On corrosion behaviour of magnesium alloy

AZ31 in simulated body fluids and influence

of ionic liquid pretreatments
Y. F. Zhang1, B. Hinton1, G. Wallace2, X. Liu2 and M. Forsyth*1
This paper reports on the corrosion of Mg alloy AZ31 in simulated body fluid (SBF) using static
immersion tests and electrochemical impedance spectroscopy. A preliminary study on the effect
of flowing SBF on the corrosion behaviour of AZ31 has also been carried out. Low toxicity ionic
liquids (ILs) trimethyl(butyl)phosphonium diphenyl phosphate P1444DPP and trihexyl(tetradecyl)-
phosphonium bis-2,4,4trimethylpentyl-phosphinate [P66614][i(C8)2PO2] have been used to provide
corrosion protection for AZ31 in SBF. Time dependent immersion tests indicate that under static
conditions, AZ31 suffers severe localised corrosion in SBF, with pits developing predominantly
beside the Al–Mn intermetallic phase in the a matrix. At longer immersion times, the corrosion
product eventually precipitates and covers the entire specimen surface. When exposed to SBF
under flowing conditions with a shear stress of 0?88 Pa, more uniform corrosion was observed.
The optical profilometry results and electrochemical impedance spectroscopy analysis suggest
that both P1444DPP and [P66614][i(C8)2PO2] pretreatments can increase the corrosion resistance of
AZ31 in SBF, in particular by decreasing the number of deeper pits found on the alloy surface.
Cytotoxic test shows that the presence of the ILs P1444DPP and [P66614][i(C8)2PO2] in cell culture
media slightly inhibits the growth of human coronary artery endothelial cells in comparison with
the good cell viability around the treated specimen. A pretreatment with IL is used in order to
improve the corrosion resistance of this alloy in SBF.
Keywords: Mg alloy AZ31, Coronary artery stents, Simulated body fluid, Electrochemical impedance spectroscopy, Flowing cell, Ionic liquids

This article is part of a special issue on the durability of biomaterials and biocorrosion.

Introduction involved in many metabolic reactions and physiological

mechanisms, excess Mg2z can be excreted in urine,1,5
Implanting artery stents into diseased arteries is considered and so the major degradation product is readily dealt
an effective treatment for coronary artery diseases.1 For with. The excellent mechanical properties of Mg alloys
some time, permanent metallic stents have been widely used, mean that they are strong enough to maintain the
but it is often claimed that the continued presence of the opening required in a coronary artery. Most alloying
stent after vessel remodelling is unnecessary.1 The stent acts additives reduce the rate of corrosion of magnesium
as a foreign body engendering long term risks and that the alloys; however, for biomedical applications, the
stent inside the coronary artery as a foreign object presents amount of these alloying additives is strictly limited,
long term risks of chronic inflammation2 in stent restenosis2 and some elements (e.g. Zr, Cd, rare earths and heavy
and late thrombosis.3 Thus, the development of biodegrad- metals) are restricted.6 In this study, Mg alloy AZ31
able coronary artery stents is very important. A biodegrad- with ,3%Al content was chosen as a candidate material.
able stent can be gradually dissolved and adsorbed after an The presence of Al not only improves the mechanical
artery is remodelled, avoiding medical complication.4 properties but also enhances the corrosion resistance.7
The use of magnesium alloys to construct biodegrad- Previous research has shown that this alloy is non-
able coronary artery stents has attracted attention allergenic in an epicutaneous patch test.8 Unfortunately,
recently.1 Mg alloys are the fourth most abundant because the oxide–hydroxide films that form on the
cation naturally present in the human body, they are surface of magnesium alloys (including AZ31) on
exposure to air provided poor protection to the under-
1
lying metal,9 these alloys corroded very rapidly in the
ARC Australia, Centre of Excellence for Electromaterials Science, IFM,
Deakin University, Burwood, Vic., Australia
body, resulting in subcutaneous hydrogen gas bubbles,10
2
ARC Australian Centre of Excellence for Electromaterials Science, IPRI, an increase in local pH10 and potential loss of the
University of Wollongong, Wollongong, NSW, Australia structural integrity of the stent before the arteries can be
*Corresponding author, email maria.forsyth@deakin.edu.au fully remoulded. Consequently, while the potential of

