Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

External oil pipeline corrosion of near-fusion heat-affected zones simulated from high-strength steels Eliyan, Faysal Fayez 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_february_eliyan_faysal.pdf [ 8.9MB ]
Metadata
JSON: 24-1.0135596.json
JSON-LD: 24-1.0135596-ld.json
RDF/XML (Pretty): 24-1.0135596-rdf.xml
RDF/JSON: 24-1.0135596-rdf.json
Turtle: 24-1.0135596-turtle.txt
N-Triples: 24-1.0135596-rdf-ntriples.txt
Original Record: 24-1.0135596-source.json
Full Text
24-1.0135596-fulltext.txt
Citation
24-1.0135596.ris

Full Text

  External Oil Pipeline Corrosion of Near-Fusion Heat-Affected Zones Simulated from High-Strength Steels   by  FAYSAL FAYEZ ELIYAN   B.Sc., Qatar University, 2009 M.A.Sc., The University of British Columbia, 2011  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)     November 2014  © Faysal Fayez Eliyan, 2014   ii  Abstract  The thesis evaluates the external pipeline corrosion at the near-fusion Heat-Affected Zones (HAZs). The HAZs were simulated by Gleeble© thermal cycles by which specific thermal conditions during pipeline welding are applied. They were applied on the high-strength steels API-X100 and API-X80. The thermal cycles were set to simulate the HAZs that are in direct adjacency to the fusion zone, where the temperature is high enough to transform the starting microstructure of the base metal into the single phase austenite. In the thermal simulations, the peak temperatures, cooling rates, and peak-temperature time periods are the parameters by which the different HAZ microstructures are simulated. For API-X80 and API-X100, the HAZ microstructural features are quantitatively and qualitatively linked, respectively, to the corrosion behavior.       The slowly-cooled ferritic HAZs were found to corrode faster than the faster-cooled bainitic and martensitic HAZs. They, however, showed a tendency to develop thicker corrosion products, which are in the long run protective against underlying dissolution. It was found that bicarbonate in low concentrations exacerbate corrosion and prevents the passivation – regardless of the environmental conditions. In higher concentrations, bicarbonate promotes the passivation indefinitely, and that was also the case with carbonate, but regardless of the concentration.     The laboratory conditions contain the corrosive constituents that drive the external pipeline corrosion upon exposure to the ground water or hydrated soils. The test solutions contain different levels of bicarbonate, carbonate, and chloride in mediums naturally-aerated, de-oxygenated, and CO2-saturated. The corrosion behavior is investigated by electrochemical methods that measure the corrosion rates, evaluate the kinetics of dissolution and cathodic reduction, and model and evaluate the growth, protectiveness, and reactivity of passive films. iii  Preface  The heat-treated API-X80 samples, investigated in chapter 5, were provided by Dr. Farzad Mohammadi. The decided thermal cycles and the microstructural analysis have received significant contributions from him. The heat-treated API-X100 samples, investigated in chapters 6, 7, 8 and 9, were given by Ms. Jennifer Reichert. She contributed in determining the thermal cycles designed to heat treat the samples.  The following journal and conference papers and presentations have been published from the research work of this thesis. Professor Akram Alfantazi has been providing the necessary guidance and support to accomplish this from all aspects.   Journal Papers: 1. Faysal Fayez Eliyan, Akram Alfantazi, Mechanisms of Corrosion and Electrochemical Significance of Metallurgy and Environment with Corrosion of Iron and Steel in Bicarbonate and Carbonate Solutions – A Review, Corrosion, in press.  2. Faysal Fayez Eliyan, Akram Alfantazi, Corrosion of the Heat-Affected Zones (HAZs) of API-X100 pipeline steel in dilute bicarbonate solutions at 90°C – An electrochemical evaluation, Corrosion Science, Volume 74, 2013, 297-307.  3. Farzad Mohammadi, Faysal Fayez Eliyan, Akram Alfantazi, Corrosion of simulated weld HAZ of API X-80 pipeline steel, Corrosion Science, Volume 63, 2012, 323-333.  Conference Papers: 1. Faysal Fayez Eliyan, Akram Alfantazi, “Corrosion Mechanisms of Iron and Pipeline Steels in Bicarbonate and Carbonate Solutions – A Review”, Proceedings of EuroCorr 2013, The European Corrosion Congress, September 1-5, Estoril, Portugal, Paper # O-1470, 5 pages, 2013.   Conference Presentations: 1. Faysal Fayez Eliyan, Akram Alfantazi, “Electrochemical significance of the passive films with the corrosion behavior of API-X100 steel HAZs”. NACE Northern Area Western Conference, Victoria, British Columbia, Canada, February 11-14, 2013.  iv  2. Faysal Fayez Eliyan, Farzad Mohammadi, Akram Alfantazi, “Microstructural analysis and electrochemical corrosion evaluation of API-X80 pipeline steel heat-affected zones in dilute bicarbonate solutions”. NACE Northern Area Western Conference, Victoria, British Columbia, Canada, 2013.                      v  Table of Contents Abstract ........................................................................................................................................... ii Preface............................................................................................................................................ iii Table of Contents………………………………………………………………………………….v List of Tables ............................................................................................................................... viii List of Figures ................................................................................................................................ ix List of Symbols ............................................................................................................................. xv Acknowledgment ........................................................................................................................ xvii Dedication .................................................................................................................................. xviii CHAPTER 1: INTRODUCTION ................................................................................................... 1 CHAPTER 2: LITERATURE REVIEW ........................................................................................ 4 2.1    Mechanisms of anodic dissolution and passivation ........................................................... 4 2.2    Significance of microstructure ......................................................................................... 12 2.3    Significance of alloying composition .............................................................................. 16 2.4    Significance of bicarbonate .............................................................................................. 18 2.4.1    CO2-saturated solutions ............................................................................................ 19 2.4.2    Naturally-aerated solutions ....................................................................................... 21 2.4.3    De-oxygenated solutions ........................................................................................... 22 2.5    Significance of carbonate ................................................................................................. 26 2.6    Significance of chloride ................................................................................................... 26 2.7    Significance of temperature ............................................................................................. 30 2.8    Significance of natural aeration ....................................................................................... 33 2.9    Significance of pH ........................................................................................................... 36 2.10  Significance of corrosion products .................................................................................. 37 2.11   Metallurgy of heat-affected zones (HAZs) ..................................................................... 43 2.12  Summary of findings in literature .................................................................................... 47 CHAPTER 3: SCOPE AND OBJECTIVES................................................................................. 50 CHAPTER 4: APPROACH AND METHODOLOGY ................................................................ 52 4.1    Corrosion test setup.......................................................................................................... 52 vi  4.2    Test materials, corrosion test samples preparation, and microstructural analysis ........... 53 4.3    Gleeble thermal HAZ simulations ................................................................................... 56 4.4    Test solutions ................................................................................................................... 62 4.5    Electrochemical tests ....................................................................................................... 63 CHAPTER 5: SIGNIFICANCE OF FERRITE, MARTENSITE-RETAINED AUSTENITE, AND BAINITE WITH CORROSION KINETICS AND PASSIVATION OF API-X80 STEEL HAZs ............................................................................................................................................. 67 5.1    Microstructural analysis ................................................................................................... 67 5.2    Potentiodynamic polarization testing ............................................................................... 71 5.3    Electrochemical impedance spectroscopy testing ............................................................ 75 5.4.   Summary .......................................................................................................................... 77 CHAPTER 6: ELECTROCHEMICAL CHARACTERISTICS, STABILITY, AND MODELING OF PASSIVE FILMS OF API-X100 STEEL HAZs IN HIGH-TEMPERATURE, DILUTE BICARBONATE SOLUTIONS ................................................................................................... 78 6.1    Potentiodynamic polarization testing ............................................................................... 78 6.2    Potentiostatic polarization testing .................................................................................... 83 6.3    Electrochemical impedance spectroscopy testing ............................................................ 85 6.4    Summary .......................................................................................................................... 94 CHAPTER 7: VOLTAMMETRIC ANALYSIS OF THE SIGNIFICANCE OF BICARBONATE CONCENTRATION WITH CORROSION REACTIONS AND PASSIVATION OF API-X100 STEEL HAZs ................................................................................................................................ 96 7.1 The corrosion reactions and voltammetric analysis ........................................................ 96 7.2 Significance of voltammetry scan rates ........................................................................... 99 7.3 Significance of bicarbonate concentration .................................................................... 101 7.4 Significance of extended cathodic reduction range ....................................................... 106 7.5 Significance of fast-cooling the HAZs .......................................................................... 107 7.6 Summary ....................................................................................................................... 111 CHAPTER 8: INTERRELATION BETWEEN BICARBONATE AND CO2 IN AFFECTING CORROSION REACTIONS AND PASSIVATION OF API-X100 STEEL HAZs ................. 114 8.1 Open-circuit potentials monitoring ............................................................................... 114 8.2  Long-range voltammetry testing .................................................................................. 116 8.3 Short-range voltammetry testing ................................................................................... 120 8.4 Significance of fast cooling the HAZs .......................................................................... 120 8.5 Summary ....................................................................................................................... 123 vii  CHAPTER 9: VOLTAMMETRIC ANALYSIS ON THE FORMATION OF Fe(OH)2 AND FeCO3, AND ON THE REACTIVITY OF PASSIVATION OF API-X100 STEEL HAZs IN CARBONATE SOLUTIONS ..................................................................................................... 125 9.1    Dissolution, passivation, and cathodic reduction in 0.05 M carbonate solution ............ 125 9.2    Significance of carbonate concentration ........................................................................ 132 9.3    Significance of the voltammetric potential range .......................................................... 134 9.4    Significance of fast cooling the HAZs ........................................................................... 136 9.5    Modeling of corrosion reactions in a dilute carbonate solution ..................................... 138 9.5     Summary ....................................................................................................................... 150 CHAPTER 10: CONCLUSION ................................................................................................. 152 CHAPTER 11: RECOMMENDATIONS FOR FUTURE WORK ............................................ 155 REFERENCES ........................................................................................................................... 156               viii   List of Tables  Table 4-1 Chemical composition of API-X100 steel (wt%) ......................................................... 53  Table 4-2 Chemical composition of API-X80 steel (wt%) ........................................................... 53  Table 4-3 Compositions of the etchants........................................................................................ 55  Table 4-4 Details of the test steels, simulated near-fusion HAZs, tests, and test solutions .......... 66  Table 5-1 Corrosion parameters and passive potential extracted from the potentiodynamic polarization profiles ...................................................................................................................... 73  Table 6-1 Equivalent circuit fitting for the EIS data of the base steel API-X100 ........................ 89  Table 6-2 Equivalent circuit fitting for the EIS data of 10oC/s HAZ ........................................... 90  Table 6-3 Equivalent circuit fitting for the EIS data of 30oC/s HAZ ........................................... 92  Table 6-4 Equivalent circuit fitting for the EIS data of 60oC/s HAZ ........................................... 94               ix  List of Figures  Figure 2- 1 Bare surface polarization curves for iron, at pH 8.8, in 0.32 M H3BO3 + 0.19 M K2B4O7 solution (in circles), and in 0.75 M KHCO3 + 0.05 M K2CO3 solution (in squares). White is for anodic currents and black is for cathodic currents [Burstein ,1980] .......................... 5  Figure 2- 2 a) The effect of bicarbonate concentration on the anodic peaks at -0.65 V vs. SCE, for a mild 1024 steel electrode rotating at 1000 rpm during 5 mV/s cyclic voltammetry at 25oC, and b) the reciprocal of current density against (ω)-0.5 in 0.1 M [Simard, 1997]. ........................... 6  Figure 2- 3 The effect of ω on the second anodic peaks at -0.4 V vs. NHE (-0.644 V vs. SCE) for a rotating iron electrode in 0.75 M KHCO3 + 0.05 M K2CO3 solutions at 25oC [Castro, 1986]. 8  Figure 2- 4 Voltammogram scans that show third anodic peaks, at -0.2 V vs. NHE (-0.444 V vs. SCE), which represent the oxidation of FeCO3 to iron hydroxides in 0.75 M KHCO3 + 0.05 M K2CO3 solutions [Valentini, 1985]. .............................................................................................. 10  Figure 2- 5 Microstructures of a spirally welded X70 at a) base steel, b) HAZ, c) HAZ/weld metal and, d) weld metal [Zhang, 2009]. ...................................................................................... 13  Figure 2- 6 Potentiodynamic polarization profiles a) scanned at 1 mV/s in 1 M NaHCO3-0.5 M Na2CO3 solutions at 23oC for different microstructures of spirally-welded X70 steel [Zhang, 2009] and b) scanned at 20 mV/min (0.333 mV/s) in 1 M NaHCO3 + 0.5 M Na2CO3 solutions at 80oC for carbon steel samples welded with tungsten inert gas welding [Stikma, 1985] .............. 14  Figure 2- 7 Significance of heat treatment with the polarization behavior of X-70 steel in 0.005 M NaHCO3 solutions [Torres-Islas, 2008]. .................................................................................. 15  Figure 2- 8 a) Passivation potentials, b) Critical currents, and c) selected 0.2 mV/s potentiodynamic polarization profiles for carbon steel (CS) and the heat-treated High Speed Steel (HSS) and chromium steel (CrS) in de-oxygenated 0.5 M NaHCO3 + 0.01 M KCl solutions at 25oC [Alves, 2002(A)]. ................................................................................................................. 17  Figure 2- 9 Speciation equilibrium of the carbonic species at ambient temperatures for CO2-saturated aqueous systems [Palacios, 1991]. ................................................................................ 19  Figure 2- 10 a) Potentiodynamic polarization curves, swept at 0.1 and 10 mV/s in 1 M NaHCO3 solutions at 23oC, and b) the relationship between bicarbonate concentration and the potentials of the first anodic peaks in carbonate-bicarbonate solutions [Parkins, 1997(A)(B)]. ....................... 20  Figure 2- 11 1.66 mV/s potentiodynamic polarization curves of pure iron in naturally-aerated solutions of different bicarbonate concentrations at 30oC,white and black data points correspond to the absence and presence of magnetic flux, respectively [Lu, 2006]. ...................................... 23  Figure 2- 12 Breakdown (Eb) potentials vs. bicarbonate concentration, retrieved from 0.33 mV/s potentiodynamic polarization scans at room temperature [Mao, 1994]. ...................................... 24 x  Figure 2- 13 Surface images of API-X80 samples in de-oxygenated bicarbonate solutions at the end of 0.33 mV/s potentiodynamic polarization scans at room temperature [Mao, 1994]. .......... 25  Figure 2- 14 a) The potentiodynamic profiles for 0.15 wt%C - 0.47 wt%Mn mild steel in de-oxygenated 0.1 M NaHCO3 solutions, and b) the critical pitting potentials decreasing with chloride concentration, except for at 90oC up to 140 ppm, at different temperatures [Jelinek, 1980]. ............................................................................................................................................ 28  Figure 2- 15 a) The current fluctuations for a carbon steel during a 0.1 mV/s potentiodynamic polarization scan in 0.5 M NaHCO3 solution containing 0.1 M NaCl, and b) the metastable (Em) and breakdown (Eb) potentials as functions of chloride concentration [Tang, 2004]. ................. 29  Figure 2- 16 a) Mott-Schottky, and b) EIS Nyquist plots of passive films formed at 0.2 V vs. SCE for 2 h, in 1 M NaHCO3 + 0.5 M Na2CO3 solutions of different chloride concentrations at 30oC [Li, 2007]. ............................................................................................................................ 31  Figure 2- 17a The 75 mV/min (1.25 mV/s) potentiodynamic polarization scans, in anodic and cathodic directions, in aerated and de-oxygenated 1 M NaHCO3 solutions at 20oC [Von Fraunhofer, 1970]. ........................................................................................................................ 34  Figure 2- 17b Schematics representing the dissolution and cathodic reactions polarization currents in the anodic direction in de-oxygenated and aerated solutions [Von Fraunhofer, 1970]........................................................................................................................................................ 35  Figure 2- 18 The corrosion rates monitored over three periods for A516 Grade 70 carbon steel at 22oC in a 0.2 M NaHCO3/Na2CO3 + 0.1 M NaCl + 0.1 M NaSO4 solution, showing the abrupt increase in the corrosion rates after 27 days of immersion [Sherar, 2010]. .................................. 36  Figure 2- 19 Corrosion products characterization for pure iron at 90oC in solutions, of 0.1-0.2 ppm O2, of 1 M NaHCO3 of a) highly crystalline morphology after immersion for one day, and b) IR spectra of the products analyzed over different periods of time [Savoye, 2001] ................ 38  Figure 2- 19 Corrosion products characterization for pure iron at 90oC in solutions, of 0.1-0.2 ppm O2, of 0.1 M NaHCO3 of c) morphology after immersion for one day, and d) IR spectra of the products analyzed over different periods of time [Savoye, 2001]. ......................................... 39  Figure 2- 20 Pourbaix diagram for Fe-bicarbonate-carbonate-water system at 325 K (52oC) [Hirnyi, 2001]. .............................................................................................................................. 40  Figure 2- 21 Pourbaix diagram of Fe–C–O system at 25oC. ........................................................ 41  Figure 2- 22a Optical image of pretreated steel surface after 55 h of anaerobic immersion, at -750 mV vs. SCE, in a 1 M NaHCO3/Na2CO3 solution [Sherar, 2010]. ............................................... 42  Figure 2- 22b Raman spectra of pretreated steel surface after 55 h of anaerobic immersion, at -750 mV vs. SCE, in a 1 M NaHCO3/Na2CO3 solution [Sherar, 2010]. ........................................ 43 xi  Figure 2- 23 Schematic diagram of the HAZ microstructure viz. the Fe-C phase diagram [Easterling, 1983]. ......................................................................................................................... 45  Figure 3- 1 Distribution of the microstructures of the base steel, HAZ, and the weldment and corresponding variation of potentials [Tsuru, 1987]..................................................................... 50  Figure 3- 2 Schematic illustration of the near-fusion HAZ, weldment, and pipeline base metal. 50  Figure 4- 1 Microstructure of as-received API-X80 steel a) etched with 2% nital solution, b) etched with LePera solution, and c) analyzed by ImajeJ (which showed 7% M-A phase). ......... 54  Figure 4- 2 The differentiation criteria for ferrite and bainite phase quantification [Tafteh, 2001]........................................................................................................................................................ 55  Figure 4- 3 Schematic representation of the strip samples used in the Gleeble© thermal simulation cycles, showing the dimensions of the sample and portion considered for the tests. . 56  Figure 4- 4 Schematic representation of the Gleeble© thermal cycles that simulate the API-X80 HAZs, in Chapter 5, by heating at 250oC/s to 770, 950, 1200, and 1350oC, held for 0.5 s, followed by controlled cooling at 30 and 100oC/s. ....................................................................... 58  Figure 4- 5 Gleeble© thermal cycles that simulate the API-X100 HAZs, in Chapter 6, by heating at 100oC/s to 950oC, held for 0.5 s, followed by controlled cooling at 10, 30 and 60oC/s. .......... 58  Figure 4- 6 Optical micrographs of the microstructures of a) base steel, b) 10oC/s HAZ, c) 30oC/s HAZ and d) 60oC/s HAZ of the API-X100 steel, considered for the investigations in Chapter 6........................................................................................................................................................ 59  Figure 4- 7 Gleeble© thermal cycles that simulate the API-X100 HAZs, for the investigations in Chapters 7, 8, and 9 by heating at 100oC/s to 1200oC, held for 15 s, followed by controlled cooling at 10 and 60oC/s. .............................................................................................................. 60  Figure 4- 8 Optical micrographs of the microstructures of a) 10oC/s HAZ and b) 60oC/s HAZ of API-X100 steel, considered for the investigations in chapters 7, 8 and 9. ................................... 61  Figure 5- 1 Ferrite and bainite contents of API-X80 steel as functions of peak temperature and cooling rate.................................................................................................................................... 69  Figure 5- 2 Hardness and M-A content of API-X80 steel as functions of peak temperature and cooling rate.................................................................................................................................... 70  Figure 5- 3 Selected microstructures of simulated API-X80 HAZs with correspondence to peak temperature and a 30oC/s cooling rate. ......................................................................................... 72  Figure 5- 4 Cyclic potentiodynamic polarization profiles for the API-X80 base steel and HAZs cooled down at a) 30oC/s and b) 100oC/s. (the environmental conditions are in Table 4-4. ........ 74 xii  Figure 5- 5 a) Nyquist and b) Bode plots for the API-X80 base steel and HAZs cooled down at 30oC/s and 100oC/s in EIS measurements carried out at the open circuit potentials. ................... 76  Figure 6- 1 0.05 mV/s cyclic potentiodynamic polarization profiles for the base steel API-X100, and 10, 30, and 60oC/s HAZs........................................................................................................ 79  Figure 6- 2 Profiles of a) the cyclic potentiodynamic polarization, and b) the cathodic polarization scanned at 0.5 mV/s for the base steel API-X100 and the 10, 30, and 60oC/s HAZs........................................................................................................................................................ 83  Figure 6- 3 Profiles of the potentiostatic polarization at 0.5 V vs. SCE for the base steel API-X100, and the 10, 30, and 60oC/s HAZs. ...................................................................................... 84  Figure 6- 4 SEM micrograph of the corrosion products formed on 10oC/s HAZ after 9000 s of immersion. .................................................................................................................................... 86  Figure 6- 5 Nyquist EIS spectra for the base steel API-X100, and the 10, 30, and 60oC/s HAZs after 9000 s of immersion. ............................................................................................................ 87  Figure 6- 6 Bode EIS spectra for the base steel API-X100, and the 10, 30, and 60oC/s HAZs after 3000, 6000, and 9000 s of immersion. .......................................................................................... 88  Figure 6- 7 The equivalent circuit fitted with the experimental EIS data for the base steel API-X100. ............................................................................................................................................. 89  Figure 6- 8 The equivalent circuits fitted with the experimental EIS data, collected after a) 3000, b) 6000, and c) 9000 s, for the 10oC/s HAZ. ................................................................................ 91  Figure 6- 9 The equivalent circuits fitted with the experimental EIS data, collected after a) 3000, b) 6000, and c) 9000 s, for 30oC/s HAZ. ...................................................................................... 93  Figure 6- 10 The equivalent circuit fitted with the experimental EIS data for 60oC/s HAZ. ....... 94  Figure 7- 1 Cyclic voltammograms of scan rates of 0.5, 5, and 25 mV/s, for 10oC/s HAZ in 0.05 M bicarbonate solutions. ............................................................................................................... 97  Figure 7- 2 Cyclic voltammograms, scanned at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, for 10oC/s HAZ in 0.2 and 0.6 M bicarbonate solutions. ............................................................................. 101  Figure 7- 3 SEM Surface micrographs of 10oC/s HAZ, after scanning at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, in 0.05, 0.2 and 0.6 M bicarbonate solutions. .................................................... 102  Figure 7- 4 Cyclic voltammograms, scanned at 0.5 mV/s from -1.3 to 1.2 V vs. SCE, for the 10oC/s HAZ in 0.2 and 0.6 M bicarbonate solutions. ................................................................. 106  xiii  Figure 7- 5 Cyclic voltammograms, scanned at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, for 60oC/s HAZ in 0.05, 0.2, and 0.6 M bicarbonate solutions. ................................................................... 108  Figure 7- 6 SEM Surface micrographs of 60oC/s HAZ, after scanning at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, in 0.05, 0.2 and 0.6 M bicarbonate solutions. .................................................... 109  Figure 7- 7 Cyclic voltammograms scanned at 0.5 mV/s from -1.3 to 1.2 V vs. SCE for the 60oC/s HAZ in 0.05, 0.2, and 0.6 M bicarbonate solutions. ....................................................... 112  Figure 8- 1 Open circuit potentials monitored for 12000 s for 10oC/s HAZ in 0.02, 0.1, 0.5, and 1 M bicarbonate solutions. ............................................................................................................. 115  Figure 8- 2 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to 1.1 V vs. SCE for 10 cycles, for 10oC/s HAZ in a) 0.1, b) 0.5, and c) 1 M bicarbonate, CO2-saturated solutions. ....................... 117  Figure 8- 3 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to -0.6 V vs. SCE for 10 cycles, for 10oC/s HAZ in a) 0.02, b) 0.1 and c) 1 M bicarbonate, CO2-saturated solutions. ...................... 121  Figure 8- 4 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to 1.1 V vs. SCE for 10 cycles for the 60oC/s HAZ in a) 0.1, b) 1 M bicarbonate, CO2-saturated solutions. ................................... 122  Figure 9- 1 Cyclic voltammetry of 10oC/s HAZ of a range from -1.1 to 0.85 V vs. SCE in 0.05 M carbonate solution. ...................................................................................................................... 126  Figure 9- 2 a) The anodic region of the second cycle, b) the anodic regions from the second to the seventh cycles and c) the cathodic region from -0.75 to -1.1 V vs. SCE from the first to the seventh cycles, of the cyclic voltammetry of 10oC/s HAZ in 0.05 M carbonate solution. ......... 128  Figure 9- 3 Cyclic voltammetry of 10oC/s HAZ of a range from -1.1 to 0.85 V vs. SCE in a) 0.1, b) 0.15, and c) 0.25 M carbonate solutions. ................................................................................ 133  Figure 9- 4 Cyclic voltammetry of the 10oC/s HAZ of a range from -1.3 to 1.2 V vs. SCE in a) 0.05, b) 0.1, c) 0.15, and d) 0.25 M carbonate solutions. ........................................................... 135  Figure 9- 5 Micrographs of the top-surface of the corrosion products formed on 10oC/s HAZ, in a) 0.05 and b) 0.25 M carbonate solutions at the end of the -1.3 to 1.2 V vs. SCE cyclic voltammetry. ............................................................................................................................... 136  Figure 9- 6 Cyclic voltammetry of the 60oC/s HAZ of a range from -1.3 to 1.2 V vs. SCE in 0.05, 0.1, 0.15, and 0.25 M carbonate solutions. ................................................................................. 137  Figure 9- 7 Schematic representation of the charge-transfer-limited reduction of water dominating the interface at a potential range from -1.1 to -1 V vs. SCE. ................................... 