INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bloedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did rx)t send UMI a complete manuscript and there are missing pages, these will t>e noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have t>een reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are availat)le for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
Bell & Howell Information arxl Learning 300 North Zeeb Road, Ann Aitx>r, Ml 48106-1346 USA 800-521-0600 UMT
FORMATION AND FUNCTION OF CHROMATE CONVERSION COATING ON AIRCRAFT ALUMINUM ALLOY PROBED BY VIBRATIONAL SPECTROSCOPY
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate School of
The Ohio State University
By
Lin Xia, B.S. *****
The Ohio State University
2000
Dissertation Committee: Approved by
Professor Richard L. McCreery, Adviser
Professor Gerald S. Frankel Adviser Professor Terry L. Gustafson Department of Chemistry UMI Number 9971664
UMI
UMI Microform9971664 Copyright 2000 by Bell & Howell Information and Leaming Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Leaming Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor. Ml 48106-1346 ABSTRACT
A Chromate Conversion Coating (CCC) is currently one of the most effective methods for protecting aluminum alloys from corrosion. Its unique “self-healing” property has been proved to be critical in corrosion prevention. During the formation process, Cr^' is “stored” in the CCC films. Under in-field conditions, most of the Cr^‘ can leach out and diffuse to local defects, and stop corrosion. However, the involvement of highly toxic Cr'^' makes CCC system environmentally hazardous. In order to find less- toxic alternatives, the formation and protection mechanisms of CCC must be imderstood.
Formation and function of CCC film are the focus of this study, and vibrational spectroscopy was chosen due to its superior structural sensitivity. First, the structure of
CCC film was characterized. The structural similarity between CCC film and a synthetic
Cr-mixed-oxide was found, and certain tests were conducted on the bulk synthetic powder which were not feasible on the thin film. All of the structural studies indicated that CCC film is mainly a Cr'“-hydroxide gel layer, which adsorbs Cr'^'-oxy species through Cr"'-0-Cr'^' chemical bonds. Further analysis revealed the reversible Cr'"-Cr'^’ adsorption-desorption equilibrium, and a mathematical model ("Langmuir" model) was established to explain the Cr^' storage-release mechanism quantitatively. In addition, the function o f Fe(CN)6^ , an additive in the coating solution, was studied. The results indicate that Fe(CN)6^' mediates the slow reaction between A1 and
Cr^', and the mediation mechanism can be illustrated as below:
Fe(CN)6^-+Al = Fe(CN)6'‘‘+
t______I, Fe(CN)6"' + Cr^' = Fe(CN)6^'+ Cr'"
In general, the formation of CCC is mediated by Fe(CN)6^', thus A1 reduces Cr^' quickly and generates Cr‘"-hydroxide on the alloy surface. The nascent Cr'"-hydroxide is chemically active enough to form chemical bonds with Cr^' from the solution, through
Cr'"-0-Cr^' bonding. Such Cr”‘-0-Cr'^* structure can form and break up reversibly according to the Langmuir model, providing mobile Cr'^' for in-field protection.
Ill To the memory o f my father.
To my mother, my husband.
and my fam ily..
IV ACKNOWLEGMENTS
The course of pursuing a dream is never easy. From that elementary school in central China to the laser lab in Newman Wolfrom, each step of this journey had been made possible by the help from a lot of people.
The first person I'd like to say "thank you" to is my advisor. Dr. Richard L.
McCreery. His support, encouragement, and enthusiasm had made this thesis possible.
His guidance and wisdom are essential in the completion of this project. Meanwhile, he provided a fnendly environment for studying and doing research, and that really helped me in appreciating the joy of scientific research.
My sincere gratitude also goes to Dr. Gerald S. Frankel, for his informative and inspiring discussions, and also for his help in several other aspects. The scientific conversions with Dr. Martin Kendig and Dr. Eiji Akiyama were also very exciting. I want to thank Dr. Terry L. Gustafson for his kindness and help, and Dr. Gordon Rankes for his suggestions at the early stage of my research.
McCreery group’s members, who had been a very important part of my life during the past few years, deserve more than a “thank you ”. The former group members,
Tze-Chi Kuo, HHY, Kristen Frost, Jun Zhao, Yi-Chun Liu, Ken Ray, Pei-Hong Chen, and Angle Horn, had been very kind to me and offered a lot of help, especially TC, who had taught me how to use almost all the spectrometers and software in the lab. Special thanks also go to Jeremy Ramsey and Stacy Du Vail, for their friendship, kindness, and understanding. I will always remember the witty conversions and the sparkling discussions with them. I appreciate the help from the other current group members. Bill
McGovern (whose vivid mimic always made me laugh), Srikanth Ranganathan (a real smart gentleman), Ilson Steidel (who is very kind and hardworking). Bill Clark (who had released me from the strenuous “hammer-bandage” procedure), and Jingya Wu (whose happiness always brighten up the sky over everyone). The new girls, Belinda Hurley, and
Yang-Yan Hu, also offered a lot of help and are sincerely appreciated.
I’m obliged to my mother, my brothers, and sisters. Without their endless support and sacrifice, all these would be impossible. I’m very grateful for everything they had done for me. Finally, but not the least, I want to thank my husband, Hai-Ying, who had been with me through good and bad times, his presence in my life makes all the hard work meaningful.
VI VITA
January 27, 1971 ...... Bom - Shanghai, China
1993...... B.S. Chemistry, Fudan University, Shanghai, China
1993-1995 ...... Graduate Research and Teaching Assistant, Fudan University, Shanghai, P. R. China
1995-present ...... Graduate Teaching and Research Assistant, The Ohio State University
PUBLICATIONS
1. L. Xia, and R. L. McCreery, “Structure and Function of Ferricyanide in the Formation of Chromate Conversion Coatings on Aluminum Aircraft Alloy”, J. Electrochem. Soc., 146(10), 3696-3701(1999)
2. L. Xia, R. L. McCreery, “Chemistry of a Chromate Conversion Coating on Aluminum Alloy AA2024-T3 Probed by Vibrational Spectroscopy”, J. Electrochem. S'oc., 145(9), 3083-3089 (1998)
3. Kang-Nian Fan, Wen-Ning Wang, Lin Xia, Shi-Lin Pan, “Studies on the Harmonic Force Field and Vibrational Spectra of y-picolinic Acid and Its Isotopic Derivatives with Ab Initio Calculations”, Huaxue Xuebao 53(7), 637-44 (1995) (Chinese)
FIELDS OF STUDY
Major Field: Chemistry
VII TABLE OF CONTENTS
page
Abstract...... ii
Dedication ...... iv
Acknowledgments ...... v
V ita...... vii
List of Tables ...... xii
List of Figures ...... xiv
List of symbols and abbreviations ...... xix
Chapters:
1. Introduction ...... 1
1.1 General introduction ...... 1
2 Fundamental corrosion science ...... 4
1.2.1 Mixed potential theory and polarization resistance method ... 4
1.2.2 “Autocatalytic” pitting mechanism ...... 8
vm 1.3 Application of spectroscopic techniques in corrosion chemistry studies...... 8
1.4 Brief introduction to Chromium chemistry ...... 15
1.5 Research objective ...... 28
2. Identification of Chromium-species in CCC on AA 2024-T3 Aluminum alloy by vibrational spectroscopy ...... 30
2.1 Introduction ...... 30
2.2 Experimental ...... 33
2.3 Experimental results ...... 46
2.4 Discussion ...... 66
2.5 Conclusions ...... 72
3. Structure and function of ferricyanide in the formation of CCC on AA 2024-T3 Aluminum alloy ...... 73
3.1 Introduction ...... 73
3.2 Experimental ...... 75
3.3 Experimental results ...... 80
3.3.1 Structure o f CN species in C C C ...... 80
3.3.2 Function o f Fe(CN)6^’in CCC form ation ...... 90
3.3.3 Rp of AA 2024-T3 and A1 with various coating thickness 104
3.4 Discussion ...... 106
3.5 Conclusions ...... 110
4. Storage and release of soluble hexavalent Chromium from Chromate- Conversion-Coatings: Equilibrium aspects of Cr^' concentration ...... I ll
IX 4.1 Introduction ...... I l l
4.2 Experimental ...... 112
4.3 Experimental results ...... 123
4.4 Discussion ...... 134
4.5 Implications to corrosion studies ...... 145
4.6 Adsorption-release model ...... 147
4.6.1 Release of Cr'"'from C C C ...... 147
4.6.2 Release of Cr'^‘ from Cr-mixed-oxide ...... 149
4.6.3 Adsorption of Cr^' to synthetic Cr“‘ hydroxide ...... 149
4.7 Conclusions ...... 150
5. Structure and property evolution of CCC-Cr^' species during aging 151
5.1 Introduction ...... 151
5.2 Experimental ...... 152
5.2.1 Isotopic substitution ...... 152
5.2.2 Synthesis of Cr‘”-Cr'^' mixed-oxides and spectroscopic studies 153
5.2.2.1 Cr‘"-Cr'^‘ mixed-oxides at constant loading level, but different [Cr'^*]t=o ...... 153
5.2.2.2 Cr'"-Cr'^’ mixed-oxides at various loading ...... 154
5.2.3 Aging treatment of CCC and Cr'"-Cr^' mixed-oxide ...... 154
5.2.4 Protection from several “artificial” films or treatm ents ...... 155
5.3 Quantum Calculation ...... 156
5.4 Results ...... 159
5.4.1 Iso topic substitution ...... 159
X 5.4.2 Quantum calculation results ...... 162
5.4.3 A stepwise study of the structural evolution during adsorption, release and heating ...... 165
5.4.4 Structural evolution during heating ...... 172
5.4.5 Protection properties of several artificial “films” and some treatment...... 186
5.5 Discussion ...... 194
5.6 Conclusions ...... 204
6. Some effects of Copper on Aluminum ...... 205
6.1 Introduction ...... 205
6.2 Experimental ...... 205
6.3 Results ...... 206
6.4 Discussion ...... 214
7. Summary & Future work ...... 217
References...... 224
XI LIST OF TABLES
Table page
1.1 Rate constants of some typical reactions involved in Cr*“ hydrolysis...... 23
1.2 pKa of some Cr‘" hydrolysis species ...... 23
2.1 Composition of Alodine 1200S coating solution ...... 31
2.2 Composition of AA 2024- T3 Aluminum alloy ...... 31
2.3 Simultaneous of measurement of [Cr‘“] and [Cr^'] of known solutions 41
2.4 Molar absorbtivity of Cr"' and Cr^' in 0.25M NaOH ...... 41
2.5 Composition of CCC and mixed-oxides ...... 47
2.6 FTIR of CCC, Cr"'-Cr^' mixed-oxide, and Al"‘-Cr''' mixed-oxide ...... 48
2.7 Raman frequencies of CCC, mixed-oxides, and Alodine solution ...... 52
2.8 Raman shift of Cr'^’-O, C r'"-0 stretching modes from standard compounds 52
2.9 Raman peak frequencies of fresh film or solid as a function of p H ...... 59
2.10 Frequencies of Cr'"'-0 vibration upon heating...... 60
3.1 Components in tests solution for growth-curve studies ...... 78
3.2 CN stretching frequencies o f Fe(CN)6^‘'‘*'derivatives ...... 84
XII 3.3 Literature values of cyano stretch of several metallocyanide compounds ... 88
3 .4 Rp of AA 2024-T3 after coating treatment in various solutions ...... 104
3.5 Rp of pure A1 after coating treatment in various solutions ...... 105
4.1 Observed equilibrium [Cr'^’] in solution under various conditions ...... 116
4.2 Equilibrium Cr^' concentration in solution after Cr“‘-hydroxide adsorption 119
4.3 Solution concentrations of Cr‘“ and Cr^' during Cr"'-Cr^' pH cycling experiment ...... 120
4.4 Solution concentrations of Cr'"* during Cr 2 0 3 -Cr'^* pH cycling experiment 121
4.5 Solution concentrations of Cr'"* during Al"'-Cr^' pH cycling experiment ...... 122
5.1 Comparison between experimental and calculated vibration frequencies of standard compounds ...... 157
5.2 Experimental freouencies of C r"^' and the calculated frequencies of hypothetical Cr‘“ ' clusters ...... 158
5.3 PCR parameters used in the analysis of the spectra from heated CCC and from Cr’*'-Cr'^' mixed-oxide ...... 178
5.4 [Cr'^'j released by heated CCC film or Cr"'-Cr^' mixed-oxide ...... 186
5.5 Protection effect of some artificial “films” ...... 187
5.6 Protection property of Cr"’ and Cr^' on AA 2024-T3 ...... 188
6.1 Rp of pure Al, pure Cu, Al-Cu pair electrodes ...... 207
xni LIST OF FIGURES
Figure Page
1.1 Schematic illustration of the Mixed Potential theory ...... 6
1.2 Pitting corrosion mechanism on stainless steel ...... 9
1.3 Schematic illustration of IR absorption, Raman scattering, and UV- Vis. absorption ...... 13
1.4 Corundum structure of CrzO] ...... 17
1.5 “Active” Cr”‘-hydroxide and its fate during aging ...... 18
1.6 Stepwise Cr‘” hydrolysis in homogeneous solution ...... 20
1.7 Time dependence of the aging in stirred and stirred suspensions at pH 5.06 and 25°C ...... 21
1.8 Distribution of Cr”* in the different oligomers as a function o f p H 22
1.9 Structure of CrOa^ , CrzO?^, and CrOa ...... 25
1.10 Crystal structure of CrgOai ...... 27
2.1 Instrumental layout of Chromex 250 spectrometer ...... 35
2.2 Instrumental layout of Kaiser spectrometer ...... 36
2.3 Instrumental layout of Bruker model 55 Equinox FT-IR spectrometer 37
xi>/ 2.4 Instrumental layout of Lambda 20 UV/VIS/NIR spectrometer ...... 40
2.5 Calibration curves of Cr 0 4 ^‘ and C r”’ (basic)...... 42
2.6 UV-VIS spectra of Cr"' and Cr^' in 0.25M NaOH solution ...... 43
2.7 Experimental setup of “migration” experiments ...... 45
2.8 IR spectra of CCC, Cr"'-Cr^' mixed-oxide, and Al‘"-Cr'^‘ mixed-oxide 49
2.9 IR spectra of K 2Cr0 4 , KzCrzO?, Cr(OH)3, and K3Fe(CN)6 solids 50
2.10 Raman spectra of Alodine, CCC, Cr'"-Cr^' mixed-oxide, and Al‘”- Cr'^^ mixed-oxide ...... 53
2.11 Raman spectra Cr-containing compounds and CCC film ...... 54
2.12 The algorithm of “splice.ab” ...... 55
2.13 Raman spectra of CCC film on AA2024-T3 at various pH ...... 56
2.14 Raman spectra of Cr'"-Cr^' mixed-oxide at various pH ...... 57
2.15 Raman spectra of Al‘"-Cr'^’ at various p H ...... 58
2.16 Raman spectra of heated CCC, Cr**'-Cr'^‘ mixed-oxide, and Al"'-Cr^' mixed-oxide ...... 61
2.17 In situ Raman spectra of Cr*"-hydroxide covered by Alodine solution 64
2.18 In situ Raman spectra of Cr'”-hydroxide covered by Cr^' solution 65
2.19 Proposed Cr^'-species strucutre in CCC film ...... 70
3.1 Possible degradation of Fe‘"(CN)6^' in acidic conditions ...... 76
3.2 Structure of Berlin Green and Prussian Blue ...... 77
3.3 Raman spectra of fresh CCC film and after various treatment ...... 81
3.4 IR spectra of fresh CCC film and after various treatment ...... 82
3.5 Raman spectra of several CN-containing standard compounds ...... 85
X"V 3.6 IR spectra of several CN-containing standard compounds ...... 86
3.7 Degradation product o f Fe(CN)6^' in pH 1.5 HNO 3 solution, with excess amount of NazSzOg ...... 89
3.8 Correlation between average 860cm’' intensity and film thickness 91
3.9 Growth curves o f CCC film on AA 2024-T3 ...... 92
3.10 Growth curves of CCC film on 99.999% Aluminum ...... 93
3.11 Raman spectra of CCC film grown on AA 2024-T3 fi'om several coating solutions ...... 95
3.12 Possible mediation reactions ...... 97
3.13 Raman monitoring of reactions between Fe(CN)6^ and Al, and between Fe(CN)6 and Cr'^’ ...... 98
3.14 Test of possible reaction between Prussian blue and Cr^', and between Berlin green and A l ...... 99
3.15 Fe(CN)6^ mediation mechanism ...... 101
3.16 Growth curves of CCC on AA2024-T3 mediated by several red-ox p airs ...... 103
4.1 UV-Vis spectra of Cr^' solution at different pH but the same total concentration of 8.30 x lO’^M ...... 113
4.2 Calibration line of Assonm vs. [Cr^' ] (pH range from 2.01 to 9 .4 7) ...... 114
4.3 “Release curves” of CCC in nanopure water under different area/volume ratio ...... 124
4.4 “Release curves” of Cr"'-Cr^' mixed oxide in nanopure water or 0.1 M NaCl solution ...... 125
4.5 Cr^' adsorption by Cr'"-hydroxide ...... 128
4.6 pH cycling of a mixture of 62mM Cr(N 0 3 ) 3 9 H2O and 9.5mM KzCrzO? m ixture ...... 131
4.7 pH cycling of a mixture of 250ml 6.6mM KzCrzO? and 0.3 8g Cr 2 0 3 132
XVI 4.8 pH cycling of a mixture of 66mM A 1(N0 3 ) 3 9 H2O and 1 ImM KzCrzO? 133
4.9 Comparison of three possible models ...... 135
4.10 Adsorption and release equilibrium ...... 137
4.11 [Cr'^'jfinai of releasing experiment and the fitting result fi-om Equation 4 -2 ...... 140
