Improving the corrosion properties of magnesium AZ31 alloy ...ijmmm.ustb.edu.cn/fileKWYJYCLXB/journal/article/ijmmm/2017/5/PD… · ported that the corrosion resistance of weldments

International Journal of Minerals, Metallurgy and Materials Volume 24, Number 5, May 2017, Page 566 DOI: 10.1007/s12613-017-1438-x

Corresponding author: M. Siva Prasad E-mail: [email protected]

© University of Science and Technology Beijing and Springer-Verlag Berlin Heidelberg 2017

Improving the corrosion properties of magnesium AZ31 alloy GTA weld

metal using microarc oxidation process

M. Siva Prasad1,2), M. Ashfaq3), N. Kishore Babu4), A. Sreekanth5), K. Sivaprasad2), and V. Muthupandi2)

1) Department of Advanced Materials Engineering, Dong-Eui University, 995 Eomgwangno, Busanjin-gu, Busan 614-714, South Korea

2) Advanced Materials Processing Laboratory, Department of Metallurgical and Materials Engineering, National Institute of Technology, Tiruchirapalli 620015, India

3) FARCAMT, Advanced Manufacturing Institute, King Saud University, Riyadh, Saudi Arabia

4) Empa, Swiss Federal Laboratories for Materials Science and Technology, Laboratory for Advanced Materials Processing, Feuerwerkerstrasse 39, CH-3602 Thun,

Switzerland

5) Department of Chemistry, National Institute of Technology, Tiruchirappalli 620015, India

(Received: 20 November 2016; revised: 17 December 2016; accepted: 19 December 2016)

Abstract: In this work, the morphology, phase composition, and corrosion properties of microarc oxidized (MAO) gas tungsten arc (GTA) weldments of AZ31 alloy were investigated. Autogenous gas tungsten arc welds were made as full penetration bead-on-plate welding under the alternating-current mode. A uniform oxide layer was developed on the surface of the specimens with MAO treatment in silicate-based alkaline electrolytes for different oxidation times. The corrosion behavior of the samples was evaluated by potentiodynamic polarization and electrochemical impedance spectroscopy. The oxide film improved the corrosion resistance substantially compared to the uncoated speci-mens. The sample coated for 10 min exhibited better corrosion properties. The corrosion resistance of the coatings was concluded to strongly depend on the morphology, whereas the phase composition and thickness were concluded to only slightly affect the corrosion resistance.

Keywords: magnesium alloys; weldments; microarc oxidation; corrosion resistance

1. Introduction

Magnesium and its alloys have become ubiquitous in aerospace, automobile, biomedical, and structural applica-tions because of their combinations of outstanding proper-ties, including high specific strength, workability, biocom-patibility, and recyclability [1]. Most of the magnesium al-loys are weldable by gas tungsten arc welding (GTAW), gas metal arc welding (GMAW), electron beam welding (EBW), and laser beam welding (LBW). GTAW is a cost-effective process that produces high-quality magnesium alloy welds [2–3]. However, these alloys suffer from severe cor-rosion issues because of their chemical reactivity and lack of a uniform protective oxide layer [4–5]. Thus, an effective surface treatment method is needed to overcome the corro-sion-associated problems of these alloys [6–7]. Among the available surface coating techniques, microarc oxidation (MAO) is a promising and environmentally friendly

aqueous electrolytic surface modification technique for the fabrication of high-quality corrosion-resistant coatings on metals such as magnesium, aluminum, titanium, zirconium, and their alloys [6,8–10]. It involves the production of a large number of microarcs under an extremely high applied voltage in an appropriate electrolyte, allowing magnesium alloys to form a ceramic oxide layer [11]. Sreekanth et al. [6] revealed that the addition of sodium tetraborate to the so-dium silicate-based electrolyte system significantly im-proved the corrosion properties of the MAO coating by de-veloping a less-porous morphology composed of Mg2SiO4 and MgO as the major phases. Many researchers have re-ported that the corrosion resistance of weldments of magne-sium and aluminum alloys was significantly improved by MAO treatment [5,12–14]. Chen et al. [12] studied the cor-rosion behavior of friction stir welded AZ31B magnesium alloy with an MAO coating and concluded that the phase composition, morphology, and corrosion properties of dif-

M. Siva Prasad et al., Improving the corrosion properties of magnesium AZ31 alloy GTA weld metal … 567

ferent zones (work zone (WZ), heat affected zone (HAZ), and base metal (BM)) of MAO-coated FSW joints were almost the same, even though they had different microstructures. Their results implied that the microstructure of magnesium alloys had no major influence on the growth of MAO coatings. Srinivasan et al. [13] arrived at a similar conclusion that MAO coatings prepared by using a sodium silicate-based electrolyte were not influenced by the grain-size differences in weldments.

