Corrosion resistance of an amino acid-bioinspired calcium phosphate coating on magnesium alloy AZ31
Corresponding authors: * E-mail address:email@example.com(R.-C. Zeng).
Received: 2019-10-8 Revised: 2020-01-4 Accepted: 2020-01-13 Online: 2020-07-15
An l-cysteine-bioinspired calcium phosphate (Ca-P) coating is prepared upon magnesium alloy AZ31 in a water bath at 60 °C. FE-SEM, FTIR, XRD, electrochemical characterization, hydrogen evolution tests and XPS were used to evaluate the microstructure, chemistry and corrosion performance of the samples. Results indicate that l-cysteine promotes the nucleation process of the coating and significantly increases its thickness. This can be attributed to the complexation of the carboxyl group and mercapto group of l-cysteine with calcium ions. Indeed, the obtained Ca-P coating possesses higher corrosion resistance than that prepared in l-cysteine-free bath.
Cite this article
Xiao-Li Fan, Chang-Yang Li, Yu-Bo Wang, Yuan-Fang Huo, Shuo-Qi Li, Rong-Chang Zeng.
In recent years, magnesium (Mg) and its alloys have become important biomedical implant materials . They are being increasingly used in clinical applications due to their ideal properties, such as low elastic modulus, high biocompatibility with bone tissue, and unique biodegradability [, , ]. Compared to other medical metal implants (e.g. titanium and its alloys), Mg alloys are bioabsorbable during the service periods and aid in easy avoidance of secondary surgery [, , ]. However, Mg alloys with low negative potentials easily corrode when used as implants, resulting in alkaline microenvironments and deterioration in mechanical strength for supporting bone-tissue healing.
In general, a calcium phosphate (Ca-P) coating is applied to the surface of Mg alloys to improve corrosion resistance and biocompatibility, which, in turn, extends the service life of Mg implants and stimulates bone regeneration [8,9]. Ca-P coating has high osteoconduction and low toxicity in physiological environments [10,11], and accordingly, has received extensive research attention in the field of biomaterials [12,13]. In fact, many studies have demonstrated that biodegradable coatings can be fabricated by the formation of Ca-P corrosion products. As a result, secondary protection is achieved in simulated body fluids containing Ca2+ and phosphates . If a Ca-P coating is mimicked and directly prepared on the surface of Mg substrates, it can positively affect degradation and biocompatibility . In previous studies, Ca-P coatings prepared under different conditions have shown different microscopic features such as particles , nanoplates [8,16], and flower-like grains . However, aggressive ions (e.g. Cl-) can penetrate the interface of Ca-P coatings/substrates via pores or channels. Hence, Ca-P coating does not often provide long-term protection for Mg alloys. Therefore, it is vital to refine the grains in, and reduce the defects of, Ca-P coating to improve its protecting effect with respect to Mg substrates.
Prior studies focused on introducing inorganic particles to Ca-P coating. Singh et al.  prepared an HA and iron oxide (Fe3O4) composite coating, which improved corrosion resistance, thereby resulting in favorable properties for bone growth and osseointegration due to the increased porosity of the coating. Batebi et al.  increased the specific gravity of the stable phase in hydroxyapatite (HA) and reduced the crystallite size through doping Ag+ and F-, thereby improving the corrosion resistance and antibacterial activity of the coating. In addition, organic additives can be added to Ca-P coatings as an either inducers or inhibitors. Our group employed EDTA , glucose  and DNA  as inducers to develop a Ca-P coating with dense and refined crystalline grain, thereby enhancing corrosion resistance as well as biocompatibility.
In addition, some organic compounds (e.g. plant extracts, chitosan and cellulose) with special functional groups or structures can act as inhibitors. The structure of an effective inhibitor usually contains nitrogen, sulfur, oxygen, phosphorus, and a multi-bond or an aromatic ring . Refs [20,21]. have shown that functional groups and the type of heteroatoms (e.g. S > O > N) can affect the inhibitory effect of inhibitors. However, the toxicity of inhibitors limits their applications. Therefore, it is necessary to develop green corrosion inhibitors.
