A Transmission Electron Microscopy Study of Constituent- Particle-Induced Corrosion in 7075-T6 and 2024-T3 Aluminum Alloys

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, APRIL 1998—1153

A Transmission Electron Microscopy Study of Constituent-Particle-Induced Corrosion in 7075-T6 and 2024-T3 AluminumAlloys

ROBERT P. WEI, CHI-MIN LIAO, and MING GAO

To better understand particle-induced pitting corrosion in aluminum alloys, thin foil specimens of7075-T6 and 2024-T3 aluminum alloys, with identified constituent particles, were immersed inaerated 0.5M NaCl solution and then examined by transmission electron microscopy (TEM). Theresults clearly showed matrix dissolution around the iron- and manganese-containing particles (suchas Al23CuFe4), as well as the Al2Cu particles. While Al2CuMg particles tended to dissolve relativeto the matrix, limited local dissolution of the matrix was also observed around these particles.These results are consistent with scanning electron microscopy (SEM) observations of pitting cor-rosion and are discussed in terms of the electrochemical characteristics of the particles and thematrix.

I. INTRODUCTION

LOCALIZED (pitting) corrosion has been observed inhigh-strength aluminum alloys in aqueous environments(electrolytes) and has been identified as a potential originfor fatigue crack nucleation.[1–4] The combination of pittingcorrosion and subsequent fatigue crack growth can signif-icantly reduce the lives of structural components made ofthese alloys in service. Mechanistic understanding of pittingin aluminum alloys and quantification of its kinetics, there-fore, are of scientific interest and technological importance.In previous studies, pitting in 7075-T6 and 2024-T3 alu-minum alloys has been identified by optical microscopy(OM) and scanning electron microscopy (SEM) with gal-vanic coupling between the matrix and the constituent par-ticles in these alloys.*[1,2] In particular, severe pit-

*The term ‘‘constituent particles’’ is used to designate the insoluble,undissolved, or precipitated coarse particles that are formed anddistributed heterogeneously in aluminum alloys from impurity elements,excess alloying elements, or improper heat treatment.

ting has been attributed to the successive interactions be-tween the alloy matrix and constituent particles in clus-ters.[5,6] In this study, a more direct observation of the natureand extent of particle-matrix interactions in an electrolytewas made with the aid of transmission electron microscopy(TEM).

Traditionally, the fundamental cause of pitting corrosionis attributed to the local breakdown of passivity of the sur-face film that forms over a metal surface, which permitssubsequent local dissolution of the metallic substrate.[7,8,9]

The resistance to pitting is believed to be determined bythe electrochemical stability of the ‘‘protective’’ passive

ROBERT P. WEI, Professor, is with the Department of MechanicalEngineering and Mechanics, Lehigh University, Bethlehem, PA 18015.CHI-MIN LIAO, Associate Scientist, is with China Steel Corporation,P.O. Box 47-29, Hsiao Kang, Kaohsiung 81233, Taiwan, R.O.C. MINGGAO, Engineering Advisor, is with Mobil Exploration & ProducingCenter (MEPTEC), Material, Corrosion and Inspection Group, FarmersBranch, TX 75381-9047.

Manuscript submitted July 1, 1997.

film and the ability of the film to repair itself in the elec-trolyte.[9] The propensity for pitting for a given metal-elec-trolyte system is characterized by its critical pittingpotential (Ep), an electrochemical measure for the instabilityof the passive film.[8,10,11] This approach is macroscopic andimplicitly assumes homogeneity for the underlying metals.Nevertheless, extensive efforts have been made to study theeffects of alloying elements, temperature, and electrolytechemistry on Ep for aluminum and its alloys.[12]

For engineering aluminum alloys (such as 7075-T6 and2024-T3), however, the critical pitting potential per sewould not serve as an adequate measure of pitting resis-tance because of microstructural heterogeneity at the alloysurface. These alloys generally contain significant amountsof constituent particles. In an electrolyte, local galvaniccells can form readily on the surface because of the differ-ence in electrochemical activity between these heteroge-neous phases and between the particles and the matrix andlead to accelerated local corrosion attack (pitting). Exam-ples of particle-induced corrosion in aluminum alloys in-clude those of iron-rich particles (such as Al3Fe),[13,14]

