Corrosion Science 50 (2008) 3569–3575 Contents lists available at ScienceDirect Corrosion Science journal homepage: www.elsevier.com/locate/corsci Evaluation of the corrosion resistance of anodized aluminum 6061 using electrochemical impedance spectroscopy (EIS) Yuelong Huang a, Hong Shih b,*, Huochuan Huang b, John Daugherty b, Shun Wu b, Sivakami Ramanathan b, Chris Chang b, Florian Mansfeld a,* a Corrosion and Environmental Effects Laboratory (CEEL), The Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, CA 90089-0241, USA b Lam Research Corporation, 4400 Cushing Parkway, Fremont, CA 94538, USA a r t i c l e i n f o Article history: Received 9 July 2008 Accepted 5 September 2008 Available online 18 September 2008 Keywords: A. Aluminum B. EIS B. SEM C. Oxide coatings a b s t r a c t The corrosion resistance of anodized Al 6061 produced by two different anodizing and sealing processes was evaluated using electrochemical impedance spectroscopy (EIS). The scanning electron microscope (SEM) was employed to determine the surface structure and the thickness of the anodized layers. The EIS data revealed that there was very little change of the properties of the anodized layers for samples that were hard anodized in a mixed acid solution and sealed in hot water over a 365 day exposure period in a 3.5 wt% NaCl solution. The specific admittance As and the breakpoint frequency fb remained constant with exposure time confirming that the hard anodizing process used in this study was very effective in providing excellent corrosion resistance of anodized Al 6061 over extended exposure periods. Some minor degradation of the protective properties of the anodized layers was observed for samples that were hard anodized in H2SO4 and exposed to the NaCl solution for 14 days. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction The corrosion resistance of aluminum and aluminum alloys can be greatly increased by forming a thick oxide layer through the anodizing process in which the aluminum sample is anodically polarized in an appropriate electrolyte to form an aluminum oxide. In the anodizing process sulfuric acid is commonly used as the electrolyte solution to grow the oxide layer to greater thickness than that of the naturally formed film [1,2]. The oxide layer formed in this process has a duplex structure consisting of the inner barrier layer and the outer porous layer. The barrier layer is very thin and dense. The outer porous layer is a much thicker, porous oxide that has a close-packed hexagonal cells structure. A sealing process is necessary to improve the corrosion resistance. The pores in the outer oxide layer can be sealed in steam and boiling water or in various cold sealing solutions such as nickel acetate and dichromate [3–5]. A study of the effects of different sealing methods on the corrosion resistance of anodized Al alloys has been conducted by Zuo et al. [5]. The authors analyzed the outer oxide layers that were sealed by boiling water, stearic acid, potassium dichromate * Corresponding authors. Tel.: +1 213 740 3016; fax: +1 213 740 7797 (F. Mansfeld), Tel.: +1 510 572 2257/299 0283 (H. Shih). E-mail addresses: hong.shih@lamrc.com (H. Shih), mansfeld@usc.edu (F. Mansfeld). 0010-938X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.corsci.2008.09.008 or nickel fluoride and concluded that the boiling water sealing method provided higher corrosion resistance [5]. Electrochemical impedance spectroscopy (EIS) is a powerful method to characterize the corrosion resistance of anodized Al alloys [6–11]. The impedance spectra can be fitted to an appropriate equivalent circuit (EC) which allows to obtain the time dependence of important properties of the anodized surface layers such as the capacitance as well as the resistance of the barrier and the porous layers. Mansfeld et al. [6] studied the protective properties of oxide layers on Al 2024, 6061 and 7075 that were produced by anodizing in H2SO4 solutions. Sealing was carried out in hot water, nickel acetate, yttrium acetate or a saturated cerium acetate solution in an attempt to replace dichromate sealing. The corrosion properties of the oxide layers for these