Read Localized Electrochemical Methods Applied to Cut Edge Corrosion text version

3654

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

Localized Electrochemical Methods Applied to Cut Edge Corrosion

K. Ogle,*,z V. Baudu, L. Garrigues, and X. Philippe

Irsid, Usinor Research, 57283 Maizières-lès-Metz, France Current density and pH mapping techniques have been used to characterize the chemical and electrochemical phenomena which occur on the cut edge of galvanized steel. pH variations between 7 and 11 were observed, primarily due to the formation of hydroxyl ions by the cathodic reaction. Zinc-based corrosion products precipitated in zones of intermediate pH were identified as ZnO and 3Zn(OH)2 2ZnCO3 for model samples. The efficiency of these corrosion products as cathodic inhibitors was demonstrated by the absence of cathodic activity at open circuit and a 300 mV negative shift of the onset potential for hydrogen formation in the affected zones. The cathodic current, localized over the steel, was independent of potential, consistent with a diffusion-limited reduction of oxygen. The anodic current, localized over the zinc, varied with potential, with a Tafel slope of 44 mV/decade for an order of magnitude decrease of potential below open circuit. The addition of SrCrO4 to the electrolyte increased the Tafel slope to 63 mV/decade, consistent with a passivating inhibitor on the anode surface. © 2000 The Electrochemical Society. S0013-4651(00)02-083-8. All rights reserved. Manuscript submitted February 22, 2000; revised manuscript received May 19, 2000. This was in part Paper 582 presented at the Honolulu, Hawaii, Meeting of the Society, October 17-22, 1999.

When a painted steel sheet is cut at the factory, the bare metal is exposed in cross section on the cut edge, where it becomes a corrosion risk regardless of the barrier properties of the paint layer on the face. Such "cut edge corrosion" is a well known problem particular to coil coated steel products1,2 and may represent a major, but largely ignored, obstacle to the development of advanced on-line coating technologies such as plasma polymerization. Therefore, there is an immediate interest in understanding the corrosion protection mechanisms which are provided by the coating layers against this form of corrosion. In a more general way, corrosion on the cut edge of galvanized steel products can be considered as a worst case scenario for sacrificial protection by zinc coatings, because of the very unfavorable anode to cathode surface area ratio. The zinc layer is on the order of 5 to 20 m thick as compared to around 800 to 1000 m for the steel substrate. Therefore secondary mechanisms of protection involving the formation of protective films from zinc corrosion products and from water soluble inhibitors in the paint2,3 play an important role on the cut edge of coil coated steel. The efficiency of chemical inhibition is highly dependent on the chemical and electrochemical environment which is found in the vicinity of the cut edge. The strong galvanic couple set up between the steel and the zinc leads to large electrical field variations, and the close proximity of the anodic and cathodic reactions leads to significant changes in the chemical environment including increases in [OH ] and [Zn 2]. The distribution of current is further altered by the precipitation of zinc-based corrosion products and the release of water soluble inhibitors present in the paint. Although extensive analytical studies on the nature of the corrosion products have been undertaken in recent years,4-7 their role in inhibiting the electrochemical reactions has been given much less attention. In this work, we have used in situ electrochemical techniques with local resolution, the scanning vibrating electrode technique (SVET) and scanning pH microscopy, to investigate the interrelationship between electrochemical activity, pH changes, and corrosion-product precipitation. In particular, localized polarization curves were obtained for the anodic and cathodic reactions in order to determine the rate-limiting processes on the cut edge and to observe the inhibiting effect of corrosion products. Preliminary results have been previously presented concerning the polarization behavior8,9 and localized pH measurements10 on the cut edge. Experimental Sample preparation and model sample design.--Cut edge phenomena were simulated in the laboratory by embedding, in an epoxy

* Electrochemical Society Active Member. z E-mail: [email protected]