138  xiv  Figure 9- 8 Schematic representation of the onset of dissolution, driven by carbonate, water and hydroxyl, and the cathodic reduction of oxygen at a potential range from -1 to -0.845 V vs. SCE...................................................................................................................................................... 139  Figure 9- 9 Schematic representation of the simultaneous dissolution and Fe(OH)2 filming at a potential range from -0.845 to -0.745 V vs. SCE. ...................................................................... 140  Figure 9- 10 Schematic representation of the simultaneous dissolution, Fe(OH)2 filming, and onset of FeCO3 precipitation at a potential range from -0.745 to -0.650 V vs. SCE. ................ 141  Figure 9- 11 Schematic representation of the growth of FeCO3 and the coverage of the active surface at a potential range from -0.65 to -0.63 V vs. SCE. ....................................................... 142  Figure 9- 12 Schematic representation of the growth of FeCO3, its oxidation to the iron oxides Fe2O3 and Fe3O4 (which simultaneously also grow directly from the surface), and to the hydrated oxide Fe(OH)3 at the end of the potential range from -0.63 to -0.5 V vs. SCE. ......................... 143  Figure 9- 13 Schematic representation of the overall thickening of the passive film and the growth of Fe(OH)3 to cover most of outer layer at -0.4 V vs. SCE. ........................................... 145  Figure 9- 14 Schematic representation of the thickening and layering of the passive film at a potential range from -0.4 to 0.7 V vs. SCE, before transpassivation starts. ............................... 146  Figure 9- 15 Schematic representation of the transpassivation and oxygen generation at a potential range from 0.70 to 0.85 V vs. SCE. ............................................................................. 147  Figure 9- 16 Schematic representation of the partial recovery of the passive film after transpassivation at a potential range from 0.85 to -0.35 V vs. SCE, before the reduction of oxygen starts. .............................................................................................................................. 148  Figure 9- 17 Schematic representation of the reduction of oxygen onto the passive film, and the reduction of the ferric-based oxides to FeCO3 at a potential range from -0.35 to -0.9 V vs. SCE...................................................................................................................................................... 149  Figure 9- 18 Schematic representation of the onset of the reduction of water, which might significantly result in partial dissolution of the passive film, and the anodic activity of the carbonate-based layer to grow as the potential becomes lower from -0.9 to -1.1 V vs. SCE. .... 149       xv  List of Symbols  Ac1 Temperature at which austenite starts to form by heating (oC) E Equilibrium potential of reaction (V) Eo Standard potential of reaction (V) Ecorr Corrosion potential (V) Epass, Ep Passivation potential (V) Etranspass Transpassivation potential (V) icorr Corrosion current density (A/cm2) icorr Critical current density (A/cm2) ipass, jp  Passivation current density (A/cm2) K1 First dissociation constant  K2 Second dissociation constant Ksp Solubility product N Constant phase element exponent Q Constant phase element (Ω.s-n/cm-2)             Qcp Corrosion products constant phase element (Ω.s-n/cm-2) Qdl Double layer constant phase element (Ω.s-n/cm-2) Qp Porosity constant phase element (Ω.s-n/cm-2) Rch Charge-transfer resistance (Ω.cm2)  Rp Porosity resistance (Ω.cm2) Rs Solution resistance (Ω.cm2) T Temperature (oC) Y Admittance (S.sn/cm2) xvi  Ydl Double layer admittance (S.sn/cm2) Ycp Corrosion products admittance (S.sn/cm2) Zre Real impedance component (Ω.cm2) Zim Imaginary impedance component (Ω.cm2) βa Anodic Tafel slope (V/dec) βc Cathodic Tafel slope (V/dec) IZI Impedance module Ω Angular frequency                xvii  Acknowledgment  I would like to express my deepest appreciation and thanks to my advisor Professor Dr. AKRAM ALFANTAZI for guiding and supporting my research. Without his supervision and knowledge, and the constant advice, patience, and motivation, accomplishing this thesis would not have been possible.  And it is also my pleasure to thank Mr. Jacob Kabel, the electron microscopist in our Department of Materials Engineering, for his help in carrying out the microscopic and chemical composition tests. My thanks are extended to the staff members of the machine shop and also to my colleagues in our research group for their technical assistance and the knowledge they shared.       I want to thank Natural Sciences and Engineering Research Council of Canada (NSERC) and Qatar National Research Fund (QNRF) for their financial support granted for this research project.            xviii    Dedication    ..to my ever beloved parents, and to my loving and sweet brothers and sisters…   1  CHAPTER 1: INTRODUCTION  Oil and gas pipelines are one of the strategic, safest means for large-scale transportation of crude oil and natural gas over long distances on land. In some countries, the pipeline networks transport crude oil and petroleum products in amounts equivalent to thousands of rail cars and tanker trucks every day [Omonbude, 2012]. They are cost-effective, require low energy to operate, and have a low carbon footprint. Pipelines are crucial to the domestic energy security and to meeting the everyday energy demands for transportation, heating, and industrial operation. They contribute to the economy as they export oil and gas and create thousands of jobs in a lot of sectors [Omonbude, 2012].  There are different types of pipelines depending on the distances they travel, the products transported, and the service diameters and thicknesses. Oil pipelines transport crude oil and natural gas liquids from producing fields to refineries, while gas pipelines in smaller diameters transport natural gas from gas wells to processing plants and distribution systems. From the preliminary concept to the time in service, designing and constructing an oil pipeline is a complicated process that involves a lot of decisions and discussions [Kennedy, 1993]. A number of factors are considered, including the distance, volume flow rate, and the type and range of the transported products. These factors determine the pipeline thickness and the number of pumping and compression stations needed to maintain the products flowing continuously and safely. Other factors are taken into account, such as the nature and terrain of the areas the pipelines go through, the nearby communities, and the seasonal changes in climate and temperature [Kennedy, 1993]. Construction of pipelines involves first surveying and staking, after which individual lengths of the pipes are brought in from stockpile sites and laid out end-to-end along 2  the way. Welders join the pipes together using manual or automated welding procedures before which the welds are inspected with ultrasonic methods.  Over the years, pipeline steels have been developed with alloying compositions that give them more strength so that pipelines can be designed with smaller thicknesses and withstand higher operation pressures at lower cost and weight. API-X100 and API-X80 are among the new-generation pipeline steels that have promising reliability for a number of future projects in North America and Europe [Bruschi, 2012].   This thesis addresses the external pipeline corrosion vulnerability of the near-fusion Heat-Affected Zones (HAZs). It is one of the most common problems that critically undermine the safety and integrity of the buried oil and gas pipelines [Singh, 2013]. It depends on the microstructures of the near-fusion HAZs which, alongside the weldment zone and the other HAZs towards the base steel, have different corrosion reactivities and constitute galvanic effects [Tsuru, 1989]. With time, the corrosion reactions localize onto the narrow HAZs and the anodic dissolution segregates from the cathodic reduction, making the corrosion reactions sustain at unequal rates and penetrate through small parts of the weldment region. This accelerates the metal loss and increases the risk for pipeline cracking and other significant failures.  In this thesis, the external pipeline corrosion at this region depends on the environment. When A pipeline coating disbonds, for example, the external surface gets in contact with hydrated soil or ground water. These environments in many of the Canadian pipeline projects are of considerable variability, as they contain different concentrations of bicarbonate, carbonate, chloride, and sulfate in mildly alkaline mediums that could be totally de-oxygenated, or contain oxygen and traces of CO2 at different temperatures [Savoye, 2001, Sherar, 2011].  3  In this thesis, the near-fusion HAZs are simulated by Gleeble© thermal cycles in which the thermal and cooling fluxes of pipeline welding at these zones are programmed with specific parameters. In regard to the environmental conditions, the investigations are carried out for near-fusion API-X80 and API-X100 HAZs in bicarbonate and carbonate solutions. They are naturally-aerated, de-oxygenated, and CO2-saturated, containing different concentrations of bicarbonate and carbonate at different temperatures. They do not necessarily simulate specific external corrosion environments in the field; rather, they contain the corrosive constituents in ranges and conditions that allow understanding the corrosion mechanisms and the significance of the corrosion products. From that, the HAZ microstructures are linked to their corrosion reactivities. This is so that their corrosion behavior can be better understood and predicted in many environmental conditions in the field, and the welding methods can be optimized to result in HAZs having lowest corrosion reactivities.                4  CHAPTER 2: LITERATURE REVIEW1  The literature review begins with consolidating the corrosion mechanisms from a wide scatter of electrochemical studies conducted in bicarbonate and carbonate solutions. The significance of the HAZ microstructure as well as its alloying composition with the properties and growth of the passive films evaluated by the potentiodynamic polarization and other modern electrochemical methods is surveyed. The significance of the environmental conditions of different bicarbonate, carbonate, and chloride concentrations, in naturally-aerated, de-oxygenated, and CO2-saturated solutions at different temperatures and pHs is reviewed. The nature of the corrosion products and the metallurgy of the HAZs are surveyed at the end of the literature review.            2.1    Mechanisms of anodic dissolution and passivation This section surveys first the roles of bicarbonate during the dissolution process of iron and steel, followed by the mechanisms of formation of FeCO3 and its reactivity to oxidize to Fe2O3, Fe3O4, and other iron oxides in bicarbonate and carbonate solutions. The mechanisms of corrosion were studied by polarization and cyclic voltammetry in most of the surveyed studies. Prior to the dissolution of iron immersed in an aqueous solution, water molecules adsorb as follows: Fe + H2O ↔ Fe.H2Oads       (2-1) For bicarbonate, it was reported that it was uninvolved with the first oxidation step which was expressed, in bicarbonate-containing solutions, as follows [Burstein, 1981]:                                                   1 A part of this chapter is published in: F. Eliyan, A. Alfantazi, Mechanisms of Corrosion and Electrochemical Significance of Metallurgy and Environment with Corrosion of Iron and Steel in Bicarbonate and Carbonate Solutions – A Review, Corrosion, in press. 5  Fe.H2Oads ↔ Fe.OHads−  + H+             (2-2) Fe.OHads- → FeOHads + e-                  (2-3) This was reported from scratching constant-potential tests of profiles shown in Figure 2-1, carried out in 0.75 M KHCO3 + 0.05 M K2CO3 solutions in comparison to tests carried out in 0.32 M H3BO3 + 0.19 M K2B4O7 solutions.     It was found that bicarbonate impedes the formation of the FeOHads monolayer as it reacts with it in a second oxidation step producing the complex FeHCO3+ ions as follows [Burstein, 1980]:  FeOHads+ HCO3- → FeHCO3++ OH-+ e-    (2-4) As a result, the linear anodic Tafel region extended to a bare-surface current density higher than that in a borate (inert species) solution, as shown in Figure 2-1 [Burstein, 1980]. The Tafel slope in both solutions was the same, 112 mV/decade, indicating that bicarbonate neither involves with nor increases the rate of the first oxidation step, and reaction (2-4) was considered a process that counteracts the kinetics of film formation.           Figure 2- 1 Bare surface polarization curves for iron, at pH 8.8, in 0.32 M H3BO3 + 0.19 M K2B4O7 solution (in circles), and in 0.75 M KHCO3 + 0.05 M K2CO3 solution (in squares). White is for anodic currents and black is for cathodic currents [Burstein ,1980] 6  Simard et al., on the contrary, considered FeHCO3+ as a product resulting from the involvement (not non-involvement [Burstein, 1980]) of bicarbonate with the dissolution steps of a rotating 1024 mild steel electrode during 5 mV/s voltammogram scans as follows [Simard, 1997]: Fe + HCO3-  → FeHCO3++ 2e-     (2-5)   This was associated with the increase of the critical currents of the first anodic peaks at -0.65 V vs. SCE, presented in Figure 2-2a, with bicarbonate concentration in solutions of pH of 9. The linear relationship of the reciprocals of the critical currents with the (rotation speed)-0.5 in 0.1 M NaHCO3 solutions, as shown in Figure 2-2b, indicated that bicarbonate ions have diffusion-limited roles in driving the dissolution reactions. In addition, Simard et al. mentioned that the FeCO3 precipitation, having higher tendencies and rates of formation with bicarbonate concentration, is governed by the localized concentrations of FeHCO3+. Castro et al. reported that Fe(OH)2, in the beginning forms (Eq. 2-6), before FeCO3 at the first anodic peak of -0.6 V vs. NHE (-0.844 V vs. SCE) in a fashion independent from the rotation speed [Castro, 1986].  Fe + 2H2O → Fe(OH)2+ 2H++ 2e-     (2-6)             Figure 2- 2 a) The effect of bicarbonate concentration on the anodic peaks at -0.65 V vs. SCE, for a mild 1024 steel electrode rotating at 1000 rpm during 5 mV/s cyclic voltammetry at 25oC, and b) the reciprocal of current density against (ω)-0.5 in 0.1 M [Simard, 1997]. 7   Their voltammogram scans, shown in Figure 2-3, were produced at 25 mV/s in solutions containing 0.05 M K2CO3. The second anodic peaks, whose locations and sizes increased with bicarbonate concentration, similar to Simard et al.’s results, were attributed to the partial removal of Fe(OH)2, not to the formation of FeCO3 as follows [Castro, 1986]:    Fe(OH)2 + HCO3- → CO32-+ OH- + Fe2+ + H2O     (2-7) Niu and Cheng reported the instability of Fe(OH)2 in de-oxygenated, dilute bicarbonate solutions with the same mechanism (Eq. 8-7), and in the presence of oxygen they suggested that FeCO3 precipitates after Fe(OH)2 and inhibits the anodic currents of API-X70 steel in near-neutral solutions [Niu, 2007]. Valentini and Moina considered the formation of Fe(OH)2 as the onset of passivation during a temperature-independent, multi-step, charge-transfer process for iron in bicarbonate-carbonate solutions [Valentini, 1985] as follows:  Fe + H2O ↔ Fe(OH) + H++ e-    (2-8) Fe(OH) ↔ Fe(OH)+ + e-             (2-9) Fe(OH)+ + OH- ↔ Fe(OH)2 → Hydrous Fe(OH)2     (2-10) In a few studies, Fe(OH)2 was reported to be an inner layer under FeCO3 forming a multilayered passivation that grows as the applied potentials get higher [Castro, 1991(A), Zhou, 2011]. Similar to another proposition by Rangel et al., the passive film grows at the first anodic peak to become a mixture (not a multilayer) of FeCO3 and Fe(OH)2 [Rangel, 1986].  The precipitation of FeCO3 as a result of direct oxidation (Eq. 8-11 [Davies, 1980]) remained, in some studies, a valid hypothesis in describing the dissolution of ferrite in bicarbonate solutions across specific ranges of the anodic polarization.  Fe + HCO3- → FeCO3+ H+ + 2e-          (2-11)                   8                    Referring again to FeHCO3+, Castro et al. reported two mechanisms [Castro, 1991(B), Castro, 1986] to explain the FeCO3 formation and the roles of bicarbonate and carbonate during which Fe(HCO3)2 could also form as follows:  FeHCO3+ ↔ Fe2++ HCO3-                     (2-12a) Fe2+ + CO32- ↔ FeCO3                         (2-12b)  and Fe2+ + CO32- ↔ FeCO3                         (2-13a) Figure 2- 3 The effect of ω on the second anodic peaks at -0.4 V vs. NHE (-0.644 V vs. SCE) for a rotating iron electrode in 0.75 M KHCO3 + 0.05 M K2CO3 solutions at 25oC [Castro, 1986]. 9        +H+ ↕ -H+                                      (2-13b)             HCO3-  + Fe2+ → FeHCO3+        (2-13c) FeHCO3+ + HCO3-  ↔ Fe(HCO3)2         (2-13d) Alternatively, Lu et al. suggested that another intermediate Fe(HCO3-) agent is involved with the dissolution steps that result in FeCO3 (which could react with bicarbonate to result in new carbon-carrying intermediates) in aerated bicarbonate solutions as follows [Lu, 2006]: Fe + HCO3- ↔ Fe(HCO3-)ads                    (2-14) Fe(HCO3-)ads↔ Fe(HCO3)ads + e-                          (2-15) Fe(HCO3)ads + OH- ↔ Fe(CO3)ads+ H2O+ e-       (2-16) Fe(CO3)ads ↔ FeCO3 metal/solution interface                 (2-17) FeCO3 metal/solution interface ↔ FeCO3                        (2-18) Most of the surveyed studies that addressed the mechanistic corrosion aspects suggested that Fe(OH)2 and FeCO3 oxidize and/or electro-oxidize to Fe2O3, Fe3O4, and other Fe3+ oxides when the test steel is polarized across specific ranges of passivation potentials, like that by [Sherar, 2010]. Lu et al. reported a role for OH- in transforming Fe(OH)2 to Fe(OH)3 and Fe2O3 as [Lu, 2006]: Fe(OH)2 + OH- ↔ Fe(OH)3+ e-                             (2-19) 2Fe(OH)3 ↔ Fe2O3+ 3H2O                                    (2-20) Castro et al. also suggested that FeCO3 oxidizes to FeOOH and Fe(OH)3, correlating that to a third anodic voltammetric peak, shown in Figure 2-4, at -0.2 V vs. NHE (-0.444 V vs. SCE) in 0.75 M KHCO3 + 0.05 M K2CO3 solutions as follows [Castro, 1991(A)]:  10  FeCO3 + 2H2O ↔ (α-FeOOH)or (γ-FeOOH) + CO32-+ 3H+ + e-    (2-21) FeCO3 + 3H2O ↔ Fe(OH)3 + CO32- + 3H+ + e-                               (2-22) The transformation of FeCO3 to Fe2O3 and/or Fe3O4 in relation to the applied potentials, pH, bicarbonate and carbonate concentrations, and temperature, was reported with many Figure 2- 4 Voltammogram scans that show third anodic peaks, at -0.2 V vs. NHE (-0.444 V vs. SCE), which represent the oxidation of FeCO3 to iron hydroxides in 0.75 M KHCO3 + 0.05 M K2CO3 solutions [Valentini, 1985]. 11  mechanisms in literature. The involvement of bicarbonate and OH- with the electrooxidation of FeCO3 to Fe3O4 and Fe2O3 was reported as follows [Lu, 2006]: 3FeCO3 + 5OH- ↔ Fe3O4 + 3HCO3 -+ H2O + 2e-            (2-23) 2FeCO3 + 4OH- ↔ Fe2O3 + 2HCO3- + H2O + 2e-            (2-24) Parkins and Zhou [Parkins, 1997(A)(B)] and Castro et al. [Castro, 1991(B)] reported the FeCO3/Fe2O3 electrooxidation as follows: 2FeCO3 + 3H2O ↔ γ-Fe2O3 + 2HCO3- + 4H+ + 2e-         (2-25) and  2FeCO3 + 3H2O ↔ γ-Fe2O3 + 2CO32- + 6H+ + 2e-           (2-26) In aerated, bicarbonate-carbonate solutions, Zhang and Cheng [Zhang, 2009] and Fu and Cheng [Fu, 2010(A)(B)] attributed the oxidation of FeCO3 to the dissolved oxygen as follows:  4FeCO3 + O2+ 4H2O → 2Fe2O3 + 4HCO3- + 4H+             (2-27)           6FeCO3 + O2 + 6H2O → 2Fe3O4 + 6HCO3- + 6H+            (2-28)       In carbonate solutions, Thomas et al. in an early work linked the first anodic peaks, which increased in size with carbonate concentration, to the formation of FeCO3 during 10 mV/min (0.167 mV/s) potentiodynamic scans as follows [Thomas, 1970]:  Fe + CO32- → FeCO3+ 2e-           (2-29) This was inferred for pure iron in aerated carbonate solutions with a pH of 11 in comparison to NaOH solutions of the same pH. At the first anodic peak of -0.58 V vs. NHE (-0.824 V vs. SCE), it was suggested that the formation of FeCO3 interfered with that of Fe3O4. They observed a color change of the passive film at a second peak of -0.22 V vs. NHE (-0.466 V vs. SCE) to report that FeCO3 electro-oxidizes to Fe2O3 as follows: 12   2FeCO3 + 3H2O ↔ Fe2O3 + 2CO32- + 6H+ + 2e-           (2-30) In de-oxygenated 0.2 mol/L NaHCO3 + Na2CO3 solutions, Lee et al. reported the significance of green rust with stabilizing the corrosion rates. However, in more concentrated solutions, they reported that carbonate complex Fe2+ ions make the FeCO3.H2O films increasingly more porous and, therefore, making the corrosion rates higher [Lee, 2006]. Detecting Fe3O4 in passive films formed in solutions with a pH of 9, it was suggested that the green rust acts as an intermediate product between Fe(OH)2 and γ-FeOOH and/or Fe3O4 in a slow process as follows: 7Fe(OH)2 + 2CO32-+ 2H2O →  4Fe(OH)2.2Fe(OH)3CO3+H2+2OH-        (2-31) 4Fe(OH)2.2Fe(OH)3CO3 + 4H+ →  Fe3O4 + 3Fe2+ + CO32-+ 8H2O          (2-32) 2.2    Significance of microstructure The significance of microstructure was investigated mainly in bicarbonate-carbonate solutions with emphasis on stress-corrosion cracking susceptibility.  At 23oC in 1 M NaHCO3 + 0.5 M Na2CO3 solutions, Zhang and Cheng reported that the acicular and bainitic-ferritic microstructures of a spirally-welded API-X70 steel had the highest passive currents and thinnest n-type passivation in comparison to a ferritic/pearlitic microstructure (of the base steel) and to a grain-boundary ferritic microstructure, which is shown with the other microstructures in figure 2-5 [Zhang, 2009]. The 1 mV/s potentiodynamic scans, carried out in the same study and displayed in Figure 2-6a, did not indicate any specific correlations between the microstructures and the corrosion behavior. This was also the case with the Local Electrochemical Impedance Spectroscope (LEIS) results at different stress levels. In the same solutions at 80oC, the slow 20  13   mV/min (0.333 mV/s) scans, with selected profiles shown in Figure 2-6b, showed no difference in the corrosion behavior among of the Heat-Affected Zone (HAZ) (grain boundary-ferritic and bainitic-ferritic), the weld (acicular-ferritic and upper-bainitic), and the base carbon steels (ferritic/pearlitic), regardless if treated with tungsten inert gas welding or with electron beam welding [Mitsui, 2008]. In another study, the martensitic sample of API-X65 in aerated 0.5 M NaHCO3 + 0.5 M Na2CO3 (pH 9.5) solutions was less susceptible to fail under stress-corrosion cracking than the ferritic-pearlitic sample, given that its higher anodic activity was attributed to the selective dissolution of its 15vol% martensite phase [Stikma, 1985]. In dilute 0.05 and 0.01 M NaHCO3 solutions at 20oC, the as-received (ferritic- Figure 2- 5 Microstructures of a spirally welded X70 at a) base steel, b) HAZ, c) HAZ/weld metal and, d) weld metal [Zhang, 2009]. 14                                             a b Figure 2- 6 Potentiodynamic polarization profiles a) scanned at 1 mV/s in 1 M NaHCO3-0.5 M Na2CO3 solutions at 23oC for different microstructures of spirally-welded X70 steel [Zhang, 2009] and b) scanned at 20 mV/min (0.333 mV/s) in 1 M NaHCO3 + 0.5 M Na2CO3 solutions at 80oC for carbon steel samples welded with tungsten inert gas welding [Stikma, 1985] 15   pearlitic) and water-quenched (martensitic) microstructures of API-X80 exhibited the minimum passive currents in comparison to quenched-and-tempered and water-sprayed microstructures  [Gonzalez-Rodriguez, 2002]. In ultra-dilute 0.005 M bicarbonate solutions, a martensitic API-X70 sample produced by water quenching after a 40-minute heating at 850oC did not passivate and it had the highest anodic currents, as shown in Figure 2-7, unlike those of the ferritic-pearlitic and acicular-ferritic microstructures [Torres-Islas, 2008]. From the surveyed studies in this section, it appears that the ferritic microstructures – along with other local active phases– Figure 2- 7 Significance of heat treatment with the polarization behavior of X-70 steel in 0.005 M NaHCO3 solutions [Torres-Islas, 2008]. 16  promote the formation of thick passive films regardless of the environmental conditions, particularly the concentrations of bicarbonate and carbonate. Microstructures that are bainitic, martensitic, or acicular-ferritic may form unstable passive films but they may not be as intrinsically active to drive the anodic reactions and cathodic reduction at high rates. The most commonly-used tests in evaluating the significance of microstructure are the potentiodynamic polarization, Mott-Schottky analysis, and EIS. 2.3    Significance of alloying composition  The significance of alloying elements with the corrosion behavior in bicarbonate and carbonate solutions has not received as wide an interest as that in CO2-saturated solutions. In aerated 0.5 M bicarbonate solutions, Alves and Brett investigated the significance of 4.5 wt% Cr in chromium steel CALMAX® and 5 wt% Mo and 6.4 wt% W in High Speed Steel (HSS) in comparison to the corrosion behavior of 1015 steel [Alves, 2002(A)]. The alloyed steels had stable passivation that had an earlier onset and low critical currents, as shown in Figures 2-8a and 8b. The high passive currents of HSS were attributed to Mo and W but without suggesting relevant reasons, and the high-potential anodic peaks of the chromium steel were attributed to specific passivation reactions that Cr drives. The passivation of both steels was much more stable against 0.01 M chloride ions than that of the carbon steel. In a different study, Alves and Brett used EIS to analyze the passivation of the three steels in the same solutions and found that the carbon steel had a broader passivation range, from 0 to 0.9 V vs. SCE, than CALMAX® and HSS [Alves, 2002(B)]. They attributed the shorter passivation potential ranges of the two latter steels to the reactivity of chromium disrupting the interfacial reactions at 0.4 V vs. SCE, as follows: CrCr + 4H2O → CrO42- + VCr3'  + 8H+ + 3e-           (2-33)  17   where CrCr represents the Cr3+ cation in the passive film and VCr3'  represents the cation vacancy. Parkins et al. reported, for ferritic steels exhibiting comparable SCC susceptibilities, that introducing 2.08 wt% Ti increases the anodic activity while introducing 4.93 wt% Cr enhances a c b Figure 2- 8 a) Passivation potentials, b) Critical currents, and c) selected 0.2 mV/s potentiodynamic polarization profiles for carbon steel (CS) and the heat-treated High Speed Steel (HSS) and chromium steel (CrS) in de-oxygenated 0.5 M NaHCO3 + 0.01 M KCl solutions at 25oC [Alves, 2002(A)]. 18  the passivation at -0.65 V vs. SCE in 1N NaHCO3 + 1N Na2CO3 solutions at 75oC [Parkins, 1981]. More specifically, Li et al. reported that introducing 3 wt% Cr to a carbon steel increased the donor density of its n-type, semi-conductive passive films by about 2 orders of magnitude. Moreover, they reported that the passive films were highly inhomogeneous and less resistant to deterioration in 1 M NaHCO3 + 0.5 M Na2CO3 solutions [Li, 2007], confirming with Alves and Brett’s findings in bicarbonate solutions [Alves, 2002(B)]. In the same solutions, Riley and Sykes reported that introducing Cr, Ni, or V to carbon steel has minor effects on increasing the inhibitive capability of the passive films, except with Mo, with a passive film that inhibited the dissolution from the early stages of the passivation growth [Riley, 1990].  The significance of alloying composition with the corrosion behavior and passivation protectiveness and stability needs further electrochemical investigations with more advanced and accurate methods. The available literature does not provide sufficient understanding and some of the available findings, especially on the significance of chromium, are critically discrepant.     2.4    Significance of bicarbonate  Bicarbonate was considered an active species of dual effects in a few early studies. In small concentrations, it exacerbates dissolution but, in higher concentrations, it is a protective agent and promotes passivation. Both the detrimental and protective mechanisms in relation to bicarbonate concentration are a subject of ongoing investigations. Brasher reported, with the use of potential/time and weight loss measurements, a 0.006 M bicarbonate concentration below which bicarbonate cannot be a protective agent [Brasher, 1968]. This section surveys the significance of bicarbonate concentration with the corrosion reactions in CO2-saturated, naturally-aerated, and de-oxygenated solutions.  19  2.4.1    CO2-saturated solutions The available studies that address the electrochemical roles of bicarbonate with the corrosion behavior and passivation in CO2-saturated solutions are few and their findings carry contradiction. The possible reasons behind this lie in the many ways an electrochemical method was used to evaluate the corrosion behavior and the variety of the environmental conditions that bicarbonate and carbonate were introduced to. In 1 M Na2CO3 solutions, Parkins and Zhou, who studied the potential bounds for intergranular and transgranular cracking, reported that the potentials of the first potentiodynamic peaks, shown in Figure 2-10a, (which represent qualitatively the onset of passivation) increased from -0.7 to -0.5 V vs. SCE with the bicarbonate concentrations of 0.2 to 1.2 M, for a 0.05 wt% C - 0.3 wt% Mn steel, at 23oC [Parkins, 1997(A)(B)]. The results, which are extracted from 10 mV/s scans, are shown in Figure 2-10b, and the significance of the scan rates was taken into account in the study. Videm  and Koren, however, reported that these potentials decreased from 0.2 to -0.6 V vs. SCE with the bicarbonate concentrations of 0.03, 0.1, to 0.5 M in carbonate-free solutions, while in the 0.01 M solution, no peaks appeared during 50 mV/min (0.833 mV/s) scans of a 0.07 wt% C - 0.03 wt% Mn steel at 25oC [Videm, 1993] . A speciation diagram for a CO2-satureated system is shown below.        Figure 2- 9 Speciation equilibrium of the carbonic species at ambient temperatures for CO2-saturated aqueous systems [Palacios, 1991]. 20                         a b Figure 2- 10 a) Potentiodynamic polarization curves, swept at 0.1 and 10 mV/s in 1 M NaHCO3 solutions at 23oC, and b) the relationship between bicarbonate concentration and the potentials of the first anodic peaks in carbonate-bicarbonate solutions [Parkins, 1997(A)(B)]. 21  2.4.2    Naturally-aerated solutions The corrosion behavior in naturally-aerated solutions received some interest but the synergistic effects of the trace oxygen concentrations were seldom linked to those of bicarbonate and carbonate. In a mechanistic study, Lu et al. correlated the passivation behavior of pure iron with bicarbonate concentration in the presence and absence of a 0.4 T magnetic flux in naturally-aerated, halides-containing solutions at 30oC [Lu, 2006].  