yingdou
TRANSCRIPT
-
8/11/2019 YingDou
1/8
45S5 bioactive glassceramic coated AZ31 magnesium alloy withimproved corrosion resistance
Ying Dou a, Shu Cai a,, Xinyu Ye a, Guohua Xu b, Kai Huang a, Xuexin Wang a, Mengguo Ren a
a Key Laboratory for Advanced Ceramics and Machining Technology of Ministry of Education, Tianjin University, Tianjin 300072, People's Republic of Chinab Shanghai Changzheng Hospital, Shanghai 200003, People's Republic of China
a b s t r a c ta r t i c l e i n f o
Article history:
Received 1 December 2012Accepted in revised form 8 April 2013
Available online 13 April 2013
Keywords:
45S5 glassceramic coating
Magnesium alloy
Corrosion resistance
Solgel
To control the biodegradation rate, 45S5 glassceramic coatingswere prepared on the commercial AZ31magnesium
alloy substrates by dip-coating method from a synthesized solgel. Sol concentration, calcination temperature and
dip-coating cycle have been optimized to prepare the compact coatings. The results showed that homogeneous
and crack-free coatings with a thickness of 0.481.00m, consisting of amorphous phase and Na2Ca2Si3O9, were
successfully fabricated on AZ31 magnesium alloys. The effects of these coatings on the corrosion behavior of the
magnesium alloy substrates have been investigated in vitro by soaking samples into modied simulated body
uid (m-SBF) for different periods. It was found that optimized 45S5 glassceramic coatings could slow down the
degradation rate and decrease the mass loss of the magnesium alloy substrate from 78.04% to 2.31% in the 7th
day test, showing a good anti-corrosion property in a certain period. Meanwhile, calcium-decient hydroxyapatite
deposition was observed on the surface of sample 3A500, indicating its biomineralization property in m-SBF. Nev-
ertheless, cracking of the coating during the immersion test is the major factor for 45S5 glass ceramic coatings to
fail to protect the magnesium alloy substrates in the later immersion period.
2013 Elsevier B.V. All rights reserved.
1. Introduction
Magnesium (Mg) and its alloys have been considered as a prom-
ising metallic material for biodegradable bone implants due to their
desirable mechanical properties, biocompatibilities and biodegrad-
abilities. The density of magnesium, 1.74 g cm3, is close to that of
natural bones (1.72.1 g cm3)[1],showing excellent biomechanical
compatibility. More importantly, compared with other metallic implant
materials such as stainless steels and Ti alloy, magnesium and its alloys
have a moderate elastic modulus closer to that of bone. Combined with
a density very close to that of bone, stress-shielding effect and the risk
of inducing osteoporosis can be effectively minimized[1,2]. Magnesium
has no noxiousness and largely presents in human body, and excess of
magnesiumcan be easilyexcreted[3]. Moreover,magnesiumplays an im-
portant role in bone metabolism and may promote the formation of new
bone tissue [4]. As biodegradable orthopaedic implant materials, it is
projected that magnesium and its alloys would remain to present in
the body and maintain their mechanical integrity over a timescale of
1218 weeks while the bone tissue heals, eventually being replaced by
natural tissue[1,5]. However, the high corrosion rate of magnesium and
its alloys in chloride containing environment, such as human body uid
or blood plasma, ledto thefast lossof mechanicalintegrity andtherelease
of hydrogen, which limited their biomedical applications [6,7]. Therefore,
it is important to improve the corrosion resistance of magnesium and itsalloys for their successful application as bone implant materials. Surface
modication with appropriate coatings on Mg-based alloys, such as
micro-arc oxidation [8,9], electrochemical deposition [10,11], plasma
electrolytic oxidation [12] and solgel [13,14] etc., is regarded as an effec-
tive way to reduce the degradation rate of these Mg alloys. Among these
methods, coatings formed through solgel method on metal surface can
be more adherent, uniform[15,16]and bioactive[17]. Meanwhile, this
methodhasseveral other advantages,such as ease of composition control,
low processing temperature, and being efcient in producing lms or
coatings on complex shaped implants and porous scaffolds[18].
Bioactive glasses and glassceramics have been widely used in bio-
medical applications as their superior bioactivities. The primary charac-
teristic of bioactive glasses and glassceramics is their rapid rate of
surface reaction when used as human body implant materials, leading
to direct and fast attachment to bone by a chemical bond [19]. There-
fore, applying bioactive glasses and glassceramics as coating materials
on magnesiumand its alloys for surface modication not only combines
the bioactivities of bioactive glasses and glassceramics with the ne
mechanical properties of metallic materials but also improves the
anti-corrosion performance. Researchers have applied bioactive glasses
on conventional metallic biomaterials, such as stainless steels [20]and
Ti-based alloys[21]for decades, and these materials have been proved
to have increased corrosion resistance. Fathi et al.[20]have reported
that the corrosion current density (85 nA/cm2) of the316 L stainless
steel coated with bioactive glass coating was smaller than that of
uncoated sample (265 nA/cm2) and its corrosion potential was 26%
Surface & Coatings Technology 228 (2013) 154161
Corresponding author. Tel.: +86 22 27425069.
