Beruflich Dokumente
Kultur Dokumente
H I G H L I G H T S G R A P H I C A L A B S T R A C T
a r t i c l e i n f o a b s t r a c t
Article history: Corrosion behaviour of polymer coated Mg alloy screw is compared with a laboratory coupon with emphasis on
Received 23 May 2017 the adhesion of coating to surface features such as threads. The corrosion behaviour of poly (ε-caprolactone)
Received in revised form 7 September 2017 (PCL) coated commercial pure Mg (CP Mg) and Mg-4Zn-0.2Ca alloy in simulated body fluid (SBF) is investigated.
Accepted 18 September 2017
CP Mg with a high corrosion rate (8.426 mmyr−1) formed large amount of corrosion products on the surface
Available online 21 September 2017
which pushed out the PCL layer. In contrast the Mg-4Zn-0.2Ca alloy with a corrosion rate of 2.481 mmyr−1
Keywords:
held the PCL coating intact for up to 72 h. Fully and partially threaded screws of the Mg-4Zn-0.2Ca alloy were fab-
Biodegradable implant ricated and coated with PCL. It was found that pH variation of the medium depends on PCL adhesion to screw,
Magnesium alloys which was better in the threaded regions than in smoother areas. As a result, the fully threaded screws showed
Corrosion slower pH rise than partial threaded ones. The crests and troughs of the threads act as anchoring points for the
Poly (epsilon-caprolactone) polymer. It was found in the current study that the increased adhesion of PCL coating to threaded Mg alloy sub-
Simulated body fluid strate dominates over the increased corrosion rates of threaded surface.
Threaded screw © 2017 Elsevier Ltd. All rights reserved.
1. Introduction
Magnesium and its alloys degrade in the human body via an electro-
chemical reaction [1–3] and are good candidates for temporary ortho-
⁎ Corresponding author. paedic fixture devices [4]. Alloying with essential elements such as Ca,
E-mail address: krr@psgias.ac.in (K.R. Ravi). Zn and Mn increases biocompatibility, corrosion resistance and reduces
https://doi.org/10.1016/j.matdes.2017.09.043
0264-1275/© 2017 Elsevier Ltd. All rights reserved.
398 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
Fig. 1. Photographs of the partial and full thread screws used in the present study with dimensions.
magnetic stirring to ensure that the solution concentration and viscosity decided to continue the experiment with 2.5% PCL solution. Coupons in
are maintained. The coupons were removed from the solution at a con- which the PCL layer was non uniform or had defects were stripped of
trolled rate of 3 mm/s with a custom made set up, and dried in a jet of the coating and the process repeated from the disc polishing onwards.
cool air for 5 min. This cycle was repeated 10 times to obtain the final The surface morphology of the coated specimen was observed in
coating on the coupons. At this point there was a visible coating on ESEM. To further assess the effect of complex geometries and sharp fea-
the coupon and hence the process was stopped. It was observed that tures on the coating adhesion and corrosion, partially and completely
the 1% solution had a very runny nature and resulted in larger pores threaded screws were fabricated from the Mg-4Zn-0.2Ca alloy with di-
on drying, due to the evaporation of large amount of solvent compared mensions as specified in Fig. 1. The screws were degreased and protec-
to the solute. The 5% solution on the other hand resulted in a more com- tive PCL layer was applied by dip coating method.
pact coating with less pores, at the cost of non-uniformity of the coating.
The 5% solution was highly viscous and dip coating resulted in the accu- 2.2. Corrosion studies on laboratory coupons
mulation of polymer on different parts of the Mg alloy coupon resulting
in a very uneven coating. Based on this experimental observation it was Simulated body fluid (SBF) was prepared according to the protocol
of Kokubo [46] and the volume of SBF used for the static immersion
test was 50 times the surface area of the test specimen. The chemical
composition of the SBF was analyzed by flame spectrophotometer,
atomic absorption spectroscopy and titration method and is given in
Table 2. pH of the medium was monitored at regular intervals using a
Eutech pH 30 meter. The samples were washed in distilled water,
dried and weighed after the corrosion test. An average of three mea-
surements was taken for each group. Percentage change in mass at dif-
ferent times is calculated using the equation:
Table 3
Average surface roughness of the specimen.
CP Mg 0.0543 ± 0.0075
Fig. 2. a) Coarse columnar microstructure of CP Mg and (b) finer equiaxed microstructure
Mg-4Zn-0.2Ca 0.0533 ± 0.0043
of Mg-4Zn-0.2Ca alloy, revealed on etching with acetic picral.
400 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
water and drying in warm air. The combined effect of mass loss due to
corrosion and mass gain due to the corrosion products is reported as a
single parameter. In addition to this, PCL when immersed in aqueous
medium tends to show an initial mass gain due to hydrolysis followed
by a reduction in mass as the polymer chain breakage occurs. These in-
dividual contributions due to the polymer, metal/alloy and corrosion
products are difficult to isolate in these experiments. The percentage
change in mass is reported as an indicator of the corrosion behaviour
which has to be understood as supplementary information to imped-
ance data. pH measurements of the simulated body fluid on progress
of corrosion are also used as an indicator of the corrosion rates in
some cases as appropriate.
Fig. 3. SEM image of (a) CP Mg, (b) Mg-4Zn-0.2Ca alloy, c) EDS analysis of different phases
in Mg-4Zn-0.2Ca alloy and (d) PCL coated Mg-4Zn-0.2Ca alloy before corrosion. Fig. 5. Percentage change in mass of specimens and corresponding pH of simulated body
fluid over 72 h of coated CP Mg and Mg-4Zn-0.2Ca alloy.