As shown in Figure 2-11, bicarbonate markedly affected the passivation behavior when increasing its concentration, indicating that critical physicochemical characteristics should be explored and linked to the electrochemical ones in future studies. In 0.5, 0.75, and 1 M bicarbonate solutions, two anodic peaks appeared at -0.6 and -0.2 V vs. SCE, whose first (at -0.6 V vs. SCE) and second critical current densities were ratios between 8 and 15. In 0.03, 0.1, and 0.2 M solutions, the ratio was nearly 1, but in 0.01 M solutions, only one broad peak appeared at 0.3 V vs. SCE. For API-X100, I evaluated the effects of 0.1, 0.5, and 0.8 M bicarbonate on the corrosion behavior at different temperatures and in the presence and absence of 3 wt% chloride [Eliyan, 2012]. With a fast 0.5 mV/s scan rate, I found that the corrosion rates increased with bicarbonate concentration but transpassivation was at the same 1 V vs. SCE, regardless of bicarbonate concentration and temperature. Moreover, with the use of EIS, the nature of the interfacial interactions did not change with bicarbonate concentration but the charge-transfer resistance increased as a result of the enhanced passivation. Introducing 3 wt% chloride made the interfacial interactions dependent on bicarbonate concentration. For API-X80 steel, Gonzalez-Rodriguez et al. reported a scheme of change in the passivation behavior from showing a continuous dissolution to single broad peaks and smaller multiple peaks of regimes of stable currents that decreased with higher bicarbonate 22  concentrations of 0.005, 0.01, 0.05, and 0.1 M at 20oC, with the use of 1 mV/s potentiodynamic scans [Gonzalez-Rodriguez, 2007]. 2.4.3    De-oxygenated solutions Most of the available studies on corrosion behaviour in bicarbonate and carbonate solutions were carried out in solutions de-oxygenated with either nitrogen or argon, simulating the corrosion conditions of an external pipeline surface in contact with ground water or simulating those at the internal pipeline surface exposed to mildly-alkaline water proportions. I investigated the corrosion behaviour of API-X100 steel in de-oxygenated solutions of 0.05, 0.1, 0.5, and 1 M bicarbonate in the presence and absence of hydrocarbon oil at 30 and 70oC with potentiodynamic polarization and EIS [Eliyan, 2013]. The passive currents increased and the passivation ranges broadened with bicarbonate concentration. Calculating the charge-transfer release of passivation indicated that it thickened with bicarbonate concentration and schematics depicting the passivated interfaces along with the significance of hydrocarbon oil were made. Moreover, the cathodic currents showed an increase during the mass-transfer-limited reduction with bicarbonate concentration. Agreeably, the corrosion rates and passive currents of AISI 4340 steel showed an increase with bicarbonate concentration at different temperatures in solutions synthesized from pure water and simulated sea water (SW, 3.5 pct NaCl) [Mohorich, 2010]. Torres-Islas et al. reported a considerable insensitivity of the dissolution and passivation currents of API-X70 during 1 mV/s scans toward bicarbonate concentrations of 0.005, 0.01, 0.05, and 0.1 M at 50oC [Torres-Islas, 2008]. On the contrary, in another study with 0.005, 0.01, and 0.02 M bicarbonate concentrations, the anodic and cathodic currents showed a noticeable increase with the use of 50 mV/s cyclic voltammetry [Niu, 2007], reflecting the significance of scan rates of potentiodynamic polarization and cyclic voltammetry behind the discrepancy of the current  23   1 M NaHCO3 0.75 M NaHCO3 0.5 M NaHCO3 0.2 M NaHCO3 0.1 M NaHCO3 0.03 M NaHCO3 0.01 M NaHCO3 Figure 2- 11 1.66 mV/s potentiodynamic polarization curves of pure iron in naturally-aerated solutions of different bicarbonate concentrations at 30oC,white and black data points correspond to the absence and presence of magnetic flux, respectively [Lu, 2006]. 24  findings on the passivation behavior. Zhou et al. and Mao et al., who used 1 and 0.33 mV/s polarization scans, respectively, studied the corrosion behavior of API-X80 in bicarbonate solutions [Zhou, 2011, Mao, 1994]. In both studies, the passivation was reported more effective with bicarbonate concentrations, and Mao et al. showed how the surfaces became pitting-immune in relation to the breakdown potentials in bicarbonate solutions of concentrations higher than 0.05 M, as presented in Figure 2-12 and Figure 2-13.   Figure 2- 12 Breakdown (Eb) potentials vs. bicarbonate concentration, retrieved from 0.33 mV/s potentiodynamic polarization scans at room temperature [Mao, 1994].  For 316L stainless steel, Park et al. reported that the pitting potentials increased from 0.4 and 0.6 to 1 V vs. SCE with bicarbonate concentrations 0.1, 0.3, and 0.5 M, in solutions 25  containing 0.5 M chloride at 20oC [Park, 1999]. The number of the pits, which showed re-passivation under (Fe,Cr)CO3 deposits, accordingly decreased.                      Figure 2- 13 Surface images of API-X80 samples in de-oxygenated bicarbonate solutions at the end of 0.33 mV/s potentiodynamic polarization scans at room temperature [Mao, 1994]. 26  2.5    Significance of carbonate The electrochemical corrosion behavior of carbon steels was seldom studied in carbonate solutions without adding bicarbonate. For pure iron, Thomas et al. studied its corrosion behavior in aerated carbonate solutions in an early study [Thomas, 1970]. With thermodynamic analysis and polarization measurements, they reported that iron carbonate first forms during dissolution with magnetite. They correlated that with the size of the first anodic peak at -0.58 V vs. NHE (-0.824 V vs. SCE), which increased with carbonate concentration during 10 mV/min (0.167 mV/s) potentiodynamic scans. A color change was observed between the first peak and a second peak at -0.22 V vs. NHE (-0.466 V vs. SCE), during which iron carbonate oxidized to form Fe2O3 and Fe(OH)3. Bicarbonate could generate from the aqueous equilibrium of carbonate to involve with the anodic and cathodic reactions during the polarization scans. The passive films become thicker and more immune against chloride deterioration with carbonate concentration, electrochemically indicated by potentiostatic polarization and by EIS. In de-oxygenated solutions, Lee et al. reported in a mechanistic and electrochemical study that increasing carbonate concentration complexes more Fe2+ ions from the FeCO3.H2O deposits to make them increasingly porous and, in turn, lead to higher corrosion rates with time [Lee, 2006].   2.6    Significance of chloride The significance of chloride in bicarbonate and carbonate solutions was mostly linked to temperature in reference to the stability of passivation and susceptibility of substrate pitting. Upon adsorption onto a passive film, chloride ions initiate cation/oxygen vacancy pairs that propagate throughout it. This occurs in an autocatalytic process of extents that depend on the chloride concentration. In an extreme case, the vacancy pairs reach the covered substrate to attack it and then to destabilize the passive film. Increasing the temperature decreases the 27  adsorbed chloride concentrations in favor of the competing bicarbonate and carbonate, or water molecules, that keep the passive film protected [Kolotyrkin, 1963]. However, up to certain temperature, the passive film becomes intrinsically unstable by, for instance, the increased solubility of Fe-Cl complexes, as suggested by Jelinek et al. [Jelinek, 1980]. They correlated, in an early work, the pitting susceptibility to chloride concentration for a mild steel in de-oxygenated 0.1 M bicarbonate solutions. As shown from the 40 mV/min (0.667 mV/s) potentiokinetic (potentiodynamic) scans in Figure 2-14a, the pitting potentials decreased with chloride concentration. Interestingly, the high-potential anodic peaks were attributed to ruptures followed by repassivations. As shown in Figure 2-14b, raising the temperature to 90oC made the passive films immune against pitting up to 140 ppm chloride at 0.8 V vs. SCE. I found that adding 3 wt% chloride exacerbates dissolution, prevents passivation, and causes substantial pitting for API-X00 in aerated 0.1, 0.5, and 0.8 M bicarbonate solutions – regardless of temperature [Eliyan, 2012]. Precisely in aerated 0.02 M bicarbonate solutions at 30oC, Lu et al. detected the transformation in the active/passive behavior from showing a unique anodic peak in a solution containing 1x10-3 M chloride to the continuous dissolution in solution containing 0.01 M chloride [Lu, 2006]. El-Naggar reported a similar sensitivity of the active/passive behavior and a decrease in the pitting potentials with chloride concentration, for a carbon steel in de-oxygenated 0.5 M bicarbonate solutions at 25oC, with 25 mV/s cyclic voltammetry [El-Naggar, 2006].  28            a b Figure 2- 14 a) The potentiodynamic profiles for 0.15 wt%C - 0.47 wt%Mn mild steel in de-oxygenated 0.1 M NaHCO3 solutions, and b) the critical pitting potentials decreasing with chloride concentration, except for at 90oC up to 140 ppm, at different temperatures [Jelinek, 1980]. 29  In the same solutions, Neshati et al. used EIS and the power spectral densities to report that the polarization and noise resistances as well as the pitting potentials decreased with chloride concentration [Neshati, 2007].  Tang and Zuo detected the onset of metastable pitting, as shown in Figure 2-15a, for currents measured in 0.5 M bicarbonate + 0.1 M chloride solution to be at -85 mV vs. SCE during 0.1 mV/s potentiodynamic scan [Tang, 2004].            b a Figure 2- 15 a) The current fluctuations for a carbon steel during a 0.1 mV/s potentiodynamic polarization scan in 0.5 M NaHCO3 solution containing 0.1 M NaCl, and b) the metastable (Em) and breakdown (Eb) potentials as functions of chloride concentration [Tang, 2004]. 30  A metastable pitting was characterized with an abrupt rise followed by a gradual decrease in currents up to -40 mV vs. SCE, after which the currents increased indefinitely, corresponding to passivation breakdown and constant pitting attack. The metastable pitting and breakdown potentials decreased with chloride concentration, as presented in Figure 2-15b. Li et al. reported that the mechanisms of chloride passivation attack did not change with chloride concentration in bicarbonate-carbonate solutions, as revealed by the Mott-Schottky analysis of selected profiles shown in Figure 2-16a [Li, 2007]. The donor density increased accordingly, though, and the resistance of passive films, calculated from EIS measurements of profiles shown in Figure 2-16b, decreased from 280 to 5 Ω.cm2 with higher chloride concentration of 0.01 to 0.5 M, in 1 M NaHCO3 + 0.5 M Na2CO3 solutions. The nature of the interfacial interactions were location-independent for stressed API-X70 samples in bicarbonate-carbonate solutions of 0.1 M chloride [Li, 2008(A)]. In carbonate solutions, Brossia and Cragnolino measured the rates and areas of the local dissolution attacks in de-oxygenated solutions of different ratios of chloride and carbonate concentrations for a carbon steel [Brossia, 2000]. The local dissolution rates were rated as minor (<1μA/cm2), moderate, and severe (>0.5mA/cm2), and the rate and nature of the attacks did not depend on the initial carbonate concentration in pH-controlled solutions. For a heat-treated sample exposed to air at 200oC preoxidation for 700 h, the surface area of crevice attacks, calculated from potentiostatic polarization tests, increased from 15 to 20 and 80 mm2 in solutions of chloride-to-carbonate ratios of 0.01, 0.02, and 0.1, respectively.       2.7    Significance of temperature Temperature in bicarbonate and carbonate solutions increases the corrosion rates but its effect on the passivation is linked to the environmental factors to various extents and sometimes contradictory with the many electrochemical methods used. Generally, increasing temperature  31                        a b Figure 2- 16 a) Mott-Schottky, and b) EIS Nyquist plots of passive films formed at 0.2 V vs. SCE for 2 h, in 1 M NaHCO3 + 0.5 M Na2CO3 solutions of different chloride concentrations at 30oC [Li, 2007]. 32  from 20 to 90oC increases the corrosion rates from 500 to 2000 μA/cm2 with the use of 0.5 mV/s potentiodynamic scans in de-oxygenated solutions.  Moreover, the passive films become more immune to deteriorate against trace chloride concentrations. At low temperature, the passive films show evidence of slow formation while at higher temperatures, they form earlier (at low anodic peak potentials) and they exhibit high-potential anodic peaks. Interestingly, though, transpassivation occurs at lower potentials with higher temperatures. At the open circuit potentials and regardless of temperature, the interfacial interactions are generally governed by charge-transfer limited processes of dissolution, adsorption, and cathodic reduction. For API-X80 steel, Li et al. reported the higher corrosion rates and the decrease of protectiveness of the n-type passive films of higher donor density with temperature in CO2-saturated bicarbonate-carbonate solutions [Li, 2008(B)]. It should be noted that the kinetic roles of anodic reactions (versus the cathodic reactions) in increasing the corrosion rates with temperature were not clearly addressed in literature. For N80 carbon steel, Louafi et al. showed that the corrosion rates increased and corrosion potentials decreased with temperature, but the Tafel anodic slopes increased in CO2-saturated 0.5 g/L bicarbonate solutions [Louafi, 2010]. Regardless of temperature, Li et al. found that the passive films formed at 0.2 V vs. SCE for 2 h in 1 M NaHCO3 + 0.5 M Na2CO3 solutions were consistently of a second-time constant under a deposited layer within which absorption and insertion phenomena take place [Li, 2007]. In addition, the passivation potential ranges decreased, the passive currents increased, and the passivation resistance decreased from 16 to 3 Ω.cm2 with temperatures between 20 and 95oC. From the surveyed studies, it seems that without linking the electrochemical findings to the mechanisms of formation and characteristics of the passive films, the influence of temperature in bicarbonate and carbonate solutions will likely remain case-specific.      33  2.8    Significance of natural aeration  Most of the corrosion studies were carried out in de-oxygenated bicarbonate and carbonate solutions and, in the few aerated-solutions studies, the significance of oxygen was not linked to the passivation or anodic reactions, if not to the cathodic reduction which oxygen takes part in. It exists in as high as 10 ppm and it was used be to linked to the shape of the polarization profiles. In an early short communication, Von Fraunhofer commented on the significance of oxygen with the polarization behavior in aerated 1 M bicarbonate solutions for a mild steel at 20oC [Von Fraunhofer, 1970].  The polarization of profiles shown in Figure 2-17a, was scanned forward and backward at 75 mV/min (1.25 mV/s). As shown in Figure 2-17a, the anodic peaks, which were slightly smaller than those in the de-oxygenated solutions, were followed by unique cathodic peaks at 0.5 V vs. SCE of critical -100 μA/cm2 currents. He explained the cathodic reactions of O2 and H2O in the schematics, shown in Figure 2-17b, and he mentioned that Fe2O3/Fe3O4 films are reduced during the backward scan of deep and broad cathodic loops which O2 reduction is partially represented with. It seems that the oxygen reduction interference with the passivation currents appears more when the forward polarization scan is slower. With slow scan rates (~of 0.05 mV/s) in dilute bicarbonate solutions, the oxygen reduction usually results in peaks that could be as broad as 400 mV, appearing as coalescence between the anodic peaks and the oxygen reduction cathodic peaks. However, with the faster scan rate of 0.5 mV/s, only single, small anodic peaks appeared for API-X100 steel at 90oC. This phenomenon is usually reported from investigations of chloride-containing and bicarbonate-carbonate solutions. Interestingly, Sherar et al. attributed the deterioration of the FeCO3-based passive films forming in 27 days of immersion to the minimal O2 traces in de-oxygenated, bicarbonate-carbonate solutions which contained some chloride and sulfate at 22oC [Sherar, 2010]. As a result, the corrosion rates 34  showed an abrupt, continuous increase, as shown in Figure 2-18. Niu et al., in a brief study, mentioned that oxygen in dilute bicarbonate solutions is necessary for effective FeCO3 precipitation [Niu, 2007].                     Figure 2- 17a The 75 mV/min (1.25 mV/s) potentiodynamic polarization scans, in anodic and cathodic directions, in aerated and de-oxygenated 1 M NaHCO3 solutions at 20oC [Von Fraunhofer, 1970]. a 35  Figure 2- 17b Schematics representing the dissolution and cathodic reactions polarization currents in the anodic direction in de-oxygenated and aerated solutions [Von Fraunhofer, 1970]. b 36  2.9    Significance of pH  The passive films in bicarbonate and carbonate solutions form faster and are more protective at a higher pH. Han et al., who studied the mesa corrosion attack of a mild steel in CO2-saturated solutions that contain bicarbonate to adjust the pH from 7 and 8, reported that, accordingly, self-passivation potentials increased from -0.55 to -0.25 V vs. Ag/AgCl, the anodic polarization peaks were larger, and the corrosion potentials were lower at 80oC [Han, 2011(B)]. Li et al. reported that increasing the pH from 7 to 11 decreased the donor density and increased the resistance, protectiveness, and stability of the n-type semi-protective passive films of API-X80   Figure 2- 18 The corrosion rates monitored over three periods for A516 Grade 70 carbon steel at 22oC in a 0.2 M NaHCO3/Na2CO3 + 0.1 M NaCl + 0.1 M NaSO4 solution, showing the abrupt increase in the corrosion rates after 27 days of immersion [Sherar, 2010]. 37  steel in bicarbonate-carbonate solutions at different temperatures [Li, 2008(B)]. It should be noted that the corrosion rates are susceptible to increase in CO2-saturated solutions with higher pH by bicarbonate, which generates from the aqueous equilibrium and acts as a corrosive agent, as reported by Han et al. [Han, 2011(A)]. In their study, increasing the pH from 4 to 8 did not substantially increase H2CO3 concentration. In the alkaline range, increasing the pH does not generally affect the corrosion behavior. Adamy and Cala reported that the corrosion rates remained at 0.2 μA/cm2 with increasing the pH from 10.4 and 11 to 12 in de-oxygenated, 0.04 M KHCO3/K2CO3 solutions at 60oC [Adamy, 1999].       2.10     Significance of corrosion products  In most studies, the corrosion products were studied for samples immersed at the open circuit potentials or with potentiostatic polarization for long periods of time in bicarbonate and carbonate solutions at high temperatures. For pure iron, the corrosion products had a crystalline-like morphology, as shown in Figure 2-19a, after immersion for 1 day in 1 M bicarbonate solutions, with 1-2 ppm O2 at 90oC [Savoye, 2001]. With the use of IR spectra, shown in Figure 2-19b, siderite (FeCO3) was a main constituent and evidence of magnetite (Fe3O4) appeared for samples immersed for longer periods of time. In 0.1 M solutions, the morphology of the corrosion products changed to appear as smaller crystals and dispersed flakes, as shown in Figure 2-19c, and evidence of other products appeared, such as iron (III) hydroxy-carbonate in the IR spectra, as shown in Figure 2-19d. For 1010 steel in de-oxygenated 1 N NaHCO3 + 1 N Na2CO3 solutions at 325 K (52oC), Hirnyi studied the thermodynamics of FeCO3, magnetite-maghemite, and H2O stability in relation to anodic hydrogenation [Hirnyi, 2001] in a Pourbaix 38  diagram, shown in Figure 2-20. Sazzadur Rahman et al. studied the corrosion behavior of ASTM A607 steel in aqueous diagram, shown in Figure 2-20.                      Figure 2- 19 Corrosion products characterization for pure iron at 90oC in solutions, of 0.1-0.2 ppm O2, of 1 M NaHCO3 of a) highly crystalline morphology after immersion for one day, and b) IR spectra of the products analyzed over different periods of time [Savoye, 2001] 39                a c Figure 2- 19 Corrosion products characterization for pure iron at 90oC in solutions, of 0.1-0.2 ppm O2, of 0.1 M NaHCO3 of c) morphology after immersion for one day, and d) IR spectra of the products analyzed over different periods of time [Savoye, 2001]. 40  Sazzadur Rahman et al. studied the corrosion behavior of ASTM A607 steel in environments in which the effects of bicarbonate, chloride, sulfate, and silicate, as well as temperature were studied by electrochemical methods [Sazzadur Rahman, 2008]. They constructed an Fe-C-O potential-pH diagram similar to one we developed for our studies, in Figure 2-21, that shows the stability of FeCO3 as being represented by a small zone - suggesting that it tends to be reduced to Fe(OH)2, which it is in equilibrium with.  Their free energy calculations of the three anodic reactions all suggest the formation of FeCO3.     Figure 2- 20 Pourbaix diagram for Fe-bicarbonate-carbonate-water system at 325 K (52oC) [Hirnyi, 2001]. 41  In optimal conditions for Fe(OH)2 to form in a bicarbonate-containing solution, they mentioned that the presence of small amounts of carbonate converts it to FeCO3. Sherar et al. reported the existence of FeCO3 as clustered crystals under the scanning electron microscope and as bright crystals under the optical microscope of the Figure 2-22a, for a pretreated A516 Grade 70 carbon steel polarized at -0.75 V vs. SCE for 55 h in 1 M NaHCO3/Na2CO3 solution at 20oC [Sherar, 2010].    Figure 2- 21 Pourbaix diagram of Fe–C–O system at 25oC. 42   Figure 2- 22a Optical image of pretreated steel surface after 55 h of anaerobic immersion, at -750 mV vs. SCE, in a 1 M NaHCO3/Na2CO3 solution [Sherar, 2010].  In another study, polarizing polycrystalline iron at -0.4 V vs. NHE (-0.644 V vs. SCE), for 60 min in 0.75 M KHCO3 + 0.05 K2CO3 solutions at 25oC made the corrosion products of dispersed clusters made up of FeCO3 and of distorted morphology with a rotation of 2580 rpm [Valentini, 1985]. In CO2-saturated, concentrated bicarbonate solutions, it is usually found that the corrosion products consist of compact, large grains at medium temperatures and as dispersed, small ones at higher temperatures. As this section comes at the end of the review, and as mentioned earlier, there remains a need for the morphological and compositional characteristics of the corrosion products to be further investigated and re-linked to their mechanisms of growth, a 43  the available electrochemical findings, the environmental conditions, and the metallurgy of iron and steel.        Figure 2- 22b Raman spectra of pretreated steel surface after 55 h of anaerobic immersion, at -750 mV vs. SCE, in a 1 M NaHCO3/Na2CO3 solution [Sherar, 2010].  2.11 Metallurgy of heat-affected zones (HAZs)  The complex physical metallurgy of welding and HAZ has received considerable research interest for a long time [Easterling, 1983, Grong, 1994]. Figure 2-23 shows a schematic diagram illustrating the relationship between the HAZ microstructure and the peak temperature – reached during the welding process – against the Fe-C phase diagram. As can be observed from Figure 2-23, the HAZ microstructure (which lies outside the fusion zone), can be defined by 3 b 44  regions: (1) Close to the fusion zone where the temperature is high enough to transform the starting microstructure completely into the single phase austenite (which is the subject of this thesis), (2) A region where the temperature is only high enough to transform the microstructure into the two-phase α + γ field, and (3) A region furthest away from the weld zone where the temperature is low enough so that the material remains at a relatively low temperature, where α + Fe3C are the phase field. For multi-phase steels such as API-X100, zone (3) can be divided into a tempered zone and another zone of the unaffected base material. In addition to the peak temperature, the heating rate and duration of the thermal cycle are critical in determining the final microstructure. Slower welding can allow for significant austenite grain growth near the weld, resulting in a decrease in nucleation sites for ferrite upon cooling, thereby increasing the potential for non-ferritic transformation products (e.g. martensite) to form [Collins, 1983].  The different regions in the HAZ experience complex and position-dependent thermal histories. The temperature rises rapidly (typically in less than 10 s) to high peak temperatures that could be as high as 1400C near the fusion line during the first weld pass, before it decreases to ambient temperatures. A typical welding process begins by heating the starting microstructure to temperatures above the Ac1 temperature, which is the temperature at which austenite starts to form [Kennedy, 1993]. Typically, for API-X80 and API-X100 steels, the starting microstructure is comprised of acicular ferrite and bainite and a mixture of other microstructures. Above the Ac1 temperature, austenite forms in a manner dependent on the heating rate. There is little literature on austenite formation in API-X100 and API-X80 steels, but the basic view holds, which is that austenite forms rapidly in pearlite colonies and then grows into ferrite at a slower rate, which is controlled by carbon diffusion [Speich, 1981, Garcia, 1981]. 45   Figure 2- 23 Schematic diagram of the HAZ microstructure viz. the Fe-C phase diagram [Easterling, 1983].  A recent work by Huang et al. showed that the process is dependent both on the starting microstructure and on the thermal interruptions during welding processes across the intercritical region [Huang, 2004]. Cordonier et al. made a preliminary attempt to model the morphology and growth of austenite under various nucleation scenarios using phase field models [Cordonier, 2005]. In regions in which austenite completely forms, grain growth of austenite is possible. A particularly important observation here is that the application of classical isothermal grain growth theories to non-isothermal situations often gives unsatisfactory results. The interaction between grain growth and precipitates such as TiN and NbC should be taken into consideration 46  [Breche, 2005]. As the temperature rises in the HAZ, it becomes possible for precipitates such as NbC to dissolve [Bowker, 1989, Palmiere, 1994], thereby reducing the pinning force on austenite grain boundaries and accelerating the austenite grain growth process. After reaching the peak temperature, the steel in the HAZ cools rapidly and decomposition of austenite occurs. The decomposition of austenite is linked to the austenite grain size that develops across a certain position of the HAZ. Several studies have reported the transformation behavior during the continuous cooling of microalloyed, high-strength steels [Thompson, 1996, Thompson, 1990]. Recently, Shome and Mohanty have observed deviations in the microstructures of a HAZ from the predictions of the standard continuous cooling transformation (CCT) diagrams [Shome, 2006]. An interesting study by Spanos et al. examined the microstructural changes in thermally-cycled API-X100 steel, simulating those occurring during a multi-pass weld [Spanos, 1995]. Their results confirmed the significance of peak temperature (which controls the austenite grain size) and the subsequent cooling rate. In terms of the modeling of austenite decomposition, Militzer and Nakata have developed a series of models for hot-rolled HSLA steels [Militzer, 2000, Nakata, 2005] which incorporate austenite grain growth and the subsequent decomposition of austenite. Their work was not directly on the microstructure evolution during a welding process, but the examined reheat conditions and cooling rates describe the microstructure evolution and the modeling methodology was validated. The mechanical properties of the HAZ have been extensively studied, such as in [Grong, 1994, Kim, 2004], which are difficult to calculate and correlate to the complex constituents of HAZs. Recently, Stewart et al. attempted to measure local mechanical properties from HAZ microstructures using shear-punch testing [Stewart, 2006]. Hudson examined the mechanical properties of different API-X100 steels welded with different procedures including single pass, tandem wire, and dual torch – reporting 47  important correlations with the welding procedures [Hudson, 2004]. That work described the thermal history of the HAZs systematically with the welding conditions by using thermocouples embedded into the HAZs.  2.12    Summary of findings in literature The literature review began by consolidating the mechanistic details that explain the roles bicarbonate and carbonate play from the early steps of dissolution till complete passivation. It surveyed and explained the significance of microstructure and alloying composition as well as those of bicarbonate, carbonate, and chloride concentrations in naturally-aerated, de-oxygenated, and CO2-saturated solutions at different temperatures and pHs. The emphasis was on the electrochemical studies in which the corrosion rates and kinetics as well as the electrochemical and physicochemical characteristics of the corrosion products were investigated. The main findings in the literature review are summarized below: 1. It is not clear what role bicarbonate plays during dissolution. In small concentrations it exacerbates dissolution while in high concentrations, it is protective and promotes passivation. In a few early studies, bicarbonate was postulated to disrupt the first oxidation step during which Fe(OH)2 forms as bicarbonate attacks the monolayer Fe(OH)2 in a second oxidation step, making the dissolving surface bare again. In this theory, therefore, bicarbonate neither drives specific first oxidation steps nor increases the rate of the Fe(OH)2 first oxidation step.  In recent voltammetric studies, however, bicarbonate is postulated to drive dissolution so that carbon-carrying adsorbents generate and drive the precipitation of FeCO3. 2. Many mechanisms were proposed for FeCO3 formation and it is generally agreed that the passivation in the beginning forms as a mixture of Fe(OH)2 and FeCO3.     48  3. The passive films undergo chemical transformations. FeCO3 oxidizes and electro-oxidizes to iron oxides and hydroxides like Fe3O4, Fe2O3, and Fe(OH)3 in many mechanisms that depend on the applied potentials, bicarbonate and carbonate concentrations, oxygen, and temperature. In carbonate solutions, FeCO3, which forms by direct surface oxidation and supersaturation, transforms to iron oxides and green rust.  4. The studies that address or indicate the significance of microstructure with the corrosion reactions are scarce. A number of the commonly-used electrochemical techniques do not link the specificity of a microstructure and the corrosion behavior. In general, the ferritic microstructures dissolve at the highest rates but have a tendency to develop thick and stable corrosion products. The passive films on bainitic, acicular-ferritic, and martensitic microstructures are thinner and not as stable.  5. The available findings on the significance of alloying composition are generally insufficient and contradictory. Especially with chromium and molybdenum, the polarization techniques do not provide reproducible and reliable results. They often contradict those of EIS and Mott-Schottky when analyzing the growth and stability of the passive films.  6. The mechanistic roles of carbonate did not receive as wide an interest as bicarbonate’s. With higher carbonate concentration, the passive films grow thicker and more resistant against chloride. Bicarbonate may generate from the aqueous equilibrium of carbonate in considerable amounts that could affect both the anodic and cathodic reactions.    7. Chloride destabilizes the passive films. Depending on its concentration, it competes with the protective bicarbonate and carbonate and adsorbs onto the passive films, initiating autocatalytic cation/oxygen pairs that spread through to destabilize them and cause 49  underlying pitting. Its significance is associated with temperature and the available studies provide a fair consensus on its roles in many environmental conditions. 