4.12 pH cycling of C r"^‘ mixture and the fitting result fi^om Equation 4-3 143
5.1 Raman spectra of Cr^' solutions in H 2O or 87.7% H20^* ...... 160
5.2 Raman spectra of Cr"'-hydroxide prepared in H 2O, 87.7%, H20^^ ...... 161
5.3 Optimized hypothetical Cr'"-0-Cr^' clusters ...... 163
5.4 Comparison of calculated Raman spectra with experimental spectmm 164 of Cr "-Cr'^’ mixed-oxide ......
5.5 Raman spectra of Cr"'-Cr^' mixed-oxide at various loading level 167
5.6 Raman spectra of Cr "-Cr^' mixed-oxides prepared under various [Cr^']pQ, but same Cr'^’ loading level and p H ...... 168
5.7 Intensity variations of 858cm ' and 943cm ' Raman peaks on CCC film during Cr^' release ...... 170
5.8 Intensity variation of 858cm ' and 945cm ' on Cr'"-Cr^' mixed-oxide during Cr^' release ...... 171
5.9 Raman spectra o f heated CCC film ...... 174
5.10 Raman spectra of heated Cr'"-Cr^' mixed-oxide ...... 175
5.11 858cm' Raman band intensity of CCC film after heating ...... 176
5.12 Raman peak intensity of Cr'"-Cr'^ mixed-oxide after heating ...... 177
5.13 PCR loading factors fi*om the Raman spectra of heated CCC ...... 180
5.14 Score variations of the loading factors for heated CCC ...... 181
5.15 PCR loading factors fi’om the Raman spectra o f heated Cr'"-Cr'^'- mixed-oxide ...... 183
xvii 5 .16 Score variations of the loading factors for heated Cr'"-Cr'^‘-mixed- oxide ...... 184
5.17 Surface images of AA2024-T3 after various treatments ...... 189
5.18 Raman spectra of regular CCC and depleted-CCC on AA 2024-T3 .... 192
5 .19 Raman spectra of regular CCC, CCC film after immersion in 0.5ml nanopure water for 3 days, and the NazSO] reduced CCC film after 2 days immersion in 0.5ml of O.IM NaCl ...... 193
5.20 Possible structural changes in CCC film during thermal treatment 198
5 .21 Possible structural changes in Cr-mixed-oxide during thermal treatment...... 199
5 .22 Crystal structure o f a Cr'”-cage compound ...... 203
6.1 CCC film growth rates on pure Al, AA2024-T3, and Al-Cu pair electrodes in Alodine solution ...... 208
6 .2 Open Circuit Potentials of pure Al, pine Cu, AA2024-T3 pair electrodes in synthetic coating solutions ...... 210
6.3 Open Circuit Potentials of AA2024-T3 electrode in synthetic coating solutions ...... 211
6.4 Open Circuit Potentials of AA2024-T3 electrodes in de-aerated synthetic coating solutions ...... 213
6.5 Possible changes when AA2024-T3 was immersed in solution-Q with or without Fe(CN)6^' ...... 216
7.1 Schematic illustration of the overall CCC formation and release 219
XVllI LIST OF SYMBOLS AND ABBREVIATIONS
symbols/abbreviations explanation
CCC Chromate Conversion Coating formed on AA2024-T3 alloy or pure Al
V vibrational frequency with unit “cm ’”
A UV/VIS absorption
E molar absorptivity (M ’ #cm ’)
r surface Cr^’ concentration = total moles of Cr^’ on each unit of surface area
rs surface Cr'^’ concentration at saturation level
V volume o f solution
S microscopic surface area of CCC film or Cr"‘-hydroxide
area/volume the ratio of CCC film's geographic surface area to the total volume of surrounding solution
[X] solution concentration of species X
[X]t=o solution concentration of species X at the beginning of the experiment
[X]equii equilibrium concentration of X in solution
XU III _vi N vi total moles of Cr'^' in the whole CCC film or Cr"‘-Cr''‘- mixed-oxide solid
equilibrium constant of Ci^”-0-Cr'^‘ formation fi-om Cr‘“- hydroxide and aqueous Cr .VI
Polarization resistance
OCP Open Circuit Potential
C r > i i / v i Cr'" -Cr^'VI mixed oxide or: hypothetical clusters containing Cr'" -O- Cr^' bonding
Cr'II! trivalent Chromium
Cr VI hexavalent Chromium
Al-Cu pair an electrode made by electrically connecting a 1x1 cm’ electrode pure Al electrode and a 1x1 cm pure Cu electrode
XX CHAPTER 1
INTRODUCTION
1.1 General Introduction
The extensive application of metal makes corrosion a common fact of our lives.
From the tiny rusty holes on the body panels of cars to the big cracks in the metallic frames of a bridge, corrosion is costing us a great deal of money, and sometimes human lives [I]. In the United States alone, annual cost of corrosion has been estimated to be more than $350 billion [2]. A recent survey shows that the US military spends more than
S3 billion each year to protect the aluminum skin of military airplanes, bodies of ships and other metallic appliances [2]. The total cost is much greater if commercial vehicles and reinforced concrete structiu’es are taken into account.
Driven by the economic impact of corrosion, an enormous amount of funding and human effort have been devoted to corrosion studies, aiming at a better understanding and more effective prevention of corrosion. Among all the metals and their alloys, Al and its alloy are one of the most studied. Aluminum and its alloys are commonly used in
industry, mainly because of their outstanding physical properties. They have low density,
high corrosion resistance and excellent thermal/electrical conductivity [1-3]. Al has
appealing appearance and fairly low toxicity [1-4]. Its ready availability also contributes
to its leading engineering applications [4].
For certain applications, pure Al is too soft and weak, therefore small amounts of alloying elements are added [5]. An example of the high strength aluminum alloy is
AA2024-T3, which is commonly used to build military aircraft structures. AA 2024-T3 was made by alloying Al with 4-5 % (w%) of Cu and a smaller amount of other elements
(Mg, Fe, Mn). After heat treatment at 920°F, the Cu-Al solid solution is rapidly cooled to room temperature by water quenching, and then aged in air. Due to the fairly low solubility of Cu in Al at room temperature, CuAh (also known as 6 phase) and AlMgCu compounds precipitate during aging[3]. The hardening of the dispersed Cu-rich phases strengthens the material [1-3, 5].
However, addition of comparatively noble Cu to highly active Al decreases the alloy's corrosion resistance. The typical forms of corrosion on AA2024-T3 include pitting/crevice, inter-crystalline corrosion, general corrosion, and SCC (Stress Corrosion
Cracking) [1-3,5]. Among them, pitting is the most common type. Both pitting and crevice corrosion are impredictable and insidious[ 1-3,5]. If not detected in time, the fast propagation of pits and crevices can cause premature failure of metallic structures [1-2].
During the long battle against corrosion, scientists and engineers have developed all kinds of corrosion prevention methods [1-3,5-6]. Manufacturing procediwes have been optimized, and special considerations have been taken in designing parts, all
2 targeted at better corrosion resistance [1-3,5]. Organic coatings are applied on top of alloys to block moisture, thus decreasing the pitting and crevice corrosion [1,2,6]. One of the most effective and economic methods is a chemical conversion coating. Chemical conversion coating is essentially the thickening of the surface oxide film by the reaction between A1 alloy and certain coating solution [2,6]. Corrosion inhibitors, such as Cr^', are usually included in the coating solution, hence the name Chromate-conversion- coatings (CCC).
CCC films are easy to apply. By simply dipping an aluminum alloy into a coating solution, a thin layer of Cr^'-containing film can be generated on the surface of the alloy
[7]. Other than promoting adhesion of the top coating [8-12], CCC itself can provide protection long after the film's formation, and the unique "self-healing" property is a lethal weapon against random pitting corrosion. “Self-healing” means that when there is any locally exposed metal due to flaws (scratches, pit, and crevices), Cr'^‘ can be released from the film and "migrates" to the flaws to stop corrosion [13-15].
Despite the excellent corrosion prevention properties of Chromate-conversion- coatings, the large amount of Cr'^' used in the coating procedure generates a tough problem for the industry. Although a minute amount of Cr*“ is essential for human health[16-19], Cr^' is a strong oxidizing agent and a toxin. Recent biological studies revealed the carcinogenic effect of Cr^' on both lab animals [20] and humans [21].
Evidence is available to explain the damages on DNA done by Cr^' [22-30], even by Cr“‘
[31-32]. The EPA has since imposed more strict regulations on Cr^''s usage and Cr^'- containing waste treatment[33]. For the long-term benefit of the industry, the environment, and ultimately human society, the Cr'^'-containing coating bath formula must be replaced by some other chemical systems with similar property but less toxicity [34]. There are countless possible ways to find an alternative. Most of the established coating formulae have been determined largely empirically without understanding the mechanism. But one of the best ways is to figure out the protection mechanism of CCC film first, then find some less toxic systems to mimic CCC's function in corrosion prevention.
In the last 10-20 years, extensive research has been conducted to achieve a better understanding of CCC film’s protection mechanism. The results of this research can be classified into two categories, which are CCC film structure characterization and anticorrosion mechanism investigation. Due to the limited space, the relevant details of previous research results will be reviewed in each chapter individually. Here, only a general introduction of the relevant theories and techniques will be provided.
1.2 Fundamental Corrosion Science
1.2.1 Mixed Potential Theory and Polarization Resistance Method
For any corrosion mechanism study, some standardized methods must be available to characterize the metal’s corrosion susceptibility. Polarization resistance is one of the simplest and most reliable techniques [1-2]. The principle of polarization resistance is closely related to the mixed potential theory [1-2]. Detailed imderstanding of corrosion electrochemistry theories and concepts can be foimd in reference 1 -2. The following is a summary of the basic concepts and theories that are relevant to this work. When metal M is corroding in an acidic solution, the following two reactions occur simultaneously on the metal surface:
M -» M"^ + n e' anodic reaction (1-1)
2H^ + 2 e —>H 2 cathodic reaction (1-2)
Each reaction has its own standard reduction potential (E°) and exchange current density
(i°). On the conductive metal surface, both reactions have to polarize and accommodate some common potential, as shown in Figure 1-1. The common potential is called the mixed potential, or corrosion potential, Ecorr- At Econ-, cathodic current density equals to anodic current density. The current density at Ecorr is defined as the corrosion current density, as shown in Equation 1-1. The anodic over-potential t|a and cathodic over potential r|c can be expressed in Equation 12a and Equation l-2b, respectively, assuming activation controlled kinetics.
ic = ia = icorr Equation 1 -1
r|a = Pa log( ^ ) Equation l-2a
Tic = Pc log( ^ ) Equation 1 -2b ia and ic are the anodic and cathodic current density, respectively. /° and / ° are the exchange current densities of the anodic reaction and the cathodic reaction, respectively.
Pa and Pc are the corresponding Tafel constants.
When an external potential different from Ecorr is applied on such a corroding surface, the current densities will be changed due to the potential polarization.
iapp,c = ic -ia Equation 1-3 iapp,a = ia -ic Equation 1-4 H*/H e
corr
*■ 1 (A ) corr
Figure 1-1. Schematic illustration of the Mixed Potential Theory [2] iapp.c or iapp,a IS the applied cathodic or anodic current density, corresponding to Eapp
Ea = pa log(^^) Equation 1-5 ^corr
Ec = Pc log(^^). Equation 1-6 ^corr
After conversion of Equation-5 and 6 to exponential form and combination with Equation
1-3, we have
iapp,c = icorrClO """^ - 10'“ ^^’) Equation 1-7
If £ is small enough, iapp,c or iapp,a, becomes linear with e. Thus,
de Equation 1-8 _^app _ 2.303,
Therefore, Rp is inversely proportional to the corrosion current, which is proportional to the corrosion rate. In other words, when Eapp is very close to Ecorr, the small changes of
Eapp will cause a corresponding current density change, and the slope of the Eapp vs. the current is Rp. Experimental observations have confirmed that a lower polarization resistance is correlated to a faster corrosion rate [ 2 ].
The experimental measurement of Rp is simple and convenient. Once the open circuit potential of the sample stabilizes at Ecorr, an Eapp in the range from Ecorr-20mV to
Ecorr+ZOmV is imposed on the sample, and the slope of Eapp vs. current density is Rp. 1.2.2 ‘‘Autocatalytic’’ Pitting Mechanism
The pitting mechanism of stainless steel has been studied extensively. It is believed that pitting on aluminum alloys happens in a similar way, following the so- called “auto-catalytic” mechanism [1-2]. For a piece of stainless steel being immersed in
NaCl solution, as shown in Figure 1-2 [2], pits initiate above Epù. Such a potential could be reached by a potentiostat or exposure to an oxidizer. Inside a just initiated pit, the anodic dissolution of Fe generates positively charged Fe^% which attracts Cl . Hydrolysis of Fe‘* drives the local pH in the pit to be more acidic, thus accelerating anodic dissolution, which in turn further concentrates Cl inside the pit. The ultimate corrosion product, Fe‘”-hydroxide, forms a “cap” on the top of the pit. The cap prevents Fe“* from getting out of the pit but allows Cl to get in. The exterior cathodic reaction couples with the interior anodic dissolution, allowing the pit to propagate.[ 2 ]
1.3 .Application of Spectroscopic Techniques in Corrosion Chemistry Studies
Corrosion is a phenomenon that occurs on an interface between a metallic base, bare or covered by certain types of coatings, and an electrolyte. The chemical reactions on such interfaces are directly related to the corrosion mechanism. Many types of surface analysis techniques have been applied to study such reactions in detail. The most popular techniques include X-ray photoemission spectroscopy (XPS)[35-42], Auger Emission
Spectroscopy (AES) [43], X-ray Absorption [13-14,44], Transmission Electron
Microscopy [45-47], Secondary Ion Mass Spectrometry (SIMS) [48-49], Scanning Probe
Microscopy (SPM) [50-52], and Atomic Force Microscopy [53-54]. Fourier Transform Cl-
Fe(OH)2 + 2H Fe
Accelerate dissolution
Figure 1-2. Pitting corrosion mechanism on stainless steel. (Modified from Figure 3-19 of Reference [1]) Infrared (FTIR) absorption [49, 55-56] and Raman Spectroscopy are gaining recognition in recent years [10, 55, 57].
XPS and AES are powerful tools in surface elemental analysis. They provide extremely high surface sensitivity [58-60]. The characteristic binding energies can help identify the element and its oxidation state. The combination with sputtering or mapping techniques greatly extended the function of XPS/AES, and lengthwise-depth distribution of elements can be obtained [58-60]. Although information about chemical bonds can be detected in some special cases, XPS/AES are usually not sufficient to distinguish minor structural changes. For instance, CrzO?^ ,Cr 0 4 ^ , and CrO] all have very similar binding energy for Cr^', although their structures are different [61]. XPS/AES are accepted as quantitative analysis techniques, but only on a semi-quantitative level [58-60].
Furthermore, the highly energetic photons or electrons can sometimes change the surface species, especially under ultra high vacuum [62].
SIMS shares the same advantages of XPS/AES. It has extremely high sensitivity, and can be used to obtain depth-profile and surface images [58-60]. But SIMS also has several disadvantages. First, the sputtering process on a sample surface depends on the specific element/matrix combination, so called "matrix effect". The quantitative correction for such matrix effect can be very complicated and time-consuming. Secondly, the sputtering breaks up the chemical bonds of siuTace species, therefore the surface may be damaged and the detailed chemical bonding information may be lost [58].
X-ray Absorption Near Edge Structure (XANES) has gained some popularity in recent years. It is a non destructive technique, and measurement can be carried out under ambient pressure. The damaging effect of X-ray becomes less significant under normal
1 0 pressure. In situ monitoring can be conducted. Similar to other techniques involving X- rays, the high-energy photons could still change the surface species after prolonged exposure. In order to extract quantitative information, tedious correction procedures are still necessary.[58]
Optical Vibrational Spectroscopic techniques, IR and Raman, are less sensitive compared with the techniques described above [58]. It may not be able to provide elemental information unless some distinctive functional groups are involved. However, vibrational spectroscopy has several unique advantages when the detailed surface structural information is concerned. First, both IR and Raman are very structure sensitive.