Amino acids are the basic unit of proteins with excellent biocompatibility . In addition, they have high purity and are relatively cheap, making them an effective substitute for toxic inhibitors . Recently, our group reported  that in vitro degradation of pure Mg in phosphate buffer saline was inhibited to some degree by the formation of Ca-P products in presence of different concentrations of alanine, glutamic acid and lysine. Ashassi-Sorkhabi et al.  incorporated different concentrations of alanine, glutamine, methionine, and aspartic acid as green inhibitors into the hybrid sol-gel coating, which improved the corrosion resistance of the coating. Among these amino acids, the presence of aspartic acid had the best corrosion resistance. Moreover, according to Ashassi-Sorkhabi et al. , the addition of alanine, glutamine, and methionine as green additives to the composite sol-gel coating improved corrosion resistance. Herein, the addition of methionine with 0.5 wt.% content possessed the best corrosion resistance; it contained two ionizable functional groups with opposite chemical properties (i.e. an amino group (-NH2) and a carboxyl group (-COOH)). Indeed, amino acids with sulfur or long hydrocarbon chains have strong inhibition characteristics [20,23]. In addition, an increase in the electron density of α-NH2 improves inhibition efficiency . In theory, l-cysteine has a relatively good ability to inhibit metal corrosion, because it contains not only -NH2 and -COOH, but also a mercapto group (-SH). In other words, it can coordinate with Mg through the nitrogen atoms of -NH2, oxygen atoms of -COOH, and sulfur atoms of -SH, forming a dense protective coating. At present, l-cysteine is widely studied as a corrosion inhibitor for various metals, such as low-carbon steel and its Al alloys [, , , ].
Mg alloy AZ31 was cut into a rectangular parallelepiped (20 mm × 20 mm × 5 mm). The substrates were polished from 150# to 2500# with silicon-carbide water sandpaper. Surface oil and impurities of the substrates were washed with absolute ethanol, and then dried with warm air. Calcium chloride (CaCl2), potassium dihydrogen phosphate (KH2PO4), and l-cysteine with analytically pure chemical reagents were purchased from Qingdao Jingke Chemical Reagent Co., Ltd., China. The simulated human body fluid (Hank's) solution was used in the electrochemical testing and hydrogen evolution testing. The composition of Hank's solution is the same as described by Xiong et al. .
2.2. Preparation of Ca-P coatings
For the Ca-P solution, 0.25 mol/L of CaCl2 and KH2PO4 were dissolved in deionized water. The samples were immersed in the Ca-P solution at 60 °C for 30 min for a water bath; 0.15 g/L of l-cysteine was added to the above mixed Ca-P solution, and the samples were treated under the same conditions. A schematic diagram of the coating preparation and the dissociation-equilibrium diagram of l-cysteine in the solution are shown in Fig. 1. When the pH of the solution (pH = 4.5 ± 0.5) was lower than the isoelectric point (pI) of l-cysteine (pI = 5.02), l-cysteine was mainly present in l-Cys+. Since the pH of the solution was close to the pI, four different forms of l-cysteine were present in the solution.
Illustration of the preparing of the Ca-P coating and Ca-PL-Cys coating on Mg alloy AZ31 and dissociation equilibrium diagram of l-cysteine in the solution.
2.3. Surface characterization
Microscopic surface morphologies and element compositions of the samples were analyzed using a scanning electron microscope (SEM, Nova Nano SEM 450, FEI Corporation, USA) equipped with energy dispersive X-ray spectroscopy (EDS). The thickness of the coatings and the elemental distribution were determined by scanning the cross section of the samples. The sample surface was subjected to a conductive treatment with a vacuum spraying gold before the SEM test. The phase composition of the coatings was analyzed by Cu-Kα (0.154 nm) radiation by using a Rigaku D/MAX 2500 PC X-ray diffractometer (XRD) from Japan. The diffraction pattern was obtained between 20° and 80° with steps of 0.02°, and the counting time per step was 1 s. The possible chemical bonding formed in the coatings was confirmed through a Fourier transform infrared (FTIR) and an X-ray photoelectron spectrometer (XPS).
2.4. Corrosion tests
Open circuit potential (OCP), electrochemical impedance spectroscopy (EIS), and potentiodynamic polarization (PDP) curves were obtained using an electrochemical workstation (Versa STAT 4, Princeton, USA) in order to evaluate the corrosion behavior of the samples with a three-electrode device . A sample with an exposed area of 1 cm2 was used as working electrode, a saturated calomel as reference electrode, and a platinum plate as counter electrode. Electrochemical testing was carried out in Hank’s solution at room temperature. First, a stable OCP was established within 600 s. EIS analysis was carried out after a short delay with an immersion of 600 s over a frequency range of 10 mHz-100 kHz at a disturbing potential of 10 mV/SCE. EIS data were fitted using ZSimpWin software. The polarization curves were recorded with a sweep rate of 2 mV/s. The electrochemical parameters-corrosion potential (Ecorr), corrosion current density (icorr), and Tafel slopes were estimated with Tafel fitting using VersaStudio and Origin Pro 9.1. A tangent is made at the cathode strongly polarized region of ΔE=60 mV and intersects with y = a (a is the value of the corrosion potential). And the abscissa of the intersection point is the icorr. The polarization resistance (Rp) was calculated using the Stern-Geary equation :
where βa and βc are the Tafel slopes of the anode and cathode, respectively. All measures were tested in triplicates in order to ensure the reproducibility of the experiment.
