@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Applied Science, Faculty of"@en, "Materials Engineering, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Pournazari, Shabnam"@en ; dcterms:issued "2018-04-18T19:27:10Z"@en, "2018"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Aluminum-copper casting alloys have relatively high strength and hardness, fatigue and creep resistances and good machinability, all of each are dependent on the copper content of the alloy. The Al-Cu casting alloy (4.2-5.0 wt% Cu), known as B206, is a potential candidate material for use in marine applications where good mechanical properties and high strength to weight ratio is desired. These properties are ideal for components of tidal-based energy generating systems. However, corrosion continues to be an issue. This dissertation presents and discusses the results of several electrochemical and microstructural investigations conducted on B206, contributing to a further understanding of the fundamental corrosion processes. Applications of this research are strongest within the marine industry field, yet are extendable to other infrastructural and engineering applications such as aerospace and military. Results of this work elucidate the mechanism of localized corrosion of B206 alloy in seawater. Focused ion beam (FIB) used to determine the subsurface microstructure at local attack sites within the corroded area reveals that localized corrosion is propagated where continuous particles are buried beneath the surface. Propagating away from the initiation sites, corrosion develops preferentially along the grain boundary network beneath the alloy surface. Retrogression and re-aging (RRA) of the alloy to modify the grain structure and render uniform the distribution of the second phase is revealed not to have a substantial effect on the corrosion susceptibility of the alloy. However, Electrochemical Impedance Spectroscopy (EIS) and Mott-Schottky tests support the feasibility of implementing anodizing and possibly anodic protection systems for B206 in specific service environments. EIS was also used to determine the effect of cathodic protection (CP) on coated B206 and reveals that its corrosion resistance with CP is superior to the situation without CP and, therefore, that the coating is compatible with CP. Due to its use in the as-cast state, the effect of casting porosity on the corrosion of B206 was investigated using a pencil electrode method. Results reveal that the corrosion can be attributed to the local chemistry inside the pores (conductivity and potential at the bottom of pores)."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/65487?expand=metadata"@en ; skos:note "CORROSION BEHAVIOR OF B206 ALUMINUM-COPPER CASTING ALLOY IN SEAWATER ENVIRONMENT: ELECTROCHEMICAL AND MICROSTRUCTURAL STUDIES by Shabnam Pournazari B.Sc., Ferdowsi University of Mashhad, 2008 M.Sc., Ferdowsi University of Mashhad, 2012 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Materials Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2018 © Shabnam Pournazari, 2018 ii Abstract Aluminum-copper casting alloys have relatively high strength and hardness, fatigue and creep resistances and good machinability, all of each are dependent on the copper content of the alloy. The Al-Cu casting alloy (4.2-5.0 wt% Cu), known as B206, is a potential candidate material for use in marine applications where good mechanical properties and high strength to weight ratio is desired. These properties are ideal for components of tidal-based energy generating systems. However, corrosion continues to be an issue. This dissertation presents and discusses the results of several electrochemical and microstructural investigations conducted on B206, contributing to a further understanding of the fundamental corrosion processes. Applications of this research are strongest within the marine industry field, yet are extendable to other infrastructural and engineering applications such as aerospace and military. Results of this work elucidate the mechanism of localized corrosion of B206 alloy in seawater. Focused ion beam (FIB) used to determine the subsurface microstructure at local attack sites within the corroded area reveals that localized corrosion is propagated where continuous particles are buried beneath the surface. Propagating away from the initiation sites, corrosion develops preferentially along the grain boundary network beneath the alloy surface. Retrogression and re-aging (RRA) of the alloy to modify the grain structure and render uniform the distribution of the second phase is revealed not to have a substantial effect on the corrosion susceptibility of the alloy. However, Electrochemical Impedance Spectroscopy (EIS) and Mott-Schottky tests support the feasibility of implementing anodizing and possibly anodic protection iii systems for B206 in specific service environments. EIS was also used to determine the effect of cathodic protection (CP) on coated B206 and reveals that its corrosion resistance with CP is superior to the situation without CP and, therefore, that the coating is compatible with CP. Due to its use in the as-cast state, the effect of casting porosity on the corrosion of B206 was investigated using a pencil electrode method. Results reveal that the corrosion can be attributed to the local chemistry inside the pores (conductivity and potential at the bottom of pores). iv Lay Summary The investigated material throughout this thesis was B206, a high strength aluminum-copper casting alloy. B206 is a potential candidate material to be used as the hub component in hydrokinetic energy generating systems due to its good mechanical properties and high strength to weight ratio. However, the marine environment is very corrosive and corrosion of this material is an issue. At present, data regarding the corrosion properties and protection systems for B206 are not available. Therefore, the motivation behind this project was to first enhance the understanding of the mechanism of localized corrosion for B206 in seawater and then to investigate the feasibility of various corrosion control methods at different environmental conditions. The results of this study contribute toward improving the corrosion resistance and the effectiveness of B206 in practical applications. v Preface The present work was conducted in Materials Engineering Department at the University of British Columbia with financial support received from the Natural Science and Engineering Research Council (NSERC) of Canada. Professors Edouard Asselin and Daan Maijer extensively helped with all aspects of the research work. Transmission Electron Microscopy (TEM) analysis was carried out at McMaster University in Hamilton by staff at the Canadian Centre for Electron Microscopy. These TEM results are presented in Figure 6-2 and Figure 6-6. The results presented in section 8.1 were gathered in collaboration with Defence Research and Development Canada (DRDC) and Dr. Shona McLaughlin. Dr. McLaughlin conducted the immersion experiments in Esquimalt Harbour, measured the mass loss and helped with immersion data analysis. The following journal articles and article in the proceedings of a professional conference have been published from the research work presented in this dissertation. Peer-reviewed Journals: 1. Sh. Pournazari, D. M, Maijer, E. Asselin, “FIB/SEM Study of Pitting and Intergranular Corrosion in an Al-Cu Alloy”, Corrosion 73 (2017), pp. 927-941. 2. Sh. Pournazari, K. M. Deen, D. M. Maijer, E. Asselin, “Effect of retrogression and re-aging (RRA) heat treatment on the corrosion behavior of B206 aluminum-copper cast alloy”, Materials and Corrosion, 69 (2018), pp. 1-18. 3. Sh. Pournazari, I. M. Gadala, A. Alfantazi, D. M. Maijer, E. Asselin, “Influence of dissolved oxygen and applied potential on the corrosion protection of B206 aluminum alloy in 6 – 25 C seawater”, [Under preparation]. 4. Sh. Pournazari, S.R. McLaughlin, D. M. Maijer, E. Asselin, “Investigation of vi cathodic disbondment of an epoxy coating on aluminum under different levels of cathodic protection potential in seawater: an electrochemical impedance spectroscopy study”, [Under preparation]. Peer-reviewed Conference Proceeding: 5. Sh. Pournazari, D.M. Maijer, E. Asselin, “Effects of Dissolved Oxygen and Temperature on the Corrosion Properties of B206 and A356 Aluminum Alloys Exposed to Seawater”, Proceedings of NACE CORROSION, Corrosion Conference and Expo, Paper No. 7561, 2016, pp. 1-9. The following table identifies the specific publication(s) upon which each corresponding chapter of this dissertation is based: Chapter number in dissertation Publication number(s) from the list above 5 1 6 2 7 3* and 5 8 4* *These manuscripts have been prepared and are in the process of submission. vii Table of Contents Abstract ............................................................................................................................... ii Lay Summary ..................................................................................................................... iv Preface ................................................................................................................................. v Table of Contents .............................................................................................................. vii List of Tables ...................................................................................................................... xi List of Figures .................................................................................................................. xiii List of Abbreviations ...................................................................................................... xxiii List of Chemical Formulae .............................................................................................. xxv List of Symbols ............................................................................................................. xxvii Acknowledgements ......................................................................................................... xxx Dedication ..................................................................................................................... xxxii 1 Introduction .................................................................................................................... 1 2 Literature review ............................................................................................................ 7 3 Objectives ...................................................................................................................... 34 viii 4 Approach and methodology ........................................................................................ 36 4.1.1 Heat treatment procedure ............................................................................... 38 4.1.2 Mass loss experiments ................................................................................... 39 4.1.2.1 Laboratory experiments in artificial seawater ......................................... 39 4.1.2.2 Field experiments in natural seawater ..................................................... 40 4.1.3 1D-artificial pit (pencil) electrode preparation .............................................. 41 4.2.1 Corrosion initiation and microstructural evaluation ...................................... 43 4.2.2 Electrochemical experiments and weight loss measurements to compare the corrosion behavior of as-cast and RRA heat-treated B206 alloy .............................. 45 4.2.3 Electrochemical test methods for corrosion and passivation processes ........ 47 4.2.4 Coating and cathodic protection (CP) application ......................................... 48 4.2.5 Electrochemical test methods for pencil electrode experiments ................... 51 5 FIB/SEM study of pitting and intergranular corrosion in B206 aluminum alloy .. 54 6 Effect of retrogression and re-aging (RRA) heat treatment on the corrosion behaviour of B206 ........................................................................................................... 78 ix 7 B206 Passive layer characterization in deaerated artificial seawater at different temperatures using EIS and Mott-Schottky Analysis ................................................ 107 7.2.1 Passive current density decay ...................................................................... 110 7.2.2 Passive layer components and structure ...................................................... 112 7.2.3 EIS ............................................................................................................... 113 7.2.4 Mott-Schottky and passive layer thickness ................................................. 120 8 Investigation of cathodic disbondment of an epoxy coating on B206 substrate under different levels of cathodic protection in seawater .......................................... 126 8.3.1 Samples without an artificial defect ............................................................ 133 8.3.2 Samples with an artificial defect ................................................................. 136 9 The effect of casting porosity on the changes in local chemistry inside the pores 153 10 Conclusion ................................................................................................................. 163 References ....................................................................................................................... 172 Appendices ...................................................................................................................... 190 x xi List of Tables Table 1-1: Mechanical properties of alloy B206 [40] ....................................................... 5 Table 4-1: Chemical composition of the B206 alloy ...................................................... 36 Table 4-2: Thermal process of the investigated B206 alloy ............................................ 39 Table 4-3: Chemical composition of the artificial seawater used in laboratory tests ..... 43 Table 5-1: Chemical composition of the indicated parts of the intermetallic particles in Figure 5-2 ......................................................................................................................... 57 Table 5-2: The changes in the chemical composition of the mentioned particles in Figure 5-1 .................................................................................................................................... 63 Table 6-1: Average values of OCP (vs. Ag/AgCl) for as-cast and RRA alloys .............. 84 Table 6-2: Electrochemical parameters obtained from polarization curves .................... 86 Table 6-3: Impedance parameters data obtained by the ZSimpWin software for RRA B206 alloy at different immersion times in artificial seawater at 10 °C .......................... 95 Table 6-4: Impedance parameters data obtained by the ZSimpWin software for as-cast B206 alloy at different immersion times in artificial seawater at 10 °C .......................... 95 Table 6-5: Corrosion rates of as-cast and RRA B206 obtained from LPR experiments at different times in artificial seawater at 10 °C ................................................................ 100 Table 7-1: Electrochemical parameters extracted from potentiodynamic polarization curves for B206 in naturally aerated and deaerated artificial seawater at 10 °C ........... 109 Table 7-2: EIS component values for B206 anodized at −0.65 V vs. Ag/AgCl in seawater solution at 6, 10, 14 and 25 °C ........................................................................ 116 Table 8-1: The coating capacitance and resistance of the B206 sample under −0.7 VAg/AgCl CP at different exposure time derived from analysis of EIS data measured .... 139 xii Table 8-2: EIS parameters fitted from the response of the coated sample with defect at −0.9 VAg/AgCl ................................................................................................................... 144 Table 8-3: Summary of the obtained electrochemical and microstructural results for the coated B206 alloy with and without an artificial defect at different cathodic potential levels .............................................................................................................................. 152 Table D-1: Volume of the hydrogen gas evolved during the potentiostatic polarization experiments of 1mm B206 pencil electrode at different potentials in artificial seawater at 22 °C............................................................................................................................... 196 xiii List of Figures Figure 1-1: Comparison of the mechanical properties of various foundry Al alloys [40] 5 Figure 2-1: Secondary (left) and Backscattered (right) electron images of (Al,Cu)x(Fe,Mn)ySi intermetallic particle in AA2024-T3 alloy after 120 min exposure to 0.1 M NaCl [20] ................................................................................................................. 9 Figure 2-2: (Top) Backscatter image of an intermetallic particle and its periphery zone and (bottom) corresponding schematic for the backscatter image [46] ........................... 10 Figure 2-3: Schematic dealloying and subsequent trenching of the S-phase in AA2024 alloy [50] .......................................................................................................................... 11 Figure 2-4: Schematic trenching of the intermetallic particle containing noble elements and trenching [50] ............................................................................................................ 12 Figure 2-5: Model of how coupled intermetallic particles contribute to the establishment of the stable pits in AA2024-T3 alloy [52] ...................................................................... 13 Figure 2-6: Al-rich corner of the Al-Cu phase diagram showing the metastable solvus boundaries for GP zones, θ\" and θ', together with the equilibrium solvus line for the phase [54] ......................................................................................................................... 16 Figure 2-7: Instantaneous corrosion rate as evaluated from hydrogen evolution (PH) during immersion testing at the open circuit potential (OCP) in 3.5% NaCl solution saturated with Mg(OH)2 for 7 days at 24 ± 1 °C for solution heat-treated Mg0.1Si specimens [30] ................................................................................................................. 25 Figure 2-8: Cathodic polarization curves measured during the 5th hour and the seventh day during immersion testing at the open circuit potential (OCP) in 3.5% NaCl solution xiv saturated with Mg(OH)2 for 7 days at 24 ± 1 °C for solution heat-treated Mg0.1Si specimens [30] ................................................................................................................. 26 Figure 2-9: Corrosion products filled inside casting porosity on HPDC AM50 after 168 h exposure [109] ............................................................................................................... 28 Figure 2-10: Potential versus current density plots for 316 L generated from linear polarization scans at different pit depths [111] ................................................................ 30 Figure 2-11: Potentiodynamic polarization curves for aluminum in 90% saturation AlCl3 solution ta 1mm depth [117] ............................................................................................ 31 Figure 2-12: Schematic drawing of the artificial crevice cell [120]................................ 32 Figure 4-1: Picture of the mold configuration ................................................................. 37 Figure 4-2: (a) Transverse slice sectioning of the B206 casting and (b) circular samples extracted from each sections ............................................................................................ 38 Figure 4-3: Picture of the samples for the laboratory mass loss experiments ................. 39 Figure 4-4: Configuration of the mass loss experiments samples in natural seawater ... 40 Figure 4-5: Picture of the fixed drill set-up used for grinding the rod B206 samples .... 42 Figure 4-6: Schematic experimental set up used for electrochemical experiments ........ 45 Figure 4-7: B206 coated sample with at artificial defect at its center ............................. 50 Figure 4-8: Schematic of the flat cell used in electrochemical experiments on coated samples with and without defect at OCP and different levels of CP ............................... 51 Figure 4-9: (a and b) Hydrogen collection experiments setup. The electrode surface faced upward. The evolved hydrogen was collected by means of the funnel and the burette above the specimen. The burette was initially full of solution, which was displaced by the evolved hydrogen, allowing easy measurement of the cathodic charge xv associated. (c) Hydrogen bubbles can be seen in the burette above the specimen. (d) Schematic top view of the experimental arrangement shown in (a and b) ...................... 52 Figure 5-1: (a) Optical image, (b) SEM micrograph of B206, (c and d) higher magnification of regions C and A, respectively ............................................................... 56 Figure 5-2: Backscatter image of some multiphase intermetallic particles (indicated by dashed circles) on the B206 surface ................................................................................. 56 Figure 5-3: Top: Backscattered electron image of a multiphase intermetallic particle. Bottom: the x-ray mapping corresponding to the shown intermetallic particle ............... 58 Figure 5-4: The EDX elemental maps show the formation of multiphase particles in the B206. The periphery of the particle is composed more from Mg and Si ......................... 59 Figure 5-5: Backscattered electron images of (a) the B206 surface, (b and c) higher magnification of selected regions (circles) after 24 h exposure to artificial seawater at 10 °C (arrows indicate the signs of localized corrosion on the surface), (d and e) higher magnification of some attacks showed by solid white arrows in b and c ........................ 62 Figure 5-6: The (a) SEM micrograph and (b) EDX elemental map for copper on a Al-Cu-Mg-Si particle in B206 alloy ...................................................................................... 63 Figure 5-7: Pit initiation on the intermetallic particles after 24 hours ............................ 64 Figure 5-8: (a) Potential-time behavior of the B206 in artificial seawater at 10 °C with an applied current of 2 µA, (b) Microstructural analysis of the surface after galvanostatic anodic polarization shown in (a). The pit was nucleated on the intermetallic surface .... 65 Figure 5-9: The overview of the investigated area (a) before immersion, (b) after 24 hours and (c and d) 30 and 48 hours after the first 24 hours, respectively ...................... 67 xvi Figure 5-10: The Backscattered (left) and Secondary electron (right) imaging of the attacked area on the surface 30 (a and b) and 48 (c and d) h of the initiation time ......... 68 Figure 5-11: (a) SEM micrograph and (b-d) EDX maps of Al, Cu and O after 48 hours of the initiation of the pits in seawater solution, respectively .......................................... 69 Figure 5-12: Schematic illustration of matrix de-alloying in the vicinity of an intermetallic particle on the surface ................................................................................. 70 Figure 5-13: (a) FIB section at location A indicated in Figure 5-10, (b) higher magnification SEM image of FIB cross-section trough the intermetallic particle after 48 hours of initiation of the pits in seawater solution (boxed area in (a)) and (c) EDX on the sponge-like morphology ................................................................................................... 71 Figure 5-14: (a) FIB milling and (b-d) SEM micrographs of corroded B206 intermetallic at various stages of milling. SEM image of FIB cross-section shows localized attack of an intermetallic particle cluster below the surface. (f and g) higher magnification of SEM image of c and e, respectively .......................................................................................... 74 Figure 5-15: SEM micrographs of the cross section of the B206 after immersion for 48 more hours after the initiation of the localized corrosion in seawater solution ............... 75 Figure 6-1: SEM and optical images of (a and c) as-cast and (b and d) RRA heat-treated B206 alloy. ....................................................................................................................... 80 Figure 6-2: Microstructural features of RRA sample (a) TEM image showing the presence of precipitates within the α-Al matrix (b) DFSTEM present the distribution of θ \" and θ ' precipitates within the matrix phase (c and d) showing DFSTEM micrographs of intermetallic particles with corresponding points EDX spectra (e) high magnification image showing the presence of round shape precipitate (f) HRTEM image xvii of precipitate illustrating the plane orientation (g) HRTEM of the α-Al matrix and inset showing the indexed SAD pattern of the matrix phase. ................................................... 82 Figure 6-3: Corrosion potential values of as-cast and RRA alloy immersed in artificial seawater at 10 °C .............................................................................................................. 84 Figure 6-4: Potentiodynamic polarization curves at different immersion times for (a) as-cast, (b) RRA and (c) both as-cast and RRA B206 after immersion in seawater solution for 120 h. .......................................................................................................................... 85 Figure 6-5: SEM micrographs of (a) as-cast and (b) RRA aluminum alloy under free corrosion condition, showing the oxide film and preferential dissolution (exposed to seawater solution after 48 hours) ..................................................................................... 86 Figure 6-6: (a) DFSTEM micrographs showing the lath shape precipitate and (b, c, d) the line EDS spectra reveals the composition variation along the precipitate ................. 88 Figure 6-7: (a) Effect of exposure time on EIS Bode plots and (b) fitting of the EIS spectra of RRA B206 ....................................................................................................... 90 Figure 6-8: Effect of exposure time on EIS Bode plots of as-cast B206 and (b) experimental and simulated Nyquist of as-cast B206 after 72h ....................................... 91 Figure 6-9: Equivalent circuit used to fit EIS data (a) RRA and (b) as-cast B206 ......... 92 Figure 6-10: Nyquist diagrams of RRA and as-cast B206 alloy after immersion for (a) 1, (b) 48 and (c) 144h of immersion in artificial seawater at 10 °C ..................................... 96 Figure 6-11: Relationship between Rp and immersion time for as-cast and RRA alloys in artificial seawater at 10 °C. .............................................................................................. 97 Figure 6-12: Pictures of the as-cast and RRA coupons after (top) 24 and (down) 144 hours ................................................................................................................................. 98 xviii Figure 6-13: Mass loss for as-cast and RRA B206 alloy in artificial seawater at 10 °C. Error bars represent the standard deviation for an average of cumulative mass losses of two examined samples ..................................................................................................... 99 Figure 6-14: (a, b and c) SEM micrographs of RRA B206 after 48 hours of immersion in seawater solution at 10 °C .............................................................................................. 102 Figure 6-15: (a, b and c) SEM micrographs of the cross section of the RRA B206 alloy after 48 hours of immersion in seawater solution showing exfoliation corrosion ......... 104 Figure 7-1: Potentiodynamic polarization curves for aluminum B206 in naturally aerated and deaerated artificial seawater at 10 °C ...................................................................... 108 Figure 7-2: Polarization curves for aluminum B206 in different temperatures in (a) deaerated and (b) naturally aerated artificial seawater. .................................................. 109 Figure 7-3: Passive current density (ipass) decay transients measured during the anodizing of B206 at −0.65 VAg/AgCl in deaerated seawater at 6, 10, 14 and 25 °C for 6 hours ............................................................................................................................... 111 Figure 7-4: (a) Nyquist impedance representations and (b) Bode phase angle and impedance moduli of the EIS impedance spectra obtained for anodized B206 specimens in deaerated seawater at temperatures between 6 and 25 °C. ........................................ 114 Figure 7-5: Proposed EEC and physical occurrences for EIS on anodized B206 in deaerated artificial seawater at 6 °C ............................................................................... 115 Figure 7-6: Proposed EEC and physical occurrences for EIS on anodized B206 in deaerated seawater at 10, 14 and 25 C .......................................................................... 119 xix Figure 7-7: Mott-Schottky plots at 1 kHz between −0.8 VAg/AgCl and −0.5 VAg/AgCl after 6 hours of anodizing in deaerated seawater at 6, 10, 14 and 25 °C (linear fits and slope values overlaid) .............................................................................................................. 121 Figure 7-8: Donor density and passive layer thickness values for B206 anodized for 6 hours in deaerated seawater at temperatures between 6 and 25 °C, calculated from Mott-Schottky and EIS results, respectively. .......................................................................... 122 Figure 8-1: B206 corrosion samples following 42 days (80 days total) exposure, before cleaning .......................................................................................................................... 128 Figure 8-2: Pictures of the coupons for after 38, 80 and 129 days (a-c) for B206 ....... 129 Figure 8-3: Cumulative mass loss against time for aluminum B206 in natural seawater. Mass loss from initial is the mass loss data obtained from the initial mass measurements (before any immersion in seawater) and mass loss from previous is the mass loss data obtained from mass measurements before and after each immersion intervals. ............ 130 Figure 8-4: Corrosion attack morphology for aluminum B206 after (a) 38 and (b) 80 days of immersion in natural seawater ........................................................................... 130 Figure 8-5: Cathodic polarization curve of B206 at 10 ºC in seawater ........................ 131 Figure 8-6: Potentiostatic polarization curves of B206 with various applied potentials in seawater solution at 10 ºC .............................................................................................. 133 Figure 8-7: Bode plots obtained from the intact coated sample (a) at OCP, (b) under CP potential (−0.7 VAg/AgCl) and (c) impedance evolution at 0.01 Hz frequency, respectively (at OCP and at −0.7 VAg/AgCl CP) for different exposure periods. ................................. 134 xx Figure 8-8: (a) Bode magnitude, (b) Bode phase angle and (c) Nyquist diagrams collected for defected coated samples after 1 h and 5, 10, 20 and 30 days of exposure to −0.7 VAg/AgCl ................................................................................................................... 137 Figure 8-9: EEC used to fit EIS data for coated samples with defect under (a) −0.7 and (b) −0.9 VAg/AgCl ............................................................................................................. 138 Figure 8-10: Protection current densities as a function of time for the coated sample with an artificial defect polarized to (a) −0.7 V Ag/AgCl and (b) −0.9 VAg/AgCl. ....................... 140 Figure 8-11: (a and b) SEM micrographs indicating some porous corrosion products partially blocked the defect area on the coated surface after 30 days of exposure to seawater solution under −0.7 VAg/AgCl CP ...................................................................... 141 Figure 8-12: (a) Bode magnitude and (b) Bode phase angle collected for defect coated samples after different times of exposure to −0.9 VAg/AgCl ............................................ 141 Figure 8-13: Example of a fitted result for a selected EIS diagram (5 days) ................ 144 Figure 8-14: Corrosion resistance as a function of (a) coating resistance, (b) coating capacitance and (c) impedance at 0.01 Hz frequencies with exposure time for the coated sample with artificial defect at −0.9 VAg/AgCl ................................................................. 145 Figure 8-15: Optical images of coated samples with an artificial defect after exposure for 30 days in artificial seawater at 10 °C under (a) −0.7 VAg/AgCl and (b) −0.9 VAg/AgCl. Coating disbondment can be seen under −0.9 VAg/AgCl CP ............................................ 147 Figure 8-16: (a) SEM micrograph of a cross-section of the coated B206 under −0.9 VAg/AgCl CP for 30 days in seawater solution and (b) Schematic representation of the delamination process ...................................................................................................... 148 xxi Figure 8-17: Electrochemical response of the coated sample with an artificial defect under −1.1 VAg/AgCl (a) Bode magnitude, (b) Bode phase and (c) Nyquist graphs ........ 149 Figure 9-1: Effect of applied potential on Qnet, Qanode and QH2 in artificial seawater at 22 °C.................................................................................................................................... 155 Figure 9-2: The change in the value of 1/i2 with time during the potentiostatic polarization at 0 VAg/AgCl in artificial seawater .............................................................. 157 Figure 9-3: The influence of applied potential on the ohmic portion of the current transients ........................................................................................................................ 158 Figure 9-4: The relationship between the estimated conductivity and bottom potential ........................................................................................................................................ 159 Figure 9-5: Current-time transition for 200 micron pencil electrode at 22 °C in artificial seawater solution ............................................................................................................ 160 Figure 9-6: \"1D-artificial pit\" electrode dissolution kinetics. (a) Comparison of the time evolution of the pore depth and (b) Decrease of the dissolution rate (µm h−1) as function of the pore depth ............................................................................................................. 161 Figure A-1: Monte Carlo electron trajectory simulations of the interaction volume in B206 ............................................................................................................................... 190 Figure B-1: EDX spectrum for spot 1 in Figure 5-2 and Table 5-1 .............................. 191 Figure B-2: EDX spectrum for spot 2 in Figure 5-2 and Table 5-1 .............................. 191 Figure B-3: EDX spectrum for spot 3 in Figure 5-2 and Table 5-1 .............................. 192 Figure B-4: EDX spectrum for spot 4 in Figure 5-2 and Table 5-1 .............................. 192 Figure B-5: EDX spectrum for spot C-spot1/0h in Table 5-2 ....................................... 193 Figure B-6: EDX spectrum for spot C-spot1/24h in Table 5-2 ..................................... 193 xxii Figure B-7: EDX spectrum for spot A-spot1/0h in Table 5-2 ...................................... 194 Figure B-8: EDX spectrum for spot A-spot1/24h in Table 5-2 .................................... 194 Figure C-1: Bode plots of bare and coated B206 after immersion for 6 days in artificial seawater at 10 °C ............................................................................................................ 