Al7Cu2Fe,[15] and MgZn2.[16]

More recent studies of particle-induced corrosion in7075-T6 and 2024-T3 aluminum alloys[1,2] indicated that theelectrochemical characters of constituent particles are dif-ferent and may be broadly divided into two groups. Parti-cles that contain Al, Cu, and Mg tend to be anodic relativeto the alloy matrix, while those that contain Al, Cu, Fe, andMn tend to be cathodic relative to the matrix. The OM andSEM observations suggested that anodic particles tended todissolve preferentially, whereas cathodic particles tended topromote dissolution of the neighboring matrix. The char-acterization of these particles in this manner, however, isnot precise. Because of observed deposition of Cu backonto the (anodic and cathodic) constituent particles duringcorrosion, most of the particles tend to behave cathodicallyand promote matrix dissolution.[1,2]

To provide a more complete understanding of the inter-actions of constituent particles and the matrix during cor-rosion, parallel studies were carried out by TEM to (1)identify the constituent particles and to determine their

1154—VOLUME 29A, APRIL 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Table I. Chemical Composition of Aluminum Alloys(in Weight Percent)

Alloy Cu Mg Si Ti Cr Fe Mn Zn Al

2024-T3 4.24 1.26 0.06 0.31 ,0.01 0.15 0.65 0.08 balance7075-T6 1.76 2.55 0.10 0.04 0.19 0.25 0.09 5.81 balance

Fig. 1—SEM micrographs of a typical surface area of a 2024-T3 aluminum alloy (a) before and (b) after immersion in 0.5M NaCl solution at roomtemperature showing cathodic behavior for essentially all of the constituent particles.

chemical composition and crystal structure and (2) deter-mine the nature and extent of their interactions followingexposure to an electrolyte. The first of these studies is re-ported in a separate article,[17] and the results of the secondstudy are described herein.

II. MATERIAL AND EXPERIMENTALPROCEDURES

A 1.3-mm-thick sheet of 7075-T6 and a 1.6-mm-thicksheet of 2024-T3 aluminum alloy were used in this study.*

*The 2024-T3 sheet was of 1960s vintage, and the 7075-T6 sheet wasof 1980s vintage.

The chemical composition of each alloy is given in TableI. The pitting characteristics of these alloys were examinedand described previously.[1,2,5,6] For the TEM studies, 3-mm-diameter disks were cut by electro discharge machiningfrom the alloys, with the disk plane parallel to the sheetsurface. The disks were mechanically thinned to a thicknessof 0.2 to 0.3 mm by manually polishing both surfaces ofthe disks through 240- to 600-grit abrasive papers. Thedisks were then electropolished in a 25 pct HNO3 1 75 pctmethanol solution at 240 7C to 250 7C and 20 V d.c. ina TENUPOL-3 jet polisher. The polishing procedure pro-duced sufficiently large, thin regions to facilitate TEM stud-ies.

The TEM studies were carried out using a PHILIPS*

*PHILIPS is a trademark of Philips Electronic Instruments Corp.Mahwah, NJ.

400T transmission electron microscope operated at 120 kV.Detailed characterizations of the constituent particles in the

alloys, along with verifications of the strengthening precip-itates, were made by a range of analytical electron micros-copy (AEM) techniques. The details are given in acompanion article.[17] For the corrosion studies, each of thethinned TEM specimens was examined by TEM to identifythe nature and location of the constituent particles of inter-est. (Because of observed ‘‘inhibition’’ of corrosion by ex-tended exposure to the electron beam, detailed examina-tions were not performed on these specimens.) The thickeredges of the specimens were masked off with stop-off lac-quer to localize corrosion in the thinned section. Each spec-imen was then immersed in a 0.5M NaCl solution (exposedto air, with pH ' 6.5 and [O2] ' 7 ppm) for up to 180minutes, quickly cleaned with distilled water and methanol,dried, and transferred into the microscope for examination.