D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410 401
Fig. 7. XRD plots of (a) CP Mg and (b) Mg-4Zn-0.2Ca alloy, before and after corrosion in
simulated body fluid for 72 h.
velocity of 0.01 mm/min on both flat and corrugated samples held to-
gether with the polymer layer in between. The flat specimens of CP
Mg and alloy of 4 mm × 8 mm area and length of 40 mm were tested.
The corrugated samples of Mg-4Zn-0.2Ca alloy of width 30 mm and
Fig. 6. SEM image of surface of (a) CP Mg (b) bare Mg-4Zn-0.2Ca alloy and (c) PCL coated thickness of 6 mm had 8 matching teeth of 3.5 mm height each. The sur-
Mg-4Zn-0.2Ca alloy after 72 h immersion in simulated body fluid. faces were polished and cleaned, and a layer of PCL was added to hold
the mating pieces in a C-clamp tightly and left to dry for 8 h. The force
impedance data was fitted with equivalent circuit parameters to model required to peel off the polymer and pull the two metallic pieces apart
the corrosion behaviour. Potentiodynamic polarization scans were car- is estimated from the stress-elongation curve as the point where the
ried out over a range of ± 500 mV with respect to the OCP at a scan stress curve starts descend. An average of two readings is presented
rate of 10 mV s−1. The data for three specimens of each sample group for CP Mg and Mg-4Zn-0.2Ca alloy.
were fitted with the Tafel approximation to the Butler-Volmer equation,
and the corrosion parameters calculated. The results are tabulated and 2.5. Corrosion studies on screw specimen
average values calculated. One set of curves alone is shown for clarity
in the results, and the data for the other two in each set is presented The static immersion studies were conducted for the coated and un-
in tabular format. coated Mg-4Zn-0.2Ca alloy screws in simulated body fluid for 336 h and
the variation in pH was monitored as an indicator of the corrosion rate
2.4. Adhesion tests of the specimens. A corrosion rate calculation based on mass loss of
the specimen was not possible in these experiments since it is difficult
Adhesion tests of polymer on the metallic sample were carried out to isolate the individual contributions due to the degradation of the
with a Universal Testing Machine (Instron 3366, USA) at a cross head polymer and alloy and deposition of corrosion products on the surface
402 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
and ternary phases respectively. The EDS analyses of these phases show
the presence of Mg and Zn in the spherical phases and Mg, Zn and Ca in
the elongated phase. The percentage of individual elements in these
phases is shown in the table adjacent to the EDS spectrum, these are
only indicative of the presence of these elements and not a measure of
the exact composition of these phases. These phases are possibly
MgZn and Ca2Mg6Zn3.
The SEM micrograph of the polymer coating formed on the Mg-4Zn-
0.2Ca alloy before corrosion is shown in Fig. 3(d) and it shows porous
surface. Analysis using ImageJ software shows that the surface has a po-
rosity of approximately 25.6% with an average pore size of 5.23 ± 0.8
μm for the smaller pores. The average pore size considering the larger
pores in the boundaries of each polymer grain is 7.24 ± 7.4 μm, which
is due to the fact that the size difference between the two types of
pores is very large. The pores on the surface of the polymer coated spec-
imens are due to the evaporation of solvent from the polymer during
dip coating and drying. The larger pores are possibly due to the shrink-
age of the polymer layer over the substrate upon drying that pulls in the
polymer chains to minimize surface area. The SEM image indicates that
there is polymer layer beneath the larger pores possibly from the earlier
layers of dip coated PCL.
Fig. 9. (a) Electrochemical impedance with (b) equivalent circuit representation of corrosion phenomenon and (c) potentiodynamic polarization plots of CP Mg, Mg-4Zn-0.2Ca alloy and
PCL coated Mg-4Zn-0.2Ca specimens in simulated body fluid.
bioactivity of the material. The absence of such phosphate layer in case 3.3. Electrochemical studies
of CP Mg and uncoated Mg-4Zn-0.2Ca alloy could be due to the flaking
off of the products in these samples and/or due to the amount of such The representative Nyquist curves (Fig. 9(a)) of the CP Mg, Mg-4Zn-
products on the surface being less than the detectable limit of XRD. 0.2Ca alloy and PCL coated Mg-4Zn-0.2Ca alloy indicate the presence of
These products are held in tact by the polymer layer and are easily de- capacitive loops at high and mid frequencies for all three specimens. In
tectable in case of the PCL coated specimen. addition there is an inductive loop at low frequency for the polymer
Table 4
Equivalent circuit analysis for electrochemical impedance.
Specimen R1 (Ω) R2 (Ω) L2 (H) C2 (F) R3 (Ω) C3 (F) R4 (Ω) C4 (F) L4 (H) W4 (Ω−1/2)
−6 −6 −6 a a a a
Mg_1 2.99 69.4 30.37e 10.37e 51.6 14.59e
Mg_2 0.68 78.6 27.65e−6 18.96e−6 49.6 12.3e−6 a a a a
PCL_1 0.05 82.06 67.0e−6 13.8e−9 164.7 3.97e−6 624.8 2.077 38.98e−6 4193
PCL_2 1.3 89 56.7e−6 12.98e−9 162 3.58e−6 589 1.05 15.6e−6 5100
PCL_3 0.5 95.6 53.62e−6 11.8e−9 159 4.99e−6 593.8 2.45 36.54e−6 3856
a
Indicates that the parameter was not used in the equivalent circuit model.
404 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
Table 5
Corrosion parameters calculated from polarization plots.