ß 2012 Institute of Materials, Minerals and Mining

Published by Maney on behalf of the Institute
Received 24 March 2012; accepted 19 April 2012
374 Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5 DOI 10.1179/1743278212Y.0000000032
 Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

using a magnesium alloy (e.g. AZ31) as the primary analysis of electrochemical impedance spectroscopy
material for biodegradable coronary artery stents is (EIS) data. The latter assists in understanding the
obvious, their use has thus far been limited.1 Further changes in the corrosion processes with this alloy in
work is required to understand the corrosion behaviour SBF as a function of time. Considering that the stents
of these alloys and the relationship between the placed into the blood vessels are in a dynamic
information obtained from in vivo corrosion tests and environment, and that flow rate may significantly affect
that from more laboratory testing through the use of the corrosion of an alloy, we also report on the influence
dynamic corrosion test protocols. of a flowing environment on the corrosion behaviour of
Controlling the corrosion rate of magnesium alloys in AZ31. A flowing environment test cell was designed to
the body is critical in the development of biodegrad- mimic, as near as possible, the physiological conditions
able magnesium alloys stents. Much research has been to which a stent is subjected in the coronary artery.
carried out to improve the corrosion resistance of This paper thus discusses the corrosion processes, in
magnesium alloys, mainly through the use surface particular the initiation of corrosion on AZ31 in SBF.
coatings.10 Typical conversion coatings (such as chro- The effect of flow on corrosion behaviour and the ability
mate conversion coating and phosphate permanganate of IL pretreatment to modify degradation have been
conversion coating) and anodising are two of the investigated.
effective processes employed.11 However, these surfaces
are unlikely to be biocompatible. Thus, the development Experimental
of an improved method to control the degradation rate
of Mg alloys in the body without any adverse effects is Alloy and specimens
desirable. One approach recently found to be promising The magnesium alloy used in this investigation was a
for controlling Mg corrosion in chloride contaminated direct strip cast cylinder of AZ31 magnesium alloy, with
aqueous media is the use of pretreatment of the alloy a nominal composition of 3%Al, 1%Zn, etc. In order for
surface with ionic liquids (ILs). An IL is a low all specimens to have similar microstructure, they were
temperature molten organic salt (melting point below taken around the circumferential edge of the cylindrical
100uC).12 The ILs have many unique properties, such as casting.
a high concentration of species reactive with a metal For the constant immersion corrosion (static) tests
surface, high stability and low volatility.13 In recent and the EIS tests, specimens were mounted in epoxy
years, the corrosion protection properties of ILs have resin to expose a test area of 1066 mm. For the flow cell
generated great interest through their potential to form corrosion (dynamic) tests, cylindrical specimens with a
protective surface films on reactive metals. Passivation diameter of 10 mm and a length of 25 mm were used.
of a reactive metal by an IL was identified by Howlett The circular section of the specimens was exposed to the
et al.14 in studies of lithium batteries using bis(trifluor- environment with the flow test.
omethylsulfonyl) (Tf2N) based IL as the electrolyte. It All the specimens were ground with SiC paper to a
was also noted by Uerdingen et al.15 that many alloys 4000 mesh surface finish under running tap water. They
showed very good corrosion resistance after contact were rinsed with distilled water and acetone and then
with ILs. Thus, if an IL is chosen that can react with an dried under a nitrogen stream. The specimens were
Mg alloy surface to provide some corrosion protection stored in a vacuum desiccator for 24 h to stabilise the
and also produce a surface, which is non-toxic to human oxide film that forms immediately after abrasion upon
cells, then IL pretreatment could also be used to control exposure to the air environment. The interface between
the biodegradation rate of Mg alloys in the human body. the specimens and the epoxy resin was sealed using nail
To explore the possibility of employing a particular polish lacquer to avoid potential crevice corrosion.
magnesium alloy for stents, an understanding of the
biocorrosion behaviour of the alloy in the human body Simulated body fluid
is essential through in vivo testing.16 However, in vivo A commercially available SBF was chosen as the test
testing has many drawbacks, such as time, cost and the solution due to its similar ionic composition to that of
potential discomfort and harm to the experimental human blood plasma. Table 1 shows the composition of
subjects,16 all of which limit its efficiency. Therefore, the SBF used in this work and a typical blood plasma.
recent research with AZ316,17,18 has been undertaken The pH of the SBF solution was 7?4¡0?5.
using simulated body fluid (SBF), which contains a
similar inorganic composition to human blood plasma. Ionic liquids
Despite this prior work, the surface corrosion processes Trimethyl(butyl)phosphonium diphenyl phosphate
(in particular the early stages) and the corrosion (P1444DPP) was prepared in the laboratory.
mechanism for AZ31 in SBF are still not well under- Trihexyl(tetradecyl)phosphonium bis-2,4,4trimethylpen-
stood. In contrast to previous published work, here, we tyl-phosphinate {[P66614][(iC8)2PO2]} was obtained from
focus predominantly on the early stages of corrosion Cytec Industries. The ILs were purified by passing
during immersion tests and a more in depth and targeted through a column containing a filter agent, alumina and

