EXFOLIATION CORROSION KINETICS OF HIGH STRENGTH ALUMINUM ALLOYS DISSERTATION. Xinyan Zhao, M.S. ***** The Ohio State University

Transcription

1 EXFOLIATION CORROSION KINETICS OF HIGH STRENGTH ALUMINUM ALLOYS DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School of The Ohio State University By Xinyan Zhao, M.S. ***** The Ohio State University 2006 Dissertation Committee: Dr. Gerald S. Frankel, Adviser Dr. Rudolph G. Buchheit Dr. Michael J. Mills Dr. Stanislav I. Rokhlin Approved by Adviser Graduate Program in Materials Science and Engineering

2

3 ABSTRACT The objectives of this study were to quantitatively study localized corrosion, especially exfoliation corrosion (EFC), of high strength aluminum alloys and to investigate the mechanism of exfoliation corrosion with a focus on the effects of alloy temper, microstructure, relative humidity (RH) and mechanical stress. A new technique, Exfoliation of Slices in Humidity (ESH) was developed for the determination of exfoliation corrosion (EFC) susceptibility and quantification of EFC kinetics. Two AA7178 plates taken from the wingskin of a retired KC135 airplane were used as test samples. Slices of the plate were pretreated by potentiostatic polarization in chloride solution to develop localized corrosion sites. Subsequent exposure to high humidity after pretreatment of properly oriented and unconstrained samples resulted in the development of EFC at the edges of the slices. The EFC kinetics was determined by measuring the width of the central unattacked region of the samples. The ESH results were representative of the different EFC behavior of the two plates during outdoor exposure. These results show the capability of the ESH test to discriminate between plates of varying susceptibility and to determine EFC rates ii

4 quantitatively. Optical microscopy and analytical TEM were used to investigate the effect of microstructure and local chemistry at grain boundary on EFC susceptibility. Alloys with more elongated grain shape are more susceptible to EFC and a high Zn content in the grain boundary precipitate free zone relates to a high susceptibility. The effects of RH, temper and applied stress on EFC kinetics of AA7178 were investigated by ESH tests. The critical RH for EFC propagation in AA7178 was found to be about 56% and the EFC kinetics increased with RH. The effect of temper in ESH test was consistent with published results from EXCO tests, but they provide a quantitative description of the temper effect. The effects of applied compressive and tensile stresses on EFC kinetics were studied using a four-point bending jig. Compression accelerated EFC significantly and tension reduced kinetics. An equation describing the effects of RH, stress, and time on EFC kinetics was developed based on the ESH results using Eyring model. In situ X-ray radiography was used to characterize intergranular and exfoliation corrosion in high strength Al alloys. The samples were either exposed to sodium chloride solution (NaCl) at a controlled potential or to high humidity after an electrochemical pretreatment in NaCl solution. Intergranular corrosion (IGC) growth rates in solution for samples encased in epoxy and exposed to solution were distributed over a range with an upper limit equal to the rates determined by the foil penetration technique. Unencased AA2024-T3 samples that were held at a controlled potential in solution exhibited IGC and then exfoliation corrosion. AA2024 and AA7178 samples encased in epoxy developed sharp intergranular fissures during exposure to 96% RH following an electrochemical pretreatment. Unencased samples given the same iii

5 treatment exhibited exfoliation in the high humidity environment. In situ X-ray radiography of intergranular corrosion attack provides a wide range of intergranular corrosion kinetics including the fastest growing sites that can be detected by foil penetration technique. This method is a good approach for visualizing the EFC process. iv

6 To my father, Tianhai Zhao and my mother, Fengqin Zhao v

7 ACKNOWLEDGMENTS I would like to express my sincere thanks to my adviser, Dr. Gerald Frankel, for his guidance, inspiring motivation, and helps during the supervision of this project. I have been fortunate to study under his tutelage. Not only his knowledge, but also his passion towards aircraft corrosion has been inspired me a lot through 5 years of graduate study. Next, I would like to thank Dr. Rudy Buchheit, for his participation on my academic advisory committee and for comments and suggests on my research. I am also grateful for a lot of good comments and suggestions given by Dr. Michael Mills and Dr. Stanislav Rokhlin, both on my advisory committee. I would like to acknowledge funding from the Aging Aircraft Division of ASC in support of the Aeronautical Enterprise Structures Strategy with a contract through S&K Technologies. My special thanks go to Dr. Bahman Zoofan from Welding department for his kind help and corporation with the use of X-ray radiography facility. Dr. Bahman Zoofan has been spending many hours working with me and contributes a great amount of work to this effort. I also want to thank Mr. Andrew Bonifas for his input to this project. He helped and designed the four point bending apparatus for studying mechanical effect on exfoliation corrosion. I also wish to thank Mr. Jim Suzel at S&K vi

8 Technology for providing the AA7178 samples and Dr. William Abbott of Battelle for performing the environmental exposure testings for these samples. I would like to thank all members of Fontana Corrosion Center: Dr. Weilong Zhang, Dr. Wenping Zhang, Dr. Qingjiang Meng, Dr. Xiaodong Liu, Dr. Tsai-Shang Huang, Ms. Hong Guan, and many current FCC members. I would also like to thank Mr. Henk Colijn, Mr. Cameron Begg and Dr. Lisa Hommel, who helped and trained me how to use TEM, FIB, SEM and XPS. I would like to thank Ms. Dena Bruedigam, Ms. Chris Putnam for their office help. I would like to acknowledge Mr. Steve Bright, Mr. Ken Kushner for their helps in preparing my experimental samples. I also wish to thank Dr. Suliman Dregia, Mr. Mark Cooper, Ms. Mei Wang, and Ms. Wendy Ranney for their kind help with my academic problems. Finally, I would like to thank my parents for their unwavering support and encouragement throughout my educational years. vii

9 CURRICULUM VITA July 1999 B.S. Materials Physics, University of Science and Technology Beijing, China. March 2003 M.S. Materials Science and Engineering, The Ohio State University present Graduate Research Associate, Materials Science and Engineering, The Ohio State University. PUBLICATIONS 1. X.Zhao, G.S.Frankel, B.Zoofan, and S.I.Rokhlin, In-Situ X-Ray Radiographic Study of Intergranular Corrosion in Al Alloys, Corrosion, 59, (2003) 2. X.Zhao, G.S.Frankel, The Visual Determination of Exfoliation Rate of Al Alloy Slices in Humidity, The Journal of Corrosion Science and Engineering, Volume 6 Paper C134, 2003, The paper was presented at the conference Corrosion Science in the 21 st Century held at UMST in July X. Zhao, T. Huang, G. S. Frankel, B. Zoofan, and S. I. Rokhlin, " Intergranular Corrosion Morphology and Growth Kinetics In High Strength Aluminum Alloys," in Critical Factors in Localized Corrosion IV, S. Virtanen, P. Schmutz, and G. S. Frankel, eds, PV , The Electrochemical Society, T.Huang, X.Zhao, Gerald Frankel, B.Zoofan and S.Rokhlin, Growth Kinetics of Intergranular and Exfoliation Corrosion in AA7178, proceedings of Triservice Corrosion Conference, Las Vegas, 2003 viii

10 5. Tsai-Shang Huang, Xinyan Zhao, and G.S.Frankel, Localized Corrosion Growth Kinetics in AA7xxx Alloys, 2005 Triservice Corrosion Conference Proceedings, G.S.Frankel, Tsai-Shang Huang, and Xinyan Zhao, Localized Corrosion Growth Kinetics in AA7178, P.Marcus, ed., Elsevier, 2005 FIELDS OF STUDY Major Field: Materials Science and Engineering ix

11 TABLE OF CONTENTS x Page ABSTRACT...II ACKNOWLEDGMENTS... VI VITA...VIII LIST OF TABLES...XIII LIST OF FIGURES...XIV CHAPTERS: 1. INTRODUCTION LITERATURE REVIEW BASIC FORMS OF CORROSION OF ALUMINUM ALLOYS PITTING EXFOLIATION CORROSION The Mechanism of Exfoliation Corrosion Factors That Influence EFC Effects of Alloying Elements on EFC Effects of The Grain Shape on Exfoliation Corrosion Effect of Heat Treatment on Susceptibility to Exfoliation Corrosion Environmental Aspects Exfoliation Corrosion Testing Immersion Test Salt Spray Tests Electrochemical Impedance Spectroscopy Deflection Technique INTERGRANULAR CORROSION IGC Mechanisms Galvanic Couple Theory Precipitate Free Zone Breakdown Model Anodic Dissolution of Grain Boundary Precipitates...32

12 2.4.2 Alloy Microstructure Intermetallic Particles Intermetallic Particles in Al-Cu & Al-Cu-Mg Alloys Intermetallic particles in Al-Zn-Mg & Al-Zn-Mg-Cu Alloys Other Intermetallic Particles Precipitate Free Zones (PFZ) INTERGRANULAR STRESS CORROSION CRACKING (IGSCC) Mechanisms of SCC Anodic Dissolution Mechanism Cathodic Mechanism Effects of Microstructure and Alloy Chemistry on IGSCC Relationship between Stress Corrosion Cracking and Exfoliation Corrosion RESEARCH OBJECTIVES QUANTITATIVE STUDY OF EXFOLIATION CORROSION: EXFOLIATION OF SLICES IN HUMIDITY TECHNIQUE INTRODUCTION EXPERIMENTAL Materials Exfoliation of Slices in Humidity (ESH) Technique RESULTS AND DISCUSSION Electrochemical and ESH Characterization Optical Microscopy TEM Analysis CONCLUSIONS EFFECTS OF RH, TEMPER AND STRESS ON EXFOLIATION CORROSION KINETICS OF AA INTRODUCTION EXPERIMENTAL Four Point Bending Technique RESULTS AND DISCUSSION Effect of RH on EFC Kinetics Effect of Temper on EFC Kinetics Effect of Stress on EFC Kinetics xi

13 4.3.4 Empirical Model for Exfoliation Corrosion CONCLUSIONS IN SITU X-RAY RADIOGRAPHY OF LOCALIZED CORROSION INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS CONCLUSION AND FUTURE WORK CONCLUSIONS FUTURE WORK APPENDIX A: FINITE ELEMENT SIMULATION A.1 PHENOMENOLOGICAL MODEL OF EXFOLIATION CORROSION A.2 MODEL DEVELOPMENT IN ABAQUS A.3 RESULTS AND DISCUSSIONS BIBLIOGRAPHY xii

14 LIST OF TABLES Table Page 3. 1 Humidity associated with saturated salt solutions Chemical composition of AA Data obtained from XPS spectra Composition of AA2024 and AA7178 plates A. 1 Stress strain input data xiii

15 LIST OF FIGURES Figure Page 2.1 Exfoliation resulting from rapid lateral attack of selective boundaries or strata forming wedges of corrosion product that force layers of metal upward giving rise to a layered appearance.[36] (Reprinted, with permission, from ASTM G34-90 Standard Test Method for Exfoliation Corrosion Susceptibility in 2XXX and 7XXX Series Aluminum Alloys (EXCO Test), copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428) Schematic graph of the variation of corrosion product wedging force as a function of time.[23] (Reprinted from Corrosion Science, Vol. 23, M.J. Robinson, The Role of Wedging Stresses in The Exoliation Corrosion of High Strength Aluminum Alloys, Pages: , Copyright (1983), with permission from Elsevier.) (a) Model of the intergranular penetration of a wrought alloy showing the boundary of the attack marked by the dotted line. (b) Improved model in which the corrosion product tapers uniformly towards the tips.[9] (Reprinted from Corrosion Science, Vol. 22, M.J. Robinson, Mathematical Modelling of Exfoliation Corrosion in High Strength Aluminum Alloys, Pages: , Copyright (1982), with permission from Elsevier.) (a) Dimensions of a thin walled hemispherical shell used to calculate the excess pressure in a blister. (b) Section through a blister diameter.[9] (Reprinted from Corrosion Science, Vol. 22, M.J. Robinson, Mathematical Modelling of Exfoliation Corrosion in High Strength Aluminum Alloys, Pages: , Copyright (1982), with permission from Elsevier.) Surface rating for 2024 in the MASTMASSIS test. (P=pitting, EA=Superficial EFC,EB=Moderate EFC,EC=Severe EFC,ED=Very Severe EFC). [1] (Reprinted from Corrosion Science, Vol. 41, M.J. Robinson and N.C. Jackson, The Influence of Grain Structure and Intergranular Corrosion Rate on Exfoliation and Stress Corrosion Cracking of High Strength Al-Cu-Mg Alloys, Pages: , Copyright (1999), with permission from Elsevier.)...64 xiv

16 2.6 (a) Corrosion product forces generated on T/2 specimens of 8090 plates. (b) Corrosion product forces generated on T/4 specimens of 8090 plate.[8] (Reprinted from Corrosion Science, Vol. 49, D.J. Kelly and M.J. Robinson, Influence of Heat Treatment and Grain Shape on Exfoliation Corrosion of Al- Li Alloy 8090, Pages: , Copyright (1993), with permission from Elsevier.) Schematic equivalent circuit for the immersion of AA7075 in EXCO solution. (Rs: solution resistance; Cdl: origin surface double layer capacitive; Rp: polarization resistance; Clf: low frequency capacitive; Rt: charge transfer resistance). [57] Schematic illustration of a mechanism for redistribution of Cu by dissolution of large Al 2 CuMg and Al 2 Cu intermetallic particles in Al alloys. [80] (Reproduced by permission of The Electrochemical Society, Inc.) Potentiodynamic polarization scans for bulk Al 2 CuMg in 0.5M NaCl solution open to air, actively aerated, and actively deaerated by sparging with N 2.[75] (Reproduced by permission of The Electrochemical Society, Inc.) Variation in corrosion potential of FeAl 3 and α-al(fe,mn)si.[91] (Reproduced by permission of The Electrochemical Society, Inc.) Anodic part of cyclic polarization curve for Al 3 Fe in deareated 0.1N NaOH solution. The curve corresponds to the first cycle starting with an as polished surface at 1.5V. In this particular run, the sweep was reversed at 0V. Continuous curve: forward sweep. Dashed curve: reverse sweep. [91] (Reproduced by permission of The Electrochemical Society, Inc.) Schematic diagram of typical crack propagation rate as a function of crack tip stress intensity behavior illustrating the regions of stage 1,2 and 3 crack propagation as well as identifying the plateau velocity and the threshold stress intensity.[12] (Reprinted with permission of ASM International. All rights reserved) Photo of AA7178 wingskin sample after 9 months of exposure at Daytona Beach. Sample was uncoated and had steel rivets attaching plates to understructure. Image provided by W. Abbott, Battelle Schematic drawing of orientation of slices for ESH test relative to elongated microstructure of AA7178 wingskin sample Polarization curves of AA7178 good and bad sample in deareated 1.0 M NaCl at a scan rate of 0.1 mv/s xv

17 3.4 Images of AA7178 wingskin samples exposed to 96% humidity following electrochemical pretreatment. (a) Bad sample. (b) Good sample Average exfoliation corrosion depth for duplicate samples of good (triangles) and bad (circles) AA7178 wingskin samples exposed to 96% RH Average exfoliation depths of good (open symbols) and bad (closed symbols AA7178 wingskin samples exposed to different humidities: circles 96%, squares 76%, triangles 65% Metallographic sections of good plate and bad plate. Also given is the terminology used for the different sections. (a) bad plate. (b) good plate Metallographic sections of AA7178 wingskin plate. (a) Bad plate. (b) Good plate. The sections are through-thickness montages, starting at the right of the top image in each pair and then wrapping around to end at the left side of the bottom image in each pair Grain size distribution and grain aspect ratios through the thickness of (a) good plate and (b) bad plate TEM micrographs of grain-boundary in AA7178 good plate. (a) Near up side of plate. (b) Near down side of the plate TEM micrographs of grain-boundary in AA7178 showing Nano EDS line profiling across two different types of the GB precipitates (a,b) near up side of the good plate and (c,d) near down side of the good plate TEM micrographs of grain-boundary in AA7178 showing Nano EDS line profiling across the GB SDZ (a) Near up side of the good plate. (b) Near down side of the good plate Nano-EDS line profile of grain boundary precipitates. (a, b) linescans across grain boundary precipitates in figure 3.11a and b near upside of the good plate. (c,d) ) linescans across grain boundary precipitates in figure 3.11c and d near downside of the good plate. The data are reported as ratio of the x-ray intensities for each element relative to that of Al Nano-EDS line profile of solute depleted zone around grain boundary. The data are reported as ratio of the x-ray intensities for each element relative to that of Al. (a) Near up side of the good plate (shown in figure 3.12a). (b) Near down side of the good plate (shown in figure 3.12b) Heat treatment of AA7178 wingskin plate xvi

18 4.2 Current response of two AA7178 samples during potentiostatic pretreatment at mv SCE in 1 M NaCl. Both samples had the same exposed area Polarization curve of AA7178 S sample in deareated 1.0 M NaCl at a scan rate of 0.1 mv/s Four point bending mode Four point bending jig Four point bending test set up SEM image of the surface of AA7178 sample after electrochemical pretreatment at 5.5 ma/cm 2 in 1M NaCl for 7 hours SEM micrograph of FIB cross section of IGC attack in the same AA7178 sample as in figure AA7178 wingskin slices after pretreatment and then 11 days in different constant humidity. The rolling direction is vertical AA7178 wingskin slices after pretreatment and then 13 days in 30% constant humidity. The rolling direction is vertical Width change of central unexfoliated region as a function of time for AA7178 wingskin in various constant RH environments (a) SEM micrographs of corrosion product of AA7178 during 66%RH exposure following electrochemical pretreatment. (b) SEM micrographs of corrosion product of AA7178 during 76%RH exposure following electrochemical pretreatment. (c) SEM micrographs of corrosion product of AA7178 during 96%RH exposure following electrochemical pretreatment X-ray EDS spectrograms of corrosion product in 56%RH. Accelerating voltage=12kv XPS spectra measured from samples of AA7178 exposed to 56%RH after electrochemical pretreatment in 1 M NaCl for 7 hours. (a) Al 2p, (b) Cl 2p, (c) O 1s, (d) Zn 2p, (e)cu 2p, (f) Mg 2p Results of cyclic exfoliation test for AA7178 wingskin samples Image of AA7178 sample exposed to 96%humidity following galvanostatic pretreatment. (a) T6 temper. (b) T7 temper Exfoliation corrosion kinetics of AA7178 samples exposed to 96% RH Image of AA7178 sample under four point bending loading exposed to 96%humidity Exfoliation corrosion kinetics under compression and tension Exfoliation corrosion kinetics under different strain conditions Predictive 3D model for EFC kinetics Average kinetics of exfoliation corrosion in 96%RH under tension xvii

19 5.1 Schematic of the microradiographic system In situ x-ray microradiographic experimental setup Metallographic sections of AA2024-T3 and AA7178-T6 wingskin, and the terminology used for the three orientations, L: longitudinal, T: long transverse, S: short transverse Metallographic cross section of ST oriented 3 mm wide AA2024-T3 cylinder exposed to 1.0 M NaCl at -580 mv SCE for 4 hr In situ x-ray radiograph of L oriented AA2024-T3 sample exposed to 1.0 M NaCl at -580 mv SCE for 19 hr Plot of the depth of various specific sites for L oriented AA2024-T3 sample as a function of time. Also shown is the curve for L oriented foil penetration samples Radiographs of 2x2 mm sample of AA2024-T3 exposed to 1.0 M NaCl at 580 mv SCE. Sample not encased in epoxy. (a) Image was taken before exposure, (b,c) images were taken after 4 h. Long axis (vertical orientation of sample) is L direction. 90 image was taken in T direction, 0 image was taken in S direction Same sample as Figure 7, taken in T direction. Exfoliation is evident In situ x-ray radiograph of L oriented AA7178 sample exposed to 1.0 M NaCl at -725 mv SCE for 3 hrs Plot of the depth of various specific sites for L oriented AA7178 wingskin sample as a function of time Samples were put in 96%RH after electrochemical treatment. Sharp IGC continued along L direction at high rate. (a) X-ray radiograph taken at 0 days. (b) X-ray radiograph taken at 48days SEM image of intergranular sharp fissure in AA7178-T6 after 7 hours in 1 M NaCl at a potential of 710mV SCE Metallographic cross-section of AA2024-T3 L sample after 10 days in the environment of RH 96%, T=22C~24C Samples were put in 96%RH after above electrochemical treatment. Sharp IGC continued along L direction at high rate. (a) 0 day in 96%RH (after 7 hours in 1 M NaCl, -710mV SCE). (b) 6 days in 96%RH. (c) 30 days in 96%RH xviii

20 5.15 Samples were put in 96%RH after above electrochemical treatment. Sharp IGC continued along T direction at high rate. (a) 0 day in 96%RH (after 7 hours in 1 M NaCl, -710mV SCE). (b) 6 days in 96%RH. (c) 48 days in 96%RH A. 1 IGC attacks are created in the sample from electrochemical pretreatment A. 2 EFC process of a sample exposed in high relative humidity. (L section) A. 3 Two dimensional finite element mesh A. 4 2D contour plot showing stress field around intergranular crack tip simulated by ABAQUS xix

21 CHAPTER 1 INTRODUCTION This dissertation consists of six chapters including this introduction. Chapter 2 is a literature review on several topics. Exfoliation corrosion is reviewed first including the effect of alloy microstructure, alloy addition, environmental condition, heat treatment and techniques for exfoliation corrosion testing. Then intergraunular corrosion and stress corrosion cracking are discussed and reviewed. Since there is a close relationship between these forms of localized corrosion, a good understanding of SCC and IGC will help us to investigate the mechanism of EFC and develop a good methodology to quantify the corrosion process. Chapters 3 through 5 cover the details of the technical findings of this dissertation. They are written as stand alone papers, and will be submitted individually for publication. Chapter 3 describes the development of a novel approach for studying EFC kinetics. Electrochemical methods such as cyclic polarization, galvanostatic and potentiostatic polarization together with materials characterization techniques such as optical metallography, SEM/EDS, TEM, FIB, and STEM/Nano-EDS were utilized in this 1

22 study. The role of grain aspect ratio and the distribution of Zn at the grain boundary on EFC susceptibility were studied and discussed. Chapter 4 describes the effect of relative humidity, mechanical stress and heat treatment on EFC kinetics. The ESH technique developed in chapter 3 was utilized for the study of these factors. A four point bending jig was designed to investigate the effect of mechanical stress on EFC. An empirical model was then developed utilizing an Erying relationship. Chapter 5 covers a study of intergranular corrosion and EFC as well as the transition from IGC to EFC. In situ X-ray radiography was utilized to study intergranular corrosion over a wide range of corrosion kinetics and visualize the process of exfoliation corrosion in immersed condition and humidity exposure. The conditions required for IGC and EFC to occur were discussed in this chapter. Chapter 6 summarizes the findings of this work and states conclusions on the kinetics of exfoliation corrosion. This chapter also covers suggestions for future work and some unresolved issues. A finite element simulation of EFC process utilizing commercial software ABAQUS is covered in Appendix A. The simulation is based on a phenomenological model of exfoliation corrosion in humidity. 2

23 CHAPTER 2 LITERATURE REVIEW INTRODUCTION High strength aluminum alloys such as AA2xxx and AA7xxx are extensively used for aircraft structures. Localized corrosion is a major cause of their failure. Among various types of localized attack that aluminum alloys undergo, exfoliation corrosion (EFC) is an important form. EFC is thought to be a particular form of intergranular corrosion (IGC) that can occur on the surface of wrought aluminum alloys with an elongated grain structure.[1] During this corrosion process, internal stresses caused by voluminous corrosion products force layers of uncorroded metal away from the body of the material, giving rise to a layered appearance.[2] An example of EFC is shown in figure 2.1. EFC can lead to metal failure by gradual disintegration and the presence of exfoliation corrosion reduces the toughness of the alloy and enhances the fatigue crack growth rate.[3-7] All of these will cause serious problems to aircraft and limit their 3

24 lifetime. EFC has been studied since the 1950 s, yet crucial phenomena remain unclear. Several papers suggested that the exfoliation corrosion results either from intergranular corrosion or alternatively from a stress corrosion mechanism.[1, 8-10] Experiments conducted by Robinson s group have shown a strong similarity between the mechanism of stress corrosion cracking (SCC) and EFC mechanism.[1, 8-10] However, exactly how this mechanism works in the case of EFC is still unknown. Wedging stresses that develop by the formation of corrosion product play a crucial role in the kinetics of EFC and make exfoliation corrosion more difficult to understand. Though many efforts have been made to enhance the resistance of aluminum alloys to EFC, the lack of a quantification method and limited laboratory tests for EFC have slowed down the development of protection methods for EFC. The purpose of this chapter is to provide an overview of previous efforts and accomplishments towards the understanding of the EFC of aluminum alloys including the critical factors that influence exfoliation corrosion and the relationship between EFC, IGC and SCC. 2.1 BASIC FORMS OF CORROSION OF ALUMINUM ALLOYS Summerson and Sprowls have classified the corrosion of aluminum alloys into two groups. The kinetics of intergranular corrosion (IGC), stress corrosion cracking (SCC) and exfoliation corrosion (EFC) are more dependent on the metallurgical structure of the alloys and are so-called structure dependent corrosion.[2] They are major problems in aircraft structures. The corrosion rate of this type of corrosion is strongly influenced by grain shape.[1, 11, 12] Uniform corrosion and other forms of localized corrosion, 4

25 mainly pitting corrosion, are more dependent on local chemistry of the environment and are included in the second group. 2.2 PITTING Pitting is an important form of localized corrosion and it has been extensively studied. Pitting can lead to IGC, EFC and SCC under certain circumstances, so it is important to understand pitting corrosion, its initiation and propagation. According to Frankel, pitting corrosion is defined as localized accelerated dissolution of metal that occurs as a result of a breakdown of the otherwise protective passive film on the metal surface.[13] The presence of specific anions such as Cl - or Br - is required for the initiation of pitting corrosion.[2] Furthermore, the distribution of pitting of Al alloys is influenced by the distribution of cathodic metallic constituents in the metal surface. Pitting can develop into IGC, which will initiate EFC and SCC when interacting with stress, eventually leading to catastrophic failure of the structure. A characteristic feature of pitting corrosion is the pitting potential. The pitting potential is the potential above which pits will initiate and below which they will not. The pitting potential is sometimes referred to breakdown potential related to the localized breakdown of oxide film.[14] Since each intermetallic phase has a unique breakdown potential in given electrolyte, the potential measurement can be used to characterize the specific phases in alloys. A phase is unattacked at potentials below its breakdown potential.[15] 5

