Effect of Molybdenum on the Corrosion of Low Alloy Steels in Synthetic Seawater

Su-Bin Shin1, Sol-Ji Song1, Young-Woong Shin1, Jung-Gu Kim1,*, Byung-Joon Park2 and Yong-Chan Suh2

1Department of Advanced Materials Science and Engineering, Sungkyunkwan University, 300 Chunchun-dong, Jangan-gu, Suwon, South Korea 2Heavy Plate R&D Team of R&D Center, Hyundai-Steel Company, 167–32 Kodae-Ri, Songak-Eup, Dangjin 343–823, South Korea

The alloying effect of Mo on the seawater immersion corrosion for low alloy steel was investigated using weight loss tests and electrochemical impedance spectroscopy (EIS) in seawater. The Mo-containing low alloy steel showed an excellent corrosion resistance by the long immersion test due to the formation of homogeneous rust layer preventing active dissolution. SEM and XPS analyses were conducted to observe cross-sectional images of rust layer and indentify chemical composition of oxide formed on the surface after immersion test. The results revealed 2− that the MoO4 ions which were oxidized from Mo form the compounds which disturb the approach of aggressive ions. [doi:10.2320/matertrans.M2016222]

(Received June 17, 2016; Accepted September 15, 2016; Published October 21, 2016)

Keywords: corrosion resistance, electrochemical impedance spectroscopy, molybdenum, seawater, X-ray photoelectron spectroscopy

1. Introduction gen reduction reaction which occurs at the interface of oxide and electrolyte is decreased.16) In other words, the electronic High strength low alloy (HSLA) steels, which provides properties of the oxide affect the rate of oxygen reduction high mechanical properties, weldability and corrosion resis- reaction and corrosion rate.17) Therefore, the study on the rust tance, are widely used as structural material such as bridge, layer is important for the design of the corrosion-resistant building, pipe line and so on. Especially, HSLA steels are steel. used in the plants and structures on severe environmental Alloying elements such as Cr, Cu, Ni and P have an effect conditions such as the ocean because it necessitates the use of on properties of the rust layer, resulting in lowered corrosion anticorrosive steels to retard corrosion.1–3) A water ballast rate.18–21) However, the role of Mo on the rust layer has not tank is a compartment lled with ballast water in ship, boat been identi ed. Mo added to stainless steels was incorporated 2− and marine construction manufactured by HSLA steel with in the Cr oxide layer in the form of MoO4 , which improved thermo-mechanical control (TMCP) process. The tank is es- localized corrosion resistance.22–24) For low alloy steel, sential to maintain ship’s balance under different operation Schultze et al. informed that the favorable effects of Mo ap- systems.4) Seawater can be used for ballast water in the tank. peared after long immersion in seawater.25) Whereas it was Therefore, inside of the tank is exposed to severe corrosive reported that Mo had no effect on the low alloy steel in sea- environment under the inuence of the seawater whose tem- water environment.5) Therefore, a detailed study is required perature reached up to 60°C by solar heat.5) in order to solve the controversy about the role of Mo-added A lifetime of water ballast tank can be extended by corro- low alloy steel. sion protection method such as protective coatings and ca- The purpose of this paper is to evaluate the alloying effect thodic protection.6–8) However, it costs a lot to operate and of Mo on the corrosion resistance of low alloy steel in the maintain the protective system.9) For these reasons, it was synthetic seawater by weight loss test, electrochemical test required to develop steels which have excellent corrosion re- and surface analyses. sistance and maintain good mechanical properties such as weldability and mechanical strength compared to commercial 2. Experimental Procedures steels. Low alloy steel was mainly used for structure of ships in- 2.1 Materials and test condition cluding the water ballast tank.10,11) The low alloy steel has Low alloy steels used in the study were produced by a ther- carbon content of less than 0.2 mass% containing mainly Cu, mo-mechanical control process (TMCP). The specimens Cr, Ni, P and Mn of a few weight percent maximum. The al- were heated and hot-rolled through a two-stage controlled loying elements are added in order to enhance corrosion re- rolling process. The rst stage rolling process was conducted sistance and improve lifetime of the steels.5,9,12,13) There have in the recrystallized austenite region and the second rolling been many developments of low alloy steel applied to marine process was implemented in the nonrecrystallized austenite structures based on alloy design. Recently, the steel plates region range above Ar3. After the process of controlled-roll- used for container ships and upper deck of ballast tank were ing, the hot-rolled plate was immediately water-cooled and reported.14,15) then air-cooled. The formation of compact and adherent rust layer is im- The chemical compositions of the steels were shown in Ta- portant to enhance corrosion resistance of low alloy steel.12) ble 1. The steel plates of 1.5 cm thickness were cut into If the steel surface is protected by the dense oxide layer, oxy- 1 cm × 1 cm pieces. The pieces of the steels were subsequent- ly polished mechanically to 600-grit silicon carbide (SiC) pa- * Corresponding author, E-mail: [email protected] per and washed with ethanol and distilled water. All experi- Effect of Molybdenum on the Corrosion of Low Alloy Steels in Synthetic Seawater 2117