resin, a commercial electrogalvanized low alloy carbon steel sample typical of that used in the automotive industry. The cut edge was exposed in cross section to form an electroactive area of 10 0.81 mm. The steel thickness was approximately 800 m, and the zinc coating ranged from 5 to 10 m. The cut edge surface was polished to a 1 m finish and particular care was taken during polishing to avoid spreading the zinc onto the steel surface. Single and double sided electrogalvanized steel were used as indicated in the text. Electrogalvanized steel was chosen for this work because of its high level of purity and reproducibility as an industrial product. For galvanic coupling experiments, a model cut edge sample was prepared. This consisted of separate steel and zinc samples mounted in the epoxy resin with approximately 2 mm of separation between them. A 250 m zinc foil (Goodfellow, 99.95 %) and a sample of low alloy carbon steel of 800 m thickness, were used. The edges of the sample were exposed to the electrolyte and the two metals were short-circuited by an external circuit. The large-area Zn/steel electrode was prepared by vapor depositing a 5 m thick Zn coating over one-third of a 5 10 mm low alloy carbon steel electrode. The deposition was done by masking the surface not to be coated by an adhesive tape (3M PTFE 5490, based on a silicone adhesive with upper temperature limit of 204 C), which was removed after the zinc deposition. The sample was sputtered with Ar ions before the deposition of the zinc in order to increase the metal/metal adherence. This sample was stored in a dissicator prior to use. The electrodes thus produced were exposed to NaCl solutions of variable concentration as described in the text. Because of a poor Zn to steel adherence, the electrodes were not submitted to any additional cleaning regime. All solutions used deionized water (>5 M cm) and reagent grade NaCl. The electrolytes were aerated, and all experiments were conducted at ambient temperature. Current density measurements.--A commercial SVET system from Applicable Electronics was used in this work.11,12 The probe consisted of an insulated Pt-Ir wire with a Pt black deposit at the tip on the order of 20 m diam. The probe is vibrated at different frequencies in the parallel and perpendicular direction to the surface, with an amplitude on the order of 20 to 40 m. The two-dimensional (2D) vibration is accomplished by use of piezoelectric wafers driven by sinewave oscillators. The parallel and perpendicular components of the local electric field are considered proportional to the parellel and perpendicular components of the current vector, Jx and Jz, and are measured by independent lock-in amplifiers. The second horizontal component of current, Jy, is not measured with this system. The calibration parameters necessary to convert the electric field data into current were determined from a one-point calibration, performed by taking a measurement at a fixed distance from a point

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

3655

source of current (usually with 60 nA imposed) for which the theoretical current density is readily calculated. The signals are digitized using a 16 bit analogue-to-digital (A/D) converter. Scanning is done with a 3D stepper motor micromanipulator. Local current.--Potential curves.--For the measurement of localized current-potential curves, a Pt wire counter electrode was added to the cell, in the form of a ring around the working electrode. A standard saturated calomel electrode was used with an agar/agar bridge made from the 0.03 M NaCl electrolyte used in the study. Under normal conditions, the electrolytic solution is connected to ground by a Pt wire reference electrode which reduces noise in the measurement. The use of a potentiostat required that the electrolyte be separated from the ground. The potential was controlled with a Solartron 1286 potentiostat. For quantitative determination of the current potential curve for the anodic reaction, the vibrating electrode was positioned over an anodic zone by locating in the x, y plane the position where Jz was a maximum and Jx was zero. The potential of the sample working electrode was adjusted, and the Jz value was monitored as a function of time until it was stable. The measurement was taken after 1 min of integration. Jx 0 was continuously verified throughout the experiment. pH microelectrodes.--pH maps were obtained using liquid membrane glass capillary pH electrodes, fabricated in house using techniques similar to previously published work.13-16 Optical and electron micrographs of the glass capillary tip are shown in Fig. 1. The electrodes were prepared from single-barreled glass capillaries with an outer diameter of 1.5 mm (WPI, Ref. TW150-4). The shape and small diameter (on the order of 5 m) of the tip was obtained using a commercial micropipette puller (Sutter Instrument Company). The capillaries were then silanized17 in an oven at 250 C by injecting 50 L of N,N-dimethyltrimethylsilane (Sigma Aldrich, Ref. 41716) in the glass preparation chamber. The ionophore was a Hydrogen I cocktail B

