Study of the corrosion product films formed on the surface of Mg–xZn alloys in NaCl solution

Corrosion Science 88 (2014) 215–225

Contents lists available at ScienceDirect

Corrosion Science journal homepage: www.elsevier.com/locate/corsci

Study of the corrosion product films formed on the surface of Mg–xZn alloys in NaCl solution Yingwei Song a,⇑, En-Hou Han a,⇑, Kaihui Dong a, Dayong Shan a, Chang Dong Yim b, Bong Sun You b a b

Environmental Corrosion Center, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China Korea Institute of Materials Science, Changwon 641-831, Republic of Korea

a r t i c l e

i n f o

Article history: Received 8 April 2014 Accepted 14 July 2014 Available online 22 July 2014 Keywords: A. Magnesium A. Alloy A. Zinc B. XPS C. Passive films

a b s t r a c t The corrosion product films formed on the surface of Mg–2Zn and Mg–5Zn alloys in NaCl solution were investigated by electrochemical measurement, scanning electron microscopy (SEM) observation and Xray photoelectron spectroscopy (XPS) analysis. It is found that a compact corrosion product film is formed in the initial stage of immersion, and then the film gradually degrades due to dissolution reaction. The product film formed on Mg–2Zn alloy presents better protection property than that on Mg–5Zn alloy, which can be attributed to the different chemical composition and microstructure of the both alloys. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Mg is a very active metal. The naturally formed oxide film in air mainly consists of MgO whose P/B ratio is below 1, which is not as compact as that formed on the metals of Al and Ti. Also, the oxide films formed in aqueous electrolytes mainly consist of Mg(OH)2 whose P/B ratio is beyond 1 [1], and they cannot still provide effective protection to the Mg substrate, which can be attributed to the following reasons: (1) the formation of Mg(OH)2 film accompanies the intense hydrogen evaluation reaction [2,3], which can degrade the compactness of the film; (2) the Mg(OH)2 film is not chemically stable and susceptible to dissolution in corrosive electrolytes [4]. Thus, the oxide films regardless of formed in air or electrolytes cannot effectively retard the corrosion of Mg substrate. Plenty of studies are carried out to adjust the chemical composition and microstructure of the oxide films by addition of alloying elements. It is found that the naturally formed oxide films on the surface of Mg alloys play a key role in the corrosion process [5–9]. Wang et al. [10] reported that the presence of Al in AZ31 and AZ91 Mg alloys can form Al2O3/Al(OH)3-containing oxide films in NaCl solution, together with the carbonate products to provide a better passivation and retard the chloride-induced corrosion. Zhao et al. [11] found that the surface films formed in corrosive mediums can be more or less effective in hindering the corrosion of AZ91. Liu et al. [12] analyzed the surface film formed in ultrapure ⇑ Corresponding authors. Tel.: +86 24 23915772; fax: +86 24 23894149. E-mail addresses: [email protected] (Y. Song), [email protected] (E.-H. Han). http://dx.doi.org/10.1016/j.corsci.2014.07.034 0010-938X/Ó 2014 Elsevier Ltd. All rights reserved.

water on pure Mg and two Mg–Al intermetallic compounds and found that the film composition on Al3Mg2 was AlMg1.4O0.2(OH)5.4 whilst on Mg17Al12 the composition was AlMg2.5(OH)8. These studies indicate that the alloying element of Al is available to improve the protection performance of the surface films in aqueous solution to some extent. Rare Earths (RE) are also very beneficial alloying elements to enhance the surface films of Mg alloys [13–15]. Pinto et al. [16,17] studied the passive behavior of RE-containing Mg alloys in borate and alkaline solutions. It is found that an amorphous yttrium oxide/hydroxide thick film is formed, which possesses higher stability when compared to magnesium oxide/hydroxide. Liu et al. [18] also found that Y in the Mg–Y binary alloys can improve the protection property of the surface films. Zn is another very popular alloying element [19–21]. There are some studies about the effect of Zn-containing second phases on the corrosion behavior of Mg alloys. Bi et al. [22] researched the corrosion property of Mg–2Dy–xZn (x = 0, 0.1, 0.5 and 1 at.%) alloys and found that the corrosion resistance first decreases and then greatly increases with Zn addition. However, the effect of Zn element on the formation of natural oxide films is paid little attention. Only Ha et al. [23] mentioned that the solid solution of Zn enhanced the protectiveness of the passive film, but the detailed characteristic of the passive film was not reported. In our previous work [24] the microstructure and protection performance of the naturally formed oxide films in air on the surface of the Mg–xZn alloys have been investigated. The results indicate that the oxide films on the Mg–2Zn and Mg–5Zn alloys present a similar chemical