Al alloys were evaluated by EIS during exposure in a 0.5 NaCl solution. The results of this study [6] revealed that the pore resistance Rpo obtained from the EIS data can be used to determine the corrosion resistance of sealed anodized aluminum alloys. The thicknesses of the barrier layer and the porous layer can be evaluated using the experimental values of the capacitance of the barrier layer (Cb) and the porous layer (Cpo), respectively. Rpo values exceeding 2  105 ohm cm2 were considered to be indicative of properly anodized and sealed samples for common applications [6]. For the semiconductor industry Rpo values for anodized aluminum used as a plasma etching chamber coating should meet or even exceed 5  106 ohm cm2 [12]. 3570 Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 Dasquet et al. [10] and Moutarlier et al. [11] indicated that EIS was an important technique for evaluating the corrosion resistance of anodized Al alloys. The results of Moutarlier’ study [11] showed that the thickness of the barrier layer obtained from the EIS method was the same as that determined from transmission electron microscopy (TEM). The extent of degradation of the anodized layers during exposure to a corrosion solution can be monitored as a function of exposure time by simple methods that do not require the recording of an entire impedance spectrum in a wide frequency range. These methods include recording of the breakpoint frequency which is defined as fb = 1/(2pCpoRpo) at the phase angle U = 45° in the frequency region in which the impedance is determined by Cpo [6,13– 15]. A single frequency impedance test for anodized aluminum is described by ASTM B 457 [16]. The specific admittance As, which is the inverse of the impedance at 1000 Hz normalized to the surface area, can be used as a qualitative measure of the coating thickness and an indicator of an oxide layer that was properly sealed [16,17]. Low values of As indicate poor sealing quality of the porous layer [16]. Otero et al. [17] used EIS to evaluate aging of cold-sealed aluminum oxide films formed on pure aluminum. They also determined As values and conclude that As = 10 lS/cm2 was indicative of a porous layer that was properly sealed in hot water [17]. For a porous layer with e = 36 a value of As = 10 lS/cm2 corresponds to a porous layer thickness of 20 lm according to As = 200/dpo (lS/ cm2) [17]. Although many studies have been conducted to improve the corrosion resistance of anodized aluminum alloys, the protective properties of the oxide layer still need to be further improved to meet the challenges faced by the semiconductor industry. In order to improve the long-term corrosion resistance of anodized Al 6061T6, two different hard anodizing (type III) processes have been evaluated. 2. Experimental methods 2.1. Materials Samples of Al 6061-T6 with dimensions of 10 cm  10 cm  0.3 cm were degreased by soaking in a detergent for 10 min followed by deionized (DI) water (2 Mohm-cm) rinsing for 2 min and hot DI water rinsing at 45 °C for 2 min. Three different types of anodized Al 6061 samples were prepared by Lam Research Corporation: three production tank samples (PT), two R&D tank samples (RDT) and two sulfuric acid hard anodized (SAHA) samples. 2.2. Anodizing procedures The oxide layers of hard anodized samples are usually produced at high current densities in a tank that contains sulfuric acid near 0 °C. The hard anodizing (type III) [18] process produces porous oxide layers that are thicker than 25 lm. The anodizing procedures used for the PT and RDT samples followed the hard anodizing process but used a mixed acid solution that mainly contained sulfuric acid (H2SO4) and oxalic acid (H2C2O4). The ramping of the anodizing voltage for the PT and RDT samples was different from that used for anodizing of the SAHA samples. The final voltage of the anodizing process stopped at about 106 V, while it stopped at 60 V for the SAHA samples. The three PT samples were anodized in a production tank which is a large size tank and allows optimization of anodizing parameters. The two RDT samples were anodized in the R&D tank which is a small size tank using the same anodizing