The majority of previous investigations involving MAO coatings focused on their corrosion behaviors when depo-sited onto a substrate and the effects of the processing pa-rameters, such as the electrolytic composition, voltage, cur-rent mode (AC or DC, unipolar or bipolar), frequency, duty cycle, and current density, on the coating performance. However, limited literatures are available on the corrosion behavior of magnesium weldments treated with the MAO process. The present investigation focuses on the effects of oxidation time on the phase composition, surface morphol-ogy, and corrosion resistance of the oxide layer formed by MAO on GTA-welded AZ31 magnesium alloy.

2. Materials and methods

2.1. Welding procedure

Rolled AZ31 magnesium alloy (96.19wt% Mg, 2.70wt% Al, 0.81wt% Zn, and 0.30wt% Mn) sheets with a thickness of 3 mm were used in the current investigation. Autogenous full penetration bead-on-plate GTA melt runs were carried out under a modified alternating current (AC) with the fol-lowing electrical parameters: arc voltage, 11 V; DCEN, 110 A, 16 ms; and DCEP, 110 A, 4 ms. An AMET VPC-450 (va-riable polarity control 450-A) power source was used to make GTA welds. Argon and a mixture of argon and helium (50vol%–50vol%) were used as the shielding gas and trail-ing gas, respectively. The flow rate was 15 L/min for the shielding gas and 20 L/min for the trailing gas. The travel speed used for the weld was 4.16 mm/s.

2.2. Development of MAO coatings

Rectangular coupons of weldments (weld, heat affected zone, and base metal together) with dimensions of 10 mm × 10 mm × 3 mm were used as the substrate material for the MAO process. Surfaces of all the samples were metallo-graphically polished to establish uniform surface finish and the samples were subsequently cleaned ultrasonically with acetone and distilled water to avoid any surface impurities prior to the MAO treatment. The electrolyte consisted of 10 g of sodium silicate (Na2SiO3·9H2O), 4 g of potassium hydrox-ide (KOH), 2 g of disodium tetraborate (Na2B4O7·10H2O),

and 1 L of distilled water. The weldment sample prepared for the MAO process was connected to the positive terminal of the power source (anode), whereas the negative terminal was connected to the stainless steel bath containing electro-lyte (cathode). During the MAO process, the electrolyte so-lution was stirred constantly by a magnetic stirrer to ensure proper dissipation of heat from the electrolyte. Also, to maintain a constant temperature, the complete electrolyte setup was enclosed in a customized bath through which cold water was circulated by an external cooler. A pulsed DC power supply with a capacity of providing voltage as high as 800 V and current as high as 20 A was used for the MAO treatment. The parameters that were selected on the basis of repeated experimental trials were as follows: current density, 85 mA/cm²; frequency, 50 Hz; and the duty cycle, 50%. The specimens were subjected to MAO treatments for three dif-ferent oxidation times: 10, 20, and 30 min. These coated samples were designated as A10, A20, and A30, respec-tively, and the uncoated one was designated as A0.

2.3 Characterization of coatings

The approximate values of the thickness for different coatings were measured using a QNIX eddy current thick-ness gauge. Ten measurements were taken, and the average value was tabulated and plotted. The phase composition of MAO coatings was analyzed using an X-ray diffractometer (Ultima III, Rigaku, Japan) equipped with a Cu Kα radiation source (λ = 0.154056 nm) operated at 40 kV and 30 mA; the samples were scanned at 18°/min and at a step size of 0.05° over the 2θ range from 20° to 80°. A scanning electron mi-croscope (S3000H, Hitachi, Japan) was used to study the surface morphology of the MAO-coated samples. SEM im-ages were processed using the Image-J software to measure the pore diameter.