216

Y. Song et al. / Corrosion Science 88 (2014) 215–225

composition, consisting of surface layer of basic magnesium carbonate and MgO following with MgO and ZnO, but the oxide film on the Mg–5Zn is thicker and has more defects than that on the Mg–2Zn. Moreover, it is found that there are visible surface product films naturally formed on Mg–xZn alloys in NaCl solution. However, the microstructure and protection property of the films are still not clear. Thus, the aim of this paper is to investigate the naturally formed corrosion product films on the Mg–2Zn and Mg–5Zn alloys in NaCl solution, and then clarify their protection ability to the Mg substrate. 2. Experimental The experimental materials used for this investigation were the extrusion Mg–2Zn and Mg–5Zn plates which were provided by Korea Institute of Materials Science (KIMS) as reported in the previous paper [24–26]. The samples were successively ground to 3000 grit paper, cleaned in alcohol, and then dried in cool air. The samples for metallographic observation were further ground to 5000 grit paper, finely polished using 1 lm diamond paste and then etched by the solution consisting of 1 g oxalic acid, 1 mL nitric acid, 1 mL acetic acid and 150 mL distilled water. The samples were immersed in 0.1 M NaCl solution for various times and then the surface morphologies were observed using a Phillips XL30 scanning electron microscopy (SEM) equipped with an energy dispersive X-ray spectroscopy (EDX). Electrochemical tests were performed in 0.1 M NaCl solution using an EG&G potentiostat model 273 (Princeton Applied Research, USA) and a model 5210 lock-in amplifier. A classical three–electrode cell was used, with a platinum plate as counter electrode, a saturated calomel electrode as reference electrode and the sample with an exposed area of 1 cm2 as working electrode. The samples were immersed in 0.1 M NaCl solution for 0, 1, 2, 5, 9 and 24 h, respectively, and then electrochemical impedance spectroscopy (EIS) measurements were carried out under open circuit potential after an initial delay of 300 s. The sample named ‘‘0 h’’ represented that immersion treatment was not carried out before EIS measurement expect for the initial delay of 300 s. The scan frequency ranged from 100 kHz to 10 mHz with a sine perturbation amplitude of 5 mV (peak to zero). The data were fitted by a ZSimpWin 3.20 software. The samples immersed in 0.1 M NaCl solution for 1, 2 and 5 h, respectively, were used for potentiodynamic measurements. The measurements started from 200 mV below open circuit potential at a constant scan rate of 0.5 mV s1 and were terminated until a final current density of approximately 10 mA cm2 after an initial delay of 300 s. At least three electrochemical measurements were performed under the same testing conditions for ensuring the reproducibility. The chemical composition of the corrosion product films formed in 0.1 M NaCl solution for 2 h was probed using an ESCALAB 250

(a)

X-ray photoelectron spectroscopy (XPS). Depth profiling was carried out from surface to inner layer. The sputtering rate was 0.1 nm s1. The data were analyzed using a Xpspeak 4.1 software. 3. Results 3.1. Microstructure of Mg–2Zn and Mg–5Zn alloys The microstructures of Mg–2Zn and Mg–5Zn alloys have been reported in our previous papers [24–26]. To facilitate understanding, the surface morphologies of the both alloys are shown in Fig. 1 again. In the case of Mg–2Zn alloy, the alloying element of Zn is completely dissolved into Mg matrix and there are no precipitation phases observed. In the case of Mg–5Zn alloy, some of Zn is dissolved into Mg matrix and the rest is precipitated on the grain boundaries. The precipitation phases were identified by the XRD in our previous work [26]. However, there were no precipitation phases detected probable due to their low volume fractions. Thus, these white precipitation phases are named as MgxZny according to EDX analysis. 3.2. Electrochemical investigation of corrosion product films The corrosion characteristics of the Mg–2Zn and Mg–5Zn alloys immersed in NaCl solution for various times are investigated by EIS measurements as shown in Figs. 2 and 3. The Nyquist plot of the Mg–2Zn alloy with 0 h immersion consists of two well–defined capacitance loops, which describe the features of double electric layer and surface film [27], respectively. At the immersion time of 1 h, the plot still consists of two capacitance loops, but the dimension of the plot increases in comparison with 0 h immersion, indicating that the corrosion resistance of the Mg–2Zn alloy is improved. When the immersion time is 2 h, the dimension of the plot reaches the maximum value. However, the shape of the plot take place great change after the immersion time of 5 h. The second capacitance loop disappears and a large inductance loop is observed. The existence of inductance loop indicates that the Mg substrate suffers attack [28]. With increasing immersion time to 9 h, the shape of the plot keeps constant, namely, a high frequency capacitance loop and a low frequency inductance, but the diameters of the plot decrease. After a longer immersion time of 24 h, the plot nearly coincides with that of 9 h immersion. In view of the Bode plot of frequency vs. impedance, the impedance values first increase and then decrease, and finally tend to be stable with increasing immersion time. It is worth to note that the impedance values of 1 and 2 h immersion are larger than that of 0 h immersion. This case can be attributed to the formation of compact product film in the NaCl solution. As for the Bode plot of frequency vs. degree, two wave crests are observed at the immersion times of 0, 1 and 2 h, implying the existence of two capacitance

(b)

MgxZny phase

Fig. 1. Microstructure of the Mg–2Zn and Mg–5Zn alloys (a) Mg–2Zn; (b) Mg–5Zn.

217

Y. Song et al. / Corrosion Science 88 (2014) 215–225

(a)

-1200

-Z'', ohm cm 2

0h 1h 2h 5h 9h 24 h

-1400 10 Hz

-1000 -800 -600

39.8 mHz

-400 -200 0 200

10 mHz

251 mHz

10 mHz

400 0

500

1000

1500

2000

2500

3000

Z', ohm cm2

(b)

3.0

-0.8

2.5

-0.6 -0.4 -0.2 0.0

2.0

1.5

0h 1h 2h 5h 9h 24 h

-1.2 -1.0

Degree

log |Z|, ohm cm 2

(c)

0h 1h 2h 5h 9h 24 h

3.5

0.2 0.4 -2

-1

0

1

2

3

4

-2

5

-1

0

1

2

3

4

5

log f, Hz

log f, Hz

Fig. 2. EIS of the Mg–2Zn immersed in 0.1 M NaCl for various times (a) Nyquist plot; (b) Bode plot of |Z| vs. Frequency; (c) Bode plot of degree vs. frequency.