III. RESULTS AND DISCUSSION

A. Particle Characterization

Previous investigations have shown that these alloys con-tained large numbers of constituent particles.[1,2] For the ma-terials used in this study, the density of constituent particlesin the rolling plane of the 2024-T3 sheet (with cross-sec-tional area greater than 1mm2) was estimated to be about3000 particles/mm2; the estimated density for the 7075-T6was about 1500 particles/mm2.[1,2] By using energy disper-sive X-ray analysis (EDS) with SEM and X-ray microprobeanalysis, the constituent particles in the 2024-T3 alloy wereseparated into those that contain Cu, Mg, and Al and thosecontaining Fe, Mn, Cu, and Al, and, in the 7075-T6 alloy,those that contain Cu, Mg, Zn, and Al and those with Fe,Mn, Cr, Cu, and Al.[1,2] The Cu- and Mg-containing parti-cles were believed to be anodic and the Fe- and Mn-con-taining particles to be cathodic, relative to the matrix.[2] Acomparison of the same area of surface of the 2024-T3alloy before and after corrosion (Figure 1), however,showed that nearly all of the particles (irrespective of type)behaved cathodically and promoted matrix dissolution. Thisbehavior will be considered in relation to the TEM obser-vations.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, APRIL 1998—1155

Table II. Corrosion Potential of Intermetallic Compounds and Aluminum Alloys

Stoichiometryor Alloy

Corrosion Potential(mV SCE) Environment Aeration Reference

Al2Cu 2700 0.5M NaCl open to air 18Al2Cu 2640 53 g/L NaCl 1 3 g/L H2O2 open to air 18Al2Cu 2621 3 pct NaCl not stated 18Al2Cu 2680 0.2M NaCl not stated 18Al2CuMg 2890 1.0M NaCl open to air 18Al2CuMg 2910 53 g/L NaCl 1 3 g/L H2O2 open to air 18Al3Fe 2470 53 g/L NaCl 1 3 g/L H2O2 open to air 18Al3Fe 2580 to 2390 3 pct NaCl not stated 18Al3Fe 2485 0.5M NaCl open to air 212024-T3 2600 ASTM G69 not stated 19, 202024-T3 2605 0.5M NaCl open to air 217075-T6 2740 ASTM G69 not stated 19, 207075-T6 2750 0.5M NaCl open to air 21Al (99.9 pct) 2730 0.5M NaCl open to air 21Al (99.999 pct) 2750 ASTM G69 not staged 19Cu (99.9 pct) 2200 0.5M NaCl open to air 19, 20Cu (99.999 pct) 0 ASTM G69 not stated 19

Fig. 2—TEM micrographs showing an Al23CuFe4 particle in a 7075-T6 aluminum alloy (a) before and (b) after 30 min and (c) 105 min (cumulative)immersion in 0.5M NaCl solution at 80 7C, along with (d) a reconstructed image of (c).

The constituent particles in the alloys were first examinedby AEM in the companion study.[17] Three principal typesof constituent particles were identified and characterized for

the 2024-T3 alloy and two for the 7075-T6 alloy. In the2024-T3 alloy, the particles are Al2CuMg, Al2Cu, and acomplex of Fe-Mn-Cu-Si-Al containing particles of the

1156—VOLUME 29A, APRIL 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3—Low- and high-magnification TEM micrographs showing a Al2Cuparticle and its environs in a 2024-T3 aluminum alloy (a) before and (b)after 180 min (cumulative) immersion in 0.5M NaCl solution at roomtemperature. Inset in (b) is an EDS spectrum that shows Cu to be theprincipal component in the corrosion product film.

Fig. 4—TEM micrographs showing a Al2CuMg particle and its environsin a 2024-T3 aluminum alloy before, and after 15 min immersion in 0.5MNaCl solution at room temperature.

form (Fe,Mn,Cu)xSi(Cu,Al)y (two possible types areAl8Fe2Si and Al10Mn3Si).[17] In the 7075-T6 alloy, the prin-cipal particles were identified to be of the Al23CuFe4 typeand ‘‘chemically inert’’ SiO2. Detailed descriptions of theseparticles are given in Reference 17.

Corrosion potential for some of these constituent parti-cles and alloys were obtained from the literature,[18–21] ormeasured, and are given in Table II. The corrosion poten-tials are referenced to the saturated calomel electrode(SCE), although other reference electrodes might have beenused for the original measurements.[18–20] Only data forNaCl solutions (with or without H2O2 additions) exposedto air (or aeration, not stated) are included.