2.4. Electrochemical corrosion studies

The corrosion properties of MAO-coated samples were assessed by potentiodynamic polarization tests and electro-chemical impedance spectroscopy (EIS) using a comput-er-controlled CH Instruments potentiostat (CHI 604D). In potentiodynamic scans, the samples (both coated and un-coated) were used as working electrodes (WE), and a plati-num electrode and a saturated calomel electrode (SCE) were used as the auxiliary electrode (AE) and the reference elec-trode (RE), respectively. After an initial delay of 30 min, the polarization scan was conducted at a sweep rate of 0.001 V/s from an initial potential of −2.0 V to a final potential of 2.0 V. EIS tests were carried out in the frequency range from 1 to 100 kHz with an AC voltage amplitude of 10 mV. All of the

568 Int. J. Miner. Metall. Mater., Vol. 24, No. 5, May 2017

corrosion tests were performed in a freshly prepared neutral 3.5wt% NaCl solution with a pH value of 6.8.

3. Results and discussion

3.1. Voltage–time characteristics and thickness of coat-ings

The voltage–time plot for MAO treatment of the magne-sium AZ31 alloy weldment in sodium silicate solution, shown in Fig. 1, shows two inflections and three stages, as previously revealed [15]. In the first stage, the voltage in-creases with the increase of time and has a very high slope of 16 V/s. Sparks are not observed on the surface during this stage, and a thin passive oxide film (MgO) is formed. This stage is similar to an anodization treatment. Oxygen gas re-sulting from the oxidation of water and hydroxyl ions is ad-sorbed uniformly over the oxide film. Thus, it forms a four-phase complex system: substrate/dielectric oxide/oxygen gas/electrolyte. Electrical conductivities of both the dielec-tric film and the gas envelope are very low, and the cell vol-tage eventually increases. At a definite point, breakdown of the dielectric film occurs and a large number of fine white sparks are visually perceived to be quickly moving over the surface of the specimen [16]. These sparks represent the second stage; during this stage, the voltage further increases with a small slope of 1.8 V/s. In the third stage, a uniform sparking is established on the metal surface and the voltage becomes relatively stable, with a diminutive slope of 0.08 V/s, because of the steady discharge and the relatively constant growth rate of the film. White sparks become more intense and turn orange in color. Gradually, the spark size increases and the density of sparks decreases. The first inflection cor-responds to the breakdown voltage (245 V), beyond which sparking phenomenon initially occurs. The time required to

Fig. 1. Variation of voltage with oxidation time during the MAO treatment.

attain the breakdown voltage is termed the ignition time and was approximately 14 s. The second point of inflection is known as the critical voltage (436 V), which is the point where the nature of the sparks changes and becomes stable [15].

Fig. 2 and Table 1 illustrate the changes in coating thick-ness and growth rate with respect to the oxidation time, which were measured using an eddy current thickness gauge. As previously mentioned, a linear increment in thickness with oxidation time is obtained, but the growth rate de-creases with the increase in oxidation time. This behavior may be due to the reduction in the rate of change of the vol-tage with respect to time, as shown in Fig. 1.

Fig. 2. Variations of the thickness and growth rate of MAO coatings with oxidation time.

Table 1. Thicknesses and growth rates of MAO-coated samples

Sample Thickness / µm Growth rate / (µm·min–1)

A10 51±4 5.10

A20 65±4 3.25

A30 84±3 2.80

3.2. Phase analysis

The X-ray diffraction results displayed in Fig. 3 clearly reveal the presence of the prevailing phases of spinel Mg2SiO4 (forsterite) and MgO (periclase) [6,17]. Previous investigations have clarified that the phase composition of both MAO-coated weldments and base metals are similar in the same electrolyte [6]. An increase in the peak intensities of the peaks of Mg2SiO4 and MgO phases is observed with the increase of oxidation time because of the increasing coating thickness. Moreover, peaks due to the magnesium substrate are also detected as a result of the thin and porous coating layer, which allows X-rays to penetrate into the sub-strate [17]. The sharpness of all the peaks suggests that the coatings are crystalline in nature [17–18].


Fig. 3. Phase compositions of different MAO-coated samples.

3.3. Surface morphology of coatings

Fig. 4 shows the morphological features of all of the MAO-coated samples with different oxidation times. Al-though the microstructure of the weld zone markedly differs from that of the base metal in terms of grain size, the MAO-coated weld zone shows a surface morphology identical to that obtained in previous studies on MAO-coated base met-als [5–6]. Fig. 4 reveals that the morphologies of the coatings differ from each other. Also, all of the coatings exhibit differ-ent porosity levels. We inferred from the surface morphologies of the coatings and the corresponding analysis of the size of micropores that a substantial difference exists in porosity

among different samples. These pores, which are more or less circular in nature, are known as microdischarge channels [19]. In addition to pores, the surface also contains craters, micro-cracks, projections, and volcano-top-like cavities.