E-mail address:[email protected](S. Cai).
0257-8972/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.surfcoat.2013.04.022
Contents lists available at SciVerse ScienceDirect
Surface & Coatings Technology
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / s u r f c o a t
http://dx.doi.org/10.1016/j.surfcoat.2013.04.022http://dx.doi.org/10.1016/j.surfcoat.2013.04.022http://dx.doi.org/10.1016/j.surfcoat.2013.04.022mailto:[email protected]://dx.doi.org/10.1016/j.surfcoat.2013.04.022http://www.sciencedirect.com/science/journal/02578972http://crossmark.crossref.org/dialog/?doi=10.1016/j.surfcoat.2013.04.022&domain=pdfhttp://www.sciencedirect.com/science/journal/02578972http://dx.doi.org/10.1016/j.surfcoat.2013.04.022mailto:[email protected]://dx.doi.org/10.1016/j.surfcoat.2013.04.022 -
8/11/2019 YingDou
2/8
higher in the normal saline solution, indicating the enhancement of
corrosion resistance.
Nevertheless, little researches have focused on the bioactive glassce-
ramic coatings on magnesium and its alloys, and the corrosion resistance
of such coatings has not been determined. In this work, crack-free 45S5
bioactive glassceramic coatings have been successfully fabricated on
AZ31 magnesium alloys through solgel dip-coating method via control-
ling the preparation parameters. The in vitro corrosion properties of 45S5
glass
ceramics coated AZ31 magnesium alloys were also investigated byimmersing the samples in modied simulated body uid (m-SBF) at
37 C for different periods.
2. Material and methods
2.1. Preparation of 45S5 glassceramic coatings on AZ31 magnesium
alloy substrates
Magnesium alloy pieces (12 12 2 mm3) were cut from the com-
mercial AZ31 magnesium alloy plate (Al 3 wt.%, Zn 1 wt.%, Mn 0.2 wt.%,
Fe b0.005 wt.%) as substrates in this study. The substrate surfaces
were ground to grits of 1000, 1500 and 2000 by SiC papers progressive-
ly, followed by ultrasonically cleaning in distilled water, ethanol and
acetone successively for 15 min, respectively, and then dried at room
temperature.
45S5 bioactive glassceramic coatings were dip-coated from a syn-
thesized sol based on previous investigation[22]. Briey, the molar ra-
tios of tetraethyl orthosilicate (TEOS), sodium nitrate (NaNO3), calcium
nitrate tetrahydrate (Ca(NO3)2 4H2O) and triethyl phosphate (TEP)
were designed according to the molar ratios of SiO2, Na2O, CaO and
P2O5in 45S5 (46.14 %, 24.35 %, 26.91 %, and 2.60%, respectively). The
sol for fabrication of 45S5 bioactive glassceramic coatings was pre-
pared by mixing two separate solutions. Solution I was prepared by
dissolving TEOS into the HNO3(0.1 M) aqueous solution at room tem-
perature and stirring for 30 min to hydrolysis. Thereupon, TEP was
added into the solution to hydrolyze for 20 min. Solution II was prepared
by dissolving NaNO3and Ca(NO3)2 4H2O in distilled water. Then the
two solutions were mixed together and stirred for 3 h to obtain a homo-
geneous, clear and transparent sol. Two sols withdifferent concentrationswere synthesized by adjusting the molar ratio of TEOS/H2O to 0.007 or
0.014, and were labeled as A and B, respectively. Then coatings were de-
posited on AZ31 magnesium alloy for 15 cycles via dip-coating tech-
nique with a withdrawal speed of 0.5 mm/s, aged at room temperature
for 24 h, dried at 60 C for 1 h and calcinated at temperature of 400 and
500 C for 90 min, respectively. Each layer in the multilayered coatings
was performed after the drying of the previous one. The processing con-
dition used in the experiment and the different samples studied are listed
inTable 1.
2.2. Coating characterization
Thesurface morphologies of thecoatings were observed byeldemis-
sion scanning electron microscope (FE-SEM, JOEL6700F, Japan), and thechemical composition of the coating and deposition on the coating after
being immersed for different period was analyzed by energy dispersive
spectrum (EDS, 7401 Oxford). Low-angle (1) X-ray diffraction (XRD,
Rigaku D, Japan) was used to examine thephase compositionof the coat-
ing. Datas were collected for2 ranging between 10 and 80 using Cu K
radiation.