B. Particle-Matrix Interactions

The simpler (relative to the type, density, and distributionof constituent particles) 7075-T6 alloy was selected for aninitial, more detailed, examination. The immersion experi-ments were carried out at about 80 7C to mitigate the an-ticipated beam-induced inhibition and to acceleratecorrosion. Based on the experience developed with the7075-T6 alloy, experiments were then conducted on the2024-T3 alloy at room temperature. The results aresummarized and discussed in terms of the principal typesof constituent particles in each of the alloys in Sections 1through 4.

1. Al23CuFe4 particles in 7075-T6Figure 2 shows representative TEM micrographs of a

typical Al23CuFe4 particle and its surrounding matrix taken

from a specimen before and after repeated immersion in0.5M NaCl solution at 80 7C. Evidence of corrosion maybe seen following 30 minutes of immersion and is in theform of hairlike corrosion products (Figure 2(b)). After acumulative total time of 105 minutes, however, the particlehad completely disappeared (i.e., fallen off), and the TEMmicrograph of the same area (at a lower magnification)showed significant dissolution of the surrounding matrix,leaving behind a thin film over a semicircular region (Fig-ure 2(c)). The thin film has been identified by electron dif-fraction as being made of randomly oriented grains ofprincipally aluminum oxide. By photographic reconstruc-tion, it may be seen that the particle fitted almost perfectlyinto its original position, nearly at the center of the corrodedregion (Figure 2(d)).

By comparing Figures 2(a) through (d), it is clear thatlocal dissolution of the matrix had resulted from galvanic

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, APRIL 1998—1157

Fig. 5—Low-magnification TEM micrographs showing a region thatcontains clusters of (Fe, Mn) containing particles in a 2024-T3 aluminumalloy (a) before and (b) after 180 min (cumulative) immersion in 0.5MNaCl solution at room temperature.

coupling between the particle and the matrix in which theparticle acted as the cathode and remained essentially un-affected. Dissolution of the matrix continued as long as theparticle retained electrical contact with the matrix. Matrixdissolution reached a region that is about 5 times the radiusof the particle. By assuming a foil thickness of 100 nm, theamount of aluminum dissolved in 105 minutes was esti-mated to be 1.6 3 10212 g. Because of the smaller arearepresented by the (cathodic) particle, dissolution would becathodically controlled. By assuming constant current den-sity over both surfaces of the particle, the current densitywas estimated to be about 550 mA/cm2 at 80 7C. Using anactivation energy of 39 kJ/mol,[22] the current density at 207C was estimated to be 36 mA/cm2. Current density overthe alloy (matrix) surface is expected to vary with distanceaway from the particle and to reflect the potential distri-bution around the particle.

2. Al2Cu particles in 2024-T3Figure 3 shows representative TEM micrographs of a

typical Al2Cu particle and its surrounding matrix takenfrom a specimen before and after immersion for a total of180 minutes in 0.5M NaCl solution at room temperature.The lower-magnification micrographs provide overall viewsof the region around the particle. The holes in these micro-graphs were produced by electropolishing during specimenpreparation, but indicate little or no overall (or general)corrosion of the 2024-T3 alloy matrix in 180 minutes. Thehigher-magnification micrographs show the microstructurearound the particle and the extent of corrosion of the sur-rounding matrix. The extent of the region of dissolutionwas again about 5 times the particle radius. The asymmetryof the corroded region reflected principally the variation (ortaper) in the thickness of the TEM specimen at the edge ofa hole. The EDS analysis (insert) showed that the darkenedappearance of the particle following corrosion reflectedplating of dissolved Cu onto the particle during corrosion.

Dissolution of the alloy matrix around the Al2Cu parti-cles was somewhat surprising in that the particle wasdeemed to be anodic relative to the matrix (with a corrosionpotential of 2700 mV SCE vs 2605 mV SCE for the ma-trix). Careful examination of the specimen (before corro-sion) showed a precipitate-free region around the particle.With the corrosion potential of pure aluminum of 2730mV SCE, galvanic coupling with the particle can initiatematrix dissolution in this precipitate-free region. Matrix dis-solution was then sustained as Cu plated out onto the par-ticle (with a corrosion potential of 2200 mV SCE), whichrendered it cathodic relative to the alloy matrix. By againassuming a foil thickness of 100 nm, the amount of alu-minum dissolved in 180 minutes was estimated to be 1.13 10211 g at a current density of about 200 mA/cm2 overthe particle surfaces.