Table 1 Inorganic ion composition of blood plasma and SBF

Inorganic ions/mM

Solution Ca2z Cl2 Naz Cr2 Mg2z Kz HCO{

3 SO2{
4 HPO2{
4

Blood plasma19 2.5 103 142 103 1.5 5 27 0.5 1

SBF 2.6 103.6 147 … 1.5 5 6 0.5 1

Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5 375

Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

2 Schematic diagram of flow cell

1 Schematic structures of a P1444 cation and DPP anion20
and b P66614 cation and (iC8)2PO2 anion21 Electrochemical impedance spectroscopy
Electrochemical impedance spectroscopy tests were
sand to remove any impurities. The schematic structures carried out using a three-electrode cell. The volume of
of the anion and cation that make up the molecule the SBF was 250 mL open to the air. An Ag/AgCl
P1444DPP and [P66614][(iC8)2PO2] are shown in Fig. 1. electrode was used as a reference electrode and a titanium
Ionic liquid pretreatment mesh as the auxiliary electrode. An EG&G PAR VMP2/Z
multichannel potentiostat was used to process the
Previous research21 has found that a specimen heated to
current–time and voltage–time data to provide the
50uC in room temperature IL for a period of time was an
impedance at different frequencies.22 The measurements
advanced IL treatment method. With this treatment, the
were taken over a frequency range of 200 kHz to 50 mHz
AZ31 was heated to 50uC using an oven. A thin layer of
with eight points per decade using a 10 mV amplitude
room temperature IL was placed over the whole sur-
perturbation. Electrochemical impedance spectroscopy
face of the specimen using a syringe. The approximate
spectra were acquired at 1 h intervals. The Zfit program
volume of IL was 1?5 mL, and the approximate thick-
in the EC-Lab software was used to fit the equivalent
ness of the IL film was 2 mm.
circuits to the EIS experimental data.
Constant immersion corrosion tests
Cytotoxic test of ILs
Plastic beakers containing 100 mL of the SBF were used
The cytotoxicity of the ILs was determined using pri-
for these tests. The SBF was maintained at 37uC in a
mary human coronary artery endothelial cell (HCAEC)
constant temperature water bath. Specimens with no IL
line. Cells were obtained from Lonza (Australia) and
pretreatment were immersed in beakers for 1 min,
maintained in growth medium containing SingleQuot
10 min, 2 h, 24 h and 168 h. The IL pretreated speci-
supplements (Lonza, Australia). Three ILs, previously
mens were immersed for 2 h.
investigated12,13 for corrosion protection of magnesium
Flow cell corrosion tests alloys, were chosen: P1444DPP, [P66614][i(C8)2PO2] and
A flow cell was designed to simulate the blood flowing in P66614DPP.
the coronary artery vessels. The schematic of the flow Stainless steel (SS) sheets were used for the tests. They
cell is shown in Fig. 2. The test channel was made of were treated with IL and exposed to ultraviolet light for
transparent polymethylmethacrylate to allow the speci- 20 min each side in a 24 well plate. The HCAECs were
men to be observed. The flow rate in the constricted added onto the plate at a density of 1?56104 cells/cm2.
section of the test channel was designed to produce a The numbers of adherent cells on the IL treated speci-
shear stress in the flowing stream of SBF of 0?88 Pa for mens were quantified using a Quant-iTTM PicoGreen
this preliminary work. The specimens were inserted into dsDNA Assay Kit after 96 h of culture. An untreated
the test channel such that the diametric plane was specimen was used as control.
exposed to the flowing SBF. Exposure times of 10 min
and 2 h were used. Results and discussion
Scanning electron microscopy and optical Behaviour of Mg alloy AZ31 in static SBF
profilometry In this initial investigation, the corrosion behaviour of
After immersion, the corroded specimens were examined AZ31 was examined with immersion of specimens in
in an SEM. Energy dispersive X-ray spectroscopy (EDX) SBF under static condition in different times. Figures 3
was used to identify the composition of the corrosion and 4 present the SEM images of corroded surfaces,
product. After immersion, the specimens were immersed which had been immersed for 1 min, 10 min, 2 h, 24 h
for 60 s in a solution of 200 g chromium trioxide, 10 g and 168 h. Initially, the corrosion product accumulated
silver nitrate and 20 g barium nitrate in 1000 mL of around the intermetallic particles (Fig. 3). After 10 min,
distilled water to remove the corrosion products. The the corrosion product generally covered these interme-
depth and number of pits were determined using a Veeco tallic particles. Energy dispersive X-ray spectroscopy
Contour GT-K1 optical profilometer. analysis of the intermetallic particles (e.g. A in Fig. 3a)