The difference in surface morphologies of the coatings may arise from different discharge events occurring at dif-ferent oxidation times of the MAO treatment. The molten oxide and oxygen gas bubbles are thrown out to the coating surface through the discharge channels and finally end up as micropores [19]. The pore diameter, pore density, and the pore area ratio of different samples, as obtained from SEM images, are tabulated in Table 2. Here the pore density, which is defined as the average number of pores per unit area of the coating, is used to express the statistical number of pores in the coating surface. However, the area of the

pores is expressed as the pore area ratio, which is the ratio of the area of pores to the total area of the coating. The number (pore density) and size of pores depend on the number and sizes of sparks at the termination of the MAO process. That is, the pore characteristics depend on the fi-nal voltage, final current, and the oxidation time. A high value of final voltage induces sizably voluminous sparks on the surface, although less in number. Thus, the sample subjected to 30-min oxidation time exhibits large pores that are less in number. Furthermore, microcracks, as shown in Fig. 5(a), are also generated on the surface; they are derived from thermal stresses due to the very high cooling rate of molten oxides and silicates. The local tem-perature of the coating surface is approximately 3000°C, and the cooling rate reaches ~10°C/s when the surface

Fig. 4. Surface morphology of MAO-coated samples: (a) A10; (b) A20; (c) A30.


contacts the relatively cold electrolyte [19]. Fig. 5 reveals the formation of large microcracks (arrow) on the surface of the MAO coating of sample A30. At a long oxidation

time of 30 min, these pores and microcracks become large in size and penetrate deep into the inner layer of the coat-ing, as displayed in Fig. 5(b).

Table 2. Diameter of pores, pore density, and pore area ratio of MAO-coated samples

Sample Average diameter of pores / µm Average pore density / (10–3 µm–2) Average pore area ratio

A10 13.6 ± 0.1 0.93 0.09

A20 17.4 ± 0.4 0.36 0.13

A30 22.8 ± 0.2 0.13 0.17

Fig. 5. Surface (a) and cross-sectional morphologies (b) of MAO-coated sample A30.

3.4. Electrochemical corrosion studies

The potentiodynamic polarization curves for coated sam-ples with different oxidation times are shown in Fig. 6, and the data obtained from the tests are summarized in Table 3. The polarization resistance and corrosion current density of the samples are generally used to characterize the anticorro-sion property of the coatings. A high polarization resistance and a low corrosion current density of a coating imply a low corrosion rate and a high uniform corrosion resistance [20]. The results in Table 3 indicate that MAO-coated samples show greater corrosion resistance than the uncoated sub-strate. The polarization resistance, Rp, of all of the MAO-treated samples is significantly higher than that of the substrate. The substrate, when exposed to the corrosive me-dium (NaCl), immediately forms a loose and permeable Mg(OH)2 layer. Increments in anodic potential lead to the acceleration of Cl− ions from the NaCl solution through this porous layer of Mg(OH)2 and promote the dissolution of the substrate. By contrast, in the case of MAO-treated samples, the complex multiphase oxide structure acts as a barrier to the transport of corrosive Cl− ions and averts their direct contact with the substrate. The sample coated using a 10-min oxidation time exhibits superior anticorrosion prop-erties (Rp = 441.16 kΩ·cm2, corrosion rate = 0.002 mm/a). Moreover, the corrosion current density of sample A10 (0.09 µA/cm2) is approximately 50 times lower than that of

sample A20 and 90 times lower than that of sample A30. The samples can be arranged on the basis of the increasing order of corrosion resistance as A0 < A30 < A20 < A10.

Fig. 6. Potentiodynamic polarization curves of MAO-coated and uncoated samples in 3.5wt% NaCl solution.

