Formation and Corrosion Inhibition Mechanisms of Chromate Conversion Coatings on Aluminum and Aa2024-T3

FORMATION AND CORROSION INHIBITION MECHANISMS OF CHROMATE CONVERSION COATINGS ON ALUMINUM AND AA2024-T3

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Wenping Zhang, M.S.

*****

The Ohio State University 2002

Dissertation Committee:

Dr. Rudolph G. Buchheit, Adviser

Dr. Gerald F. Frankel Approved By Dr. Henk Verweij

Dr. Richard L. McCreery ______Adviser

Dept. of Materials Science & Engineering

ABSTRACT

Chromate conversion coatings (CCCs) are applied to aluminum alloys to enhance their resistance to localized corrosion and to increase paint adhesion. However, chromate is toxic and suspected carcinogen. To develop environmentally friendly alternative coatings, a detailed and accurate understanding of CCC formation and breakdown is needed. Several studies on CCC formation and breakdown were conducted in this regard.

A first set of experiments was aimed at studying CCC formation and breakdown on 25-element Al electrode arrays. Results from coating formation experiments show that the coating process occurs in two stages. The first stage is characterized by intense electrochemical activity on the array and last from 20 to 30 seconds. The second stage occurs under electrochemical quiescence and little measurable current flows among elements in the electrode array. Raman spectroscopy shows that the coating continues to adsorb Cr6+. Anodic polarization of conversion coated arrays in chloride solutions led to several important findings. First, it was found that pitting potential increases as coating time increases through both stage one and stage two coating formation. Therefore, it can be concluded that CCCs do inhibit anodic reactions. These results also show that changes in coating structure and chemistry occur during the electrochemically quiescent second stage of coating formation. Further analysis showed that pitting potentials were higher on

ii electrode elements that were net cathodes during first stage CCC formation than on electrode elements that were net anodes. Related experiments were conducted by forming

CCCs on electrode arrays in conversion coating baths where the activating agent, NaF, and the accelerating agent K3Fe(CN)6 were withheld either individually or together.

Coatings formed in these modified solutions were then subject to anodic polarization in chloride solution. Results showed that these supplemental ingredients are essential to

CCC formation and contribute greatly to increasing the corrosion protection provided by the coating.

A second set of experiments was aimed at characterizing the effect of aging on

CCC structure and properties. CCCs are dynamic due to the fact that they continue to polymerize after they are removed from the coating bath. Using cathodic polarization experiments carried out in aerated chloride solutions, it was found that CCCs less than 48 hours old inhibited cathodic reactions. With increased aging time in ambient lab air, cathodic inhibition was lost. This loss of cathodic inhibition was attributed to coating dehydration and continued polymerization, which led in turn to the development of shrinkage cracking and loss in Cr6+ leachability. The relative humidity of the environment in which coatings aged was also found to have a significant effect on the

CCC aging process. After aging in dry air (RH ~ 0%), CCCs contained significant shrinkage cracking. Cr6+ release from such coatings was severely inhibited and corrosion resistance was poor in electrochemical testing. CCCs aged in ambient lab air (RH ~ 50%) exhibited less shrinkage cracking, a 2-order of magnitude increase in Cr6+ release, and considerably greater corrosion resistance. These results support the notion that CCC aging is strongly influenced by coating dehydration.

iii

Dedicated to my wife Qing and my daughter Sherry

ACKNOWLEDGMENTS

I am indebted to many people who have helped me during my five years of study and research at Fontana Corrosion Center. First and foremost, I would like to thank my adviser, Dr. Rudy Buchheit, for his excellent advice, inspiring encouragement, financial support, and his patience with me. His guidance has helped me to learn advanced knowledge in the field of corrosion and electrochemistry and to develop experience and skills to perform research in the same fields for my future career. I treasure this invaluable experience here and will always remember that I am from Fontana Corrosion

Center. I would also like to thank the members of my Dissertation Committee, Dr. Jerry

Frankel, Dr. Richard McCreery and Dr. Henk Verweij for their comments and suggestions on my research and dissertation. I also want to thank Dr. Akbar for his comments on my research at my dissertation overview.

My thanks also go to the staff in the Department of Materials Science and

Engineering for their help with instruments and facilities usage. They are: Mr. Gary

Dodge, Mr. Ken Kushner, Mr. Cameron Begg, Mr. Steve Bright, and Mr. Henk Colijin.

During the course of my study, I also received help from current or former FCC group

members, especially from: Dr. Jian Zhang, Dr. Donghui Lu, Dr. Valerie Laget, Dr.

Patrick Leblanc, and Dr. Christian Pagila. I would like to thank Ms. Belinda Hurley from

Department of Chemistry for helping me with Raman Spectroscopy. I have had many useful and intriguing discussions with Mr. Qingjiang Meng in my research.

Administrative help from Ms. Dena Bruedigam and Ms. Cindy Flores is also acknowledged.

VITA

October 10, 1968...... Born – Shandong, P.R. China

July, 1992 ...... B. S. Materials Science, Tsinghua University, Beijing, P.R. China

1992 - 1997...... Materials Engineer, Beijing Electric Power Research Institute, Beijing, China

1997 – present ...... Graduate Research Associate The Ohio State University

2000...... M.S. Materials Science and Engineering, The Ohio State University

PUBLICATIONS

1. W. Zhang, B. Hurley, and R. G. Buchheit, “Characterization of Chromate Conversion Coating Formation and Breakdown Using Electrode Arrays.” J. Electrochem. Soc., 149, B357, (2002).

2. W. Zhang and R. G. Buchheit, “Hydrotalcite coatings formed in oxidizing chemistries on 2024-T3.” Corrosion, 58, 591, (2002).

3. R. Leggatt, W. Zhang, R. G. Buchheit, and R. Taylor, “Performance of hydrotalcite conversion treatments used within a coating system on 2024-T3.” Corrosion, 58, 322, (2002).

4. R. Leggatt, W. Zhang, R. Taylor, and R. G. Buchheit, “Performance of field applied chromate conversion coatings for 2024-T3.” Corrosion, 58, 283, (2002).

vii

FIELDS OF STUDY

Major Field: Materials Science and Engineering

viii

TABLE OF CONTENTS

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita...... vii

List of Figures ...... xi

Chapters:

1. Introduction...... 1

2. Literature Review ...... 5 2.1 Introduction ...... 5 2.2. Processing and Formation Mechanisms of CCCs ...... 6 2.2.1. Chromating Processes...... 6 2.2.2. Coating Formation Mechanism ...... 8 2.2.3. Microstructural Heterogeneity in Aluminum Alloys and its Effects on Corrosion and Coating Formation ...... 15 2.3. Structure, Composition and Properties of CCC...... 19 2.3.1. Coating Morphology and Structure ...... 19 2.3.2. Composition of CCCs...... 20 2.3.3. Coating Corrosion Resistance...... 27 2.3.4. Paint Adhesion...... 29 2.3.5. Other Properties...... 30 2.4. Corrosion Protection Mechanisms of CCCs...... 31 2.4.1. CCC as a Barrier Layer ...... 32 2.4.2. Bipolar Membrane Mechanism ...... 32 2.4.3. Active Corrosion Protection Mechanism ...... 33 2.4.4. Chromate in Solution...... 37 2.5. Development of Chromate-Free Conversion Coatings ...... 40 2.5.1. Hydrotalcite Coatings ...... 41

2.5.2. Ce-Based Conversion Coatings ...... 43 2.5.3 Other Alternative Coating Systems ...... 44 2.6. Key Issues to be Addressed Regarding the Formation and Corrosion Inhibition Mechanisms of Chromate Conversion Coatings ...... 45

3. Characterization of Chromate Conversion Coating Formation And Breakdown Using Electrode Arrays ...... 72 3.1 Introduction ...... 72 3.2 Experimental Procedures...... 74 3.3 Results ...... 78 3.4 Discussion...... 85 3.5 Summary...... 94

4. The Effect Of Ambient Aging On Inhibition Of Oxygen Reduction By Chromate Conversion Coatings...... 124 4.1 Introduction ...... 124 4.2 Experimental Procedures...... 126 4.3 Results and Discussion ...... 127 4.4 Summary...... 133

5. Effect Of Ambient Temperature Aging In Dehumidified Air On The Properties Of Chromate Conversion Coatings (CCCs)...... 147 5.1 Introduction ...... 147 5.2 Experimental Procedures...... 149 5.3 Results ...... 152 5.4 Discussion...... 158 5.5 Summary...... 163

6. Conclusions, Relevance of Results, and Future Work...... 180

Bibliography...... 192

LIST OF FIGURES

Figure Page

3- 2.1 Proposed Fe(CN)6 mediation mechanism. Arrows represent redox cross reactions...... 58

2.2 Mechanism of coating growth...... 59

2.3 Microstructure of AA2024-T3 (a) SEM image of constituent particles. (b) TEM image of precipitates at matrix and grain boundaries...... 60

2.4 Mud-cracking morphology of CCC formed by a 3 min. immersion in Alodine 1200S on Al 7075...... 61

2.5 Depth profile of a CCC on Al. (a) Al, Fe, Cr, O (b) Al, N, C, F...... 62

2.6 A structure model of CCC on Al proposed by Treverton et al...... 63

2.7 Sputtering profiles measured by Auger electron spectroscopy in a fresh Alodine 1200S coating on Al 2024...... 64

2.8 Coating structure model of CCC on Al proposed by Townsend et al...... 65

2.9 Structure model of CCC on Al 2024 suggested by Hughes et al...... 66

2.10 Possible structure for the Cr(III)/Cr(VI) mixed oxide present in CCC. The bold line indicates an amorphous trivalent chromium hydroxide solid - 2- containing various amount of H2O and OH . Cr(VI) exist as Cr2O7 2- (shown) or CrO4 ...... 67

2.11 The differences in XANES between (a) Cr(III) standard (Cr2O3) and (b) Cr(VI) standard (K2CrO4)...... 68

2.12 Schematic of components that make up an aircraft coating system...... 69

2.13 Schematic drawing of artificial scratch cell. The larger piece is CCC coated metal panel and smaller one is bare metal...... 70

2.14 Polarization curves of Al2024-T3 in an aerated, oxygen stirred 1M NaCl base solution with dichromate...... 71

3.1 Photograph of electrode array used in this study ...... 100

3.2 (a) A frame from a current evolution movie during coating formation. (b) a movie from a 21-Al-4Cu array. (c) a movie from a 24Al-1-Cu array. (d) a movie from a 25-Al array...... 101

3.3 Representative current vs. time behavior for an electrode that exhibited distinct current oscillations during early CCC formation. In the coating solution notation, Cr, F and Fe stand for CrO3, NaF and K3Fe(CN)6, respectively. The same notations are used for the remaining figures...... 102

3.4 Representative current vs. time behavior for electrodes exhibiting persistent anodic behavior during early CCC formation...... 103

3.5 Representative current vs. time behavior for an electrode exhibiting persistent cathodic behavior during early CCC formation...... 104

3.6 The effect of F- on current evolution on Al during early CCC formation...... 105

3- 3.7 Effect of Fe(CN)6 on current evolution on Al during CCC formation...... 106

3.8 The current vs. time curve on one electrode segment during immersion at OCP in 0.5M NaCl when no Cu was present in the electrode array...... 107

3.9 Detail of the larger current spike in Figure 3.7 ...... 108

3.10 The current vs. time curve on one electrode segment during immersion at OCP in 0.5M NaCl when one Cu wire was included in the electrode array...... 109

3.11 A metastable pit that lasts 70s during immersion in 0.5M NaCl at OCP when a Cu wire was included in the electrode array...... 110

3.12 Typical anodic polarization curve on Al wire electrode coated with CCC in 0.5M NaCl solution...... 111

xii

3.13 Effect of coating time on CCC breakdown potential distribution. Coating solution: CrO3 + NaF + K3Fe(CN)6...... 112

3.14 Effect of polarity during coating formation on coating breakdown potential distribution. Coating solution: CrO3 + NaF + K3Fe(CN)6. Coating time: 3min...... 113

3.15 Relative contributions of coating bath components to coating breakdown potential. Coating time: 2min...... 114

3.16 Effect of K3Fe(CN)6 on coating breakdown potential distribution...... 115

3.17 Effect of NaF on coating breakdown potential...... 116

3.18 Effect of Fe(NO3)3 on coating breakdown potential distribution. Coating time: 2min...... 117

3.19 Comparison of breakdown behavior of CCCs formed in simulated Alodine and an Alodine 1200S solution prepared according to manufacturer’s specifications...... 118

3.20 The intensity of the 860cm-1 band in Raman spectra increases with coating 6+ time, indicating the build-up of Cr in CCCs. Coating solution CrO3 + NaF + K3Fe(CN)6...... 119

3.21 Normalized intensity of the 859cm-1 Cr3+-O-Cr6+ bonds as a function of coating time. A representative current versus time trace is shown for comparison...... 120

3.22 Polarization response of pure Al in simulated Alodine solution with and without F- additions...... 121

3.23 Morphology of coatings formed in different chemistries. (a: bare Al. b: in Cr + Fe. c: in Cr + Fe + F. Coating time: 2min...... 122

3.24 Anodic polarization curves on 1cm2 Al samples in 0.5M NaCl...... 123

4.1 Cathodic polarization curves for conversion coated 2024-T3 exposed to aerated 0.5 M NaCl solution. Aging time refers to time of exposure in laboratory air prior to collection of the polarization curve...... 137

4.2 Change in corrosion rates of CCC-coated samples with aging time...... 138

4.3 Cathodic polarization behavior of CCC-coated 2024-T3 in 0.1M Na2SO4 +

xiii

0.005M NaCl...... 139

4.4 An optical macrogaph of conversion coated 2024-T3 illustrating color changes due to Cr6+ leaching in solution. Samples were aged in ambient lab air for the times indicated, then subject to cathodic potentiodynamic polarization...... 140

4.5 Scanning electron micrograph of shrinkage cracking in a CCC on 2024-T3. ..141

4.6 CCC on 2024-T3 imaged with optical microscope after different aging times in ambient air...... 142

4.7 FIB cross section of a CCC on 2024-T3 illustrating the mophology of shrinkage cracks...... 143

4.8 An optical macrogaph of conversion coated 2024-T3 samples illustrating color changes due to Cr6+ leaching in solution. Samples were heat treated prior to cathodic polarization for the times and temperatures indicated, then subject to cathodic potentiodynamic polarization...... 144

4.9 Cathodic polarization curves for conversion coated 2024-T3 subject to elevated temperature exposure at the conditions indicated in the plot...... 145

4.10 Cathodic polarization curve of conversion coated 2024-T3 in 0.5M NaCl + -4 -6 10 or 10 M H2CrO4 solution. CCCs were aged in ambient air for various times as labeled in the plot before polarization test...... 146

5.1 SEM image of indent introduced with diamond pyramid indenter on CCC-coated 2024-T3. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl. Image collected after exposure to 0.5M NaCl for 192 hours...... 166

5.2 Optical images of CCC on 2024-T3 after aged in ambient air for different times. (a) 30 minutes. (b) 1 day (c) 4 days...... 167

5.3 Optical images of CCC on 2024-T3 after aged in dry air for different times. (a) 5 minutes. (b) 8 hours...... 168

5.4 Discoloration of CCCs on 2024-T3 aged in different environments after 20 days exposure to 0.5M NaCl solution. S/V=2cm-1. (a) Ambient air aged CCC. (b) Dry air aged CCC...... 169

5.5 Representative Raman spectra showing Cr6+ before and after exposure to NaCl on CCCs aged under different conditions. (a) Ambient air

xiv

aged CCC after exposed to 0.5M NaCl for 20 days. (b) Dry air aged CCC after exposed to 0.5M NaCl for 20 days...... 170

5.6 Cr6+ released from CCCs into surrounding 0.5M NaCl solutions after different aging treatments. S/V=2cm-1...... 171

5.7a Nyquist plots of CCCs after different treatments. Collected after 10 days of exposure to corresponding solutions. S/V = 2cm-1...... 172

5.7b Bode plots of CCCs after different aging treatments. Collected after 10 days of exposure to corresponding solutions. S/V = 2 cm-1...... 173

5.8a Schematic of cross section of CCCs aged in dry air and the equivalent circuit model used to fit the impedance spectra ...... 174

5.8b Experimental impedance spectra (dots) collected on dry air aged CCCs and the fitting results (solid lines) with the equivalent circuit model in Figure 5.8a...... 175

5.9 Pits observed on dry air aged CCCs after exposure to 0.5M NaCl. (a). SEM image of pit on CCC-coated 2024-T3 after 5 days’ exposure. (b) Optical images of pits on CCC-coated 2024-T3 after 10 days’ exposure...... 176

5.10 Corrosion resistance variation of CCC-coated 2024-T3 in damage resilience test. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl. S/V = 2cm-1...... 177

5.11 Corrosion resistance variation of CCC-coated 2024-T3 in damage resilience test. CCC aged in dry air for 1 day before exposed to 0.5M NaCl. S/V = 2cm-1...... 178

5.12 Corrosion resistance variation of CCC-coated 2024-T3 in damage resilience test. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl + 10-4M Cr6+...... 179

CHAPTER 1

INTRODUCTION

Aluminum alloys are widely used in aerospace and defense industries because of their relatively low densities and moderate strength. Aluminum and aluminum alloys are in their passive state in the pH range of 4-8.5 due to a naturally formed barrier oxide film on their surface. This oxide film is 2-3 nm thick and is strongly bonded to the substrate.

The oxide film can re-form immediately when damaged in most environments. However, when exposed to aggressive environments, aluminum and its alloys are susceptible to localized corrosion, such as pitting and crevice corrosion. Chromate conversion coatings are usually applied to aluminum alloys to enhance corrosion resistance and improve paint adhesion. Despite its efficiency and versatility, the use and emission of hexavalent chromium (Cr6+) has been under increasingly strict regulation because of its high toxicity and carcinogenic effect. It is believed the use of chromate will be completely banned eventually although the current trend is that its usage will be significantly reduced. This, along with other pressures, has led to the search for alternative protection methods that are environmentally friendly and yet efficient in corrosion protection. Numerous coating

systems have been developed to replace chromate conversion coatings (hereafter referred to as CCCs). However, none of them is able to match the performance and ease of application of CCCs. This has been partly attributed to the lack of deep understanding of the chromate conversion coating formation process and its corrosion inhibition mechanisms.

Therefore, the overall objectives of this study are to obtain a better understanding of chromate conversion coatings in terms of the coating formation process and corrosion inhibition mechanisms and to provide some guidance in the development chromate-free coating system. To reach the above objectives, the following aspects of chromate coatings have been addressed: the two stage chromate coating formation process on aluminum, the effects of coating bath chemistry on coating formation and breakdown, inhibition of anodic dissolution of Al by chromate coatings, inhibition of oxygen reduction by chromate coatings, and the effects of storage conditions on coating properties. The results of these studies are described and discussed in this thesis.

This dissertation consists of six chapters including this introduction. Each chapter focuses on one aspect of chromate coatings. Chapter 1 (current chapter) gives a short description of the objectives of this project and outlines the organization of this thesis.

Chapter 2 is a brief review of previous studies on chromate conversion coatings.

A rich body of literature has been developed by decades of studies on chromate coatings.

Reports on coating formation, coating morphology and structure, coating chemical composition, and corrosion inhibition mechanisms are reviewed and discussed. Some development of chromate free coatings is also covered. Through this review, several key

issues regarding chromate coatings are identified and proposed as the foci of the present study.

Chapters 3 through 5 cover the experimental results obtained from this study on chromate coating formation and corrosion inhibition. Chapter 3 focuses on the chromate coating formation process on Al and the effects of coating bath chemistry on coating formation process and coating breakdown behavior. A new electrochemical instrument called multichannel microelectrode analyzer (MMA) was utilized in this study to monitor current evolution during coating formation and during coating breakdown. Raman spectroscopy was employed to characterize Cr6+ in chromate coatings. The roles of minor components in the coating bath in the formation of corrosion resistant coatings were emphasized in this study. The inhibition of Al anodic dissolution by chromate coatings is also examined.

Chapter 4 details a study on the inhibition of oxygen reduction by chromate coatings. Cathodic polarization experiments were performed on chromate coatings after aging for different periods. The coating morphology was examined with optical microscopy and focus ion beam (FIB) cross sectioning. The results from this study clarified confusion over whether chromate coatings inhibit oxygen reduction.

Chapter 5 describes a study on the effects of aging in dry air on chromate coating properties. Chromate coatings were aged in ambient air or in desiccator to examine the differences in aging behavior in two different environments. The coating morphology evolution during aging was examined with optical microscopy. Coating corrosion resistance was assessed with electrochemical impedance spectroscopy (EIS) and

exposure to chloride solution. Cr6+ release from CCC into surrounding solution was examined with inductively coupled plasma optical emission spectroscopy (ICP-OES).

The results showed significant effects of aging in dry air on coating corrosion resistance.

Chapter 6 summarizes the results from chapters 3-5 and some conclusions were drawn regarding coating formation, the effects of minor ingredients in coating bath on coating formation and breakdown, the anodic inhibition and cathodic inhibition by chromate coatings, and the effects of dry air aging on coating properties. A better understanding on the coating formation and corrosion inhibition mechanisms is obtained.

Some suggestions are provided for the future study of chromate coatings and the development of chromate-free coating systems.

CHAPTER 2

LITERATURE REVIEW

2.1 Introduction

It has been more than ninety years since the first use of CCC on aluminum alloys.

Significant progress in the understanding of the properties and corrosion inhibition mechanisms of CCCs has been made in the most recent twenty years or so, although some issues still remain unresolved. In this review, several aspects of CCCs are examined based on the literature published over the last forty years. First, the coating processes and coating formation mechanisms are discussed. The roles played by fluoride ion and ferricyanide in coating formation are reviewed. Complications in coating formation associated with microstructural heterogeneity of aluminum alloys are discussed.

Secondly, the properties of CCCs are described. The morphology, structure, composition and their relationships with corrosion protection, paint adhesion and other properties are discussed with emphasis on the coating composition because it is critical to corrosion protection. Thirdly, the proposed mechanisms of corrosion protection by CCCs are reviewed. The active corrosion protection mechanism is discussed in detail. Mounting evidence suggests that this unique property contributes significantly to the corrosion

protection of CCC coatings. The development of replacement coating systems is discussed in the fourth section. Some promising coating systems including hydrotalcite coatings are described in this chapter. Several key issues in the transition from chromate to chromate-free coatings are discussed in the fifth section. Future work and approaches are proposed in the final part of this chapter. Most of the information in this review is related to CCC coatings on aluminum and its alloys although some work about CCC coatings on zinc is also cited.

2.2 Processing and Formation Mechanisms of CCC Coatings

2.2.1 Chromating Processes

A conversion coating, by definition, is a thin film that results from the reactions between a metal surface and a suitable chemical solution during simple immersion. The metal surface is converted from an originally active condition to an inert film, hence the name conversion coating 1.

Chromate coatings are applied by contacting the metal surface with solutions containing hexavalent chromium (Cr6+) and other components. Before this step the metal parts are usually cleaned to remove contaminants and oxide film from the metal surface.

Alkaline cleaners are used to remove grease, oil and soils. Acidic deoxidizers are used to reduce the oxide film and to activate the surface for chromating 2. These pretreatment steps can have significant effects on the properties of CCCs 3,4, but the chromating step is most important in producing CCCs with desirable properties. During this step, several factors are critical in controlling the coating properties. These factors include time,

temperature, chromate concentration, pH, and accelerator concentration 2. The final step is drying. Drying is also important to the coating performance because it affects the chromate release characteristics, which will be discussed in the section on active corrosion protection.

The composition of the chromating solution depends on the metal to be treated, the application method and the requirements for the final product. Components in the coating bath can be classified into three categories: attacking agent, inhibiting agent and film forming agent 5,6. According to the nature of attacking agent used, the coating solutions are of two kinds: alkaline and acidic 7. An example of an alkaline-type coating process is found in the Modified Bauer-Vogel (MBV) Process, in which a solution containing sodium carbonate and sodium chromate is used. Acidic coating solution contains chromates, fluorides and either phosphates (in chromate-phosphate coating process) or ferricyanides (in accelerated chromate coating process). During coating formation, fluorides act as activators and ferricyanides act as accelerators. When both phosphates and ferricynanides are absent, the coating formation is very slow and the process is not of technological significance.

CCCs are mostly applied by immersion or spraying methods. Other methods of application such as brushing, coil coating, and squeegee can also be used. The simplicity, low cost, high efficiency of corrosion protection and versatility of application method has made CCCs popular and widely used in many industries. The main fields of application for CCCs on aluminum alloys include aircraft and aerospace structural components, coil extrusions, heat exchanger parts, and containers 3.

2.2.2 Coating Formation Mechanism

Historically, the formation of CCC has been described as a process resulting from a redox reaction between chromate ions (or dichromate ions) in solution and aluminum.

- 2- At low pH ( pH<=2 ) chromate exists as either HCrO4 or Cr2O7 . Both of these two ions are strong oxidizing agents and have high reduction potentials 8:

- + - 3+ HCrO4 + 7H + 3e → Cr + 4H2O Eqn. 2. 1

E0 = +1.35V (vs. SHE)

2- + - 3+ Cr2O7 + 14H + 6e → 2Cr + 7H2O Eqn. 2. 2

E0 = +1.23V (vs. SHE)

In the presence of these oxidizers, aluminum is oxidized while hexavalent chromium is reduced to trivalent chromium. Other reduction reactions such as hydrogen evolution and/or oxygen reduction may also occur:

+ - + - 2H + 2e → H2 and/or O2 + 4H + 4e → 2H2O Eqn. 2.3

All these reactions consume protons and lead to an increase in pH near the metal surface.

This pH increase results in the formation of an amorphous mixture of hydrated aluminum and chromium oxide 3. This reaction has been suggested to occur as follows 9, 10:

2- 0 + Cr2O7 + 2Al + 2H + H2O → 2CrOOH↓ + 2AlOOH↓ Eqn. 2.4

However, this reaction is not consistent with the fact that little aluminum is found in

CCC. Other reactions paths have also been suggested, for example11, 12:

0 2- + 3+ 2 Al + Cr2O7 + 8H → 2Cr(OH)3 + H2O + 2Al Eqn. 2.5

+ - 0 3+ 8H + 2HCrO4 + 2Al → 2Al + Cr2O3⋅3H2O + 2H2O Eqn. 2.6

The presence of fluoride ion (F-) is important for coating formation. Film growth is very slow in the absence of F-. Several roles played by F- in coating formation have been suggested 8,9. First, it has been suggested that F- dissolves the oxide film initially present on the aluminum surface and activates the surface for the deposition reactions to proceed.

This etching process occurs through this following reaction:

Al2O3 + 6HF → AlF3 (soluble) + 3H2O Eqn. 2.7

Secondly, it has been suggested that F- can dissolve a portion of the growing film, which would allow the solution to penetrate the film and make contact with metal surface and prevent the surface from complete passivation. These reactions are supposed to follow the equations below:

AlOOH + 3HF → AlF3 (soluble) + 2H2O Eqn. 2.8

CrOOH + 3HF → CrF3 (soluble) + 2H2O Eqn. 2.9

Fluoride ion has been described as a unique monodentate ligand that enhances the dissolution of aluminum oxide. Finally it has been observed that the incorporation of F- in the oxide film on pure aluminum can increase both the ionic and electronic conductivity, which might increase the rates of the electrode reactions. The F- incorporated in CCC coating during coating formation is suggested to have a similar function 9.

Ferricyanide in the coating solution has been characterized as an “accelerator”.

When ferricyanide is added to the solution, the coating corrosion resistance is greatly improved; the coating weight, coating thickness and coating formation rate are increased

1, 3, 8, 10, 13. Some mechanisms have been proposed regarding how ferricyanide affects coating formation. Treverton et al. 12 suggested that an absorbed monolayer of chromium ferricyanide might accumulate on the precipitated chromium oxide particles during coating formation. This monolayer of absorbed ferricyanide on the particle surface will prevent or reduce the absorption of chromates on the particle surface. Thus more chromates will be available for the redox reaction and the reaction rate is increased.

