A.M. Mierisch and S.R. Taylor
Center for Electrochemical Science and Engineering, University of Virginia, Charlottesville, VA 22903
An understanding of the events leading to the local breakdown of organically coated alloys requires the use of local electrochemical and chemical methods. Intrinsic (i.e. naturally occurring) breakdown events on AA2024-T3 coated with a series of neat resins (vinyl, polyurethane, and epoxy) were examined using Local Electrochemical Impedance Mapping/Spectroscopy (LEIM/S) along with supportive techniques. Several revealing events and processes were identified. Several types of breakdown sites were identified that differed in both electrochemical and underfilm chemistry characteristics. The time evolution of these defects was observed as a rise in the local impedance in the first stages of observation. Metastable behavior defects have also been observed. Present effort is focussed on theoretical corroboration of the interpretation of these events.
The initial sites of corrosion on organically coated alloy substrates are likely governed by either chemical heterogeneities within the organic coating , heterogeneities within the alloy (e.g. intermetallics) [2-4], or a juxtaposition of the two, assuming a physically uniform film. An understanding of the relative contributions from these processes to coating breakdown relies on the ability to obtain information on the local chemical and electrochemical processes.
Traditionally, the electrochemical investigation of organic coating behavior has used global techniques to provide insight into the general performance of a coating [5-10]. However, these measurements are averaged over the entire surface and do not provide information about individual sites of corrosion initiation. It is believed that information about these sites will help govern the development of better surface pre-treatment and coating chemistries. Thus, recent efforts have turned towards developing methods that can measure the important electrochemical events at these breakdown sites[11-16].
This paper presents several observations of local electrochemical events associated with the breakdown of intrinsic defects (i.e. naturally occurring) on organically coated AA2024-T3 using Local Electrochemical Impedance Mapping (LEIM) and Spectroscopy (LEIM) in conjunction with other supportive methods. The literal translation of these data provides insight into the initiation and evolution processes of localized underfilm corrosion.
The LEIM/S system used in this investigation employed the five-electrode (split micro-reference electrode) configuration originally designed by Lillard, Isaacs, and Moran , and further developed for specific use on local defects in coated alloys  as described in detail by Mierisch . The probe uses a 5-electrode configuration with two teflon coated Ag/AgCl micro-reference electrodes, each 125mm in diameter, in addition to the standard working, counter and reference electrodes. In mapping, a single, small amplitude (ca. 15 mV) sine wave perturbation of optimal frequency (generally 700 Hz) is applied to a DC potential set at the global open circuit potential of the substrate. The local current density is measured by measuring the potential gradient in solution above the substrate between the two micro-reference electrodes and then converting this potential difference through Ohm�s law to a local current density. The local impedance (or admittance) is determined by comparing the global voltage to the local current density, and an LEI map is generated by moving the probe to discrete sites across the sample and plotting the local admittance as a function of x,y position. LEI spectroscopy operates by holding the probe over a particular x,y position and scanning the sine wave excitation frequency, generally between 10KHz and 1Hz. In both cases the voltage signal from the split micro-reference electrode is amplified to minimize the required excitation signal and maintain a nondestructive character.
All electrochemical measurements were performed on 1 mm thick AA2024-T3 cut into 6.5x6.5 cm panels coated with a neat organic resin. One of several coating chemistries were applied depending on the particular variable to be studied. These coatings were (a) vinyl VYHH copolymer of polyvinyl chloride and polyvinyl acetate, (b) polyurethane coatings made with either a 100% polyether polyurethane or a polyurethane composed of 50% polyether and 50% polyester (both mixed with isocyanate, a dibutyl tin dilaurate catalyst and methyl ethyl ketone solvent), (c) a two-part polyester polyurethane comprised of Desmophen polyester polyol and desmodur aliphatic polyisocyanate, and (d) a two-part epoxy comprised of bisphenol A-epichlorohydrin-based epoxy resin solution with an Epi-cure fatty acid-polyethylenepolyamine based polyamide mixture curing agent. Butyl cellosolve, in varying amounts, was used as a solvent to adjust viscosity in coatings (c) and (d).
All panels were cleaned prior to coating by the following procedure and then left to air dry: scrubbed with acetone, scrubbed with Alconox, cleaned ultrasonically in hexanes for at least 5 minutes, then rinsed with isopropanol. Ultrapure water was used for rinsing between steps. The appropriate coating was then spin cast onto the substrate producing a uniform dried film thickness of 5-20 microns depending on the experiment. Freshly coated samples were dried in a dessicator for at least 48 hours prior to use. A glass reservoir was affixed at its base perimeter to the coating surface with RTV silicone. Samples were exposed primarily to 0.6 M NaCl of ambient aeration, but other solutions including 0.1 M NaCl and 0.1 M KCl were used as needed.
