Home

People

Purpose

Research

Employment

SEAS Home

UVa Home

Associate Professor and Center Co-Director Phone: 434.982.5783 Fax: 434.982.5799 Email: rgkelly@virginia.edu

Curriculum Vitae Kelly AIChE JCSE Paper DC Short Course Rodman Scholars Program Graduate Student Handbook The Science of Localized Corrosion (a collaborative research project) Gordon Conference Talk (July 7-12, 2002) Local Probes of Corrosion and Organic Coating Failure on Aluminum Alloys Kevin Cooper's PhD papers and presentations Materials directorate, Air Force Research Lab of Wright-Patterson Air Force Base COURSE DOCUMENTS ENGR 141R-1 ENGR 141R-2 ME 339 MSE 606

JCSE
This paper is a pre-print of paper number 4 in the 1995 edition of JCSE.
Comments may be sent to: R.A.Cottis@umist.ac.uk

An Electrochemical Investigation of the Corrosion Behavior of Aluminum Alloy AA5052 in Methanolic Solutions

L. A. Pawlick, R. G. Kelly
Center for Electrochemical Science and Engineering
Department of Materials Science and Engineering
University of Virginia Charlottesville, VA 22903

Abstract
The electrochemical behavior of aluminum alloy AA5052 in methanolic solutions containing low concentrations of acid, chloride, sulfate and water has been studied. In all solutions investigated, the alloy exhibited spontaneous passivity. The addition of 1 mM acid increased the corrosion potential dramatically, whereas 1 mM chloride decreased the pitting potential and 1mM sulfate had no measurable effects. Water appears to decrease the pitting potential due to the enhancement of aluminum ion hydrolysis at incipient pits. No important second order interactions were observed. The effects of acid on the corrosion potential and rate are explained in terms of mixed potential theory.

INTRODUCTION
As the use of alcohol in automotive fuels increases due to environmental concerns, so does the need to characterize and understand the corrosion of metals and alloys in alcoholic solutions. Of particular importance are the effects of low levels of the impurities commonly expected in these fuels such as acid, chloride, sulfate, and water. In recent work on iron in methanolic solutions, Brossia and Kelly first developed a quantitative characterization of the electrochemical phenomenology of iron in methanolic solutions including the effects of common impurities [1] which led to a description of the mechanisms underlying the accelerating effects of acid and the inhibitory effects of water on the open circuit corrosion rate [2]. One of the key elements of that work involved the use of statistical experimental design [3] to allow a quantitative analysis of the main (first-order) effects of the various solution additions, second-order interactions and an estimate of the experimental error. The error estimation permitted the statistical significance of each effect to be quantitatively assessed.

Brossia et al. [2] were able to explain the two primary empirical observations concerning the effect of acid on the corrosion of iron in methanol [4-6]: large increases in corrosion rate (by a factor of 15 due to a 1 mM acid addition [4]) and the corrosion potential (by 200 mV due to a 1 mM acid addition [5]). Brossia et al. [2] showed that the effects of acid were due to two separable phenomena; acid activates the originally passive iron surface, and proton reduction has substantially faster kinetics than oxygen reduction in acidified methanol. This latter effect dominates, leading to an increase in the corrosion potential with a subsequent increase in corrosion rate.

The addition of water inhibits the corrosion of iron in acidic methanol dramatically [1-6]. Brossia et al. [2] determined that proton reduction was under substantial mass transport control at the corrosion potential of iron. They then showed that the addition of water to acidified solutions reduced the corrosion rate predominantly by inhibiting the mobility of the proton, thus reducing the proton diffusivity and hence the diffusion limited current density of the dominant cathodic reaction in acidified methanol. This reduction in diffusion limited current density led directly to the decrease in the corrosion rate. The reduction in proton mobility is due to the preferential solvation of protons by water relative to methanol [7]. At low concentrations of water, this limits the proton hopping that is important for conductivity.

In an extension of that work, the electrochemical behavior of aluminum alloy AA5052 has been studied and analyzed. By studying a valve metal such as aluminum which passivates with a thick film, information on the effect of the nature of the passive film on the corrosion of materials in methanol can be gained. In addition, AA5052 is being considered as a construction material for future automobile fuel tanks. Thus, an assessment of the effects of impurities on its corrosion behavior would have practical importance as well.