However, this mechanism lacks strong evidence. Hagans and Haas 8 proposed a mechanism for the acceleration of CCC formation on Al-Cu alloy 2024-T3. AES analysis of the coating composition found that ferricyanide (as monitored by nitrogen) was present over the entire surface of the CCC coating. However, larger concentrations were localized in the coatings formed on the Cu-containing intermetallic phases, which indicated that ferricyanide was preferentially deposited onto these regions. Moreover, in

the depth profiles nitrogen was distributed only in the top 200 Å of the coating on the matrix (total film thickness 1000Å) while N was found through the entire coating thickness on the intermetallics. This suggests that some ferricyanide might have interacted with Cu and formed insoluble Cu4Fe(CN)6 or Cu2Fe(CN)6. It was claimed that the benefit of this reaction was that it reduced the magnitude of galvanic couple that existed between the matrix and the intermetallics. A recent effort trying to clarify the function of ferricyanide in CCC formation has been made by Xia et al. 14 A redox mediation mechanism was proposed, as illustrated in Figure 2.1. It was observed that

6+ 3+ 3- direct reduction of Cr to Cr was very slow. However, during coating, Fe(CN)6

4- 6+ rapidly oxidizes 2024-T3 and Fe(CN)6 rapidly reduces Cr under conditions similar to

3- those in a coating bath. Thus the redox mediation action by Fe(CN)6 appears possible.

3- 3+ 3- During coating formation, Fe(CN)6 first oxidizes Al to Al while Fe(CN)6 is reduced

4- 4- 3- 6+ 6+ to Fe(CN)6 . Then Fe(CN)6 is oxidized back to Fe(CN)6 by Cr while Cr is reduced to Cr3+. The result is Cr6+ reduction to Cr3+ and Al oxidation to Al3+. CCC formation is

3- greatly accelerated through the mediation of Fe(CN)6 . Xia et al. suggested that, in principle, any redox system with a redox potential between that of Cr6+/Cr3+ and Al/Al3+ and also with fast redox kinetics with these two systems can be a mediator for CCC

3-/2- 3+/2+ 3+/2+ formation. Based on this idea, they tested IrCl6 , Fe and V as mediators. They found that they all accelerated the CCC formation and the coatings formed had similar properties as conventional coatings, which further substantiated their claim that ferricyanide acted as a mediator.

The discussions above addressed the chemical aspects of coating formation.

Several physical models on coating formation and growth have also been developed and will be discussed below.

Katzman et al. 9 proposed a model of coating growth based on their AES and XPS study of CCC coating on commercially pure aluminum. At the onset of coating formation, the initial oxide film is removed by hydrogen fluoride. Then the exposed aluminum reacts with dichromate resulting in the precipitation of hydrated aluminum and chromium oxide. The film starts to grow. The AlOOH in the mixed oxide dissolves and leaves the less soluble CrOOH on the surface. This process continues, as the film grows outward from the aluminum surface. Aluminum and aluminum oxide are dissolved wherever they are in contact with solution and are replaced by amorphous hydrated chromium oxide. As coating continues to grow, more aluminum is oxidized at the aluminum-coating interface by the oxidizing action of the chromate ions absorbed on the coating surface. Al3+ formed by oxidation diffuses through the coating to the surface and dissolves into the solution. This model is illustrated in Figure 2.2. Due to the dissolution of coating by hydrogen fluoride, the coating cannot grow indefinitely. When film formation and dissolution reach equilibrium, the coating will cease growing. Another factor limiting film growth is that as coating thickens to a certain value, the oxidizing power of absorbed chromate is no longer sensed at the metal surface and aluminum oxidization stops. Some chromate ions are strongly absorbed on the coating surface and remain unreduced when the coating step finishes. This model cannot explain layered

particulate feature observed in the CCC coating structure and is in conflict with the fact that there is little aluminum found in the coating.

While Katzman et al. assumed a uniform film growth during coating formation,

Brown et al. have proposed a different model for coating growth based on their extensive electron microscopy studies 15-20. They assumed that the removal of initial oxide film on aluminum surface by F- is incomplete. For aluminum of low purity they suggest that the aluminum surface is heterogeneous due to the segregation of impurities on grain and cellular boundaries. The regions where these boundaries intersect the surface constitute minor heterogeneities that are potential sites for flaw formation. The remaining oxide film after hydrogen fluoride attack at these flaw sites provides easy paths for electronic conduction while the oxide film on the general Al surface is relatively insulating. During coating formation, these sites become the preferential sites for the reduction of chromate to hydrated chromium oxide while anodic reaction occurs over the aluminum surface in general. The chromium oxide deposits grow laterally and perpendicularly and finally cover the aluminum surface. For aluminum of high purity (>99.9995%), the impurity segregation becomes insignificant and the aluminum surface is generally homogeneous.

In this case, the coating growth will proceed over the general aluminum surface by tunneling of electrons through the oxide film. The fluoride attack to the initial oxide film is not uniform. At the sites where the film is thinner, the reduction of chromate occurs and results in the deposition of chromium oxide. The areas where the oxide film thickness is greater are not covered with chromium deposit. During subsequent coating

growth, some of these uncovered areas will remain as fixed and persistent anodic sites while deposit-covered areas act as cathodic sites. The final result is a relatively uniform conversion coating with some holes extending from the outer surface to the locally corroding aluminum substrate. Brown does not explain how the tunneling of electrons was still possible after thick deposit layer of chromium oxide had formed while electron tunneling was not possible at the persistent anodic sites. Recent examination 21of cross sections of CCCs did not show the existence of holes associated with anodic sites as proposed by Brown .

Both the Katzman’s and Brown’s coating formation models fail to explain how layered and particulate morphologies develop on Al and aluminum alloys. The layered and particulate morphology has been frequently reported in the literature 22-27.

Arrowsmith et al. 23 have proposed a coating formation model that describes how a layered particulate structure might develop. Initially, the oxide film is dissolved and anodic dissolution of aluminum occurs. Reduction of chromate proceeds at cathodic sites to form strands of hydrated chromium oxide and then the strands grow into spherical particles. These particles grow and merge to form a monolayer of particulates with inter particle gaps. These gaps serve as paths for transport of fresh solution to the metal for continued reaction. This model does not explain how later layers form after the Al surface is covered with the first layer of particles.

Recently, CCC formation has been described from a sol-gel perspective 28, 29. In this model, Cr6+ is first reduced to Cr3+. Cr3+ then hydrolyzes and condenses to form Cr3+ monomers, dimers, trimers and tetramers. With the increase in pH near the substrate

surface region, these species polymerize and precipitate onto the substrate to form a hydrated oxide film. The precipitated Cr3+ oxide forms the “backbone” of CCCs. Cr6+ species in solution then are adsorbed onto the chromium oxide backbone. This model

- 3- does not address the presence of F or Fe(CN)6 in the coating.

These models all agree that the reduction of chromate to form hydrated chromium oxide is the key reaction in the coating formation. They differ in how coating formation proceeds. Arrowsmith’s model is more consistent with experimental observations than the other two models but does not address the Cr6+ in CCCs. The sol-gel model successfully explained why both Cr3+ and Cr6+ exist in the coating and is probably the most accurate coating formation model so far.

2.2.3 Microstructural Heterogeneity in Aluminum Alloys and the

Effects on Corrosion and Coating Formation

Although pure aluminum has good corrosion resistance, it is of limited use in structural applications because of its low strength. Alloying elements such as Cu, Mg,

Mn, Zn, Si, and Li are added to aluminum to increase its strength. Besides these alloying elements, some elements also are present as impurities, for example, Fe and Si. These elements exist either in solid solution and/or in second-phase particles. These second- phase particles constitute the heterogeneity in microstructure and greatly affect the mechanical, corrosion and other properties of aluminum alloys. AA2024-T3 is used as an example to illustrate how second phase particles affect the corrosion behavior of

aluminum alloys and how microstructural heterogeneity affects CCC formation. on aluminum alloys.

AA2024-T3 is a heat treatable alloy. The concentration ranges of the main alloying elements in AA2024 are 3.8-4.9% Cu, 1.2-1.8% Mg, 0.3-0.9% Mn, with balance

Al 30. It also contains 0.5% Fe and Si. A range of different studies has shown that 2024-

T3 has a complex, multiphase structure. Buchheit et al. 33 have characterized the second phase particle distribution by size and chemistry. In this study, it was shown that the second phase particles cover about 4.2% of the alloy surface and S phase (Al2CuMg) is the predominant type of particle, as it covers 2.7% of the alloy surface and constitutes

60% of the total particle population. Chen et al. 34 and Cawley et al. 35 have found that the average population density of particles with a projected surface area > 1µm2 is around

300,000/cm2. The population of particles can be divided into two categories: constituent particles and precipitates. Constituent particles include (Mn,Fe)3SiAl12, Al2CuMg (S phase), reaction product during solution heat-treatment Cu2FeAl7, and fine dispersoid

31 Cu2Mn3Al20 . Precipitates includes S phase and intermediate precipitation products, S’

(Al2CuMg). These rod-shaped precipitates are distributed both in matrix and at grain boundaries. Due to the precipitation of Cu-containing particles at grain boundaries, a Cu depleted zone may be observed at grain boundaries. Figures 2.3a and 2.3b show the constituent particles and precipitates in AA 2024-T3 32.

Localized corrosion of 2024-T3 has been investigated in many studies 33--41.The heterogeneity in its microstructure, especially the Cu-containing particles, contributes significantly to the corrosion susceptibility and corrosion protection ineffectiveness.

These second-phase particles have different electrochemical activities from pure aluminum and aluminum solid solution and the activities are also different among different types of particles. Buchheit 36 has compiled the corrosion potentials of some Al- based intermetallics in different environments. These second phase particles are expected to be either anodic or cathodic to the aluminum matrix. For example, some Al-Cu-

(Fe,Mn) particles are cathodic (more noble) to matrix, while Al2CuMg is initially anodic

(more active) to aluminum matrix, but becomes cathodic with increasing exposure time in sodium chloride solution due to dealloying. Small galvanic couples form between second phase particles and matrix across the surface due to their different activities, which has been used to explain the accelerated corrosion of aluminum alloys, pitting corrosion, intergranular corrosion 36. Studies also have revealed that localized corrosion associated with intermetallics is more complicated than the simple galvanic cell model.

For example, Al2CuMg is anodic to the aluminum matrix, however, the matrix at the periphery of these particles has been found dissolved. Two explanations have been given

33, 34, 39 in literature. One is the dealloying of Al2CuMg . It was suggested that preferential dissolution of Mg from Al2CuMg leaves behind Cu-rich remnants, which are cathodic to the matrix. The Cu-rich remnants can result in two types of pits: pits in the matrix at the periphery of Al2CuMg particles and pits left by decomposition of Cu-rich remnants. Cu clusters formed by decomposition of Cu-rich remnants can detach from the alloy leading to non-Faradaic Cu dissolution. Guillaumin et al. 39 also give another explanation. They found that around coarse Al2CuMg particles there exists a dispersoid-free zone. This zone is depleted in Cu, thus it is anodic to both Al2CuMg and aluminum matrix. This results in

pitting of the matrix around Al2CuMg particles. Al-Cu-(Fe, Mn) type particles also show complicated pitting behavior. Matrix trenching can occur around these particles because they are cathodic to the matrix. However these particles also exhibit nonuniform dissolution within themselves because they are heterogeneous in nature 37, 38.

Some efforts have been made to understand the effects of second phase particles on the coating formation. Hagans and Haas. 8 observed that during coating formation, the coating growth rate is lower on intermetallics than on the solid solution matrix for the first 3 minutes of coating formation. The coating growth rate increases in the following order: (Fe,Mn)3SiAl12 < Al2CuMg < aluminum matrix. They observed that after five minutes, the coating thickness on matrix and intermetallics was almost the same. This observation does not appear to be consistent with recent results, which show that the coatings on large intermetallic particles are much thinner than on matrix even after 5 minutes of immersion 21, 42. The coating composition is also different on the matrix and on intermetallics. Ferricyanide is present through the entire coating thickness on intermetallics but only present in surface region of the coating on matrix. Waldrop et al.

27 studied the nucleation of CCC coating on 2024-T3 by AFM. They found that nucleation rates on the matrix and two types of intermetallics are different. The film deposition rate on Cu-Mg-Fe-Al type of intermetallics (cathodic to matrix) is much faster than on Al matrix while the film deposition rate on Al2CuMg (anodic to matrix) is considerably slower than on matrix. The latter result is in direct conflict with the observations of Hagans and Haas. In a study on a chromate-free Ce-Mo-based conversion coating 43, it has been found that the coating morphology on particles is different than that

on the matrix. Additionally, the distribution of Ce and Mo is not uniform. It is high on

Fe-Mn-Cu -containing intermetallics and low on matrix and Al2CuMg particles. A more recent study has concluded that the Cr concentration and distribution on intermetallics is different from that on aluminum matrix. The amount of mixed Cr3+/Cr6+ oxide is inversely proportional to the Cu content of intermetallics 44. A model has been proposed to explain why CCCs on intermetallics are thinner than matrix. In this model, the CN- group of ferricyanide is strongly adsorbed onto Cu-containing intermetallics and inhibits further reduction of ferricyanide, thus reducing the redox mediation function of ferricyanide and resulting in a thinner coating. However, this model cannot explain why normal CCCs are formed on smaller intermetallic particles, regardless of particle type 21.

2.3 Structure, Composition and Properties of CCC

In the last 20 years, CCCs have been studied by various surface sensitive characterization techniques. These techniques include: X-ray Photoelectron Spectroscopy

(XPS), Auger Electron Spectroscopy (AES), Fourier Transformation Infrared (FTIR)

Spectroscopy, Secondary Ion Mass Spectroscopy (SIMS), Raman Spectroscopy, X-ray

Absorption Near Edge Structure (XANES), Extended X-ray Absorption Fine Structure

(EXAFS), Atomic Force Microscopy (AFM). More conventional tools such as Scanning

Electron Microscopy (SEM), Energy Dispersive Spectroscopy (EDS), Transmission

Electron Microscopy (TEM), X-ray Diffraction (XRD), Thermogravimetic Analysis

(TGA) have also been applied in the investigation of CCCs. Characterization of CCCs

with these tools has greatly enhanced the understanding of the morphology, structure, and composition of CCCs.

2.3.1 Coating Morphology and Structure

CCCs are amorphous thin films. The amorphous nature has been confirmed by

XRD and electron diffraction 45, 46. When observed under SEM, the CCCs show a mud- cracked pattern 46, 47, like that shown in Figure 2.4. The appearance suggests that the cracking is due to shrinkage. High resolution SEM and TEM have revealed that the coating body has a multi-layer structure 22-24. The thickness of a layer is between 30 nm to 50 nm 22. Each layer comprises of ellipsoid-shaped particles with diameter ranging from 30 nm to 50 nm 22. The morphology and structure of CCC coatings play important roles in the promotion of paint adhesion 22, 24.

2.3.2 Composition of CCC coatings

The composition of CCC coatings has been investigated thoroughly because it is important for a full understanding of the corrosion inhibition mechanism. Coating compositions depend on substrate, cleaning procedures, coating bath composition, coating processing parameters, and aging conditions. Although it is generally accepted that the coating is composed of complex chromium compounds, some components of the coating bath, and elements from the substrate 2, opinions differ on the types of compounds, the relative concentration of different components and distribution of the elements in the coatings.

Coating Composition

Many compounds have been proposed to exist in CCCs, including Cr2O3,

Cr2O3⋅xH2O, Cr(OH)3, CrOOH, Cr(OH)CrO4, Crx(CrO4)y, CrF3, CrFe(CN)6, AlOOH,

12 Al2O3, AlF3. Treverton et al. studied CCC coatings formed on aluminum using a bath containing chromic, hydrofluoric acids and potassium ferricyanide. Based on XPS data and elemental depth profiling (Figure 2.5), a coating structure model was proposed. As shown in Figure 2.6, the surface layer is composed of Cr2O3⋅xH2O and CrFe(CN)6. The bulk of the film is Cr2O3⋅xH2O. At the interface between the coating and aluminum are

6+ AlF3, AlOF, and Al2O3. No Cr was detected in their experiments. The authors attributed this absence of Cr6+ to aging in air during which the Cr6+ species react with metal and form Cr3+ species. It is worthy to note that F- reached its peak concentration at the interface region between coating and aluminum substrate. In a later study by the same authors 13, Cr6+ was present at 15 at. % of total measured chromium content. CCC coatings formed on aluminum in CrO3 + Na2Cr2O7 + NaF bath was investigated by

Katzman et al. 9 A film consisting of a layer of chromium oxide over a layer of mixed chromium-aluminum oxide was suggested based on their findings. Cr6+ was detected in the top 120Å layer and the ratio of Cr(VI)/Cr(III) was about 1/2. Asami et al. 48 studied the composition of the near surface regions of conversion-coated aluminum using a bath chemistry similar to Katzman’s. It was found that the surface regions are composed of

6+ hydrated chromium oxyhydroxide (CrOx(OH)y⋅nH2O) and a small amount of Cr species. However, they found no aluminum in the near surface region, which is in contrast with Katzman’s finding of 3 at% Al. Lytle et al. 47 investigated the coating

formed on 2024-T3 in Alodine 1200S bath. Their results were similar to those of

Katzman. Figure 2.7 shows the depth profiles of some elements in the coating. They found that the Cr6+ fraction in the top 1000 Å is about 23%, with the balance as Cr3+. Al and other alloying elements were not components of the top 20 Å layer, which suggested that the intermetallic compounds of the original alloy were totally covered by CCC.

Figure 2.8 illustrates their proposed coating model. The profile after salt spray test shown

45 in this figure will be discussed later. A study by Yu et al. on coatings formed in CrO3 +

NaF + K3Fe(CN)6 on aluminum concluded that the coatings contained mainly γ-

- 3- 6+ CrOOH⋅nH2O. F and Fe(CN)6 were present throughout the coating. Cr species were absorbed on the outer surface layer of the coatings. However, the amorphous nature of

49 CCC calls into question the existence of γ-CrOOH⋅nH2O. Townsend et al. proposed a model for coating on aluminum, as shown in Figure 2.7. Their measurement suggested the presence of Cr0 at the coating-substrate interface. But this observation of metallic Cr layer at the interface is now believed to be an artifact caused by ion beam damage 50. The influence of second phase particles on CCC composition was studied by Hagans and

Haas 8. CCCs formed on 2024-T3 in two bath chemistries were investigated. Depth profiles of coatings formed after immersion in CrO3 + NaF solution for 3 minutes showed differences between the coatings formed on solid solution matrix and on intermetallics. In coatings on the matrix, Cr and O are present through the entire coating thickness and Al was not present at surface but was found in the rest of the coating. An enriched Cu layer that was present on the metal surface before immersion was intact after coating formation. In coatings on Cu-containing intermetallics, appreciable amounts of Cu and Al

still existed at surface, indicating a slow covering by the CCC. XPS analysis showed that

6+ 60% of Cr at the surface region was Cr for the coatings formed in NaF + K3Fe(CN)6 +

CrO3 solution. Another study of CCC coating on 2024-T3 was carried out by Hughes et al. 10. Based on XPS, SEM and EDS analyses, a coating model as shown in Figure 2.9 was proposed. The external surface layer consists of hydrated chromium oxide, chromium fluorides and possibly ferricyanide and a small amount of chromate. The bulk of the coating was composed of chromium oxide, chromium oxyhydroxide, chromium fluoride and chromium ferrocyanide. The interface region between coating and substrate contained alumina, chromium and aluminum oxyfluorides, and copper. This model is largely in agreement with Treverton’s model but with some modifications. Xia et al. 51 compared the CCC (formed in Alodine 1200S solution) and a synthetic mixed oxide of

3+ 6+ Cr /Cr made with Cr(NO3)3, K2Cr2O7 and NaOH. The CCC and Cr mixed oxide showed a similar Cr3+/Cr6+ ratio and similar shifts in Raman frequencies when pH and temperature were varied, indicating the CCC contains a material similar to the synthesized Cr mixed oxide. Polymeric chromium hydroxide has been suggested to be a component of the freshly formed CCC. Xia et al. suggested that the binding between

- chromium hydroxide and chromate is covalent. They proposed that the OH or H2O in the

2- 2- chromium hydroxide polymer was substituted with Cr2O7 or CrO4 and a bond of

Cr(III)-O-Cr(VI) was formed. A possible structure for this arrangement is shown in

Figure 2.10. Also shown is the proposed hydrolysis of this mixed oxide through which H+ and chromate ions are released.

From the above discussions it can be concluded that the main components of

CCCs are Cr3+-containing compounds and Cr6+-containing compounds, with Cr6+ species present at the external surface and Cr3+ species making up bulk of the coating. Some

- 3-/4- components from the coating bath such as F and Fe(CN)6 are present in the coating as minor components. The coatings generally contain some elements from the substrate.

Water is also an important component, mainly present in hydrated chromium compounds.

Due to the complexity in the interpretation of XPS results, the local chemical environments of coating constituents cannot be unequivocally determined. It is also possible that there is variability in composition from one coating to the next. Therefore different interpretations lead to different coating structural models. Because the coating is highly hydrated and amorphous, it is also possible that the composition may not have a discrete stoichiometry 8.

Accurate Determination of Cr(VI) in CCC Coatings

As will be discussed later, hexavalent chromium plays an important role in the corrosion protection of CCC coating. Accurate measurement of Cr(VI) is essential. XPS measurements have been widely used in the determination of Cr(VI) contents since XPS can reveal the oxidation state of an element. However the photoreduction of Cr(VI) species in the ultrahigh vacuum (UHV) environments during XPS measurement has been a problem in the accurate determination of Cr(VI). The photoreduction of Cr(VI) species in CCCs was first mentioned by Asami et al. 48, who noticed that there was a color change after the CCC was exposed to XPS analysis. Another phenomenon that shows

photoreduction of Cr(VI) species is that Cr(VI) content decreases with time when coating is exposed to MgKα12 radiation. Four hundred minutes of exposure results in a nearly complete reduction of Cr(VI) species 50. This photoreduction of Cr(VI) species has been characterized by Halada et al.50 and Schrott et al.52.

To overcome the limitations of XPS, such as photoreduction of Cr6+ and the effect of high vacuum environment, X-ray Absorption Spectroscopy (XAS) has been used. This technique can be performed at ambient pressure and the coatings are less susceptible to photoreduction by higher energy photons. The X-ray absorption spectrum can be roughly divided into two parts 53: extended x-ray absorption fine structure (EXAFS), which refers to the fine structure in the spectrum starting past an absorption edge and extending typically 1000eV further, and x-ray absorption near-edge structure (XANES), which covers the range near the threshold for absorption. Both have found applications in the study of chromate related coatings. Figure 2.11 (b) shows a complete X-ray absorption spectrum of Cr6+ 54.

There are two differences in the XANES of Cr(III) and Cr(VI). First the absorption edge of Cr(VI) is shifted several eV relative to Cr(III) edge. Second, there is a distinct pre-edge peak in the Cr(VI) spectrum while no such peak shows up in Cr(III) spectrum 11, 54. This pre-edge peak is ascribed to a bound state transition which is allowed in Cr(VI) but not allowed in Cr(III). The XANES of Cr(VI) and Cr(III) are shown in

Figure 2.11. Linear combination of spectra from Cr(VI) and Cr(III) oxides allows a quantitative determination of the ratio of Cr(VI) to Cr(III) or Cr(VI) to total Cr. XANES has been used to study the oxidation state of chromium in CCC coating on aluminum

alloys, in anodic films on aluminum formed in chromate electrolyte, and anodized films on aluminum grown in non-chromate solution and then sealed in chromate solution.

Hawkins et al. 54, 55 studied the interactions between chromate and aluminum supporting oxide films of different thickness using XANES. They found that the Cr(VI) in the films after immersion in chromate/chloride solution strongly depended on the thickness of initial oxide film. The films formed on samples initially supporting air-formed oxide

(~2.5 nm) contain only Cr(III) after immersion in chromate solution. The films on samples with initial anodized layer of 6 – 36 nm contained both Cr(VI) and Cr(III) after immersion in chromate solution, with thicker initial oxide containing more Cr(VI).

Wainright et al. 56 found that only Cr(VI) was incorporated into the films when samples initially covered by a 20 µm anodic layer were immersed in a hot dichromate solution.

Anodic films formed by anodizing aluminum in chromate solution also contain Cr(VI) and Cr(III). Cr(VI) is largely associated with the outermost film regions 57. Results of

Cr(VI) fraction in CCC coatings measured by XANES have been reported by several authors. Kendig et al. 11 found that Cr(VI) in CCC coatings formed on 2024-T3 after 5 min immersion in Alodine 1200S solution was roughly 20% of total chromium content.

They also found that this ratio increased with time of immersion up to 5 min and then reached a limit at 20%. Lytle et al. 47 also investigated the CCC coatings formed on 2024-

T3 in Alodine 1200S solution. The ratio of Cr(VI) to total chromium was found to be unchanged at 23±2% for Alodine treatment time 1 min and 3 min. A study by Wan et al.

58 showed that 30% of the total chromium is Cr(VI) in the CCC coatings formed on pure aluminum in solution containing CrO3 + Na2Cr2O7 + NaF.

In the EXAFS range, there are some oscillations in the absorption spectrum. By analyzing this part of the spectrum, one can get the information about the types and numbers of atoms surrounding the absorber atom. EXFAS probes the environment that is within a 6Å range around the absorbing atom 59. EXAFS has been used to study the coordination environment of chromium in CCCs and in chromate-sealed anodized films.

It has been determined that the Cr(VI) in a CCC is in tetrahedral coordination and Cr(III) is in octahedral coordination 47, 56 and the Cr-O bond distances of Cr(VI) and Cr(III) are

1.71±0.03Å and 1.99±0.01Å, respectively 47.

To summarize the discussions about Cr(VI) content in CCC coatings: Cr(VI) species have been found present in CCCs. XPS and XANES analysis show that the ratio of Cr(VI) to total Cr range from 6% to 60%, with a typical value of 30%. The Cr6+ component of CCCs is important for the self-healing function. This will be discussed later.

2.3.3 Corrosion Resistance

CCCs provide excellent corrosion protection to aluminum and its alloys, zinc, cadmium, and magnesium and its alloys. For example, AA2024-T3, which is normally poorly resistant, coated with CCCs can easily withstand 336 hours of salt spray test in accordance with ASTM B117-97 60 and MIL-C-5541E 61. For comparison, bare AA2024-

T3 will be severely corroded after 24 hours salt spray test. Exposure of CCC coated

AA2024-T3 in a marine environment for one year showed very little corrosion. Only some extremely shallow pits appeared 62. The corrosion protection of CCCs has been

attributed to its balanced composition and structure. CCC coating contains Cr(III) species which are insoluble and Cr(VI) which are somewhat soluble in water. The insoluble portion serves as a barrier to aggressive environment. The soluble Cr(VI) species provide a self-healing function. Self-healing refers to the ability of a CCC to leach Cr(VI) species, which can migrate to a damaged site and prevent further corrosion.

The protectiveness of a CCC depends on the environment to which it is exposed, the metal substrate, coating bath composition, coating processing parameters, pre- and post-treatments 63, 64.

In general, corrosion resistance of CCC coatings increases with coating thickness.

It has been shown that the corrosion resistance of CCC coatings on zinc increases with the total chromium content in the coating 64. There are contradictory results about the effects of hexavalent chromium content on corrosion resistance. In one study it has been shown that corrosion resistance increases with Cr(VI) content. Removing the Cr(VI) content by bleaching or making it insoluble by heating at elevated temperature decreases the corrosion resistance as measured by salt spray test 65. In a study by Williams 66 of

CCC coating on electroplated zinc, it was concluded that Cr(VI) content does not correlate well with the corrosion resistance in either salt spray test or in exposure to urban environment. He suggested that the leachability of Cr(VI) is more important in determining the corrosion resistance.