195 xxiii List of Abbreviations 1D One dimensional AFM Atomic Force Microscopy ASTM American Society for Testing and Materials CP Cathodic Protection CPE Constant phase element CR Corrosion Rate DFSTEM Dark Field Scanning Transmission Electron Micrograph DO Dissolved Oxygen EDX/EDS Energy-dispersive X-ray Spectroscopy EEC Electrical equivalent circuit EIS Electrochemical Impedance Spectroscopy FIB Focused Ion Beam GPB Gunier-Preston-Bagaryatsky HE Hydrogen embrittlement HER Hydrogen evolution reaction HRTEM High Resolution Transmission Electron Microscopy HPDC High-pressure die cast HVN Hardness Vickers Number FIB Focused Ion Beam IGC Intergranular Corrosion LPR Linear Polarization Resistance NACE National Association of Corrosion Engineers xxiv OCP Open circuit potential ORR Oxygen Reduction Reaction PDP Potentiodynamic Polarization PFZ Precipitate Free Zone PSP Potentiostatic Polarization ppb Parts per billion ppm Parts per million RC Rheocasting RRA Retrogression and re-aging SAD Selected Area Diffraction SCC Stress corrosion cracking SCE Saturated calomel electrode SEM Scanning Electron Microscope/Microscopy SHE Standard hydrogen electrode TPM Through Process Model UTS Ultimate Tensile Strength SDAS Secondary Dendrite Arm Spacing xxv List of Chemical Formulae Ag Silver AgCl Silver Chloride Al Aluminum Al3+ Aluminum (III) ions AlCl3 Aluminum Chloride Al(OH)2+ Aluminum aqueous cations Al(OH)2Cl Aluminum Chloride Dihydroxide Ar Argon B Boron C Carbon C3H6O Acetone CaCl2 Calcium Chloride Cl− Chloride Cu Copper Fe Iron H Elemental hydrogen HNO3 Nitric acid H+ Hydrogen ion (proton) H2 Diatomic hydrogen, gas H3BO3 Boric Acid KCl Potassium Chloride KBr Potassium Bromide Mg Magnesium MgCl2 Magnesium Chloride Mn Manganese N2 Diatomic nitrogen, gas NaCl Sodium chloride xxvi NaF Sodium Fluoride NaHCO3 Sodium Bicarbonate NaOH Sodium hydroxide Na2SO4 Sodium Sulfate O Elemental oxygen O2 Diatomic oxygen, gas OH− Hydroxide Pt Platinum Si Silicon SiC Silicon carbide SrCl2 Strontium Chloride Ti Titanium Zn Zinc xxvii List of Symbols A Electrode surface area (cm2) C Capacitance (general) (F) Cdl Double layer capacitance (equivalent from corresponding CPE) (F cm−2) Csc Space-charge layer capacitance (F cm−2) df Thickness of the anodized passive layer (nm) e Elementary charge (C) E Electrode potential, general (V) Eapp Applied potential (V) Ebot Potential at the bottom of the pore (V) Ecorr Corrosion Potential (V) Efb Flat-band potential (V) Epit Pitting Potential (V) Erp Repassivation Potential (V) F Faraday’s constant (C/equiv) f Frequency (Hz) i Current density, general (A cm−2) I Current, general (A) ianode Net anodic current density (A cm−2) iH2 Hydrogen evolution current density (A cm−2) inet Net current density (A cm−2) icorr Corrosion current density (A cm−2) ilim Limiting current density (A cm−2) ipassive Passive current density (A cm−2) ix Stability product (A cm−1) M Atomic weight (g) n Number of electrons in a reaction, or charge of an ion (equiv/mol) Nd Donor density of n-type semiconductors (cm−3) xxviii Qanode Total anodic charge (C) Qdl Constant phase double layer element (Ω−1 sn cm−2) Qf Constant phase film element (Ω−1 sn cm−2) QH2 Charge associated with hydrogen evolution reaction (C) Qnet Net charge (C) R Resistance, general (Ω) Rca Ratio of the local cathodic current density to the anodic current density Rct Charge transfer resistance (Ω cm2) Rf Film resistance (Ω cm2) Rp Polarization resistance (Ω cm2) Rpore Pore resistance (Ω cm2) Rs Average specific resistance of the electrolyte in the pore (Ω cm) Rsol Solution resistance (Ω cm2) V Volume of the hydrogen gas evolved (cm3) WA Warburg constant (Ω−1 s0.5) wt% Percent mass fraction (aka weight %) Z Impedance, general (Ω) ZCPE CPE impedance (Ω cm2) Zim of Z\" Impedance, imaginary component (Ω cm2) Zre of Z' Impedance, real component (Ω cm2) ZW Warburg Impedance (Ω cm2) |Z| Impedance magnitude (Ω cm2) τ Time constant (s) θ Phase angle in EIS (degree) δ Pore depth (cm) 𝜌 Density, general (g cm−3) ε Dielectric constant 0 Vacuum permittivity (F cm−1) xxix 𝜎 Solution Conductivity (Ω−1 cm−1) 𝜔 Frequency of alternating current (radians) 𝜒2 Chi-squared xxx Acknowledgements First and foremost, I wish to sincerely thank my supervisors and mentors, Professors Edouard Asselin and Daan Maijer for their instrumental role in the completion of this work. It has truly been a privilege to work under their supervision and guidance. I appreciate their patience, generous financial support, wise judgment, and continual encouragement throughout my doctoral program. Funding for the research presented in this thesis has mainly been provided through STPGP Grant 430111-12 from the Natural Sciences and Engineering Research Council (NSERC) of Canada. Special gratitude goes to all my colleagues in the Corrosion Group at UBC for their valuable friendship, scientific input, and positive influence, in particular: Davood Nakhaie, Kashif Mairaj Deen, Yu Liu and Drs. Masyam Mohammadi, Pooya Hosseini Benhangi and Min Xu. I also sincerely appreciate the assistance of Harshmeet Singh in some of the experiments. I also thank the members of the Materials Engineering staff at UBC for their help and patience during this work, in particular: Dr. Carl Reilly, Michelle Tierney, Norma Donald, Sherry Legislador, Mary Jansepar, Glenn Smith, Marlon Blom, Kate Yu, Ross McLeod, Carl Ng, Kim Wonsang and David Torok. Special gratitude and thanks go to Jacob Kabel for his technical assistance during SEM imaging. I would like to express my grateful thanks to my special friends Dr. Jing Liu, Dr. Ibrahim Gadala and Maryam Mohammadi. Their friendship, support and belief in me were a treasure. They also gave me many precious memories from my PhD program. xxxi Next, I would like to thank my aunt and her husband for their endless help, support and encouragement when I had just arrived here in Vancouver to start my program. No words can describe how grateful I feel for the love and backing of all my family members. I am greatly indebted to them all, from my mother Shahla for her unconditional love and concern, my father, Professor Mohammadreza Pournazari, for his wise advice and encouragement, to my beautiful sister Maryam for her affection when living abroad and my brother Ebrahim Pournazari. I only wish I am as good a daughter and sister as they are parents sister and brother. Last, but certainly not least, I must acknowledge with tremendous and deep thanks my fiancé, Dr. Mahmood Ebadian. Through his love, patience, support and unwavering belief in me, I have been able to complete this long dissertation journey. He has taken care of whatever needed tending to without complaining, just so I could focus on completing my dissertation. xxxii Dedication To my loving family and fiancé 1 1 1. Introduction Aluminum casting alloys have seen service in marine applications, where high strength, good mechanical properties and better resistance to corrosion are desired. These properties are ideal for components of tidal-based energy generating systems. From a materials performance outlook, a critical component in such systems is the hub that must be able to transfer the load from the turbine blades to the shaft of the generator. Aluminum–copper casting alloys offer great mechanical properties, such as high strength and hardness, low-cycle fatigue and creep resistances and good machinability [1–6]. The first and most widely used aluminum alloys are those containing 4–10 wt.% copper [4]. The Al-Cu alloy (4.2 wt% to 5.0 wt% Cu), known as B206, is one of the strongest and toughest aluminum casting alloys [7,8]. Hence, further research into the potential of high strength B206 alloy as components of marine hydrokinetic clean energy systems and its reliability and integrity in marine environment is undoubtedly of great importance to governments, energy companies, the general public, and the environment. Impact of B206 corrosion and protection studies Of the many reasons for and contributors to localized corrosion mechanism, microstructural features play a very significant role. It is generally understood that the corrosion of wrought Al-Cu alloys relates to the composition and distribution of a range of intermetallic particles such as Al2Cu, Al2CuMg, and AlCuFeMn [4,7,9–27]. These intermetallic particles are normally considered to be the initiation sites for localized corrosion as a result of the galvanic coupling between them and the surrounding matrix [20]. When the particles form clusters, then stable pitting can develop from the cluster 2 [28]. The most frequent form of localized corrosion in seawater for aluminum alloys is pitting corrosion. Experience shows that when pitting corrosion occurs, it will develop through the first weeks of exposure [2]. Intermetallic particles may also increase the susceptibility of the alloy to intergranular corrosion (IGC) as a result of their precipitation at the grain boundaries [29]. Localized corrosion around intermetallic particles in wrought aluminum-copper alloys has been an area of intense interest; however, there are no public data on the diversity of the intermetallic particles and their effects on the localized corrosion mechanism of cast aluminum-copper alloys. Investigation on the effect of subsurface microstructure at local attack sites would also be of importance to comprehensively understand the mechanism of propagation of the localized corrosion. Therefore, investigation of the effect of microstructure as it pertains to the corrosion behavior of B206 in marine environments will shed lights on the mechanism of corrosion. It will also provide valuable information on the possible thermal treatments to modify the grain structure and distribution of the second phase throughout the grains, which may result in better corrosion properties of the B206. Defects, such as casting porosity, may also increase the weight loss by increasing the real exposed surface area compared to the geometric surface area or by increasing localized corrosion associated with the breakdown of the passive film. Two samples identical in every way except porosity may exhibit very different corrosion characteristics [30]. The limited interchange of the solution inside the pores with that outside the pores might also increase the corrosion rate [30,31]. The study of the effect of casting defects on the corrosion behavior of B206 is a new area of research, particularly as it is linked to 3 the changes in local chemistry inside the defect i.e. the porosity (a major casting defect) has not yet been evaluated in terms of its impact on corrosion. Mechanistic information aside, it is generally known that bare Al alloys will corrode in seawater. One of the protection methods for Al alloys is the use of the passive region within its own electrochemical behavior to create a protective passive layer. This can be done through anodizing the surface before installation in the marine environment, or by ensuring the continuous incidence of a passive condition with an anodic protection system [32]. A combination of protective coating(s) and cathodic protection is another effective and economical method of protection [33,34]. On such a coated structure subjected to cathodic protection, the metal surface at the bottom of the coating defects and pores is protected by the cathodic current. However, application of cathodic protection would enhance coating failure, such as coating disbondment [33]. Therefore, the protection effect will only be obtained if the cathodic protection system is compatible with the chosen coating. Studies on this topic would advance our knowledge of the practical considerations needed for in service conditions. Relevance of this research to tidal energy systems Currently, various technologies are utilized to harvest electricity from marine currents. Tidal energy has two major advantages as compared to wind and solar technology: good predictability based on the gravitational force of the moon and high energy density. Early tidal-energy capture systems, such as Nova Scotia’s Annapolis Royal Generating Station, were based on
 direct adaptation of conventional hydroelectric dam technology in which movable floodgates allowed
 the incoming tide to fill a reservoir and the outgoing tide to turn electrical-generating turbines [35]. Newer 4 hydrokinetic technologies focus on the placement of underwater devices in tidal-current channels using technology based on wind turbines. A critical component in these systems, from the standpoint of materials performance, is the hub that transfers load from the turbine blades to the shaft of the generator. Potential candidate materials include aluminum, titanium, and steel alloys, and carbon-fibre based composites. Of these, aluminum alloys provide good cost-to-performance benefits assuming that a shape casting process can be used and the component lifetime can be optimized. In contrast, the cost-of-materials and manufacturing difficulties for titanium and composites remain too great for cost-effective energy generation. In addition, it remains unclear whether composite materials can even withstand the harsh aqueous environment seen by tidal turbines. For example, a wind-derived prototype placed in New York’s Hudson River in 2007 used composites as rotors and connectors. After six months, the connectors failed due to the marine environment and were replaced with a cast aluminum alloy [36]. Thus, further research into the potential of high strength aluminum alloy castings as components of marine hydrokinetic clean energy systems is warranted. Material properties of B206 In the aluminum foundry industry, the high-strength aluminum-copper 206 alloy offers new opportunities for design engineers to create products that are lighter in weight. The mechanical properties of 206 are comparable to those of cast iron and carbon steel [7,8,37], which makes this alloy very attractive as a substitute material where improved strength-to-weight ratio together with a high toughness are the major design goals. However, due to limitations associated with stress-corrosion, only T4 and T7 tempers are currently available for 206. The T4 temper offers maximum toughness and a very high 5 ductility. The T7 temper, on the other hand, may be applied when the maximum strength and hardness is required. Alloy B206 is a recently developed variant of 206 which is grain modified to enhance its mechanical properties [38–40]. The B206 alloy contains a lower amount of Ti as compared to the 206 variant, which in turn, improves its response to Ti-B based grain refiners and renders it less susceptible to hot tearing [40]. The properties of the B206 alloy are given in Table 1-1. Table 1-1: Mechanical properties of alloy B206 [40] Elastic Modulus Shear Modulus Tensile Strength Yield Strength Shear Strength 68.9 GPa 26.9 GPa 400 MPa 360 MPa 300 MPa A comparison of the strength and ductility of other foundry aluminum alloys against the cast alloy B206 as a function of the secondary dendrite arm spacing (SDAS) is shown in Figure 1-1. The graph clearly shows a superior strength and toughness of B206 for both T4 and T7 tempers compared to the common Al-Si-Cu and Al-Si-Mg alloy families. Figure 1-1: Comparison of the mechanical properties of various foundry Al alloys [40] Business'Confidential' 2013%!! 4! Figure 2. Comparison of the mechanical properties of various foundry Al alloys as a function of the secondary dendrite ar spacing [4]. 5. Hub Design 5.1. Locking Device Selection The locking device is used to connect the hub to the shaft. Given the shaft OD from Table 1 (6.25”) and referring to Ameriloc Internal Locking Device (the locking device manufacturer) technical datasheets, the appropriate locking device is a 3020A 150/200 product. Its characteristics are shown in Table 5. As the locking device is sized in metric, the shaft will have to be machined slightly (approx. 8 mm) to fit inside the bore of the locking device. The 3020A product can transmit a torque of 24,200 Nm (Table 5, variable Mt) per locking device. The safety factor for the rotor torque can then be calculated using this information, along with the turbine power and speed data provided in table 2 as: (e qs. 2) where Tr is the rotor torque, P is the nominal power, ω is the angular velocity, Thub is the hub torque, and S.F.RT is the safety factor for the rotor torque. The hub torque is calculated as one-third the rotor torque since three hubs are used to connect the turbine blades to the shaft in the Mavi Mi1 design. As shown in the calculation, the safety factor for this design is 41.46. 6 Motivation This dissertation focuses on the corrosion behavior of an Al-Cu casting alloy (4.2 wt% to 5.0 wt% Cu), known as B206, in seawater. The broad motivation behind this research is the advancement of a fundamental understanding of the microstructural influences on B206 localized corrosion mechanism. The information gained from this analysis would allow for use in through-process models that ultimately result in better components. The marine environment is one of the most corrosive and diminishment of structural integrity is directly linked to corrosion. Hence, improving the corrosion resistance of B206 by applying different methods such as heat treatment, anodizing and a combination of coating and cathodic protection is also studied. This research is geared toward the marine industry, however, understanding the corrosion of aluminum-copper alloys in aqueous chloride solutions and low temperatures with and without further protection is of benefit in numerous other applications. This includes the automobile, aerospace and military industries. 7 2 2. Literature review Aluminum–copper casting alloys have relatively high strength and hardness, fatigue and creep resistances and good machinability, all of which are dependent on the copper content of the alloy [1–6]. The first and most widely used of these aluminum alloys were those containing 4–10 wt.% Cu [4]. The Al-Cu alloy (4.2 wt% to 5.0 wt% Cu), known as B206, is one of the strongest and toughest aluminum alloys, and exhibits high strength produced by precipitation hardening, good low-cycle fatigue properties, and excellent ductility [7,8]. B206 has a number of potential benefits in engineering applications, such as reducing the weight of important aerospace [8]. It is also a potential candidate material for use in marine applications where good mechanical properties and high strength to weight ratio are desired [41]. The high strength in the 206 alloy (436 MPa tensile stress in T7 condition [1]) is achieved by the presence of copper as an alloying element. These properties are ideal for components of novel tidal-based energy generating systems. However, corrosion continues to be an issue. The following is a review of previous studies on the corrosion process of aluminum alloys in chloride containing solution. Investigations of microstructural dependence of initiation and propagation of the localized corrosion, casting porosity effects and alloy modifications such as thermal treatments, anodizing and a combination of coating and cathodic protection (CP) are discussed due to their influence on corrosion processes and, ultimately material integrity. Microstructure dependence of localized corrosion in Al alloys It is well known that alloying elements are added to aluminum to improve mechanical properties such as the strength and toughness [1]. These alloying elements in 8 the bulk alloy are usually present in concentrations far exceeding their equilibrium solubility under solid conditions. Therefore, exposure to elevated temperatures during solidification and heat treatment leads to the formation of precipitates and a heterogeneous structure [1,4,42]. It is generally understood that the corrosion of Al-Cu alloys relates to the composition and distribution of a range of intermetallic particles such as Al2Cu (θ phase), Al2CuMg (S-phase), and AlCuFeMn [10–21,43–45]. Intermetallic particles are generally considered to be the initiation sites for localized corrosion as a result of the galvanic coupling between them and the surrounding matrix [20]. When the particles form clusters, then stable pitting can develop from the cluster [28]. Intermetallic particles may also increase the susceptibility of the alloy to IGC as a result of their precipitation in the grain boundaries [29]. Intermetallic particles can be classified into two groups: cathodic particles consist of noble elements (such as Cu and Fe) and anodic intermetallic particles with active elements (such as Mg, Zn and Li). Intermetallic particles that are cathodic with respect to the matrix are frequently found to be surrounded by a thin, peripheral zone of matrix which has been dissolved. This type of localized attack is known as trenching [20]. Figure 2-1 shows a typical trenching around the periphery of the AlCuFeMnSi type intermetallic particle after 120 min exposure to 0.1 M NaCl solution [20]. 9 Figure 2-1: Secondary (left) and Backscattered (right) electron images of (Al,Cu)x(Fe,Mn)ySi intermetallic particle in AA2024-T3 alloy after 120 min exposure to 0.1 M NaCl [20] Regarding the intermetallic particles in AA2024 alloys, a paper by Buchheit et al. [15] is often used as a guide. They characterized the second-phase particles greater that 0.2 µm by size and those with diameter greater that 0.5 to 0.7 µm by chemistry in AA2024-T3 alloy. They suggested that the majority of the particles in 2024-T3 are the S phase rather than θ phase. The small and round S-phase particles were found to be approximately 60% of the particles by number, which corresponded to 2.7% of the total surface area. They have also reported a range of compositions including Al7Fe2Cu, Al6Mn, (Al,Cu)6Mn, and a number of undetermined compositions in the class of Al6(Cu,Fe,Mn). The S-phase is active to the matrix, and it was suggested that first Mg and then Al dissolve out of the S-phase particles and leave behind Cu-rich remnants. In the other more recent compositional studies Boag et al. [46] and Hughes et al. [47] reported a variety of intermetallic particles with different compositions in AA2024-T3. They stated that the microstructure of AA2024-T3 is very complicated: it displays multiphase particles, periphery phases around composite particles, and clustering. Boag et al. [46] stated that approximately 40% of the total numbers of intermetallic particles 10 are anodic with respect to the surrounding matrix in AA 2024-T3 alloys, while the remaining 60% of particles are cathodic. However, the reason for this difference with an earlier report by Buchheit et al. [15] was not clear. The second and third most common particles were the Al-Cu-Mn-Fe (e.g. Al6(Cu,Fe,Mn)) and Al-Cu-Fe type particles (e.g. Al7Cu2Fe), respectively. They also discovered that the intermetallic particles usually contain multiphase phases and periphery phases around them and, therefore, the shape-based classification of the particles might not be accurate to categorize them as the anodic or cathodic particles. Figure 2-2 shows a backscattered image and corresponding schematic of a particle with surrounding periphery zone. Figure 2-2: (Top) Backscatter image of an intermetallic particle and its periphery zone and (bottom) corresponding schematic for the backscatter image [46] 11 Hughes, et al. [47] analyzed 82000 compositions in approximately 18000 intermetallic particles in AA2024-T3. They reported nine different compositions including the matrix and a periphery composition around S-phase/θ-phase composites as well as other intermetallic particles containing Al, Cu, Fe, Mn, and Si. The consequences of these different microstructures on the corrosion resistance are not clear. Other reports [16,20,45,48,49] have shown that for S-phase particles, which contain both active elements of Al and Mg and the more noble element Cu, the preferential dissolution of Al and Mg leads to an increase in the concentration of Cu in these particles which is the cause of their subsequent cathodic behavior. Therefore, initially, S-phase particles corrode quickly by dealloying leaving behind a sponge like Cu-rich remnant. After conversion to cathodic Cu-rich remnants, corrosion will continue by the dissolution of the matrix in the vicinity of the particle and trenching. The schematic of this mechanism can be seen in Figure 2-3 [50]. These Cu-enriched particles along with other cathodic particles then caused severe subsurface grain boundary attack [51]. Figure 2-3: Schematic dealloying and subsequent trenching of the S-phase in AA2024 alloy [50] However, when the intermetallic particles are cathodic with respect to the matrix 12 the adjacent matrix will dissolve at the particle/matrix interface, which results in trenching in the vicinity of the particle [45,50]. This mechanism is shown in Figure 2-4. Figure 2-4: Schematic trenching of the intermetallic particle containing noble elements and trenching [50] Boag et al. [52] studied the establishment of stable pits on AA2024-T3 containing a variety of intermetallic particles in 0.1 M NaCl solution. They noticed that there was localized attack of S-phase particles or their remnants and of the adjacent Al-Cu-Fe-Mn particles at their centers. They pointed out that the presence of the Al-Cu-Fe-Mn intermetallic in the vicinity of the S-phase remnants suggests the electrical coupling of the two particles which leads to establishment of the stable pit site. Based on the obtained results, they proposed a model where the coupling between the anodic and cathodic intermetallic particles explained the formation of the stable pits (Figure 2-5). Phase I in Figure 2-5 shows neighboring intermetallic particles with different electrochemical characteristics. Al-Cu-Fe-Mn particles are cathodic with respect to the matrix leading dissolution of the matrix in their vicinity and trenching. In Phase II, the S-phase particles have undergone through dealloying leaving behind a sponge like Cu-rich remnant. Cu-rich remnants act as cathodes which, again, results in trenching and, therefore, stable pit formation. In Phase III, the increased cathodic activity leads to 13 further attack on nearby anodic sites such as the grain boundary leading to propagation into the alloy and subsurface attack. Therefore, the distribution and clustering of intermetallic particles would also affect the corrosion process as well as their composition (which in turn also affects the corrosion process). Figure 2-5: Model of how coupled intermetallic particles contribute to the establishment of the stable pits in AA2024-T3 alloy [52] In another work, Zhang and Frankel [10] studied the transition between pitting and intergranular corrosion in AA2024 in 1M NaCl solution and determined the 14 breakdown potentials associated with pitting and intergranular corrosion. They reported that when two breakdown potentials were observed, the more active of the two was found to be related to the transient dissolution of S phase Al2CuMg particles leading to pitting while the least noble of the two was thought to result primarily from initiation and growth of IGC. In addition to the time of exposure, the galvanic relations between the matrix and intermetallic particles seem to change as environmental conditions are altered. For instance, Al20Cu2Mn3 is cathodic with respect to the matrix at room temperature. However, it becomes anodic as temperature increases to 50 °C, displaying a dealloying behavior such as it is usually seen for the S-phase [45]. Thus, there appears to be a difference in the mechanism of attack for any two different microstructures. The compositional differences observed in the aforementioned studies alone suggest that, to have a better understanding of the corrosion properties for a given material, the correlation between the composition and distribution of the intermetallic particles and the mechanism of localized corrosion should be further investigated. While the localized corrosion around intermetallic particles in wrought aluminum-copper alloys has been an area of intense interest, there are no public data on the diversity of the intermetallic particles and their effects on the localized corrosion mechanism of cast aluminum-copper alloys. Furthermore, studies on the effect of subsurface microstructure at local attack sites would advance our understanding of localized corrosion propagation. As already reviewed, the composition and distribution of intermetallic particles play an important role in the corrosion of Al-Cu alloys. Therefore, heat treatment of the 15 alloy to modify the grain structure and distribution of the second phase throughout the grains might result in better corrosion properties. The mechanical properties and corrosion resistance of aluminum alloys are dependent on the heat treatment process [53]. Variations in the thermal treatments are known to affect the corrosion resistance of aluminum alloys due to significant changes in the microstructure [2]. The first step in the heat treatment of Al alloys is a high temperature solution treatment that allows the alloying elements i.e. Cu, Mg, etc. to dissolve homogeneously into the solid solution. After solution treatment, the alloy is quenched into water/oil to trap the dissolved alloying elements at concentrations beyond their solid solubility limits. With time under ambient conditions (natural aging process) or if the aging is done at a temperature below 220 °C (artificial aging), some of the supersaturated copper combines with vacancies to form extremely fine disk-shaped precipitates, called GP zones. The homogeneous distribution of these precipitates improves the mechanical strength of the alloy [54]. During subsequent artificial aging (heating the water quenched alloy at temperatures above room temperature), precipitates form at grain boundaries, where diffusion is more rapid. As a consequence, the supersaturated copper is depleted along the grain boundaries, which become anodic with respect to the higher copper metal in the centers of the grains and leads to corrosion failure [55]. It has been reported that the pitting potential of aged Al-4wt%Cu was 0.1V lower than that of the water quenched alloy [11]. At longer artificial aging times the supersaturated copper in the center of the grains is consumed by the growing precipitates, and the compositional difference in dissolved copper between the grain boundary and grain center disappears. The material once again becomes resistant to corrosion [55]. Castings subjected to long aging times or simply 16 higher aging temperatures are said to be over-aged and are designated to be in the T7 temper. However, it has been reported that the strength of the over-aged aluminum alloys (7xxx series) is decreased by 5% to 10% [56,57]. Figure 2-6 shows the Al-rich side of the equilibrium Al-Cu phase diagram. Figure 2-6: Al-rich corner of the Al-Cu phase diagram showing the metastable solvus boundaries for GP zones, θ\" and θ', together with the equilibrium solvus line for the phase [54] It has been proposed [42,54] that the decomposition sequence in this system contains one or more of the following processes: Supersaturation Solid Solution → GP zones → θ\" → θ′ → θ {E- 2-1} Various steps in this process may be suppressed by aging at temperatures close to or above the solvus temperatures. To simultaneously maintain high strength and improve 17 the corrosion resistance of aluminum alloys, an alternative heat treatment process has been employed: it is referred to as Retrogression and Re-aging (RRA) treatment [57]. The RRA treatment involves heating the material to high temperature below the solvus line for a short period of time. This step of the process is called retrogression. The material is then re-aged back to its peak strength, and to improve corrosion resistance, by performing artificial aging. This process was first reported by Cina [58] in 1974 and has been shown to enhance the corrosion resistance of 7xxx series aluminum alloys (Al-Zn-Mg-Cu) while maintaining the strength at levels of the T6 condition [59]. The artificial aging causes uniform distribution of precipitates within the grains, which may improve the corrosion resistance of the material. Previous reports have pointed out the occurrence of different reactions during the RRA process and have identified a mechanism of stress corrosion cracking (SCC) resistance [26,27]. The microstructural features such as precipitate free zones (PFZ) adjacent to the grain boundaries, dispersion of precipitate particles along grain boundaries and solute concentration in the grain boundaries have been reported to increase the susceptibility to SCC. Several studies [22,23,26] have reported that the main microstructural changes during retrogression in 7xxx series alloys are the partial dissolution of GP zones and formation of fine η´ MgZn2 precipitates in the aluminum matrix that are re-precipitated during the re-aging, while the η (MgZn2) precipitates at the grain boundaries are allowed to form and grow. The microstructure resulting from the RRA is a very fine distribution of η´ MgZn2 in the aluminum matrix grains while η (MgZn2) precipitates at the grain boundaries distributed similarly to the T7 temper. Therefore, the RRA heat treatment provides for large grain boundary precipitates and coherent matrix precipitates which are responsible for enhanced SCC resistance and 18 high strength, respectively, in 7xxx series aluminum alloys [62]. As it is mentioned above, there are many investigations that agree on the beneficial effect of RRA treatment on the SCC resistance of 7075 aluminum alloys. However, in 1982, Swanson et al. [63] stated that the SCC resistance of the RRA 7075 is no better than the conventional T6 temper. They evaluated the SCC behavior of 7075 in 3.5% NaCl solution at 50 °C on notched samples. They emphasized that the RRA samples resistance to initiation of the crack is better than the T6 samples, however they had the same resistance to the propagation of the crack. Therefore, RRA heat treatment was suggested to be used with caution in that it might not afford the properties it was intended to confer. To date, only a few reports, such as [8,64], have investigated the effects of RRA heat treatment on the corrosion behavior of B206. Singh [64], in a thesis, employed different heat treatments including T7 and RRA (at different temperatures and times) on B206 and investigated the changes in hardness and corrosion behavior to optimize both the mechanical properties and the corrosion resistance in seawater. Based on hardness and potentiodynamic polarization tests, Singh concluded that RRA heat-treated alloys possessed better mechanical properties with slightly lower corrosion current densities than T7 samples. However, he did not employ other useful electrochemical methods to examine the changes in the corrosion behavior of the RRA heat-treated B206 alloy; whilst again, comparatively little is known regarding the corrosion response of RRA B206 in seawater solution. Improving the corrosion resistance of Al-Cu alloys Increasing use of aluminum alloys in seawater has promoted further research into their corrosion behavior under conditions relevant to the marine industry. The effect of 19 oxygen on the corrosion rates of active-passive metals, such as aluminum, can be quite variable. It has been reported that for these alloys, high oxygen concentrations tend to promote healing of the passive film and thus retard the initiation of pitting corrosion [6]. Conversely, high oxygen also favors a vigorous cathodic reaction and tends to increase the rate of pit and crevice propagation after initiation [6]. Additionally, the relatively low and constant temperature (changing from 2 to 12 °C) of seawater, especially in deep layers, is another factor for the large variance in the corrosion behavior of aluminum alloys [65]. It has been pointed out in previous reports that the effects of DO on the corrosion rate are often stronger than those of temperature [6]. Nevertheless, there is a lack of publicly available information for the detailed corrosion behavior of B206 in seawater. As it is mentioned in section 2.1, variations in thermal treatments are known to affect the corrosion resistance of aluminum alloys because of the changes brought about in microstructure [2]. However, further protection is necessary to limit corrosion rates to acceptable levels. One of the protection methods for Al alloys is the use of the passive region within its own electrochemical behavior to create a protective passive layer. This can be done through anodizing the surface before installation in the marine environment, or by ensuring the continuous incidence of a passive condition with an anodic protection system [32]. Anodized aluminum or aluminum alloys have been shown to exhibit an excellent resistance to localized corrosion compared to their as-received counterparts, especially in aqueous neutral chloride solution [66]. The thickness, porosity, breakdown resistance, and semi-conductive behavior of the anodized passive layer are all important 20 properties which determine its performance in preventing further corrosion. EIS is one of the techniques used for characterizing the passive layer(s) formed on different passivating alloys [67–69]. The use of EIS for this purpose can be divided into primary and more advanced analysis methods. A primary method of characterizing passive layer behavior is through single-stage modelling of the charge-transfer resistance in the electrochemical system, which provides a direct correlation to the bulk corrosion rate occurring on the alloy surface [70,71]. On the other hand, secondary or advanced methods involve two or more stages of modelling of the EIS spectra to yield estimates of more specific parameters such as passive layer thickness [72–75]. The protective performance of a passive layer against corrosion is linked with its electronic properties. This correlation has been established over a broad range of metal alloys, including aluminum [76], nickel [77], copper [78], and common iron-based alloys such as carbon steel [79–81] or stainless steel [82,83]. Therefore, methods which can study the electronic properties of passive films can ultimately characterize the protectiveness of passive layer(s), yielding