Limited previous work on the corrosion of aluminum alloys in methanolic solutions has been performed. Mansfeld [8] and Palit et al. [9] focussed on the corrosion of aluminum alloys in strongly acidified methanolic solutions. Upon anodic polarization of AA6061 above 0 V(SCE) in 0.1 N sulfuric acid, Mansfeld observed pitting with gas evolution from the pits. He proposed that the pitting was due to the presence of the sulfate ion. Palit et al. [9] also observed pitting of pure Al in acidic methanol in solutions open to air. Hronsky [10] measured weight loss for pure Al in 0.3 M HCl. Severe pitting was observed under open circuit conditions. Wing et al. [11] and Lash [12] also used coupon testing to investigate the behavior of various cast aluminum alloys in M85 (15% fuel, 85% methanol with 1.1mM acid, 0.1% water, and 0.07mM chloride). They found substantial weight losses and pitting of the cast Al alloys.

In order to better understand the corrosion of aluminum alloys in methanol, the present study was conducted. The corrosion and electrochemical behavior of AA5052 in a variety of methanolic solutions was studied and the effects of the addition of several impurities were quantified and explained in terms of the previous work on iron [2]. Previous observations in this area are also rationalized within the framework of mixed potential theory.

EXPERIMENTAL
Materials - Aluminum alloy AA5052 was supplied by Kobe Steel in the form of 0.50 mm thick sheet. Its composition was (in wt.%) 2.46 Mn, 0.22 Fe, 0.19 Cr, 0.1 Si, 0.02 Ti, remainder Al. Disk electrodes were cut from the sheet and wet polished with successively finer silicon carbide paper to a 600 grit finish.

Solutions - All test solutions were based on spectrophotometry grade, "Photrex" reagent methanol (J.T. Baker). All solutions contained 0.1 M anhydrous sodium perchlorate (Aesar) as a supporting electrolyte. Other solution additions included water, chloride, sulfate, and/or acid. The chloride was added as anhydrous lithium chloride (Fisher), the sulfate as anhydrous sodium sulfate (Aldrich), and acid as sulfuric acid. The water content of the solutions was measured via Karl Fischer titration (Mettler DL-18). A full factorial design was used to investigate the effects of the various impurities. Two levels of concentration were used, zero and 1 mM, except for water. The inherent water level of the methanol and that added from perchlorate led to a minimum water content for solutions of < 0.06 wt.%. The effects of water were studied by the addition of water to a concentration of 0.5 wt.%. No efforts were made to remove dissolved molecular oxygen as this would tend to evaporate large quantities of methanol, changing the concentrations of the dissolved substances. Previous work [2] has identified the relevant cathodic reactions at occurring in the different solutions. Testing Procedures - All experiments were conducted at room temperature. Electrochemical measurements were conducted with an E.G.&G. Princeton Applied Research (PAR) Versastat(TM) controlled by the PAR Model 352 software. A silver/silver chloride (SSC) electrode immersed in a compartment containing methanol with 0.1 M LiCl and 1.5 mass% water was used as the reference electrode. The reference electrode was separated from the working electrode solution by a Vycor(TM) frit. The SSC electrode has a potential of -29 mV vs. an aqueous saturated calomel electrode. The SSC was found to be very stable with respect to time and minimized the liquid junction potential [13]. Cyclic polarization scans were conducted at a scan rate of 0.5 mV/s starting at an initial potential of -0.8 V(SSC). A vertex current density of 5 mA/cm2 was used. Automatic current interruption was used for on-line correction for ohmic drop.

Analysis Procedures - Each curve was analyzed for corrosion potential, corrosion current density and pitting potential. The corrosion current density was determined by use of PARcalc(TM) and checked manually by Tafel extrapolation. To quantitatively determine the significance of the effects of added impurities and the experimental error, the results of the full factorial design were analyzed with Number Cruncher Statistical Software (NCSS) ver.5.03 (licensed by J. L. Hintze). Performing duplicate experiments for each condition and comparing the results of all experiments containing the solution species of interest to all experiments that did not contain that species allowed for the separation of effects due to single impurities as well as any synergistic effects between species. Upon completion of the tests, each effect was statistically analyzed to determine its significance as well as to determine the experimental error.