(a) -800

0h 1h 2h 5h 9h 24 h

10 Hz

-Z'', ohm cm2

-600 -400 -200 0 631 mHz

200 400

10 mHz

0

500

1000

1500

2000

Z', ohm cm2

3.6

0h 1h 2h 5h 9h 24 h

log |Z|, ohm cm 2

3.4 3.2 3.0 2.8 2.6 2.4 2.2

(c)

-1.2

0h 1h 2h 5h 9h 24 h

-1.0 -0.8 -0.6

Degree

(b)

-0.4 -0.2 0.0 0.2

2.0

0.4

1.8

0.6

1.6 -2

-1

0

1

2

3

4

5

log f, Hz

-2

-1

0

1

2

3

4

5

log f, Hz

Fig. 3. EIS of the Mg–5Zn immersed in 0.1 M NaCl for various times (a) Nyquist plot; (b) Bode plot of |Z| vs. Frequency; (c) Bode plot of degree vs. frequency.

loops. There are one wave crest and one trough at the immersion times of 5, 9 and 24 h, implying the existence of one capacitance and one inductance. The variation trends of the EIS spectra of Mg–5Zn alloy present a great difference in comparison with that of Mg–2Zn alloy as shown in Fig. 3. It is visible with a short inductance loop at 0 h immersion. Moreover, the diameter of the medium frequency capacitance loop is very small. At the immersion time of 1 h, the plot still consists of one high frequency capacitance loop, one medium frequency loop and one low frequency inductance loop, but

the dimension of the plot is larger than that of 0 h immersion. With increasing immersion time to 2 h, the dimension of the plot reduces. Also, the medium frequency capacitance loop disappears and a large inductance loop is found in the low frequency regions, indicating the occurrence of corrosion [28]. In the meantime, the surface morphology of this sample after EIS test is observed optically. It is truth that black corrosion products can be seen. During the immersion times of 5–24 h, the plots still consist of one capacitance loop and one inductance loop, and the dimensions of the plots hardly change.

218

Y. Song et al. / Corrosion Science 88 (2014) 215–225

Q dl Rs

Mg-2Zn – Mg-5Zn

2100

Qf

1800

Rt, ohm cm2

Rt Rf

(a)

1500 1200 900 600

Q dl Rs

300

Qf

Rt

0

Rf

0

5

10

15

20

25

Immersion time, h Fig. 5. Changes of Rt value with immersion time in 0.1 M NaCl.

RL

L

(b) Qdl Rs

R ct

RL

L

(c) Fig. 4. Equivalent circuits of the EIS spectra (a) the Mg–2Zn with 0, 1 and 2 h immersion; (b) the Mg–5Zn with 0 and 1 h immersion; (c) Others.

The Bode plots of frequency vs. impedance show that the impedance values quickly drop from approximately 900–200 X cm2 after 2 h immersion, indicating the rapid failure of the surface film. Then the impedance values nearly keep constant. It is in agreement with the dimension variations of the Nyquist plots. The Bode plots of frequency vs. degree at the immersion times of 0 and 1 h consist of three time constants, namely two capacitance loops and one inductance, whereas there are two time constants (one capacitance and one inductance) at the immersion times of 2–24 h. The EIS results mean that the Mg–5Zn exhibits the best corrosion resistance at 1 h immersion. In order to obtain the further information from the EIS measurements, the spectra are fitted according to the equivalent circuits shown in Fig. 4. The fitting results are listed in Table 1. Rs represents the solution resistance. Constant phase element Q is used in place of a capacitor to compensate the non-homogeneity, which is defined by two values, Y and n. If n is equal to 1, Q is identical to a

capacitor. Rt represents the charge transfer resistance, while Qdl represents the electric double layer capacity at the interface of Mg substrate and electrolyte solution, which is defined by Y0 and n. Qf and Rf represent the capacity and resistance of the corrosion product film formed in NaCl solution, respectively. Qf is defined by Yf and nf. RL and L represent the inductance resistance and inductance, respectively, which are used to describe the low frequency inductance loop. The plots of Mg–2Zn with immersion times of 0, 1 and 2 h are fitted by the equivalent circuit in Fig. 4a. The plots of Mg–5Zn alloy with immersion times of 0 and 1 h can be described by the equivalent circuit in Fig. 4b. In comparison with Fig. 4a, RL and L are added to character the short inductance loop. The existence of the second short capacitance loop means that the surface film can still protect the Mg substrate to some degree. The rest plots, containing one high frequency capacitance loop and one low frequency inductance loop, can be fitted by the equivalent circuit in Fig. 4c. The disappearance of medium frequency capacitance loop indicates that the product film degrades completely. The corrosion resistance of the Mg–2Zn and Mg–5Zn alloys is compared based on the Rt values. It is known that Rt values reflect the dissolution rate of Mg substrate [28]. Usually, the higher Rt values mean the better corrosion resistance. Fig. 5 shows the variation trends of the Rt values. In the case of Mg–2Zn alloy, the Rt reaches the maximum value of 2068 X cm2 at 2 h immersion and then reduces gradually, and finally it keeps a stable level of approximately 270 X cm2. The Rt values of Mg–5Zn alloy exhibit a similar variation trend to that of Mg–2Zn, but it reaches the maximum value at 1 h immersion. In addition, the Rt values of the Mg–2Zn alloy are higher than that of Mg–5Zn alloy at all immersion durations. This result indicates that the corrosion resistance of Mg–2Zn alloy is superior to that of Mg–5Zn alloy.