An important observation made in early investigations  was that several different types of defects occurred within a single panel. These defects types were differentiated initially by the color that they attained upon long-term exposure. Two such defects were classified as �red� and �black� blisters. LEIM was used on examples of such blisters in order to evaluate the electrochemical differences between them. Maps of a red and black blister are shown in Figure 1.
Figure 1: LEIM plots showing the differences in electrochemical activity between black (left) and red (right) blisters. Note the higher electrochemical activity of the red blister. These AA2024-T3 panels were coated with vinyl VYHH to 10mm thickness and exposed to 0.6 NaCl. (larger image)
It should be noted that these types of blister (i.e. red and black) were observed on all samples when exposed to chloride solutions regardless of the coating chemistry, however differences in rate of formation, and extent may vary. It is clear from the figure that the red blister is electrochemically more active, while the black blister is relatively passive. After several weeks, the coating was removed from the substrate and the alloy surface at the blister sites examined by an optical microscope. As suspected from the LEIM results, the red blister had indeed incurred significant metal loss, whereas the black blister had caused little damage. An example of the pitting caused by a red blister is shown in an electron micrograph in Figure 2.
Figure 2: Electron micrograph of a pitted region of the substrate beneath the initiation site of a red blister. This substrate had previously been coated with a 10mm thick polyester polyurethane coating. The coating was removed for viewing in the electron microscope. (larger image)
It was hypothesized that LEIM/S could predict which initial defects might become the severe red blisters or the more innocuous black blisters by comparison of initial electrochemical activity to the eventual blister formed. LEIS was performed on defects which were identified through mapping in early, still microscopic stages. The results are shown in Figure 3. The defect which would later become a red blister was even then significantly more electrochemically active than the eventual black blister.
Figure 3: LEIS shows that even in early stages, the red blister is more electrochemically active than the black blister. The smaller circle diameter on this Nyquist plot represents a lower pore resistance. (larger image)
In addition to differences in electrochemical activity, the blisters were also examined for chemical differences. In one experiment, small quantities of solution were drawn from several examples of each type of blister and the ionic content of the underfilm solution was analyzed using capillary electrophoresis (CE). Cu2+, Al3+, and Mg2+ ions were found in the solution of the red blisters whereas only small amounts of Mg2+ and Zn2+ were found in the black blister solution. The cations present in the red blister solution suggest that active dealloying of the substrate or a particular phase within the substrate has taken place [2-4,17]. These results demonstrate that the electrochemical and chemical nature within very small (<1 mm) blisters can be quantified and characterized.
Differences in underfilm pH and potential also differentiate these blister types and substantiate the LEIM/S findings. In a similar experiment, local pH values of the underfilm solutions were obtained for several examples of each type of blister. This was done by inserting an Ir/IrO microelectrode into the blister. Prior to these tests, the electrochemical activity of each blister was mapped by LEIM. Some of the red blisters were found to be low in electrochemical activity, while most were found to be high in activity as expected. In fact, the blisters which were high in electrochemical activity were found to contain solution at pH values of 3-5, whereas the blisters with low electrochemical activity, presumably temporarily repassivated, contained solution of pH 8-9. Local pH measurement provides valuable information about the underfilm solution, but it is also a destructive technique. Once the coating has been breached, the defect will no longer develop as it would have. However, this problem can be avoided through use of the nondestructive LEIM technique.
Figure 4: LEIM of the evolution of a red blister. Note the fluctuation in electrochemical activity representative of metastability. At 52 hours, the open circuit potential dropped dramatically, suggesting a possible breach in the coating. (larger image)
Red blisters generally show metastable behavior, with periods of temporary repassivation or decreased activity occurring between periods of increasing electrochemical activity. Such metastable behavior has been suggested in the literature but has not been specifically investigated. It has been suggested  that the biphasic nature parallels capacitance changes in the coating. Metastability could also be caused by accumulation of corrosion product periodically dispersed by the convection of hydrogen evolution. The former possibility will be investigated by comparing LEIM of a gold disk electrode with and without a coating. Reference standard gold disk electrodes have been created using Si/SiO wafers with embedded gold electrodes. These standards will be discussed in a future paper.
In later stages, metastability is visually observable as growth spurts. The blisters experience rapid growth (visible growth within 1-2 hours) followed by lengthy periods of no outward growth (12-24 hours). In earlier stages, however, when these changes are not visually apparent, such a localized event can only be measured by local electrochemical methods. The observed changes in growth and in pH value of the solution render the possibility unlikely that metastability of the blisters might be due to changes in coating properties such as water uptake. However, preliminary studies have shown that a low (3-5) pH solution increases the rate of transfer of ions across the coating interface. It is possible that variations in pH affect the activity of the blisters and could contribute to metastability.