The ability of a CCC to protect an alloy also depends on the substrates. For example, Cu-containing aluminum alloys AA2024 (3.8-4.9%Cu) and AA7075 (1.2-2.0%

Cu) are less corrosion-resistant than other Cu-free aluminum alloys when they are coated

with CCC. In a work done by Ketchman 62, all AA6061 (0.15-0.40%Cu) samples coated with CCC withstood 960 hrs salt spray test while all CCC-coated AA2024 panels passed

240 hours but most failed before 480 hours. CCC-coated Al 7075 showed intermediate corrosion resistance between AA6061 and AA2024. All AA 7075 passed 480 hours but most failed to pass 960 hours in the same test. It seemed that the corrosion resistance of

CCC decreased with increasing copper content of the aluminum alloys. The corrosion resistance of CCCs also depends on the environments to which the coatings are exposed.

CCCs usually perform very well in mild neutral environments but poorly in acidic

67 environments, for example in SO2-containing environments .

2.3.4 Paint Adhesion

A typical aircraft coating system consists of three coating layers: a conversion coating layer, a primer layer and a top coat, as shown in Figure 2.12 68. One important function of a CCC is to provide a good base for the subsequently applied organic coating because the natural oxide layer cannot provide enough adhesion between the primer and the substrate. CCCs are strongly bonded to metal substrate and the organic coatings adhere well to the chromated surface thus the paint adhesion is enhanced. It is necessary to distinguish between initial adhesion and durability of the adhesion during exposure to a corrosive environment. The initial adhesion is mainly related to the topography 22 of the

CCCs while the durability of adhesion is related to the corrosion resistance of CCCs.

Several explanations have been given regarding how the CCCs improve the paint adhesion. Removal of grease and other surface contaminants during coating processing

leads to enhanced adhesion because these materials are known to destroy paint adhesion when present. As discussed earlier, CCCs are composed of layers of ellipsoidally or spherically shaped particles. The particles increase surface area that can contact with the organic coatings, leading to larger interaction between organic coating and conversion coating 22. The gaps between the particles may also provide mechanical bonds such as capillary pores and cavities for the paint 69. The high corrosion resistance of CCCs is responsible for the durability of paint adhesion. The loss of adhesion after exposure to corrosive environment has been attributed to corrosion under the organic coating caused by water vapor and oxygen that penetrates the organic coating 5, 70. The presence of soluble chromate can prevent or suppress this kind of corrosion and render the adhesion more durable. CCCs can also prevent reactions between paint constituents and base metals, which may be detrimental to adhesion 71.

2.3.5 Others Properties

Electrical Contact Resistance: Some electrical and electronic equipment is made of aluminum. This equipment requires corrosion protection. CCCs have low electrical contact resistance and can be used for this purpose. Electric contact resistance is a parameter used to characterize the ohmic resistance between two coated components. It is different from the intrinsic electronic resistance. To measure electric contact resistance, a 200 lb load is applied on coated surface through a polished copper rod with a cross section area of 1cm2. The ohmic resistance between the copper rod and the coated

substrate is the electric contact resistance. Electrical contact resistance of CCCs is usually below 5mΩ·in2 2.

Color, Thickness and Coating Weight: The color of CCC coatings on aluminum can be clear, iridescent yellow or brown, depending on the processing conditions. The yellow color is an indication of the existence of Cr(VI) species. Thus the color can be used to roughly estimate Cr(VI) content. The thickness of CCCs is in the range of 10 nm to 1 µm. Coating weight is often used to monitor quality of CCCs in industry. Normal coating weight per unit area on aluminum ranges from 15 to 200 mg/ft2 2.

2.4 Corrosion Protection Mechanism of CCCs

As discussed earlier, CCCs are very effective in the corrosion protection of aluminum and aluminum alloys. The protection mechanisms of CCCs have been a research subject for decades. Significant progress has been made in the understanding of protection mechanism of CCC coatings, especially since late 1980s with the advance of the analytical techniques. Several mechanisms have been proposed to explain how and why CCCs protect the substrate metals from corrosion. These mechanisms include barrier layer protection, bipolar membrane mechanism, active corrosion protection, anodic inhibition, and cathodic inhibition. Active corrosion protection has been addressed by many researchers and it will be discussed in some detail in this section. Corrosion

inhibition by chromate in solution is a topic related to CCC coatings and will also be briefly reviewed.

2.4.1 CCC Coating as a Barrier Layer

CCCs can act as a physical barrier layer to the environment for substrate metals.

This barrier layer is inert and impervious and provides a high degree of isolation between the metal and the corrosive environment. This barrier function mainly comes from the insoluble trivalent chromium species. The hydrated, amorphous nature makes the film a stable layer. There is evidence that supports the barrier function of CCC films. It has been found that after the soluble hexavalent chromium species are leached out of the coatings, the insoluble portion of the film still provides a high level of corrosion protection in salt spray test 72. Other evidence for barrier protection is that after the soluble hexavalent chromium was transformed to the insoluble form by heating to 100ºC, only a limited decrease in corrosion resistance was observed 71. These results indicate that the insoluble portion of CCC coatings provide part of the corrosion protection.

2.4.2 Bipolar Membrane Mechanism

Sato 73, 74 proposed a bipolar membrane model for the passivity of iron group metals and stainless steel. He suggested that as metal dissolution proceeded, a gel-like or porous hydrated metal oxide film formed on the metal surface. Such films are usually

2- 2- anion-selective. However, when non-aggressive oxyanions such as CrO4 , MoO4 are absorbed on the precipitate film, the outmost layer is converted to a cation-selective layer

while the inner layer remains anion-selective. Thus the precipitate film becomes bipolar in nature and ion selective. The innermost layer of the film is positively charged due to excess of metal cation or oxygen vacancy and the outmost layer is negatively charged due to excess oxygen or metal vacancies. The ion transport through this bipolar film is asymmetric and allowed to occur only in the forward direction. Thus the anodic ion transference from metal to environment is constrained. Furthermore, the outmost cation- selective layer impedes the absorption of aggressive anion such as Cl- and makes the passive film corrosion resistant. The structure of CCCs suggests that this model might also be applicable in CCCs 4. The structure of CCCs matches the predictions based on this model very well. Recently it was observed that Cr6+ adsorbed on anodized aluminum lowered the pH of zero charge and discouraged the adsorption of chloride ions. This result also supports the applicability of this bipolar membrane mechanism in CCCs 44.

2.4.3 Active Corrosion Protection Mechanism

Active corrosion protection, or self-healing, is regarded as an important mechanism that makes CCC coatings so effective in corrosion protection. Active corrosion protection refers to the ability of CCC coatings to self-repair when damaged.

This process is thought to proceed as follows. The soluble chromate retained in the coatings dissolves and moves to the active corrosion sites through diffusion or electromigration. Then the chromate is adsorbed onto gel corrosion products and stifles the anodic or cathodic reactions. Thus the corrosion at the damaged area is inhibited 9, 11,

47, 56. Compelling evidence suggests this mechanism operates in CCC coatings. One well-

known phenomenon is that when a CCC-coated metal panel (e.g., Al) is scribed and then subjected to salt spray test, no corrosion is observed in the scribe. It has also been shown that after exposure to NaCl solution, Cr6+ was both released from the surface layer 9, 47, 58 and reduced to Cr3+ 47, 58. It has been assumed that any successful replacement for CCCs must also bear this property 75.

A recent work by Zhao et al. 76 has confirmed the self-healing nature of chromate in CCCs. An “artificial scratch” cell as shown in Figure 2.13 was designed to simulate the scribe on CCCs. In this cell a bare metal sample was brought close to a CCC-coated panel, with the 1.8 mm gap between them filled with sodium chloride solution. After 96 hours, only a few small pits were observed on the bare sample. The polarization resistance of the bare sample was increased by two orders of magnitude. XPS and Raman spectra revealed the existence of some materials chemically similar to CCCs in or near the pits formed on the originally bare sample. These results indicate that the chromate in

CCCs must have dissolved into the solution in the gap and migrated to the active corrosion sites. The migrating chromate was then either reduced to form insoluble precipitates or was absorbed onto previously formed corrosion products, or both, at corrosion sites. The reduction products and/or absorbed chromate prevent further corrosion. One surprising fact is that the concentration of chromate released from the

CCCs was estimated at 8 ×10-5M and it was enough to protect the Al alloy in 0.1 M

NaCl. Other researchers have found much higher chromate concentration is required to inhibit pit growth at similar conditions 4, 77, 78.

The release characteristics of soluble chromate from CCCs are important for the self-healing action to occur. If the release rate is too low, there may not be enough chromate to heal the defects. If the release rate is too high, the protection life of CCCs may be shortened. Release kinetics have been studied by several authors 11, 44, 79. Glass 79 studied the leaching rate of chromate using a radioactive tracer method. It was found that the leaching of chromate from CCCs was very rapid at the beginning but decreased with immersion time in NaCl solution. Finally, the chromate concentration reached a plateau.

These characteristics have been confirmed by recent studies 44. A mechanism for the release of chromate from CCCs has been proposed 14, 44. This mechanism is based on the release and adsorption of Cr (VI) species on insoluble Cr(III) compounds. The proposed reaction can be represented by following equation 44:

- + Cr(III)-O-Cr(VI)O3H + H2O → Cr(III)-OH + HCrO4 (aq) + H Eqn. 2.10

Where Cr(III)-OH represents insoluble chromium hydroxide, Cr(VI)aq is soluble chromate and Cr(III)-O-Cr(VI) is the mixed chromium oxide in CCCs. The release of

Cr6+ from CCCs was well described with a modified Langmuirian adsorption-desorption model 80. Kendig et al.11 took a different approach to study the release of chromate. They measured the Cr(VI) content in CCCs after the coatings had been exposed to corrosive environment for varying times. The decay of Cr(VI) content in CCCs can be approximated by an exponential function:

-t/τ y = y0 ·e Eqn. 2.11

where y is concentration of Cr(VI) in CCCs, t is time and τ is a time constant. Several factors can influence the chromate release from CCCs. Aging of CCCs at room temperature decreased the leaching rate and total amount of chromate released 79, 81.

Aging at elevated temperature had similar but more pronounced effect. The coatings aged at 100ºC for 2 hours did not release detectable chromate. It is suggested that the dehydration during aging causes the loss of Cr6+ leachability. There is evidence that supports this argument. For example, when CCCs were covered with a strippable film and then heated at elevated temperature, the reduction in Cr6+ leachability was much less evident 79. Thermogravimetric analysis clearly showed the water loss when CCCs were heated 1. It was suggested that hydration of Cr(VI) species was necessary for the solubility of these species 47. Accompanying the changes of leaching kinetics is the variation in the corrosion resistance. Aging at room temperature for an extended time deteriorated the coating performance in salt spray test. The extent of this deterioration increased with aging time 82. Aging at 75ºC, 100ºC, 150ºC and 200ºC for 2 hours resulted in progressive damage in salt spray exposure 81, 83.

The results presented above indicate that retained chromate plays an important role in the protection provided by CCCs. It increases the corrosion protection through self-healing action.

2.4.4 Chromate in Solution

Chromates are very effective corrosion inhibitors when added to solution. The behavior of chromate in solution is relevant to the corrosion protection by CCCs, because the retained chromate can dissolve into the local environment and provide part of the protection offered by CCCs.

Chromate influences several aspects of pitting corrosion and repassivation process of corroding metals. Bohni et al. 77 found that the pitting potential of pure aluminum in dilute chloride solution was shifted in the noble direction when chromate was added. The logarithm of critical activity of chromate required to inhibit pitting corrosion of Al in aerated solution was linear with the logarithm of the activity of chloride ion. Chromate also affects metastable pitting behavior, such as nucleation frequency, peak pit current, apparent pit radii and apparent pit current densities 84. Through decreasing the metastable pit nucleation rate and minimizing pit growth rate, chromate reduced the chance for pit stabilization thus reduced pitting corrosion. Chromate also affects the growth of existing pits. The growth rate of pits decreased when the [Cr6+]/[Cl-] ratio was above 2:1 85, 86.

The repassivation process of aluminum in NaCl solution was found to be accelerated by the addition of chromate 87. This acceleration of passivation was due to the reduction of

Cr(VI) to Cr(III).

The characteristics of chromate make it an effective corrosion inhibitor. First, chromate is a strong oxidizer. Second, the reduction products are insoluble, impervious and passive in nature. One monolayer of Cr3+ oxide can effectively inhibit oxygen reduction reaction. Finally chromate is speciated as an oxoanion over pH 0~14 88. Some

mechanisms regarding the corrosion inhibition by chromate have been proposed. The existence of many models suggests that there is no widely accepted explanation about the action of chromate as an inhibitor. Some of the mechanisms are briefly discussed below.

Chromate as a Cathodic Inhibitor 4, 86: This mechanism suggests that chromate mainly affects the cathodic reaction. Polarization curves in Figure 2.14 show that the cathodic reaction rate is reduced at the presence of 10-4 to 10-3 M chromate while anodic reaction is not affected when tested in 0.5M NaCl. This is a little bit surprising given the fact that chromate is a strong oxidizer. An explanation is given as following.

Reduction of chromate leads to the formation of a mixed aluminum-chromium mixed oxide film, which is insoluble and has higher electrical resistance than alumina film. This mixed oxide film stifles the cathodic activity and blocks the cathodic partial reaction.

With limited cathodic current, the anodic reaction, which is the dissolution of metal, will slow down. Thus the metal is protected from corrosion. Recent results 89 suggest that even a monolayer of chromium oxide can effectively shut down the oxygen reduction reaction by blocking active sites for O2 adsorption and reducing electrons tunneling rate through this thin chromium oxide film. These results provide strong evidence for the cathodic inhibition argument.

Chromate as an Anodic Inhibitor: The anodic inhibition function of chromate depends on the environment. When tested in 0.5 M NaCl, the anodic polarization behavior was not affected by the presence or absence of chromate ions at a

90 concentration up to 0.05 M . However, when tested in 0.1 M Na2SO4 + 0.005 M NaCl solution, the pitting potential was increased when 0.01 M chromate was added into solution 91. Although chromate might reduce localized corrosion initiation, it does not effectively inhibit localized corrosion growth at [Cr6+]/[Cl-] below 2:1 90, 92.

Competitive Absorption Theory 77, 91, 93: This theory suggests that chromate ions are specifically and strongly absorbed on the oxide film on metals, preventing the absorption of chloride ion and their entry into the oxide. Thus the film on the metal is kept stable and protective. There is a linear relationship between the logarithm of the activity of chloride ions and the logarithm of the activity of chromate ions that was required to shift the pitting potential of Al by an arbitrary value. Recent studies by Kendig et al. 94 suggest that chromate adsorbed on Al oxide lowered the zeta potential and the pH of zero charge of Al oxide surfaces. This should inhibit chloride adsorption.

Blocking Defects in Oxide Film 54, 92: In this model, it is suggested that the original oxide film on Al contains flaws. When Al supporting a flaw-containing film is immersed in chromate solution, these flaws allow easy access of chromate to Al. The mixed, hydrated Al/Cr oxide resulting from the reduction of chromate and oxidation of

Al will block these flaws, preventing further exposure of substrate to the environment.

This model readily explained the fact that Cr(VI) content in passive film increases with the thickness of initial oxide film after samples supporting oxide film were immersed in

chromate solution. Because flaws are more numerous in thin films than in thick films, the reduction of absorbed chromate is more complete in thin films.

Inhibition of Hydration of Oxide Film 95: This theory suggests that the deterioration of the oxide film on Al is due to hydration caused by the penetration of water molecules or OH- into the oxide lattice when the oxide film is immersed in water.

Chromate ions can be adsorbed on the oxide surface and form a monolayer. This monolayer hinders the penetration of water or OH- into oxide and stabilizes the oxide film.

From the above discussion it can be concluded that the inhibition function of chromate ion results from either its reduction to form insoluble compounds or its absorption on the oxide surface, or perhaps both.

2.5 Development of Chromate-Free Conversion Coatings

The high toxicity and carcinogenic nature of Cr6+ has limited the future usage of chromate in the corrosion protection field. Considerable effort has been devoted to the development of alternative inhibitors and coating systems 68, 96, 97. As a result, many inhibitors and coating systems have been proposed as possible candidates for the replacement of chromate. Some of them are very promising. In this section, two coating systems will be briefly discussed. These include hydrotaclite coatings and cerium-based coatings. These two coatings were selected because the effort that had been devoted to these two coating systems and the availability of literature.

2.5.1 Hydrotalcite Coatings

Buchheit et al. developed a lithium-aluminum-carbonate-hydroxide hydrate

(hydrotalcite) as a corrosion resistant coating for aluminum alloys 98-100. This coating system was based on the formation of a passive film on Al and its alloys when they are exposed to alkaline solution containing lithium. Several researchers have observed the unexpected passivity of Al and Al alloys in alkaline Li-containing solutions 101-104.

Fernandes et al. 101 found that an active-passive transition occurred when Al was anodically polarized in a solution containing Li2CO3. However, this transition did not

+ occur in Na2CO3 solution. They attributed this behavior to the incorporation of Li into the passive film. Perrotta 102 and Buchheit 103 reported that Al was also passivated under open circuit condition when Al exposed to lithium hydroxide and/or lithium carbonate solutions. They found the passive film had a hydrotalcite-like structure. The passive film formed was shown to have high corrosion resistance against chloride ion attack 104.

Buchheit 98-100 developed a corrosion resistant coating process on Al and Al alloys utilizing this passivation phenomenon. This new coating process was procedurally similar to that of CCC coatings, involving a simple immersion of metal parts into a formulated solution. The coatings impart high corrosion resistance to substrate metals. Additionally, the chemicals used in the coating process were not toxic and posed negligible health and environmental risk. During immersion, a passive film gradually formed accompanied by

Al dissolution and hydrogen evolution. The coating was formed by a coprecipitation of

Al dissolution products and components in the coating bath. The initial coating bath

chemistry they worked with was Li2CO3 plus LiOH. The composition, structure, and corrosion resistance of the coating formed in this chemistry have been studied. It was determined that the coating was mainly comprised of hydrotalcite-like compounds. In this case, it was the carbonate variant of these compounds: Li2[Al2(OH)6]·CO3·nH2O. The coating had a plate-like morphology, with plates intersecting each other. Cross section showed the coating had a double layer structure, with outer layer being crystallites and inner layer being amorphous. The coating corrosion resistance depended on substrate alloy type. The coating performed very well on AA1100 and AA6061 and was able to pass the 168 hours salt spray test. However, this carbonate variant of hydrotalcite coating

2- did not work well on Cu-containing Al alloys. Recent work has found that if CO3 in the

- coating bath was replaced by NO3 , coating corrosion resistance was increased by one order of magnitude as measured by EIS 100. The coatings formed on Cu-containing

AA2024-T3 in solution containing lithium nitrate salts were able to pass the 168 hrs salt spray test. This improvement in corrosion resistance was believed to be due to the

- - oxidizing ability of NO3 . It has been found that NO3 was reduced to other species during coating process. The main protection action by hydrotalcite coating was its barrier function. It was not clear if it also inhibited cathodic kinetics. Recent results have illustrated that Ce- modified hydrotalcite coating (by dipping hydrotalcite coating into

Ce-containing solution) showed active corrosion protection 105,106. This active corrosion protection needs to be explored further.

2.5.2 Ce-Based Conversion Coatings

Ce salts as corrosion inhibitors and as conversion coating forming agents have been studied extensively 107-114. Hinton et al. 107 first studied Ce salts as corrosion inhibitors for Al alloys. It was found that the corrosion rate of AA7075 in 0.1 M NaCl solution was reduced by a factor of 10 as measured by weight loss when 100-1000 ppm

CeCl3 was added to the solution. Resistance to pitting corrosion was also increased in the

3+ presence of CeCl3. Polarization experiments indicated that Ce acted as a cathodic inhibitor. Cathodic reaction, either oxygen reduction or hydrogen evolution resulted in an increase in pH at cathodic sites. Ce oxides precipitated at high pH and formed a film on cathodic sites. As this process continued, the film will cover the entire surface. Oxygen reduction was suppressed on this film thus the anodic dissolution of Al was inhibited 107,

108. The film mainly contained Ce3+ species and, upon extended exposure to NaCl solution, Ce3+ was oxidized to Ce4+ 110, 111. This Ce3+ to Ce4+ transition showed that this cerium-containing oxide film is electronically insulated from the aluminum substrate.

This is consistent with the model of Ce3+ as a cathodic inhibitor.

Much effort has also been put into developing a cerium-based conversion coating

102-107. The incorporation of Ce into a protective film met some difficulty. Producing Ce- based coatings by open circuit immersion required prolonged periods, on the order of weeks. For example, exposure of AA 7075-T6 to NaCl solution containing 100 ppm for

20 days only grew a film of 120 nm thick 107. For the film to be protective on AA 7075-

T6, an open circuit immersion of one week in 1000 ppm CeCl3 solution was necessary

109. The film formed by this lengthy immersion showed increased corrosion resistance in

chloride solution for certain alloys. Mansfeld et al. 112 found that AA 6061 and AA7075-

T6 that had been immersed in 1000ppm CeCl3 solution for one week did not suffer from pitting corrosion in subsequent immersion in aerated 0.5M NaCl solution for 3 weeks.

However, a treatment of this long time is impractical. Some accelerated methods have been developed. Hinton et al. 109 developed a process in which a cathodic potential was applied during immersion of Al 7075-T6 in 1000ppm CeCl3 solution or in a non-aqueous

113, 114 Ce(NO3)3 solution. Another process was developed by Mansfeld et al. . Their process included immersions in boiling Ce(NO3)3 solution and boiling CeCl3 solution for

2 hrs each. Then the sample was anodically polarized in a deaerated molybdate solution for 2hrs. The coating formed on AA 6061 by this treatment did not show any signs of corrosion after 60 days of immersion in 0.5 M NaCl solution. A panel with a scribe on coating surface did not show any corrosion either after 25 days immersion in NaCl solution. Again this coating process was not feasible for industry application. The big challenge in the application of Ce-based coatings is to develop a practical coating procedure.

2.5.3 Other Alternative Coating Systems

Many other coating systems have been developed with the intention to replace

CCC coatings 68, 96, 97. Some of them have been reported to have good corrosion resistance, for example, coatings formed in potassium permanganate baths, and cobalt/molybdenum-based coatings. But the processes either have too many steps or

involve high temperature steam or boiling water, which make them expensive and sometimes impractical.

A trivalent chromium conversion coating process that uses trivalent chromium salts in coating bath has also been developed 115. The coatings produced withstand 500 hours salt spray test. But this coating still involves Cr(VI), because after the trivalent chromium coating is formed, it is immersed in H2O2 to oxidize some Cr(III) to Cr(VI).

Molybdate has been used as an inhibitor but the effort to make conversion coatings with molybdate is not successful. The coatings developed with molybdate have not been able to provide satisfactory corrosion resistance.

2.6 Key Issues to be Addressed Regarding the Formation and

Corrosion Inhibition Mechanisms of Chromate Conversion

Coatings

Efforts to find a non-toxic substitute for chromate as corrosion protection date to at least the early 70s 1. However no replacement technology that can match chromate in application simplicity and in effectiveness of corrosion protection has been developed so far. The difficulty of developing replacement technology for chromate is in part due to chromate’s rare or maybe unique properties. However, lack of understanding or lack of depth in understanding regarding chromate formation and protection mechanism also hinders the transition from chromate to chromate-free protection. Although much has been known about the chromate inhibition, some important questions remain.

CCC formation process on Al has been studied extensively. However, available coating formation models have major flaws. For examples, Katzman’s model cannot explain why there is little aluminum in CCCs. Brown’s model predicts the existence of deep holes in isolated anodic sites, which are not found in the cross section of CCCs.

Arrowsmith’s model does not explain how coating growth is still possible after the substrate is covered with the first layer of CCC. Therefore a better understanding of coating formation process is needed. The newly available multichannel microelectrode analyzer (MMA) makes it possible to study the current evolution during coating formation and provide new information on coating formation process. The functions of the minor components in coating bath have been documented, but how they affect the coating resistance to localized corrosion has not been addressed. Study of this subject could provide guidance to the development of chromate-free coating systems.

The second issue to be addressed is the cathodic inhibition function of chromate conversion coatings. There are conflicting reports on this subject. Some results imply that chromate coatings should strongly inhibit cathodic reaction (oxygen reduction). But there are also reports that conclude chromate coatings do not inhibit oxygen reduction. Some results indicate that the cathodic inhibition by chromate coatings changes with coating aging time. The fundamental importance of the cathodic inhibition of CCC warrants a clarification on this subject.

It is known that chromate coatings are dynamic systems. Several changes can occur during aging after coatings are formed on substrate. These changes include dehydration, development of shrinkage cracks, transition from hydrophilicity to

hydrophobicity, and loss of mobility of Cr6+ in coatings and corrosion resistance. The effects of aging time and temperature on coating properties have been studied previously.

A factor that has been neglected is the humidity of the aging environment. The studies of the dramatic effect of aging in dry air on coatings properties have both fundamental and practical significances. The results shed new light on what happens during aging process and could be used in practice to preserve the corrosion resistance of chromate coatings.

REFERENCES

1 N. J. Newhard, Jr., Metal Finishing, 49, 66 (1972)

2 K. A. Korinek (1987). Chromate Conversion Coatings. Metals Handbook.

Metals Park, OH, ASM International. Vol. 13.

3 P. L. Hagans and C. M. Hass (1994). Chromate Conversion Coatings. ASM

Handbook. Metals Park, ASM International. Vol.5: 405.

4 G. S. Frankel (1998). Mechanism of Al Alloy Corrosion and the Role of

Chromate Inhibitors. Columbus, OH, The Ohio State University.

5 J. F. Murphy (1963). Chemical Conversion Coatings for Aluminum Alloys. The

Finishing of Aluminum. G. H. Kissin. New York, Reinhold Publishing

Corporation.

6 R. M. Burns and W. W. Bradley (1967). Protective Coatings for Metals. New

York, NY, Reinhold Publishing Corporation.

7 S. Wernick, R. Pinner and P. G. Sheasby (1987). The Surface Treatment and

Finishing of Aluminum and its Alloys. Metals Park, OH, ASM International.

8 P. L. Hagans and C. M. Hass, Surface and Interface Analysis, 21, 65 (1994)

9 H. A. Katzman, Malouf, G. M., Bauer, R. and Stupian, G. W., Applications of

Surface Science, 2, 416 (1979)

10 A. E. Hughes, R. J. Taylor and B. R. W. Hinton, Surface and Interface Analysis,

25, 223 (1997)

11 M. W. Kendig, Davenport, A. J. and Isaacs, H. S., Corrosion Science, 34, 41

(1993)

12 J. A. Treverton and N. C. Davis, Metals Technology, 480 (1977)

13 J. A. Treverton and N. C. Davis, Surface and Interface Analysis, 3, 194 (1981)

14 L. Xia and R. L. McCreery, Journal of The Electrochemical Society, 146, 3696

(1999)

15 G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson and G. C. Wood,

Corrosion Science, 34, 1045 (1993)

16 G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson and G. C. Wood,

Corrosion Science, 33, 1371 (1992)

17 G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson and G. C. Wood,

Corrosion Science, 35, 253 (1993)

18 G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson and G. C. Wood,

Corrosion Science, 34, 2099 (1993)

19 M. F. A. Rabbo, J. A. Richardson and G. C. Wood, Corrosion Science, 18, 117

(1978)

20 K. Shimizu, G. M. Brown, K. Kobayashi, G. E. Thompson and G. C. Wood,

Corrosion Science, 34, 1853 (1993)

21 Q. Meng (2003). The Effect of Cu on Corrosion Behvior and Chromate Coating

Formation on AA 7XXX Alloys. Materials Science and Engineering. Columbus,

OH, The Ohio State University.