Table 1 Chemical composition of the low alloy steels (mass%). using monochromatic Al Kα energy source. All of specimen Specimen C Si Mn P S Mo Fe surfaces were observed after the immersion for 30 days in the Blank steel 0.07 0.3 1 0.012 0.003 - Balance synthetic seawater. 0.05 Mo steel 0.07 0.3 1 0.012 0.003 0.05 Balance 3. Results and Discussion 0.1 Mo steel 0.07 0.3 1 0.012 0.003 0.1 Balance 0.2 Mo steel 0.07 0.3 1 0.012 0.003 0.2 Balance 3.1 Corrosion properties The effect of Mo concentration in low alloy steel was ex- amined in the synthetic seawater at 60°C. The weight loss ments were carried out at the temperature of seawater in a tests were conducted for 15, 30 days of immersion period. water ballast tank, 60°C which was always maintained using The corrosion rate was calculated by the following equa- water bath, and under an aerated condition. A synthetic sea- tion:28) water solution for the experiments was prepared by a method 87,600W according to the ASTM D1141 standard.26) Corrosion rate(mm/y) = (1) Atρ 2.2 Weight loss tests where W is the weight loss (g), A is the area of exposure Weight loss tests were implemented on low alloy steels in (cm2), t is the immersion time (hour) and ρ is the density (g/ 27) 3 accordance with ASTM G31-72. The initial mass (mi) of cm ). the steel pieces (1 cm × 1 cm × 1.5 cm) was measured. Each The average corrosion rates with error bars calculated by specimen was immersed in the synthetic seawater solution for weight loss tests are shown in Fig. 1. After 15 days of immer- 15 and 30 days by hanging with a plastic wire. After the im- sion period, all of Mo-containing steels enhanced corrosion mersion period of 15 and 30 days, the specimens cleaned with resistance beside blank steel. However, the corrosion rate of ethanol and distilled water were pickled in a solution for 0.2 mass% Mo steel was only lower than that of blank steel 10 min. The cleaning solution was a mixture of 3.5 g hexam- after 30 days of immersion period. The corrosion rates of 0.05 ethylene tetramine and 500 mL HCl to which distilled water and 0.1 mass% Mo steel were similar to that of blank steel. was added to make it 1000 mL. The specimens were subse- Therefore, it is obvious that corrosion resistance of low alloy quently degreased in ethanol using an ultrasonic cleaner for steel was improved by the addition of Mo (0.2 mass% or 10 min, then cleaned with distilled water and dried by drying more). machine. Finally, the nal mass (mf) of these specimens was measured. To ensure reproducibility of the test, the experi- 3.2 Electrochemical behavior ment was repeated at least two times. Figure 2 shows EIS Bode plots of blank steel and three steels with Mo contents for 30 days in the synthetic seawater 2.3 Electrochemical tests at 60°C. The Bode plot represented impedance parameter (Z) A three-electrode electrochemical system was employed and phase angle as a function of the frequency.29) In Bode for electrochemical tests. The system was comprised of a plot, the Z of high frequency implies solution resistance, steel specimen with the surface area of 0.8 cm × 0.8 cm, two whereas the Z of low frequency is total resistance including graphite rods and a saturated calomel electrode (SCE) as the solution and polarization resistance. The solution resistance working, counter and reference electrodes, respectively. An of all specimens during 1 day has similar values and the little open-circuit potential (Eocp) of the specimen was measured differences are due to the distance between specimen and ref- for about 2 hours to reach the stable electrochemical state be- erence electrode. Also, the increase of solution