3. Al2CuMg particles in 2024-T3Representative TEM micrographs of a typical Al2CuMg

particle and its surrounding matrix, taken from a specimenbefore and after immersion for 15 minutes in 0.5M NaClsolution at room temperature, are shown in Figure 4. Themicrographs show that after 15 minutes, the particle wascompletely dissolved (except for a small region marked byc). The corrosion products were identified by electron dif-fraction as a mixture of Cu and Cu2O. A small region (re-gion a) of the matrix was also dissolved during this period.These observations support the findings from the previousSEM study, which suggested the anodic nature of Al2CuMgand the preferential dissolution of Al and Mg from theseparticles. This localized dissolution of the matrix is consis-tent with the influence of copper (resulting from dealloyingof the particle) over the precipitate-free zone around theparticle (region a). The amount of Al2CuMg dissolved in15 minutes was estimated to be 2.6 3 10211 g at a currentdensity of about 180 mA/cm2 over the particle surfaces.

4. Fe-Mn-Cu-Si-Al particles in 2024-T3The composition and crystal structure of these constitu-

ent particles in this 2024-T3 are extremely complex. Theparticles are generally of the form (Fe,Mn,Cu)xSi(Cu,Al)y.They tend to be clustered, and the clusters appear in locallydense groups. Figure 5 shows TEM micrographs (at a lowmagnification) of a representative grouping of clustered

1158—VOLUME 29A, APRIL 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 6—TEM micrographs of region a in Fig. 5 (at a higher magnification) (a) before and (b) after 180 min (cumulative) immersion in 0.5M NaCl solutionat room temperature.

Fig. 7—Comparison of anodic and cathodic current densities as a functionof cathode-to-anode surface area ratio for an Al (anode)-Al3Fe (cathode)couple in 0.5M NaCl solution at room temperature.

Fig. 8—SEM micrograph showing the morphology of a typical particle-induced corrosion pit at the surface of a 2024-T3 aluminum alloy in 0.5MNaCl solution at 80 7C for 24 h.

particles and their surrounding matrix taken from a speci-men before and after immersion for 180 minutes in 0.5MNaCl solution at room temperature. Companion higher-magnification TEM micrographs for regions a, before andafter corrosion, are shown in Figure 6.

The low-magnification micrographs show the complexpatterns of particle-induced matrix dissolution surroundingthe particle clusters (see regions a, b, and c in Figure 5).Other regions, presumably as a result of particle-inducedlocal dissolution, have also become transparent to the elec-tron beam. The extent of corrosion is consistent with theprevious suggestion that matrix dissolution around the par-ticle clusters and the propagation of corrosion from clusterto cluster is the principal mechanism for the formation and

growth of severe corrosion pits.[1,2,5,6] Figure 6 shows theextent of corrosion around the cluster of particles at a ingreater detail. After 180 minutes, a part of the cluster ofconstituent particles had ‘‘fallen’’ away, and the foil in thatregion was also ‘‘broken.’’ Because of the complex natureof the particles and the potential influence of neighboring(or ‘‘hidden’’) particle clusters, only a rough estimate ofthe current density could be made. The estimate is about50 mA/cm2 over the particle surfaces.

C. Some Further Observations and Discussion

To better understand particle-induced corrosion, an in-vestigation of the galvanic coupling between a model com-pound (Al3Fe) and high-purity aluminum has beeninitiated.[19] Preliminary measurements at room temperaturein a 0.5M NaCl solution (exposed to air) show that the

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, APRIL 1998—1159

Fig. 9—Conceptual models of particle-matrix interactions (localcorrosion) for (a) a cathodic and (b) an anodic particle.

galvanic current density varied with the ratio of surfaceareas of Al3Fe (cathode) and Al (anode) (Figure 7). Whenthe cathodic area is small, current flow is limited by rateof reactions at the cathodic surface and is reflected by aconstant cathodic current density. When the cathodic sur-face is large relative to the anodic surface, on the otherhand, current limitation is transferred to the anode (or Al)and the process proceeds at a constant anodic current den-sity. For the Al3Fe-Al couple, the limiting cathodic and an-odic current densities were found to be about 38 and 31mA/cm2, respectively.[21] The cathodic current density isconsistent with those estimated for the (Fe,Mn) containingconstituent particles from the TEM studies.