relatively comparable trends to the results of EIS and other fundamental electrochemical methods. The Mott-Schottky technique is commonly used in identifying the semiconductive properties of thin films by measuring electrode capacitance as a function of potential [84,85]. These findings highlight the need for more investigations to characterize the corrosion protection performance and identify the donor density of the passive layer formed on anodized B206 in seawater at different temperatures. As mentioned above, investigating corrosion protection methods for Al alloys in marine environments remains very important. Cathodic protection (CP), which is 21 generally accomplished by using sacrificial anodes or impressed current [86], has been the primary corrosion control methodology for the submerged portion of offshore structures and pipelines for the past 60 years [87]. In impressed current systems, a source of direct current is applied using a power supply and an external anode located some distance away but having an electrical connection to the protected structure. It is generally understood that the current required for the cathodic protection of aluminum alloys in seawater is about an order of magnitude lower than steel, which has been attributed to the formation of the oxide film [88]. In 1989 Gundersen and Nisancioglu [88] studied the cathodic protection of 3 different aluminum alloys at a range of potentials (E = −0.8, −0.9, −1.0. −1.1, −1.2 and −1.3 VSCE) in seawater. They emphasized the effect of intermetallic compounds in the cathodic protection process as follows: oxygen reduction and hydrogen evolution reactions happened at the surface of the cathodic intermetallic particles and lead to an increase in the local pH (and subsequently a decrease in the protectiveness of the oxide film) and corrosion of the adjacent matrix. The trenches that formed between the particle and the surrounding matrix became filled with Al(OH)3, which electrically isolated the particle from the metal substrate. Therefore, both anodic and cathodic processes were slowed, which favored the formation of the protective oxide layer on the matrix surface facing the detached particle. Thus, they concluded that cathodic protection of aluminum alloys removes most of the cathodic intermetallic particles from the surface without exposing fresh particles from the underlying matrix. The formation of a more continuous insulating oxide leads to an improved passivity and low current densities. A combination of protective coating and cathodic protection is the most effective 22 and the most economical method of protection [33,34]. The benefits of utilizing a CP system in conjunction with an appropriate coating system have been recognized since 1950 [89]. The current necessary for CP is significantly decreased in such systems and a substantial power saving is expected [90]. However, it should be noted that the protection effect will be obtained only if the cathodic protection system is compatible with the chosen coating [33]. Large voltages can have disadvantages such as coating disbondment. Disbondment occurs when moisture penetrates a coating and hydrogen is generated at the metal surface beneath the coating [6]. Coating disbondment is recognized as a major corrosion threat, and is commonly associated with several forms of localized corrosion [34]. Coating disbondment commences normally at coating defects where the electrochemical reactions involve an anodic reaction in which the exposed substrate is dissolved (Al → Al3+ + 3e−) and a cathodic reaction (O2 + 2H2O + 4e− → 4OH− (E° = +0.401 VSHE at pH ≥ 7) in less negative potentials or 2H2O + 2e− → H2 + 2OH− (E° = −0.820 VSHE at pH ≥ 7) in more negative potentials) which generates hydroxyl ions takes place [91]. The hydroxyl ions, which are formed underneath the disbonded coating, will destabilize the local charge neutrality and thereby create an electric field, which will facilitate the transport of cations to the cathodic areas. The cations will primarily be sodium ions because the major salt component in seawater is sodium chloride. Consequently, a highly alkaline environment will arise underneath the coating [92]. It is known that during the propagation of coating disbondment, the pH beneath the disbonded area is greater than 14 [93]. Thus, the rate of cathodic disbondment should be affected by the level of CP potential. Martinez et al. [33] investigated the disbonding of an epoxy coating subjected to 23 varying levels of cathodic protection (low (E = −0.98 VAg/AgCl), moderate (E = −1.10 VAg/AgCl) and high (E = −1.40 VAg/AgCl)) on low alloy steel in seawater. The result was that the epoxy coating was capable of withstanding normal levels of cathodic protection that are widely accepted and prescribed by the international standards for steel (potentials between −0.8 and −1.1 VAg/AgCl). After 30 days, with the sample polarized to E = −1.40 VAg/AgCl, the coating was detached from the metal, forming cracks and blisters. In a long-term study on the cathodic protection of coated steel Knudsen and Steinsmo [94] observed that the coating disbonded completely in less than 2 years, which lead to an increase in the current demand (and associated cost) for cathodic protection. Several tests are proposed in the literature to evaluate the compatibility of a coating with CP. As an example, the NACE standard TM0115 [95] recommends the creation of an artificial defect, namely a hole (referred to as a holiday) through the coating and then to immerse panels under a high cathodic potential (E = −1.38 ± 0.02 VAg/AgCl) for 4 weeks. According to this standard, at the end of the test, cathodic disbondment of the coating is evaluated by visual inspection. A knife is also used to physically peal the coating away from the substrate to check for loss of adhesion around the holiday. Therefore, this method is somewhat subjective and dependent on the experimentalist. Generally, a limitation of the visual inspection methods is that they are ex-situ assessment methods that cannot be used to determine the initiation and propagation of the coating disbondment. Therefore, techniques that could provide in-situ monitoring of the system under CP would be very beneficial. Hence, to get more reliable information about the degradation process of the coating, electrochemical tests may be helpful and, particularly, EIS is expected to be an adequate technique. 24 EIS has been used for about 3 decades to measure and monitor the degradation of coatings by quantitatively measuring the resistances and capacitances of coated electrodes in an electrochemical cell [93,96–105]. However, the behavior of EIS measurements when a cathodic protection potential is present is also critical for in situ measurement of cathodic disbondment. Margarit et al. [105] studied the impedance behavior of steel electrodes coated with defect-containing fusion bonded epoxy coatings. The authors asserted that a decrease in the impedance values during exposure time was attributable to the coating disbondment, although a quantitative relationship between the disbonded area and the electrochemical parameters could not be established. Recently, Mahdavi et al. [103] reported that EIS is able to measure and monitor the cathodic disbondment of organic coatings under simulated pipeline CP conditions. However, a clear correlation between electrochemical parameters and coating disbondment area was available only for a relatively short exposure period of 15 days. As it is mentioned above, there have been many investigations on the application of cathodic protection to coated structures, and this protection method is highly relevant to industry. Thus, the compatibility of the CP method with the applied coating and the possible disbondment of the coating are important issues. Effect of casting porosity on the corrosion behavior Corrosion properties of aluminum alloys, which owe their corrosion resistance to the formation of passive films on their surface, might be affected by processing defects such as porosity. These defects result in the localized breakdown of the passive film and pitting corrosion [106]. Moreover, the presence of porosity will increase the exposed surface area in the solution [107]. 25 Song et al. [31,108] studied the corrosion behavior of die cast AZ91D magnesium alloy and reported that, in addition to increased real exposed surface area, micro-pores that originate from defects in the alloy can behave as active sites for corrosion reactions. The corrosion inside the pores might also increase because of limited interchange of the solution inside the pores with that outside the pores. Song et al. [31] have also emphasized that there is a significant and increased risk of localized and crevice corrosion associated with porosity. The influence of casting porosity on the corrosion behavior of Mg-0.1Si alloys has been investigated by Cao et al. [30]. They conducted electrochemical and immersion experiments on high purity Mg pieces with large, small and no porosities (designated as ‘LH’, ‘SH’ and ‘NH’, respectively) to evaluate the corrosion behavior from both potentiodynamic polarization graphs and the hydrogen volume evolved. Figure 2-7 shows the corresponding instantaneous corrosion rate evaluated from the instantaneous hydrogen evolution. Figure 2-7: Instantaneous corrosion rate as evaluated from hydrogen evolution (PH) during immersion testing at the open circuit potential (OCP) in 3.5% NaCl solution saturated with Mg(OH)2 for 7 days at 24 ± 1 °C for solution heat-treated Mg0.1Si specimens [30] 26 It can be seen that for samples with large porosities corrosion rate increases first and tends to a steady state value at the end of exposure. The corrosion rates of the samples with small porosities increase constantly and the final values are similar to those of large porosities. However, for samples with no porosities the corrosion rate was constant and low. However, in one of the ‘NH’ samples the corrosion rate increases after 4 days of exposure. It was pointed out that for this sample, the corrosion during the first days of exposure was sufficient so that the corroding surface intersected with a pore, and thereafter the corrosion accelerated over the whole surface. Figure 2-8 shows the cathodic polarization behavior of the samples with different-sized porosities. Figure 2-8: Cathodic polarization curves measured during the 5th hour and the seventh day during immersion testing at the open circuit potential (OCP) in 3.5% NaCl solution saturated with Mg(OH)2 for 7 days at 24 ± 1 °C for solution heat-treated Mg0.1Si specimens [30] It can be seen that the corrosion potential (Ecorr) of each specimen increased after 7 days immersion. The corrosion current density (icorr) also increased significantly for each specimen after 7 days, except for NH-2, for which there was no significant change after 7 days. They argued that in the presence of casting porosity the real surface area 27 exposed to the solution is larger than the geometric surface area, because of the extra surface area of the internal porosity. They have also reported that the most important impact of activating the sample surface with increased porosity was the breakdown of the protective surface film and micro-galvanic acceleration of the corrosion by Fe-rich particles. Recently, Esmaily et al. [109] have studied the atmospheric corrosion behavior of magnesium-aluminum AM50 alloys produced by different casting methods: rheocasting (RC) and high-pressure die cast (HPDC) with low and high fraction of casting pores, respectively, in cyclic conditions. They reported that the metal loss results, in combination with morphological studies and topographical imaging, showed that RC AM50 exhibited significantly better corrosion resistance than the HPDC material. This effect was explained by differences in the solidification microstructures. They have concluded that an increase in the corrosion rate (mg cm−2) by increasing the porosity, besides the microstructural effects, can be partially attributed to the larger surface area being exposed to electrochemical attack. They have also noted that some pores were filled with corrosion product (Figure 2-9), which implies that the corrosion rate was higher within the pores. This may also be a result of a slower electrolyte evaporation process in the pores. 28 Figure 2-9: Corrosion products filled inside casting porosity on HPDC AM50 after 168 h exposure [109] The above limited research was dedicated to study the effect of casting porosity in magnesium alloys. To the author’s knowledge, there is no public research on the influence of casting porosity on the corrosion behavior of aluminum alloys. Furthermore, in the above-mentioned studies, local chemistry of the solution inside the pores with different size and shapes which might play an important role in kinetic parameters of the corrosion process, are not thoroughly understood. One-dimensional artificial pit (1D-artificial pit) electrodes have found widespread use in the study of localized corrosion. Galvele [110] studied the kinetic parameters for a 1D- artificial pit model of the anodic dissolution process, based on pit depth (x) and current density (i). He considered that dissolution was followed by hydrolysis to generate an acidic environment adjacent to the dissolving interface. By considering transport of ionic species in and out of the pit he calculated a critical value of ix (which is called the pit stability product) above which the pH at the bottom of the pit could maintain anodic dissolution, and the metal remained in the active state. In his study, the critical pH value for aluminum was reported as pH=5 which was reached for ix values lower that 10−6 A 29 cm−1. Therefore, he concluded that since the current density inside the pit, at pit initiation, is at least 1 A cm−2, the necessary acidification can be obtained in pits as small as 10−6 cm in depth. Valuable information on the effect of the pit stability product under a salt film and repassivation potential (Erp) for stainless steels [111–115] in the literature supply supporting evidence for the 1D- artificial pit model of Galvele. The general experimental procedure in localized corrosion studies for pit growth kinetics for stainless steels [111–113] is usually started by applying anodic potentials potentiostatically for different times to grow 1-D pits of different depth. After each potentiostatic hold the potential is scanned in the cathodic direction. In this way, the pit stability product and the repassivation potential of the steel samples could be determined from the same experiment. The potential at which the current changed polarity is taken as the repassivation potential of the sample for that pit depth and the value of (ix)saltfilm is determined from an analysis of the current density plateau at high potentials as shown in Figure 2-10. 30 Figure 2-10: Potential versus current density plots for 316 L generated from linear polarization scans at different pit depths [111] This plateau is the diffusion-limited current density of the metal dissolution at the base of the pit. The diffusion-limited current density is used to calculate the pit stability product under a salt film following {E- 2-2} where iL is the diffusion-limited current density, d is the depth of pit, n is the number of electrons transferred, F is the Faraday’s constant, D is the diffusion coefficient, and ΔC is the difference in concentration of metal cations between the surface of the pit and the bulk solution [112]: 𝑖𝐿𝑑 = 𝑛𝐹𝐷∆𝐶 = (𝑖. 𝑥)𝑠𝑎𝑙𝑡𝑓𝑖𝑙𝑚 {E- 2-2} However, the 1-D pencil electrode technique (and the electrochemical experimental procedure mentioned above) has not been widely used for aluminum and aluminum alloys most likely due to experimental complications associated with hydrogen evolution within the pit cavity [116–118]. Beck [117] studied salt film formation during corrosion of pure aluminum using rods with 5 mm2 cross-sectional areas. He reported that 31 aluminum artificial pit electrodes show only ohmic growth (pit growth in the salt free condition) at low chloride concentration and low potentials as the convection induced by hydrogen evolution prevents the formation of a continuous salt film. However, he showed that high anodic potentials (ca. 4 VSCE) and concentrated aluminum chloride solutions (> 80% of saturation) may result in a mass transfer-controlled current density and the formation of a continuous salt film. Potentiodynamic polarization curves for aluminum in concentrated AlCl3 solution are shown in Figure 2-11 [117]. Figure 2-11: Potentiodynamic polarization curves for aluminum in 90% saturation AlCl3 solution ta 1mm depth [117] Ohmic growth has also been reported for aluminum by others [118–120]. More recently, Cook et al. [116] studied pit propagation in pure aluminum using the 1D-artificial pit technique in HCl solution and reported mass transport control at high potentials, ohmic growth at intermediate values followed by a transition to a further region of potential-independent dissolution at ca. 100 mV prior to repassivation for pits 32 of 50 to 250 µm. However, for 25 µm pits dissolution proceeded under mass-transport control. Findings in their study supported the hypothesis that continuous pit propagation requires a high degree of saturation, near 100%, in AlCl3. Therefore, one of the major experimental limitations encountered for aluminum was that the pore is frequently blocked with hydrogen bubbles during testing [120] and that the Al anodic dissolution current cannot be measured directly from the potentiodynamic polarization graphs. However, Akiyama and Frankel [120] used an experimental set up to collect hydrogen gas simultaneously from the pure Al foil electrodes during aluminum dissolution as is shown in Figure 2-12 to calculate the charge associated with the cathodic reaction (the primary cathodic reaction that takes place on the Al electrode is hydrogen evolution). Cao et al. [30] also used more or less the same experimental set up to calculate the volume of evolved hydrogen in their study on Mg alloys. Figure 2-12: Schematic drawing of the artificial crevice cell [120] Based on the findings of the aforementioned studies, the pencil electrode method 33 can be used to study the effect of the solution inside the pores on the kinetic parameters of the corrosion process of alloys such as B206. 34 3 3. Objectives Aluminum alloy B206 is a recently developed high strength foundry alloy that has strong potential for use in automotive, aerospace and marine applications. In the multi-faceted review of relevant previous literature presented above, it is apparent that the corrosion properties of B206 have not been formerly investigated. Therefore, the present study aims to investigate the marine corrosion and protection of B206. Considering the deficiency of available fundamental information on the above-mentioned aspects of this alloy, the goal of this work is to contribute a comprehensive understanding of the corrosion mechanism of B206 in seawater solution. The fulfillment of the key technical objectives outlined in section 3.1 contributes to the improvement of general knowledge around aluminum alloy corrosion mechanisms and corrosion control methods such as anodizing and cathodic protection. Key technical objectives {1} Examine how corrosion develops on B206 during the exposure to seawater solution. Study the early stages of localized corrosion initiation and propagation and the role of intermetallic particles buried beneath the surface. {2} Due to the potential significant impact of microstructure on corrosion, determine whether the RRA heat treatment has a substantial effect on the corrosion susceptibility of B206 in seawater. {3} Characterize the behavior of the passive film formed on B206 at different temperatures in seawater including the thickness and semi-conductive properties. Can the passive film be relied-on to protect this alloy? 35 {4} Examine the compatibility of the chosen coating with the cathodic protection system and the possible disbondment of the coating under cathodic protection. In the absence of acceptable thermal or passivation processes, will coating/CP provide a solution for the use of this alloy system in seawater? {5} Study, for the first time, the effect of B206 porosity on corrosion outcomes. This information could eventually be used for through-process models of B206 castings. 36 4 4. Approach and methodology In order to achieve the proposed objectives listed in section 3.1, this work includes a comprehensive set of electrochemical experiments and microstructural studies. Each set of laboratory experiments is performed using a combination of the following electrochemical techniques: Open Circuit Potential (OCP), Linear Polarization Resistance (LPR), Potentiodynamic Polarization (PDP), Potentiostatic Polarization (PSP), EIS under OCP or PSP conditions, and Mott-Schottky tests. Optical Microscopy, Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) are used to determine the microstructural features of the B206 specimen as well as the microstructural evolution due to the heat treatment and disbondment of the coating. Microstructural analyses are conducted using the following techniques: Energy-dispersive X-ray Spectroscopy (EDX), Focused Ion Beam (FIB), Selected Area Diffraction (SAD) pattern, Dark Field Scanning and High Resolution Transmission Electron Micrograph (DFSTEM and HRTEM). Casting procedures and sample preparation Ingots of aluminum alloy B206 with the chemical composition indicated in Table 4-1 were melted in a resistance furnace and degassed with Ar for 10 minutes. Table 4-1: Chemical composition of the B206 alloy Elements (wt.%) Cu Si Fe Mn Mg Ti Zn Al 4.68 0.03 0.05 0.27 0.31 0.02 0.01 balance 37 Following degassing, liquid metal was manually poured at 700 °C into a fireclay mould with a water-cooled copper chill at the bottom. A picture of the mould assembly is shown in Figure 4-1. Figure 4-1: Picture of the mold configuration The mould is insulated on the sides and designed to directionally solidify the casting from the bottom to the top, thereby providing a range of solidification conditions and avoiding macroporosity formation. To obtain microstructural consistency between the samples, the castings were cut in transverse sections relative to the solidification direction. Therefore, it is expected that that microstructure would not vary appreciably from one location to another at each section (Figure 4-2 (a)). Circular specimens with 1 cm2 surface area were then cut from sections of the casting as shown in Figure 4-2 (b). The line intercept method was used to measure the average grain size. The specifics of the sample preparation and the types of electrochemical experiments performed on the samples will be described in more detail in the following sections. 38 Figure 4-2: (a) Transverse slice sectioning of the B206 casting and (b) circular samples extracted from each sections 4.1.1 Heat treatment procedure In order to study the effect of heat treatment on the corrosion properties of the B206 alloy, some of the specimens were subsequently subjected to the RRA thermal process. As previously mentioned in chapter 2.1, RRA is a multistep heat treatment process, which involves the material being solution heat treated, quenched, artificially aged, retrogressed and then aged again. Singh [64], in a thesis, employed a 3-step artificial aging (low-high-low) which he called an RRA heat treatment. The high temperature aging step, which was called retrogression in that study, was performed at three different temperatures: 200, 210 and 220 °C. The ensuing changes in hardness and corrosion behavior in seawater were then investigated. Potentiodynamic polarization on B206 (in naturally aerated artificial seawater) retrogressed at various temperatures showed very similar icorr values (5, 3 and 4 µA cm-2 for 200, 210 and 220 °C, respectively). Thus, the retrogression temperature did not have a significant effect on the corrosion resistance of the studied alloy. However, the RRA B206 retrogressed at 220 °C showed the highest hardness values ranging from 150-170 Hardness Vickers Number 39 (HVN). Therefore, in the present study, the retrogression step was performed at 220 °C. It is important to note that this study did not aim to demonstrate that the heat treatment used for B206 resulted in classical RRA microstructure (such as what has been previously described by Cina [58] for 7xxx series alloys). It should also be mentioned that “RRA” was chosen as the name for this process following the Singh [64] study. The thermal process is described in more detail in Table 4-2. Table 4-2: Thermal process of the investigated B206 alloy Heat treatment Homogenized Artificial aging * As cast --- --- RRA 515 °C, 2 h + 525 °C, 8 h + 65 °C water quenching 155 °C, 20 h + 220 °C, 0.08 h + 155 °C, 20 h *Note: Intermediate quenching at 65 °C is carried out between the first two artificial aging steps. 4.1.2 Mass loss experiments 4.1.2.1 Laboratory experiments in artificial seawater For the mass loss measurements, B206 coupons (both as-cast and RRA alloys) of 20×10×2 mm with a hole at one end (Figure 4-3) were used. Figure 4-3: Picture of the samples for the laboratory mass loss experiments 40 The mass loss samples were cut in the same direction as the other circular samples (transverse sections relative to the solidification direction). The electrolyte and the preparation procedure for mass loss measurements are described in section 4.2 and 4.2.2, respectively. 4.1.2.2 Field experiments in natural seawater For the mass loss measurements in natural seawater, rectangular bar samples were machined from the B206 alloy casting. The samples ranged from 20 to 23 mm in length, with a cross-sectional area of about 85 mm2. A 5-mm diameter hole was drilled at one end of each sample. The samples were attached to a Teflon plate using plastic cable ties with sufficient slack that they would be free to move slightly, and spaced sufficiently far apart that they would not come into contact with each other. The plate was enclosed in a wire cage to protect the samples from incidental contact as shown in Figure 4-4. Figure 4-4: Configuration of the mass loss experiments samples in natural seawater The samples were immersed in clean, natural seawater close to the inlet of Esquimalt Harbor, off of the Strait of Juan de Fuca in Victoria, BC. The cage was suspended approximately 3 m from shore and at a depth of 1 to 3 m (depending on tide). The entire set of samples was removed for mass loss measurements after eight exposure 41 periods: 38 days, 42 days (80 days total), 49 days (129 days total), 51 days (180 days total), 44 days (224 days total), 49 days (273 days total), 45 days (318 days total) and 48 days (366 days total); following cleaning, measurements, and examination, the samples were returned to the exposure site. Once removed from the seawater, the white corrosion product on the sample surface was gently scraped off. The samples were then cleaned through ultrasonic agitation in tap water for about 10 minutes, followed by a light scrubbing with a toothbrush under running water to remove the bulk of the corrosion product and any biological growth. The final cleaning step was ultrasonic cleaning at room temperature in 67-70% TraceMetal grade HNO3 for 10 minutes, followed by ultrasonic cleaning in water, and immersion in ethanol to displace any water. The samples were then dried in air overnight, and weighed. 4.1.3 1D-artificial pit (pencil) electrode preparation In order to study the effect of casting porosity, a 1D-artificial pit (pencil) electrode is used. To manufacture B206 pencil electrodes, the casting was cut into rod samples with 1 mm diameter and 2 cm length. The obtained samples were ground to produce electrodes of different diameters of approximately 1 mm to 200 microns using a fixed drill set-up as shown in Figure 4-5. The samples were turned in the drill and SiC paper was applied to the sample to reduce the diameter. 42 Figure 4-5: Picture of the fixed drill set-up used for grinding the rod B206 samples The electrodes were made by mounting the different sized samples in epoxy resin housed within a PVC tube so that one end could be exposed as the working electrode. Prior to any experiment, the electrode was abraded with 120 grit silicon carbide (SiC) paper and washed with deionized water. Since the cross-sectional area varied somewhat along the length of the electrode, the surface area of each electrode was measured before and after each experiment using an optical microscope. Test environments and methods Laboratory tests in this study are conducted in ASTM artificial seawater solution (ASTM D-1141-98 [121]) at pH 8 containing different dissolved oxygen (DO) concentrations and at different temperatures. The chemical composition of the artificial seawater is presented in Table 4-3. The specifics of each test environment depend on the corrosion and microstructural aspects being investigated, and will be described in each 43 corresponding section separately. Moreover, test routines are designed with various electrochemical procedures, microscopy methods and characterization techniques to achieve the key technical objectives outlined in section 3.1. Table 4-3: Chemical composition of the artificial seawater used in laboratory tests Chemical Composition (g L−1) NaCl MgCl2 Na2SO4 CaCl2 KCl NaHCO3 KBr H3BO3 SrCl2 NaF 24.53 5.20 4.09 1.16 0.695 0.201 0.101 0.027 0.025 0.003 4.2.1 Corrosion initiation and microstructural evaluation Final polishing was performed on the Buehler Phoenix Beta grinder/polisher via a three-step process: 6 then 1 µm diamond paste followed by 0.05 µm colloidal silica, rinsed with distilled water and air dried. Polished specimens of B206 were immersed at 10 C (the corresponding approximate in-service temperature of the previously mentioned tidal energy generators) for 2.5, 10, 15, 30, 60 minutes and 2, 15, 24 and 48 hours in artificial seawater. The samples were then removed, rinsed in deionized water and allowed to dry in laboratory air. After each immersion interval, the microstructural changes were observed on the same regions of the B206 samples. Scanning electron microscopy (SEM) was performed on a Hitachi S3000N utilizing a conventional tungsten electron gun equipped with an EDX detector. Backscattered and secondary electron imaging was performed using beam energies of 15 kV and probe currents of approximately 1 nA. EDX elemental maps were conducted to characterize the composition of different phases and EDX spot analysis results were used to compare the changes in the chemical composition of these particles after each immersion time. The EDX results were converted to weight percent using the QUARTZ 44 XOne software (the PhiRhoz (PRZ) calculation was used by the software for this conversion). Monte Carlo electron trajectory simulation was developed using the Casino program to represent the condition used to image structures in SEM. The interaction volume in the B206 sample is shown as a function of a beam energy of 15 kV in Figure A-1. An electron-beam interaction volume of 1 µ3 was previously reported for the AA2024 alloy at 20 kV accelerating voltage [15]. Cross-sectional SEM examination on as-cast B206 samples was performed on a Zeiss Sigma instrument utilizing a Schottky field emission gun. Backscattered imaging was performed using beam energies of 15 kV. The surface of the corroded sample was sectioned with the focused ion beam milling (FIB) method using a FEI Helios NanoLab 650 DualBeam. The SEM imaged the newly milled surface with a current of 0.2 nA at 2 kV accelerating voltage and a working distance of 4 mm. The specimen was carbon coated with 15 nm of carbon with the Leica EM MED020 high vacuum coating system. The coated samples were mounted on standard 25 mm aluminum stubs using electrically conducting adhesive carbon and colloidal silver. In order to identify the preferential nucleation sites of pitting, anodic galvanostatic experiments were carried out at a constant current density of 2 µm cm−2 and the potential changes as a function of time were monitored. The first maximum spike indicated pit nucleation. After galvanostatic experiments, the sample was carefully rinsed in distilled water and dried in air and the microstructural analysis was performed using SEM. 45 4.2.2 Electrochemical experiments and weight loss measurements to compare the corrosion behavior of as-cast and RRA heat-treated B206 alloy Electrochemical tests were conducted in a conventional jacketed 1 L glass cell with the B206 alloy as the working electrode, a silver/silver chloride (Ag/AgCl, 4 M KCl) electrode as the reference, and a graphite rod as the auxiliary electrode at 10 °C. The cell was connected to a VWR cooling and heating bath with a programmable temperature controller. A thermometer in the test cell ensured that the temperature of the test electrolyte was consistently at equilibrium with the coolant temperature. The electrochemical tests were performed using a Princeton Applied Research (PAR) VersaSTAT3 potentiostat/galvanostat. The experimental set up is shown in Figure 4-6. Figure 4-6: Schematic experimental set up used for electrochemical experiments All the potentials reported in this thesis are with respect to the Ag/AgCl reference electrode (0.197 VSHE). The working electrode was connected to a copper wire by conductive silver epoxy, and the assembly was mounted in epoxy resin and positioned in 46 the cell. Before each experiment, the working electrode was immersed in the electrolyte until the corrosion potential reached a steady state condition (change of less than 5 mV in OCP over a 10 minute period [122]). Electrochemical impedance spectroscopy (EIS) tests at the free corrosion potential were conducted at different immersion times (4 hours followed by every 12 h for 6 days). The perturbation signal had an AC excitation signal of 10 mV and a frequency range of 105–10−2 Hz. Potentiodynamic polarization was performed from −0.4 to 0.6 V vs. OCP with a scan rate of 0.166 mV s−1. The measurements mentioned above were conducted at least twice to confirm good reproducibility. The Linear polarization resistance (LPR) experiments were carried out from −10 to +10 mV versus the OCP at a scan rate of 0.166 mV s–1. The polarization resistance (Rp) values were calculated from the slope of the linear portion of the current–potential plot. For the mass loss measurements, B206 coupons were immersed in artificial seawater for different durations. Once removed from the solution, the samples were cleaned by ultrasonically agitating in tap water for about 10 minutes, followed by a light scrubbing with a toothbrush under running water to remove the bulk of the corrosion product. The final cleaning step consisted of sonication at room temperature in HNO3 for 10 minutes, followed by ultrasonic cleaning in water, and immersion in ethanol. The samples were then dried in air, and weighed. In each case duplicate experiments were conducted and showed that the second results were within ±1% of the first. SEM was performed with a Hitachi S3000N utilizing a conventional Tungsten electron gun. Backscattered and secondary electron imaging was performed using beam energies of 15 kV and probe currents of approximately 1 nA. TEM characterizations were 47 performed at McMaster University. Electron transparent samples for TEM characterization were prepared by dimpling using a Dimple Grinder from Gatan. Then, ion milling of the dimpled sample was done using Low Angle Ion Milling & Polishing System from Fischione. TEM images were obtained from Jeol 2010F at 200 kV, equipped with an EDX system from Oxford Instruments. 