8. The corrosion rates in bicarbonate and carbonate solutions increase with higher temperature. However, the influence of temperature on passivation seems to be case-specific. More investigations are needed to link the physicochemical characteristics of passivation to the current electrochemical findings so that a clearer understanding on the influence of temperature can be made in relation to bicarbonate and carbonate concentrations and pH.      9. The mechanistic significance of oxygen in naturally-aerated bicarbonate and carbonate solutions was seldom studied. In many cases, it did not seem to markedly affect the corrosion behavior if compared to that in de-oxygenated solutions.   10. It is widely agreed that increasing the pH enhances the passive films to be thick, protective, and inhibitive of the corrosion rates. 11. The transformations the HAZs undergo near the fusion line depend on the peak temperatures and cooling rates. The resulting microstructures critically depend on the austenite grain sizes and grain growth rates, which affect the dynamics of the evolution of ferrite and other microstructures and the significance of carbide and nitride precipitates.               50  CHAPTER 3: SCOPE AND OBJECTIVES  The external pipeline corrosion at the HAZs is a complex and critical problem. Different from the ordinary pipeline corrosion cases, it depends on the metallurgy of the pipeline steels in a unique way. It is dependent on many microstructures that are confined in a small zone having different corrosion reactivities and in galvanic coupling, as illustrated in Figure 3-1. This makes the corrosion reactions segregated onto anodic and cathodic zones and localizes and accelerates the anodic dissolution.   Analysing the first stages of the corrosion phenomena by the electrochemical methods and studying the HAZs most sensitive to the thermal welding conditions, the near-fusion HAZs (shown in Figure 3-2), are the first step towards understanding the problem, which is what this thesis is undertaking.     The electrochemical assessment on the cathodic reduction, dissolution, passivation, and reactivity of the passive films is linked to the microstructures of different near-fusion HAZs.  In Figure 3- 1 Distribution of the microstructures of the base steel, HAZ, and the weldment and corresponding variation of potentials [Tsuru, 1987]. Figure 3- 2 Schematic illustration of the near-fusion HAZ, weldment, and pipeline base metal. 51  laboratory environmental conditions containing different concentrations of bicarbonate and carbonate, key electrochemical correlations will reveal how certain near-fusion HAZ microstructures react in comparison to the base API-X80 and API-X100 steels. This will provide clues on their tendency to sustain the anodic versus the cathodic reactions and the roles of the external corrosion products in the field. The laboratory environmental conditions do not simulate specific environmental field conditions. They contain the corrosive constituents with specifically-selected ranges that make the electrochemical assessments more comprehensive and help to better understand the corrosion phenomena in many conditions in the field.  The objectives of the investigations carried out in this thesis are: 1. Determining the significance of the quantified microstructural features of the simulated near-fusion API-X80 HAZs with the corrosion kinetics and passivation in bicarbonate-chloride solutions.  2. Determining the microstructural significance qualitatively of the ferritic, acicular ferritic, and martensitic API-X100 HAZs with the cathodic kinetics of O2 and the stability and thickness of the passive films in high-temperature, dilute bicarbonate solutions.  3. Determining the roles of bicarbonate with the cathodic reduction, dissolution, passivation, and the stability and reactivity of the passive films in naturally-aerated and CO2-saturated bicarbonate solutions. 4. Determining the roles of carbonate with the same considerations mentioned in point 3 in naturally-aerated solutions. 5. Modelling the corrosion reactions in carbonate solutions. 6. Linking the electrochemical significances of bicarbonate and carbonate qualitatively with microstructural features of near-fusion API-X100 HAZs, slow- and fast-cooled from 1200oC.  52  CHAPTER 4: APPROACH AND METHODOLOGY  In this chapter, the details of the corrosion test setup, the test materials API-X80 and API-X100, the corrosion samples preparation, the microstructural analysis, and the Gleeble© thermal simulations are first outlined. They are followed by the specifications of the simulated HAZs, the laboratory environmental conditions, and the electrochemical tests in each of the upcoming chapters.  The electrochemical methods include the potentiodynamic polarization, potentiostatic polarization, electrochemical impedance spectroscopy, open-circuit potentials monitoring, and cyclic voltammetry. The morphology of the corrosion products was studied by scanning electron microscopy.  The methodology is based on interlinking the corrosion kinetics, rates, and products with the microstructural features of near-fusion HAZs. This is so that their behavior in the field can be predicted and the welding methods can be optimized to result in HAZs that are more resistant to the external pipeline corrosion conditions and having low metallurgical corrosion causes. The laboratory environmental conditions take into account the significance of bicarbonate, carbonate, chloride, natural aeration, de-oxygenation, CO2, and temperature. The near-fusion HAZ microstructures are simulated by thermal cycles in which the peak temperatures (above which the austenitic transformation begins), peak-temperature holding time periods, and cooling rates were the variables of the programmed thermal treatments.      4.1    Corrosion test setup The electrochemical corrosion tests were carried out in a standard 600 mL, three-electrode, multi-port glass jacketed cell. The working electrodes were fabricated from API-X100 and API-X80 steel sheets into flat disks fitted into a Tefzel sample holder. The potentials were 53  measured against the Saturated Calomel Electrode (SCE) of +0.240 V vs. SHE, which was isolated in a salt bridge and maintained at room temperatures. Some experiments were carried out at high temperatures for which a circulator was used to pump a high-temperature water flow through plastic tubes connected to a chamber jacketing the test cell. The de-oxygenation by N2 or CO2 was achieved by a continuous flow of them by a bubbler fitted into the test cell. At high temperatures a condenser was used to maintain the water content. The counter electrode was made of a thin graphite rode. A Versastat 4 potentiostat was used to carry out the experiments which were set, controlled, monitored, and analyzed by a VersaStudio software program.  4.2    Test materials, corrosion test samples preparation, and microstructural analysis The test samples were made of API-X80 and API-X100. The API-X80 steel was supplied by Evraz Inc. and the API-X100 was supplied by CANMET Materials Laboratories. Their chemical compositions, presented in Tables 4-1 and Table 4-2, were analyzed by the International Plasma Labs Limited (IPL).     Table 4-1 Chemical composition of API-X100 steel (wt%)  C Mn Mo Ni Cu Ti Al Cr N Nb P Fe 0.1 2.19 0.2 0.13 0.3 0.01 0.02 0.02 0.006 0.043 0.03 Bal.   Table 4-2 Chemical composition of API-X80 steel (wt%)  C Mn Mo Ni Cu Ti Al Nb N Fe 0.06 1.65 0.24 0.4 0.13 0.012 0.088 0.034 0.005 Bal.  The test samples were wet ground sequentially with 120, 320, and 600 grit emery papers, ultrasonically degreased in ethyl alcohol, rinsed in distilled water, and dried in a hot air stream 54  before immersion in the test solutions. For the microstructural analysis, selected samples were wet-ground up to 1200 grit finish, polished with 6 and 1 μm diamond suspensions, immersed in a 2% nital etchant, of the composition presented in Table 4-3, and treated with alcohol swapping, and dried in a hot air stream. The 2% nital etchant was prepared by mixing 2 mL nitric acid with 98 mL ethanol in which the etching time was between 15 and 30 s. The microstructures were studied by a Nikon EPIPHOT 300 series optical microscope and scanning electron microscopy. To quantify the specific phase martensite-retained austenite (M-A) for the investigations in chapter 5, the LePera solution, of the composition presented in Table 4-3, was used along with Clemex Vision PE image analyzing software. As shown in Figure 4-1b, the M-A phase appears white upon exposure to the LePera Solution [LePera, 1979], and in Figure 4-1c it appears bright white with the use of the image analyzer ImageJ, which quantified it to be 7%.         a b c Figure 4- 1 Microstructure of as-received API-X80 steel a) etched with 2% nital solution, b) etched with LePera solution, and c) analyzed by ImajeJ (which showed 7% M-A phase). 55  The LePera solution was made by mixing two solutions together. The first solution was made by dissolving 4 g picric acid in 100 mL ethanol and the second solution was made by dissolving 1 g sodium metabisulfite in 100 mL deionized water. The two solutions were mixed in a 1:1 ratio immediately before etching. Each sample was etched in freshly prepared LePera solutions for a time period between 35 and 60 s.  Table 4-3 Compositions of the etchants  Etchant Nitric acid Sodium metabisulfite Water picric acid Ethanol 2% nital 2 mL - - - 98 mL LePera - 1 g 100 mL 4 g 100 mL  Phase quantification for the remaining ferrite and bainite was made by using 2% nital as the etchant. The differentiation criteria for ferrite and bainite phases in the microstructure are shown in Figure 4-2.   Figure 4- 2 The differentiation criteria for ferrite and bainite phase quantification [Tafteh, 2001].  One of the differences between bainite and ferrite is that the prior austenite grain boundaries are not observed in the ferritic structures, while they are in bainitic structures. Acicular ferrite: Featureless grains  with non-smooth prior austenite boundaries 56  Depending on the cooling rate, three bainite types or a combination of them may develop. The image analysis procedures, both for LePera and 2% nital etchants, were repeated on at least ten pictures captured from each sample under identical imaging processes. Therefore, each data point on the phase quantification figures, Figures 5-1 and 5-2, in Chapter 5, is the average of at least ten individual measurements. The microhardness was measured by Micromet®3 microhardness tester by Buehler. The load was set to 500 g and the loading time was adjusted to 15 s. Each experiment was repeated ten times to ensure the accuracy and reproducibility of the results. 4.3    Gleeble thermal HAZ simulations  The near-fusion HAZs were simulated by the Gleeble© 3500 thermomechanical processing machine (Dynamic System Inc., Poestenkill, NY). The thermal treatments were carried out with computer-programmable heating and cooling cycles at 1.3x10-3 Pa vacuum. The heating was provided by an electric current running through the samples and the temperature (up to 1350oC) was measured and controlled by NiCr-Ni (type K) thermocouples. A compressed Helium gas flow was used to cool down the samples at controlled rates. Strip samples, of the dimensions shown below in Figure 4-3, were used in the thermal simulations.          Figure 4- 3 Schematic representation of the strip samples used in the Gleeble© thermal simulation cycles, showing the dimensions of the sample and portion considered for the tests. 95 mm 10 mm 1 mm Portion cut out for microstructural analysis and corrosion experiments 10 mm Direction of heat flux  57  The strip samples were 95 mm x 10 mm x 1 mm, with holes to fit through the sample holder of the simulator. This length allowed the temperature to achieve a uniform distribution across the middle part (of ~ 15 mm length, at which four thermocouples were attached) of the heat-treated sample. The 1 mm thickness did not contribute for significant temperature gradients in its direction.  The portions considered for the microstructural analysis and corrosion tests were cut off at 10 mm across the middle of the strip samples. The sections studied were 10 mm x 1 mm, and they were cold mounted and prepared for the microstructural analysis and corrosion experiments according to the procedures outlined in section 4.2.  The Gleeble© thermal parameters were selected in reference to previous temperature readings for HAZs developing from welding high-strength steels with Gas Metal Arc Welding (GMAW) method [Rao, 2010].  The significance of HAZ microstructures is quantitatively linked, in Chapter 5, to the corrosion behavior of API-X80 in low-temperature, bicarbonate-chloride solutions. In the following chapters the significance of HAZ microstructures is qualitatively linked to the corrosion behavior of API-X100 in naturally-aerated and CO2-saturated bicarbonate solutions, in Chapters 6, 7, and 8, and in naturally-aerated carbonate solutions in Chapter 9.  In Chapter 5, the near-fusion HAZs are simulated by eight thermal cycles in which the heating rate is the same, 250oC/s, and the peak temperatures reached are 770, 950, 1200 and 1350oC. They are held for 0.5 s before the samples are cooled at two controlled rates 30 and 100oC/s. The thermal cycles are displayed in the schematic representation in Figure 4-4. Selected HAZ microstructures of these thermal cycles are shown in Chapter 5 along with the microstructural analysis. In Chapter 6, the cooling rates 10, 30, and 60oC/s from the peak temperature 950oC, illustrated in the schematic diagram in Figure 4-5, are considered in 58  simulating three types of near-fusion HAZs which are heated at 100oC/s. The microstructures of the base API-X100 steel and the three HAZs are shown in Figure 4-6.    Figure 4- 4 Schematic representation of the Gleeble© thermal cycles that simulate the API-X80 HAZs, in Chapter 5, by heating at 250oC/s to 770, 950, 1200, and 1350oC, held for 0.5 s, followed by controlled cooling at 30 and 100oC/s.   Figure 4- 5 Gleeble© thermal cycles that simulate the API-X100 HAZs, in Chapter 6, by heating at 100oC/s to 950oC, held for 0.5 s, followed by controlled cooling at 10, 30 and 60oC/s.59  BF F PF Pearlite AF M M a b c d Figure 4- 6 Optical micrographs of the microstructures of a) base steel, b) 10oC/s HAZ, c) 30oC/s HAZ and d) 60oC/s HAZ of the API-X100 steel, considered for the investigations in Chapter 6. 60  The microstructure of the base steel, in Figure 4-6(a), consists of bainitic ferrite (BF) and it shows dispersed ferrite (F). The microstructure of the simulated HAZ cooled at 10oC/s (10oC/s HAZ), in Figure 4-6(b), contains polygonal ferrite (PF) and dispersed pearlite (P). The microstructure of the simulated HAZ cooled at 30oC/s (30oC/s HAZ), in Figure 4-6(c), shows acicular ferrite (AF), and that cooled at 60oC/s (60oC/s HAZ), in Figure 4-6(d), is martensitic.  In Chapters 7, 8, and 9, the near-fusion HAZs are simulated by thermal cycles that heat them at 100oC/s to the peak temperature 1200oC, held for 15 s before cooling them down at controlled rates 10 and 60oC/s, as illustrated in the schematic diagram in Figure 4-7.  The higher temperature and the extended holding time austenizes the simulated HAZs further and affects upon cooling the contents and distributions of the ferritic and bainitic structures [Yi, 2007].           Figure 4- 7 Gleeble© thermal cycles that simulate the API-X100 HAZs, for the investigations in Chapters 7, 8, and 9 by heating at 100oC/s to 1200oC, held for 15 s, followed by controlled cooling at 10 and 60oC/s. The microstructures of the 10 and 60oC/s HAZs are shown in Figure 4-8. The 10oC/s HAZ microstructure shows some of the prior austenite grain boundaries and ferrite, while that of the 60oC/s HAZ contains a mixture of acicular ferrite, bainite, and martensite.61  BF BF      a a b F Prior austenite grain boundaries M AF 500 µm 500 µm Figure 4- 8 Optical micrographs of the microstructures of a) 10oC/s HAZ and b) 60oC/s HAZ of API-X100 steel, considered for the investigations in chapters 7, 8 and 9. 62  4.4    Test solutions The laboratory environmental conditions were prepared to evaluate the corrosion behavior of the simulated HAZs in aqueous solutions that contain bicarbonate, carbonate and chloride. These species exist in the hydrated soils and ground water that get in contact with the external surfaces of the buried pipelines and drive the corrosion reactions. The experiments are carried out to determine the significance of low bicarbonate concentrations in certain environmental conditions, and later the significance of bicarbonate and carbonate concentrations with the corrosion reactions and the morphology of the corrosion products. The experiments therefore are carried out for fundamental evaluations and the laboratory environmental conditions do not necessarily simulate the external corrosion conditions in the field. The experiments ultimately interlink the electrochemical roles of bicarbonate and carbonate with the HAZ microstructures’ corrosion kinetics, rates and products; by which the corrosion behavior in the field can be better understood and predicted in a variety of conditions.    The solutions are synthesized from double-distilled, deionized water in which bicarbonate, carbonate, and chloride were introduced in the form of analytical-grade, Fisher-procured NaHCO3, Na2CO3, and NaCl. The pH values in bicarbonate solutions were between 8.1 and 8.6 and in carbonate solutions were between 9.8 and 10.4, in solutions of different concentrations of bicarbonate and carbonate, temperatures, test time periods, and which were naturally-aerated, de-oxygenated, and CO2-saturated.  In chapter 5, the significance of low 0.1 M bicarbonate concentration and 1 wt% chloride in de-oxygenated solutions at 25oC was studied. Bicarbonate and chloride exist in relatively low concentrations in the external pipeline corrosion environments and the concentrations considered lie within the range [Geldreich, 1996]. The investigations in that Chapter were an initial step to 63  understand the roles of bicarbonate with dissolution and passivation. In Chapter 6, the significance of bicarbonate in a smaller 0.05 M bicarbonate concentration at 90oC in naturally-aerated solutions was investigated; taking into account the significance of possible high temperatures of the external pipeline surfaces and the trace amounts of oxygen in the environment [Singh, 2013]. In chapter 7, the significance of bicarbonate concentrations 0.05, 0.2, and 0.6 M, and in chapter 8, the 0.02, 0.1, 0.5, and 1 M concentrations are respectively investigated in naturally-aerated and CO2-saturated solutions at 25oC. In chapter 9 the significance of carbonate concentrations 0.05, 0.1, 0.15, and 0.25 M with the corrosion reactions in naturally aerated solutions is investigated at 25oC. The environmental conditions simulated for the investigations in each of the upcoming chapters are shown in Table 4-4, along with the test steels, the simulated near-fusion HAZs, and the electrochemical tests.          4.5    Electrochemical tests The electrochemical tests carried out are the potentiodynamic polarization, electrochemical impedance spectroscopy, potentiostatic polarization, open-circuit potentials monitoring, and cyclic voltammetry. They are used to evaluate the corrosion kinetics and rates, and evaluate and model the growth and stability of the passive films and corrosion products. The experiments are carried out at least three times to ensure the reproducibility of the results. The presented electrochemical profiles are the average of three profiles, as so are the measurements and calculations. At the beginning of each electrochemical test, a potentiostatic cathodic conditioning is applied at -2 V vs. SCE to dissolve any air-formed oxides, in time periods between 1000 and 2000 s.     64  In chapter 5, the potentiodynamic polarization and electrochemical impedance spectroscopy tests were carried out. The potentiodynamic scans were carried out at the standard scan rate 0.167 mV/s [ASTM, 1993] from -0.9 to -0.1 V vs. SCE. The scan at -0.1 V vs. SCE was reversed to achieve hysteresis loops that intersect with the original upward passivation currents. The potential range, along with the scan reversal, was selected to investigate the cathodic reduction of bicarbonate, the corrosion kinetics (corrosion current densities and potentials), dissolution, passivation, and the protectiveness of the passive films. The electrochemical impedance spectroscopy tests were carried out to evaluate the interfacial interactions and the interfacial roles of the passive films at the open-circuit potentials. They were carried out at a frequency range from 10 kHz to 0.01 Hz with a sampling rate of 10 points per decade.      In Chapter 6, the potentiodynamic polarization, potentiostatic polarization, and electrochemical impedance spectroscopy tests were carried out. The potentiodynamic polarization was carried out at 0.05 and 0.5 mV/s scan rates from -1.3 to 1.2 V vs. SCE. The scan rate 0.05 mV/s was considered to make the electrochemical reactions sustain slow enough to reveal the significance of oxygen with the cathodic reduction and passivation, in comparison with the potentiodynamic scans carried out at 0.5 mV/s. The scan was reversed at 1.2 V vs. SCE to evaluate the protectiveness of the passive films. The potentiostatic polarization was carried out at 0.5 V vs. SCE for 10000 s, by which the growth and stability of the passive films were evaluated. The electrochemical impedance spectroscopy was carried out at an extended range from 10 kHz to 0.001 Hz, after 3000, 6000, and 9000 s of immersion at the open-circuit potentials in order to model the growth of the passive films with equivalent electric circuits by ZSimpWin© software.    65  In Chapter 7, the cyclic voltammetry was used to model the corrosion reactions starting from the cathodic reduction, anodic dissolution, passivation, reactivity of passivation, transpassivation, and the stability of the passive films – in relation to the bicarbonate concentration. They were carried out at different scan rates of 0.5, 10, and 25 mV/s to characterize the main voltammetric features of the corrosion reactions. They were carried out from -1.3 to 1.2 V vs. SCE and from -0.9 to 1.2 V vs. SCE for ten cycles. The first potential range was selected to evaluate the significance of water reduction with the corrosion reactions and passive films stability, and the second range discounts the significance of water reduction with the passivation process. The scan rate selected in the analysis was 0.5 mV/s, which is the scan rate conventionally selected in most of the available studies in literature – as mentioned in the literature review – and by which the results can be compared with. The top-surface morphology of the corrosion products is studied by scanning electron microscopy.    In Chapter 8, the open-circuit potentials were monitored for 12000 s to evaluate the interrelation between the hydrated CO2 and bicarbonate with the corrosion reactions and with the passivation process. In independent tests, the cyclic voltammetry was used to model the corrosion reactions and evaluate the significances of transpassivation and bicarbonate concentration across potential ranges from -1.1 to 1 V vs. SCE and from -1.1 to -0.6 V vs. SCE, for ten cycles at 1 mV/s. The 1 mV/s scan rate is among the conventionally selected scan rates for cyclic voltammetry carried out in similar environmental conditions.  In chapter 9, the cyclic voltammetry was used to evaluate and model the corrosion reactions in carbonate solutions, across the potential ranges from -1.1 to 0.85 V vs. SCE, and from -1.3 to 1.2 V vs. SCE, for seven cycles at 1 mV/s. The top-surface morphology of the corrosion products was studied by the scanning electron microscopy.   66  Table 4-4  Details of the test steels, simulated near-fusion HAZs, tests, and test solutions  * The details of the tests carried out are outlined in section 4.5.  Chapter Steel Simulated near-fusion HAZs Tests* Test solutions Chapter 5 API-X80 -Heating rate: 250oC/s. -Peak temperatures: 770, 950, 1200 and 1350oC.  -Peak temperature holding time: 0.5 s. -Cooling rates: 30 and 100oC/s.  -Potentiodynamic polarization. -Electrochemical impedance.   0.1 M bicarbonate and 1 wt% chloride 25oC De-oxygenated solutions Chapter 6 API-X100 -Heating rate: 100oC/s. -Peak temperatures: 950oC. -Peak temperature holding time: 0.5 s. -Cooling rates: 10, 30 and 60oC/s. -Potentiodynamic polarization. -Potentiostatic polarization. -Electrochemical impedance.   0.05 M bicarbonate  90oC Naturally-aerated solutions  Chapter 7 API-X100 -Heating rate: 100oC/s. -Peak temperatures: 1200oC. -Peak temperature holding time: 15 s. -Cooling rates: 10 and 60oC/s. Cyclic voltammetry 0.05, 0.2, and 0.6 M bicarbonate 25oC Naturally-aerated solutions Chapter 8 API-X100 -Heating rate: 100oC/s. -Peak temperatures: 1200oC. -Peak temperature holding time: 15 s. -Cooling rates: 10 and 60oC/s. Cyclic voltammetry 0.02, 0.1, 0.5 and 1 M bicarbonate 25oC CO2-saturated solutions Chapter 9 API-X100 -Heating rate: 100oC/s. -Peak temperatures: 1200oC. -Peak temperature holding time: 15 s. -Cooling rates: 10 and 60oC/s. Cyclic voltammetry 0.05, 0.1, 0.15, and 0.25 M carbonate 25oC Naturally-aerated solutions 67  CHAPTER 5: SIGNIFICANCE OF FERRITE, MARTENSITE-RETAINED AUSTENITE, AND BAINITE WITH CORROSION KINETICS AND PASSIVATION OF API-X80 STEEL HAZs2    The electrochemical investigations in this chapter were carried out to link between the quantities of selected microstructural features, of the simulated near-fusion HAZs, and their corrosion behavior. The HAZs were made of API-X80 steel by the thermal cycles represented in Figure 4-4. The details of the thermal cycles carried out are outlined in section 4.3. The electrochemical tests were the potentiodynamic polarization and electrochemical impedance spectroscopy, with the parameters outlined in section 4.5. They were used to evaluate the cathodic reduction, corrosion kinetics, dissolution, the passivation process, and the stability of the passive films – in comparison to the base API-X80 steel – in de-oxygenated solutions containing 0.1 M bicarbonate and 1 wt% chloride at 25oC.  5.1    Microstructural analysis The final phases in the steel microstructure depend on the applied thermal cycles. Three main factors have been reported determining the formation of the different phases in the high-strength low-alloy (HSLA) steels [Olasolo, 2011]. The prior-austenite grain size is the first factor since the grain boundaries provide the preferred sites for the nucleation of new phases. Therefore, the peak temperature affects the final phase content by affecting the austenite grain size and consequently the grain boundaries. The second factor is the cooling rate. Transformation of austenite to ferrite takes place by diffusion. Ferrite nucleates on the austenite grain boundaries and grows into the grains [Porter, 2000], which, like all diffusional transformations, is highly affected by the cooling rate. The third factor, which is significant in the HSLA steels, is the state                                                  2 A part of this chapter is published in: F. Mohammadi, F. Eliyan, A. Alfantazi, Corrosion of simulated weld HAZ of API X-80 pipeline steel, Corrosion Science, Volume 63, 2012, 323-333. 68  of Nb, whether it is in the solution or precipitated. It was reported for API-X80 that the presence of Nb in the precipitated form generally results in higher austenite decomposition temperatures [Tafteh, 2001]. The higher the peak temperature, the higher the Nb present in the solution.  Figure 4-4 shows schematics of the thermal cycles carried out by the Gleeble© machine for the simulations of the near-fusion HAZs. For both the cooling rates 30 and 100oC/s, the temperature at which austenite decomposition occurs by diffusional processes gets constantly less by increasing the peak temperature – in relation to the austenite grain size and Nb content. Increasing the peak temperature increases the grain size of the austenite phase. This in turn decreases the concentration of grain boundaries in the microstructure and as a result decreases the number of the preferred sites for the ferrite nucleation; requiring larger undercooling rates for the transformation to take place. In addition, increasing the peak temperature generally increases the Nb content in the solution (austenite), and this increases the transformation temperature [Tafteh, 2001]. Figure 5-1 shows the results of the image analysis for ferrite and bainite phase quantification. The highest ferrite percentage was observed at 770C with the 30C/s cooling rate. This sample also contained the lowest amount of bainite compared to the others. Increasing the peak temperature increased the bainite content cooled both at 30 and 100oC/s. The ferrite content, however, continuously decreased with higher peak temperature. This is ascribed to the effect of temperature on the grain size and, thus, the grain boundary concentration. Increasing the peak temperature increases the austenite grain size constantly and the number of sites for ferrite nucleation decreases and austenite tends to decompose through less diffusive processes. Therefore, at high temperatures, formation of bainite is favored and a higher percentage of bainite is detected. The effect of the cooling rate on the percentages of ferrite and bainite phases is not significant. 69                However, except for the 770C peak temperature, the 100C/s cooling rate resulted in slightly higher bainite content in the samples. This means that both cooling rates are high enough to favor the formation of bainite, given the austenite grain size becomes large enough. Figure 5-2 shows the results of M-A phase quantification along with the measured hardness values. The highest M-A phase percentage was in samples cooled down from the lowest peak temperature (770C), at 30C/s, producing more M-A than that cooled down at 100C/s. This can be related to the effect of the cooling rate on the ferrite formation. When the cooling rate is low, austenite first transforms to ferrite and, therefore, become enriched in carbon due to the rejection of carbon from ferrite.  Figure 5- 1 Ferrite and bainite contents of API-X80 steel as functions of peak temperature and cooling rate. 70               Therefore, the remaining austenite becomes of higher tendency to transform to martensite as the temperature is lowered. However, at higher cooling rates, conditions are more suitable for bainite formation, which results in the remaining austenite being less enriched in carbon and less prone to transform to martensite, compared to the initial case of the ferrite formation. The effect of the peak temperature on the M-A content can be described with the same phenomenon and reasons. At lower peak temperatures the austenite grains are small and the concentration of grain boundaries is fairly high. Therefore, formation of ferrite from the austenite is highly favorable to enriching the remaining austenite in carbon. Therefore, as the temperature is lowered, martensite can form from the remaining austenite. On the contrary, when the peak temperature is higher, the austenite grains are enlarged and thus the concentration of grain boundaries is less, which favors Figure 5- 2 Hardness and M-A content of API-X80 steel as functions of peak temperature and cooling rate. 71  the formation of bainite. Therefore, higher M-A contents corresponds to lower peak temperatures and lower cooling rates. As is apparent in Figure 5-2, the hardness of the samples is not a function of their martensite content. However, if the results are compared with their bainite contents in Figure 5-1, it can be concluded that the hardness of the samples increased with the bainite content. The measured values of the hardness were in the middle range of previously reported values for hardness of bainite and ferrite [De, 2003]. Between 1200C and 1350C, the bainite contents of the samples were relatively similar for the 100C/s cooling rate, however, the hardness values differed significantly. Although the as-received sample contained a relatively smaller percentage of ferrite, its hardness was comparable to those of simulated HAZs heated up to 770C and 950C.  