376 Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5

 Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

3 Corrosion product morphology of AZ31 after immersion in SBF for a 1 min and b 10 min

revealed that they contained mostly Al, Mn and Mg. With future immersion times of up to 2, 24 and 168 h,
The Al–Mn particles are shown to be present in the non-uniform and highly cracked corrosion product
magnesium a matrix of AZ31.23 These particles are more layers were clearly visible over the entire surface of
cathodic than both the b phase and the Mg matrix.23 AZ31. Mud cracking was particularly evident on the
Therefore, the corrosion of AZ31 in SBF initiated from specimens exposed for 24 and 168 h. The greater extent
these Al–Mn particles as a result of the galvanic of mud cracking indicates a thicker layer of corrosion
difference between them and the surrounding matrix. products. The EDX data obtained after 24 h immersion
With increased immersion time, general corrosion of the in SBF (Fig. 5b) show the presence of O, Mg, Ca and P
Mg matrix occurred as a result of breakdown of the on the surface of the corroded AZ31. This is consistent
passive film [mainly a mixture of MgO and Mg(OH)2], with the observation reported previously.6 The presence
with the corrosion product eventually covering the of O and Mg indicates a corrosion product of MgO and
whole surface. In Fig. 3b, ‘mud cracking’ is evident on Mg(OH)2, while Ca and P arise from components of the
many of the surfaces. These ‘mud cracks’ developed SBF solution. The relation between Mg corrosion and
during drying of corrosion product in the SEM. the deposition of calcium phosphate has previously been

4 Images (SEM) of AZ31 specimens after immersion in SBF for a 2 h, b 24 h and c 168 h

5 Energy dispersive X-ray spectroscopy data obtained from a particle A in Fig. 3a and b corrosion products on surface
of specimen in Fig. 4b after immersion in SBF for 24 h

Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5 377

Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

8 Equivalent circuit for AZ31 immersion in SBF (note: Rt,

charge transfer resistance; Cd, double layer capacity;
Qf, capacity associated with surface film; Rf, surface
resistance; Re, electrolyte resistance; average value of
constant phase element a parameter50?830)