Specimen Ecorr (V, Vs Ag/AgCl) Icorr (μAcm−2) βc (mV decade−1) βa (mVdecade−1) Corrosion rate (mmpy) Rp (Ω cm2) (fitted value) R⁎p ( Ω cm2) (calculated)
Rp (Ω cm2) (fitted value): Fitted for default values of βa and βc of 120 mV, and E = ±25 mV using EC-Lab® software V10.44 (Bio-Logic-Science Instruments)
R⁎p (Ω cm2) (calculated): Calculated using the equation: Rp;E ¼
corr
ðβa βc Þ
Icorr ðβa þβc Þ ln 10
coated specimen. The diameter of the high and mid frequency loops are calculated at E = Ecorr with the “Rp fit” feature in the software and
indicative of the surface inhomogeneity and charge transfer resistance also with Eq. (3).
of the porous layer on the surface. For the bare alloy this contribution
is mainly due to the surface corrosion products, and for the coated ðβa βc Þ
Rp;Ecorr ¼ ð3Þ
specimen, this is due to the polymer coating itself. The polarization Icorr ðβa þ βc Þ ln 10
resistance (Rp) of the bare alloy is calculated as the sum of the charge
transfer resistance and layer resistance, and is the x-intercept of the The results are summarized in Table 6 and Fig. 10(a–b) and clearly
Nyquist graph at low frequency (ideally at dc voltage or frequency = shows that the corrosion potential (Ecorr) shifts to more positive poten-
0 Hz). The polarization resistance of the PCL coated Mg-4Zn-0.2Ca tial in the order PCL coated Mg-4Zn-0.2Ca N Mg-4Zn-0.2Ca N CP Mg as
specimen is calculated to be 898.74 Ω which is four times larger expected. At the same time, the values of the corrosion current (Icorr)
than that of the bare alloy (223.36 Ω). A mechanism involving pene- of the polymer-coated sample is less than that of the bare Mg-4Zn-
tration of the simulated body fluid through the polymer layer that 0.2Ca alloy, which in turn is less than that of CP Mg. The linear polariza-
causes pitting in the underlying alloy is proposed in the discussion tion resistance increases in the order of CP Mg b Mg-4Zn-0.2Ca alloy
section of this paper. b PCL coated Mg-4Zn-0.2Ca alloy. These results corroborate the finding
The equivalent circuit parameters for the samples are fitted with EC- that the PCL coating enhances the corrosion resistance of magnesium
Lab® software V10.44 (Bio-Logic-Science Instruments) and are summa- alloy in this study.
rized in Table 4 for the circuit given in Fig. 9(b)). R1 is the solution resis-
tance of the medium, R2 and C2 indicate the resistance and charge effect 3.4. Adhesion tests
offered by the inhomogenities in the porous layer of corrosion products
on the surface and inductance L2 is used to model the pores in the layer. The representative stress-extension curves obtained from the uniax-
The medium frequency loop can be modelled with the parallel combina- ial loading of the samples tested for adhesive strength are plotted in
tion of R3 and C3, which represents the charge transfer resistance of the Fig. 11(a) and the values of stress at polymer peel off are tabulated in
material itself. For the Mg-4Zn-0.2Ca alloy, it was found that an addi- Table 7. The stress at peel off for the CP Mg and Mg-4Zn-0.2Ca alloy
tional time constant is involved possibly due to the relaxation of mass are almost comparable. This is expected since the adhesion of a polymer
transport through the layer of corrosion product and is modelled by to a metal depends mainly on the surface roughness of the metal, its
an additional R3 and C3 parameter. It was not possible to achieve an cleanliness and surface pre-treatments. Since both the materials were
exact fit for the curve especially below a frequency of 1 Hz with this subject to the same sample preparation procedure, the adhesive
model. For the PCL coated samples, the diffusion through the porous strengths are similar. Slight increase in value of adhesive strength for
polymer layer and the pitting below the underlying corrosion layer is the CP Mg samples could be attributed to the higher surface roughness
modelled with the combination of inductance L4 and Warburg imped- as presented in Table 3.
ance W4, along with the charge transfer resistance R4 and dielectric ef- The adhesion tests further prove that the tendency of the coating on
fect of the polymer, C4. CP Mg to peel off comes from the inherently faster corrosion rates of CP
The corrosion parameters based on Tafel equation (Eq. (2)) for the Mg as compared to the Mg-4Zn-0.2Ca alloy in spite of similar adhesive
CP Mg, Mg-4Zn-0.2Ca alloy and PCL coated Mg-4Zn-0.2Ca alloy calculat- strengths in the dry state. The significant difference in the adhesion in
ed from the polarization curves (Fig. 9(c)) are summarized in Table 5. immersed state occurs due to the faster H2 evolution and formation of
E−Ecorr E−Ecorr
corrosion products as a result of higher corrosion rate of CP Mg that re-
I ¼ Icorr 10 βa −10− βa ð2Þ sults in polymer layer peel off.
The stress required for polymer peel off in corrugated Mg-4Zn-0.2Ca
where, Ecorr is the corrosion potential, Icorr the corrosion current, βa and alloy sample is higher than that of the flat samples by approximately
βc are the anodic and cathodic Tafel coefficients for the anodic and ca- 17% (Fig. 11(d)) since the crests and troughs on the surface
thodic reactions respectively. The polarization resistance Rp is (Fig. 11(c)) offer a better adhesion than the flat surface (Fig. 11(b))
Table 6
Average values of corrosion parameters calculated from polarization plots.
Specimen Ecorr (V, Vs Ag/AgCl) Icorr (μA cm−2) Corrosion rate (mmpy) Rp (Ω cm2)
Table 7
Adhesion strength of PCL to metallic surface tested by uniaxial loading.