RESULTS
In the base solution, AA5052 did not exhibit an active-passive transition as shown in Figure 1. This lack of an active-passive transition near the corrosion potential can be termed spontaneous passivity, and this term is used throughout this manuscript to refer to such behavior. This passive region was limited at +0.35 V(SSC) by pitting as shown in the photomicrograph of Figure 2. The effects on the polarization behavior of the addition of individual impurities (1 mM acid or 1 mM sulfate or 1 mM chloride or 0.5 wt.% water) are shown in Figure 3. Acid increased the open circuit potential and decreased the pitting potential of the AA5052, and increased the open circuit corrosion rate slightly. The addition of water led to a decrease in the pitting potential, while sulfate had little effect. Chloride decreased the pitting potential.

A summary of the first order effects of the impurities on the electrochemical parameters that characterize the corrosion process in both low and high water contents is shown in Figures 4 and 5. None of the solution species had a dramatic effect on the corrosion rate, either alone or in combination. No significant synergistic effects were observed. In all solutions studied in which acid was present, the corrosion potential rose dramatically. Acid was the only solution addition studied which affected the corrosion potential to a statistically significant degree. A solution containing acid and water led to an increased OCP and a decreased pitting potential as shown in Figure 6. None of the solution additions had a significant effect on the repassivation potential, with all values close to -0.44 V(SCE).

In order to investigate the importance of hydrolysis on the stabilization of pits, a study was performed in which AlCl3 was added and its effects on the acid concentration and the corrosion of the AA5052 were studied. The initial water concentration was 500 ppm, and the open circuit potential of the AA5052 was -0.5 V(SSC). Upon addition of the 0.05 M AlCl3, a white precipitate formed and the solution fizzed. The open circuit potential of the AA5052 fell to -1.1 V(SSC) and the polarization resistance decreased from 897 to 4.3 ohms-cm2. Post-test inspection of the surface showed that the attack was pitting, not uniform corrosion. The diffusion limited current density on Pt increased after the addition of the AlCl3 to approximately 1.3 mA/cm2, indicative of a acid concentration of approximately 60 mM. Water was then added to the solution to increase its concentration to 0.5 wt.%. This addition resulted in the reduction in the diffusion limited current density for proton reduction on Pt to 0.929 mA/cm2. The corrosion potential of the AA5052 increased to -0.722 V(SSC) and the polarization resistance also increased to 491ohms-cm2.

DISCUSSION
In all solutions studied, AA5052 exhibits spontaneous passivity, indicating that the air-formed film is stable at and near the open circuit potential (Figure 1). The correspondence of the measured pitting potential to the stabilization of localized attack was established via post-test examination (Figure 2). While the effects of sulfate, chloride, acid and water on the open circuit corrosion parameters (i.e., corrosion rate and corrosion potential) were qualitatively similar to those observed for iron previously [1,2], the effects are much smaller quantitatively. Of particular interest was the small effect of low levels of acidity. At these same levels of acid, iron spontaneously activated under open circuit conditions [2]. These levels might reasonably be expected in a methanol fuel [11]. Thus, it might appear that AA5052 tanks would perform better from a corrosion standpoint than steel in such an application. However, it should be noted that the presence of 1 mM acid did increase the corrosion potential of the AA5052 substantially (ca. 200 mV) while lowering its pitting potential by 100 mV (Figure 3). This decrease in the range of passivity increases the probability of stable pitting occuring under open circuit conditions.

As was also observed for iron [2], sulfate had no effect on the corrosion behavior of AA5052 (Figures 4 and 5). This is contrary to the conclusions of Mansfeld [8] who speculated that sulfate was the cause of the pitting of a variety of alloys (including AA6061) in acidified methanol. The experiments in that work were not designed to separate the effects of acid from those of the sulfate. In the present work, the use of sodium sulfate and a statistical experimental design allowed the individual effects of each species to be determined, indicating that AA5052 will pit in neutral methanolic solutions in the absence of sulfate upon sufficient anodic polarization. The low pitting potential observed by Mansfeld of 0 V(SCE) was most likely due to the high acidity (0.2 M) present in his solutions.