Figure 5-3 shows the microstructures of the simulated HAZs heated up to 770C, 950C, 1200C, and 1350C, and cooled down at 30C/s. The microstructures are identified in reference to the microstructural study in [Chen, 2014], and to the criteria in [Tafteh, 2011]. The primary austenite grain boundaries are clearly visible for the 1200C and 1350C HAZs, and the grain sizes are larger for the 1350C HAZ. The microstructures of these samples include mostly bainite with small fractions of ferrite. The 950C HAZ microstructure contains a mixture of bainite and ferrite of relatively equal percentages, while the 770C HAZ microstructure is mostly ferritic with small fractions of bainite.  5.2    Potentiodynamic polarization testing Figure 5-4 shows the cyclic potentiodynamic polarization profiles of the base API-X80 steel and the HAZs cooled down at 30 and 100C/s. The potentiodynamic polarization was carried out at 0.167 mV/s from -0.9 to -0.1 V vs. SCE. From -0.9 V vs. SCE to the corrosion 72  potentials, the cathodic reduction current densities representing the cathodic reduction of bicarbonate were decreasing and increasingly becoming under charge-transfer control.    The simulated HAZs exhibited similar corrosion kinetics both during the cathodic reduction and anodic dissolution. The corrosion kinetics data were extracted from the diagrams and are shown in Table 5-1. The corrosion potentials (Ecorr), for the simulated HAZs were BF Bainite Prior austenite grain boundaries F F Figure 5- 3 Selected microstructures of simulated API-X80 HAZs with correspondence to peak temperature and a 30oC/s cooling rate. 73  apparently comparable, around -0.820 V vs. SCE. The corrosion current densities (icorr) did not vary significantly, which were nearly 5 µA/cm2. In comparison to the base steel, the corrosion potentials of the HAZs were all lower than that of the base API-X80 steel, of about -0.77 V vs. SCE, but its corrosion current density was nearly as low as 1 µA/cm2. This indicates that the simulated near-fusion HAZs have tendency to drive the anodic dissolution, in a possible galvanic coupling with the base steel that have tendency to drive the cathodic reduction. The anodic, cathodic and passivation reactions are outlined in detail in chapter 6, which addresses the corrosion reactions in a similar bicarbonate-based environment. Table 5-1Corrosion parameters and passive potential extracted from the potentiodynamic polarization profiles   The anodic and cathodic Tafel slopes (βa) and (βc) did not indicate at significant variations between the mechanisms of anodic dissolution, cathodic reduction, and the microstructures of the tested samples. The passivation started at nearly -0.63 V vs. SCE, corresponding to anodic peaks appeared after anodic dissolution. Their intensities did not vary significantly both for the simulated HAZs cooled at 30 and 100oC/s. For 30oC/s HAZs, only the ones cooled from 770 and 950oC exhibited a stable passivation with low passive currents, attributed to the higher ferrite and M-A contents in their microstructures which would dissolve to enhance the interfacial constituents that drive better passivation [Zhang, 2009].  74                          a a b Figure 5- 4 Cyclic potentiodynamic polarization profiles for the API-X80 base steel and HAZs cooled down at a) 30oC/s and b) 100oC/s. (the environmental conditions are in Table 4-4. 75  The base steel and the HAZs cooled from 1200 and 1350oC did not exhibit stable passivation regimes above -0.5 V vs. SCE. For HAZs cooled down at 100oC/s, their corrosion behavior and kinetics were in general similar to those cooled down at 30oC/s, and they tended to show stable passivation. It seems that both for 30 and 100oC/s HAZs, the higher the peak temperatures the higher the passive current densities, corresponding to thinner passive films [Zhang, 2009] and indicating that the bainitic structures are not as effective as the ferritic ones in promoting the passivation process.    5.3    Electrochemical impedance spectroscopy testing Figure 5-5 shows the Nyquist plots for the HAZs cooled at 30oC/s and the phase Bode diagrams for the HAZs cooled at 100oC/s. From the diagrams, the corrosion mechanisms and the nature of the interfacial interactions both for the base steel and the HAZs did not vary significantly. The interfacial interactions, at the open-circuit potentials, were of two-time constant nature as a passive film formed under which the corrosion reactions were under mass transfer control. In the Nyquist plots, this is characterized by semi-circle plots followed by straight line at the low-frequency regions, corresponding to mass-transfer controlled processes governed by the Warburg element [Hamadou, 2005]. It is difficult to correlate the EIS results to the effectiveness of the passive films as the interfacial processes could also be governed by diffusion, regardless of the of the natures of the passive films. In a previous study by Davies and Burstein, they reported that the corrosion reactions for pipeline steels at the open-circuit potentials get governed at later stages both by the passive films and the intermediate complexes that migrate out from the interface to the bulk solution [Linter, 1999]. The EIS results are in agreement with the polarization results, as the size of the Nyquist plots of the HAZs are smaller 76  than that of the base steel, corresponding to a lower charge-transfer resistance, or higher corrosion current density.                                         a b Figure 5- 5 a) Nyquist and b) Bode plots for the API-X80 base steel and HAZs cooled down at 30oC/s and 100oC/s in EIS measurements carried out at the open circuit potentials. 77  5.4. Summary Samples of the API-X80 pipeline steel were heated up to four peak temperatures and cooled down at 30 and 100°C/s to simulate a number of near-fusion HAZs. Both cooling rates were high enough to produce bainitic microstructures. The lowest peak temperature resulted in the highest amounts of M-A phase in the microstructure. The as-received samples had the lowest corrosion current densities. The presence of M-A phase in the microstructure broadened the passive region to higher potentials. Microstructures with higher bainite content exhibited higher passive current densities.                     78  CHAPTER 6: ELECTROCHEMICAL CHARACTERISTICS, STABILITY, AND MODELING OF PASSIVE FILMS OF API-X100 STEEL HAZs IN HIGH-TEMPERATURE, DILUTE BICARBONATE SOLUTIONS3  The main objective in this chapter is to evaluate the anodic and cathodic corrosion kinetics and evaluate and model the passivation process in high-temperature, dilute bicarbonate solutions, for three types of API-X100 HAZs. They were simulated according to the schematic thermal cycles illustrated in Figure 4-5. Their microstructures, along with the base API-X100 steel microstructure, are shown in Figure 4-6, and they are explained in section 4.3. The electrochemical tests, outlined in section 4.4, were carried out in naturally-aerated solutions containing 0.05 M bicarbonate at 90oC. The electrochemical methods, outlined in section 4.5, include the potentiodynamic polarization, potentiostatic polarization, and EIS.  6.1    Potentiodynamic polarization testing The microstructure-dependent passive currents and the interference of oxygen reduction with the active-passive transition are the two main features revealed by the slow 0.05 mV/s potentiodynamic scan, as shown in Figure 6-1. The cathodic currents, dominantly representing water reduction at -1.3 V vs. SCE, were 5 mA/cm2 as high, except for the acicular-ferritic 30oC/s HAZ, of nearly 130 µA/cm2. The anodic dissolution, which was of high currents below -0.7 V vs. SCE, incorporates bicarbonate and hydroxyl species – in mechanisms reported as seemingly independent from the minimal oxygen traces [Rangel, 1989, Davies, 1980]. The nature of dissolution (given it was controlled and slow), as indicated by the current densities, was independent from a HAZ microstructure type. In literature, the extent of involvement of bicarbonate and hydroxyl, in such relatively dilute bicarbonate solutions, is unclear. At the                                                  3 A version of this chapter is published in: F. Eliyan, A. Alfantazi, Corrosion of the Heat-Affected Zones (HAZs) of API-X100 pipeline steel in dilute bicarbonate solutions at 90°C – An electrochemical evaluation, Corrosion Science, Volume 74, 2013, 297-307. 79  active-passive transition, the nature of the dissolution products and the proportions of their constituents of Fe(OH)2 and FeCO3, outlined in Figure 6-2, in a morphology in which they can be multilayered or mixed had received an earlier interest, but the findings carried some controversy. Electrochemically, it is mainly on whether it is the bicarbonate or hydroxyl initially (or more dominantly) drives the charge-transfer steps after which the growth of the corrosion Figure 6- 1 0.05 mV/s cyclic potentiodynamic polarization profiles for the base steel API-X100, and 10, 30, and 60oC/s HAZs. 80  products depend at higher potentials. Reported from voltammetric scans for iron and mild steels in bicarbonate-based solutions of pH of no less than 8.4, Fe(OH)2 forms as a defective, hydrous layer during the “first stage” of oxidation [Valentini, 1985, Castro, 1986,  Rangel, 1989]. Taking first into account the thermodynamic tendency of formation; the temperature-independent, charge-transfer steps of Fe(OH)2 formation were reported as: Fe + H2O ↔ Fe(OH) + H+ + e-                                                      (6-1a) Fe(OH) ↔ Fe(OH)++ e-                                                                 (6-1b) Fe(OH)+ + OH- ↔ Fe(OH)2 → Hydrous Fe(OH)2                        (6-1c) Lu et al. reported step (1b) as rate-determining, in a mechanism hydroxyl acts as an active adsorbent, and in which Fe(OH) is a precursor of dissolution, reporting the overall reaction of Fe(OH)2 formation as [Lu, 2006]:   Fe + 2OH- → Fe(OH)2+ 2e-           (6-2)                              On the other hand, Simard et al. reported, as concluded from voltammetric scans of rotating mild steel electrodes in bicarbonate solutions, that bicarbonate directly oxidises iron (Eq. 6-3) in proportion with its concentration [Simard, 1997]. Fe + HCO3-  → FeHCO3++ 2e-    (6-3)  That was associated to the higher critical currents of the first anodic peaks, at -0.65 V vs. SCE, with the bicarbonate concentration - higher critical currents whom Castro et al., on the other hand, attributed to the partial dissolution of Fe(OH)2 by bicarbonate as [Castro, 1986]: Fe(OH)2 + HCO3- → CO32-+ OH- + Fe2+ + H2O        (6-4) In general, and regardless of the controversy on the dissolution steps, the interfaces are considered to be passivated, more rapidly in our case at 90oC, with a mixture of Fe(OH)2 and 81  FeCO3. FeCO3, which Niu et al. reported to form in aerated bicarbonate solutions [Niu, 2007], and as being stable in Pourbaix systems [Pourbaix, 1965, Hirnyi, 2001, Sazzadur Rahman, 2008], can grow mainly by carbon-carrying intermediates [Lu, 2006], or, as an outer layer [Simard, 1997], forming as a result of Fe(OH)2 reacting with bicarbonate as follows [Rangel, 1989]: Fe(OH)2 + HCO3- → FeCO3+ H2O + OH-               (6-5) The appearance of relatively broad peaks extending from about -0.77 to -0.5 V vs. SCE is attributed to the intersection of the cathodic currents of oxygen reduction (Eq. 6-6) with the passive currents at least at two potentials.          O2+2H2O + 4 e-→ 4OH-       (6-6)     They can be viewed as a coalescence between the first anodic and cathodic peaks (which occurred in a highly reproducible manner), except for 30oC/s HAZ, which initially showed an anodic peak below multiple, smaller peaks below -0.2 V vs. SCE. The cathodic behavior, which showed the minimum currents in that case, was not reported as distinct for similar acicular-ferritic microstructures when polarized at higher scan rates [Zhang, 2009, Gonzalez-Rodriguez, 2007]. The minimum cathodic activity of such a microstructure was reported from harsher corrosive conditions [Ramírez, 2008]. Above the mixed-control regions at almost -0.2 V vs. SCE, the significance of microstructure appeared the most during passivation, which is of electronic and morphological properties on whom largely could depend [Zhang, 2009, Kwok, 2007]. The as-received and 10oC/s HAZ, of the lowest 3µA/cm2 passive currents, are of considerably large ferrite grains that dissolve much more readily. This facilitates the corrosion products to precipitate into the conductive, anchoring cementite (Fe3C) platelets as continually thickening, compact passive 82  films. For the 30 and 60oC/s HAZs, however, the greater lattice defects of their more rapidly cooled, ferritic-bainitic, and martensitic microstructures seemed to make the passive films more active and of greater donor density. As a result, they become thinner, whose currents were on average 35 and 10 µA/cm2, respectively, showing also a gradual decrease with the upscan. Transpassivation occurred between 0.45 and 0.6 V vs. SCE. There were peaks that appeared more reproducibly for 30 and 60oC/s HAZs at 0.9 V vs. SCE, which can be ascribed, most likely, to further oxidations of the FeCO3-based films to Fe2O3 and/or Fe3O4. Although such transformations were found evident to take place in a number of electrochemical and characterization studies, the reactions and, in turn, the responsible species for driving the transformations were many and different. This actually adds – given that their correlation to a microstructure type is very difficult to establish for such conditions – to the original difficulty in dealing with the chemical synergism imposed by HCO3- , O2, H2O, and OH-. These species can be involved in active roles at different potentials at which different paths for a certain transformation can be proposed. Considering, for example, Fe2O3, below is shown reactions proposed of its transformation from FeCO3 in electrochemical studies carried out in bicarbonate-based solutions: 4FeCO3 + O2 + 4H2O → 2Fe2O3 + 4HCO3- + 4H+     (6-7) [Fu, 2010(A)(B)]   2FeCO3+3H2O → Fe2O3+2CO32-+6H++2e-                 (6-8) [Parkins, 1997(A)(B)] 2FeCO3+ 4OH- → Fe2O3 + 2HCO3- + H2O + 2e-        (6-9) [Lu, 2006] After reversing the currents, at the vertex 1.2 V vs. SCE, all samples exhibited negative hysteresis loops, suggesting considerably stable, pitting-immune passive films. In a short communication, Von Fraunhofer attributed the negative loops to the reduction of O2 to OH- [Von Fraunhofer, 1970], although that the reduction of Fe2O3 and/or Fe3O4 can still not be discounted 83  in our similar case. The 0.5 mV/s polarization profiles are shown in Figure 6-2a. The oxygen-reduction cathodic peaks did not appear and the passivation behavior of all samples was relatively comparable and the cyclic loops were negative. It is worth mentioning that the corrosion potentials, of almost between -0.8 and -0.77 V vs. SCE, were in few millivolts difference with those of the slower 0.05 mV/s potentiodynamic scan. In independent tests, the significance of a microstructure type with the cathodic currents is shown by 0.5 mV/s potentiodynamic scans from -2 V vs. SCE to the corrosion potentials, of profiles shown in Figure 6-2b. The reduction was initially under mass-transfer control of currents of 45 and 70 mV/s, for the 30oC/s HAZ, and the 10 and 60oC/s HAZs, respectively. Further studies are encouraged to study the interrelation between a HAZ microstructure and the rate of the cathodic reactions, with slower scans especially when the reduction becomes under a mixed control of charge and mass transfer.  6.2    Potentiostatic polarization testing The potentiostatic polarization experiments were carried out to measure the currents, and observe the manner they change with, over 10000 s, at 0.5 V vs. SCE – a potential at which all Figure 6- 2 Profiles of a) the cyclic potentiodynamic polarization, and b) the cathodic polarization scanned at 0.5 mV/s for the base steel API-X100 and the 10, 30, and 60oC/s HAZs. 84  samples underwent a passive state. The lowest 6 µA/cm2 currents, as shown in Figure 6-3, of fluctuations limited to between 3 and 10 µA/cm2 over time, were exhibited by the 10oC/s HAZ.   Figure 6- 3 Profiles of the potentiostatic polarization at 0.5 V vs. SCE for the base steel API-X100, and the 10, 30, and 60oC/s HAZs.  The currents of the as-received sample were initially as high as 1.6 mA/cm2 during the first 2500 s before a continuous decay to reach 100 µA/cm2. The more stable currents of the 10oC/s HAZ suggest, as explained in section 5.1, that the passive films grew to be thicker/compact in a direct dependence on the initial dissolution of its more abundant ferrite content. The 30 and 60oC/s HAZs exhibited a similar behavior during which the potentiostatic 85  currents showed a decay, and the intensity of the fluctuations was dramatically dampening over the time to be almost 20 and 10 µA/cm2, respectively. The shape of these fluctuations, in which the currents abruptly increased, and slowly decreased, can be ascribed to a repetitive mode of breakdowns and repassivations [Tang, 2004] of the passive films, under which the substrate dissolution was simultaneously inhibited. For all samples, the final potentiostatic currents were generally of the same order of magnitude of the potentiodynamic currents.  6.3    Electrochemical impedance spectroscopy testing The EIS tests were carried out to evaluate the time-dependent electrochemical influence of the growing passive films on the interfacial interactions, in reference to a HAZ microstructure, at the open circuit potentials. The charge-transfer resistance, and the resistance and capacitance of the passive films, fitted with equivalent circuits, were calculated by ZSimpWin software for EIS data measured after 3000, 6000, and 9000 s of immersion. In the first place, the rate of dissolution of ferrite, the surface area covered by the active adsorbent complexes [Lee, 2006], and, in turn, the steps (and rate-determining steps) of the passive film formation [Lu, 2006] are the main factors that can correlate its thickness, essentially, over a period of time, to a microstructure type [Schmitt, 1999, Al-Hassan, 1998]. In our case, the passive films appeared as an intact, dark precipitation only on 10oC/s HAZ, as shown in Figure 6-4, while they appeared as slightly transparent, brownish films on the other samples after 9000 s.  It is shown in Figures 6-5 and 6-6 the Nyquist spectra at the end of the immersion, and the Bode phase spectra after specific time intervals. Dissolution of ferrite and subsequent precipitation and growth of the passive films of the as-received and the 10oC/s HAZ could have proceeded initially, due to their relative microstructural resemblance, with similar manners. That 86  is, as explained by Dugstad [Dugstad, 1998], the vulnerability of ferrite for the two samples to selectively dissolve allow Fe3C to be an active, unattacked area from becoming, with time, a region onto which the adsorbent intermediates drive precipitation in an accelerating process, to a region onto which the cathodic reduction is slower and reaction-controlling.  While it is difficult, in our case of mildly alkaline bicarbonate solutions, to determine on the role(s) of the cathodic reduction of O2, HCO3- , and H2O (unlike with the case of the predominant reduction of H+ in lower-pH, CO2-saturated solutions), the homogeneity and thickness of the passivation growing into Fe3C seemed to dominantly affect the interactions, and Figure 6- 4 SEM micrograph of the corrosion products formed on 10oC/s HAZ after 9000 s of immersion. 87  in turn the charge-transfer resistance. For the as-received sample, although that the depth of Fe3C (between 35 to 55 µm) could not have been as large as that of 10oC/s HAZ, the thickness of film precipitation seemed not to change substantially over time. The passive film, as confirmed by fitting the EIS data with the {R(Q((QR)R))} equivalent circuit, formed a mixture with Fe3C at a thickness beyond which part of Fe3C remained a porous layer, as represented in Figure 6-7.   Figure 6- 5 Nyquist EIS spectra for the base steel API-X100, and the 10, 30, and 60oC/s HAZs after 9000 s of immersion. 88  To account for the interfacial heterogeneities [Juttner, 1990], a constant phase element (CPE) was used for the double layer and the passive film, whose admittance (Y) is expressed as [Boukamp,1989]:    2sin2cosnπ ω Yj  nπ ω YY nQnQ         (6-10) Figure 6- 6 Bode EIS spectra for the base steel API-X100, and the 10, 30, and 60oC/s HAZs after 3000, 6000, and 9000 s of immersion. 89  where ω is the angular frequency and n is the CPE exponent.            The solution resistance (Rs) was almost 11 Ω.cm2 and as shown in Table 6-1, the double layer, as indicated from nch values, was pseudo-capacitive. As shown in Figure 6-6, the phase values beyond the low-frequency peaks at nearly 0.02 Hz increased with time, associated, as demonstrated mathematically by Mansfeld [Mansfeld, 1981], and experimentally by Kinsella et al. [Kinsella, 1998], with the increase of charge-transfer resistance (Rch). The resistance of the porous Fe3C (Rp) did not change from an average of about 130 Ω.cm2.  Table 6-1 Equivalent circuit fitting for the EIS data of the base steel API-X100   Figure 6- 7 The equivalent circuit fitted with the experimental EIS data for the base steel API-X100. 90   For the 10oC/s HAZ, the interface transformed over time to make the interactions multi-time-constant based, as indicated from the changes with the size and number of the phase peaks in Figure 6-6. During immersion, the passive film grew to occupy Fe3C to constitute a second time constant in a {R(QR)(QR)} circuit, which is shown in Figure 6-8. This circuit was used by Alves et al. in more concentrated bicarbonate solutions to study the passivation of a number of steels at different potentials [Alves, 2002(B)]. The passive film showed evidence (see Figure 6-4) to grow further, forming a deposit layer, to account for a third-time constant as shown in Figure 6-8c, which was already used by Li et al., Hamadou et al., and Zhou et al., to account for the absorption and insertion phenomena which occur specifically within [Li, 2007, Hamadou, 2005, Zhou, 2011]. Since the morphology of, and especially cohesion between the inner and deposit layers were not to be investigated for this case, the inner layer is denoted in Figure 6-8c as a slightly extended interface under the deposit layer. Rch increased with time, as shown in Table 6-2, and the resistance of the passive films (Rf) were almost between 600 and 800 Ω.cm2. For the 30oC/s HAZ, the interface, as depicted from the final Nyquist and Bode spectra, was of no more than two-time-constant based. Table 6-2 Equivalent circuit fitting for the EIS data of 10oC/s HAZ      91                         Figure 6- 8 The equivalent circuits fitted with the experimental EIS data, collected after a) 3000, b) 6000, and c) 9000 s, for the 10oC/s HAZ. 92  Fitting the EIS data provided an evidence for unstable, or defective, passive films which could have underwent an intermittent mode in covering the substrate. The acicular-ferritic microstructure in our case, as explained earlier, and as validated by Zhang et al. [Zhang, 2009], therefore did not facilitate with the passive film growth to be thicker or more compact. The role of Fe3C is difficult to articulate on, long period in the middle of immersion time, to have been porous through which the solution was in contact with the substrate. Such an interface was proposed by Qin et al. [Qin, 2004] and Lee et al. [Lee, 2006], and the equivalent circuit {R(Q((QR)R))}, as shown in Figure 6-9b, was applicable to simulate it, and the circuit achieved a fairly precise fitting with the EIS data. It is important to indicate to the relatively high charge-transfer resistance of the 30oC/s HAZ, as shown in Table 6-3 – intrinsically in relation to its microstructure; highly confirming with the potentiodynamic results that showed the minimum anodic and cathodic currents.  Table 6-3 Equivalent circuit fitting for the EIS data of 30oC/s HAZ    The passive film of 60oC/s HAZ seemed to be more stable, and its interface, as suggested from the Bode spectra, was two-time-constant based, fitted with the equivalent circuit {R(QR)(QR)} shown in Figure 6-10. The phase values shifted to lower frequencies with time, 93  associated to a slight continual increase in the charge-transfer resistance, as shown in Table 6-4, by a passive resistance of 160 Ω.cm2.                           Figure 6- 9 The equivalent circuits fitted with the experimental EIS data, collected after a) 3000, b) 6000, and c) 9000 s, for 30oC/s HAZ. 94            Table 6-4 Equivalent circuit fitting for the EIS data of 60oC/s HAZ    6.4    Summary In this chapter the electrochemical study investigated the corrosion behavior for a number of HAZs, made of API-X100 pipeline steel, in naturally aerated, 0.05 M bicarbonate solutions at 90oC. It addressed the corrosion susceptibility and passivation stability of near-fusion HAZs, of Figure 6- 10 The equivalent circuit fitted with the experimental EIS data for 60oC/s HAZ. 95  welded pipeline segments, whose microstructures depend on the cooling rates from 950oC. The HAZs were simulated by Gleeble© thermal cycles by which the samples were cooled with controlled rates of 10, 30, and 60oC/s from 950oC. The slow potentiodynamic scans revealed the important roles of ferrite, in the 10oC/s HAZ, and acicular ferrite and martensite in the 30 and 60oC/s HAZs, respectively, with the passivation process and the cathodic currents. The 10oC/s HAZ and as-received samples exhibited the lowest passive current densities, while those of 30 and 60oC/s HAZs showed an evidence of thin, unstable passive films. Oxygen reduction currents intersected with the passivation regimes, and appeared as broad cathodic peaks. The mechanisms of anodic dissolution and the active-passive transition, which were independent from the HAZ microstructures, were outlined in the beginning. The 30oC/s HAZ, notably, exhibited the minimum cathodic current densities. The potentiostatic currents of the 30 and 60oC/s HAZs at 0.5 V vs. SCE suggested a slow growth of the passive films which underwent repetitive breakdowns and repassivations. By measuring the impedance over 3000, 6000, and 9000 s of immersion at the open circuit potentials, an electrochemical evidence of the growth of the passive films was found. Taking into account the roles of the porous Fe3C, and the changing thickness of the passive films, a set of equivalent circuits and interfaces were proposed to fit the multi-time-constant EIS data, in reference to a HAZ microstructure. The charge-transfer resistance decreased with time as a result of the growing passive films, but the 30oC/s HAZ showed an intrinsic high charge-transfer resistance, confirming with the potentiodynamic polarization results.      96  CHAPTER 7: VOLTAMMETRIC ANALYSIS OF THE SIGNIFICANCE OF BICARBONATE CONCENTRATION WITH CORROSION REACTIONS AND PASSIVATION OF API-X100 STEEL HAZs   The cyclic voltammetry is used in this chapter to simulate the corrosion reactions, starting from the cathodic reduction, to anodic dissolution, and passivation of two types of simulated near-fusion API-X100 HAZs. The significance of bicarbonate concentration with these processes as well as with the stability of the passive films is studied. The HAZs, simulated with thermal cycles outlined section 4.3 and represented in Figure 4-7, were produced by heating at 100oC/s to the peak temperature 1200oC, held for 15 s and cooled at the controlled rates 10 and 60oC/s. Their microstructures are presented in Figure 4-8 and explained in section 4.3. The tests, outlined in section 4.5, were carried out in naturally-aerated solutions, outlined in section 4.4, containing 0.05, 0.2, and 0.6 M bicarbonate at 25oC.        7.1 The corrosion reactions and voltammetric analysis The cyclic voltammogram tests were carried out to study the electrochemical and chemical reactions of the cathodic reduction of water, oxygen and bicarbonate, the early anodic dissolution steps, the pre-passive filming process, the transformations of the passive films as they dissolve, react, and thicken, transpassivation, oxygen evolution, cathodic reduction of the corrosion products, and these same reactions, sustained onto and through the corrosion products with 10 cycles of voltammetric scanning. Different scan rates of 0.5, 5, and 25 mV/s, as shown in Figure 7-1, were applied. The scan rate considered for analyzing the corrosion behavior is 0.5 mV/s, by which the charge transfer-controlled and diffusion-controlled anodic and cathodic reactions, the corrosion products growth, and transport across the passive films are slow. This was to emulate the slow nature of corrosion, albeit its overall rate represents intertwined 97  reactions of different rates, which could at the HAZ accelerate more under the galvanic coupling than by the environmental and hydrodynamic effects.   Figure 7- 1 Cyclic voltammograms of scan rates of 0.5, 5, and 25 mV/s, for 10oC/s HAZ in 0.05 M bicarbonate solutions.  Another cyclic parameter considered was the range of the cathodic reduction, to -1.3 and to -0.9 V vs. SCE, under which the cathodic reduction differently affects the stability of the corrosion products and, in turn, the activity of the interfacial forward-scan processes with higher 98  number of cycles. The electrochemical analysis considers the shapes and sizes of the shoulders and inflections, as well as the intensity, location, and number of peaks, across which certain reactions take place during the forward and backward scans. Bicarbonate was introduced in concentrations of 0.05, 0.2, and 0.6 M, by which it differently affects the aqueous equilibrium and, in turn, affects the active roles of hydroxyl, water, and oxygen, and during which bicarbonate becomes of different significances with the reactions of forward and backward scans. Although the electrochemical tests here are different from those of the open-circuit potentials monitoring [Eliyan, 2012, Torres-Islas, 2009], the reactions of cyclic voltammetry similarly depend on the natures of the interface. They, i.e., depend on the microstructure and surface being initially bare and later corroded and covered with corrosion products whose properties change with increased cycling. During that, the areas of the anodic versus cathodic regions onto which the interfacial concentrations of the adsorbing active species – transported from the bulk solution or generated from the reactions – affect the tendencies for specific reactions to take place at certain scanned potentials [Stikma, 1985, Staicopolus, 1963, Al-Hassan, 1998, Remita, 2008]. The amounts of the active constituents driving anodic dissolution, which is ferrite in different phases, and cementite (Fe3C) which is conductive for the cathodic reduction and anchoring of the corrosion products [Berntsen, 2013, Videm, 1996, Palacios, 1991], are considerably different in 10 and 60oC/s HAZs. And this is a cause of possibly significant differences in the mechanisms, steps, and rate-determining steps of the anodic and cathodic reactions – onto the polarized surfaces across a wide range of potentials – to ultimately result in corrosion products having different thicknesses, morphologies, and properties.     