The position of an IR absorption band or Raman band depends not only on the elements and their oxidation states, but also their local symmetry and their interaction with the surrounding matrix. As mentioned before, XPS cannot distinguish CrzO?^ from CrO^" , because the oxidation state of Cr atom in a dichromate cluster or a chromate cluster is the same. However, the IR and Raman features of CrzO?^ and CrOa^'are quite different, due to the subtle differences in bond lengths, bond angles, and the dramatic difference in symmetry. Such high structural sensitivity is important for corrosion mechanism studies, since sometimes small changes in structure may result in significant difference in corrosion properties [63]. Secondly, IR and Raman are both nondestructive and can be conducted under ambient pressure. Although water and CO 2 can cause severe interference in IR, they are almost transparent in the most commonly used Raman shift ranges. This makes in situ Raman monitoring possible. For most corrosion studies, the metal alloy surface is chemically heterogeneous. Sophisticated IR microscopy and scanning Raman spectrometer make surface mapping of such samples possible, and the
11 combination of mapping with in situ provides even more information [58,64-66]. In
addition, the previous comprehensive vibrational spectroscopic studies offer a spectral database of countless chemicals interested in corrosion studies [67-68]. All in all, vibrational spectroscopy is a valuable asset in studying the structural variations during corrosion and protection.
The theoretical principles for IR and Raman are quite different [65, 6 6 , 6 8 ]. When molecules are illuminated by incident light, the interaction between the electromagnetic field and the molecules results in absorption, elastic scattering, and inelastic scattering of the incident photons. A schematic illustration of these processes is shown in Figure 1-3.
If the incident photon energy lies in the rtdd-injrared region, the molecules can undergo transitions between vibrational energy levels within the ground electronic state. For example, a transition from V'=0 to V'=l will absorb the incident photons with energy E =
(E2-E 1). The quantitative comparison between initial power, l°(v), and the power after passing through the sample, l(v), obeys Beer's Law:
r(i/) = Equation 1-9
A(y) = -lo g r(v ) = log Equation 1-10 /(V)
Transitions between ground electronic state and excited electronic state are possible, if the incident photons have energy in the UV/VIS region. The transitions are usually n-> n* or 7C n*. The same Beer's Law is followed, and this is called UV-
Visible Spectroscopy [69].
1 2 V ' = ? Excited electronic V = 1 state
V = 0
Virtual state
V ' = 2 Ground electronic V'=l (E,) state V' = 0 (E,) GO > - § I c/3I I c/3 il I c 'S < s i A I jac I LÜ
Figure 1-3. Schematic illustration of IR absorption, Raman scattering, and UV-Vis. absorption
13 Raman is the result of inelastic scattering. As shown in Figure 1-3, most of the scattered photons are from Rayleigh scattering, in which the scattered photon energy is the same as the incident photons. Only a tiny fraction, 10"^ - 10*’°, of the incident photons are inelastically scattered. If we define the energy of inelastic scattered photon as Vj, the incident photon energy as v°, and the energy difference between two vibrational energy levels as v%, then
Vs = (v° ± Vx) Equation I -11
Usually, the (v°-Vx) branch is called Stokes Raman, and (v°+Vx) branch is called anti-
Stokes Raman. Commonly discussed “Raman” is Stokes Raman. (v°-Vs) is called the
Raman Shift, which reveals the target molecule's vibrational frequency.
The selection rules for IR and Raman are also different.
• a vibrational mode with a net dipole moment change is IR active
• a vibrational model with net polarizability change is Raman active
For any molecule with a symmetry center, or symmetry plane, or S 4 symmetry, the IR and Raman spectra will not overlap. For other molecules, the two spectra usually overlap partially.
Raman is much weaker than Infrared absorption; the detection of the Raman signal becomes even more difficult with the overwhelmingly strong Rayleigh scattering superimposed on the Raman signals. This has hindered the application of Raman, despite the fact that Raman was discovered earlier than IR absorption [58,65,66]. It was not until recently, with technical advancements in several aspects, that Raman re-gained recognition. The discovery of the laser provided Raman an ideal source with very high monochromaticity, very low divergence, high radiance and coherence [70]. The 14 development of highly effective laser line rejection filters simplified Raman spectrometer design[58], while double or triple monochromators still show their advantages when extremely low Raman shifts are interested [58]. Another technological advancement is the CCD (charge-coupled device) array detector. Liquid N 2 cooled CCD detectors eliminate moving parts from a Raman spectrometer and provide high sensitivity [58]. On the other hand, along with FT-IR, FT-Raman has been developed to achieve greatly improved signal/noise ratio and resolution [58]. Nowadays, for most common applications, including many samples for corrosion studies, both IR and Raman spectra can be measured with ease, speed and quality. Partially for this reason, the research in this dissertation was done mainly by using Raman, IR and UV-Vis. spectroscopy.
1.4 Brief Introduction to Chromium Chemistry
The major targets of this dissertation are the structure and function of chromium species in CCC film. Therefore, a brief review of Cr chemistry is necessary.
Chromium element was first discovered in 1797 by French chemist Nicolas-Louis
Vauquelin in Siberian red lead, PbCr 0 4 - It was named chromium (from Greek chroma, color) because of its colorful compounds [71]. Cr can have oxidation states from -2 to +6, but the most stable and important state is Cr'" [4].
Cr'" has a valence electronic structure 3d^ 4s', and usually exists in octahedral complexes, with t:g levels singly occupied. The most commonly used Cr'" compound for this study is CrfNO])] «QHzO, in which Cr'" exists as hexaquo ion, Cr(OH 2 >6 ^ [4].
Cr(OH2)6^^ is a regular octahedron, which occurs in aqueous solution and several other hydrated Cr'"-salts, such as [Cr(H20)6]Cl3. Cr(OH2)6 ^^ solution is usually acidic (pKa=
15 4.3 [87]). When the pH o f the solution becomes more basic, the aquo ions condense to
form dimeric, trimeric, and oiigmeric hydroxo-bridged species. The ultimate product of
such a polymerization process is amorphous Cr*”-hydroxide. Cr‘“-hydroxide can be
converted to crystalline CriOs, by further cross-linking and losing all the water. The
crystal structure of Cr^Os [72] is shown in Figure 1-4.
The process of Cr(OH 2)6^ hydrolysis to generate chromium hydroxide has been
studied extensively. Thermodynamic and kinetic details of the hydrolysis reactions are
available, mostly for the early stages of the whole polymerization process. Relevant
results from past work are briefly reviewed here. Hydrolysis of Cr(OH 2 )6^ has been
studied in both aqueous suspension and homogeneous aqueous solution. In the former
case, the nascent form of chromium hydroxide solid, “active” chromium hydroxide, was
found upon quick addition of base to aqueous Cr(OH 2)6^ [73]. “Active” chromium
hydroxide has an octahedral layered structure, and one-third of the octahedral sites are occupied. The “cross-linking” in “active” chromium hydroxide does not involve any bridging hydroxide ligand, it is solely hydrogen-bonding [73-75], as shown in Figure 1-5.
“Aging” of the “active” hydroxide, in solution or in solid, transforms it into hydroxide cross-linked amorphous chromium hydroxide [73-75]. The hydrolysis of Cr(OH 2 )6 ^ in homogeneous aqueous solution involves p-OH formation. At pH 3.5-5.0, Cr(OH 2 )6^ \ and deprotonated aquo ions, Cr(OH)(OH 2)5^^, Cr(OH)2(OH2 )4 , react with each other to form dimers according to Equation 1-3 [76]:
Cr(OH)r(OH2)6-/" ‘/2 Cr2 (p-OH) 2(OH)s(OH2)'' ' + mH" (1-3)
16 o o o o o o o \ I / \ I / Cr Cr y I \ y I \ o o o o o o o \ I y \ I y Cr Cr y I y y I \ o o o o o o o
Cr Cr y I \ y I \ o o o o o o o
Figure 1-4. Corundum structure of Cr^O^ (Modified from Figure 1.7 of [74] )
17 H -mH,0 O High Oligomers ^ Cr O ' H
Cr(OH)n(OH2)6.,(3.")+ -2 H2O
dissolve -H" + H+
solution
solid 'OH — O C r O H O ^ I
Cr
-H,Q
Cr Cr High oligomers
Figure 1-5. “Active” Cr'"-hydroxide (1), and its fate during aging (Modified from Scheme 1 of reference [75])
18 Further cross-linking generates trimers [77-78], tetramers[79], hexamers[78]. The
structures of various dimers[80-83], trimers, [84-86], and tetramers [87-91], and their
properties in solution [92] were also studied. The structure of the hydrolysis product
depends on the formation conditions, such as pH, ionic strength, concentration, and aging
time. In general, as shown in Figure 1-6, p-OH serves as bridge between two adjacent
Cr'" centers, forming various sized Cr"‘-hydroxy clusters. The comparatively fast “inter-
conversion”[88] allows different configurations for each cluster size. The polymerization
rate is significantly faster between deprotonated species compared with their protonated
analogues [76-81, 84, 92, 93], and the stability of larger oligomers improves when pH becomes higher [94]. Higher concentration and longer aging time also favor higher oligomers [73,92]. Figure 1-7, and Figure 1-8 illustrate the effect of aging time, pH, and total Cr"' concentration on the oligomers’ distribution [73, 92]. The rate constants of some typical hydrolysis reactions are listed in Table 1-1, and Table 1-2 shows the pKa of some of the small oligmers.
19 H O C r(O H 2>,3- -► (HzO)4Cr^ ^Cr(OHz)^ O ^
H ^ 2 ? H Re-arrange I ^ Q X (HjO)4CrC: %: CKC ^ Cr(OHz)4 ^ O I O
Chromium hydroxide
Figure 1-6. Stepwise Cr'” hydrolysis in homogeneous solution (Partially adapted from reference [78-93])
2 0 %Crg
80 High oligomers
60 Monomer
Stirred unstirred 40 suspension suspension
20
Low oligomers
1 2 3 4 Time (s)
Figure 1-7. Time dependence o f the aging in stirred (full line) and unstirred (broken line) suspensions at pH 5.06 and 25 °C Low oligomers include dimers, trimers, and tetramers. High oligomers covers higher oligomers. (Modified from Figure 2 of reference [75])
2 1 100
eo
60
40
20
2 3 4 5
100 I 80 0 c 60 1 I 40
c i 20 & 0 2 3 4 5
Monomer IIII Tetramer Dimer Pentamer Trimer
Figure 1-8. Distribution of chromium in the different oligomers as a function of pH for solutions with (a) [CrJjot =0.1M, (b) [Cr],j,t =0.00 IM (Modified from Figure 1 and Figure 2 of reference [94])
2 2 Reaction Rate constant Reference
Cr^^ +Cr(OH)^‘^—> dimmers 6x10^ M 's ' [76]
Cr(OH)-^+Cr(OH)^^-> dimmers 2 x 10-^M 's' [76]
Cr(OH)^+Cr2(p-OH) 2(OH)^^^ trimers 3 xlO-^M 's ' [77]
Cr(OH )2*+Cr 2(p-OH) 2(OH)2^^^ trimers 3 M-'s"' [77]
Cr(OH)“"+Cr3(|i-OH)4(OH)^^^ tetramers 2 x ^M '‘s*' [78] Cr(OH)^^+Cr3(p-OH)4(OH)2^^—► tetramers 5 M-'s"' [78]
2 Cr2(p-OH) 2(OH)^^-> tetramers 2 x 10"‘M ''s‘ [79]
2 Cr2(p-OH) 2(OH)2'^-> tetramers 0.8 M 's ' [79]
2 Cr3(p-OH) 4 (OH)^"—> hexamers 8 X 10"^ M 's ' [78]
2 Cr3(p-OH) 4 (OH)2^^-> hexamers 5 x 10 ' M 's ' [78]
Water exchange with 2.4 X lO’^s*' [76]
Water exchange with Cr 2(p-OH) 2(OH)^^ ~ 10"^ s ' [84] (I=1.0M with L1C104 or NaC 104, pH within 0-4)
Table 1-1. Rate constants of some typical reactions involved in Cr(III) hydrolysis
Species pKa reference
Cr^^ (aq) 4.3 [87]
Cr2(p-OH) 2"^ (aq) 3.7 [87] Cr3(p-OH>4^* (aq) 4.4 [87]
Cr4(p-OH) 5 (OH)^" (aq) 3.5 [88]
Table 1-2. pKaOf some Cr"' hydrolysis species
23 Other than cross-linking with itself to form chromium hydroxide, the water ligand
in Cr(GH2)6 ^^ can exchange with other ligands to form a whole group of Cr'"-containing
complexes. Ligand-displacement reactions of Cr*“ complexes are kinetically slow, as
shown in Table 1-1, with halftimes o f more than several hours [4, 95]. Such chemical
“inertness” makes it possible to isolate some Cr'"-complexes even under conditions where they are thermodynamically unstable. The huge surface area of chromium hydroxide makes such adsorption/replacement more favorable. Chromium hydroxide can bind to various anions through Cr"-O-L chemical bonds, where L is a foreign ligand, such as S0 4 ^’ [97], C 2 0 4 ^’ and CH3COO [71]. As a matter of fact, most anions, except for
NO3 and CIO4 ’, can replace one or more water ligand from CrfOH^)^ [71,96].
Another group of important Cr chemicals are chromium'^' oxy compounds. The common forms are C 1O 3, Cr0 4 ^ -salt, and Cr20?^ -salt. In most Cr^'-oxy solids and solutions, Cr^' exists as tetrahedral C 1O 4 Chromic acid, Cr 0 3 , is a linear polymer of such tetrahedrons sharing O atoms. Dichromate and chromate are the corresponding dimer and monomer. The structures of the three chemicals are shown in Figure 1-9, with their bond angles and bond lengths marked [98]. In aqueous solution, several possible equilibria coexist, depending on pH and total [Cr^'][71]. Usually, HCr 0 4 , Cr0 4 ^, and Cr2 0 7 “' are the prominent species, and they obey equilibria as below:
HC1O 4 — Cr0 4 ^’ + H^ pKa= 5.9 (ionic strength =3)[71]
Cr20?^ + H2O ^ 2 HC1O 4 pK = 2.2 (ionic strength =3)[71 ]
The solubility of chromate salt varies, depending on the cation. Most alkaline earth chromâtes have low Ksp, such as SrCr 0 4 (Ksp = 3.0 x 10'^ ) [71].
24 0O o Crvi
CrO^^
2- C r ^ O ?
CrO 1.59 Â 1.60 Â
Figure 1-9. Structure of CrO^^, CrjOy-', and CrOj (Modified from Figure 4 of reference [98])
25 Due to its strong oxidizing ability, Cr'^‘ is commonly used in organic synthesis.
But the details of Cr'^'^—> Cr"' redox reactions are not totally clear. It is believed that intermediates such as Cr'^ and Cr'^ are involved in the reduction of Cr^' to Cr'", or the reverse reaction [71]. Both Cr'^ and Cr'^ are usually unstable, especially in aqueous environments, and can undergo further electron transfer to generate Cr'" or Cr^'. [71] In solid phase, Cr'"^Cr^' and Cr^' Cr"' can be achieved by high temperature treatment, with or without the presence of O 2 - Under controlled conditions, a group of Cr'"^ ' mixed valence compounds was generated and their structures were characterized by X-ray crystallography [99-104]. The specific structural parameters depend on the counter ions and the experimental conditions. But in general, Cr'" atoms are surrounded by six oxygen atoms in octahedrons, and Cr^' atoms exist in C 1O 4 tetrahedrons. The Cr '^ '0 4 units are cross-liked with Cr'"06 by Cr "'-O-Cr^' bridging bonds. Figure 1-10 shows the crystal structure of CrgO:, determined by using XRD[99].
26 o o
C r'" Cr'"
©
Figure 1-10. Crystal structure of Crg02, (modified from Figure 5 of reference [101])
27 1.5 Research Objective
The research results that will be presented in this dissertation mainly deals with the formation mechanism of CCC film on AA2024-T3 by using vibrational and electronic spectroscopic techniques. It is the author's motive that these results can provide some hints in finding non-toxic coating systems for metals.
Chapter 2 is dedicated to Cr-species identification, since Cr'^' in the CCC film is the center of the whole CCC-film-protecting-AJ-alloy scenario. Both Raman and Infrared spectroscopy were used to determine the generic structure of the Cr-species in CCC. The strong resemblance between CCC film and synthetic Cr'"-Cr^' was observed, and detailed spectroscopic evidences were provided for a Cr''^* -O- Cr"‘ structure.