4.2.3 Electrochemical test methods for corrosion and passivation processes Figure 4-6 shows a schematic set up used to acquire all the electrochemical measurements. To investigate the effect of temperature on corrosion behaviors of the B206 specimen, experiments were carried out at 6, 10, 14 and 25 C (±1 C). The working electrode was connected to a copper wire by conductive silver epoxy, and the assembly was mounted in epoxy resin and positioned in the cell. Before each experiment, the working electrode was immersed in the electrolyte until the corrosion potential reached a steady state condition (change of less than 5 mV in open circuit potential (OCP) over a 10 minute period [122]). Potentiodynamic polarization was performed from −0.25 V vs. OCP up to 1.0 VAg/AgCl at a scan rate of 0.166 mV s–1. To investigate the temperature-dependent development and performance of the B206 passive layer, a multi-stage test routine was conducted in deaerated artificial seawater. Only seawater with practically zero DO (N2-deaerated) based on ref [86], results in approximately 1 ppb dissolved oxygen could be achieved) was used because no stable passivation of B206 occurs in the presence of DO from natural aeration. Furthermore, since it is anticipated that B206 turbines in marine environments will not be installed near the seawater surface, the DO is expected to be low. In fact, knowledge of the absence of notable passivation for B206 in DO-containing environments may support 48 the choice of deeper placement locations for such turbines during their design, especially if an uncoated or anodized B206 material is to be used in the construction. In the multi-stage test routine for evaluating passive layer behavior, following a 2 h OCP segment to stabilize the corrosion processes occurring on the specimen surface, a potentiostatic anodizing session was run for 6 h at a potential of −0.65 VAg/AgCl to grow the passive layer(s) to be evaluated. This potential was predetermined based on the passive region measured during previous potentiodynamic polarization experiments. This was immediately followed by EIS and then Mott-Schottky, without removal of the specimen in between. EIS is considered to be a non-destructive electrochemical technique [123], and since it was conducted at the same anodizing potential of −0.65 VAg/AgCl and lasted less than 5 minutes in each routine, it was considered to not affect the passive layer structure for the subsequent Mott-Schottky test. EIS was run at a frequency range between 104 and 5×10−2 Hz, an AC disturbance signal of 10 mV, and a sampling frequency of 10 points per decade. The subsequent Mott-Schottky tests were scanned in the −0.8 VAg/AgCl to −0.5 VAg/AgCl potential range within the passive region of the previous potentiodynamic polarization plots. The Mott-Schottky scans were conducted at a frequency of 1 kHz, an AC amplitude of 10 mV, and with a step height of 25 mV. 4.2.4 Coating and cathodic protection (CP) application Electrochemical experiments conducted to determine the corrosion protection potential range (PSP on the bare B206 samples) were performed in the setup shown in Figure 4-6. The potentiostatic polarization experiments were carried out at OCP toward cathodic potentials with steps of 0.05 V up to −0.95 VAg/AgCl and conducted for 3 hours to obtain a stable current at 10 C (the most probable corresponding in-service temperature 49 of the previously mentioned tidal energy generators). This experiment shows the effects of concentration polarization due to an oxygen reduction reaction (ORR) and activation polarization due to hydrogen generation as follows, respectively: O2 + 2H2O + 4e− → 4OH− {E- 4-1} 2H2O + 2e− → H2 + 2OH− {E- 4-2} Since pigmented epoxy coatings applied on nickel aluminum bronze propellers successfully reduced corrosion rates [124], this coating system is used here. A commercial 2-component epoxy paint from AkzoNobel (Intershield 300) was able to increase the corrosion resistance of B206 by up to 4 orders of magnitude for short immersion times of about 5 days in artificial seawater. The related EIS measurements can be found in Figure C-1 in Appendix C. Therefore, after determining the potential range for cathodic protection, the B206 was coated with the epoxy paint using an adjustable film applicator (Elcometer 3570) in one layer of 100 µm thickness and according to the manufacturer’s specifications. Prior to all experiments, the substrate was cleaned based on SSPC (the Society for Protective Coatings)-SP1 (removal of all visible oil, grease, soil and other soluble contaminants from surface with solvent) and polished with a 120-grit paper. Samples were dried at room temperature for 7 days before being used in testing. A digital Elcometer gauge was also used to measure the dry film thickness. The reason for applying a thin coating is to get satisfactory EIS measurements (because of the nonconductive nature of the coating). To make a coating defect an artificial hole of 1 mm diameter was introduced at the center of the coated samples by drilling using an end mill to reach the metal surface (Figure 4-7). 50 Figure 4-7: B206 coated sample with at artificial defect at its center Since the main purpose of this component of this investigation was to examine the compatibility of the cathodic protection and the protective coating, the main experimental parameter was cathodic polarization potential. The electrochemical measurements of the coated samples were carried out in artificial seawater using an electrochemical workstation (Princeton Applied Research (PAR) VersaSTAT3) with a three-electrode configuration flat cell (Figure 4-8). Cathodic polarization to different potential levels of −0.7, −0.8, −0.9 and −1.1 VAg/AgCl (hydrogen evolution region) was applied to coated samples and prolonged for 30 days for the first three potentials (in the oxygen evolution region) and 17 days (408 hours) for the potential of −1.1 VAg/AgCl. To obtain a general understating of the coating’s electrochemical response, with and without cathodic potentials, electrochemical experiments were conducted on an intact coating surface (without any artificial defect) at OCP and the cathodic potential of −0.7 VAg/AgCl for 30 days. The coating appearance and disbondment were studied microscopically (both optical and SEM). 51 Figure 4-8: Schematic of the flat cell used in electrochemical experiments on coated samples with and without defect at OCP and different levels of CP 4.2.5 Electrochemical test methods for pencil electrode experiments Electrochemical tests were conducted in a conventional 100 mL glass-jacketed cell with the B206 pencil electrode as the working electrode, Ag/AgCl reference electrode and a platinum coated Ti mesh as a counter electrode. The cell was connected to a VWR cooling and heating bath with a programmable temperature controller to maintain the temperature of the electrolyte at 22 °C. Due to the limited number of samples the electrochemical experiments were performed at 22 °C (to compare the obtained data with the results reported in the literature) rather than 10 °C. Figure 4-9 shows the experimental setup during which the evolved hydrogen was collected in a burette above the specimen. Coating Coated aluminum sample 52 Figure 4-9: (a and b) Hydrogen collection experiments setup. The electrode surface faced upward. The evolved hydrogen was collected by means of the funnel and the burette above the specimen. The burette was initially full of solution, which was displaced by the evolved hydrogen, allowing easy measurement of the cathodic charge associated. (c) Hydrogen bubbles can be seen in the burette above the specimen. (d) Schematic top view of the experimental arrangement shown in (a and b) The electrode surface faced upward during the hydrogen collection experiments. The charge associated with the hydrogen evolution reaction (HER) QH2 was determined from the volume of the hydrogen gas collected. This assumes a collection efficiency of 100%, which is an overestimate since some of the evolved hydrogen gas will dissolve into solution or remain as bubbles in the pore or burette walls. We also made an assumption that the charge consumption associated with the ORR was negligible. The B206 artificial pencil electrodes were potentiostatically polarized over a range of 0 to 1 VAg/AgCl. The current density and net charge, Qnet, were recorded from the Princeton Applied Research (PAR) VersaSATAT3 potentiostat/galvanostat. The cathodic charge was determined from Faraday’s law: Funnel Pt coated Ti mesh counter electrode Reference Electrode Ag/AgCl Pencil Electrode (d) (a) (b) (c) 53 𝑄𝐻2 =𝑉𝜌𝑛𝐹𝑀 {E- 4-3} where V is the volume of the hydrogen gas evolved (cm3), ρ is the density (g/cm3), n is the number of electrons transferred (equiv/mol), F is Faraday’s constant (96487 C/equiv) and M is the atomic mass of hydrogen in this equation. The net charge passing through the potentiostat, Qnet is the difference between the total anodic charge, Qanode, and the charge of the local cathodic reaction on the B206 pencil electrode, QH2, because significant cathodic reactions take place on the B206 electrodes. Therefore, the charge associated with Al dissolution at the pore, Qanode, is determined as follows: 𝑄𝑛𝑒𝑡 = 𝑄𝑎𝑛𝑜𝑑𝑒 − 𝑄𝐻2 {E- 4-4} In order to grow pores of different depth, potentiostatic polarization experiments were conducted on pencil electrodes with different diameters at 0 VAg/AgCl for 600 seconds. The sum of the anodic charge densities passed is directly proportional to the pit depth δ, given in {E- 4-5}, based on Faraday’s law [112]: 𝛿 =𝑀𝑛𝐹𝜌∫ 𝑖𝑑𝑡 {E- 4-5} 54 5 5. FIB/SEM study of pitting and intergranular corrosion in B206 aluminum alloy1 Aluminum-Copper casting alloys offer good mechanical properties such as high strength and hardness, fatigue and creep resistances, and good machinability, all of which are dependent on the copper content of the alloy [1–6]. However, corrosion continues to be an issue. The microstructural differences and their influences on the corrosion behavior mentioned in chapter 2.1 alone suggest that, to have a better understanding of the corrosion properties for a given material, the correlation between the composition and distribution of the intermetallic particles and the mechanism of localized corrosion should be further investigated. While the localized corrosion around intermetallic particles in wrought aluminum-copper alloys has been an area of intense interest [10,14,20,28,29,45–47,52,125–129], there are no public data on the diversity of the intermetallic particles and their effects on the localized corrosion mechanism of cast aluminum-copper alloys. The objective of this chapter is to investigate the corrosion of a B206 Al-Cu casting alloy in artificial seawater. The subsurface microstructure at local attack sites is studied by milling cross sections using a FIB. This allows detailed subsurface examination to be performed on a site-selective basis. The roles of microstructure and de-alloying of the intermetallic particles, copper dissolution and re-deposition, and grain boundary attack are addressed. This chapter is based on a paper [130] which was published as part of the research work leading towards this PhD thesis. 1 Sh. Pournazari, D. M. Maijer, E. Asselin, Corrosion, vol. 73, No. 8, pp. 927-941, August 2017. 55 Composition of the intermetallic particles Figure 5-1 shows the optical and backscattered SEM micrographs of the Al-Cu alloy, which shows the intermetallic particles in the Al matrix. With regards to shape, there are two main types of particles present: those that are more or less spherical with a diameter of 4–6 µm and those that are irregular in shape, present in the alloy in the form of clusters of the continues particles. Higher magnification SEM micrographs of the regions A and C, for which the corrosion response will be discussed afterwards, are shown in Figure 5-1 (c) and (d), respectively. As it is shown in Figure 5-2 at higher magnification, some particles are multiphase. Mg, Mn, Si, and Fe were indicated on multiphase particles as well as main compositional elements Al and Cu. The chemical composition of the particles marked in Figure 5-2 is shown in Table 5-1. The related EDX spectra can be found in Figures B-1 to B-4 in Appendix B. To gain a better understanding of the composition of the intermetallic particles, elemental EDX maps were generated on an area that includes different types of multiphase particles on the alloy surface. According to Figure 5-3, the EDX maps show that the matrix is α-Al with dissolved Mg, Cu, and Si while it is depleted in Fe and Mn. In the matrix, there are three main types of intermetallics which are described as follows: 56 Figure 5-1: (a) Optical image, (b) SEM micrograph of B206, (c and d) higher magnification of regions C and A, respectively Figure 5-2: Backscatter image of some multiphase intermetallic particles (indicated by dashed circles) on the B206 surface (a) (d) (c) (b) 1 1 C A B D E 1 2 4 3 57 Table 5-1: Chemical composition of the indicated parts of the intermetallic particles in Figure 5-2 Spot Composition (wt%) Al Cu Si Mg Fe Mn 1 47.17 52.34 0.49 - - - 2 76.3 13.82 5.37 4.51 - - 3 52.51 22.46 6.31 - 13.68 5.04 4 46.74 52.79 0.47 - - - The most abundant intermetallic for Al-Cu alloys is of nominal composition Al2Cu. This compound can be found both in particles formed at the grain boundaries during solidification and in the eutectic phase. According to the Al-Cu phase diagram, the solubility of Cu in Al is maximum at 548 °C (equal to 5.8 wt%). A decrease in temperature decreases this value to less than 1 wt% at room temperature [131]. Solidification of the Al-~5 wt% Cu alloy starts at 650 °C by nucleation of Al rich phase and ends at 560 °C and the result is a single-phase α-Al matrix. However, thermodynamically, the α-Al matrix with ~5 wt% Cu in solid solution is not stable below 530 °C and hence, based on the phase diagram, Cu deposits as the Al2Cu intermetallic more preferentially at grain boundaries. These can be seen as the elongated particles at the grain boundaries of the matrix in a binary alloy. In these types of alloys with Cu less than 5.8 wt%, thermodynamically, the eutectic reaction is not possible. However, in the studied alloy, under real conditions, because of the existence of other alloying elements (Mg, Si, Mn, Fe, Zn, …) and also non-steady-state cooling, the critical copper solubility and eutectic points shift. Therefore, non-equilibrium eutectic Al2Cu phase will form [132,133]. 58 Figure 5-3: Top: Backscattered electron image of a multiphase intermetallic particle. Bottom: the x-ray mapping corresponding to the shown intermetallic particle Al-Cu-Mn-Fe-Si secondary phase Mg and Si containing phase Al2Cu Al-Al2Cu eutectic 59 The other particle type, which is also an intermetallic compound at the grain boundaries, contains Mg and Si elements. A well-known type of these precipitates in aluminum alloys is Mg2Si. The maps in Figure 5-3 show that Mg and Si-containing particles can be found at the grain boundaries and between Al2Cu particles. Some reports show that the formation of Al-Cu-Mg-Si particles, such as Al5Cu2Mg8Si6, is also possible during the solidification of this alloy [37]. The formation of these compositions in the periphery of the particles is more obvious in the EDX maps shown in Figure 5-4. Figure 5-4: The EDX elemental maps show the formation of multiphase particles in the B206. The periphery of the particle is composed more from Mg and Si The other precipitate that can be found at the grain boundaries and at the interface of Al2Cu phase and matrix contains mainly Fe, Si, and Mn as well as Al and Cu. The presence of these types of particles has been reported in AA2024-T3 aluminum alloys 60 with the composition of Al20(Cu,Fe,Mn)5Si including isolated domains of Si-free Al3(Cu,Fe,Mn) [46]. The formation of these particles can be attributed to the coring phenomenon during solidification. As a result of coring, the less soluble elements in the liquid (e.g., Fe and Mn) are pushed to the solidification front. This changes the undercooling for solidification, and also changes the accumulation of the elements at the grain boundaries of the solid material. This accumulation leads to the formation of an intermetallic phase. One of the possible complex compounds, which has been identified during the solidification of this alloy, is Al32(Cu,Mn,Fe)8(Al,Si)4Si2. It forms according to the following reaction [37]: L → Al + Al2Cu + Al32(Cu,Mn,Fe)8(Al,Si)4Si2 It is important to remember that because of the nature of this alloy, and the fact that it has Al as the base metal and Cu as the main alloying element, every intermetallic precipitate that forms is considered to contain these two elements. The aforementioned particles are the most abundant and observable phases in the microstructure of this alloy. However, there are possibilities for formation of different phases and particles. The presence of different phases (and consequently interfaces) when coupled with the different metallic elements from highly active elements like aluminum in the electromotive series to noble elements like copper, suggests that a complicated and diverse type of corrosion mechanisms might occur when subjected to a corrosive environment. Initiation of the localized attack In a recent report [20], the authors showed extensive localized corrosion on aluminum AA2024-T3 after just 2.5 min of immersion in 0.1 M NaCl solution. However, 61 no evidence of corrosion initiation on the alloy surface was observed after 2.5, 5, 10, 15, 30, and 60 min and 2 h and 15 h in this study. Additionally, EDX spot analysis on the intermetallic particles showed no changes in the composition of the intermetallic particles in all of the studied regions for these immersion times. The first signs of localized attack on the surface were observed after 24 h of exposure in seawater solution. Figure 5-5 (a) shows an overview of the attack on the surface after 24 h. Copper was evident at scattered areas in a very low proportion on the surface and accumulated on the intermetallic particles (bright white spots). Previous studies on de-alloying and corrosion of Al alloy 2024 [16,44,49,134] reported that there are two possible mechanisms (which will be discussed more in detail later) for the copper rich nature of the particles. One is that the copper ions generated by the dissolution of the alloy matrix re-deposited on the cathodic particles. Second, the copper enrichment was due to selective dissolution of Al from the particles. Higher magnification SEM micrographs of the regions C and A in Figure 5-5 (b) and (c) indicated the potential sites for initiation of localized attack (showed by arrows). The onset of pitting can occur at the surface of intermetallic particles as labeled by the dashed arrows in Figure 5-5 (b) and (c). Table 5-2 indicates the changes in the chemical composition of each element (the same particles marked in Figure 5-1 on regions A and C). The EDX spectra can be found in Figures B-5 to B-9 in Appendix B. Oxygen peaks show the initiation of the corrosion on the surface. The decrease in the weight percent of Al in the particles (region A and C-spot 1) suggests the dissolution of intermetallic particles at pit initiation sites. 62 Figure 5-5: Backscattered electron images of (a) the B206 surface, (b and c) higher magnification of selected regions (circles) after 24 h exposure to artificial seawater at 10 °C (arrows indicate the signs of localized corrosion on the surface), (d and e) higher magnification of some attacks showed by solid white arrows in b and c C A C (a) (b) (c) A (e) (d) 63 Table 5-2: The changes in the chemical composition of the mentioned particles in Figure 5-1 Regions/Time Chemical composition in wt% Al Cu Si Mg Fe Mn O A-Spot 1/0h 46.53 52.47 0.51 0.49 - - 2.57 A-Spot 1/24h 41.12 47.33 0.56 0.49 - - 10.5 C-Spot1/0h 46.9 51.98 0.59 0.52 - - 2.18 C-Spot1/24h 39.36 45.27 0.64 2.22 - - 12.5 In order to prove whether the dissolution and re-deposition of Cu occurs from the particle surface, EDX elemental maps of the particle are shown in Figure 5-6. After 24 h of immersion in the solution, copper enrichment can be observed on the surface of the particle in the EDX map. Figure 5-6: The (a) SEM micrograph and (b) EDX elemental map for copper on a Al-Cu-Mg-Si particle in B206 alloy It has been reported that Al alloys can release Cu particles in a non-faradaic process producing Cu metal. These Cu particles are then oxidized by dissolved O2 resulting in copper ions that are subsequently re-deposited on the alloy surface, resulting in enrichment above and beyond enrichment arising simply from the dissolution of Al [45,49,135]. The aforementioned results, which indicate the presence of Cu on both the (a) (b) 64 matrix and intermetallic particles, can provide evidence to support non-faradaic release of Cu impurities from the alloy and the subsequent re-deposition on to the alloy surface. Buchheit, et al. [45,49,135] have studied Cu ion formation by de-alloying Al2CuMg phases and they stated that Cu ion generation from Al2CuMg under free corrosion conditions is reasonable. Although it is not quantitatively identified that the studied particles are S phase (determining the stoichiometry of the intermetallic particles is not the aim of this chapter), the proposed mechanism for pitting at the surface of the intermetallic particles can be justified based on the aforementioned results. The solid white arrow in Figure 5-5 shows the initiation of the corrosion attack at the interface of the intermetallics with the matrix or the periphery phase. The higher magnification SEM micrographs on initial sites of localized attack are given in Figure 5-5 (d) and (e) as well. Figure 5-7 indicates the localized attack at some other intermetallic particles in the investigated area. Figure 5-7: Pit initiation on the intermetallic particles after 24 hours 65 The possible preferential starting point of the pit formed on this intermetallic particle can be the formation of a microcrack at the interface between the Al matrix and the intermetallic particles. The attack then might concentrate on the particle, resulting in a pit. However, there is no obvious microstructural factor that affects whether peripheral matrix pitting or selective particle dissolution occurs. In order to identify the preferential nucleation sites of pitting, a constant anodic current has been imposed at the electrode surface (galvanostatic anodic polarization) and the potential changes as a function of time were monitored. The first maximum spike after 10 s of polarization in Figure 5-8 indicates pit nucleation. The microstructural analysis of the surface after this experiment indicates that the pits are more likely to initiate on the surface of the particles rather than the matrix/particle interface. Figure 5-8: (a) Potential-time behavior of the B206 in artificial seawater at 10 °C with an applied current of 2 µA, (b) Microstructural analysis of the surface after galvanostatic anodic polarization shown in (a). The pit was nucleated on the intermetallic surface Localized corrosion propagation SEM micrographs of the alloy surface in artificial seawater for 4, 8, and 16 more h did not display any changes in morphology, size, or propagation of the attacked sites on the surface. However, some dark regions were identified in the backscattered SEM (b) (a) 66 imaging after 30 h from initiation of the attack. Figure 5-9 shows the overview of the investigated area after different times of exposure in the solution. It is interesting that except for the localized attack, which is indicated by a dashed arrow at the top left side of Figure 5-5 (c), there is no significant change in the size or propagation of the pits after this exposure time. Those pits are apparently passivated. Since the contrast in SEM micrographs originates from the compositional and atomic number differences, it is evident that there is a propagation of the localized corrosion indicated by the white circle in Figure 5-9. The presence of dark regions is attributed to the formation of corrosion products (aluminum oxide/hydroxide) surrounding the pit [28,136]. As it can be seen, some parts of the continuous particle, which was attacked after 30 h, dissolved completely away by increasing immersion time. 67 Figure 5-9: The overview of the investigated area (a) before immersion, (b) after 24 hours and (c and d) 30 and 48 hours after the first 24 hours, respectively The higher magnification images of this attack in Figure 5-10 show the particle edge and trench in detail. The corrosion products are evident at the center and surrounding area of the particle indicating the corrosion of the matrix in the vicinity of the corroded intermetallic particle. However, it is difficult to determine exactly how corrosion developed. Al, Cu, and oxygen can be seen in the EDX maps after 48 h of exposure to the solution (Figure 5-11). (a) (d) (b) (c) 68 Figure 5-10: The Backscattered (left) and Secondary electron (right) imaging of the attacked area on the surface 30 (a and b) and 48 (c and d) h of the initiation time This seems to indicate an anodic behavior of the intermetallic, which subsequently switches to cathodic behavior, enabling the development of localized corrosion in the adjacent matrix. The resulting anodic process must be the selective dissolution of Al and/or Mg. The cathodic response of this, which is the reduction of O2 and/or H2O, has been reported to take place on the de-alloyed Al-Cu-Mg particles at an acceptable rate [15,134,137]. (a) (d) (c) (b) 69 Figure 5-11: (a) SEM micrograph and (b-d) EDX maps of Al, Cu and O after 48 hours of the initiation of the pits in seawater solution, respectively ORR occurs with the following reaction, which results in an increase in the pH of the electrolyte in the neighborhood of local cathodes: O2 + 2H2O + 4e− → 4OH− {E-5-1} As pH increases to about the range 9-9.5, the passive oxide on the surface of the alloy matrix will chemically dissolve and the bare aluminum will begin to selectively dissolve via the soluble AlO2 anion according to the reaction, Al2O3 + 2OH−(adsorbed) → 2AlO2−(aqueous) + H2O {E-5-2} Thus, the particle remnants directly supported the oxygen reduction and further contributed to the local alkaline condition and consequently enabled the matrix de-alloying which, in its turn, cause the more local enrichment of copper. Schematic (a) (c) (d) (b) A B 70 illustration of matrix de-alloying is shown in Figure 5-12. As illustrated in this figure, the de-alloyed second phase serves as cathode active to oxygen reduction and the surrounding matrix undergoes Al de-alloying causing a local enrichment of copper. Figure 5-12: Schematic illustration of matrix de-alloying in the vicinity of an intermetallic particle on the surface This contribution of the de-alloyed particle in de-alloying of the surrounding matrix has been shown in Figure 5-13. A dual beam FIB/SEM was used to examine the subsurface microstructure at selected local attack sites within the corroded area. Figure 5-13 shows the cross section of the corrosion site at the location A indicated in Figure 5-11. Cross-sectional examination at this position revealed that de-alloying was accompanied by dissolution of the adjacent matrix. Figure 5-13 clearly supports the aforementioned mechanism. EDX analysis on that sponge-like morphology at low accelerating voltage showed high intensity of Cu with respect to Al and Mg. In the work of Vukmirovic, it has been reported that the matrix de-alloying process results in the agglomeration of solid solution Cu into Cu particles that also serve as cathodic sites available to support yet more oxygen reduction [16]. Once the intermetallic particle becomes more enriched in copper and its character become clear cathode, the cathodic 71 process of reduction takes place over the intermetallic and the associated anodic response of the oxidation of aluminum takes place over the neighboring matrix. Figure 5-13: (a) FIB section at location A indicated in Figure 5-10, (b) higher magnification SEM image of FIB cross-section trough the intermetallic particle after 48 hours of initiation of the pits in seawater solution (boxed area in (a)) and (c) EDX on the sponge-like morphology Cross-sectional examination at position B (Figure 5-14) clearly revealed that localized corrosion is associated with the large intermetallic particle which is buried beneath alloy surface, but probably connected to the surface. It can be seen from Figure 5-14 (b) that the intermetallic particle had been partially converted to sponge at the interface with the matrix. However, by further milling, sponge-like morphology appeared in the particle along a narrow, straight path, beginning at the particle/matrix boundary (Figure 5-14 (c)). Recently, Zhang, et al. [128] have studied the de-alloying behavior of the multi-phase intermetallic particles in AA2024-T351 aluminum alloy and (a) (b) (c) 72 reported that this banding structure was developed due to the preferential attack of the stacking faults in θ phase particle. These places are marked with black arrows in Figure 5-14 (c). From Figure 5-14 (b) and (c) it can be seen that the matrix/grain boundary attack began at the very earliest stages of de-alloying. Furthermore, it seems that de-alloying began preferentially at the particle/matrix interface and then it proceeded down the interface and along certain directions within the particle. It is also obvious in Figure 5-14 (c) that there is the remnant of the intermetallic particle (sponge-like morphology) at the surface just below the corrosion products cap, which is connected with a dark-etched layer of the matrix to the buried particle. It might provide an example of cooperative behavior between intermetallic particles or IGC. However, this is not clear. King, et al. [28,127] found a similar preferentially interface/matrix dissolution of the S-phase particle in AA2024 under a seawater droplet. It is also evident in Figure 5-14 (d) that the morphology of the particle had become partially porous and the other parts of the particle retained the typical morphology of the intermetallic particle. Accurate EDX analysis of the specific features in the de-alloyed particle remnant is not possible because of the relatively large interaction volume at the acceleration voltage required for the detection of copper with respect to the fine size of the features. Qualitative EDX analysis of the porous region of the particles detected significantly increased Cu/Al and Cu/Mg ratios compared with the intact particle, confirming that parts of the particle had undergone de-alloying of magnesium and aluminum to leave Cu-rich remnants. Some dark areas (indicated by arrows in Figure 5-14 (e), (f) and (g)) were occasionally observed along the interface of the particle and matrix and in the matrix. These dark areas demonstrated that dissolution of the 73 adjacent matrix occurs in conjunction with the particle dealloying [127]. As the matrix beneath the alloy surface at the periphery of intermetallic particles was attacked during immersion there is the possibility that IGC occurred in the vicinity of the intermetallic particle. However, it is difficult to establish the direct link between IGC and the intermetallic particle-induced localized corrosion site from Figure 5-14. Figure 5-15 shows the images of cross sections of the alloy surface. The dissolution of the alloy matrix immediately adjacent to the particle is obvious in all backscattered images. The remnant of an intermetallic particle can be seen in Figure 5-15 (d). In Figure 5-15 (c) through (f) the grain boundary appeared etched out, indicating that corrosion of the alloy matrix propagated in the form of IGC. It is also more evident from Figure 5-15 (f) that IGC connects to the localized corrosion at the surface of the intermetallic particle, suggesting that IGC initiated from the corrosion pit bottom, and developed into the large network buried underneath the alloy surface. In Figure 5-15 (c) and (f) a crack was found to form into the bulk material. The development of acidic conditions at the reaction front may have been the cause of this behavior. However, it was not clear whether the corrosion products had dissolved (aluminum hydroxides are soluble in strong acids [28]) or not formed in the first place. Beneath the alloy surface, the matrix at the vicinity of intermetallic particle was also attacked. The development of corrosion in Figure 5-15 (e) formed at the edge of the specimen shows the big network of IGC. As the grain boundary network spreads in all directions, i.e. laterally as well as in depth, IGC also propagated in all directions along the grain boundary network. As indicated in Figure 5-15 (g), IGC, which developed deeply into the bulk alloy to the depth of 160 µm, shows a severe attack. 74 Figure 5-14: (a) FIB milling and (b-d) SEM micrographs of corroded B206 intermetallic at various stages of milling. SEM image of FIB cross-section shows localized attack of an intermetallic particle cluster below the surface. (f and g) higher magnification of SEM image of c and e, respectively (a) (c) (b) (e) (g) (f) (d) Remnant of the particle Remaining particle 75 Figure 5-15: SEM micrographs of the cross section of the B206 after immersion for 48 more hours after the initiation of the localized corrosion in seawater solution(a) (b) (c) (d) (e) (f) (g) 76 Summary In this chapter, SEM with EDX was used to study the formation of stable pits on the B206 aluminum-copper casting alloy in artificial seawater at 10 °C. The polished specimens were immersed for various times and the development of corrosion around intermetallic particles was monitored. It is demonstrated that alloy microstructure plays a defining role in the corrosion of the B206 alloy in artificial seawater. The EDX maps show that the microstructure of the alloy is complicated, exhibiting multiphase particles. Although not quantified in this study, there are three main types of intermetallic particles which could be related to Al2Cu particles containing small amounts of Mg and Si or Al-Cu-Mg-Si, and Al-Cu-Mn-Fe-(Si). SEM studies on polished surfaces reveal that there is no sign of localized corrosion in the first 24 h of exposure in seawater solution at 10 C. The onset of pitting can occur at the surface of intermetallic particles or at the interface of the intermetallic particles and matrix. However, galvanostatic anodic polarization shows that the preferential nucleation sites of pitting are on the surface of the particles. The localized corrosion observed at the surface of intermetallic particles can be divided into two types: localized corrosion associated with a small amount of corrosion product (not propagating during the course of immersion) and localized corrosion that develops through subsurface particle clusters connected to the surface particles. Copper deposits at specific sites on the particles, beginning mainly around the edges. The dissolution and re-deposition of the copper on the intermetallic particles supports the idea that de-alloying of the Al2Cu or Al-Cu-Mg particles occurs followed by the onset of pits on the surface of these particles. Thus, the intermetallic particles initially present an anodic behavior with respect to the matrix, giving rise to a process of selective 77 dissolution of Al and Mg. As a result of this, there is an increase in the concentration of Cu in these intermetallics, which is the cause of their subsequent cathodic behavior. The results from the FIB-SEM/EDX characterization demonstrate that the subsurface particles are able to contribute to activities at the surface. Below local attack sites, de-alloying of the particles and matrix/particle interfacial attack were observed. The de-alloyed particle remnant with the copper layer on its surface can provide effective cathodic support for the preferential anodic dissolution of the alloy matrix adjacent to the particle as long as the remnant is connected to the alloy matrix. Intergranular attack was not generally present over the alloy surface. Corrosion was found to initiate at a cluster of intermetallic particles at the surface and to develop into a network of IGC. The grain boundaries played a linking role between nearby intermetallic particles. Intergranular attack penetrated deep (160 µm) into the alloy. 