with immersion time. It is thought that26 the high

frequency loop is associated with the film of the corrosion
6 Typical Nyquist impedance spectra obtained at 1 h
product, and the medium frequency loop could be due to
interval during immersion of AZ31 in SBF
charge transfer resistance and double layer capacitance,
discussed, which suggested that Mg corrosion had an which form between the corroded layer and the metal
influence on the precipitation process of Ca and P from substrate. At the metal interface, when the solution
the SBF.24 The accumulation of calcium phosphate has reaches the metal surface, a charged double layer is
also been found during in vivo tests.25 For an Mg based established. The fact that there was an increase in high
stent, calcium phosphate deposits would be undesirable frequency loops indicates that the corrosion product was
and could under some circumstances threaten the life of a kept on accumulation on the surface during the 24 h
patient with an implant. Large precipitates of calcium immersion. The accumulation of corrosion product would
phosphate deposits could block an artery near the stent or increase the charge transfer resistance (Rt). Hence, the
travel to the other parts of the body. The precipitation of medium frequency loop increased with immersion time.
the corrosion product is associated with the local increase Figure 7 shows the corrosion product layer with a certain
in pH due to cathodic reaction accompanied by the thickness formed on the surface of AZ31 after immersion
dissolution of the Mg alloy matrix. A reduction in in SBF solution for 24 h. We should note that there were
corrosion rate would mean a smaller increase in local pH, cracks existing in the corrosion product layer, so the SBF
which in turn would influence this precipitated process. could still diffuse through these cracks to the metal surface
and enable corrosion.
Electrochemical impedance spectroscopy An equivalent circuit was fitted to the Nyquist data. A
To investigate in more detail the electrochemical corrosion schematic of this circuit is presented in Fig. 8. This
mechanism, EIS was conducted with an AZ31 electrode circuit consists of a resistor representing the electrolyte
surface in SBF, and the responses over a wide frequency resistance Re, and two RC units in series that describe
range were recorded. Analysing these responses and using the interfacial properties of working electrode: a charge
an appropriate equivalent circuit to fit the EIS data transfer resistance Rt in parallel with the double layer
provided information on the corrosion process. Figure 6 capacity, Cd, and a capacity associated with the surface
shows successive Nyquist impedance plots obtained at 1 h film of corrosion product Qf, in parallel with a surface
intervals during immersion of AZ31 in SBF under static resistance of corrosion product Rf. Here, the constant
conditions. The shape of these data is characterised by two phase element was used to represent a ‘non-ideal’
‘depressed’ capacitive loops at high and medium frequen- capacitance. The circuit provided a good fit, with x2
cies. With increased immersion time, the general shape of being ,0?004. One fitting example is shown in Fig. 9.
these plots did not significantly change, indicating that the
corrosion process did not significantly change with
immersion. Both real and imagery impedances for high
and medium frequency loops were observed to increase

7 Scanning electron microscopy observation of cross-

section of layer formed on surface of AZ31 in SBF for
24 h 9 Example of fit to data (x256?922e23)

378 Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5

 Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

10 Corrosion product morphology of AZ31 after immersion in flowing SBF for a 10 min and b 2 h

This circuit is consistent with the physical observation as which had a porous structure, leading to a large exposed
observed in the SEM, that is, corrosion occurring metal area in the aggressive solution. With SBF, the
beneath surface. existence of Ca and P in the corrosion layer could
Table 2 presents the parameters obtained from fitting increase the density of the corrosion product layer, and
the equivalent circuit to the data. It can be observed that thus, the exposed area in the solution was decreased with
Cd decreased with the increased immersion time. time. Therefore, the charge transfer resistance Rt in the
Meanwhile, Rt increased. The changes of Cd and Rt with SBF would have been larger than that in Na2SO4.
immersion could be explained by the high corrosion
reactivity of the AZ31 in SBF. As observed from Figs. 3 Effect of flowing SBF
and 4, the corrosion product progressively covered The appearance of the corroded surface after immersion
the whole surface with increasing immersion time, and in flowing SBF for 10 min and 2 h is shown in Fig. 10.
the accumulation of corrosion product would increase the In comparison with specimens exposed to static SBF, a
charge transfer resistance (Rt). The increase in Rt significant difference in corrosion morphology was
diminished the charge accumulation on the electrical evident. In flowing SBF (Fig. 10a), little corrosion
conductors of double layer capacitor, leading to a product was observed around the intermetallic particles,
reduction in double layer capacitance Cd. The increase while in the static environment, well defined precipitates
in Rt indicates a decrease in the corrosion rate resulting of corrosion product around intermetallic particles were
from the corrosion product accumulation on the surface, evident (Fig. 3b). The reason for this difference lies in
so a thick corrosion product layer provides some the fact that under flowing SBF, the specimen was
protection for AZ31 in SBF. Both Cf and Rf increased constantly flushed by the circulation of the test solution
with immersion time, which is also due to the progressive over the specimen surface; this would have prevented a
accumulation of corrosion products. The accumulation rise in local pH, thus reducing localised precipitation of
of corrosion product increases not only the thickness of corrosion product around the intermetallic particles,
the corrosion product layer (d) but also the coverage area promoting a more uniform corrosion product layer.
(A). It is clear that with time, A increased at a faster rate However, in the static environment, the corrosion
than d; hence, Cf, which is proportional to the A/d, product precipitated around the intermetallic particles
increased. and led to localised corrosion.
It is interesting to compare the EIS data with the pure The circled area in Fig. 10a shows an area on the
Mg in aerated Na2SO4 by other works.27 AZ31 in SBF corroded surface, which may have been removed by the
had a similar evolution of impedance behaviour with flow of liquid. After 2 h immersion in the flow cell, a
pure Mg in aerated Na2SO4; this suggests that the more uniform corrosion product film was visible on the
corrosion process of AZ31 in SBF was dominated by the surface of AZ31, in comparison with the film formed on
dissolution of the Mg matrix. However, the change in the specimen exposed to static SBF environment.
medium frequency loops was significantly different. These SEM observations suggest that with a flowing
With AZ31 in aerated Na2SO4, the medium frequency SBF (shear stress50?88 Pa), a more uniform corrosion
loops slightly changed with immersion time. One product morphology developed. At this stage, no data
possible explanation is that the corrosion product layer are available to allow a comparison of corrosion rate
of AZ31 in Na2SO4 mainly consisted of Mg(OH)2, between the static and the flowing SBF condition. Also,