The long-term corrosion resistance of the samples was characterized by performing a hydrogen evolution test in Hank's solution for 168 h at 37.5 ± 0.5 °C. The hydrogen evolution set-up design was tested by placing the sample with full surface exposure in Hank's solution at 37 ± 0.5 °C under an inverted funnel connected to a graduated burette . The volume of hydrogen evolution in the burettes was intermittently recorded during immersion for 7 days. The hydrogen evolution rate (HER) was calculated using the following formula:
where v denotes the hydrogen evolution volume (mL), s and t denote the exposed area and immersion time, respectively.
3.1. Surface analysis
The microtopography of the (a, b, c) Ca-P coating and (d, e, f) Ca-PL-Cys coating is presented in Fig. 2. Indeed, the surface of the Ca-P coating has many crystal clusters, which are unevenly distributed, moreover, there are fewer regular radial grains in the coating. The microtopography of the Ca-PL-Cys coating surface is clearly different from that of the Ca-P coating. In particular, the Ca-PL-Cys coating surface is almost entirely covered with regular radial grains, with cluster-like grains almost disappearing. The atomic ratio of the surface elemental composition of the coatings and the results are shown in Table 1. The elements Ca, P, C, O, and Mg are detected on both coating surfaces. Al is detected on the surface of the Ca-P coating, which is derived from Mg alloy AZ31. This suggests that the Ca-P coating is defective, which is consistent with the surface morphology analysis. In addition, the presence of N and S elements indicates the presence of l-cysteine in the Ca-PL-Cys coating. Moreover, the Ca/P ratio of the Ca-PL-Cys coating (0.81 ± 0.01) is roughly 1.14 times larger than that of the Ca-P coating (0.71 ± 0.12), which theoretically suggests that the Ca-PL-Cys coating has better bone-tissue compatibility.
SEM micrographs of the (a, b, c) Ca-P coating and (d, e, f) Ca-PL-Cys coating.
Table 1 Chemical composition of Ca-P coating and Ca-PL-Cys coating (at.%).
The cross-section and the element mapping images of (a) Ca-P coating, (b) Ca-PL-Cys coating are shown in Fig. 3. Note that the thicknesses of the three different locations are taken and their average is used to indicate the thickness of the coatings. The thickness of the Ca-P coating is roughly 9.67 ± 4.16 μm, and that of the Ca-PL-Cys coating is roughly 18.67 ± 1.52 μm, both of which have a distinct dense inner layer and a loose outer layer. The dense inner layer of the Ca-P coating is roughly 2 μm thick; the outer layer is significantly rough and loose. The interface between the Ca-P coating and the substrate is rough, which means that the substrate corroded before the Ca-P coating could form. The loose outer layer of the Ca-PL-Cys coating is only 3 μm thick; the entire coating is denser and smoother than that of the Ca-P coating. From the above results, it is evident that the addition of l-cys increases the thickness and compactness of the Ca-P coating.
Cross-section and elemental mapping images of (a) Ca-P coating and (b) Ca-PL-Cys coating.
The Ca-PL-Cys coating is smoother than that of the Ca-P coating, as shown in Fig. 3. Accordingly, the surface roughness of two coatings is further tested using a three-dimensional topography detector. The three-dimensional morphology of both coatings is shown in Fig. 4, where the color bars to the left of the image indicate different depth values; the darker the color, the deeper the pothole. The surface roughness of the Ca-P coating is 6.61 ± 0.76 μm; for Ca-PL-Cys coating, it is 2.41 ± 0.23 μm. Indeed, the roughness of the Ca-PL-Cys coating is markedly lower than that of the Ca-P coating, which is consistent with the results in Fig. 3. This suggests that the addition of l-cysteine reduces the surface roughness of the Ca-P coating.
Surface roughness of (a) Ca-P coating and (b) Ca-PL-Cys coating.