Repassivation, or �healing� has been suggested to occur on chromate conversion coated (CCC) aluminum alloys and cited as an asset of these coatings . LEIM was used to study the effect of chromate added to solution on the activity of red blisters. In Figure 5, LEIM shows the change in electrochemical activity before and after chromate solution is added.
Figure 5: LEIM of a red blister before and after a chromate solution was added to the bulk electrolyte. 1.2 M Na2CrO4 was added to 0.6 M NaCl. After chromate is added, the admittance decreases, in agreement with a rise in open circuit potential. After 24 hours, the admittance begins to rise, although open circuit potential alone does not indicate this change. (larger image)
Once the chromate was added, the open circuit potential rose to a more noble value as expected, and LEIM showed a decreased electrochemical activity in the blister. LEIS also showed a decrease in electrochemical activity. However, while the open circuit potential remained relatively constant suggesting a continuing repassivation of the blister, LEIM and LEIS showed that, in fact, electrochemical activity began again to increase within the blister. Future studies will focus on metastability of defects on substrates coated first with CCC�s and then with organic coating. Although several investigations have studied CCC coated substrates [20-22], the effect of an organic coating on CCC coated substrates has not been investigated from a local electrochemistry standpoint.
In addition to characterizing overall properties of individual blisters, LEIM was also used to distinguish between anodic and cathodic regions of red blisters. It was observed that blisters initiate at a given site, and then branch into one or more secondary lobes that grow away from the initiation site. LEIM showed that the initiation site is electrochemically active, while the secondary lobes are relatively inactive. An example of this is shown in Figure 6.
Figure 6: An optical micrograph of a red blister with initiation site (lighter region to right) and secondary lobe (darker region to left). LEIM shows the higher activity in the initiation region. Local underfilm open circuit potentials, also shown, confirm the higher activity in the initiation region. (larger image)
Energy dispersive spectroscopy (EDS) also revealed a higher amount of Cu and lower amount of Al and Mg in the general area of initiation (with Cu deposited just outside the initiation pit region). Local open circuit measurements, made with a Ag/AgCl microelectrode similar to the one used in pH measurements, showed a more anodic potential within the initiation region and a more noble potential in the secondary lobe. Although further studies are required to draw an absolute conclusion, it can be presumed that the initiation region is the anode and the secondary lobe/lobes the cathode. The contributing factors to the blister front advancement are presently under investigation.
Whereas later stages of blister development can be characterized by compiling information obtained both from local electrochemical and chemical methods and from optical and global methods, the earliest stages of development and initiation of a blister require the use of local electrochemical methods. By stepping the LEI probe over the surface of a coated alloy panel, LEIM was able to reveal microscopic regions of defect initiation. During this early stage of development, it was found that the impedance had actually increased relative to the surrounding region. LEIM of an example of this is shown in Figure 7.
Figure 7: An example of the initial decrease in admittance, or increase in impedance, in earliest stages of defects. This AA2024-T3 panel was coated with vinyl VYHH to 10 mm thickness and exposed to 0.6 M NaCl solution.
Several explanations for this increase in impedance have been proposed. Although it was suspected that nonuniform field distributions emanating from the defect may have given rise to the increased impedance, this explanation was found to be improbable. The expected field from an equipotential disk was derived and calculated for dimensions similar to those in the experiment. It was shown that the micro-reference electrodes would have to be very close to the surface compared to the radius of the disk (or size of the disk is large relative to the height of the micro-reference electrodes from the surface -unlikely in early stages). If the probe were indeed within this realm, a dip in admittance would originate from an increased base admittance value and substantial peaks from edge effects at the perimeter of the disk would be apparent. A representative plot of theoretical results that would result in an admittance dip is shown in Figure 8. The normal field over an equipotential disk is calculated for different heights above the surface relative to the radius of the disk. Actual admittance plots of defects did not display any of these features. Two additional explanations for the increased impedance are water nucleation under the coating and build-up of corrosion product at the initiation site .
Figure 8: The normal field is calculated theoretically for an equipotential disk at various heights (z) from the surface relative to the radius of the disk (a). Note that a dip does occur when the probe is very close to the surface, but this dip is accompanied by significant edge effects and the dip does not extend below the baseline of the surface.
The local nature of corrosion of coated alloys necessitates the use of local electrochemical and chemical techniques toward understanding individual breakdown events. LEIM/S is capable of characterizing in situ the initiation of and changes in electrochemical activity of a single defect without disturbing blister development. LEIM/S of local breakdown sites has also confirmed previous observations of metastable behavior.