resistance fore measurements. with immersion time is caused by the change of resistivity of Electrochemical impedance spectroscopy (EIS) measure- solution. As the conductive ions such as chloride, sulfate in ments for the interpretation of corrosion behavior were car- the solution react with specimen, the concentration of con- ried out using an EG&G PAR VMP2 potentiostat/galvanos- ductive ions is decreased. For all of specimens, the polariza- tat. The impedance tests were conducted after recording tion resistance generally decreased as the immersion time open-circuit potential during 10 min of the exposure of the increased, which means a reduction in corrosion resistance. It specimen to the synthetic seawater. The impedance spectra shows different results of weight loss tests. Zou et al. investi- were measured over a frequency range from 100 kHz to gated great deviations of electrochemical estimate from 30) 10 mHz with an amplitude (Va) of 10 mV. The operation was weight loss result after long-term immersion. It was ex- done once in 1 day, 15 days and 30 days and repeated at least plained that β-FeOOH in inner rust layer with higher reduc- two times for the reliability of experimental data. tion reactivity can attribute to the strengthening of cathodic reduction reaction, which increases corrosion reaction due to 2.4 Surface analyses small polarization for electrochemical test.31) Therefore, the The surface analyses were carried out to trace the mecha- Z value did not precisely reect corrosion resistance of the nism for the alloying effect of Mo-added low alloy steel. The specimens for EIS test in the synthetic seawater due to the analyses were conducted to observe surface morphology of interference of rust. the specimen and investigate element distribution in rust layer However, EIS is one of the useful techniques in long im- using a scanning electron microscope (SEM, S-3000H, Hita- mersion tests, because it does not interfere with the experi- chi). The chemical composition of rust layer was identi ed by mental system.32) Especially, the Bode plot provides a plain X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo) explanation of behavior dependent on electrochemical fre- 2118 S.-B. Shin, et al. quency.33) Through interpretation of EIS Bode plot, it pro- rust on Mo-containing steel. The extent of splitting decreased vides information of surface/electrolyte solution interface. as Mo contents increased. Especially, it seemed that the phase The spectra of all specimens showed two time constants after spectra of 0.2 mass% Mo have one time constant despite the 15 days, indicating that the rust layer was formed at the time increase in immersion time. by a reaction between the steel and electrolyte.34) The split- Figure 3 is the equivalent electrical circuit models for t- ting in the phase angle-frequency curve indicates the pres- ting of the EIS data. The circuit with one time constant is ence of other phase maximum.35) Also, the phase angle max- shown in Fig. 3(a), and the one with two time constants is ima for all specimens decreased away from −90° and shoul- shown in Fig. 3(b).38) The circuit was used for EIS data in 1 der on the phase angle shifted to lower frequency with in- day, 15 days and 30 days of immersion times, respectively. creasing of immersion times. It means non-homogeneity for The circuit of one time constant has the following elements: specimens increased and the rust layer formed by uniform corrosion was thicken.36,37) The phase spectra of Mo-containing steels were not separated more distinctly than that of Mo- free steel (Fig. 2). This is due to the relative homogeneity of