22 J. A. Treverton and M. P. Amor, Transactions of the Institute of Metal

Finishing, 60, 92 (1982)

23 D. J. Arrowsmith, J. K. Dennis and P. Sliwinski, Transactions of the Institute of

Metal Finishing, 62, 117 (1984)

24 J. A. Treverton and M. P. Amor, Journal of Materials Science, 23, 3706 (1988)

25 J. S. Crompton, P. R. Andrews and E. McAlpine, Surface and Interface

Analysis, 13, 160 (1988)

26 J. A. Treverton, A. Bosland and J. M. Brown, Corrosion Science, 30, 1159

(1990)

27 J. R. Waldrop and M. W. Kendig, Journal of The Electrochemical Society, 145,

L11 (1998)

28 L. Xia (2000). Formation and Function of Chromate Conversion Coating on

Aircraft Aluminum Alloy Probed by Vibrational Spectroscopy. Department of

Chemistry. Columbus, OH, The Ohio State University.

29 J. S. Osborne, Progress in Organic Coatings, 41, 280 (2001)

30 P. A. Schweitzer (1996). Corrosion Engineering Handbook. New York, NY, M.

Dekker, Inc.

31 J. E. Hatch (1984). Aluminum: Properties and Physical Metallurgy. Metals Park,

OH, ASM.

32 W. Zhang (2001). Localized Corrosion Kinetics In High Strength AA2024

Alloys. Materials Science and Engineering. Columbus, OH, The Ohio State

University.

33 R. G. Buchheit, R. P. Grant, P. F. Hlava, B. Mckenzie and G. L. Zebnder,

Journal of The Electrochemical Society, 144, 2621 (1997)

34 G. S. Chen, M. Gao and R. P. Wei, Corrosion, 52, 8 (1996)

35 N. R. Cawley and D. G. Harlow, Journal of Materials Science, 31, 5127 (1996)

36 R. G. Buchheit, Journal of The Electrochemical Society, 142, 3994 (1995)

37 P. Schmutz and G. F. Frankel, Journal of The Electrochemical Society, 145,

2285 (1998)

38 P. Schmutz and G. F. Frankel, Journal of The Electrochemical Society, 145,

2295 (1998)

39 V. Guillaumin and G. Mankowski, Corrosion Science, 41, 421 (1999)

40 C. Blanc, B. Lavelle and G. Mankowski, Corrosion Science, 39, 495 (1997)

41 K. Kowal, J. Deluccia, J. Y. Josefowicz, C. Laird and G. C. Farrington, Journal

of The Electrochemical Society, 143, 2471 (1996)

42 W. R. McGovern, P. Schmutz, R. G. Buchheit and R. L. McCreery, Journal of

The Electrochemical Society, 147, 4494 (2000)

43 A. E. Hughes, J. D. Gorman and P. J. K. Paterson, Corrosion Science, 38, 1957

(1996)

44 G. S. Frankel (1999). Mechanism of Al Alloy Corrosion and the Role of

Chromate Inhibitors. Columbus, OH, The Ohio State University.

45 Z. Yu, H. Ni, G. Zhang and Y. Wang, Applied Surface Science, 62, 217 (1992)

46 L. Kintrup and L. d. Riese-Meyer, Inst. Phys. Conf. Ser., 119, 257 (1991)

47 F. W. Lytle, R. B. Greegor, G. L. Bibbins, K. Y. Blohowiak, R. E. Smith and G.

D. Tuss, Corrosion Science, 37, 349 (1995)

48 K. Asami, M. Oki, G. E. Thompson, G. C. Wood and V. Ashworth,

Electrochimca Acta, 32, 337 (1987)

49 H. E. Townsend and R. G. Hart, Journal of The Electrochemical Society, 131,

1345 (1984)

50 G. P. Halada and C. R. Clayton, Journal of The Electrochemical Society, 138,

2921 (1991)

51 L. Xia and R. L. McCreery, Journal of The Electrochemical Society, 145, 3083

(1998)

52 A. G. Shrott, G. S. Frankel, A. J. Davenport, H. S. Isaacs, C. V. Jahnes and M.

A.Russak, Surface Science, 250, 139 (1991)

53 D. C. Koningsberger and R. Prins (1988). X-ray Absorption Principles,

Applications, Techniques of EXAFS, SEXAFS and XANES. New York, NY,

John Willey & Sons, Inc.

54 J. K. Hawkins, H. S. Isaacs, S. M. Heald, J. Tranquada, G. E. Thompson and G.

C. Wood, Corrosion Science, 27, 391 (1987)

55 J. K. Hawkins, H. S. Isaacs, S. M. Heald, J. Tranquada, G. E. Thompson and G.

C. Wood (1986). Proceedings (The Electrochemical Society) 86-11, The

Electrochemical Society.

56 J. S. Wainright, O. J. Murphy and M. R. Antonio, Corrosion Science, 33, 281

(1992)

57 S. W. M. Chung, J. Robinson, G. E. Thompson, G. C. Wood and H. S. Isaacs,

Philosophical Magazine B, 63, 557 (1991)

58 J. Wan, G. E. Thompson, K. Lu and C. J. E. Smith, Physica B, 208&209, 511

(1995)

59 L. C. Feldman and J. W. Mayer (1986). Fundamentals of Surface and Thin Film

Analysis. New York, NY, Elsevier Science Publishing Co., Inc.

60 ASTM B117-97 "Standard Practice for Operating Salt Spray (Fog) Apparatus",

ASTM (1997)

61 MIL-C-5541E " Chemical Conversion Coatings on Aluminum and Aluminum

Alloys" (1990).

62 S. J. Ketcham (1963). Properties of Chemical Films on Aluminum Alloys. The

Finishing of Aluminum. G. H. Kissin. New York, NY, Reinhold Publishing

Corporation.

63 L. F. Spencer, Metal Finishing, 58, 58 (1960)

64 I. Suzuki (1989). Corrosion Resistant Coatings Technology. New York, NY,

Marcel Dekker, Inc.

65 Effect of Hexavalent Chromium Content on Performance of Chromate Films.

Philadelphia, PA, Aeronautical Materials Lab, Naval Engineering Center

(1964).

66 L. F. G. Williams, Surface Technology, 7, 113 (1978)

67 V. S. Agarwala, ACS Symposium Series, 211 (1986)

68 R. L. Twite and G. P. Bierwagen, Progress in Organic Coatings, 33, 91 (1998)

69 C. Wick (1983). Tool and Manufacturing Engineers Handbook. Dearborn, MI,

Society of Manufacturing Engineers.

70 R. A. Dickie, Progress in Organic Coatings, 25, 3 (1994)

71 T. Biestek and J. Weber (1976). Electrolytic and Chemical Conversion

Coatings. Poland, Portcullis Press Limited-Redhill.

72 W. E. Pocock, Metal Finishing, 52, 48 (1954)

73 N. Sato, Corrosion, 45, 354 (1989)

74 N. Sato, Corrosion Science, 31, 1 (1990)

75 G. S. Frankel (1996). Mechanism of Al Alloy Corrosion and the Role of

Chromate Inhibitors. Columbus, OH, The Ohio State University.

76 J. Zhao, G. Frankel and R. L. McCreery, Journal of The Electrochemical

Society, 145, 2258 (1998)

77 H. Bohni and H. H. Uhlig, Journal of The Electrochemical Society, 116, 906

(1969)

78 G. S. Frankel (1997). Mechanism of Al Alloy Corrosion and the Role of

Chromate Inhibitors. Columbus, OH, The Ohio State University.

79 A. L. Glass, Materials Protection, 26 (1968)

80 L. Xia, E. Akiyama, G. S. Frankel and R. L. McCreery, Journal of The

Electrochemical Society, 147, 2556 (2000)

81 A. Gallaccio, F. Pearlstein and M. R. D'ambrosio, Metal Finishing, 64, 50

(1966)

82 P. Trathen, B.Hinton, K. Hammon, T. Hughes and R. Taylor (1994). The

Variability of Chromate Conversion Coating Process and the Effect on Corrosion

Protection Performance. Clayton Victoria, Australia, CSIRO.

83 L. F. G. Williams, Surface Technology, 5, 105 (1977)

84 S. T. Pride, J. R. Scully and J. L. Hudson, Journal of The Electrochemical

Society, 141, 3028 (1994)

85 C. S. Jeffcoate, H. S. Isaacs, J. Hawkins and G. E. Thompson (1998).

Electrochemical Society Proceedings 98-17.

86 G. S. Frankel (1999). The Influence of Inhibitor Ions and Conversion Coatings

on Localized Corrosion of Al and Al Alloys. Columbus, OH, The Ohio State

University.

87 R. J. Cinderey and G. T. Burstein, Corrosion Science, 33, 493 (1992)

88 R. G. Buchheit, G. S. Frankel, H. L. Fraser, R. L. McCreery, J. S. Beatty and M.

S. Donley (1999). Critical Factors for the Transition from Chromate to

Chromate-Free Protection. Columbus, OH, The Ohio State University.

89 W. J. Clark and R. L. McCreery, Journal of The Electrochemical Society, 149,

B379 (2002)

90 A. Sehgal, D. Lu and G. S. Frankel, Journal of The Electrochemical Society,

145, 2834 (1998)

91 G. O. Illevbare and J. Scully, Corrosion, 56, 227 (2001).

92 A. Sehgal, G. S. Frankel, B. Zoofan, and S. Robin, Journal of The

Electrochemical Society, 147, 140 (2000).

93 M. A. Heine and M. J. Pryor, Journal of The Electrochemical Society, 114,

1001 (1967)

94 M. Kendig, R. Addison and S. Jeanjaquet, Journal of The Electrochemical

Society, 146, 4419 (1999)

95 S. K. Konno, H. Takahashi and M. Nagayama, Corrosion Science, 22, 913

(1982)

96 B. R. W. Hinton, Metal Finishing, 89, 15 (1991)

97 S. M. Cohen, Corrosion, 51, 71 (1995)

98 R. G. Buchheit, M. D. Bode and G. E. Stoner, Corrosion, 50, 205 (1994)

99 C. A. Drewien, M. O. Eatough, D. R. Tallant, C. R. Hills and R. G. Buchheit,

Journal of Materials Research, 11, 1507 (1996)

100 W. Zhang and R. G. Buchheit, Corrosion, 58, 591 (2002)

101 J. C. S. Fernandes and M. G. S. Ferreira, Electrochimca Acta, 37, 2659 (1992)

102 A. J. Perrotta, S. Y. Tzeng, W. D. Imbrogno and R. Rolles, Journal of Materials

Research, 7, 3306 (1992)

103 R. G. Buchheit, F. D. Wall, G. E. Stoner and J. P. Moran, Corrosion, 51, 417

(1995)

104 C. M. Rangel and M. A. Travassos, Corrosion Science, 33, 327 (1992)

105 R. G. Buchheit, S. B. Mamidipally, P. Schmutz and H. Guan, Corrosion, 58, 3

(2002)

106 S. B. Mamidipally (1999). Active Corrosion Protection by Cerium-Modified

Hydrotalcite Coatings on Aluminum Alloys. Materials Science and Engineering.

Columbus, OH, The Ohio State University.

107 B. R. W. Hinton, D. R. Arnott and N. E. Ryan, Metals Forum, 7, 211 (1984)

108 B. R. W. Hinton, Journal of Alloys and Compounds, 180, 15 (1992)

109 B. R. W. Hinton, D. R. Arnott and N. E. Ryan, Materials Forum, 9, 162 (1986)

110 A. J. Davenport, H. S. Isaacs and M. W. Kendig, Corrosion Science, 32, 653

(1991)

111 A. J. Davenport and H. S. Isaacs, Corrosion Science, 31, 105 (1990)

112 F. Mansfeld, S. Lin, S. Kim and H. Shih, Corrosion, 45, 615 (1989)

113 F. Mansfeld, Y. Wang and H. Shih, Journal of The Electrochemical Society,

138, L74 (1991)

114 F. Mansfeld, Y. Wang and H. Shih, Electrochimca Acta, 37, 2277 (1992)

115 F. Pearlstein and V. S. Agarwala, Plating and Surface Finishing, 81, 50 (1994)

FIGURES

3- Figure 2.1 Proposed Fe(CN)6 mediation mechanism. Arrows represent redox cross reactions 14.

Figure 2.2 Mechanism of coating growth 9.

20µm ( a )

( b )

Figure 2.3 Microstructure of AA 2024-T3. (a) SEM image of constituent particles.

(b) TEM image of precipitates at matrix and grain boundaries 32.

20 µm

Figure 2.4 Mud-cracking morphology of CCC coating formed by 3min immersion in Alodine 1200S on Al 7075 47.

( a ) ( b )

Figure 2.5 Depth profile of a CCC coating on Al 12. (a) Al, Fe, Cr, O (b) Al, N, C, F

Figure 2.6 A structure model of CCC coating on Al proposed by Treverton et al 12.

Figure 2.7 Sputtering profiles measured by Auger electron spectroscopy in a fresh Alodine 1200S coating on Al 2024 47.

Figure 2.8 Coating structure model of CCC coating on Al proposed by Townsend et al 49.

Figure 2.9 Structure model of CCC coating on Al 2024 suggested by Hughes et al 10.

Figure 2.10 Possible structure for the Cr(III)/Cr(VI) mixed oxide present in CCC coating.The bold line indicates an amorphous trivalent chromium - hydroxide solid. containing various amount of H2O and OH . Cr(VI) 2- 2- 51 exist as Cr2O7 (shown) or CrO4 .

( a )

( b )

Figure 2.11 The differences in XANES between (a) Cr(III) standard (Cr2O3) and 54 (b) Cr(VI) standard (K2CrO4) .

Topcoat

Primer

Conversion coating

Substrate

Figure 2.12 Schematic of components that make up an aircraft coating system 68.

Figure 2.13 Schematic drawing of artificial scratch cell. The larger piece is CCC- coated metal panel and smaller one is bare metal 76.

Figure 2.14 Polarization curves of Al2024-T3 in an aerated, oxygen stirred 1M NaCl base solution with dicromate 90.

CHAPTER 3

CHARACTERIZATION OF CHROMATE CONVERSION COATING FORMATION AND BREAKDOWN USING ELECTRODE ARRAYS

3.1 Introduction

The structure and chemical composition of CCCs have been the subject of numerous studies 1-11. The findings of these studies show CCCs to be a mixture of chromium oxides, other components from coating bath, and components from substrate.

Chromium is present in CCC as both Cr3+ and Cr6+, with Cr6+ predominantly in the outer layer. Several mechanisms have been proposed to explain the excellent corrosion protection provided by CCCs. Among them are the barrier layer protection mechanism

12,13, the bipolar membrane mechanism 14,15, and the active corrosion protection mechanism 3,11,16,17. Other studies have been carried out on the formation of CCCs.

Commonly CCC formation is described as a redox reaction between chromate ions and substrate metals 3,8,9. Chromate ions are reduced to non-soluble chromium oxide, which forms on the substrate as a protective layer. In accelerated chromium chromate coating

2,8,9,18,19 6+ 3+ 0 3+ formulations, K3Fe(CN)6 is present as an accelerator for the Cr /Cr -Al /Al

redox couple. NaF is present as an activator 3,8 that dissolves any air-formed surface film, and allows the conversion reaction to proceed with greater intensity than would otherwise be possible.

Although the chemistry and structure of CCCs have been investigated extensively, only a few studies have focused on the relationship between formation of

CCCs and subsequent breakdown behavior in Cl- solutions under potential control 20,21.

The functions of K3Fe(CN)6 and NaF in the coating bath have also been studied, but how these minor additions affect coating breakdown is not clear. CCC performance is usually assessed by salt spray or field exposure, but electrochemical impedance spectroscopy

(EIS) has been used with increasing regularity in recent years 22. Anodic polarization methods have not been widely used to evaluate CCC breakdown behavior perhaps because chromate conversion coatings on large area electrodes (Area ≥ 1.0cm2) do not usually show much improvement in pitting potential in concentrated chloride solutions.

Improvements in pitting potentials of CCC-coated 2024-T3 have been reported in 0.1M

20,21 Na2SO4 + 0.005M NaCl solutions however .

In this study, a multichannel microelectrode analyzer (MMA) was used to monitor the electrochemical activity on aluminum electrode arrays during coating formation. The coatings were allowed to dry and were then subject to potentiodynamic polarization in 0.5M NaCl solution until breakdown was detected. The effects of coating time, K3Fe(CN)6 and NaF on coating formation and breakdown were studied using this approach. Using the MMA, it was possible to directly study the relationships between the coating formation process, as indicated by the current evolution, and coating breakdown.

The results from these experiments shed new light on the CCC formation process and the relationship between CCC processing and bath chemistry. These findings may also provide some guidance to the development of chromate-free coating systems.

3.2 Experimental Procedures

Materials and electrode construction. To study CCC formation and breakdown, aluminum wires, 0.5 mm in diameter, with a purity of 99.999% were used to build 5 × 5 electrode arrays. To study the effects of Cu on coating formation and breakdown behavior, pure Cu wire of 0.5 mm diameter was also introduced into to the electrode arrays. If one Cu wire was used, it was located at the center of the electrode array. When two Cu wires were used, the wires were located at the third row, the second and the fourth segments. The electrode arrays used in the study were all aluminum arrays if not specified. A photograph of electrode array is shown in Figure 3.1. The distance between two adjacent electrodes in a row or a column is around 1 mm. At this distance, diffusion fields associated with individual electrode elements do not overlap and there is no chemical interaction between array elements during the coating formation process.

Assuming a diffusion coefficient of 5 x 10-6 cm2/s 23 for ions in the coating bath and a maximum coating time of 300 s, the diffusion length can be estimated using:

L = (Dt)1/2 Eqn. 3.1

The calculated diffusion length is 0.4 mm, which is less than the distance between the electrodes in immediate neighboring rows or columns. During subsequent anodic potential scanning experiments in which individual electrode elements developed pitting, chemical interactions cannot be ruled out on the basis of a diffusion argument because the running time of anodic polarization test is much longer than the coating time. However, pitting on an array appeared to occur randomly − that is that pitting of one element did not appear to accelerate or delay pitting of its immediate neighbors compared to the entire population of electrodes in the array.

Since these electrode segments are connected through zero resistance ammeters

(ZRAs), it is necessary to estimate the ohmic resistance between adjacent electrodes.

Using a solution provided by Newman for a disk electrode 24, the ohmic resistance between neighboring electrodes was calculated to be about 55 Ω in 0.5 M NaCl and 140

Ω in the coating bath.

It should be noted that similar experimental approaches have been used by Lunt et al. to study the interactions among localized corrosion sites on a 5 x 5 array of 316 stainless steel wires 25. Array-based studies have also been conducted by Tan to study the heterogeneous electrochemical processes on steel surfaces due to water droplet corrosion

26.

All the chemicals used in the study were ordered from commercial vendors and were of reagent grade. The distilled water, with a resistivity of 18 MΩ⋅cm, was used to make up all solutions.

Conversion coating and breakdown. A Model 900 MMA (Scribner and

Associates, Southern Pines, NC) was used to monitor the current on each electrode during CCC formation and breakdown. The MMA was used to measure each electrode element in the array individually. During coating formation, the current on each electrode was measured on a separate dedicated ZRA capable of measuring currents up to 1 µA with a resolution of 33 pA. All electrodes were electronically connected so that under open circuit conditions the net array current was zero. Data acquisition was controlled by software installed in a personal computer.

The coating bath used in this study contained 5.4 g/L CrO3, 0.9g/L K3Fe(CN)6 and 0.9g/L NaF, which is close to the Alodine 1200S bath 27. This chemistry was used to make all CCCs unless otherwise indicated.

In preparation for conversion coating, electrode arrays were polished to 600 grit using SiC paper. Arrays were then immersed in dilute nitric acid for 1minute to obtain a clean deoxidized surface. The electrode array was then connected to the MMA and immersed in the coating solution for various lengths of time. The current on each electrode was sampled at a rate of 50 Hz. After coating, the array was then rinsed in distilled water and dried with warm flowing air. After chromate coatings were formed on the electrode arrays, the coatings were allowed to dry in ambient air for 24 hours and then were connected to MMA and immersed in 0.5M NaCl. The current on each segment was monitored either at open circuit potential (OCP) or under anodic potentiodynamic polarization control. For experiments at OCP, Cu wire was present or absent as indicated in the result section.

In breakdown experiments, the entire array was operated as a working electrode in a three-electrode anodic polarization experiment. A built-in potentiostat was used to polarize the array. Measurements were made in a 0.5M NaCl solutions at a scan rate of

0.2 mV/s starting from OCP. All potentials reported are quoted versus the saturated calomel electrode (SCE). Breakdown (pitting) events were recorded on each electrode during the measurement until all electrodes in the array broke down.

To prevent crevice corrosion, a low viscosity epoxy (EPO-THIN by Buehlerâ) was used to mount the aluminum wire array. After the polarization, the electrode array was examined under an optical microscope. If breakdown occurred at the perimeter area of the electrode, that breakdown potential was discarded to exclude edge effects in the data sets. It should be noted that breakdown in the perimeter area did not necessarily correspond to a low breakdown potential in these electrodes.

Raman spectroscopy. The use of the peak at 859 cm-1 due to Cr6+-O-Cr3+ stretch for examining CCCs has been thoroughly documented 10. This peak was measured ex situ to study the evolution of CCCs. Raman spectra of CCCs were collected using a Chromex

2000 spectrometer, with a standard interference band reject filter and EEV 15-11deep depletion CCD. A 785 nm excitation and 180º backscattered sampling geometry were employed to obtain the Raman spectra. The instrument was frequency-calibrated with 4- acetamidophenol (Tylenol) and the intensity was calibrated with a glass sample that has known intensity-frequency curve. The area under the 859 cm-1 peak after baseline correction was used to indicate the amount of Cr6+ in CCCs.

3.3 Results

Coating formation. Current transients during CCC formation exhibit two distinct stages: 1) an initial 30 second period of intense electrochemical activity characterized by large net currents on electrode elements, and 2) a subsequent stage characterized by electrochemical quiescence in which net currents were very small. The polarity of the net current during the first stage of coating growth varied from electrode to electrode and indicated the dominant reaction on the electrode as well as the progress of coating formation. Some electrodes showed extensive, almost periodic oscillations between anodic and cathodic polarities. When Cu wire was introduced in the electrode array, the current evolution showed much less oscillation and fewer changes of polarity and the whole system appeared to be more stable during coating formation process. Figure 3.2 shows a frame from a movie collected on an electrode array that contained 4 Cu electrodes. In this frame, the red color indicates a net anodic current and the blue color indicates a net cathodic current. The intensity of the color shows the magnitude of the current. The square at the center of segment records the peak current that segment has experienced. Also in Figure 3.2 are links to three movies of current evolution collected on electrode arrays that contain different numbers of Cu wires. Figure 3.3 shows the current evolution during coating formation on an electrode exhibiting pronounced current oscillations. Other electrodes exhibited more or less persistent anodic or cathodic activity during the initial stage of CCC formation. The I-t behavior of these electrodes is shown in

Figure 3.4 and Figure 3.5, respectively.

Effect of supplemental bath ingredients on coating formation. NaF additions to the coating bath have a significant effect on the current evolution during coating formation. Figure 3.6 shows the current evolution when NaF is present or absent. When

NaF is absent, the maximum current observed is smaller than when NaF is present. More significantly, the current decreases more rapidly when NaF is absent. The current drops two orders of magnitude in 2-3 s, compared with 20-30 s needed for the same decrease when F- is present. This suggests a more rapid cessation of coating growth.

The effect of K3Fe(CN)6 additions on current evolution is shown in Figure 3.7.

The addition of K3Fe(CN)6 to CrO3 + NaF bath does not cause significant change in current vs. time behavior. It appeared that the time required for electrodes to passivate was decreased when K3Fe(CN)6 was present in the coating bath.

Effects of Cu wire on the behavior of coatings during exposure to 0.5 M NaCl soloution. During the immersion of CCC-coated electrode arrays in 0.5M NaCl at OCP, the current on each electrode was monitored with the MMA. When the electrode array consisted of all-aluminum wires, the OCP of the electrode array fluctuated in the range from –790 to –740mV (SCE), and there were very few current spikes observed on the background noise current which was on the order of 10-10A. Figure 3.8 shows the current vs. time behavior on one segment of an all aluminum electrode. In the 25 Al wire array, only three segments showed one or two current peaks on the background. Figure 3.9 is a detail of the taller peak shown in Figure 3.8. This peak showed a sharp rise in 1 s and

gradually returned to background current level in 4 s. This spike lasted about 5 s with a maximum current of 13.6 pA. These features are consistent with those of metastable pitting of aluminum. All spikes on the three segments in this array showed features similar to the one shown in Figure 3.9 and the current spikes observed lasted less than 6s, with a transient height range of 2 to 14 pA. Significantly different behaviors were observed when one Cu wire was introduced into the electrode array. The first difference was that the OCP of the electrode array ranged from –820mV to –650mV (SCE), wider than the OCP range of the all-aluminum electrode arrays. The second difference was that all the aluminum wire segments except one showed current spikes on current vs. time curve and the number of spikes was much greater than that on the aluminum segment in the all-aluminum array. Figure 3.10 shows a typical current vs. time curve on Al electrode when Cu was connected in the array. The third difference is that the peak current values had a much wider distribution, ranging from 10 pA to 300 pA. The duration of these spikes were generally less than 10 s, but spikes as along as 100 s were also observed. Figure 3.11 shows the current spikes in more detail. They have features very similar to those observed in all-aluminum arrays, except that the peak currents and lifetimes are different. Another interesting observation is that at the moment when one segment was at the peak of its current transient, all the remaining segments showed net cathodic current. When a stable pit finally developed on one segment, all the other segments exhibited net cathodic current, and there were no further current transients observed after that on any segment.

Effect of coating time on CCC breakdown. Anodic polarization curves were collected separately for each element in the array after the conversion coating was dried in air for 24 ± 1hrs. During potentiodynamic polarization, the CCC breaks down locally and stable pitting develops on the substrate. Pitting is detected as a sharp break in the polarization curve of the electrodes. Metastable pitting was not usually detected. The potential at which this break occurs is termed the “breakdown” potential. Figure 3.12 shows a typical anodic polarization curve on one electrode element. The breakdown potentials of two electrode arrays totaling more than 40 measurements were collected for each distinctive coating condition.

Each set of breakdown potential data was plotted as cumulative probability versus potential. Because breakdown potential data are usually scattered, the cumulative probability plotting approach is a good way to illustrate the distribution in the measurement population 20,21,28. Some otherwise indistinguishable trends in the breakdown potentials can be clearly seen in these types of plots.

Coating time has a significant effect on measured breakdown potential distribution (BPD). CCCs are usually formed by 1 to 3 minutes of immersion. In Figure

3.13, it can be seen that breakdown potentials increase with coating time up to 2 minutes.

Increasing coating time from 2 min to 5 min does not increase the breakdown potential significantly. Even a few seconds of immersion showed marked shift in the BPD to more positive potentials. Dramatic increases in breakdown resistance were achieved within the first 30 s of immersion, where the median breakdown potential increased 0.25 V from –

0.74 V for bare Al to –0.49 V. By comparison, the median breakdown potential increased by only an additional 90 mV when coating time was increased from 30 s to 300 s.

Effect of formation current polarity on CCC breakdown. During the first stage of coating formation, two primary reactions occur on each electrode: aluminum oxidation and chromate reduction. When aluminum oxidation is dominant, the electrode acts as a net anode. When chromate reduction is dominant, the electrode acts as a net cathode. Therefore, the difference in polarity of the current on each electrode during coating formation may be associated with a change in coating composition or thickness and thus a change in the resistance to breakdown. The data in Figure 3.14 support this idea. In this figure, breakdown potential data from conversion-coated electrodes were segregated according to the net current characteristics observed during the first stage of a

3-minute immersion in a CCC bath. “Net anodes” exhibited predominantly anodic current during the first 30 seconds of coating formation, “Net cathodes” exhibited predominantly cathodic current, and “Mixed character” electrodes exhibited significant currents of both polarity. Comparison of these BPDs shows that net cathodes are possibly more resistant to breakdown than electrodes of mixed current character, and are much more resistant than electrodes that were net anodes.