78 6 6. Effect of retrogression and re-aging (RRA) heat treatment on the corrosion behaviour of B2062 Having identified the corrosion processes that occur when the B206 alloy is exposed to seawater solution, in the previous chapter, it is clear that the alloy is susceptible to corrosion and, hence, may require surface treatments in practical applications. Furthermore, IGC was observed in Figure 5-14 and Figure 5-15 to propagate well beneath the alloy surface. As mentioned in section 2.1, to maintain high strength and improve the corrosion resistance of aluminum alloys (7xxx series) simultaneously, a heat treatment process has been employed; it is referred to as Retrogression and Re-aging (RRA) treatment [57]. The RRA treatment involves heating the material to high temperature below the solvus line for a short period of time. This step of the process is called retrogression. The material is then re-aged back to its peak strength by performing artificial aging and to improve corrosion resistance. To date, only a few reports, such as [8,64], have investigated the effects of RRA heat treatment, which have been successful in 7xxx series aluminum alloys, on the corrosion behavior of B206. Singh [64], in a thesis, employed different heat treatments including T7 and a three-step artificial aging referred to as RRA (at different temperatures and times) on B206 and investigated the changes in hardness and corrosion behavior to optimize both the mechanical properties and the corrosion resistance in 2 Sh. Pournazari, K. M. Deen, D. M. Maijer, E. Asselin, Materials and Corrosion, 69 (2018), pp. 1-18, January 2018 79 seawater. Based on hardness and potentiodynamic polarization tests, Singh concluded that RRA heat-treated alloys possessed better mechanical properties with slightly lower corrosion current densities than T7 samples. However, they did not employ other useful electrochemical methods to examine the changes in the corrosion behavior of the RRA heat-treated B206 alloy. For this reason, the present chapter is aimed at determining whether this heat treatment (detail of the heat treatment is described in 4.1.1) may have a substantial effect on the corrosion susceptibility of B206 in seawater. Thus, the corrosion resistance of as-cast and RRA B206 is evaluated in artificial seawater at 10 °C by means of electrochemical experiments such as EIS and potentiodynamic polarization, weight loss measurements and microstructural studies. The changes in the corrosion behavior of both alloys are monitored for different immersion times. This chapter is based on a paper [138] which was accepted for publication in Materials and Corrosion as part of the research work leading to this PhD thesis. Microstructural evolution and electrochemical behavior comparison of as-cast and RRA B206 alloy Figure 6-1 shows SEM and optical micrographs of as-cast and heat-treated B206. It seems that the RRA treatment has little effect on the grain size. The average grain size for RRA and as-cast samples are 53 and 70 µm, respectively. The grain boundary phase is thought to be Al2Cu. Intermetallic particles were also observed within the grains in the as-cast microstructure. The same microstructural features were observed in the SEM micrographs of Figure 6-1 (c) and (d). Both samples exhibit microstructure consisting of precipitates along the grain boundaries and within the α–Al matrix. It seems that the grain boundary area in the as–cast sample was decorated with continuous and thick layers of 80 precipitates compared to the RRA sample, which contains a relatively low amount of precipitates at the grain boundaries. However, it is difficult to achieve detailed microstructural information about the formation of the early stage intermetallic particles and transition phases during RRA treatment from the SEM images. Figure 6-1: SEM and optical images of (a and c) as-cast and (b and d) RRA heat-treated B206 alloy. Therefore, TEM characterization combined with EDX and diffraction analysis of RRA B206 were preformed at McMaster University and the obtained results are presented in Figure 6-2. The presence of precipitates with different shapes is evident within the α–Al matrix, as shown in Figure 6-2 (a). The dark field scanning transmission electron micrograph (DFSTEM) (Figure 6-2 (b)) represents the differentiation in the chemically and structurally sensitive phases formed within the matrix (α–Al) during RRA treatment. The uniformly dispersed fine θ\" precipitates were revealed, which are expected 81 to nucleate from GP zones during the aging process. However, some heterogeneous θ' precipitate may also nucleate within the grains especially aligned along the dislocations sites [139]. EDX analysis carried out to determine the composition variation across the precipitates showed the presence of precipitates with different compositions. EDX spectra of the particles in Figure 6-2 (c) show high concentration of Mn, Fe and Si in addition to Al and Cu. However, for the lath shape particles shown in Figure 6-2 (d), Si peaks cannot be seen. The high resolution TEM (HRTEM) micrographs in Figure 6-2 (f) and (g) show the precipitates and matrix α-Al phase. Confirmation of the plane orientation within the precipitate was ambiguous due to very large inter-planer spacing. However, the dominant (200) plane orientation was in agreement with the d–spacing (0.204 nm) of the α-Al phase in the HRTEM. The inset of Figure 6-2 (g) presents the indexed selected area diffraction (SAD) pattern of α–Al which also validated the {200} and {220} plane orientations. It should be noted that an attempt was made to characterize an as-cast B206 sample by TEM. However, it was not possible to suitably prepare the sample for TEM because of large precipitates and voids that exceeded the thickness of the TEM sample. 82 Figure 6-2: Microstructural features of RRA sample (a) TEM image showing the presence of precipitates within the α-Al matrix (b) DFSTEM present the distribution of θ\" and θ' precipitates within the matrix phase (c and d) showing DFSTEM micrographs of intermetallic particles with corresponding points EDX spectra (e) high magnification image showing the presence of round shape precipitate (f) HRTEM image of precipitate illustrating the plane orientation (g) HRTEM of the α-Al matrix and inset showing the indexed SAD pattern of the matrix phase. (a) (c) (b) (d) (e) (f) 0.024 nm (200) 1.23 nm f) (g) ΘΘ 83 The OCP values of the B206 alloys are shown in Figure 6-3. The potential of the RRA alloy varied over time with an increase during immersion to a peak value of −0.565 V at approximately 48 hours. This could be attributed to a longer incubation period (greater than 48 hours) for the localized corrosion of the RRA alloy, and may be as a result of a relatively stable and protective oxide film. The potentials then remained relatively stable at ~ −0.595 V until the end of the exposure period. For the as-cast sample, no significant changes occurred in the corrosion potential during the immersion and it stabilized around −0.55 V at the end of the immersion period. The fluctuation of the OCP values with time for the as-cast alloy during the immersion period is probably because of the corrosion product layer induced competition between localized corrosion and passivation [140]. The increasing trend in OCP but relatively negative value of the RRA sample, compared to the as-cast sample, indicates its higher corrosion tendency [141]. This behavior was associated with the development and stability of intrinsic passive film, which is directly related with the microstructural features of the RRA sample. The average OCP values with standard deviations for the RRA and as-cast samples are shown in Table 6-1. The value of the potential is close to −0.550 V for the as-cast alloy. However, the potential of the RRA sample reaches a stable value of −0.590 V after 120 h immersion. The difference observed in the OCP of the as cast and RRA samples during long immersion is thus around 40 mV. 84 Figure 6-3: Corrosion potential values of as-cast and RRA alloy immersed in artificial seawater at 10 °C Table 6-1: Average values of OCP (vs. Ag/AgCl) for as-cast and RRA alloys Alloys Average OCP of 120 h (V) As-cast −0.549 ± 0.007 RRA −0.590 ± 0.02 To further investigate the corrosion behavior of the as cast and RRA alloys, potentiodynamic polarization experiments were conducted after different immersion times, as shown in Figure 6-4. As cast B206 exhibits similar polarization behavior at different times of immersion in artificial seawater. Figure 6-4 (a) shows that the cathodic branch is under diffusion control. At potentials just below the corrosion potential, the cathodic polarization curve showed rapid kinetics on the electrode surface, which approached limiting current densities at high negative overpotential. The potentiodynamic polarization curves for the RRA alloy (Figure 6-4 (b)) had almost the same cathodic polarization behavior and corrosion potential after different immersion times. In contrast to the cathodic branch, the anodic branches were quite different. The potentiodynamic 85 polarization curves obtained for the RRA and as-cast alloys after 120 hours of immersion in solution are presented in Figure 6-4 (c). The continuous increase in the current during anodic polarization of the alloy samples suggest their uniform corrosion with the possibility of progressive localized attack. It has been reported that pitting of this alloy naturally occurs at its corrosion potential [41]. Figure 6-4: Potentiodynamic polarization curves at different immersion times for (a) as-cast, (b) RRA and (c) both as-cast and RRA B206 after immersion in seawater solution for 120 h. SEM micrographs of the as cast and RRA alloys at this point (Figure 6-5) show the dissolution of the particles or trenches and edges at the periphery of some particles and confirm the localized corrosion of the studied alloys at their OCP. 86 The values of the corrosion potential (Ecorr) and corrosion current density (icorr) (calculated by extrapolation of the Tafel slopes back to Ecorr) of the alloys are listed in Table 6-2. According to Figure 6-4 and Table 6-2, there is no significant change in Ecorr for both RRA and as-cast samples. With respect to the icorr values, there is no significant difference between the as-cast and RRA samples. Figure 6-5: SEM micrographs of (a) as-cast and (b) RRA aluminum alloy under free corrosion condition, showing the oxide film and preferential dissolution (exposed to seawater solution after 48 hours) Table 6-2: Electrochemical parameters obtained from polarization curves McCafferty [142] discussed the validity and limitations of the Tafel extrapolation method for the determination of corrosion rates and pointed out that the Tafel method works better for uniform corrosion than for localized attack. However, localized corrosion does occur at OCP for both RRA and the as-cast sample (Figure 6-5). This is likely why no significant difference in the icorr values is observed from the polarization Alloy Immersion time (h) Ecorr (V vs. Ag/AgCl) icorr (µA cm–2) RRA 40 −0.59 6.5 120 −0.61 5.6 As-cast 40 −0.51 6.3 120 −0.47 5.2 87 curves. It has been observed that the severity of localized attack on the RRA sample is higher than for the as cast sample. The severe dissolution along the grain boundaries of the RRA samples is certainly related to the high concentration of the intermetallic precipitates after 48 hours of immersion. During the aging process, the formation of the small but dense coherent phases (precipitates) could grow in size along the grain boundaries leaving the depleted matrix. This could enhance the localized intergranular dissolution due to the development of local galvanic cells. Ralston et al. [143] reported that after extended aging the decrease in number density and coarsening of the Guinier-Preston-Bagaryatsky (GPB) zones and S–phases could increase the pitting corrosion of Al-Cu-Mg alloy. This, then, is a possible reason for the significant corrosion and intergranular dissolution observed on the RRA samples (as evident in Figure 6-5 (b)). However, the presence of the S–phase was not identified in RRA samples and, therefore, the mechanism of localized attack is still not clear. Muller et al. [55] have reported that the decrease in the pitting potential of the aged Al-Cu alloys at different aging times is related to the Cu depletion around the precipitates. Therefore, it could be assumed that the decrease in Cu and other alloying additions within the matrix at the interface of intermetallic particles (as shown in Figure 6-6 (a)–(d)) may enhance the pitting tendency of the aged samples. 88 Figure 6-6: (a) DFSTEM micrographs showing the lath shape precipitate and (b, c, d) the line EDS spectra reveals the composition variation along the precipitate Although polarization curves have been used widely for corrosion analysis, the use of EIS (described below) provides further insight into the corrosion process. The EIS spectrum of the RRA and as-cast alloys was measured after the first hour of immersion followed by every 12 hours, up to 144 h in artificial seawater at 10 °C. The results of the 4, 12, 24, 48, 72 and 144 hour immersion experiments are shown in Figure 6-7 and Figure 6-8. (The EIS response of the RRA samples was quite similar for 96, 120 and 144 hours of immersion and thus only the results for 144h are shown here). The impedance behavior at high frequency corresponds to the electrolyte resistance. The surface roughness and microstructural variation over the surface could lead to non-homogeneous charge distribution within the double layer. This behavior can be represented by a Aluminum Copper Manganesbcd 89 constant phase element (CPE), as is the case in our analysis. The low frequency impedance trends can be related to the charge transfer reactions at the metal/electrolyte interface. Figure 6-8 (b) shows an example of the Nyquist plots recorded for as-cast B206 after 72 hours of immersion in artificial seawater. The capacitive loop at relatively high and intermediate frequencies could also be related to the charge distribution in the double layer and charge transfer through adsorption/desorption of reaction intermediate species leading to localized dissolution of samples. 90 Figure 6-7: (a) Effect of exposure time on EIS Bode plots and (b) fitting of the EIS spectra of RRA B206 (a) (b) 91 Figure 6-8: Effect of exposure time on EIS Bode plots of as-cast B206 and (b) experimental and simulated Nyquist of as-cast B206 after 72h Also, the ingress of Cl− within the localized reaction site i.e. pit could further enhance dissolution by the following reaction mechanism ({E- 6-1} through {E- 6-4}) [144–146]: Al → Al3+ + 3e− {E- 6-1} 2H2O + 2e− → 2OH− + H2 {E- 6-2} Intermediate reactions and diffusion of Cl− occurring within a pit: Al3+ + 2OH− → Al(OH)2,ads+ {E- 6-3} Al(OH)2,ads+ + Cl− → Al(OH)2Cl {E- 6-4} Figure 6-9 shows the equivalent electrical circuits (EEC) used to fit the EIS data of RRA B206 samples. (a) (b) 92 Figure 6-9: Equivalent circuit used to fit EIS data (a) RRA and (b) as-cast B206 As seen from Figure 6-7, they have two (not well-defined) time constants and the protection of the material from corrosion by heat treatment is represented by the following values: Rsol corresponds to the solution resistance and is expressed at high frequencies; Rporous/oxide and Rct are the resistance of the porous and barrier (inner) layers, respectively, and Qporous/oxide and Qdl are the frequency independent real constants associated with the porous layer and the electrochemical double layer, respectively. The bode plots of the as-cast sample contain one time constant and the Nyquist plots contain a high frequency semi-circle and a low frequency region where the real and (a) (b) 93 imaginary parts increased linearly with one another or plateaued. This behavior is often associated with mass-transport-limited reactions [147]. However, for longer immersion times this low frequency region was no longer seen in the Nyquist plots. Several models of circuits to fit our experimental data were attempted. The best results (best agreement between experiment and fitting for as-cast samples) were obtained using the EEC in Figure 6-9 (b) in which, Rct and Cdl are the charge transfer resistance and the constant phase element are attributed to the surface distribution of double layer charge. W is a finite Warburg diffusion component corresponding to the diffusion of ionic species (Cl−, OH−) within the localized defects. The impedance spectra were fitted to these circuits and the extracted impedance parameters analyzed by ZSimpWin software from EIS plots are listed in Table 6-3 and Table 6-4. In the present cases the χ2 values, which indicate the best fit of the experiment data to the simulated models, was in the range of 10−3 to 10−4. Figure 6-7 (b) and Figure 6-8 (b) show the data fitting for as-cast and RRA samples at the different immersion times. As it can be observed from Table 6-3, Rsol remains almost constant. It is also evident that there is no clear trend for the Qporous/oxide values, however Rporous/oxide values show a tendency to lower values as immersion time increases. This may be explained by an increase of the pitted area or by increased acidification of the inner pit environment. The Qdl values tend to increase significantly with immersion time with the exception of the value obtained after 144 h. Increasing CPE of the porous layer during the course of immersion may be due to the growth of pits and subsequently increase in the real surface area or, as mentioned above, due to acidification of the pit environment [148]. It should also be noted that, initially, Qdl had a value of 20.5 Ω−1 sn cm−2 ×10−6, but with time this 94 increased by an order of magnitude and the n values decreased. This is also reported [147] to be a characteristic of increased surface roughness and increased inhomogeneity in the current density across the sample’s surface caused by pitting corrosion. The decrease in Qdl and increase in Rct by the end of the experiment might be attributed to the formation of a surface layer of corrosion products that partially block the pits, providing an extra protection to the surface. For as-cast samples (Table 6-4) it can be observed that Qdl increases with increasing time but the Rct remains approximately unchanged. However, there is an increase in Rct after 144 hours of immersion, which may again be explained by the formation of the corrosion products layer on the surface. The decrease in W may be due to adsorption or precipitation of reaction products with the pit/defects. The increase in Rsol with increasing immersion time may be due to saturation of reaction products near the surface. 95 Table 6-3: Impedance parameters data obtained by the ZSimpWin software for RRA B206 alloy at different immersion times in artificial seawater at 10 °C Sample Time (h) Rsol (Ω cm2) Rporous/oxide (kΩ cm2) Qporous/oxide (Ω−1 sn cm−2 ×10−6) nporous/oxide R2 (kΩ cm2) Qdl (Ω−1 sn cm-2 ×10−6) ndl RRA 4 7.6 4.79 9.6 0.92 5.02 20.5 0.95 12 8.0 3.27 14 0.89 3.02 63.3 0.92 24 8.08 2.87 44.6 0.79 1.94 99.9 0.98 48 8.38 2.90 28.1 0.80 1.85 244.5 0.84 72 7.93 2.33 65.8 0.73 1.31 326.3 0.84 96 8.0 2.03 44.7 0.72 2.33 212.2 0.76 120 8.2 1.81 73.8 0.74 1.91 301.6 0.55 144 8.03 1.19 41.5 0.72 3.38 129 0.55 Table 6-4: Impedance parameters data obtained by the ZSimpWin software for as-cast B206 alloy at different immersion times in artificial seawater at 10 °C Sample Time (h) Rsol (Ω cm2) Rct (kΩ cm2) Qdl (Ω−1 sn cm−2 ×10−6) ndl W (Ω−1 s0.5) × 10−3 As-cast 4 6.87 4.12 38.1 0.81 6.2 12 8.97 4.7 22.6 0.84 6 24 8.98 4.3 30.3 0.8 4.6 48 8.96 4.38 39.8 0.76 4.7 72 8.96 4.12 48.9 0.74 3.8 96 8.95 3.88 58.3 0.72 2.8 120 9.03 3.9 65.5 0.71 3.2 144 10.42 5.71 73 0.68 2.4 96 Comparing Table 6-3 and Table 6-4 and the Nyquist diagrams of the as-cast and RRA samples at various immersion times, presented in Figure 6-10, it is clear that the polarization resistance for the RRA sample is higher than that of the as-cast samples up to almost 48 hours of immersion. Figure 6-10: Nyquist diagrams of RRA and as-cast B206 alloy after immersion for (a) 1, (b) 48 and (c) 144h of immersion in artificial seawater at 10 °C It is well known that the higher the polarization resistance, the better the corrosion resistance [149]. Therefore, it can be concluded that, initially, the RRA samples had a lower corrosion rate and resisted the onset of localized corrosion (diameter of the semicircle in the Nyquist plot remained larger for the first 2 days). However, once localized corrosion started the corrosion rate of the RRA alloy was higher than that of the as cast. Therefore, the RRA treatment increases the corrosion resistance of the B206 alloy only for short immersion times. However, prolonged immersion could decrease the (a) (b) (c) 97 corrosion resistance of RRA and the resistance may get even worse with respect to the as-cast samples. The decrease in the size of the capacitive arc with time can also suggest a weakening of the protective action of the film covering the metallic surface, probably due to progressive dissolution or deterioration of the air-formed film in contact with the aggressive artificial seawater. The relationship between polarization resistance (Rp) and the immersion time for as-cast and RRA B206 is plotted in Figure 6-11. Figure 6-11: Relationship between Rp and immersion time for as-cast and RRA alloys in artificial seawater at 10 °C. Rp values for the RRA sample are higher for the first 20 hours than those of the as-cast B206. This may be due to the size and the amount of precipitates at the grain boundaries and also within the grains. Rp values for the RRA samples decreased with immersion time but the as-cast samples had relatively constant Rp. The Rp values for both samples are quite close after 48 hours of immersion. These values are in a good agreement with the results obtained from EIS (Table 6-3 and Table 6-4 and Figure 6-11) discussed above. 98 Weight loss measurements Weight loss measurements of as-cast and RRA alloys in artificial seawater were made after four exposure periods: 24 h, 24 h (total 48h), 144 h (total 192 h) and 72 h (total 264 h) in artificial seawater at 10 °C. When removed from the solution, a white corrosion deposit was observed on the surface of both samples (Figure 6-12). Figure 6-12: Pictures of the as-cast and RRA coupons after (top) 24 and (down) 144 hours After the first 24 hours of immersion, the amount of deposit on the as-cast samples was larger than the product formed on the RRA samples. Prolonging immersion, the deposit was concentrated within the holes of the RRA samples, whereas those on the As-cast RRA 99 as-cast samples covered a large surface area. The cumulative mass loss as a function of time for both alloys is shown in Figure 6-13. The points refer to the average mass loss of the two samples after subsequent exposure time. Figure 6-13: Mass loss for as-cast and RRA B206 alloy in artificial seawater at 10 °C. Error bars represent the standard deviation for an average of cumulative mass losses of two examined samples The mass loss is found to increase for both alloys. It is also clear that the mass loss of the RRA samples is lower than the as–cast alloy in the initial 24 hours. However, according to the Nyquist plots (Figure 6-10 (b)), the RRA samples are more corrosion resistant than the as-cast samples for the first 48 hours of exposure. The results shown in Figure 6-13 are the cumulative mass loss and the mass loss at each point in this figure is the delta from the initial weight of the samples. The results shown were also the average of the mass loss measurements for two samples. These might be the reasons for the difference between the results of mass loss and EIS measurements. It can also be seen that the increase in the mass loss of the RRA samples after 192 hours is comparable to that of the as-cast samples. It seems that the corrosion products formed on the surface partially protect it, which leads to similar mass loss increase as compared to the as-cast samples. 100 In order to confirm the corrosion tendency of these samples under ambient conditions during immersion, the corrosion current density of these samples was also measured from accelerated electrochemical test (linear polarization resistance; LPR) method by using following equation: 𝑖𝑐𝑜𝑟𝑟 =𝐵𝑅𝑝 {E- 6-5} where icorr is the current density in A cm−2, Rp represents polarization resistance in Ω cm2 and B is the approximated value as 0.026 V [86]. Table 6-5 provides the quantitative information about the corrosion current density obtained from the electrochemical experiments. Table 6-5: Corrosion rates of as-cast and RRA B206 obtained from LPR experiments at different times in artificial seawater at 10 °C Alloy Time (h) LPR icorr (μA cm−2) As-cast 48 5.5 72 4.6 144 4.2 RRA 48 5.4 72 5.7 144 5.9 Surface morphology of the exposed RRA and as-cast samples The EIS experiments showed that during initial 48 hours of exposure, the corrosion behavior of the RRA sample is better than the as-cast one (Figure 6-10 and Figure 6-11). After 48 h, the icorr reaches higher values than that of as-cast samples (Table 6-5). As mentioned above, the main difference in the structure of these two 101 samples is the variation in the microstructural features i.e. the morphology of the precipitates along the grain boundaries and within the grains. Figure 6-14 presents SEM micrographs of RRA B206 after 48 hours of immersion in seawater solution. The localized dissolution can be clearly seen along the surface of the sample. Birbilis et al. [43] demonstrated the electrochemical behavior of various intermetallic particles in chloride media. They also reported the Ecorr values of various intermetallic compounds. They found that the compounds containing Cu, Fe and Ti were nobler (having more positive Ecorr) than the α–Al matrix and do not affect the stability of the passive film. On the other hand, the intermetallic compounds composed of Mg, Si, and Zn were found to be more active (presented more negative Ecorr) than the matrix and may corrode preferentially and could disrupt the continuity of the passive film. This electrochemical heterogeneity in the microstructure of the RRA sample is quite evident in the Figure 6-14. Based on the experimental evidence and with support from the literature, the high pitting corrosion tendency of the RRA sample is expected by the formation of more electrochemically active intermetallic phases in its microstructure. It is therefore suggested that electrochemical heterogeneity in the microstructure would promote the development of localized galvanic cells at the surface of RRA sample and the large cathode to anode site area ratio could accelerate the pitting corrosion. This behavior could also be predicted from the decrease in the Rp (Figure 6-11). In the previous chapter on the mechanism of localized corrosion of as-cast B206, it has also been demonstrated that pits are more likely to initiate on the surface of the intermetallic particles. Therefore, some intermetallic particles initially behave as anodic regions with respect to the matrix, hence giving rise to selective dissolution of Al alloy. 102 As a result of this preferential dissolution of active intermetallic particles, there is an increase in the concentration of Cu enriched intermetallic phases within the microstructure, which found to be more cathodic than the Al matrix. Figure 6-14: (a, b and c) SEM micrographs of RRA B206 after 48 hours of immersion in seawater solution at 10 °C In other words, the corrosion of these alloys will follow the initial localized dissolution of active phases within the matrix depending on the electrochemical nature and concentration of the intermetallic particles. Hence, when the fraction of the precipitates is decreased by RRA treatment it is reasonable to expect better corrosion resistance. However, after a certain amount of time, which may be considered as an incubation period for the dissolution of particles and/or the particle/matrix boundary [150], the corrosion proceeded more rapidly as compared to as-cast sample. One of the (a) (b) (c) 103 main features of the RRA treated sample is the corrosion along grain boundaries. Intergranular corrosion, rather than isolated pits, provides fresh surface for solution penetration and reaction [151]. It can also be noted that the smaller grain size of the RRA samples reduces the distance between the particles and the matrix, which lead to an increase in the corrosion rate. When the cross section of the immersed sample is examined after 48 hours of immersion, the exfoliation corrosion was observed as shown in Figure 6-15. There has been extensive work on the exfoliation of Al alloys [53,150,152–156]. In general, exfoliation corrosion of these alloys is expected if they have higher susceptibility to intergranular corrosion [155]. Although the RRA B206 is an equiaxed material, the exfoliation corrosion of RRA can be seen Figure 6-15, which presents the preferential intergranular dissolution of the subsurface region after exposure to artificial seawater. Intergranular progression of the attack beneath the surface may lead to exfoliation corrosion of this alloy, due to the short distances between the hardening particles at the boundaries. Generally, as the immersion time increases and the corrosive attack proceeds, the corroded areas get interconnected beneath the surface. The corrosion products can accumulate at the grain boundaries and exert pressure between the grains causing delamination of thin surface layers. A microstructural study on the RRA heat treated 7075 aluminum alloy [62] mentioned that trapping hydrogen at the matrix/particle interface could result in hydrogen-assisted intergranular corrosion. In previous studies on 2024 aluminum alloy [157–159], it has been demonstrated that hydrogen produced during the corrosion process diffuses towards the bulk of the material and is trapped at preferential locations in the 104 interior of the aluminum alloy and subsequently gives rise to hydrogen embrittlement. Microstructural observations and decreasing corrosion resistance during the course of immersion in the present work could be the indirect evidence of hydrogen-assisted intergranular corrosion in RRA B206. Such a mechanism would proceed via transport of corrosive solution deep into the material through an intergranular network, where it would react to produce hydrogen. Thus, hydrogen could be generated, for example, at the bottom of pits and may spread to the adjacent unaffected material, establishing a hydrogen diffusion region below the corrosion zone. However, it should be noted that further work is required to definitely prove this mechanism. Figure 6-15: (a, b and c) SEM micrographs of the cross section of the RRA B206 alloy after 48 hours of immersion in seawater solution showing exfoliation corrosion (a) (b) (c) 105 Summary Electrochemical, weight loss and microstructural studies are performed on the RRA and as-cast B206 alloy to evaluate the effect of heat treatment on the corrosion behavior of B206 alloy at different immersion times in seawater solution. The uniformly dispersed fine θ\" precipitates were revealed in RRA B206, based on TEM analysis, which are expected to nucleate from GP zones during the aging process. According to EDX analysis, RRA samples contain Al-Cu-Mn-Fe and Al-Cu-Mn-Fe-Si precipitates. One of the main features of the RRA sample is the corrosion along grain boundaries. Chloride ions in seawater solution can easily destroy the surface film to induce pitting corrosion, and finally form an occluded cell. Thus, an acidic medium generated in the corrosion zone due to the hydrolysis of the aluminum cations leads to grain-boundary attack. When the cross section of the immersed sample is examined, IGC, more specifically, exfoliation corrosion seems to be the mechanism of corrosion propagation. EIS results for the RRA and as-cast alloys and, in particular, their evolution with time, indicate that applying RRA heat treatment increases the corrosion resistance of the B206 alloy at short immersion times. However, prolonging immersion could change this behavior and the resistance may get even worse with respect to the as-cast samples. Results obtained from LPR experiments and weight loss measurements also confirm this behavior. This study was begun hoping that RRA treatment could decrease the corrosion susceptibility of B206. However, it seems that the aging characteristics of this alloy are 106 different from the 7xxx series alloys. Although the long-term corrosion behavior of the heat-treated samples is more or less the same as the as-cast samples, it may be beneficial to apply heat treatment on the B206 in terms of having better mechanical properties. It has been reported that the hardness of the RRA B206 alloy (150-170 Hardness Vickers Number (HVN)) is almost 80 HVN more that as-cast samples (80-90 HVN) [64]. 107 7 7. B206 Passive layer characterization in deaerated artificial seawater at different temperatures using EIS and Mott-Schottky Analysis Previous chapters in this thesis have discussed in detail the corrosion mechanism and the effect of RRA heat treatment on the corrosion behavior of the B206 alloy. One of the protection methods for Al alloys is the use of the passive region to create a protective layer. This can be done through anodizing the surface before installation in the marine environment, or by ensuring the continuous presence of a passive condition with an anodic protection system [32]. Anodized aluminum or aluminum alloys have been shown to exhibit an excellent resistance to localized corrosion compared to their as-received counterparts, especially in aqueous neutral chloride solution [66]. The thickness, porosity, breakdown resistance, and semi-conductive behavior of the anodized passive layer are all important properties, which determine its performance in preventing further corrosion. In this chapter, EIS is used to characterize the corrosion protection performance of anodized B206 in artificial deaerated seawater of 4 different temperatures between 6 and 25 C – both direct correlations of charge transfer resistance with corrosion rate and determinations of passive layer thickness are presented. The Mott-Schottky technique is used to identify the donor density of the passive layer formed on anodized samples in deaerated seawater at the above-mentioned temperatures. Electrochemical studies at different temperatures and DO conditions Figure 7-1 displays the polarization curves for the B206 alloy in artificial seawater with different DO conditions (naturally aerated, and deaerated) at 10 C. 108 Figure 7-1: Potentiodynamic polarization curves for aluminum B206 in naturally aerated and deaerated artificial seawater at 10 °C The Ecorr value of B206 in the naturally aerated solution is higher than that in deaerated solution; the difference is ~0.5 V. The change in the DO condition influences the cathodic process and subsequently, the corrosion potential. During potentiodynamic polarization, the pitting potential (Epit) is normally defined as the potential at which the anodic current increases significantly with further applied potential [86], as labeled in Figure 7-1. Above the Epit, the oxide layer ruptures at random heterogeneities therein. The barrier layer created by the oxide is unable to repair itself, and hence localized corrosion develops at these points [6]. It is also obvious in Figure 7-1 that the anodic polarization curve of B206 presents an active-passive-breakdown transition in the deaerated environment, while the corresponding curve in the naturally aerated environment indicates that localized corrosion of B206 can occur under open circuit conditions. The change in cathodic reaction by introducing oxygen in the solution causes an increase in the corrosion potential close to the pitting potential. Deaerated solution Naturally aerated solution 109 The average values of the electrochemical parameters extracted from the polarization curves are Ecorr, Epit, icorr, and the passive current density (ipass). The Ecorr and icorr parameters were measured by Tafel extrapolation of the anodic and cathodic lines. The values for all the extracted parameters are listed in Table 7-1. Table 7-1: Electrochemical parameters extracted from potentiodynamic polarization curves for B206 in naturally aerated and deaerated artificial seawater at 10 °C Alloy Deaerated Aerated B206 Ecorr (V vs.Ag/AgCl) Epit (V vs.Ag/AgCl) ipassive (µA cm−2) Ecorr (V vs.Ag/AgCl) icorr (µA cm−2) −1.15 −0.49 2.64 −0.55 5.92 Comparing results of naturally aerated and deaerated conditions shows that corrosion potential in naturally aerated solution is very close to the pitting corrosion potential. The effect of seawater temperature on the polarization curves of B206 was investigated in deaerated and naturally aerated conditions. Figure 7-2 (a) and (b) show the polarization curves for the B206 alloy at temperatures ranging from 6 to 25 °C with deaerated (Figure 7-2 (a)) and naturally aerated (Figure 7-2 (b)) artificial seawater. Figure 7-2: Polarization curves for aluminum B206 in different temperatures in (a) deaerated and (b) naturally aerated artificial seawater. From Figure 7-2 (a), it is evident that the temperature has little effect on Epit. (a) (b) 110 However, the passive current generally increases with increasing temperature in deaerated solution. It is also noted that the increase in passive current density appears more pronounced at lower temperatures; whereas clear separation of the passive region is observed between the 6 °C and 10 °C profiles, no significant difference is seen between the 14 °C and 25 °C profiles even though the latter two profiles have a larger temperature difference than the former profile pair. It should be stated though that this may simply be an artifact of the x-axis scaling. On the order of 10−6 A cm−2, because a logarithmic scale will not show graphical separation of higher current densities (e.g. between 8 and 9) as much as lower current densities (e.g. between 2 and 3), comparative differences appear greater than they really are. This interpretation is confirmed by the upcoming EIS results of the anodized passive layer, which identifies a difference in the impedance (hence the passive current density) of the 14 °C versus the 25 °C environments. Passive layer behaviour The discussion in sections 7.2.1-7.2.4 below are based on the results of the multi-stage anodizing, EIS, and Mott-Schottky test routines previously described in 4.2.3; These tests were conducted to evaluate the protective structure and behavior of B206’s passive layer anodized in deaerated artificial seawater at temperatures between 6 and 25 C. 7.2.1 Passive current density decay The current density decay transients measured during anodizing at −0.65 VAg/AgCl are shown in Figure 7-3. 111 Figure 7-3: Passive current density (ipass) decay transients measured during the anodizing of B206 at −0.65 VAg/AgCl in deaerated seawater at 6, 10, 14 and 25 °C for 6 hours These are considered passive current densities (ipass) since the anodizing potential falls within the passive region of the corresponding potentiodynamic polarization curves in Figure 7-2 (a). Clear separation in the ipass transients in Figure 7-3 is observed based on the temperature of the solution. The magnitude of ipass increased at higher temperatures of the same electrolyte and DO level, indicating decreased passive layer protection and increased corrosion rate with temperature. The smoothest and most uniform transient was that which corresponded to the 6 C solution, reaching a stable current density of approximately 3 A cm−2 after around 6000 s of anodizing. The passive current density in 10 C solution reached a slightly higher current density, yet only stabilized towards the end of the 6 h anodizing period. This longer stabilization period also existed for the 14 C transient, in addition to visibly more unsteadiness in the ipass measurement especially towards the beginning of the anodizing process. This is indicative of instability of the passive layer formation, an occurrence which manifested itself the most in the 25 C current density transients during the first 10,000 s of anodizing. Hence, not only do the final ipass values reached after 6 h 112 of anodizing show the less robust passive layer(s) formed at higher temperatures, the clear instability in the transients during the earlier development stages allude to more sluggish and possibly less uniform passive layer growth. 