Fig. 1 Average corrosion rates by weight loss tests for 15, 30 days of im- Fig. 3 Equivalent circuit models for interpretation of the EIS data: (a) 1 mersion periods. day, and (b) 15 days and 30 days.

Fig. 2 Bode plots with immersion time: (a) Blank, (b) 0.05 mass% Mo, (c) 0.1 mass% Mo, and (d) 0.2 mass% Mo. Effect of Molybdenum on the Corrosion of Low Alloy Steels in Synthetic Seawater 2119 the solution resistance (Rs), the constant phase elements immersion tests in the synthetic seawater, as shown in Fig. 5. (CPE1), charge transfer resistance (Rct) and Warburg imped- In rust layer, the epoxy and corrosion product are co-existed: ance (W) which means oxygen diffusion controlled system. black area is epoxy and grey area is corrosion product. The Whereas, the circuit of two time constants has the following specimens without Mo and with 0.05 mass% Mo and components: the solution resistance (Rs), the constant phase 0.1 mass% Mo had considerably porous rust layer on the steel elements (CPE1, CPE2), the rust resistance (Rrust) and charge surface. Whereas comparatively dense rust layer was ob- transfer resistance (Rct). The Warburg impedance is rarely ap- served on the specimen with 0.2 mass% Mo, as shown in peared in 15 days and 30 days due to the change of metal Fig. 5(d), which means that the homogeneity of the rust layer surface state. The CPE substitutes the capacitance to consider was greater than that of other steels. These results support the depression angle and the impedance of the CPE can be EIS results, indicating that 0.2 mass% Mo steel had the lower expressed in the following equation:39,40) depression angle. The 0.2 mass% Mo steel can be expected to improve corrosion resistance due to the formation of nonpo- 1 n 1 Z = ( jω) − (2) rous and protective rust layer on the surface. CPE Y 0 Figure 6 shows the XPS peaks of the specimens after im- where Y0 is the magnitude of the CPE and n is the depression mersion tests for 30 days in synthetic seawater. The analysis parameter (0 ≤ n ≤ 1). The CPE displays ideal capacitance was conducted to search the chemical composition of the rust when n = 1. In other words, the parameter means the devia- layer. The spectra for Mo-containing steel indicated that Fe, tion from an ideal capacitive behavior of the interface.41) The O and Mo ions existed on the steel surface. The Fe peaks depression angle (α) can be expressed as:42) were detected in all of the specimens including blank steel but the peak shift was observed depending on Mo contents. The 2 (1 n) α = − (3) Mo-containing steels show the peak at the lower binding en- π ergy. Table 2 indicates possible chemical compounds in rust An increase of depression angle has been ascribed to the layer and binding energy by the analysis of XPS peaks. For heterogeneity of the surface, which means the increase of sur- Mo-containing steels, not only the two peaks for Fe2p ap- 43–47) face roughness and the formation of porous layers. Fig- peared at 710.5 eV (2p3/2) and 724.1 eV (2p1/2) which indi- ure 4 shows the calculated depression angle of rust layers cates the FeMoO4, but also other ferrous and ferric oxidation with immersion time. The depression angle of blank steel, states are detected which are the same as Mo-free steel. It can 0.05 mass% Mo and 0.1 mass% Mo steel increased slightly be inferred from these data that the iron oxides such as FeO, for 30 days. The depression angle of 30 days was less than FeOOH, Fe2O3 and Fe3O4 existed in rust layer for all of the that of 15 days for 0.2 mass% Mo steel, indicating that rela- specimens.39,49–51) Unlike the Fe peaks, the Mo peaks were tive nonporous rust layer was formed on the steel surface. only detected in the 0.2 mass% Mo steel noticeably. These Hashimoto et al. mentioned that the addition of Mo produced results suggest that Mo compounds were suf ciently precipi- the formation of Mo (VI) oxyhydroxide or molybdate at ac- tated in the rust layer for 0.2 mass% Mo steel, which im- tive sites, which renders a bene cial effect of decrease of the proved corrosion resistance of the steel effectively. In Table 2, dissolution rate. The activity of active sites decreased by the the two peaks of Mo3d spectra only appeared at 231.9 eV adsorption of Mo compounds leads to the formation of homo- (3d5/2) and 235.1 eV (3d3/2) in 0.2 mass% Mo steel, in which geneous layer on steel surface.48) oxidized molybdenum existed in the hexavalent state such as 2− 52) molybdate (MoO4 ) and MoO3. According to the Pour- 3.3 Surface analysis baix diagram of Mo, the Mo (VI) could form molybdate ion The SEM image represents the cross section of rust layer for the Mo-free steel and Mo-containing steels after 30 days