procedure that was used for anodizing of the PT samples. The R&D tank does not allow complete optimization of tank condition control [19]. The two SAHA samples were produced following the standard type III anodization process. After anodizing, the samples were first rinsed by cool DI water for 5 min (2 Mohm-cm DI water), hot DI water at 45 °C for 5 min (2 Mohm-cm DI water), and then high purity DI water (18 Mohm-cm) for 10 min. 2.3. Sealing procedures After complete rinsing, the samples were immersed in a hot high purity DI water tank for sealing. The sealing time was about 3 min per lm in oxide thickness. The bath temperature was controlled at 98 °C or higher and the pH was maintained between 5.7 and 6.2. After hot DI water sealing, the samples were moved into a class 100 cleanroom. Isopropyl alcohol (IPA) was used to wipe the samples with class 100 cleanroom wipes. High purity DI water (18 Mohm-cm) was used for additional rinsing for 5 min followed by drying with nitrogen gas that was filtered with a 0.1 lm filter. The samples were then baked in a class 100 cleanroom compatible oven at 110 °C for 30 min, removed from the oven and cooled down to room temperature [19,20]. 2.4. Test methods Impedance spectra were collected at the open-circuit potential (OCP) in a three-electrode cell in which the test sample was placed on the bottom with an exposed area of 20 cm2. A stainless steel electrode was used as the counter electrode and a SCE as the reference electrode. The impedance spectra were obtained with a BSAZahner IM6 unit using a frequency range between 1 MHz to 1 mHz and an ac signal amplitude of 10 mV. Seven samples (three PT, two RDT and two SAHA) have been tested by EIS during exposure to 3.5% NaCl for 14 (RDT and SAHA samples) or 365 days (PT samples). The surfaces of exposed and unexposed samples have been evaluated using SEM (Cambridge Model Stereoscan 360). Observations of the cross sections of the sample were also made to determine the thickness of the oxide layer. 3. Results and discussion Fig. 1 shows some of the impedance spectra that were obtained for the three types of anodized Al 6061 samples for an exposure time of 365 days or 14 days in the Bode plot format in which the logarithm of the impedance modulus |Z| and the phase angle U are plotted vs. the logarithm of the frequency f of the applied ac signal. Very little difference can be detected in the spectra for the PT and RDT samples and there was hardly any change of the impedance for each sample with exposure time. The spectra for the SAHA samples differed from those determined for the other samples since a third time constant appeared at intermediate frequencies. The third time constant might be due to fine cracks and defects in the oxide layers that extend into the base metal. The spectra for the PT and RDT samples are representative for samples that have been properly anodized and sealed in hot water. The very high impedance values at intermediate frequencies for the PT and the RDT samples indicate that the outer oxide layer is very well sealed. The spectra for all PT and RDT test panels for a given treatment were almost identical and did not change much with exposure time. These results indicate that the anodizing process was very reproducible and that the sealing process was very effective. Figs. 2 and 3 show the open-circuit potential (OCP) values for the three types of samples. The OCP for the three PT samples increased in the first 14 days of exposure and then remained very stable at values close to zero V vs. SCE (Fig. 2). The OCP values 3571 Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 a b 9 9 8 8 7 7 6 6 5 RDT 5 PT 4 3 2 1 -4 -3 -2 -1 1d 2d 3d 5d 7d 9d 11d 14d 4 1d 7d 40d 142d 260d 365d 3 2 1 0 1 2 3 4 5 6 -90 -90 -80 -80 -70 -70 -60 -60 -50 -50 -40 -40 -30 -30 -20 -20 -10 -10 -3 -2 -1 0 1 2 3 4 5 6 -3 -2 -1 0 1 2 3 4 5 6 0 0 -4 -3 -2 -1 0 c 1 2 3 4 5 6 9 SAHA 1d 2d 3d 5d 7d 9d 11d 14d 8 7 6 5 4 3 2 1 (c) -3 -2 -1 0 1 2 3 4 5 6 -3 -2 -1 0 1 2 3 4 5 6 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 Fig. 1. Bode-plots for PT (a) RDT (b), and SAHA (c) samples for an exposure period of 14 days and 365 days respectively in 3.5 