2.3. Immersion experiment
In this study, in vitro immersion tests were carried out at 37 C in
m-SBF (pH = 7.40) prepared by Oyane et al. [23]for intervals from 1 to
7 days to investigate the corrosion properties of the uncoated and coated
samples. The volume of solution was calculated based on a volume-
to-sample area of 20 mL/cm2
, according to ASTM G31-72[24]. After
predetermined periods of time, the samples were removed from the
m-SBF,washed gently withdistilled water and dried in air at room tem-
perature for the use of surface observation by FE-SEM. The residual so-
lutions were used to determine the pH values by a pH meter (PB-10,
China). To evaluate the mass loss of AZ31 magnesium alloy substrates,
samples immersed for different periods were cleaned in a chromic
acid solution (K2Cr2O7+ H2SO4) to remove the coatings and corrosion
products formed on the samples, and then rinsed with distilled water,
cleaned ultrasonically in ethanol and dried in air. The mass loss was cal-culated as follows:
Mass Lossm0ml
m0 100 1
where m0 is the substrate mass before immersing, andml is the massof
immersed sample after being cleaned by chromic acid. An average of
three measurements was used for evaluating the sample mass varia-
tion. The surface observation of the immersed samples through FE-
SEM technology has been taken without washing by chromic acid
solution.
3. Results and discussion
3.1. Surface morphologies and phase composition of 45S5
glassceramic coatings
The corrosion resistance of magnesium alloy substrates coated with
solgel coatings is ascribed to the coating physical barrier properties
and coating solubility related to the chemical composition, so forming
a homogenous and crack-free coating is important to improve the
anti-corrosion performance of such a coated sample. To prepare an in-
tact coating without cracks, the sol concentrations, calcinated tempera-
tures and coating thickness (dip-coating cycles) were optimized in this
work.
Surface morphologies of different coated samples prepared by using
different molar ratio of TEOS/H2O and followed drying at 60 C and
calcinating at different temperatures (400 and 500 C) are shown in
Fig. 1.FE-SEM images display that samples obtained from both sol Bcracked when calcinated at 400 C and 500 C (Fig. 1a and b). On the
surface of sample B500 (Fig. 1b), reticular cracks sized among 13 m
can be observed, larger than those of sample B400. For coatings derived
from sol A, small size of cracks also was found on the sample surfaces
after calcinated at 400 C (Fig. 1c), while a relative smooth and uniform
solgel coating can be obtained when the calcinate temperature is
500 C, as shown inFig. 1d. It is considered that the surface morphology
of the solgel coating is affected by the properties of the precursory sols,
such as the viscosity, surface tension, stability and etc. In this case, the
concentration of the 45S5 glassceramic precursor sols increases with
the enhancement of the molar radio of TEOS/H2O, which simultaneously
leads to the increase of viscosity of the precursor sols. The surface tension
is high for the coating prepared from sol with high viscosity[25]and the
tension would deteriorate thesurface qualityand integrality of thesolgelcoating [26], thus it can be deduced that the obvious cracks formed on the
Table 1
Processing conditions used for preparing different 45S5 bioactive glassceramic coatings.
Sample name TE OS/H2O Calcination
temperature/C
Dip-coating cycle
A400 0.007 400 1
A500 0.007 500 1
2A500 0.007 500 2
3A500 0.007 500 3
4A500 0.007 500 4
5A500 0.007 500 5
B400 0.014 400 1
B500 0.014 500 1
155Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
-
8/11/2019 YingDou
3/8
surface of samples B400 and B500 are attributed to the high viscosity of
the precursory sol. Many researchers have demonstrated that coatings
with obvious cracks could not prevent the underlying magnesium alloy
substrate from being exposed to the surrounding immersion medium
due to the permeation of soaking medium through cracks[27]. Thereby
the optimal TEOS/H2O ratio of 0.007 should be used afterwards.
By comparing the surface morphologies of coated samples de-served from solA (TEOS/H2O ratio of 0.007) calcinated at temperatures
of400 Cand500 C (Fig. 1c andd), it can be observed that cracks sized in
several hundreds nanometers appeared on the surface of A400 (Fig. 1c).