LEIM has identified several different types of defects on coated AA2024-T3 with the two most prevalent examples being a red and a black blister. Through LEIM and LEIS the red blister was found to be more electrochemically active both in early and later stages of development than the black blister. CE analysis revealed that the underfilm solution of the red blisters contained Cu2+, Al3+, and Mg2+ ions whereas the black blister only contained small amounts of Mg2+ and Zn2+ ions. LEIM showed that red blisters often repassivated during development and that this repassivation was found to correspond to an increase in local pH from an aggressive environment of 3-5 to a more benign environment of 8-9. Red blisters caused a visibly severe amount of damage to the substrate as compared to black blisters.
The metastable nature of defects on the coated panels was monitored by LEIM. Several possible explanations for the metastability, including changes in coating properties, accumulation of corrosion product, and variation in local pH were discussed. The addition of chromates to the bulk electrolyte caused an initial repassivation of a red blister. Whereas open circuit potential measurements remained steady after the addition of chromates, LEIM showed that repassivation was only temporary.
Red blisters were observed to initiate in a given region and branch into secondary lobes. LEIM showed that the initiation sites were anodic to the secondary lobes, suggesting a possible cathodic head in growth. Local OCP measurements and EDS measurements of the composition of the affected substrate confirmed the differentiation of the red blisters into an active initiation region with corrosion taking place and relatively passive secondary lobes.
LEIM revealed an initial increase in impedance of blisters in the earliest stages of development. Possible explanations, including nonuniform field distribution, water nucleation, and corrosion product accumulation were presented. Nonuniform field distributions were found to be an unlikely cause. Defects in such early, microscopic stages were located through LEIM of the substrate surface.
The authors would like to thank the Air Force Office of Scientific Research (AFOSR) for their sponsorship of this project as well as Jiangnan Yuan and Robert Kelly for their contributions of CE analysis.
1. J.E.O. Mayne, D.J. Mills, J. Oil Col Chem Assoc., 58:155 (1975).
2. G.S. Chen, M Gao and R.P. Wei, Corrosion, 52(1):8 (1996).
3. P.Schmutz and G.S. Frankel, J. Electrochem. Soc., 145(7):2295 (1998).
4. K. Nisancioglu, J. Electrochem. Soc., 137(1):69 (1990).
5. H. Leidheiser, Jr., Corrosion, 39(5):189 (1983).
6. F. Mansfeld, M.W. Kendig, S.Tasi, Corros. Sci., 23(4):317 (1983).
7. S.R. Taylor, IEEE Trans. Elec. Insul., 24(5):787 (1989).
8. S. Haruyama, M. Asari, Tsuru, �Corrosion Protection by Organic Coatings�, The Electrochemical Society, Proc. 87-2 (1987), p.197.
9. S. Hirayama, S. Haruyama, Corrosion, 47(12):953 (1991).
10. C.T. Chen and B.S. Skerry, Corrosion, 47(8):598 (1991).
11. I. Annergren, D. Thierry, F. Zou, J. Electrochem. Soc., 144, 1208 (1997).
12. E. Bayet, F. Huet, M. Keddam, K. Ogle, and H. Takenouti, J. Electrochem. Soc., 144(4),L87 (1997).
13. H. S. Isaacs, Y. Ishikawa, Corrosion 85, paper 55, Nace International, Houston, TX (1985).
14. R. S. Lillard, P.J. Moran and H.S. Isaacs, J. Electrochem. Soc., 139(4), 1007 (1992).
15. A. M. Mierisch, S.R. Taylor, in Electrochemically Based Microstructural Characterization, Vol. 500, Ed. By R.A. Gerhardt, M.A. Alim, S.R. Taylor, Materials Rsearch Society, Warrendale, PA, p.35 (1998).
16. M. W. Wittman and S.R. Taylor, in Advances in Corrosion Protection by Organic Coatings II, Ed. By J.D. Scantlebury and M.W. Kendig, The Electrochem. Soc., PV 95-13:158-168 (1995).
17. R.G. Buchheit, R.P. Grant, P.F. Hlava, B. Mckenzie, G.L. Zender, J. Electrochem Soc., 144(8):2621 (1997).
18. J.A. Grandle, S.R. Taylor, Corrosion, 50(10):792 (1994).
19. J.E. Castle, in Advanced Monitoring and Analytical Techniques, Ed. By F. Mansfeld, W.H. Smyrl, Proceeding of Corrosion, (1997).
20. J. Zhao, G. Frankel, R.L. McCreery, J. Electrochem. Soc., 145(7):2258 (1998).
21. G.P. Halada, C.R. Clayton, M.J. Vasquez, J.R. Kearns, M.W. Kendig, S.L. Jeanjaquet, G.G. Peterson, G.S. McCarthy, G.L. Carr, in Critical Factors in Localized Corrosion III, Electrochemical Society, PV98-17 (1998).
22. A.E. Hughes, R. Taylor, B. Hilton, Surface and Interface Analysis, 25(4):223 (1997).