(Fluka, Ref. 95293) which has a functional pH range of 5.5 to 12. The internal solution consisted of a neutral buffer composed of 0.01 M KH2PO4 in 0.1 M KCl. The liquid membrane was introduced at the tip of the probe with a length of about 20 to 30 m. A silver chlorinated wire was immersed into the internal solution to serve as the internal reference electrode. A conventional Ag/AgCl/3 M KCl electrode was used as an external reference electrode. The microelectrodes were mounted on the SVET system to control the position and program the sweeping of the microelectrode. A special preamplifier of 1015 input impedance was used to measure the potential. Potential acquisition was performed with a delay of 4 s at each measuring position (a typical scan of 12 20 points would require approximately 16 min). Electrodes were calibrated before and after each experiment using a range of commercial pH buffers allowing a conversion of the measured potential to pH. The electrodes were discarded if the calibration slope was less than 45 mV. Of the electrodes used, the mean calibration slope was found to be 57.3 7.3 mV in good agreement with the Nernstian response of 59 mV, although with a rather large dispersion. Results Validity of localized measurements in the vicinity of the cut edge.--Large pH and electric field variations may be expected to occur in the vicinity of a cut edge, and it is therefore important to verify the validity of SVET measurements in the presence of pH gradients due to electrochemical reactions, as well as the sensitivity of pH measurements to ionic currents. In this section we describe an empirical approach to validation which makes use of a series of point-source electrodes in which the ionic currents and pH gradients either occur simultaneously, or can be separated. Model point-source electrodes are commonly used to calibrate and test the vibrating electrode.11,12 If we consider an ideal point source of electrical current located at the origin of the x, y, z coordinates, and suspended in a homogeneously conductive medium, the magnitude of the current density vector, |J| (in amperes/unit area), at any point (x, y, z) in the medium will be given by the total current of the point source, I (amperes), divided by the area of a sphere around the point source, upon which the point x, y, z is located: |J| I /4 (z2 x2 y2). For the experiment, z is the vertical dimension, and the x, y plane is parallel to the sample surface. The x axis is experimentally defined by the direction of the parallel vibration of the vibrating electrode. It follows that the distribution of the x and z components of current density, Jx and Jz are given by J x ( x, y) | J | sin 4 (z J x ( x, y) | J | cos 4 (z I x

2

x2 I x

y2 ) 2

3

[1]

2

x2

y2 ) 2

3

[2]

Figure 1. Electron (upper) and optical (lower) micrographs of pH electrode tip.

where is the angle of the vector from x, y, z to the origin with respect to the x axis. An equation analogous to Eq. 1 and 2 could be written for the y component. Note that for a point source on a nonconducting plane, the current is distributed over a hemisphere and the 4 is replaced by 2 in Eq. 1 and 2. "Model" point source electrodes used in the laboratory will only approximately simulate the ideal case. In addition to the finite size and other geometrical constraints, the passage of electricity into the electrolyte will involve the generation of ionic species which can affect the electrical field measurement by locally changing the conductivity, or by introducing a parasitic chemical potential term into the measured electrical potential.18 The reverse problem may be true for pH measurements. The electric fields generated by the corrosion activity could be detected and interpreted as variations in pH. This problem has led some researchers to use double-barrel capillary electrodes,14,15 although for the large pH gradients measured in this work, the electric-field component should be negligible. In order to verify that the basic assumptions of these techniques are correct, model electrodes were used in which the electric fields

3656

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

and the chemical changes could be separated. Therefore, two types of electrodes were constructed. First, metal point source electrodes consisting of a polymer-coated stainless steel needle with only the extreme tip exposed (r 10 m). This is the standard system used for calibration of the SVET. For this electrode, the electric-field distribution due to ionic currents and pH gradients appear simultaneously in the vicinity of the electrode. Second, glass capillary point-source electrode consisting of a Ag/AgCl wire placed within a glass capillary of about 5 m inner tip diameter. In this way, only electric-field variations are produced in the vicinity of the capillary tip, which serves as a funnel to form an electric-point source far removed from the faradaic currents at the surface of the Ag wire. Furthermore, the cathodic reaction occurs by reduction of AgCl, distributed over a large surface area and occurs without changing the pH. Figure 2 shows the variation of pH measured with the micro-pH electrode in the vicinity of the metal point source (left) and the glass capillary point-source electrodes (right). For this experiment, I 100 nA, which theoretically would give rise to a current density on the order of 80 A/cm2 at a distance of 100 m. As expected, significant pH changes are detected around the metal point electrode, varying from pH 6.8 to 9.8. However, no significant potential changes were detected with the pH electrode in the vicinity of the glass capillary electrode. This confirms that the contribution of the electric fields to the pH measurement is insignificant for the orders of magnitude of current used in this work (between 10 and 1000 A/cm2). Unfortunately a reproducible point source of hydroxide ions in the absence of electrical current was not available for this work. Therefore we had to rely on a more indirect approach to validate the SVET measurements in the presence of pH gradients. This involved simply verifying that the current distribution closely followed Eq. 1 and 2 for both the metal point and the glass capillary electrodes. If pH gradients did affect the measured SVET signal, one would expect that the current distribution would show systematic deviations from ideality. Figure 3 shows typical current distributions measured over a point source (in this case the glass capillary electrode). The points represent the experimental data, and the solid curve is a curve fit for Eq. 1 and 2. To verify the validity of the measurements, similar profiles were measured at different heights and treated using the calibration data obtained from the 70 m scan. In all cases, the forms of the curves were in good agreement with Eq. 1 and 2. Figure 4 shows Jz(max)/I as a function of 1/4 z2 where Jz(max) is the current at x and y 0. Again, linear behavior is observed with a slope of 1 as predicted by Eq. 1. In addition to the two point sources described above, a flat metal point source electrode (FMPS) was used in this curve, which consisted of a 10 m stainless steel wire mounted in an epoxy resin and polished. This is similar to the metal point-

