2022 Volume 63 Issue 8 Pages 1164-1169
To reduce the electrochemical measurement error of magnesium alloys, surface treatment of AZ91 alloys in 0.1 M NaCl solution at pH = 11 at several potential values were investigated. Scanning electron microscope (SEM) was used to observe the surface conditions. Elemental distribution was investigated using an electron probe micro-analyzer (EPMA). The thickness and chemical composition of the coatings could not be accurately determined from the observations. The electrochemical impedance spectroscopy (EIS) results showed that the electrochemical stability and damage of the surfaces treated at −2.2 V were better than those treated at −1.5 V. This result suggests that surface treatment at −2.2 V prior to polarization test measurements results in smaller experimental errors. For the polarization curve measurements, the surface treatment at −2.2 V was effective to reduce the experimental error. This is also reflected in the standard deviation values for each corrosion property. The standard deviation of the corrosion potential of the surface treatment at −2.2 V was 0.009 V, which was clearly smaller than the standard deviation value of 0.0024 V for the surface treatment at −1.5 V. Therefore, it can be concluded that the surface treatment at −2.2 V reduced the experimental error in the measurement of the polarization curve.
Fig. 5 Polarization curves of AZ91 alloy measured after immersion in 0.1 M NaCl (pH = 11) solution for 2 h at surface treatment potentials of −1.5 V (a) and −2.2 V (b).
Magnesium alloys have a density of 1.738 g/cm3, the lowest of all practical metals.1) In addition to this value being approximately two-thirds the density of aluminum and one-quarter the density of iron, they have a number of advantages over other metals, such as a high natural resource content.2) Therefore, these alloys have been widely used in the automotive, aerospace, and mobile phone industries.3,4) There is also a trend to replace existing structural materials with magnesium alloys; thus, the scope of the magnesium alloys’ applicability is expanding.5) On the other hand, magnesium is electrochemically active, with the lowest standard electrode potential among all practically used metals.6) Therefore, to expand the applicability of magnesium alloys, it is necessary to improve their corrosion resistance.7–12) Polarization curve measurements are very effective for evaluating short-term corrosion resistance, and have been used in many studies.13–18) The main advantage of polarization curve measurements is the ability to quantify the relative rates of anodic and cathodic reactions at various potentials. This is essential for elucidating the corrosion mechanisms of Mg alloys and providing an engineering basis for alloy design and selection. It can also be used to determine the individual contribution of each phase to the overall corrosion mechanism of the alloy.19) However, magnesium is very electrochemically active and, moreover, the reactions are instantaneous. Since the surface state changes significantly during the polarization curve measurements of magnesium alloys, the obtained data likely to contain a large number of experimental errors. Experimental factors which affect experimental errors include the concentration of solution components, scan rate and immersion before scanning. In this paper, we investigated suitable surface treatment conditions to reduce experimental errors in electrochemical measurements on magnesium alloys. As a result of the study, it was considered that the experimental error could be reduced by conducting the measurement in a highly alkaline solution and keeping the potential of the sample surface constant before scanning. In this experiment, the sample surface potential was set at −1.5 V and −2.2 V for comparison. The fine settings of the surface treatment conditions were chosen empirically, taking care that the initial state of the surface immediately before scanning was stable for each measurement and that the surface state did not deviate from the initial state during scanning. The aim of this study was to check whether experimental errors can be reduced by surface treatment of magnesium alloys in a high alkaline solution at a constant potential prior to polarization curve measurements. For confirmation, electrochemical impedance spectroscopy (EIS) was employed, which can non-destructively quantify surface film formation and thickness changes over time.16) An electron probe micro-analyzer (EPMA) was also used to observe the elemental distribution of the formed surface film. A scanning electron microscope (SEM) was used to observe the surface conditions. The surface film data from these experiments and the actual polarization curves were used to determine the appropriate surface treatment potential for immersion prior to the start of polarization measurements.
AZ91 magnesium alloy was used as the sample in this experiment. Its chemical composition is shown in Table 1. All specimens were wet polished with emery paper and buffed with diamond paste particles of 3 µm and 1 µm thickness for a mirror finish. Supersonic cleaning was carried out with ethanol and acetone in that order for 10 min each. For electrochemical measurements, a 10 mm × 10 mm electrode was left on the sample surface, insulated with silicone rubber and used as the working electrode. A platinum electrode was used as the counter electrode and an Ag/AgCl (3.33 kmol×m−3 KCl) electrode as the reference electrode. Unless otherwise stated, the measured potentials discussed in this paper were obtained using the Ag/AgCl (3.33 kmol×m−3 KCl) electrode. All electrochemical measurements were performed in a 0.1 M NaCl solution adjusted to pH = 11 adding NaOH. Based on the potential-pH equilibrium diagram20) of Mg in this aqueous solution, it is considered that the main corrosion product of Mg is the hydrated oxide film Mg(OH)2. The test solution was sufficiently degassed with high purity N2 gas in the electrochemical cell before the electrochemical measurements. During the test, the electrochemical cell was kept at room temperature (298 K).