Table 2 Parameter determined from Randles circuit fit to data

Time/h Re*/V cm22 Qf/mF cm22 a Rf/V cm22 Cd/mF cm22 Rt/V cm22 x2

0 40 22 0.84 62 4.1 17 4.561023

1 40 22 0.84 216 2.0 92 2.661023
6 40 28 0.83 495 1.6 252 4.561023
11 40 30 0.83 806 1.2 469 5.961023
16 40 34 0.82 1.26103 1.0 611 5.361023
21 40 34 0.82 1.56103 1.0 721 4.261023
*Assumed value.

Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5 379

Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

11 Human coronary artery endothelial cell morphology after 72 h culture with a P1444DPP pretreated SS, b
[P66614][i(C8)2PO2] pretreated SS, c P66614DPP pretreated SS and d control SS

the effect of flow rate (shear stress) is yet to be immersion, and the depth of corrosion was measured
determined. This effect will be reported in the future using optical profilometry. The three-dimensional optical
publication. profiles of the corroded surfaces are shown in Fig. 13,
which reveal the variation in the corroded area on the
Effect of IL pretreatment on corrosion differently surface treated specimens. For the P1444DPP
Cytotoxicity of ILs pretreated AZ31, this figure clearly shows that the
To test the cytotoxicity of ILs, SS specimens treated with corroded area was reduced. For the [P66614][i(C8)2PO2]
ILs were immersed in the cell culture for 72 h. Figure 11 pretreated AZ31, the corroded area was relatively larger
shows the HCAEC morphology after 72 h culture. Live than the P1444DPP pretreated AZ31, but it was still
cells were observed around the surface of specimens smaller than the control sample. Hence, qualitatively,
pretreated by P1444DPP and [P66614][i(C8)2PO2]. However, pretreatment of these two ILs can offer some protection
compared with the control specimen (no IL pretreat- for the AZ31 from corrosion in SBF.
ment), there was a significant reduction in cell density. To better compare the protective properties of the IL,
Less live cells were observed around the surface of the pit number, pit depth and volume lost were recorded
P66614DPP pretreated specimen. The statistics of live cells from an area of 937?561250?0 mm on each of the
in Fig. 12 confirmed the observation. Therefore, speci- specimens shown in Fig. 13. The histogram in Fig. 14
mens pretreated by P1444DPP and [P66614][i(C8)2PO2] are shows the distribution of pit depths with corresponding
relatively low toxic to HCAEC. pit numbers. Zero pit depth corresponds to areas of no
corrosion, and values on the x axis a little greater than
Difference of corrosion behaviour
zero correspond to peaks associated with the grinding
In order to qualify the protection performance of the IL damage from specimen preparation. It is immediately
treatment, the corrosion product was removed after apparent that the distribution of pit depths on the IL
treated specimens differed significantly from the
untreated surface. In particular, larger pits were absent
on the surface treated with P1444DPP (,3 mm deep and a
fairly sharp distribution), while some pit depths on the
[P66614][i(C8)2PO2] pretreated specimen approached
9 mm deep. In contrast, the untreated sample appeared
to have a broader distribution of pits, and these were
skewed to larger pit depths. There were a greater
number of pits on the control specimen compared with
the IL treated specimens. The volume of alloy removed
from the surface due to pitting was also calculated and
presented in Table 3. These data also clearly demon-
12 Human coronary artery endothelial cell viabilities after strate that, with these two IL pretreatments, a smaller
96 h culture with IL treated SS {IL1: P1444DPP, IL2: volume of alloy was removed by corrosion during the
[P66614][(iC8)2PO2], IL3: P66614DPP} immersion for 24 h in SBF, suggesting that such