26 2.3 EXFOLIATION CORROSION The Mechanism of Exfoliation Corrosion Exfoliation corrosion is often observed in 2xxx (Al-Cu-Mg) and 7xxx (Al-Zn-Cu- Mg) series high strength alloys.[16] This type of corrosion is characterized by lamellar surface attack of alloys containing a highly directional grain structure. Robinson has pointed out that exfoliation corrosion is generally intergranular in nature due to the galvanic interaction between grain boundary precipitates and the adjacent matrix.[1] If the precipitates are active to the matrix, they may be preferentially dissolved. If they are noble to the matrix, then attack at the solute depleted zones around the precipitates is probable. Intergranular corrosion can proceed along multiple intergranular paths parallel to the surface. The hydrated aluminum oxide corrosion product has a higher volume than the alloy matrix that converts into the corrosion product. Consequently, relative large wedging stresses are produced at the IGC front and lift the surface grains, promoting further attack. The corrosion product of aluminum alloys is assumed to be hydrated alumina.[17, 18] During localized corrosion in chloride environments, hydrolysis of aluminum ions can be described as following reaction: AlCl 3 +3H 2 O=Al(OH) 3 +3HCl The aluminum hydroxide gel is not stable and crystallizes with time to convert into the rhombohedral monohydrate (Al 2 O 3 H 2 O or boehmite), the monoclinic trihydrate (Al 2 O 3 3H 2 O or bayerite) and another monoclinic trihydrate(hydrargilite).[19] Generally speaking, three factors which are required for exfoliation to occur[10, 20]: 1. Highly directional microstructure. 6

27 2. A preferential anodic path along grain boundaries resulting from the electrochemical action between precipitates and adjacent solute denuded zones.[21] 3. A Corrosive environment. The driving force for EFC is the wedging stress that is produced by the insoluble corrosion product. Pickering clearly stated that the wedging stress could be produced when the volume of insoluble corrosion product is larger than the metal from which it formed and the product is formed in a restricted region.[22] The corrosion product in 18-8 stainless steel was calculated to be 4000 psi to 7000 psi, providing all the energy required for stress corrosion cracking.[22] Robinson et al. have measured the corrosion product force in aluminum alloys.[10, 23, 24] They proposed that the stress corrosion cracking mechanism is probably the predominant factor that determines the kinetics of exfoliation corrosion.[9] The role of wedging forces on exfoliation corrosion of aluminum alloys was studied by using a special jig with a load cell that transmitted the force from the propagating exfoliation blister.[23] The wedging force generated by insoluble corrosion products in aluminum alloy L95 (W)(Zn 5.8, Mg 2.5, Cu 1.6, Cr 0.15) developed during immersion in EXCO solution was measured. It was found that there is an incubation period, during which the corrodant penetrated the surface layer of wrought aluminum grains.[23] After the incubation time, the force increased linearly until it attained a maximum value of Newtons in EXCO when observations showed that the entire area of the surface had suffered exfoliation corrosion. This process is shown in figure 2.2. The period before the maximum value was reached was considered to be due to the build-up of compressive stresses overcoming the propagation of exfoliation. The corrosion product pressure was 7

28 also calculated to be 2MN m -2 when the surface of the specimen was affected by exfoliation. The highest wedging pressure within the corrosion product is 4.8MNm -2. The stress intensity at the developing corrosion front in exfoliation corrosion was then calculated by crack opening displacement measurement (2.9MN m -3/2 ) and it had a comparable value with the K ISCC (<=2.6MNm -3/2 ) for alloy L95 in EXCO derived from double cantilever beam specimen measurements, indicating that SCC is a possible mechanism. Kelly also measured the corrosion product force of Al-Li alloy AA8090 using the same experimental arrangement as in Robinson s experiment.[10, 23] The magnitude of maximum corrosion product force generated in AA8090 with different aging time was ranked in the following order: under aged > over aged > peak aged condition for the plate and sheet material. He concluded that the EFC susceptibility of the precipitation hardened Al alloys was dependent on the magnitude of the force generated by corrosion product. Since the corrosion product force resulted in a net tensile stress at the crack tip during SCC, it is then a question whether the corrosion product force can generate the stress that is necessary to propagate an SCC crack. To answer this question, Kelly compared the maximum corrosion product force with lowest K ISCC of AA8090 at different aging times. The comparable results supported the proposition that exfoliation proceeded by a stress-assisted corrosion mechanism.[10] In conventional alloys, overaging increases the resistance to EFC. The increased susceptibility to IGC in over-aged AA8090 could be attributed to the enhanced galvanic action promoted by the formation of Al 2 CuMg and Al 2 CuLi. 8

29 More recently, the corrosion product force was measured in 7xxx alloys (AA7150- T651, AA7055-T7751, AA7449-T79) by McNaughtan et al.[24] The magnitude of maximum force generated in EXCO solution is in the order of: AA7150-T651 (25-50% recrystallized) > AA7150-T651 (80% recrystallized)> AA7449-T79> AA7055-T7751 indicating their decreasing susceptibility. The increase of corrosion product wedging in AA7150-T651 (with 25-50% recrystallized) is faster than in AA7150-T651 (with 80% recrystalized) indicating a relationship between the grain aspect ratio and wedging stress. It was also found that first drying out and then re-wetting in EXCO solution increased the corrosion product force. McNaughtan et al. explained that this is because the corrosion product was a hydrated alumina gel.[17] On drying out, the corrosion product changed to crystalline Al 2 O 3 3H 2 O with higher mechanical strength and this gave rise to a further increase in corrosion product force. To investigate the correlation between stress corrosion cracking and exfoliation corrosion, the stress corrosion resistance was determined by measuring K ISCC values using DCB specimens. An inverse linear relationship between the corrosion product forces and the K ISCC values for stress corrosion cracking indicates their common corrosion mechanism and low K ISCC corresponds to high susceptibility to SCC. This study also shows the strong dependence of corrosion product force on the IGC rate as well as the susceptibility of the grain boundaries to IGC, which is controlled by the composition and distribution of grain boundary precipitates. Considerable effort has been spent on proving that exfoliation propagates by an SCC mechanism. The fundamental question as to what electrochemical interactions are responsible for the generation of corrosion products has also been explored. Some 9

30 investigators have suggested that the cause of exfoliation in 7xxx aluminum alloys is an electrochemical potential developed between primary or secondary particles and the surrounding AlZnMg solid solution.[20, 25, 26] Mattsson et al. studied the mechanism of EFC in AlZn5Mg1 by investigating the phases occurring in the alloy in the naturally aged and artificially aged conditions.[20] In naturally aged condition, GP zones are the main phases. EFC is caused by galvanic corrosion cell consisting of cathodic α-al(fe,me)si particles along the extrusion or rolling direction and narrow zones of anodic Zn-, Mg rich matrix. This narrow zone of Zn and Mg rich matrix will be anodically attacked leading to EFC along boundary. The MgZn 2 phase is much less noble and subject to dissolution, and the matrix is low in Zn and Mg so that it is cathodically protected. Since MgZn 2 is evenly distributed in the artificially aged condition, the attack is general and the alloy is not prone to EFC. Polarization of these phases showed that artificially aged material is more noble than naturally aged materials, further confirming his model. In summary, the exfoliation process is a synergic combination of chemical and mechanical actions. Papers containing mechanistic studies on exfoliation corrosion show that electrochemical mechanisms underlie intergranular corrosion and the wedging stress generated from intergranular corrosion decides the exfoliation characteristics. The wedging action of corrosion product is essential to continued EFC propagation by a stress corrosion mechanism.[8, 9, 22, 23, 27-29] The extent of exfoliation is affected by many factors, especially the alloy composition, grain aspect ratio and heat treatment, which will be discussed in detail in the following. 10

31 2.3.2 Factors That Influence EFC Effects of Alloying Elements on EFC Mattsson et al. have stated that an increase in Fe content decreases resistance to exfoliation corrosion in Al-Zn-Mg alloys.[20] In studying the EFC mechanism of Al- 5Zn-Mg1, they proposed that the main factor governing the amount of the α-al(fe,me)si phase is the Fe content. Increasing Fe content thus markedly increases the amount of α- Al(Fe,Me)Si. This phase is active as a cathode so that increasing Fe will increase the susceptibility to EFC. Increasing Mn content will also increase the amount of α- Al(Fe,Me)Si and hence increase the susceptibility to EFC. They suggested that Si and Cr might also have a similar effect. An influence of chromium has been reported elsewhere, the susceptibility being higher with increased chromium content.[30] In AlMgSi (6000) series alloys, Mg, Fe and Cu at concentration levels about , and , respectively, develop high susceptibility to EFC, while alloys with low level of these elements did not show a tendency to EFC.[5] Robinson also mentioned that the difference in copper/magnesium ratio between AA2024 and AA2014 results in the formation of different grain boundary phases.[8] CuAl 2 is predominant in alloy AA2014 and CuMgAl 2 is predominant in AA2024. The preferential dissolution of CuMgAl 2 in AA2024 leads to fast rates of IGC and EFC. It was assumed that Mn, Fe and Zr diffuse slowly and can maintain stable concentration so that they play a primary role in creating exfoliation paths, while Zn and Mg play a secondary role because they diffuse readily at normal homogenizing temperatures.[20, 25, 26, 31, 32] Evans and Jeffrey found that only those alloys 11

32 containing both Fe and Mn in AA7004 develop a high sensitivity to EFC and the corrosion path is at grain boundaries where there is depleted in Mn.[31] They proposed that the Mn-rich and Mn-depleted regions form an electrochemical cell and these regions are elongated during fabrication, resulting in exfoliation path. This is in agreement with the report by Bassi et al. that Mn content was a primary factor in EFC.[25] They further stated that Mn could replace Fe in compounds such as α-alfesi so that the formation of α-alfesi will result in depletion of Mn in surrounding areas. Posada et al. have stated that, in AA2024 aluminum body skins on aging aircraft, structural differences might control the extent of exfoliation corrosion by control of boundary precipitation and elemental depletion.[33] They carried out an analysis of precipitation within the grains and grain boundaries and elemental depletion profiles along grain boundaries. Mg was somewhat uniformly distributed and no Mg containing precipitates were observed even though the previous studies showed that Mg and Zn depletion zones were involved in the attack of 7xxx alloys.[34] They also did not observe the systematic depletion or enrichment of Cu, Si, Mn, and Fe in grain boundary zones Effects of The Grain Shape on Exfoliation Corrosion Elongated grains are an important requirement for EFC.[21] It was reported that the severity of exfoliation corrosion depends on the grain aspect ratio of the material.[1, 8, 9, 24] The greater the grain aspect ratio, the more severe the exfoliation attacks. Robinson developed a mathematical model of the exfoliation process.[9] In his model, the effect of grain shape on exfoliation susceptibility was studied. Figure 2.3 shows a 12

33 schematic of intergranular corrosion through the grains of a wrought aluminum alloy. In this model, the aluminum skin was assumed to deform into a section of a hemispherical blister, the pressure within the blister is assumed to be uniform. The height of blisters was calculated from the depth of unattacked metal from which the corrosion product was formed. Robinson calculated the strain that was generated in an exfoliation blister by corrosion product through the equation of ε c =(L -L)/L; the meaning of L and L are shown in figure 2.4. An iterative process was used to obtain several values for blister strain that develops with a certain volume of metal converted to corrosion product. When a blister strain reached the fracture strain of 11%, the blister was said to fracture. Alloys with larger grain aspect ratios take longer time to reach the 11% fracture strain than alloys with smaller grain aspect ratios, and a larger diameter of blister develops for the former. More elongated grains convert more material to reach the surface fracture strain and result in a more severe surface attack on the surface. The depth of penetration was shown to be unrelated to the shape of the grains. The model investigated exfoliation mainly from the mechanical aspect and many assumptions were introduced to simplify the model. The real situation is more complicated and electrochemical reactions are an essential element that cannot be neglected throughout the EFC process. More recently, Robinson and Jackson performed exfoliation tests on Al-Cu-Mg alloys AA2014-T651 and AA2024-T351.[1, 8] Samples with different sections from each plate material were exfoliated using an acidified intermittent salt fog test.[35] The corrosion was assessed by rating the surface appearance with reference to a series of standard photographs given in ASTM G34[36] and by measuring the depths of EFC attack. The severity of exfoliation corrosion was shown to be related to the grain aspect ratio of the 13

34 material, with the most advanced attack occurring on the mid-section of the plate containing more grains with higher grain aspect ratio. The severity ranked in the order: T/2 > T/4 >surface. More elongated grains result in larger surface blisters before fracture. The depth of penetration was also in the same rank order, which contradicts their previous finding that the depth of penetration is unrelated to the grain shape. It is possible that AA2024 and AA2014 could have different grain boundary chemistry, which has some effect on their susceptibility to IGC. The surface ratings of different sections from the two materials are shown in figure 2.5. The influence of grain shape on the depth of exfoliation corrosion was measured by both a four-point bend method[4] and optical metallography.[37] The result is in agreement with the observations of the surface ratings. AA2024-T351 with higher grain aspect ratio exfoliated faster than AA2014-T651. Constant load stress tests were performed in EXCO solution for the same material to investigate the SCC behavior. It was shown that the cracks propagated most rapidly on the T/2 section where the grains were most elongated. The authors concluded that the grain structure has a similar effect in both exfoliation and SCC, providing further evidence for a link between the two forms of corrosion. The same results were also obtained by corrosion tests on lithium aluminum alloys AA8090.[37] The susceptibility of AA8090 to EFC was ranked as T/2 > T/4 > Surface.[37] EXCO and MASTMAASIS tests were conducted on the AA8090 plates with different degrees of recrystallization. It was shown that the surface rating of exfoliation ranked as fully partially unrecrystallized sheet > recrystallized sheet > recrystallized sheet, which means the susceptibility to EFC decreases in this sequence. The microstructure of the plate affects the corrosion product force. More elongated 14

35 grains introduced a longer and more tortuous corrosion path, slow development of the corrosion product force and higher maximum value of corrosion product force. In a rolled or extruded product, variations in exfoliation susceptibilities and attack rates can be often attributed to differences in thickness of the recrystallized layer usually present on the outside surface.[38] Exfoliation corrosion of aluminum alloys HE10, (Al-Cu-Cd) and HE15 fully aged materials were also extensively studied by Liddiard et al.[4] Extrusions were tested, with recrystallized surface layers and elongated grains below the surface. Atmospheric exposure tests for 8 years and salt spray tests in the lab were performed. The specimen was first corroded in a salt spray cabinet and then loaded in a deflection jig. The approximate average material loss, or the change in the effective cross section was obtained by measuring the deflection of the specimen under load. Three distinctive periods for the loss of metal were observed. The first stage, the incubation period, was explained by intergranular penetration of the recrystallized surface. The second period involved rapid metal loss and corresponded to the attack of the elongated grains beneath the recrystallized layer, with lifting of the metal. The third period showed an even faster attack, explained by the authors as a consequence of the highly directional longer grains found at the center of the extrusion. An elongated grain structure is an important prerequisite for exfoliation. Intergranular corrosion does occur in materials with equiaxed grains, but this tends to result in pitting rather than the formation of exfoliation. 15

36 Effect of Heat Treatment on Susceptibility to Exfoliation Corrosion The heat treatment condition of the alloy is thought to be important in exfoliation corrosion because heat treatment changes the microchemistry.[4, 39, 40] Winifred carried out a full investigation on the susceptibility of HE15 (Al-Cu- Mg-Mn-Si-Fe) to EFC considering the influence of copper content and aging treatment on susceptibility to exfoliation corrosion.[40] The laboratory tests clearly showed that extended aging prevented the exfoliation corrosion. The aging temperature also has an effect on the extent of EFC. The specimens fully aged at 200C showed more severe attack than those fully aged at 165C or 185C. Higher copper content in the material is desirable for beneficial effect of extended aging to resistance to EFC. The extended aging results in reduction in tensile stress as well as hardness. Metallographic studies on HE15 were conducted and provided an explanation that as the aging time increased, a more apparent subgrain structure could be observed. In over aged material, corrosion takes place entirely at the substructure boundaries and does not lead to EFC. The grain boundaries of an Al-4%Cu alloy with different aging duration were observed using TEM by Garner and Tromans.[41] The underaged material showed intact Al 2 Cu precipitates after immersion in test solutions, but the grain boundary area had been attacked. At slightly longer aging times, a precipitate free zone was seen, which appeared sensitive to attack. Peak aging would seem to give rise to the copper containing phases in the matrix as well as at the grain boundaries, so attack at this temper was seen to consume matrix material as well as that at the grain boundaries. The underaged material was most susceptible to intergraunular corrosion, because of the continuous anodic zone formed. 16

37 Liddiard et al. also studied aging effect on the exfoliation corrosion of HE10 and HE15.[4] Aging was observed to confer some resistance to exfoliation corrosion, because of increased subgrain boundary precipitation for longer aging times, allowing less attack at the grain boundaries and relieving the corrosive process there. The susceptibility can be reduced by increasing aging time at 190C from 1h to 24h, though overaging can result in a loss in mechanical properties. It was also proposed by Lifka that over-aging tends to reduce an alloy s susceptibility to EFC by virtue of larger precipitates giving a more broken anodic path, and reducing the potential difference between the grain boundaries and the matrix.[29] In Li-bearing alloys, the relationship between heat treatment and susceptibility to exfoliation is not as straightforward as in the case of AA7xxx series. Work done on Al- Li alloys indicates that, in alloys containing Cu and Mg, the susceptibility to exfoliation increases with over-aging. Sheppard et al. have found that the susceptibility to exfoliation increases with increased aging time. They attributed this to the increased precipitation of T1 (CuAl 2 Li) at the grain boundary.[42] The increased precipitation enhances the galvanic action between the matrix and the active T1 (CuAl 2 Li) phase.[10] This is supported by studies on bulk T1 phase indicating that it is active with respect to the matrix.[43, 44] Recent studies on AA8090 alloys by Kelly showed that corrosion product forces were ranked in the order: under aged > over aged > peak aged, which in turn determine the exfoliation susceptibility as shown in figure 2.6.[10] The authors explained that the high susceptibility of over aged temper in AA8090 was attributed to the formation of anodic 17

38 phases (Al-Li, S phase, and T1 phase Al 2 CuLi) that precipitated at grain boundaries during the over aging.[45] Several combinations of tempers were shown to increase the resistance to exfoliation corrosion. Lifka et al. conducted corrosion tests in a laboratory accelerated environment (acidified 5% NaCl intermittent spray at 49C) as well as seacoast and inland industrial atmospheres on high strength AA2xxx and AA7xxx aluminum alloys with different temper combinations.[29] They found several compositions and tempers that exhibit virtual immunity to EFC and SCC: AA2024-T851, AA2219-T851 and AA7075-T7351. AA2020-T7351, AA2024-T851, AA2219-T851, AA7001-T7551, AA7075-T7351 and AA7178-T7651 were free of exfoliation in laboratory tests. For T4 and T3 tempers, decreasing the plate thickness increased the degree of susceptibility. The 3.5% NaCl alternate immersion test on these materials showed that the AA7075-T651 and AA7079- T651 were susceptible to SCC. Similar results were obtained from seacoast atmosphere exposure tests at Point Judith, Rhode Island. AA7xxx series alloys in T76 or T73 tempers exhibited relative resistance to EFC. T6 7xxx series were susceptible to EFC. AA2024-T3 plates exhibited severe exfoliation, whereas less exfoliation corrosion was observed for the artificially aged T8 condition.[46] Environmental Aspects EFC is generally observed under specific environmental conditions such as high humidity and salt containing environments.[2, 4, 33, 47] Lifka and Sprowls found that the corrosivity of an atmosphere is controlled by factors relating to the moisture on a corroding surface and by atmospheric pollutants such as chlorides.[48] It was also shown 18

39 by Carter and Campbell that the weather conditions during the initial period of exposure, particularly as they affect the time of wetness, would have a critical influence on the pattern of corrosion that develops on aluminum alloys.[49] Exfoliation Corrosion Testing It was reported that exfoliation propagated at a nearly linear rate.[50] Robinson s study on exfoliation extent by measuring depth of penetration during MASTMAASIS test[35] also showed that rate of penetration of EFC was almost linear.[1] From the discussion above, we know that, by mechanical processing or heat treatment to eliminate the preferential corrosion path or change grain structures, we can improve the resistance to exfoliation corrosion of aluminum alloy. However, their effects on EFC rate have only been studied qualitatively. Previous test methods were limited to natural environments, but these are time consuming and accelerated tests are often required.[51] A good exfoliation test should show enough sensitivity to distinguish between susceptible and non-susceptible alloys and ideally show results during a short exposure period, which should correlate well with extended tests such as atmospheric exposure to marine environments. Efforts have been devoted to investigate the extent or the rate of exfoliation corrosion. The common exfoliation corrosion tests in laboratory that have been reported are: immersion test method, salt spray method, electrochemical corrosion test and mechanical tests. The first two methods usually involve in the interaction between alloys and acidic solution to determine the susceptibility of certain alloy. The resulting surface is often compared to the environmental exposure data. 19

40 Immersion Test The EXCO test according to ASTM G34 is widely used to evaluate the exfoliation corrosion behavior of 2xxx and 7xxx series aluminum alloys.[36] This accelerated test method can help people to predict the performance of the products after many years of service by using an aggressive corrosion solution containing 4.0M NaCl, 0.5M KNO 3 and 0.1 M HNO 3. The solution has an initial ph of 0.4. After constant immersion in the EXCO solution for a period of 48 to 96hrs, depending on the alloy system, the susceptibility to exfoliation is determined by visual assessment of the surface. The surface is compared to a series of standard photographs of sample surface ratings from N (No appreciable attack), P(pitting), to various degrees of exfoliation attack EA- ED, which represent conditions from superficial exfoliation to very severe exfoliation. During the test, the ph of the solution will increase from 0.4 to about 3. The reason for the increase in ph was attributed to the hydrolysis of aluminum ions.[51] At low ph, following reaction occurs: Al+3H + Al 3+ +3/2H 2 [51] After part of the acid is consumed, the aluminum hydrolyzes and has following equilibrium reactions: Al 3+ +H 2 O Al(OH) 2+ +H + Al 3+ +2H 2 O Al(OH) H + Al 3+ +3H 2 O Al(OH) 3 +3H + Until the ph for the deposition of Al(OH) 3 is reached.[51] Consequently, the attack will continue at a lower rate. EXCO test successfully distinguishes between the resistant T76 temper and the susceptible T6 temper for AA7075. Exfoliation 20

41 performances of AA7075 and AA7178 alloys in atmospheric exposure and laboratoryaccelerated tests were compared. The EXCO test was successful in predicting the resistance to exfoliation of AA7075 and AA7178 in seacoast and industrial atmospheric environments.[7] However, the EXCO test is generally considered to be unreliable for predicting the exfoliation corrosion behavior of Al-Li alloys due to the poor correlation with atmospheric exposure data.[52] One limitation for the EXCO test method is that it is not identical to the service environment. The EXCO solution is too corrosive for some materials in the T7 tempers and does not always accurately predict the resistance of AA2024 and AA7x50 products.[52] The EXCO test could not reveal intermediate exfoliation corrosion resistance for AA2xxx and AA7xxx series aluminum alloys. Lee and Lifka proposed a modified EXCO solution with ph adjusted to 3.2 by the addition of HCl and aluminum ions.[46] The solution contains 2.96g/L AlCl 3, 230g/L NaCl, and 60.7g/L KNO 3. Results of the modified EXCO solution were in good agreement with the performance of the alloys in actual service conditions.[46, 52] The modified EXCO test reproduced the atmospheric data better than standard EXCO test. A good correlation between the performance in natural environments and in modified EXCO exposure was reported for Al-Cu-Mg AA2024 and Al-Cu-Li alloy AA2090. However, the Al-Li alloys seemed to suffer more severe exfoliation than in the standard EXCO solution. Other immersion tests such as the ASSET test was used by Sprowls et al. to look at the exfoliation behavior of AA5xxx and AA7xxx aluminum alloys.[1, 38] A solution containing ammonium chloride and ammonium nitrate with additions of ammonium 21

42 tartrate and hydrogen peroxide was devised for the ASSET test. This test was found to give a good correlation with natural exfoliation Salt Spray Tests The SWAAT test (Sea Water Acetic Acid Test) was first developed by Romans.[53] The 5% synthetic seawater was used and was acidified to ph 2.8 by the addition of glacial acetic acid. The test solution is introduced onto the surface of aluminum specimens using a salt spray cabinet. The cabinets enable a cycle of wetting and drying to take place. These tests have been deemed successful in that the corrosion product does not form away from the site of corrosion attack as with some constant immersion tests. The corrosion product may therefore cause surface lifting of grains and blister formation. The SWAAT test can distinguish between alloys not susceptible to exfoliation and those having a degree of susceptibility under specialized conditions of exposure, and susceptible alloys under normal exposure. A period of one-week duration was thought to be sufficient to give optimum results for surface analysis. The standard wetting and drying cycle was 30 minutes for spraying the specimens, and 90 minutes allowed for drying. A prolonged drying period was seen to increase the surface ratings for exfoliation attack, whereas a longer wetting period had no effect on the exfoliation ratings. Cyclic acidified salt fog MASTMAASIS testing is another accelerated test method developed by Lifka and Spowls.[21] The solution for MASTMAASIS is 5 percent NaCl in distilled water buffered to ph 3 with acetic acid.[21] A commercial fog cabinet at 95F is used to permit automatic cycling. One-spray cycle consists of a 45min spray period, a 22