Table 3. Potentiodynamic polarization data for MAO-coated and uncoated samples

SampleEcorr /

mV

icorr /

(µAcm–2)

Rp /

(kΩcm2)

Corrosion rate /

(mma–1)

A0 –1435 153.50 0.21 3.540

A10 –1444 0.09 441.16 0.002

A20 –1537 4.67 9.29 0.108

A30 –1542 8.34 5.28 0.192


EIS is a valuable technique for examining the corrosion behavior of MAO-treated metals. Fig. 7 shows the Nyquist plot, and the Bode plots are projected in Figs. 8 and 9 for both uncoated (A0) and coated samples (A10, A20, and A30). The equivalent circuit for the samples in corrosive aqueous solution is shown in Fig. 10. It is composed of solution resistance offered by NaCl (Rs), pore resistance (Rpo), inner-barrier-layer resistance (Rb) of the oxide film, constant phase elements (CPEs) corresponding to pores (CPEpo), and the inner barrier layer (CPEb). CPE is defined as ZCPE = [(jω)CPE-nCPE-T]–1,where CPE-T is the constant of the CPE element, jω is the complex variable for sinu-soidal perturbations (ω = 2πf), and CPE-n is the exponent of CPE with values between –1 and +1. Table 4 shows the corresponding circuit parameter values obtained from the EIS plots.

The presence of two time constants in the equivalent circuit indicates a two-layered structure of the oxide coat-ing — specifically, the inner dense layer and the outer por-ous layer. The value of Rb is 1.5–4 times greater than the value of Rpo, which indicates that the inner layer plays a sig-nificant role in protecting weldments from aggressive Cl− ions. Thus, Rb is a dominant factor in the corrosion behavior of MAO-treated samples [21]. As evident from the EIS results,

Fig. 7. Electrochemical impedance spectroscopy (Nyquist plot) of MAO-coated and uncoated samples in 3.5wt% NaCl solu-tion.

Fig. 8. Electrochemical impedance spectroscopy (Bode plot: impedance/frequency) of MAO-coated and uncoated samples in 3.5wt% NaCl solution.

Fig. 9. Electrochemical impedance spectroscopy curves (Bode plot: theta/frequency) of MAO-coated and uncoated samples in 3.5wt% NaCl solution.

Fig. 10. Equivalent electrical circuit for modeling the beha-vior of coated samples (A10, A20, and A30) (a) and the un-coated sample (A0) (b).

Table 4. Electrochemical impedance data for MAO-coated and uncoated samples

Sample CPEpo-T / (µFcm–2) CPEpo-n Rpo / (Ωcm2) CPEb-T / (µFcm–2) CPEb-n Rb / (Ωcm2)

A0 619.400 0.68 13.45 — — —

A10 1.791 0.75 20677 0.175 0.87 25730

A20 10.890 0.44 2352 0.394 0.79 5542

A30 20.170 0.20 754 1.149 0.64 3764


the outer porous layer resistance and the inner dense layer resistance of the A10 sample are very high compared to those of the other samples. Hence, it exhibits superior anti-corrosion properties. Furthermore, the CPEpo-T and CPEb-T values of the A10 sample are the lowest among the investi-gated samples. If the porosity of the coating is high, the sur-face area will be high and the capacitance will follow the same trend [6]. Thus, the low CPEpo-T value of the A10 sample indicates a low porosity of the coating, which is also evident from its surface morphology. Sample A10, with low values of CPEpo-T and CPEb-T and high values of Rpo and Rb, exhibits a low porosity and, therefore, better corrosion

resistance. By contrast, CPEpo-T and CPEb-T values increase and Rpo and Rb values decrease with the increase of oxidation time. Previous investigations revealed that the fine-grained structure of the weldment facilitates the formation of a denser layer, providing superior anticorrosion performance [22]. Hence, the oxide layer over the weldment is denser than that over the base metal. The relation between CPEpo-T and poros-ity (i.e., the diameter of the pores, pore density, and the pore area ratio) is illustrated in Fig. 11. Changes in CPEpo-T, pore diameter, and pore area ratio as functions of oxidation time follow similar behaviors, i.e., the value of CPEpo-T increases with the increase of pore diameter and pore area.

Fig. 11. Variation of the CPEpo-T value, diameter of pores, pore density, and pore area ratio with oxidation time.