However, inthe caseof A500, only smallerand narrowercracks can beob-
served, as given in the inset in Fig. 1d. During the preparation processesof
the solgel coatings, crack formation is mainly associated with the resid-
ual stress appeared in the coatings. Residual stress is the sum of intrinsic
stress and thermal stress, and intrinsic stress in the solgel coatings is in-
duced by densication of coatings under constraint. As soon as the gel
coatings formed on the magnesium alloy substrate, gel coatings shrink
due to solvent evaporation as well as polymerization and cross-linking
of the gel, densifying the coating materials. In this process, the coatings
stuck to the substratecannot shrink in the direction of parallel to the sub-
strate surface; therefore tensile stress evolves in the coatings. Neverthe-less, the formation of thermal stress is mainly ascribed to the mismatch
of thermal expansion coefcients between AZ31 magnesium alloy sub-
strate and 45S5 glassceramic coating. When theresidual stress appeared
in the coatings exceeds the stability (inner coherence) of the material,
cracking is observed. If coating stress would be relieved by viscous ow
dueto the transition andsofteningof glassyphase or structural relaxation,
the coatings would be crack-free [28,29]. Theglass transition temperature
(Tg) of 45S5 glasses is conrmed to be approximately 500 C by previous
document [30], thus the fewer cracksformed in sample A500 arepresum-
ably as a result of the release of residual stress due to the transition and
softening of the glassy phase at 500 C[19,31]. Therefore, to obtain a
good anti-corrosion performance and avoid cracking of the coating, the
annealing temperature 500 C was selected in this work to fabricate the
45S5 glass
ceramic coatings on AZ31 magnesium alloy substrates.
It has been reported that coating thickness, which is mainly related to
the dip-coating cycles, also plays an important role to effectively protect
the magnesium alloy substrates[32]. As reported by Zhu and coworkers
[33], the diffusion perpendicular to the surfaces is greatly restricted by
the coating thickness, the thicker the coating is, the less Mg2+ releases.
Fig. 2shows the surface morphologies of the uncoated and coated
AZ31 magnesium alloy substrates for 2 and 3 cycles, as well as thecross-sections of sample 3A500. The uncoated sample shows amounts
of scratches formed during the grinding process of the substrate during
surface preparation, while absence of scratches on the surface of the coat-
edones (Fig. 2b and c) indicated the successful preparation of dense con-
tinuous coating. Crack-free, dense and continuous coatings can only be
obtained when dip-coated for up to 3 cycles; whereas for the samples
coated for 4 and 5 cycles, cracks appeared in the coatings. In this work,
the coating thickness increases with the repeat of dip-coating process
and the thickness for 3A500 is measured to be 1.00 m (as shown in
the inset inFig. 2c). Meanwhile, no cracks across the coating from the
surface to the AZ31 magnesium alloy substrates could be observed for
3A500, indicating the integrality and uniformity of the prepared coat-
ing, which is favorable to enhance the corrosion resistance of AZ31
magnesium alloy substrates. Corkovic et al.[34] have calculated the re-sidual stress in the solgel coatings according to Stoney's equation and
the results evidenced that residual stress would decrease with the in-
crease of coating thickness when the thickness of the coating ranged
from 0.20 to 1.00 m. In present work, all the coated samples pre-
pared without cracking, compared with samples 3A500, A500 and
2A500 whose thickness is approximately 0.48 and 0.63 m, respective-
ly, have been deduced to accumulate more residual stressand to be eas-
ier to generate signicant cracks. From this point of view, cracking
behavior of these coatings would show a great difference throughout
the immersion tests, which will be discussed in theSection 3.2.
Besides themicrostructure(such as pores and cracks) andthicknessof
the coating, the corrosion behavior of the AZ31 magnesium alloy sub-
strateswith coatings is greatlydependent on thecomposition of theinitial
coatings and the changes of coating structure during the degradation
Fig. 1.FE-SEM images of surface morphologies of different 45S5 bioactive glass ceramic coatings: (a) B400, (b) B500, (c) A400 and (d) A500.
156 Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
-
8/11/2019 YingDou
4/8
process, such as cracking and peeling off of the coatings. Therefore, the
phase composition of the coating with the optimized microstructure
is characterized by XRD. The low-angle (1) XRD patterns of 45S5 glass
ceramic powders and coated sample 3A500 are shown in Fig. 3. The
45S5 glassceramics powder (Fig. 3a) is composed of amorphous phase
and crystalline phase Na2Ca2Si3O9 (according to the JCPDS card No.
22-1455), which is consistent with the result of Qian et al. [31]. For the
coated sample 3A500 (Fig. 3b), as peaks of Mg (according to the JCPDS
card No. 35-0821) at 2 = 34.4, 36.6, 57.5, 63.1 and 69.7 were
detected, some peaks of Na2Ca2Si3O9 were overlapped and could not
be detected. However, the diffraction pattern of amorphous phase and
Na2Ca2Si3O9phase inFig. 3b is similar to that inFig. 3a, displaying the
successful formation of glass
ceramic coating on AZ31 magnesiumalloy substrate.