43 2 hour purge period and a 3 hour soak period at high relative humidity due to the solution maintained in the cabinet. Samples are suspended at an angle of 45 o. The result then is compared to standard photographs for determination of EFC extent.[35] The test period depends on the susceptibility of alloys. A good correlation was also found between results of outdoor exposure and MASTMAASIS for Al-Li alloys. The MASTMAASIS test produces better results for long-term outdoor exposure. In contrast to the simple, rapid EXCO test, however, cyclic acidified salt fog testing is complex, requiring specialized apparatus and relatively long testing periods. Comparison of accelerated and atmospheric exposure tests for corrosion of aluminum alloys was conducted by Dedamborenea et al.[51] Two Al-Cu alloys (AA2024-T4 and AA7075-T7351) and two Al-Li alloys (AA2091-T84 and AA8090- T8171) were exposed for 2 years in a moderately aggressive marine environment, and their corrosion behavior was compared with that in accelerated tests for intergranular corrosion and exfoliation. The results in the aggressive EXCO solutions did not agree well with those from outdoor exposure, even using the modified solution. The EXCO test resulted in a much greater extent of attack than outdoor exposure for AA8090-T8171 and AA2091-T84. None of the solutions reproduced the pitting or intergranular attack that developed during outdoor exposure.[51] Braun also conducted laboratory exposure tests on these alloys.[52] His research showed that the cyclic acidified salt fog test reproduced the outdoor corrosion data for the Al-Li alloys AA8090-T81 and AA2091-T84 and marine exposure results reported for the conventional alloys AA2024-T351 and AA7075-T7351. The standard EXCO test indicated better exfoliation corrosion behavior of the alloys investigated except for 23

44 AA8090-T6 and AA7075-T7351 plates. In the modified EXCO test, AA7075-T7351 panels were susceptible to pitting, whereas the other alloys studied generally suffered more severe EFC than in the standard EXCO test.[52] Immersion test and salt spray tests are useful for research and development purposes but should not be used as a method for quality acceptance or quantification. The discrepancy of the results between accelerated test and environmental exposure tests as well as the different results among several accelerated tests indicates the limitation of these tests. The appropriate synthetic environment for use in accelerated exfoliation corrosion tests is still an open question. Another drawback is the visual assessment of the corrosion attack, which is subjective. Furthermore, visual assessment cannot give sufficient evidence of the depth of penetration or kinetics. To get the full information of EFC, metallographic sectioning is often required Electrochemical Impedance Spectroscopy In an attempt to obtain improved predictive ability for EFC behavior, Keddam et al. examined the ability of EIS to distinguish exfoliation corrosion from other forms of corrosion.[54] AA2024-T351, AA2219-T87 and AA6013-T6 were divided into two groups in terms of their different impedance behavior. In AA2024-T351 and AA6013-T6, a marked feature corresponding to potential distribution inside pores or lamellae by the corrosion was observed, indicting the formation of exfoliation or intergranular corrosion. No such feature was observed in AA2219. They concluded that EIS could distinguish between formation of pits and deep lamellae due to exfoliation or intergranular 24

45 corrosion.[54] Some other works have also used EIS to characterize the exfoliation corrosion process of aluminum alloys.[55-57] Conde and Damborenea performed EIS measurement on AA8090 in EXCO solution and found that impedance technique made it possible to detect the different stages of localized attack of the alloy. One single time constant in EIS spectra related to a corrosion process under activation control with hardly any detachment of the surface grain. A depression in the capacitive arc and narrowing of the phase angle indicated that the material was being attacked. Two time constants was a consequence of the great amount of surface grains lost by the effect of the delamination.[56] Similar work was done on Al-Li-Cu-X and Al-Mg-Li-X alloys in EXCO solution. The earlier appearance of two-time constants for Al-Li-Mg-X indicates that this alloy has higher susceptibility to EFC than AA2091 alloy, where two-time constants became evident after 96hrs.[55] More recently, Cao et al. studied the exfoliation corrosion of aluminum alloy AA7075 by EIS method. They used the equivalent circuit shown in figure 2.7 to analyze the EIS data.[54, 56] Similar to Keddam s findings, they concluded that the decrease of low frequency resistance and increase of low frequency capacitance might originate from the appearance of pitting on the surface of the corroding surface.[57] A decrease in C lf indicates the occurrence of exfoliation corrosion. When EFC occurs, the low frequency inductive arc disappears and the Nyquist plot is composed of two capacitive arcs.[57] Deflection Technique A non-destructive method based on deflection has been used as a quantitative measurement of exfoliation corrosion.[4, 8] In this method, the deflection of specimens 25

46 under four point loading condition is related to the thickness of the specimen by equation δ/w=k/d 3 where δ is the deflection, W is the load, d is thickness, K can be determined initially from the sample dimensions.[4] By following the change in the deflection per unit load, the changes in the effective cross-section of the specimen can be determined. During the test, it is assumed that corrosion causes loss in one dimension only. Changes in this dimension with time can then be followed. In Liddiard s work, the specimens were corroded in the unstressed condition, and the thickness change with time due to the exfoliation corrosion can be determined by intermittent measurement. Robinson et al. also used four point bending method to measure the mean depth of attack of the AA2024 and AA2014 specimens after salt fog cabinet or coastal exposure tests.[1] However, this technique requires a complicated experimental setup. The assumption of uniform thinning of the specimen during corrosion was met after long exposure times when the entire specimen surface underwent exfoliation. It was less valid in the earlier stages of the test when some areas of the surface were free of attack or where the corrosion was in the form of localized pitting. In conclusion, exfoliation corrosion is a type of intergranular corrosion. The mechanism of exfoliation corrosion is strongly related with that of IGC and intergranular tress corrosion cracking. To understand exfoliation corrosion, it is essential to understand these two types of corrosion. In the following, intergranular corrosion and stress corrosion cracking as well as their relation with exfoliation corrosion will be reviewed. 26

47 2.4 INTERGRANULAR CORROSION Intergranular corrosion is a selective attack of grain boundaries or closely adjacent regions without appreciable attack of the grains themselves.[16] Most heat treatable high strength aluminum alloys are susceptible to IGC to some extent due to the formation of precipitates along grain boundaries during heat treatment. It has been reported that there is a correlation between the susceptibility to IGC and exfoliation in some Al alloys.[58, 59] In this section, the mechanisms of IGC, the relationship between IGC and exfoliation and the causes of susceptibility to IGC in high strength aluminum alloys will be reviewed IGC Mechanisms The mechanisms of IGC are electrochemical.[38] Generally speaking, there are three major accepted explanations of IGC in aluminum alloys. They are: 1. Galvanic couple theory. 2. Precipitate free zone breakdown model. 3. Anodic dissolution of grain boundary precipitates Galvanic Couple Theory Dix et al. first proposed the galvanic couple theory.[60] According to this theory, IGC susceptibility is attributed to the local galvanic corrosion between the noble grain matrix and the anodic grain boundary regions. In Al-Mg and Al-Mg-Zn alloys, IGC is considered to be a result of galvanic corrosion between anodic grain boundary particles (Al 8 Mg 2, MgZn 2 ) and the noble grain matrix.[38] In Al-Cu, Al-Cu-Mg and Al-Zn-Mg-Cu 27

48 alloys, IGC is explained as a result of galvanic corrosion between copper-depleted (precipitate free) zones and the grain matrix.[38] The galvanic couple theory was verified and developed by the work of Mears and Brown.[38] They developed a technique to measure the corrosion potentials of the grain boundary regions and the grain centers of large grained Al-4%Cu alloy in NaCl-H 2 O 2. The grain boundary regions were anodic relative to the grain bodies and the potential difference reached a maximum after aging for about 4 to 8 hrs at 191 C. It was found that the Al-4%Cu was most susceptible to intergranular attack when the corrosion potential difference was a maximum. With extended aging, the precipitation within the grain center began to approach that at the grain boundaries until complete precipitation occurred and the difference in potential decreased to almost zero. The anodic grain boundary regions were copper depleted zones. TEM studies by Hunter et al. successfully related the path of IGC attack to microstructural features of Al-Cu-Mg alloy 2024.[61] Brown et al. tried to explain IGC in Al-Cu based alloys on galvanic coupling between the grain boundary constituents and matrix. They stated that the susceptibility of peak-aged temper to IGC was due to the difference in corrosion potentials of the grain boundary and the grain interior. Because of Cu depletion, the grain boundary region was 200 mv more active than the grain interior.[62] Studies on Al-Zn-Mg alloys were also performed. The potential difference between the grain boundary region, which includes precipitate free zone and the grain boundary precipitates, and the matrix was only 5 mv. Though the potential difference is small, the material failed along the grain boundaries with the application of stress.[60, 63] Dix et al. explained this behavior based on the fact that the η(mgzn 2 ) phase is active to 28

49 the matrix and the preferential dissolution of the η phase leads to susceptibility to intergranular and IGSCC.[60, 63] Many experiments have shown that IGC and SCC are not seen in solutions that do not contain Cl - ions. The galvanic couple theory fails to explain the role of chloride ions and other halide ions on intergranular corrosion of Al alloys Precipitate Free Zone Breakdown Model. The galvanic couple mechanism does not explain why halide ions are necessary for intergranular corrosion. Galvele and DeMicheli proposed a second theory, the precipitate Free Zone (PFZ) Breakdown Model, which explain the effect of chloride ion on IGC behavior in Al-Cu alloys.[8] They proposed that the difference in breakdown potential between the grain matrix and the solute depleted zones adjacent to the grain boundaries produces the IGC susceptibility of aged Al alloys. In studying the mechanism of IGC of Al-4%Cu aged alloys, solutionized Al- 4%Cu, Al 2 Cu and Al-0.2%Cu were used to simulate the behavior of the grain bodies, intermetallic Al 2 Cu phase and copper depleted zone along the grain boundaries in aged Al-4%Cu, respectively.[64] Anodic polarization was performed in deareated NaCl solution with various concentrations to determine the breakdown potentials. The critical potential of Al-4%Cu solid solution, Al 2 Cu intermetallic phase, Al-0.2%Cu and pure Al is related to activity of NaCl. The breakdown potentials of the Al-4%Cu solid solution alloy were similar to that of the intermetallic Al 2 Cu. There were no differences in the breakdown potentials of the Al-0.2%Cu and high purity aluminum. The breakdown 29

50 potential for the Cu rich phases were 100 mv higher than the values for lower Cu content specimens. Galvele and DeMicheli found a clear relationship between pitting and intergranular corrosion. Al-4%Cu in its peak-aged condition exhibits two breakdown potentials when anodically polarized in Cl - containing solution.[64-66] They explained that the two breakdown potentials in this alloys are due to the copper depleted PFZ present along the grain boundaries of the peak-aged Al-4%Cu. This Cu depleted zone approximate to Al-0.2%Cu has a lower breakdown potential than the grain matrix and grain boundary precipitate (Al 2 Cu). When the potential is higher than the PFZ breakdown potential but lower than the matrix breakdown potential, localized attack on the PFZ results in IGC; when the applied potential is above the more noble breakdown potential, it showed both IGC and pitting in the matrix. Consistent with this explanation, the experiment showed that the over aged Al-4%Cu had only one breakdown potential because the grain body is mainly a two-phase structure and the increase in anodic area will favor generalized corrosion rather than IGC. They found that the IGC and pitting of Al only occurred in the presence of certain anions such as Cl -, Br -, I -, ClO - 4. Intergranular attack in AA7xxx (Al-Zn-Mg-Cu) series was studied by Maitra et al.[15, 67] They found that the alloys showed different localized corrosion behavior dependent on the artificial aging. The anodic polarization curve in deareated NaCl solution for W temper showed only one breakdown potential at 800 mv SCE, which corresponded to pitting corrosion in the single-phase solid solution matrix. The alloys were susceptible to intergranular corrosion and exhibited two breakdown potentials in T6 temper. The more noble breakdown potential 725mV SCE corresponded to the pitting 30

51 of solid solution matrix (containing less Zn and Mg) and the breakdown potential 800mV SCE was assumed to relate to the pitting of the solute enriched grain boundary region. IGC was observed between two breakdown potentials in T6 temper. T7 temper exhibited decreased susceptibility to IGC and the polarization curve consisted of one breakdown potential at 0.765mV SCE. Pitting was observed above this potential. They mentioned that if the microstructure contains more phases with distinct breakdown potentials, the polarization curves would consist of a number of plateaus. The phase will dissolve if the potential is above its breakdown potential. The authors explained that the difference is due to the segregation of different elements to the grain boundary during tempering. During T6 tempering, Zn and Mg solute atoms diffuse to the grain boundary, which results in lowering the breakdown potential of the grain boundary.[15] The reason for only a single breakdown potential in the T7 temper is attributed to the fact that Cu is removed from solid solution and incorporated in eta phase (Mg (Al, Cu, Zn)) resulting in the shift of second breakdown potential to negative value while the solute enriched grain boundary region becomes narrower and less continuous. They also stated that the increased susceptibility of AA7075-T7351 to IGC is due to increased differences between pitting potentials of different regions of the microstructure. The addition of nitrates to the aqueous solution can increase this difference and hence increases the susceptibility.[67] Meng et al. studied the corrosion behavior of AA7xxx series aluminum alloys.[68] Two breakdown potentials were found in Cu-containing 7xxx alloys in deareated chloride solution. They explained that the first breakdown potential corresponded to transient dissolution associated with attack of the fine hardening 31

52 particles and the surrounding solid solution in a thin surface layer. The copper content in these hardening particles controls the first breakdown potential. The second breakdown potential was associated with combined intergranular and selective grain attack. Guillaumin and Mankowski studied the corrosion behavior of AA2024-T351 in chloride media and proposed that the first breakdown potential was related to the dissolution of Al 2 CuMg particles or dealloying of Mg from the particles and the second breakdown potential corresponded to the matrix breakdown potential. Both pitting and intergranular corrosion developed above second breakdown potential.[69] In their study on AA6065-T6 in 1 M NaCl solution, they found that intergranular corrosion and pitting are dependent on each other. Pits first developed and then led to intergranular corrosion. They proposed that the mechanism of AA6056-T6 alloys consists of preferential dissolution of the anodic Cu and Si depleted zone along grain boundaries.[70] Anodic Dissolution of Grain Boundary Precipitates The anodic dissolution of grain boundary precipitates was proposed as an explanation for intergranular stress corrosion and IGC behavior in Al-Li-Cu and Al-Mg alloys. This theory suggests that the existence of a preferential anodic path along grain boundaries results in anodic dissolution of grain boundary precipitates. In the wrought AA5xxx(Al- Mg) series alloys, it is generally accepted that Mg 5 Al 8 particles can form a continuous intergranular path, which is active to metal matrix [58] and is corroded preferentially. Work on AA5083 alloys showed that IGSCC of the alloy was related to the anodic dissolution of the continuous grain boundary Mg 2 Al 3 when Mg content was greater than 32

53 3wt%.[59] In copper free AA7xxx (Al-Zn-Mg) series alloys, the anodic path is generally considered to be the anodic zinc and magnesium bearing constituents at the grain boundary.[71] Selective T1 phase dissolution for IGC in Al-Li-Cu alloy was first proposed by Rinker when he studied the corrosion and stress corrosion behavior of AA2020.[43] T1 phase (CuAl 2 Li) is believed to be the most active phase present in the subgrain boundaries and matrix.[43] Buchheit et al. used anodic dissolution theory to explain the IGC of Al-Li-Cu alloys.[43, 44, 72] They conducted electrochemical measurements on T1 phase and α-al-matrix to determine the mechanism of IGC. The Cu-depleted zone in AA2090 alloy was modeled by pure Al. The anodic polarizations of Al 2 CuLi and pure Al in deareated 0.6 M NaCl solution showed a larger dissolution rate at E corr of Al 2 CuLi phase (10-4 A/cm 2 ) than that of Cu depleted zone and α-al-matrix, which were less than 10-6 A/cm 2. Anodic polarization of T1 phase and SHT AA2090 in deareated 3.5wt% NaCl solution showed that the T1 phase was more active than alloy matrix. They concluded that the intergranular corrosion and pitting were controlled by selective dissolution of T1 phase, which resulted in the creation of an aggressive occluded environment leading to the continued grain boundary attack. In summary, there are essentially three different theories proposed to explain IGC susceptibility of Al alloys from electrochemical point of view. It is generally accepted that susceptibility to IGC results from the precipitation of intermetallic phases in the vicinity of the grain boundaries. The electrochemical roles of the grain boundary precipitates and the adjacent PFZ zones vary from one alloy system to another and will be reviewed in the following. 33

54 2.4.2 Alloy Microstructure It is known that the susceptibility to IGC, EFC, and IGSCC of aluminum alloys is strongly affected by alloy microstructure such as the dispersion of various intermetallic particles and precipitate free zones. A clear understanding of the electrochemical behavior of these microstructural factors will help us to understand EFC Intermetallic Particles The addition of alloying elements to Al leads to the dispersion of various intermetallic particles in grain bodies and grain boundaries. The formation of these particles is due to the low solubility of alloying elements in aluminum. They are two groups of the intermetallic particles. First group consists of the particles formed from the melt during solidification and they usually contain impurities. The second group consists of particles formed during heat treatment in heat treatable alloys. Heat treatable alloys are strengthened by precipitation hardening. Heat treatment develops large differences in the behavior of the grain boundary precipitate and precipitates free zones[73], hence affecting the susceptibility to localized corrosion. Many papers show that the anodic dissolution of intermetallic particles at the grain boundaries can lead to IGC and SCC.[43, 44] Al alloys that do not form intermetallic particles at grain boundaries, or those in which the intermetallic particles have potential similar to the grain matrix, have high resistance to localized corrosion. The electrochemical behavior of intermetallic particles in aluminum alloy plays an important role on localized corrosion process.[74] The review below will discuss the major intermatillic particles that exist in Al-Cu and Al-Zn alloys. 34

55 Intermetallic Particles in Al-Cu & Al-Cu-Mg Alloys Al 2 Cu (θ phase) and Al 2 CuMg (S phase) are the most important intermetallic particles in Al-Cu and Al-Cu-Mg alloys. The electrochemical behavior of these phases has been studied in detail.[66, 74-78] It is well known that Al 2 Cu phase has a more noble corrosion potential than the Al matrix.[64, 66] A study on the effect of solution aeration on the OCP of Al 2 Cu showed that OCP ranges from 0.59 to 0.7 V for aerated and deareated solutions with chloride concentrations ranging from 0.2 to 1M.[65] The susceptibility to IGC of alloys is influenced primarily by copper concentration gradients in the Al-Cu solid solution in the grain boundary regions.[38] Brown et al. claimed that the retention of 5 percent copper in solid solution changes the electrode potential of aluminum about 200 mv in the cathodic direction.[38, 62] Mazurkiewicz et al. studied the electrochemical properties of Al 2 Cu using potentiodynamic polarization on stationary electrodes and rotating ring disk-electrodes to study the mechanism of dissolution of Al 2 Cu.[79] The dissolution of Al 2 Cu leads to the plating of copper. They stated that Al 2 Cu dissolution followed the following reactions: Al 2 Cu Cu o +2Al 3+ +6e - Al 2 Cu 2Al +3 +Cu 2+ +8e - and Cu 2+ +2e Cu o Potential dynamic polarization of Al 2 Cu in sulphate solutions showed that when the anodic potential of copper dissolution was reached, the maximum current was reached.[79] Above the current maximum, the anodic process is controlled by aluminum hydroxide film in low and neutral ph solutions because copper dissolves at a relatively high rate. In alkaline solution, it is controlled by aluminum hydroxide films and the 35

56 stability of copper before the potential of formation of CuO 2-2 is reached. The anodic polarization of Al 2 Cu and Al in 1 M NaCl solution showed that the breakdown potential of Al was 0.2V lower than that of Al 2 Cu. Guillaumin et al. concluded that the large potential difference between Al and Al 2 Cu leads to the galvanic dissolution of aluminum anodes at OCP. Al 2 CuMg (S phase) is a key contributor to the localized corrosion in Al-Cu-Mg alloys. Electrochemical measurements on bulk S phase showed that it has an open circuit potential ranging from 925 to 935 mv SCE in 0.5 M NaCl solution independent of solution aeration.[75] S phase is more active than the Cu depleted zone and α-al matrix phase (Al-4Cu solid solution). It undergoes strong anodic polarization, which stimulates dealloying of the particles and nonfaradaic liberation of Cu clusters, consequently changing the galvanic relationship with the surrounding microstructure.[76] Buchheit et al. investigated the dealloying of S phase.[76] They stated that Al 2 CuMg (S) particles are initially electrochemically active with respect to the AA2024-T3 matrix. S phase is dealloyed in aggressive solutions leaving Cu rich remnants that behave like noble sites. X-ray microanalysis of corroded surface and X-ray line scans across an S phase exposed to a ph 4.2 chloride solution provides the proof that Al and Mg are selectively dissolved from S phase and leave Cu rich remnants behind. Two types of pit morphologies were found in the matrix as a result of S phase dealloying. The first type of pits was observed at the periphery of the remnants. The second type was observed far away from the primary pits. For the second type, Buchheit proposes that the particle remnant from selective dissolution decomposes into 10 to 100 nm Cu clusters, which detach from the 36

57 alloy surface and redistributed across the surface. These clusters act as local cathodes and induce peripheral matrix pitting. Further investigations using a rotating ring disk collection experiment and stripping voltammetry provide the evidence for Cu ion formation by dissolution and dealloying the Al 2 CuMg intermetallic compound.[80] The cyclic voltammogram for the collection experiment conducted in aerated solution with Al 2 CuMg shows three peaks that correspond to the oxidative formation of CuCl(s) from Cu, formation of Cu + at conditions far from equilibrium and the formation of Cu 2+ due to oxidation of CuCl (s). Metallic Cu was detected by X-ray microchemical analysis of the ring surface. Buchheit et al. concluded that the Cu ion generation mechanism contributed to the poor corrosion resistance and poor conversion coating characteristics of Al 2 CuMg bearing alloys. Figure 2.8 shows a schematic illustration of the mechanism for redistribution of Cu by dissolution of Al 2 CuMg and Al 2 Cu intemallic particles. The particles are anodically polarized by the surrounding Al matrix phase at an OCP of 0.93~-0.915V, which is negative of E 2+. The Cu cluster remnants are captured by the corrosion product gel, Cu / Cu detach from the particle remnant and move away from the corrosion site due to solution movement. The OCP of Cu in near neutral chloride solutions is positive to E 2+ so that the oxidation of the cluster is possible to generate Cu ions. Figure 2.9 shows potentiodynamic polarization curves for Al 2 CuMg in air sparged and N 2 sparged 0.5 M NaCl solution.[75] The positive shift in OCP is due to the increasing oxygen dissolved in the solution, which intensifies the Cu enrichment process causing a greater shift in OCP.[75] On the return scan, the corrosion potential was detected to return to 0.92 to 0.94V indicating that Cu enrichment is eliminated by 37 Cu / Cu

58 anodic polarization. The returning value of OCP also supports the previous study on nonfaradaic liberation of Cu clusters from dealloyed Al 2 CuMg in which the detached Cu cluster did not directly contribute to measurable electroactivity.[76] Other studies on AA2024-T3 showed that another type of intermetallic particle in Al-Cu-X alloys is AlCu(Fe,Mn).[81, 82] Al-Cu(Fe, Mn) particles have a potential of 675mV SCE, which is 200mV more noble than the matrix of Al-Cu-Mg based alloys having an OCP of 0.65V SCE in 0.1 M NaCl Intermetallic particles in Al-Zn-Mg & Al-Zn-Mg-Cu Alloys The formation of Zn and Mg bearing intermetallic particles can create differing electrochemical potentials in AA7xxx aluminum alloys.[32] It has been reported that eta phase in 7xxx series alloys is active to the Al alloy matrix as it has an OCP of 1100mV SCE in 1 M NaCl.[20] The electrochemical behavior of Mg (Zn, Cu, Al) 2 in AA7150 was studied by Ramgopal et al. by using thin film composition analogs prepared by flash evaporation.[83] Since eta phase has considerable solubility for Cu and Al, the generic composition can be considered to be Mg (Zn, Cu, Al) 2.[84] The electrochemical behavior of eta phase as a function of Cu and Al was studied in deareated neutral and high ph 0.5 M NaCl. It was shown that MgZn 2 has OCP of 1400mV SCE and a breakdown potential of 1140mV SCE in deareated 0.5 M NaCl. The addition of 17 and 27 atom % Cu results in an increasing corrosion potential by 250 and 300 mv and breakdown potential by 150 and 210 mv, respectively, while the breakdown potential of eta phase is relatively unaffected by Al addition up to 10 atom%. OCP playback experiments 38

59 provided the relevance to localized corrosion of eta phase. The low Cu containing intermetallic compounds completely dissolved in chloride solution while high Cu containing intermetallic compounds dealloyed, resulting in the enrichment of Cu, which can have a different effect on the corrosion behavior of AA7xxx. Unlike the uniform attack in neutral solution, the eta phase thin film analog was susceptible to more classic localized breakdown in high ph due to the enhanced stability of Mg (OH) 2 in high ph. Ramgopal investigated the mechanism of IGC in AA7150 and explained the higher resistance to IGC of T7 temper than T6 temper. He used analytical transmission electron microscopy to study grain boundary phases of AA7150 and found that the particles in T7 temper had much higher Cu concentration and lower Zn concentration than those in T6 temper.[84] The composition of the matrix and PFZ for two tempers is similar and has similar electrochemical behavior in the same solution. Electrochemical tests on thin film compositional analogs of T6 and T7 grain boundary precipitates indicated that the breakdown potential for both tempers is below the breakdown potential of the alloy, indicating that grain boundary precipitate does not initiate IGC directly. However, the dissolution of precipitates of two tempers can create different microchemistries. Dissolution of T7 grain boundary precipitates Mg (Zn, Cu, Al) results in higher content of Cu ions in solution, hence preventing the dissolution of PFZ and resulting in a high resistance to IGC Other Intermetallic Particles Intermetallic particles in aluminum alloys commonly contain iron because iron always exists in the alloys as an impurity. The influence of the Al 3 Fe phase and various 39