A pair of scanning electron micrographs of the same areaof surface of the 2024-T3 alloy before and after corrodingfor 24 hours in a 0.5M NaCl solution are shown in Figure1. The large particle in the center is associated with the(Fe,Mn) containing cathodic particles.[1,2] The smaller roundparticles are identified with the anodic (Al2Cu and Al2CuMg)particles.[1,2] A comparison between the two surfaces clearlyshows that the sizes of corrosion pits induced by the par-ticles are nearly always larger than the particles themselves.This observation is consistent with cathodic behavior ofnearly all of the constituent particles in the alloy and withthe TEM observations. The morphology of a typical cor-rosion pit around a constituent particle at the surface isshown in Figure 8. The pit is nearly hemispherical in shape.The extent of matrix dissolution is consistent with the TEMobservations and suggests that the particle maintained elec-trical contact for an extended period of time.

Based on the TEM and SEM observations, it is clear that

pitting corrosion at the surface of these alloys resulted fromthe galvanic coupling between the constituent particles andthe alloy matrix. Conceptually, corrosion around a singlecathodic constituent particle proceeded under cathodic con-trol at a constant current density over the particle surface,as illustrated schematically in Figure 9(a). The dissolutionrate decreased with distance away from the particle as ex-pected, with electrical continuity maintained underneath theparticle; a more quantitative model of the process is beingdeveloped. A similar schematic model for dissolution of ananodic particle is shown in Figure 9(b). Dissolution occursat constant (anodic) current density over the constituent par-ticle (principally, Al2CuMg).

Corrosion around a single particle can expose other sub-surface particles to the electrolyte and allow further localdissolution of the matrix. If the process exposes a suffi-ciently large cluster or clusters of particles, then a severepit can develop, as proposed previously.[1,2,5,6] Matrix dis-solution within such a pit is expected to be enhanced bythe same particle-matrix galvanic couple, although the cur-rent densities would reflect the local electrochemical con-ditions within the pit. Because of the realization, throughthis study, that Al2Cu and Al2CuMg particles can also ex-hibit cathodic response, it is necessary to include the con-tributions of these particles into quantitative models forpitting corrosion of aluminum alloys.

IV. SUMMARY

A TEM study of the electrochemical interactions be-tween constituent particles and the alloy matrix was con-ducted on 2024-T3 and 7075-T6 aluminum alloys to betterunderstand particle-induced pitting corrosion in these al-loys. Thin foil specimens were examined by TEM beforeand following immersion in aerated 0.5M NaCl solution toidentify the constituent particles and characterize the re-sulting interactions (or corrosion). The results clearly con-firmed the previous observations by SEM and showed matrixdissolution around the Fe- and Mn-containing particles (suchas Al23CuFe4) as a result of galvanic coupling between theparticle and the matrix. Extensive matrix dissolution was alsoobserved around the nominally anodic Al2Cu particles as aresult of plating of Cu back onto the particles during cor-rosion. Although the Al2CuMg particles dissolved rapidly asa result of galvanic coupling to the matrix, some matrix dis-solution was noted from preferential dissolution of Al andMg from these particles, or Cu deposition.

Dissolution current densities were estimated based on theestimated amounts of material removed by galvanic corro-sion and reflected anodic or cathodic control by the parti-cles. For the anodic (Al2CuMg) particles, the estimatedcurrent density was about 180 mA/cm2 at room temperature.The estimated values were about 200 and 40 mA/cm2, re-spectively, for Al2Cu and the (Fe, Mn) containing particles;the latter value is consistent with that of a Al3Fe-Al couple.For the cathodic behaving particles, matrix dissolution ex-tended out to a distance about 5 times the particle radius.Based on these and the previous SEM observations, a con-ceptual model for corrosion induced by a single particle isproposed. A more quantitative model, incorporating poten-tial distribution around the particle, is being developed andwill be integrated into a model for severe pitting that in-volves clusters of constituent particles.