On the basis, the uniform and crack-free coatings of 45S5 glass
ceramics canbe prepared on AZ31 magnesiumalloys followed thecondi-
tionof preparing theprecursorysol at TEOS/H2O ratio of 0.007,dip-coated
13 cycles, and then calcinated at 500 C. Thereafter, the uncoated and
coated samples withdifferent dip-coating cycles wereused to investigate
the corrosion properties byin vitroimmersion tests in m-SBF at 37 C in
the following work.
3.2. In vitro corrosion properties and biomineralization
The immersion experiment can provide information with respect to
the longer term protection afforded by the glassceramic coatings. To in-
vestigate the corrosion behavior, the mass loss of magnesium alloy sub-strates and pH variation of the immersing medium for different systems
were examined and illustrated inFigs. 4 and 5, respectively. It can be ob-
served that after one day of immersion the mass loss of theuncoated and
coated samples A500, 2A500 and 3A500 was 2.31%, 1.68%, 1.40% and
0.82% respectively, and all testing solutions with coated samples showed
lower pH values compared to the m-SBF with the uncoated sample
(9.68). High mass lossand pH variation represent low anti-corrosion per-
formance.The coatings canhinderthe permeationof surrounding immer-
sion mediumand protect thesubstrate.It shouldnot be neglected that the
pH of the m-SBF with 3A500(8.71) is lower than that with A500(9.06) or
2A500 (9.02), suggesting that the glassceramic coatings with different
thickness can hinder the corrosion reaction in the initial immersion peri-
od, whereas the thick coating showed better protective effect. Increasing
the immersion time to 3 days, the mass loss of samples A500, 2A500 and3A500 only showed a slight increase while the uncoated AZ31 magne-
sium alloy substrate displayed sharp increase to 31.21%, accompanying
the similar tendency in pH variation. In the following immersing stage,
remarkable increase of mass loss for samples A500 (50.63%) and
2A500 (40.35%) after 5 days immersion was detected, and the pH
value of m-SBF for A500 and 2A500 underwent a notable increase
and reached 10.51 and 10.83 respectively, nearly to that of the
uncoated sample (52.33% for mass loss and 11.01 for pH value).
While the mass loss and pH value of m-SBF for sample 3A500 were
Fig. 2.FE-SEM images of surface morphologies of (a) uncoated AZ31 magnesium alloy,
(b) 2A500 and (c) 3A500 (the inset is the cross-section FE-SEM graph of 3A500).
Fig. 3.Low-angle (1) X-ray diffraction pattern of (a) 45S5 glassceramic powders and
(b) sample 3A500.
Fig. 4. Mass loss of the uncoated and coated samples immersed in m-SBF for different
periods.
157Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B4%80http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B3%80http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B2%80 -
8/11/2019 YingDou
5/8
2.02% and 9.29 respectively, much lower than the other two coated
samples. Continuing to extend the immersion time to 7 days, the
mass loss for samples A500 (72.72%) and 2A500 (72.24%), slightly
lower than that of naked sample (78.04%), presented the similar ten-
dency, and their pH values reached 11.03 and 11.04, respectively. Al-
most the same with the uncoated samples (11.05) suggesting that
the coatings on samples A500 and 2A500 could not effectively pre-
vent the corrosion attack of the substrates after 7 days immersion.
However, sample 3A500 still showed the low mass loss (2.31%) and suit-
able pH value (9.67) after 7 days immersion, indicating that the 3A500
coating canprovide a relatively long-term protectionto AZ31 magnesium
alloy substrates. It can be concluded that the thickness of the coating
played an important role in theprotection of theMg alloy substrate.Sam-
ple 3A500 with the thickest thickness (~1.00 m) displayed the optimal
anti-corrosion behavior.
In order to determine the corrosion and protection mechanisms
responsible for the behaviors observed, the surface morphologies and
cross-sections of samples after being immersed for different days were
examined by FE-SEM and the corresponding results are depicted inFigs. 6 and 7. After being immersed for 1 day, plate-like Mg(OH)2layers
formed on the surface of the uncoated sample (Fig. 6a) dueto theelectro-
chemical reaction of Mg withwater, demonstrating the corrosion of AZ31
magnesium alloy. With regard to the coated samples, the above analysis
of the mass loss and the pH variation demonstrated that coatings pro-
vided a barrier effect to retard the contact between substrate and im-
mersion medium in the initial immersion period. The surface changes
of the uncoated and coated sample also conrmed the protective ef-
fects. As the surface morphology of the three coated samples after
being immersed for 1 day was observed to be similar by FE-SEM, the
image for sample 3A500 (Fig. 6b), was chosen to represent the three
samples. As shown inFig. 6b, after being immersed in m-SBF for 1 day
the coating cracked, while the coating still tightly adheredto AZ31 mag-
nesium alloy substrates and did not seriously peel off, which could pro-vide further protection for substrates. Cross-sections of sample A500
(Fig. 7a) in this period showed that some vertical cracks formed on
the surface of the coating, while cracks were not through the coating.