60 AlFeSi phases on the pitting[85-89] and exfoliation[5, 20, 90] of aluminum alloys has been extensively studied. It is generally known that the presence of iron in such forms is deleterious to the corrosion resistance of aluminum alloys. Nisancioglu investigated the electrochemical properties of the phases Al 3 Fe, α-al(fe, Mn)Si and δalfesi in NaOH solutions.[91] An alkaline environment was chosen as the test medium because diffusion layers of high ph are known to form around intermetallics acting as local cathodes on a corroding alloy in aqueous solutions. Figure 2.10 shows the variation of corrosion potential of Al 3 Fe and α-al(fe, Mn)Si with time in deaerated 0.1N NaOH solution.[91] The rapid shift in potential for Al 3 Fe was also observed by Golubev and Ronzhin.[92] According to their interpretation, aluminum undergoes selective dissolution at the more negative potential, and the surface of Al 3 Fe becomes richer in Fe component, which shifts corrosion potential from the initial value of 1270 mv SCE to 1180 mv SCE. α- Al(Fe, Mn)Si follows the same process at a much slower rate. At an applied potential of 1.2V SCE, the current density-time behavior of Al 3 Fe was unstable due to the detachment of Fe rich flakes after dealloying. Similar treatment on α-al(fe, Mn)Si particles also resulted an enrichment of surface with Mn and Si in addition to Fe. Cyclic polarization of Al 3 Fe shows several peaks, figure The first peak at around -1.15V ~ -1.2V was correlated with the selective dissolution of aluminum followed by inhibition of aluminum dissolution by the formation of a passive FeOH monolayer.[88, 92, 93] Other peaks at more noble potentials were related to oxidation of iron into successively higher valence states.[91] The polarization curves of α-al(fe, Mn)Si and δalfesi show two superimposed anodic peaks, which are also controlled by aluminum dissolution and the formation of passive iron oxide film. The addition of Si to alloys reduced 40

61 significantly the selective dissolution rate of aluminum in Fe containing particles and caused a negative shift in the corrosion potential. Mn has same effect as Si. Though the dealloying of Fe-rich compounds catalyzes the cathodic reaction, which is detrimental to the corrosion behavior of aluminum alloys, Mn or Si addition can suppress this reaction. The Mg 2 Si phase is the basis for precipitation hardening in Al-Si-Mg series alloys. This phase is anodic to aluminum and very reactive in acidic solutions with open circuit potential of 1590 mv SCE.[65] In addition to Mg 2 Si, Al 3 Fe, α-al-fe-mn-si and MnAl 6 are common particles present in Al-Si-Mn-Fe aluminum alloys. These particles are slightly cathodic to aluminum and tend to increase the frequency of pitting.[2] MnAl 6 has a corrosion potential of 760 mv SCE in a NaCl-H 2 O 2 environment, which is close to that of the aluminum matrix. Wrought materials, especially those containing more than 3wt% Mg, have Al 8 Mg 5 and Al 2 Mg 3 distributed on grain boundaries and are susceptible to corrosion. Al 3 Mg 2 has a corrosion potential of 1.15V SCE and Al 8 Mg 5 has a potential of 1.05mV SCE in 1M NaCl-H 2 O 2.[94] At high Zr contents, primary crystals of the tetragonal phase Al 3 Zr are easily formed. This phase has been found to be cathodic to pure aluminum and supports cathodic reactions.[77] Precipitate Free Zones (PFZ) Alloys that are age hardenable tend to produce PFZ parallel to grain boundaries. The formation of PFZ is generally thought to be related to two mechanisms.[20] One is due to the vacancy depletion in these areas giving too low diffusion rates for the 41

62 formation of precipitate nuclei.[95] The other mechanism is that the PFZ is formed through solute depletion caused by the preferential precipitation in the grain boundaries.[73, 96] Which mechanism is important depends on the heat treatment of the material. The width of the PFZ depends on many factors such as alloy composition, solution treatment temperature, quenching rate, aging temperature and period of natural aging before artificial aging. Many investigators have studied the effect of cooling rate during quenching on the resistance to intergranular corrosion and to stress corrosion cracking.[97-99] Higher solution treatment temperature, faster quenching rate, and lower aging temperature produce narrower PFZs.[38] The existence of PFZs has been reported to cause high susceptibility to IGC in many Al alloy systems, such as Al-4%Cu[64], AA2024 (Al-Cu-Mg)[69], AA2090 (Al- Li-Cu)[100], AA6056-T6 (Al-Mg-Si-Cu)[70], AA6013-T6 (Al-Mg-Si-Cu)[70], and Al- Zn-Mg-Cu (7xxx)[15]. The anodic PFZs along grain boundaries in these alloys contribute to their high susceptibility to intergranular corrosion. 2.5 INTERGRANULAR STRESS CORROSION CRACKING (IGSCC) According to Jones, Stress corrosion cracking is the brittle failure at relatively low constant tensile stress of an alloy exposed to a corrosive environment.[16] Stress corrosion cracking can be intergranular or transgranular. In contrast to the transgranular cracking which is associated with mechanical fractures, IGSCC usually propagates along electrochemically active ground boundary path. The IGSCC mechanism has been proposed as one of the possible mechanisms of exfoliation corrosion.[1, 8, 9, 23] Hence, a fundamental knowledge of the mechanism of IGSCC in high strength aluminum alloy 42

63 will help us to better understand the phenomenon of exfoliation corrosion. The literature review here will be confined to the two main proposed mechanisms of SCC. The effect of materials microstructure and microchemistry on stress corrosion cracking will also be reviewed Mechanisms of SCC Three conditions must be present simultaneously to produce SCC:[16] 1. Critical environment. 2. A susceptible alloy. 3. Some component of tensile stress. The sequence of events involved in the SCC process is divided into three stages:[12] 1. Crack initiation and propagation. 2. Steady-state crack propagation. 3. Final failure stage. These three stages are shown in figure 2.12, a schematic diagram of typical crack propagation rate as a function of crack tip stress intensity behavior. The magnitude of the stress distribution at the crack tip is quantified by the stress intensity factor, K, for the specific crack and loading geometry. No crack propagation is observed below K ISCC, which is determined by the alloy, its environment and metallurgical condition. The rate of crack propagation exceeds the plateau velocity as the stress intensity level approaches K Ic.[12] Though SCC has been studied for many years, the initiation of cracks is not well understood. Many papers provide the evidence that cracks can initiate at surface 43

64 discontinuities, at pits, or from intergranular corrosion sites.[12] The mechanism of IGSCC propagation has also been argued for many years, yet there is no one mechanism universally accepted. General speaking, the mechanisms proposed for SCC propagation can be classified into two basic categories: 1. Anodic mechanism, which involve stress assisted anodic dissolution of an active path. 2. Cathodic mechanism, or hydrogen embrittlement Anodic Dissolution Mechanism The anodic mechanism can be described as the active dissolution and removal of material from the crack tip. Two main anodic dissolution models have been proposed. The first is active path intergranular SCC model. In this model, IGSCC occurs by preferential corrosion of electrochemical active path. The second is film rupture model.[12] This model assumes that the stress acts to open the crack and rupture the protective surface film. Some reports have evidence to support this model.[ ] However, this theory is not generally accepted as a mechanism of transgranular SCC.[12] Most researchers proposed that the mechanism of SCC for AA2xxx alloys (both Al-Cu and Al-Cu-Li alloys) is the anodic dissolution along the grain boundaries.[107] Izu, Shiokawa, and Sato consider local anodic dissolution as the controlling factor in the SCC of AA2017 alloys. They showed that the time to failure of stressed alloy was dependent on the applied current density, not on the applied stress.[108] Gruhl stated that Al-Cu-Mg alloys did not truly stress corrosion crack, but rather exhibit intergranular corrosion, 44

65 which could occur without stress but was accelerated under stress.[109] In the study of AA2020 alloys under varying aging treatments, Rinker et al. proposed that the variation in SCC was due to preferential corrosion of the T 1 phases.[110] Meletis studied the SCC behavior of Al-2.9Cu-2.2Li-0.12Zr in the peak-aged and over-aged conditions using DCB specimens and TEM analysis. He found that peak-aged temper was the most resistant to SCC and over-aged temper is more susceptible to SCC. Meletis explained that preferential T 1 precipitation in over-aged condition produced localized strain at the grain boundary, which promoted either dissolution or hydrogen embrittlement along the grain boundaries.[111] Some researchers also considered anodic dissolution mechanism as the mechanism of SCC in the AA7xxx alloys.[15, ] Joshi et al. found that there were greater amounts of Cu, Mg and Zn along grain boundaries in T6 temper than T73 temper, which can explain their different corrosion behavior.[116] Sedriks et al. studied the electrochemical behaviors of the intermetallic phase MgZn 2 and Al-Zn-Mg alloys in NaCl and AlCl 3 solutions.[112] They proposed that a crack was formed by the lateral coalescence of pits due to the dissolution of MgZn 2 precipitates and propagated by ductile fracture and active dissolution of MgZn 2 particles. The process is repeated when the crack tip reaches the next MgZn 2 particle. However, Poulose et al. have a different point of view regarding the role of MgZn 2. They found that the SCC crack velocity was inversely proportional to the volume of MgZn 2. Hence they proposed that the grain boundary precipitates acted as sacrificial anodes to retard IGSCC.[117] 45

66 Cathodic Mechanism Since hydrogen evolution is always observed during localized corrosion of Al alloys, hydrogen embrittlement is accepted by the majority researchers as the cause of intergranular stress corrosion cracking in Al-Zn-Mg-Cu system.[107] Considerable research work indicated that the presence of cathodically charged internal hydrogen could degrade the ductility of Al-Zn-Mg alloys.[ ] Some researchers suggested that the existence of hydrogen causes crack blunting[122], reduction in cohesive strength[123], and microvoid nucleation[124]. Works have been done to provide the evidence of hydrogen induced cracking in 7xxx alloys. Hardie et al. found that embrittlement occurred in tensile tested AA7179-T651 by presoaking the materials in 70 o tap water for enough time and recovered by drying out in lab air or vacuum.[125] Scamans observed crack arrest markings on the fracture faces of failed DCB and C-ring samples of AA7071 and AA7078.[126] Chu et al. have studied the dynamic processes of the nucleation and propagation of hydrogen induced delayed cracking (HIDC) and SCC of AA7075. They observed the similar fracture features between SCC and HIDC and concluded that the mechanism of SCC of high strength aluminum alloys in aqueous solution was hydrogen induced delayed cracking.[127] Tong, Lin and Hsiao found that the atomic binding energy was reduced by charging hydrogen into pure aluminum[123] and water vapor accelerated the fatigue crack propagation rate in pure aluminum due to the rapid diffusion of hydrogen through dislocations into the interior of the crystal.[128] The path of hydrogen diffusion into interiors of alloys has been investigated. Albrecht et al. studied the SCC of AA7075 alloys and concluded that dislocations can 46

67 transport hydrogen deep into the alloy and cause brittle intergranular fracture.[129] Grugl suggested that hydrogen transport was by the grain boundaries.[109] He proposed that increasing concentration of Zn in solid solution would decrease the solubility of hydrogen in the grain boundaries and hence increased the time to failure. On the other hand, tensile stress normal to grain boundaries opened the lattice and helped the diffusion of hydrogen into grain boundaries. Talianker and Cina used TEM to study the dislocation density of the grains and grain boundaries in AA7075, AA7050 and AA7278 in T6 temper and in retrogression and re-aged treatment. They observed high density of dislocations in the T6 temper, which disappeared after RRA treatment. Their study supports both of the above theories that grain boundary dislocations transport hydrogen.[130] Thompson has found out that mode I loading on AA7075 led to the much higher susceptibility to SCC due to a hydrogen embrittlement mechanism.[131] Other studies showe that alloys containing large fraction of coherent particles and few active slip planes have high susceptibility to SCC.[132] Since the copper rich alloys have lower volume fraction of coherent particles, these alloys have less susceptibility to SCC Effects of Microstructure and Alloy Chemistry on IGSCC SCC susceptibility strongly depends on alloy microstructure. Chemical composition and metallurgical treatment can change the microstructure and markedly influence the resistance of an alloy to SCC. Kent investigated the effect of quench rate on the microstructure and SCC resistance of Al-Zn-Mg alloys.[133] Slow quenching rate resulted in an increase in the size and spacing of the precipitates at high angle grain boundaries, an increase in the 47

68 width of PFZ and consequently a longer time to fail. He claimed that an increase in the PFZ width and an increase in the size and spacing of grain boundary precipitate lead to excellent SCC resistance. However, a later investigation conducted by De Ardo indicated that there was no correlation between PFZ width and the susceptibility, though the size and spacing of grain boundary precipitate was very important in determining the SCC susceptibility.[134] Rajan et al. have suggested that coarse grain boundary precipitates act as hydrogen traps and can create hydrogen bubbles.[122] Speidel reviewed the literature on SCC of aluminum alloys and concluded that the over aged temper T76 and T73 provided high strength aluminum alloys with improved stress corrosion resistance.[132] Puiggali et al. investigated the susceptibility to stress corrosion cracking of Al-Zn- Mg-Cu alloy of different temper (under aged T351, peak aged T651 and over aged T7451) by slow strain rate testing.[135] They found that T7451 alloy was relatively resistant to corrosion and immune to SCC and T651 alloy was resistant to corrosion to a similar extent but susceptible to SCC. They explained that during aging, the GP zones might develop into MgZn 2 η and η precipitates whose volume increases with prolonged aging. Large MgZn 2 precipitates became obstacles for the crack initiation and trapped a great amount of hydrogen, thereby decreasing its interstitial concentration below a critical value. As a result, the susceptibility to SCC was decreased.[136] T351 alloy had the smallest precipitates and showed high susceptibility to pure corrosion, which seemed to decrease the relative risk of SCC. The pitting potential was the most negative for T351 temper alloy and the most positive for T7451 temper alloy indicating their decreasing tendency to pitting from which the cracks may initiate. 48

69 Lin et al. have studied the effect of minor alloying addition of Mn, Cr, Zr and Ti on SCC of Al-Zn-Mg-Cu alloys. They concluded that Mn, Cr, Zr and Ti could all reduce the susceptibility to SCC but by different mechanisms. Hardwick, Thompson and Bernstein catholically charged Al-6Zn-2Mg alloys, AA7050 and low copper AA7050 alloys with hydrogen to determine the effect of copper and microstructure on the hydrogen embrittlement of these alloys. The results showed that AA7050 was susceptible to SCC only for under aged condition and was highly resistant to SCC for peak aged and over aged conditions. Low copper AA7075, on the other hand, was embrittled for all the tempers, which indicated a beneficial effect of copper additions for increasing SCC resistance.[137] Scamans and Holroyd proposed that the magnesium promotes hydrogen diffusion into aluminum.[126, 138] They stated, SCC only occurs when cracks can be nucleated on grain boundary precipitates [139]. The enrichment of magnesium increases corrosion activity and enhanced hydrogen entry by forming magnesium hydride.[138] However, Pickens and Langan reported different results.[140] They conducted SCC tests on boltloaded DCB specimen and measured the grain boundary composition. No correlation between Mg segregation and SCC susceptibility was found in their study. Microstructural features of aluminum alloys play an important role in SCC, yet many issues remain unresolved. There is no agreement on the composition of the solute depleted zone and role of grain boundary precipitates in SCC, both of which are extremely important in the understanding of SCC. 49

70 2.5.3 Relationship between Stress Corrosion Cracking and Exfoliation Corrosion IGSCC and EFC in high strength aluminum alloys have been thought to have close relationship.[10] They both can initiate from intergranular corrosion and both propagate by a stress assisted intergraunlar corrosion mechanism.[1, 9, 10, 23] In the case of SCC, the stress is generally applied through an external load whereas in exfoliation the stress arises internally from the wedging effect of the voluminous corrosion products. Both stress corrosion cracking and exfoliation are strongly influenced by their copper content and heat treatment conditions. It has been shown that some factors that lead to an increase in the rate of grain boundary corrosion in exfoliation also have a similar effect on SCC. For example, SCC also propagates more rapidly on more elongated grain structure. At low stress intensities, SCC velocities are comparable with the rate of intergranular penetration for exfoliation corrosion.[23] 2.6 RESEARCH OBJECTIVES While the qualitative effects of grain shape and heat treatment condition on exfoliation susceptibility are well known, there has been little attempt to measure these effects quantitatively. The reason is that no satisfactory method to quantify the extent of exfoliation exists. What is the best accelerated environment for laboratory testing of EFC is still an open question. The first objective of this work is to develop a laboratory testing method and to quantitatively investigate the effect of grain shape and heat treatment on exfoliation corrosion kinetics. 50

71 It has been noticed that some materials can be free from exfoliation during exposure in one location for many years, and attacked by exfoliation within several months in another place. This can be explained by a threshold humidity, below which the exfoliation cannot occur. Little attempt has been made to investigate the threshold humidity of exfoliation corrosion and few reports about the effects of relative humidity on susceptibility to exfoliation corrosion exist. The effect of humidity is studied in this work and the threshold humidity for EFC propagation is determined. To understand the nature of exfoliation, it is important to understand the role of the grain boundary constituents, namely the behavior of the PFZ and the precipitates, in determining EFC susceptibility. From the literature review above, materials having same grain structure could have different susceptibility to EFC.[1] The question can probably be solved by looking at the microchemistry in the grain interior or at the grain boundary. In the proposed research, TEM and FIB are utilized to characterize the grain boundary precipitates and PFZ to investigate the correlation between microchemistry and EFC susceptibility. The existence of exfoliation corrosion on aircraft structure is detrimental to its fracture and fatigue behavior.[3, 6] Similarly, mechanical stress also affects the propagation of exfoliation corrosion. Airplane wingskin plates experience both compressive and tensile stress either during flight or on the ground. It is important to understand the effect of mechanical stress on exfoliation corrosion. In this work, a special apparatus is designed to apply surface strain. The effects of compression and the tension on exfoliation corrosion kinetics are quantitatively studied. 51

72 In order for airplane maintenance to be based on need rather than time, a corrosion model is required for predicting the corrosion extent under various conditions. The last objective of this work is to develop a model that will predict how existing corrosion will grow as a result of the total exposure conditions such as relative humidity and stress. 52

73 REFERENCES 1 M.J. Robinson and N.C. Jackson, Corrosion Science, 41, (1999). 2 T.J. Summerson and D.O. Sprowls, Corrosion Behavior of Aluminum Alloys, in International Conference in Celebration of the Centennial of the Hall-Heroult Process University of Virginia, Charlottesville,Virginia. 3 J.P. Chubb, T.A. Morad, B.S. Hockenhull, and J.W. Bristow, International Journal of Fatigue, 17(1), (1995). 4 E.A.G. Liddiard, J.A. Whittaker, and H.K. Farmery, Journal of the Institute of Metals, 89, ( ). 5 J. Zahavi and J. Yahalom, Journal of The Electrochemical Society, 129(6), (1982). 6 P.C. Conor, A.D. James, and R.N. Collier, "Evaluation of Exfoliation Corrosion Damage to 7075-T6 Aluminum Alloy Aircraft Structural Components", Materials and Structures Group, Defense Technology Agency (2004). 7 D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, "Simplified Exfoliation Testing of Aluminum Alloys", in Localized Corrosion-Cause of Metal Failure, ASTM STP 516, American Society for Testing and Materials, (1972). 8 M.J. Robinson and N.C. Jackson, British Corrosion Journal, 34(1), (1999). 9 M.J. Robinson, Corrosion Science, 22(8), (1982). 10 D.J. Kelly and M.J. Robinson, Corrosion, 49(10), (1993). 11 W. Zhang and G.S. Frankel, Electrochemical and Solid-State Letters, 3(6), (2000). 12 R.H. Jones, Stress Corrosion Cracking. Materials Park, Ohio: ASM International, 1-41 (1992). 13 G.S. Frankel, Journal of Electrochemical Society, 145(6), (1998). 53

74 14 E.H. Hollingsworth and H.Y. Hunsicker, "Corrosion of Aluminum and Aluminum Alloys", in Metals Handbook, Materials Park, OH: ASM International, (1990). 15 S. Maitra and G.C. English, Metallurgical Transactions A, 12A, (1981). 16 D.A. Jones, "Principles and Prevention of Corrosion", 2nd ed. Upper Saddle River, NJ: Prentice-Hall, Inc. (1996). 17 H.P. Godard and W.E. Cooke, Corrosion, 16, (1960). 18 L.L. Shreir, F.R.I. C, and F.I. M, "Corrosion", 2nd ed., ed., L.G.Shreir. Newnes- Butterworth (1976). 19 E. Ghali, "Aluminum and Aluminum Alloys", in Uhlig's Corrosion Handbook, ed., R.W. Revie, New York, NY: John Wiley & Sons, INC., (2000). 20 E. Mattsson, L.O. Gullman, L. Knutsson, R. Sundberg, and B. Thundal, British Corrosion Journal, 6, (1971). 21 B.W. Lifka and D.O. Sprowls, Corrosion, 22, 7-15 (1966). 22 H.W. Pickering, F.H. Beck, and M.G. Fontana, Corrosion, 18, (1962). 23 M.J. Robinson, Corrosion Science, 23(8), (1983). 24 D. McNaughtan, M. Worsfold, and M.J. Robinson, Corrosion Science, 45, (2003). 25 G. Bassi and J.J. Theler, Zeitschrift fur Metallkunde, 60(3), (1969). 26 D. Adenis and A. Guilhaudis, Memoires Scientifiques de la Revue de Metallurgie, 64(10), (1967). 27 K. Ebtehaj, D. Hardie, and R.N. Parkins, British Corrosion Journal, 24(3), (1989). 28 R. Ambat and E.S. Dwarakadasa, Corrosion Science, 33(5), (1992). 29 B.W. Lifka, D.O. Sprowls, and J.G. Kaufman, Exfoliation and Stress-Corrosion Characteristics of High Strength, Heat Treatable Aluminum Alloy Plate, in Joint Conference of North Central and Northeast Regions, National Association of Corrosion Engineers, 1967, Columbus, Ohio. 30 W. Rosenkranz, Metall, 15(5), (1961). 54

75 31 D.G. Evans and P.W. Jeffrey, Exfoliation Corrosion of AlZnMg Alloys, in U.R.Evans Conference on Localized Corrosion, NACE-3, 1974, Houston,TX: Alcan International Limited Research Center. 32 H.A. Holl, Journal of the Institute of Metals, 97(200), (1969). 33 M. Posada, L.E. Murr, C.-S. Niou, D. Roberson, D. Little, R. Arrowood, and D. George, Materials Characterization, 38, (1997). 34 B.W. Lifka, "Corrosion of aluminum alloys", in Corrosion Engineering Handbook, P.A.Schweitzer, Marcel Dekker, New York: (1996). 35 G.S. Haynes and R. Baboian, "Modified Salt Spray (Fog) Testing, Laboratory Corrosion Tests and Standards", in ASTM Standard G85-85 A2. STP866, Philadelphia, PA: ASTM, (1985). 36 "Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)", in Annual Book of ASTM Standards, Philadelphia, PA: The American Society for Testing and Materials, (1990). 37 D.J. Kelly, "Exfoliation and Stress Corrosion Cracking of the Aluminum Lithium Alloy 8090", Cranfield University. (1991). 38 B.W. Lifka and D.O. Sprowls, "Significance of Intergranular Corrosion in High Strength Aluminum Alloy Products", in Localized Corrosion : Cause of Metal Failure. ASTM STP 516, p. 120 (1972). 39 J. Onoro and C. Ranninger, British Corrosion Journal, 28(2), (1993). 40 W.A. Bell and H.S. Campbell, Journal of the Institute of Metals, 89, ( ). 41 A. Garner and D. Tromans, Corrosion, 35(2), (1979). 42 T. Sheppard and N.C. Parson, Materials Science and Technology, 3, (1987). 43 R.G. Buchheit, J.P. Moran, and G.E. Stoner, Corrosion Science, 50(2), (1994). 44 R.G. Buchheit, J.P. Moran, and G.E. Stoner, Corrosion Science, 46(8), (1990). 45 H.M. Flower and P.J. Gregson, Materials Science and Technology, 3, (1987). 55

76 46 S. Lee and B.W. Lifka, "Modification of the EXCO Test Method for Exfoliation Corrosion Susceptibility in 7XXX, 2XXX and Aluminum-Lithium Alloys", in New Methods for Corrosion Testing of Aluminum Alloys, ASTM STP 1134, eds., V.S.Agarwala and G.M.Ugiansky, Philadelplia, PA: American Society for Testing and Materials, 1-19 (1992). 47 E. Tankins, J. Kozal, and E.W. Lee, Journal of Metals, 9, (1995). 48 D.O. Sprowls, T.J. Summerson, and F.E. Loftin, "Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys-Interim Report on Atmospheric Exposure Tests", in American Society for Testing and Materials, Philadelphia, PA: (1973). 49 V.E. Carter and H.S. Campbell, Journal of the Institute of Metals, 89, (1961). 50 J.R. Davis, Corrosion of Aluminum and Aluminum Alloys. Materials Park, OH: ASM International, p.68 (1999). 51 J.D. Damborenea and A. Conde, British Corrosion Journal, 30(4), (1995). 52 R. Braun, British Corrosion Journal, 30(3), (1995). 53 H.B. Romans, Materials Research & Standards,, (1969). 54 M. Keddam, C. Kuntz, H. Takenouti, D. Schuster, and D. Zuili, Electrochimica Acta, 42(1), (1997). 55 A. Conde and J. De Damborenea, Corrosion Science, 42, (2000). 56 A. Conde and J. De Damborenea, Electrochimica Acta, 43(8), (1998). 57 F.H. Cao, Z. Zhang, J.F. Li, Y.L. Cheng, J.Q. Zhang, and C.N. Can, Materials and Corrosion, 55(1), (2004). 58 M.O. Speidel and M.V. Hyatt, "Stress-Corrosion Cracking of High-Strength Aluminum Alloys", in Advances in Corrosion Science and Technology, eds., M.G. Fontana and R.W. Staehle, New York-London: Plenum Press, p.118 (1972). 59 D.R. Baer, C.F. Windisch, M.H. Engelhard, M.J. Danielson, and R.H. Jones, Journal of Vacuum Science and Technology A, 18(1), (2000). 60 E.H. Dix Jr., Transactions, American Institute of Mining and Metallurgical Engineers, 137, (1940). 61 M.S. Hunter, G.R. Frank, and D.L. Robinson,, in Second International Congress on Metallic Corrosion National Association of Corrosion Engineer. 56

