts as a barrier to prevent magnesium from further absorbing Ca


and P from SBF. Hence, it becomes almost impossible for calcium
phosphate to deposit on pure magnesium surfaces. Nevertheless, the
precipitation of crystalline hydroxyapatite was found in the SBF
solution (Fig. 4). These results conclusively suggests that it is important
to depress/eliminate Mg(OH)2 formation on sample surfaces so as to
have calcium phosphate deposited.
After 3 days of soaking in SBF, needle-shaped clusters were
formed on the sample surfaces (Fig. 5). The number of clusters
increased from the 3rd day to the 7th day, but disappeared after
14 days of immersion. The surface composition of the samples at
days 7 and 21 as well as the cluster composition at day 3 are
presented in Table 2. EDX results revealed that the needle-shaped
clusters were rich in Cl, suggesting that they may consist of MgCl2.
Fig. 4. XRD spectra of sample surface composition at different time points and
Chloride ions in SBF can break down the Mg(OH)2 film and
the composition of precipitation in SBF. 1 = Mg(OH)2; 2 = HA; 3 = Mg.
accelerate the corrosion of magnesium by forming MgCl2. MgCl2 is
more soluble in water than Mg(OH)2, so it was hard to collect
sufficient amount of MgCl2 for XRD examination. Due to the
the film-free regions [13,14]. Due to the formation of Mg(OH)2 film,
conversion of Mg(OH)2 to MgCl2 and the dissolution of the later,
OH− was consumed. As a result, CR was found abruptly dropped at the
the amount of Mg(OH)2 was too low to be detected by XRD at the
5th day, but the pH increase in SBF slowed down between the 3rd and
7th day. Thus, only magnesium was identified by XRD. At the
the 5th day.
early stage of immersion, the formation of MgCl2 was faster than
The precipitation of hydroxyapatite in SBF is a spontaneous process
its dissolution. The formation of the Cl-rich needle-shaped clusters
after a certain length of immersion [10]. The dissolution of magnesium
increased with time in the first 7 days. With the further dissolution
results in pH increase, which favors hydroxyapatite precipitation [15].
of Mg, both the thermodynamic driving force and the rate of MgCl2
Ca and P components were found on the sample surface from the 3rd
formation decreased. Therefore, not enough MgCl2 was built up
day to the 7th day. However, it has been reported that magnesium ions
locally to form clusters after 14 days of immersion. However,
retard or inhibit crystallization of hydroxyapatite and other calcium
chloride was detected on sample surface at all time points,
phosphates under different conditions [16–21]. Thus, the deposition of
suggesting that MgCl2 always coexists with Mg(OH)2 in the
Ca and P did not lead to the nucleation and growth of hydroxyapatite
corrosion product.
on magnesium surfaces.
Cracks and pits were observed from the cross-section of the
After soaked for 21 days, the Ca and P were hardly detected on the
samples. Deposits were found in the pits but disappeared after
sample surface that was completely covered by a Mg(OH)2 layer. This
etching. Table 3 shows the composition of the deposits in the pits at
the 7th and 21st day. Cl was detected at the 7th day but not at the
21st day, suggesting that chloride-rich deposit is apt to re-dissolve
as the immersion time prolongs. Thus, the localized buildup of
chloride ions to form more soluble MgCl2 can be a major cause of
pitting formation on pure Mg. It was also discovered in our study
that neither the cracks nor the pits originate from or proliferate
along the grain boundaries (figure not shown). Magnesium has an
extremely low electropotential that all the impurities in it would act
as cathodes while electrochemical corrosion takes place. Impurities
tend to segregate at grain boundaries, leading to the increase in
electropotential of these areas and the corrosion of the adjacent
magnesium.

Table 3
Composition of deposit in pits at different time points (wt.%)
Time, days O Mg P Ca Cl
7 61.74 16.98 10.37 7.21 3.72
21 72.39 27.61 0 0 0
Fig. 5. Needle-like clusters on the corroded surface.
2184 Y. Wang et al. / Materials Letters 62 (2008) 2181–2184

4. Conclusions [3] C.D. Mario, H. Griffiths, O. Goktekin, N. Peeters, J. Verbist, M. Bosiers,


K. Deloose, B. Heublein, R. Rohde, V. Kasese, C. Ilsley, R. Erbel, J. Interv.
Cardiol. 17 (2004) 391.
In summary, pure magnesium heterogeneously corrodes in [4] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006)
SBF. The corrosion rate of the samples decreased with 1728.
increasing immersion time. Network-like cracks and pits were [5] F. Witte, V. Kaese, H. Haferkamp, E. Switzer, A.M. Lindenberg, C.J.
formed during the sample dissolution, but not originated from Wirth, H. Windhagen, Biomaterials 26 (2005) 3557.
[6] F. Witte, J. Fischer, J. Nellesen, H.A. Crostack, V. Kaese, A. Pisch, F.
grain boundaries. Porous Mg(OH)2 deposit formed at the early
Beckmann, H. Windhagen, Biomaterials 27 (2006) 1013.
stage of corrosion process acted as a barrier for the deposition of [7] Al-Abdullat, S. Tsutsumi, N. Nakajima, M. Ohta, H. Kuwahara, K.
Ca and P, but the presence of Cl− ions contributed to the Ikeuchi, Mater. Trans. 42 (2001) 1777.
breakdown of the Mg(OH)2 deposit and thereby prompted pit [8] L.C. Li, J.C. Gao, Y. Wang, Surf. Coat. Technol. 185 (2004) 92.
formation. [9] H.Y. Lopez, D.A. Cortes, S. Escobedo, D. Mantovani, Key Eng. Mater.
309-311 (2006) 453.
[10] T. Kokubo, H. Takadama, Biomaterials 27 (2006) 2907.
Acknowledgements [11] ASTM-G31-72: Standard Practice for Laboratory Immersion Corrosion
Testing of Metals.
MW would like to acknowledge National Science Founda- [12] G. Song, A. Atrens, Adv. Eng. Mater. 5 (2004) 837.
tion (DMI 0500269) and the other authors would like to [13] G. Baril, G. Galicia, C. Deslouis, N. Pebere, B. Tribollet, V. Vivier, J.
acknowledge Chinese National Natural Science Foundation Electrochem. Soc. 154 (2007) C108.
[14] G. Song, A. Atrens, D. Stjohn, J. Nairn, Y. Lit, Corros. Sci. 39 (1997) 855.
(No. 30670562) for their support of the research. [15] G. Vereecke, J. Cryst. Growth 104 (1990) 820.
[16] K.S. TenHuisen, P.W. Brown, J. Biomed. Mater. Res. 36 (1997) 306.
[17] A.L. Boskey, A.S. Posner, Mater. Res. Bull. 9 (1974) 907.
References [18] F. Abbona, A. Baronnet, J. Cryst. Growth 165 (1996) 98.
[19] F. Barrere, P. Layrolle, C.A. van Blitterswijk, K. de Groot, Bone 25 (1999)
[1] P. Zartner, R. Cesnjevar, H. Singer, M. Weyand, Catheter. Cardiovasc. 107S.
Interv. 66 (2005) 590. [20] M.H. Salimi, J.C. Heughebaert, G.H. Nancollas, Langmuir 1 (1985) 119.
[2] D. Schranz, P. Zartner, I. Michel-Behnke, H. Akinturk, Catheter. [21] W. Kibalczyc, J. Christoffersen, M.R. Christoffersen, A. Zielenkiewicz, W.
Cardiovasc. Interv. 67 (2006) 671. Zielenkiewicz, J. Cryst. Growth 106 (1990) 355.

You might also like