Fig. 5 SEM images of cross sections of the rust layer after immersion tests: (a) Blank, (b) 0.05 mass% Mo, (c) 0.1 mass% Mo, and (d) 0.2 mass% Fig. 4 Depression angle with immersion time. Mo. 2120 S.-B. Shin, et al.

Fig. 6 XPS spectra of the specimens after 30 days immersion tests: (a) Fe2p, and (b) Mo3d.

Table 2 Analysis of the XPS peaks for the surface of the specimens. in weakly alkaline solution.53) Analyses of the XPS spectra Product Binding energy (eV) The effect of molybdate ion on the corrosion inhibitor has been studied extensively in literature.49,52,54–56) Figure 7 Spectrum of Fe2p FeO 710.0, 709.3 shows a schematic representation of the molybdate ion acting FeOOH 711.5, 724.3 as a barrier of electrochemical reaction for Mo-containing Fe O 711.0, 724.0, 710.8 2 3 steel. In initial state, oxygen is diffused to steel surface and it Fe O (Fe2+) 708.3 3 4 dominates the corrosion rate as oxygen diffusion-controlled. 3+ Fe3O4 (Fe ) 710.2 During the anodic reaction, the steel is corroded, then the Fe FeMoO4 710.5, 724.1 ions and Mo ions are dissolved into solution. The Fe ions Spectrum of Mo3d FeMoO4 231.7, 232, 235 generate the iron oxide layer on steel surface and it weakly Na2MoO4 231.9 protects the steel.54–56) As the Mo ions are dissolved suf - MoO3 232.1 ciently, the Mo compounds were adsorbed on oxide layer, − 2− which prevented ion penetration such as Cl and SO4 ions and the layer acts as cation selective phases.49,57) From results of XPS studies, the Mo compounds in the rust layer were 50,58) identi ed as FeMoO4 and Na2MoO4. Especially, the FeMoO4 formed by the reaction with molybdate ion and dissolved Fe2+ ion is an insoluble compound which acts as a stable passive layer.56,59)

4. Conclusions

This study investigated the alloying effect of molybdenum on the corrosion of low alloy steel in the synthetic seawater at 60°C using weight loss tests, electrochemical impedance spectroscopy (EIS) and surface analysis. It comes down to the following conclusions: (1) Weight loss tests revealed that only the 0.2 mass% Mo-containing steel presented much higher corrosion resistance than Mo-free steel. (2) EIS interpretation of the specimens suggested that Mo-containing steel has relative homogeneous rust layer, showing that the spectra of the phase angle were not separated more distinctly. Especially, the depression angle of 0.2 mass% Mo-containing steel decreased during immer- Fig. 7 Schematic of the molybdate ion acting as a barrier of electrochemi- sion periods, which means the steel had nonporous corro- cal reaction in seawater. sion rust layer on the steel surface. (3) Through SEM and XPS analyses, the rust layer of 0.2 mass% Mo-containing steel was much tighter and denser than that of Mo-free, 0.05 and 0.1 mass% Mo-containing steels. The 0.2 mass% Mo-containing steel com- Effect of Molybdenum on the Corrosion of Low Alloy Steels in Synthetic Seawater 2121