Figure 3. The x and z components of the current vector measured by the SVET along the x axis above a metallic point source electrode in 30 mM NaCl. The points are experimental data, and the solid curves are curve fits to the theoretical point source response given by Eq. 1 and 2.

source electrode except that it is positioned on a flat surface and thus simulates the conditions of practical measurement. The absence of any systematic deviation of the SVET response from that predicted by Eq. 1 strongly suggests that the vibrating probe is not sensitive to pH changes, at least when detecting currents at this order of magnitude. Localized Observations at the Open-Circuit Potential Current distributions on technical samples.--Figure 5 compares the current distribution measured by the SVET experiment with an optical micrograph of the cut edge surface at the open-circuit potential, obtained after 20 min of exposure (the current cartography required approximately 10 min to obtain). In the optical micrograph to the left, the zinc surface is seen as two white lines running vertically down the image with the steel surface sandwiched between the zinc layers. Precipitated corrosion products form semicircular patterns on the edge of the steel surface. The white rectangle outlines the area over which the current cartography to the right was measured.

Figure 2. pH distribution along x axis over a constant current point source consisting of a stainless steel wire coated with polymer (10 m tip exposed) (left); Ag/AgCl wire immersed in a glass capillary (right). The electrolyte was 0.5 M NaCl for both experiments.

Figure 4. Variation of the Jz/I as a function of 1/z2 for different types of point source electrodes. The solid line is the theoretical current variation (slope 1). The point source electrodes are described in the text.

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

3657

Figure 5. Current distribution over the cut edge of double-sided galvanized steel mounted in an epoxy resin and exposed to 0.03 M NaCl. The rectangle on the left shows the area scanned.

The current density correlates well with the visual aspect of the surface. Anodic currents, indicated by positive values of the current density, were localized over the zinc surface where the anodic dissolution of Zn metal to form Zn 2 ions occurs. Similar localization of the anodic sites in NaCl solution has been reported for the cut edge of galvanized steel19 as well as Zn-coated steel with a scratch defect to the steel.20 It is noted that the white precipitate forms an outline around the anodic zones. The cathodic current is more homogeneously distributed in the center of the steel surface where a band of nearly constant electrochemical activity is detected. The zinc surface is 5-10 m thick, below the spatial resolution of the system, and appears as a point or line source of anodic current. pH distribution on technical samples.--Figure 6 shows the variation of pH during a similar cut edge experiment as in Fig. 5, also after waiting for about 20 min for the beginning of the scan. The total measurement time was approximately 20 min. In this case however, the precipitated corrosion products form a Y-shaped band in the center of the sample. The resulting pH cartography is shown superimposed as a contour plot of iso-pH lines. Very high pH values are observed in the upper center of the Y. The pH varied from about 7.6 to 10 under the conditions of this experiment. The presence of corrosion products correlates neatly with a zone of nearly constant pH between 8.0 and 8.2. Elevated pH values are observed in the upper part of the sample, and although they are shifted slightly to the right side, the elevated pH zone is clearly centered over the steel surface rather than the zinc. An area of slightly lower pH is identified at the lower right side of the surface, which is clearly centered over the zinc surface. pH distribution on model cut edge.--Because of the limited spatial resolution of the scanning pH technique, these results were compared with model samples in order to isolate the phenomena associated with the different surfaces. A model cut edge sample was prepared with the zinc and steel surfaces separated by several millimeters, such that corrosion-product precipitation would not occur on the steel surface. The geometry of the electrode and the resulting pH profiles obtained 20 min after a short circuit, are shown in Fig. 7. Over the steel surface, significant pH increases were detected, obtaining a maximum value of 11.2 shifted toward the right of the pH profile (the side closest to the zinc electrode). These results confirm the association of the alkaline pH with the steel surface and, therefore, the cathodic reaction. Over the zinc sur-

face, the pH remained nearly constant, with only a very slight increase toward the left position. The pH variation from left to right was less than 0.04 pH units. This result demonstrates that acidification due to the hydrolysis of Zn 2 ions is negligible in this medium, even when the anode and cathode are well separated. In principle, a 0.1 M solution of Zn 2 should give rise to a solution of pH of 5.3 according to the equilibrium Zn

2

H2O [ ZnOH

H

pK

9.67

Acidification might be expected for the situation where the anodic reaction is highly localized, as it is for the technical cut edge samples, or for the situation where the anodic reaction occurs under the polymer layer.