EIS measurements were performed using a potentiostat (Biologic, SP150) to investigate corrosion properties. EIS measurements were performed continuously during the first 24 hours of immersion to observe changes in the time constant. EIS measurements were performed with the working electrode held at respective potentials of −1.5 V or −2.2 V (VS. Ref) until the end of the test. The frequency was measured from 1 MHz to 1 Hz with 10 points per decade and the sinusoidal excitation voltage to the working electrode was 10 mV (VS. −2.2 V or −1.5 V). All the electrochemical tests were repeated at least three times to ensure reproducibility. The EIS results were fitted with Zfit software (Biologic, EC-Lab V11.33) using an equivalent circuit. The polarization curves were measured using a potentiostat (Biologic, SP150) as in the EIS measurements. The working electrode was held at a potential of −1.5 V or −2.2 V for 2 h as surface preparation prior to polarization curve measurements and then scanned from −2.2 V to 0 V. All polarization curves were recorded at a scan rate of 10 mV/s. All the electrochemical tests were repeated at five times to ensure reproducibility. The corrosion potential (Ecorr), corrosion current density (Icorr) and pitting potential (Epit) were obtained from the results of the polarization curve measurements. The corrosion current density was obtained using the Tafel extrapolation method.
The EIS measurements were performed continuously at each surface treatment potential (−2.2 V or −1.5 V) for 24 h after the start of immersion. Nyquist plots for various immersion times are displayed in Fig. 1(a) and 2(a), and Bode plots for various immersion times in Figs. 1(b) and 2(b). Figure 1(d) and 2(d) show the values of resistance of surface film Rf after fitting the EIS results with the Zfit software (EC-Lab V11.33) using the respective equivalent circuits (Fig. 1(c) and 2(c)). However, as summarized in Ref. 21) of Wang et al. EIS measurements of magnesium are controversial in the analysis and interpretation of EIS results for Mg. The thickness of the surface film was calculated by EIS fitting, assuming that the high-frequency time constant originates from the surface film only.21) Without going into depth on the analysis and interpretation, only the values of the resistance of surface film Rf were used in this paper. The Nyquist plot in Fig. 1(a) is characterized by two well-defined capacitive loops, at high and medium frequencies, followed by an inductive loop in the low frequency (LF) range. The two capacitive loops gradually expand with time. The Rf values in Fig. 1(d) are extremely low compared to the values in Fig. 2(d), so the formed film is considered to be extremely thin. The small change in the Rf value suggests that the film thickness is almost constant within the measurement time. This result may indicate that the electrochemical activity of the Mg alloy is suppressed and, accordingly, the rate of film formation is almost balanced by the rate of film dissolution. The Nyquist plot in Fig. 2(a) is characterized only by a capacitive loop at high frequencies. As shown in Fig. 2(d), the value of Resistance of surface film Rf increased rapidly during the first 3 h after immersion and continued to increase with increasing immersion time, reaching a maximum value at 24 h. This can be attributed to the increase in film thickness. The rapid increase of the Rf value is also considered to increase the experimental error of the polarization curve measurement. The EIS results suggest that immersion at a surface treatment potential of −2.2 V is more suitable than immersion at a surface treatment potential of −1.5 V before polarization curve measurements.
EIS results of AZ91 after various immersion time in 0.1 M NaCl (pH = 11 NaOH) of Surface treatment potential −2.2 V: (a) Nyquist plots; (b) Bode plots; (c) Equivalent circuit; (d) Film resistance Rf after fitting.
EIS results of AZ91 after various immersion time in 0.1 M NaCl (pH = 11 NaOH) of Surface treatment potential −1.5 V: (a) Nyquist plots; (b) Bode plots; (c) Equivalent circuit; (d) Resistance of surface film Rf after fitting.
Figure 3 shows SEM micrographs of the cross-section and surface of a sample after immersion for 24 hours at surface treatment potentials of −2.2 V and −1.5 V. The blue curve in the figure shows the boundary between the sample surface and cross-section. The length of the gap between the blue and red curves indicates the depth of corrosion. Corrosion damage was observed near the cross-section of the sample surface and greater damage was observed when immersed at a surface treatment potential of −1.5 V. The length of the corroded area was measured to be 20 µm at a surface treatment potential of −1.5 V and 10 µm at a surface treatment potential of −2.2 V. The results show that the surface condition differs significantly depending on the surface treatment potential. In Fig. 3(a), (b), the Mg(OH)2 film is dehydrated and too thin to be observed. Under alkaline solution as in Ref. 22), the thickness of the film is on the order of nanometres. Therefore, even if the film thickness can be predicted indirectly by EIS, it is difficult to observe it directly by SEM under these conditions. These results suggest that setting the surface treatment potential to −1.5 V is not appropriate for the state of immersion prior to polarization curve measurement.