77 62 R.H. Brown, W.L. Fink, and M.S. Hunter, Transactions, American Institute of Mining and Metallurgical Engineers, 143, (1941). 63 E.H. Dix Jr., Transactions of the American Society for Metals, 42, (1950). 64 J.R. Galvele and S.M. De De Micheli, Corrosion Science, 10, (1970). 65 R.G. Buchheit, Journal of The Electrochenical Society, 142(11), (1995). 66 K. Sugimoto, K. Hoshino, M. Kageyama, and Y. Sawada, Corrosion Science, 15, (1975). 67 S. Maitra and G.C. English, Metallurgical Transactions A, 13A, (1982). 68 Q. Meng and G.S. Frankel, Surface and Interface Analysis, 36, V. Guillaumin and G. Mankowski, Corrosion Science, 41, (1999). 70 V. Guillaumin and G. Mankowski, Corrosion Science, 42, (2000). 71 T. Ramgopal, "Intergranular Corrosion of AA7XXX Alloys", MSE, The Ohio State University: Columbus. (2001). 72 R.G. Buchheit, F.D. Wall, G.E. Stoner, and J.P. Moran, Corrosion, 51(6), (1995). 73 A.J. Cornish and M.K.B. Day, Journal of the Institute of Metals, 97, (1969). 74 T. Ramgopal, P. Schmutz, and G.S. Frankel, Journal of The Electrochenical Society, 148(9), B348-B356 (2001). 75 R.G. Buchheit, L.P. Montes, M.A. Martinez, J. Michael, and P.F. Hlava, Journal of The Electrochenical Society, 146(12), (1999). 76 R.G. Buchheit, R.P. Grant, P.F. Hlava, B. McKenzie, and G.L. Zender, Journal of The Electrochemical Society, 144(8), (1997). 77 J.R. Scully, T.O. Knight, R.G. Buchheit, and D.E. Peebles, Corrosion Science, 33, (1993). 78 Katsuhiko Urushino and Katsuhisa Sugimoto, Corrosion Science, 19, (1979). 79 B. Mazurkiewicz and A. Piotrowski, Corrosion Science, 23(7), (1983). 57

78 80 R.G. Buchheit, M.A. Martinez, and L.P. Montes, Journal of The Electrochemical Society, 147(1), (2000). 81 P. Schmutz and G.S. Frankel, Journal of The Electrochemical Society, 145(7), (1998). 82 G.S. Chen, M. Gao, and R.P. Wei, Corrosion, 52(1), 8-15 (1996). 83 L.I. Maissel and R. Glang, "Handbook of Thin Film Technology", New York: McGrawHill (1970). 84 T. Ramgopal, P.I. Gouma, and G.S. Frankel, Corrosion, 58(8), (2002). 85 M. Zamin, Corrosion, 37(11), (1981). 86 E.I. Storchai and A.V. Turkovskaya, Protection of Metals, 6, 626 (1970). 87 K. Sugimoto and Y. Sawada, Journal of the Japan Institute of Metals, 37, 148 (1973). 88 R. Grauer and E. Wiedmer, Werkstoffe Korros., 31, 550 (1980). 89 A.P. Bond, G.F. Bolling, H.A. Domian, and H. Biloni, Journal of The Electrochemical Society, 113, 773 (1966). 90 M.C. Reboul and J. Bouvaist, Werkstoffe Korros., 30, 700 (1979). 91 K. Nisancioglu, Journal of The Electrochemical Society, 137(1), (1990). 92 A.I. Golubev and M.N. Ronzhin, "Corrosion of Metals and Alloys", eds., N.D.Tomashov and E.N.Mirolyubev. London: Oldbourne Press (1966). 93 G.T. Burstein and G.W. Ashley, Corrosion, 39(6), (1983). 94 W.W. Binger, E.H. Hollingsworth, and D.O. Sprowls, "Aluminum", ed., K.R.Van Horn. Metals Park, OH: American Society for Metals (1967). 95 J. Embury and R. Nicholson, Acta Metallurgica, 13, (1965). 96 N. Ryum, B. Haegland, and T. Lindtveit, Zeitschrift fur Metallkunde, 58, (1967). 97 E.H. Dix Jr., in Physical Metallurgy of Aluminum Alloys, Cleveland: American Society for Metals, 214 (1949). 98 S.J. Ketcham and F.H. Haynie, Corrosion, 19(7), 242t (1963). 99 S.J. Ketcham, Corrosion Science, 7(6), 305 (1967). 58

79 100 C. Kumai, J. Kusiniski, G. Thomas, and T.M. Devine, Corrosion, 45, 294 (1989). 101 D.A. Vermilyea, Journal of The Electrochemical Society, 119, (1972). 102 H.L. Logan, Journal of Research of the National Bureau of Standards, 48, 99 (1952). 103 E.W. Hart, "Surface and Interfaces II", Syracuse University Press, p.120 (1968). 104 H.J. Engle, "The Theory of Stress Corrosion Cracking in Alloys", North Atlantic Treaty Organization (1971). 105 J.C. Scully, Corrosion Science, 15, 207 (1975). 106 R.W. Staehle, in the Theory of Stress Corrosion Cracking in Alloys, North Atlantic Treaty Organization, p.223 (1971). 107 T.D. Burleigh, Corrosion, 47(2), (1991). 108 M. Izu, M. Shiokawa, and E. Sato, Journal of the Japan Institute of Metals, 34(11), (1984). 109 W. Gruhl, Zeitschrift fuer Metallkunde, 75(11), (1984). 110 J.G. Rinker, M. Marek, and T.H. Sanders, Materials Science and Engineering, 64(2), (1984). 111 E.I. Meletis, Materials Science and Engineering, 93(1-2), (1987). 112 A.J. Sedriks, J.A.S. Green, and D.L. Novak, Corrosion Processes and Solution Chemistry Within Stress Corrosion Cracks in Aluminum Alloys, in National Association of Corrosion Engineers, 1971, Houston, TX. 113 A.H. Le and R.T. Foley, Corrosion, 40(4), (1984). 114 Y. Choi, H.-C. Kim, and S.-I. Pyun, Journal of Materials Science, 19(5), (1984). 115 M.B. Hinz, Scripta Metallurgica, 19(12), (1984). 116 A. Joshi, C.R. Shastry, and M. Levy, Metallurgical Transaction A: Physical Metallurgy and Materials Science, 12A(6), (1981). 117 P.K. Poulose, J.E. Morral, and A.J. Mevily, Metallurgical Transactions, 5(6), (1974). 118 I.L. Muller and J.R. Galvele, Corrosion Science, 17, 179 (1977). 59

80 119 R.J. Gest and A.R. Troiano, Corrosion, 30, (1974). 120 T.F. Klimowicz and R.M. Latanision, Metallurgical Transaction A, 9A, (1978). 121 J. Albrecht, A.W. Thompson, and I.M. Bernstein, Metallurgical Transaction A, 10A, (1979). 122 K. Rajan, W. Wallace, and J.C. Beddoes, Journal of Materials Science, 17, (1982). 123 Z.-X. Tong, S. Lin, and C.-M. Hsiao, Metallurgical Transactions A, 20A, (1989). 124 C.P. You, M. Dollar, A.W. Thompson, and I.M. Bernstein, Metallurgical Transaction A, 22A, (1991). 125 D. Hardie, N.J.H. Holroyd, and R.N. Parkins, Materials Science, 12(11), (1979). 126 G.M. Scamans, Metallurgical Transactions A: Physical Metallurgy and Materials Science, 11A(5), (1980). 127 W.-Y. Chu, Y.-B. Wang, and C.-M. Hsiao, Corrosion, 38(11), (1982). 128 Z.-X. Tong, S. Lin, and C.-M. Hsiao, Metallurgical Transactions A, 20A, (1989). 129 J. Albrecht, I.M. Bernstein, and A.W. Thompson, Metallurgical Transactions A, 13A, (1982). 130 M. Talianker and B. Cina, Metallurgical Transactions A: Physical Metallurgical and Materials Science, 20A(10), (1989). 131 A.W. Thompson, M.A. Mueller, and I.M. Bernstein, Metallurgical Transactions A, 24A, (1993). 132 M.O. Speidel, "Interaction of Dislocations with Precipitates in High Strength Aluminum Alloys and Susceptibility to Stress Cracking", in Fundamental Aspects of Stress Corrosion Cracking, ed., R.W.Staehle, (1967). 133 K.G. Kent, Journal of the Institute of Metals, 97, (1969). 134 A.J. De Ardo Jr. and R.D. Townsend, Metallurgical Transactions, 1, (1970). 135 M. Puiggali, A. Zielinski, J.M. Olive, E. Renauld, and D. Desjardins, Corrosion Science, 40(4/5), (1998). 60

81 136 L. Christodoulou and H.M. Flower, Acta Metallurgica, 28, (1980). 137 D.A. Hardwick, A.W. Thompson, and I.M. Bernstein, Corrosion Science, 28(12), (1988). 138 N.J.H. Holroyd and G.M. Scamans, Scripta Metallurgica, 19(7), (1985). 139 G.M. Scamans, N.J.H. Holroyd, and C.D.S. Tuck, Corrosion Science, 27(4), (1987). 140 J.R. Pickens and T.J. Langan, Metallurgical Transactions A: Physical Metallurgical and Materials Science, 18A(10), (1987). 61

82 FIGURES Figure 2. 1 Exfoliation resulting from rapid lateral attack of selective boundaries or strata forming wedges of corrosion product that force layers of metal upward giving rise to a layered appearance.[36] (Reprinted, with permission, from ASTM G34-90 Standard Test Method for Exfoliation Corrosion Susceptibility in 2XXX and 7XXX Series Aluminum Alloys (EXCO Test), copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428) Figure 2.2 Schematic graph of the variation of corrosion product wedging force as a function of time.[23] (Reprinted from Corrosion Science, Vol. 23, M.J. Robinson, The Role of Wedging Stresses in The Exoliation Corrosion of High Strength Aluminum Alloys, Pages: , Copyright (1983), with permission from Elsevier.) 62

83 Figure 2.3 (a) Model of the intergranular penetration of a wrought alloy showing the boundary of the attack marked by the dotted line. (b) Improved model in which the corrosion product tapers uniformly towards the tips.[9] (Reprinted from Corrosion Science, Vol. 22, M.J. Robinson, Mathematical Modelling of Exfoliation Corrosion in High Strength Aluminum Alloys, Pages: , Copyright (1982), with permission from Elsevier.) Figure 2.4 (a) Dimensions of a thin walled hemispherical shell used to calculate the excess pressure in a blister. (b) Section through a blister diameter.[9] (Reprinted from Corrosion Science, Vol. 22, M.J. Robinson, Mathematical Modelling of Exfoliation Corrosion in High Strength Aluminum Alloys, Pages: , Copyright (1982), with permission from Elsevier.) 63

84 Figure 2.5 Surface rating for 2024 in the MASTMASSIS test. (P=pitting, EA=Superficial EFC,EB=Moderate EFC,EC=Severe EFC,ED=Very Severe EFC). [1] (Reprinted from Corrosion Science, Vol. 41, M.J. Robinson and N.C. Jackson, The Influence of Grain Structure and Intergranular Corrosion Rate on Exfoliation and Stress Corrosion Cracking of High Strength Al-Cu-Mg Alloys, Pages: , Copyright (1999), with permission from Elsevier.) 64

85 Figure 2.6 (a) Corrosion product forces generated on T/2 specimens of 8090 plates. (b) Corrosion product forces generated on T/4 specimens of 8090 plate.[10] (Reprinted from Corrosion Science, Vol. 49, D.J. Kelly and M.J. Robinson, Influence of Heat Treatment and Grain Shape on Exfoliation Corrosion of Al-Li Alloy 8090, Pages: , Copyright (1993), with permission from Elsevier.) 65

86 Figure 2.7 The schematic equivalent circuit for the immersion of AA7075 in EXCO solution. (Rs: solution resistance; Cdl: origin surface double layer capacitive; Rp: polarization resistance; Clf: low frequency capacitive; Rt: charge transfer resistance). [57] Figure 2. 8 Schematic illustration of a mechanism for redistribution of Cu by dissolution of large Al 2 CuMg and Al 2 Cu intermetallic particles in Al alloys. [80] (Reproduced by permission of The Electrochemical Society, Inc.) 66

87 Figure 2. 9 Potentiodynamic polarization scans for bulk Al 2 CuMg in 0.5M NaCl solution open to air, actively aerated, and actively deaerated by sparging with N 2.[75] (Reproduced by permission of The Electrochemical Society, Inc.) Figure Variation in corrosion potential of FeAl 3 and α-al(fe,mn)si.[91] (Reproduced by permission of The Electrochemical Society, Inc.) 67

88 Figure Anodic part of cyclic polarization curve for Al 3 Fe in deareated 0.1N NaOH solution. The curve corresponds to the first cycle starting with an as polished surface at 1.5V. In this particular run, the sweep was reversed at 0V. Continuous curve: forward sweep. Dashed curve: reverse sweep. [91] (Reproduced by permission of The Electrochemical Society, Inc.) Figure 2.12 Schematic diagram of typical crack propagation rate as a function of crack tip stress intensity behavior illustrating the regions of stage 1,2 and 3 crack propagation as well as identifying the plateau velocity and the threshold stress intensity.[12] (Reprinted with permission of ASM International. All rights reserved) 68

89 CHAPTER 3 QUANTITATIVE STUDY OF EXFOLIATION CORROSION: EXFOLIATION OF SLICES IN HUMIDITY TECHNIQUE 3.1 INTRODUCTION Exfoliation corrosion (EFC) is a particular form of intergranular corrosion that is often observed on the surface of wrought aluminum alloys with an elongated grain structure.[1] When intergranular corrosion proceeds along intergranular paths parallel to the surface, the hydrated aluminum oxide corrosion product has a higher volume than the alloy matrix from which the product formed. Consequently, relatively large wedging stresses are produced, which lift the surface grains, giving rise to a layered appearance. EFC is significant source of life-limiting degradation in airframes.[2] It consumes load bearing cross section and increases the stress in the remaining intact material, resulting in a loss in mechanical properties of Al alloys. The susceptibility of Al alloys to EFC can be assessed by exposure testing. ASTM G-34, known as the EXCO test, involves exposure to an oxidizing acidic chloride solution and comparison of the resulting surface to standard photographs.[1, 3] Other 69

90 tests involving exposure to aggressive environments have been devised, including the ASSET and MASTMAASIS tests. In these methods, salt solutions are alternately sprayed onto the surface of the specimens and the resulting surfaces are evaluated to determine the EFC susceptibility.[4-7] The behavior of Al alloys in these accelerated environments has been correlated to long-term exposure in less-aggressive natural environments.[8-10] As a result, these tests are very useful for assessing susceptibility to EFC attack. However, they do not provide quantitative measurements of susceptibility or growth kinetics, which are required for predictive modeling of corrosion development. Liddiard and coworkers have used a deflection technique to quantify exfoliation extent and determine EFC kinetics.[11] In this technique, the effective remaining load-bearing section of specimens having undergone EFC was determined from their compliance under four-point bending. The rate of EFC can be assessed from periodic measurements. The deflection technique is valid only when the thinning of the specimen during corrosion is uniform. The alloy grain shape is thought to be an important factor in exfoliation corrosion. Robinson et al. found that the severity of exfoliation corrosion is related to the grain aspect ratio of the material.[12-14] More elongated grain shape results in severe exfoliation corrosion. Besides elongated grain structure, grain boundary chemistry also plays an important role in the EFC susceptibility. It is well known that grain boundaries are often more susceptible to corrosion than the grain interiors because of the microstructrual heterogeneity. The grain boundary region of a typical Al alloy can contain precipitates on the grain boundary and a precipitate free zone (PFZ) next to the grain boundary, which have very different electrochemical behaviors than the grain 70

91 interior region.[15] Maitra and English have attributed the IGC susceptibility of AA7075- T6 to Mg and Zn solute segregation or enrichment in the grain boundary region.[16] Ramgogal et al. proposed that IGC in 7075-T6 temper was caused by anodic dissolution of Mg(ZnCuAl) 2.[15, 17] Other investigations have suggested that the cause of exfoliation is an electrochemical potential developed between primary or secondary particles,which contain some or all of the elements Al, Fe, Si, Mn, Mg, and Zn, and the surrounding AlZnMg solid solution.[18, 19] Evans proposed the possibility of an electrochemical cell occurring between Mn-rich and Mn-depleted regions. The Mn bearing particles precipitate preferentially in the center of the grains compared to the region near grain boundaries, resulting in differing electrochemical potential.[5] Determination of EFC kinetics is critical for the development of predictive failure models. The aim of this work was to develop a technique and analysis methodology that can provide quantitative measurement of exfoliation corrosion kinetics. The EFC rates of different AA7178 wingskin plates were determined and related to the alloy microstructure and compositions of the grain boundary constituents determined from analytical TEM techniques. 3.2 EXPERIMENTAL Materials Test samples were machined from a piece of a wingskin of a retired KC135 airplane. The wingskin section contained two AA7178-T6 plates attached to an underlying support beam by steel rivets. The wingskin section was divided into two parts. One part of the wingkin section was exposed in the uncoated condition at an atmospheric 71

92 exposure test site near Daytona Beach, FL by W. Abbott of Battelle. The other part of the section was used to perform laboratory testing. Figure 3.1 shows the section after 9 months exposure at Daytona Beach. The two plates in the section exhibited vastly different exfoliation behavior during the exposure near the seacoast. One of the plates exfoliated badly next to the steel rivets and the other plate only developed cosmetic surface attack. They are referred to as good and bad plates in reference to their EFC susceptibility. Samples were machined from the good and bad plates in the shape of rectangular slices, as illustrated in figure 3.2. The slices were 3-4 cm long, and oriented such that the long axis of each slice was in the longitudinal orientation of the microstructure (along the rolling direction). The slice thickness was around 1 mm, and the slice thickness was oriented in the plate transverse direction. The width of each slice was the full plate through-thickness in the short transverse direction: 4.1 and 4.6 mm for the good and bad plates, respectively. The slice edges, which were the original outer surfaces of the plate, were lightly polished. All other faces were ground in ethanol to 800 grit finish, cleaned ultrasonically in ethanol, and finally dried by an air stream. Detailed metallographic analysis was performed on T (long transverse), S (short transverse), and L (longitudinal) sections of the good and bad plates. Samples were polished to 1 micron and chemically etched with Keller s reagent (2mL HF, 3mL HCl, 5mL HNO 3 and 190 ml H 2 O) to reveal the microstructure. Optical microscopy methods were used to examine metallographic sections in three perpendicular orientations. The grain dimension in three orientations as a function of the position along through thickness direction was determined with Clemex Vision image analysis software. The through 72

93 thickness cross section was divided into 15 zones. In each zone, 30 grains were measured and the mean values were calculated to represent the grain size in each zone. Analytical transmission electron microscopy (TEM) was performed to investigate the local chemistry at the grain boundaries. A 10 mm long sample was cut from a T section of the good plate. The width of the sample was the whole thickness of the plate. The sample surface was polished to 0.5 µm by alumina suspended in ethanol and slightly etched with acid reagent (6mL HF, 40mL H 2 SO 4, 360 ml H 2 O) for 20 seconds to reveal the grain boundaries on the surface. Cross sectional TEM samples were prepared by an FEI Strata Dual Beam 235M scanning electron microscope (SEM) / focused ion beam (FIB) tool using a 30 kev Ga ion beam and a 5 kev electron beam. The TEM samples were made such that they contained a grain boundary located 10~30 µm away from the surface. TEM samples were made close to the upside surface (corresponding to the surface that was exposed to the Daytona Beach environment) and close to the downside surface (corresponding to the surface that was not exposed to the Daytona Beach environment). Prior to sectioning, specific sites near the specimen edge were located and identified by SEM/EDS in the FIB and then covered by deposition of a 1.5 µm-thick Pt layer to protect the specimen surface during the FIB sectioning. The membrane had an area of 15 µm x 5 µm and was thinned in the FIB to a thickness of 100 nm for electron transparency. The membrane was plucked out of the bulk sample under an optical microscope using a sharp Pyrex needle of 1 um in diameter and placed on a 200 mesh Au TEM grid with a formvar/ carbon support film for TEM analysis. The TEM observation was made in the region of the membrane above a hole in the mesh grid. TEM characterization of the FIB sectioned membranes was conducted using an FEI Tecnai 73

94 TF20 scanning transmission electron microscope (STEM) operating at 200 kv. The probe size was <2 nm and the step size was 5 nm. Electrochemical polarization measurements were performed on samples ground to 1200 grit. 1 M NaCl solution was deareated with Ar gas to decrease the corrosion potential and allow for clear observation of the breakdown potentials. Potentiodynamic scans were performed at a rate of 0.1 mv/s. A Pt counter electrode and saturated calomel reference (SCE) were used. All potentials in this paper are referenced vs. SCE Exfoliation of Slices in Humidity (ESH) Technique Samples sliced from the plates as described above were given a potentiostatic electrochemical pretreatment in 1 M NaCl at a potential of 710 mv SCE for 7 h. This potential is above the second breakdown potential for this material in this environment, which is -725 mv SCE as will be shown below. The purpose of the pretreatment was to initiate localized corrosion attack and develop an aggressive environment in the corrosion sites. Following the pretreatment, the sample was rinsed with DI water and placed in a humidity chamber, consisting of a sealed beaker containing a saturated salt solution at room temperature (22-25 C). The samples were held in the air space above the solution and a graph paper was placed behind them to facilitate determination of EFC extent. Different saturated salt solutions were used to create a range of constant humidity: sodium sulfate (Na 2 SO 4 ), ammonium chloride (NH 4 Cl), potassium iodide (KI), potassium carbonate (K 2 CO 3 2H 2 O), and calcium chloride (CaCl 2 6H 2 O). The relative humidities above saturated solutions of these salts measured by an RH meter were similar to handbook values[20], as is reported in Table 3.1. Pretreated samples developed EFC 74

95 upon subsequent exposure to high humidity. The EFC started at the outer edges, corresponding to the original plate surfaces, and moved inward. Images of the samples were recorded by digital photography through the glass walls of the humidity chamber. The contrast between the boundary of the outer exfoliated and inner unattacked regions was sufficient to allow tracking of the EFC kinetics by analysis of the photographs. The width change of the inner unattacked region of an ESH sample can be related to the material loss due to exfoliation corrosion. The photographs of the ESH samples were analyzed to measure the unattacked width as a function of time. One measurement of the unattacked width was made on each sample at a position 10 mm from the top of the sample. This value was subtracted from the original width at this position, d 0, to determine the depth of metal consumed by EFC: d = d 0 d, where i=1,2,.day. i i Since the exfoliation occurred at the two edge surfaces, one half of the width change d i 2 represents the average EFC depth. 3.3 RESULTS AND DISCUSSION Electrochemical and ESH Characterization Potentiodynamic polarization curves of the good and bad plates were obtained in deareated 1 M NaCl at a potential sweep rate of 0.1 mvs -1, figure 3.3. Both plates exhibited two breakdown potentials as has been observed for AA7xxx alloys in the underaged and peak aged conditions.[21] The values of the breakdown potentials are similar for the two plates, so the difference in exfoliation susceptibility cannot be attributed to a difference in breakdown potentials. There was, however, a difference in 75

96 the magnitude of both the transient peak above the first breakdown potential and the current flowing at higher potentials; both were smaller for the good plate. This reflects some differences in the microstructure. Samples of the good and bad plates from the AA7178 wingskin section were ESH-tested by electrochemical pretreatment and then exposure to 96% RH. Figure 3.4 shows images of the samples over a period of time from 0 (as-pretreated) to 52 days of high humidity exposure. The samples are oriented such that the rolling direction is vertical and the plate thickness is horizontal, with the original plate surfaces on the left and right edges of the samples. After several days, EFC was evident on both edges of the slice from the bad plate and it continued to progress more or less evenly on both edges over the exposure period of almost 2 months, figure 3.4a. The behavior of the slice removed from the good plate behaved quite differently. One side of the sample (the left side of the images in figure 3.4b) exhibited severe EFC, while the other side of the sample was practically unattacked. The side of the slice that was not attacked in the ESH test corresponds to the side of the good plate that was exposed upwards to the environment at Daytona Beach. So the high resistance of the good plate to EFC was reproduced in the lab in that the surface exposed at Daytona Beach reacted slowly in the ESH test. This resistance was only representative of one surface; the other side of the good sample corresponding to the side of the plate that faced downward and was not exposed to the elements at Daytona Beach, showed rapid attack in the ESH test. It is expected that the good plate would have exfoliated rapidly at Daytona Beach if the plate had been oriented such that the down side had been facing upward. Overall, the bad sample was attacked much faster than the good sample. 76

97 As seen in figure 3.4, corrosion product exuded out of the surfaces (corresponding to transverse face of the microstructure) of both good and bad samples upon exposure to 96% RH. This generation of corrosion product is evidence of continued corrosion at localized sites from the aggressive environment deposited in the microstructure during the electrochemical pretreatment. However, the EFC proceeded inward from the edges. The kinetics of EFC can be determined by measuring the change in width of the central unexfoliated region. Figure 3.5 shows the average exfoliation depth, which is equal to half of the change in width of the inner unexfoliated region. The good sample exfoliated primarily on one side, so the real EFC depth for the attacked side of the good sample is approximately twice the value given in the figure 3.5. Considering this fact, the depth of EFC on the susceptible side of the good sample was approximately the same as the depth on both sides of the bad sample, which is evident in figure 3.4. Figure 3.5 shows that the rate of EFC was not constant with time. For 2 samples, it started out rapidly, with a change of about 80 µm in the first day and then slowed to about 4 µm/day. The other two samples did not experience the initial rapid EFC and started off with a slow rate. It is interesting that the two good and two bad samples did not behave the same at the beginning. One of each type started out quickly and one started slowly. After about 2 weeks, the EFC rate for the bad samples and one of the good samples increased to a higher value, about 20 µm/day. The rate for each of these samples then decreased again to a lower value of about 2 µm/day. The other good sample (the one shown in figure 3.4b) did not exhibit an increase again until after 35 days. These results indicate that there are zones of the plate through the thickness with varying susceptibility and EFC kinetics. The apparent resistance of the good plate during outdoor exposure 77