Figure 6. pH microscopy over the cut edge of double-sided galvanized steel. Each iso pH line represents 0.2 pH units. The electrolyte was 0.5 M NaCl.

3658

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

nificant effect on the nature of the reactions since only the cathodic reaction occurs on the steel surface. It is reasonable to expect that the rate of the cathodic reaction would be reduced however. In situ Raman spectroscopy was used to identify the major species present as ZnO and 3Zn(OH)2 2ZnCO3. Ex situ analysis indicated that the ZnO was present as an inner layer, and 3Zn(OH)2 2ZnCO3 as an outer layer in agreement with the results of atmospheric corrosion.4,22 Localized Current-Potential Relations Rate limiting processes near the open-circuit potential.--The spatial separation of the anodic and cathodic reaction on the cut edge electrode makes possible the determination of the current-potential characteristics of each half-reaction simultaneously but independently. Figure 9 gives a typical example of how the current distribution on the cut edge varies with applied potential for a galvanized steel sample with a zinc coating on only one side. Curve a in bold, gives the current distribution at the open-circuit potential ( 965 mV). The subsequent curves show the distribution obtained at a progressive cathodic polarization down to 1035 mV. The active surface of this sample can be divided into three areas of differing electrochemical behavior as follows: the anodic area (A) localized over the zinc surface; the cathodic area (C) localized over the steel farthest from the zinc surface; and an area of low electrochemical activity (B) between the two metal surfaces. Precipitated corrosion products were present in area (B). Single-sided galvanized steel was chosen for this work because the zone affected by the corrosion products is better defined than when a zinc coating is present on both sides. The dotted lines show the current obtained after polarization in successive 10 mV intervals. The anodic reaction decreases rapidly with potential, indicating that the rate-limiting step is the transfer of charge. By contrast, the cathodic reaction is independent of potential, indicative of a limitation by diffusion. Even after the potential has been decreased to 1035 mV, no cathodic current is observed in zone B, which strongly suggests that the cathodic reaction is inhibited in this region. Anodic dissolution kinetics near the open-circuit potential.--Figure 10 gives the variation of the anodic current with potential below the open-circuit potential for a double side galvanized steel sample. As described in the experimental section, these experiments were conducted by positioning the probe at the anodic maximum (with Jx 0) and integrating for several minutes, as the line scans in Fig. 5 are too noisy for this type of analysis. Particular care was taken to

Figure 7. pH distributions obtained over a model cut edge electrode with the zinc and steel separated by approximately 3 mm; 20 min after short circuit of the two electrodes in 0.5 M NaCl. Each distribution was performed in separate experiments. The steel sheet and zinc foil were 800 and 250 m thick, respectively. Note the different pH scales used for the two surfaces. The scan direction was from right to left for both experiments.

Large-surface area model electrode studies.--In the same manner as above, it was of interest to isolate the steel surface covered with zinc corrosion products. The method employed was similar, but on a smaller scale, to a method proposed by Baldwin et al.21 for the measurement of the sacrificial protection as a function of distance. A model electrode was constructed by vapor depositing a thin film of zinc metal over one-third of a 5 10 mm rectangular disk of low alloy steel. The experimental geometry and pH profiles obtained at variable times of exposure in 0.5 M NaCl are shown in Fig. 8. During the exposure of this electrode to the electrolyte, the steel surface was progressively covered by a white precipitate which moved as a front from the steel/zinc interface. As shown in the figure, a significant pH gradient is observed at the interface between the corrosion products and the intact steel. High pH values were observed on the steel side of the steel/corrosion product interface, and a near-neutral pH over the corrosion products, consistent with the results obtained for technical samples. The pH drops from 11.5 over the steel surface to near 8.0 over a distance of less than 1 mm. The pH neutralization can be explained by the buffering effect of the corrosion products as well as their role as cathodic inhibitors when adsorbed on the steel surface. It is possible that residual adhesive material may have been present on the steel surface; however, this is unlikely to have a sig-

Figure 8. pH distribution as a function of distance from Zn/steel interface for PVD model sample. The position of the pH gradient was observed to correlate closely with the growth of the corrosion product/steel interface as depicted in the figure. The relative distance between the two profiles is approximate.

Figure 9. Current distributions on the cut edge of single-sided galvanized steel embedded in an epoxy resin, zinc surface to the right. The horizontal line indicates zero current. The distance calibration is shown with the sample geometry below. The bold face letters indicate different spatial regions: A, anodic zone; B, inhibited area (correlates with the presence of zinc corrosion products); C, cathodic zone free from zinc corrosion products. The applied potentials (/mV vs. SCE) are a 965 (open-circuit potential); b 975; c 985; d 1005; e 1015; f 1035.