Effect of supplemental bath ingredients on CCC breakdown. Of the three main components of CCC bath, NaF has been classified as activator, and K3Fe(CN)6 as

18 an accelerator . Figure 3.15 shows the effect of NaF and K3Fe(CN)6 additions on the

BPDs for a fixed coating time of 2 minutes. CCCs formed in a CrO3-only solution increase the median breakdown potential only slightly; approximately 0.01V over bare

Al. When NaF is added to the bath the median breakdown potential increases by about

0.23 V over the CrO3-only median. When K3Fe(CN)6 is added, the increase is about 0.04

V. These results illustrate the importance of these supplemental bath ingredients on CCC formation. Without these additions, it is likely that coatings with useful levels of corrosion protection do not form.

Both NaF as an activator and K3Fe(CN)6 as a redox accelerator might be expected to exert their greatest effect on corrosion protection during the first stage of CCC formation where apparent electrochemical activity is greatest. To examine this possibility, CCCs were formed in the presence and absence of NaF and K3Fe(CN)6 for 30 s when all of the electrochemical activity takes place. Complementary coating experiments were performed where the coating time was fixed at 2 minutes. In this case, most of the coating growth is expected to occur during the electrochemically quiescent stage of growth. BPDs were measured on coatings formed in these experiments and are shown in Figures 3.15 and 3.16. In these figures, the BPD for bare Al and the 30-second and 2-minute coatings containing CrO3, NaF, and K3Fe(CN)6 are the same data sets. In

Figure 3.16, the BPDs are identical for coatings formed in the absence of NaF suggesting that the action of K3Fe(CN)6 to improve corrosion protection is complete in 30s. This is not the case for NaF. Figure 3.17 shows BPDs for CCCs formed in the absence of

K3Fe(CN)6. The 2-minute BPD is shifted considerably to more positive potentials suggesting that NaF acts over the entire coating formation period to increase resistance to

breakdown. NaF and K3Fe(CN)6 clearly work together to improve resistance to breakdown as BPDs for coatings formed in the "full" chemistry exhibit the most noble

BPDs by far.

Alternate accelerators. To explore the role of the accelerator in CCC formation further, K3Fe(CN)6 was replaced with another possible accelerator, Fe(NO3)3. In this case, the Fe3+/Fe2+ couple was intended to serve as the redox mediator. In this coating

3+ 3- bath the molar concentration of Fe was made identical to that of Fe(CN)6 and the

BPDs of the coatings formed in these two chemistries were compared. The results are shown in Figure 3.18. The coatings formed in Fe(NO3)3-containing bath have much lower breakdown potentials than those formed in K3Fe(CN)6 -containing bath and show only very limited improvement over bare Al. Compared with coatings formed in CrO3 + NaF bath, Fe(NO3)3 shows adverse effect on coating breakdown potentials.

Effect of minor bath ingredients on CCC breakdown. It is of interest to know how the coatings formed in simulated Alodine bath perform compared with those formed in actual Alodine 1200S bath. By comparing the BPD from the simulated Alodine coatings to coatings made from commercial product we were able to assess the influence

27 of other minor ingredients, such as KBF4 and K2ZrF6 . Figure 3.19 shows the BPD of coatings formed in simulated and actual Alodine 1200S solution. The data show that the coatings formed in simulated Alodine are in fact better than coatings formed in actual

Alodine 1200S; at least in terms of breakdown potentials. It has also been observed that

CCCs formed in simulated chemistry perform better than those formed with Alodine

20 1200S in salt spray tests . In these experiments, the minor ingredients such as KBF4 and

K2ZrF6 in Alodine 1200S do not appear to have a significant influence on the breakdown resistance of CCCs.

Cr6+ Concentration in CCC Determined by Raman Spectroscopy. Figure 3.20 shows Raman spectra of conversion coatings on pure Al in the 859 cm-1 region. Figure

3.21 shows the peak area integral, which was taken as a measure of the scattering intensity of the Cr6+-O-Cr3+ stretch in the CCC structure. A representative net current transient measured during CCC formation is superimposed on the plot. Data for CCCs on pure and Al and AA2024-T3 substrates are reported. Each data point in Figure 3.21 is the average of four measurements on the same sample at different locations. As coating time increases, the amount of Cr6+ in the CCCs increases. This finding is generally consistent with XANES results, which also show that the total Cr and Cr6+ concentration in the CCC increases over this time frame 21,29. The data in Figure 3.21 indicate significant CCC evolution (growth and/or chemistry change) in the absence of measurable electrochemical current, suggesting the possibility that a major episode of CCC formation may not be electrochemical in nature.

3.4 Discussion

CCC Formation Process on Al. CCC formation on Al is commonly described as an electrochemical process involving oxidation of Al and reduction of Cr6+ to Cr3+ 3,11,30:

Al → Al3+ + 3e- Eqn. 3.2

+ - Al + 2H2O → AlOOH + 3H +3e Eqn. 3.3

2- + - Cr2O7 + 8H + 6e → 2Cr(OH)3↓ + H2O Eqn. 3.4

Coating growth occurs when Cr3+ hydrolyzes, condenses and polymerizes on the aluminum surface to form an amorphous, hydrous layer 10. While the trivalent chromium hydroxide forms on the electrode surface from the coating bath, chromates are adsorbed onto it 10. Chromates are known to adsorb strongly onto many oxides and hydroxides 31-

33, and adsorption of chromate on the Cr3+ hydroxide is likely to be favored in the acidic environment of coating bath because the adsorption reaction consumes protons 10. In service environments which are less acidic, desorption is favored, leading to self-healing characteristics 11,17,29,34,35.

Figure 3.21 raises the possibility that a significant component of CCC growth can be chemical in nature. Specifically, the figure shows that intense net currents are measured on the array for only about 30 seconds during coating formation. However, the

859 cm-1 Raman band intensifies over the entire 300 second coating formation interval.

Arguably, this result is equivocal with respect to non-electrochemical film growth in later stages of the coating process because the current measured on an electrode is a net current. In other words, no or low net current does not necessarily mean no or low electrochemical activity on an electrode. Nonetheless, the result of CCC formation is electrode passivation. Figure 3.21 suggests that electrochemical passivation may be largely complete early in the coating process. Provided that sufficient Cr3+ is produced by electrochemical reduction and retained in the electrolyte near the metal surface in the

early stages of the coating process, continued CCC growth by Cr3+ hydrolysis, polymerization and condensation, combined with adsorption of chromate 10 would account for continued evolution of the CCC in the latter stages of the coating process even though the Al surface is electrochemically passive. In any case, these results enable

CCC formation to be divided into two distinct stages: one characterized by intense measurable electrochemical activity, and a second that occurs under comparatively quiescent conditions.

Lateral coating heterogeneity. The results in Figure 3.14 shows that net cathodes are more resistant to breakdown than net anodes. Further characterization of the coatings formed on net cathodes and net anodes is necessary to understand why they behave differently. However, it is expected that net cathodes support Cr6+ reduction at higher rates and are therefore richer in Cr hydroxide than net anodes. Cr hydroxide enrichment might reasonably be expected to translate into increased corrosion resistance.

If this is so, CCC formation on engineering alloys probably occurs unevenly because anodic and cathodic activity is localized by microstructural heterogeneity. This may result in regions of differing corrosion resistance. In microtomed cross sections on Al alloys, Brown et al. have found Cr-rich deposits on isolated regions of Fe surface

36 enrichment after conversion coating in CrO3-NaF solutions .

Effect of Cu wire on coating formation and breakdown and interactions among pits. When Cu electrodes were introduced into the electrode array, the current evolution

showed less oscillation and fewer polarity changes during coating formation. Essentially, electrodes tended to exhibit a single polarity throughout coating formation. Considering the fact that net cathodes and net anodes have different resistance to breakdown, the introduction of Cu may result in less uniform formation and thus less uniform coating properties. When a Cu wire was present in a conversion coated electrode array, the coated

Al electrodes exhibited much higher tendency for metastable pitting. The number of metastable pitting events, the lifetimes of these events and the heights of the associated current transients were increased by the presence of a Cu element in the electrode array.

The main reason for this difference is that the OCP of the whole conversion coated electrode array was increased when Cu wire was present. Without Cu wire, the OCP ranged from –790 mV to –740 mV during the immersion period. The OCP range changed to –820 mV to –650 mV when Cu wire was included. This brought the OCP closer to the pitting potential of coated Al and enhanced metastable pitting events. This result may partially explain why the protectiveness of CCC decreases with Cu content in aluminum alloys.37 As the Cu content in an aluminum alloy increases, the OCP generally increases under the same environmental conditions. This increase in OCP decreases the span between OCP and coating breakdown potential and increases the chances for metastable pitting and subsequent stable pitting.

Another phenomenon observed in this study was the interaction between electrodes that showed metastable pits or stable pits and the remaining electrode segments. When a metastable pit was at its peak current or a stable pit developed on one segment, the remaining segments were rendered as cathodes. The effects of active pits on

subsequent pitting events in nearby regions have been studied by Lunt et al. 38. Three types of interactions have been proposed: the ohmic potential drop near an active pit, the composition change in local solution, and the action on the passive film by the altered solution. The ohmic potential drop near an active pit inhibits pitting in surrounding regions. The change in local solution composition and the change in nearby passive film due to the locally altered solution enhance subsequent pitting. The current observations were consistent with ohmic potential drop mechanism. When a metastable pit is at its peak current (anodic), the remaining area provides the cathodic current that is needed to balance the anodic pit current. It appears that the magnitude of the balancing current on an individual electrode segment does not depend on the distance between that electrode and the metastable pits. When a stable pit developed on one electrode segment, all other electrode segments in the array were providing the supporting current and were under cathodic protection by the stable pit. This persistent cathodic current due to a stable active pit elsewhere hindered pitting development.

Effects of Coating Bath Chemistry on Coating Formation and Breakdown

Behavior. The functions of NaF and K3Fe(CN)6 in coating formation have been studied by several groups and theories have been put forward describing their role in CCC formation. It has been observed that without F- coating formation is very slow. It has been suggested that F- dissolves the oxide film initially present on Al surface and activates the surface for chromate reduction 3,8. The reaction has been proposed to occur as:

Al2O3 + 6HF → 2AlF3 + 3H2O ( eq. 5 ).

During coating growth, F- may also delay film formation in a manner that sustains electrochemical reactions. Without F-, the surface rapidly passivates, and film growth stops having only formed a very thin film with limited corrosion protection. The results from this experiment support the above arguments. As shown in Figure 3.6, when F- is absent, the peak current during coating formation is generally lower and the current drops very sharply in the first several seconds of coating formation. The smaller area under the curve measured in the absence of F- indicates that the extent of the coating formation reaction is reduced compared to when F- is present. Once formed, this thin layer prevents further contact between coating solution and Al surface and coating growth slows or ceases. Apparently, the thin coating that is formed is not very protective in Cl- solutions.

The role of F- in promoting electrochemical reactions can be seen very clearly in potentiodynamic polarization experiments. As shown in Figure 3.22, the corrosion rate at the OCP when F- is present is three orders of magnitude greater than that when F- is absent. A feature in the cathodic polarization curve in the presence of F- is that the current decreases with increased cathodic overpotential, suggesting coating formation on the Al surface as the scan is carried out. This was confirmed by SEM observations.

Figure 3.23 shows the surface morphology of pure Al after immersion in different coating

bath chemistries. Without F- in the coating bath, there is little change on Al surface. With all three ingredients, the familiar mud-cracking pattern of CCCs was observed on Al.

It has been reported that when ferricyanide is added to coating bath, the coating weight, coating thickness, formation rate, and coating corrosion resistance are increased

2,8,9,18,39 3- . Comparison of the BPDs with and without Fe(CN)6 indicates that the coating

3- corrosion resistance is indeed increased when Fe(CN)6 is added to coating bath. The function of ferricyanide has been examined by Xia 19 who suggests that the sluggish

3- oxidation of Al by chromate is greatly increased because Fe(CN)6 rapidly oxidizes Al.

3- 4- The reduction product of Fe(CN)6 , Fe(CN)6 , reduces chromate to complete the mediation cycle.

3- The results of BPD measurement with and without Fe(CN)6 in the coating bath

3- are consistent with this mediation mechanism. According to this mechanism, Fe(CN)6 mainly affects electrochemical reactions, and it might be expected that its effect on coating BPD would be more pronounced for coatings formed in 30 s than coatings formed in 2 minutes since electrochemical activity mainly occurs in the first 30 s. Results in Figure 3.17 indicate that the difference in the median breakdown potentials due to the

3- presence or absence of Fe(CN)6 for 30-second coatings is about 0.07 V larger than that between 2-minute coatings. Xia further suggests that any redox system with a redox potential between that of Cr6+/Cr3+ and Al0/Al3+ and fast kinetics with those two systems can act as mediator. A suggested list of possible mediators included Fe3+/Fe2+. The results from this study suggest that Fe3+/Fe2+ system does not improve the corrosion resistance of

CCC. On the contrary, it seems that the addition of Fe3+ has detrimental effect on coating

performance, which may be due to precipitation of Fe(OH)2 in locally neutral or alkaline conditions. This indicates additional requirement for a redox mediator is that both states

3- 4- of the mediator must be highly soluble. This is true for Fe(CN)6 /Fe(CN)6 but not true for Fe3+/Fe2+, because the solubility of Fe2+ is low under the conditions that exist near the substrate surface.

From the results shown in Figures 3.14-16 and the discussion above concerning the effects of NaF and K3Fe(CN)6 on CCC formation, it is concluded that, to form a corrosion resistant conversion coating, the addition of appropriate supplemental ingredients to the coating bath chemistry are as important as the primary film forming agent itself. In the case of CCCs, the film-forming agent, CrO3, is necessary to form

- 3- CCC. But without the addition of F and Fe(CN)6 (or other chemicals in different CCC processes), a corrosion resistant coating will not form. This idea might be important in developing chromate-free coating systems.

Certain types of cerium conversion coatings are examples of this idea already in practice. Soluble Ce is known to form protective coatings on Al alloys 40~44. However, the formation of films with latent corrosion protection requires tens to hundreds of hours of exposure to Ce solutions. Chemical acceleration with H2O2 additions is extremely effective in reducing coating time. Corrosion resistant Ce coatings can be formed in a

43 matter of minutes by immersion in H2O2-modified Ce coating baths . Although the chemistry of Ce conversion coating formation is very different from that of CCCs, it seems that chemical "acceleration" and "activation" might be quite useful as general concepts in conversion coating development.

CCC breakdown, small area electrodes and anodic inhibition by CCCs.

Pitting potentials (or breakdown potentials) are naturally dispersed. Increasingly, pitting and breakdown potentials are being represented with distribution plots 20,21,28. In order to construct such a plot, many tests are necessary to accurately describe the distribution. The

MMA is useful in this regard because it functions essentially as a multiplexer allowing many polarization curves to be collected simultaneously.

Coating breakdown and pitting are dominated by surface defects. These defects may be in the substrate or in the coating, but in any given sample, they exist with some characteristic areal density. Assuming a characteristic areal defect density among similarly prepared electrodes, BPDs are expected to shift to more positive potentials for decreasing electrode area as the likelihood of having a defect that initiates breakdown at a low potential decreases 45,46. Essentially, the chances of breakdown at low potentials are greater for large area electrodes than for small area electrodes. The BPDs shown in this chapter are believed to capture the breakdown behavior of large area electrodes measured in polarization experiments at the "foot" of the distributions, which all tend to converge

(e.g. Fig. 8). To explore this idea, replicate anodic polarization curves were measured for

1 cm2 bare Al and Al electrodes coated with a 2-minute CCC. Anodic polarization curves were collected in 0.5 M NaCl and are shown in Figure 3.24. The breakdown potential for

Al is about -0.75 V and that for the CCC sample is -0.740 V. These values agree well with low end of the BPDs in Figure 3.13. The breakdown potentials are also within 10 mV of one another. It is interesting to speculate on whether anodic inhibition has been

unfairly overlooked as a contribution to CCC corrosion protection due to the use of large area electrodes in polarization testing.

Evaluation of BPDs in the manner described here to understand corrosion protection by CCCs may be more closely related to evaluation methods like electrochemical impedance spectroscopy (EIS) 22, or visual examination of samples subject to cabinet exposure testing 47. In these evaluations, results reflect the contributions of the entire surface and not just the first breakdown event.

Another important aspect of breakdown potential testing is environmental aggressiveness. Chloride ion concentration may be an important factor in determining whether evidence of anodic inhibition is detected or not. One may conclude that CCCs do not inhibit anodic reactions on the basis of the polarization responses in Figure 3.24, which were obtained in 0.5M NaCl solution. However, anodic inhibition by CCCs on bulk samples is supported in results reported by Ilevbare, where the experiments were

20, 21 carried out in 0.1M Na2SO4 plus 0.005M NaCl solutions . These findings illustrate the need to consider environmental aggressiveness when interpreting evidence for or against anodic inhibition by CCCs.

3.5 SUMMARY

The main findings of this study are summarized as follows:

- In electrochemical measurements, CCC formation occurs in two stages. The first

stage occurs in the first 30 seconds of CCC formation and is characterized by

measurable electrochemical activity. The second stage occurs over the remainder of the

coating formation period and is characterized little measurable electrochemical activity.

Coating evolution continues through both of these stages as indicated by increases in the 859cm-1 Raman scattering band due to Cr3+-O-Cr6+ bonding in the CCC.

- The resistance to breakdown of an electrode in a conversion coated array is related to the polarity of the current during the first stage of coating formation.

Resistance to breakdown decreases in the order:

Net cathodes ~ Mixed polarity > Net anodes.

- These results confirm earlier findings that CCCs inhibit anodic reactions (20,

21). Increasing coating time increases anodic inhibition as indicated by increasing

BPDs. Most of the improvement in inhibition occurs in the first 30s of coating time.

-NaF and K3Fe(CN)6 both have significant positive effects on CCC breakdown resistance. NaF appears to have the larger effect of the two ingredients. Together, these ingredients vastly improve the latent corrosion protection properties of CCCs.

REFERENCES

1. J. A. Treverton and N. C. Davis, Met. Technol., Oct., 480 (1977).

2. J. A. Treverton and N. C. Davis, Surf. Interface Anal, 3, 194 (1981).

3. H. A. Katzman, G. M. Malouf, R. Bauer and G. W. Stupian, Applica. Surf.

Sci., 2, 416 (1979).

4. K. Asami, M. Oki, G. E. Thompson, G.C. Wood and V. Ashworth,

Electrochim. Acta, 32, 337 (1987).

5. F. W. Lytle, R. B. Greegor, G. L. Bibbins, K. Y. Blohowiak, R. E. Smith and

G. D. Tuss, Corros. Sci., 37, 349 (1995).

6. Z. Yu, H. Ni, G. Zhang, Y. Wang, Appl. Surf. Sci., 62, 217 (1992).

7. H. E. Townsend and R. G. Hart, J. Electrochem. Soc., 131, 1345 (1984).

8. P. L. Hagans and C. M. Haas, Surf. Interface Anal., 21,65 (1994).

9. A. E. Hughes, R. J. Taylor and B. R. W. Hinton, ibid, 25, 223 (1997).

10. L. Xia and Richard McCreery, J. Electrochem. Soc., 145, 3083 (1998).

11. M. W. Kendig, A. J. Davenport and H. S. Isaacs, Corros. Sci., 34, 41 (1993).

12. T. Biestek and J. Weber, translated from the Polish by A. Kozlowski,

Electrolytic and Chemical Conversion Coatings, Portcullis Press Limited-

Redhill, Poland (1976).

13. W. E. Pocock, Met. Finish., 52, n12, 48 (1954).

14. N. Sato, Corrosion, 45, 354 (1989).

15. N. Sato, Corros. Sci., 31, 1 (1990).

16. J. S. Wainright, O. J. Murphy and M. R. Antonio, Corros. Sci., 33, 281

(1992).

17. J. Zhao, G. Frankel, and R. L. McCreery, J. Electrochem. Soc., 145, 2258

(1998).

18. N. J. Newhard, Jr., Met. Finish., n7, p. 49, n8, p.66 (1972).

19. L. Xia, and McCreery, J. Electrochem. Soc., 146, 3696 (1999).

20. G. O. Ilevbare, J. R. Scully, J. Yuan, and R. G. Kelly, Corrosion, 56, 227

(2000).

21. G. O. Ilevbare, and J. R. Scully, J. Electrochem. Soc., 148, B196 (2001).

22. R. G. Buchheit, M. Cunningham and H. Jensen, Corrosion, 54, 61 (1998).

23. A. J. Bard, and L. R. Faulkner, Electrochemical Methods: Fundamentals and

applications, John Wiley, New York (2001).

24. J. Newman, J. Electrochem. Soc., 113, 501 (1966).

25. T. T. Lunt, V. Brusamarello, J. R. Scully, and J. L. Hudson, Electrochem.

Solid-State Lett., 3, 271 (2000).

26. Y. Tan, Corros. Sci., 41, 229 (1999).

27. Material Safety Data Sheet for Alodine 1200S, Henkel Corporation.

28. T. Shibata and T. Takeyama, Corrosion, 33, 243 (1977).

29. M. W. Kendig, A. J. Davenport and H. S. Isaacs, Corros. Sci. 34, 41 (1993).

30. J. R. Waldrop and M. W. Kendig, J. Electrochem. Soc., 145, n1, L11 (1998).

31. J. M. Zachara, C. E. Cowan, R. L. Schmidt and C. C. Ainsworth, Clays Clay

Miner., 36, n4, 317 (1988).

32. C. H. Weng, J. H. Wang and C. P. Huang, Water Sci. Technol., 35, n7, 55

(1997).

33. C. H. Weng, C. P. Huang, H. E. Allen, P. B. Leavens and P. F. Sanders,

Environ. Sci. Technol., 30, 371 (1996).

34. A. L. Glass, Mater. Prot., July, 26 (1968).

35. L. Xia, E. Akiyama, G. S. Frankel, and R. L. McCreery, J. Electrochem. Soc.,

147, 2556 (2000).

36. G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson, G. C. Wood,

Corrosion Sci., 33, 1371 (1992).

37. S. J. Ketcham (1963). Properties of Chemical Films on Aluminum Alloys.

The Finishing of Aluminum. G. H. Kissin. New York, NY, Reinhold

Publishing Corporation.

38. T. T. Lunt, J. R. Scully, V. Brusamarello, A. S. Mikhailov, and J. L. Hudson,

J. Electrochem. Soc., 149, B163 (2002).

39. P. L. Hagans and C. M. Hass, “ Chromate Conversion Coatings”, p.405, ASM

Handbook, Vol. 5, Surface Engineering, ASM International, Metals Park, OH

(1994).

40. B. R. W. Hinton, D. R. Arnott and N. E. Ryan, Met. Forum, 7, 211 (1984).

41. B. R. W. Hinton, J. Alloys Compd., 180, 15 (1992).

42. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Corrosion, 45, 615 (1989).

43. F. Mansfeld, Y. Wang, and H. Shih, J. Electrochem. Soc., 138, L74 (1991).

44. F. Mansfeld, Y. Wang, and H. Shih, Electrochim. Acta, 37, 2277 (1992).

45. H. Bohni, T. Suter, and A. Schreyer, Electrochim. Acta, 40, 1361 (1995).

46. G. T. Burstein, and G. O. Ilevbare, Corros. Sci., 38, 2257 (1996).

47. MIL-C-5541E, “Chemical Conversion Coatings on Aluminum and Aluminum

Alloys” (1990).

FIGURES

~1mm center

500µm dia.

Figure 3.1 Photographs of electrode array used in this study.

100

( a ) ( b )

( c ) ( d )

Figure 3.2 (a) A frame from a current evolution movie during coating formation. (b) a movie for a 21-Al-4-Cu array. (c) a movie for a 24-Al-1-Cu array. (d) a movie for a 25Al array.

101

8 10 -7 Electrode: Al

-7 6 10 Coating Solution: Cr + F + Fe

4 10 -7

2 10 -7 Current (A) Current 0

-2 10 -7 start of immersion

-4 10 -7

25 30 35 40 45 50 55 60 65 Time (s)

Figure 3.3 Representative current vs. time behavior for an electrode that exhibited distinctcurrent oscillations during early CCC formation. In the coating solutio notation, Cr, F and Fe stand for CrO3, NaF and K3Fe(CN)6, respectively. The same notations are used for the remaining figures of this chapter.

102

1 10 -6

Electrode: Al 8 10 -7 Coating Solution: Cr + F + Fe

6 10 -7

4 10 -7 Current (A) 2 10 -7

0 Start of immersion

-2 10 -7

0 50 100 150 200 250 Time (s)

Figure 3.4 Representative current vs. time behavior for electrodes exhibiting persistent anodic behavior during early CCC formation.

103

2 10 -7

-7 -2 10 Start of immersion

-4 10 -7

-7 Current (A) Current -6 10 Electrode: Al

-8 10 -7 Coating Solution: Cr + F + Fe

-1 10 -6 10 15 20 25 30 35 40 45 50 Time (s)

Figure 3.5 Representative current vs. time behavior for an electrode exhibiting persistent cathodic behavior during early CCC formation.

104

1 10-6

8 10-7 Coating solution: Cr + Fe + F 6 10-7

4 10-7 Current (A) 2 10-7

0 Co ating so lu tio n: C r + Fe S tart of imme rsio n -2 10-7 0 5 10 15 20 25 30 35

Time (s)

Figure 3.6 The effect of F- on current evolution on Al during early CCC formation.

105

1 10-6

8 10-7 Coating solution: Cr + F 6 10-7

4 10-7 Current (A) 2 10-7

Coating solution: Cr + F + Fe 0

Start of immersion -2 10-7 0 5 10 15 20 25 30 35 40 Time (s)

3- Figure 3.7 Effect of Fe(CN)6 on current evolution on Al during CCC formation.

106

-8 1.4 10

-8 1.2 10

-8 1.0 10

-9 8.0 10

-9 6.0 10

-9

Current ( A ) ( Current 4.0 10

-9 2.0 10

0 0.0 10

-9 -2.0 10 0 10000 20000 30000 40000 50000 60000 70000 80000

Exposure time ( s )

Figure 3.8 The current vs. time curve on one electrode segment during immersion at OCP in 0.5M NaCl when no Cu was present in the electrode array.

107

-8 1.4 10

-8 1.2 10

-8 1.0 10

-9 8.0 10

-9 6.0 10 Current ( A ) -9 4.0 10

-9 2.0 10

0 0.0 10 41230 41240 41250 41260 41270 41280

Exposure time ( s )

Figure 3.9 Detail of the larger current spike in Figure 3.7.

108

Figure 3.10 The current vs. time curve on one electrode segment during immersion at OCP in 0.5M NaCl when one Cu wire was included in the electrode array.

109 -7 1.0 10

-8 8.0 10

-8 6.0 10

-8 4.0 10 Current ( A ) ( Current

-8 2.0 10

0 0.0 10

-8 -2.0 10 38020 38040 38060 38080 38100 38120 38140

Exposure time ( s )

Figure 3.11 A metastable pit that lasts 70s during immersion in 0.5M NaCl at OCP when a Cu wire was included in the electrode array.

110

1.2 10 -6

1 10 -6

8 10 -7

6 10 -7

-7 Current (A) Current 4 10

-7 2 10 E pit

-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4

Potential vs SCE (V)

Figure 3.12 A typical anodic polarization curve on Al wire electrode coated with CCC in 0.5M NaCl solution.

111

1.2 bare Al 4s 15s 30s 60s 120s 300s

0.8

0.6

0.4 Cumulative Probability.

0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Breakdown potential (V vs SCE)

Figure 3.13 Effect of coating time on CCC breakdown potential distribution. Coating solution: CrO3 + NaF + K3Fe(CN)6.

112

0.8 Net anodes

0.6

Ne t ca tho des 0.4

Cumulative Probability Mixed character 0.2

0 -0.7 -0.65 -0.6 -0.55 -0.5 -0.45 -0.4 -0.35 -0.3 Breakdown potential (V vs. SCE)

Figure 3.14 Effect of polarity during coating formation on coating breakdown potential distribution. Coating solution: CrO3 + NaF + K3Fe(CN)6. Coating time: 3min.

113

1.2 Coating solution: Coating solution: Bare Al Cr + Fe Cr + Fe + F

30s 0.8

30s 2min 0.6

2min 0.4 CumulativeProbability

0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2

Breakdown potential (V vs SCE)

Figure 3.15 Effect of NaF on coating breakdown potential distribution.