7.2.2 Passive layer components and structure Aluminum oxide (Al2O3) forms immediately upon exposure of Al with monatomic or polyatomic oxygen [32,65,160]. In atmospheric air containing diatomic oxygen, this proceeds according to Reaction {E- 7-1}. The presence of Al2O3 on the surface of the B206 specimens hence precedes their immersion in the seawater solution of the test cells. Through anodizing in the alkaline seawater solution, the surface of the aluminum alloy becomes covered with a duplex structure consisting of an inner anhydrous Al2O3 layer adjacent to the substrate surface and an outer hydrous layer of aluminum hydroxide (Al(OH)3) corrosion product [161,162]. Both components of this dual-structure thicken with continued anodizing at the passive −0.65 VAg/AgCl potential in these alkaline solutions [161–164], leading to a decrease in the anodic corrosion (Reaction {E- 7-2}) of the base Al in the B206 alloy. 4Al + 3O2 → 2Al2O3 {E- 7-1} Al → Al3+ + 3e− {E- 7-2} In deaerated conditions, the formation of an Al(OH)3 corrosion product follows a three-step process involving intermediary Al(OH)2+ and Al(OH)2+ complexes as shown in Reactions {E- 7-3} to {E- 7-5} [161,164]. The corresponding cathodic reaction is the reduction of water as shown in Reaction {E- 7-6}. The formation and continual thickening of the Al(OH)3 corrosion product increases the impedance of the overall dual- 113 layer barrier, enhancing the corrosion protection of the aluminum alloy substrate. The strength of this Al(OH)3 corrosion product is dependent on the rate of its hydroxide-accelerated dissolution through Reaction {E- 7-7} versus its ongoing formation culminating in Reaction {E- 7-5} [162]. Al + OH− ↔ Al(OH)2+ + 3e− {E- 7-3} Al(OH)2+ + OH− ↔ Al(OH)2+ {E- 7-4} Al(OH)2+ + OH− ↔ Al(OH)3 {E- 7-5} 3H2O + 3e− → 3OH− +32 H2 {E- 7-6} Al(OH)3 + OH− ↔ Al(OH)4− {E- 7-7} The stabilization of the ipass transients shown in Figure 7-3 can therefore be explained as the slow decay of the ratio between the formation and dissolution rate of Al(OH)3 to unity following a period where the formation rate exceeded the dissolution rate. However, since the ipass transients at all temperatures tested eventually reach horizontal stability, the differences in final ipass values for different temperatures cannot be explained by the formation vs. dissolution ratios. Rather, it is suggested here to be the result of structural (i.e. thickness and porosity) and electronic (i.e. semiconductive) differences in the duplex-layer as evidenced by the EIS and Mott-Schottky results below, respectively. 7.2.3 EIS Figure 7-4 (a) shows the Nyquist impedance representations of the EIS impedance spectra for anodized (6 hours) B206 specimens in deaerated seawater at temperatures between 6 and 25 C. 114 Figure 7-4: (a) Nyquist impedance representations and (b) Bode phase angle and impedance moduli of the EIS impedance spectra obtained for anodized B206 specimens in deaerated seawater at temperatures between 6 and 25 °C. Figure 7-4 (b) illustrates the corresponding Bode phase angles and impedance moduli for the same EIS tests. At the coldest temperature of 6 C, it is seen that the Nyquist profile manifests as a nearly complete depressed semicircle, with the real component of the impedance (Zre) reaching a value slightly greater than 3 x 104  cm2 at the lowest frequency. Although the adsorption of Cl− on the surface of anodized Al specimen in Cl− containing electrolytes is an established phenomenon [161,165,166], both the Nyquist and Bode plots at 6 C reveal the absence of significant auxiliary processes such as adsorption or induction at lower frequencies at this temperature. Based on the duplex structure of the passive layer known to form on Al in alkaline media [161–164], as described in the previous section, an electrical equivalent EEC with film/layer resistance and CPE (Rf and Qf, respectively) is proposed to model the EIS result at this temperature. This is shown in Figure 7-5 with a parallel charge-transfer resistance (Rct) and double-layer capacitance (Cdl) in series to model the solid-electrolyte interface at which electron transfer occurs. A schematic of the corresponding proposed physical system is also illustrated. (a) (b) 115 Figure 7-5: Proposed EEC and physical occurrences for EIS on anodized B206 in deaerated artificial seawater at 6 °C CPEs were used instead of ideal capacitor elements due to the surface heterogeneities. The fitting results of the EEC in Figure 7-5 for the 6 C EIS spectrum are shown in Table 7-2 and are overlaid on Figure 7-4 (a) and (b). 116 Table 7-2: EIS component values for B206 anodized at −0.65 V vs. Ag/AgCl in seawater solution at 6, 10, 14 and 25 °C Components Solution temperature [°C] 6 10 14 25 Rsol [Ω cm2] 6.14 14.2 4.58 8.41 Qf [Ω−1 sn cm−2] 1.83 x 10−6 2.10 x 10−5 2.46 x 10−5 9.55 x 10−5 nn 1 0.92 0.93 0.96 Cf [F cm−2] 1.83 x 10−6 8.67 x 10−6 11.35 x 10−6 61.30 x 10−6 Rf [Ω cm2] 2.96 x 104 2.44 x 104 2.32 x 104 2.19 x 104 Qdl [Ω−1 sn cm−2] 1.15 x 10−3 1.43 x 10−3 2.14 x 10−3 1.01 x 10−2 ndl 0.84 1 1 0.92 Rct [Ω cm2] 2.44 x 104 1.56 x 103 8.58 x 102 8.12 x 102 W [Ω−1 s0.5] / 1.47 x 108 8.99 x 107 1.7 x 107 𝝌𝟐 4.22 x 10−4 1.10 x 10−3 1.34 x 10−3 2.15 x 10−3 % error in fit < 2.06 < 3.32 < 3.66 < 4.64 The results show a very good correlation between the EEC and the experimental measurements, determined by 𝜒2 value of 4.22 x 10−4 and a maximum error in fit below 2.06%. It is noted that an EEC with a nested parallel adsorption (i.e. {R(Q(R(QR))}) was attempted in modelling the 6 C EIS spectrum, yet resulted in no significant fitting correlation difference compared to the EEC in Figure 7-5. However, when a simpler {R(QR)} as in [167] was attempted, a drop in fitting accuracy was observed. This indicates that a two time-constant EEC best suits the EIS spectrum measured at 6 C. Figure 7-4 (a) also shows the Nyquist and Bode plots for the EIS conducted at 10, 14, and 25 C. In the Nyquist plots at these temperatures, much more noticeable supplemental electrochemical processes can be identified at low frequencies < 0.1 Hz. This clearly indicates a change in governing mechanism; based on the characteristic linear slope segments at low frequencies in the Nyquist representations, this is likely due 117 to diffusion-controlled behavior. Combined with an apparent decrease in impedance values in these spectra compared to that of the 6 C condition, it strongly appears that a change in the structure of the passive layer occurs at higher temperatures, allowing the mass-transfer of electro-active species therein. It has been described earlier that a duplex passive layer structure exists on the anodized B206 specimen, where a compact inner anhydrous layer of Al2O3 resides between the alloy substrate and a porous outer hydrous Al(OH)3 layer. In Cl−-containing electrolytes, the adsorption of Cl− onto the outer Al(OH)3 layer has been established on numerous accounts by other researchers [161,165,166]. In a recognized work by Nguyen and Foley [168], the reaction activation energy between Cl− and the intermediary Al(OH)2+ complex of Reaction {E- 7-3} was shown to be lower than that between Cl− and Al3+. Hence, adsorbed Cl− on the outer layer of the passive film reacts preferentially with the former versus the latter, ultimately thinning and breaking down the Al(OH)3 layer. The ensuing breakdown of the Al(OH)3 layer under the influence of Cl− results in AlCl4− through the multi-step process of Reactions {E- 7-8} to {E- 7-11} shown by McCafferty in [169]: Al(OH)3 + Cl− ↔ Al(OH)2Cl + OH− {E- 7-8} Al(OH)2Cl + Cl− ↔ Al(OH)Cl2 + OH− {E- 7-9} Al(OH)Cl2 + Cl− ↔ AlCl3 + OH− {E- 7-10} AlCl3 + Cl− ↔ AlCl4− {E- 7-11} In a fundamental investigation by Kolics et al. [170], it was shown that this Cl−-induced adsorption on the Al passive layer is pH dependent, where the Cl− concentration of the surface decreases with increasing pH. Although the pH of the electrolytes in the 118 present work do not change substantially with temperature, an increase in temperature still increases the rate of Al(OH)3 dissolution into AlCl4− simply based on the temperature influence in the Arrhenius law. As such, even at a constant anodizing potential of −0.65 VAg/AgCl and a constant Cl− concentration in the anodizing solution, at increased temperatures the thinning and pitting of the outer Al(OH)3 ensues at a faster rate; This is shown to drastically decrease the film resistance [161], increase opening in pores allowing the diffusion of electro-active species therein, and hence increase the corrosion rate of the Al metal or alloy substrate [171]. Based on the manifestation of mass-transfer controlled behavior at 10, 14, and 25 C in the EIS spectra of Figure 7-4, the EEC of Figure 7-6 is proposed for modelling the electrochemical behavior at these temperatures. 119 Figure 7-6: Proposed EEC and physical occurrences for EIS on anodized B206 in deaerated seawater at 10, 14 and 25 C A corresponding schematic of the proposed physical occurrences is also shown in Figure 7-6. Here, the combined influence of increased temperature and aggressive Cl− ions in the alkaline solution are suggested to thin and pit the Al(OH)3 outer layer, creating larger channels for the diffusion of key electroactive species during the ensuing corrosion of the substrate. The Nyquist spectra at 10, 14, and 25 C show near characteristic Warburg diffusion element (W) behavior. 120 The fitting results of the EEC in Figure 7-6 for the 10, 14, and 25 C EIS spectra are shown in Table 7-2 and are overlaid on Figure 7-4. Again, the results show a good correlation between the EEC and the experimental measurements, with 𝜒2 on the order of low 10-3 values and a maximum error in fit below 4.64%. Overall, steady decreases in Rct and Rf values with increased temperature demonstrate the increased corrosion caused by the decreased passive layer protectiveness at higher temperatures. 7.2.4 Mott-Schottky and passive layer thickness The Mott-Schottky plots of anodized B206 in seawater solutions at 6, 10, 14, and 25 C are shown in Figure 7-7. The Mott-Schottky technique relies on the assumption that at high applied frequencies, like the 1 kHz used in this work, the capacitance at the passive layer’s interface with the electrolyte mainly represents the space-charge layer capacitance (Csc) [172]. This is because Csc is considered to be much less than the capacitance of the Helmholtz layer. Therefore, the Csc−2 vs. potential representation of the Mott-Schottky plot in Figure 7-7 can be used to describe the semi-conductive behavior of the depletion region in the passive layer anodized on B206 at the four temperatures shown. Figure 7-7 clearly depicts increasing Csc with temperature and a linear relationship between the capacitance and the applied potential in the passive potential region investigated. The positive slopes (S) of the Mott-Schottky plot at every temperature tested indicate the n-type semi-conductive behavior of the passive layer, agreeing with the vast majority of electrochemical studies on Al2O3 including [173–175]. 121 Figure 7-7: Mott-Schottky plots at 1 kHz between −0.8 VAg/AgCl and −0.5 VAg/AgCl after 6 hours of anodizing in deaerated seawater at 6, 10, 14 and 25 °C (linear fits and slope values overlaid) For n-type semiconductors, the charge distribution as a function of applied potential can be determined through {E- 7-12}, where Nd represents the donor density, Efb is the flat-band potential, k is the Boltzmann constant, T is the temperature in Kelvins,  is the dielectric constant of the passive layer ( ≈ 10 for Al alloys [176]), 0 is the vacuum permittivity (8.85 x 10−14 F cm−1), and e is the elementary charge (1.6 x 10−19 C). 1Csc2 =2εε0eNd(E − Efb −kTe) {E- 7-12} Thus, Nd for the passive layer anodized on B206 in deaerated seawater can be determined from the corresponding 2𝜀𝜀0𝑒𝑁𝑑 term or slope S in the experimental Csc−2 vs. potential plot of Figure 7-7. Using the linear fit slope values overlaid on the figure, Nd values are calculated and plotted in Figure 7-8. 122 Figure 7-8: Donor density and passive layer thickness values for B206 anodized for 6 hours in deaerated seawater at temperatures between 6 and 25 °C, calculated from Mott-Schottky and EIS results, respectively. The Nd values are seen to increase with the temperature of the anodizing solution. These Nd values are comparable to the densities on the order of 1020 cm−3 measured for the passive layers of other Al alloys in the literature [167,176,177]. Based on the point defect model [178,179], increased Nd indicates an increase in defects in the anodized passive layer with temperature. In light of the established attribution of n-type behavior to metal interstitials and oxygen vacancies in the oxide layer [165,176,177], the trend in Figure 7-8 specifically reveals an increased density of oxygen vacancies in the Al2O3 inner layer at higher temperatures. This increased oxygen vacancy density at higher temperature is directly correlated to increased conductivity and thus decreased protectiveness of the Al2O3 layer [180], agreeing with the trend of passive current density results of Figure 7-2 (a) and Figure 7-3, and the decreased Rf results with temperature shown in Table 7-2. Thus, it is shown here that increasing the temperature of the 123 anodizing solution not only weakens the Al(OH)3 outer layer on B206, it also leads to a less resistive Al2O3 inner layer due to a more defective electronic structure containing a higher density of oxygen vacancies. The role of increased temperature in the reduction of the anodized film’s protection is further supported through calculation of the passive layer thickness. As overlaid on the Nd results of Figure 7-8, a decrease in the thickness of the passive layer with temperature was found through calculations using the Power-Law (P-L) model [75,181–183]. The model converts the CPE parameter for the passive layer (Qf in Table 7-2) to an effective capacitance for the layer represented by Cf, through a relationship involving the largest measured frequency ( 𝑓𝑚𝑎𝑥 ), and the exponent n describing the capacitive idealness of the CPE [183]: Cf = 𝐠Qf(2π𝑓𝑚𝑎𝑥)n−1 𝐠 = 1 + 2.88(1 − n)2.375 {E- 7-13} (a) {E- 7-13} (b) The 𝐶𝑓 values corresponding to the 𝑄𝑓 results at each temperature determined from modeling the EIS results are shown in Table 7-2. The thickness of the anodized passive layer (df) on B206 after 6 hours is then calculated through {E- 7-14}, where A is the effective surface area of the specimen and the remaining 𝜀, 𝜀0, and 𝐶𝑓 parameters are the same as what has been previously described: df =εε0ACf {E- 7-14} The decreasing df with increasing temperature can be seen in Figure 7-8 which corroborates previously discussed results from Mott-Schottky, EIS, and polarization of both the potentiostatic and potentiodynamic type. 124 Summary The properties and growth of corrosion products on the B206 sample in deaerated artificial seawater at different temperatures are studied using EIS, i decay analysis, and Mott-Schottky. Electrochemical studies on B206 demonstrated that the effect of dissolved oxygen is more significant than that of temperature. The anodic polarization curves present an active-passive-breakdown transition in deaerated environment. While in an aerated environment, localized corrosion of B206 can occur under open circuit conditions. The anodic polarization curves present an active-passive-breakdown transition in a deaerated environment. While in an aerated environment, localized corrosion of B206 can occur under open circuit conditions. The protectiveness of the outer Al(OH)3 layer formed through anodizing B206 in deaerated seawater decreases with increased temperature in the 6 – 25 °C range due to the higher rate of Cl−-induced Al(OH)3 dissolution. Increased diffusion of electroactive species with temperature suggests the enhanced breakdown of Al(OH)3 at higher temperatures, forming larger sized pores reaching the Al2O3 inner layer. Increasing the temperature of the deaerated seawater anodizing solution increases the donor density (mainly oxygen vacancies) of the Al2O3 inner layer, further contributing to the reduced resistivity and protectiveness of the overall duplex-structure passive layer on B206. The temperature-dependent trend in the passive layer thickness supports the decreased protectiveness (due to smaller thickness) at higher temperatures in the 6 – 25 °C range. Therefore, the results in this chapter may support the choice of anodizing for hub components to reduce corrosion. However, the degree of protection 125 afforded by anodization will clearly depend on the turbine placement location (colder and deeper when the amount of dissolved oxygen is negligible). 126 8 8. Investigation of cathodic disbondment of an epoxy coating on B206 substrate under different levels of cathodic protection in seawater The results in chapter 6 on the effect of retrogression and re-aging (RRA) heat treatment, which has been shown to improve the corrosion and mechanical properties of 7xxx series aluminum alloys [57,58,60,61], on B206 have shown that heat-treated samples are as susceptible to corrosion as the as-cast ones over long immersion periods. Therefore, heat treatment is not a promising route to increase the corrosion resistance of these alloys, thus, further protection is necessary to limit corrosion rates. On the other hand, as mentioned in chapter 7, in a naturally aerated environment, localized corrosion of B206 can occur under open circuit conditions and the anodic polarization curve of B206 does not show the active-passive transition. Furthermore, the thickness of the passive layer decreased with increasing temperature. Therefore, anodizing the surface to protect the alloy from corrosion would only be useful if the temperature of the seawater and the depth of the turbine were such as to preclude elevated temperatures and high oxygen concentrations. As mentioned in section 2.2, a combination of cathodic protection and protective coating(s) is the most effective and the most economical method of corrosion control [33,34]. Applied coatings are not perfect and will contain defects, which threaten the proper protection of the substrate. On such a coated structure subject to CP, the cathodic current protects the metal surface at the bottom of the coating defect. However, application of CP in these situations could accelerate the deterioration of the coating and cause coating disbondment [33]. Therefore, the protection effect will only be obtained if 127 the CP system is compatible with the chosen coating. Thus, the choice of cathodic protection potential becomes very important. In this chapter, first, results obtained from the immersion of bare B206 samples in natural seawater (Esquimalt harbor, Victoria, BC, Canada) for 1 year are presented to demonstrate the extent of the corrosion problem for this alloy. Second, we investigate the effect of various CP potentials on coating disbondment using electrochemical measurement techniques and microstructural analysis. EIS is used to characterize the corrosion protection performance of an epoxy coating combining with CP. The coating disbondment under different levels of CP is addressed both with electrochemical techniques and optical and SEM characterization. Corrosion of the unprotected samples immersed in natural seawater Mass loss and corrosion rate measurements of the B206 samples subjected to a natural seawater environment were conducted on the samples following the total immersion time of one year (sample collection for weighing occurring at 38, 80, 129, 180, 224, 273, 318 and 366 days). When removed from the seawater, white corrosion deposits were observed on the surfaces of the samples (Figure 8-1). 128 Figure 8-1: B206 corrosion samples following 42 days (80 days total) exposure, before cleaning The deposits on the B206 samples were small, and numerous, covering a large surface area of the sample. Even after scrubbing with a toothbrush, adherent corrosion products remained on the surface and in the pits of these samples. Following cleaning, pitting was generally observed under where these deposits had been. Although not quantified in the present work, there did appear to be an increase in both the size and number of pits following each exposure period, as shown in Figure 8-2. 129 Figure 8-2: Pictures of the coupons for after 38, 80 and 129 days (a-c) for B206 After each immersion interval, the samples were cleaned and weighed, and then returned to the seawater for additional exposure. The cumulative mass loss as a function of time for the B206 is shown in Figure 8-3. With each subsequent exposure period, the mass loss is found to increase almost linearly. Examination with an optical microscope showed that the corrosion attack on the B206 was confined to the dendrites (Figure 8-4). The corrosion products, which were scraped off of the samples analyzed by EDX, consisted primarily of aluminum, oxygen, and silicon. (a) (b) (c) 130 Figure 8-3: Cumulative mass loss against time for aluminum B206 in natural seawater. Mass loss from initial is the mass loss data obtained from the initial mass measurements (before any immersion in seawater) and mass loss from previous is the mass loss data obtained from mass measurements before and after each immersion intervals. Figure 8-4: Corrosion attack morphology for aluminum B206 after (a) 38 and (b) 80 days of immersion in natural seawater General corrosion rates can be calculated from the mass loss data however, as it can be seen in Figure 8-4, localized attack occurred at the surface of B206 in natural seawater. Therefore, the calculation of the corrosion rates from the mass loss data and comparing the obtained data with the electrochemical experiments in artificial seawater would not result in an accurate comparison [184]. Nevertheless, the steep increase in the measured weight loss data with increasing immersion time shows that B206 alloy will (a) (b) 131 corrode rapidly in natural seawater and hence, would require a protection method such as CP to have a better long-term in-service performance. Cathodic corrosion protection potential studies Figure 8-5 shows the cathodic polarization curve of the B206 at 10 ºC in artificial seawater solution. Figure 8-5: Cathodic polarization curve of B206 at 10 ºC in seawater It can be seen that, as the potential shifted from the corrosion potential to the cathodic direction, a section of concentration polarization resulting from the DO reduction reaction and a section of active polarization resulting from hydrogen gas generation appeared as follows, respectively: O2 + 2H2O + 4e− ↔ 4OH− {E- 8-1} 2H2O + 2e− ↔ H2 + 2OH− {E- 8-2} The current density is relatively stable between these two regimes. If the electrode potential is lower than the potential of hydrogen evolution, the cathodic current density Total current O2 reduction H2 Evolution 132 will increase rapidly. The intersection of the Tafel extrapolation for the HER and ilim for the ORR (~10−5 A cm−2) showed that this potential is ~ −1.02 VAg/AgCl. Thus, based on the results above, the range for the corrosion protection potentials can be identified from −0.65 VAg/AgCl (which is close to the corrosion potential) to ~ −1 VAg/AgCl in which the ORR appears to be the dominant cathodic reaction. The current densities in this potential range had very low values of ~10−5 A cm−2. Figure 8-6 shows the current density change with time obtained by conducting potentiostatic experiments in the cathodic region in artificial seawater. The current density at the potential of −0.60 VAg/AgCl showed a tendency of slow decrease, however, as this potential is very close to OCP, this current is still under activation control. It can be seen that there is a big difference between the current density values at potentials of −0.60 and −0.65 VAg/AgCl. The current densities for the potentials of −0.70, −0.75, −0.80, −0.85 VAg/AgCl were similar and just the results for −0.85 VAg/AgCl are shown in Figure 8-6. Generally, the current densities for the applied potentials in this range rapidly increased at the beginning of the experiment and stabilized at around −20 to −22 µA cm−2 after 6000 seconds. It is clear that the reaction in this range is limited by the diffusion of oxygen to the surface and the cathodic current density is independent of the potential. As discussed for the cathodic polarization curve (Figure 8-5), it is considered that the current densities increased under the effect of activation polarization from −1 VAg/AgCl which is a turning point between the concentration polarization due to ORR and the activation polarization by hydrogen gas generation. However, based on the potentiostatic polarization results, it seems that the current densities began to increase at potentials around ~ −0.95 VAg/AgCl. Potentiostatic polarization curves at more cathodic potentials 133 clearly showed that the cathodic reaction is dominated by the HER. As can be seen in Figure 8-6, the current density at −1.05 VAg/AgCl increases and reaches ~ −26 µA cm−2. It can be concluded that the corrosion protection potentials for B206 should be −0.70 to −0.90 VAg/AgCl as the oxygen reaction is the dominating cathodic reaction in this potential range. Figure 8-6: Potentiostatic polarization curves of B206 with various applied potentials in seawater solution at 10 ºC Long-term EIS measurements on coated B206 samples 8.3.1 Samples without an artificial defect Figure 8-7 (a) and (b) show Bode plots obtained from the coated sample without artificial defect at OCP and under cathodic potential of −0.7 VAg/AgCl for different exposure periods. 134 Figure 8-7: Bode plots obtained from the intact coated sample (a) at OCP, (b) under CP potential (−0.7 VAg/AgCl) and (c) impedance evolution at 0.01 Hz frequency, respectively (at OCP and at −0.7 VAg/AgCl CP) for different exposure periods. The impedance data for all samples were collected periodically during 30 days of exposure to seawater solution. A decrease in the magnitude of the impedance can be seen with exposure time at OCP. However, the impedance of the coated samples under CP did not change significantly with the exposure time. The total impedance values measured at low frequencies are usually used to evaluate the corrosion resistance of the system in coating and CP studies [90,98,185,186]. In both cases, Bode plots show a relatively high impedance (impedance value of 108-107 Ω cm2 at low frequencies) at the beginning of exposure. The large magnitude of the impedance (>1 × 107 Ω cm2) and the 90° phase angle are considered to be characteristics of a good coating in the literature [93]. However, it should be mentioned that the main goal of this chapter was to determine the effect of CP on the protective properties and any possible disbondment of the coating at 135 different cathodic potentials. Therefore, the results of this study should not be implied to infer a method of determining the adhesion, physical properties or quality of the particular coating that has been used. As it can be seen in Figure 8-7 (b), the impedance values at low frequencies and the phase angle remain close to 108 Ω cm2 and 90°, respectively, which show that the coating keeps its resistivity and capacitance characteristics at −0.7 VAg/AgCl CP for the entire immersion period. The system resistance was determined from the |Z|at10mHz which indicates the total resistance (solution, charge transfer and coating resistance) of the system (Figure 8-7 (c)). As seen from Figure 8-7 (c), the degradation of the system is faster at OCP and the impedance decreases to values close to 106 Ω cm2 while it remains almost constant at −0.7 VAg/AgCl CP. Therefore, with increasing immersion time, the coating with CP showed better corrosion resistivity and the difference between |Z|at10mHz values for coated samples with and without CP are nearly one order of magnitude at the end of the experiments. The large amplitude of the impedance (>1×107 Ω cm2) even after 30 days of exposure suggests that almost no corrosion products were produced and accumulated at the metal/coating interface when CP was applied. EIS data collected for coatings without an artificial defect but with CP showed that long immersion periods were needed to obtain coating degradation and, thus to properly evaluate the compatibility between coatings and cathodic protection. Therefore, electrochemical measurements were conducted on the coated samples with an artificial defect. 136 8.3.2 Samples with an artificial defect The Bode magnitude measured for the defect samples at −0.7 VAg/AgCl in Figure 8-8 (a) shows that the impedance is lower than that of the intact coated samples (obviously because of the artificial 1 mm diameter defect). Yet, this difference is relatively small (one order of magnitude) and suggests that a larger contribution from the coating itself (rather than the bare substrate) was seen in the impedance data of the defect sample. The impedance of the coated metal in the frequency range of 0.1-0.01 Hz with artificial defect increased up to 10 days of exposure and then decreased slightly to the end of the exposure period. Another observation regarding the Bode magnitude graphs is that the impedance response of the defect samples did not change after 20 days of exposure to the solution. Increasing |Z|at10mHz values up to 10 days show the compatibility of the coating and the CP leading to postponement of the disbondment. The decrease of |Z|at10mHz can be attributed to a scale deposited form the seawater upon alkalization from CP. It is also notable that the decrease in |Z|at10mHz is not significant. Thus, it can be assumed that the artificial defect in the coated sample was partially blocked due to the deposits. This would protect the substrate and prevent a significant decrease in the sample impedance, which will be further discussed. The phase angle is a very sensitive indicator of coating damage. Poor separation of time constants in the Bode phase plot can be seen after 1 day of exposure, as shown in Figure 8-8 (b). However, after 5 days of exposure, the separation of the time constants became more obvious at high and medium frequencies. 137 Figure 8-8: (a) Bode magnitude, (b) Bode phase angle and (c) Nyquist diagrams collected for defected coated samples after 1 h and 5, 10, 20 and 30 days of exposure to −0.7 VAg/AgCl The proposed circuit shown in Figure 8-9 (a) was used to fit the obtained EIS data. According to Figure 8-9 (a), Rsol represents the solution resistance (not considered as an important parameter in the analysis of coating performance). Rcoat (coating resistance), which is often used in a general equivalent circuit for coated metals, indicates coating resistance at a constant phase element (CPEcoat). Rd indicates resistance at the bare substrate at the artificial defect in parallel with the corresponding CPEd. CPEs were used instead of ideal capacitors to account for frequency dispersion due to surface heterogeneities. (a) (b) (c) 138 Rsol: Simulated seawater solution resistance Rcoat: Coating resistance CPEcoat: Constant phase element corresponding to coating capacity Rd: Resistance of the corrosion products at artificial defect CPEd: Constant phase element corresponding to the corrosion products inside the artificial defect Rdis: Polarization resistance at disbonded area Cdis: Capacitance corresponding to the corrosion products at disbonded area Figure 8-9: EEC used to fit EIS data for coated samples with defect under (a) −0.7 and (b) −0.9 VAg/AgCl To convert a constant phase element, the coating and double layer capacitances are calculated using the following equation [149]: 𝐶 = 𝑃1𝑛 𝑅1−𝑛𝑛 {E- 8-3} In the above expressions P is the magnitude of the CPE and n is the deviation 139 parameter. Table 8-1 summarizes electrochemical parameters obtained by analyzing EIS data. Table 8-1: The coating capacitance and resistance of the B206 sample under −0.7 VAg/AgCl CP at different exposure time derived from analysis of EIS data measured -0.7 V vs. Ag/AgCl Time (hour) Ccoat(µF cm−2) Rcoat(Ω cm2) 𝝌𝟐 24 0.52 6.1 x 105 3.43 x 10-3 120 0.22 1.9 x 104 3.46 x 10-3 240 0.12 2.8 x 104 1.27 x 10-3 480 0.03 2.1 x 104 1.02 x 10-3 720 0.03 1.9 x 104 8.66 x 10-3 It can be seen from Table 8-1 that Rcoat decreased with the passage of time, which indicates the degradation of the coating with time. The coating capacitance data shown in Table 8-1 reveals that there is a slight decrease in the capacitance of the coated samples up to 10 days. However, the coating capacitance decreased rapidly from 10 to 20 days. This rapid decrease has been reported to be attributed to the filling of defects by corrosion products [185]. The protection current was recorded during the exposure and the results are presented in Figure 8-10 (a). The decrease of the current densities might be related to the formation of deposits at the defect as a consequence of the alkalinity. It may be concluded that the variation of the current results from the competition of the two opposite processes; the increase in the corrosion process due to the coating degradation and the decrease of the bare metal area due to the formation of a visible deposit having limited protective properties. 140 Figure 8-10: Protection current densities as a function of time for the coated sample with an artificial defect polarized to (a) −0.7 V Ag/AgCl and (b) −0.9 VAg/AgCl. After around 2 weeks of polarization at −0.7 VAg/AgCl, visible white deposits were observed at the hole in the coated sample. SEM micrographs of this sample after the whole exposure period (Figure 8-11) showed that the corrosion damage is largely confined to the surface of the hole, with the start of some damage on the hole wall. It can also be seen that the porous corrosion products partially blocked the defect area. These micrographs also confirm the assumptions that have been made to explain the increase in the impedance and decrease in the current densities during the course of exposure. 141 Figure 8-11: (a and b) SEM micrographs indicating some porous corrosion products partially blocked the defect area on the coated surface after 30 days of exposure to seawater solution under −0.7 VAg/AgCl CP The EIS results for the coated samples at a CP potential of −0.8 VAg/AgCl showed little or only a slight difference compared to the ones at −0.7 VAg/AgCl, suggesting that the corrosion response of the coated sample will not change with an increase of 0.1 VAg/AgCl in CP. Figure 8-12 shows the Bode magnitude and phase plots for the holiday samples at −0.9 VAg/AgCl (the potential close to the limit potential of −1 VAg/AgCl) at different times of exposure in artificial seawater. It can be found in Figure 8-12 that the impedance behavior is different for the defect coated sample under more cathodic potentials. Figure 8-12: (a) Bode magnitude and (b) Bode phase angle collected for defect coated samples after different times of exposure to −0.9 VAg/AgCl Regarding Bode phase diagrams, the poor separation of time constants in the Bode phase plot after 1 day of exposure at CP of −0.7 VAg/AgCl has been changed to two (a) (b) (b) (a) 142 obvious separate time constants. The main difference between the impedance behavior of the defect coated sample at −0.7 and −0.9 VAg/AgCl is the presence of a third time constant at low frequencies of ~ 0.1 Hz with an increase in test duration. Generally, the appearance of more than one time constant reflects the diversity of the phenomena that occur in the system. Coated systems are usually defined as models in which the high frequency part characterizes the coating behavior while the lower frequencies describe the behavior of the metal in direct contact with the electrolyte at the bottom of the pores or at defects or disbonded areas [187]. Therefore, the appearance of the third time constant after immersion for 5 days, can be interpreted to correspond to a corrosion reaction under the coating, suggesting that the coating was disbonded and that the corrosion process occurred under the disbonded coating. This is supported by the observation of coating disbondment in the optical and SEM micrographs and will be discussed later. It should also be noted that the time constants (after 5 days) at high frequencies are not well defined and as obvious as the ones at −0.7 VAg/AgCl. At the early stage of exposure, the impedance suddenly drops to lower values, resulting in a loss of protection, and then shifts upward in Bode magnitude (the Bode phase plot disclosed two time constants at this stage). Afterwards, an increase in the Bode magnitude indicates an improvement in corrosion resistance. Between 5 and 30 days of exposure, the Bode phase behavior and the impedance magnitude at low frequencies do not show a significant change. However, at 30 days of exposure a slight decrease in impedance can be seen. As mentioned above, the |Z|at10mHz values indicate the total resistance. It might be deduced that the corrosion occurred substantially at defects in the first day of exposure and the accumulation of the corrosion products (up to 5 days) at defects decreased the corrosion 143 rate. It can also be inferred that at the initial stage of exposure (between 0 and 5 days) of the coated specimen with holiday at −0.9 VAg/AgCl the coating disbondment started and the corrosion resistance decreased. However, with an increase of the exposure time to 5 days, accumulation of the corrosion products on the bare substrate under the disbonded area led to a decrease in the rate of the corrosion process. As it is mentioned above, the difference between the electrochemical behavior of the samples at −0.7 VAg/AgCl and −0.9 VAg/AgCl CP is the appearance of a third time constant at low frequencies as a result of a new reaction at −0.9 VAg/AgCl. Thus, using EEC shown in Figure 8-9 (a) (used for samples at −0.7 VAg/AgCl) would not lead to a good fit for samples at −0.9 VAg/AgCl and it requires to add another RC. Therefore, the EEC shown in Figure 8-9 (b) was used to fit the obtained EIS results at −0.9 VAg/AgCl. In this model, Rdis and Cdis indicate the resistive and capacitive response of the disbonded area. The suitability of the EEC was indicated by the Chi-squared (χ2) value (corresponding fitting values had the lowest χ2 values (lower than 10−4) for EEC shown in Figure 8-9 (b)) as well as the good agreement between the fitted and experimental EIS plots (Figure 8-13). This indicates that the equivalent circuit {Rsol(CPEcoat(Rcoat(CPEd(Rd(CdisRdis)))))} is suitable for the investigated systems. The results are presented in Table 8-2. 