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

3659

choose a point where the anodic currents were well isolated from the cathodic currents. The curve on the left of Fig. 10, gives the results obtained in the 30 mM NaCl solution. The open-circuit potential was 990 mV as indicated by the dotted line. A linear relation is observed between log I and potential for more than an order of magnitude variation in log I, with a Tafel slope of 44 mV/decade, which is within the range of values given in the literature for zinc in alkaline solution.23,24 A second series of experimental points was obtained in 30 mM NaCl with the addition of SrCrO4, a well-known zinc corrosion inhibitor often used as a pigment in organic coatings for galvanized steel. The effect of the inhibitor is to shift the open-circuit potential in the positive direction by about 100 mV, and to increase the Tafel slope to 63 mV/decade. This is consistent with the formation of a passivating film on the zinc surface. Note that two data sets are shown for the inhibited solution, the electrode being polished between experiments. The open-circuit potential of the cut edge electrode differed by 25 mV, although no significant difference in the Tafel slopes was detected. The inhibition of hydrogen formation by corrosion products.--The upper series of curves in Fig. 11 shows the progressive polarization from 1065 to 1265 mV. The cathodic current distribution changes very little between 965 and 1200 mV, which is consistent with a cathodic reaction limited by oxygen diffusion. At polarizations below 1200 mV, the magnitude of the cathodic current increases monotonically with potential, even though the cathodic maximum remains localized on the edge of the steel electrode. This increase is attributed to a change from diffusion-limited oxygen reduction to electron-transfer-limited formation of hydrogen. The lower series of curves in Fig. 11 shows the progressive polarization from 1325 to 1465 mV. The second cathodic peak continues to intensify with decreasing potential, and, at the most negative potentials, cathodic activity is detected in zones B and A with an onset between 1325 and 1365 mV. The reactivity in this region is clearly illustrated by curve l, at 1465 mV, shown in boldface. The onset of a potential sensitive cathodic current was shifted by 300 mV in zone A/B with respect to zone C, indicative of cathodic inhibition in this region. Polarization of steel in the absence of zinc.--It is of interest to compare the current distribution obtained for single-sided electrogalvanized steel with the identical experiment using a steel sample without the zinc coating. The results are shown in Fig. 12. In this case, the anodic reaction was localized at the extremities of the cut edge and may be due to crevice corrosion between the epoxy resin and the metal.11 The cathodic reaction is independent of potential up to 1035 mV, and is homogeneously distributed across the surface, consistent with the diffusion-limited reduction, as in Fig. 10. Below

1035 mV, the cathodic current increases with potential again consistent with the water reduction reaction, but this time the reaction is homogeneously distributed across the surface. Discussion These results suggest that zinc corrosion products precipitate at an intermediate zone between the increased pH produced by the cathodic reaction and the increased Zn 2 ion concentration produced by the anodic reaction. The pH in the area affected by the corrosion products assumed a nearly constant value. This may be due to the buffering effect of the products or to the fact that they inhibit the cathodic reaction. The cathodic formation of ZnO thin films by reduction of oxygen in Zn 2-containing solutions has also been shown to be driven by interfacial augmentation of the pH.25 These results are consistent with literature values for the solubility of various zinc based corrosion products. ZnO, Zn(OH)2, and various hydroxychlorides all show solubility minima between pH 9 and 11.25 Likewise, the spontaneous pH of water saturated with Zn(OH)2 is 8.40 and 8.85 for amorphous and -Zn(OH)2, respectively,26 although at higher pH the solubility increases markedly due to the formation of zincate ions. Although the inhibiting effect of the zinc corrosion products has been demonstrated, no attempt has been made to trace its origin. While the inhibited zone correlated with the presence of visible corrosion products, it has not been ascertained that the inhibition was actually due to these products and it is likely that much thinner, more compact films are formed underneath the precipitated corrosion products which are not visible in the microscope. Maeda et al.27 have proposed Zn(OH)2 as a more effective inhibitor than ZnO, and according to Graedel,4 Zn(OH)2 is the first species formed during the atmospheric corrosion of zinc. Because of the very slight difference in equilibrium potential (3 mV) between Zn(OH)2 and ZnO,28 the transformation will generally occur, although it is thought that certain zinc alloys favor Zn(OH)2 27 formation and are thus more resistant to atmospheric corrosion. ZnO is an n-type semiconductor, and its presence has been invoked to explain the light-enhanced corrosion of zinc.4,29 The bandgap can vary between 2.2 and 3.6 eV depending on the presence of impurities and amorphous phases.30

Figure 10. Localized polarization behavior over the zinc surface on the cut edge of galvanized steel in 0.03 M NaCl in the presence and absence of chromate ions. The dashed lines indicate the spontaneous corrosion potential.