Chapter 3 focuses on the function of ferricyanide in CCC film formation. In order to achieve a thorough understanding of the mechanism, the structure of CN species in
CCC film was studied first. The possible effect o f other additives in CCC formation was also considered, and ferricyanide was found to be the one responsible for "accelerating"
CCC film growth. Then, Raman was used to study the possible reactions between Cr'’ * and Al, between Cr'^‘ and ferrocyanide, and between A1 and ferricyanide. The in situ results support the ferricyanide mediation mechanism.
Chapters 2 and 3 are devoted to basic studies of CCC film structure and general formation mechanism. Chapter 4 is focused on applying these structures and models to answer questions firom the corrosion point of view. A Langmuir adsorption model was established to explain the Cr'^' “storage’Vrelease quantitatively. The model was based on the results from Chapter 2, and is consistent with the experimental data in general.
28 More detailed spectroscopic and theoretical studies of the Cr-species structure are discussed in Chapter 5. Both experimental data and Ab intio results support the idea that the major Cr'^' species is monmeric. Based on cturent results, hypotheses were proposed to explain the “fixation” of Cr'^' upon aging.
Some interesting experiments and the results about Cu -Al coupling are described in Chapter 6 . Although not completed, these test results could be valuable for future researchers in the same field.
29 CHAPTER 2
IDENTIFICATION OF CHROMIUM-SPECIES IN CCC ON AA2024-T3
ALUMINUM ALLOY BY VIBRATIONAL SPECTROSCOPY
2.1 Introduction
The formation mechanism of chromate-conversion-coating on AA2024-T3 alloy is a challenging topic, since both the coating solution and the alloy are very complicated.
The coating baths usually contain Cr^% and several additives are present for better formation and performance. The composition of one commercial coating solution,
Alodine 1200S, is listed in Table 2-l[107]. The AA2024-T3 aluminiun alloy studied here is currently used in building military and commercial aircrafts. The composition and microstructure of the alloy are so complicated that it can be a separate topic itself. In addition to Al, several alloying elements are added during manufacturing process. The phase precipitation during aging results in various kinds of “Cu-rich” or “Fe-rich” intermetallic inclusions [1-3,5], and these intermetallic phases make the alloy surface
30 heterogeneous. The overall composition of 2024-T3 can be measured and is listed in
Table 2-2 [108]. Since the CCC film is the product of reactions between two complex
mixtures, the alloy and the Alodine, its structure is often finstratingly complicated.
Cr'^‘ Chemicals 4 ( from fCjCrjC?) NaF K3Fe(CN)6 KBF KiZrFo
Concentration (mM) 3 8 -4 5 9-18 2-3 12-18 1 - 2
Table 2-1. Composition of Alodine 1200S coating solution [107]
Element Fe Mn Mg Cu Al
Wt% 0.29 0.62 1.4 4.4 93.29
At% 0.14 0.31 1 . 6 1.9 96.02
Table 2-2. Composition o f AA 2024-T3 Aluminum alloy [108]
31 The practical procedure for CCC film formation on Al alloy is surprisingly simple and quick. Dipping the alloy in the coating solution for only 15-30 seconds is enough to
form a golden-yellow thin film on top of the alloy surface [7]. The coating can provide soluble Cr^' to protect any locally damaged area, even when the area is not covered by
CCC film[13-15].
For many years, scientists were puzzled by the CCC’s outstanding anti-corrosion properties. But it was not imtil the last 10 to 20 years, when the EPA imposed more strict regulations on the usage and disposal of Cr'^’-containing chemicals [33], had an enormous amount of funding and human effort been involved in CCC protection mechanistic studies. Currently, the research on CCC film anti-corrosion mechanism mainly focuses on the film structure characterization and the protection mechanism studies. The clarification of its protection mechanism will provide a logical route for finding non-toxic alternatives. In order to imderstand the formation mechanism, the film structure is essential, since such structural details provide the starting point for imderstanding how the film “behaves” during corrosion protection.
As mentioned in Chapter 1, a variety of structural probes have been used on CCC.
The CCC formed on AA2024-T3 and AA7075 alloys by Alodine 1200S was found to contain Cr, of which, 23 ± 2% is Cr^' with the balance Cr’"[43]. Examination of CCC film growth on AA2024-T3 with X-ray absorption near edge structme (XANES) revealed an increase in Cr^'iCr"' ratio during early exposure to the chromating bath, reaching a constant value o f 0.25 after 5 minutes [20% of total Cr as Cr'^'j, with the Cr'" present as a hydrated chromiiun hydroxide [14]. CCC films formed on high purity aluminum were found to contain little Al'", with dominant components being chromium and oxygen [48].
32 The Cr'^' was tetrahedrally coordinated by oxygen atoms, with a Cr-O distance of 1.71 Â, and Cr‘” was octahedrally coordinated with a Cr-O bond distance of 1.99Â[43]. FTIR of
CCC films revealed M-O and O-H bonds and significant quantities of C =N arising from the Fe(CN)6^' in the treatment bath [43,57]. Raman spectroscopy also has revealed that the Cr-O stretch fi-om chromium species in CCC differed fi’om that in either Cr 0 4 ^', or
CriO?^ Zhao et al. observed that chromate species were released fi-om the CCC upon exposure to water or salt solution [15]. XANES has shown that Cr^' is lost more rapidly than Cr'" from CCCs on aluminum alloys [14].
The research work presented in this chapter was directed toward structural identification of CCC films on AA2024-T3 alloy, using primarily Raman and FTIR spectroscopy. The spectra of CCC films were compared to those of synthetic mixed oxides made from pure Al"'-, Cr'"-, and Cr'^'-containing compounds. Generic structures of the chromium species in CCC films on AA2024-T3, formed from Alodine, are proposed. The general identification of CCC film Cr-species, and some of its chemical behavior in aqueous solution, will be the focus of this chapter.
2.2 Experimental
Alodine 1200S solution was prepared by dissolving about 1.9 ± 0.1 g Alodine
1200S powder (Henkel Surface Technologies, Madison Heights, MI) in 250ml water, adjusted to pH -1.3 with concentrated HNO 3 . Water for all experiments was “Nanopure” from Bamstead, with resistivity higher than 18MQ. Aluminum alloy AA 2024-T3 made by Aluminum Company of America was obtained from Joseph T. Ryerson and Son, Inc.
1 cm xlcm AA2024-T3 coupons were moimted in epoxy (Buehler Ltd., Lake Bluff, IL)
33 and polished in water with a succession o f240,400, 600, 800 and 1200 grit sandpaper
(Buehler Ltd., Lake Bluff, IL). Polished and air dried (24h) samples were immersed in
room-temperature Alodine 1200S solution for 60 seconds, then rinsed thoroughly with
more than 100 ml running nanopure water before air drying for periods indicated below.
For electrochemical tests, some AA 2024-T3 coupons were attached to conductive wires by silver-epoxy (SPI Supplies, West Chester, PA) on the backside before mounted into epoxy.
Raman spectra were obtained with 514.5 nm excitation, 180° backscattered sampling geometry. A Chromex 250 spectrograph (Chromex, Albuquerque, NM), with
50 mm focusing and collection lens, a CdS crystal (Cleveland Crystal, Inc.) band rejection filter, and front illuminated Photometries charge coupled device (CCD) detector system was used to collect high Raman shift region (~ 500-5000 cm'*). Raman spectra within the range of 300-1600 cm * were collected by using a Holographic Kaiser spectrograph, with 50 mm focusing and collection lens, a holographic transmission grating with linear dispersion of 3.0 nm/mm, a holographic super notch plus band rejection filter (Kaiser), and fi-ont illuminated Photometries charge coupled device (CCD) detector. The schematic layouts of the two Raman systems are shown in Figure 2-1 and
Figure 2-2. Frequencies were calibrated with 4-acetamidophenol (Tylenol), but the intensity was not corrected for instrumental response.
FTIR spectra were acquired in transmission mode on a Bruker model 55 Equinox spectrometer. The instrumental setup is shown in Figure 2-3. Powder samples were spread on a glass slide, and dried in air for about 24 hours before mixing with KBr.
Freshly prepared CCC films (Alodine 1 minute) were scraped off the metal surface, and
34 Sample holder He-Ne laser for alignment
F/4.0 Spectrograph IT / Mirror
LAu» CdS crystal Cylindrical lens
Ar+ ion laser Mirror LI - Focusing lens Band pass filter L2 - Objective lens
Figure 2-1 Instrumental layout of Chromex 250 spectrometer Band pass filter for 5l4.5nin
Mirror
Ar Laser
Video camera
Mirror
Focusing Movable Objective lens mirror lens F/1.8 spectrograph
Sample stage
Figure 2-2. Instrumental layout of Kaiser spectrometer (HBR - Holographic super notch plus band rejection filter) FT-IR Macro part Source 1 "(Transmittance)
Sample holder (in a N; purged chamber)
Optics for Optics for DIGS U> best focusing best focusing Detector Mirror (condenser) Sample (pellet)
Figure 2-3. Instrumental layout of Bruker model 55 Equinox FT-IR spectrometer
(continued on next page) (continued)
Liquid N^c^en FT-IR Microscope
MCT Detector
Human eye
Mirror objective Ocular
Movable mirror Movable mirror
Aperture Beam splitter
Mirror objective Source 2 (Reflection)
Sample stage
Mirror (condenser)
Figure 2-3. Instrumental layout of Bruker model 55 Equinox FT-IR spectrometer
38 dried in air for a few hours, then the scrapings were compressed with KBr. The sample/KBr weight ratio was -1/20 for all pellets.
The Cr"'-Cr'^' mixed oxide samples were prepared by two different routes:
“simultaneous” and “sequential”. For the simultaneous method, about 4 g of
CrfNOsl OHzO and 0.5 g of KiCrzO? were dissolved in 50 ml of Nanopure water. 1.0 M
NaOH was added dropwise, while the solution was agitated by a magnetic stirrer. The pH was monitored during NaOH addition, and precipitate started to form at approximately pH 4. The brown precipitate was collected by suction filtration, rinsed with more than
50ml of Nanopure water, and air-dried. For the sequential method, about 4 g of
CrfNGs) 9 H2O was dissolved in 25 ml water and 1.0 M NaOH was added until the pH reached 4 - 5 . For easier filtration or higher yield, higher NaOH concentration (such as 5
M), higher collection pH (such as 5-6), and longer equilibration time were necessary. The green Cr'"- hydroxide was collected by suction filtration, then a portion of the solid was immersed in 50 ml of 0.03 M K2Cr20? for 30 minutes. As the solution was agitated with a spatula, the green solid turned brown. At this point, the solid was collected, washed, and air-dried.
The Al'"-Cr^' mixed oxide was prepared from AlfNOsla XH 2O (x -9) and
K2Cr20?, with an initial molar ratio of Al:Cr being approximately 3:1. The simultaneous method described above was used, after substitution of CrfNO])] «9H20 with AlfNOs)]
•xHiO (X -9).
Quantitative analysis o f Cr'" and Cr'^‘ in powders and films was performed using
UV/VIS absorption spectroscopy in basic solution with a Perkin Elmer Lambda 20 spectrometer. The schematic layout of the spectrometer is shown in Figure 2-4.
39 Halogen M1, M4 and M5 = Plane Mirror lamp M2 = Torroidal Mirror M3 = Spherical Mirror M2
Ml Reference Detector M5 Filter wheel Deuterium o lens lamp M3 k Slit 1 Beam Splitter
Slit 2
M4 Grating Sample lens Detector
Figure 2-4. Instrumental layout of Lambda 20 UV/VIS/NIR spectrometer Calibration curves for Cr 0 4 ^' (^max=373 nm) and Cr^" (>-max=588 nm), as shown in Figure
2-5, were constructed from CrCNOs) 9 H2O and K2C1O 4 solutions in 0.25 M NaOH. The
UV-Vis spectra of CrCNOs) 9 H2O and K2Cr0 4 in 0.25 M NaOH are shown in Figure 2-6.
Separate tests on Cr"'-Cr^' mixture solution indicate that there is no significant
interaction between Cr"' and Cr'^' imder such basic conditions (see Table 2-3). So [Cr"'] and [Cr^'] can be measured simultaneously. As shown in Table 2-4, the interference of
[Cr"'] from Cr^' is negligible, since is O.OCX). Although the overlapping at
339~373nm region is significant, the maximum error from such overlapping is less than
1% for [Cr'"], when [Cr'"]:[Cr'"] =1:3.
Known concentrations and ratios measured results
[Cr'"] (mM) [Cr'"'] (mM) Cr'":Cr'" [Cr'"] (mM) [Cr'"] (mM) Cr'":Cr'"
64.05 19.27 3.324 62.99 19.27 3.269 32.89 2.753 11.95 35.88 2.821 12.54
Table 2-3. Simultaneous measurement of [Cr'"] and [Cr'^'] of known solutions
EfSSnm E373nm E339nm Species (M'.cm ') (M".cm') (M '.c m ')
Cr'" 24.25 16.00 4.35 Cr'" 0 . 0 0 0 4884 1479
Table 2-4. Molar absorptivity of Cr'" and Cr'^' in 0.25M NaOH
41 3 '373nm = 4884 mol'«L*cm '
s
1
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 [CfV'] (mM)
0.3 S88nm = 24.25 moi‘'»L»cm‘*
0.2
E I <
0.0
24 6 80 10 12 [Cr"'] (mM)
Figure 2-5. Calibration curves of CrO^^ (top) and Cr"' ( bottom)
42 373 nm 0.024 mM Crvi .12 2.1mM Cr'" .10
.08 588 nm < .06 w .04
.02
300 400 500 600 700 800 X (nm)
Figure 2-6. UV-VIS spectra of Cr'" and Cr^' in 0.25M NaOH solution An air-dried (Ih) CCC film was dissolved directly in 25.00 ml of 1 M NaOH and
diluted to 100.00 ml before spectrophotometric analysis. The Cr and Al-Cr mixed oxides
were first dissolved in a small amount of concentrated HNO 3 before adjusting to pH 13
with NaOH, to a total volume of 50.00 ml. In cases of very diluted Cr'" solution, the
sensitivity of A sssnm is not high enough for precise analysis. Therefore, excess amount of
H2O2 was added into the Cr'" solution after adjusting solution pH to basic, and the solution was allowed to sit for about 24 hours before spectrophotometric measurement at
373nm. A separate test was conducted to confirm the total oxidation by H 2O2 : 4 ml of
0.15M H2O2 was added into 1.00 ml of 1.67 mM Cr'" solution, which was changed to basic beforehand. The mixture was then diluted to 50.00 ml by 0.25 M NaOH, and allowed to stand for about 0.5-1 hour at room temperature. The resulting [Cr^'] was measured at 373nm, and about 95% of the Cr'" had been converted to Cr'^'.
“Migration” experiments were conducted in the way described by Zhao et al. [15].
A polished sample of AA 2024-T3 was "sandwiched" with a CCC coated AA 2024-T3 sample, or -0.1 g Cr'"-Cr'^' mixed oxide, or - O.lg Al'"-Cr'^' mixed oxide for 24 hours.
The setups are shown in Figure 2-7. The coated and the bare AA 2024-T3 aluminum alloy were separated by an O-ring, which held about 0.5 ml of 0.1 M NaCl solution between the two surfaces [15]. Mixed-oxide samples were wrapped in a filter paper bag, and hung in the solution by a piece of thread. About 50 ml of O.IM NaCl solution was used, and there was no direct contact between the mixed-oxides and bare AA2024-T3 alloy. RpOf the samples were measured in O.IM NaCl solution by using Gamry software, with Ag/AgCl reference electrode and Pt wire auxiliary electrode.
44 Plain polished AA2024-T3
0.5 ml O.IM NaCl solution
CCC covered AA2024-T3
O-rings
B Plain polished AA2024-T3
50ml O.IM NaCl Filter-paper wrapped C r* "-Cr'^’-mixed-oxide or Al"*-Cr'^-mixed-oxide
Figure 2-7. Experimental setup of “migration” experiments (A). “Sandwich” setup (B). “Migration” test by using mixed-oxides and AA2024-T3
45 2.3 Experimental Results
Since the composition of the solids formed by reacting Al‘“ and Cr'" with Cr' ' is not well defined at this point, they are referred to as “mixed-oxides”. Thus, the Cr'" +
Cr'' ' product is “Cr'"-Cr^' mixed oxide” and the Al'" + Cr^' product is “Al'"-Cr^' mixed oxide”. Inductively Coupled Plasma (TCP) analysis of the Al'"-Cr^' mixed oxide, synthesized under the above mentioned conditions, yielded a 4.0/1.0 molar ratio of the Al to Cr. ICP analysis of scrapings fi'om a CCC on AA2024-T3 showed a much lower aluminum content, with Al/Cr molar ratio o f about 0.01. Table 2-5 summarizes the ICP results, along with the iron content in the CCC.
The Cr'":Cr'^' ratio in the CCC and the mixed oxides was assessed by dissolving the solid and performing UV-VIS spectroscopy. The concentrations of Cr'" and Cr^' can be obtained simultaneously by using Agggnm and A 373nm- The resulting Cr'":Cr^' ratio of
CCC film and mixed-oxide are listed in Table 2-5.