CP Mg 92.5 ± 4.9
Mg-4Zn-0.2Ca 86.0 ± 7
Mg-4Zn-0.2Ca corrugated 105.0 ± 3.4
was observed that the pH of the medium increased in all samples, the
increase was highest in case of the uncoated screws of both partially
and fully threaded design. In case of coated screws, the partially
threaded configuration demonstrated higher corrosion rates compared
to the fully threaded one.
The cross sectional image of the PCL coated screw in the threaded re-
gion (Fig. 13) shows a layer of 126.5 ± 10.2 μm thickness over the alloy.
The SEM of the cross section of the coated screw after 336 h of immer-
sion in simulated body fluid is presented in Fig. 14 and it is interesting to
note that the PCL layer is still attached at some point, in spite of the
bulky layer of corrosion products of approximately 190 μm. The PCL
layer was too thin to get any useful information in the EDS analysis,
only peaks of Au sputtered on it for conductivity in SEM imaging was
obtained. The alloy showed peaks of Mg and O, whereas the corrosion
products consisted mainly of Mg, Ca and O.
4. Discussion
Fig. 11. (a) Uniaxial tensile stress vs. extension of polymer bonded metallic specimens, (b) photographic image of flat specimen, (c) photographic image of corrugated specimen and
(d) comparison of adhesive strengths calculated based on uniaxial loading tests.
406 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
Fig. 12. (a) pH variation of medium during corrosion of coated and uncoated screw specimens in simulated body fluid (b) threaded portion of uncoated screws are corroded more after
336 h (c) comparison of uncoated and coated partially threaded screws and (d) comparison of uncoated and coated fully threaded screws on immersion in simulated body fluid for 336 h.
polymer is no longer there. This effect manifests initially at sharp cor- not only on the absence of porosity in the coating but also on the adhe-
ners and edges of the coated specimen, and hence the layer peels off sion of polymer to the alloy surface. This is further demonstrated in the
from the edges. From this event forth, there is an increase in pH of the current work on threaded screws undergoing corrosion in simulated
coated CP Mg specimen. The mass change is slower since it is the com- body fluid for a period of 336 h.
bined effect of polymer swelling/degradation, corrosion mass loss of the The corrosion of polymer coated specimen is a complex combination
metal and deposition of corrosion products (Fig. 5). of one or more of these processes: Polymers such as PCL undergo both
In case of coated Mg-4Zn-0.2Ca alloy, the corrosion rate is slower, and surface and bulk degradation in simulated body conditions [37]. A sche-
hence the underlying metallic layer does not degrade at a much faster matic representation of the mechanism of corrosion of PCL coated Mg
rate. The slower corrosion, sluggish pH change, lesser precipitates and re- alloy shown in Fig. 16(a) clearly indicates these different aspects. If
duced H2 evolution from the surface enables the polymer layer to cling on the rate of hydrolytic cleavage of surface bonds occurs faster than the
longer to the alloy. There is also mass gain in case of the polymer coated penetration of liquid into the polymer bulk, surface degradation or ero-
alloy that can be a combined effect of swelling of the polymer and the cor- sion occurs, and the coating material becomes thinner. The reverse is
rosion products. Photograph of the specimen after exposure to simulated true in bulk degradation, where the rapid penetration of water into
body fluid for 72 h are seen in Fig. 15(c) which supports this proposition.
Surface analysis of the coated specimen after corrosion using SEM
shows swelling in the polymer due to hydrolysis and the pores have be-
come more rounded with visibly thinner edges and opened up as seen
in Fig. 6(c) and inset of Fig. 14. This will further increase the exposure
of the alloy surface to the simulated body fluid. It has to be considered
that the volume of the medium seen by the alloy surface is small and
there is also a lesser chance of the ions released as part of the corrosion
to diffuse into the bulk of the medium due to the barrier offered by the
polymer layer. Hence there is a stagnation zone between the polymer
and the alloy with a higher pH medium due to the corrosion of the mag-
nesium alloy. This high alkalinity can result in the precipitation of the
corrosion products, mainly hydroxides and phosphates of Mg and Ca.
It is posited that these products along with hydrogen gas evolved
from the magnesium corrosion result in the bulging of the PCL layer
[20,29,35] away from the surface of the Mg-4Zn-0.2Ca alloy eventually
peeling off the layer. Park et al. [35], proposed the use of molten poly-
mer as an alternative to the solvent based method to avoid the forma-
tion of pores in the coating. It can be safely assumed that the long Fig. 13. Cross sectional SEM of the PCL coating on Mg-4Zn-0.2ca alloy screw before
term presence of polymer coating on the alloy surface is dependent corrosion.
D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410 407
Fig. 14. Cross sectional SEM of coated Mg-4Zn-0.2ca alloy screw after 336 h corrosion with EDS analysis.
the entire bulk of the polymer results in random cleavage of ester bonds similar in the underling alloy material and the bare alloy. The mid fre-
within the bulk. The increase in the number of carboxyl end groups quency time constant is associated with the charge transfer resistance
(\\COOH) [48] formed as a product of the ester bond cleavage in the Rct of the material and the double-layer capacitance, Cdl, at the interface
polymer bulk increases the acidic content within (Fig. 16(b)). Diffusion of the coating and base alloy and is maximum for the PCL coated speci-
of surface carboxyl groups (\\COOH) into the surrounding medium, re- men, since the contribution is mainly from the polymer layer itself. In
sults in the degradation of the material from inside out. It is difficult to the CP Mg and Mg-4Zn-0.2Ca alloy, the partially protective surface corro-
isolate these two modes of degradation occurring in the PCL coating sion layer offers only a small resistance to the corrosion process. The size
on Mg specimen in our study. The degradation of the polymer layer re- of the capacitive loops at mid frequency can be taken as a representation
sults in the penetration of the simulated body fluid to the metallic sur- of the corrosion resistance of the uncoated and coated specimen in simu-
face and its subsequent corrosion. The acidic pH resulting from the lated body fluid. There appears to be a third capacitive loop in the low fre-
polymer degradation itself may affect the corrosion rate of the substrate. quency region which could be indicative of the relaxation of mass
The pH variation of the medium is the effect of the degradation of poly- transport through the corrosion product layer or polymer layer for the
mer and that of the alloy. It is difficult to isolate these two parameters in coated specimen, a supporting evidence for the partially protective nature
a coated sample. Due to the thin layer of PCL and higher corrosion rate of of the coating or product layer. These multiple mass and charge transfer
the Mg alloy itself, the effect of metal corrosion will be dominant. The processes occur simultaneously and interdependently in the material.