Table 1 Fitting results of the EIS spectra. h

Rs (X cm2)

Y0 (X1 cm2 s1) 5

n

Rt (X cm2)

Yf (X1 cm2 s1) 3

2Zn 0 1 2 5 9 24

48.12 47.32 49.04 46.43 50.63 47.75

1.098  10 1.436  105 1.253  105 3.943  105 13.63  105 17.01  105

0.9408 0.9113 0.9346 0.8389 0.8586 0.8752

1197 1647 2068 766.4 279.6 269.4

1.777  10 1.306  103 1.602  103

5Zn 0 1 2 5 9 24

47.56 48.53 47.66 46.46 48.16 46.48

1.468  105 2.343  105 14.637  105 18.21  105 18.82  105 20.15  105

0.9271 0.8895 0.8721 0.9045 0.9372 0.8359

976.8 1362 509.4 252.1 212.1 268

2.487  103 7.591  104

nf

Rf (X cm2)

0.7444 0.7865 0.7643

428.8 309.9 825.7

0.5683 0.982

166 194.7

L (H cm2)

RL (X cm2)

3519 1862 8181

627.1 190 610.8

5261 863 1059 1519 5699 1727

835.6 1681 165.6 502.9 306.2 760.6

219

Y. Song et al. / Corrosion Science 88 (2014) 215–225

Mg-2Zn 0 h Mg-2Zn 1 h Mg-2Zn 2 h Mg-2Zn 5 h

-1.3 -1.4

(b)

5h

-1.0

2h

Breakdown

-1.2

1h

Potential, VSCE

Potential, VSCE

(a)

0h

-1.5 -1.6 -1.7 -1.8

Mg-5Zn 0 h Mg-5Zn 1 h Mg-5Zn 2 h Mg-5Zn 5 h

Breakdown

-1.4 -1.6 1h2h

-1.8

-1.9 -7

10

-6

10

-5

10

-4

-3

10

10 -2

Current density, A cm

-2.0 -8 10

5h

0h -7

10

-6

10

-5

10

-4

-3

10

10 -2

Current density, A cm

Fig. 6. Polarization curves of the Mg–2Zn and Mg–5Zn alloys immersed in 0.1 M NaCl solution for 0, 1, 2 and 5 h, respectively (a) Mg–2Zn; (b) Mg–5Zn.

Fig. 7. Surface morphologies of the Mg–2Zn alloy immersed in 0.1 M NaCl for various times (a) (b) 30 min; (c and d) 1 h; (e) (f) 2 h; (g) (h) (i) 5 h; (j) (k) (l) 24 h.

220

Y. Song et al. / Corrosion Science 88 (2014) 215–225

Fig. 7 (continued)

Fig. 6 shows the polarization curves of the Mg–2Zn and Mg–5Zn alloys immersed in 0.1 M NaCl solution for 0 [26], 1, 2 and 5 h, respectively. The corrosion potentials of the both alloys after immersion treatment greatly rise in comparison with that of without immersion treatment. In the meantime, they exhibit the lower anodic dissolution rate but faster cathodic reaction. Especially, the lower anodic dissolution rate can be attributed to the formation of corrosion product films in NaCl solution. The higher cathodic reaction rate is a confusing phenomenon. Maybe the formation of corrosion product films on the surface of Mg alloys can reduce the overpotential of hydrogen evolution reaction. The curves with various immersion times of 1, 2 and 5 h present the different characteristics. The Mg–2Zn alloy shows almost the same cathodic sides at the immersion times of 1 and 2 h, while the sample of 5 h immersion exhibits a higher cathodic reaction rate, corresponding to quicker hydrogen generation rate [29,30]. In addition, the anodic sides are visible with great difference. There exists an ambiguous film breakdown point at the potential of 1.48 VSCE for the sample of 1 h immersion. Distinct film breakdown point is observed at the potential of 1.41 VSCE for the sample of 2 h immersion, but the film breakdown point disappears at the immersion time of 5 h. The polarization curves indicate that the compactness of the surface film formed on the surface of Mg–2Zn alloy in NaCl solution is improved with increasing immersion time to 2 h and the film degrades after 5 h immersion accompanying quick cathodic hydrogen evolution reaction. As for the Mg–5Zn alloy, obvious passivate trend at the anodic side is only observed at the sample of 1 h immersion and then this trend is replaced by quick anodic dissolution reaction after increasing immersion time to 2 and 5 h. Correspondingly, the cathodic hydrogen evolution reaction is accelerated after the degradation of the surface film, which is in accordance with the case of Mg– 2Zn that the corrosion process is controlled by the cathodic hydrogen evolution reaction after the breakdown of surface films. The polarization curves indicate that the optimum passivation trends are observed at the immersion of 2 h and 1 h for Mg–2Zn and Mg–5Zn, respectively. Then the anodic dissolution reaction occurs with increasing immersion time. These results are in good agreement with the EIS measurements.

3.3. Surface morphologies of the corrosion product films The above electrochemical results prove the existence of corrosion product films formed in NaCl solution. Fig. 7 shows the surface morphologies of the Mg–2Zn alloy immersed in NaCl for 30 min, 1 h, 2 h, 5 h and 24 h, respectively. At the immersion time of 30 min, scratches can be observed on the surface of Mg substrate from the low magnification morphology (Fig. 7a), and the corrosion product film can be clearly seen from the high magnification morphology (Fig. 7b). The film consisting of many needle-like particles is compact but does not cover the Mg surface completely. At the immersion time of 1 h (Fig. 7c and d), the Mg substrate is covered by the compact surface film completely. With increasing immersion time to 2 h (Fig. 7e and f), there are not obvious change observed. Only it seems that the film becomes more compact according to the high magnification morphology. However, corrosion is visible in the local regions of the Mg substrate at the immersion time of 5 h (Fig. 7g). The compact product film can still be observed on the undamaged areas (Fig. 7h), but the needle–like particles observed from the high magnification morphology (Fig. 7i) are changed to the flat status, which can be attributed to the dissolution reaction of the product film in NaCl solution. After the Mg–2Zn alloy is immersed in NaCl for 24 h, most of the Mg substrate suffers attack and displays a filiform corrosion feature (Fig. 7j). The severe corrosion regions are visible with loose and flocculent particles (Fig. 7i). These loose corrosion products cannot provide good protection to the Mg substrate. The unattacked areas are still visible with compact surface film (Fig. 7k), but these particles seem to be more flat due to the further dissolution reaction. Fig. 8 shows the surface morphologies of the Mg–5Zn alloys immersed in 0.1 M NaCl for various times. After 30 min immersion, the corrosion product film consisting of many needle-like particles covers the Mg substrate completely (Fig. 8a and b). In comparison with the product film formed on the surface of Mg–2Zn alloy at the same immersion time, the needle-like particles are longer and thinner. After immersion for 1 h (Fig. 8c and d), the needle-like particles grow to a larger dimension but they seem to be looser than that on Mg–2Zn alloy. After immersion for 2 h, corrosion pits are visible on the surface of Mg–5Zn alloy as shown in Fig. 8e. In the