As shown in Figure 3, the addition of 1 mM acid increased the corrosion potential of AA5052 by 100 mV. This increase results from the introduction of an additional cathodic reaction (proton reduction) with faster kinetics than oxygen reduction (which is the predominant cathodic reaction in neutral methanol [2]). Unlike iron, the small acid addition does not activate the surface of the AA5052, so the increase in corrosion rate is quite modest and due solely to the slight increase in passive current density with potential (Figure 1). The addition of 1 mM acid also decreases the pitting potential of the AA5052 by 200 mV, which combined with the 100 mV increase in the corrosion potential, significantly decreases the passive range of the alloy (Figure 6). It would be expected that sufficient increase in the acid concentration would lead to the equivalence of the corrosion and pitting potentials, leading to pitting under open circuit conditions as observed by Hronsky [10] for pure Al in strongly acidified methanol. The high chloride content of Hronsky's solutions would also be expected to contribute to a decrease the pitting potential, as shown in this work. Wing and Evarts [11] also observed pitting of AA356 under open circuit conditions in a simulated 85% methanol fuel containing low levels of chloride, acid and water.

Water has a very small effect on the open circuit corrosion rate of AA5052, even in the presence of acid. This is not unexpected, due to the fact that the alloy surface remains passive. Under these conditions, the decrease in the limiting current for proton reduction due to the addition of water [2] will not have an effect on the corrosion rate, i.e., the corrosion potential of AA5052 is in the activation control region of the proton reduction reaction in methanol.

The deleterious effect of water on the pitting potential was unexpected. Generally, water is considered to improve the ability of the passive film to form in nonaqueous solutions, particularly under acidic conditions [1-6]. One explanation for the present results showing a clearly deleterious effect on water on the pitting potential is an increase in the stabilization of pitting that could result from hydrolysis of Al ions produced at metastable pits. It has become generally accepted that the pitting potential observed in potentiodynamic polarization scans is related to the conditions under which pits can propagate stably, rather than the potential at which they can initiate. Under such conditions, the high level of acidity required for localized corrosion site propagation must be achieved by hydrolysis of Al3+ at the incipient pit. As the water content of the solution increases, the hydrolysis becomes less reactant limited and stabilization of pitting occurs at lower potentials.

This proposal was tested by the study of the addition of AlCl3 on the acid content and the corrosion of the AA5052. The addition of the AlCl3 caused fizzing and the formation of a white precipitate. The hydrolysis of the aluminum salt would occur via: AlCl3 + 3H2O = Al(OH)3 + 3H+ + 3Cl-.

The reaction proceeded as written as evidenced by the large increase in acid content (from 0 to 60 mM) for the 50 mM AlCl3 added. The large decrease in the open circuit potential and spontaneous pitting of the AA5052 upon the addition of the AlCl3 supports the contention that the stabilization of the pitting occurs by hydrolysis. The 500 ppm of water initially present is more than sufficient to allow the hydrolysis to occur to completion. The 60 mM of acid (and 150 mM chloride) led to the formation of stable pits which propagated at a very high rate due to the reduction on protons on the areas which did not pit. The massive decrease in the polarization resistance could not be used to calculate a penetration rate, due to the localized nature of the attack, but in conjunction with the post-test inspection and drop in the open circuit potential, can be used to support the contention that the pitting intiated upon addition of the AlCl3. The addition of further water did decrease the corrosion rate, as signified by the increase in polarization resistance from 4.3 to 491 ohms-cm2.

Thus, water plays several roles in organic solutions. It can decrease the mobility of the proton and thereby decrease corrosion rates when active corrosion occurs, as has been shown for iron [2] and for the actively pitting AA5052 in this study. In addition, for metals which strongly hydrolyze, water additions can lower the pitting potential in neutral solutions by enhancing the stability of metastable pits. The extent of the effect of water on the pitting potential in neutral organic solutions would be expected to depend upon the hydrolysis rates of the different metal ions in methanolic solutions, as the diffusion limited current density for proton reduction will be independent of substrate. Based on previous work on iron [2], in which neither water nor acid had deleterious effects on the pitting potential, it appears that hydrolysis effects are more important for aluminum. It would be of interest to use methanol solutions as test vehicles for studies of metal ion hydroylsis and its relation to localized corrosion.