Fig. 5 shows the (I) FTIR spectra and (II) XRD patterns of the samples. The FTIR curves of the two samples have no overt differences in the number and position of peaks, and general characteristic peaks of the Ca-P coating appear. Groups of PO43- and H2PO4- can be found in the peaks ranging from 800-1200 cm-1 (Fig. 5(I)) . In particular, C=O and C—N are present at 1738 cm-1 and 1476 cm-1, respectively, indicating the presence of l-cysteine in the Ca-PL-Cys coating. Characteristic peaks of tricalcium phosphate (TCP, Ca3(PO4)2) , calcium-deficient hydroxyapatite (CDHA) , and dicalcium phosphate dihydrate (DCPD, CaHPO4·2H2O)  also appear in the XRD spectrum. Indeed, the XRD results suggest that different forms of Ca-P products exist in the coatings, which correspond to different reactions (Eqs. (8)-(11)).
(I) FTIR spectra of (a) Ca-P coating and (b) Ca-PL-Cys coating and (II) XRD patterns of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating.
3.2. Corrosion behavior
OCPs of Mg alloy AZ31, the Ca-P coating, and the Ca-PL-Cys coating in Hank's solution are shown in Fig. 6. When the potential difference between two dissimilar metal electrodes exceeds 50 mV, the risk of galvanic corrosion increases . At the beginning of the reaction, the potential difference between the Ca-P coating and the substrate is roughly 57.97 mV, indicating that galvanic corrosion is possible. As the reaction time is prolonged, the OCP difference between the Ca-P coating and the substrate gradually decreases below 50 mV. Indeed, the OCP of the Ca-P coating is lower than that of the substrate, suggesting the chemical dissolution of the Ca-P coating and the formation of corrosion products, which, in turn, results in an increase in the OCP of the Ca-P coating. For the Ca-PL-Cys coating, the OCP difference from the substrate is maintained at 270 mV or more, indicating that the Ca-PL-Cys coating has constant galvanic corrosion.
OCP curves of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating.
Electrochemical impedance spectroscopy (EIS) measurements were performed to characterize the corrosion resistance of the coating. The Nyquist and Bode plots of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating are illustrated in Fig. 7(I-IV). At the same time, the corresponding equivalent circuit models are introduced to analyze the EIS curves. According to the corrosion process of the substrate and coated samples, three equivalent circuit models can be selected corresponding to three Nyquist curves. In the circuit diagram, Rs represents the solution resistance; CPE1 and CPE2 indicate the constant phase components; Rct denotes the charge transfer resistance; RL and L express the inductance resistance and inductance, respectively; and R1 denotes the coating resistance. The Nyquist curves of bare Mg alloy AZ31 and Ca-P coating are composed of a medium-high frequency capacitor loop and a low frequency inductive loop. For the Mg alloy AZ31 substrate, the medium-high frequency capacitive loop is usually attributed to the charge transfer Rct, and the low-frequency induction loop is related to the dissolution and pitting of α-Mg, which is represented by RL and L [, , ]. Indeed, the Rct of the Ca-P coating (2980 Ω cm2) is larger than that of the AZ31 Mg alloy (550 Ω cm2), indicating that the Ca-P coating can protect the Mg alloy matrix. At the same time, the induction loop in the low-frequency region corresponds to pitting corrosion in corrosive media, wherein the RL of the Ca-P coating (2950 Ω cm2) is significantly smaller than that of the AZ31 Mg alloy (7210 Ω cm2), which means that the degree of Ca-P coating pitting corrosion is weaker than that of the AZ31 Mg alloy substrate. Indeed, the Rct of the Ca-PL-Cys coating 19,860 Ω cm2) is six times larger that of the Ca-P coating (2980 Ω cm2); moreover, RL and L completely disappear. Furthermore, the size of the capacitor ring of the Ca-PL-Cys coating is the largest, showing effective corrosion resistance. In addition, a large |Z|0.01Hz indicates optimal corrosion resistance. In Fig. 7(IV), it is apparent that |Z|0.01Hz of Ca-PL-Cys coating is the largest [39,40]. These results all indicated that the corrosion resistance of Ca-PL-Cys coating was the best. The specific fitted data are listed in Table 2.
(I, II, III) Nyquist curves and corresponding equivalent circuits, (IV) bode plots of (a) AZ31 Mg alloy, (b) Ca-P coating and (c) Ca-PL-Cys coating.