114

1.2 Coating solution: C oating solution: Bare Al Cr + F Cr + F + Fe 1

30s 0.8 30s 2min

0.6

2min 0.4 Cumulative probability 0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Breakdown potential (V vs SCE)

Figure 3.16 Effect of K3Fe(CN)6 on coating breakdown potential.

115 1.2 Co ating solution Bare Al C r o nly C r + Fe C r + F C r + F + Fe 1

0.8

0.6

0.4 Cumulative Probability

0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Breakdown Potential (V vs SCE)

Figure 3.17 Relative contributions of coating bath components to coating breakdown potential. Coating time: 2min.

116 1.2 Coa ting solution Bare Al Cr + F + Fe(NO) Cr + F Cr + F + Fe 3 3 1

0.8

0.6

0.4 Cumulative Probability

0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Breakdown Potential (V vs SCE)

Figure 3.18 Effect of Fe(NO3)3 on coating breakdown potential distribution. Coating time: 2min.

117

1.2 Simulated Alodine 1200S Bare Al Alodine 1200S Alodine 30s 2min 30s 1

0.8

0.6

Simulated 0.4 Alodine 2min Cumulative Probability

0.2

0 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 Breakdown Potential (V vs SCE)

Figure 3.19 Comparison of breakdown behavior of CCCs formed in simulated Alodine and an Alodine 1200S solution prepared according to manufacturer’s specifications.

118

0.0045

0.004 5 min 0.0035 3 min 0.003 30s 1 min 0.0025 15s Intensity (AU) 0.002

0.0015

0.001 700 800 900 1000 1100

Raman shift (cm-1)

Figure 3.20 The intensity of the 860cm- band in Raman spectra increases with coating time, indicating the build-up of Cr6+in CCCs. Coating solution CrO3 + NaF + K3Fe(CN)6.

119

1 10-6 0.5

-7 8 10 0.4

on pure Al (AU) Intensity Raman

-7 6 10 0.3

-7 4 10 0.2 Current (A) on 2024-T3 -7 2 10 0.1

0 100 0

-50 0 50 100 150 200 250 300 350

Time (s)

6+ Figure 3.21 Cr in CCC formed in different times. Coating solution: CrO3 + NaF + K3Fe(CN)6.

120

0.4

0.2

0 Solution: Cr + Fe + F -0.2

-0.4

-0.6

Potential (V vs.SCE) Solution: Cr + Fe -0.8

-1 10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3

Current (A)

Figure 3.22 Polarization response of pure Al in simulated Alodine solution with and without F- additions.

121

( a ) ( b )

5µm

( c )

Figure 3.23 Morphology of coatings formed in different chemistries. (a: bare Al. b: in Cr + Fe. c: in Cr + Fe + F. Coating time: 2min.)

122

-0.5

-0.6

-0.7

CCC coated Al -0.8

-0.9 bare Al

Pot ential (V vs. SCE) -1

-1.1

-1.2 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2

Current density (A/cm2)

Figure 3.24 Anodic polarization curves on 1cm2 Al samples in 0.5M NaCl.

123

CHAPTER 4

THE EFFECT OF AMBIENT AGING ON INHIBITION OF OXYGEN

REDUCTION BY CHROMATE CONVERSION COATINGS

4.1 Introduction

Despite the wide usage of CCCs, elements of chromate corrosion protection have not been fully described. For this reason, there has been increased interest in understanding how chromates and CCCs promote corrosion resistance and paint adhesion so that they might be replaced with functionally equivalent, environmentally friendly substitutes.

On the basis of recent studies, CCC corrosion protection appears to be due to several different components. These include suppression of anodic reactions 2,3, suppression of cathodic reactions 4, and self-healing 5. CCCs consist largely of insoluble chromium hydroxides, which prevent contact of the environment with the substrate. In this sense they also provide barrier protection 6.

Soluble chromates inhibit oxygen reduction on aluminum and lower the corrosion rate under free corrosion conditions in aerated aqueous chloride solutions 7. Oxygen reduction reaction (ORR) inhibition appears to be due to monolayer coverage of metal

124

surfaces by Cr(III) hydroxide, which is adsorbed and reduced from solution containing soluble Cr(VI) 2,8,9.

Reports on ORR inhibition by CCCs are mixed. Clark et al. observed that a monolayer of chromate reduction product suppressed ORR 8, and it seems reasonable to assume that CCCs would also inhibit ORR since CCCs are normally several thousand times thicker than a chemisorbed monolayer. Ilevbare suggested that CCCs on 2024-T3 substrates do not inhibit ORR in an aerated 0.1 M Na2SO4 + 0.005M NaCl solution based on experiment results on CCCs that were aged between 48 –120 hours 10. Hughes has shown that whether CCCs inhibit ORR or not depends on coating age 4. In the Hughes study, cathodic polarization curves were collected on conversion coated 2024-T3 substrates aged in air then tested in aerated 3.46% NaCl solution. ORR inhibition increased during the first 40 hours after CCC formation, then decreased. After 165 hours of air aging, ORR inhibition was almost completely absent.

Hughes noted that an increase in shrinkage cracking with aging and speculated that loss in ORR inhibition and corrosion resistance was related to gradual immobilization of Cr6+. Loss in Cr6+ leaching eliminated healing of emerging shrinkage cracks and led to lower corrosion resistance.

In the present study, cathodic polarization curves on chromate conversion coated

AA2024-T3 (Al-4.4Cu-1.5Mg-0.6Mn) were collected systematically as a function of coating age where coatings were aged in ambient laboratory air or at elevated temperature. The cathodic polarization response is generally consistent with earlier studies in that freshly formed CCCs were found to inhibit the ORR, and that ORR

125

inhibition decreased with increased aging time. The transition from inhibition to loss of inhibition occurred within the time frame of one week. On the basis of CCC morphology and an examination of Cr6+ leaching, the loss in ORR inhibition is attributed mainly to the evolution of shrinkage cracking in the coating due to CCC dehydration.

4.2 Experimental procedures

Aluminum alloy 2024-T3 sheet (Al-4.4Cu-1.5Mg-0.6Mn, solution heat treated, cold water quenched and naturally aged) with a thickness of 3 mm was used as the substrate in all experiments. Samples with dimensions of 20 x 20 mm were mounted in epoxy. Mounted samples were polished in water to 600 grit and then polished in alcohol with 800 grit and 1200 grit abrasive paper. The polished samples were ultrasonically cleaned in alcohol before being immersed in the conversion coating bath.

The conversion coating bath was obtained by dissolving 7.55 g Alodineâ 1200S in 1 liter distilled water, which is the chemistry recommended by the manufacturer. The pH of the solution was about 1.6, which is in the range that produces good CCCs. The

2024-T3 samples were immersed in the coating bath for 3 minutes and then thoroughly rinsed with overflowing distilled water. The CCCs were dried with gentle air stream immediately following the water rinse.

CCCs were aged in laboratory air (RH = 55 – 65%, T = 19 - 23oC. The emphasis on ambient air is not trivial, as relative humidity greatly affects dehydration of CCC and its corrosion resistance 12 .) for various times from 1 to 168 hours before being cathodically polarized. Cathodic polarization curves were collected in aerated 0.5 M

126

NaCl or 0.1 M Na2SO4 + 0.005 M NaCl solutions at a scanning rate of 0.2 mV/s. All tests were carried out using a three-electrode cell configuration, and all potentials are reported versus a saturated calomel electrode scale (0Vsce = +0.242 Vshe). Prior to the start of cathodic polarization, the samples were exposed to test solutions for 10 minutes or 1.5 hours, as indicated.

CCC cross sections were prepared using a FEI DB-235 dual beam focussed ion beam / SEM instrument. This instrument is outfitted with an ion beam focussing column for ion milling and scanning ion imaging, and an electron-focusing column for scanning electron imaging. To prepare cross sections, a thin metallic Pt layer was deposited over the region of interest to preserve a sharp edge and the coating surface. Then cross sectioning was carried out by milling a square-edged crater several micrometers deep in the sample using Ga+ ions. CCC cross sections were revealed on the faces of the crater, which were examined by scanning electron microscopy.

4.3 Results and discussion

ORR inhibition and CCCs. Cathodic polarization curves of CCCs aged in laboratory air different times are shown in Figure 4.1. The cathodic polarization was carried out in 0.5M NaCl. For comparison, the cathodic polarization curve of bare 2024-

T3 is also shown. The polarization curve of bare 2024-T3 is consistent with diffusion limited oxygen reduction from the corrosion potential at about -0.63V to -1.00V. At more negative potentials, there is a change in reaction control and the form of the curve suggests that the water reduction is the dominant reaction.

127

For conversion-coated surfaces aged for 1 and 20 hours, the shape of the polarization curve is considerably different than that of bare 2024-T3. These curves indicate that the reduction reaction is not strictly under diffusion control. At the corrosion potential, the reduction reaction rates are lower than for bare 2024-T3. Correspondingly, the corrosion rates are lower and the corrosion potential is shifted in the negative direction compared to the bare alloy. Reaction kinetics are somewhat slower for the sample aged for 20 hours than the 1 hour sample, but overall the results support the notion that CCCs suppress the ORR on Al alloys under free corrosion conditions. At increasingly negative applied potentials reduction reaction rates quickly exceed the diffusion limited reaction rate for bare 2024-T3. The origin of this response has not yet been determined.

After further aging, the cathodic polarization behavior changes considerably. The diffusion limitation in the oxygen reduction reaction, which is absent in the 1 and 20 hour samples, is evident in the sample aged for 120 hours and is pronounced in the one aged for 168 hours. As aging time increases, the corrosion potential and the corrosion rate increase. The diffusion limited reaction rate after 168 hours aging is essentially the same, if not slightly greater than that for bare 2024-T3 suggesting a complete loss in cathodic inhibition after this amount of aging time.

The change in corrosion rate during the above aging process is shown in Figure

4.2. These corrosion rates were determined by Tafel extrapolation applied to the cathodic polarization curve using Corrview software. The linear region on the cathodic polarization curve was fitted with straight lines and the current at open circuit potential

128

on the fitted line was taken as the corrosion rate. Figure 4.2 shows that initially the corrosion rates on CCC coated samples were below that on bare 2024-T3. However, the corrosion rates of coated samples reached or exceeded the corrosion rate of bare 2024-T3 after CCCs were aged for about one week, indicating the loss of effective corrosion inhibition under these conditions.

In experiments similar to those depicted in Figure 4.1, cathodic polarization curves were collected as a function of aging time over a 168-hour period with about 12- hour resolution. In these experiments (data not shown), the loss in ORR inhibition is readily apparent in the shape of the polarization curve after aging for 48 to 72 hours, and all evidence of inhibition generally disappears on CCCs for aging times over 144 hours.

Although the time of the transition of a CCC from its ORR inhibiting to non-inhibiting condition varies from experiment to experiment, the transition is always observed.

Cathodic polarization curves were also collected on selected samples that were exposed to solution for 1.5 hours to ensure that the short pre-conditioning period (10 minutes) did not create transient conditions such as changing open circuit potential on the electrode surface that affected the cathodic polarization response. In these experiments, there was no appreciable change in the polarization response suggesting that a 10-minute stabilization time is sufficient in to evoke a response similar to that of a longer exposure test.

In the above experiments, cathodic polarization curves were collected in 0.5 M

NaCl. In order to determine if this loss of cathodic inhibition is unique to this solution or not, tests were also carried out in 0.1 M Na2SO4 + 0.005M NaCl, which was used by

129

Ilevbare et al. in their study 10. The results are shown in Figure 4.3. The trend observed in

0.5M NaCl solution was observed again. The cathodic inhibition gradually disappeared with increasing aging time in air.

Shrinkage cracking and Cr6+ leachability. It is known that CCCs are a dynamic system 11. After coatings are formed on the substrate, several changes can occur during subsequent aging in air. These changes include dehydration, loss of mobility of Cr6+, and the development of shrinkage cracks.

It has been suggested that cathodic inhibition by CCCs originates with Cr6+ leached from the coating 4. In this scenario, Cr6+ is leached into solution, transported to local sites of high net cathodic activity, and reduced and adsorbed strongly enough to inhibit chemisorption and oxygen reduction 2,9,10,13. In this context, loss in ORR inhibition with ambient aging is attributed primarily to immobilization of Cr6+ and the inability to heal shrinkage cracks that form in the CCC 4. This notion is supported by visual inspection experiments that qualitatively assess leaching of Cr6+. Essentially,

CCCs retain their original golden color if Cr6+ has not leached, and become light yellow or clear if Cr6+ has been leached or reduced. Figure 4.4 shows optical macrographs of the samples after various lengths of ambient aging time and cathodic polarization.

Examination shows a gradual progression from significant color loss for short aging times to little color loss for longer aging times. These samples do not give any indication of the extent of shrinkage cracking, though the incidence of shrinkage cracking tends to increase as aging time increases.

130

A freshly formed CCC is crack-free. As a CCC ages, it dehydrates. As water is lost, the shrinkage across the coating thickness is different. The water loss in the outer layer is greater than inner layer, thus a tensile stress develops at the coating surface. This tensile stress leads to the formation of cracks 15. Cracked and uncracked CCCs can be identified in scanning electron micrographs or optical micrographs illustrating the change in CCC condition. Mud-cracked CCCs are easily resolved in scanning electron micrographs because the microscope vacuum ensures dehydration (Figure 4.5).

Observing crack-free coatings is considerably more difficult for the same reason. To avoid this difficulty, crack free CCCs were recorded by imaging freshly formed CCCs with optical microscope in this study (Figure 4.6).

Morphological characterization of aged coatings suggests that shrinkage cracking facilitates the ORR by exposing the underlying substrate. Figure 4.7 shows a cross section of a dehydrated conversion coating containing shrinkage cracking. This image is a scanning electron micrograph of a coating cross section prepared by focussed ion beam

(FIB) milling. This micrograph shows shrinkage cracking penetrating through the thickness of the conversion coating to the coating-substrate interface. The image also suggests that cracks propagate along the interface exposing additional surface area. Once exposed, this area can support oxygen reduction at enhanced rates.

To separate the contributions to loss in ORR inhibition from Cr6+ leaching and shrinkage cracking, two experiments were carried out on CCC coated samples. In the first experiment, conversion coated samples were subject to 15 minute to 2 hour-treatments at temperatures of 75o or 95oC, which increases shrinkage cracking and strongly suppresses

131

Cr6+ leaching 14. The loss in leachability is illustrated in Figure 4.8, which shows the optical macrogaphs of heat-treated samples that were subject to cathodic polarization. In each sample there is little if any color loss indicating little Cr6+ leaching or reduction.

Control experiments were carried out with bare 2024-T3 samples to ensure that heat treatment did not induce artificial aging of the 2024-T3 that appreciably affected the corrosion behavior substrate or the cathodic polarization response of the substrate. In fact, the cathodic polarization response of the bare alloy was not detectably altered in experiments where 2024-T3 samples were treated at 95°C for 2 hours.

Figure 4.9 shows cathodic polarization curves for bare 2024-T3 and from conversion coated samples thermally treated under various conditions as indicated in the plot. About one-half of the 20 samples that were heat-treated showed evidence of ORR inhibition. For certain treatments, ORR inhibition was enhanced by heat treatment, in others ORR inhibition was lost. In these experiments, Cr6+ leachability was significantly reduced in all samples, but cathodic inhibition was only sporadically observed. These results suggest that ORR is only weakly dependent on availability of leached Cr6+.

A second experiment provides additional evidence that the relationship between

Cr6+ leachability and cathodic inhibition is weak. Figure 4.10 shows the cathodic polarization curves of CCC-coated samples that were aged for about two weeks in ambient air to induce loss in ORR inhibition. Experiments were carried out in 0.5 M

-4 -6 NaCl solutions containing 10 M H2CrO4 and 10 M H2CrO4, which are the steady state chromate concentrations in solution (S/V = 2 cm-1) for a CCC aged one day in ambient air and a CCC heat treated at 80°C for 15 minutes, respectively 11. The cathodic

132

polarization curves of bare 2024-T3 and an aged CCC in 0.5M NaCl are also included for comparison. There is no evidence of ORR inhibition in these curves, and the results show that even the presence of 10-4M chromate in solution, ORR inhibition cannot be restored in CCCs that have developed shrinkage cracking due to aging. On the basis of these findings, it appears that the loss in ORR inhibition is more closely related to the presence of shrinkage cracking than any loss in Cr6+ leachability.

Taken together, these observations explain why cathodic inhibition by CCCs is not always detected in cathodic polarization experiments. If cathodic polarization experiments are performed on CCCs that are more than a few days old, inhibition is likely to be lost due to the formation of shrinkage cracks. For this reason, a close accounting of CCC age is necessary to properly understand the cathodic polarization response of chromate conversion coated 2024-T3.

4.4 Conclusion

These results show that the ability of a CCC to inhibit oxygen reduction varies with coating aging time. Freshly formed coatings and coatings aged up to 48 hours inhibit the ORR, but inhibition is gradually lost as the coating ages beyond 48 hours. ORR inhibition may be completely lost after aging in ambient air for 168 hours in agreement with earlier findings by Hughes 4. Loss in ORR inhibition appears to be due to shrinkage cracking of CCCs. Shrinkage cracking exposes the substrate, which can eventually support oxygen reduction at rates comparable to the unprotected alloy. This idea is supported by direct observation of coating cross sections by scanning electron microscopy. Cathodic polarization experiments with heat treated and aged conversion

133

coatings in chloride-chromate solutions argue against a strong relationship between ORR inhibition and Cr6+ leaching. In these experiments, the ORR is not inhibited on cracked

CCCs, even when chromate is available in solution.

134

REFERENCES

1. Wernick, S., R. Pinner, and P.G. Sheasby, The Surface Treatment and Finishing

of Aluminum and its Alloys, 5th ed. Vol. Vol. 1. 1987, Metals Park, OH: ASM

International. p. 220.

2. G. O. Ilevbare and J.R. Scully, J. Electrochem. Soc., 2001. 148: p. B196.

3. W. Zhang and R.G. Buchheit, J. Electrochem. Soc., 2002. 149: p. B357

4. Hughes, A.E., R.J. Taylor, and B.W.R. Hinton, Surface and Interf. Anal., 1997.

25: p. 223.

5. J. Zhao, G. Frankel, and R.L. McCreery, J. Electrochem. Soc., 1998. 145: p. 2258.

6. Pocock, W.E., Metal Finishing, 1954. 52: p. 48.

7. A. Sehgal, D. Lu, and G. S. Frankel, J. Electrochem. Soc., 1998. 145: p. 2834.

8. Clark, W.J. and R.L. McCreery, J. Electrochem. Soc., 2002. 149: p. B379.

9. G. O. Ilevbare and J. R. Scully, Corrosion, 2001. 57: p. 134.

10. G. O. Ilevbare and J. R. Scully, Corrosion, 2000. 56: p. 227.

11. V. Laget, H. S. Isaacs, C. S. Jeffcoate, and R. G. Buchheit, J. Electrochem. Soc.,

2002. in review.

12. W. Zhang and R. G. Buchheit, submitted to J. Electrochem. Soc., 2002.

13. W. J. Clark, J. D. Ramsey, and R.L. McCreery, J. Electrochem. Soc., 2001. 149:

p. B179.

14. Glass, A.L., Mat. Protect., 1968 (July): p. 26.

135

15. Brinker, C.J. and G.W. Scherrer, Sol-Gel Science: The Physics and Chemistry of

Sol-Gel Processing. 1990, New York: Academic Press, Inc. p.493.

136

Figures

-0.5

-0.7 )

sce -0.9

-1.1

Potential (V bare 2024-T3 1 hour 20 hours -1.3 120 hours 168 hours

-1.5 10-9 10-8 10-7 10-6 10-5 10-4 10-3

Current Density (A/cm2)

Figure 4.1 Cathodic polarization curves for conversion coated 2024-T3 exposed to aerated 0.5 M NaCl solution. Aging time refers to time of exposure in laboratory air prior to collection of the polarization curve.

137

1.8 10-6

1.6 10-6 Corrosion rate on bare 2024-T3

1.4 10-6

1.2 10-6 ) 2 1 10-6 (A/cm cor r I -7 CCC-coated 2024-T3 8 10

6 10-7

4 10-7

2 10-7 0 50 100 150 200 Aging time (hr)

Figure 4.2. Change in corrosion rates of CCC coated samples with aging time.

138

-0.4

bare 2024-T3 -0.6

22hrs -0.8 2hrs 47hrs 198hrs

90hrs -1 Potentialvs. (V SCE)

-1.2

10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3

Current (A/cm2)

Figure 4.3. Cathodic polarization behavior of CCC coated 2024-T3 in 0.1M Na2SO4 + 0.005M NaCl.

139

Figure 4.4. An optical macrogaph of conversion coated 2024-T3 illustrating color changes due to Cr6+ leaching in solution. Samples were aged in ambient lab air for the times indicated, then subject to cathodic potentiodynamic polarization.

140

20 µm

Figure 4.5. Scanning electron micrograph of shrinkage cracking in a CCC on 2024-T3.

141

( a )

40µm

( b )

Figure 4.6. CCC on 2024-T3 imaged with optical microscope after different aging times in ambient air.

142

deposited Pt layer

CCC

substrate

400 nm

2 µm

Figure 4.7. FIB cross section of a CCC on 2024-T3 illustrating the mophology of shrinkage cracks.

143

Figure 4.8 An optical macrogaph of conversion coated 2024-T3 samples illustrating color changes due to Cr6+ leaching in solution. Samples were heat treated prior to cathodic polarization for the times and temperatures indicated, then subject to cathodic potentiodynamic polarization.

144 -0.5

-0.7

) -0.8 sce

-1.0 1hr at 75°C

Potential (V 15min at 95°C -1.2 Bare 2024 2hr at 95°C -1.3 6hr at 23°C

-1.5 10-10 10-9 10-8 10-7 10-6 10-5 10-4 Current density (A/cm2)

Figure 4.9 Cathodic polarization curves for conversion coated 2024-T3 subject to elevated temperature exposure at the conditions indicated in the plot.

145

-0.500

-0.750 CCC aged in air for 360hrs, in Bare 2024 in 0.5M NaCl -6 6+ 0.5M NaCl + 10 M Cr

-1.000

CCC aged in air for 168hrs, in 0.5M NaCl Potential (V vs. SCE) vs. (V Potential -1.250 CCC aged in air for 408hrs, in 0.5M NaCl + 10 -4M Cr 6+

-1.500 -9 -8 -7 -6 -5 -4 -3 -2 10 10 10 10 10 10 10 10

Current density (A/cm 2)

Figure 4.10 Cathodic polarization curves of conversion coated 2024-T3 in 0.5M NaCl + -4 -6 10 or 10 M H2CrO4 solution. CCCs were aged in ambient air for various times as labeled in the plot before polarization test.

146 CHAPTER 5

EFFECT OF AMBIENT TEMPERATURE AGING IN DEHUMIDIFIED AIR ON

THE PROPERTIES OF CHROMATE CONVERSION COATINGS (CCCS)

5.1 Introduction

Numerous studies have been carried out to characterize CCCs and understand the corrosion inhibition mechanisms 1-10. It is now widely accepted that CCCs are mainly composed of hydrated, chromium mixed oxides, along with some minor components from coating bath and alloy substrate. The amount of Cr6+ accounts for about 30% of the total chromium in CCCs and is mainly concentrated in the outer surface region. Cr6+ can be released to the surrounding aqueous solution upon exposure and migrate to defects or damaged sites to inhibit further corrosion through adsorption or reduction-deposition or both 11. This characteristic of CCCs is termed “ self-healing”, and is regarded as one of the key attributes that make CCCs protective. Two elements are required for this protection scheme to work: first, there must be a stable reservoir of Cr6+ in the coating; second, the Cr6+ must be mobile and releasable. The chromium oxides in CCCs are sufficiently stable and inert and serve as a physical barrier that separates the environment and the alloy substrate. CCCs have been shown to inhibit both anodic reactions 12, 13 and 147 the oxygen reduction reaction (ORR) 14, at least for several days after coating formation.

However as CCCs age, coating morphology, structure and chemistry change to inhibit the release kinetics of Cr6+; and therefore influence the corrosion resistance of CCCs. It has been shown that aging at room temperature for extended time decreases the corrosion resistance 15. Aging at elevated temperature greatly accelerates the aging process, resulting in pronounced deceases in corrosion resistance 16.

Several physical processes have been proposed to explain the aging response 9, 16,

17. These include dehydration, development of shrinkage cracks, reduction in Cr6+ leachability, and transition from hydrophilic and hydrophobic character. All but the last contribute to loss in corrosion resistance. Therefore the study of aging phenomenon of

CCCs has both scientific and practical importance.

In this chapter, we report the effects of humidity of aging environment on the coating properties. The evolution of coating morphology of CCCs during aging was examined by optical microscopy. The amount Cr6+ in CCCs before and after exposure to

NaCl solution was characterized by Raman spectroscopy. The release of chromate from

CCC was measured with inductively coupled plasma optical emission spectroscopy (ICP-

OES) of solutions in contact with conversion-coated substrates. The corrosion resistance was evaluated with electrochemical impedance spectroscopy and by examining the pit morphology after exposure to 0.5 M NaCl solutions. The effect of aging in dry air on the self-healing ability of CCCs was assessed with a methodology developed by Hunter 18 known as the damage resilience test. Overall, the results show that aging in dry air has dramatic effects on CCCs properties that are not widely discussed in the literature. These

148 results shed new light on what happens during the coating aging process and show that aging conditions have a significant effect on CCC properties.

5.2 Experimental

Coating process. Coatings were formed in a coating bath obtained by dissolving

7.55 g Alodine 1200Sâ powder in 1 liter distilled water. The pH of the coating bath was adjusted to 1.6 with NaOH or HCl as necessary. AA2024-T3 sheets with a thickness of 3 mm were used as substrates in all experiments. Samples were cut into 2 x 5 in. rectangular coupons and then polished in water to 800 grit with abrasive paper. The polished samples were ultrasonically cleaned in alcohol before being immersed in the conversion coating bath. The polished samples were immersed in coating bath for 3 minutes and then thoroughly rinsed with overflowing distilled water. CCCs were dried with gentle air stream immediately following water rinse.

Some CCC-coated samples were aged in ambient air (RH = 55 – 65%, T = 19 -

23oC. This condition is termed “ambient air” hereafter), and some were aged in a functional desiccator (RH ≈ 0%, T = 19 - 23oC. This condition is termed as “dry air” hereafter) for various times before examination of coating morphologies and testing for corrosion resistance. Ambient air aged samples were used for comparative purposes.

Coating Morphology Characterization. The coating morphology development in ambient air and in dry air was examined by optical microscopy. An optical microscope was used to avoid the vacuum environment in SEM, which affects the coating aging process by promoting rapid dehydration and shrinkage cracking. When examining pit 149 morphology, a Philips XL-30 FEG SEM was used because dehydration was no longer a concern.

Raman Spectroscopy. The use of the peak at 859 cm–1 due to Cr6+-O-Cr3+ stretch for characterizing CCC Cr6+ chemistry has been thoroughly documented 10. In this study,

Raman spectroscopy was used to quantify the Cr6+ in CCCs before and after exposure to

NaCl solution. Raman spectra were collected using a Chromex 2000 spectrometer, with a standard interference band reject filter and EEV 15-11 deep depletion charge-coupled device (CCD). A 785 nm excitation and 180° backscattered sampling geometry were employed to obtain the Raman spectra. The instrument was frequency-calibrated with 4- acetamidophenol (Tylenol) and the intensity was calibrated with a glass that has known intensity-frequency curve. The area under 859 cm–1 peak after baseline correction indicates the amount of Cr6+ in CCCs.

Chromate Release Measurements. ICP-OES was used to measure the chromate released from CCCs. ICP-OES has sensitivity in the part-per-billion (ppb) range and was very suitable for this experiment because the chromate released from dry air aged CCCs was expected to be very low. CCCs subject to different aging treatments were exposed to

0.5 M NaCl for different lengths of times. After contact with conversion-coated surfaces, aliquots of the 0.5 M NaCl solutions were collected, acidified with HNO3, and analyzed with ICP-OES. The ratio of exposed surface area to the volume of the solution was 2.0 cm-1 in all experiments.