380 Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5

 Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

14 Histogram results from optical profilometry analysis of

control AZ31 and IL pretreated AZ31 after immersion in
SBF for 24 h {IL1: P1444DPP, IL2: [P66614][(iC8)2PO2]}

15 Film resistance of AZ31 versus exposure time

as ‘uniform’ corrosion. For the corrosion of coronary

artery stents in the body, uniform corrosion is optimal.
The IL pretreatment can increase the possibility of
uniform corrosion.
The protective effect of the IL was also accessed using
data from the EIS test. Figure 15 shows the variation in
film resistance with immersion time, which is extracted
13 Corroded surface of a P1444DPP pretreated AZ31, b
from the EIS spectra. These data clearly indicate that the
[P66614][i(C8)2PO2] pretreated AZ31 and c control observed
film resistance of both IL pretreated specimen and
by optical profilometry (coloured version available online)
control specimen increased rapidly after immersion in
SBF. The significant modification of surface film
pretreatments can increase the corrosion resistance of resistance during immersion can be attributed to high
AZ31 in SBF. Furthermore, the IL treatment appeared reactivity of the AZ31 in SBF. The existence of a
to decrease the depth of pitting and led to a greater protective IL passive film on AZ31 is confirmed by a
number of shallow pits, which could ultimately appear higher film resistance at the beginning of immersion
(0 h). The film resistance on the IL treated specimen did
not increase with time as much as the untreated
Table 3 Net volume lost of alloys removed during
specimen. This indicates that the presence of IL reduced
corrosion {IL1: P1444DPP, IL2: [P66614][(iC8)2PO2]}
the amount of corrosion product, which developed on
IL1 treated IL2 treated Control the surface, and that is reducing corrosion rate.
However, corrosion still occurred in the presence of
AZ31/mm3 AZ31/mm3 AZ31/mm3 IL, and this could be due to a heterogeneous and/or
defective film of IL. For a film to provide protection on
Volume lost 6.56105 11.76105 27.96105
an Mg alloy for coronary artery stent, a homogeneous

Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5 381

Zhang et al. Corrosion behaviour of AZ31 and influence of liquid pretreatments

film with uniform coverage across the entire surface will References were renumbered.
be required. Such a film may have an ‘ideal’ rate of
change of film resistance in SBF, as shown by the dashed 1. H. Hermawan, D. Dubé and D. Mantovani: Acta Biomater., 2010,
6, 1693–1697.
line in Fig. 15. This slower rate of increase would 2. R. S. Schwartz, N. A. Chronos and R. Virmani: J. Am. Coll.
represent a slow rate of corrosion that the body can Cardiol., 2004, 44, 1373–1385.
tolerate, making sure that the coronary artery stent can 3. R. L. Moreno, C. Fernández, R. Hernández, F. Alfonso, D. J.
keep its mechanical integrity before the diseased vessel Angiolillo, M. Sabate, J. Escaned, C. Bañuelos, A. Fernández-
Ortiz and C. Macaya: J. Am. Coll. Cardiol., 2005, 45, 954–959.
returns to its original shape, and be gradually degraded 4. R. Waksman, R. Pakala, P. K. Kuchulakanti, R. Baffour,
and absorbed in the body. D. Hellinga, R. Seabron, F. O. Tio, E. Wittchow, S. Hartwig,
C. Harder, R. Rohde, B. Heublein, A. Andreae, K. H. Waldmann
and A. Haverich: Cathet. Cardiovasc. Interv., 2006, 68, 607–617.
Conclusions 5. J. Vormann: Mol. Aspects Med., 2003, 24, 27–37.
6. Y. Song, D. Shan, R. Chen, F. Zhang and E.-H. Han: Mater. Sci.
Corrosion of Mg alloy AZ31 in static SBF initiates beside Eng. C, 2009, C29, 1039–1045.
Al–Mn intermetallic particles in the Mg matrix. With 7. Polmear I. J: ‘Light alloys’, 5th edn, 237–297; 2005, Oxford,
increased immersion time, a thick and non-homogeneous Butterworth-Heinemann.
corrosion product layer eventually covers the whole 8. F. Witte, N. Hort, C. Vogt, S. Cohen, K. U. Kainer, R. Willumeit and
F. Feyerabend: Curr. Opin. Solid State Mater. Sci., 2008, 12, 63–72.
surface. The EDX analysis indicates the presence of Mg, 9. M. Forsyth, P. C. Howlett, S. K. Tan, D. R. MacFarlane and
O, Ca and P in the corrosion product layer. Mg and O are N. Birbilis: Electrochem. Solid-State Lett., 2006, 9, B52–B55.
most likely Mg(OH)2, and Ca and P rise from the 10. G. Song and S. Song: Adv. Eng. Mater., 2007, 9, 298–302.
components in SBF. Electrochemical impedance spectro- 11. H. E. Friedrich and B. L. Mordike: ‘Magnesium technology –
scopy analysis confirms the growth process of the metallurgy, design data, applications’, 2006, Secaucus, NJ,
Springer-Verlag.
corrosion product layer on the surface of AZ31 in SBF. 12. M. Forsyth, W. C. Neil, P. C. Howlett, D. R. MacFarlane, B. R.
The corrosion product layer can provide corrosion W. Hinton, N. Rocher, T. F. Kemp and M. E. Smith: ACS Appl.
protection for AZ31 in SBF, but as there are cracks in Mater. Interfaces, 2009, 1, 1045–1052.
the corrosion product layer, corrosion of AZ31 still 13. P. C. Howlett, T. Khoo, G. Mooketsi, J. Efthimiadis, D. R.
MacFarlane and M. Forsyth: Electrochim. Acta, 2010, 55, 2377–2383.
continues. In the flowing condition, the circulation of the 14. P. C. Howlett, N. Brack, A. F. Hollenkamp, M. Forsyth and D. R.
SBF over the surface of the specimen inhibits the local pH MacFarlane: J. Electrochem. Soc., 2006, 153, A595–A606.
increase, thus avoiding localised corrosion and promot- 15. M. Uerdingen, C. Treber, M. Balser, G. Schmitt and C. Werner:
ing more uniform corrosion product morphology. Green Chem., 2005, 7, 321–325.
16. N. T. Kirkland, N. Birbilis and M. P. Staiger: Acta Biomater., 2011,
The cytotoxic test shows that SS pretreated by
8, (3), 925–936.
P1444DPP and [P66614][i(C8)2PO2] is relatively low toxic 17. H. Wang, Y. Estrin and Z. Zuberova: Mater. Lett., 2008, 62, 2476–
to HCAEC. Using these two ILs that pretreat the AZ31, 2479.
the IL pretreatment can increase the corrosion resistance 18. Z. Wen, C. Wu, C. Dai and F. Yang: J. Alloys Compd, 2009, 488,
of AZ31 in SBF and the possibility of uniform corrosion. 392–399.
19. T. Kokubo and H. Takadama: Biomaterials, 2006, 27, 2907–2915.
However, IL pretreatment cannot provide good corro- 20. J. Sun, P. C. Howlett, D. R. MacFarlane, J. Lin and M. Forsyth:
sion protection for AZ31 in SBF. One possible reason is Electrochim. Acta, 2008, 54, 254–260.
that the IL film formed by this treatment method is 21. P. C. Howlett, S. Zhang, D. R. MacFarlane and M. Forsyth: Aust.
defective and/or heterogeneous. J. Chem., 2007, 60, 43–46.
22. R. Cottis and S. Turgoose: ‘Electrochemical impedance and noise’;
1999, Houston, TX, NACE International.
Acknowledgements 23. G. L. Song and A. Atrens: Adv. Eng. Mater., 1999, 1, 11–33.
24. J. E. Gray-Munro and M. Strong: J. Biomed. Mater. Res. A, 2009,
The authors would like to acknowledge Boston Scientific 90A, 339–350.
and ACEs for funding the linkage project no. LP0990621. 25. F. Witte, V. Kaese, H. Haferkamp, E. Switzer, A. Meyer-
Special thanks to S. Moulton (University of Wollongong) Lindenberg, C. J. Wirth and H. Windhagen: Biomaterials, 2005,
26, 3557–3563.
and J. Weber and T. Schenermann (Boston Scientific) for 26. G. Baril, C. Blanc and N. Pebere: J. Electrochem. Soc., 2001, 148,
advice and organisation and J. Sun (Monash University B489–B496.
and Cytes Industrials) for providing the IL used. 27. G. Baril and N. Pebere: Corros. Sci., 2001, 43, 471–484.

382 Corrosion Engineering, Science and Technology 2012 VOL 47 NO 5

Copyright of Corrosion Engineering, Science & Technology is the property of Maney Publishing and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.