Table 2 Fitting results of EIS spectra.
|Samples||Rs (kΩ cm2)||CPE1|
(Ω-1 sn cm-2)
|n1||R1 (kΩ cm2)||CPE2|
(Ω-1 sn cm-2)
|n2||Rct (kΩ cm2)||L (102H cm2)||RL (kΩ cm2)|
|AZ31 substrate||0.08 ± 0.01||1.58 × 10-5 ± 5.5 × 10-6||0.89 ± 0.02||-||-||-||0.55 ± 0.16||4.63 ± 0.83||7.21 ± 0.02|
|Ca-P coating||0.12 ± 0.02||4.38 × 10-6 ± 4.47 × 10-7||0.68 ± 0.03||10.67 ± 0.55||8.13 × 10-6 ± 4.29 × 10-6||0.87 ± 0.04||2.98 ± 0.32||1.67 ± 0.82||2.95 ± 7.63|
|Ca-PL-Cys coating||0.09 ± 0.01||1.15 × 10-5 ±7.28 × 10-6||0.65 ± 0.12||280.22 ± 152.57||1.3 × 10-5 ±1.27 × 10-5||0.80 ± 0.16||19.86 ± 2.93||-||-|
The polarization curves of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating are shown in Fig. 8. The fitting results of the corrosion potential (Ecorr) and the corrosion current density(icorr) obtained by the Tafel curve extrapolation method are listed in Table 3 . Indeed, the icorr of the Ca-P coating (7.21 × 10-6 A cm-2) is in the same order of magnitude as that of the bare Mg alloy AZ31 (5.57 × 10-6 A cm-2), while Ecorr (-1.51 V/SCE) is lower than that of the bare Mg alloy AZ31 (-1.48 V/SCE), indicating an unsatisfactory corrosion resistance with respect to the Ca-P coating. However, the blunt potential (Eb) of the Ca-P coating (-1.26 V/SCE) is higher than that of the Mg alloy substrate (-1.29 V/SCE), indicating that the Ca-P coating can protect the Mg alloy. Furthermore, the icorr of the Ca-PL-Cys coating (4.19 × 10-7 A cm-2) is an order of magnitude smaller than that of the AZ31 Mg alloy (5.57 × 10-6 A cm-2); moreover, Ecorr (-1.41 V/SCE) significantly improves. In addition, the Rp values are arranged in the following order: AZ31 Mg alloy (2.53 kΩ cm2) < Ca-P coating (4.30 kΩ cm2) < Ca-PL-Cys coating (40.6 kΩ cm2). This further proves that the Ca-PL-Cys coating has optimal corrosion resistance.
Polarization curves of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating.
Table 3 Electrochemical parameters of the polarization curves of the samples.
|Ecorr (V/SCE)||icorr (μA cm-2)||βa (102 mV/dec)||-βc (102 mV/dec)||Rp (kΩ cm2)|
|AZ31 substrate||-1.48 ± 0.05||5.57 ± 2.49||1.51 ± 0.76||1.49 ± 0.04||2.53 ± 1.83|
|Ca-P coating||-1.51 ± 0.03||7.21 ± 2.96||1.70 ± 0.98||1.27 ± 0.21||4.30 ± 1.65|
|Ca-PL-Cys coating||-1.41 ± 0.03||0.42 ± 1.84||1.87 ± 0.57||1.55 ± 0.17||40.6 ± 2.09|
Hydrogen evolution tests of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating were used to characterize the long-term corrosion resistance of the samples, as shown in Fig. 9. It is evident that the HER curves as a function of time have an analogous trend (Fig. 9(I)). The HER values sharply decrease at the initial stage of less than 10 h, then increase up to 70 h of immersion, and slowly decrease. However, the change in pH value is completely opposite to that in HER curve. The first drop in HER can be attributed to the formation of the corrosion products of the samples being corroded by weak alkaline Hank's solution , leading to a small increase in pH value. Thereafter, the coating is destroyed and defects are generated; and glucose in Hank’s solution is converted to gluconic acid , resulting in enhanced corrosion rates and decreased pH values. However, as the corrosion progresses, the corrosion products were gradually deposited on the substrate surface, and resume the protection to some degree. The HER values of the three samples can be arranged in the following order: Mg alloy AZ31 (6.08 × 10-4 mL cm-2 d-1) > Ca-P coating (5.30 × 10-4 mL cm-2 d-1) > Ca-PL-Cys coating (2.68 × 10-4 mL cm-2 d-1). And in the final stage, the pH value of the Ca-PL-Cys coating was always the maximum. The results suggest that the Ca-PL-Cys coating can provide effective long-term protection for the Mg alloy substrate. To be sure, the above HER results are lower than what the human body can withstand (2.25 mL cm-2 d-1) . Therefore, the prepared coatings are able to meet the corrosion-rate requirements of Mg alloys in the human body.
(I) Hydrogen evolution rates, (II) pH-Time curves of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating.