wt% NaCl solution. for the two RDT and two SAHA samples were stable with exposure time of 14 days indicating the high corrosion resistance of these anodized surfaces (Fig. 3). For the SAHA samples the OCP values were close to 0.6 V (Fig. 3) which is similar to the OCP of bare 3572 Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 0.6 Cpo (nF/cm 2) 0.1 E ocp (V) 0 -0.1 0.5 0.4 PT1 PT1 PT2 PT2 PT3 PT3 0.3 0 -0.2 0 100 200 300 50 100 150 200 250 300 350 Time (days) 400 Time (day) 0.7 Fig. 2. Open-circuit potential (OCP) as a function of exposure time for three PT samples. PT1 PT2 Cb ( µ F/cm 2) 0.6 0.2 RDT1 RDT2 0 SAHA1 E ocp (V) SAHA2 PT3 0.5 0.4 -0.2 -0.4 0.3 0 50 100 150 200 250 300 350 250 300 350 250 300 Time (days) -0.6 7.5 -0.8 4 6 8 10 12 14 Time (days) Fig. 3. Open-circuit potential (OCP) as a function of exposure time for two RDT and two SAHA samples. Al 6061. Apparently the substrate metal was exposed to the corrosive solution through fine cracks that run through the anodized layers. Additional differences in the measured OCP values for the three types of anodized Al 6061 samples might be due to different amounts of MgSi2 and MgSiFe second phases present in the anodized layers [19,20]. The impedance spectra were fitted to the equivalent circuit (EC) shown in Fig. 4 using the software ANODAL [6,21]. Cb and Cpo are the capacitance of the inner barrier layer and the outer porous layer, respectively and Rb is the resistance of the barrier layer. Zpo = K(jx)n, where K and n are fit parameters and x = 2pf, is a constant phase element that is used to account for the variations of the properties of the pores in the outer porous layer such as pore diameter, pore depth and degree of sealing [22]. The fit parameter K is used as a measure of the pore resistance Rpo. Cpo Cb PT1 PT2 LogRpo (ohm.cm 2) 2 PT3 7 0 50 100 150 200 Time (days) 12 PT1 PT2 LogRb (ohm.cm 2) 0 PT3 11 10 9 0 Rs 50 100 150 200 350 Time (days) Zpo Rb Fig. 4. Equivalent circuit for the analysis of the impedance spectra for different anodized Al samples. Fig. 5. Time dependence of fit parameters as a function of exposure time for three PT samples. 3573 Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 Cpo (nF/cm 2) 1 0.5 SAHA1 SAHA2 RDT1 RDT2 0 0 2 4 6 8 10 12 14 Tim e (day) 1 The fit parameters Cb, Cpo, Rb and Rpo for three PT samples are plotted in Fig. 5. Cpo had values between 0.3 and 0.6 nF/cm2 and Cb remained between 0.4 and 0.6 lF/cm2. The barrier and porous layer thickness d can be estimated based on 0.5 ð1Þ SAHA1 SAHA2 RDT1 RDT2 0 0 2 4 6 8 10 12 14 10 12 14 12 14 Time (day) 7 LogRpo (ohm.cm 2) where x = po or b; eo = 8.85  1012 F/m and A is the exposed area. For an average value of Cb = 5  107 F/cm2 and e = 10 [23] the thickness db of the barrier layer is estimated to be 18 nm. The thickness of the anodized layer obtained by the SEM image of the cross section for the PT sample was 63 lm (Fig. 6). For an average value of Cpo = 4.6  1010 F/cm2 the dielectric constant e of the porous layer of the PT sample is estimated to be 33. The observed oxide thickness values can be compared with the thickness of the porous layer of 20 lm and the thickness of the barrier layer of 20 nm for Al samples anodized in H2SO4 and sealed in hot water (SA/HWS) [6]. The very high and stable values of Rb and Rpo (Fig. 5) indicate that the sealing process was very effective for the mixed acid anodizing method. Rpo values exceeding 200 kohm cm2 are considered to be indicative of properly anodized and sealed samples [17]. The spectra for the two RDT samples are shown in Fig. 1b. Cpo had final values between 0.3 and 0.5 nF/cm2, while the final values of Cb were between 0.3 and 0.5 lF/cm2 (Fig. 7). Since the two RDT samples were anodized using the same anodization procedure as the two PT samples, the dielectric constant of the porous layer e = 33 can be used to estimated the thickness of the porous layer. The calculated average values of dpo and db were 68 lm and 22 nm, respectively. The very high and constant values of Rpo indicate that the sealing process was very effective (Fig. 7).The fit parameters for the two SAHA samples are also shown in Fig. 7. The Rb and Rpo