Nevertheless, 3A500 displayed fewer cracks and even no obviouscracks
can be observed from the cross-section FE-SEM micrograph shown in
Fig. 7b. Therefore, the three coated samples displayed slight corrosion.
From the EDS spectrum of the coating (Fig. 6c), it can be found that
the coating degraded and no CaP deposits were formed. After 5th day
immersion, as coating seriously detached from the substrate for sample
A500, amounts of corrosion products formed in the cracks and severe
corrosion appeared, shown inFig. 6d. From the image ofFig. 7c, serious
crack and break of coating were observed for A500. Moreover, corrosion
products were formed on the exposed substrates without coating, indi-
cating severe corrosion behavior. From the result of EDS spectrum of
the deposit in Fig. 7d,the molar ratio ofMg/O equalsto 0.5,it can bede-
duced that amounts of Mg(OH)2have been deposited on the broken
coating, resulting in the increase of coating thickness shown in Fig. 7c.
However, for sample 3A500, cracks gradually formed through the coat-
ing(Fig. 7e), while the cracked coatingstill adhered to thesubstrate and
did not peeloff(Figs. 6e and 7e), suggesting that it can providea certain
degree of protection for the magnesium alloy substrates in the later
immersion process. EDS spectrum (Fig. 7f) of the deposited layer dem-
onstrated that Ca
P deposit withCa/P molar ratio of almost 1.60 formedon the surface of sample 3A500. Moreover, spherical deposits further
deposited on the surface of sample 3A500 after immersing for 7 days
(Fig. 6f) and EDS spectrum demonstrated that the molar radio of Ca/P
was almost 1.60 (Fig. 6g), evidencing the deposition of calcium-decient
hydroxyapatite and the biomineralization of sample 3A500.
To better understand the corrosion behavior of the coated samples in
immersion test, the corrosion process for the samples A500 and 3A500 is
schematically illustrated inFig. 8. According to above discussions, no ob-
vious cracks were found in the coatings before the immersion tests,
while it is should not be neglected that the residual stress would be accu-
mulated in the coatings during the preparation processes and sample
3A500 possessed the lowest stress as discussed in the Section 3.1. During
theimmersiontest, thepermeationand physical washing of m-SBF would
cause signicant strain generation and residual stress release, leading to
the cracking of the coating. As reported by Yang and coworkers [35],
cracks are easy to form on theedge where stress concentrated. For coated
samples, corrosion may start at surface defects in the coating, and after
corrosion initiation at defects, fracture and aking off of the coating
could take place[36]. In this work, corrosion was favored to appear on
the edge of the coated sample and led to slight increase in mass loss
and pH value in the initial 1 day immersion, as given in Figs. 4 and 5.
Meanwhile, for both samples A500 and 3A500, cracks formed on the sur-
face of the coatings but not through the coatings (Figs. 6b,7a and b),
which also might contributeto the slight corrosion. However, compared
with sample 3A500, cracks resulted fromstress release easily formed
in the coating of sample A500 owing to the higher accumulated re-
sidual stress when immersed into m-SBF. Compared with sample
3A500 (Fig. 7b), more cracks on sample A500 can be observed from
the images of cross-section in Fig. 7a. It could be deduced that thecrack propagation for sample A500 is easier than that of sample 3A500.
Thus after 3 days immersion, cracks across the coating from the surface
to the AZ31 magnesium alloy substrates appeared for sample A500,
whereas for sample 3A500 fewer cracks can be observed on the surface,
as illustrated in Fig. 8; therefore, corrosionof sampleA500 wasslightly se-
verer than that of 3A500. In the following stages, immersion mediums
permeated into sample A500 through the cracks and reacted with the
substrate,causing obvious peelingoff of the coatingand seriouscorrosion
of substrate after 5-day of immersion (Figs. 6d,7c and8), and the sharp
increase in mass loss and pH value (Figs. 4 and 5). Additionally, sample
2A500 displayed the similar tendency with A500 and the coatings were
inadequate to protect the substrate due to the formation of through
cracks during the immersion test. Coatings with low thickness were con-
sidered to be broken more easily under external force, while thick coat-ings might display preferable performance in a relative long term. Thus
forsample 3A500 although cracksexpanded acrossthecoatingand corro-
sion occurred in the cracking zones after being immersed for 5 days
(Fig. 7d), coatings did not detach from the AZ31 magnesium alloy
(Figs. 6e and 8) and still acted as a physical barrier to retard the evolution
of the corrosion process in the 7 days test. And for sample 3A500 the pH
value of m-SBF showed the smallest change and did not surpass 10.00,
which is favorable for the formation of apatite, as reported in document
[37]. Meanwhile, no apatite was observed on the surface of samples
A500 or 2A500, it might be related with the rapid degradation of
magnesium alloy substrates and excess release of magnesium ions
into m-SBF, which could inhibit the apatite formation [38]. However,
with the prolongation of immersion time, it cannot be neglect that
cracks on 3A500 could reach to the substrate and the coating is
Fig. 5.pH variation of the uncoated and coated samples immersed in m-SBF for different
periods.