Y. Song et al. / Corrosion Science 88 (2014) 215–225

221

Fig. 8. Surface morphologies of the Mg–5Zn alloy immersed in 0.1 M NaCl for various times (a and b) 30 min; (c and d) 1 h; (e–g) 2 h; (h and i) 5 h; (j) 24 h.

meantime, the verges of the needle-like particles become ambiguous in view of the high magnification morphology in Fig. 8f. After immersion for 5 h, the surface of the Mg–5Zn alloy suffers more severe attack and the corrosion product film is very loose. After immersion for 24 h, the surface of the Mg–5Zn alloy is completely covered by loose and thick corrosion products.

According to the observation of corrosion morphologies, it can be found that the Mg–5Zn alloy takes place corrosion at 2 h immersion while it is visible with corrosion pits on the surface of Mg–2Zn alloy at 5 h immersion. The corrosion product films on the surface of Mg–2Zn alloy are more compact than that on Mg–5Zn alloy.

Y. Song et al. / Corrosion Science 88 (2014) 215–225 5

8.0x10

4

6.0x10

4

Counts, s

(a) 1.0x10

0s 50 s 1460 s 2060 s

C 1s

4.0x10

4

2.0x10

4

-

2-

OH -CO3

2060 s 1460 s

(b)

5

1.6x10

0s 50 s 1460 s 2060 s

Zn 2p

5

Counts, s

222

1.4x10

2060 s 1460 s

5

1.2x10

50 s 5

1.0x10

0s

0.0

50 s

297

294

291

288

285

282

0s

4

8.0x10

1050 1045 1040 1035 1030 1025 1020 1015

279

Binding Engergy, eV

Binding Engergy, eV 4

2.2x10

4

5

0s 50 s 1460 s 2060 s

Cl 2p

2.0x10

4

Counts, s

1.8x10

Cl

4

1.6x10

4

1.4x10

2060 s

4

1.2x10

1460 s

4

1.0x10

50 s

3

8.0x10

(d) 4.0x10

5 5

3.0x10

5

2.5x10

5

2.0x10

5

1.5x10

3

6

4

540

537

534

531

528

525

522

Binding Engergy, eV 0s 50 s 1460 s 2060 s

Mg 1s

5

8.0x10

Counts, s

0s

5.0x10

Binding Engergy, eV

(e) 1.0x10

2060 s 1460 s 50 s

5

1.0x10

0s

6.0x10 210 207 204 201 198 195 192 189

0s 50 s 1460 s 2060 s

O 1s

3.5x10

Counts, s

(c)

ZnO

5

6.0x10

1460 s 5

4.0x10

5

2.0x10

2060 s

50 s 0s

1308

1305

1302

1299

1296

Binding Engergy, eV Fig. 9. XPS analysis of the high resolution spectrum of various elements in the corrosion product film on the surface of the Mg–2Zn alloy after immersed in 0.1 M NaCl solution for 2 h (a) C 1s; (b) Zn 2p; (c) Cl 2p; (d) O 1s; (e) Mg 1s.

3.4. XPS analysis of corrosion product films The chemical composition of the naturally formed corrosion product films on the Mg–2Zn and Mg–5Zn alloys in NaCl solution was detected by depth profiling of XPS. The corrosion product films are too thick. The sputtering is terminated after the Zn element is detected. Fig. 9 shows the XPS depth sputtering of C 1s, Zn 2p, Mg 1s, O 1s and Cl 2p on the Mg–2Zn alloy. Correction is not carried out to the spectra. In the case of C 1s spectra, there is one peak between 288 and 282 eV observed at the sputtering times of 0 and 50 s, which can be attributed to the adventitious carbon signal. Another new peak between 293 and 290 eV appears at the sputtering time of 1460 s. This peak should originate from the constituents in the corrosion product film. According to the standard XPS spectra [31], the C 1s peaks between 293 and 290 eV correspond to basic carbonates. Then the intensity of this peak enhances at the sputtering time of 2060 s. It indicates that the basic carbonate compounds locate in the inner layer of the corrosion product film. There is no Zn element detected at the sputtering times of 0 and 50 s. The Zn peaks are visible until the sputtering time of 1460 s. There exite two Zn 2p peaks, 2p3/2 and 2p1/2. Zn 2p3/2 is the energy peak and Zn 2p1/2 is the satellite peak [31]. The Zn 2p3/2 peak at the binding energy of approximately 1021.4 eV indicates the presence of zinc oxides. Then the intensity of the Zn peak increases at the sputtering time of 2060 s. The variation trends of