The freshly prepared CCC was dissolved by immersion in 1M NaOH for 10-60 seconds. The mixed-oxides were dissolved in a small amount of HNO 3 (to aid dissolution). All three samples were adjusted to pH 13 and diluted to a final volume of
50.00 ml. Neither the dissolution time nor the use of HNO 3 affected the observed
Cr'":Cr^' ratio. As indicated in Table 2-5, the CCC and Cr mixed oxides both yield a 3:1 ratio of Cr'" to Cr^' upon dissolution, while the Al'"-Cr^' mixed oxide yielded no Cr'".
The results for the CCC film are consistent with previous XPS and XANES results showing 20-25% Cr^' content [14]. In a separate series of experiments, the CCC and air- dried Cr'"-Cr'^' mixed oxide were immersed in pH 4.02 HNO 3 for 30 minutes, then the pH was measured before being adjusted to pH 13 to assess the Cr'" and Cr^'
46 From ICP Concentration of elements (mg/g) Al’’’:Cr'" molar ratio Cr Al Fe
CCC 262.8 1.44 28.5 0 . 0 1 0 Al"'-Cr'^’ 99 207 - 4.0 mixed oxide
From UV-Vis. Weight [Cr’"] [Cr'"’] Cr"’:Cr'" (mg)' (mM) (mM) molar ratio CCC - 0.346 0.116 3.0±0.6“ Cr mixed oxide 5.0 0.791 0.267 3.0±0.4= Al"‘-Cr^‘ 8 . 6 0 . 0 0 0 0.408 0 . 0 0 mixed oxide i t T r - J* J : Standard deviation of three replicate samples. Standard deviation of four replicate samples.
Table 2-5. Composition of CCC and mixed-oxides
concentrations. 5.00 ml aliquots were diluted to 25.00 ml by 1 M NaOH. Only UV-VIS peaks from Cr^' were observed. The possible [Cr'"] in the solution was measured by adding extra amount of H 2O2 to another 5.00 ml aliquot after changing pH to basic. The solution was allowed to sit for 24 hours before spectrophotometric measurement. The
Cr"’ concentration in solution was negligible, and the molar ratio of released [H'] to released [Cr'^'j was approximately 0.97-1.07.
Powder X-ray diffraction of the mixed oxides yielded no observable peaks, indicating amorphous structmes. The CCC and mixed oxides were exposed to solutions of varying pH to visually monitor stability. The CCC and mixed oxides remained yellow in pH range 3.8 to -9, but all three decomposed at higher pH.
47 FTIR spectra of the CCC and mixed oxides are shown in Figure 2-8, for samples mixed with KBr and pressed into pellets. Reference spectra of KzCrzO?, K 2C 1O 4 ,
Cr(OH>3, A 1(0 H)3, and K3Fe(CN)6 are shown in Figure 2-9. Assignments for observed peaks are listed in Table 2-6. The CCC film shows IR features characteristic of Cr mixed oxides, as well as a 2083 cm ' band for C=N, presumably derived from the Fe(CN)&^' in the Alodine coating solution. (More detailed studies related to the CN structure and function in formation will be discussed in Chapter 3.) Although Cr'^'-O vibrations are apparent in all three samples shown in Figure 2-8, the Cr^' frequencies are shifted from those in pure KiCr 0 4 , or KzCrzO?. The absence of CN in the Raman spectra of mixed- oxides is as expected, since no CN-containing chemical was involved in syntheses.
IR peak frequencies (cm' -) I \a CCC Tentative assignment C r " W Al"'-Cr'" 3600-3000 3600-3000 3600-3000 Anti-symmetric and symmetric OH stretching
2154(sh).2083 - - C=N stretching 1621 1621 1636 HOH bending of lattice water
1400,1384 1384 I384(vw) NO3 960(sh),919,817 961(sh),919,817 924(br),797 Cr(Vl)-0 vibration
592 - Fe-C vibration in Fe-CN
- 559 Al"‘-OH vibration 526,5 lO(sh) 492 Cr'"-OH vibration a. s=strong, br=broad, sh=shouider, m=moderate b. All samples were air-dried for about two months hours before spectroscopic measurement.
Table 2-6. FTIR of CCC, Cr'"-Cr^' mixed oxide, and Af'-Cr""III f^VI mixed oxide
48 w 00 K) CCC film scrapings
% Ï
W 00 S O Cr"'-Cr^'-mixed-oxide V / 00 vo « è p ^ s n >T=
AI"'-Cr^'-mixed-oxide
0 0 5 0 30 5 00 1000 1500 2000 2500 3000 3500500 Wavenumber (cm')
Figure 2-8. IR spectra of CCC, Cr'"-Cr''' mixed oxide, and Al'"-Cr''‘ mixed oxide oo KjCrjOy LA LA V HI l a O n ' w LA
to KjCrO^ to Ui
Cr(OH), L^ 00 Ul 00 o AI(OH), L^w
OK) K3Fe(CN),
500 1000 1500 2000 2500 3000 3500 4000 Wavenumber (cm ')
Figure 2-9. IR spectra of K2C1O4, K2Cr20y, Cr(OH)j, and KjFe(CN)(, solid Raman spectra of the CCC and the mixed oxides are shown in Figure 2-10, and the Raman spectra of several Cr'^'-containing standard chemicals are shown in Figure 2-
11. The tentative assignments of the Raman signals in Figure 2-10 are listed in Table 2-7, and those frequencies in Figure 2-11 are listed in Table 2-8. For each spectrum shown in
Figure 2-10, two spectra of the same sample were taken by the two spectrometers, Kaiser and Green-Chromex, and the two segments were normalized to the same scale by using the overlapping peaks, and then combined by using “splice.ab” in Grams/32 software, which automatically offset the difference in linear background. An example of the combination is shown in Figure 2-12. The Kaiser spectrometer was used to obtain all the other Raman spectra shown in this chapter. Since the spectra shown in Figure 2-10 are the combination of two segments, and the two segmental spectra were not individually intensity-corrected before combining, the intensity of the peaks shown in this figure may not reveal the bands’ intrinsic cross section or relative amount.
As shown in Figure 2-13, the spectrum of the CCC changed very little from that of Figure 2-10 over the pH range 3.8 to ~9, but the 858cm'' band disappeared at around pH 9, as did the golden-yellow color. The Raman spectra of the chromium mixed oxides, shown in Figure 2-14, behaved qualitatively similar as the pH increased, with normally prominent 858-859 cm ' feature absent above pH - 9.0. However, the peak frequency of the -860 cm ' Raman band decreased from 870 to 845 cm ' for the Al'"-Cr^' mixed oxide as the pH increased from 4.0 to 9.2. The Raman spectra of these Al'"-Cr^' mixed oxides are shown in Figure 2-15. Over the same pH range, the peak frequencies for the CCC and the Cr mixed oxide remained constant at 858-859 cm '. The pH results are listed in Table
2-9, and indicate the similarity between the CCC and the chromium mixed oxides.
51 Raman shift (cm ')
CCC Cr"'-Cr'" Al'”-Cr'^' Alodine solution assignment OH stretch of 3600-3000(br) 3600-3000(br) 3600-3000(br) 3600-3000(s,br) H 2O 2097, 2144 2134 C=N in Fe-CN 1709 (w,br) 1708(br) 1648 (br) HOH bending
1050 NO3 858 (s) 859 (s), 379, 903, 369 (m) 944(sh), 906(s), 454(m) 373(m) Cr'^'-O a. All samples were air-dried for about two months before spectroscopic measurement. b. s=strong, br=broad, sh=shoulder, m=moderate
Table 2-7. Raman frequencies of CCC, mixed oxides, and Alodine solution
compounds Raman Shift (cm ') “ Assignment
KzCriO? 911 (s),945(m ) ^
K2Cr0 4 (solid) 853(s), 869(m), 880(m), 906(m) KiCr.O? (aq) " 901(s), 944(sh) Cr'^'-O stretching K.CrOj (aq) 843(s), 885(sh)
C1O 3 (solid) 978(s), 1003(m) CrzO] (solid) 550 (w) Cr"'-0
Cr(GH)3 (solid) 528-530 (w) Cr'"-OH a. The frequencies listed here are corresponding to stretching modes only, frequencies from other vibrational modes will be discussed in Chapter 5. b. The concentrations were around 40-60mM.
Table 2-8. Raman Shift of Cr-O stretching fi’equencies from several standard compounds
52 n z X n § Z - to 2 w 00 y ; : /
00 Alodine solution Vl 00 to to O n vo® f. T" 00 ’ CCC film on AA2024-T3
WLA Cr"'-Cr''‘ mixed-oxide
Al'"-Cr'^‘ mixed-oxide
500 1000 1500 2000 2500 3000 3500 4000 4500
Raman Shift (cm ')
Figure 2-10. Raman spectra of Alodine, CCC, Cr'"-Cr^' mixed oxide, and Al'"-Cr^' mixed oxide _L KjCrjO, (solid) JL K2Cr0 4 (solid)
A _JL CrO; (solid)
JL KjCrjO, (aq)
K2 C1O 4 (aq)
Cr2 0 , (solid)
Cr(OH)3 (solid)
CCC (film) "400 600 800 iooo 1200 MOO
Raman shift (cm ')
Figure 2-11. Raman spectra of several Cr-containing compounds and CCC film i .1.
Normalize the two spectra by i the 902cm ’ band intensity
b X 4
i Offset the two spectra’s background
b X 4
Average the overlapping region
a + bx4
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Raman Shift (cm ')
Figure 2-12 The algorithm of “splice.ab” a. Spectrum from Kaiser spectrometer; b. Spectrum from Chromex spectrometer
55 2
pH=1.0
pH=1.5
pH=2.0
pH=2.5
pH=5.0
pH=7.0
pH=9.0
pH=I I
400 600 800 1000 1200 1400 1600 Raman Shift (cm'*)
Figure 2-13 Raman spectra of CCC film on AA2024-T3 at various pH (• stray light)
56 I
pH=3.80
pH=4.64
pH=4.76
=6.51
pH=8.27
400 600 800 1000 1200 1400 1600 Raman Shift (cm ')
Figure 2-14 Raman spectra of Cr'"-Cr^' mixed-oxide at various pH
57 i pH=3.9
pH=4.0
pH=4.6
^ pH=5.1
pH=5.5
pH=6.3
- pH=6.7
pH=6.9
pH=8.1
pH=9.2
pH =10
—I------1------1------1------1------1------1— 400 600 800 1000 1200 1400 1600 Raman Shift (cm*‘)
Figure 2-15 Raman spectra of Al"'-Cr^' at various pH
58 CCC film Cr'-'-Cr"" Al'-'-Cr'"
pH Raman shift pH Raman shift pH Raman shift (c m ‘) (cm ') (cm ‘)
3.0 858 (s) 3.8 859 (s) 3.9 867 (s), 364 (w) 4.0 870 (s), 361 (w) 4.6 859 (s) 4.6 870 (s), 361 (w) 5.0 858 (s) 5.1 874 (s), 365 (w) 5.5 864 (s), 357 (w) 6.5 859 (s) 6.3 853 (s), 354 (w) 6.7 850 (s), 353 (w) 7.0 858 (s) 6.9 851 (s), 354 (w) 8.3 859 (s) 8.1 845 (s), 345 (w) 8.8 859 (s) 9.2 845 (s), 345 (w)
Table 2-9. Raman peak frequencies of fresh film or solid as a function of pH
The effect of heating in air on the CCC and the mixed oxide powders is shown in
Figure 2-16 and Table 2-10. Each sample was heated for Ih at 50, 100, 150°C, etc., and
Raman spectra were acquired between successive heating periods. The CCC spectra showed little change up to 200°C, with the 859 cm * band unaltered in position or bandwidth, but minor shape changes. Above 200°C, the band shape change becomes significant, and the 551 cm ' peak characteristic of CrzO] appeared. Cr mixed oxide
59 behaved similarly, with clear decomposition above 200°C. The Al“'-Cr'^‘ mixed oxide showed a substantial shift in the 858cm'^ band with increasing temperature, up to 918cm ' at 254°C. The spectral and structural variations of both CCC film and Cr-mixed-oxide will be discussed in further details in Chapter 5.
Treatment Raman shift of Cr^'- ■O(cm') Raman shift of Cr '-O (cm ') Criii/vi CCC Cr'“-Cr''' Al"'.CrV CCC
Fresh/wet 859 859 864,351 Air-dried for 859 859 894, 364 3-4 days
Air-dried for 859 859 904, 370 2-3 months
50°C/lh 859 859 IOO°C/lh 859 859 150°C/lh 859 859 200°C/lh 859 859, 1004(w, hr) 545
250°C/lh 859,1003(br) 859(w,br), 918,378 551 542 1004(w, hr)
300°C/lh 859,1003(br) 551 542
350"C/lh 730(br), 861, 551 551 1003(br)
Table 2-10. Frequencies of Cr'^'-O vibration upon heating
60 00 ^ 00 Ui o LA 00 00 u> ^ Air-dried/lday , ...A ~ ~ '——-
50 “C/lh
100 "C/1 h
o\ ^ 200"C/lh
250«C/lh
300 "C/1 h
400 800 1200 1600 400 800 1200 1600 Raman Shift (cm ') Raman Shift (cm ')
Figure 2-16. Raman of heated (A) CCC, (B) Cr'"-Cr^' mixed oxide, and (C) Al'"-Cr^' mixed oxide (* stray light) (continued on next page) (continued)
00 2^
LAW A Fresh 00 2 W S Air-dried/3-4days S I W air-dried/2-3 months sO
heating 245“C/lh 400 800 1200 1600 Raman Shift (cm ')
Figure 2-16 (continued) (C) Raman of heated Al'"-Cr^' mixed oxide Since the Raman scattering from water is quite weak, it is possible to monitor
Raman spectra o f aqueous solutions in situ, while reactions occur. About 1 -2 g of
freshly synthesized Cr‘"-hydroxide was spread on a glass slide, to a depth of approximately 1-2 mm. A Raman spectrum of the Cr""-hydroxide was collected first, then
-0.1 ml of various Cr^'-solutions (concentration 40 - 50 mM) were added, while one spectrum of the surface was collected every minute, with an exposure time of 5 seconds.
As shown in Figure 2-17, Cr*“-hydroxide shows a characteristic band at 526-528cm‘', along with a residual nitrate band at 1047-1049cm’’. When Alodine solution was added, a new band appeared at 858cm ', similar to that observed in the CCC. The sharp narrow
Raman band at 902 cm ' is from solution Cr^'. The addition of either KiCrO^ (pH - 7) or
KiCriO? (pH - 4) solution to fresh Cr'“-hydroxide precipitate had a similar effect, as shown in Figure 2-18, producing the same 858-859 cm ' band. Since the absorbance of the solution and the solids can strongly affect the sampling depth, the relative intensities of the 1049, 859, and 528 cm ' bands are only semi-quantitatively meaningful. However, it is clear that the 859cm ' band may be formed from Cr'"-hydroxide and Cr^', without aluminum present as Al° or Al^^.
63 ■ - 600 «
■ 400 £
- 200
IOOO
Raman shift (cm ‘)
Figure 2-17. In situ Raman spectra of Cr'"-hydroxide covered by Alodine solution B 2 00 K) On Cr(OH), in water
lOmin after adding Cr VI
C/lON
20min after adding Cr'"
—I------1------1------1------1------1------1— —I------1------1------1------1------1------1— 400 600 800 1000 1200 1400 1600 400 600 800 1000 1200 1400 1600 Raman shift (cm ') Raman shift (cm ')
Figure 2-18. In siiu Raman spectra of Cr'"-hydroxide covered by (A) dichromatc solution or (B) chroniate solution 2.4 Discussion
The low Al concentration in the CCC reported previously, and confirmed here
(Table 2-5), indicates that the Al"'-Cr^' mixed oxide is not a major component of the
CCC. In fact, Ramsey et al. have shown that the Al'"-Cr^' mixed oxide is structurally
similar to the corrosion products inside a pit [59]. Furthermore, the shift in Raman peak
frequencies observed for Al^'-Cr'^' mixed oxide v/ith both pH and temperature differ from
those for the CCC, further confirming the absence of this mixed oxide in the CCC in
observable amounts. In contrast, the Cr mixed oxide is very similar to the CCC in
Cr'":Cr^ ' ratio, and shows trends similar to the CCC in Raman frequencies with pH and
temperature variations. Based on Cr oxidation state and Raman spectral behavior, the
CCC contains a material very similar to the Cr mixed oxide made from Cr'" nitrate and
Cr^' salts.
The differences between the FTIR spectra of the CCC and the Cr mixed oxide are attributable to the incorporation of FefCNje^* in some form. The ICP results and the
2083cm ' FTIR band for the CCC indicate significant levels of Fe and CN, with the ICP implying a Fe:Cr molar ratio of about 1:10. The Fe and CN present in the CCC did not appear to perturb its spectroscopic behavior or pH and temperature stability, compared to the Cr mixed oxide. More details about the structure and function of Fe-CN species will be discussed in Chapter 3.