The results of potentiodynamic polarization and EIS stu-dies are in good agreement with the results obtained from morphological analyses. The MAO coating with low poros-ity is reasonably understood to exhibit superior corrosion re-sistance. Thus, we concluded that very low porosity and im-proved corrosion resistance can be achieved with shorter oxidation times. However, the anticorrosion property of ce-ramic coatings has previously been demonstrated to be de-cided by a joint effect of its thickness, morphology (porosi-ty), and chemical composition [20]. However, in the present case, coating thickness and phase composition exert no such significant influence on corrosion properties. The thicker coating with a high porosity exhibits poorer corrosion resis-tance than the thinner coating with a low porosity. The dif-ference in the corrosion resistance of MAO coatings for dif-ferent oxidation times should be related to their different surface and cross-sectional morphologies, which means that the coating with low porosity and denser inner layer exhibits better corrosion resistance. This behavior is related to the structural differences of MAO coatings with different thicknesses. The thicker coating requires a higher break-down voltage, which permits greater energy accumulation and temperature, along with the formation of large pores,

craters, and microcracks, as shown in Fig. 5. As the energy accumulation and local temperature of the coating is rela-tively higher, greater stresses developed during the genera-tion of the thicker coating compared to that developed dur-ing the generation of the thinner one. These defects act as passages for corrosive ions to enter the coating and subse-quently adversely affect the corrosion resistance of the oxide coating. In some cases, it may even result in the coated samples exhibiting lower corrosion resistance than uncoated samples [20,23]. The decrease in corrosion resistance of weldments with the increase of oxidation time is due to the enlargement of pores and microcracks and it is well unders-tood from the EIS data. Thus, we inferred that, with a short-er oxidation time of microarc oxidation, less-porous ceramic coatings with good corrosion resistance can be fabricated.

4. Conclusions

(1) MAO coating thicknesses increase with the increase of oxidation time; however, the growth rate decreases grad-ually toward the end of the process.

(2) The MAO coatings produced in the silicate solution consist of Mg2SiO4 and MgO phases for all oxidation times.


(3) With the increase in oxidation time, the area of pores increases but the number of pores decreases.

(4) From potentiodynamic polarization and electrochem-ical impedance spectroscopy, with the increase in oxidation time and irrespective of the phase composition and thickness of the coating, the corrosion resistance of MAO-coated samples is improved compared to that of the substrate. However, the denser coating with fine pores exhibits supe-rior anticorrosion properties. Deterioration of the corrosion resistance with the increase in oxidation time is observed and is attributed to the enlargement of pores and cracks.

Acknowledgements

The authors would like to acknowledge the Department of Metallurgical and Materials Engineering, National Insti-tute of Technology, Tiruchirapalli for providing the funds and facilities to conduct this research work.

References

[1] B.L. Mordike and T. Ebert, Magnesium: properties — appli-cations — potential, Mater. Sci. Eng. A, 302(2001), No. 1, p. 37.

[2] A. Munitz, C. Cotler, A. Stern, and G. Kohn, Mechanical properties and microstructure of gas tungsten arc welded magnesium AZ91D plates, Mater. Sci. Eng. A, 302(2001), No. 1, p. 68.

[3] N.K. Babu and C.E. Cross, Grain refinement of AZ31 mag-nesium alloy weldments by AC pulsing technique, Metall. Mater. Trans. A, 43(2012), No. 11, p. 4145.

[4] B.A. Shaw, Corrosion Resistance of Magnesium Alloys, ASM Handbook, 2003, p. 692.

[5] P.B. Srinivasan, S. Riekehr, C. Blawert, W. Dietzel, and M. Koçak, Slow strain rate stress corrosion cracking behaviour of as-welded and plasma electrolytic oxidation treated AZ31HP magnesium alloy Autogenous laser beam weldment, Mater. Sci. Eng. A, 517(2009), No. 1-2, p. 197.

[6] D. Sreekanth, N. Rameshbabu, and K. Venkateswarlu, Effect of various additives on morphology and corrosion behavior of ceramic coatings developed on AZ31 magnesium alloy by plas-ma electrolytic oxidation, Ceram. Int., 38(2012), No. 6, p. 4607.

[7] D. Sreekanth, N. Rameshbabu, K. Venkateswarlu, C. Sub-rahmanyam, L.R. Krishnac, and K.P. Rao, Effect of K2TiF6 and Na2B4O7 as electrolyte additives on pore morphology and corrosion properties of plasma electrolytic oxidation coatings on ZM21 magnesium alloy, Surf. Coat. Technol., 222(2013), p. 31.

[8] L.O. Snizhko, A.L. Yerokhin, A. Pilkington, N.L. Gurevina, D.O. Misnyankin, A. Leyland, and A. Matthews, Anodic processes in plasma electrolytic oxidation of aluminium in al-kaline solutions, Electrochim. Acta, 49(2004), No. 13, p. 2085.