the Zn 2p peaks with increasing sputtering time prove that the zinc oxides locate in the inner layer of the corrosion product film. Chlorides are also found in the corrosion product film. In the initial stage of sputtering (0 and 50 s), the Cl 2p peaks are weak. Then the peak is strengthened after the sputtering time of 1460 s. Based on the above analysis, the basic carbonates, zinc oxides and chlorides mainly locate in the interior of the corrosion product film. There is one new peak observed at the low binding energy regions of O 1s with sputtering time of 50 s. It indicates the existence of at least two oxides. Because Zn oxides and basic carbonates are not detected at the surface layer, the two oxides are probably MgO and Mg(OH)2. In view of the fact that the Mg 2p and O 1s peaks are observed at all the sputtering time, the Mg-containing oxides exist though the whole corrosion product film [7]. The above XPS results show that the basic carbonates, Mg-containing oxides, zinc oxides and chlorides mainly locate in the inner layer of the corrosion product film, and the outer layer consists of Mg-containing oxides as well as a small amount of chlorides. The chemical composition of the corrosion product film on Mg–5Zn alloy is also detected by XPS. Fig. 10 shows the XPS depth sputtering of C 1s, Zn 2p, Cl 2p, O 1s and Mg 1s. Besides the C 1s peak between 288 and 282 eV which is associated with the adventitious carbon signal, a new C 1s peak between 293 and 290 eV is observed at the sputtering time of 1500 s, then this peak disappears with increasing sputtering time.

223

Y. Song et al. / Corrosion Science 88 (2014) 215–225

4

7x10

0s 300 s 1500 s 2700 s

C 1s

4

Counts, s

6x10

4

5x10

0s

4

-

4x10

2-

OH -CO3

2700 s

4

3x10

(b)

5

2.4x10

4

5

5

1.5x10

2700 s 1500 s

5

1.2x10

300 s

4

1x10

300 s

1050 1045 1040 1035 1030 1025 1020 1015

Binding Engergy, eV

Cl

4

Counts, s

0s 300 s 1500 s 2700 s

Cl 2p

1.8x10

Binding Engergy, eV

2700 s 4

1.5x10

4

1.2x10

1500 s

(d)

5

4.0x10

0s 300 s 1500 s 2700 s

O 1s

5

3.5x10

5

Counts, s

4

0s

4

9.0x10

296 294 292 290 288 286 284 282 280 278

(c) 2.1x10

ZnO

1.8x10

1500 s

2x10

0s 300 s 1500 s 2700 s

Zn 2p

5

2.1x10

Counts, s

(a)

3.0x10

5

2.5x10

5

2.0x10

2700 s

5

1.5x10

1500 s

5

3

9.0x10

0s

300 s

1.0x10

0s

4

5.0x10

210

207

204

201

198

195

192

189

300 s

540 538 536 534 532 530 528 526 524 522

Binding Engergy, eV

(e) 1.0x10

6

Binding Engergy, eV

0s 300 s 1500 s 2700 s

Mg 1s

Counts, s

5

8.0x10

5

6.0x10

300 s 5

4.0x10

1500 s 2700 s

5

2.0x10

0s

1308

1306

1304

1302

1300

Binding Engergy, eV Fig. 10. XPS analysis of the high resolution spectrum of various elements in the corrosion product film on the surface of the Mg–5Zn alloy after immersed in 0.1 M NaCl solution for 2 h (a) C 1s; (b) Zn 2p; (c) Cl 2p; (d) O 1s; (e) Mg 1s.

As mentioned before, the new C 1s peak corresponds to the basic carbonates. The basic carbonates mainly locate in the medium layer of the corrosion product film. The Zn element is not found at the sputtering times of 0–1500 s. It appears until the sputtering time of 2700 s. Similarly, the Zn peak corresponds to zinc oxides. But the zinc oxides locate in the more inner layer of the corrosion product film in comparison with that in the Mg–2Zn alloy based on the sputtering time. Cl 2p peaks with similar intensity are visible at all sputtering times. It indicates that the chlorides run through the whole corrosion product film [7]. The full width at half maximum (FWHM) of O 1s peaks with longer sputtering time is dramatically wider than that of sputtering 0 s, indicating the presence of at least two oxides [32] such as MgO and Mg(OH)2 due to in the absence of Zn oxides and basic carbonates at the outer layer. The XPS results imply that there is slight difference at the chemical composition of the corrosion product films formed on the both alloys. 4. Discussion The SEM and electrochemical results together prove that there are corrosion product films formed on the surface of the Mg–2Zn and Mg–5Zn alloys in NaCl solution. The films gradually become compact in the initial stage of immersion and then they degrade due to the dissolution reaction after longer immersion time. The

product film formed on the Mg–2Zn at 2 h immersion presents the best protection ability, while the film shows the best protection ability to the Mg–5Zn at 1 h immersion. Moreover, the corrosion product films on the Mg–2Zn can more effectively slow down the corrosion rate than that on the Mg–5Zn. The possible formation and dissolution processes of the surface films are suggested as follows. There are oxide films naturally formed on the surface of the Mg–2Zn and Mg–5Zn alloys in air [24]. But the naturally formed oxide film in air is loose and porous. After the Mg–2Zn and Mg– 5Zn alloys are immersed in NaCl, the chloride ions can penetrate the film along these micro–pores. Then the exposed Mg substrate under the micro–pores will take place slow dissolution reaction in the light of Eqs. (1) and (2) [33].