posed of mainly of Fe, Fe oxides and Mo compounds 26) ASTM D1141-98: Standard Practice for the Preparation of Substitute Ocean Water, (2013). such as FeMoO4. The positive effect of Mo on the corrosion resistance was 27) ASTM G31-72: Standard Practice for Laboratory Immersion Corrosion Testing of Metals, (2002). presented when the Mo was added about 0.2 mass% as an 28) D. A. Jones: Principles and Prevention of Corrosion, (Prentice Hall, alloying element. Therefore, it is reasonable to add at least New Jersey, 1996) p. 34. 0.2 mass% Mo in low alloy steel in order to increase the life- 29) R. Cottis, S. Turgoose: Electrochemical Impedance and Noise, (NACE time of water ballast tank. Also, the effect of Mo is expected International, Houston, 1999) pp. 36–37. to become more apparent as time goes on. 30) Y. Zou, J. Wang and Y.Y. Zheng: Corros. Sci. 53 (2011) 208–216. 31) H. Antony, S. Perrin, P. Dillmann, L. Legrand and A. Chaussé: Electro- chim. Acta 52 (2007) 7754–7759. Acknowledgements 32) S.T. Keera and M.A. Deyab: Colloids Surf. A Physicochem. Eng. Asp. 266 (2005) 129–140. This research was supported by the Hyundai Steel Compa- 33) K.H. Kim, S.H. Lee, N.D. Nam and J.G. Kim: Corros. Sci. 53 (2011) ny. 3576–3587. 34) N.D. Nam, M.J. Kim, Y.W. Jang and J.G. Kim: Corros. Sci. 52 (2010) 14–20. REFERENCES 35) W.A. Badawy, S.S. El-Egamy and A.S. Azab: Corrosion 53 (1997) 842–851. 1) R.A. Ricks, P.R. Howell and G.S. Barritte: J. Mater. Sci. 17 (1982) 36) D.Y. Lee, W.C. Kim and J.G. Kim: Corros. Sci. 64 (2012) 105–114. 732–740. 37) N.D. Nam and J.G. Kim: Corros. Sci. 52 (2010) 3377–3384. 2) K. Prasad and D.K. Dwivedi: Int. J. Ade. Manuf. Tech. 36 (2008) 475– 38) H. Li, H. Yu, T. Zhou, B. Yin, S. Yin and Y. Zhang: Mater. Des. 84 483. (2015) 1–9. 3) W. Liu, Q. Zhou, L. Li, Z. Wu, F. Cao and Z. Gao: J. Alloy. Compd. 598 39) J.H. Hong, S.H. Lee, J.G. Kim and J.B. Yoon: Corros. Sci. 54 (2012) (2014) 198–204. 174–182. 4) A. Heyer, F. D’Souza, C.F. Leon Morales, G. Ferrari, J.M.C. Mol and 40) M.A. Amin, S.S.A. El-Rehim, E.E.F. El-Sherbini and R.S. Bayoumi: J.H.W. de Wit: Ocean Eng. 70 (2013) 188–200. Electrochim. Acta 52 (2007) 3588–3600. 5) D.W. Kim and H. Kim: Kor. J. Met. Mater. 48 (2010) 523–532. 41) S. Sathiyanarayanan, K. Balakrishnan, S.K. Dhawan and D.C. Trivedi: 6) S.J. Kim, M. Okido and K.M. Moon: Kor. J. Chem. Eng. 20 (2003) Electrochim. Acta 39 (1994) 831–837. 560–565. 42) H.J. Guadalupe, E. G-Ochoa, P.J.M. Rivas, J. Cruz and T. Pandiyan: J. 7) D.C. Cook: Corros. Sci. 47 (2005) 2550–2570. Electroanal. Chem. 655 (2011) 164–172. 8) P.A. Sørensen, S. Kiil, K. Dam-Johansen and C.E. Weinell: J. Coatings 43) K. Jüttner: Electrochim. Acta 35 (1990) 1501–1508. Technology 6 (2009) 135–176. 