values for these anodized surfaces were similar for those observed for the RDT samples, however a slow decrease of Rpo was found for the SAHA samples. Cpo and Cb for the two SAHA samples remained more or less constant during the entire exposure time. For sample SAHA1 Cpo and Cb had average values of about 7  1010 F/cm2 and 8.2  107 F/cm2, respectively. The dpo and db values were estimated as 46 lm and 11 nm, respectively. Cpo and Cb for the SAHA2 sample had average values of about 6.3  1010 F/cm2 and 8.2  107 F/cm2, respectively corresponding to values of dpo = 50 lm and db = 11 nm. Otero et al. [17] used the specific admittance As as a qualitative measure of coating thickness and an indicator of an oxide layer that was properly sealed in hot water. A value of As = 10 lS/cm2 was considered to be indicative of an oxide layer properly sealed in hot water. As = 10 lS/cm2 corresponds to an oxide layer thickness of 20 lm according to As = 200/dpo lS/cm2 for an oxide layer with e = 36 [17]. SAHA1 SAHA2 RDT1 RDT2 6.5 0 2 4 6 8 Time (day) 10 LogRb (ohm.cm 2) C x ¼ o x A=dx Cb (µ F/cm 2) Fig. 6. SEM images of the cross section of a PT sample. 9 SAHA1 SAHA2 RDT1 RDT2 8 0 2 4 6 8 10 Time (day) Fig. 7. Time dependence of fit parameters as a function of exposure time for two SAHA and two RDT samples. 3574 Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 500 3 SAHA1 SAHA2 400 A s (µ S/cm 2) RDT1 fb (Hz) RDT2 300 200 PT1 PT2 100 PT3 2 0 0 50 100 150 200 250 300 350 0 2 4 6 8 10 12 14 Time (days) Time (days) Fig. 8. Time dependence of specific admittance As as a function of exposure time for three PT samples. Fig. 10. Time dependence of break point frequencies fb as a function of exposure time for two RDT and two SAHA samples. 6 80 SAHA2 fb (Hz) As (µ S/cm 2) 60 SAHA1 4 RDT1 RDT2 40 PT1 20 PT2 PT3 2 0 2 4 6 8 10 12 14 0 Time (days) 0 Fig. 9. Time dependence of specific admittance As as a function of exposure time for two RDT and two SAHA samples. Since jZ j ¼ 1=xC po It follows that which for As ¼ 1=Z 1000 A ¼ xC po =A ¼ 2pf o =dpo  ¼ 36 leads to 2 As ¼ 200=dpo ðlS=cm Þ ð2Þ ð3Þ ð4Þ where Z1000 is the impedance at 1000 Hz, Cpo is the capacitance of the porous layer and A = 20 cm2 is the exposed area. Based on the calculated value e = 33 for the PT sample, As = 183/dpo (lS/cm2). Figs. 8 and 9 show the time dependence of As for the 7 samples. These data suggest that the porous layer has a thickness of about 71 lm for the RDT and PT samples that did not change much with time. This thickness value is in general agreement with the thickness of the porous layer calculated from the Cpo data. For the two SAHA samples an average value of dpo = 40 lm is determined from the average As values. The time dependence of the properties of the anodized layers during exposure to a corrosive solution can also be estimated from the impedance spectra using the concept of the breakpoint frequency which is defined as fb = 1/(2pCpoRpo) [6,13–15]. For the two samples treated in the SAHA process fb slightly increased with exposure time (Fig. 10) most likely due to the slow decrease of Rpo with exposure time (Fig. 7) which suggests that some conductive paths and defects had developed in the porous layers during the exposure period [24,25]. For the PT and RDT samples fb values were stable for exposure time of 365 or 14 days, respectively indicating that there was very little change of the porous layers (Figs.10 and 11). Figs. 10 and 11 confirm the conclusions based on the analysis of the impedance spectra that the anodized layers for the five sam- 50 100 150 200 250 300 350 400 Time (days) Fig. 11. Time dependence of break point frequencies fb as a function of exposure time for three PT samples. ples that were hard anodized in the mixed acid (R&D tank and the production tank) were very stable and provided excellent corrosion resistance over extended exposure periods. 