pH of uncoated alloy varies from 7.4 to 9 whereas that of the coated The low frequency inductive loop in the coated Mg-4Zn-0.2Ca alloy
alloy is below 8. This stable value of pH indicates that the coating is in- specimen indicates that there is possible pitting corrosion that occurred
tact. However, the presence of corrosion products within the layer indi- below the polymer layer. This is possible due to the penetration of the
cates that corrosion has occurred in the interface. simulated body fluid through the pores in the PCL layer, and subsequent
corrosion of the alloy in a highly concentrated medium confined to a
4.2. Electrochemical studies small space. Though pitting corrosion is not seen in the uncoated Mg-
4Zn-0.2Ca alloy, the chances of its occurrence in the coated specimen
Electrochemical impedance spectroscopy analysis is a steady state is higher due to the localised exposure of the specimen surface to the
technique that provides valuable insight into the mechanism of corro- simulated body fluid that penetrates the pores of the polymer coating.
sion of coated Mg-4Zn-0.2Ca alloy. The two capacitive time constants The low volume of chloride ion rich SBF medium trapped in the inter-
indicate that there are two rate controlling processes that occur in this face of the polymer and alloy surfaces cause pitting in coated specimen.
case due to the material and a partially protective layer on the surface.
The high frequency loop could be indicative of the surface non homoge- 4.3. Corrosion of screw specimen
neity of the alloy and is almost similar for the CP Mg and coated/uncoat-
ed Mg-4Zn-0.2Ca alloy. This loop includes no contribution from the An observation from the photographs in Fig. 12(c–d) is that the ad-
polymer layer and it means that the corrosion mechanism is almost hesion of the coating is better in the fully threaded specimen (specimen
Fig. 15. Photographic image of (a) coated specimen, (b) Mg specimen with coating peeled off with dense layer of corrosion products on its surface and (c) Mg-4Zn-0.2Ca coated specimen is
intact after 72 h of immersion test, localised corrosion products visible in the substrate.
408 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
Fig. 16. (a) Degradation mechanism of PCL coated Mg alloy in simulated body fluid, (b) cleavage of ester bonds and formation of carboxyl end groups.
3 in Fig. 12(d)), on account of the enhanced roughness due to the ma- portion of the corroding alloy is now exposed to a larger volume of
chined threads that act as anchoring points for the coating. The smooth the medium, the effect of local rise in pH that happens usually under
portion of the partially threaded screw has a higher tendency of corro- the coating is absent here and therefore lesser corrosion products are
sion due to the smoother and plain geometry (specimen 3 in observed in this case.
Fig. 12(c)), in which the medium can penetrate easily and peel off the There are however unresolved issues at this point not addressed in
coating. In contrast the coating in the threaded area stays intact longer. this work. It is expected that under in vivo conditions, since there is a
For the threaded specimen, the increased number of crests and troughs steady pH maintained due to the replenishment of the tissue fluid in
also acts as points of steadfast contact of the coating with the metal over the human body and diffusion of released ions into the tissue, the corro-
its entirety. Hence the exposure of the specimen at any one thread is iso- sion behaviour is different from the stagnant tests carried out in vitro.
lated from the rest of the length and the overall corrosion rate is much However, the high pH at the interface of the alloy and polymer coating
slower than the coated partially threaded screw. Also since a small can occur under such circumstances also, with a chance of large amount
D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410 409
of corrosion products precipitated between the polymer and alloy. [2] S. Agarwal, J. Curtin, B. Duffy, S. Jaiswal, Biodegradable magnesium alloys for ortho-
paedic applications: a review on corrosion, biocompatibility and surface modifica-
Wong et al. [40] demonstrated that a PCL coated AZ91 alloy showed tions, Mater. Sci. Eng. C (2016)https://doi.org/10.1016/j.msec.2016.06.020.
no toxicity in vitro and cell attachment occurred in 3 days. The specimen [3] G. Song, A. Atrens, Recent insights into the mechanism of magnesium corrosion and
implanted in rabbit model on removal after 2 months of implantation research suggestions, Adv. Eng. Mater. 9 (2007) 177–183, https://doi.org/10.1002/
adem.200600221.
showed signs of bone tissue formation. However no mention was [4] F. Witte, The history of biodegradable magnesium implants: a review, Acta
made of corrosion products on the interface of AZ91 alloy and PCL coat- Biomater. 6 (2010) 1680–1692, https://doi.org/10.1016/j.actbio.2010.02.028.
ing. Additional influencing parameters occur when the material is im- [5] X.N. Gu, Y.F. Zheng, A review on magnesium alloys as biodegradable materials,
Front. Mater. Sci. Chin. 4 (2010) 111–115, https://doi.org/10.1007/s11706-010-
planted in an animal model and these are considered beyond the 0024-1.
scope of the current work and needs further evaluation. [6] Y. Ding, C. Wen, P. Hodgson, Y. Li, Effects of alloying elements on the corrosion be-
Based on these observations, it is the authors' opinion that the re- havior and biocompatibility of biodegradable magnesium alloys: a review, J.