44) J. Flis, H.W. Pickering and K. Osseo-Asare: Electrochim. Acta 43 9) F. Blekkenhorst, G.M. Ferrari, C.J. Van Der Wekken and F.P. Ijsseling: (1998) 1921–1929. Bri. Corros. J. 21 (1986) 163–176. 45) U. Rammelt and G. Reinhard: Corros. Sci. 27 (1987) 373–382. 10) M.T. Gudze and R.E. Melchers: Corros. Sci. 50 (2008) 3296–3307. 46) D.A. López, S.N. Simison and S.R. de Sánchez: Electrochim. Acta 48 11) J.K. Paik, A.K. Thayamballi, Y.I. Park and J.S. Hwang: Corros. Sci. 46 (2003) 845–854. (2004) 471–486. 47) M. Lebrini, M. Lagrenée, H. Vezin, M. Traisnel and F. Bentiss: Corros. 12) M. Morcillo, K. Diaz, B. Chico, H. Cano and D. de la Fuente: Corros. Sci. 49 (2007) 2254–2269. Sci. 83 (2014) 6–31. 48) K. Hashimoto, K. Asami and K. Teramoto: Corros. Sci. 19 (1979) 13) R.E. Melchers: Corros. Sci. 46 (2004) 1669–1691. 3–14. 14) S. Suzuki, R. Muraoka, T. Obinata, S. Endo, T. Horita and K. Omata: 49) C.R. Clayton and Y.C. Lu: J. Electrochem. Soc. 133 (1986) 2465–2473. JFE Tech. Rep. 2 (2004) 41–48. 50) Z. Zhang, C. Hu, M. Hashim, P. Chen, Y. Xiong and C. Zhang: Mater. 15) K. Shiotani and M. Nakamura: JFE Tech. Rep. 20 (2015) 40–46. Sci. Eng. B 176 (2011) 756–761. 16) M. Stratmann and H. Streckel: Corros. Sci. 30 (1990) 697–714. 51) A.P. Grosvenor, B.A. Kobe, M.C. Biesinger and N.S. Mclntyre: Surf. 17) M. Stratmann and J. Müller: Corros. Sci. 36 (1994) 327–359. Interface Anal. 36 (2004) 1564–1574. 18) T. Misawa, T. Kyuno, W. Suëtaka and S. Shimodaira: Corros. Sci. 11 52) Y.C. Lu and C.R. Clayton: J. Electrochem. Soc. 132 (1985) 2517–2518. (1971) 35–48. 53) M. Pourbaix: Atlas of Electrochemical Equilibria in Aqueous Solutions, 19) H.E. Townsend: Corrosion 57 (2001) 497–501. (NACE International, Houston, 1974) pp. 272–279. 20) X. Chen, J. Dong, E. Han and W. Ke: Mater. Lett. 61 (2007) 4050– 54) M. Meziane, F. Kermiche and C. Fiaud: Bri. Corros. J. 33 (1998) 302– 4053. 308. 21) K. Asami and M. Kikuchi: Corros. Sci. 45 (2003) 2671–2688. 55) E.A. Lizlovs: Corrosion 32 (1976) 263–266. 22) N. Anselmo, J.E. May, N.A. Mariano, P.A.P. Nascente and S.E. Kuri: 56) W.D. Robertson: J. Electrochem. Soc. 98 (1951) 94–100. Mater. Sci. Eng. A 428 (2006) 73–79. 57) M.S. Vukasovich and J.P.G. Farr: Polyhedron 5 (1986) 551–559. 23) M. Seo, G. Hultquist, C. Leygraf and N. Sato: Corros. Sci. 26 (1986) 58) V.I. Nefedov, M.N. Firsov and I.S. Shaplygin: J. Electron Spectrosc. 949–960. Relat. Phenom. 26 (1982) 65–78. 24) K. Sugimoto and Y. Sawada: Corrosion 32 (1976) 347–352. 59) J.S. Kim, P.J. Xiang and K.Y. Kim: Corrosion 61 (2005) 174–183. 25) W.A. Schultze and C.J. Wekken: Bri. Corros. J. 11 (1976) 18–24.