158 Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B5%80 -
8/11/2019 YingDou
6/8
possible to detach from the substrate, so the substrate lost protec-
tiveness. Overall sample 3A500 here was evidenced to provide a rela-
tive long-term protection to AZ31 magnesium alloy substrates and
present a favorable anti-corrosion performance in a certain period.
The current results in this study revealed that the uniform and crack-
free 45S5 glassceramic coating with the thickness of 1.00 m could in-
crease the corrosion resistance of the AZ31 magnesium alloy in a certain
period and was believed to have potential application for protecting the
magnesium alloys.
4. Conclusions
In order to increase biocorrosion resistance of AZ31 magnesium alloy,
45S5 glass
ceramic coatings were produced on the magnesium alloy
Fig. 6.Surface morphologies of (a) uncoated AZ31 magnesium alloy and (b) 3A500 immersed for 1 day (the inset is the enlarged image) and (c) reveals EDS spectrum of area A in
(b); Micrographs of (d) A500 and (e) 3A500 immersed for 5 days, (f) 3A500 immersed for 7 days (the inset is the enlarged image of the deposition indicated by the white cycles)
and (g) reveals the EDS spectrum of area B in f.
159Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B6%80 -
8/11/2019 YingDou
7/8
substrates by dip-coating method from a synthesized solgel. The results
obtained in this work demonstrated that homogeneous and intact coat-
ings with the thickness of 0.481.00m, composed of amorphous phase
and Na2Ca2Si3O9, were fabricated in the condition of TEOS/H2O ratio
0.007, calcination temperature 500 C and dip-coating cycles 13.
The inuence of thecoatings on thecorrosion behavior wasstudiedby
in vitro immersion testsin m-SBF at 37 C for different periodsand the re-
sults suggested that the coated samples could offer a physical barrier
to inhibit the contact of AZ31 magnesium alloy substrate and m-SBF,
which improved the corrosion resistance. Immersion tests evidenced
Fig. 7.Cross-section FE-SEM micrographs and EDX microanalysis after immersion test: (a) A500 and (b) 3A500 after 1 day immersion, (c) A500 after 5 days immersion, (d) EDS
spectrum of area A in c, (e) 3A500 after 5 days immersion and (f) EDS spectrum of area B in e.
Fig. 8.Schematic illustration of the corrosion processes for A500 and 3A500 in m-SBF.
160 Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161
http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B8%80http://localhost/var/www/apps/conversion/tmp/scratch_7/image%20of%20Fig.%E0%B7%80 -
8/11/2019 YingDou
8/8
that cracks across the coating from surface to the substrate appeared for
samples A500 and 2A500 in the 7 days test, leading to the peeling off
the coating and serious corrosion of AZ31 magnesium alloy substrate.
While the mass loss of sample 3A500 (2.31%) was lower than those of
samples A500 (72.71%), 2A500 (72.24%) and AZ31 magnesium alloy sub-
strate(78.04%), along with a lower pH variation of m-SBF after 7 days
immersion in m-SBF, suggesting that sample 3A500 showed favorable
protectiveness to the magnesium alloy substrates in a certain period.
Meanwhile, calcium-de
cient hydroxyapatite deposited on the surfaceof 3A500 after 7 days immersion, indicating its biomineralization proper-
ty in m-SBF and potential applications in biomedical elds. Moreover,
crack formation in the coating during the immersion test is the main fac-
tor for 45S5 glassceramic coatings to fail in protecting the magnesium
alloy substrate from being corroded, and further works need to be done
to avoid cracking in the immersion test.
Acknowledgments
Authors acknowledge thenancial support by China NaturalScience
Foundation (Grant No. 51072129), Tianjin Natural Science Foundation
(Grant No. 11JCYBJC02600).
References
[1] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006) 1728.[2] N.T. Kirkland, Corros. Eng. Sci. Technol. 47 (2012) 322.[3] N.-E.L. Saris, E. Mervaala, H. Karppanen, J.A. Khawaja, A. Lewenstam, Clin. Chim.
Acta 294 (2000) 1.[4] P.J. Marie, R. Travers, E.E. Delvin, Calcif. Tissue Int. 35 (1983) 755.[5] F. Witte, V. Kaese, H. Haferkamp, E. Switzer, A. Meyer-Lindenberg, C.J. Wirth, H.
Windhagen, Biomaterials 26 (2005) 3557.[6] M.B. Kannan, R.K. Raman, Biomaterials 29 (2008) 2306.