144 Figure 8-13: Example of a fitted result for a selected EIS diagram (5 days) Table 8-2: EIS parameters fitted from the response of the coated sample with defect at −0.9 VAg/AgCl −0.9 V vs. Ag/AgCl Time hour Qcoat (Ω−1 sn cm−2) n Rcoat (Ω cm2) Qd (Ω−1 sn cm−2) n Rd (Ω cm2) Cdis (F cm-2) Rdis (Ω cm2) 24 1.5 x 10−6 0.73 26280 1.6 x 10−5 0.95 2.4 x 105 - - 120 1.1 x 10−6 0.67 11770 2.6 x 10−7 0.81 3.5 x 105 1.1 x 10−5 3.8 x 105 240 9.6 x 10−7 0.64 8265 2.6 x 10−7 0.79 4.4 x 105 1.3 x 10−5 4.2 x 105 480 2.8 x 10−9 0.99 976 1.4 x 10−6 0.52 1.8 x 105 1.0 x 10−7 9.5 x 105 720 4.2 x 10−9 0.99 890 1.7 x 10−6 0.54 7.9 x 104 6.1 x 10−8 1.1 x 106 The corrosion resistance values of the tested samples were determined from the magnitude of the impedance data at 0.01 Hz as a function of immersion time as well as the coating resistance and capacitance, and the values are presented in Figure 8-14. 145 Figure 8-14: Corrosion resistance as a function of (a) coating resistance, (b) coating capacitance and (c) impedance at 0.01 Hz frequencies with exposure time for the coated sample with artificial defect at −0.9 VAg/AgCl 146 The coating resistance is reduced significantly from 26280 to 978 Ω cm2 over 20 days of exposure, indicating that the electrolyte was able to penetrate through the coating. However, in the last 10 days of exposure the rate of decrease became insignificant and the coating resistance stayed constant. On the other hand, the coating capacitance decreased first followed by a slight increase in the last 10 days of exposure. The increase in coating capacitance is the result of electrolyte penetration into the coating. Both of these findings suggest a general degradation of the coating after 20 days of immersion at −0.9 VAg/AgCl CP. The resistance of the B206 coated sample at −0.9 VAg/AgCl (Figure 8-14 (c)) shows a relatively complex trend: it decreases gradually in the initial 24 h of exposure, then it increases sharply with exposure time up to 120 h of exposure followed by a gradual increase to 630 hours and falls again at 720 h. In the early stage of exposure in artificial seawater, corrosion at the defect and degradation of the coating under cathodic polarization results in the reduction of impedance. However, with passage of time, the bare surface of the alloy at the defect and damaged interface area of the substrate/coating can be covered with corrosion products, which can protect the system from further corrosion. The relatively stable impedance values from 408 to 636 hours can be due to this process. However, it seems that after immersion for 636 h, the corrosion products were penetrated by the electrolyte, resulting in the sudden decrease of impedance. The protection current was recorded during the exposure of the coated sample with an artificial defect polarized to −0.9 VAg/AgCl and the results are presented in Figure 8-10 (b). Within the first day of exposure, the protection current densities showed a large decrease to approximately 0.6 µA cm2 followed by a sharp increase (indicating 147 decrease of the coating resistance) and then reached relatively steady state values (~ 0.8 µA cm2) after almost 10 days. The decrease of the current densities might be related to the formation of the corrosion products as a consequence of the alkalinity. According to the EIS experiments, it is proposed that disbondment of the coating occurred under −0.9 VAg/AgCl CP. The optical micrographs of the coated samples under −0.7 and −0.9 VAg/AgCl after 30 days of exposure to solution (Figure 8-15) revealed that no coating disbondment occurred at −0.7 VAg/AgCl CP. However, for the sample at −0.9 VAg/AgCl CP it seems that corrosion phenomena started from the defect and the interface between coating and substrate and propagated through the substrate under the coating as indicated by color changing to red around the defect. Figure 8-15: Optical images of coated samples with an artificial defect after exposure for 30 days in artificial seawater at 10 °C under (a) −0.7 VAg/AgCl and (b) −0.9 VAg/AgCl. Coating disbondment can be seen under −0.9 VAg/AgCl CP 148 The cross-sectional back-scattered electron micrograph of the coated sample at −0.9 VAg/AgCl in Figure 8-16 shows the coating disbondment and corrosion product development beneath the coating in the vicinity of the artificial defect. Figure 8-16: (a) SEM micrograph of a cross-section of the coated B206 under −0.9 VAg/AgCl CP for 30 days in seawater solution and (b) Schematic representation of the delamination process In the EIS measurements shown in Figure 8-12, the formation of these corrosion products was detected as the presence of a third time constant at low frequencies. Figure 8-16 (b) shows the schematic representation of the disbondment process. Finally, in order to evaluate the coating behavior at over protection potentials, the coated sample was exposed at −1.1 VAg/AgCl (in the hydrogen evolution region). The EIS response is shown in Figure 8-17. Low impedance values (<1 × 106 Ω cm2) can be seen from the beginning. The Nyquist diagrams of the sample at this potential show the formation of an elaborate semicircle that is consistent with a coating time constant at high frequencies, implying access of solution to the metal surface and thus the occurrence of coating disbondment. The high frequency time constant (related to the coating properties) eventually exceeded the measured frequency range. This can be easily seen in the Bode phase plot of the coated sample after 408 hours of exposure. This evolution suggests that the barrier properties of the coating generally degraded with time. It can be seen that the 149 second time constant relating to electrochemical reactions at metal surface (at low frequencies) can be well distinguished from the very early stage of cathodic disbondment and the second semi-circle on the Nyquist plot is recognizable. This implies more advanced cathodic disbondment occurred at more negative potential [93]. Comparing the Bode phase angle diagrams for the sample at different exposure times, the elimination of the time constant at high frequencies is observable. This indicates that extending the exposure time probably resulted in the decrease in the contribution of the coating impedance to the overall impedance and thus the impedance data from the disbonded metal surface becomes dominant. The magnitude of impedance at this stage drops to values around 3 × 104 Ω cm2. Figure 8-17: Electrochemical response of the coated sample with an artificial defect under −1.1 VAg/AgCl (a) Bode magnitude, (b) Bode phase and (c) Nyquist graphs 150 Summary Data obtained from weight loss measurement for 1 year in natural seawater showed that B206 would corrode rapidly and, therefore, needs further protection to safely be used in-service conditions. EIS, as an in-situ monitoring technique, has been used rather than the visual inspection methods to get more reliable data. The obtained results were confirmed with SEM micrographs. Therefore, EIS can be used as an adequate technique for determination of the disbondment of an epoxy coating under different levels of CP potential. Table 8-3 summarizes the obtained results regarding the electrochemical behavior of the B206 coated samples with and without a holiday in artificial seawater at different levels of CP. The results indicate that application of CP improves the protective property of the epoxy coating as the coating |Z|at10mHz values stayed around 108 Ω cm2 for the entire test period. The effect of CP potential levels on the degradation of the coating containing an artificial defect has also been investigated using EIS. The cathodic disbondment has been found to be highly dependent upon the applied CP potential. During the exposure of the coated sample with an artificial defect at −0.7 VAg/AgCl CP, EIS results showed two time constants (at high and medium frequencies), which indicates that the coated sample shows a stable corrosion resistance and the substrate under the coating did not corrode at −0.7 VAg/AgCl during the whole test. However, application of −0.9 VAg/AgCl on the coated sample results in coating disbondment as evidenced by a third time constant appearing at the low frequencies, corresponding to a corrosion reaction and a reduction in the protective properties of coating. The disbonded area was also observed via optical and SEM micrographs. 151 Comparing the Bode phase angle diagrams for the sample at different exposure times at −1.1 VAg/AgCl, the elimination of the time constant at high frequencies is observable which indicates that the impedance data from the disbonded metal surface becomes dominant and HER becomes significant. The findings in this chapter show the performance and the compatibility of CP with the chosen coating in the protection potential range of −0.7 to −0.9 VAg/AgCl. Coating disbondment at −0.9 VAg/AgCl was shown based on both EIS and SEM results. No disbondment was observed at −0.7 VAg/AgCl. Therefore, the combination of CP at −0.7 VAg/AgCl and the coating can be used as an effective corrosion protection method for the hub component in practical applications. 152 Table 8-3: Summary of the obtained electrochemical and microstructural results for the coated B206 alloy with and without an artificial defect at different cathodic potential levels Potential (VAg/AgCl) With defect Without defect OCP ---  |Z|at10mHz value of ca. 108 Ω cm2 at low frequencies at the beginning of exposure  Decrease in the magnitude of the impedance from 7×107 to 3×106 Ω cm2 in 744h  The degradation of the system is faster compared to the system with CP −0.7  Two time constants in Bode phase plots at high (corresponding to the coating behavior) and medium frequencies (corresponding to the capacity of the coating/substrate)  The decrease in |Z|at10mHz is not significant  The coated sample shows a stable corrosion resistance and the coating disbondment under it did not corrode at −0.7 VAg/AgCl based on optical and SEM micrographs  |Z|at10mHz values stayed around 108 Ω.cm2 for the whole test period  |Z|at10mHz values and the phase angle stayed ca. 108 Ω cm2 and 90°, respectively, for the whole test period show that the coating keeps resistivity and capacitance characteristics at −0.7 VAg/AgCl CP for the entire immersion period −0.9  Appearance of the third time constant in Bode phase plots at low frequencies  Coating disbondment by showing a third time constant, corresponding to corrosion reaction  The disbonded area also observed in optical and SEM micrographs --- −1.1  Low impedance values ca. 1×106 Ω cm2  Elimination of the time constant at high frequencies indicates that the impedance data from the disbonded area becomes dominant and hydrogen evolution reaction becomes significant --- 153 9 9. The effect of casting porosity on the changes in local chemistry inside the pores Corrosion properties of aluminum alloys, which owe their corrosion resistance to the formation of passive films on their surface, might be affected by processing defects such as porosity. These defects result in the localized breakdown of the passive film and pitting corrosion [106]. Moreover, the presence of porosity will increase the exposed surface area in the solution [107]. In order to study the effect of casting porosity on the corrosion rate of B206, potentiodynamic polarization experiments were tried on specimens that contained different size pores (qualitatively smaller and larger pores) from different sections of the casting. However, those measurements did not show any significant difference in corrosion behavior of the specimens. As noted in section 2.3, it seems that the 1D-artificial pit technique (pencil electrode) can be used to simulate the casting porosity to understand the effect of these defects on corrosion properties. However, one of the major experimental limitations encountered for aluminum was that the pit is frequently blocked with hydrogen bubbles during testing [120]. Therefore, as shown in Figure 2-11, the anodic dissolution process for aluminum and aluminum alloys may result in mass transport control (formation of a salt layer) only at high anodic potentials in highly concentrated AlCl3. Therefore, at low chloride concentration, the pit growth follows the ohmic law and, thus, the anodic current cannot be determined from the potentiodynamic polarization experiments. However, using the experimental set-up shown in Figure 4-9 to collect the evolved hydrogen gas during the potentiostatic polarization of the B206 pencil electrodes at different anodic potentials can help to determine the aluminum anodic dissolution current using {E- 4-4}. 154 Therefore, this research, for the first time, will show how the corrosion behavior of the B206, might be affected by porosity. The aim of this work is to measure the effect of the porosity on B206 dissolution kinetics. The attribution of the corrosion behavior to the changes in local chemistry inside the pores is a new field of research, which will represent a fundamental contribution to corrosion science. Determining the QH2/Qanode, conductivity and potential of the pore bottom Potentiodynamic polarization curves for B206 pencil electrode in seawater solution at 22 °C did not show any mass transport limited conditions even at high potentials of 3 VAg/AgCl. This behavior is reasonable since it has already been reported in the literature that the convection induced by hydrogen evolution prevents the formation of a continuous salt film and, therefore, the pit growth will be limited only by ohmic resistance in the solution and not mass transport [116,117]. Thus, the charge associated with anodic dissolution of aluminum (Qanode) cannot be calculated from the potentiodynamic polarization curves. However, as it is mentioned above, it can be determined from {E- 4-4} (𝑄𝑛𝑒𝑡 = 𝑄𝑎𝑛𝑜𝑑𝑒 − 𝑄𝐻2in which QH2 can be determined using the experimental set up shown in Figure 4-4). (The volume of the hydrogen gas evolved at different potentials can be found in Table D-1 in Appendix D.) Figure 9-1 shows the effect of applied potential on Qnet, QH2 and Qanode (the same as what has been described in section 4.2.1) for the 3600 s potentiostatic polarization in artificial seawater at 22 °C. This figure can be used to calculate the ratio of the iH2/ianode. As seen in Figure 9-1, the iH2/ianode for B206 pencil electrodes can be calculated to be around 30%. In literature [117,120,188], the ratio of iH2/ianode was estimated to vary from 10% to 15% for pure 155 aluminum. However, the aforementioned studies used pure aluminum as the pencil electrode. The difference between the materials that have been used might be the reason. They assumed that the charge associated with oxygen reduction is negligible. However, as discussed earlier in chapter 0, the intermetallic particles in B206 directly support the oxygen reduction at their surface. Therefore, although, we made the same assumption in this study, the contribution of the ORR to the cathodic reactions for B206 might be higher with respect to pure aluminum in the previous studies. Figure 9-1: Effect of applied potential on Qnet, Qanode and QH2 in artificial seawater at 22 °C Under pure ohmic control, the net current density is given by the ratio of the voltage drop to the resistance: 𝑖𝑛𝑒𝑡 = 𝛥𝛦𝐴𝑅=𝛥𝛦𝑅𝑠𝛿 {E- 9-1} where inet (A cm−2) is the net current density, E (V) is the potential difference that is creating the ohmic potential drop, A is the surface area in cm2, R (Ω) is the average solution resistance in the pore, Rs (Ω cm) is the average specific resistance of the 156 electrolyte in the pore, and δ (cm) is the depth of the pore. The net current density must be considered instead of total anodic current density because the ohmic potential drop results only from the net current flowing out of the pore. The ratio of the local cathodic current density to the anodic current density is labelled Rca, (which is shown in Figure 9-1 to be around 30%). Therefore, inet = (1- Rca) ianode [120]. The total anodic current density can be defined by Faraday’s law, as shown in {E- 4-5}. Therefore: 𝑖𝑛𝑒𝑡 = (1 − 𝑅𝑐𝑎)𝑖𝑎𝑛𝑜𝑑𝑒 = (1 − 𝑅𝑐𝑎)(𝑑𝛿/𝑑𝑡)(𝜌𝑛𝐹/𝑀) {E- 9-2} where t (s) is time. Rearranging {E- 9-1} and {E- 9-2} gives: 𝛿𝑑𝛿 = △ 𝐸𝑀𝑑𝑡(1 − 𝑅𝑐𝑎)𝑛𝐹𝑅𝑠𝜌 {E- 9-3} Integration of {E- 9-3} yields: 𝛿2 =2𝛥𝛦𝑀𝑡(1 − 𝑅𝑐𝑎)𝑛𝐹𝑅𝑠𝜌 {E- 9-4} Substituting{E- 9-4} back into {E- 9-1}: 𝑖𝑛𝑒𝑡 =𝛥𝛦𝑅𝑠𝛿= [(1 − 𝑅𝑐𝑎)𝜌𝑛𝐹𝛥𝛦2𝑅𝑠𝑀]1/2𝑡−1/2 {E- 9-5} Rearranging yields: 1𝑖𝑛𝑒𝑡2⁄ =2𝑅𝑠𝑀𝑡(1 − 𝑅𝑐𝑎)𝜌𝑛𝐹𝛥𝛦 {E- 9-6} The value of 1/i2net as a function of time is shown in Figure 9-2. In the initial stage of polarization, 1/i2 is linearly related to polarization time, in accordance with ohmic control. After 400 seconds or more, there is a transition to a steeper slope, and then later there is another transition where the slope decreases. Since conductivity is the reverse of resistivity, {E- 9-6} can be rearranged to give [120]: 157 𝑖𝑛𝑒𝑡2 𝑡 =(1 − 𝑅𝑐𝑎)𝜌𝑛𝐹ΔΕ2𝑅𝑠𝑀= (1 − 𝑅𝑐𝑎)𝜌𝑛𝐹𝜎(𝐸𝑎𝑝𝑝 − 𝐸𝑏𝑜𝑡)2𝑀 {E- 9-7} where σ (Ω−1 cm−1) is the average conductivity inside of the pore, Eapp is the applied potential relative to the reference electrode in the bulk solution and Ebot is the potential at the bottom of the artificial pore. The values of i2net t for the initial stage of dissolution (the reciprocal of the slope in Figure 9-2) are plotted in Figure 9-3. By fitting lines to the data in Figure 9-3, and applying {E- 9-7}, it is possible to extract values for the pore conductivity, σ, and the potential at the bottom of the pore, Ebot. The fitting lines are assumed to be linear with reasonable R2 values in order to estimate the slopes. However, it cannot be claimed that the best-fitting lines for this experiment are necessarily linear or if they are linear, the slopes of the lines will be the same as the ones estimated here. Due to the complexity of the experimental conditions and sample preparation, collection of more data was not possible. Figure 9-2: The change in the value of 1/i2 with time during the potentiostatic polarization at 0 VAg/AgCl in artificial seawater 158 Figure 9-3: The influence of applied potential on the ohmic portion of the current transients The relationship between the estimated conductivity inside the pore and the potential at the bottom of the pore is shown in Figure 9-4. It can be seen that the potential at the bottom of the pore decreased as the conductivity of the solution inside the pore decreased. The same observation was revealed by Akiyama and Frankel [120]. They pointed out that this observation is reasonable since the bottom potential will decrease with increasing ohmic potential drop. Another finding is the change in the potential at the bottom of the pore and the conductivity of the solution inside the pore. It can be seen that the potential at the bottom of the pore is decreasing with increasing pore size. Therefore, the difference between the potential at the bottom of the pore and the surface (ohmic drop) is higher for the larger pores. Thus, the ohmic dissolution of the metal surface at the bottom of the pore and, accordingly, the total current is increasing with increasing pore size. On the other hand, the conductivity of the solution inside the pore is lower for larger pores. The insulating nature of the hydrogen bubbles [120] my be the reason for the lower conductivity in the bigger pores. Akiyama [120] reported that the motion of the hydrogen bubbles in the pit or crevice will cause considerable mixing in the crevice and will 159 prevent the increase in ion concentration and resultant increase in conductivity. The lower conductivity of the solution inside the bigger pores can result in decreasing the transportation of the ions from inside and outside of the pore, which causes the solution inside the bigger pores to be more aggressive with respect to the one in the smaller pores. This could be another reason for the higher corrosion rate of the samples with larger porosities. It should be noted that the measurements, which lead to the results in Figures 9-3 and 9-4, were conducted only twice (due to the limited number of samples) and to confirm reproducibility further experiments would be needed. Figure 9-4: The relationship between the estimated conductivity and bottom potential Pore geometry and dissolution rates It was possible to calculate the dissolved pore length using the true anodic current (using the previously described method and equations in section 4.1.3). In reality, casting pores can have complex geometries, which are often different from the simple hemispherical case usually considered in pit investigation studies. In our simplified system, the time evolution of a typical net current transient can be seen in Figure 9-5. Noise can be attributed to the hydrogen bubble motion inside the pores. 160 Figure 9-5: Current-time transition for 200 micron pencil electrode at 22 °C in artificial seawater solution As mentioned above, from the net current time variation it was possible to extract the anodic current corresponding to the dissolution of the B206 pencil electrode and to plot the change in the pore depth deduced from Faraday’s law as function of time (Figure 9-6 (a)). From Figure 9-6 (b), it can be seen that the dissolution rate decreases linearly with increasing pore depth. It can also be seen that the changes in the dissolution rate along the pore depth are slower for larger pores (e.g. the dissolution rate is almost constant along the whole pore depth for the 1 mm pore diameter). It can be deduced that the conductivity of the solution inside the pores decreases by increasing the pore depth, which may lead to a prevention of the passage of the anodic current through the pore depth. It should also be mentioned that for the narrower pores (200 and 300 µ diameter) the dissolution rate does not change significantly after a specific depth. For aluminum alloys, it has been previously mentioned [118] that the hydrogen bubbles stir the electrolyte at the corroding pit or crevice, which allows renewal of the solution and migration of the ions from inside or outside of the dissolving pit or crevice. Thus, from 161 the result of this study it might be concluded that at specific depth, the precipitation of the corrosion product or formation of the salt film cover the pore bottom, which would influence the anodic dissolution and, therefore, corrosion rate. Therefore, the morphology of the pore might influence the corrosion process however, it should be noted that the result of this study cannot definitely prove this and further work is required. Figure 9-6: \"1D-artificial pit\" electrode dissolution kinetics. (a) Comparison of the time evolution of the pore depth and (b) Decrease of the dissolution rate (µm h−1) as function of the pore depth Summary The effect of casting porosity on the corrosion of B206 in artificial seawater was simulated using the 1D-pencil electrode technique. It is concluded that the chemistry of 162 the solution inside the pore greatly affects the corrosion behavior. Higher corrosion rates in the presence of the larger pores is shown to be related to the lower conductivity of the solution inside the pore compared to the smaller pores. The potential at the bottom of the pore also depends on the size of the pore. As a result, the potential drop in the larger pores is greater with respect to smaller ones. It seems that the metal surface dissolution kinetics are also dependent on the morphology of the pores. For the narrower pores, the decrease in the dissolution rate along the pore depth was higher compared to wider pores. It is also observed that the dissolution rate for narrower pores tends to reach a steady state at a specific depth which might be the result of precipitation of the corrosion products or formation of the salt film that covered the pore bottom. However, the results of this chapter cannot definitely prove this. This chapter builds on the knowledge of using pencil electrode in studying the kinetic parameters of aluminum anodic dissolution. The results of this chapter show that the use of the pencil electrode method can be extended to specific alloys rather than just pure aluminum. 163 10 10. Conclusion Key contribution and broad implication The investigated material throughout this thesis was B206, a high strength aluminum-copper casting alloy; being of the newer candidate materials to be used in the clean hydrokinetic energy generating systems as the hub component. At present, data regarding the corrosion properties and protection systems for B206 are not available. Therefore, this novel study of the surface/subsurface microstructural-induced corrosion of B206, contributes to further understanding of the corrosion mechanisms associated with this new alloy system. Investigation of feasible corrosion protection methods at various environmental conditions (e.g. seawater temperature and dissolved oxygen) adds further value toward improving the corrosion resistance and the effectiveness of B206 in practical applications. This study also provides some of the first long term corrosion data for B206 in a tidal marine environment. Overall, the contributions and implications to the field of corrosion science and protection made by the results of this work can be summarized in the following: 1. Throughout this thesis, the microstructural aspects of pitting and intergranular corrosion of B206 have been shown. Although cast aluminum alloys have a broad application in different industries, the roles of microstructure and de-alloying of the intermetallic particles in the corrosion properties of these alloys are largely ignored in previous work [7,8]. The dependence of the localized corrosion initiation, dealloying of the intermetallic particles and propagation of the localized attack on the subsurface microstructure are of importance. A few recent studies 164 have addressed this subject [28,127] and should warrant much more detailed attention.  The earliest stages of attack start with pit initiation at the surface of intermetallic particles resulting in de-alloying. De-alloying of the particle and matrix/particle interfacial attack produced a sponge-like remnant and cracks at the interface, respectively. Intergranular attack was not generally present over the alloy surface. However, propagating away from the initiation sites, corrosion developed preferentially along the grain boundary network beneath the alloy surface. This finding shows the significant role of the subsurface microstructure in corrosion studies, which has been given attention in limited literature just recently. This conclusion addresses objective 1, which were discussed in chapter 5.  Implications: The consequences of the different microstructures on the corrosion resistance of the aluminum alloys are broad. As corrosion is influenced by microstructure and thus processing, this thesis lays the groundwork for better understanding microstructure-corrosion relationship in such alloys. In future, such understanding will guide foundry engineers to modify the alloy microstructure, especially, with respect to the alloying additions to produce high quality castings from B206 in terms of reducing its susceptibility to corrosion. The subsurface examination at the localized attack sites can help to design a heat treatment to modify the grain structure and distribution of the second phase particles throughout the grains and hence, to reduce the corrosion susceptibility of the alloy. 165 2. Previous considerations of the corrosion properties of the B206 in the regarding the effect of microstructural evolution during this heat treatment process are incomplete and lacking.  The electrochemical results in chapter 6 show that although the corrosion resistance of the RRA heat treated alloy is much higher than that of as-cast microstructure for the short exposure period in seawater (48 hours according to EIS measurements), the RRA alloy corrodes more rapidly for longer immersion times. This thesis is the first to examine the microstructure evolution of B206 during RRA heat treatment. Implication: The RRA heat treatment may improve the mechanical performance of B206, however, the findings of this study imply that RRA should not be considered an appropriate solution to the corrosion protection of B206 in seawater. In fact, the findings of this study can contribute to adjust the RRA heat treatment parameters, in particular the time and temperature of the retrogression step, in such a way to improve both corrosion and mechanical properties of this alloy simultaneously. 3. The experimental results in chapters 7 and 8 indicate some corrosion protection methods for B206 alloy at different marine service environments to meet the objectives 3 and 4 in chapter 3.  The influence of anodizing the surface of B206 in deaerated seawater indicates the development of a more protective passive layer in colder seawater. The protectiveness of the outer Al(OH)3 layer formed through anodizing B206 in deaerated seawater decreases with increased temperature due to the higher rate of 166 Cl− induced Al(OH)3 dissolution. This thesis is the first to identify the passive layer behavior of the B206 alloy. However, in naturally aerated seawater, the electrochemical behavior of B206 shows that localized corrosion occurs on the alloy at open circuit potential. Therefore, once placed in a DO-containing marine service environment, B206 will be susceptible to localized corrosion and will need further protection such as the combination of coating and cathodic protection (CP). Although standards are available for cathodic protection systems for various alloys, the choice of CP potential is controversial in the literature in that the CP potential should be adjusted according to the environment/material relationship (e.g. temperature of the solution and type of coating). Throughout this thesis EIS, as an in-situ monitoring technique, has been used rather than the visual inspection methods. The findings in chapter 8 show the effectiveness and the compatibility of CP with the chosen coating. The combination of CP/coating at −0.7 VAg/AgCl can be used as the most effective corrosion protection method for the hub component.  Implication: The findings presented in this study support the feasibility of implementing anodizing and possibly anodic protection systems for B206 in specific service environments. Knowledge of the absence of notable passivation for B206 in DO-containing environments may support the choice of deeper placement locations for such turbines during their design, especially if an uncoated or anodized B206 material is to be used in the construction. A combination of CP and the chosen protective coating can be successfully applied on B206 alloy in practical applications. Therefore, in order to protect the surface 167 in real service conditions even relatively thin coatings (100 µm) can be applied on the surface of the hub component and further protected by impressed current CP. Furthermore, the findings throughout chapters 7 and 8 can be useful to develop the experiment-based corrosion model to evaluate the integrity of the hub component in varying marine environments and optimize the protection parameters (e.g. protection current and potential) with the choice of placement location (e.g. temperature of the seawater and depth). 4. The results in chapter 9 are the first to use the pencil electrode method to investigate the effect of the chemistry of the solution inside the pore on the corrosion of an aluminum-copper casting alloy.  The results show that the shape and size of the pores play an important role in the corrosion process of the alloy. For bigger pores, the conductivity of the solution inside the pore is lower and the difference between the potential at the surface and the bottom of the pore is higher. Therefore, the higher corrosion rate of the surfaces with bigger pores is not only due to an increase in the real exposed surface area, but also due to the chemistry of the solution inside the bigger pores. This finding has been largely overlooked in most studies regarding the effect of casting porosity prior to this work. Implication: The attribution of the corrosion behavior to the changes in local chemistry inside the pores is a new field of research, which can represent a fundamental contribution to corrosion science. This can be accomplished by linking the experimental data to a Through Process Model (TPM) to predict the 168 in-service life of the hub component and hence can be very valuable to prevent the unexpected failure of the materials. In summary, as corrosion is heavily influenced by microstructure and thus processing, prior to commencing any electrochemical experiments, the microstructure-induced localized corrosion was fully characterized. Susceptibility to pitting and intergranular corrosion was investigated using well-established techniques. Then, firstly, the RRA heat treatment was examined to assess whether this thermal process has any substantial effect on the corrosion properties of B206. Moving forward, to mitigate in-service hub component corrosion, the effectiveness of anodizing and CP/coating systems were examined. Finally, the pencil electrode method was used to simulate the effect of casting porosity on the corrosion behavior. Ultimately, the findings in this thesis can be used to generate a matrix of corrosion outcomes based on service environment and microstructure. This matrix will form a key component of the TPM for B206 which will be enable the concurrent casting process and component design for this high strength alloy. This optimization will result in both reduced costs for clean energy power generation and in expanding Canada’s capacity to shape cast high- strength aluminum alloys for a wide range of industrial applications. 169 Suggestions for future work Corrosion and protection of B206 is a complex process, which requires further investigation and research to understand and control. The following points are suggested as direction and areas for future research and development:  A comprehensive study on the composition and stoichiometry of the second phase particles in the B206 matrix should be conducted using quantitative EDX or chemical extraction techniques followed by X-ray diffraction (XRD). In this way, the electrochemical behavior of the second phase particles (whether they are anodic or cathodic with respect to the matrix) can also be determined prior to immersion in the corrosive media.  In chapter 5, Figure 5-15, a crack was found to form in the bulk material after immersion in artificial seawater solution. FIB/SEM could be performed on the as-received B206 (before any exposure to solution) to provide evidence of whether this microcrack was in the initial state of the material or had been created as a result of the corrosion process.  In order to determine the changes in the microstructure as a result of heat treatment, comprehensive microstructural studies using SEM or TEM should be performed at different steps in the RRA heat treatment.  In chapter 6 Figure 6-14, the dissolution of the particles at the grain boundaries was shown for the RRA samples. FIB/SEM cross-sectional examination should be performed to show how the localized corrosion propagated through the grain boundaries for RRA B206. 170  It was shown that the B206 alloy in RRA state is at least as corrosion resistant (not more) as the as-cast for long term immersion periods. However, it is beneficial to apply heat treatment on the B206 in terms of having better mechanical properties. Changing the parameters of the RRA heat treatment such as time and temperature of the retrogression and re-aging steps would influence the microstructure of the alloy and might affect the corrosion behavior of the heat-treated alloy in a better way. Working on the aforementioned parameters in conjunction with the microstructural studies would be beneficial in terms of improving the mechanical and corrosion properties of the B206, simultaneously.  