99  7.2 Significance of voltammetry scan rates In Figure 7-1, the cyclic voltammograms of 10oC/s HAZ in 0.05 M solutions are shown for scan rates of 0.5, 5, and 25 mV/s, across long ranges of potentials from -1.3 to 1.2 V vs. SCE. Regardless of the scan rate, the surfaces were extensively pitted and covered fully with rust in these dilute solutions. During 0.5 mV/s voltammetry, up to nearly -0.75 V vs. SCE, the anodic currents increased sharply within a short range of potentials, right from the first cycle, without exhibiting anodic plateaus within which unique peaks could appear in correspondence to transitions from dissolution to pre-passivation [Rangel, 1986, El-Naggar, 2004]. Moreover, the cycles of these voltammograms did not show coincidence at any region, given though the slow scan supposedly seemed to allow bicarbonate, even in low concentrations [Videm, 1993, Lu, 2006] , to stabilize for a protective carbonate-based passive film – so that it becomes a source of reproducibility (or voltammetric coincidence). As previously explained in potentiodynamic polarization studies [Eliyan, 2012, Fu, 2010], and later in this chapter, higher bicarbonate concentrations enhance the formation of protective passive films. The first peaks, at nearly -0.17 V vs. SCE, during the first three cycles, increased in intensity from 500 to 650 μA/cm2 and the passive currents beyond also increased to the end of the forward scans. The first peaks were followed by broader, by as much as 400 mV, second peaks of characteristics that did not significantly change with increased cycling. The intensity of the first peak, starting from the third cycle, decreased from 650, 450, 350, and 250 to 200 μA/cm2 and it became increasingly blunt. Rarely a third peak, corresponding to a transpassivation process possibly interfering with oxygen evolution appeared, at nearly 1 V vs. SCE. The cathodic currents during the backward cycles did not change with increased cycling, except at potentials lower than -0.9 V vs. SCE, at which the reduction of H2O, oxygen, and bicarbonate became more significant with higher currents that 100  interfere with those of the corrosion products reactivity [Von Fraunhofer, 1970]. The cathodic currents at low potentials decreased with the increased cycling and the size of the last cycles became increasingly bigger. This indicates that the extensive dissolution was the dominant phenomenon by which the extents of the anodic and cathodic reactions were controlled by the increasing proportion between the surface areas of the attacked (and later covered) regions versus the unattacked regions after each cycle.  At scan rates of 5 and 25 mV/s, the forward currents particularly at potentials higher than -0.7 V vs. SCE were nearly two times higher and the peaks appeared noticeably smoother than those of 0.5 mV/s scans. These were a result of the greater share of the charging current density, which is proportional with the scan rate and the interfacial capacitance [Bard, 2001], with the measured currents. For 5 mV/s voltammetry, the size of the positive loop gradually decreased until it disappeared at the sixth cycle, and the surface completely corroded and got covered with rust. With increased cycling, the first peak decreased in intensity from 1150, 1000, and 900 to 550 μA/cm2, appearing at decreasing potentials from -50 to -200 mV vs. SCE, indicating the thickening of the corrosion products. The 25 mV/s voltammetry, most commonly used in bicarbonate/carbonate solutions [Castro, 1991, Rangel, 1989, Castro, 1991], was similar in shape to that of 5 mV/s, its peaks decreased in intensity and appeared at lower potentials from 200 to 0 mV vs. SCE, and the currents were higher and having smoother profiles. For the 5 and 25 mV/s voltammetries, the first shoulder in the backward scan from -0.4 to -0.9 V vs. SCE corresponds to the reduction of oxygen (Eq. 7-1), and the following shoulder from -0.9 to -1 V vs. SCE indicates at the anodic activity of the corrosion products and partially represents the reduction of bicarbonate (Eq. 7-2), while that from -1 to -1.3 V vs. SCE, corresponds to the reduction of water (Eq. 7-3) which obliterates the rest of the cathodic currents.  101  O2 + 2H2O + 4e- → 4OH-             Eo = 0.159 V vs. SCE                 (7-1) HCO3- + e- → Hads+ CO32-             Eo = -0.851 V vs. SCE                (7-2) 2H2O + 2e- → H2 + 2OH-            Eo = -1.07 V vs. SCE                  (7-3) 7.3 Significance of bicarbonate concentration The cyclic voltammograms of 10oC/s HAZ in 0.2 and 0.6 M solutions are shown in Figure 7-2, of 0.5 mV/s scan across a potential range from -0.9 to 1.2 V vs. SCE. Bicarbonate in higher concentrations played protective roles and increased the rates of the cathodic reduction in these solutions. In the 0.2 M solution, as shown in Figure 7-3, the surface was semi-protected with a transparent, brownish film with tiny rust clusters – unlike the surface in the 0.05 M solution.   Figure 7- 2 Cyclic voltammograms, scanned at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, for 10oC/s HAZ in 0.2 and 0.6 M bicarbonate solutions.   102                      0.05 M  0.2 M  0.6 M  Figure 7- 3 SEM Surface micrographs of 10oC/s HAZ, after scanning at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, in 0.05, 0.2 and 0.6 M bicarbonate solutions. 103  The anodic regime began from -0.725 V vs. SCE, the same potential with increased cycling, from which the dissolution currents showed high coincidence up to -0.6 V vs. SCE, at which the first anodic peak appeared. Prior to that peak, with similar potentials previously reported with different scan rates [Burstein, 1980, Lu, 2006, Davies, 1980], the consecutive charge-transfer adsorption and dissolution steps take place to form a hydroxide-based layer as [Castro, 1986, Valentini, 1985]: Fe + H2O ↔ Fe(OH) + H+ + e-                                                       (7-4a) Fe(OH) ↔ Fe(OH)++ e-                                                                  (7-4b) Fe(OH)+ + OH- ↔ Fe(OH)2 → Hydrous Fe(OH)2                         (7-4c) The temperature-independent (4b) step was reported as rate-determining in aerated bicarbonate solutions [Lu, 2006], during which Fe(OH) is a main precursor of dissolution, which is overall expressed as:  Fe + 2OH- → Fe(OH)2 + 2e-                (7-5) Upon increasing the forward potential, Fe(OH)2 becomes a reactive layer with HCO3-, by which it partially dissolves (Eq. 7-6) [Castro, 1986], and simultaneously becomes an inner layer [El-Naggar, 2004], between the substrate (Fe) and FeCO3 (which forms as an outer layer) (Eq. 7-7a, b, and c) [Zhang, 2009] , with special kinetics that increases the anodic passive current densities: Fe(OH)2+ HCO3- → CO32- + OH-+ Fe2++ H2O   (7-6) Fe + HCO3- ↔ FeHCO3-                                      (7-7a) FeHCO3- → FeHCO3+ e-                                    (7-7b) FeHCO3 + OH- → FeCO3 + H2O + e-               (7-7c) 104  When the kinetics of FeCO3 precipitation surpasses that of dissolution, an appreciably complete surface coverage abruptly decreases the currents to result in the first -0.6 V vs. SCE peaks. In this case, the currents after the peaks decreased under the influence of oxygen reduction [Von Fraunhofer, 1970], temporarily to increase again and form broad second peaks extending from -0.35 to 0.15 V vs. SCE. They correspond to a multiple of reactions, involving oxygen (Eq. 7-8 [Fu, 2010(A)(B)]) and hydroxyl (Eq. 7-9 [Lu, 2006]), during which FeCO3 transforms to iron oxides at higher potentials: 4FeCO3 + O2 + 4H2O → 2Fe2O3 + 4HCO3- + 4H+             (7-8)  2FeCO3+ 4OH- → Fe2O3 + 2HCO3- + H2O + 2e-                (7-9) The passive currents remained relatively stable afterwards, between 20 and 40 μA/cm2 over increased cycling. At nearly 0.95 V vs. SCE, the currents increased abruptly as a result of oxygen evolution and transpassivation, until 1 V vs. SCE at which the slope of the currents decreased as a result of some interfering transpassivation processes, of different kinetics, which results in FeO42- [Beck, 1985]. Upon reversing the scan, the currents were 2 to 4 times less than those of the forward scan, during which the corrosion products partially get reduced from F3+ to F2+ states. The currents became negative, almost at the same potential of 0 V vs. SCE, regardless of increased cycling, at which the reduction of oxygen started. With increased cycling the intensity of the first peaks increased from 35 to 45 μA/cm2, indicating that a portion of the active (unattacked or film-free) substrate remained contributing to the interdependent dissolution and FeCO3 precipitation processes (Eq. 7-7a, b, and c) at each cycle. The intensity of the broad second peaks decreased accordingly. This relates to the continuous increase in the protectiveness of the passive films, of FeCO3 that continued to oxidize to iron oxides (Eqs. 7-8 and 7-9). With 105  increased cycling, the passive profiles became smoother during which a mixture of products remained transforming among Fe(OH)2, Fe3O4, and Fe2O3 as [El-Naggar, 2004]: Fe + 2Fe(OH)2+ 4OH- ↔ Fe3O4 + 4H2O + 4e-      (7-10)  2Fe3O4 + H2O ↔ 3Fe2O3 + 2H+ + 2e-                    (7-11)                   2Fe2++ 3H2O ↔ Fe2O3 + 6H+ + 2e-                        (7-12) The significance of these reactions, in comparison with FeCO3/Fe3O4/Fe2O3 reactions, with the morphology and compactness of the corrosion products needs future investigations. From the sixth cycle, an inner peak (a unique decrease in the cathodic currents) appeared in the cathodic backward scan at -0.65 V vs. SCE. It corresponds to an anodic activity (which coincides with the location of the first anodic peaks in the forward scans) of the corrosion products.  In 0.6 M solution, the surface appeared protected as it was covered with a brownish transparent film, not rust or solid corrosion products, as shown in Figure 7-3. The voltammetry profiles, as shown in Figure 7-2, were smoother and both the anodic and cathodic currents were notably higher than those in 0.2 M solution. The anodic regime began from -0.725 V vs. SCE up to anodic peaks appearing at a higher -0.55 V vs. SCE potential, with comparable critical currents of nearly 85 μA/cm2. This indicates that bicarbonate, in increased concentrations, catalyzed the anodic process during which the hydroxide film dissolved more (Eq. 7-6) with stronger kinetics, only initially, before the dissolution was later suppressed by the effective formation of bicarbonate-based passive films. Regarding the onset of the higher cathodic currents during the backward scans from a higher 0.2 V vs. SCE potential; given that bicarbonate increases the local alkalinity [Parkins, 1997] to make the cathodic potentials of the oxygen reduction lower (not higher in this case), the high cathodic currents are likely attributed to the better quality (in comparison to those in 0.05 and 0.2 M solutions) of the surface onto which the 106  oxygen reduction and hydrogen generation took place. Also, those special cathodic inner peaks, which appeared at the last cycles in 0.2 M solution at -0.65 V vs. SCE, did not appear in 0.6 M solution as the much cleaner substrate – onto which no significant corrosion products formed to dissolve at low potentials – sustained the cathodic reduction of the bulk species at high rates.            7.4 Significance of extended cathodic reduction range  Extending the cathodic potential ranges by 400 mV, to -1.3 V vs. SCE, resulted in noticeable changes with the cathodic reduction and passivation currents, as shown in Figure 7-4. The voltammograms exhibited a high degree of coincidence in a situation the high concentrations of bicarbonate protected the surfaces with passive films which seemed, though, not effective enough to resist the extended cathodic reduction at the end of each cycle. As a result, the kinetics of dissolution and passivation during the forward scans with increased cycling had semi-reproducible currents, of reactions sustained onto surfaces not significantly changing accordingly. In 0.2 M solution, the first peak, at -0.6 V vs. SCE, appeared as a small shoulder and the second peak appeared sharper with intensity decreasing from 160 to 100 μA/cm2 with increased cycling.  Figure 7- 4 Cyclic voltammograms, scanned at 0.5 mV/s from -1.3 to 1.2 V vs. SCE, for the 10oC/s HAZ in 0.2 and 0.6 M bicarbonate solutions. 107  The cathodic currents were accordingly decreasing from -900 to -500 μA/cm2 at the beginning of each cycle as a result of the accumulating layers of rust and passivation that affected the kinetics of reduction to be slower. In 0.6 M solution, the passivation, both during the forward and backward scans, seemed more enhanced and stable, as shown in Figure 7-4. The cathodic currents, catalyzed by the reduction of bicarbonate, were remarkably as higher as -2000 μA/cm2. Due to the large scale of the current densities in the y-axis, the anodic peaks do not appear clearly, but the transpassivation, starting from 1 V vs. SCE, was a main feature of these semi-reproducible cycles.          7.5 Significance of fast-cooling the HAZs  The microstructure of the 60oC/s HAZ consists of different phases that mainly are ferrite, mostly in acicular form, and retained austenite. Their active roles collectively seemed to result in comparable anodic behavior, as shown in Figure 7-5, during dissolution and passivation to that of the 10oC/s HAZ (which is considerably more ferritic). However, the martensite phase, along with acicular ferrite, resulted in low cathodic activity, both in passive films/corrosion products reduction and hydrogen generation, and as a result the interface became more under cathodic control. The low cathodic activity of a similar microstructure, cooled at 30oC/s from a peak 950oC temperature, was previously reported in low- and high-temperature aerated bicarbonate solutions, bicarbonate-carbonate solutions, and in harsh low-pH solutions. In 0.05 M solution, the substrate was heavily corroded, as shown in Figure 7-6, as bicarbonate acted aggressively with the increased polarization before it could effectively be involved in a competitive, protective process (Eq. 7-7a, b, and c) – before the first peaks at -0.45 V vs. SCE. Their intensity increased, by 10 μA at each cycle, from 140 to 270 μA/cm2. The first peaks were instantly followed by second peaks, with a decreasing intensity from 460 and 430 to 200 μA/cm2, for 108  reasons explained in section 3.1.2.; pertaining to the electroactivity and amounts of FeCO3 after each cycle. The passive currents, up to transpassivation, decreased in a fashion that indicates a continuous layering of the corrosion products that increasingly suppressed the underlying  Figure 7- 5 Cyclic voltammograms, scanned at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, for 60oC/s HAZ in 0.05, 0.2, and 0.6 M bicarbonate solutions.   109                                 0.05 M  0.6 M  0.2 M  Figure 7- 6 SEM Surface micrographs of 60oC/s HAZ, after scanning at 0.5 mV/s from -0.9 to 1.2 V vs. SCE, in 0.05, 0.2 and 0.6 M bicarbonate solutions. 110  substrate dissolution, and the difference between the forward and backward passive currents was small. Inner cathodic peaks, similar to those of the 10oC/s HAZ in 0.2 M solution, appeared after the third cycle. In 0.2 M solution the currents, as shown in Figure 7-5, were much less and the first peaks, which were similar with intensities decreasing with increased cycling, appeared at lower potentials, which indicate the early onset of the formation of effective passive films – which contained numerous but small rust clusters, as shown in figure 7-6. In 0.2 M solution the currents, as shown in Figure 7-5, were much less and the first peaks, which were similar with intensities decreasing with increased cycling, appeared at lower potentials, which indicate the early onset of the formation of effective passive films – which contained numerous but small rust clusters, as shown in Figure 7-6. They were followed by valleys, corresponding to oxygen reduction, and by second peaks as a result of a mix of oxidation reactions (Eqs. 7-8 to 7-12) the passive films/corrosion products undergo in bicarbonate solutions. The passivation features across this range could also correspond to the diversity of the active phases of the 60oC/s HAZ which could exhibit their anodic characteristics at different potentials. The notably low cathodic activity during the backward scan contributed to protecting the accumulating corrosion products to retain a strong anodic activity, showing inner, similar large cathodic peaks at -0.6 V vs. SCE. In 0.6 M solution, the surface was highly protected, as shown in Figure 7-6, on which a transparent film formed, containing a few rust clusters at their early stages. Bicarbonate was protective and the shapes of the voltammograms changed, as shown in Figure 7-5, in which the anodic currents were less by almost two times than those in 0.2 M solution. The second peak diminished from the third cycle and the passive currents were coincidental and reproducible, being nearly 1.25 μA/cm2. The transpassivation currents were remarkably high, but attributing 0.6 M  111  this to the increased bicarbonate concentration requires separate mechanistic investigations. But given that the local alkalinity increases, the faster generation of ferrate (Eq. 7-13 [El-Naggar, 2004]) at potentials higher than 1 V vs. SCE might partially explain the high transpassivation currents.  Fe + 4OH- → FeO42- + 4H++ 6e-                          (7-13) Rangel et al. argued that the charge release associated with the generation of “superficial” ferrate ions, regardless of  the mechanisms, is relatively small to correspond to specific peaks; rather, Rangel et al. attributed them to the decomposition of ferrate [Rangel, 1989] – which is oxygen-releasing as follows [Beck, 1985]:  4FeO42- + 2H2O → 4FeO2- + 3O2+ 4OH-             (7-14) As shown in Figure 7-7, extending the cathodic potentials to -1.3 V vs. SCE did not markedly change the major anodic characteristics in 0.05 M solution of 60oC/s HAZ, which was heavily corroded. It seemed, however, it affected the kinetics of dissolution and passivation at the early potentials of the forward scan at which the first peaks hardly appeared in this solution, as well as in 0.2 and 0.6 M solutions of profiles appeared similar to those of 10oC/s HAZ.       7.6 Summary The research in this chapter investigated the significance of bicarbonate concentration with the corrosion reactions, simulated by cyclic voltammetry, with two types of API-X100 near-fusion HAZs, cooled down at 10 and 60oC/s from 1200oC. The following, in brief, is found:  1. Bicarbonate in increased concentrations catalyzes anodic dissolution but increasingly involves in counteractive kinetics that drives the precipitation of FeCO3. As a result, passivation forms and the surface becomes protected, or heavily attacked, in concentrated and dilute solutions, respectively. This applied for both the 10 and 60oC/s HAZs. 112      Figure 7- 7 Cyclic voltammograms scanned at 0.5 mV/s from -1.3 to 1.2 V vs. SCE for the 60oC/s HAZ in 0.05, 0.2, and 0.6 M bicarbonate solutions. 113  2. With increased voltammetric cycling, indications appeared about the gradual involvement of active (unattacked) surface with anodic dissolution (and pre-filming), and also about diminishing amounts of FeCO3 (to transform to oxides) after each cycle in the concentrated solutions. Also, an indication appeared about the continuous thickening of the corrosion products after each cycle both in dilute and concentrated solutions. 3. Bicarbonate increases the rates of the cathodic reduction, directly in increased concentration, and also indirectly as the cathodic reduction of all species takes place onto cleaner (not rust-covered), protected surfaces.  4. It is recommended to cool down the welded pipeline joints quickly when heated to very high temperatures (~1200oC). The 60oC/s HAZ exhibited low anodic activity, better passivation, and low cathodic activity, in comparison to the slower-cooled 10oC/s HAZ. In a field situation, the 60oC/s HAZ is highly likely to sustain slow cathodic reactions, under which with the base steel the galvanic effects are mitigated and the corrosion products are stable and protective.                114  CHAPTER 8: INTERRELATION BETWEEN BICARBONATE AND CO2 IN AFFECTING CORROSION REACTIONS AND PASSIVATION OF API-X100 STEEL HAZs   In this chapter the significance of bicarbonate concentration in CO2-saturated solutions is investigated with the corrosion behavior of the two types of API-X100 HAZs examined in Chapter 7. The interrelation between bicarbonate and H2CO3 with the filming process during free immersion and with the corrosion reactions was investigated by open-circuit potentials and cyclic voltammetry, with parameters outlined in section 4.5. The experiments were carried out in CO2-saturated solutions containing 0.02, 0.1, 0.5, and 1 M bicarbonate at 25oC, as explained in section 4.4. 8.1 Open-circuit potentials monitoring    The open circuit potentials were monitored to analyze the significance of bicarbonate concentration (versus that of CO2 hydration which dissociates more to H2CO3 in more dilute bicarbonate solutions) with the passivation process during free immersion. As shown in Figure 8-1, for the 10oC/s HAZ, the open circuit potentials were lower with the bicarbonate concentration, which were respectively -0.705, -0.715, -0.74, and -0.745 V vs. SCE. The OCP profiles exhibited trends that reflect the effectiveness and rate of formation of the passive films, which initially depended on the fashions the dissolution and cathodic reduction occurred with. To elucidate, both bicarbonate and H2CO3 drive the anodic and cathodic reactions [Simard, 1997, Davies, 1980, Lu, 2006, Linter, 1999]. The extents, however, for each of them to drive either more the anodic or the cathodic reactions are difficult to estimate. Reported in earlier studies, bicarbonate in small concentrations in aerated and de-oxygenated solutions (by N2 and Ar, not by CO2) exacerbates dissolution and dissolves the initial monolayer of passivation [Burstein, 1981]. In higher concentrations, it is a protective species which promotes the formation of effective 115  passive films and decreases the interfacial significance of H2CO3. Therefore, in the concentrated 0.5 and 1 M solutions the OCPs decreased quickly within the first 2000 s period as a result of the vigorous, but only initial, exacerbation of dissolution by bicarbonate, followed by an interfacial saturation with adsorbents that resulted from the dissolution and drove the formation of thick passive films. In 0.02 and 0.1 M solutions, the rates of the cathodic reduction were higher and the passive films – probably as a result – were thinner and porous, but they continued forming and suppressing dissolution at later stages, as the OCPs by then showed a gradual increase.   Figure 8- 1 Open circuit potentials monitored for 12000 s for 10oC/s HAZ in 0.02, 0.1, 0.5, and 1 M bicarbonate solutions.    116  8.2  Long-range voltammetry testing The long-range cyclic voltammetry was carried out to investigate the significance of bicarbonate concentration with the corrosion reactions that begin with the cathodic reduction, followed by dissolution and passivation. It was scanned at 1 mV/s for 10 cycles. The corrosion reactions sustaining onto the surface and later through and onto the passive films, as well as the growth, reactivity and stability of the passive films were analyzed. In Figure 8-2, it is shown the cyclic voltammograms, of a range between -1.1 to 1.1 V vs. SCE, of 10oC/s HAZ in 0.1, 0.5 and 1 M bicarbonate solutions. In 0.1 M solution, the currents across most of the voltammogram were reproducible with increased cycling. This was especially the case when the cathodic currents decreased across the region in which dissolution began, which was followed by the formation of a passive film that covered the surface and resulted in the sole, first anodic peak at -0.345 V vs. SCE, as well as in a slight inflection from -0.55 V vs. SCE to -0.1 V vs. SCE. Beyond -0.1 V vs. SCE, it was the forward-scan passive currents which reflected the changes in passivation with increased cycling.         In the beginning, water molecules dominate the ionic influx towards the polarized bare surface, getting reduced under charge-transfer control [Eliyan, 2012(A)] between -1.1 and -1 V vs. SCE with currents that decreased from -100 to -50 μA/cm2. During this stage, hydrogen is generated and local alkalinity increases. As the potential rises from -1 to -0.75 V vs. SCE, during which the surface becomes totally under the anodic control, the migrating hydroxyl ions (in addition to the already resulted from water reduction) as well as water molecules (which have anodic significance [Lu, 2006]), drive dissolution in anodic steps that result in dispersed Fe(OH)2 clusters. H2CO3 and bicarbonate have the thermodynamic capacity across this stage in driving  117   Figure 8- 2 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to 1.1 V vs. SCE for 10 cycles, for 10oC/s HAZ in a) 0.1, b) 0.5, and c) 1 M bicarbonate, CO2-saturated solutions. Potential (V vs. SCE)-1.0 -0.5 0.0 0.5 1.0Current (A/cm2)-50005001000Potential (V vs. SCE)-1.0 -0.5 0.0 0.5 1.0Current (A/cm2)-500050010001500Potential (V vs. SCE)-1.0 -0.5 0.0 0.5 1.0Current (A/cm2)02004006008001000aForward scansBackward scansbc118  the cathodic reduction, but increasingly so the anodic reactions as the applied potentials rise. They drive dissolution in mechanisms different from those of hydroxyl and water. They result in FeCO3 which precipitates electrochemically or under supersaturation – in association to local alkalinity and carbon-carrying species concentration [Thomas, 1970, Nagayama, 1963, Moiseeva, 2005].  The dissolution continues and increases in a heterogeneous process, driven both by H2CO3 and bicarbonate, with current densities reaching 800 μA/cm2.  Bicarbonate has a higher concentration in driving and accelerating dissolution, which as a result enhances the precipitation of FeCO3, disrupts the formation of Fe(OH)2, and probably significantly dissolves it [Niu, 2007, Sherar, 2010]. By the first anodic peak –with critical currents of 900 μA/cm2 that did not change much with increased cycling – the passive film covers most of the surface with appreciable thickness. Beyond that, the currents abruptly decreased within 200 mV to 250 μA/cm2, at -0.55 V vs. SCE, showing a small inflection across which the currents decreased less quickly until they stabilized at -0.1 V vs. SCE. During this stage, the passive film grows as it starts to contain iron oxides, such as Fe2O3 and Fe3O4. They could have formed directly from the substrate or by the reactivity of the FeCO3-based passive film during which it changes in color having a new layer of iron oxides. This results in further suppression of the underlying dissolution, causing the passive currents to decrease and stabilize across a wide region prior to the end of voltammetry. The passive currents were between 20 and 60 μA/cm2 and tended to be higher with increased cycling. In regard to transpassivation, the voltammetry range was not wide to show the transpassivation currents except with some slight, abrupt increases in the currents at the end of voltammetry in the first cycles. The backward passive currents were insensitive to the increased cycling, being between 1 and 5 μA/cm2 until 0.05 V vs. SCE, at which the currents became cathodic only for 300 mV before a large, cathodic “inner” peak, of positive currents, 119  appeared coinciding with the anodic peaks of the forward scan. The maximum currents of these peaks with increased cycling did not show a clear trend to reflect on any relationship between the increased reactivity of the passive film (which was getting thicker) and the activity of the substrate which accordingly resulted in higher current densities.  Increasing bicarbonate concentration to 0.5 and 1 M resulted in noticeable voltammetric changes, as represented in Figures 8-2b and 9-2c. These changes resulted from thicker passive films. In 0.5 M solution, the course of the cathodic reduction was not much different from that in 0.1 M solution, but the cathodic currents were high in the first cycles, indicating the reduction being catalyzed by bicarbonate, but later the currents showed a continuous decrease indicating the sluggishness of the cathodic reduction taking place onto increasingly thicker passive films.  This explains the lower cathodic currents in 0.5 M solution than those in 0.1 M solution during the last four cycles. The maximum current densities of the first anodic peaks with increased cycling took an interesting trend. They increased to 1.5 mA/cm2 at the first cycle beyond which they decreased to 400 μA/cm2 at the tenth cycle. This refers to the increasing proportion between the growing passive film versus the exposed surface area which was being covered – but dissolving increasingly through the porous passive film with increased cycling. Both the forward and backward passive currents were comparable with those in 0.1 M solution. The inner cathodic peaks were noticeably smaller than those in 0.1 M solution, and getting so with increased cycling, confirming that these peaks, especially given their bigger size in more dilute solutions, are correlative to the reactivity of the surface to passivation, not to the reactivity of the passivation to oxidation.  In 1 M solution, the cathodic currents were high only during the first two cycles before they decreased as a result of a thick passivation that formed quicker. It did not only impede the 120  cathodic reduction, but also inhibited the dissolution earlier than that in 0.1 and 0.5 M solutions. The first anodic peaks were getting smaller from the third cycle, and so were the cathodic peaks. The passive currents during the forward and backward scans were nearly two times lower than those in 0.1 and 0.5 M solutions. The cathodic regime of the backward scans started at higher potentials, likely indicating the higher tendency of the thicker passive films to get reduced before the onset of the reduction of the migrating species.          8.3 Short-range voltammetry testing The short-ranged cyclic voltammetry was carried out to evaluate the significance of increased cycling during the cathodic reduction and anodic dissolution. It was scanned at 1 mV/s from -1.1 to -0.6 V vs. SCE – a range covering the cathodic reduction and dissolution up to before the first anodic peak appears. The voltammograms in 0.02, 0.1 and 1 M bicarbonate solutions are shown in Figure 8-3. Increasing bicarbonate concentration increased the cathodic reduction but later it decreased, more noticeably in the concentrated solutions, with increased cycling. The anodic regime, however, during both forward and backward scans did not exhibit sensitivity with increased cycling in all solutions. It is worth noting that the dissolution currents in 0.1 M solution were higher than those in 0.02 M solution, but they were also higher than those in 1 M solution. This shows that bicarbonate only at an optimum concentration exacerbates dissolution beyond which it facilitates (at higher concentrations) the formation of effective passive films at lower anodic potentials.    8.4 Significance of fast cooling the HAZs Long cyclic voltammetry was carried out to study the microstructural significance of 60oC/s HAZ, which consists of a mixture of martensite, acicular ferrite, and martensite-retained  121   Figure 8- 3 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to -0.6 V vs. SCE for 10 cycles, for 10oC/s HAZ in a) 0.02, b) 0.1 and c) 1 M bicarbonate, CO2-saturated solutions. Potential (V vs. SCE)-1.1 -1.0 -0.9 -0.8 -0.7 -0.6Current (A/cm2)-500-400-300-200-1000100Potential (V vs. SCE)-1.1 -1.0 -0.9 -0.8 -0.7 -0.6Current (A/cm2)-300-200-1000100Potential (V vs. SCE)-1.1 -1.0 -0.9 -0.8 -0.7 -0.6Current (A/cm2)-100-50050100abc122  austenite, in 0.1 and 1 M solutions, of voltammograms presented in Figure 8-4. In 0.1 M solution, both the anodic and cathodic currents were comparable to those of the 10oC/s HAZ.                  The passivation regime, however, showed a number of differences. The first anodic peaks, of comparable maximum currents with increased cycling, appeared at 200 mV higher potentials and with higher critical currents than those of the 10oC/s HAZ. Moreover, the currents Figure 8- 4 Cyclic voltammetry, scanned at 1 mV/s from -1.1 to 1.1 V vs. SCE for 10 cycles for the 60oC/s HAZ in a) 0.1, b) 1 M bicarbonate, CO2-saturated solutions. 