Mater. Chem. B 2 (2014) 1912–1933, https://doi.org/10.1039/C3TB21746A.
sults obtained from laboratory scale experiments on coated corrosion
[7] Y. Sun, B. Zhang, Y. Wang, L. Geng, X. Jiao, Preparation and characterization of a new
coupons of simple geometry cannot be extended to the final product. biomedical Mg-Zn-Ca alloy, Mater. Des. 34 (2012) 58–64, https://doi.org/10.1016/j.
In an uncoated Mg alloy screw, the sharp geometrical features such as matdes.2011.07.058.
thread crests are stress raisers and have higher tendency to corrode. [8] G. Song, Recent progress in corrosion and protection of magnesium alloys, Adv. Eng.
Mater. 7 (2005) 563–586, https://doi.org/10.1002/adem.200500013.
Hence in an uncoated partially threaded geometry, the smooth surface [9] X. Liu, J. Sun, F. Zhou, Y. Yang, R. Chang, K. Qiu, Z. Pu, L. Li, Y. Zheng, Micro-alloying
will corrode slower than the threaded portion. This scenario changes with Mn in Zn-Mg alloy for future biodegradable metals application, Mater. Des. 94
when a polymer coating is applied as is demonstrated in the current (2016) 95–104, https://doi.org/10.1016/j.matdes.2015.12.128.
[10] W. Zhang, M. Li, Q. Chen, W. Hu, W. Zhang, W. Xin, Effects of Sr and Sn on micro-
work. The polymer coating tends to peel off faster from a smooth surface structure and corrosion resistance of Mg – Zr – Ca magnesium alloy for biomedical
than a threaded one, and hence a reversal of scenario occurs. The applications, Mater. Des. 39 (2012) 379–383, https://doi.org/10.1016/j.matdes.
smoother area under the polymer layer now corrodes faster than the 2012.03.006.
[11] G. Ben Hamu, D. Eliezer, L. Wagner, The relation between severe plastic deformation
threaded area under the polymer. Thus our experiments in this work microstructure and corrosion behavior of AZ31 magnesium alloy, J. Alloys Compd.
show that the sharp edges of the threads in spite being at higher risk 468 (2009) 222–229, https://doi.org/10.1016/j.jallcom.2008.01.084.
for corrosion (Fig. 12(b)) also anchored the polymer coating better [12] H.R. Bakhsheshi-Rad, E. Hamzah, M. Medraj, M.H. Idris, A.F. Lotfabadi, M.
Daroonparvar, M.A.M. Yajid, Effect of heat treatment on the microstructure and cor-
(specimen 3 of Fig. 12(d)) and served to reduce the degradation of the
rosion behaviour of Mg-Zn alloys, Mater. Corros. 65 (2014) 999–1006, https://doi.
coated fully threaded screw. The result is a slower rise in the pH of the org/10.1002/maco.201307492.
medium compared to the bare specimens. Within a component, [13] J.R. Smith, D.A. Lamprou, Polymer coatings for biomedical applications: a re-
view, Trans. IMF. 92 (2014) 9–19, https://doi.org/10.1179/0020296713Z.
threaded and smooth portions corrode at dissimilar rates and affect
000000000157.
the polymer attachment differently as seen from the partially threaded [14] S. Abela, Protective coatings for magnesium alloys, in: F. Czerwinski (Ed.), Magnes.
screw. Hence the effect of bulk and surface geometrical features of the Alloy. Surf. Treat 2011, p. 344.
end use component on the attachment of the coating has to be individ- [15] S. Ehtemam-Haghighi, K.G. Prashanth, H. Attar, A.K. Chaubey, G.H. Cao, L.C. Zhang,
Evaluation of mechanical and wear properties of Ti\\xNb\\7Fe alloys designed for
ually investigated in each component designed for implant applications. biomedical applications, Mater. Des. 111 (2016) 592–599, https://doi.org/10.1016/
j.matdes.2016.09.029.
[16] M. Geetha, A.K. Singh, R. Asokamani, A.K. Gogia, Ti based biomaterials, the ultimate
choice for orthopaedic implants – a review, Prog. Mater. Sci. 54 (2009) 397–425,
5. Conclusion https://doi.org/10.1016/j.pmatsci.2008.06.004.
[17] X. Liu, P.K. Chu, C. Ding, Surface modification of titanium, titanium alloys, and relat-
Corrosion rate of coated Mg-4Zn-0.2Ca alloy decreased from 4.52 to ed materials for biomedical applications, Mater. Sci. Eng. R. Rep. 47 (2004) 49–121,
https://doi.org/10.1016/j.mser.2004.11.001.
0.269 mmpy on application of PCL coating. Infiltration of the simulated [18] H. Attar, S. Ehtemam-haghighi, D. Kent, I.V. Okulov, H. Wendrock, M. B., A.S. Volegov,
body fluid through the pores of the polymer and its bulk degradation re- M. Calin, J. Eckert, M.S. Dargusch, Nanoindentation and wear properties of Ti and Ti-
sult in corrosion of the substrate. Uniaxial adhesion strength of PCL on TiB composite materials produced by selective laser melting, Mater. Sci. Eng. A 688
(2017) 20–26, https://doi.org/10.1016/j.msea.2017.01.096.
CP Mg and Mg-4Zn-0.2Ca are 92.5 ± 4.9 and 86.0 ± 7 kPa respectively
[19] M. Vert, Degradable and bioresorbable polymers in surgery and in pharmacology :
in dry state, but coating on CP Mg peels off faster in simulated body fluid beliefs and facts, J. Mater. Sci. Mater. Med. 20 (2009) 437–446, https://doi.org/10.
due to higher corrosion rate of substrate. Pitting corrosion is also ob- 1007/s10856-008-3581-4.