Some additional clues about the Cr-species structure are available from CCC and mixed oxide chemistry. The release of Cr'^' into water from the CCC has been reported[15], and observed here for both the CCC and the Cr mixed oxide. Furthermore, the release of Cr'^' is accompanied by [H^ increase for the air-dried Cr mixed oxide.
66 Once Cr^' was released into water, it can be Cr^O?^, HC 1O 4 , or Cr 0 4 ^', depending on pH and concentration, but interestingly, the observed [H^:[Cr'^‘] ratio was close to 1.0 (0.98-
1.07). Finally, spectroscopically similar products resulted when Cr(OH 2)6^" and CriO?' were treated with base in the same solution (“simultaneous” synthesis), or when Cr'“- hydroxide was formed first, then Cr^' was added (“sequential” synthesis). A working hypothesis, which incorporates these findings, is given by Equation 2-1.
H2O + [Cr'"3(0Cr'''03H)(0H)x(0H2)y]^**^
^ + HCr0 4 + [Cr'“3(OH)x+i(OH 2)y]^®”‘^ Equation 2-1
The nature of the bonding in the Cr mixed oxide on the left side of the Equation 2-1 will be the focus o f Chapter 5, and the form of Cr^' in solution depends on the concentration and pH. The 3:1 ratio of Cr'":Cr^' content and the release of H^ are supported by current observations. Furthermore, there is general agreement that Cr"' hydroxide is built up by
CrOô octahedral imits, while Cr^' usually exists as Cr 0 4 tetrahedral clusters[4,73]. Both the forward and reverse reactions of Equation 2-1 are demonstrated by the current results, in which the mixed oxide may release Cr'^' and Cr‘"-hydroxide may absorb Cr^'. The 3:1
Cr"':Cr^' ratio observed in the CCC and Cr mixed oxide may represent saturation of chromium hydroxide with Cr^'. A lower Cr^' loading is certainly possible. The Cr- mixed-oxides at various Cr^' loadings were studied and will be discussed in Chapter 5.
Two prominent possibilities exist for the nature of the binding between a Cr'" hydroxide and a Cr^' species. The “sequential” and “simultaneous” synthesis results imply that Cr'" chemically binds to an existing Cr‘"-hydroxide species. Polymeric Cr'" hydroxides are well known, consisting of Cr'" octahedrals bridged by OH or H 2O groups
[75-99]. Each Cr'" center is surrounded by H 2O and OH, completing the octahedra
67 coordination. The OHzHzO ratio and the size of the oligomers increase with solution pH
[95-96]. For example, a Cr’" solution at a pH of 5.0 has a significant concentration of
[Cr4(OH)6(OH2)io]^^[94]. The interaction of Cr 0 4 ^', CrzO? or HC 1O 4 with a polymeric
Cr'"- hydroxide could be electrostatic or covalent [109]. In the electrostatic model, positive charge on the Cr’" hydroxide matrix attracts negatively charged Cr'^’ anions.
Such an interaction is expected to be quite pH sensitive, as pH will affect the surface charge of the polymeric hydroxide. Changes in pH or hydration shift the Cr 0 4 ^ /HCr0 4 /
CriO?"’ equilibrium for the adsorbed Cr^’, causing the shift in the overall Cr^’-O Raman band.
The CCC and Cr mixed oxide show different behavior from electrostatic binding.
Their Raman spectra vary little with pH in the range of 4-8, and there is a stoichiometric relationship between released H^ and released chromate upon the reaction of Cr mixed oxide with water. An alternative to electrostatic interaction is the formation of Cr ”’-0-
Cr^ ’ linkages by substitution of OH or H 2O on the Cr’" hydroxide polymer with some
Cr'^’-oxy clusters. Several precedents for such bonding exist in the crystal structure of Cr- mixed-valence compounds [101-106]. In these compounds, Cr’" octehedra and Cr^’ tetrahedra share one or more bridging oxygen atoms. The Cr’’’-0 bond length (1.93 -
2.01 Â) and Cr^’-O bond length (1.55 - 1.79 Â) observed in such crystalline compounds[ 101-106] are similar to those observed in the CCC (1.99 and 1.71 Â )[43].
The specifics of these structures are slightly different: some have OH as p-ligand between two Cr’’’ centers[101], some have mono-dentate Cr 0 4 [101-103,106], bi-dentate
Cr0 4 ligands [106], even dichromate and tetra-chromate ligands[101]. An example of these compoimds is shown in Figure 1-10. As a possibility, consider the structure of
68 Figure 2-19. The reaction of a polymeric Cr*” hydroxide with soluble Cr^' might yield the lower structure of Figure 2-19, reaction of this Cr "-O-Cr^^ mixed oxide with one water molecule could hydrolyze the Cr‘“-0-Cr'^' bonds, yielding fC and Cr^' in the mole ratio of 1:1, and leaving one additional OH on the Cr"" hydroxide polymer. Therefore, the hypothetical structures of Figure 2-19 are consistent with Equation 2-1, and the experimentally observed stoichiometry. The OH/H 2O content in the Cr mixed oxide will depend on pH. The Cr may exist as HC 1O 4 , Cr0 4 ^‘, or CrzO?", depending on formation conditions and follow-up treatment. So the structure in Figure 2-14 is generic. However, the existence of Cr"'-0-Cr'^* bonds is consistent with experimental observations, and has several precedents in other well-characterized materials. More in-depth studies of the
Cr^' structure issue will be discussed in Chapter 5.
The synthesis of the Cr mixed oxide establishes that a material very similar to the major components of a CCC film may be formed without aluminum. It is generally accepted that chromate species in chromating baths such as Alodine are reduced by A1 metal after the aluminum oxide (passive film) is destabilized by fluoride ion in the bath
[110]. The results showm in Figure 2-17, 2-18 indicate that, under Alodine conditions, the
Cr mixed oxide forms spontaneously once Cr'" hydroxide is exposed to Cr' ' solution. So exposure of active A1 to chromate coating bath first reduces chromate to Cr'" hydroxide, then chromate binds to the insoluble Cr"' hydroxide to form a mixed oxide similar to that in Figure 2-19. Further studies, shown in following chapters and by other studies done by other researchers[59], indicate the special condition of CCC formation determines the dominance of Cr mixed oxide.
69 H H
À
OH
+ HjO J
Figure 2-19. Proposed structure of Cr''*-species in CCC film
70 It should be emphasized that the structure in Figure 2-19 is not the only possibility
for a mixed oxide that is consistent with the observations, but any such oxide must
exhibit a Cr'"-0-Cr^' bond, and reversible Cr'"' binding. The reversibility of the binding
is important to both the formation of the CCC, and to subsequent release of Cr'
Quantitative details of this binding will be addressed in Chapter 4.
The importance of the Cr-mixed-oxide to the anticorrosion properties of CCC's
lies in the “self-healing” process. Kendig et al.[14] concluded that the CCC provides a
“timed-release source” of hexavalent chromium for repairing defect sites. The Cr mixed oxide is a reversible storage mechanism for Cr'^’, capable of releasing Cr'^' into water or salt solution. The results of Zhao et al. demonstrated that CCC can release Cr'^', which can then protect an initially untreated alloy surface[15]. Even low concentrations of Cr' '
(-10“* M) can concentrate in pits and preventing pit growth or passivating intermetallics[15]. The Cr mixed oxide described here can release chromate via Equation
2-1, to provide the low concentration of Cr'^‘ required for self-healing. The apparent stability of the mixed oxide implies that such release may occur long after the initial formation of the CCC.
71 2.5 Conclusions
The IR spectra and Raman spectra indicate that a Cr‘”-Cr'^' mixed oxide is a significant component of the chromate conversion coating on AA2024-T3 surface. This mixed oxide may be formed by the reduction of dichromate by aluminum during CCC formation. The CCC or Cr mixed oxide can act as a repository for Cr^' species, which release soluble Cr^' upon exposure to water. The spectroscopic results are consistent with the reversible formation of a covalent Cr^'-O-Cr'^' bond, which binds the Cr^' species to the insoluble Cr"' hydroxide. These results support the previous proposal that CCC films undergo self-healing by releasing Cr^' from the film, which then stop pitting[15]. A related Al'"-0-Cr^' mixed oxide is not a major component of the CCC, but may form in a reaction between Cr^' and aluminum hydroxide.
72 CHAPTER 3
STRUCTURE AND FUNCTION OF FERRIC Y ANIDE IN THE FORMATION OF
CCC ON AA 2024-T3 ALUMINUM ALLOY
3.1 Introduction
The previous chapter described the structure of Cr'"-Cr'^' species in CCC film.
Although the CCC film is the product of Cr^' reduction on a surface made of more than
90% Al, the CCC film does not contain much A1 [14,43,48]. XPS results indicate the existence of Fe, C, and N in the film [37, 41, 108]. Both Lytle’s FTIR spectra [43] and the IR-Raman spectra shown in previous chapter indicate signals from certain type of cyano-species. In addition, a few reports about the function of each additive in Alodine solution deserve attention. The fairly high concentration of fluoride is believed to be crucial for the removal of passive AlO* film, thus maintaining an active surface[42,47-
48,50] The Al passive film can be removed by HF, and the highly stable AlFg^' complex makes such a reaction thermodynamically favorable. However, NaF alone can not lead to a satisfactory CCC film formation and another additive, Fe(CN)6^', has to be used.
73 Physical tests proved that Fe(CN)6^' actually results in faster CCC film formation [111],
hence the name “accelerator”. However, the specific nature of ferricyanide species in the
CCC and the mechanism of “acceleration” were not clearly understood in the past.
There were several hypotheses attempting to explain the Fe(CN)6^ -acceleration
mechanism, mostly based on XPS results[37,41,108]. Although XPS is a powerful tool in
determining elemental identity and oxidation states, it can not provide enough structural
sensitivity in this Fe(CN)6^ -acceleration mechanism study. One obstacle is the reduction
effect on Fe'" by high energy X-rays, which makes the oxidation state of Fe uncertain.
Furthermore, the binding energy of either Fe'" or Fe" does not strongly depend on the
identity of the first shell ligands [63], therefore not enough structural information can be
obtained by XPS.
However, structural sensitivity is one of the biggest advantages of vibrational
spectroscopy. In the case of transitional metal cyano-compounds, the CN stretch is both
IR and Raman active, and its fi-equency (vcn) is quite sensitive to local bonding status
and counter ion. For example, vcn of Cr 4 [Fe(CN)6]3 and K4 Fe(CN)6 differ by 54cm '.
Furthermore, the comprehensive studies done by several generations of inorganic
chemists have provided a “database” for structural characterization [70]. Sophisticated
theoretical models are available to predict the CN vibration fi-equency for certain rare or
unstable cyano-compounds [70, 112]. In addition, FTIR and Raman are nondestructive, and under some circumstances in situ Raman spectroscopy can provide structural
information dynamically, during reactions.
74 In the case of Fe(CN)6^'in Alodine solution, structural sensitivity is essential, because of the possible complicated chemical reactions and the various products with quite similar structures. Fe(CN)6^ is unstable in a strongly acidic aqueous solution.
Decomposition of Fe(CN)6^ generates HCN and aquocyanoferrate ions, as shown in
Figure 3-l[l 13]. Further degradation can result in a whole series of aquocyano ferrate ions, including aquo-Fe^^[l 13-116]. The cross-reactions between various decomposition products produce Fe 2 '"(CN),o% Fe"Fe“'(CN)io^’, Berlin green, and Prussian blue [113].
Both Berlin green and Prussian blue are polymers built up by Fe and CN units, and their structures are schematically shown in Figure 3-2 [117]. Note that all the CN ligands are bonded on both ends. In Prussian blue, the “carbon” end always binds directly to Fe” and the “nitrogen” end binds to Fe'”.
3.2 Experimental
Alodine coating solution and 2024-T3 aluminum alloy coupons (Icmxlcm) or electrodes were prepared in the manner described in Chapter 2. Pure aluminum coupons/electrodes were prepared in the same way as AA2024-T3 alloy samples.
All the test solutions for growth curve studies were prepared from nanopure water and analytical grade KzCrzO?, K 3 pe(CN) 6 , NaCN, KBF4, KzZrF^, and HCIO4 with the final concentrations listed in Table 3-1. Three AA 2024-T3 electrodes, or pure Aluminum samples, were coated in each test solution for 1,3,5 minutes, then washed and dried in air before any further measurement. The voliune of test solutions used for treatment was
-50 ml, and a given test solution was used for all three samples.
75 Pe"'(CN)6-^- —> Berlin-Green
hv Fe"'(CNV' ^ Fe'"(CN)5(0H/- Fe"'(0HJ3+
i (hv) hv hv Fe"(CN)5(OH,)3------> •••------> Fe"(0H.)2+
Prussian Blue F e'"(C N ))-
Figure 3-1. Possible degradation of Fe'"(CN)(,^’ in acidic conditions
(Modified from Scheme 1 of Reference [113]) CN O Fe" • Fe"i A A / I
o x i d i z e
•-4 r
Berlin Green Prussian Blue
Figure 3-2. Structure of Berlin Green and Prussian Blue. For clarity, sonic CN ligands arc shown as lines, (Modified from Figure 22.5 of [117)) Solution [Cr'^'l [NaF] [NaCN] [K3Fe(CN)6] [KBF4 ] [K2ZrF6] pH' (mM) (mM) (mM) (mM) (mM) (mM) A 42.2 14.4 0 0 0 0 1.3-1.6 B 42.2 14.4 3 0 0 0 1.3-1.6 C 42.2 14.4 0 3 0 0 1.3-1.6 D 42.2 14.4 0 3 14.6 2 1.3-1.6
Alodine'’ 42.2 14.4 0 2.5 15 1.5 - 1.6 (38-45) (9-18) (2-3) (12-18) (1-2) a. pH was adjusted with HNOj. b. The composition of Alodine coating solution is proprietary; therefore only approximate concentrations are listed.
Table 3-1. Components in test solution for growth -curve studies
FTIR spectra were obtained as described previously in Chapter 2, except that IR spectra were obtained in reflection mode from coated surfaces. Broad range Raman spectra (-500-4000 cm ') were collected using a Chromex spectrometer, with a 514.5 nm excitation laser. All the spectra for quantitative Raman band intensity measurement were obtained on a Kaiser spectrometer, also with a 514.5 nm excitation laser but Raman shift from 300-1700 cm"'.
Prussian blue was prepared by adding 0.5 g Fe(NH 4 )(S0 4 ) 2 I 2 H2O to 20 ml of
0.05M K*Fe(CN)6 3 H2O solution ([Fe"]:[Fe"'] =1:1). After brief stirring, the dark blue precipitate was collected and washed [118-119]. Berlin green was prepared in a similar way, except that 20 ml of 0.05M K 3Fe(CN)6 solution was used, instead of
K4Fe(CN)ô.3H20[l 18-119]. Fe"-CN-Cr'" polymer was prepared by the method of 78 Brown, et al. [112]. In a N 2 purged volumetric flask, Cr"" was reduced to Cr" with Zn
metal, then K4 Fe(CN) 6 was added and stirred for 2 hours. The nearly white precipitate
was collected by filtration, then dried in air to a light yellow solid. Cu“-ferricyanide and
Cu"-feiTocvanide were made by adding 0.9 g K 3pe(CN )6 or 1 g K^FefCbOe SHiO into
20 ml of 0.1 M CUSO 4 solution while stirring [120-122]. The precipitates were collected
and washed. Fe(CN)6 ^ -Cr(OH)3 and Fe(CN)6^-Cr(OH)3 were prepared either
simultaneously or sequentially. In the simultaneous method, ~2g Cr(N 0 3 ) 3 9 H2 O and
0.5g K 3Fe(CN)ô or 0.7g K4Fe(CN)6.3HzO were dissolved in 20ml o f nanopure water,
then 1 M NaOH was added dropwise while agitating the solution. The precipitate was
collected at approximately pH 5-7, then washed with nanopure water. In the sequential
method, Cr(OH )3 was prepared from Cr(N 0 3 )3 .9 H2 0 and NaOH solutions, washed, and
then immersed in 0.05M of K 3Fe(CN) 6 or K 4Fe(CN)6 3 H2O solution, with agitation. The
solid was collected by filtration. Fe(CN)6^'-Al(OH)i and Fe(CN) 6‘*'-Al(OH )3 were
prepared by the simultaneous method as above, except that A 1(N0 3 ) 3 XH2O (x~9)
replaced Cr(N 0 3 ) 3 9 H2O.
“Migration” between various treated AA 2024-T3 and plain polished AA 2024-T3 were conducted in the way described by Zhao et al. [15] All the Rp reported here were obtained by using Gamry software, with Pt as auxiliary electrode and BAS Ag/AgCl as reference electrode.