123  beyond the first anodic peaks decreased more abruptly, without showing the inflections that appeared at -0.55 V vs. SCE during the 10oC/s HAZ voltammograms. Interestingly, the passive currents afterwards to the end of the voltammetry were nearly three times lower. The backward cathodic peaks were not as coincident with the first anodic peaks as they appeared at lower potentials. All these features suggest that the reactivity of the 60oC/s HAZ during dissolution was not high and therefore did not develop effective passivation [Videm, 1993]. A similar behavior was reported in aerated and de-oxygenated bicarbonate and carbonate solutions for HAZs of similar microstructures. During passivation, the iron oxides might have formed from the direct oxidation of the substrate, rather than of the FeCO3-based film for microstructural reasons. The low passive currents of 60oC/s HAZ are likely more indicative of the low reactivity of the underlying substrate than of the effectiveness of the passive film in suppressing it.    Increasing the bicarbonate concentration to 1 M made the bicarbonate have an environmental effect of significance larger than that of 60oC/s HAZ microstructure. As a result, the cyclic voltammograms of the 10 and 60oC/s HAZs in 1 M solutions showed similarity. Compared to that in 0.1 M solution, the passive film seemed thicker and more effective as the anodic peaks appeared at lower potentials and had lower critical currents, as generally are the currents across the forward scan with increased cycling.  8.5 Summary In this chapter, the open-circuit potentials monitoring and cyclic voltammetry were used to simulate and analyze the corrosion behavior of two types of API-X100 HAZs, as investigated in chapter 7, in CO2-saturated bicarbonate solutions. Across a long potential range from -1.1 to 1.1 V vs. SCE, and a short range from -1.1 to -0.6 V vs. SCE, with a slow scan rate of 1 mV/s for 124  10 cycles, the cathodic reduction, dissolution, as well as the growth, reactivity and stability of the passive films were investigated. Before the cyclic voltammetry tests, the significance of bicarbonate concentration with the passivation process during open-circuit immersion was studied. In dilute solutions containing 0.02 and 0.1 M bicarbonate, the passive films formed slowly and the passivated interfaces were more under the cathodic control of H2CO3, which dissociates more from CO2 hydration. Increasing bicarbonate concentration to 0.5 and 1 M restricted CO2 hydration, and bicarbonate catalyzed dissolution, initially, before it facilitated the formation of effective passive films that made the open-circuit potentials low. During cyclic voltammetry, increasing bicarbonate to an optimum concentration catalyzes anodic dissolution beyond which it enhances the passivation process. In addition, the cathodic currents increased with bicarbonate concentration during the first cycles before they decreased as a result of the growth of the passive films – with increased cycling – which made the cathodic process more sluggish. With increased cycling and bicarbonate concentration, the reactivity of the substrate to passivate was less than the reactivity of the passive film to oxidize. Increasing bicarbonate concentration decreases the significance of HAZ microstructure. Cooling down the near-fusion HAZs at slow rates makes them more reactive and more likely to develop thick corrosion products. Cooling them at high rates makes them more corrosion resistant but more likely to develop thinner corrosion products in low-bicarbonate environments.         125  CHAPTER 9: VOLTAMMETRIC ANALYSIS ON THE FORMATION OF Fe(OH)2 AND FeCO3, AND ON THE REACTIVITY OF PASSIVATION OF API-X100 STEEL HAZs IN CARBONATE SOLUTIONS  In this chapter cyclic voltammetry is used to simulate, analyze and model the corrosion reactions in naturally-aerated carbonate solutions for the same simulated near-fusion API-X100 HAZs tested in chapters 7 and 8. The simultaneous processes of the dissolution, Fe(OH)2 formation, and FeCO3 precipitation in carbonate solutions are examined, and some of the early findings on Fe(OH)2 are revaluated in relation to its kinetics of formation which interferes with that of FeCO3. The formation and electroreactivity of FeCO3 are investigated in detail. The electrochemical significance of the thickness, protectiveness, stability, and reactivity of the passive films are investigated in relation to carbonate concentration, cyclic voltammetry scans, and the cathodic reduction of oxygen and water. With the increased cycling and higher carbonate concentration, the cathodic currents were found indicative of the increasing thickness and reactivity of the passive films. The tests, outlined in section 4.4, were carried out in naturally-aerated solutions of 0.05, 0.1, 0.15, and 0.25 M carbonate at 25oC. The cyclic voltammetry parameters are outlined in section 4.5. The corrosion reactions in the dilute 0.05 M carbonate solution are modeled.   9.1    Dissolution, passivation, and cathodic reduction in 0.05 M carbonate solution The cyclic voltammetry of 10oC/s HAZ, with a range from -1.1 to 0.85 V vs. SCE and scan rate of 1 mV/s, in 0.05 M carbonate solution is shown in Figure 9-1. In 0.05 M solution, each cycle began with cathodic current densities as comparable as -30 μA/cm2, corresponding first to the reduction of water (Eq. 9-1), which was under charge-transfer control.  2H2O + 2e- → H2 + 2OH-            Eo = -1.07 V vs. SCE                  (9-1) 126   Figure 9- 1 Cyclic voltammetry of 10oC/s HAZ of a range from -1.1 to 0.85 V vs. SCE in 0.05 M carbonate solution.  They steeply decreased to -10 μA/cm2 at -1 V vs. SCE, as the cathodic process became increasingly controlled by the reduction of O2 (Eq. 9-2), and as the kinetics of the anodic dissolution started to account in the measured currents.  O2 + 2H2O + 4e- → 4OH-             Eo = 0.159 V vs. SCE                 (9-2) The cathodic currents more gradually decreased until nearly -0.95, at which the surface, in the first cycle, became totally under anodic control. That potential became higher to -0.85 V vs. SCE in the second cycle and with increased cycling to -0.83 V vs. SCE. The anodic region, from the beginning to the first anodic peak at -0.63 V vs. SCE, and the second peak, at -0.43 V vs. SCE, was correlative to the increased cycling, as the currents were correspondingly lower after each cycle. In that region, the anodic dissolution, hydroxide-based pre-filming, and iron 127  carbonate precipitation and growth occur nearly simultaneously. Thermodynamically, they depend on the interfacial hydroxyl and carbonate concentrations, but the significance of the active substrate remains; explaining the decrease in the anodic currents after each cycle. Starting from the second cycle, that region, which is shown in Figure 9-2b, took reproducible patterns with slight differences with increased cycling. In this dilute solution of 0.05 M, hydroxyl initially drives dissolution in a multi-step process before the first peak, which involves specific adsorbents that drive at least one rate-determining step, like Fe(OH)+ in Eq. 9-3b [Lu, 2006], as follows [Valentini, 1985, Castro, 1986]:    Fe + H2O ↔ Fe(OH) + H+ + e-                                                      (9-3a) Fe(OH) ↔ Fe(OH)++ e-                                                                 (9-3b) Fe(OH)+ + OH- ↔ Fe(OH)2 → Hydrous Fe(OH)2                        (9-3c) The mechanism of Fe(OH)2 formation could be different (in the presence of the as-active carbonate), given that the process above was reported to occur in bicarbonate and bicarbonate-carbonate solutions, in which bicarbonate has its specific roles (which are better established in literature) with kinetics that depend and compete with that of Fe(OH)2 formation. In the few available studies in literature, the significance of Fe(OH)2 was not physicochemically investigated, rather, its formation was indicated electrochemically. Disregarding the formation of Fe(OH)2, Thomas et al. utilized their famous potentiostatic-thermodynamic data on the passivation of iron [Thomas, 1967] to conclude that it is magnetite (Fe3O4) that firstly (and mainly) forms during dissolution in carbonate solutions [Thomas, 1970].   128              Figure 9- 2 a) The anodic region of the second cycle, b) the anodic regions from the second to the seventh cycles and c) the cathodic region from -0.75 to -1.1 V vs. SCE from the first to the seventh cycles, of the cyclic voltammetry of 10oC/s HAZ in 0.05 M carbonate solution. 129  They linked that to the first anodic peak that appeared at -0.58 VNHE (-0.83 V vs. SCE) during a forward potentiodynamic scan of 120 mV/min (2 mV/s) in an aerated 0.1 M carbonate solution. They then, however, proposed that FeCO3 forms, but it interferes with the formation of Fe3O4, as the size of the first anodic peak increased with carbonate concentration.  In our case, additional mechanistic analysis is made to outline i) the formation of Fe(OH)2, ii) the underlying reactivity of the substrate (Fe/Fe2+), iii) the reactivity of FeCO3, and iv) the viability of other oxides to form, besides the hydrated γ-ferric oxide Fe(OH)3 which Thomas et al. only proposed to form by a second peak at -0.22 VNHE (-0.472 V vs. SCE) in their potentiodynamic scans.  The electrochemical formation of Fe(OH)2 (Eq. 9-4) is thermodynamically more favourable than that of FeCO3 (Eq. 9-5) to drive first, but not necessarily dominantly, the anodic dissolution (depending on its concentration versus that of carbonate), which exhibits its first slope that is shown in Figure 9-2a. Fe + 2OH- → Fe(OH)2 + 2e-                Eo = -1.117 V vs. SCE        (9-4)   Fe + CO32- → FeCO3 + 2e-                    Eo = -0.999 V vs. SCE        (9-5)                              When the currents followed a higher second slope from 4 to 7 μA/cm2, the Fe(OH)2 clusters achieved an appreciable coverage of the surface to decelerate the currents. The third slope corresponds to the carbonate-driven dissolution that results in FeCO3 (Eq. 9-5), which is more effective, in kinetics different from that of Fe(OH)2, in covering the surface, i.e. in suppressing dissolution to result in the first peak at -0.63 V vs. SCE, which is shown in Figure 9-2a. By that peak, FeCO3 could have formed onto the Fe(OH)2 layer, or directly onto the surface during when carbonate might desorb the first-adsorbed hydroxyl or attack Fe(OH)2 – plausible processes that might result in partial dissolution of Fe(OH)2 and in inhomogeneous precipitation 130  of FeCO3. After the first peak, the currents continued to increase, especially during the first cycles, corresponding to the electroactivity of FeCO3 for nearly 200 mV before the second peak at -0.43 V vs. SCE. It was larger than the first peak and its size did not change with increased cycling. Between the two peaks, FeCO3 was oxidizing to iron oxides in processes that were of comparable thermodynamic tendencies to occur and with kinetics that are governed by the surface state and the amount of FeCO3. Considering the thermodynamic tendency of FeCO3 to transform to α-Fe2O3:  E = 1.12 - 0.177 pH + 0.059 log [CO32-]                (V vs. SCE), at 25oC    (9-6) to γ-Fe2O3:  E = 1.24 - 0.177 pH + 0.059 log [CO32-]                (V vs. SCE), at 25oC     (9-7) and to hydrated γ-Fe2O3 (Fe(OH)3) [Thomas, 1970]:  E = 1.45 - 0.177 pH + 0.059 log [CO32-]                (V vs. SCE), at 25oC      (9-8) The calculated potentials, in relation to 0.05 M carbonate and 11.2 pH, are -0.94 V vs. SCE for α-Fe2O3, -0.82 V vs. SCE for γ-Fe2O3, and -0.61 V vs. SCE for Fe(OH)3. Thomas et al. attributed the second -0.47 V vs. SCE peak, which was smaller than our second -0.43 V vs. SCE peak, to the formation of Fe(OH)3. This was based on i) the colour change (from brown to green between the first and second peaks), ii) the charge release of that peak in comparison to that of cathodic peaks of a reverse scan, and iii) the corresponding potential of Fe(OH)3 that lies between the potentials of the first and second peaks. Between the two peaks, FeCO3 is suspected to have a larger cause with the second peak to also transform slowly and continuously to Fe2O3. Even if FeCO3/ Fe(OH)3 is of a greater potential (to particularly associate to high-potential second anodic peaks), FeCO3 might not have 131  existed in sufficient amounts to transform to Fe2O3 so quickly by the first peak and then in another reaction transforms to Fe(OH)3 by the second peak.  That seemed difficult for Thomas et al. to verify with relatively fast 2 mV/s potentiodynamic scans that were applied only once, not in cycles. As evident in Figure 9-2, the currents continued to increase after the first peak (a feature that did not appear in Thomas et al.’s scans) to correspond to the electro-oxidation of FeCO3 to Fe2O3, which has a stronger thermodynamic drive than that of FeCO3 to Fe(OH)3. With increased cycling, the currents between the first and second peaks were getting lower, as the proportion between FeCO3 (forming directly from the substrate (Eq. 9-5)) and Fe2O3 was decreasing.  After the second peak, the passive currents increased and then decreased gradually to form a third, broad peak of decreasing passive currents that started to stabilize at 5 μA/cm2, from 0.5 V vs. SCE until transpassivation. Across that region, the passive film thickens in a process that is driven by the transformation of FeCO3 and/or Fe3O4 to Fe2O3 as follows [Parkins, 1997(A)(B)]: 2Fe3O4 + H2O ↔ 3Fe2O3 + 2H+ + 2e-        (9-9)               or to γ-Fe2O3, but at higher potentials.  Transpassivation started at comparable potentials of 0.75 V vs. SCE, regardless of increased cycling. It occurred at lower potentials than those in bicarbonate solutions (which have a lower pH of 8). In carbonate solutions, however, carbonate enhances the passive films, but the higher alkalinity makes oxygen generate (Eq. 9-10) at lower potentials.         2H2O → O2+ 4H++ 4e-       Eo = 0.989 V vs. SCE        (9-10)      Moreover, the generation of ferrate species (FeO42-) during transpassivation [El-Naggar, 2004, Beck, 1985, Rangel, 1989] becomes more anodic in carbonate solutions. It is difficult to determine by how much the passive film was superficially affected by the two processes, given 132  that oxygen generation seemed to obliterate important transpassivation features that could have appeared at potentials higher than 0.75 V vs. SCE. The maximum currents of transpassivation decreased from 50 to 35 μA/cm2 with increased cycling.  During the backward scan, the cathodic currents showed reproducibility, regardless of increased cycling. They were decreasing gradually, showing no negative peaks. They were decreasing from 4 μA/cm2, at 0.74 V vs. SCE, to -0.33 V vs. SCE, at which the reduction of oxygen (Eq. 9-2) started onto the passive film, which part of it was susceptible to dissolve by the end of each backward scan. As Figure 9-2c shows, a large cathodic peak appeared between -0.75 and -1 V vs. SCE, followed by a smaller, inner peak between -1 to -1.1 V vs. SCE. The large peak decreased in size with increased cycling, corresponding to the increased anodic activity of the thickening carbonate-based film after each cycle (to make the negative currents more positive), before the currents increased (getting more negative) again to form a small inner peak as a result of the reduction of water (Eq. 9-1).     9.2    Significance of carbonate concentration The cyclic voltammetries of the 10oC/s HAZ in 0.1, 0.15, and 0.25 M carbonate solutions are presented in Figure 9-3. It shows that the cycles are bigger with carbonate concentration. In the forward scan, increasing carbonate concentration increased the rates of anodic dissolution to enhance as a result the precipitation of FeCO3. The reactivity of FeCO3 in higher amounts increased to transform more to the iron oxides Fe2O3, Fe3O4, and Fe(OH)3, making the passive films thicker. They suppress the underlying dissolution further, but they become at the same time of higher reactivity, making the overall electrochemical behavior difficult to correlate to the passive currents. The reactive, thicker passive films react with oxygen, even though its 133  concentration is less in the more concentrated carbonate solutions, [Zhang, 2009, Fu, 2010(A)] as follows:  4FeCO3 + O2 + 4H2O → 2Fe2O3+ 4HCO3-+ 4H+         (9-11)                   Figure 9- 3 Cyclic voltammetry of 10oC/s HAZ of a range from -1.1 to 0.85 V vs. SCE in a) 0.1, b) 0.15, and c) 0.25 M carbonate solutions. 134  6FeCO3 + O2 + 6H2O → 2Fe3O4+ 6HCO3-+ 6H+         (9-12) 4Fe(OH)2 + O2 → 2Fe2O3+ 4H2O                                 (9-13) The higher reactivity of FeCO3 with increased carbonate concentration can be linked to the intensities of the second anodic peaks, which accordingly increased from 25 to 45 μA/cm2 at potentials that increased from -0.4 to -0.35 V vs. SCE. Interestingly, transpassivation occurred at the same potentials of 0.8 V vs. SCE, regardless of carbonate concentration. The oxygen reduction during the backward scan commenced at higher potentials with carbonate concentration at -0.35, -0.3, -0.25, and -0.20 V vs. SCE in 0.05, 0.1, 0.15, and 0.25 M solutions, respectively. The cathodic currents at low potentials in 0.25 M solution showed the most noticeable decrease with increased cycling. This is attributed to the same reason mentioned for the behavior in 0.05 M solution in section 7.1. It is the increasing anodic activity of the passive films that were getting thicker with increased cycling (in 0.05 M solution) and with the higher carbonate concentration of 0.25 M. Bicarbonate might substantially generate from the carbonate aqueous equilibrium in concentrated carbonate solutions to have anodic and cathodic (Eq. 9-14) roles. HCO3- + e- → Hads+ CO32-             Eo = -0.851 V vs. SCE               (9-14)  9.3    Significance of the voltammetric potential range Figure 9-4 shows the cyclic voltammograms of 10oC/s HAZ in 0.05, 0.1, 0.15, and 0.25 M solutions, of anodic ranges extended 350 mV, to 1.2 V vs. SCE, and of cathodic ranges extended 200 mV, to -1.3 V vs. SCE. The passive films are under two opposing effects. Extending the forward scans to higher potentials might thicken the passive films, but making the cathodic potentials lower make the passive films more susceptible to dissolve by water reduction 135  onto them. The first and second anodic peaks appeared at lower potentials and with higher critical current densities than those of the short-ranged voltammetry. Moreover, the passive currents of the forward scans were higher and the cathodic currents, at potentials lower than -0.5 V vs. SCE, were lower than those of the short-ranged voltammetry.  b c d a Figure 9- 4 Cyclic voltammetry of the 10oC/s HAZ of a range from -1.3 to 1.2 V vs. SCE in a) 0.05, b) 0.1, c) 0.15, and d) 0.25 M carbonate solutions. 136   These features indicate that the passive films became thicker with the extended-range voltammetry and they confirm that the passive films become in turn more reactive to be of higher passive currents. Interestingly, extending the voltammetry ranges seems to be more significant than increased cycling in thickening the passive films. Transpassivation occurred at almost the same potentials of the short-ranged voltammetry. Figure 9-5 shows the top-surface of the corrosion products that formed in 0.05 and 0.25 M solutions, reflecting the thicker passivation in 0.25 M solution. 9.4    Significance of fast cooling the HAZs The voltammetric profiles of the 0.05, 0.1, 0.15, and 0.25 M solutions are shown in Figure 9-6. They were of similar features to those of 10oC/s HAZ. Given that the microstructure c Figure 9- 5 Micrographs of the top-surface of the corrosion products formed on 10oC/s HAZ, in a) 0.05 and b) 0.25 M carbonate solutions at the end of the -1.3 to 1.2 V vs. SCE cyclic voltammetry. 137  of the 60oC/s HAZ consists of a mixture of retained austenite and martensite, its anodic behavior in the dilute carbonate solutions seemed different from that in the concentrated ones. However, it consistently exhibited low cathodic activity. In 0.05 and 0.1 M solutions, its voltammetry was not as smooth as that of the 10oC/s HAZ. It hardly exhibited the first anodic peaks and it was of a higher anodic activity in the regions prior to the second peaks, seeming to be of thin passive films, as reported for similar microstructures in bicarbonate and carbonate-bicarbonate solutions.  Figure 9- 6 Cyclic voltammetry of the 60oC/s HAZ of a range from -1.3 to 1.2 V vs. SCE in 0.05, 0.1, 0.15, and 0.25 M carbonate solutions. a b c d 138  9.5    Modeling of corrosion reactions in a dilute carbonate solution The corrosion reactions of the 10oC/s HAZ in 0.05 M carbonate solution are modelled in this section. The cyclic voltammetry started at -1.1 V vs. SCE, at which the cathodic reactions were first under the charge-transfer-controlled reduction of water molecules (Eq. 9-15) which migrate towards the polarized substrate, as depicted in Figure 9-7, amongst carbonate and hydroxyl.   2H2O + 2e- → H2 + 2OH-            Eo = -1.07 V vs. SCE                  (9-15)  Hydrogen is generated and the local alkalinity increases as a result. The generated interfacial hydroxyl ions could be more readily available to drive the earliest adsorption-dissolution steps than the migrating hydroxyl ions from the bulk to the interface. The mechanisms of hydrogen generation, especially in relation to the fashion that the adsorbed hydrogen atoms interact to generate the hydrogen bubbles, are likely different from the Figure 9- 7 Schematic representation of the charge-transfer-limited reduction of water dominating the interface at a potential range from -1.1 to -1 V vs. SCE. 139  commonly-reported mechanisms in the low-pH, CO2-saturated solutions [De Waard, 1975, Ogundele, 1987]. This is due to the high alkalinity and high concentrations of carbonate which adsorbs onto the surface and impedes the migration of H2O to change its reduction mechanisms. This decelerates the reduction rates to be lower cathodic currents – lower than those in the mildly-alkaline, bicarbonate solutions at the same potentials that are scanned with comparable sweep rates. As the potentials become higher than -1 V vs. SCE, the reduction of water becomes increasingly of lower thermodynamic tendency. The water molecules continue migrating, as depicted in Figure 9-8, to engage in an active adsorption that drives dissolution, along with carbonate and hydroxyl at higher potentials, and also drives the reduction of oxygen (Eq. 9-16). O2 + 2H2O + 4e- → 4OH-             Eo = 0.159 V vs. SCE                 (9-16)  Figure 9- 8 Schematic representation of the onset of dissolution, driven by carbonate, water and hydroxyl, and the cathodic reduction of oxygen at a potential range from -1 to -0.845 V vs. SCE.  The two processes continue, but heterogeneously [Puiggali, 2002, Chen, 2002], over the substrate that becomes of diminishing cathodic regions, and of other regions onto which the concentrations of the hydroxyl-based and carbon-carrying adsorbents increase [Bessone, 1977, 140  Doig, 1977]. These adsorbents drive new kinetics that counteracts dissolution, which results in the precipitation of reactive phases [Castro, 1991(B), Ninh Pham, 2006].  The surface became under total anodic control at -0.845 V vs. SCE, as shown in Figure 9-9, during which the interface is saturated with the migrating species as well as with the complex intermediates that generate from the surface.   Figure 9- 9 Schematic representation of the simultaneous dissolution and Fe(OH)2 filming at a potential range from -0.845 to -0.745 V vs. SCE.  Prior to the first peak, the anodic regime consisted of three distinct slopes that correspond to the simultaneous, but considerably consequential processes of dissolution, Fe(OH)2 formation, and FeCO3 precipitation. Figure 9-9 highlights the first slope during which Fe(OH)2 starts to form in dispersed clusters that cover part of the dissolving substrate. Fe(OH)2 (Eq. 9-17) formation is thermodynamically more favourable than that of FeCO3 (Eq. 9-18) in covering the active surface areas [Castro, 1986].      Fe + 2OH- → Fe(OH)2 + 2e-                Eo = -1.117 V vs. SCE         (9-17)   141  Fe + CO32- → FeCO3 + 2e-                    Eo = -0.999 V vs. SCE         (9-18)                              The hydrated Fe(OH)2 was reported to form in a multi-step process (Eqs. 9-19a, b, and c) in the less alkaline, carbonate-bicarbonate solutions [Rangel, 1989, Valentini, 1985], where adsorbents like Fe(OH)+ drive a rate-determining step as [Lu, 2006]: Fe + H2O ↔ Fe(OH) + H+ + e-                                                      (9-19a) Fe(OH) ↔ Fe(OH)++ e-                                                                  (9-19b) Fe(OH)+ + OH- ↔ Fe(OH)2 → Hydrous Fe(OH)2                         (9-19c) As illustrated in Figure 9-10, Fe(OH)2 grows further to cover more of the substrate to suppress its dissolution across a second slope from -0.745 to -0.65 V vs. SCE. Carbonate starts to have significance as it drives dissolution as well as FeCO3 precipitation either under the direct oxidation (Eq. 9-18) or by combining with Fe2+ under supersaturation.  Figure 9- 10 Schematic representation of the simultaneous dissolution, Fe(OH)2 filming, and onset of FeCO3 precipitation at a potential range from -0.745 to -0.650 V vs. SCE.  142  In special kinetics [Muñoz, 2005], FeCO3 grows further across a third, narrow steep slope from -0.65 to -0.63 V vs. SCE covering most of the surface, as depicted in Figure 9-11. When the passivation kinetics surpasses that of dissolution, the first anodic peak at -0.63 V vs. SCE appears.   Figure 9- 11 Schematic representation of the growth of FeCO3 and the coverage of the active surface at a potential range from -0.65 to -0.63 V vs. SCE.  The currents after the first peak continued to increase, indicating at the growth and reactivity of FeCO3. It becomes a second layer atop Fe(OH)2 through which the underlying substrate reactions and the diffusion of Fe2+ depend on its stability [Castro, 1991(A), Simard, 1997,  Zhou, 2011]. FeCO3 could precipitate directly onto the surface, after when carbonate attacks the Fe(OH)2 monolayer in a manner similar to that of bicarbonate [Burstein, 1980] and afterwards drives the electro-oxidation that forms FeCO3. Between the first peak at -0.63 V vs. SCE and the second peak at -0.43 V vs. SCE, the FeCO3-based passivation was subject to 143  electrochemical transformations that changed its thickness, colour and chemistry, as modelled in Figure 9-12. The significance of the ionic flux might not be different from that at the lower potentials, but the passivated interface becomes more under the thermodynamic influence that changes its composition. The three schematics in Figure 9-12 correspond to the growth of FeCO3, the formation of the iron oxides Fe3O4 and Fe2O3 and, at higher potentials, the formation of iron (III) hydroxide Fe(OH)3.   Figure 9- 12 Schematic representation of the growth of FeCO3, its oxidation to the iron oxides Fe2O3 and Fe3O4 (which simultaneously also grow directly from the surface), and to the hydrated oxide Fe(OH)3 at the end of the potential range from -0.63 to -0.5 V vs. SCE.  Under the increased activity of the underlying substrate, the porous, FeCO3-based passive film [Flis, 1992] grows further within which the diffusing carbonate ions interact with the residing ferrous ions to precipitate FeCO3 – at rates possibly much faster than those in the CO2-saturated brines [Palacios, 1991] due to the high alkalinity. Simultaneously, Fe3O4 (which has a high thermodynamic drive to form at low anodic potentials [Thomas, 1967, Thomas, 1970]) forms directly onto the surface at locations where the porous passive film is hydrated enough and has a high local alkalinity, as follows:       144  3Fe + 4H2O → Fe3O4 + 8H+ + 8e-                (9-20)  The overall anodic activity of the passive film increases as a result (to be of rising passive currents between the two peaks) during which Fe3O4 and Fe2O3 collectively grow to form a conductive “root” between the outer film layer and the substrate surface. The significance of the iron oxides conductivity, to be noted, did not receive interest in the literature of corrosion in bicarbonate and carbonate solutions in which the emphasis mostly was on the environmental factors that affect the passivation behavior overall [Atkinson, 1985, Chang, 1972].  The colour changed with the upscan to light red, as depicted in Figure 9-12. Across the same range of potentials, Thomas et al. reported, with a faster scan rate, that the passive film colour changed from green to brown [Thomas, 1970]. Simultaneously, as FeCO3 grows, it electro-oxidizes more to the iron oxides and other intermediate products that cover it and change therefore the colour. Within similar potential ranges overlapping with that highlighted in Figure 9-12, the FeCO3/oxide transformations were reported to occur in carbonate solutions [Thomas, 1970], carbonate-bicarbonate solutions [Castro, 1986, Rangel, 1989, Valentini, 1985, Zhang, 2010], and bicarbonate solutions [Lu, 2006, Eliyan, 2012]. However, it should be noted that the thermodynamic drive for these reactions decreases respectively in these solutions as their pHs are accordingly lower. Interestingly, in literature, carbonate was not reported to drive these reactions; rather, it was as the carbon-carrying bicarbonate, a product, not a reactant. For Fe3O4, it was reported to form from FeCO3 as follows: 3FeCO3 + 4H2O → Fe3O4 + 3HCO3- + 5H+ + 2e-        (9-21) 3FeCO3 + 5OH- → Fe3O4 + 3HCO3- + H2O + 2e-        (9-22) For Fe2O3, Thomas et al. raised not the significance of the ferrous ions to drive its transformation from FeCO3 (which was reported though in an earlier study by Nagayama et al. 145  [Nagayama, 1963] to occur) due to the low solubility limits in the Thomas et al.’s carbonate solutions [Thomas, 1970]. Rather, they reported the electrochemical transformation – similarly to Parkins et al. who reported it to occur in bicarbonate-carbonate solutions [Parkins, 1997(A)(B)] as follows: 2FeCO3 + 3H2O → Fe2O3+ 2CO32-+ 6H++ 2e-           (9-23)   Lu et al. reported it to occur in bicarbonate solutions as follows [Lu, 2006]:     2FeCO3 + 4OH- → Fe2O3+ 2HCO3- + H2O + 2e-       (9-24) Before the second peak, at -0.43 V vs. SCE, a new product is attributed for the change in the passivation behavior. Thomas et al., based on thermodynamic calculations, reported that the hydrated γ-Fe2O3, Fe(OH)3, which is depicted to grow and cover FeCO3 in Figures 9-12  and 9-13, formed by a second peak during a 2 mV/s potentiodynamic scan.    Figure 9- 13 Schematic representation of the overall thickening of the passive film and the growth of Fe(OH)3 to cover most of outer layer at -0.4 V vs. SCE.  The peak was smaller than this case’s, and it is suspected that it was due to only the formation of Fe(OH)3. Rather, FeCO3 might not have existed in sufficient amounts to 146  significantly transform only into Fe(OH)3, but it was growing and simultaneously undergoing multiple oxidation reactions to result in a mixture of products that were slowly and continuously covering the reacting FeCO3.           As shown in Figure 9-14, the passive currents increased slightly after the second peak and then decreased to form a third 500 mV-broad peak, during which the passive film was thickening and increasingly suppressing the dissolution of the underlying substrate.   Figure 9- 14 Schematic representation of the thickening and layering of the passive film at a potential range from -0.4 to 0.7 V vs. SCE, before transpassivation starts.  