According to the results of the electrochemical test and hydrogen-evolution test, the coating experiences galvanic corrosion. Since the Ca-P coating is not dense, the corrosive medium reaches the substrate through the holes, causing corrosion in the early stages of immersion. With time, the galvanic-corrosion efficiency decreases because the coating pores are blocked by the corrosion products. Therefore, the results of the polarization curve suggest that the Ca-P coating corrosion resistance is inferior to that of the matrix; however, the long-term corrosion resistance of the Ca-P coating is better than that of the substrates.
Fig. 10 displays the digital-camera photographs and SEM morphologies after a 168-h immersion period. The elemental composition and atomic ratio of the corresponding points in Fig. 10 are shown in Table 4. From Fig. 10(a), it is evident that the Mg alloy AZ31 substrate is seriously corroded. The cracks on the substrate surface can be found in the SEM images in Fig. 10(b) and (c). According to the EDS results, the corrosion products around the corrosion pit mainly consist of Mg(OH)2 and Ca-P products; moreover, a part of the Ca-P coating is dissolved, as can be seen in Fig. 10(d). From Fig. 10(e) and (f), the Ca-P coating has an obvious corrosion pit after immersion and the surface microscopic morphology is visibly missing. However, the macroscopic surface of the Ca-PL-Cys coating is relatively complete, with only a few corrosion marks at the sample boundary. Indeed, there is no corrosion pit in the microscopic morphology of the Ca-PL-Cys coating, with the surface coating relatively intact. Unlike the stereoscopic morphology of the Ca-P coating, the Ca-PL-Cys coating has a flat morphology. The Ca-P “crouch” can provide optimal protection for the sample surface.
Digital camera photographs and corrosion SEM morphologies after a 168 h immersion for (a, b, c) Mg alloy AZ31, (d, e, f) Ca-P coating and (g, h, i) Ca-PL-Cys coating.
Table 4 EDS analysis of the samples after immersion for 168 h (at.%).
Fig. 11 shows the macroscopic morphology and surface roughness of the samples after removing the corrosion products. Pitting corrosion emerges on the AZ31 substrate (a) and the Ca-P coating (b); while the corrosion degree of Ca-PL-Cys coating is the lowest. The surface roughness proves that the Ca-P coating has a deeper corrosion pit than the substrate. When the sample was immersed in the Hank’s solution, the corrosion pits are quickly initiated beneath the Ca-P coating on the α-Mg matrix next to intermetallic compounds (i.e. AlMn) in the substrate, resulting in the peeling off of the Ca-P coating and the exposure of the substrate to the solution. In addition, the exposed part of the sample and the neighboring parts of the sample created a galvanic cell due to the presence of an initial potential difference of approximately 60 mV (Fig. 6), and thus galvanic corrosion occurred, resulting in a corrosion pit with depth of > 0.40 mm (Fig. 11(b)). Regardless of the big difference in initial potential of the Ca-PL-Cys coating and its substrate, the prepared Ca-PL-Cys coating is thicker and more uniform than the Ca-P coating, and hence provides better protection to the substrate. This finding is consistent with the results of the electrochemical and hydrogen evolution tests.
Digital camera photo and surface roughness of (a) Mg alloy AZ31, (b) Ca-P coating and (c) Ca-PL-Cys coating after immersion 168 h and removing corrosion products.
4.1. Effects of organic additives on Ca-P coating
In our previous study, it was found that organic compounds with specific functional groups can effectively induce the formation of Ca-P coating. For example, glucose was added to the Ca-P conversion solution to prepare a Ca-P coating on pure Mg . The presence of glucose makes the coating more compact without increasing the thickness. Accordingly, the Rp of Ca-P coating will be three times larger than that of pure Mg. In the present study, l-cysteine was added to the Ca-P conversion solution to prepare the Ca-P coating for the AZ31 Mg alloy. Indeed, the Ca-PL-Cys coating had a denser microscopic morphology (Fig. 2) than the Ca-P coating and its thickness increased by a factor of roughly two (Fig. 3). Moreover, the Rp of the Ca-PL-Cys coating increased by a factor of roughly sixteen, relative to the AZ31 Mg alloy. This can be attributed to the conversion of glucose to gluconic acid in the solution, thereby containing only -COOH, whereas l-cysteine contained -COOH and -SH. Indeed, both -COOH and the -SH can effectively bond with Ca2+ and Mg2+ in the solution, thereby promoting nucleation of the Ca-P compound.