Anodic reaction : Mg ! Mg2þ þ 2e

ð1Þ

Cathodic reaction : 2H2 O þ 2e ! 2OH þ H2

ð2Þ

Chemical reaction : Mg2þ þ 2OH ! MgðOHÞ2

ð3Þ

The magnesium ions will migrate outward through the initial oxide film, resulting in the deposition of magnesium hydroxides in the outer layer as well as the inner layer. In the meantime, the magnesium oxides in the naturally formed oxide film in air can react with water to form magnesium hydroxides. With increasing immersion time, more and more magnesium hydroxides are deposited to block the micro-pores and increase the film thickness. Thus, the compactness of the corrosion product film is improved in

224

Y. Song et al. / Corrosion Science 88 (2014) 215–225

the initial stage of immersion (Fig. 7b, d and f). Because the oxide film formed in air on the surface of Mg–5Zn has more defects than that on Mg–2Zn [24], the formation rate of the magnesium hydroxides on the Mg–5Zn is faster. As a result, the optimum film is obtained on Mg–2Zn at 2 h immersion while it needs a shorter time of 1 h immersion for Mg–5Zn. After longer immersion time, the surface layer of the product film can take place dissolution reaction in some weak sites due to the attack of chloride ions (Fig. 7i and k), resulting in the presence of local corrosion. This case is proved by the existence of chlorides in the corrosion product films. Then the exposed Mg substrate in the corrosion pits is corroded quickly accompanying the fast hydrogen evolution reaction. As a result, the formed corrosion products are very loose and cannot provide effective protection to the Mg substrate again (Fig. 7g and l). Finally, the whole surface of Mg alloys is damaged gradually (Figs. 7j and 8j). In addition, it is found that the protection ability of the product film on the Mg–2Zn is superior to that on the Mg–5Zn, which can be attributed to their different microstructure and chemical composition. Firstly, there are plenty of MgxZny phases in the grain boundaries of the Mg–5Zn alloy. Also, cracks are visible on the naturally formed product film in air [24]. As a result, the corrosion product films formed on the precipitation phases and cracks exhibit uneven microstructure. Corrosion is susceptible to initiation from these special spots. Especially, the existence of plenty of MgxZny phases can accelerate the cathodic reaction. The quick hydrogen evolution rate can impact the film and lower its compactness. Secondly, the chemical composition of the corrosion product films formed on the Mg–2Zn and Mg–5Zn is slightly different. The concentration of the chlorides in the surface layer of the product film on the Mg–2Zn is lower than that on the Mg–5Zn (Fig. 9). Correspondingly, slightly different microstructure can be observed from the high magnification morphologies of the both films (Figs. 7f and 8g). The film formed on Mg–2Zn is regular and compact, while the film formed on the Mg–5Zn seems to take place dissolution reaction in view of the vague boundary of the pin-like particles (Fig. 8g). Also, the zinc oxides locate in the more inner layer of the corrosion product film in Mg–5Zn than that in Mg–2Zn alloy. Zinc oxides are more chemically stable than that of magnesium oxides. As a result, the surface layer of the corrosion product film on the Mg–5Zn is more prone to corrosion. Based on the above analysis, the corrosion product film formed on the Mg–2Zn alloy presents better protection property than that on the Mg–5Zn alloy.

5. Conclusions There are obvious corrosion product films formed on the surface of Mg–2Zn and Mg–5Zn alloys in NaCl solution. The films gradually become compact in the initial stage of immersion, and then the films degrade with increasing immersion time, finally the Mg substrate suffers attack at the weak sites of the product films. The electrochemical and SEM results indicate that the protection ability of the corrosion product film on the surface of Mg–2Zn alloy is superior to that on Mg–5Zn alloy. The chemical composition of the product films on the both alloys is greatly associated with the naturally formed oxide films in air, namely, the zinc oxides locate in the inner layer, the Mg-containing oxides and chlorides run through the whole film. But the concentration of the chlorides at the outer layer of the film on the Mg–2Zn alloy is lower, and the zinc oxides locate in the more inner layer of the film on the Mg–5Zn alloy. The chemical composition and microstructure of Mg substrates have a great effect on the protection property of the corrosion product films.