For the last 300 mV before transpassivation, the passive currents stabilized to 5 μA/cm2, indicating at the electrochemical homogeneity of the passive film through which oxygen was reported to diffuse. As a result, further transformations were reported to occur during that stage of passivation [Parkins, 1997(A)(B), Zhang, 2009, Beck, 1985], as follows:   4FeCO3 + O2 + 4H2O → 2Fe2O3 + 4HCO3- + 4H+       (9-25) 2FeCO3 + 2H2O → 3γ-Fe2O3 + 2H++ 2e-                     (9-26) 147  At potentials higher than 0.75 V vs. SCE, the oxygen generation (Eq. 9-27) and transpassivation commenced, which are depicted in Figure 9-15. They are driven by the intense ionic flux of the water molecules and high alkalinity.  2H2O → O2 + 4H+ + 4e-               (9-27)       Figure 9- 15 Schematic representation of the transpassivation and oxygen generation at a potential range from 0.70 to 0.85 V vs. SCE.  The current densities sharply increased to as high as 50 μA/cm2, mostly under oxygen generation that seemed to obliterate important transpassivation features which could correspond, as previously reported from scans in carbonate-bicarbonate solutions, to the ferrate (FeO42-) generation [Rangel, 1989, Beck, 1985, El-Naggar, 2004] and its decomposition to FeO4- and oxygen (Eq. 9-28 [Beck, 1985]).   4FeO42- + 2H2O → 4FeO2- + 3O2+ 4OH-             (9-28) The passive film dissolves partially by the two processes but it is difficult to determine which one has the greatest significance. In literature, this is still being investigated in a variety of 148  solutions of different pHs.  Upon the scan reversal, which is represented in Figure 9-16, the difference between the backward and forward passive currents increased gradually until the reduction of oxygen onto the passive film started at -0.38 V vs. SCE. During that stage, no peaks appeared to correspond to passivation reactions taking place significantly in reverse [Von Fraunhofer, 1970] or to compare their charge release to those of the anodic peaks of the forward scan. However, some of the passive film loss that occurred during transpassivation could have been slightly recovered, but that requires special validations.   Figure 9- 16 Schematic representation of the partial recovery of the passive film after transpassivation at a potential range from 0.85 to -0.35 V vs. SCE, before the reduction of oxygen starts.  As illustrated in Figure 9-17, the surface became totally under cathodic control at -0.38 V vs. SCE, onto which the oxygen reduction started with kinetics that depends on the quality of the passive film. The rate of the oxygen reduction increased with the negative scan and the colour of the passive film changed to become greenish-brown as the oxides transform back to FeCO3. This depends on the rate of the cathodic polarization (1 mV/s, in our case) and the local alkalinity.  149    Figure 9- 17 Schematic representation of the reduction of oxygen onto the passive film, and the reduction of the ferric-based oxides to FeCO3 at a potential range from -0.35 to -0.9 V vs. SCE.    Figure 9- 18 Figure 9-18 Schematic representation of the onset of the reduction of water, which might significantly result in partial dissolution of the passive film, and the anodic activity of the carbonate-based layer to grow as the potential becomes lower from -0.9 to -1.1 V vs. SCE. 150  The cathodic currents started to decrease to form a large cathodic peak with an intensity of -35 μA/cm2, at -0.9 V vs. SCE corresponding to abrupt anodic activity of the passive film, as depicted in Figure 9-18. A considerable thickening of the passive film could happen as a result.  At -1 V vs. SCE, the charge-transfer-controlled reduction of water raised again the cathodic currents to make an inner small peak, and then the hydrogen generation takes place onto the passive film. At this last stage of voltammetry, the passive film is under two opposing effects, one that thickens and another that dissolves it, but the net significance on the stability of the passive film depends on the applied low potentials and the rates of the hydrogen generation.  9.5     Summary The corrosion behavior for two types of simulated near-fusion API-X100 HAZs in naturally-aerated carbonate solutions was examined by cyclic voltammetry. The anodic dissolution, passivation, cathodic reduction, and the simultaneously-occurring processes of dissolution, Fe(OH)2 formation, and the precipitation of FeCO3 were examined. New findings were outlined and some of the earlier findings, reported by Thomas et al. in 1970 [Thomas, 1970], were re-examined and refuted. The electrochemical significance of the thickness, stability, protectiveness, and reactivity of the passive films were studied, and the following, in brief, is concluded:    1. Fe(OH)2 forms in carbonate solutions. This occurs with kinetics that depends on carbonate concentration and interferes with the kinetics of FeCO3 formation. 2. FeCO3 forms afterwards, but more significantly to suppress the underlying dissolution. It becomes a reactive layer that slowly and continuously transforms to iron oxides with processes of comparable thermodynamic tendencies. 151  3. With the increased voltammetric cycles, the passive films become thicker and of high current densities. They are more indicative of the reactivity of the passive films than of their effectiveness in suppressing the underlying dissolution. 4. Increasing carbonate concentration increases the thickness of the passive films. 5. The cathodic reduction currents of oxygen and water are indicative of the increasing thickness of the passive films after each cycle. 6. Extending the ranges of the voltammetric scans increases the thickness of the passive films in a manner more significant than that by the increased cycles.  7. Cooling down the welded pipeline joints at high rates make them have low cathodic activity and develop thin protective films.                             152  CHAPTER 10: CONCLUSION   This thesis investigated the external pipeline corrosion behavior of near-fusion HAZs simulated by Gleeble© thermal simulation cycles. For API-X80 and API-X100 steels, the microstructural features were, respectively, quantitatively and qualitatively linked to the corrosion behavior. The electrochemical methods were applied to analyze the cathodic reduction, dissolution, passivation, stability, reactivity, and protectiveness of the passive films. The laboratory environmental conditions do not simulate specific environmental corrosion conditions in the field. They contain the corrosive constituents with specific ranges that make the electrochemical assessments comprehensive and help to understand the corrosion behavior in many conditions in the field. The following, in brief, is concluded. 1. The corrosion kinetics, rates, and mechanisms do not depend on the HAZ microstructures. This was found for eight different API-X80 HAZs tested by potentiodynamic polarization and electrochemical impedance spectroscopy in 0.1 M bicarbonate – 1wt% chloride solutions at 25oC. 2. The passivation process depends on the HAZ microstructures. Cooling down the simulated HAZs mentioned in point 1 from low peak temperatures make them have more ferrite and martensite-retained austenite contents, to which an earlier onset of formation and more protectiveness of the passive films were attributed.   3. At high temperatures, the ferritic HAZs develop thick, stable, and protective passive films. The acicular-ferritic and martensitic HAZs develop thinner and less stable passive films and they have lower anodic and cathodic activities. This was revealed by potentiodynamic and potentiostatic polarizations and electrochemical impedance 153  spectroscopy in naturally-aerated, dilute 0.05 M bicarbonate solutions at 90oC for the API-X100 HAZs.      4. In small concentrations, bicarbonate exacerbates dissolution and prevents passivation in naturally-aerated and CO2-saturated solutions, regardless of the microstructures of the API-X100 HAZs. In higher concentrations, bicarbonate is a protective agent that promotes passivation and increases its thickness and protectiveness. 5. Fe(OH)2 forms in carbonate solutions with kinetics interfering with that of FeCO3 and is dependent on carbonate concentration. 6. Increasing carbonate concentration protects the surface, promotes passivation, and increases the thickness and reactivity of the passive films, regardless of the HAZ microstructures.  7. The cathodic reduction currents in bicarbonate and carbonate solutions indicate the thickness and reactivity of the passive films, which increase with increased voltammetric cycling and bicarbonate and carbonate concentrations.   In the field, it is generally recommended welding the new-generation pipeline steels with procedures that allow them to cool off at fast rates so that the welded joints develop near-fusion HAZs that are more martensitic and acicular ferritic. They have low anodic and cathodic reactivities, have a low tendency to develop galvanic effects, but develop thin corrosion products. In harsh external pipeline corrosion conditions in which the corrosion products could play protective roles because of their thickness, the welded joints should be cooled at more controlled rates so that they can develop more ferritic structures. They could initially dissolve at high rates but, in the long run, the corrosion products that develop are protective in environments 154  that contain low bicarbonate concentrations, chloride, or considerable oxygen traces, regardless of temperature.                      155  CHAPTER 11: RECOMMENDATIONS FOR FUTURE WORK   The HAZ corrosion vulnerability can be investigated in future studies in simulated environmental conditions with more proximity to the corrosion conditions in the field. In regard to the metallurgy, the HAZs have complex microstructures whose significance with the corrosion behavior, across a narrow region in the pipeline, cannot be fully understood from HAZs simulated by Gleeble© thermal cycles. Carrying out electrochemical tests on HAZs taken from welded segments provides better clarity on other aspects that govern their corrosion behavior in many environmental conditions. The galvanic coupling between the HAZs, welding fusion, and base pipeline steel can be investigated in future studies that describe the corrosion behavior of the system overall – from which i) the distribution of the anodic and cathodic regions can be determined and ii) the susceptibility for localized corrosion and tendency for thicker corrosion products to form can be predicted. The HAZs in the field are subject to localized corrosion that could result in accelerated thinning that raise the stress and cracking susceptibility. Therefore, carrying out corrosion tests in high-pressure autoclave solutions and investigating the synergistic influence of applied stresses in stress-corrosion-cracking tests will reveal more details about the electrochemical process that govern the corrosion behavior of the HAZs. The corrosion products and passive films of the HAZs should be further studied with a methodology that correlates their properties more comprehensively to the environmental conditions and the HAZ microstructures. From that, the pipelines welding processes can ultimately be optimized and the HAZ microstructures can be better controlled for metallurgical systems that have lower corrosion susceptibility.             156  REFERENCES   Adamy S., Cala F. (1999), inhibition of pitting in ferrous materials by carbonate as a function of temperature and alkalinity, Corrosion, 55, 825-839.   Al-Hassan S., Mishra B., Olson D., Salama M. (1998), Effect of microstructure on corrosion of steels in aqueous solutions containing carbon dioxide, Corrosion, 54, 480-491.  Alves V., Brett C. (2002) (A), Influence of alloying on the passive behavior of steels in bicarbonate medium, Corrosion Science, 44, 1949-1965.  Alves V., Brett C. (2002) (B), Characterisation of passive films formed on mild steels in bicarbonate solution by EIS, Electrochimica Acta, 47, 2081-2091.  ASTM Standard G 5-82 (1993), Standard Practice for Standard Reference Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements, 1993 Annual Book of Standards, Vol. 3.02 (Philadelphia, PA: ASTM).  Bard A., Faulkner L. (2001), Electrochemical Methods – Fundamentals and Application, second ed., Wiley, New York.   Beck F., R. Kaus, M. Oberst (1985), Transpassive dissolution of iron to ferrate (VI) in concentrated alkali hydroxide solutions, Electrochimica Acta, 30, 173-183.  Bowker J., Ng-Yelim J., Malis T. (1989), Effect of weld thermal cycle on behaviour of Ti–Nb carbonitrides in HSLA steel Materials Science and Technology, 5, 1034-1036.  Boukamp B. (1989), Equivalent Circuit User’s Manual, 2nd ed., University of Twente, Einshede. 157   Brasher D. (1968), Stability of the oxide film on metals in relation to inhibition of corrosion. II. Dual role of the anion in the mild inhibition of the corrosion of mild steel, British Corrosion Journal, 4, 122-128.  Breche Y., Militzer M. (2005), A note on grain size dependent pinning, Scripta Materialia, 52, 1299-1303.  Brossia C., Cragnolino G. (2000), Effect of environmental variables on localized corrosion of carbon steel, Corrosion, 56, 505-514.  Bruschi R. (2012), From the longest to the deepest pipelines, Proceedings of the Twenty-second (2012) International Offshore and Polar Engineering Conference.  Burstein G., Davies D. (1981), The electrochemical behavior of scratched iron surfaces in aqueous solutions, Journal of the Electrochemical Society, 128, 33-39.  Burstein G., Davies D. (1980), The effects of anions on the behaviour of scratched iron electrodes in aqueous solutions, Corrosion Science, 20, 1143-1155.  Castro E., Vilche J. (1991) (A), Electrooxidation/electroreduction processes at composite iron hydroxide layers in carbonate-bicarbonate buffers, Journal of Applied Electrochemistry, 21, 543-551.  Castro B., Vilche J., Arvia A. (1991) (B), Iron dissolution and passivation in K2CO3-KHCO3 solutions. Rotating ring disc electrode and XPS studies, Corrosion Science, 32, 37-50.  Castro E., Valentini C., Moina C., Vilche J., Arvia A. (1986), The influence of ionic composition on the electrodissolution and passivation of iron electrodes in potassium carbonate-bicarbonate solutions in the 8.4-10.5 pH range at 25°C, Corrosion Science, 26, 781-789,791-793. 158   Chang R., Wagner Jr. J. (1972), Direct-current conductivity and iron tracer diffusion in hematite at high temperature, Journal of the American Ceramic Society, 55, 211-213.  Chen X., Qiao G., Han X., Wang X., Xu F., Liao B. (2014), Effects of Mo, Cr and Nb on microstructure and mechanical properties of heat affected zone for Nb-bearing X80 pipeline steels, Materials & Design, 53, 888-901.   Chen W., King F., Jack T., Wilmott M. (2002), Environmental aspects of near-neutral pH stress corrosion cracking of pipeline steel, Metallurgical and Materials Transactions A, 33, 1429-1436.   Collins L., Godden M., Boyd J. (1983), Microstructures of linepipe steels, Canadian Metallurgical Quarterly, 22, 169-179.  Cordonier J., Militzer M., Jacot A. (2005), in: Solid-to-Solid Phase Transformations in Inorganic Materials, 2005, TMS (The Minerals, Metals and Materials Society), 735-740.  Davies D., Burstein G. (1980), Effects of bicarbonate on the corrosion and passivation of iron, Corrosion, 36, 416-422.  De A., Speer J., Matlock D. (2003), Color tint-etching for multiphase steels, Advanced Materials and Processes, 161, 27-30.  De Waard C., Milliams D. (1975), Carbonic acid corrosion of steel, Corrosion, 31, 177-181.  Dugstad A. (1998), Mechanism of protective film formation during CO2 corrosion of carbon steel, CORROSION/98, Paper No. 31, NACE.  Easterling  K. (1983), Introduction to the Physical Metallurgy of Welding, Butterworths and Co, Ltd. London, UK, pp. 105-155. 159   Eliyan F., Mahdi E., Farhat Z., Alfantazi A. (2013), Interpreting the passivation of HSLA steel from electrochemical corrosion investigations in bicarbonate-oil aqueous emulsions, International Journal of Electrochemical Science, 8, 3026-3038.  Eliyan F., Mahdi E., Alfantazi A. (2012), Electrochemical evaluation of the corrosion behaviour of API-X100 pipeline steel in aerated bicarbonate solutions, Corrosion Science, 58, 181-191.  El-Naggar M. (2006), Effects of Cl-, NO3- and SO42- anions on the anodic behavior of carbon steel in dearated 0.50 M NaHCO3 solutions, Applied Surface Science, 252, 6179-6194.  El-Naggar M. (2004), Cyclic voltammetric studies of carbon steel in deaerated NaHCO3 solution, Journal of Applied Electrochemistry, 34, 911-918.  Flis F. and T. Zakroczymski (1992), Enhanced hydrogen entry in iron at low anodic and low cathodic polarizations in neutral and alkaline solutions, Corrosion, 48, 530.  Fu A., Cheng Y. (2010) (A), Electrochemical polarization behavior of X70 steel in thin carbonate/bicarbonate solution layers trapped under a disbonded coating and its implication on pipeline SCC, Corrosion Science, 52, 2511-2518.  Fu A., Cheng Y. (2010) (B), Effects of alternating current on corrosion of a coated pipeline steel in a chloride-containing carbonate/bicarbonate solution, Corrosion Science, 52, 612-619.  Garcia C., DeArdo A. (1981), Formation of austenite in 1.5 pct Mn steels, Metallurgical and Materials Transactions A, 12, 521-530.  Geldreich E. (1996), Microbial Quality of Water Supply in Distribution Systems, 1st ed., (New York:  CRC-Press).  160   Gonzalez-Rodriguez J., Casales M., Salinas-Bravo V., Albarran J., Martinez L. (2002), Effect of microstructure on the stress corrosion cracking of X-80 pipeline steel in diluted sodium bicarbonate solutions, Corrosion, 58, 584-590.  Hamadou L., Kadri A., Benbrahim N. (2005), Characterisation of passive films formed on low carbon steel in borate buffer solution (pH 9.2) by electrochemical impedance spectroscopy, Applied Surface Science, 252, 1510-1519.  Han J., Zhang J., Carey J. (2011) (A), Effect of bicarbonate on corrosion of carbon steel in CO2 saturated brines, International Journal of Greenhouse Gas Control, 5, 1680-1683.  Han J., Nešić S., Yang Y., Brown B. (2011) (B), Spontaneous passivation observations during scale formation on mild steel in CO2 brines, Electrochimica Acta, 56, 5396-5404.  Hirnyi S. (2001), Anodic hydrogenation of iron in a carbonate-bicarbonate solution, Materials Science, 37, 491-498.  Huang J., Poole W., Militzer M. (2004), Austenite formation during intercritical annealing, Metallurgical and Materials Transactions A, 35, 3363-3375.  Hudson M. (2004), Welding of X100 Linepipe, Cranfield University Cranfield, 400.  Jelinek J., Neufeld P. (1980), Temperature effect on pitting corrosion of mild steel in de-aerated sodium bicarbonate-chloride solutions, Corrosion Science, 20, 489-496.  Juttner K. (1990), Electrochemical impedance spectroscopy (EIS) of corrosion processes on inhomogeneous surfaces, Electrochimica Acta, 35, 1501-1508.  Kennedy J. (1993), Oil and gas pipeline fundamentals, 2nd ed., Pennwell, Oklahoma.   Kim S., Im Y., Lee S., Lee H., Kim S., Hong J. (2004), Effects of alloying elements on fracture toughness in the transition temperature region of base metals and 161  simulated heat-affected zones of Mn-Mo-Ni low-alloy steels, Metallurgical and Materials Transactions A, 35, 2027-2037.    Kinsella B., Tan Y., Bailey S. (1998), Electrochemical impedance spectroscopy and surface characterization techniques to study carbon dioxide corrosion product scales, Corrosion, 54, 835-842.  Kolotyrkin J. (1963), Pitting corrosion of metals, Corrosion, 19, p. 261t-268t.  Lee C., Qin Z., Odziemkowski M., Shoesmith D. (2006), The influence of groundwater anions on the impedance behaviour of carbon steel corroding under anoxic conditions, Electrochimica Acta, 51, 1558-1568.  LePera F. (1979), improved etching technique for the determination of percent martensite in high-strength dual-phase steels, Metallography, 12, 263-268.  Li M., Cheng Y. (2008) (A), Corrosion of the stressed pipe steel in carbonate–bicarbonate solution studied by scanning localized electrochemical impedance spectroscopy Electrochimica Acta 53, 2831-2836.  Li J., Zuo J. (2008) (B), Influences of temperature and pH value on the corrosion behaviors of X80 pipeline steel in carbonate/bicarbonate buffer solution, Chinese Journal of Chemistry, 26, 1799-1805.  Li D., Feng Y., Bai Z., Zhu J., Zheng M. (2008) (C), Photo-electrochemical analysis of passive film formed on X80 pipeline steel in bicarbonate/carbonate buffer solution, Applied Surface Science, 254, 2837-2843.  Li D., Feng Y., Bai Z., Zhu J., Zheng M. (2007), Influence of temperature, chloride ions and chromium element on the electronic property of passive film formed on carbon steel in bicarbonate/carbonate buffer solution, Electrochimica Acta, 52, 7877-7884. 162   Linter R., Burstein G. (1999), Reactions of pipeline steels in carbon dioxide solutions, Corrosion Science, 41, 117-139.  Louafi Y., Ladjouzi M., Taibi K. (2010), Dissolved carbon dioxide effect on the behaviour of carbon steel in a simulated solution at different températures and immersion times, Journal of Solide State Electrochemistry, 14, 1499-1508.  Lu Z., Huang C., Huang D., Yang W. (2006), Effects of a magnetic field on the anodic dissolution, passivation and transpassivation behaviour of iron in weakly alkaline solutions with or without halides, Corrosion Science, 48, 3049-3077.  Mansfeld F. (1981), Recording and analysis of AC impedance data for corrosion studies, Corrosion, 36, 301-307.  Mao X., Liu X., Revie R. (1994), Pitting corrosion of pipeline steel in dilute bicarbonate solution with chloride ions, Corrosion, 50, 651-657.   Militzer M., Hawbolt E., Meadowcroft T. (2000), Microstructural model for hot strip rolling of high-strength low-alloy steels, Metallurgical and Materials Transactions A, 31, 1247-1259.  Mitsui H., Takahashi R., Asano H., Taniguchi N., Yui M. (2008), Susceptibility to stress corrosion cracking for low-carbon steel welds in carbonate-bicarbonate solution, Corrosion, 64, 939-948.  Mohorich M., J. Lamb, D. Chandra, J. Daemen, R. Rebak (2010), Electrochemical studies on silicate and bicarbonate ions for corrosion inhibitors, Metallurgical and Materials Transactions A, 41,  2563-2574.    Moiseeva L. (2005), Carbon dioxide corrosion of oil and gas field equipment, Protection of Metals, 41, 76-83. 163   Nakata N., Militzer M. (2005), Modeling of microstructure evolution during hot rolling of a 780 MPa high strength steel, ISIJ International, 45, 82-90.  Nagayama M., Cohen M. (1963), The anodic oxidation of iron in a neutral solution II. Effect of ferrous ion and pH on the behavior of passive iron, Journal of the Electrochemical Society, 110, 670-680.  Neshati J., Masiha H., Mahjani M., Jafarian M. (2007), Study of corrosion of carbon steel API 5L (X60) in NaHCO3/NaCl solutions by electrochemical noise and impedance measurements, Corrosion Engineering, Science and Technology, 42, 371-376.  Ninh Pham A., Rose A., Feitz A., Waite T. (2006), Kinetics of Fe(III) precipitation in aqueous solutions at pH 6.0–9.5 and 25°C, Geochimica et Cosmochimica Acta, 70, 640-650.  Niu L., Cheng Y. (2007), Corrosion behaviour of X-70 pipe steel in near-neutral pH solution, Applied Surface Science, 253, 8626-8631.  Ogundele G., White W. (1987), Observations on the influences of dissolved hydrocarbon gases and variable water chemistries on corrosion of an API-L80 steel, Corrosion 43, 665-673.  Olasolo M., Uranga P., Rodriguez-Ibabe J., López B. (2011), Effect of austenite microstructure and cooling rate on transformation characteristics in a low carbon Nb–V microalloyed steel, Materials Science and Engineering A, 528, 2559-2569.  Omonbude E. (2012), Cross-border oil and gas pipelines and the role of the transit country: economics, challenges and solutions, 1st ed., Palgrave Macmillan, New York.  Palacios C., Shadley J. (1991), Characteristics of corrosion scales on steels in a CO2-saturated NaCl brine, Corrosion, 47, 122-127.  164   Palmiere E., Garcia C., DeArdo A. (1994), The influence of niobium supersaturation in austenite on the static recrystallization behavior of low carbon microalloyed steels, Metallurgical and Materials Transactions A, 25, 277-286.  Park J., Pyun S., Lee W., Kim H. (1999), Effect of bicarbonate ion additives on pitting corrosion of Type 316L stainless steel in aqueous 0.5 M sodium chloride solution, Corrosion, 55, 380-387.  Parkins R., Slattery P., Poulson B. (1981), The effects of alloying additions to ferritic steels upon stress corrosion cracking resistance, Corrosion, 37, 650-664.  Parkins R., Zhou S. (1997) (A), The stress corrosion cracking of C-Mn steel in CO2-HCO3-- CO32- solutions. I: Stress corrosion data, Corrosion Science, 39, 159-173.  Parkins R. , Zhou S.  (1997) (B), The stress corrosion cracking of C-Mn steel in CO2-HCO3--CO32-solutions. II: Electrochemical and other data, Corrosion Science 39 175-191.  Porter A., Easterling K. (2000), Phase transformations in metals and alloys, 2nd ed., Stanley Thorns, Cheltenham.  Pourbaix M., Vandervelden F. (1965), Intensiostatic and potentiostatic methods. Their use to predetermine the circumstances for corrosion or non-corrosion of metals and alloys, Corrosion Science, 5, 81-111.   Puiggali M., Rousserie S., Touzet M. (2002), Fatigue crack initiation on low-carbon steel pipes in a near-neutral-ph environment under potential control conditions, Corrosion, 58, 961-970. 165   Qin Z., Demko B., Noel J., Shoesmith D., King F., Worthingham R., Keith K. (2004), Localized dissolution of millscale-covered pipeline steel surfaces, Corrosion, 60, 906-914.   Ramírez E., González-Rodriguez J., Torres-Islas A., Serna S., Campillo B., Dominguez-Patiño G., Juárez-Islas J. (2008), Effect of microstructure on the sulphide stress cracking susceptibility of a high strength pipeline steel, Corrosion Science, 50, 3534-3541.  Rangel C., Leitão R. (1989), Voltammetric studies of the transpassive dissolution of mild steel in carbonate/bicarbonate solutions, Electrochimica Acta, 34, 255-263.  Rangel C., Fonseca I., Leitão R. (1986), Some aspects of the electrochemical behavior of mild steel in carbonate/bicarbonate solutions, Electrochimica Acta, 31, 1659-1662.  Rao Z., Hu J., Liao S., Tsai H. (2010), Modeling of the transport phenomena in GMAW using argon–helium mixtures part II – the metal, International Journal of Heat and Mass Transfer, 53,  5722-5732.  Remita E., Tribollet B., Sutter E., Vivier V., Ropital F., Kittel J. (2008), Hydrogen evolution in aqueous solutions containing dissolved CO2: quantitative contribution of the buffering effect, Corrosion Science, 50, 1433-1440.  Riley A., Sykes J. (1990), The active-passive transition in low alloy steels in carbonate solutions, Electrochimica Acta, 35, 35-45.  Savoye S., Legrand L., Sagon G., Lecomte S., Chausse A., Messina R., Toulhoat P. (2001), Experimental investigations on iron corrosion products formed in bicarbonate/carbonate-containing solutions at 90°C, Corrosion Science, 43, 2049-2064. 166   Sazzadur Rahman M., Divi S., Chandra D., Daemen J. (2008), Effect of different salts on the corrosion properties of friction type A607 steel rock bolt in simulated concentrated water, Tunnelling and Underground Space Technology, 23, 665-673.  Schmitt G., Mueller M., Papenfuss M., Strobel-Effertz E. (1999), Understanding localized CO2 corrosion of carbon steel from physical properties of iron carbonate scales, CORROSION/99, Paper No. 38, NACE, Houston, TX.   Sherar B., Keech P., Shoesmith D. (2011), Carbon steel corrosion under anaerobic–aerobic cycling conditions in near-neutral pH saline solutions – Part 1: Long term corrosion behavior, Corrosion Science, 53, 3636-3642.  Sherar B., Keech P., Qin Z., King F., Shoesmith D. (2010), Nominally aerobic corrosion of carbon steel in near-neutral pH saline environments, Corrosion, 66, 45001-1-45001-11.  Shome M., Mohanty O. (2006), Continuous cooling transformation diagrams applicable to the heat-affected zone of HSLA-80 and HSLA-100 steels, Metallurgical and Materials Transactions A, 37, 2159-2169.  Simard S., Drogowska M., Meânard H. (1997),  Electrochemical behaviour of 1024 mild steel in slightly alkaline bicarbonate solutions, Journal of Applied Electrochemistry, 27, 317-324.  Singh R. (2013), Pipeline Integrity Handbook: Risk Management and Evaluation, 1st ed., Gulf Professional Publishing, New York.  Spanos G., Fonda R., Vandermeer R., Matuszeski A. (1995), Microstructural changes in HSLA-100 steel thermally cycled to simulate the heat-affected zone during welding, Metallurgical and Materials Transactions A , 26, 3277-3293. 167   Speich G., Demarest  A., Miller R. (1981), Formation of Austenite During Intercritical Annealing of Dual-Phase Steels, Metallurgical and Materials Transactions A, 12A, 1419-1428.  Staicopolus D. (1963), The role of cementite in the acidic corrosion of steel, Journal of the Electrochemical Society, 110, 1121-1124.  Stewart G., Elwazri A., Varano R., Pokutylowicz N., Yue S. (2006), Jonas J., Shear punch testing of welded pipeline steel , Materials Science and Engineering: A, 20, 115-121.  Stikma J., Bradford S. (1985), Stress corrosion cracking of dual-phase steel in carbonate/bicarbonate solutions, Corrosion, 41, 446-450.  Tafteh R. (2011), Austenite decomposition in an X-80 linepipe steel, MSc. Thesis, The University of British Columbia, 2011.  Tang Y., Zuo Y. (2004), The metastable pitting of mild steel in bicarbonate solutions, Materials Chemistry and Physics, 88, 221-226.   Thomas J., Nurse T., Walker R. (1970), Anodic passivation of iron in carbonate solutions, British Corrosion Journal, 5, 87-92.   Thomas J., Nurse T. (1967), The Anodic Passivation of iron in solutions of inhibitive anions, British Corrosion Journal, 2, 13-20.  Thompson S., Colvin D., Krauss G. (1996), Austenite decomposition during continuous cooling of an HSLA-80 plate steel, Metallurgical and Materials Transactions A, 27, 1557-1571. 168   Thompson S., Colvin D., Krauss G. (1990), Continuous cooling transformations and microstructures in a low-carbon, high-strength low-alloy plate steel, Metallurgical and Materials Transactions A, 21, 1493-1507.  Torres-Islas A., González-Rodríguez J. (2009), Effect of electrochemical potential and solution concentration on the SCC behaviour of X-70 pipeline steel in NaHCO3, International Journal of Electrochemical Science, 4, 640-653.   Torres-Islas A., Gonzalez-Rodriguez J., Uruchurtu J., Serna S. (2008), Stress corrosion cracking study of microalloyed pipeline steels in dilute NaHCO3 solutions, Corrosion Science, 50, 2831-2839.  Tsuru Y., Savva C., Aust T. (1987), Construction and use of a scanning micro-electrode to study the localized corrosion of a low-C-Mn steel weldment in 10-3 M sulfuric acid, Corrosion, 45, 415-419.  Valentini C., Moina C. (1985), The electrochemical behavior of iron in stagnant and stirred potassium carbonate-bicarbonate solutions in the 0-75oC temperature range, Corrosion Science, 25, 985-997.  Videm K., Kvarekvaal J., Perez T., Fitzsimons G. (1996), Surface effects of the electrochemistry of iron and carbon steel electrodes in aqueous CO2 solutions, CORROSION/1996. Paper no. 1, NACE.  Videm K., Koren A. (1993), Corrosion, passivity, and pitting of carbon steel in aqueous solutions of HCO3-, CO2 , and Cl- , Corrosion, 49, 746-754.  Von Fraunhofer J. (1970), The polarization behaviour of mild steel in aerated and de-aerated 1M NaHCO3, Corrosion Science, 10, 245-251. 169   Yi H., Du L., Wang G., Liu X. (2007), Influence of holding time after deformation on bainite transformation in niobium microalloyed steel, Journal of Iron and Steel Research International, 14, 62-65.   Zhang G., Cheng Y. (2009), Micro-electrochemical characterization and Mott-Schottky analysis of corrosion of welded X70 pipeline steel in carbonate/bicarbonate solution, Electrochimica Acta, 55, 316-324.  Zhou J., Li X., Du C., Pan Y., Li T., Li Q. (2011), Passivation process of X80 pipeline steel in bicarbonate solutions, International Journal of Minerals, Metallurgy, and Materials, 18, 178-185.    

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0135596/manifest

Comment

Related Items