Residual stress in pulsed dc microarc oxidation treated AZ31 alloy

Microarc oxidation coating preparation

The substrate for the MAO coating deposition was made of AZ31 specimen 20×10×1 mm in size. AZ31 has the following chemical compositions: 2·5–3·5 wt-%Al, 0·7–1·3 wt-%Zn, 0·2–1·0 wt-%Mn, 0·05 wt-%Si, 0·01 wt-%Cu and Mg balance. A few substrates were successively ground on various grade SiC papers to a roughness R_a of ∼1·6 μm, ultrasonically cleaned and dried. The MAO coating was processed in the MAO20 equipment (Chengdu PULSETECH Electrical Co., China), which has an adjustable dc pulse source up to 50 kW and comprises a stainless steel container with a sample holder, stirring and cooling system. The AZ31 substrate and the walls of the stainless steel container were used as anode and cathode respectively. An electrolyte prepared from 10 g L⁻¹ Na₃PO₄ solution in distilled water was kept at room temperature during the entire treatment procedure. The coatings were obtained under a constantly applied voltage of 325 V for 5 min at four pulse frequencies of 300, 500, 1000 and 3000 Hz respectively. The pulse/time ratio of all the pulse frequencies is set to be 3∶7 (i.e. 1∶2·333, 0·6∶1·4, 0·3∶0·7 and 0·1∶0·2333 ms respectively).

Coating characterisations

Scanning electron microscopy (Netherlands Quanta200) was used to examine the surface morphologies of the MAO coated specimens, while X-ray diffraction (XRD) (X'Pert PRO MRD, PANalytical B. V., The Netherlands; Cu K_a radiation, 45 kV and 40 mA) was used to determine the coating composition.

Potentiodynamic polarisation tests

The porosity of the MAO coating was assessed by potentiodynamic polarisation tests, in which a computer controlled potentiostat/frequency response analyser (Corrosion Cell Kit, Gamry Instruments, Inc.) was used. The tests were conducted in a typical three-electrode cell in which the coated sample served as a working electrode, a saturated calomel electrode (SCE) as the reference electrode and a graphite rod as the counter electrode. All the tests were carried out in a 1000 mL simulated body fluid (SBF) with ionic concentrations per the compositions: Na⁺, 142·0; K⁺, 5·0; Ca²⁺, 2·5; Mg²⁺, 1·5; Cl⁻, 147·8; , 1·0; , 4·2; and , 0·5 mmol L⁻¹.20 Potentiodynamic polarisation tests were performed after 10 min of exposure at open circuit potential at a scanning rate of 3 mV s⁻¹, starting at −2·1 to −0·5 V after exposing the specimens with a surface area of 4·18 cm² to the SBF for 0·5 h.

Residual stress measurement

The residual stress in the MAO coating was measured by the XRD sin2 ψ method21 using the diffracted peak (220) at 2θ = 62·306° (which corresponds to the MgO phase, Fig. 1). This comparatively high angle diffraction peak (220) was chosen to reduce the measurement error and to ensure high sensitivity to the strain.14, 22 The measurements were carried out at five different values of ψ (0, 7·5, 15, 22·5 and 30°) in the 2θ range from 61·75 to 63·00° with a step size of 0·0125°.

OPEN IN VIEWER

Figure 1

X-ray diffraction patterns for MAO coated AZ31 alloy at different frequencies

The residual stress developed in MAO coatings on the AZ31 substrate was calculated by23 (1) where σ_φ is the residual stress, and E and v are the Young's modulus and Poisson's ratio of the coating layer respectively. For MgO, E = 248 GPa and v = 0·187.24 The parameters θ and θ₀ represent Bragg's angle at the diffraction peak with and without the residual stress, and ψ is the sample tilt angle.

The change in the lattice spacing d at different ψ values results in a corresponding peak shift. The measured 2θ can be plotted against sin2 ψ. Therefore, the residual stress in the coating can be determined by the slope of the linear regression in the 2θ versus sin2 ψ plot.

Surface phase analyses

The XRD patterns of the coatings in each of the four samples, as obtained at different pulse frequencies shown in Fig. 1, suggest that the processing frequency has little influence on the phase composition, and that the coating in each sample has the same compositions: Mg, MgO, MgAl₂O₄ and Mg₃(PO₄)₂. The Mg phase, attributed to the AZ31 substrate, exhibits a different extent of peak in the XRD pattern from sample to sample, while the MgO phase is barely influenced by the pulsed frequency of the dc source. The MgAl₂O₄ phase has a low intensity but can still be detected with peaks at 2θ = 74·133° for the 300 and 500 Hz samples, and becomes undetectable for both 1000 and 3000 Hz samples because (1) a very low amount of Al exists in AZ31, and (2) the discharge at 300 Hz, which is comparatively continuous and prolonged, results in a relatively high temperature in the discharge channels, propitious to the formation of this low symmetry oxide. At higher frequencies, a large amount of Mg₃(PO₄)₂ was observed, suggesting that the electrolyte ions participated in the growth process of the coating.25

Surface morphology

The effect of different pulse frequencies for MAO deposition on the surface morphology of the resultant MAO coatings is illustrated by the SEM images in Fig. 2a–d. Micropores and cracks were observed on the coating surface for all the four frequencies tested, albeit different porosities at each frequency. During the MAO deposition, the pores and cracks served as the microarc discharging channels through which the oxide leaves the surface and then experiences intensive oxygen evolution, which partly resulted in the porous structure of ceramic coatings. The number of pores on the surfaces of the specimen coated at 300, 500 and 1000 Hz is similar (Fig. 2a–c), while for the 3000 Hz sample, the number seems to be the largest. The average pore size and the average crack size decrease with the increase in pulse frequency.