98 at Daytona Beach was caused by the seemingly chance location of a resistant zone at the upward-facing surface of that plate. It is interesting that zones of varying EFC susceptibility exist through the thickness of a plate. This has never been reported, but other techniques, such as the EXCO tests, do not have the sensitivity to find these zones. This varying susceptibility could be caused by variations in microstructure or residual stress in the plate. ESH testing was performed on good and bad samples under a range of constant humidity, figure 3.6. The two plates exhibited different exfoliation behavior in different humidities. In the lowest humidity (30%, CaCl 2 ), the surfaces of both samples were still shiny after 60 days exposure, and no EFC was evident. This humidity is apparently below the critical humidity of the environment deposited in the localized corrosion sites created by pretreatment. This environment dried completely, preventing further attack. In about 50% RH, some product exuded out of the surface and EFC was seen on the downside of the "good" sample after 3 days. However, the EFC ceased after that time, perhaps because it took some time for the environment in the pretreatment sites to equilibrate with the humidity in the chamber and dry out. Above around 50% RH, samples exfoliated more severely with increasing humidity. The average EFC rate was 3.0, 4.0, and 14.0 µm/day for the bad samples in 66, 76, and 96% RH, respectively. For good samples, the values were 0.6, 0.8, and 5.0 µm/day, respectively, for the same humidities. The effect of humidity and other factors on EFC rate will be covered in detail in a subsequent communication.[22] However, it is clear from these data that the rate of EFC increases with increasing humidity above a critical humidity. 78

99 3.3.2 Optical Microscopy The microstructures of the good and bad plates are given in figure 3.7, along with the convention utilized for the various sections. Figure 3.8 is a montage of highermagnification micrographs of the T sections of the two plates covering the full cross section from one side of the plate to the other. Each section is given in two parts, with some overlap of the top left and bottom right. The grains in the bad plate are larger and more elongated in the L direction than in the good plate. In particular, the near-surface regions of the good plate exhibit much smaller grains with a smaller aspect ratio. The grain sizes of the good and bad plates as a function of the position through thickness direction were obtained by the analysis of L, T, and S cross-sections and are shown in figure 3.9. The grain aspect ratios in longitudinal and transverse orientations are reported in figure 3.9 with the L/S ratio shown above the T/S ratio. The bad plate has a higher grain aspect ratio than the good plate. In particular, the size of grains in the L direction is about 3 times greater in the bad plate. It is known that a microstructure with grains that are less-elongated should be more resistant to exfoliation (but should exhibit faster IGC in the S direction). The larger grain size and higher grain aspect ratio of the bad plate contribute to its higher EFC susceptibility. There is a small difference in the grain aspect ratio near the upside and downside of the good sample. However, it is unlikely that the slightly smaller grain aspect ratio at the upside is the sole explanation for the much better EFC resistance there. Furthermore, the grain size and grain aspect ratio near the downside of the good sample are much smaller than those of the bad sample, yet the EFC rate was similar. Therefore, factors other than grain size and shape must play a 79

100 determining role in the EFC susceptibility. Possibilities include residual stress in the plate and local variations in grain boundary microchemistry TEM Analysis Analytical transmission electron microscopy (TEM) was used to investigate the composition of the grain boundary constituents in the good plate. Figure 3.10 shows transmission electron micrographs of samples taken from near the up and downsides of the good plate. Both micrographs exhibit particles in the matrix as well as on the grain boundaries and a precipitate free zone (PFZ) on either side of the grain boundaries. The width of the PFZ near the upside of the good plate was around 90~100 nm, which is somewhat larger than the PFZ width near the downside, less than 50 nm. There was no clear difference in the size, distribution, and composition of the large precipitates in the matrix near the two surfaces. EDS line profiling was performed across grain boundaries at locations that include precipitates and locations away from precipitates to analyze the PFZ. Four different precipitates and five different PFZs along grain boundaries near upside and downside of the good plate were analyzed. Figure 3.11 and 3.12 show representative precipitates and PFZs that exist near upside and downside of the plate. The lines in the higher resolution images in these figures show the paths of the line profiles in one measurement. The concentration profiles in the form of concentration ratios for various elements relative to Al are shown in figure 3.13 for the line-scans across the grain boundary precipitates in figure 3.11, and in figure 3.14 for the line scans across the grain boundary precipitate free zones in figure The concentration ratios were determined using the appropriate k factors and the x-ray intensity ratios. Two kinds 80

101 of precipitates are found along grain boundaries near both upside and downside of the plate. One is Al-Zn-Mg-Cu-Cr precipitate (concentration profiles across these precipitates are shown in figure 3.13a near upside of the plate and in figure 3.13c near the downside of the plate) and the other is Cu-Al precipitate (concentration profiles across these precipitates are shown in figure 3.13b near upside of the plate and figure 3.13d near the downside of the plate). There are no obvious differences in the precipitates near the up- and downsides of the plate. Furthermore, it is unlikely that a difference in the composition of grain boundary precipitates can explain the difference in EFC behavior because the precipitates are isolated along the boundaries in this material, i.e. they do not form a continuous pathway on the grain boundaries. The concentration profiles in the form of concentration ratios for the line-scans in figure 3.12 across PFZ regions are shown in figure Zn is seen to be depleted in the PFZ near the grain boundary at the upside surface relative to the nearby lattice, figure 3.14a. In contrast, there is no apparent Zn depletion in the PFZ near the grain boundary at the downside surface, figure 3.14b. As noted above, only a limited number of different grain boundaries were analyzed (five near the upside and five near the downside). It is always possible to question the validity of using a limited number of very local TEM measurements to explain bulk properties. Nonetheless, the profiles shown in figure 3.14 are representative of the regions studied, and a difference in Zn concentration in the PFZ regions near the upside and downside was clearly evident. It is known that Zn, the main alloying element in 7xxx alloys, diffuses readily at normal homogenizing temperatures. This element has high diffusion rates and low solubility in Al, resulting in the formation of Mg-Zn containing phases along grain 81

102 boundaries and the depletion of Mg and Zn in the area adjacent to the grain boundaries.[5] Park and Ardell have used analytical transmission electron microscopy to show that the grain boundary region in AA7150-T6 exhibited considerable Zn depletion relative to the matrix.[23] Many studies have shown that the alloying of aluminum with Zn reduces the pitting potential and repassivation potentials.[24-27] Ramgopal et al. investigated the dissolution kinetics of binary Al alloys containing Cu, Zn and Mg in chloride solution and found out Zn addition decreased the repassivation potential by increasing the dissolution kinetics at a given potential. [28] The TEM results in this study suggest that the difference in Zn concentration in the PFZs near the upside and downside surfaces of the good plate is at least partly responsible for the difference in EFC behavior. This different distribution of Zn might be caused by details of the plate processing. The Cr and Cu enrichment in GB precipitates results in depletion of these two elements in grain boundary area. The PFZ is anodic to the rest of the grain and is preferentially attacked. In PFZs near the upside surface of this plate, the absence of solute Zn increased the corrosion resistance of the PFZ and reduced the susceptibility to IGC and thus EFC. On the other hand, the presence of Zn in the PFZ near the downside surface sharpened the electrochemical heterogeneity between the grain boundaries and the adjacent regions and hence enhanced EFC. The grain boundary chemistry seems to be an important factor controlling the susceptibility to EFC. 3.4 CONCLUSIONS 1. The Exfoliation of Slices in Humidity (ESH) test was used to assess the susceptibility of AA7178 samples to exfoliation corrosion (EFC) and to determine 82

103 the rate of EFC. In this test, samples sliced in a particular orientation are given an electrochemical pretreatment and then exposed in a constant humidity environment. 2. The ESH test was able to reproduce exfoliation behavior of plates observed during outdoor exposure. 3. EFC was not observed below 50% RH and the EFC kinetics increased with increasing humidity above this critical humidity. 4. EFC susceptibility was found to depend on grain size and shape as well as grain boundary composition. 5. A high susceptibility to EFC was associated with a high Zn content in the precipitate free zone at grain boundaries. 83

104 REFERENCES 1 "Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)", in Annual Book of ASTM Standards, Philadelphia, PA: The American Society for Testing and Materials, (1990). 2 J.P. Chubb, T.A. Morad, B.S. Hockenhull, and J.W. Bristow, International Journal of Fatigue, 17(1), (1995). 3 S. Lee and B.W. Lifka, "Modification of the EXCO Test Method for Exfoliation Corrosion Susceptibility in 7XXX, 2XXX and Aluminum-Lithium Alloys", in New Methods for Corrosion Testing of Aluminum Alloys, ASTM STP 1134, V.S.Agarwala and G.M.Ugiansky, Philadelplia: American Society for Testing and Materials, 1-19 (1992). 4 B.W. Lifka and D.O. Sprowls, Corrosion, 22, 7-15 (1966). 5 D.G. Evans and P.W. Jeffrey, Exfoliation Corrosion of AlZnMg Alloys, in U.R. Evans Conference on Localized Corrosion, NACE-3, 1974, Houston, TX: Alcan International Limited Research Center. 6 G.S. Haynes and R. Baboian, "Modified Salt Spray (Fog) Testing, Laboratory Corrosion Tests and Standards", in ASTM Standard G85-85 A2. STP866, Philadelphia, PA: ASTM, (1985). 7 D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, "Simplified Exfoliation Testing of Aluminum Alloys", in Localized Corrosion-Cause of Metal Failure, ASTM STP 516, American Society for Testing and Materials, (1972). 8 R. Braun, British Corrosion Journal, 30(3), (1995). 9 D.O. Sprowls, T.J. Summerson, and F.E. Loftin, "Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys-Interim Report on Atmospheric Exposure Tests", in American Society for Testing and Materials, Philadelphia, PA: (1973). 10 B.W. Lifka and D.O. Sprowls, "Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Series Aluminum Alloys with Exposure to a Seacoast Atmosphere", in ASTM SPEICIAL TECHNICAL PUBLUCATION 558, (1973). 84

105 11 E.A.G. Liddiard, J.A. Whittaker, and H.K. Farmery, Journal of the Institute of Metals, 89, ( ). 12 M.J. Robinson and N.C. Jackson, Corrosion Science, 41, (1999). 13 M.J. Robinson and N.C. Jackson, British Corrosion Journal, 34(1), (1999). 14 M.J. Robinson, Corrosion Science, 22(8), (1982). 15 T. Ramgopal, P.I. Gouma, and G.S. Frankel, Corrosion, 58(8), (2002). 16 S. Maitra and G.C. English, Metallurgical Transactions A, 12A, (1981). 17 T. Ramgopal, P. Schmutz, and G.S. Frankel, Journal of The Electrochemical Society, 148(9), B348-B356 (2001). 18 E. Mattsson, L.O. Gullman, L. Knutsson, R. Sundberg, and B. Thundal, British Corrosion Journal, 6, (1971). 19 G. Bassi and J.J. Theler, Zeitschrift fur Metallkunde, 60(3), (1969). 20 R.C. Weast, M.J. Astle and W.H. Beyer, CRC Handbook of Chemistry and Physics, Boca Raton, FL: CRC Press, Inc. (1987). 21 Q. Meng and G.S. Frankel, Journal of The Electrochemical Society, 151(5), B271-B283 (2004). 22 X. Zhao, Chapter 4 of this dissertation. 23 J.K. Park and A.J. Ardell, Acta Metallurgica et Materialia, 39(4), (1991). 24 I.L. Muller and J.R. Galvele, Corrosion Science, 17, (1977). 25 F. Sato and R.C. Newman, Corrosion, 54(12), (1998). 26 F. Sato and R.C. Newman, Corrosion, 55(1), 3-9 (1999). 27 P.L. Bonora, G.P. Ponzano, and V. Lorenzelli, British Corrosion Journal, 9(2), (1974). 28 T. Ramgopal and G.S. Frankel, Corrosion, 57(8), (2001). 85

106 TABLES AND FIGURES Salt Expected %RH (from CRC) Measured %RH at RT Na 2 SO 4 93 at 20 o C 96 NH 4 Cl 79.3 at 20 o C 76.5 KI 56.2 at 100 o C 65.1 K 2 CO 3 2H 2 O 43 at 24.5 o C 49~50 CaCl 2 6H 2 O 32.3 at 20 o C 30.1 Table 3. 1 Humidity associated with saturated salt solutions.[20] 86

107 bad good Figure 3.1 Photo of AA7178 wingskin sample after 9 months of exposure at Daytona Beach. Sample was uncoated and had steel rivets attaching plates to understructure. Image provided by W. Abbott, Battelle. Figure 3.2 Schematic drawing of orientation of slices for ESH test relative to elongated microstructure of AA7178 wingskin sample. 87

108 "good" plate Potential (mv SCE) "bad" plate Current Density (A/cm 2 ) Figure 3.3 Polarization curves of AA7178 good and bad sample in deareated 1.0 M NaCl at a scan rate of 0.1 mv/s. 88

109 4.07mm 0day 7days 15days 27days 52days (a) 3.86mm 0day 2days 6days 29days 52days (b) Figure 3.4 Images of AA7178 wingskin samples exposed to 96% humidity following electrochemical pretreatment. (a) Bad sample. (b) Good sample. 89

110 0.5 Average EFC Depth (mm) "bad" samples "good" samples Time (days) Figure 3.5 Average exfoliation corrosion depth for duplicate samples of good (triangles) and bad (circles) AA7178 wingskin samples exposed to 96% RH. 90

111 Average EFC Width (mm) "good" "bad" 96% RH 76% RH 65% RH Time (days) Figure 3.6 Average exfoliation depths of good (open symbols) and bad (closed symbols AA7178 wingskin samples exposed to different humidities: circles 96%, squares 76%, triangles 65%. 91

112 Figure 3.7 Metallographic sections of good plate and bad plate. Also given is the terminology used for the different sections. (a) bad plate. (b) good plate. 92

113 S 0.2 mm L a. UP SIDE b. Figure 3.8 Metallographic sections of AA7178 wingskin plate. (a) Bad plate. (b) Good plate. The sections are through-thickness montages, starting at the right of the top image in each pair and then wrapping around to end at the left side of the bottom image in each pair. 93

114 Grain Size (um) S T L Grain Aspect Ratio L/S T/S Grain Size (um) 0 T1/15 T8/ upside Position Through Thickness (4.1mm) (a) S T L Grain Aspect Ratio T15/15 L/S T/S downside T1/15 T8/15 T15/15 upside Position Through Thickness (4.6mm) (b) downside Figure 3.9 Grain size distribution and grain aspect ratios through the thickness of (a) good plate and (b) bad plate. 94

115 (a) (b) Figure 3.10 TEM micrographs of grain-boundary in AA7178 good plate. (a) Near up side of the good plate. (b) Near down side of the good plate. 95

116 (a) (b) 96

117 (c) (d) Figure TEM micrographs of grain-boundary in AA7178 showing Nano EDS line profiling across two different types of the GB precipitates (a,b) near up side of the good plate and (c,d) near down side of the good plate. 97

118 (a) (b) Figure 3.12 TEM micrographs of grain-boundary in AA7178 showing Nano EDS line profiling across the GB PFZ (a) Near up side of the good plate. (b) Near down side of the good plate. 98

119 Concentration Ratio Profile (Ci/CAl) Concentration Ratio Profile (Ci/CAl) 0.25 CZn/Al CMg/Al 0.2 CCu/Al CCr/Al CFe/Al CTi/Al CSi/Al 0.15 CMn/Al Zn Cr Mg Cu Position (nm) (a) 0.7 CZn/Al 0.6 Cu CMg/Al CCu/Al CFe/Al 0.5 CCr/Al CSi/Al CTi/Al 0.4 CMn/Al Mg Zn Position (nm) (b) 99

120 Concentration Ratio Profile (Ci/CAl) Concentration Ratio Profile (Ci/CAl) Mg Cr Zn Zn/Al Mg/Al Cu/Al Cr/Al Fe/Al Ti/Al Si/Al Mn/Al Position (nm) (c) Cu Cu CZn/Al CMg/Al CCu/Al CFe/Al CCr/Al Csi/Al CTi/Al CMn/Al Zn 0.1 Mg Position (nm) (d) Figure 3.13 Nano-EDS line profile of grain boundary precipitates. (a, b) linescans across grain boundary precipitates in figure 3.11a and b near upside of the good plate. (c,d) ) linescans across grain boundary precipitates in figure 3.11c and d near downside of the good plate. The data are reported as ratio of the x-ray intensities for each element relative to that of Al. 100

121 Concentration Ratio Profile (Ci/CAl) Concentration Ratio Profile (Ci/CAl) PFZ CZn/Al CMg/Al CCu/Al CCr/Al CFe/Al CTi/Al CSi/Al 0.08 CMn/Al 0.06 Zn 0.04 Mg 0.02 Cu Position (nm) (a) 0.3 CZn/Al CMg/Al CCu/Al 0.25 PFZ CFe/Al CCr/Al CSi/Al 0.2 CTi/Al CMn/Al 0.15 Zn 0.1 Mg 0.05 Cu Position (nm) (b) Figure 3.14 Nano-EDS line profile of solute depleted zone around grain boundary. The data are reported as ratio of the x-ray intensities for each element relative to that of Al. (a) Near up side of the good plate (shown in figure 3.12a). (b) Near down side of the good plate (shown in figure 3.12b). 101

122 CHAPTER 4 EFFECTS OF RH, TEMPER AND STRESS ON EXFOLIATION CORROSION KINETICS OF AA INTRODUCTION Exfoliation corrosion (EFC) is a particular kind of intergranular corrosion often found in the wrought Al alloys (AA2024 and AA7178) used in wingskins and fuselages of airplanes.[1] It proceeds along elongated grain paths parallel to the surface. The internal stresses caused by voluminous corrosion products force metal grains away from the bulk giving rise to a layered appearance. EFC usually does not directly lead to failure of Al components on aircraft, but it can decrease the load-bearing cross section and lead to stress corrosion cracking or corrosion fatigue.[2, 3] A great deal of work has been devoted to address the importance of elongated grain structure on EFC propagation.[4, 5] However, a lack of quantitative understanding of the environment and applied stress conditions hinders the development of predictive models and protection schemes to prevent EFC. To assure aircraft structural integrity, it 102

123 is important to understand the EFC of aircraft materials including the effects of microstructure, environment and stress on the corrosion progress. Exfoliation corrosion is affected by environmental conditions such as humidity to a marked degree.[6, 7] In this work, EFC tests were performed in different humidities to simulate changing RH found under real conditions. Predictive models for EFC of airplane wingskins also require a quantitative understanding of the effects of stress because aircraft wings are subjected to both compressive and tensile stress, whether on the ground or in flight. Exfoliation tests were carried out in this study using a four-point bending apparatus to study the effect of stress on EFC. Susceptibility of AA7178 materials to exfoliation corrosion is influenced by heat treatment. The resistance to exfoliation corrosion is usually improved in alloys with extended aging time.[8-12] For instance, the EXCO test on 7xxx alloys show that the T7 temper is more resistant to EFC than the T6 temper.[13-15] However, for some alloys, the EXCO test is too corrosive which markedly reduces the ability to distinguish between all the various commercial tempers.[14] Exfoliation tests were carried out on T7 and T6 samples to investigate the effect of temper on exfoliation corrosion. 4.2 EXPERIMENTAL An AA7178 plate of thickness 0.89 cm cut from the wing of a decommissioned airplane was studied. The nominal composition of AA7178 (in wt %) was determined by Wavelength Dispersive X-ray Fluorescence (at the OSU Department of Geological Sciences) and is shown in table 4.1. Samples were machined from the plates in the shape of rectangular slices. The slices were oriented such that the long axis of each slice was in 103

124 the longitudinal orientation of the microstructure (along the rolling direction). The sample thickness, oriented in the plate transverse direction, was around 2-3 mm. The width of each sample was the full plate through-thickness in the short transverse direction. The sample edges, which were the original outer surfaces of the plate, were lightly polished to remove the surface coating. All other faces were ground to 800 grit in ethanol, cleaned ultrasonically in ethanol, and finally dried by an air stream. Samples were primarily tested in the as-received condition. The original temper of the AA7178 plate was probably T6, but the natural aging over the decades of service might have altered the microstructure and properties. To test the effect of temper on EFC susceptibility, the as-received AA 7178 plates were re-heat-treated to the T6 and T7 tempers.[16] Details of the heat treatments are summarized in Figure 4.1. T, S, L sections of the T6 and T7 temper were metallographically cross-sectioned. The samples were polished to 1 µm and etched in Keller s reagent. Electrochemical polarization measurements were performed on samples ground to 1200 grit in ethanol. 1 M NaCl solution was deareated with Ar gas to decrease the corrosion potential and allow for clear observation of the breakdown potentials. Potentiodynamic scans were performed at a rate of 0.1 mv/s. A Pt counter electrode and saturated calomel reference (SCE) were used. All potentials in this paper are referenced vs. SCE. The Exfoliation of Slices in Humidity (ESH) test was used to measure EFC kinetics. Slices of Al alloy plates prepared as described above were electrochemicallypretreated and then exposed to a high relative humidity environment where exfoliation occurred. In earlier work, a potentiostatic pretreatment in chloride solution was used to 104

125 initiate corrosion attack in the samples.[17] The applied potential was a relatively high value above the second breakdown potential. However, it was subsequently found that the results of the ESH were sometimes quite scattered using a potentiostatic pretreatment because this method can create different amounts of pre-corrosion if the samples have slightly different electrochemical properties. Figure 4.2 shows the current measured during a potentiostatic experiment at -710 mv SCE in 1 M NaCl on two nominally identical AA7178 samples with the same exposed area. The current measured, the charge passed, and thus the damage created, were different for the two samples. In this work, a galvanostatic pretreatment was used to obtain more reproducible results. The pretreatment current density was selected based on the potentiodynamic polarization curve measured on AA7178-T6 in 1 M NaCl, Figure 4.3. The polarization curve of AA7178-T6 sample exhibited two breakdown potentials, and the second breakdown potential was found to be about -722 to -730 mv SCE. A nominal current density associated with a potential just above this breakdown potential was desired to get sustained intergranular corrosion.[18] A current density of 5.5 ma/cm 2 was selected, as it results in a potential between 0.69 and 0.72 V SCE. The electrochemical pretreatment for the ESH test was thus 7 hours in 1 M NaCl at 5.5 ma/cm 2. Following the pretreatment, the sample was rinsed with DI water, dried with a stream of air and placed in a humidity chamber, consisting of a sealed beaker containing a saturated salt solution at room temperature (22-25C). Several salts were used to create a range of constant humidity: sodium sulfate (K 2 SO 4 ), ammonium chloride (NH 4 Cl), potassium iodide (KI), and magnesium nitrate (Mg(NO 3 ) 2 ) for creating constant relative humidity of 96%, 76%, 66%, 56%, respectively. During RH exposure, digital 105

126 photographs were taken of the T face of the samples. The EFC started at the outer edges of the samples, corresponding to the original plate faces, and moved inward. The contrast at the boundary of the outer exfoliated and inner unattacked regions was sufficient to allow tracking of the EFC kinetics by analysis of digital photographs of the sample taken through the glass walls of the humidity chamber. The EFC kinetics can be determined by measuring the width change of central unattacked region of the sample. In this work, the width of the unattacked region was measured at 30 evenly distributed positions along the long axis of the sample. From these measurements, the average width change as a function of time were determined from d 0j d ij = d ij, i = 1,2,3 days, j = 1,2, 30 positions. The mean value is calculated by: 30 j= 1 d ij d i =, i = 1,2,3... days. 30 d i Since the exfoliation occurred at two surfaces, one half of the width change 2 represents the average amount of material consumed on the two sides. X-ray photoelectron spectrometry measurements were also performed on the corroded transverse surface of a sample exposed to 56%RH after electrochemical pretreatment using an AXIS Ultra spectrometer controlled by a VISON data system. A monochromatic Al Ka X-ray line with energy of 1300 ev and 130 W was used as the incident radiation source. The binding energies of the measurements of interest were calibrated with respect to the C 1s line at ev. 106

127 4.2.1 Four Point Bending Technique The deflection of specimens under four point loading conditions is a function of the dimensions of the sample and stressing jig.[8] Figure 4.4 is a schematic figure of a symmetrically loaded specimen supporting four concentrated loads P. The geometry of the specimen is represented in the figure. From elastic theory, the deflection δ under four point loading between two loading points A and B is given by [19]: Pa 2 2 δ = (3Lx 3x a ) (1) 6EI where δ is the deflection in the y direction, x is the distance from the support O ( a x L a ), L is the separation of the stressing points on the tensile side and a is the separation distance of the opposing loading points (Figure 4.4). The maximum normal stress in the specimen is given by h M ( ) y Pah σ 2 max = = I 2 (2) I where M y is the bending moment in the y direction; I is the moment of inertia, and h is the width of the specimen.[19] Since the bending moment between the two loading points A and B is constant, the maximum normal stress is also constant in this region. Considering the deflection at point A or B, where x=a or x=l-a, equation (1) becomes: Pa 2 δ = (3La 4a ) (3) 6EI Solving equation (2) for Pa/I and substituting into equation (3), the relationship between deflection and maximum stress is obtained: max 2 δ = σ (3La 4a ) (4) 3hE 107