79 3.3 Experimental Results
3.3.1 Structure of CN species in CCC
The IR and Raman spectra of a CCC film and the peak assignments were shown
in Chapter 2. In addition to the Cr-O and OH vibrations, the IR spectrum shows a Fe-CN peak at 592 cm ' and a CN peak at 2083 cm*'. Raman spectra of the CCC show CN vibrations at 2098 cm*' and 2145 cm*' plus the prominent 858 cm*' band attributed to the
Cr'"-Cr''^' mixed oxide. Some spectroscopic changes were obtained when the CCC was aged in air or water. Figure 3-3 shows Raman spectra of a fresh CCC and one aged for 17 days in air or immersed in Nanopure water. Air exposure causes minor changes in the
Raman spectrum, with a slight decrease in the H-O-H band intensity (~I650cm*') presumably due to dehydration. Water exposure causes a major decrease in the 858 cm*' band, to about 12% of its original area, and some changes in the CN stretch region. With the large decrease in 858 cm*' intensity, the Cr"-OH stretch at 528 cm ' becomes visible.
IR spectra (Figure 3-4) also show little change during air aging, with some decrease in the 3400 cm ' water band, and a significant decrease in the CN stretch intensity during aging in water. As shown in the insets in Figure 3-4, the decrease in 2083 cm ' IR absorption for the sample immersed in water is associated with the relative intensity changes in the two component bands at 2065 and 2089 cm*'.
There are several possibilities for the form of CN and Fe(CN)6 in the CCC film.
Since the CN stretching frequency is quite sensitive to local chemical environment and bonding, it may be used as a fingerprint to compare the fi-equency observed in the CCC
80 K) to 120 eVsec/mW \0 V,
Conventional CCC
120 e/sec/m W
Air aged 17 days
I 15 e‘/sec/m W Water aged 17 days
Synthetic Cr(OH) 30 eVsec/mW
500 1000 1500 2000 2500 3000 3500 4000 300 500 700 Raman shift (cm ')
Figure 3-3. Raman spectra of fresh CCC film and after various treatment (* stray light) Regular CCC
to to o 00 w00
to to I 00 00 to s> Ol VO o U)00
Water-: 17 days
to VO00 1500 2000 2500 3000 3500 Wavenumber (cm' )
Figure 3-4. IR spectra of fresh CCC film and the films after various treatment with that o f known materials. Some of the possible forms o f Fe(CN)6^"‘‘ in the CCC include trapped or physisorbed Fe(CN)6^' or Fe(CN)6^', Cuz Fe(CN)6, Cuj [Fe(CN)6]z, salt of Cr'" and Fe(CN)6^' or Fe(CN)e^', and the degradation products resulting from
Fe(CN)6^' hydrolysis. The latter include Prussian blue (Fe^'^-CN-Fe^O and Berlin green
(Fe^"-CN-Fe^^, which are insoluble in water.
Raman spectra of the CN stretch region of several synthetic compounds and ferricyanide or ferrocyanide are compared to that of a CCC in Figure 3-5. FTIR spectra of the same materials also show C=N stretch bands, as shown in Figure 3-6. The observed
Raman and IR frequencies are listed in Table 3-2. The IR and Raman frequencies of CCC do not match those of K 3Fe(CN) 6 , K4 Fe(CN)6, Berlin green, or Prussian blue, except for the close correspondence of the Berlin green IR frequency (2089 cm ') and one of the
CCC IR bands (2089 cm '). However, incorporation of Fe(CN)6^' into Cr(OH )3 by either the sequential or simultaneous syntheses yields Raman features at 2098, 2132, and
2145cm '. These match the Raman bands of the CCC(2099, 2145) and free Fe(CN)6^'
(2132). IR spectra of the Fe(CN) 6 ^*-Cr(OH )3 sample show a CN stretch at 2065cm ', the same as the IR band observed for the CCC (2065 cm '). IR spectra of Fe(CN)6^'adsorbed onto synthetic Cr'"-Cr^' mixed oxide, or Cr(OH )3 in the presence of NaF, showed identical CN stretch frequencies. Therefore, the CN stretching features in the air-dried
CCC correspond to Fe(CN)6^' physisorbed on Cr(0H)3 for the Raman spectra, and to
Berlin green plus physisorbed Fe(CN)o^ for the IR spectra. The Raman band for Berlin green at 2152cm ' is apparently too weak to be observed in the CCC Raman spectrum, although its IR absorption (2089 cm*') is quite strong.
83 Chemicals v,R,cm‘' VRaman. CUl
K3Fe‘"(CN) 6 solid 2118 2132 K4Fe"(CN)6.3H20 solid 2044 2025, 2066, 2093
Fe'”(CN)6^'solution - 2132 Fe"(CN)6‘‘'solution - 2056, 2095
Berlin green solid 2089 (s) 2152(w) Prussian blue solid 2078 (s) 2159 (w)
CCC (air, 17 days) 2083 2099,2145 CCC (water, ITdays) 2065, 2089 -2099,-2145 (vw)
Cr(OH)3 + Fe'"(CN)6^' 2065 2098,2132,2145
Cr(GH)3 + Fe"(CN)6^- - 2067,2110
A1(0H)3 + Fe"‘(CN)6^' 2133 - A1(0H)3 + Fe"(CN>6'‘' 2058 -
Cu"3[Fe"'(CN)6]2 2100(s),2172(w) 2188 Cu":[Fe"(CN)6] 2098 2116,2153
Table 3-2. Observed CN stretching frequencies o f Fe(CN)6^’^'^‘derivatives
84 \0 IT) tNO CMO
K^Fe(CN)g solution
- I T K3Fe(CN)g solution o o o (N tN ÏN rsis
K^FeCCN)^ (solid)
K^FeCCN)^ (solid)
CCC
Cr(OH)3 - Fe(CN)e- (solid)
Cr(OH)3 - Fe(CN)/- (solid)
Berlin green (solid)
Prussian blue (solid)
1600 1800 2000 2200 2400 2600 2800 Raman shift (cm ')
Figure 3-5. Raman spectra of several CN-containing standard chemicals (• Nj)
85 Wave-number (cm*) 1800 2000 2200 2400 2600
00
i i
CA
Prussian blue
00 o CN Berlin green
o\ 00
Figure 3-6. IR spectra of several CN-containing standard chemicals
86 It is noteworthy that depletion of Cr^' from the CCC by water exposure [with
accompanying loss of the 858 cm’’ Cr'^' band] results in Raman and IR spectra consistent
with Berlin green and Fe(CN) 6 ^' physisorbed on Cr(OH) 3 .
Finally, it was also possible to generate BerUn green by slow hydrolysis of
K3Fe(CN)ô at pH 1-1.5, in the presence of an excess amount o f an oxidizing agent,
NaiSiOg. An IR spectnun of the resulting solid is shown in Figure 3-7. The 2089 cm ' LR
feature can be produced from Fe(CN)&^' without the involvement of chromium species.
Other combinations of Fe, Cu, Cr, CN-, and Fe(CN) 6 ^'or Fe(CN) 6^' were considered, but did not correspond to observed CCC features. In particular, the Al(OH) 3+
Fe(CN)6 ^ or Fe(CN) 6"* and Cu"-Fe(CN) 6^' /Fe(CN)6'*' salts listed in Table 3-2, do not match the observed CCC features. Additional cyano frequencies from the literature are listed in Table 3-3, for several relevant metal/cyano combinations. For mixed metal complexes such as Cr’”-CN-Fe”, the material can exist as isomeric Fe"-CN-Cr'", with a different vcn[1 12]. In the case of the Fe"-CN-Cr'" isomers, experimental values of vcn were unavailable, but well-established empirical rules predict vcn values higher than
Fe(CN)6^' itself [70]. ER absorption for Fe"-CN-Cr'" polymer (2098cm ') is close to the
CCC values o f2083 (air-dried) and 2089 (water-aged), but not as close as Fe‘"-CN-Fe‘“
(2089 cm'). Cr'" complexes of this type are slow to form due to the substitution inertness of Cr^*(aq)[l 12]. Furthermore, Fe"-CN-Cr'" should be susceptible to oxidation by Cr'^' and should be unstable in the CCC. Finally, the IR spectrum of synthetic Fe"-CN-Cr"‘ polymer prepared as described in the literature showed a vcn of 2097cm''[112]. While other cyano complexes are possible, particularly as minority species, the spectroscopic
87 results are consistent with the major product being physisorbed Fe(CN) 6^ and Berlin green (Fe‘"-CN-Fe‘" polymer) produced by partial hydrolysis of Fe(CN)e^' during CCC formation.
Compoimd v c n ( I R ) , cm ' v c n ( I R ) , cm ' vcN(Raman), cm ' (observed) (predicted) (observed)
Cr"'-CN-Fe" 2170[112] --
Cr"'-CN-Fe'" - >2128 ^ - Fe"-CN-Cr'" 2098[112] -2090 [123] 2 0 9 7 '»
Fe"'-CN-Cr"' - >2118“ -
K 4 F e (C N > 6 3 H2O = 2044 [70] 2094, 2058 [127] 2044 [110] 2098, 2062 [129] 2044 [125] 2098, 2062 [126] 2044 [126] 2044 [129]
K 3 F e (C N > 6 2118[123] 2131 [128] 2118 [129] 2135[129] 2118[125] 2132[127] 2132[130] Cr"'-CN 2128 [70] - -
Cu"-CN 2125 [70] - -
Cu'-CN 2094 for Cu(CN) 3^ [70] - - 2076 for Cu(CN)3^'[70] Free CN 2080[124]
2080 [70] ■ a. Predicted as described in Reference 70 b. Obtained for material synthesized during the current work. c. Solid for IR, solution for Raman
Table 3-3. Literature values of cyano stretch of several metallocyanide compoimds. 88 Wave-number (cm ')
1600 1800 2000 2200 2400
§ 3
nI
II
Figure 3-7. Degradation product of Fe(CN)^^- in pH 1.5 HNO3 solution, with excess Na 2 S2 0 g
89 To summarize the structural results, the observed IR and Raman spectra of the
CCC on AA20224-T3 indicate the formation of Berlin green (Fe“'-CN-Fe”* polymer) and
physisorption of Fe(CN)6^'to the polymeric Cr(OH )3 matrix. Prolonged exposure to water
( 17 days) caused a loss of 80-90% of the Cr'^‘ in the CCC, and leaves behind the
insoluble Cr(OH )3 polymeric backbone containing Berlin green and Fe(CN)ô^ • If cyano
compounds other than Berlin green and physisorbed Fe(CN)6^‘ are present in the CCC,
they were not observed with either FTIR or Raman spectroscopy.
3.3.2 Function of FefCN)^^' in CCC formation
Now that the major products of Fe(CN)6^‘ in the CCC are identified, the question
of the role of Fe(CN)6^‘ in the CCC formation may be answered. It is well known that
Fe(CN)6^' increases the rate of CCC formation on aluminum alloys, hence the term
“accelerator”. In the film growth study, special coating solutions were made, as listed in
Table 3-1. AA2024-T3 surfaces were studied by Raman quantitatively after the treatments described in the Experimental section. Several spectra were taken from each sample by focusing laser on different position. Since the typical sampling depth covers the whole film, when spectroscopic conditions were accurately reproduced (2.5mW power and lOxls scan average), the average 860 cm ’ band intensity indicates the film thickness. McGovern et al. confirmed the linear correlation between CCC film thickness and the intensity of 860cm ’ by profilometry [131]. An example of current film growth rate study is shown in Figure 3-8, where longer exposure to Alodine coating solution results in higher average intensity of 860 cm ' band. The effect of Fe(CN)&^' and other additives on film growth rate is shown clearly in Figure 3-9 and Figure 3-10, which
90 Average intensity of 858cm-' band Treatment 1200
1200 Alodine I Smin 1090 Alodine / 3min D Alodine / Imin < 700 I
As-poiished AA2024-T3
400 800 1200 1600 Raman shift (cm ')
Figure 3-8. Correlation between average 860 cm ' band intensity and film thickness (♦ stray light) CfV, Fe(CN)/-, NaF, KBF^, KjZrF,, HNO3 (D)
1600 r^', Fe(CN)&3 , NaF, HNO3 (C) 1400
1200 Alodine, acidified with HNO c 1000
800
600 S Cr'", NaCN, NaF, HNO3 (B) W)00 400 Alodine, acidified with HCIO 00 / CrV', NaF, HNO3 (A) 200
T 0 50 100 150 200 250 300 350 Bath immersion time (second)
Figure 3-9. Growth curves of CCC films on AA2024-T3 CrV, Fe(CN)é^-, NaF, KBF„ K^ZrP^,, HNOj(D) 1800 Cr\", Fe(CN)6^ , NaF, HNO)(C) 1600 1400 1200 Alodine, acidified with HNO I 1000 1 800 S 600 u 400 00 Cr'/', NaCN, NaF, HNO3 (B) •r» 00 200 0 ^CfV, NaF, HNOj(A)
0 50 100 150 200 250 300 350
Bath immersion time (seconds)
Figure 3-10. Growth curves of CCC films on 99.999% Aluminum shows the growth of the 858 cm ' band intensity of the CCC as a function of immersion
time in various bath solutions. The error bars represent the standard deviation of the
858cm ' peak intensity on each sample surface. Separate tests indicate that the 858cm '
peak intensity deviation from sample to sample lies in the same error range when spectral
and preparation conditions were accurately reproduced. Figure 3-9 shows the growth of
the 858 cm ' band on AA 2024-T3 for several of these synthetic solutions, all at room
temperature. KBF 4 and KaZrFe have little impact on film growth, and HCIO 4 may be
substituted for HNO 3 with no apparent effect. Growth is much slower if Fe(CN) 6^' is omitted, with the 858 cm ' Raman band reaching a level only a few percent of its value
for the Alodine solution. Replacement of KsFefOOe with NaCN did not accelerate growth in the absence of Fe(CN)e^', so free CN resulting from FefCN)^^ decomposition does not appear sufficient for rapid film growth. Similar behavior was observed for
99.999% aluminum (Figure3-10), although the deposition rate in the absence of
Fe(CN)6 ^‘ was faster than that observed for AA 2024-T3. The difference between
AA2024-T3 and pure A1 will be discussed in Chapter 6 . The Raman spectra for the CCC films grown in solutions containing Fe(CN)ô^'are shown in Figiu-e 3-11, and indicate that the absence of KBF 4 , RzZrF^, or HNO 3 does not lead to spectroscopically observable changes in film structure. Taken as a whole, the results depicted in Figure 3-9 and 3-10 lead to the conclusion that the true “accelerator” is either Fe(CN) 6 ^' itself or one of its degradation products, such as Berlin green or Prussian blue.
In the current context, a redox mediation mechanism would be possible if the direct reduction of Cr^' to Cr'" by the alloy is quite slow. Figure 3-9 and 3-10 provide
94 CCC
CrV'/NaF/HN0)/Fe(CN)6^ (C)
Cr'''/NaF/HN0j/Fe(CNV-/Bp^-/ZrP^2- (Q)
Ui
CrV/NaF/HCl0 4 /Fe(CNV VBF^/ZrF,
Cr''‘/NaF/HNO,/lrCL^/BF//ZrF^2-
1000 15002000 2500 Raman Shift (cm ')
Figure 3-11. Raman spectra of CCC films on AA2024-T3 after immersion in several coating solutions strong evidence that this is the case, since CCC growth is very slow in the absence of
Fe(CN)ô^*, indicating a slow reduction of Cr^' by the aluminum or the alloy surface. For redox mediation to increase the reduction rate, the reduced form of the mediator must reduce Cr^' rapidly, and the alloy must reduce the oxidized form of the mediator rapidly, as shown schematically in Figure 3-12. If both of these reactions are fast enough compared to direct reduction of Cr^' by the alloy, the overall reduction rate of Cr^' is increased in the presence of the mediator. The resulting rapid formation of Cr”' results in more rapid CCC growth.