[20] Y. Chen, Y. Song, S. Zhang, J. Li, C. Zhao, X. Zhang, Interaction between a high purity
served under the polymer layer. A fully threaded screw fabricated
magnesium surface and PCL and PLA coatings during dynamic degradation, Biomed.
from the alloy and coated with PCL shows slower pH rise in the medium Mater. 25005 (2011)https://doi.org/10.1088/1748–6041/6/2/025005.
than a partially threaded screw of same major dimensions when tested [21] P. Tian, D. Xu, X. Liu, Mussel-inspired functionalization of PEO/PCL composite coat-
in simulated body fluid for a period of 336 h. Partial threaded configura- ing on a biodegradable AZ31 magnesium alloy, Colloids Surf. B: Biointerfaces
(2016)https://doi.org/10.1016/j.colsurfb.2016.02.004.
tion results in peel off of PCL coating in smooth regions and rapid in- [22] H.R. Bakhsheshi-Rad, E. Hamzah, A.F. Ismail, M. Daroonparvar, M.A.M. Yajid, M.
crease in pH. It is found in the current study that increased adhesion Medraj, Preparation and characterization of NiCrAlY/nano-YSZ/PCL composite coat-
of PCL coating to threaded Mg alloy substrate dominates over the in- ings obtained by combination of atmospheric plasma spraying and dip coating on
Mg-Ca alloy, J. Alloys Compd. 658 (2016) 440–452, https://doi.org/10.1016/j.
creased corrosion rates of threaded surface. jallcom.2015.10.196.
[23] L.H. Li, T.S.N. Sankara Narayanan, Y.K. Kim, Y.M. Kong, I.S. Park, T.S. Bae, M.H. Lee, De-
position of microarc oxidation-polycaprolactone duplex coating to improve the cor-
rosion resistance of magnesium for biodegradable implants, Thin Solid Films 562
Acknowledgements (2014) 561–567, https://doi.org/10.1016/j.tsf.2014.04.004.
[24] J. Wang, Y. He, M.F. Maitz, B. Collins, K. Xiong, L. Guo, Y. Yun, G. Wan, N. Huang, A
The authors would like to acknowledge funding support for part of surface-eroding poly(1,3-trimethylene carbonate) coating for fully biodegradable
magnesium-based stent applications: toward better biofunction, biodegradation
this work by the Department of Science and Technology, Government
and biocompatibility, Acta Biomater. 9 (2013) 8678–8689, https://doi.org/10.
of India [Project No. SR/WOS-A/ET-79/2013(G), dated 29.5.2014]. 1016/j.actbio.2013.02.041.
The authors would also like to thank Mr. G. Sathish kumar (Adhesion [25] A. Zomorodian, M.P. Garcia, T. Moura, J.C.S. Fernandes, M.H. Fernandes, M.F.
Tests), Ms. Anju Gupta, Dr. N. Meenakshi Sundaram, Dr. R. Montemor, Biofunctional composite coating architectures based on
polycaprolactone and nanohydroxyapatite for controlled corrosion activity and en-
Sivasubramanian and Dr. V. Udhayabanu for valuable discussion at var- hanced biocompatibility of magnesium AZ31 alloy, Mater. Sci. Eng. C 48 (2015)
ious stages of this study. 434–443, https://doi.org/10.1016/j.msec.2014.12.027.
[26] X.N. Gu, Y.F. Zheng, Q.X. Lan, Y. Cheng, Z.X. Zhang, T.F. Xi, D.Y. Zhang, Surface mod-
ification of an Mg-1Ca alloy to slow down its biocorrosion by chitosan, Biomed.
References Mater. 4 (2009) 44109, https://doi.org/10.1088/1748-6041/4/4/044109.
[27] C. Wu, Z. Wen, C. Dai, Y. Lu, F. Yang, Fabrication of calcium phosphate/chitosan coat-
[1] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Magnesium and its alloys as orthopedic ings on AZ91D magnesium alloy with a novel method, Surf. Coat. Technol. 204
biomaterials: a review, Biomaterials 27 (2006) 1728–1734, https://doi.org/10.1016/ (2010) 3336–3347, https://doi.org/10.1016/j.surfcoat.2010.03.045.
j.biomaterials.2005.10.003.
410 D.B. Prabhu et al. / Materials and Design 135 (2017) 397–410
[28] J. Ma, M. Thompson, N. Zhao, D. Zhu, ScienceDirect similarities and differences in hydroxyapatite composite scaffolds for bone tissue engineering, Mater. Des. 114
coatings for magnesium-based stents and orthopaedic implants, J. Orthop. Transl. (2017) 149–160, https://doi.org/10.1016/j.matdes.2016.10.054.
2 (2014) 118–130, https://doi.org/10.1016/j.jot.2014.03.004. [40] H.M. Wong, K.W.K. Yeung, O.K. Lam, V. Tam, P.K. Chu, K.D.K. Luk, K.M.C. Cheung, A
[29] A. Alabbasi, S. Liyanaarachchi, M.B. Kannan, Polylactic acid coating on a biodegrad- biodegradable polymer-based coating to control the performance of magnesium
able magnesium alloy: an in vitro degradation study by electrochemical impedance alloy orthopaedic implants, Biomaterials 31 (2010) 2084–2096, https://doi.org/10.
spectroscopy, Thin Solid Films 520 (2012) 6841–6844, https://doi.org/10.1016/j.tsf. 1016/j.biomaterials.2009.11.111.