4.2. Film formation mechanism
Fig. 12 demonstrates the corresponding microscopic topographies at different times during the formation of the Ca-P coating and Ca-PL-Cys coating. The elemental composition of the coatings, which were soaked for 10 s, 30 s and 60 s, are shown in Table 5. Monitoring the formation of coatings helps us to understand the mechanism behind it. From Fig. 12(a) and (g), it is evident that the two coatings form a flower-like crystal nucleus, ultimately growing into crystal grains. In addition, the nucleation of the Ca-PL-Cys coating is faster and smaller than that of the Ca-P coating. The presence of N and S elements (Table 5) detected on Ca-PL-Cys coating after 10 s, 30 s and 60 s of formation confirms the idea that l-cysteine was involved in Ca-PL-Cys coating formation. In addition, with respect to the Ca-P coating, grain agglomeration is evident for the following time range: 1-20 min. The Ca-P crystal grains show an irregular sheet form based on the flower crystal nucleus, as can be seen in Fig. 12(b)-(f). Moreover, there are many voids between the crystal grains. The crystal agglomeration during the formation of Ca-PL-Cys coating is less intense than that during the Ca-P coating formation, and most of the grain growth is dominated by irregular sheets, as seen in Fig. 12(h)-(l). Unlike the Ca-P coating, the crystal grains in the Ca-PL-Cys coating are essentially parallel to the surface, with less voids between them. Based on this information, it is evident that l-cysteine can promote nucleation, refine grains, and reduce grain agglomeration during the Ca-P coating formation process.
Corresponding microscopic topographies at different time during the formation of (a-f) Ca-P coating and (g-l) Ca-PL-Cys coating.
Table 5 EDS analysis of corresponding microscopic topographies at different time during the formation of coatings (at.%).
|Ca-P coating 10 s||#1||15.45||--||76.9||4.26||1.4||2.00||--||--||0.33|
|Ca-P coating 30 s||#4||20.81||67.31||6.30||2.87||2.72||0.46|
|Ca-P coating 60 s||#7||70.25||9.79||11.04||8.91||0.81|
|Ca-PL-Cys coating 10 s||#10||60.68||20.94||3.03||7.35||6.01||1.53||0.46||0.82|
|Ca-PL-Cys coating 30 s||#13||66.32||13.06||8.80||6.59||5.00||0.23||0.75|
|Ca-PL-Cys coating 60 s||#16||65.26||6.37||13.41||12.68||2.28||0.95|
To understand the effect of l-cysteine on Ca-P coating formation, XPS testing was performed on different stages of the Ca-PL-Cys coating. Fig. 12 shows the XPS survey spectrum of the Ca-PL-Cys coating, and Fig. 13 shows the XPS spectra of (a, d) C 1s, (b, e) N 1s and (c, f) S 2p. At 10 s, C, N, S, Ca and P already exist on the Ca-PL-Cys coating surface, indicating that Ca-P was produced and l-cysteine was involved in the reaction. The C, N and S elements were fitted with peaks accordingly, as can be seen in Fig. 14. According to the C1s spectrum (Fig. 14(a, d)), the C elements in the Ca-PL-Cys coating are all derived from l-cysteine, which are C—C (284.8 eV), C—O (287.6 eV) and C=O (288.8 eV) in HSCH2CHNH2COOH [15,45,46]. Furthermore, both Mg2+ and Ca2+ are linked to -COOH. According to the N 1s spectrum (Fig. 14(b, e)), N is derived from C—N (399.7 eV) and N—H (400.4 eV) in HSCH2CH(NH2)COOH . From the S 2p spectrum (Fig. 14(c, f)), S exists in the form of S—O (sulfur oxide) , because -SH of l-cysteine is easily oxidized. Since there was no oxidant in the solution and the oxidation of O2 in the air was weak, -SH only slightly oxidized into sulfur oxide. Accordingly, -SH bonded to the metal ions (Mg2+ and Ca2+) in the form of S—O. According to the results of XPS analysis, both -COOH and -SH in l-cysteine can attract and bind metal ions (Mg2+ & Ca2+) in solution, resulting in a better effect of cysteine modification than additive with only -COOH.
XPS spectra of Ca-PL-Cys coating (a, d) C 1s, (b, e) N 1s, (c, f) S 2p.
Fig. 15 shows a schematic diagram of the Ca-P coating and the Ca-PL-Cys coating formation mechanism, from which it is evident that the α-Mg dissolved into Mg2+ and that Mg2+ reacted with H2O in the solution to generate Mg(OH)2, which was deposited onto the Mg-alloy surface in order to produce a thin protective layer. Indeed, Mg(OH)2 continued to grow throughout the coating process. The presence of Mg(OH)2 is demonstrated by the FTIR results. The reaction equation at this stage is as follows:
Schematic diagram of (a-c) Ca-P coating and (d-f) Ca-PL-Cys coating formation mechanism.