SUMMARY AND CONCLUSIONS
1. This work has provided additional support for the mechanisms proposed by Brossia et al. [2] concerning the deleterious effects of acid and the inhibitory effects of water on the corrosion rate of materials in methanol. The deleterious effects of acid were shown to be due to the addition of a more oxidizing cathodic reaction (proton reduction) which has faster kinetics than oxygen reduction, despite the much higher oxygen concentration (68 vs. 1 ppm). Water additions to acidified methanol inhibit this increase in corrosion rate for actively corroding materials by reducing the rate of proton hopping due to preferential solvation of the proton by water. For materials such as AA5052 which do not activate in 1 mM acid, the quantitative effects of water are much smaller than materials such as iron which actively corrode in 1 mM acid.

2. While the effects of the 1 mM chloride, sulfate and acid were qualitatively the same for AA5052 as previously found for iron, all of the effects were smaller in magnitude for the AA5052, indicating that the aluminum alloy would be more robust to minor impurities present in methanol-based fuels.

3. The proposed mechanisms for the accelerating effects of acid and the inhibitory effects of water on active corrosion have been used to rationalize the previously reported work on the corrosion of Al alloys in methanol. Much of the previous work did not interprete their results in terms of underlying mechanisms. In some cases, this work has been used to show that effects previously attributed to sulfate ions were in fact due to the presence of high concentrations of acid.

4. The deleterious effect of water on the pitting potential has been related to the competition between its stabilization effects on incipient pits by Al3+ hydrolysis and its inhibition of the diffusion limited current for proton reduction. Thus, while water is generally found to be a good inhibitor for general corrosion in acidified organic solvents, the addition of small amounts water to an acid-containing methanol solution would lead to an increased likelihood of pitting for AA5052. At higher water contents, the stabilization of the pitting is restricted by the inhibition of the predominant cathodic reaction occuring on the passive surface. This effect on the cathodic reaction overcomes the effects of water on hydrolysis and water acts as an inhibitor at higher concentrations.

ACKNOWLEDGEMENTS
The authors would like to acknowledge the technical discussions with C.S. Brossia (University of Virginia) as well as the financial support of Kobe Steel and equipment support from EG&G Princeton Applied Research. The assistance of Jean Reese in the preparation of the HTML version of this manuscript is also gratefully acknowledged.

REFERENCES
[1] C. S. Brossia, R. G. Kelly, Proc. of the 6th Automotive Corrosion and Prevention Conf., p. 173, Society of Automotive Engineers, Warrendale, PA (1993).

[2] C. S. Brossia, E. Gileadi, R. G. Kelly, Corros. Sci., in press.

[3] G. E. P. Box, W. G. Hunter, J. S. Hunter, Statistics for Experimenters, Wiley, NY (1978).

[4] F. Bellucci, G. Capobianco, G. Faita, C. A. Farina, G. Farina, F. Mazza, S. Torchio, Corros. Sci., 28, 371 (1988).

[5] P. L. de Anna, Corros. Sci., 25, 43 (1985).

[6] F. Mazza, S. Torchio, N. Ghislandi, Int. Cong. on Metallic Corrosion, Vol. 1, p. 465, National Research Council of Canada, Ottawa (1984).

[7] R. DeLisi, M. Goffredi, Electrochim. Acta, 16, 2181 (1971).

[8] F. Mansfeld, in Galvanic and Pitting Corrosion, ASTM STP 576, R. Baboian, W. D. France, L. C. Rowe, Jr., J. F. Rynewicz, eds., Amer. Soc. for Testing and Materials, p. 180 (1976).

[9] G. C. Palit, K. Elayaperumal, Corros. Sci., 18, 169 (1978).

[10] P. Hronsky, Corrosion, 37, p. 161 (1981).

[11] L. Wing, G. L. Evarts, D. M. Tramontana, J. Mat. Eng., p. 26 (April, 1992).

[12] R. J. Lash, Proc. of the 6th Automotive Corrosion and Prevention Conf., p. 153, Society of Automotive Engineers, Warrendale, PA (1993).

[13] C. S. Brossia, R. G. Kelly, submitted to Electrochim. Acta.