OPEN IN VIEWER

Figure 2

Surface morphology of MAO coatings on AZ31 alloy formed by pulse frequency at a 300 Hz, b 500 Hz, c 1000 Hz and d 3000 Hz

The differences in the extents of microcracks and micropores in the four coatings can be attributed to the intensity of energy used in the MAO process.26 Although all the four pulse frequencies were operated at the same pulse/time ratio, the 3000 Hz case only has a discrete pulse on time of 0·1 ms, as opposed to 0·3 ms (1000 Hz), 0·6 ms (500 Hz) or 1 ms (300 Hz). This indicates that the pulse frequency at 3000 Hz delivers the least intensity of energy (among the four cases tested) and thus resulted in a coating that has relatively smaller sized micropores and microcracks. At 300 Hz, larger pores are produced (Fig. 2a) because of the relatively prolonged pulse on time (1 ms), which is much longer than the lifetime of the spark and thus results in continuous breakdown of longer period. In addition, at 300 Hz, the continuous discharge during the pulse on time leads to higher temperature in the coating layer, and the heavier quenching effects induce microcracks.26

Porosity measurements of MAO coatings

The vulnerability of a coating against an aggressive environment is directly in connection with the per cent of the pores in the MAO coating;27, 28 therefore, porosity measurement for the coating is necessary. The most often used method for measuring the porosity in the coating is equation (2) introduced by Tato and Landolt29 (2) where R_pm is the polarisation resistance of the uncoated sample, R_p is the polarisation resistance of the MAO coated sample, ΔE_corr is the difference of corrosion potential between the coated and uncoated samples and β_a is the anodic Tafel slope of the uncoated sample.

Figure 3 compares the potentiodynamic polarisation curves of AZ31 Mg alloy and the specimens treated by the MAO process at various pulse frequencies. The corrosion potential (E_corr), corrosion current density (I_corr) and anodic/cathodic Tafel constants (β_a and β_c) were extracted directly from the potentiodynamic polarisation curves by the Tafel fit method using the Gamry Echem Analyst software. The polarisation resistance (R_p) can then be obtained per the Stern–Geary equation30 (3) All the parameters are summarised in Table 1.

OPEN IN VIEWER

Figure 3

Tafel curves of MAO coated samples and bare AZ31 alloy after immersion in SBF for 0·5 h

OPEN IN VIEWER

Table 1

Electrochemical data for untreated and MAO coated samples from potentiodynamic polarisation curves

Sample	Ecorr/V	Icorr/μA cm−2	βa/v/decade	βc/v/decade	Rp/Ω cm2
300 Hz	−1·13	6·22	0·890	0·330	16 828
500 Hz	−1·12	6·12	0·888	0·324	16 865
1000 Hz	−1·04	4·60	0·704	0·337	21 541
3000 Hz	−1·05	4·41	0·698	0·333	22 227
AZ31 alloy	−1·41	294·00	2·494	1·124	1146

Using the data in Table 1, the coating porosity (F) was calculated per equation (2), and the results are listed in Table 2. The results show that the porosity (%) proportionally increases with the pulse frequency. These results are in the range of 3·70–5·26%, similar to Cai's report.31 It was found that the 3000 Hz sample has the lowest porosity of 3·7%, meaning that the densest coating in this study was produced at 3000 Hz.

OPEN IN VIEWER

Table 2

Calculated porosity of MAO coatings by equation (2)

Frequency/Hz	300	500	1000	3000
Porosity/%	5·26	5·20	3·78	3·70

Residual stresses

Table 3 lists the calculated residual stresses in the MAO coatings. The stresses are compressive and their magnitudes decrease with the pulse frequency.

OPEN IN VIEWER

Table 3

Residual stresses in MAO coatings deposited on AZ31 substrates at different pulse frequencies

Frequency/Hz	300	500	1000	3000
Residual stress/MPa	−1272±63	−1178±37	−1032±26	−860±6

Because the porosity changes the equivalent Young's modulus and Poisson's ratio of the thin film the residual stress in the thin film is accordingly amended. Stoney equation is used to relate the thin film stress σ_f to the mechanical properties of the film and the substrate32 (4) where the subscripts f and s denote the film and substrate respectively, E and v are the Young's modulus and Poisson's ratio respectively, the parameter k is the curvature of the film–substrate composite and h is the thickness. In this experiment, samples produced at different frequencies have the same thickness h_f. Moreover, the Young's modulus E_s, Poisson's ratio v_s and thickness h_s of the substrate are the same for all coatings. With these constant parameters, equation (4) describes that σ_f is proportional to k. Furthermore, according to the Euler–Bernoulli beam theory, the curvature k is inversely proportional to the equivalent Young's modulus under a fixed bending. The equivalent Young's modulus of a porous material (such as the porous MAO coating in this study) can be related to its porosity by33 (5) where E_p is the equivalent Young's modulus, E is the Young's modulus of the bulk material and F is the porosity. Therefore, equations (4) and (5) together indicate that the residual stress σ_f is inversely proportional to E_p; the smaller the porosity is, the larger the E_p and thus the less the curvature, and accordingly, the less the residual stresses. Table 3 also shows that the residual stress decreases with the increase in frequency. On the other hand, the calculated porosity F per equation (2) suggests that the porosity decreases with the increase in pulse frequency. Therefore, Stoney equation can be applied to explain the experimental data and to extrapolate the residual stress based on the measured porosity. The residual stress should be determined by the porosity instead of the pore size alone.