At the same time, H2PO4- has the following dissociation equilibrium in solution:
Both cations and anions are present in the solution, mutually attracting each other to form the Ca-P coating. In the early stages of coating formation, Ca-P nucleated and scattered on the substrate surface, and as the reaction continued, Ca-P gradually covered the entire surface. The following reactions can be carried out simultaneously in the solution:
The Ca-PL-Cys coating differs from the Ca-P coating in that cysteine appears in different patterns in the solution depending on the pH level. The negatively charged functional group of l-cysteine binds to Ca2+ and Mg2+, whereas the functional group with a positive charge attracts the anion in the solution. On one hand, it promotes the dissolution of the Mg matrix. On the other hand, it accelerates the Ca-P nucleation. The Ca-PL-Cys coating is thus more stable, compact and uniform than the Ca-P coating.
A Ca-P coating was prepared upon the surface of Mg alloy AZ31. The corrosion performance of the coating was improved by the addition of l-cysteine. The conclusions were outlined below:
(1)The addition of l-cysteine significantly improved the uniformity and integrity of the coating's micro-morphology, resulting in grain growth with respect to the Ca-PL-Cys coating. The roughness of the coating decreased from 6.61 ± 0.76 μm to 2.41 ± 0.23 μm. At the same time, its thickness increased from 9.67 ± 4.16 μm to 18.67 ± 1.52 μm.
(2)Corrosion current density, i.e. icorr of the Ca-PL-Cys coating was an order of magnitude smaller than that of the Ca-P coating. HER of the Ca-PL-Cys coating (2.68 × 10-4 mL cm-2 d-1) was three times lower than that of the substrate, significantly enhancing corrosion resistance.
(3)Under the dual action of -COOH and -SH, Mg2+ and Ca2+ were quickly adsorbed, thereby accelerating Ca-P nucleation and refining the Ca-P crystal grains. The -SH existed in the form of S—O and bonded to the metal ions.
This work was financially supported by the National Natural Science Foundation of China (No. 51571134), the SDUST Research Fund (No. 2014TDJH104) and the Science and Technology Innovation Fund of SDUST for graduate students (No. SDKDYC190301).
To overcome the defect of high degradation rate of magnesium (Mg), bioactive coatings with compact structure, sufficient bonding strength and enhanced corrosion resistance are essential for Mg-based biodegradable implants. In this study, a dense Mg-substituted beta-tricalcium phosphate and magnesium hydroxide (beta-TCMP/Mg(OH)2) composite coating was prepared on AZ31 alloy via one-step hydrothermal method. The influences of hydrothermal temperature on its composition, microstructure of the surface and interface, bonding strength and corrosion behavior were evaluated. The results showed that the compact composite coating synthesized at 140 degrees C not only possessed a crack-free bilayered structure with an adequate bonding strength (more than 20.88+/-1.60MPa), but also got an extreme high impedance (1197.003+/-152.817kOmegacm(2)) so that significantly enhanced the corrosion resistance and inhibited the formation of pitting corrosion. Furthermore, the in vitro immersion test suggested that the composite coating slower the initial degradation rate of Mg alloys and enhanced its surface bioactivity to some extent.
A biodegradable Ca-deficient hydroxyapatite (Ca-def HA) coating has been directly prepared on Mg-Zn-Ca alloy by pulse electrodeposition to improve its corrosion resistance and biocompatibility. However, the formation mechanism of such a Ca-def HA coating on magnesium substrate is still not clear. In this study, the microstructure evolution of the coating was characterized using x-ray diffractometer, x-ray photoelectron spectroscopy and scanning electron microscopy. Thermodynamic and kinetic studies of the precipitation of hydroxyapatite (HA), octacalcium phosphate (OCP) and dicalcium phosphate dihydrate (DCPD) in the used electrolyte were also carried out. Theoretical analyses illustrate that the precipitation of HA, OCP and DCPD are all possible when the electrolyte pH is higher than 6 at 80 C, and that the higher the pH value, the more favorable is the formation of HA. Nevertheless, there is mainly poor crystalline Ca-def HA on the substrate when pulse electrodeposition lasts for 5 min, and its crystallinity increases with duration time The direct formation of the Ca-def HA coating on Mg-Zn-Ca alloy is closely dependent on the phase composition and microstructure of the substrate, the deposition parameters and Mg2-1 ions substitution in HA structure. 2014 Elsevier B.V.