128 Assuming elastic behavior and substituting σ max = Eε into equation (4), the initial max outer fiber strain of the specimen, ε max, can be determined from the specimen geometry and deflection in y direction, δ: 3δh ε max = (5) 2 3La 4a A four-point bend test frame was constructed from stainless steel as shown in figure 4.5. Ceramic rods were used to apply stress to the sample, and a dial gauge was attached to the screw used to adjust the separation of the load frame plates. One revolution of the screw resulted in a 1.27 mm displacement of the plates and the accuracy of the displacement based on the dial gauge markings, was 1 or 3.53 µm. The samples were cut from the AA7178 plate so that the length (about 60 mm) was along the L or rolling direction, the width was the full thickness of the plate in the S or through thickness direction (about 7.5 mm), and the thickness was 2 mm in the transverse direction. The surfaces of the original plate were coated, one side with gray paint and the other side with a chromate conversion coating. These coatings were removed from the samples by light polishing prior to the test, and the identities of the two sides of the original plate were tracked through the tests. The other surfaces were ground in alcohol to 1200 grit. The samples were given a galvanostatic pretreatment as described above, rinsed with DI water, and placed in the frame so that the bending axis was along the short transverse axis of the plate. The bending stress was applied by displacing the rod support plates by an amount of δ using the dial gauge as a monitor. The maximum strain, ±ε max, was determined from equation (5) where h was approximately 7.5 mm, L was 63.5 mm, and a = (L L )/2 where L, the separation of stressing points on the compressive side, 108

129 was mm, so a = mm. Previously reported values for the modulus of elasticity, E, and yield strength, σ Y, in compression and tension for AA7178-T6 were used: 71 GPa, 530 MPa, and 540 MPa, respectively.[20, 21] Using Hooke s Law, the yield strain in compression and tension was calculated as and , respectively. The applied maximum strains were selected to be ±0.004, ±0.005, and ±0.006 to stay in the elastic region. The loaded frame was placed in a 96% RH chamber so that the sample length was vertical, similar to the unstressed ESH tests, as shown in figure 4.6. Digital photographs of the T face were taken every other day through the wall of the glass chamber. Because one side of the stressed sample was in tension and the other side was in compression, they exfoliated at different rates. It was therefore necessary to distinguish the EFC rate on each side, rather than use the average rate calculated from the change in width of the inner unattacked region. The sample was surrounded by mm scale during the RH exposure. The inward movement of the boundary between exfoliated region on each side and central unattacked region could be tracked by the horizontal scale giving the exfoliation rate on each side. 4.3 RESULTS AND DISCUSSION Effect of RH on EFC Kinetics Samples cut from the as-received AA7178 wingskin plate were ESH tested in a range of RH. Figure 4.7 shows the corrosion morphology after the electrochemical pretreatment at 5.5mA/cm 2 in 1 M NaCl for 7 hours. The corrosion morphology appears to be a combination of IGC and selective grain attack.[22] To check the depth of attack, 109

130 the same sample was cross-sectioned using a focused ion beam (FIB) tool. As shown in figure 4.8, intergranular attack can be seen in the underlying region in the FIB-etched section. Figure 4.9 shows photographs of samples in 56, 66, 76.6, and 96 %RH on the 11th day of exposure. There is a clear effect of RH on EFC behavior. Below 56% RH, samples did not exfoliate at all and the surface remained shiny and unattacked. Figure 4.10 shows a sample exposed to 30%RH at room temperature for 13 days; no change at all was observed. The sample exposed at 56 %RH did not exhibit sustained EFC. Only a small amount of corrosion product was observed on the sample surface. The sample exhibited some initiation of EFC at the top edge during the first several days of exposure, but the exfoliation stopped growing and did not progress for the rest of the exposure time. The critical RH for EFC appears to be close to 56%. As the RH increased above this critical humidity, the extent of EFC increased and more corrosion product was found to decorate the transverse surface. The width change of central unexfoliated region was determined as a function of time in various constant humidity environments, and the average exfoliation depth, equal to half of the width change, is shown in Figure Replicate tests were performed for each humidity, and the average values from each test are shown as symbols. The results were rather reproducible for samples pretreated galvanostatically. The lines in Figure 4.11 represent average values from the replicate runs. Two sets of error bars are shown in Figure 4.11, representing the standard error for each measurement and the standard error in the whole population. Clearly, the EFC rate increased with RH. 110

131 Prior work on other AA7178 plates indicated that the rate of EFC, given by the slope of the EFC depth curve, did not always vary smoothly.[17] In those plates, the rate increased and decreased with time as the EFC propagated through regions of the plates with varying susceptibility. The plate used in the current study seems to be more uniform, and the rate of EFC was rather constant as the attack proceeded into the plate. ESH tests in 96%, 76%, 66% and 56%RH were continued until 100 days of exposure. The samples were removed from the humidity chambers and the sample surfaces (T face) were observed by SEM. For the sample exposed to 56%RH, most corrosion sites formed during the pretreatment were free from corrosion product. However, the sample exposed to 66%RH clearly shows corrosion product exuding out of the initiation sites, Figure 4.12a, and the amount of corrosion product on the sample surface increased with increasing RH. The sample exposed to 76%RH exhibited a high density of corrosion product and the surface was rather rough in appearance, Figure 4.12b. The sample exposed to 96 % RH exhibited fewer but larger corrosion product aggregates that sat on top of what appears to be a uniform layer of corrosion product as shown in Figure 4.12c. EDS (shown in figure 4.13) and XPS analysis of an AA7178 sample exposed to 56%RH (shown in figure 4.14) shows that the corrosion product consisted of Al, O and Cl. Note that the sample was rinsed with DI water after pretreatment, so the Cl was not a remnant of the NaCl pretreatment solution. The EDS and XPS results indicate that the corrosion product formed during pretreatment, or dried to form hydrated aluminum chloride or aluminum oxychloride. The detailed results obtained from analysis of the XPS spectra are listed in table 4.2. The critical humidity for partially hydrolyzed 111

132 aluminum chloride Al(OH) 3-x Clx has been reported to be in the range between 40-95%RH, depending on the degree of hydrolysis.[23] This is consistent with our observation that exfoliation corrosion only propagated at humidity above 56%RH. At lower humidities, the localized corrosion environment generated by the pretreatment completely dried, preventing further attack. In real exposure environments, the RH is not constant with time and might vary above and below the critical humidity. It is of interest to know if EFC will continue to propagate when a sample is moved from a high to a low RH environment and, if after drying out at low RH, it can rehydrate and resume EFC at the same rate as exhibited previously. To study these factors, samples were exposed to cyclic RH conditions. Figure 4.15 shows the results of a cyclic RH ESH test where the RH was switched between 96% and 56% by quickly moving the sample from one chamber to another. EFC initiated as usual during the first exposure to the 96%RH environment. After 10 days, the sample was moved to 56%RH. The exfoliation continued for one or two days and then almost stopped growing for the rest of exposure period of 20 days. When the sample was moved back to high humidity again, exfoliation resumed very quickly and developed with the same kinetics as in a continuous 96% RH exposure test. In fact, during the test shown in figure 4.15, the sample was exposed to 96% RH for a total of 34.5 days and the total amount of EFC, about 0.25 mm, is equal to the amount of EFC developed during an exposure to a constant RH of 96% for that same period. The extent of EFC was controlled by the total exposure time at high RH. This indicates that the EFC environment quickly equilibrates with the exterior environment, which is very different from the behavior of sharp intergranular corrosion fissures that form in this alloy under 112

133 similar conditions.[22] The speed of the equilibration of the local and exterior environments might be partly the result of the configuration of the samples. The throughplate slices exposed the entire transverse surface to the environment, whereas a real component is primarily exposed at the S surface. Nonetheless, it is interesting that the corrosion product in the EFC cracks responsible for creating the wedging stress that propagates attack can dehydrate and rehydrate quickly to create similar conditions resulting in similar rates of EFC Effect of Temper on EFC Kinetics Metallographic cross sections show that the microstructures of various sections of AA7178, i.e. the grain size and shape, are similar for the T6 and T7 temper and there was no recrystallization during heat treatment. Figure 4.16 shows digital photographs of T6 and T7 tempers on the 6 th, 11 th and 17 th day of exposure in 96%RH. Less corrosion product is evident on the T7 sample surface than on the T6 sample. The width change of the unattacked region of these samples was determined, and the EFC rate for T7 was found to be lower than for T6, figure The kinetics of the T7 temper in 96%RH was a little faster than for T6 temper in 76%RH. These ESH test results are consistent with reported EXCO test results, which also indicated that the T6 temper is more susceptible to EFC than T7.[13-15] Effect of Stress on EFC Kinetics A sample mounted in the four point bending jig as described above was exposed in 96%RH for a month. Figure 4.18 shows a stressed sample with applied maximum strain 113

134 of Within days, exfoliation was apparent on both sides of the T face of the sample. The images in Figure 4.18 are oriented such that the tensile side is on the left and the compressive side is on the right. It is clear that the extent of EFC on the compressive side was greater than on the tensile side, and this is supported by a detailed analysis of the images, Figure The exfoliation rate at the compressive side increased with increasing strain, while the exfoliation rate at the tensile side was not affected much by tensile stress. The total amount of exfoliation on both sides, as measured by the total width change of the unexfoliated region, is shown in Figure The total amount of EFC was similar for all applied strains, and was similar to the unstressed case. Unstressed ESH tests on samples from the same plate showed that the material exfoliated similarly on both sides, indicating that the two sides of the sample have the same EFC susceptibility. Therefore, the different exfoliation rates on the two sides result from the effects of the applied strain (and stress). The acceleration of exfoliation on the compression side was more or less balanced by the decreased exfoliation on the tension side so that the total exfoliation rate on a bent sample was independent of strain. It is known that the wedging stresses introduced by insoluble corrosion product at the grain boundary can provide the energy required for lifting the surface gains and creating a layered attack.[24-26] Due to the Poisson effect, the compressive side of the 4-point bend specimen experienced lateral expansion normal to the direction of compression, which helped push up the surface grains, consequently accelerating the exfoliation corrosion rate.[19] Conversely, the tensile side of the specimen experienced lateral contraction normal to the direction of the tension, which counteracted the wedging effect of the corrosion product, resulting in lower exfoliation kinetics. 114

135 4.3.4 Empirical Model for Exfoliation Corrosion Accelerated corrosion testing helps corrosion engineers to obtain information about corrosion susceptibility of materials within a compressed time. To estimate the corrosion kinetics at normal stress levels based on the accelerated life testing data, extrapolation is required. (Note that the term stress here is used generically to represent all parameters that might accelerate degradation, including mechanical, environmental and thermal factors.) In accelerated life testing, the nominal life-time is often related to stress levels by an acceleration equation that predicts the service lifetime of the material or component based on the lifetime under stress and the level of the stress.[27] In the following, the effects of time, RH, and mechanical stress will be considered in turn. It was shown that EFC would propagate above a critical humidity of 56%RH. The high humidity assures that a wetted film forms and remains on the aluminum surfaces throughout the test period. As shown in figure 4.11, ignoring the data for the first few days, the EFC kinetics were found to be approximately constant with time, so the depth increased linearly with time. The first days seem to be a transition period similar to that shown in the cyclic test experiments in figure Therefore, the ESH data following the transition period were fitted to d = Kt, resulting in K values of 0.2, 1.2, and 6.2 µm/day for 66, 76 and 96% RH, respectively. An empirical model including RH effects can be developed using the Eyring model of acceleration. The Eyring model has been utilized to model acceleration process in many fields[27-32] as this model has a solid theoretical foundation in chemistry and quantum mechanics.[28] When thermal stress (temperature) is the main acceleration variable in the 115

136 process, an Eyring relationship is often considered. The Eyring model also can be used for stress variables other than temperature such as relatively humidity and mechanical stresses. Considering temperature, T, and one additional non-thermal stress, S 1, the time to failure, t f, takes the general form[33]: α H C t f = AT exp{ + ( B + ) S1} kt T (1) where A, B, C and α are constants, H is the enthalpy change, and k is Boltzmann s constant. Additional stresses can be added into this equation with more terms in the exponential. In this work, temperature was assumed to be constant during humidity exposure. H So exp( ) kt can be included into parameter A. The Eyring model can be simplified to the following form, including the time dependence: d = A t exp(b RH) (2) This exponential dependence on RH has also been applied to predict the failure rate of integrated circuits.[34] Fitting the ESH results to this equation gives the following values and standard deviations of the constants A and B for d in mm, RH in fractions, and t in days: A = 1.48E-05 mm/day (σ = 2.96 E-06), B = 6.44 (σ = 0.21) A fitted surface representing Eqn 2 is shown in Figure 4.21 along with the ESH data. Mechanical stress effects may be included in the relationship by assuming that A in Eqn 2 is stress dependent. The ESH results show that compressive and tensile stresses have different effects on exfoliation corrosion kinetics. Compression accelerated the 116

137 EFC kinetics, whereas tension decreased EFC kinetics and was independent of the magnitude of tension. A conditional function can be constructed to describe the different effects of stress. It is expected that corrosion rate under different stress conditions should also follow a linear relationship with respect to time for long exposure period. The conditional function f (ε ) is constructed as follows: C exp( D ) when ε < 0 (Compression) 1 ε A = f (ε ) = C 2 when ε = 0 (Unstressed condition) C 3 exp( E) when ε > 0 (Tension) where C i, D, and E are constants. Combining this function into equation 2: d = f (ε ) t exp(b RH) (3) When ε = 0, the sample is in the unstressed condition. For the unstressed condition, as shown above, C 2 = A = 1.48E-05 mm/day and B = When ε < 0, the sample is under compression and Equation 3 becomes: d = C t exp( B RH + D ) (4) 1 ε In our work, ESH tests under stress were carried out only at 96%RH, so RH=0.96. B has same value as when ε = 0, which is Fitting the ESH data to this equation with ε in fractional units results in C 1 = 0.88E-05 mm/day (σ = 0.29E-05) and D = 0.149E+03 (σ = 0.6E+02). Since the 4-point bend ESH tests were only performed in one RH, the fitted parameters might not be accurate at other RH values. However, we can still get general information of corrosion kinetics from this model. When ε > 0, the sample was under tensile stress, the corrosion kinetics were found to be independent of stress levels. As 117

138 shown in figure 4.22, the average value under tension follows: d = t. Equation (3) becomes: d = C t exp( B RH ) (5) 3 E When RH = 0.96, B = 6.44, so: d = t exp(e') (6) where E =( ln C E). Fitting the average values under tensile stress into the model, we get E = 3.9 (σ=0.014). Again for accurate values of the constants, tests should be performed at different RH values. 4.4 CONCLUSIONS Exfoliation corrosion (EFC) of AA7178 was studied using the exfoliation of slices in humidity test, in which a through-thickness slice of a plate is pretreated electrochemically and then exposed in a high humidity. The followings were found: 1. EFC rate increased with humidity above the critical humidity of around 56% and followed a linear relationship with time. 2. The effect of temper was quantitatively measured, and the EFC rate of AA7178- T6 was found to be higher than that of AA7178-T7. 3. The effects of stress were determined using a 4-point bend apparatus. The EFC rate increased with increasing compressive stress. Tension decreased the rate of EFC, independent of the stress level over the range studied. 4. An empirical model was developed to describe the combined effects of RH and stress. The model is more accurate at longer time and higher humidity. 118

139 REFERENCES 1 M. Posada, L.E. Murr, C.-S. Niou, D. Roberson, D. Little, R. Arrowood, and D. George, Materials Characterization, 38, (1997). 2 J.P. Chubb, T.A. Morad, B.S. Hockenhull, and J.W. Bristow, International Journal of Fatigue, 17(1), (1995). 3 K. Ebtehaj, D. Hardie, and R.N. Parkins, British Corrosion Journal, 24(3), (1989). 4 M.J. Robinson and N.C. Jackson, Corrosion Science, 41, (1999). 5 M.J. Robinson and N.C. Jackson, British Corrosion Journal, 34(1), (1999). 6 D.O. Sprowls, T.J. Summerson, and F.E. Loftin, "Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys-Interim Report on Atmospheric Exposure Tests", in American Society for Testing and Materials, Philadelphia, PA: (1973). 7 T.J. Summerson and D.O. Sprowls, Corrosion Behavior of Aluminum Alloys, in International Conference in Celebration of the Centennial of the Hall-Heroult Process, 1986, University of Virginia, Charlottesville,Virginia. 8 E.A.G. Liddiard, J.A. Whittaker, and H.K. Farmery, Journal of the Institute of Metals, 89, ( ). 9 B.W. Lifka and D.O. Sprowls, Corrosion, 22, 7-15 (1966). 10 W.A. Bell and H.S. Campbell, Journal of the Institute of Metals, 89, ( ). 11 H.B. Romans, Materials Research & Standards,, (1969). 12 D.J. Kelly and M.J. Robinson, Corrosion, 49(10), (1993). 13 B.W. Lifka and D.O. Sprowls, "Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Series Aluminum Alloys with Exposure to a Seacoast Atmosphere", in ASTM SPEICIAL TECHNICAL PUBLUCATION 558, (1973). 119

140 14 S. Lee and B.W. Lifka, "Modification of the EXCO Test Method for Exfoliation Corrosion Susceptibility in 7XXX, 2XXX and Aluminum-Lithium Alloys", in New Methods for Corrosion Testing of Aluminum Alloys, ASTM STP 1134, eds., V.S.Agarwala and G.M.Ugiansky, Philadelplia: American Society for Testing and Materials, 1-19 (1992). 15 D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, "Simplified Exfoliation Testing of Aluminum Alloys", in Localized Corrosion-Cause of Metal Failure, ASTM STP 516, American Society for Testing and Materials, (1972). 16 M.O. Speidel and M.V. Hyatt, "Stress-Corrosion Cracking of High-Strength Aluminum Alloys", in Advances in Corrosion Science and Technology, eds., M.G. Fontana and R.W. Staehle, New York-London: Plenum Press, p.118 (1972). 17 X. Zhao, Chapter 3 of this dissertation. 18 Q. Meng and G.S. Frankel, Journal of The Electrochemical Society, 151(5), B271-B283 (2004). 19 J.M. Gere and S.P. Timoshenko, "Mechanics of Materials", Fourth Edition, Boston, MA: PWS Publishing Company (1997). 20 J.M. Holt, H. Mindlin, and C.Y. Ho, "Structural Alloys Handbook", West Lafayette,IN:CINDAS/Purdue University,c J.R. Davis, Metals Handbook. Materials Park, OH: ASM International, (1998). 22 T. Huang and G.S. Frankel, "Sharp Intergranular Corrosion Fissures in AA7178", Submitted to Corrosion Science, 11/ N. LeBozec, D. Persson, A. Nazarov, and D. Thierry, Journal of Electrochemical Society, 149(9), B403-B408 (2002). 24 H.W. Pickering, F.H. Beck, and M.G. Fontana, Corrosion, 18, (1962). 25 M.J. Robinson, Corrosion Science, 22(8), (1982). 26 M.J. Robinson, Corrosion Science, 23(8), (1983). 27 P. Qiu, The Indian Journal of Statistics, 62, (2000). 28 S. Glasstone, K.J. Laidler, and H. Eyring, "Electrochemical Processes", in The Theory of Rate Processes, Frick Chemical Laboratory, Princeton University: McGRAW-HILL BOOK COMPANY, Inc., (1941). 29 G.C. Montanari and A. Motori, Journal of Physics D: Applied Physics, 28, (1995). 120

141 30 G.C. Montanari and A. Motori, Journal of Applied Physics D: Applied Physics, 24, (1991). 31 G. Mazzanti, G.C. Montanari, and A. Motori, Journal of Applied Physics D: Applied Physics, 27, (1994). 32 L. Bitman, Y.T. Choi, S.B. Choi, and N.M. Wereley, Smart Materials and Microstructures, 14, (2005). 33 P. Tobias, "Assesing Product Reliability", in NIST/SEMATECH e-handbook of Statistical Methods, Statistical Engineering Division, 34 B. Reich and E.B. Hakim, Solid State Technology, September, (1972). 121

142 TABLES AND FIGURES Alloy Si Fe Cu Mn Mg Cr Zn Ti Al AA Bal. Table 4. 1 Chemical composition of AA7178. Element Compound Oxidation BE (ev) FWHM % State (ev) Al 2p Al 2 O 3 /Al(OH) 3 Al Al 2p Unknown compound, Al perhaps Al (OH) 3-x Cl x Al 2p Al(OH) 3 Al O 1s Al(OH) 3 Al O 1s Al 2 O 3 /Al(OH) 3 O 2- /OH Cl 2p 3/2 Cl exits in two different Cl Cl 2p 3/2 chemical states Cl Table 4. 2 Data obtained from XPS spectra. 122

143 Figure 4.1 Heat treatment of AA7178 wingskin plate Current (A) Time (h) Figure 4.2 Current response of two AA7178 samples during potentiostatic pretreatment at -710 mv SCE in 1 M NaCl. Both samples had the same exposed area. 123

144 Potential (mv SCE) Current Density (A/cm 2 ) Figure 4.3 Polarization curve of AA7178 S sample in deareated 1.0 M NaCl at a scan rate of 0.1 mv/s. Figure 4.4 Four point bending mode. 124

145 Figure 4.5 Four point bending jig. Figure 4.6 Four point bending test set up. 125

146 Figure 4.7 SEM image of the surface of AA7178 sample after electrochemical pretreatment at 5.5 ma/cm 2 in 1M NaCl for 7 hours. Figure 4.8 SEM micrograph of FIB cross section of IGC attack in the same AA7178 sample as in figure

147 Figure 4.9 AA7178 wingskin slices after pretreatment and then 11 days in different constant humidity. The rolling direction is vertical. 127

148 Figure 4.10 AA7178 wingskin slices after pretreatment and then 13 days in 30% constant humidity. The rolling direction is vertical. 128

149 Half of Width Change of Unattacked Region (mm) %RH 76%RH 66%RH Standard error in single measurement Standard error in whole population Time (days) Figure 4.11 Width change of central unexfoliated region as a function of time for AA7178 wingskin in various constant RH environments. 129

150 (a) (b) (c) Figure 4.12 (a) SEM micrographs of corrosion product of AA7178 during 66%RH exposure following electrochemical pretreatment. (b) SEM micrographs of corrosion product of AA7178 during 76%RH exposure following electrochemical pretreatment. (c) SEM micrographs of corrosion product of AA7178 during 96%RH exposure following electrochemical pretreatment. 130

151 Al Counts O 500 Cl X-Ray Energy (KeV) Figure 4.13 X-ray EDS spectrograms of corrosion product in 56%RH. Accelerating voltage=12kv. 131

152 Al 2p Cl p 3/2 Intensity (Counts) Intensity (Counts) O 1s 700 Binding Energy (ev) (a) Binding Energy (ev) (b) 6500 Zn 2p Intensity (Counts) Intensity (Counts) Intensity (Counts) Binding Energy (ev) (c) Cu p 3/ Cu p 1/ Bending Energy (ev) Mg Binding Energy (ev) (d) Binding Energy (ev) (e) Binding Energy (ev) (f) Figure 4.14 XPS spectra measured from samples of AA7178 exposed to 56%RH after electrochemical pretreatment in 1 M NaCl for 7 hours. (a) Al 2p, (b) Cl 2p, (c) O 1s, (d) Zn 2p, (e)cu 2p, (f) Mg 2p. 132

153 Half Width Change of Central Unattacked Region (mm) %RH 56%RH 96%RH 56%RH 96%RH 96%RH Time (days) Figure 4.15 Results of cyclic exfoliation test for AA7178 wingskin samples. 133

154 (a) (b) Figure 4.16 Image of AA7178 sample exposed to 96%humidity following galvanostatic pretreatment. (a) T6 temper. (b) T7 temper. 134

155 Half of Width Change of Unattacked Region (mm) T6, 96% RH T7, 96% RH T6, 76% RH Time (days) Figure 4.17 Exfoliation corrosion kinetics of AA7178 samples exposed to 96% RH. 135

156 Figure 4.18 Image of AA7178 sample under four point bending loading exposed to 96%humidity. 136

157 EFC Depth (mm) compression unstressed Time (days) tension Figure 4.19 Exfoliation corrosion kinetics under compression and tension. 137

158 Sum of EFC Extent of Both Sides (mm) Time (days) unstressed Figure 4.20 Exfoliation corrosion kinetics under different strain conditions. 138

159 Figure 4.21 Predictive 3D model for EFC kinetics. Solid square: ESH data in 96%RH; Solid circle: ESH data in 76%RH; Solid triangle: ESH data in 65%RH. 139

160 0.12 EFC Depth on Tension Side (mm) Time (days) d= t Figure 4.22 Average kinetics of exfoliation corrosion in 96%RH under tension. 140

161 CHAPTER 5 IN SITU X-RAY RADIOGRAPHY OF LOCALIZED CORROSION 5.1 INTRODUCTION Intergranular corrosion (IGC) and exfoliation corrosion (EFC) are often found in the high strength Al alloys (e.g. AA2024 and AA7178) used in wingskins and fuselages of airplanes. IGC and EFC are significant sources of life limiting degradation in airframes. A great deal of effort has been devoted over the years to the study of intergranular and exfoliation corrosion of aluminum alloys. The foil penetration technique recently has been used to quantify IGC growth rate.[1] When performed on foils with a range of thickness, foil penetration experiments determine the growth kinetics of the fastest growing localized corrosion sites.[2, 3] However, there are many other sites in such samples that do not penetrate, and those sites have a set of different growth kinetics. It is possible that many sites simply grow slower than the fastest sites. Others might repassivate and stopped growing before they penetrate the foil. Still other sites might have started growing after the initiation of the experiment. The latter two types of sites could have had faster growth kinetics than what is determined by the foil penetration 141