150 Corrosion Resistance Evaluation. Electrochemical impedance spectroscopy

(EIS) was used to quantitatively characterize the corrosion resistance of CCCs.

Conversion coated samples were exposed to aerated 0.5 M NaCl solution using a flat cell with a 15 cm2 window for the coated surface. Impedance spectra were collected using a

Princeton Applied Research Model 273A potentiostat and Solartron Model 1250 frequency response analyzer after different exposure times in solution. A sinusoidal 10 mV voltage perturbation with frequencies ranging from 10000 - 0.001 Hz was used in all experiments. The measurement was controlled by Scribner Associates Zplot™ impedance software installed in a personal computer. Impedance spectra were modeled using an equivalent circuit analysis and complex non-linear least squares fitting of the data to a suitable equivalent circuit. The values of charge transfer resistance determined by fitting were used as an indication of coating corrosion resistance.

Damage Resilience Test. This test was recently developed by Hunter et al. 18 to test the self-healing function of inhibitors in paints. We modified the test protocol to examine for evidence of self-healing in CCCs. In this test, samples were first exposed to

NaCl solution for four days, and impedance spectra were collected at regular intervals.

After about 100 hours of exposure, coatings were intentionally damaged by making an array of indents with a diamond pyramid microhardness indenter. A load of 500 g was applied and this load was held for 20 s. Indents of diamond shape with a diagonal length of 80 µm were created on 2024-T3 with above loading conditions. Figure 5.1 shows an indent typical of this process. These indents showed very good reproducibility in terms of depth and shape. A 3 x 3 indent array was made on each CCC coated sample. The

151 distance between vertical and lateral neighboring indents was 1 mm. After indenting, samples were re-exposed to solution and impedance spectra were collected again several times over a 100-hour period. These impedance spectra were fitted with suitable equivalent circuit models as discussed previously. Coating corrosion resistance (Rcorr) determined by fitting was plotted versus time to examine for changes indicating the extent of self-healing. An increase in Rcorr with time was taken as a sign of self-healing.

A decrease in Rcorr was taken as an indication that self-healing was not occurring.

5.3 Results

The Evolution of Coating Morphology under Different Conditions. Figure 5.2 shows optical micrographs of CCCs aged in ambient air for increasing lengths of time.

The images in Figure 5.2 are blurred because the thickness of the nearly transparent film is about the same as the depth of field of the microscope at this magnification. At short times, there were few cracks on the surface, but after 2 days, the extent of cracking increased and cracks began to connect to form a continuous network in some local areas.

The fully mud-cracked pattern observed in scanning electron micrographs, which is often regarded as a characteristic CCC morphology, was not seen even after 7 days aging in ambient air. This suggests that the mud-cracking pattern is just an artifact caused by the high vacuum environment of SEM. This idea is supported by examination of CCCs aged in dry air. Figure 5.3 shows images of CCCs aged in a dessicator at a relative humidity approaching 0%. A partially developed mud-cracking pattern was observed after just 5 min aging. This cracking pattern was comparable to CCCs aged in ambient air for several days. A fully developed cracking pattern was obtained after a few hours in the desiccator.

152 CCCs aged in ambient air did not reach this stage even after aging for 7 days. This clear difference in coating morphology evolution suggests that dry air dramatically accelerated the CCC dehydration process.

In these experiments it was also noted that characteristic colors developed as a result of the various aging treatments. Coatings aged in ambient air retained the original golden yellow color; while coatings aged in dry air showed brownish yellow color, very similar to coatings that have been heated.

The Leachability of Cr6+ in CCCs after Different Aging Treatments. Cr6+ in

CCCs can be released into aqueous solution upon contact leading to self-healing 3, 5. The results indicated that aging in dry air not only resulted in accelerated shrinkage crack development; it also caused significant reduction in the leachability of Cr6+. Figure 5.4 shows the discoloration (indicated in these images as a change in contrast) of CCCs after exposure to 0.5 M NaCl solution at S/V=2 cm-1 for 20 days. The panel on the left was aged in ambient air for one day prior to exposure to 0.5 M NaCl solution. The panel on the right was aged in dry air for one day before exposure. The left panel shows significant discoloration with almost complete loss of the characteristic golden yellow color. The panel that was aged in dry air showed very little discoloration, retaining a dark yellow color. Since the yellow color of CCCs originates from the presence of Cr6+, the discoloration is a qualitative indication of how much Cr6+ is released. Consequently, less

Cr6+ is expected to be released from dry air aged CCCs than from ambient air aged samples under the same exposure conditions.

153 The extent of Cr6+ release was confirmed by Raman spectroscopy. Figure 5.5 shows Raman spectra of the 859 cm-1 scattering band characteristic of Cr3+-O-Cr6+ bonds in CCCs 10 before and after 20 days exposure to NaCl solution. It can be seen that Cr6+ in the ambient air aged CCCs is almost completely leached by exposure. Calculations using the area under the 859 cm-1 peak after background correction indicated that more than

90% of Cr6+ was released into solution. For CCCs aged in dry air, the peak changed little due to exposure to NaCl for 20 days. The fraction of Cr6+ released was less than 10%. It is worth noting that the 859 cm-1 peak in Raman spectra collected before and after aging in dry air did not show significant changes in position or intensity.

The dependence of Cr6+ leaching on CCC age directly affects the chromate concentration in solutions contacting conversion-coated surfaces, which, in turn, affects self-healing. Figure 5.6 compares the chromate concentrations in 0.5 M NaCl solutions that have been exposed to either CCCs aged in ambient lab air or CCCs aged in dry air at

S/V = 2 cm-1. The chromate concentration in solution exposed to lab air aged CCCs is on the order of 10-4 M, which is consistent with the values measured with UV absorption methods under comparable conditions 19. The chromate in solution exposed to dry air aged CCCs is about two orders of magnitude lower. This value is very close to the chromate concentration released from CCCs that have been heat-treated at elevated temperature. These results show the significant effects of aging in dry air on the leaching of Cr6+ from CCCs.

The Effects of Different Aging on Coating Corrosion Resistance. Aging in dry air resulted in accelerated shrinkage crack development and reduced leachability of Cr6+

154 in CCCs. Both of these two factors are expected to affect the corrosion resistance of

CCCs. Figures 5.7a and 5.7b shows the representative impedance spectra for samples aged in dry and ambient air. On average, the corrosion resistance of CCCs aged in ambient air is about one order of magnitude greater than those aged in dry air, even after

10 days exposure in aerated 0.5 M NaCl. Figure 5.7a shows EIS data from both types of samples plotted in the complex plane. The EIS response is largely capacitive in all cases, but the effects of aging are easily distinguishable.

Differences in the capacitive responses are evident in Bode phase angle plots in

Figure 5.7b. The sample aged in dry air exhibits a smaller phase shifts at all frequencies and a larger capacitive impedance. There is a large difference in the phase shift at frequencies greater than about 10 Hz indicating that the sample aged in ambient air exhibits more a more purely capacitive behavior than either of the dry air aged samples.

This difference in impedance response is believed to be related to the presence of cracks in dry air aged CCCs. The impedance spectra of dry air aged CCC can be fitted with the equivalent circuit model shown in Figure 5.8a. Typical values used to fit the circuit model in Figure 5.8a are listed in Table 5.1. This equivalent circuit was modified from a model used for the oxide film on anodized aluminum 20. In this modified model used in this study, CCCs are considered to be consisted of two layers: the porous outer layer with many shrinkage cracks and the relative dense layer with a few cracks that reach the substrate. Figure 5.8b shows the experimental results and the fitting curves using the equivalent circuit in Figure 5.8a. A very good fitting was obtained, indicating the equivalent circuit model accurately reflected the coating structures.

155 In order to isolate the contribution of reduced Cr6+ leachability on the decrease in corrosion resistance, some CCCs were aged in dry air and then exposed to 0.5 M NaCl +

-4 10 M Na2CrO4 solution. This solution was used to simulate the solution that developed upon exposure of ambient air aged CCCs to chloride solutions. Representative impedance spectra from these experiments are also shown in the complex plane (Fig. 5.7a) and in

Bode plots (Fig. 5.7b). The addition of chromate to the solution slightly increases the low frequency impedance response of a dry air aged coating, but does not restore the loss in corrosion protection. In the Bode phase angle plot there is a strong increase in the phase shift at low frequencies and a modest one at frequencies between 1 ad 100 Hz perhaps indicating some healing of the shrinkage cracking structure.

Examination the pit development coated samples supports the notion of reduced corrosion resistance in dry air aged samples. On dry air aged CCCs, visible pits started to appear after two to three days’ exposure to 0.5M NaCl. After 10 days of immersion, dozens of pits developed on an exposed area of 15 cm2. Figure 5.9 shows micrographs of pits on dry air aged CCCs after exposure to 0.5M NaCl for 5 days and 10 days. On ambient lab air aged CCCs, visible pits were not seen even after 20 days of exposure.

-4 When dry air aged CCCs were exposed to 0.5 M NaCl + 10 M H2CrO4 solution, visible pits were observed after 10 day exposure, which indicated moderate increase in pitting resistance over dry air aged CCCs exposed to 0.5 M NaCl.

Damage Resilience Test Results. The damage resilience test is intended to measure self-healing in coatings or coating systems 18. If a coating is self-healing, the corrosion resistance Rcorr is expected to recover after damage is induced. Figure 5.10

156 shows the variation in the coating resistance through the course of this test on CCC that

6 2 was aged in ambient air for 24 hours. A steady state Rcorr value of about 9 × 10 Ω⋅cm was measured during the 95 hour interval between the start of exposure and the time the coating was damaged by indentation. EIS measurements made immediately after indentation damage showed a sharp drop in the Rcorr value and an equally sharp recovery.

6 2 The Rcorr values after recovery of about 7-8 × 10 Ω⋅cm suggest that the coating damage healed to a significant extent. The drop-and-recovery characteristic in this test is taken as an indication of self-healing. Figure 5.11 shows the Rcorr evolution during the damage resilience test on CCC that was aged in dry air for 24 hours. The coating resistance was steady at about 2-3 × 106 Ω⋅cm2 before the indentation damage. After indents were induced on the coating, the coating resistance dropped to about 5 × 105 Ω⋅cm2 and stayed at that value for the remainder of the exposure period without recovering. This trend in

Rcorr indicates that self-healing is not operating in dry air aged CCCs. When the test on

-4 dry air aged CCCs was carried out in 0.5 M NaCl + 10 M H2CrO4 solution, the drop- and-recovery signature was observed, as shown in Figure 5.11. In this case the spike associated with the introduction of coating damage appears to be superimposed on an overall increasing trend in the Rcorr value. This background trend may be associated with healing due to chromate added to solutions. In any case, this test illustrates the effects of aging in dry air on the self-healing ability of CCCs and the importance of soluble chromate for promoting repair of minor amounts of coating damage.

157 5.4 Discussion

The Nature of Dehydration. CCCs are formed by a series of reactions involving chromate and the Al alloy substrate 10. In this process, chromate is first reduced to Cr3+ in

3+ 2+ + solution. Cr then hydrolyzes to form Cr(OH) and Cr(OH)2 . These species condense and polymerize leading to the formation of a thin gelatinous film of hydrated Cr(OH)3.

The hydrated chromium oxide forms what is referred to as the “backbone” of CCCs 10.

Chromates in coating solutions are adsorbed onto the chromium oxide backbone through

Cr(III)-O-Cr(VI) bonds 10. This hydrated, mixed Cr(III)-Cr(VI) oxide is not stable, and will spontaneously dehydrate after removal from the coating bath into air. Dehydration of

21 Cr(OH)3⋅3H2O can occur due to an oxolation reaction described by Eqn. 5.1 .

Cr(OH)3(OH2)3 → Cr(OH)3 + 3H2O Eqn. 5.1

Water loss during dehydration process in ambient air has been studied by Kearns et al. 21, and Laget et al. 15. Kearns et al. 22 determined with FT-IR that the water content in CCCs decreased by 16% after 1 day aging in ambient air with a RH of 55-65%. Laget et al. 15 studied CCCs dehydration process with TGA and DGA and observed an endothermic reaction at 100°C due to water loss. Both Laget 15 and Xia 22 have observed that thermal treatment accelerated the aging process. Due to differential dehydration across the coating thickness, a tensile stress parallel to coating surface is developed and this stress causes shrinkage cracking in CCCs 23.

On the basis of the results in the present study it is evident that the dehydration process is greatly accelerated by aging in dry air. In terms of dehydration and its attendant effects on CCC structure and properties, it appears that aging in dry air is

158 similar to exposure at elevated temperatures. This is understandable when Eqn. 5.1 is considered. The low partial pressure of water in a dessicator should provide a large driving force for the reaction to shift to the right.

The Relationship between Dehydration and Cr6+ Leachability. It has been suggested that dehydration immobilizes the Cr6+ in CCCs 16, 19, 24. It is known that after

CCCs are aged in lab air at room temperature for extended time, slightly less Cr6+ is released and the release kinetics slow 19. Several studies have shown that after CCCs were thermally treated at elevated temperatures below 600°C, significantly less Cr6+ was released compared to ambient air aged CCCs under comparable exposure conditions 16, 23.

For example, Gallaccio et al. 16 observed that CCCs on cadmium, zinc and aluminum after heat treatment at 100°C for 2 hours released less than 3% of the Cr6+ that untreated

CCCs released. While there may be other factors that contribute to the reduced leachability of Cr6+ in thermally treated CCCs, the current results strongly support the idea that dehydration is the main cause of reduced Cr6+ leachability. During aging in dry air, no elevated temperature was involved, eliminating the complications that might be caused by thermal activation (for example, Cr6+ reduction by Al 19).

Several mechanisms have been proposed to explain why dehydration causes Cr6+ in CCC less leachable. These explanations are based on the fact that dehydration causes structural changes in CCCs. Based on extended X-ray absorption fine structure (EXFAS) analysis, Kearns 22 found that there were structural changes in CCCs associated with aging in ambient air at room temperature. Laget et al. 15 studied the structural changes during thermal treatment of CCCs by EXAFS. Although CCCs do not crystallize due to 159 low temperature or ambient temperature treatment, the distance between neighboring Cr centers was decreased after heat treatment at 80°C for 15 minutes. In CCCs, Cr(III) ions are octahedrally coordinated with hydroxyls or water molecules. These octahedral units are connected together by sharing corners or edges. It was postulated that after heat treatment the octahedral units changed from corner sharing to edge sharing. The effect of this reorganization was to reduce the amount of open space in the inorganic polymer structure thereby blocking pathways for Cr6+ release. A similar argument has been

19 proposed by Xia . She suggested that the continued polymerization of Cr(OH)3 in CCCs could result in closed cage structures. These structures could trap Cr6+ and prevent it from being released to the surrounding environment.

The structural changes are understandable when considering the nature of the dehydration process. When the water molecules in CCCs are removed, structural rearrangements are necessary to satisfy the coordination requirements of Cr cations.

There has been little study on the changes in the bond between Cr6+ and oxygen after aging, but Xia et al. 19 have suggested that some surface Cr6+ species, such as the one in

3+ 6+ 6+ Cr -O-Cr O3H, might be subject to further anchoring and form bidentate Cr -O ligands during heating. It was further suggested that new peaks at 780 cm-1 and 896-898 cm-1 in the Raman spectra of CCC after thermal aging arise from these bidentate Cr6+ species.

The resulting multiple bonds might fix Cr6+ preventing its release upon subsequent exposure. A point worth noting is that the loss of Cr6+ leachability appears to be irreversible, at least at room temperature. In the current study, Cr6+ in CCCs after aging in dry air for 1 day remained insoluble even after 20 days of immersion in solution. This suggests that CCCs cannot be re-hydrated at room temperature after dehydration.

160

The Relationship between Dehydration and Corrosion Resistance. Several studies have shown that after thermal treatment the corrosion resistance of CCCs is decreased 15, 16. Gallaccio et al. 16 found that thermal treatment at 50°C for 2 hours imparted a slight adverse effect on CCC corrosion resistance in salt spray testing.

Thermal treatment at increasingly higher temperature resulted in progressive damage to coating corrosion resistance. A two-hour treatment at 150°C caused a complete loss of the protectiveness of CCCs. More recently, Laget et al. 15 showed that aging at room temperature in ambient air for extended periods also degraded the corrosion resistance as evaluated with EIS. When samples were exposed at elevated temperature, the coating degradation was accelerated. The corrosion resistance of conversion-coated samples measured by EIS was reduced to that of a bare Al alloy surface after treatment at 80°C for 15minutes. The decrease in corrosion resistance of CCCs after thermal treatment has been attributed to either the loss of leachability of Cr6+ or the development of shrinkage cracks. Due to the loss of Cr6+ leachability, Cr6+ in the coating is no longer leachable and is therefore no longer available to provide self-healing. This leads to reduced corrosion resistance. The development of shrinkage cracks due to dehydration exposes the substrate to the corroding environment, resulting in a loss of cathodic inhibition, and an overall decrease in corrosion resistance.

The present results show that both shrinkage cracking and loss of chromate leachability contribute to the degradation in corrosion resistance. This is evident from comparison of the results of ambient air aged and dry air aged CCCs exposed to 0.5 M

-4 NaCl solution and dry air aged CCC exposed to 0.5 M NaCl + 10 M Na2CrO4 solution. 161 However, it is important to note that the impedance results indicate that shrinkage cracks,

-4 once formed, are not healed by the addition of 10 M Na2CrO4 such that original corrosion resistance is resorted. Results from this study also confirm a speculation by

Lytle et al. 5, who suggested that hydration was necessary for CCCs to be protective.

Additionally, the current results also support the assertion made by Laget 15 that the degradation of CCCs after heat treatment was due to the loss of water, and not due to thermal exposure itself.

The relationship between the amount of Cr6+ in CCCs and the performance of

CCCs has been the subject of several studies 25, 26, however consensus has not been reached in describing this relationship. In one study carried out in the 1960s 25, it was found that the corrosion resistance of CCCs increases with releasable Cr6+ content in the coating. Removing Cr6+ by bleaching or making it insoluble by heating at elevated temperature decreases the corrosion resistance in salt spray testing. Other studies find no such relationship. Williams 26 studied CCCs on zinc and concluded that the Cr6+ content did not correlate well with the performance of CCCs either in salt spray test or in exposure to urban environment. He suggested that the leachability of Cr6+ was also important in determining the corrosion resistance of CCCs. Results of this study support

Williams’ conclusion. The amount of Cr6+ in dry-air aged CCCs is the same as in ambient air aged ones. However, their corrosion resistance differs greatly from each other partly because of different Cr6+ leachability. This indicates that the Cr6+ must be releasable in order to provide the self-healing function and effectively inhibit corrosion.

162 5.5 Conclusions

Through the study of the effects of aging in dry air on the properties of CCCs, we have demonstrated that:

The humidity of the aging environment has dramatic effects on CCCs properties, including coating morphology, the leachability of Cr6+ from the coatings, and coating corrosion resistance.

Dehydration of CCCs results in shrinkage cracking and decrease in Cr6+ leachability. Both factors contribute to the degradation in corrosion resistance of CCCs.

Current results showed that Cr6+ in CCCs must be releasable in order to be effective.

163

REFERENCES

1. J. A. Treverton and N. C. Davis, Met. Technol., Oct., 480 (1977).

2. J. A. Treverton and N. C. Davis, Surf. Interface Anal, 3, 194 (1981).

3. H. A. Katzman, G. M. Malouf, R. Bauer and G. W. Stupian, Applica. Surf.

Sci., 2, 416 (1979).

4. K. Asami, M. Oki, G. E. Thompson, G.C. Wood and V. Ashworth,

Electrochim. Acta, 32, 337 (1987).

5. F. W. Lytle, R. B. Greegor, G. L. Bibbins, K. Y. Blohowiak, R. E. Smith and

G. D. Tuss, Corros. Sci., 37, 349 (1995).

6. Z. Yu, H. Ni, G. Zhang, Y. Wang, Appl. Surf. Sci., 62, 217 (1992).

7. H. E. Townsend and R. G. Hart, J. Electrochem. Soc., 131, 1345 (1984).

8. P. L. Hagans and C. M. Haas, Surf. Interface Anal., 21,65 (1994).

9. A. E. Hughes, R. J. Taylor and B. R. W. Hinton, ibid, 25, 223 (1997).

10. L. Xia and R. L. McCreery, J. Electrochem. Soc., 145, 3083 (1998).

11. J. D.Ramsey and R. L. McCreery, J. Electrochem. Soc., 146, 4076 (1999).

12. G. O. Ilevbare, J. R. Scully, J. Yuan, and R. G. Kelly, Corrosion, 56, 227

(2000).

13. W. Zhang, B. Hurley, and R. G. Buchheit, J. Electrochem. Soc., 149, B357

(2002).

164 14. W. J. Clark, J. D. Ramsey, and R.L. McCreery, J. Electrochem. Soc., 149,

B179 (2002).

15. V. Laget, C. S. Jeffcoate, H. S. Isaacs, and R. G. Buchheit, submitted to J.

Electrochem. Soc., in review (2002).

16. A. Gallaccio, F. Pearlstein, and M. R. D’ambrosio, Met. Finish. 64 (8), 50

(1966).

17. K. A. Korinek, ASM Handbook, Vol. 5, p389, ASM International (1987).

18. N. Hunter, J. H. Osborne, and R. S. Taylor, Corrosion, 56 (10), 1059 (2000).

19 L. Xia, Ph.D. dissertation, The Ohio Sate University (2000).

20. J. Hitzig, K. Juttner, W. J. Lorenz, and W. Paatsch, Corrosion. Sci., 24, 945

(1984).

21. J. P. Jolivet, Metal Oxide Chemistry and Synthesis, John Wiley & Sons, West

Sussex, England (2000).

22. G. S. Frankel, MURI Third Annual Report, The Ohio Sate University, 179

(1999).

23. C. J. Brinker and G. W. Scherer, Sol-gel science: the physics and chemistry of

sol-gel processing, Academic Press, Boston, MA (1990).

24. A. L. Glass, Mat. Prot., 7, 26 (1968).

25. Report NAEC-AML 2065, “Effect of hexavalent chromium content on

performance of chromate films”, Aeronautical Materials Lab, Naval

Engineering Center, Philadelphia, PA (1964).

26. L. F. G. Williams, Surf. Technol. 7, 113 (1978).

165

FIGURES

20µm

Figure 5.1 SEM image of indent introduced with diamond pyramid indenter on CCC coated 2024-T3. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl. Image collected after exposure to 0.5M NaCl for 192 hours.

166

( a ) ( b )

20µm

( c )

Figure 5.2 Optical images of CCC on 2024-T3 after aged in ambient air for different times. (a). 30 minutes. (b). 1 day (c). 4 days

167

( a )

( b ) 20µm

Figure 5.3 Optical images of CCC on 2024-T3 after aged in dry air for different times. (a). 5 minutes. (b). 8 hours.

168

2 mm

( a) ( b )

Figure 5.4 Discoloration of CCCs on 2024-T3 aged in different environments after 20 days exposure to 0.5M NaCl solution. S/V=2cm-1. (a). Ambient air aged CCC. (b). Dry air aged CCC.

169

0.006

0.005 unexposed area 0.004

0.003

0.002 Intensity (au) Intensity

0.001 exposed area

0 50 0 60 0 70 0 80 0 90 0 1 00 0 1 10 0 1 20 0 1 30 0

-1 Raman shift (cm )

( a )

0.008

0.007 unexposed area

0.006

0.005 exposed area

Intensity (au)Intensity 0.004

0.003

0.002 500 600 700 800 900 1000 1100 1200 1300

-1 Raman shift (cm )

( b)

Figure 5.5 Representative Raman spectra showing Cr6+ before and after exposure to NaClon CCCs aged under different conditions. (a). Ambient air aged CCC after exposed to 0.5M NaCl for 20 days. (b). Dry air aged CCC after exposed to 0.5M NaCl for 20 days.

170

-3 10

-4 10

aged in ambient air for 1day

-5 10 aged in dry air for 1day concentration (M) 6+

Cr -6 10

-7 10 0 50 100 150 200

E xpos ure tim e (h r)

Figure 5.6 Cr6+ released from CCCs into surrounding 0.5M NaCl solutions after different aging treatments. S/V=2cm-1.

171

7 -1. 2 10

7 -1. 0 10

Ambient air aged CCC in 0.5M NaCl

6 -8. 0 10 ) 2

6 -6. 0 10 Z" (ohm-cmZ"

6 -4. 0 10 -4 6+ Dry air aged CCC in 0.5M NaCl + 10M Cr

6 -2. 0 10 Dry air aged CCC in 0.5M NaCl

0 0.0 10 0 6 6 6 6 7 7 0.0 10 2.0 10 4.0 10 6.0 10 8.0 10 1.0 10 1.2 10

2 Z' (ohm -cm )

Figure 5.7a. Nyquist plots of CCCs after different treatments. Collected after 10 days of exposure to corresponding solutions. S/V = 2cm-1.

172

108

107 Ambient air aged, in 0.5M NaCl 6 Dry air aged, in 0. 5M NaCl + 10-4MCr 6+ ) 10 2 Dry air aged, in 0. 5M NaCl 105

104

IZI (ohm-cm IZI 103

102

101

100 10-2 10-1 100 101 102 103 104 Frequency (Hz)

-90

-80

-70

-60 Theta (degree) Theta

Amb ient air a ged, in 0.5M NaCl -50 Dry air aged, in 0. 5M NaCl + 10-4M Cr6+ Dry air aged, in 0. 5M NaCl -40 10-2 10-1 100 101 102 103 104 Frequency (Hz)

Figure 5.7b. Bode plots of CCCs after different aging treatments. Collected after 10 days of exposure to corresponding solutions. S/V = 2 cm-1. 173

R1 C1

C2 Rcorr Cdl

Figure 5.8a Schematic of cross section of CCCs aged in dry air and the equivalent circuit model used to fit impedance spectra.

174

-2.0 10 5

5 -1.5 10

5 -1.0 10 Z" (ohm)

-5.0 10 4

0.0 10 0

0.0 10 0 5.0 10 4 1.0 10 5 1.5 10 5 2.0 10 5

Z' (ohm)

106 -85 -80 105 -75 104 Theta (degree) -70

103 -65

IZI (ohm) IZI -60 102 -55 101 -50 100 -45 10-2 10-1 100 101 102 103 104

log(frequency) (Hz)

Figure 5.8b Experimental impedance spectra (dots) collected on dry air aged CCCs and the fitting results (solid lines) with the equivalent circuit model in Figure 5.8a.

175

100 µm

( a )

3 mm

( b )

Figure 5.9 Pits observed on dry air aged CCCs after exposure to 0.5M NaCl. (a). SEM image of pit on CCC coated 2024-T3 after 5 days’ exposure. (b). Optical images of pits on CCC coated 2024-T3 after 10 days’ exposure. 176

1.2 10 7

Indents introduced after this point 1.0 10 7 ) 2 8.0 10 6 (ohm-cm corr R 6.0 10 6

4.0 10 6

2.0 10 6 0 50 100 150 200 250 Time (hours)

Figure 5.10 Corrosion resistance variation of CCC coated 2024-T3 in damage resilience test. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl. S/V = 2cm-1.

177

3.0 10 6 Indents introduced after this point

2.5 10 6

2.0 10 6 ) 2

1.5 10 6 (ohm-cm corr R 1.0 10 6

5.0 10 5

0.0 10 0 0 50 100 150 200 Time (hours)

Figure 5.11 Corrosion resistance variation of CCC coated 2024-T3 in damage resilience test. CCC aged in dry air for 1 day before exposed to 0.5M NaCl. S/V = 2cm-1.

178

1.2 10 7

1.0 10 7 ) 2 8.0 10 6 Indents introduced after this point (ohm-cm

corr 6.0 10 6 R

4.0 10 6

2.0 10 6 0 50 100 150 200 250 Time (hour)

Figure 5.12 Corrosion resistance variation of CCC coated 2024-T3 in damage resilience test. CCC aged in ambient air for 1 day before exposed to 0.5M NaCl + 10-4M Cr6+. S/V = 2cm-1.