4. Summary and conclusions EIS has provided valuable information concerning the properties of the inner barrier layer and the outer porous layer of Al 6061 that was hard anodized using two different processes and their changes during exposure to a corrosive solution. The impedance spectra obtained for samples that were anodized in a mixed acid process and sealed in hot water were in very good agreement with spectra that are normally observed for samples that had been properly anodized and sealed [6,17]. There were very little changes of the spectra with exposure time to 3.5% NaCl for the individual samples which suggests that these surfaces were very corrosion resistant. There were also very little differences in the spectra for the different samples in one group, i.e. the three PT samples, which were treated in the same manner. The thickness of the porous layers was higher than that commonly found for Al alloys that were anodized in H2SO4 and sealed in hot water [6,17]. The impedance spectra as well as the As and fb values for the three PT samples were very stable for an exposure time of 365 days showing that these surfaces were very corrosion resistant. The very high and stable values of Rpo for these three samples indicate that the newly developed anodizing and sealing process was very effective. Y. Huang et al. / Corrosion Science 50 (2008) 3569–3575 The impedance spectra and the parameters As and fb for the two RDT samples were very stable during 14 days exposure. The high values of Rpo reflect the high corrosion resistance and the effectiveness of the hard anodizing and hot water sealing processes. The impedance spectra for the two SAHA samples suggest that the oxide layers formed by hard anodizing in H2SO4 were more complex than those for the other samples. The OCP data and the slow decrease of Rpo suggest that the base metal was exposed to the corrosive test solution through fine cracks and other defects. Acknowledgment Y. Huang and F. Mansfeld acknowledge financial support by the Lam Research Corporation. References [1] S.D. Cramer, B.S. Covino Jr, Aluminum Anodizing ASM Handbook 13A Corrosion: Fundamentals Testing and Protection, vol. 736, Materials Park, 2003. [2] L.L Sheir, R.A. Jarman, G.T. Burstein, Coatings Produced by Anodic Oxidation, Corrosion, Corrosion Control, vol 2, third edition, Butterworth-Heinemann, 15 (3) 2000. [3] S. Wernick, R. Pinner, P.G. Sheasby, The Surface Treatment and Finishing of Aluminum and its Alloys, sixth ed., vol. 2, ASM International, 2001. [4] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, National Association of Corrosion Engineers, 1974. [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] 3575 Y. Zuo, P. Zhao, J. Zhao, Surf. Coat. Technol. 166 (2003) 237. F. Mansfeld, G. Zhang, C. Chen, Plat. Surf. Finish. 84 (Dec.) (1997) 72. J. Hitzig, K. Juettner, W.J. Lorenz, W. Paatsch, Corros. Sci. 24 (1984) 945. F. Mansfeld, M.W. Kendig, Corrosion 41 (1984) 490. F. Mansfeld, M.W. Kendig, J. Electrochem. Soc. 135 (1988) 828. J.-P. Dasquet, D. Caillard, E. Conforto, J.-P. Bonino, R. Bes, Thin Solid Films 371 (2000) 183. V. Moutarlier, M.P. Gigandet, B. Normand, J. Pagetti, Corros. Sci. 47 (2005) 937. H. Shih, D. Outka, J. Daugherty, Specification for Hard Anodized Aluminum Coatings Using Mixed Acids for Critical Chamber Components, Lam Research Confidential Specification, January 17, 2006. F. Mansfeld, C.H. Tsai, Corrosion 47 (1991) 958. J.R. Scully, J. Electrochem. Soc. 136 (1989) 979. F. Mansfeld, L.T. Han, C.C. Lee, G. Zhang, Electrochim. Acta 43 (1998) 2933. ASTM B 457, Standard test Method for Measurement of Impedance of Anodic Coatings on Aluminum ASTM Annual Book of Standards, vol. 192, Springer, 2003. E. Otero, V. Lopez, J.A. Gonzales, Plat. Surf. Finish. 83 (1996) 50. J. Edwards, Coating and Surface Treatment Systems for Metals, vol. 38, ASM International, Finishing Publications Ltd, 1994. H. Shih, Private Communications, April and May, 2008. H. Shih, T.C. Huang, J. Daugherty, Lam Research Confidential Technical Report, October 10, 2006. F. Mansfeld, H. Shih, H. Greene, C.H. Tsai, ASTM STP 1188 (1993) 37. F. Mansfeld, Analysis and interpretation of EIS data for metals and alloys, Schlumberger Technical Report 26 (1993). J. Hitzig, K. Juttner, W.J. Lorenz, W. Paatsch, Corros. Sci. 24 (1984) 945. J. Hubrecht, J. Vereecken, M. Piens, J. Electrochem. Soc. 131 (1984) 2010. J.D. Scantlebury, A. Guiseppi-Elie, D.A. Eden, L.M. Callow, Corrosion 39 (1983) 108.