Acknowledgements Thanks for the financial supported by Korea Institute of Materials Science and the National Natural Science Foundation of China (No. 51171198) and National Key Basic Research Program of China (No. 2013CB632205). References [1] G.L. Song, Corrosion and protection of Mg alloys, Chemical Industry Press, Beijing, 2006 (in Chinese). [2] G.L. Song, A. Atrens, X.L. Wu, B. Zhang, Corrosion behavior of AZ21, AZ501 and AZ91 in sodium chloride, Corros. Sci. 40 (1998) 1769–1791. [3] J. Chen, J.Q. Wang, E.H. Han, W. Ke, ESEM observation of the process of hydrogen generation around the micro-droplets forming on AZ91 magnesium alloy, Electrochem. Commun. 10 (2008) 577–581. [4] N. Hara, Y. Kobayashi, Formation and breakdown of surface films on magnesium and its alloys in aqueous solutions, Corros. Sci. 49 (2007) 166–175. [5] M. Santamaria, F.D. Quarto, S. Zanna, P. Marcus, Initial surface film on magnesium metal: a characterization by X-ray photoelectron spectroscopy (XPS) and photocurrent spectroscopy (PCS), Electrochim. Acta 53 (2007) 1314–1324. [6] S. Feliu Jr., C. Maffiotte, A. Samaniego, J.C. Galván, V. Barranco, Effect of naturally formed oxide films and other variables in the early stages of Mg alloy corrosion in NaCl solution, Electrochim. Acta 56 (2011) 4554–4565. [7] S. Feliu Jr., C. Maffiotte, A. Samaniego, J.C. Galván, V. Barranco, Effect of the chemistry and structure of the native oxide surface film on the corrosion properties of commercial AZ31 and AZ61 alloys, Appl. Surf. Sci. 257 (2011) 8558–8568. [8] D. Wang, J.S. Zhang, J.D. Xu, Z.L. Zhao, W.L. Cheng, C.X. Xu, Microstructure and corrosion behavior of Mg–Zn–Y–Al alloys with long-period stacking ordered structures, J. Magnesium Alloys 2 (2014) 78–84. [9] A. Atrens, G.L. Song, F.Y. Cao, Z.M. Shi, P.K. Bowen, Advances in Mg corrosion and research suggestions, J. Magnesium Alloys 1 (2013) 177–200. [10] L. Wang, T. Shinohara, B.P. Zhang, XPS study of the surface chemistry on AZ31 and AZ91 magnesium alloys in dilute NaCl solution, Appl. Surf. Sci. 256 (2010) 5807–5812. [11] M.C. Zhao, M. Liu, G.L. Song, A. Atrens, Influence of the b-phase morphology on the corrosion of the Mg alloy AZ91, Corros. Sci. 50 (2008) 1939–1953. [12] M. Liu, S. Zanna, H. Ardelean, I. Frateur, P. Schmutz, G.L. Song, A. Atrens, P. Marcus, A first quantitative XPS study of the surface films formed, by exposure to water, on Mg and on the Mg–Al intermetallics: Al3Mg2 and Mg17Al12, Corros. Sci. 51 (2009) 1115–1127. [13] R. Arrabal, E. Matykina, A. Pardo, M.C. Merino, K. Paucar, M. Mohedano, P. Casajús, Corrosion behaviour of AZ91D and AM50 magnesium alloys with Nd and Gd additions in humid environments, Corros. Sci. 55 (2012) 351–362. [14] N. Birbilis, M.A. Easton, A.D. Sudholz, S.M. Zhu, M.A. Gibson, On the corrosion of binary magnesium-rare earth alloys, Corros. Sci. 51 (2009) 683–689. [15] K. Zhang, X. Zhang, X. Deng, X.G. Li, M.L. Ma, Relationship between extrusion, Y and corrosion behavior of Mg–Y alloy in NaCl aqueous solution, J. Magnesium Alloys 1 (2013) 134–138. [16] R. Pinto, M.G.S. Ferreira, M.J. Carmezim, M.F. Montemor, Passive behavior of magnesium alloys (Mg–Zr) containing rare-earth elements in alkaline media, Electrochim. Acta 55 (2010) 2482–2489. [17] R. Pinto, M.G.S. Ferreira, M.J. Carmezim, M.F. Montemor, The corrosion behaviour of rare-earth containing magnesium alloys in borate buffer solution, Electrochim. Acta 56 (2011) 1535–1545. [18] M. Liu, P. Schmutz, P.J. Uggowitzer, G.L. Song, A. Atrens, The influence of yttrium (Y) on the corrosion of Mg–Y binary alloys, Corros. Sci. 52 (2010) 3687–3701. [19] W.C. Neil, M. Forsyth, P.C. Howlett, C.R. Hutchinson, B.R.W. Hinton, Corrosion of magnesium alloy ZE41-the role of microstructural features, Corros. Sci. 51 (2009) 387–394. [20] Y.W. Song, D.Y. Shan, R.S. Chen, E.H. Han, Effect of second phases on the corrosion behaviour of wrought Mg–Zn–Y–Zr alloy, Corros. Sci. 52 (2010) 1830–1837. [21] G.B. Hamu, D. Eliezer, K.S. Shin, The role of Si and Ca on new wrought Mg–Zn– Mn based alloy, Mater. Sci. Eng. A 447 (2007) 35–43. [22] G.L. Bi, Y.D. Li, S.J. Zang, J.B. Zhang, Y. Ma, Y. Hao, Microstructure, mechanical and corrosion properties of Mg–2Dy–xZn (x = 0, 0.1, 0.5 and 1 at.%) alloys, J. Magnesium Alloys 2 (2014) 64–71. [23] H.Y. Ha, J.Y. Kang, J. Yang, C.D. Yim, B.S. You, Limitations in the use of the potentiodynamic polarisation curves to investigate the effect of Zn on the corrosion behaviour of as-extruded Mg–Zn binary alloy, Corros. Sci. 75 (2013) 426–433. [24] Y.W. Song, E.H. Han, K.H. Dong, D.Y. Shan, C.D. Yim, B.S. You, Microstructure and protection characteristics of the naturally formed oxide films on Mg–xZn alloys, Corros. Sci. 72 (2013) 133–143. [25] Y.W. Song, E.H. Han, K.H. Dong, D.Y. Shan, C.D. Yim, B.S. You, The role of second phases in the corrosion behavior of Mg–5Zn alloy, Corros. Sci. 60 (2012) 238– 245. [26] Y.W. Song, E.H. Han, K.H. Dong, D.Y. Shan, C.D. Yim, B.S. You, The effect of Zn concentration on the corrosion behavior of Mg–xZn alloys, Corros. Sci. 65 (2012) 322–330.

Y. Song et al. / Corrosion Science 88 (2014) 215–225 [27] J. Chen, Y.W. Song, D.Y. Shan, E.H. Han, Study of the corrosion mechanism of the in situ grown Mg–Al–CO2 3 hydrotalcite film on AZ31 alloy, Corros. Sci. 65 (2012) 268–277. [28] C.N. Cao, Principles of corrosion electrochemistry, Chemical Industry Press, Beijing, 2004 (in Chinese). [29] G.S. Frankel, A. Samaniego, N. Birbilis, Evolution of hydrogen at dissolving magnesium surfaces, Corros. Sci. 70 (2013) 104–111. [30] N.T. Kirkland, G. Williams, N. Birbilis, Observations of the galvanostatic dissolution of pure magnesium, Corros. Sci. 65 (2012) 5–9.

225

[31] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics, Inc., Minnesota, 1995. [32] Q. Zhou, Electron Energy Spectroscopy, Nankai University Press, Tianjin, 1995 (in Chinese). [33] N. Dinodi, A.N. Shetty, Electrochemical investigations on the corrosion behaviour of magnesium alloy ZE41 in a combined medium of chloride and sulphate, J. Magnesium Alloys 1 (2013) 201–209.