2012.07.090. [41] Z. Li, X. Gu, S. Lou, Y. Zheng, The development of binary Mg-Ca alloys for use as bio-
[30] S.C. Woodward, P.S. Brewer, F. Moatamed, A. Schindler, C.G. Pitt, The intracellular degradable materials within bone, Biomaterials 29 (2008) 1329–1344, https://doi.
degradation of poly(e-caprolactone), J. Biomed. Mater. Res. 19 (1985) 437–444, org/10.1016/j.biomaterials.2007.12.021.
https://doi.org/10.1002/jbm.820190408. [42] U. Thormann, V. Alt, L. Heimann, C. Gasquere, C. Heiss, G. Szalay, J. Franke, R.
[31] J.N. Li, P. Cao, X.N. Zhang, S.X. Zhang, Y.H. He, In vitro degradation and cell attach- Schnettler, K.S. Lips, The biocompatibility of degradable magnesium interference
ment of a PLGA coated biodegradable Mg – 6Zn based alloy, J. Mater. Sci. 45 screws: an experimental study with sheep, Biomed. Res. Int. (2015) https://doi.
(2010) 6038–6045, https://doi.org/10.1007/s10853-010-4688-9. org/10.1155/2015/943603.
[32] X. Xu, P. Lu, M. Guo, M. Fang, Cross-linked gelatin/nanoparticles composite coating [43] D. Bormann, A. Lucas, B. Denkena, C. Hackenbroich, A. Meyer-lindenberg, Influence
on micro-arc oxidation film for corrosion and drug release, Appl. Surf. Sci. 256 of different surface machining treatments of magnesium-based resorbable implants
(2010) 2367–2371, https://doi.org/10.1016/j.apsusc.2009.10.069. on the degradation behavior in rabbits, Adv. Eng. Mater. 11 (2009) 47–54, https://
[33] W.F. Ng, M.H. Wong, F.T. Cheng, Stearic acid coating on magnesium for enhancing doi.org/10.1002/adem.200800273.
corrosion resistance in Hanks' solution, Surf. Coat. Technol. 204 (2010) [44] J. Reifenrath, N. Angrisani, N. Erdmann, A. Lucas, H. Waizy, J.M. Seitz, A. Bondarenko,
1823–1830, https://doi.org/10.1016/j.surfcoat.2009.11.024. A. Meyer-lindenberg, Degrading magnesium screws ZEK100: biomechanical testing,
[34] M. Yazdimamaghani, M. Razavi, D. Vashaee, L. Tayebi, Microstructural and mechan- degradation analysis and soft-tissue biocompatibility in a rabbit model, Biomed.
ical study of PCL coated Mg scaffolds, Surf. Eng. 30 (2014) 920–926, https://doi.org/ Mater. 45012 (2013)https://doi.org/10.1088/1748–6041/8/4/045012.
10.1179/1743294414Y.0000000307. [45] M. Ezechieli, M. Ettinger, C. König, A. Weizbauer, P. Helmecke, R. Schavan, A. Lucas,
[35] M. Park, J.E. Lee, C.G. Park, S.H. Lee, H.K. Seok, Y. Bin Choy, Polycaprolactone coating H. Windhagen, C. Becher, Biomechanical characteristics of bioabsorbable
with varying thicknesses for controlled corrosion of magnesium, J. Coatings Technol, magnesium-based (MgYREZr-alloy) interference screws with different threads,
Res, 2013https://doi.org/10.1007/s11998-013-9474-6. Knee Surg. Sports Traumatol. Arthrosc. 24 (2016) 3976–3981, https://doi.org/10.
[36] J. Degner, F. Singer, L. Cordero, A.R. Boccaccini, S. Virtanen, Electrochemical investi- 1007/s00167-014-3325-6.
gations of magnesium in DMEM with biodegradable polycaprolactone coating as [46] T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity?
corrosion barrier, Appl. Surf. Sci. 282 (2013) 264–270, https://doi.org/10.1016/j. Biomaterials 27 (2006) 2907–2915, https://doi.org/10.1016/j.biomaterials.2006.01.
apsusc.2013.05.115. 017.
[37] M.A. Woodruff, D.W. Hutmacher, The return of a forgotten polymer — [47] S. Gautam, A. Kumar, N. Chandra, Fabrication and characterization of PCL/gelatin
Polycaprolactone in the 21st century, Prog. Polym. Sci. 35 (2010) 1217–1256, composite nano fi brous scaffold for tissue engineering applications by
https://doi.org/10.1016/j.progpolymsci.2010.04.002. electrospinning method, Mater. Sci. Eng. C 33 (2013) 1228–1235, https://doi.org/
[38] H.M. Wong, P.K. Chu, F.K.L. Leung, K.M.C. Cheung, K.D.K. Luk, K.W.K. Yeung, 10.1016/j.msec.2012.12.015.
Engineered polycaprolactone-magnesium hybrid biodegradable porous scaffold [48] J.H. Gao, X.Y. Shi, B. Yang, S.S. Hou, E.C. Meng, F.X. Guan, S.K. Guan, Fabrication and
for bone tissue engineering, Prog. Nat. Sci.: Mater. Int. 24 (2014) 561–567, https:// characterization of bioactive composite coatings on Mg-Zn-Ca alloy by MAO/sol-gel,
doi.org/10.1016/j.pnsc.2014.08.013. J. Mater. Sci. Mater. Med. 22 (2011) 1681–1687, https://doi.org/10.1007/s10856-
[39] X. Li, S. Zhang, X. Zhang, S. Xie, G. Zhao, L. Zhang, Biocompatibility and physicochem- 011-4349-9.
ical characteristics of poly(Ɛ-caprolactone)/poly(lactide-co-glycolide)/nano-