Figure 11. Continuation of Fig. 9 at more negative potentials. The applied potentials (/mV vs. SCE) are (upper series) g 1065; h 1205; i 1265; (lower series) j 1325; k 1425; l 1465 (bold). Horizontal lines mark the zero for each series of curve.

3660

Journal of The Electrochemical Society, 147 (10) 3654-3660 (2000)

S0013-4651(00)02-083-8 CCC: $7.00 © The Electrochemical Society, Inc.

Localized polarization curves were obtained on technical cut edge samples by measuring the current distribution at various applied potentials. The anodic reaction was activation controlled, with a Tafel slope of 44 mV/decade and linear log I-E behavior for more than an order of magnitude below the open-circuit potential. By contrast, the cathodic current was independent of applied potential, consistent with the diffusion-limited reduction of oxygen. At more negative potentials, an increase in cathodic current with potential was indicative of the reduction of water. The onset of this reaction was shifted in the cathodic direction by nearly 300 mV in the zone affected by the corrosion products, demonstrating their efficiency as cathodic inhibitors. The effect of chromate addition to the electrolyte on the anodic reaction was also investigated. The corrosion potential increased by 100 mV, and the Tafel slope increased to 63 mV/decade. These results are consistent with the formation of a passivation layer as expected for an anodic inhibitor such as chromate. Acknowledgments The authors would like to thank Eric Karplus (Sciencewares) and Alan Shipley (Applicable Electronics) for their counsel and aid in fabricating the pH microelectrodes, Alan Lamande (Irsid) for the preparation of the model Zn/steel galvanic couple electrode by vapor deposition, and Dr. Norma Philips (Irsid) for the Raman spectroscopic analysis.

Irsid assisted in meeting the publication costs of this article.

Figure 12. Current distributions during the cathodic polarization of an low alloy steel cut edge without zinc coatings, immersed in 0.03 M NaCl. The applied potentials (/mV vs. SCE) are: a 557 (open circuit); b 577; c 597; d 617; e 637; f (dashed), 877; g 1077; h 1117; i 1137; j 1157; k 1177; l 1197 mV.

Consideration of these phenomena could shed light on the variations of corrosion rate with the nature of the zinc alloy used for galvanization, as the presence of trace levels of impurities can have considerable effect on the conductivity of the oxide, and therefore its efficiency as a cathodic inhibitor. Furthermore, the presence of water-soluble inhibitors in the paint would have a marked effect, entirely changing the nature the corrosion products and their electronic properties. In fact, the most commonly used inhibitors such as chromate and phosphate form compounds with Zn 2 which are even less soluble than ZnO at high pH. The incorporation of redox couples (such as Cr(VI) or Fe(III) from surface treatments or from alloying elements with the zinc) could lead to a potential buffering effect, stabilizing the surface as is observed for iron scales.31 Conclusions This work has demonstrated the usefulness of scanning electrochemical techniques in characterizing the chemical environment and inhibition phenomena on the cut edge of galvanized steel in immersion. The spatial resolution of these techniques is well adapted for cut edge studies. The steel surface on technical cut edge samples is on the order of 800 m, which is within the spatial resolution of the SVET and scanning pH measurements. The zinc surface is 5 to 10 m thick, below the spatial resolution of the system. By using a series of model point-source electrodes, the validity of the pH measurements in the presence of electrical fields generated by the corrosion reactions was directly confirmed. The validity of the current-density measurement with a vibrating probe was indirectly confirmed in the presence of pH gradients. Under the conditions of this study, the zinc was anodic, and the steel cathodic, and in the electrolyte used here (0.03 M NaCl) no anodic reaction was ever observed on the steel over the course of 8 h. The pH was found to be highly variable, ranging from near neutral to above 10. Experiments with model electrodes have demonstrated that the high pH values are due to the formation of hydroxide by the cathodic reaction. No evidence of acidification by the hydrolysis of Zn 2 released from the zinc was observed, even when the zinc and steel surfaces were separated. Corrosion-product precipitation was visually observed in intermediate zones between the anodic and cathodic activity. Raman measurements on large-surface-area model samples with vapordeposited zinc under similar conditions indicated that the steel surface was covered by a mixture of ZnO and 3Zn(OH)2 2ZnCO3.