Full Factorial Experimental Design and Analysis A two-level, full factorial design was implemented in these experiments to compare the effects of four different species on the corrosion parameters. A two-level, full factorial design consists of 2k experiments where k is the number of factors each with a high and low value. In this context a factor is an experimental variable, and a result is the quantitative measure of the parameter of interest. For example, in the study of corrosion , a factor may be the acid concentration in the solution, and a result would be the corrosion potential. The relevant statistical effects of each factor are found by comparing the results from all of the experiments with the high value of a factor to the results of all of the experiments with the low values. For example, for a four-factor analysis, the eight experiments with a high value of one factor are compared to the eight experiments with the low value. In this manner, it is also possible to quantify higher order (or combined) effects. Table 1 shows a sample two-level three factor design. (Note: The design matrix and the given results are not all of the actual data collected in this experiment. The values have been simplified for the purpose of this example.) Table 1 Experiment Factors Results A B C Rep 1 Rep 2 (Water (Acid (Chloride (Corrosion (Corrosion Content) conc.) conc.) Potential) Potential) 1 0 0 0 (1) -426 -314 2 1 0 0 a -226 -236 3 0 1 0 b -349 -514 4 1 1 0 ab -235 -287 5 0 0 1 c -388 -372 6 1 0 1 ac -469 -503 7 0 1 1 bc -218 -249 8 1 1 1 abc -289 -290 The mean effect of factor A alone for this three factor analysis can be found by First order effect (A) = avg(a+)- avg(a-) This main effect is the mean of all of the results where A was a factor minus the average of all of the results where A was not a factor. The second and third order effects can be found by the equation Second order effect (AB)= [Mean effect of a(b+)- Mean effect of a(b-)]/2 = [Mean effect of b(a+)- Mean effect of b(a-)]/2 In the terms of our example , a second order effect would be the average water content effect with acid minus the average water content effect without acid divided by two. The other second and third order interactions are found in a similar manner. Figures 1a and 1b are the cube plots for this example. Figure 1a shows the factorial design, and Figure 1b shows the first, second and third order interactions. Table 2 gives the calculated effects. Table 2 Experiment Mean Effect Error (Grand Mean) (-335.3) 13.3 a 37 26.5 b 63 26.5 ab 20.5 26.5 c -23.8 26.5 ac -117.8 26.5 bc 108.8 26.5 abc 4.8 26.5 When duplicate experiments are run, the variance can be determined for the entire set of data and the standard error may be found. Figure 2a shows a sample printout of the sum of the squares, the mean squares, the F ratios and the probabilities for this factorial analysis. From these values, the true effects can be separated from the effects which are within the experimental error. For this set of data, the acid content is a true effect. The second order acid effects are also strong effects. Furthermore, the same method employed to find the relevant effects in this three-factor analysis can be expanded and applied to a four-factor analysis. References Box, Hunter, and Hunter, Statistics for Experimenters. John Wiley & Sons, New York: 1978. Hintze, Jerry L. Number Cruncher Statistical Software, Experimental Design, Version 5.4. Kaysville, Utah: 1989. Statmost Statistical Analysis and Graphics, User's Handbook. DataMost Corporation, Salt Lake City, Utah: 1994.

Figure1: Cyclic polarization curve for AA5052 in the base solution (anhydrous methanol containing 0.1 M sodium perchlorate).

Figure 2: Photomicrograph of AA5052 surface after polarization scan shown in Figure 1. Note the corrosion pits in an otherwise as-polished surface.

Figure 3: Polarization behavior for AA5052 in (a) the base solution and upon the addition of individual impurities: (b) 1 mM acid or (c) 1 mM sulfate or (d) 1 mM chloride or (e) 0.5wt.% water. Note the lack of an effect of sulfate, while acid, water and chloride all lead to a decrease in the pitting potential.

Figure 4: Summary of the effects of the impurities (a) alone and (b) in combination on the corrosion potential and open circuit corrosion rate for AA5052. All solutions contained 0.1 M sodium perchlorate and <0.06 mass% water.

Figure 5: Summary of the effects of the impurities (a) alone and (b) in combination on the corrosion potential and open circuit corrosion rate for AA5052. All solutions contained 0.1 M sodium perchlorate and 0.5 mass% water.

Figure 6: Polarization behavior for AA5052 in (a) the base solution and (b) upon the addition of 1 mM acid and 0.5 wt.% water. Note the decrease in the passive range upon the addition of the two impurities.