[9] A.L. Yerokhin, X. Nie, A. Leyland, and A. Matthews, Cha-racterisation of oxide films produced by plasma electrolytic oxidation of a Ti–6Al–4V alloy, Surf. Coat. Technol.,

130(2000), No. 2-3, p. 195. [10] E. Matykina, R. Arrabal, P. Skeldon, G.E. Thompson, P.

Wang, and P. Wood, Plasma electrolytic oxidation of a zirco-nium alloy under AC conditions, Surf. Coat. Technol., 204(2010), No. 14, p. 2142.

[11] Y.G. Ko, S. Namgung, and D.H. Shin, Correlation between KOH concentration and surface properties of AZ91 magne-sium alloy coated by plasma electrolytic oxidation, Surf. Coat. Technol., 205(2010), No. 7, p. 2525.

[12] T.F. Chen, W.B. Xue, Y.L. Li, X.L. Liu, and J.C. Du, Corro-sion behavior of friction stir welded AZ31B magnesium alloy with plasma electrolytic oxidation coating formed in silicate electrolyte, Mater. Chem. Phys., 144(2014), No. 3, p. 462.

[13] P.B. Srinivasan, R. Zettler, C. Blawert, and W. Dietzel, A study on the effect of plasma electrolytic oxidation on the stress corrosion cracking behaviour of a wrought AZ61 mag-nesium alloy and its friction stir weldment, Mater. Charact., 60(2009), No. 5, p. 389.

[14] K.P. Rao, G.D.J. Ram, and B.E. Stucker, Improvement in corrosion resistance of friction stir welded aluminum alloys with micro arc oxidation coatings, Scripta Mater., 58(2008), No. 11, p. 998.

[15] H.F. Guo, M.Z. An, H.B. Huo, S. Xu, and L.J. Wu, Micro-structure characteristic of ceramic coatings fabricated on magnesium alloys by micro-arc oxidation in alkaline silicate solutions, Appl. Surf. Sci., 252(2006), No. 22, p. 7911.

[16] H.F. Guo, M.Z. An, S. Xu, and H.B. Huo, Formation of oxy-gen bubbles and its influence on current efficiency in mi-cro-arc oxidation process of AZ91D magnesium alloy, Thin Solid Films, 485(2005), No. 1-2, p. 53.

[17] A. Ghasemi, V.S. Raja, C. Blawert, W. Dietzel, and K.U. Kainer, The role of anions in the formation and corrosion re-sistance of the plasma electrolytic oxidation coating, Surf. Coat. Technol., 204(2010), No. 9-10, p. 1469.

[18] L.S. Wang and C.X. Pan, Characterisation of microdischarge evolution and coating morphology transition in plasma elec-trolytic oxidation of magnesium alloy, Surf. Eng., 23(2007), No. 5, p. 324.

[19] H. Chen, G.H. Lv, G.L. Zhang, H. Pang, X.Q. Wang, H.J. Lee, and S.Z. Yang, Corrosion performance of plasma elec-trolytic oxidized AZ31 magnesium alloy in silicate solutions with different additives, Surf. Coat. Technol., 205(2010), Suppl.1, p. S32.

[20] L. Wang, L. Chen, Z.C. Yan, H.L. Wang, and J.Z. Peng, Ef-fect of potassium fluoride on structure and corrosion resis-tance of plasma electrolytic oxidation films formed on AZ31 magnesium alloy, J. Alloys Compd., 480(2009), No. 2, p. 469.

[21] Y.H. Gu, C.F. Chen, S. Bandopadhyay, C.Y. Ning, Y.J. Zhang, and Y.J. Guo, Corrosion mechanism and model of pulsed DC microarc oxidation treated AZ31 alloy in simu-lated body fluid, Appl. Surf. Sci., 258(2012), No. 16, p. 6116.

[22] M.B. Kannan, W. Dietzel, R. Zenga, R. Zettler, and J.F. dos Santos, A study on the SCC susceptibility of friction stir welded AZ31 Mg sheet, Mater. Sci. Eng. A, 460-461(2007), p. 243.

[23] T.B. Wei, F.Y. Yan, and J. Tian, Characterization and wear- and corrosion-resistance of microarc oxidation ceramic coat-ings on aluminum alloy, J. Alloys Compd., 389(2005), No. 1-2, p. 169.

Improving the corrosion properties of magnesium AZ31 alloy ...ijmmm.ustb.edu.cn/fileKWYJYCLXB/journal/article/ijmmm/2017/5/PD… · ported that the corrosion resistance of weldments

Documents