References

1. I. Dehri, R. L. Howard, and S. B. Lyon, Corros. Sci., 41, 141 (1999). 2. R. L. Howard, I. Zin, S. B. Lyon, and J. D. Scantlebury Abstract 105, The Electrochemical Society Meeting Abstracts, Vol. 98-1, San Diego, CA, May 3-8, 1998. 3. F. Zou, C. Barreau, R. Hellouin, D. Quantin, and D. Thierry, in Proceedings of the 3rd International Converence on Zinc and Zinc Alloy Coated Steel Sheet (Galvatech '95), p. 837, Japan (1995). 4. T. E. Graedel, J. Electrochem. Soc., 136, 193C (1989). 5. I. Odnevall and C. Leygraf, in Atmospheric Corrosion, ASTM STP 1239, W. W. Kirk and H. H. Lawson, Editors, American Society for Testing and Materials, Philadelphia, PA (1994). 6. M. C. Bernard, A. Hugot-Le Goff, and N. Phillips, J. Electrochem. Soc., 142, 2162 (1995). 7. V. R. Grauer, Werkstoffe Korros., 31, 837 (1980). 8. V. Baudu, L. Garrigues, and K. Ogle, in SF2M Journées d'Automne '96, Proceedings in Rev. Métall., Paris, JA96, p. 66 (1996). 9. K. Ogle, R. Nicolle, L. Garrigures, H. Lavelaine, and V. Baudu, in Proceedings of the 4th International Converence on Zinc and Zinc Alloy Coated Steel Sheet (Galvatech '98) p. 388, Chiba, Japan (1998). 10. L. Garrigues, V. Baudu, K. Ogle, and X. Philippe, in Proceedings of Eurocorr '98, Utrecht (1998). 11. H. S. Issacs, A. J. Davenport, and A. Shipley, J. Electrochem. Soc., 138, 390 (1991). 12. H. Isaacs, J. Electrochem. Soc., 138, 722 (1991). 13. K. T. Brown and D. G. Flaming, Neuroscience, 2, 813 (1977). 14. J. O. Park, C. Paik, and R. C. Alkire, J. Electrochem. Soc., 143, L174 (1996). 15. C. Wei, A. J. Bard, G. Nagy, and K. Toth, Anal. Chem., 67 (1995). 16. E. Klusmann and J. W. Schultze, Electrochim. Acta, 42, 3123 (1997). 17. D. Ammann, Ion-Selective Microelectrodes, Springer-Verlag, Berlin (1986). 18. E. Bayet, F. Huet, M. Keddam, K. Ogle, and M. Takenouti, Electrochim. Acta, 44, 4117 (1999). 19. S. Baum, H. N. McMurray, S. M. Powell, and D. A. Worsley, Electrochim. Acta, 45, 2165 (2000). 20. H. Issacs, A. J. Aldykiewicz, Jr., D. Thierry, and T. C. Simpson, Corrosion, 52, 163 (1996). 21. K. R. Baldwin, M. J. Robinson, and C. J. E. Smith, Corros. Sci., 36, 1115 (1994). 22. J. R. Vilche, F. E. Varela, E. N. Codaro, and C. A. Gervasi Abstract 155, p. 212, The Electrochemical Society Meeting Abstracts, Vol. 96-2, San Antonio, TX, Oct 6-11, 1996. 23. J. O'M. Bockris, Z. Nagy, and A. Damjanovic, J. Electrochem. Soc., 119, 285 (1972). 24. Yu-Chi Chang, Electrochim. Acta, 41, 2425 (1996). 25. S. Peulon and D. Lincot, J. Electrochem. Soc., 145, 864 (1998). 26. M. Pourbaix, Atlas des èquilibres èlectrochimiques, p. 411, Gauthier-Villars, Paris (1963). 27. Y. Miyoshi, J. Oka, and S. Maeda, Trans. ISIJ, 23, 974 (1983). 28. K. Wippermann, J. W. Schultze, R. Kessel, and J. Penninger, Corros. Sci., 32, 205 (1991). 29. E. Juzeliunas, P. Kalinauskas, and P. Miecinskas, J. Electrochem. Soc., 143, 1525 (1996). 30. P. Scholl, X. Shan, D. Bonham, and G. A. Prentice J. Electrochem. Soc., 138, 191 (1991). 31. M. Stratmann and J. Müller, Corros. Sci., 36, 327 (1994).

Information

Localized Electrochemical Methods Applied to Cut Edge Corrosion

7 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

74298

You might also be interested in

BETA
Microsoft Word - NJNJDOT Proj_Bridge Appurtenance1-24-05.doc
Microsoft Word - PEA_27_5_2009_555_564.doc
untitled
14FZ.fm.save