179 CHAPTER 6

CONCLUSIONS AND FUTURE WORK

6.1 Contributions from this study to a better understanding of CCCs

A significant volume of information regarding the chromate coating formation, composition, structure, and corrosion inhibition mechanisms has been generated as a result of recent research on chromate and chromate conversion coatings. The addition of the results from the present study provides a clearer picture from coating formation to coating inhibition mechanisms and coating degradation during aging. In this section, the current results are summarized in the context of previous findings to demonstrate how the current results improve the understanding of chromate coatings.

Recent studies show that inorganic polymerization is the chemical process that accounts for much of the phenomenology associated with chromate coatings, from coating formation process to the aging responses of CCC. CCC formation has been described as a hydrolysis, condensation and polymerization of chromate reduction products and adsorption of chromates onto these inorganic polymeric chromium hydroxides 1, 2. The development of shrinkage cracks and loss of leachability of Cr6+ in

CCCs are the results of continuing polymerization that occurs as coatings age. The

180

present results are united with previous findings through this inorganic polymerization theme.

The chromate coating formation process has been the subject of many studies 1-8.

The most recent description of this process is from a sol-gel perspective 1, 2. HF in the coating bath dissolves the original aluminum oxide film on aluminum. Freshly exposed

Al surface is oxidized by chromate and chromate is reduced to form Cr3+ species. This

3+ oxidation reduction reaction is accelerated by the mediation action of K3Fe(CN)6. Cr then hydrolyzes, condenses and polymerizes to form hydrated hydroxides. These chromium hydroxides provide the backbone of the CCC and chromates in the coating bath are adsorbed onto this backbone through chemisorption. The present results indicate that the electrochemical reduction of chromate and oxidation of Al occur primarily in the first 30s of alloy exposure to the treatment bath. However, the chemisorption of chromates takes place during the whole coating formation process. This result combined with previous studies shows that the CCC formation is dominated by the inorganic polymerization process and chemical adsorption and electrochemical reactions only play major roles in the early period of the coating formation process. During the first 30s, the electrochemical reactions generate large amount of Cr3+ species near the aluminum surface. After 30s, the aluminum surface is covered with the solid products of the inorganic polymerization process. It is possible that the inorganic polymerization of the remaining Cr3+ species still occurs even after the surface is already covered with

181

polymerization products. This may help explain the continued thickening of CCCs after

30 s.

The current results are consistent with the proposed roles played by NaF and

- K3Fe(CN)6 during coating formation. When F is absent, the surface quickly passivates and no significant film formation process occurs. When K3Fe(CN)6 is present, the coating formation process is accelerated. Furthermore, the current results emphasize the contributions of these two minor ingredients to the coating corrosion resistance. These two components are critical in the formation of corrosion resistant coatings. These results suggest that in the development of chromate free coating systems, identifying appropriate minor ingredients for coating bath chemistries are also important.

The anodic inhibition function of CCCs has not been well characterized. The present results clearly show that the pitting potential of CCC coated pure Al was increased by as much as 500 mV when polarized in 0.5 M NaCl, which clearly indicates the inhibition of Al breakdown by CCCs.

CCCs are dynamic systems and several changes occur after coatings are formed on Al surfaces. Notable changes include loss of water content in the coating, development of shrinkage cracking and loss of leachability of Cr6+ in the coatings. These changes result from continued inorganic polymerization and have profound effects on coating corrosion protection, such as the loss of cathodic inhibition function with increasing aging time and decrease in corrosion resistance.

182

There has been some controversy on whether CCCs inhibit oxygen reduction.

Clark et al.9 demonstrated that a monolayer of chromate reduction product was able to inhibit oxygen reduction. CCCs have a thickness equivalent to a thousand layers of chromium oxide and are expected to show strong inhibition of oxygen reduction.

However, ORR inhibition of oxygen reduction by CCCs was not always observed in cathodic polarization testing 10, 11. The current results indicate that freshly formed CCC

(within 48hrs after formation) inhibits oxygen reduction, but with increasing aging time, inhibition gradually disappears. Carefully designed experiments and FIB characterization indicate that the loss of cathodic inhibition is due to the development of shrinkage cracking, not due to the loss of leachability of Cr6+ in CCCs. These results are consistent with the observations that CCCs do not inhibit oxygen reduction, because in those experiments, the coatings were not tested until several days after coating formation 10,11.

According to the results of the present study, ORR inhibition would have been lost in those specimens. By that time, cathodic inhibition might have already disappeared. Thus the controversy over the cathodic inhibition by CCC is successfully resolved. Resolution of the controversy over cathodic inhibition lies in the fact that this component of corrosion protection is temporary. Caution must be used in conducting corrosion experiments with CCCs since their protection behavior is dynamic. Caution should also be used in interpreting results from studies where coating age was not well controlled or specified.

183

The chromium hydroxide in CCCs is very stable and inert and provides barrier protection for the substrate. For this protection to be effective, the coating should be intact and cover the area that is exposed to corrosive environments. Chromate in CCCs provides an important element of the corrosion inhibition by CCCs: self-healing. For this mechanism to be operating, there must be a Cr6+ reservoir in CCC; the Cr6+ in CCC must be releasable upon exposure to aqueous solution 12. During aging, several changes that occur within the chromate coatings interrupt these two protection actions. The development of shrinkage cracks expose substrate surface and results in the loss of cathodic inhibition. The decrease in Cr6+ leachability eliminates self-healing. These two actions will decrease the corrosion resistance of chromate coatings. It is known that aging in ambient air for extended periods degrades the coating corrosion resistance and thermal treatment greatly accelerates the aging process 13, 14. The present results demonstrate that the relative humidity of aging environment has a similar effect as thermal treatment on coating properties. Aging in dehumidified air at room temperature for 5 minutes resulted in shrinkage cracks that required four days to develop in ambient air. After aging in dehumidified air for one day, the Cr6+ released from the CCC was two orders of magnitude lower than that released from an ambient air aged CCCs. The self-healing of dry air aged CCCs was lost in damage resilience test. As a result, the corrosion resistance of CCCs after aging in dehumidified air for 1 day was decreased by a factor of 5 compared to ambient air aged CCCs under the same exposure conditions when assessed by EIS. The time required to develop visible pits on dry air aged CCCs was 2-3 days,

184

compared to 20+ days for ambient air aged CCCs when exposed to 0.5 M NaCl. These results indicate that the storage conditions such as RH of CCCs can affect its corrosion resistance.

These aging responses of CCCs to relative humidity are further indications that

CCCs are subject to continued inorganic polymerization during aging. Dehydration of

CCCs results in tensile stress on the CCC surface and causes shrinkage cracks. Continued inorganic polymerization may cause several structural changes in CCCs, such as the formation of closed “cage” structures that may trap Cr6+, the formation of bidentate Cr6+-

O ligands that may lock Cr6+, and the transition from corner sharing to edge sharing by chromium octahedras 1,14.

6.2 Summary of key results

The objectives of this study have been to develop a better understanding of the chromate coating formation process and its corrosion inhibition mechanisms, and to provide some guidance to the development of chromate free coating systems. Research in three areas has been carried out to achieve those objectives: coating formation process on

Al and effects of coating bath components on coating formation, corrosion inhibition mechanisms of chromate coatings, and the response of coating properties to aging. New understanding regarding the coating formation and corrosion inhibition mechanisms has been reached through this study. Specifically:

185

1. Chromate coating formation on Al has been found to consist of two overlapping

stages: intensive electrochemical reaction in the first 30 s and the adsorption of

chromate from coating bath onto chromium oxide during the whole coating

process. Coating formation across Al surface is not uniform. Areas where

cathodic reaction (chromate reduction) dominates have higher breakdown

potentials.

2. The breakdown potentials of CCC coated Al are higher than bare Al when

anodically polarized in 0.5 M NaCl and increase with coating time up to 5

minutes. The median breakdown potential of 30 s coating increases by more than

250 mV compare to bare Al. Increasing coating time from 30 s to 5 minutes only

results in additional 100 mV increase in median breakdown potential. These

results indicate that CCCs inhibit Al anodic dissolution in 0.5 M NaCl, which has

not been observed on CCC-coated bulk (1 cm2) Al samples.

3. Minor components in the CCC coating bath, NaF and K3Fe(CN)6, play critical

roles in the formation of coatings with corrosion resistance. Inclusion of these two

components in the coating bath significantly increases pitting potentials of coated

Al and enhances coating corrosion resistance. Without NaF, Al surface quickly

passivates and no significant coating formation occurs. Addition of K3Fe(CN)6 to

coating bath increases coating breakdown potentials. These results provide

important guidance to the development of chromate coating systems in that

186

current results emphasize the important roles played by minor components in

coating bath.

4. The cathodic inhibition by CCCs changes with coating aging time. Coatings aged

for less than 48 hours in ambient air exhibit cathodic inhibition. The cathodic

inhibition gradually disappears with increasing aging time. After one week aging

in ambient air, the cathodic inhibition is completely lost. The loss of cathodic

inhibition by CCCs with aging time is attributed primarily to the development of

shrinkage cracking, not to the loss of Cr6+ leachability, although this does occur.

These results indicate that the ability of CCCs to inhibit oxygen reduction is

temporary.

5. Aging in dehumidified air has dramatic effects on CCC properties. The

development of shrinkage cracking is greatly accelerated by aging in dry air and

the leachability of Cr6+ from CCCs is significantly reduced. The amount of

chromate released from a CCC aged in dry air for 24 hours is about two orders of

magnitude lower than that of a CCC aged for the same time in ambient air.

Damage resilience tests show that a CCC aged in dry air for 1 day lost their self-

healing ability, while coatings aged in ambient air retain self-healing

characteristics. As a result, the corrosion resistance of a CCC aged in dry air for 1

day is decreased compared to that of CCC aged in ambient, as evidenced in the

lower electrochemical impedance and more rapid development of pits during

exposure to 0.5M NaCl. These results indicate that aging of a CCC is mainly a

187

dehydration process and the storage conditions of CCC affect their corrosion

resistance.

6.3 Future work

Present work has opened up some issues that are of interest for future study. The following topics are suggested for future work.

1. This study has found that coatings formed on net cathodes during

coating formation are more resistant to pitting than those on net anodes.

Characterization is needed to understand the reasons behind this

phenomenon. Coating thickness, composition, and structure are to be

examined to help explain this difference. FIB, AES, MMA, Raman

spectroscopy can be employed to characterize the coatings.

2. In cathodic polarization experiments, conversion-coated 2024-T3

showed different behavior from bare 2024-T3 when tested within 2 days

after coating formation. On conversion-coated samples the reactions are

under activation control while on bare 2024-T3 the reactions are under

diffusion control. The reactions on conversion-coated samples are not

well understood. Several reactions may occur: hydrogen reduction, Cr6+

reduction, and oxygen reduction. Identification of the reaction(s) will

help better understand the corrosion inhibition mechanisms.

188

3. This study showed the dramatic effects of aging in dry air on coating

properties. The structural, compositional changes of CCC during aging

in dry air need to be characterized using vibrational spectroscopy,

XANES, and EXAFS to understand why aging in dry air causes

accelerated cracking and reduced Cr6+ leachability.

4. In this study coating formation and breakdown on pure Al wires were

examined with MMA. It will be of interest to know how AA 2024-T3

wires will behave in similar experiments. The roles of intermetallics on

coating formation and breakdown may be examined with MMA.

5. The results from this study showed the critical roles played by minor

ingredients on CCC formation and their contributions to coating

resistance to breakdown. These results may be applied to guide the

development of chromate-free coating systems. For example, vanadate,

molybdate, and cerium chloride all have been tried for chromate-free

coating development. If appropriate minor ingredients could be

identified, these coating bath chemistries may produce corrosion

resistant coatings on aluminum alloys.

189

REFERENCES

1. L. Xia, Formation and Function of Chromate Conversion Coating on Aircraft

Aluminum Alloy Probed by Vibrational Spectroscopy, in Department of

Chemistry. 2000, The Ohio State University: Columbus, OH.

2. J.S. Osborne, Progress in Organic Coatings, 41, 280 (2001).

3. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 33, 1371 (1992).

4. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 34, 1045 (1993).

5. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 35, 253 (1993).

6. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 34, 2099 (1993).

7. H.A. Katzman, Malouf, G. M., Bauer, R. and Stupian, G. W., Applications of

Surface Science, 2, 416 (1979).

8. D.J. Arrowsmith, J.K. Dennis, and P. Sliwinski, Transactions of the Institute of

Metal Finishing, 62, 117 (1984).

190

9. W.J. Clark and R.L. McCreery, Journal of The Electrochemical Society, 149,

B379 (2002).

10. G.O. Ilevbare and J.R. Scully, Corrosion, 56, 227 (2001).

11. G.O. Ilevbare and J.R. Scully, Corrosion, 57, 134 (2001).

12. R.G. Buchheit, G.S. Frankel, H.L. Fraser, R.L. McCreery, J.S. Beatty, and M.S.

Donley, "Critical Factors for the Transition from Chromate to Chromate-Free

Protection", The Ohio State University (1999).

13. A. Gallaccio, F. Pearlstein, and M.R. D'ambrosio, Metal Finishing, 64, 50 (1966).

14. V. Laget, H.S. Isaacs, C.S. Jeffcoate, and R.G. Buchheit, Journal of The

Electrochemical Society, In review, (2002).

191

BIBLIOGRAPHY

1. "ASTM B117-97 "Standard Practice for Operating Salt Spray (Fog) Apparatus"",

ASTM (1997)

2. "MIL-C-5541E " Chemical Conversion Coatings on Aluminum and Aluminum

Alloys" (1990)"

3. "Effect of Hexavalent Chromium Content on Performance of Chromate Films",

Aeronautical Materials Lab, Naval Engineering Center (1964)

4. V.S. Agarwala, ACS Symposium Series, 211 (1986).

5. D.J. Arrowsmith, J.K. Dennis, and P. Sliwinski, Transactions of the Institute of

Metal Finishing, 62, 117 (1984).

6. K. Asami, M. Oki, G.E. Thompson, G.C. Wood, and V. Ashworth, Electrochimca

Acta, 32, 337 (1987).

7. T. Biestek and J. Weber, "Electrolytic and Chemical Conversion Coatings",

Poland: Portcullis Press Limited-Redhill (1976).

8. C. Blanc, B. Lavelle, and G. Mankowski, Corrosion Science, 39, 495 (1997).

192 9. H. Bohni and H.H. Uhlig, Journal of The Electrochemical Society, 116, 906

(1969).

10. C.J. Brinker and G.W. Scherrer, "Sol-Gel Science: The Physics and Chemistry of

Sol-Gel Processing", New York, NY: Academic Press, Inc. (1990).

11. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 33, 1371 (1992).

12. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 34, 1045 (1993).

13. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 35, 253 (1993).

14. G.M. Brown, K. Shimizu, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 34, 2099 (1993).

15. R.G. Buchheit, Journal of The Electrochemical Society, 142, 3994 (1995).

16. R.G. Buchheit, M.D. Bode, and G.E. Stoner, Corrosion, 50, 205 (1994).

17. R.G. Buchheit, G.S. Frankel, H.L. Fraser, R.L. McCreery, J.S. Beatty, and M.S.

Donley, "Critical Factors for the Transition from Chromate to Chromate-Free

Protection", The Ohio State University (1999).

18. R.G. Buchheit, R.P. Grant, P.F. Hlava, B. Mckenzie, and G.L. Zebnder, Journal of

The Electrochemical Society, 144, 2621 (1997).

193 19. R.G. Buchheit, S.B. Mamidipally, P. Schmutz, and H. Guan, Corrosion, 58, 3

(2002).

20. R.G. Buchheit, F.D. Wall, G.E. Stoner, and J.P. Moran, Corrosion, 51, 417

(1995).

21. R.M. Burns and W.W. Bradley, "Protective Coatings for Metals", New York, NY:

Reinhold Publishing Corporation (1967).

22. N.R. Cawley and D.G. Harlow, Journal of Materials Science, 31, 5127 (1996).

23. G.S. Chen, M. Gao, and R.P. Wei, Corrosion, 52, 8 (1996).

24. S.W.M. Chung, J. Robinson, G.E. Thompson, G.C. Wood, and H.S. Isaacs,

Philosophical Magazine B, 63, 557 (1991).

25. R.J. Cinderey and G.T. Burstein, Corrosion Science, 33, 493 (1992).

26. W.J. Clark and R.L. McCreery, Journal of The Electrochemical Society, 149,

B379 (2002).

27. W.J. Clark, J.D. Ramsey, R.L. McCreery, and G.S. Frankel, Journal of The

Electrochemical Society, 149, B179 (2002).

28. J.S. Crompton, P.R. Andrews, and E. McAlpine, Surface and Interface Analysis,

13, 160 (1988).

29. A.J. Davenport and H.S. Isaacs, Corrosion Science, 31, 105 (1990).

194 30. A.J. Davenport, H.S. Isaacs, and M.W. Kendig, Corrosion Science, 32, 653

(1991).

31. R.A. Dickie, Progress in Organic Coatings, 25, 3 (1994).

32. C.A. Drewien, M.O. Eatough, D.R. Tallant, C.R. Hills, and R.G. Buchheit,

Journal of Materials Research, 11, 1507 (1996).

33. L.C. Feldman and J.W. Mayer, "Fundamentals of Surface and Thin Film

Analysis", New York, NY: Elsevier Science Publishing Co., Inc. (1986).

34. J.C.S. Fernandes and M.G.S. Ferreira, Electrochimca Acta, 37, 2659 (1992).

35. G.S. Frankel, "Mechanism of Al Alloy Corrosion and the Role of Chromate

Inhibitors", The Ohio State University (1996).

36. G.S. Frankel, "Mechanism of Al Alloy Corrosion and the Role of Chromate

Inhibitors", The Ohio State University (1997).

37. G.S. Frankel, "Mechanism of Al Alloy Corrosion and the Role of Chromate

Inhibitors", The Ohio State University (1998).

38. G.S. Frankel, "Mechanism of Al Alloy Corrosion and the Role of Chromate

Inhibitors", The Ohio State University (1999).

39. G.S. Frankel, "The Influence of Inhibitor Ions and Conversion Coatings on

Localized Corrosion of Al and Al Alloys", The Ohio State University (1999).

40. A. Gallaccio, F. Pearlstein, and M.R. D'ambrosio, Metal Finishing, 64, 50 (1966). 195 41. A.L. Glass, Materials Protection, 26 (1968).

42. V. Guillaumin and G. Mankowski, Corrosion Science, 41, 421 (1999).

43. P.L. Hagans and C.M. Hass, "Chromate Conversion Coatings", Metals Park:

ASM International (1994).

44. P.L. Hagans and C.M. Hass, Surface and Interface Analysis, 21, 65 (1994).

45. G.P. Halada and C.R. Clayton, Journal of The Electrochemical Society, 138, 2921

(1991).

46. J.E. Hatch, "Aluminum: Properties and Physical Metallurgy", Metals Park, OH:

ASM (1984).

47. J.K. Hawkins, H.S. Isaacs, S.M. Heald, J. Tranquada, G.E. Thompson, and G.C.

Wood, "Proceedings (The Electrochemical Society) 86-11", eds., p. 345, (1986).

48. J.K. Hawkins, H.S. Isaacs, S.M. Heald, J. Tranquada, G.E. Thompson, and G.C.

Wood, Corrosion Science, 27, 391 (1987).

49. M.A. Heine and M.J. Pryor, Journal of The Electrochemical Society, 114, 1001

(1967).

50. B.R.W. Hinton, Journal of Alloys and Compounds, 180, 15 (1992).

51. B.R.W. Hinton, D.R. Arnott, and N.E. Ryan, Metals Forum, 7, 211 (1984).

52. B.R.W. Hinton, D.R. Arnott, and N.E. Ryan, Materials Forum, 9, 162 (1986).

196 53. E.H. Hollingsworth and H.Y. Hunsiker, "Corrosion of Aluminum and Aluminum

Alloys", Metals Park, OH: ASM International (1987).

54. A.E. Hughes, J.D. Gorman, and P.J.K. Paterson, Corrosion Science, 38, 1957

(1996).

55. A.E. Hughes, R.J. Taylor, and B.R.W. Hinton, Surface and Interface Analysis, 25,

223 (1997).

56. N. Hunter, J.H. Osborne, and R.S. Taylor, Corrosion, 56, 1059 (2000).

57. G.O. Ilevbare and J.R. Scully, Corrosion, 56, 227 (2001).

58. G.O. Ilevbare and J.R. Scully, Corrosion, 57, 134 (2001).

59. G.O. Ilevbare and J.R. Scully, Journal of The Electrochemical Society, 148, B196

(2001).

60. C.S. Jeffcoate, H.S. Isaacs, J. Hawkins, and G.E. Thompson, "Electrochemical

Society Proceedings 98-17", eds., p. 192, (1998).

61. J.P. Jolivet, "Metal Oxide Chemistry and Synthesis", West Sussex, England: John

Wiley & Sons (2000).

62. H.A. Katzman, Malouf, G. M., Bauer, R. and Stupian, G. W., Applications of

Surface Science, 2, 416 (1979).

63. M.W. Kendig, Davenport, A. J. and Isaacs, H. S., Corrosion Science, 34, 41

(1993). 197 64. S.J. Ketcham, "Properties of Chemical Films on Aluminum Alloys", New York,

NY: Reinhold Publishing Corporation (1963).

65. L. Kintrup and L.d. Riese-Meyer, Inst. Phys. Conf. Ser., 119, 257 (1991).

66. D.C. Koningsberger and R. Prins, "X-ray Absorption Principles, Applications,

Techniques of EXAFS, SEXAFS and XANES", New York, NY: John Willey &

Sons, Inc. (1988).

67. S.K. Konno, H. Takahashi and M. Nagayama, Corrosion Science, 22, 913 (1982).

68. K.A. Korinek, "Chromate Conversion Coatings", Metals Park, OH: ASM

International (1987).

69. K. Kowal, J. Deluccia, J.Y. Josefowicz, C. Laird, and G.C. Farrington, Journal of

The Electrochemical Society, 143, 2471 (1996).

70. V. Laget, H.S. Isaacs, C.S. Jeffcoate, and R.G. Buchheit, Journal of The

Electrochemical Society, In review, (2002).

71. F.W. Lytle, R.B. Greegor, G.L. Bibbins, K.Y. Blohowiak, R.E. Smith, and G.D.

Tuss, Corrosion Science, 37, 349 (1995).

72. S.B. Mamidipally, Active Corrosion Protection by Cerium-Modified Hydrotalcite

Coatings on Aluminum Alloys, in Materials Science and Engineering. 1999, The

Ohio State University: Columbus, OH.

73. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Corrosion, 45, 615 (1989).

198 74. F. Mansfeld, Y. Wang, and H. Shih, Journal of The Electrochemical Society, 138,

L74 (1991).

75. F. Mansfeld, Y. Wang, and H. Shih, Electrochimca Acta, 37, 2277 (1992).

76. E. McCafferty, Journal of The Electrochemical Society, 137, 3731 (1990).

77. J.F. Murphy, "Chemical Conversion Coatings for Aluminum Alloys", New York:

Reinhold Publishing Corporation (1963).

78. N.J. Newhard, Jr., Metal Finishing, 49, 66 (1972).

79. A.J. Perrotta, S.Y. Tzeng, W.D. Imbrogno, and R. Rolles, Journal of Materials

Research, 7, 3306 (1992).

80. W.E. Pocock, Metal Finishing, 52, 48 (1954).

81. S.T. Pride, J.R. Scully, and J.L. Hudson, Journal of The Electrochemical Society,

141, 3028 (1994).

82. M.F.A. Rabbo, J.A. Richardson, and G.C. Wood, Corrosion Science, 18, 117

(1978).

83. M.F.A. Rabbo, J.A. Richardson, G.C. Wood, and C.K. Jackson, Corrosion

Science, 16, 677 (1976).

84. J.D. Ramsey and R.L. McCreery, Journal of The Electrochemical Society, 146,

4076 (1999).

199 85. C.M. Rangel and M.A. Travassos, Corrosion Science, 33, 327 (1992).

86. N. Sato, Corrosion, 45, 354 (1989).

87. N. Sato, Corrosion Science, 31, 1 (1990).

88. P. Schmutz and G.F. Frankel, Journal of The Electrochemical Society, 145, 2285

(1998).

89. P. Schmutz and G.F. Frankel, Journal of The Electrochemical Society, 145, 2295

(1998).

90. P.A. Schweitzer, "Corrosion Engineering Handbook", New York, NY: M.

Dekker, Inc. (1996).

91. A. Sehgal, D. Lu, and G.S. Frankel, Journal of The Electrochemical Society, 145,

2834 (1998).

92. K. Shimizu, G.M. Brown, K. Kobayashi, G.E. Thompson, and G.C. Wood,

Corrosion Science, 34, 1853 (1993).

93. A.G. Shrott, G.S. Frankel, A.J. Davenport, H.S. Isaacs, C.V. Jahnes, and M.

A.Russak, Surface Science, 250, 139 (1991).

94. L.F. Spencer, Metal Finishing, 58, 58 (1960).

95. I. Suzuki, "Corrosion Resistant Coatings Technology", New York, NY: Marcel

Dekker, Inc. (1989).

200 96. H.E. Townsend and R.G. Hart, Journal of The Electrochemical Society, 131, 1345

(1984).

97. P. Trathen, B.Hinton, K. Hammon, T. Hughes, and R. Taylor, "The Variability of

Chromate Conversion Coating Process and the Effect on Corrosion Protection

Performance", CSIRO (1994).

98. J.A. Treverton and M.P. Amor, Transactions of the Institute of Metal Finishing,

60, 92 (1982).

99. J.A. Treverton and M.P. Amor, Journal of Materials Science, 23, 3706 (1988).

100. J.A. Treverton, A. Bosland, and J.M. Brown, Corrosion Science, 30, 1159 (1990).

101. J.A. Treverton and N.C. Davis, Metals Technology, 480 (1977).

102. J.A. Treverton and N.C. Davis, Surface and Interface Analysis, 3, 194 (1981).

103. R.L. Twite and G.P. Bierwagen, Progress in Organic Coatings, 33, 91 (1998).

104. J.S. Wainright, O.J. Murphy, and M.R. Antonio, Corrosion Science, 33, 281

(1992).

105. J.R. Waldrop and M.W. Kendig, Journal of The Electrochemical Society, 145,

L11 (1998).

106. J. Wan, G.E. Thompson, K. Lu, and C.J.E. Smith, Physica B, 208&209, 511

(1995).

201 107. S. Wernick, R. Pinner, and P.G. Sheasby, "The Surface Treatment and Finishing

of Aluminum and its Alloys", Metals Park, OH: ASM International (1987).

108. C. Wick, "Tool and Manufacturing Engineers Handbook", Dearborn, MI: Society

of Manufacturing Engineers (1983).

109. L.F.G. Williams, Surface Technology, 5, 105 (1977).

110. L.F.G. Williams, Surface Technology, 7, 113 (1978).

111. L. Xia, Formation and Function of Chromate Conversion Coating on Aircraft

Aluminum Alloy Probed by Vibrational Spectroscopy, in Department of

Chemistry. 2000, The Ohio State University: Columbus, OH.

112. L. Xia and R.L. McCreery, Journal of The Electrochemical Society, 145, 3083

(1998).

113. L. Xia and R.L. McCreery, Journal of The Electrochemical Society, 146, 3696

(1999).

114. Z. Yu, H. Ni, G. Zhang, and Y. Wang, Applied Surface Science, 62, 217 (1992).

115. W. Zhang and R.G. Buchheit, Journal of The Electrochemical Society, Submitted,

(2002).

116. W. Zhang, B. Hurley, and R.G. Buchheit, Journal of The Electrochemical Society,

149, B357 (2002).

202 117. J. Zhao, G. Frankel, and R.L. McCreery, Journal of The Electrochemical Society,

145, 2258 (1998).

203