Indirect evidence of the hydrogen-assisted intergranular corrosion in RRA B206 was shown in this thesis. Hydrogen can be generated, for example, at the bottom of the pits and spread to the adjacent unaffected material, establishing a hydrogen diffusion zone. Future investigation of the hydrogen generation and subsequent HE is required to definitively prove this mechanism.  The thickness of the passive layer formed on B206 surface at different temperatures was calculated using electrochemical measurements which was in a good agreement with previous studies on aluminum-copper alloys. However, methods such as X-ray photoelectron spectroscopy (XPS) or Auger electron spectroscopy (AES) could be used to confirm these calculations.  The compatibility of an epoxy coating with CP in a potential range was investigated in this dissertation. Methods such as Atomic Force Microscopy 171 (AFM) can be used to determine the adhesion, physical properties or the quality of the coating.  In order to study behavior of the coating at different levels of CP, localized electrochemical techniques such as Scanning Electrochemical Microscopy (SECM) and Localized Electrochemical Impedance Spectroscopy (LEIS) can be used to monitor the disbondment of the coating in-situ, at a greater level.  In Chapter 8, the composition of the deposits formed in the holiday were not characterized. This characterization could be performed using XPS or AES.  The results of the effect of the local chemistry inside the pore in this thesis revealed useful information. This can be accomplished by linking the experimental data to a Through Process Model (TPM) to predict the in-service life of the hub component.  In Chapter 9, electrochemical experiments were conducted on a few pencil electrode samples and only a few data points were collected. As a result, it was not possible to determine the conductivity of the solution inside- and potential at the bottom of- the different sized pores. 172 11 References [1] American Institute of Hydrology Committee, ASM Handbook, Volume 2-Properties and Selection: Nonferrous Alloys and Special-Purpose Materials. (Materials Park, OH: ASM International, 1990). [2] C. Vargel, M. Jacques, M.P. Schmidt, Corrosion of Aluminium. (London: Elsevier, 2004). [3] Pierre. R. Roberge, Handbook of Corrosion Engineering Library of Congress Cataloging-in-Publication Data. (New York, NY: Marcel Dekker, 1999). [4] American Institute of Hydrology Committee, ASM Handbook, volume 15-Casting. (Materials Park, OH: ASM International, 2010). [5] Y. Kim, R.G. Buchheit, A characterization of the inhibiting effect of Cu on metastable pitting in dilute Al-Cu solid solution alloys, Electrochim. Acta, 52 (2007) 2437-2446. [6] J. Davis, J. Destefani, H. Frissell, Metals Handbook, Volume 13-Corrosion, 9th ed. (Materials Park, OH: ASM International, 1987). [7] D.O. Northwood, M. Manivannan, J.H. Sokolowski, H.J. Flitt, D. P. Schweinsberg, Corrosion and Prevention, A Potentiodynamic Polarization Study of Heat-Treated B206 Aluminum Casting Alloy, (2009), paper No. 059, 1–11. [8] M. Manivannan, J.H. Sokolowski, D.O. Northwood, Corrosion and Prevention, An Investigation of the Effect of Various Heat Treatments on the Corrosion Resistance of a High- Strength Aluminum-Copper Casting Alloy, (2008), paper No. 059, 1–10. [9] W.R. Osório, J.E. Spinelli, I.L. Ferreira, A. Garcia, The roles of macrosegregation and of dendritic array spacings on the electrochemical behavior of an Al-4.5 wt.% Cu alloy, Electrochim. Acta. 52 (2007) 3265–3273. [10] W. Zhang, G.S. Frankel, Transitions between pitting and intergranular corrosion in AA2024, Electrochim. Acta. 48 (2003) 1193–1210. [11] J.R. Galvele, S.M. de De Micheli, Mechanism of intergranular corrosion of Al-Cu alloys, Corros. Sci. 10 (1970) 795–807. 173 [12] J.F. Li, Z.Q. Zheng, S.C. Li, W.J. Chen, W.D. Ren, X.S. Zhao, Simulation study on function mechanism of some precipitates in localized corrosion of Al alloys, Corros. Sci. 49 (2007) 2436–2449. [13] J.F. Li, Z. Ziqiao, J. Na, T. Chengyu, Localized corrosion mechanism of 2xxx series Al alloy containing S(Al2CuMg) and θ(Al2Cu) precipitates in 4.0% NaCl solution at pH 6.1, Mater. Chem. Phys. 91 (2005) 325–329. [14] P. Leblanc, G.S. Frankel, A Study of Corrosion and Pitting Initiation of AA2024-T3 Using Atomic Force Microscopy, J. Electrochem. Soc. 149 (2002) B239. [15] R.G. Buchheit, Local Dissolution Phenomena Associated with S Phase (Al2CuMg) Particles in Aluminum Alloy 2024-T3, J. Electrochem. Soc. 144 (1997) 2621. [16] M.B. Vukmirovic, N. Dimitrov, K. Sieradzki, Dealloying and Corrosion of Al Alloy 2024-T3, J. Electrochem. Soc. 149 (2002) B428. [17] T. Suter, R.C. Alkire, Microelectrochemical Studies of Pit Initiation at Single Inclusions in Al 2024-T3, J. Electrochem. Soc. 148 (2001) B36. [18] M. Dhondt, I. Aubert, N. Saintier, J.M. Olive, Effects of microstructure and local mechanical fields on intergranular stress corrosion cracking of a friction stir welded aluminum-copper-lithium 2050 nugget, Corros. Sci. 86 (2014) 123–130. [19] A. Seyeux, G.S. Frankel, N. Missert, K.A. Unocic, L.H. Klein, A. Galtayries, P. Marcus, ToF-SIMS Imaging Study of the Early Stages of Corrosion in Al-Cu Thin Films, J. Electrochem. Soc. 158 (2011) C165. [20] A. Boag, A.E. Hughes, A.M. Glenn, T.H. Muster, D. McCulloch, Corrosion of AA2024-T3 Part I: Localised corrosion of isolated IM particles, Corros. Sci. 53 (2011) 17-26. [21] R. Grilli, M. a. Baker, J.E. Castle, B. Dunn, J.F. Watts, Localized corrosion of a 2219 aluminium alloy exposed to a 3.5% NaCl solution, Corros. Sci. 52 (2010) 2855–2866. [22] Z.X. Liang, B. Ye, L. Zhang, Q.G. Wang, W.Y. Yang, Q.D. Wang, A new high-strength and corrosion-resistant Al-Si based casting alloy, Mater. Lett. 97 (2013) 104–107. 174 [23] R. Arrabal, B. Mingo, a. Pardo, M. Mohedano, E. Matykina, I. Rodríguez, Pitting corrosion of rheocast A356 aluminium alloy in 3.5wt.% NaCl solution, Corros. Sci. 73 (2013) 342–355. [24] S. Tahamtan, A. Fadavi Boostani, Quantitative analysis of pitting corrosion behavior of thixoformed A356 alloy in chloride medium using electrochemical techniques, Mater. Des. 30 (2009) 2483–2489. [25] S. Tahamtan, A. Fadavi Boostani, Evaluation of pitting corrosion of thixoformed A356 alloy using a simulation model, Trans. Nonferrous Met. Soc. China (English Ed) 20 (2010) 1702–1706. [26] J.M. Bastidas, A. Forn, M.T. Baile, J.L. Polo, C.L. Torres, Pitting corrosion of A357 aluminium alloy obtained by semisolid processing, Mater. Corros. 52 (2001) 691–696. [27] C. Park, S. Kim, Y. Kwon, Y. Lee, J. Lee, Mechanical and corrosion properties of rheocast and low-pressure cast A356-T6 alloy, Mater. Sci. Eng. A. 391 (2005) 86–94. [28] P.C. King, I.S. Cole, P.A. Corrigan, A.E. Hughes, T.H. Muster, S. Thomas, FIB/SEM study of AA2024 corrosion under a seawater drop, part I, Corros. Sci. 53 (2011) 1086–1096. [29] C. Luo, X. Zhou, G.E. Thompson, A. E. Hughes, Observations of intergranular corrosion in AA2024-T351: The influence of grain stored energy, Corros. Sci. 61 (2012) 35–44. [30] F. Cao, Z. Shi, G.-L. Song, M. Liu, M.S. Dargusch, A. Atrens, Influence of casting porosity on the corrosion behaviour of Mg0.1Si, Corros. Sci. 94 (2015) 255–269. [31] G.L. Song, D.H. St John, T. Abbott, Corrosion behaviour of a pressure die cast magnesium alloy, Int. J. Cast Met. Res. 18 (2005) 174–180. [32] D.A. Jones, Principles and Prevention of Corrosion, (Upper Saddle River: Prentice Hall, 1995). [33] S. Martinez, L.V. Žulj, F. Kapor, Disbonding of underwater-cured epoxy coating caused by cathodic protection current, Corros. Sci. 51 (2009) 2253–2258. 175 [34] F. Varela, M.Y.J. Tan, M. Forsyth, Understanding the effectiveness of cathodic protection under disbonded coatings, Electrochim. Acta. 186 (2015) 377–390. [35] S. Sheth, M. Shahidehpour, Tidal Energy in Electric Power Systems, IEEE (2005). [36] M. Adonizio, R. Smith, Roosevelt Island Tial Energy (RITE) Environmental Assessment Project, (March 2011) [Online]. Available at : https://tethys.pnnl.gov/sites/default/files/publications/RITE_Environmental_Impact_Statement.pdf. Accessed on: 29.Nov.2017. [37] H. Kamguo Kamga, D. Larouche, M. Bournane, A. Rahem, Solidification of aluminum-copper B206 alloys with iron and silicon additions, Metall. Mater. Trans. A. 41 (2010) 2844–2855. [38] G.K. Sigworth, F. DeHary, S. Millhollen, Use of high strength aluminum castings alloys in automotive applications, Light Met. (2001) 313–322. [39] G.K. Sigworth, F. DeHart, Recent Development in the High Strength Aluminum-Copper Casting Alloy A206, AFS Trans. 111 (2003) 341–354. [40] J.F. Major, G.K. Sigworth, Chemistry/property relationships in AA 206 alloys, AFS Trans. 114 (2006) 117–128. [41] Sh. Pournazari, D. Maijer, E. Asselin, Effects of Dissolved Oxygen and Temperature on the Corrosion Properties of B206 and A356 Aluminum Alloys Exposed to Seawater Corrosion Conference and Expo, (2016) Paper No. 7561, 1–9. [42] D. Porter, K. Easterling, Phase transformations in metals and alloys. (London: Chapman and Hall, 1992) [43] C. Luo, S.P. Albu, X. Zhou, Z. Sun, X. Zhang, Z. Tang, et al., Continuous and discontinuous localized corrosion of a 2xxx aluminium-copper-lithium alloy in sodium chloride solution, J. Alloys Compd. 658 (2016) 61–70. [44] T. Hashimoto, X. Zhang, X. Zhou, P. Skeldon, S.J. Haigh, G.E. Thompson, Investigation of dealloying of S phase (Al2CuMg) in AA 2024-T3 aluminium alloy using high resolution 2D and 3D electron imaging, Corros. Sci. 103 (2015) 157–164. 176 [45] J. Li, B. Hurley, R. Buchheit, Effect of temperature on the localized corrosion of AA2024-T3 and the electrochemistry of intermetallic compounds during exposure to a dilute NaCl solution, Corrosion. 72 (2016) 1281–1291. [46] A. Boag, A.E. Hughes, N.C. Wilson, A. Torpy, C.M. MacRae, A.M. Glenn, How complex is the microstructure of AA2024-T3?, Corros. Sci. 51 (2009) 1565–1568. [47] A.E. Hughes, C. MacRae, N. Wilson, A. Torpy, T.H. Muster, A.M. Glenn, Sheet AA2024-T3: a new investigation of microstructure and composition, Surf. Interface Anal. 42 (2010) 334–338. [48] R.G. Buchheit, L.P. Montes, M. a Martinez, J. Michael, P.F. Hlava, The electrochemical characteristics of bulk-synthesized Al2CuMg, J. Electrochem. Soc. 146 (1999) 4424–4428. [49] R G. Buchheit, M.A. Martinez, L.P. Montes, Evidence for Cu Ion Formation by Dissolution and Dealloying the Al2CuMg Intermetallic Compound in Rotating Ring‐ Disk Collection Experiments, J. Electrochem. Soc. 147 (2000) 119–124. [50] A. Hughes, T.H. Muster, A. Boag, aA. M. Glenn, C. Luo, X. Zhou, Co-operative corrosion phenomena, Corros. Sci. 52 (2010) 665–668. [51] J. Li, J. Dang, A Summary of Corrosion Properties of Al-Rich Solid Solution and Secondary Phase Particles in Al Alloys, Metals. 7 (2017) 1-19. [52] A. Boag, R.J. Taylor, T.H. Muster, N. Goodman, D. McCulloch, C. Ryan, Stable pit formation on AA2024-T3 in a NaCl environment, Corros. Sci. 52 (2010) 90–103. [53] Y. Peng, S. Li, Y. Deng, H. Zhou, G. Xu, Z. Yin, Synergetic effects of Sc and Zr microalloying and heat treatment on mechanical properties and exfoliation corrosion behavior of Al-Mg-Mn alloys, Mater. Sci. Eng. A. 666 (2016) 61–71. [54] S.P. Ringer, K. Hono, Microstructural Evolution and Age Hardening in Aluminium Alloys : Atom Probe Field-Ion Microscopy and Transmission Electron Microscopy Studies, Mater. Charact., 44 (2000) 101-131. [55] L. Mullert, J.R. Galvele, Pitting Potential of High Purity Binary Aluminum alloys–I. Al-Cu Alloys Pitting and Intergranular Corrosion, Corros. Sci., 17 (1977) 179–193 177 [56] J.F. Li, N. Birbilis, C.X. Li, Z.Q. Jia, B. Cai, Z.Q. Zheng, Influence of retrogression temperature and time on the mechanical properties and exfoliation corrosion behavior of aluminium alloy AA7150, Mater. Charact. 60 (2009) 1334–1341. [57] A.F. Oliveira, M.C. de Barros, K.R. Cardoso, D.N. Travessa, The effect of RRA on the strength and SCC resistance on AA7050 and AA7150 aluminium alloys, Mater. Sci. Eng. A. 379 (2004) 321–326. [58] B. M. Cina, Reducing the susceptibilty of alloys, particulary aluminium alloys, to stress corrosion cracking, United States Patent, 3 856 584, (1974). [59] R.T. Holt, M.D. Raizenne, W. Wallace, D.L. DuQuesnay, RRA Heat Treatment of Large Al 7075-T6 Components, paper presented at RTO AVT Workshop. New Met. Mater. Struct. Aging Aircraft, (1999) 7.1-7.11. [60] J.K. Park, A.J. Ardell, Effect of retrogression and reaging treatments on the microstructure of Ai-7075-T651, Metall. Mater. Trans. A., 15 (1984) 1531–1543. [61] N.C. Danh, K. Rajan, W. Wallace, A TEM study of microstructural changes during retrogression and reaging in 7075 aluminum, Metall. Mater. Trans. A. 14 (1983) 1843–1850. [62] K. Rajan, W. Wallace, J.C. Beddoes, Microstructural study of a high-strength stress-corrosion resistant 7075 aluminium alloy, J. Mater. Sci. 17 (1982) 2817–2824. [63] R.E. Swanson, I.M. Bernstein, A.W. Thompson, Stress Corrosion Cracking of 7075 Aluminum in the T6-RR Temper, Scr. Metall. 16 (1982) 321–324. [64] H. Singh, Master Thesis, The Corrosion Behavior of Aluminum Alloy B206 in Seawater, The University of British Columbia, 2016. [65] V.S. Sinyavskii, V.D. Kalinin, Marine corrosion and protection of aluminum alloys according to their composition and structure, Prot. Met. 41 (2005) 317–328. [66] K.H. Na, S.I. Pyun, H.P. Kim, Analysis of electrochemical noise obtained from pure aluminium in neutral chloride and alkaline solutions, Corros. Sci. 49 (2007) 220–230. 178 [67] F. Mohammadi, T. Nickchi, M.M. Attar, A. Alfantazi, EIS study of potentiostatically formed passive film on 304 stainless steel, Electrochim. Acta. 56 (2011) 8727–8733. [68] Y. Huang, H. Shih, H. Huang, J. Daugherty, S. Wu, S. Ramanathan, Evaluation of the corrosion resistance of anodized aluminum 6061 using electrochemical impedance spectroscopy (EIS), Corros. Sci. 50 (2008) 3569–3575. [69] C. Boissy, C. Alemany-Dumont, B. Normand, EIS evaluation of steady-state characteristic of 316L stainless steel passive film grown in acidic solution, Electrochem. Commun. 26 (2013) 10–12. [70] A. Nagiub, F. Mansfeld, Evaluation of microbiologically influenced corrosion inhibition (MICI) with EIS and ENA, Electrochim. Acta. 47 (2002) 2319–2333. [71] R. Rosliza, H.B. Senin, W.B.W. Nik, Electrochemical properties and corrosion inhibition of AA6061 in tropical seawater, Colloids Surf A Physicochem. Eng. Asp. 312 (2008) 185–189. [72] M.E. Orazem, N. Pébère, B. Tribollet, Enhanced Graphical Representation of Electrochemical Impedance Data, J. Electrochem. Soc. 153 (2006) B129–B136. [73] C.H. Hsu, F. Mansfeld, Technical Note: Concerning the Conversion of the Constant Phase Element Parameter Y0 into a Capacitance, CORROSION. 57 (2001) 747–748. [74] G.J. Brug, A.L.G. van den Eeden, M. Sluyters-Rehbach, J.H. Sluyters, The analysis of electrode impedances complicated by the presence of a constant phase element, J. Electroanal. Chem. Interfacial Electrochem. 176 (1984) 275–295. [75] B. Hirschorn, M.E. Orazem, B. Tribollet, V. Vivier, I. Frateur, M. Musiani, Determination of effective capacitance and film thickness from constant-phase-element parameters, Electrochim. Acta. 55 (2010) 6218–6227. [76] R. Khatami, A. Fattah-alhosseini, M.K. Keshavarz, Effect of grain refinement on the passive and electrochemical behavior of 2024 Al alloy, J. Alloys Compd. 708 (2017) 316–322. [77] H. Wu, Y. Wang, Q. Zhong, M. Sheng, H. Du, Z. Li, The semi-conductor property and corrosion resistance of passive film on electroplated Ni and Cu–Ni alloys, J. Electroanal. Chem. 663 (2011) 59–66. 179 [78] O. Imantalab, A. Fattah-alhosseini, Electrochemical and Passive Behaviors of Pure Copper Fabricated by Accumulative Roll-Bonding (ARB) Process, J. Mater. Eng. Perform. 24 (2015) 2579–2585. [79] Y. Cheng, J. Luo, A comparison of the pitting susceptibility and semiconducting properties of the passive films on carbon steel in chromate and bicarbonate solutions, Appl. Surf. Sci. 167 (2000) 113–121. [80] Y.F. Cheng, J.L. Luo, Electronic structure and pitting susceptibility of passive film on carbon steel, Electrochim. Acta. 44 (1999) 2947–2957. [81] I.M. Gadala, A. Alfantazi, A study of X100 pipeline steel passivation in mildly alkaline bicarbonate solutions using electrochemical impedance spectroscopy under potentiodynamic conditions and Mott–Schottky, Appl. Surf. Sci. 357 (2015) 356–368. [82] L. Choudhary, W. Wang, A. Alfantazi, Electrochemical Corrosion of Stainless Steel in Thiosulfate Solutions Relevant to Gold Leaching, Metall. Mater. Trans. A. 47 (2016) 314–325. [83] Z. Feng, X. Cheng, C. Dong, L. Xu, X. Li, Passivity of 316L stainless steel in borate buffer solution studied by Mott–Schottky analysis, atomic absorption spectrometry and X-ray photoelectron spectroscopy, Corros. Sci. 52 (2010) 3646–3653. [84] K. Gelderman, L. Lee, S.W. Donne, Flat-Band Potential of a Semiconductor: Using the Mott–Schottky Equation, J. Chem. Educ. 84 (2007) 685–688. [85] R. De Gryse, W.P. Gomes, F. Cardon, J. Vennik, On the Interpretation of Mott‐Schottky Plots Determined at Semiconductor/Electrolyte Systems, J. Electrochem. Soc. 122 (1975) 711–712. [86] E.E. Stansbury, R. Buhcanan, Fundamentals of Electrochemical Corrosion, (Materials Park, OH: ASM International, 2000). [87] W.H. Hartt, 2012 Frank Newman Speller Award: Cathodic Protection of Offshore Structures—History and Current Status, Corrosion. 68 (2012) 1063–1075. [88] R. Gunderson, K. Nisancioglu, Cathodic Protection of Aluminium in Seawater, Corrosion. 46 (1990) 279–285. 180 [89] W.L. Crosby, The use of cathodic protection in conjunction with paint coatings, paper presented at Sixth Annual Conference, NACE, 6 (1950) 383–388. [90] Q. Le Thu, H. Takenouti, S. Touzain, EIS characterization of thick flawed organic coatings aged under cathodic protection in seawater, Electrochim. Acta. 51 (2006) 2491–2502. [91] N. Kamalanand, G. Gopalakrishnan, S.G. Ponnambalam, J. Mathiyarasu, P. Subramaniam, N. Palaniswamy, et al., Role of hydrogen and hydroxyl ion in cathodic disbondment, Anti-Corrosion Methods Mater. 45 (1998) 243–247. [92] E.L. Koehler, Mechanism of Cathodic Disbondment of Protective Oeganic Coatings- Aqueous Displacement at Elevated pH, Corrosion. 40 (1984) 5–8. [93] F. Mahdavi, M. Forsyth, M.Y.J. Tan, Understanding the effects of applied cathodic protection potential and environmental conditions on the rate of cathodic disbondment of coatings by means of local electrochemical measurements on a multi-electrode array, Prog. Org. Coatings. 103 (2017) 83–92. [94] O. Knudsen, U. Steinsmo, Effects of cathodic disbonding and blistering on current demand for cathodic protection of coated steel, Corrosion. (2000) 256–564. [95] NACE INTERNATIONAL, TM0115-2015 Standard Test Method Cathodic Disbondment Test for Coated Steel Structures Under Cathodic Protection, (2015). [96] Y. Liu, J. Wang, L. Liu, Y. Li, F. Wang, Study of the failure mechanism of an epoxy coating system under high hydrostatic pressure, Corros. Sci. 74 (2013) 59–70. [97] D. Nguyen Dang, B. Peraudeau, S. Cohendoz, S. Mallarino, X. Feaugas, S. Touzain, Effect of mechanical stresses on epoxy coating ageing approached by Electrochemical Impedance Spectroscopy measurements, Electrochim. Acta. 124 (2014) 80–89. [98] I.C.P. Margarit-Mattos, F.A.R. Agura, C.G. Silva, W.A. Souza, J.P. Quintela, V. Solymossy, Electrochemical impedance aiding the selection of organic coatings for very aggressive conditions, Prog. Org. Coatings. 77 (2014) 2012–2023. [99] P. Berce, S. Skale, M. Slemnik, Electrochemical impedance spectroscopy study of waterborne coatings film formation, Prog. Org. Coatings. 82 (2015) 1–6. 181 [100] M. Kendig, J.R. Scully, Basic aspects of electrochemical impedance application for the life prediction of organic coatings on metals, Corrosion. 46 (1990) 22–29. [101] F. Mansfeld, Use of electrochemical impedance spectroscopy for the study of corrosion protection by polymer coatings, J. Appl. Electrochem. 25 (1995) 187–202. [102] O. Schneider, R.G. Kelly, Localized coating failure of epoxy-coated aluminium alloy 2024-T3 in 0.5 M NaCl solutions: Correlation between coating degradation, blister formation and local chemistry within blisters, Corros. Sci. 49 (2007) 594–619. [103] F. Mahdavi, M.Y.J. Tan, M. Forsyth, Electrochemical impedance spectroscopy as a tool to measure cathodic disbondment on coated steel surfaces: Capabilities and limitations, Prog. Org. Coatings. 88 (2015) 23–31. [104] S. Papavinasam, B. Arsenault, M. Attard, R.W. Revie, Metallic under-layer coating as third line of protection of underground oil and gas pipelines from external corrosion, Corrosion. 68 (2012) 1146–1153. [105] I. Margarit, O. Mattos, About coatings and cathodic protection: Electrochemical features of coatings used on pipelines, J. Coatings Technol. 73 (2001) 61–65. [106] L.L. Shreir, R.A. Jarman, G.T. Burstein, Corrosion, Metal/ Environment Reactions, Vol. 1, 3rd edition, (Oxford: Butterworth, Heineman, 1994). [107] J.C. Riviere, S.Myhra, Handbook of surface and interface analysis (New York: CRC Press, 2009). [108] G. Song, Infuence of microstructure on the corrosion of diecast AZ80D, Corros. Sci. 41 (1999) 249–273. [109] M. Esmaily, J.E. Svensson, M. Halvarsson, L.G. Johansson, Corrosion Behavior of Alloy AM50 in Semisolid Cast and High-Pressure Die Cast States in Cyclic Conditions, Corrosion, 71 (2015) 737–748. [110] J.R. Galvele, Transport Processes and the Mechanism of Pitting of Metals, J. Electrochem. Soc. 123 (1976) 464. [111] J. Srinivasan, M.J. McGrath, R.G. Kelly, A High-Throughput Artificial Pit Technique to Measure Kinetic Parameters for Pitting Stability, J. Electrochem. Soc. 162 (2015) C725–C731. 182 [112] M.T. Woldemedhin, J. Srinivasan, R.G. Kelly, Effects of environmental factors on key kinetic parameters relevant to pitting corrosion, J. Solid State Electrochem. 19 (2015) 3449–3461. [113] J. Srinivasan, C. Liu, R.G. Kelly, Geometric Evolution of Flux from a Corroding One-Dimensional Pit and Its Implications on the Evaluation of Kinetic Parameters for Pit Stability, J. Electrochem. Soc. 163 (2016) C694–C703. [114] M.H. Moayed, R.C. Newman, The Relationship Between Pit Chemistry and Pit Geometry Near the Critical Pitting Temperature, J. Electrochem. Soc. 153 (2006) B330. [115] M.H. Moayed, R.C. Newman, Deterioration in critical pitting temperature of 904L stainless steel by addition of sulfate ions, Corros. Sci. 48 (2006) 3513–3530. [116] A.B. Cook, D.L. Engelberg, N.P. Stevens, N.J. Laycock, S. White, M. Ghahari, Pit Propagation in Pure Aluminum Investigated via the 1D Artificial Pit Technique: Growth Regimes, Surface Morphology and Implications for Stability Criteria, ECS. Trans. 41 (2012) 121–132. [117] T.R. Beck, Salt film formation during corrosion of aluminum, Electrochim. Acta. 29 (1984) 485–491. [118] Q. Meng, T. Ramgopal, G.S. Frankel, The Influence of Inhibitor Ions on Dissolution Kinetics of Al and Mg Using the Artificial Crevice Technique, Electrochem. Solid-State Lett. 5 (2002) B1–B4. [119] J. Sweitzer, J. Scully, R. Bley, J. Hsu, Nanocrystalline Al87Ni8.7Y4.3 and Al90Fe5Gd5 alloys that retain the localized corrosion resistance of the amorphous state, Electrochem. Solid-State Lett. 2 (1999) 267–270. [120] E. Akiyama, The Influence of Dichromate Ions on Aluminum Dissolution Kinetics in Artificial Crevice Electrode Cells, J. Electrochem. Soc. 146 (1999) 4095. [121] Standard Practice for the Preparation of Substitute Ocean Water - D1141-98, ASTM (Reapproved 2013), 1–3. [122] R.G. Kelly, J.R. Scully, D.W. Shoesmith, R.G. Buchheit, Electrochemical Techniques in Corrosion Science and Engineering (New York: MARCEL DEKKER 2003). 183 [123] M.E. Orazem, B. Tribollet, Electrochemical Impedance Spectroscopy, Wiley, 2008. [124] T. Huber, Y. Wang, Effect of propeller coating on cathodic protection current demand: Sea trial and modeling studies, Corrosion. 68 (2012) 441–448. [125] X. Zhang, T. Hashimoto, J. Lindsay, X. Zhou, Investigation of the de-alloying behaviour of θ-phase (Al2Cu) in AA2024-T351 aluminium alloy, Corros. Sci. 108 (2016) 85–93. [126] A.M. Glenn, T.H. Muster, C. Luo, X. Zhou, G.E. Thompson, A. Boag, Corrosion of AA2024-T3 Part III: Propagation, Corros. Sci. 53 (2011) 40–50. [127] A. E. Hughes, A. Boag, A. M. Glenn, D. McCulloch, T.H. Muster, C. Ryan, Corrosion of AA2024-T3 Part II: Co-operative corrosion, Corros. Sci. 53 (2011) 27–39. [128] P.C. King, I.S. Cole, P.A. Corrigan, A.E. Hughes, T.H. Muster, S. Thomas, FIB/SEM study of AA2024 corrosion under a seawater drop, part II, Corros. Sci. 55 (2012) 116–125. [129] X. Zhou, C. Luo, T. Hashimoto, A.E. Hughes, G.E. Thompson, Study of localized corrosion in AA2024 aluminium alloy using electron tomography, Corros. Sci. 58 (2012) 299–306. [130] Sh. Pournazari, D. Maijer, E. Asselin, FIB / SEM study of pitting and intergranular corrosion in an Al-Cu alloy, Corrosion. 73 (2017) 927–941. [131] American Institute of Hydrology Committee, ASM Handbook, Volume 3-Alloy Phase Diagrams. (Materials Park, OH: ASM International, 1992). [132] N. Haghdadi, A.B. Phillion, D.M. Maijer, Microstructure Characterization and Thermal Analysis of Aluminum Alloy B206 During Solidification, Metall. Mater. Trans. A. 46 (2015) 2073–2081. [133] H. Kamali, M. Emamy, A. Razaghian, The influence of Ti on the microstructure and tensile properties of cast Al-4.5Cu-0.3Mg alloy, Mater. Sci. Eng. A. 590 (2014) 161–167. [134] R.G. Buchheit, L.P. Montes, M. a Martinez, J. Michael, P.F. Hlava, The electrochemical characteristics of bulk-synthesized Al2CuMg, J. Electrochem. Soc. 146 (1999) 4424–4428. 184 [135] J. Li, N. Birbilis, R.G. Buchheit, Electrochemical assessment of interfacial characteristics of intermetallic phases present in aluminium alloy 2024-T3, Corros. Sci. 101 (2015) 155–164. [136] J. D. Gorman, A. E. Hughes, D. Jamieson, P. K. J. Paterson, Oxide formation on aluminum alloys in boiling deionised water and NaCl, CeCl3 and CrCl3 solutions, Corros. Sci. 45 (2003) 1103–1124. [137] N. Dimitrov, Copper Redistribution during Corrosion of Aluminum Alloys, J. Electrochem. Soc. 146 (1999) 98. [138] Sh. Pournazari, K. M. Deen, D. M. Maijer, E. Asselin, “ Effect of retrogression and re-aging (RRA) heat treatment on the corrosion behavior of B206 aluminum-copper cast alloy”, Materials and Corrosion [accepted, Dec.2017]. [139] R.K. Wyss, R.E. Sanders, Microstructure-property relationship in a 2xxx aluminum alloy with mg addition, Metall. Trans. A. 19 (1988) 2523–2530. [140] W. Liu, F. Cao, A. Chen, L. Chang, J. Zhang, C. Cao, Effect of Chloride Ion Concentration on Electrochemical Behavior and Corrosion Product of AM60 Magnesium Alloy in Aqueous Solutions, Corrosion. 68 (2012) 1–14. [141] M.C. Zhao, M. Liu, G.L. Song, A. Atrens, Influence of pH and chloride ion concentration on the corrosion of Mg alloy ZE41, Corros. Sci. 50 (2008) 3168–3178. [142] E. McCafferty, Validation of corrosion rates measured by the Tafel extrapolation method, Corros. Sci. 47 (2005) 3202–3215. [143] K.D. Ralston, N. Birbilis, M. Weyland, C.R. Hutchinson, The effect of precipitate size on the yield strength-pitting corrosion correlation in Al-Cu-Mg alloys, Acta Mater. 58 (2010) 5941–5948. [144] J.B. Bessone, D.R. Salinas, C.E. Mayer, M. Eberet, W.J. Lorenz, An EIS Study of Aluminum Barrier-Type Oxide-Films Formed in Different Media, Electrochim. Acta. 37 (1992) 2283–2290. [145] J.A. Lyndon, R.K. Gupta, M.A. Gibson, N. Birbilis, Electrochemical behaviour of the β-phase intermetallic (Mg2Al3) as a function of pH as relevant to corrosion of aluminium-magnesium alloys, Corros. Sci. 70 (2013) 290–293. 185 [146] N. Birbilis, R.G. Buchheit, Electrochemical Characteristics of Intermetallic Phases in Aluminum Alloys, J. Electrochem. Soc. 152 (2005) B140–B151. [147] J.J. Pang, F.C. Liu, J. Liu, M.J. Tan, D.J. Blackwood, Friction stir processing of aluminium alloy AA7075: Microstructure, surface chemistry and corrosion resistance, Corros. Sci. 106 (2015) 217–228. [148] J.A. Moreto, C.E.B. Marino, W.W. Bose Filho, L.A. Rocha, J.C.S. Fernandes, SVET, SKP and EIS study of the corrosion behaviour of high strength Al and Al-Li alloys used in aircraft fabrication, Corros. Sci. 84 (2014) 30–41. [149] Sh. Pournazari, M.H. Moayed, M. Rahimizadeh, In situ inhibitor synthesis from admixture of benzaldehyde and benzene-1,2-diamine along with FeCl3 catalyst as a new corrosion inhibitor for mild steel in 0.5M sulphuric acid, Corros. Sci. 71 (2013) 20–31. [150] T. Ramgopal, P.I. Gouma, G.S. Frankel, Role of grain boundary precipitates and solute depleted zone on the intergranular corrosion of aluminium alloy 7150, Corrosion. 58 (2002) 687–697. [151] G.N. Haidemenopoulos, N. Hassiotis, G. Papapolymerou, V. Bontozoglou, Hydrogen Absorption into Aluminum Alloy 2024-T3 during Exfoliation and Alternate Immersion Testing, Corrosion. 54 (1998) 73–78. [152] J.F. Li, Z.Q. Jia, C.X. Li, N. Birbilis, C. Cai, Exfoliation corrosion of 7150 Al alloy with various tempers and its electrochemical impedance spectroscopy in EXCO solution, Mater. Corros. 60 (2009) 407–414. [153] F.H. Cao, Z. Zhang, J.F. Li, Y.L. Cheng, J.Q. Zhang, C.N. Cao, Exfoliation corrosion of aluminum alloy AA7075 examined by electrochemical impedance spectroscopy, Mater. Corros. 55 (2004) 18–23. [154] T. Marlaud, B. Malki, A. Deschamps, B. Baroux, Electrochemical aspects of exfoliation corrosion of aluminium alloys: The effects of heat treatment, Corros. Sci. 53 (2011) 1394–1400. [155] S. Chen, K. Chen, G. Peng, L. Jia, P. Dong, Effect of heat treatment on strength, exfoliation corrosion and electrochemical behavior of 7085 aluminum alloy, Mater. Des. 35 (2012) 93–98. 186 [156] S.D. Liu, B. Chen, C.B. Li, Y. Dai, Y.L. Deng, X.M. Zhang, Mechanism of low exfoliation corrosion resistance due to slow quenching in high strength aluminium alloy, Corros. Sci. 91 (2015) 203–212. [157] H. Kamoutsi, G.N. Haidemenopoulos, V. Bontozoglou, P. V. Petroyiannis, S.G. Pantelakis, Effect of prior deformation and heat treatment on the corrosion-induced hydrogen trapping in aluminium alloy 2024, Corros. Sci. 80 (2014) 139–142. [158] P. V. Petroyiannis, E. Kamoutsi, A.T. Kermanidis, S.G. Pantelakis, V. Bontozoglou, G.N. Haidemenopoulos, Evidence on the corrosion-induced hydrogen embrittlement of the 2024 aluminium alloy, Fatigue Fract. Eng. Mater. Struct. 28 (2005) 565–574. [159] H. Kamoutsi, G.N. Haidemenopoulos, V. Bontozoglou, S. Pantelakis, Corrosion-induced hydrogen embrittlement in aluminum alloy 2024, Corros. Sci. 48 (2006) 1209–1224. [160] A. Kuznetsova, J.T. Yates, G. Zhou, J.C. Yang, X. Chen, Making a Superior Oxide Corrosion Passivation Layer on Aluminum Using Ozone, Langmuir. 17 (2001) 2146–2152. [161] H.J. Lee, I.-J. Park, S.R. Choi, J.G. Kim, Effect of Chloride on Anodic Dissolution of Aluminum in 4 M NaOH Solution for Aluminum-Air Battery, J. Electrochem. Soc. 164 (2017) A549–A554. [162] K.C. Emregül, A.A. Aksüt, The behavior of aluminum in alkaline media, Corros. Sci. 42 (2000) 2051–2067. [163] S.M. Moon, S.I. Pyun, The formation and dissolution of anodic oxide films on pure aluminium in alkaline solution, Electrochim. Acta. 44 (1999) 2445–2454. [164] K. Ishii, R. Ozaki, K. Kaneko, H. Fukushima, M. Masuda, Continuous monitoring of aluminum corrosion process in deaerated water, Corros. Sci. 49 (2007) 2581–2601. [165] P.M. Natishan, W.E. O’Grady, Chloride Ion Interactions with Oxide-Covered Aluminum Leading to Pitting Corrosion: A Review, J. Electrochem. Soc. 161 (2014) C421–C432. 187 [166] N. Lampeas, P.G. Koutsoukos, The importance of the solution pH in electrochemical studies of aluminum in aqueous media containing chloride, Corros. Sci. 36 (1994) 1011–1025. [167] S.O. Gashti, A. Fattah-alhosseini, Y. Mazaheri, Electrochemical Behavior of Passive Films Formed on the Surface of Coarse-, Fine- and Ultra-fine-Grained AA1050 Based on a Modified PDM, Acta Metall. Sin. (English Lett.). 29 (2016) 629–637. [168] T.H. Nguyen, R.T. Foley, The Chemical Nature of Aluminum Corrosion II- The Initial Dissolution Step, J. Electrochem. Soc. 129 (1982) 27–32. [169] E. McCafferty, Sequence of steps in the pitting of aluminum by chloride ions, Corros. Sci. 45 (2003) 1421–1438. [170] A. Kolics, J.C. Polkinghorne, A. Wieckowski, Adsorption of sulfate and chloride ions on aluminum, Electrochim. Acta. 43 (1998) 2605–2618. [171] S. Krakowiak, K. Darowicki, K. Jurak, Cyclic analysis of thermal impedance of a passive layer of aluminium in a neutral borate buffer solution, Anti-Corrosion Methods Mater. 59 (2012) 285–290. [172] J.F. Dewald, The charge distribution at the zinc oxide-electrolyte interface, J. Phys. Chem. Solids. 14 (1960) 155–161. [173] J.O. Bockris, Y. Kang, The protectivity of aluminum and its alloys with transition metals, J. Solid State Electrochem. 1 (1997) 17–35. [174] C.L. Chang, S.K.R.S. Sankaranarayanan, M.H. Engelhard, V. Shutthanandan, S. Ramanathan, On the Relationship between Nonstoichiometry and Passivity Breakdown in Ultrathin Oxides: Combined Depth-Dependent Spectroscopy, Mott−Schottky Analysis, and Molecular Dynamics Simulation Studies, J. Phys. Chem. C. 113 (2009) 3502–3511. [175] J.C.S. Fernandes, R. Picciochi, M. Da Cunha Belo, T. Moura e Silva, M.G.S. Ferreira, I.T.E. Fonseca, Capacitance and photoelectrochemical studies for the assessment of anodic oxide films on aluminium, Electrochim. Acta. 49 (2004) 4701–4707. 188 [176] Y. Liu, G.Z. Meng, Y.F. Cheng, Electronic structure and pitting behavior of 3003 aluminum alloy passivated under various conditions, Electrochim. Acta. 54 (2009) 4155–4163. [177] L. Jinlong, L. Hongyun, Effect of surface burnishing on texture and corrosion behavior of 2024 aluminum alloy, Surf. Coatings Technol. 235 (2013) 513–520. [178] D.D. Macdonald, On the Existence of Our Metals-Based Civilization I. Phase-Space Analysis, J. Electrochem. Soc. 153 (2006) B213–B224. [179] A. Fattah-alhosseini, Passivity of AISI 321 stainless steel in 0.5M H2SO4 solution studied by Mott–Schottky analysis in conjunction with the point defect model, Arab. J. Chem. 9 (2016) S1342–S1348. [180] J. Zhang, W. Zhang, C. Yan, K. Du, F. Wang, Corrosion behaviors of Zn/Al–Mn alloy composite coatings deposited on magnesium alloy AZ31B (Mg–Al–Zn), Electrochim. Acta. 55 (2009) 560–571. [181] B. Hirschorn, M.E. Orazem, B. Tribollet, V. Vivier, I. Frateur, M. Musiani, Constant-Phase-Element Behavior Caused by Resistivity Distributions in Films I. Theory, J. Electrochem. Soc. 157 (2010) C452–C457. [182] B. Hirschorn, M.E. Orazem, B. Tribollet, V. Vivier, I. Frateur, M. Musiani, Constant-Phase-Element Behavior Caused by Resistivity Distributions in Films II. Applications, J. Electrochem. Soc. 157 (2010) C458–C463. [183] M. Mohammadi, L. Choudhary, I.M. Gadala, A. Alfantazi, Electrochemical and Passive Layer Characterizations of 304L, 316L, and Duplex 2205 Stainless Steels in Thiosulfate Gold Leaching Solutions, J. Electrochem. Soc. 163 (2016) C883–C894. [184] NACE INTERNATIONAL, TM0169-2000, Standard Test Method Laboratory Corrosion Testing of Metals, (2000). [185] M. Behzadnasab, S.M. Mirabedini, M. Esfandeh, Corrosion protection of steel by epoxy nanocomposite coatings containing various combinations of clay and nanoparticulate zirconia, Corros. Sci. 75 (2013) 134–141. [186] M. Golabadi, M. Aliofkhazraei, M. Toorani, A.S. Rouhaghdam, Evaluation of La containing PEO pretreatment on protective performance of epoxy coating on magnesium, Prog. Org. Coatings. 105 (2017) 258–266. 189 [187] R.M. Souto, M.L. Llorente, L. Fernández-Mérida, Accelerated tests for the evaluation of the corrosion performance of coil-coated steel sheet: EIS under cathodic polarisation, Prog. Org. Coatings. 53 (2005) 71–76. [188] R. Bonzom, R. Oltra, Intergranular corrosion propagation rate of 2024 alloy investigated via the “one-dimensional artificial pit” technique, Corros. Sci. 111 (2016) 850–855. 190 12 Appendices Appendix A: supplementary figures for chapter 4 Figure A-1: Monte Carlo electron trajectory simulations of the interaction volume in B206 191 Appendix B: supplementary figures for chapter 5 Figure B-1: EDX spectrum for spot 1 in Figure 5-2 and Table 5-1 Figure B-2: EDX spectrum for spot 2 in Figure 5-2 and Table 5-1 192 Figure B-3: EDX spectrum for spot 3 in Figure 5-2 and Table 5-1 Figure B-4: EDX spectrum for spot 4 in Figure 5-2 and Table 5-1 193 Figure B-5: EDX spectrum for spot C-spot1/0h in Table 5-2 Figure B-6: EDX spectrum for spot C-spot1/24h in Table 5-2 194 Figure B-7: EDX spectrum for spot A-spot1/0h in Table 5-2 Figure B-8: EDX spectrum for spot A-spot1/24h in Table 5-2 195 Appendix C: supplementary figure for chapter 8 Figure C-1: Bode plots of bare and coated B206 after immersion for 6 days in artificial seawater at 10 °C 196 Appendix D: supplementary table for chapter 9 Table D-1: Volume of the hydrogen gas evolved during the potentiostatic polarization experiments of 1mm B206 pencil electrode at different potentials in artificial seawater at 22 °C Applied Potential (V vs. Ag/AgCl) Volume of the hydrogen gas evolved (cm3) 0 0.52 0.1 0.1 0.22 0.2 0.2 0.12 0.2 0.3 0.2 2.1 x 104 1.02 x 10-3 0.5 0.4 1 0.8 "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2018-05"@en ; edm:isShownAt "10.14288/1.0365752"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Materials Engineering"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@* ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@* ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Corrosion behavior of B206 aluminum-copper casting alloy in seawater environment : electrochemical and microstructural studies"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/65487"@en .