[7] E. Zhang, L. Xu, G. Yu, F. Pan, K. Yang, J. Biomed. Mater. Res. A 90 (2009) 882.[8] H.F. Guo, M.Z. An, H.B. Huo, S. Xu, L.J. Wu, Appl. Surf. Sci. 252 (2006) 7911.[9] H.F. Guo, M.Z. An, Appl. Surf. Sci. 246 (2005) 229.
[10] Y. Song, S. Zhang, J. Li, C. Zhao, X. Zhang, Acta Biomater. 6 (2010) 1736.[11] C.-Y. Zhang, R.-C. Zeng, C.-L. Liu, J.-C. Gao, Surf. Coat. Technol. 204 (2010) 3636.[12] Z. Yao, Y. Xu, Y. Liu, D. Wang, Z. Jiang, F. Wang, J. Alloys Compd. 509 (2011) 8469.[13] A.J. Lpez, E. Otero, J. Rams, Surf. Coat. Technol. 205 (2010) 2375.[14] A.J. Lpez, J. Rams, A. Urea, Surf. Coat. Technol. 205 (2011) 4183.[15] S. Shadanbaz, G.J. Dias, Acta Biomater. 8 (2012) 20.[16] J.E. Gray, B. Luan, J. Alloys Compd. 336 (2002) 88.[17] P. Li, K.D. Groot, J. SolGel. Sci. Technol. 2 (1994) 797.
[18] S. Hofacker, M. Mechtel, M. Mager, H. Kraus, Prog. Org. Coat. 45 (2002) 159.[19] J. Liu, X. Miao, Ceram. Int. 30 (2004) 1781.[20] M. Fathi, A. Doostmohammadi, J. Mater. Process. Technol. 209 (2009) 1385.[21] W. Zhang, X. Chen, X. Liao, Z. Huang, X. Dan, G. Yin, J. Mater. Sci. - Mater. Med. 22
(2011) 2261.[22] Q.Z. Chen, Y. Li, L.Y. Jin, J.M. Quinn, P.A. Komesaroff, Acta Biomater. 6 (2010) 4143.[23] A. Oyane, H.M. Kim, T. Furuya, T. Kokubo, T. Miyazaki, T. Nakamura, J. Biomed.
Mater. Res. A 65A (2003) 188.[24] ASTM Standard G31-72, Standard Practice for Laboratory Immersion Corrosion
Testing of Metals, ASTM Standards, Philadelphia, PA, USA, 2004.[25] I. Holowacz, H. Podbielska, J. Bauer, A. Ulatowska-Jarza, Opt. Appl. 35 (2005) 691.[26] Y. Zhang, K. Bai, Z. Fu, C. Zhang, H. Zhou, L. Wang, S. Zhu, S. Guan, D. Li, J. Hu, Appl.
Surf. Sci. 258 (2012) 2939.[27] A. Roy, S.S. Singh, M.K. Datta, B. Lee, J. Ohodnicki, P.N. Kumta, Mater. Sci. Eng.
B-Adv. 176 (2011) 1679.[28] J.H. Hu, C.L. Zhang, B.H. Cui, K.F. Bai, S.K. Guan, L.G. Wang, S.J. Zhu, Appl. Surf. Sci.
257 (2011) 8772.[29] Z. Feng, Y. Liu, G.E. Thompson, P. Skeldon, Surf. Interface Anal. 42 (2010) 306.[30] V. Rajendran, A.N. Begum, M.A. Azooz, F.H. El Batal, Biomaterials 23 (2002) 4263.
[31] J. Qian, M. Xu, W. Zhang, J. Chin. Chem. Soc. 39 (2011) 1472.[32] L.P. Xu, A. Yamamoto, Colloids Surf., B 93 (2012) 67.[33] Y.Y. Zhu, Q. Zhao, Y.H. Zhang, G.M. Wu, Surf. Coat. Technol. 206 (2012) 2961.[34] S. Corkovic, R.W. Whatmore, Q. Zhang, J. Appl. Phys. 103 (2008) 084101.[35] J.X. Yang, Y.P. Jiao, F.Z. Cui, I.-S. Lee, Q.S. Yin, Y. Zhang, Surf. Coat. Technol. 202
(2008) 5733.[36] H. Hornberger, S. Virtanen, A.R. Boccaccini, Acta Biomater. 8 (2012) 2442.[37] X. Lu, Y. Leng, Biomaterials 26 (2005) 1097.[38] S. Cai, J.X. Li, G.H. Xu, X.D. Li, X.J. Ye, W. Jiang, Mater. Chem. Phys. 131 (2011) 462.
161Y. Dou et al. / Surface & Coatings Technology 228 (2013) 154161