1160—VOLUME 29A, APRIL 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

ACKNOWLEDGMENTS

This research was supported by the Air Force Office ofScientific Research, Grant No. F49620-93-1-0426, and theFederal Aviation Administration, Grant No. 92-G-0006.The authors express their appreciation to Dr. Vinod Sikka,Oak Ridge National Laboratory, for providing the Al3Feintermetallic compound used in this study.

REFERENCES

1. G.S. Chen, M. Gao, D.G. Harlow, and R.P. Wei: FAA/NASA Int.Symp. on Advanced Structural Integrity Methods for AirframeDurability and Damage Tolerance, NASA Conference Publication3274, Langley Research Center, Hampton, VA, Sept. 1994, pp. 157-73.

2. G.S. Chen, M. Gao, and R.P. Wei: Corrosion, 1996, vol. 52 (1), pp.8-15.

3. S. Chen, K.-C. Wan, M. Gao, R.P. Wei, and T.H. Flournoy: Mater.Sci. Eng., 1996, vol. A219, pp. 126-32.

4. R.S. Piascik and S.A. Willard: Fat. Fract. Eng. Mater. Struct., 1994,vol. 17 (11), p. 1247.

5. Chi-Min Liao, Jean Marc Olive, Ming Gao, and Robert P. Wei:Corrosion, 1998, in press.

6. Chi-Min Liao, Gim S. Chen, and Robert P. Wei: Scripta Mater., 1996,vol. 35 (11), pp. 1341-46.

7. Metals Handbook, 9th ed., vol. 13, Corrosion, ASMINTERNATIONAL, Metals Park, OH, 1987, p. 113.

8. Z. Szklarska-Smialowska: Corrosion, 1971, vol. 27, p. 223.9. J. Zahavi and M. Metzger: Localized Corrosion, Dec. 6–10, 1971,

R.W. Staehle, B.F. Brown, J. Kruger, and A. Agrawal, eds.,International Corrosion Conf. Series, NACE-3, Houston, TX, 1974,p. 547.

10. T.P. Hoar: Localized Corrosion, Dec. 6–10, 1971, R.W. Staehle, B.F.Brown, J. Kruger, and A. Agrawal, eds., International CorrosionConference Series, NACE-3, Houston, TX, 1974, p. 112.

11. A. Broli and H. Holtan: Corr. Sci., 1973, vol. 13, p. 237.12. H. Bohni and H.H. Uhlig: J. Electrochem. Soc., 1969, vol. 116, p.

906.13. W.K. Johnson: Br. Corros. J., 1971, vol. 6, p. 200.14. C.H. Paik: Ph.D. Dissertation, University of Illinois, Urbana-

Champaign, Urbana, 1993.15. T.G. Dunford and B.E. Wilde: in Field Metallography, Failure Analysis,

and Metallography, M.E. Blum, P.M. French, and G.F. Van der Voort,eds., ASM INTERNATIONAL, Metals Park, OH, 1987, p. 263.

16. E. Lunarska, E. Trela, and Z. Szklarska-Smialowska: Corrosion,1987, vol. 43, p. 219.

17. Ming Gao, C.R. Feng, and Robert P. Wei: Metall. Mater. Trans. A,1998, vol. 29A, pp. 1145-51.

18. R.G. Buchheit: J. Electrochem. Soc., 1995, vol. 142 (11), pp. 3994-96.

19. T.D. Butleigh, R.C. Rennick, and F.S. Bovard: Corrosion, 1993, vol.49 (8), pp. 683-85.

20. ASM Specialty Handbook: Aluminum and Aluminum Alloys, ASMINTERNATIONAL, Materials Park, OH, 1994, p. 581.

21. Chi-Min Liao: Ph.D. Dissertation, Lehigh University, Bethlehem, PA,1997.

22. Raymond M. Burynski, Jr., Gim-Syang Chen, and Robert P. Wei:1995 ASME Int. Mechanical Engineering Congress and Exposition onStructural Integrity in Aging Aircraft, San Francisco, CA, November12-17, 1995, C.I. Chang and C.T. Sun, eds., ASME, New York, NY,1995, pp. 175-83.