SEM photographs of AZ91 magnesium alloy after 24 h immersion at surface treatment potentials of (a) −1.5 V and (b) −2.2 V, (Solution: 0.1 M NaCl (pH = 11) Temperature: 298 K, tilt angle 45°).
Figure 4(a), (b) shows the EPMA elemental mapping (Mg, Al, O) results of the sample surface after immersion for 24 h at surface treatment potentials of −1.5 V and −2.2 V, respectively. EPMA was performed for analyzing the elemental distribution in the corroded sample, and for confirming the formation of a protective film. As a result, the most Al-rich β-phase region, the surrounding Al-rich α-phase region, and the most Mg-rich α-phase region were identified on the sample surface.23) In a highly alkaline solution, Mg(OH)2 is mainly formed on the magnesium alloy surface.22) Based on the analysis results, the distribution of oxygen was clearly observed in both cases, in which the potential values were set to −2.2 V and −1.5 V. In an alkaline solution (pH = 11), Mg(OH)2 is always present as a protective film, because the magnesium alloy is in the passive region. Therefore, the distribution of O on the surface detected by EPMA corresponded to the distribution of Mg(OH)2 (corrosion product film). However, owing to the formation of a thin protective film, sufficient characteristic X-rays of O could not be obtained. Therefore, although the protective film’s formation was confirmed by EPMA, it was not possible to quantify the thickness of the formed protective film.
EPMA elemental mapping of AZ91 magnesium alloy surfaces immersed in 0.1 M NaCl (pH = 11) solution for 24 h at surface treatment potentials of −1.5 V (a) and −2.2 V (b).
Figure 5(a), (b) shows the polarization curves obtained after immersion for 2 hours at surface treatment potentials of −1.5 V and −2.2 V, respectively. The vertical axis in the figures is the logarithm of the potential, while the horizontal axis is the logarithm of the current density. Table 2 shows the mean values and standard deviations of the corrosion potential (Ecorr), corrosion current density (Icorr) and pitting potential (Epit) corrosion properties. The overall shape of the polarization curve in Fig. 5(a) shows that there is a large variation between each measurement. The measurement results contain a large experimental error, which is clearly shown in the value of the standard deviation when the surface treatment potential is set to −1.5 V in Table 2. On the other hand, the overall shape of the polarization curves in Fig. 5(b) shows little variation between each measurement. The standard deviation values for each corrosion characteristic in Table 2 are also very small. The standard deviation values for Icorr, which is generally regarded as an evaluation criterion for corrosion resistance, differ by a factor of about 10. The results show that the experimental error can be significantly reduced by starting the measurement of the polarization curve after immersion at a surface treatment potential of −2.2 V. This factor can be attributed to the stability of the surface condition. The results of the previous experiments are summarized and discussed below. When the surface treatment potential was set to −1.5 V, the film formed was thicker, but the corrosion attack was also more active, and the surface area and other factors were considered to have changed significantly during immersion due to localized corrosion damage. The initial state of the surface before the polarization curve measurement started differed greatly from one measurement to the next, which may have increased the systematic error factor in the experimental error. On the other hand, when the surface treatment potential was set to −2.2 V, the film formed was thin, but the corrosion attack was also suppressed and the corrosion damage was relatively small. This is thought to have stabilized the initial state of the surface before the start of the polarization curve measurement and reduced the experimental error value. Thus, immersion at a surface treatment potential of −2.2 V prior to polarization curve measurement is considered to effectively suppress experimental error.
Polarization curves of AZ91 alloy measured after immersion in 0.1 M NaCl (pH = 11) solution for 2 h at surface treatment potentials of −1.5 V (a) and −2.2 V (b).
Surface treatments were applied to AZ91 alloys to reduce experimental errors in electrochemical measurements, and the factors involved were investigated from the results of each experiment. The thickness and chemical composition of the coatings could not be accurately determined from observations. On the other hand, the results of the EIS measurements suggested that: 1) the coating formed during immersion at a surface treatment potential of −2.2 V was very thin and the thickness was almost constant during the measurement time; 2) during immersion at a surface treatment potential of −1.5 V, the coating on the sample surface was thick and the thickness could continue to increase during the measurement time; 3) Immersion at a surface treatment potential of −2.2 V before measuring the polarization curve is more suitable than at −1.5 V, as the experimental error is smaller. The results of actual surface treatment and polarization curve measurements show that immersion in a surface treatment potential of −2.2 V before polarization curve measurement can effectively reduce experimental error.
This study was financially supported by the Grant-in-Aid for Scientific Research (C) (No. 18K04774).