162 technique. It is clearly of interest to obtain information regarding the full population of corrosion sites, not just the fastest growing sites. Exfoliation is a form of intergranular corrosion that can occur on the surface of wrought aluminum alloys with elongated grain structure.[4] However, few reports of quantitative measurements of the kinetics of exfoliation in Al alloys exit, and little is known about how EFC develops under a wide range of environmental conditions.[5-7] Predictive models of the effects of corrosion on the structural integrity of aircraft require information and prediction of the growth kinetics of the various forms of localized corrosion. X-ray radiography technology is one approach for providing this information. The essence of radiography is modulation of the x-ray intensity due to discontinuities or thickness change caused by corrosion.[8] Owing to the relatively low density of water, x-ray images can be obtained on metal samples in situ, i.e. immersed in solution. Radiography can of course be performed in air so, unlike electrochemical techniques, it can be used on samples exposed to atmospheric conditions, which are more related to the environment of aircraft components. In this work, in situ x-ray radiography was used to address localized corrosion kinetics and morphology of AA2024-T3 and AA7178 in different orientations relative to the rolling direction. By orientating the sample properly, based on the understanding generated in prior work, it is possible to image defects as they are growing down into a sample. 5.2 EXPERIMENTAL The schematic of the developed micro-radiographic experimental setup for micro radiography of corroded samples is shown in figure 5.1. It is located in the Welding 142

163 Engineering Building at OSU. The microfocal radiographic system used in this study had a 225 kv, 5 µm ( in.) x-ray microfocal source. Samples were held with a specially designed chuck that allowed for precise positioning between the x-ray source and x-ray film holder as shown in figure 5.2. The chuck allows for accurate repositioning so that intermittent x-ray radiography can be performed on a sample over an extended period of time. Eastman Kodak AA film was used to reduce the exposure time as it is fast and high contrast. The exposed and processed films were back-illuminated and digitized by illuminating a CCD camera and a Data Translation frame grabber (DT3155).[ 8, 9] Two different Al alloys, AA2024-T3 of thickness 1.9cm and AA mm thick plate removed from the wingskin of a retired KC135 airplane were tested. Their compositions are shown in Table 5.1. Figure 5.3 shows the anisotropic microstructure of the two materials and terminology used for the three different sections. The L section is perpendicular to L direction, which is along the rolling direction. The S section is perpendicular to S direction, which is through the thickness of the plate. Special cells and sample configurations were developed in this study to allow the real time in situ radiography. Pillars of dimension 0.8 mm x 2 mm x 30 mm were cut from AA2024-T3 such that the long dimension of the sample was oriented along the plate rolling direction (longitudinal, L), the intermediate dimension was along the plate through-thickness direction (short transverse, S), and the short dimension was along the plate transverse direction (long transverse, T). This orientation allowed for IGC or EFC growth in the L direction. Some samples were oriented such that their long dimension was along the T direction and the short dimension was along the L direction, allowing IGC or EFC growth in the T direction. Some samples were machined to have a square 2 mm x 2 mm 143

164 cross-section rather than a rectangular section. The long direction of the sample was positioned vertically and sealed through the bottom of a plastic cell with epoxy. Some AA2024-T3 samples were first encased in epoxy, except for the top face, which was polished clean. IGC was then formed by polarization in 1 M NaCl at a potential of 580 mv SCE. The attack was constrained to IGC in the encased samples. Other samples were freely exposed to the solution with no epoxy encasement. Freelyexposed samples did not have the physical constraint of the epoxy, and the attack was in the form of exfoliation rather than IGC. These samples had square cross sections with their long axis orientated either in the L or LT direction. IGC and EFC attack (on epoxy-encased and freely-exposed samples, respectively) were studied by in situ x-ray radiography on samples exposed to a high humidity environment after initiation in chloride solution. These samples were AA7178 pillars cut from 8.9 mm thick plate removed from the wingskin of a retired KC135 airplane. Samples were machined using a band saw or electric discharge machining (EDM) into long pillars with their long axis orientated in either the L or T direction. Some samples had square cross-sections of 2 mm x 2 mm and were exposed freely in the solution to promote EFC. Other samples had cross-sections of 0.8 mm x 2 mm and were encased in epoxy leaving only L face polished and exposed to promote IGC rather than EFC. The sample surface was first ground in alcohol to 800 grit. The cell was the same as was used for the AA2024-T3 samples. A 7 hour potentiostatic pretreatment in 1 M NaCl at -725 and -710 mv SCE was performed on the encased and free standing samples, respectively, to initiate intergranular corrosion. From foil penetration experiments, it is expected that this electrochemical treatment will generate attack about 0.46 mm in the AA

165 samples.[10] Following the pretreatment, the samples were rinsed with DI water and placed in a sealed desiccator containing saturated Na 2 SO 4 solution at room temperature. This saturated solution created an environment with constant RH of 96%. Intermittent x- ray radiography was performed on these samples to obtain information on how IGC and EFC propagated in high humidity. 5.3 RESULTS AND DISCUSSION Before showing the results of the real time radiography, an example of metallographic sectioning results will be given. Figure 5.4 shows an example of a metallographic section for a 3 mm thick cylindrical AA2024-T3 sample with axial orientation in the S direction after a 4 h exposure at -580 mv SCE in 1 M NaCl. The attack was in the L-T plane of the sample, i.e. along the grain boundaries. Samples were exposed for different times and sectioned. The kinetics determined from the longest site in each section were slower than those found for L or T oriented samples determined by the foil penetration technique, which measures the fastest growing site.[1] This proves that metallographic sectioning does not determine the kinetics of the fastest growing sites. Figure 5.5 shows an in situ x-ray radiograph of a 2 mm thick AA2024-T3 sample oriented such that the L direction was vertical. The sample was encased in epoxy with one face, the L face, exposed to the solution. This sample was exposed for 19 h to the same conditions as the last sample. The dark lines in the image are corroded grain boundaries. The depths of several selected growing sites were measured on the films at various times. Analysis of a series of images taken from this sample at varying exposure times resulted in the data shown in figure 5.6. The depth of specific IG sites is shown as 145

166 a function of time in that figure. The rates comprise a band that falls just beneath the line representing an extrapolation of the data from foil penetration experiments.[1] These results show that in situ x-ray radiography provides information on the full range of growth kinetics, including those for the fastest growing sites. It should be noted that the circular features in this image are hydrogen bubbles. Localized corrosion of Al alloys always results in the formation of hydrogen. Aggressive dissolution of the Al alloys for a long period resulted in the formation of a foam in the solution. The low density of the hydrogen bubbles causes them to appear dark, and the bright line along their perimeter is from phase contrast, which highlights the boundary between the bubble and solution.[11-14] Figure 5.7 shows radiographs of a similarly oriented sample with no epoxy encasement exposed under the same conditions. This sample had a square 2x2 mm crosssection. Figure 5.7a shows a radiograph prior to the electrochemical treatment. After 4 hr, the attack is apparent in the radiographs, figures 5.7b and 5.7c. Since the sample was free standing with no epoxy encasement, attack proceeded from both the L face on the top and the free T faces on the sides. The period of exposure was too short to generate EFC, so the attack was IGC in nature. However, the IGC propagated inward from three of the five exposed surfaces of the pillar sample: the top L surface and the two side T surfaces. The orthogonal views in figures 5.7b and 5.7c show the morphology of attack. In figure 5.7b, the sample is oriented so that the x-rays are parallel to the long transverse direction. IGC can be seen as sharp dark lines. The lines are darker at the top of the sample because the IGC penetrated from the top along the whole thickness. The x-rays integrate the attack throughout the whole thickness of the samples. The fact that the IGC 146

167 propagated from three faces is clear in figure 5.7c, which is taken at 90º to figure 5.7b, i.e. with the x-rays parallel to the through-thickness or S orientation of the microstructure. The attack is seen from the top L surface and the two side T surfaces. The front of attack coming from all faces of the sample was evident, and is highlighted in figure 5.7c by the dotted lines. With time, the attack on this free-standing sample developed into EFC. Figure 5.8 shows the radiograph after 22 hours of polarization at -580 mv SCE. The orientation of this image is the same as in figure 5.7b, i.e. with the x-rays oriented along the T direction of the microstructure. The many dark lines indicate that the IGC increased in severity. The higher magnification image in figure 5.8b shows that grains are delaminating at the edge of the sample, which is characteristic of EFC. Figure 5.9 shows an in situ x-ray radiograph of a 2 mm thick AA7178 sample oriented such that the L direction was vertical. The sample was encased in epoxy with one face, the L face, exposed to the solution. This sample was exposed for 3 h at - 725mV SCE in 1 M NaCl. The growth kinetics of selected corrosion sites along the L direction were measured from in situ radiographs recorded during the whole 7 hr potentiostatic treatment, and it was found that the general law d=at n was obeyed, as shown in figure The value of n was about 0.33 and A was for t hr and d in mm. This is consistent with results from the foil penetration method.[10] Figure 5.11a shows an in situ x-ray radiograph of the same AA7178 sample with long axis oriented in the L direction. This radiograph was taken after 7 h exposure at mv SCE in 1 M NaCl. After 7 hours in 1 M NaCl at 725mV SCE, the sample was rinsed with DI water and placed in 96%RH. Sharp IGC was evident in radiographs after 147

168 several days of exposure in the high humidity and it continued to grow at high rate. Figure 5.11b shows a radiograph of the same sample after 48 days in the high humidity environment. Tight fissures grow several mm into the sample. Figure 5.12 shows SEM image of sharp intergranular fissure on AA7178. The fissure is around 1-2 um wide and filled of corrosion product. Such sharp fissures can also be found in metallographic cross-sections for AA2024-T3 samples after a similar treatment, as shown in figure This sample was encased in epoxy, pretreated in 1 M NaCl at 580 mv SCE for 7 h, and placed in 96 %RH for 10 days. Figure 5.14 and figure 5.15 shows x-ray radiographs for a free-standing AA7178 samples with long direction oriented in L and T direction respectively. After the electrochemical treatment (7 h at -710 mv SCE in 1 M NaCl), IGC is evident, as was found for the free-standing AA2024-T3 sample after short times. This sample was exposed to the 96% RH environment, and rapid EFC is evident. These radiographs clearly visualize the exfoliation corrosion and record the progress of the corrosion. As is shown in the figures, the sample with long axis along the L direction exfoliated much more severe than the sample with long axis along the T direction along. This is due to the higher length to width ratio in L oriented sample than T oriented sample. 5.4 CONCLUSIONS The intergranular nature of the attack in AA2024-T3 generated different radiographs depending on the orientation of the sample and the x-rays relative to the elongated microstructure. IGC appears as broad and diffuse features when looking in the 148

169 S direction, and as focused sharp features in the L or T direction. Novel cells that utilized samples with specific size, shape, and orientation were used to perform in situ, real time radiographic measurements. In situ radiography of such samples generated information about the growth kinetics of a large number of sites with a range of penetration rates, not just the fastest sites. In situ X-ray radiography is a good approach in studying immersed samples exposed to different controlled conditions. It has great advantages in visualizing the corrosion progress and hence makes it possible to quantify the exfoliation corrosion. 149

170 REFERENCES 1 W. Zhang and G.S. Frankel, Electrochemical and Solid-State Letters, 3(6), (2000). 2 F. Hunkeler and H. Bohni, Corrosion, 37(11), (1981). 3 F. Hunkeler and H. Bohni, Corrosion, 40(10), (1984). 4 "Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)", in Annual Book of ASTM Standards, Philadelphia, PA: The American Society for Testing and Materials, (1990). 5 M.J. Robinson and N.C. Jackson, Corrosion Science, 41, (1999). 6 M.J. Robinson, Corrosion Science, 22(8), (1982). 7 D.J. Kelly and M.J. Robinson, Corrosion, 49(10), (1993). 8 B. Zoofan and S.I. Rokhlin, Materials Evaluation, 52(2), (1998). 9 A. Sehgal, G.S. Frankel, B. Zoofan, and S.I. Rokhlin, Journal of The Electrochemical Society, 147(1), (2000). 10 T.S. Huang and G.S. Frankel, Localized Corrosion Growth Kinetics in Al Alloys, in 6th Joint FAA/DoD/NASA Aging Aircraft Conference-Sept , T.J. Davis, D. Gao, T.E. Gureyev, A.W. Stevenson, and S.W. Wilkins, Nature, 373(16), (1995). 150

171 12 S.W. Wilkins, T.E. Gureyev, D. Gao, A. Pogany, and A.W. Stevenson, Nature, 384(28), (1996). 13 R. Fitzgerald, Physics Today, 53, (2000). 14 A. Singirev and I. Snigireva, Review of Scientific Instruments, 66(12), (1995). 151

172 TABLES AND FIGURES Alloy Si Fe Cu Mn Mg Cr Zn Ti Al AA Bal. AA Bal. Table 5. 1 Composition of AA2024 and AA7178 plates 152

173 Figure 5.1 Schematic of the microradiographic system. Figure 5.2 In situ x-ray microradiographic experimental setup 153

174 Figure 5.3 Metallographic sections of AA2024-T3 and AA7178-T6 wingskin, and the terminology used for the three orientations, L: longitudinal, T: long transverse, S: short transverse. Figure 5.4 Metallographic cross section of S oriented 3 mm wide AA2024-T3 cylinder exposed to 1.0 M NaCl at -580 mv SCE for 4 hr. 154

175 250 µm Figure 5.5 In situ x-ray radiograph of L oriented AA2024-T3 sample exposed to 1.0 M NaCl at -580 mv SCE for 19 hr. 155

176 2 Depth (mm) From foil penetration: d = t 1/2 d = t 1/2 slowest growing site Time (h) Figure 5.6 Plot of the depth of various specific sites for L oriented AA2024-T3 sample as a function of time. Also shown is the curve for L oriented foil penetration samples [1]. 156

177 t = 0 t = 4 h 2 mm 90 0 a b c Figure 5.7 Radiographs of 2x2 mm sample of AA2024-T3 exposed to 1.0 M NaCl at 580 mv SCE. Sample not encased in epoxy. (a) Image was taken before exposure; (b,c) images were taken after 4 h. Long axis (vertical orientation of sample) is L direction. 90 image was taken in T direction, 0 image was taken in S direction. 157

178 a b Figure 5.8 Same sample as figure 5.7, taken in T direction. Exfoliation is evident. 158

179 Figure 5.9 In situ x-ray radiograph of L oriented AA7178 sample exposed to 1.0 M NaCl at -725 mv SCE for 3 hrs. 159

180 d=0.34t 1/3 Depth (mm) d=0.23t 1/ Time (hrs) Figure 5.10 Plot of the depth of various specific sites for L oriented AA7178 wingskin sample as a function of time. 160

181 Figure 5.11 Samples were put in 96%RH after electrochemical treatment. Sharp IGC continued along L direction at high rate. (a) X-ray radiograph taken at 0 days. (b) X-ray radiograph taken at 48days. 161

182 Figure 5.12 SEM image of intergranular sharp fissure in AA7178-T6 after 7 hours in 1 M NaCl at a potential of 710mV SCE. 162

183 Figure 5.13 Metallographic cross-section of AA2024-T3 L sample after 10 days in the environment of RH 96%, T=22C~24C. 163

184 Figure 5.14 Samples were put in 96%RH after above electrochemical treatment. Sharp IGC continued along L direction at high rate. (a) 0 day in 96%RH (after 7 hours in 1 M NaCl, -710mV SCE). (b) 6 days in 96%RH. (c) 30 days in 96%RH. 164

185 Figure 5.15 Samples were put in 96%RH after above electrochemical treatment. Sharp IGC continued along T direction at high rate. (a) 0 day in 96%RH (after 7 hours in 1 M NaCl, -710mV SCE). (b) 6 days in 96%RH. (c) 48 days in 96%RH. 165

186 CHAPTER 6 CONCLUSIONS AND FUTURE WORK 6.1 CONCLUSIONS In this work, the exfoliation corroaion (EFC) of high strength Al alloys was studied quantitatively. The effects of alloy temper, microstructure, local chemistry at the grain boundary, relative humidity and applied mechanical stress on EFC behavior was characterized. A new technique: exfoliation of slices in humidity (ESH) was developed and X-ray radiography was utilized to quantify the corrosion behavior under various conditions. A four point bending jig was used to study the effect of the mechanical stress on corrosion behavior. This study provides good methods to quantitatively investigate exfoliation corrosion kinetics. It has also generated a better understanding on how environmental conditions and the alloy microstructure affect the exfoliation corrosion kinetics. The followings are the main findings: 1. The initiation and propagation of exfoliation corrosion strongly depends upon environmental condition, especially relative humidity. EFC of sliced samples of 166

187 AA7178 was not observed below a critical relative humidity of about 56% and the EFC kinetics increased with increasing humidity above this critical humidity. 2. EFC susceptibility was found to depend on grain size and shape as well as grain boundary composition. The more elongated grains or higher grain aspect ratio and bigger grain size result in higher susceptibility to EFC, while small grain aspect ratio will more likely develop pitting attack. Samples with similar grain aspect ratio but different grain boundary composition will show different susceptibility to EFC. In the AA7178 wingskin sample studied, a high susceptibility to EFC was associated with a high Zn content in the precipitate free zone at grain boundaries. 3. The effect of temper on exfoliation corrosion has been qualitatively studied. In this study, a quantitative evaluation of the effect of T6 and T7 temper on AA7178 was performed by utilizing ESH technique. The EFC rate of AA7178-T6 was found to be higher than that of AA7178-T7, which is consistent with EXCO test results. 4. To simulate the stress condition of airplane wingskin, a 4-point bend apparatus was designed. A specimen under four point bending condition is under compression on one side and under tensile stress on the other side. The effects of surface strain on exfoliation corrosion kinetics were determined by placing 4- point bend apparatus in constant humidity chamber. The EFC rate was observed to increase with increasing compressive stress. Tension decreased the rate of EFC, but level of tension stress over the range that was studied has little effect. The summation of exfoliation corrosion of both sides showed that the total 167

188 exfoliation rate on a bent sample was independent of strain because the acceleration effect of compression stress was offset by the opposite effect of tension stress on the other side. 5. An empirical model for EFC propagation was developed by extrapolating the experimental data from ESH tests in order to describe the combined effects of RH and stress. An Eyring relationship was used. The model is more accurate at longer time and higher humidity. 6. Exfoliation corrosion is a particular kind of intergranular corrosion, which is observed on the surface of the wrought materials. To study the correlation of intergranular corrosion and exfoliation corrosion, in-situ X-ray radiography was utilized. The results on AA2024 and AA7178 showed that in situ radiography of rectangular samples generated information about the growth kinetics of a large number of sites with a range of penetration rates, not just the fastest sites that can be only measured by foil penetration technique. In situ X-ray radiography is a good approach in studying samples exposed to different controlled conditions such as immersed condition or humidity exposure conditions. It has great advantages in visualizing the corrosion progress and hence can be used to observe the transition between intergranular corrosion and exfoliation corrosion. 6.2 FUTURE WORK This work has successfully introduced a new technique and methodology to study the exfoliation corrosion kinetics leading to a better understanding of quantitative 168

189 mechanism of exfoliation corrosion. The following procedures may be followed as an extension to this study. 1. In this study, several relative humidity: 30%, 56%, 66%, 76%, 96% were used to investigate the threshold humidity for propagation of EFC in AA7178. The critical relative humidity was determined roughly based on the results of the exposure in these humidity values at room temperature. ESH tests on more RH values at varied temperature are necessary in order to get detailed information of threshold humidity and the effect of relative humidity on EFC. 2. Because of the rapid response of the EFC to RH, the empirical model developed in this study can be used to predict the exfoliation extent for a sample exposed to varying RH conditions. However, the responsiveness of this system to changes in the environment might result from the specific sample geometry and configuration used in the experiment. The full section of the slices was exposed to the environment, whereas a real structure is typically only exposed on the top surface. In future work, different sample geometry should be tested in ESH test to clarify how the geometry will affect the EFC kinetics. 3. In order to obtain a complete predictive model for exfoliation corrosion as a function of stress, more stress level should be used for determining the stress effect on exfoliation corrosion. The exposure time in humidity under stressed condition should be extended in order to get a reliable time vs. kinetics relationship under stressed condition. 169

190 4. This work confirmed the effect of T6 and T7 temper on exfoliation corrosion. The quantitative effect of other heat treatment can be evaluated by using ESH technique. 5. An empirical model of exfoliation corrosion was developed in this study. This model describes a relationship between the exfoliation corrosion kinetics and intergranular sharp fissure growth. To realize the model, a numerical simulation is necessary. Understanding of the development of wedging stress is essential to perform this simulation. 170

191 APPENDIX A FINITE ELEMENT SIMULATION A.1 PHENOMENOLOGICAL MODEL OF EXFOLIATION CORROSION A phenomenological model is proposed to explain EFC and IGC process in this work. Figure A.1 illustrates a 3D rectangular sample after an electrochemical pretreatment: selective grain attacks are created during pretreatment. Subsequent exposure to humidity results in the growth of IGC from these sites. Figure A.2 shows an L cross-section of a specimen exposed in relative humidity illustrating the EFC process. The phenomenological model is developed based on the following assumptions: 1. Selective grain attack along T direction was created during pretreatment (figure A.2a). The depth of attack can be obtained from foil penetration experiments: a(µm) = t(h) 1/3.[1] 2. After pretreatment for time t pre, the unattacked region along T direction has the dimension of 2c 0 =[b-2a(t pre )], where b is the original sample dimension in the T direction. Subsequent exposure to humidity results in the growth of IGC from these sites. IGC 171

192 grows faster at the edge of sample (figure A.2b). We assume that the kinetics of end grain attack is linear: d e = k e t, where d e is the extent of end grain corrosion attack in the T direction and k e is a constant. It is further assumed that all grains other than at the end of the sample exhibit sharp IGC fissure growth at the much slower rate given by the IGC sharp fissure kinetics of constrained samples: d T (µm) = 5.14 t(h) 0.58 (along T direction).[2] When IGC at the two edges growing inwards in the T direction meet each other (figure A.2c), the outermost layer disconnects from the bulk material as exfoliated material shown in A.2d. The second outermost grain layer then becomes the end grain and starts to follow end grain kinetics. The process is repeated for each underlying layer. We assume each exfoliated layer has the thickness of one-grain dimension along the S direction. For the first layer, the distance that end grain sharp fissure travels along T direction is c 0, which is the unattacked thickness right after the pretreatment. The unattacked thickness for each underlying layer decreases with time due to sharp fissure growth and can be calculated by: d ei =c t 0.58, i=1,2,3 n. A.2 MODEL DEVELOPMENT IN ABAQUS A finite element model was developed and analyzed using the commercial finite element program ABAQUS. The meshes consist of node-plane-strain quadrilateral, reduced integration elements. Meshes of a quarter of the L section of ESH sample are shown in figure A.3. The model consisted of 5 cracks opened at one side of the section, each of which has one-grain size in width and 400um in length. In each crack, a U shape rigid body was inserted to simulate the corrosion product. At room temperature, the AA7178 has an elastic modulus E=71.7GPa, Poisson s raio ν=0.33, and 172

193 materials density ρ= 2.8E-9g/cm 3. Stress strain relationship in plastic zone is determined by the Holloman equation and the calculation is shown as follows: For AA7178-T6, yield stress σ y = 538MPa. Ultimate tensile stressσ UTS = 607MPa. Elongation at breaking e f = p n Assume: σ = σ 0 + k ( ε ) for σ σ 0 (a) Letσ = σ 0 y To determine k and n, two conditions were used as following: 1. σ ε p = σ f where σ f is true stress at the necking which is given by f σ f = σ UTS ( 1+ e f ) σ 2. p = σ p ε f f ε For AA7178-T6 at the necking: σ f = 538 ( ) = 597MPa σ e f Elastic true strain: ε = = E f Total true strain: ε ln( 1+ e ) = f = f p e Plastic strain: ε = ε ε = For condition 1, we get: f f f p n k ( ε ) σ σ (1) f = f 0 For condition 2, we get: nk( σ (2) p n ε f ) 1 = f Equation (2) divided by equation (1), we get 173

194 σ f p n = ε f (3) σ σ f 0 Substituting equation (2) into equation (1), we get: k σ f σ 0 = (4) n ( ε ) p f Solving the above equations, we obtain n = 0.88 and k = 523 Mpa. The stress strain relationship can be determined by equation (a). The input stress-strain data is listed in table A.1. During finite element simulation, the boundary condition is semi-infinite condition. The left side and top side of the model was fixed in x and y directions. A force of 0.9N calculated from measured wedging stress was applied on the U-shape, forcing it to move towards the crack tip. This mimics the wedging action of the corrosion product, forcing the crack to open. A.3 RESULTS AND DISCUSSIONS The Von-Mises stress was used to show the stress field. The von-mises stress is a scalar measure of the stress state (the normal and shear stresses) at any point within a body. It is a stress quantity that is proportional to the strain energy density associated with a change in shape (with a zero volume change) at a material point: 1 2 zx σ vm = ( σ x σ y ) + ( σ y σ z ) + ( σ z σ x ) + 6( τ xy + τ yz + τ 2 In figure A.4, a 2D contour plot from finite element simulation shows that high stresses develop around the crack tip at the edge during EFC. The stress around the crack tip is decreasing towards the center of the sample. The Von Mises stress at the crack tip of the 174 )

195 edge can be as high as 500MPa. While the Von Mises stress for the inner crack tip is only between 40-90MPa. The finite element simulation of the stress field during EFC is consistent with the experimental results which showed that unconstrained sample developed exfoliation corrosion. 175

196 REFERENCES 1 T. Huang and G.S. Frankel, "Effect of Temper and Potential on Localized Corrosion Kinetics of AA7075", Submitted to Corrosion, 11/05. 2 T. Huang and G.S. Frankel, "Sharp Intergranular Corrosion Fissures in AA7178", Submitted to Corrosion Science, 11/

197 TABLES AND FIGURES Stress(Mpa) Strain Table A. 1 Stress strain input data. 177

198 Figure A. 1 IGC attacks are created in the sample from electrochemical pretreatment. 178

199 (a) Electrochemical pretreatment creates IGC attacks. (b) IGC continues growing during RH exposure. (c) IGC grows fastest at the edge of sample. (d) Layers of materials exfoliate. Figure A. 2 EFC process of a sample exposed in high relative humidity. (L section) 179

200 Figure A. 3 Two dimensional finite element mesh. 180

201 Figure A. 4 2D contour plot showing stress field around intergranular crack tip simulated by ABAQUS. 181