The possibility that Fe(CN) 6^^ or Berlin green/Prussian blue may act as redox mediators exists [132-143], and was tested by observing their reactivity toward Cr^' and
AA2024-T3. Fe(CN)6 ^ and Fe(CN)6‘*' are easily distinguishable by Raman spectroscopy, so relative concentrations can be monitored in solution, as shown in Figure 3-13. First,
6 mM Fe(CN)e^ was rapidly (< 30 s) and completely oxidized by 6 mM CrzO?' Second, polishing AA2024-T3 in Fe(CN)ô^ solution for 2 minutes generated Fe(CN) 6^' in the solution. In addition, immersion of AA 2024-T3 in a solution of Fe(CN)a^ /NaF/HNO]
(pH - 1. 6 ) for 20 minutes also generated an observable concentration of Fe(CN) 6^’-
Similar experiments were performed with the Berlin green/Prussian blue redox system with quite different results (see Figure 3-14). Prussian blue was not detectably oxidized by 6 mM CrzO?" upon stirring for 24 hours. Polishing AA20224-T3 in a solution saturated with Berlin green showed no reduction to Prussian blue. These experiments demonstrate that the redox mediation scheme depicted in Figure 3-12 is feasible for
Fe(CN)6 ''^'but much less likely for Berlin green or Prussian blue. Since the kinetics will
96 Red.
Fast cycle
Ox. AI»
Fe(CN)6^*^'*' or Berlin-green/Prussian-blue
Figure 3-12. Possible mediation reactions
97 ow wo t-o ^ G\ \0 w H^O (truncated)
Fe(CN)/ solution
Polish AA2024-T3 in Fe(CN)^^ solution vo 00
Cr^' solution + Fe(CN)/ solution
AA2024-T3 in Fe(CN)<,^ solution with NaF/HNOj
1000 1500 2000 2500 3000 3500 Raman shift (cm ')
Figure 3-13. Raman monitoring o f reactions between Fe(CN)<,’‘ and Al, and between Fe(CN),,'*' and Cr^' Solution after Cr'^‘ + Prussian-blue for 24 hours
N
s
Solid after Cr^' + Prussian-blue for 24 hours
T 1000 1500 2000 2500 3000 3500 Raman Shift (cm ')
Figure 3-14. (A). Raman spectra of Prussian blue suspension after mixing with Cr^' for 24 hours
(continued on next page) (Continued)
Wavenumber (cm ‘) 1000 15002000 2500 3000 3500
00
I Figure 3-14.(continued) (B) IR spectrum of the solid remaining from polishing AA2024-T3 (mounted in epoxy) in Berlin green supernatant solution depend strongly on local concentrations and mass transport in the solution near CCC, the observations do not permit detailed kinetic analysis. However, it is clear that the two redox cross reactions in Figure 3-15 are much faster than the direct reduction of Cr^' by
AA2024-T3. O f all species known to be present during CCC formation, Fe(CN) 6 ^^^ is the only which fulfils the requirements of a redox mediator. Furthermore, the acceleration of
CCC growth on 99.999% aluminum and the spectroscopic similarity of the resulting film imply that a similar mediation mechanism applies on pure Al as well as AA2024-T3.
Cr207 -2 Fe(CN),
fast fast
C r": Fe(CN), AI«
Figure 3-15. Ferricyanide mediation mechanism
101 In principle, any redox system with a redox potential between that of Cr^'/Cr"'
and Al/Al"’ and also has fast redox kinetics with these two systems should be able to act
as a mediator or accelerator. The Cr'^VCr*" and Al/Al"' potentials vary significantly with
local conditions, but are approximately 2.0V apart under Alodine conditions. IrClo"" "
(E° = 1.02V vs. NHE), Fe^^^^ (E° = 0.77V vs. NHE), and V^^^^ (E° = -0.26V vs. NHE)
have redox potentials in the required range (compared to Fe(CN) 6^"^^* at +0.36V vs.
NHE), and were tested as mediators. Growth curves for the 858cm ' band are shown in
Figure 3-16 for solutions in which one of these mediator candidates was substituted for
Fe(CN)6 ^’. Fe^\ V^^, and IrCle'" all accelerated CCC growth, although not as well as
Fe(CN)6 ^‘- The film formed when IrCle^' was substituted for FefCN)^^' had a Raman
spectrum very similar to that of the conventional CCC, except for the lack of the CN
bands (Figure 3-11). Furthermore, other properties of the film formed with IrCU"’ are
similar to a conventional CCC. The IrClo^ film released soluble Cr^' like a CCC, and protected AA 2024-T3 from visual corrosion in 0.1 M NaCl. The polarization resistance
increases from <10‘* to >10^ Qcm^ upon formation of the IrCle''* catalyzed film. Finally, the film formed with IrCU^* can protect a nearby sample o f untreated AA2024-T3 alloy
from visible corrosion in the “migration” experiment.
102 E«(V) ox./red. 1400 ( « SUE» Fe(CN),^ (D) g 1200 1.33 Cr20/2/Cr+3 < J 1000 1.02 IrClft-VlrCV^ irci*; 1.00 V(OH)//VO+^ I 800 C V3+ s 600 0.77 Fe+3/Fe+2 x> Fe3+ r 400 0.34 Fe(CN)/3/Fe(CN)(,+4 2 I » 200 00 0 No accelerator (A)
"T------1------1------1------1------r }------1 - 1.66 ArVAl 0 50 100 150 200 250 300 350
Bath immersion time (seconds) -2.07 AIF.VAI
Figure 3-16. Growth curves of CGC on AA2024-T3 mediated by several red-ox systems 3.3.3 Rp of AA 2024-T3 and AI with various coating thickness
Rp of the coated AA2024-T3 or pure AI samples were tested in 0.1 M NaCl, and
listed in Table 3-4, Table 3-5 respectively. In general, the brief exposure to Cr'^' containing solution improved the corrosion resistance, despite the big difference in film thickness. There is some difference between AA2024-T3 and pure aluminum, mainly in
Open-Circuit-Potential (OCP) values. OOP of AA2024-T3 dropped more than 100 mV after exposure to Cr^', with or without film formation; while for pure aluminum, the OCP is approximately 60-100 mV higher than bare surfaces.
Coating Rp GCP (mV) Coating solution time (min) (G cm ) vs. Ag/AgCl
None 0 1.03 X 10^ ~-500 A Iodine I 3.36 X 10^ -854 Alodine 3 1.935 X 10^ -619
.Alodine 5 2.51 X 10^ -648
A Cr^'/NaF/HNO] 1 6.19 X 10^ -739 A Cr'^/NaF/HNOa 3 2.28 X 10^ -684 A Cr^'/NaF/HNOs 5 2.78 X 10^ -655 B Cr'^'/NaF/NaCN/HNGs 3 2.74 X 10® -651
B Cr'^'/NaF/NaCN/HNGs 5 3.97 X 10® -608
C Cr^‘/NaF/Fe(CN)6^VHNG3 1 1.96 X 10® -664
C Cr''‘/NaF/Fe(CN)6 ^VHNG3 3 3.47 X 10® -684
C Cr'"‘/NaF/Fe(CN)6^VHNG3 5 2.84 X 10® -673
Table 3-4. Rp of AA2024-T3 after coating treatment in various solution
104 Coating OCP (mV) Coating solution Rp , time (min) (Q cm ) vs. Ag/AgCl)
None 0 7.94 X 10^ -882
Alodine 1 8.60 X 10^ -822
Alodine 3 1.51 X 10^ -843
Alodine 5 8.70 X 10^ -882
A: Cr^'/NaF/HNOs 1 6.42 X 10^ -820
A: Cr'"/NaF/HN0 3 3 1.57 X 10® -771
A: Cr'^‘/NaF/HN0 3 5 1.19 X 10® -780
B: Cr'"‘/NaF/NaCN/HN0 3 3 2.03 X 10® -768
B: Cr'''/NaF/NaCN/HN0 3 5 1.30 X 10® -776
C: Cr'",T4aF/Fe(CN)6^VHN03 1 9.55 X 10® -819
C: Cr'"/'NaF/Fe(CN)6^7 HN0 3 3 1.79 X 10® -816
C: Cr'"/NaF/Fe(CN)6 ^VHN0 3 5 1.86 X 10® -829
Table 3-5. Rp of pure Al after coating treatment in various solution
105 3.4 Discussion
The principle structural conclusions available from the Raman and IR results are
that the CCC contains Fe(CN) 6^’ adsorbed to Cr(OH) 3 , and Berlin green, a complex hydrolysis product containing Fe*“ bridging by CN groups. The shift in vcn downwards by 35cm ' when Fe(CN)6^' is adsorbed on Cr(OH )3 implies a fairly strong interaction.
Experimental and theoretical evidence has been reported for similar vcn shifts in compounds of the type X-CN—H-O-Y, where X and Y are either inorganic or organic molecules. Such interactions weaken the C=N bond by redistribution of electron density, thus lowering vcn [144-148]. Therefore, the observed vibrational spectra support the conclusion that much of the cyano species in the CCC is in the form of Fe(CN) 6^' hydrogen bonded to Cr-OH groups in the Cr(OH )3 matrix. The additional cyano component of the CCC is Berlin green, with the generic structure of Fe'"-CN-Fe‘". Berlin green is a three-dimensional lattice with Fe^^ ions located in octahedral sites bonded to the carbons of CN groups, and also in complementary octahedral sites bonded to the nitrogen ends (see Figure 3-2). The CN stretching frequency, vcn, for pure Berlin green is similar to that for Berlin green in CCC, imply that the solid hydrolysis product is co precipitated in Cr(OH )3 without strong interactions with the Cr(OH )3 matrix. Since Berlin green has fairly low solubility, it might have formed tiny crystals on the porous Cr(OH )3 surface. Raman and FTER are less sensitive to Fe-O species and such groups may exist in the CCC, but cyano species other than H-bonded Fe(CN> 6^' and Berlin green are unlikely.
If significant amounts of other cyano species are present, their C=N stretching vibrations are not apparent in the IR and Raman spectra.
106 The reduction of Cr^' to Cr'” by aluminum, whether direct or indirect, has long
been accepted as a critical step in CCC formation. The role of ferricyanide in the process
is less clear from the literature, although interference with Cr'"‘ adsorption [41] and
interactions with Cu containing intermetallics [108] have been proposed. The current
results confirm that direct reduction of Cr^' in the absence of Fe(CN) 6 ^' is quite slow,
particularly for the AA 2024-T3 alloy. The results also establish that ferricyanide can act
as a redox mediator, through rapid reduction to ferrocyanide by the alloy, and rapid
oxidation of ferrocyanide by Cr'^’, as depicted in Figure 3-15. Fe-CN containing
chemicals had been used previously as redox mediators for a variety of redox processes,
ranging from small molecules to redox enzymes [132-143]. During CCC formation, several other redox systems can substitute for Fe(CN) 6 ^ as accelerators, although not as effectively. The rates of CCC film growth for the mediators tested follow the trend
Fe(CN)6^* > IrClô^’ > >Fe^^. However, this series does not correspond to the order of
values [IrC^^"^'" >Fe^^ ' > Fe(CN)6^'^‘*‘ > V^'^'^^]. This lack o f correspondence implies that the kinetics of the cross reactions between Cr^', mediator and alloy have a larger effect on CCC growth rate than any tendency of the mediator to control the alloy potential at a value near the mediator’s E°'.
It is worth noticing that CCC film grows much faster on pure A1 than on AA2024-
T3. The reason probably is the dramatic difference in their Open-Circuit-Potential. For more details, please refer to Chapter 6 .
While the acceleration of CCC formation through redox mediation by Fe(CN) 6 ^’ is strongly supported by the current results, the possibility of other roles for Fe(CN) 6^'
107 in corrosion protection should be considered [108]. Given that the final film contains
physisorbed Fe(CN) 6 ^‘ and Berlin green, are these components important to film behavior
and corrosion protection? The chromate film formed with IrCU^' mediator is
spectroscopically and visually very similar to a conventional CCC film. It can release
Cr^' into water, a process believed to be important to “self-healing” of CCC [15]. The
chromate film formed with IrCle^' also increases Rp and decreases the corrosion potential
of AA 2024-T3, in much the same manner as a conventional CCC. Finally, the IrCU-
formed film exhibits active protection by migration of released Cr^' to an initially
untreated AA 2024-T3 sample. At least by those tests, the chromate film formed in the
absence of Fe(CN) 6^' ^*’ is visually, structurally, and functionally very similar to a
conventional CCC. The current results do not rule out actions other than redox mediation
by Fe(CN)6 ^' ^*', but they do establish that redox mediation accounts for the current
observ ations, Fe(CN) 6^'remaining in the CCC as Berlin green or physisorbed Fe(CN) 6^' is
not necessary for higher polarization resistance, slower visual corrosion, or self-healing
in the artificial scratch cell.
The behavior of various intermetallic compounds (IMCs) is often invoked in
explanations of corrosion mechanisms [149-161] and protection by a CCC [162-164],
McGovern et al.’s study [131, 164] indicates that the FefCN)^^ mediated CCC film
formation was slowed down on Cu- or Fe-rich intermetallics, and possible reasons were proposed, such as CN binding to intermetallics and preventing further film deposition[164]. In addition, the film formed by IrCle^' mediation is more uniform, with more deposition on intermetallics compared with a Fe(CN) 6^ mediated CCC film [164].
108 In this work, the Rp of IrCU-CCC is about one order of magnitude lower than regular
CCC film on AA2024-T3. Those differences all point at the possibility of
Cu<—>Fe(CN)6 ^ compound interaction and their possible impact on the overall corrosion protection mechanism. The formation of Cu-Fe(CN )6 compoimds at intermetallic compounds (IMCs) was proposed to decrease their reactivity and reduce corrosion[108]. The current spectroscopic observations were not resolved spatially, and hence the spectra represent spatial averages over 100-500 microns of a CCC or alloy surface. IMCs are present at levels of a few percent in particles smaller than this sampling area, so spectroscopic features resulting from IMC interactions with the CCC may contribute to the observed spectra in minor ways. It is very possible that spectral features of the IMCs are lost in the relatively strong signals from the CCC itself. Both the IR and
Raman spectra reflect major components of the CCC, and minor contributions may quite possibly be unobservable. The current results do not rule out the possible effect related to the IMCs, which represent a small fiaction of the total IR or Raman spectra, may be important to corrosion protection by the CCC.
Another important issue, which might not be related to Fe(CN) 6 ^', is the Rp values of AA2024-T3 or pure A1 after being treated in various coating solutions (see
Table 3-4, Table 3-5). For both AA2024-T3 and pure Al, simple exposure to Cr^' solution is enough to provide protection, no thick film is required, and the Rp does not correlate with film thickness at all. One thing is certain from the results: the high Rp and low OCP after chromate treatment do not require a thick film, like a conventional CCC film. Previous studies[15] also indicate that very diluted Cr^^ in solution phase is enough
109 to stop corrosion. The similarity between AA2024-T3 and Al in this issue implies that, whatever Cr^' is doing, it can act directly on Al and its corrosion sites. No intermetallic sites are necessary. Of course, it is also possible that Cr^' protection mechanism on
AA2024-T3 could be very different from that on pure aluminum. One important point is that a thick film provides a storage mechanism (discussed in Chapter 4), but a thin film is adequate to establish a high Rp.
3.5 Conclusions
Raman and IR spectra of CCC film revealed CN species in the CCC film. The structures of the CN species were identified as follow: most of the CN in the CCC film is
Fe(CN)6 ^’, which is physically adsorbed onto Cr(OH )3 solid through hydrogen bonding; small amount of the CN exists as Berlin Green precipitate, which is the decomposition product of Fe(CN) 6^' in a highly acidic condition.
Among all the additives in Alodine coating solution, Fe(CN) 6^‘ is the only one that can “accelerate” the CCC film formation. The acceleration mechanism is shown in
Figure 3-15. Fe(CN) 6 ^'or Fe(CN) 6'*’ can react with Al or Cr^', respectively. The total reaction rate of such mediation reaction cycle is much faster than that of the direct reaction between Al and Cr^', thus accelerating the film formation rate.
Rp values of AA2024-T3 or pure Al samples covered by CCC films were studied.
The Rp values do not correlate with film thickness, and a thin film is adequate to establish a high Rp
110 CHAPTER 4
STORAGE AND RELEASE OF SOLUBLE HEXAVALENT CHROMIUM FROM
CHROMATE CONVERSION COATINGS:
EQUILIBRIUM ASPECTS OF CHROMIUM(VI) CONCENTRATION
4. 1 Introduction
It is generally accepted that mobile Cr^' in chromate conversion coatings (CCCs) and in SrCr0 4 containing primers are the critical components in corrosion protection for coated Al alloy components used in aerospace applications. The ability of a CCC to release Cr'^* as soluble chromate species is likely to be important to “self healing” exhibited by CCC, in which Cr^' or related species can migrate to defects or corrosion sites and to inhibit further damage [13-15]. As described in previous chapters, Cr"'-Cr^' mixed oxide has been identified as a major CCC component. Furthermore, release of
Cr^' from a CCC has been demonstrated, as has protection of an initially untreated alloy surface by dilute Cr0 4 ^‘ in a chloride solution [15]. The mechanism o f corrosion protection by chromate is currently being debated, but storage and release of Cr'^' by a
CCC appear to be essential for its long-term protection property. In addition, release of
111 chromate from sparingly soluble SrCr 0 4 in primers may also provide a source of dilute chromate for self-healing.
The current investigation addresses the storage and release of Cr'"’ in more quantitative detail. The release of Cr^% from a CCC and SrCr 0 4 , into water and salt solution was monitored quantitatively with UV-Vis spectroscopy in order to examine solution concentrations, saturation (if any), release rate, and possibly storage mechanism.
By considering a variety of conditions, a quantitative model for Cr^' storage and release was formulated, and its implications to corrosion protection were considered.
4.2 Experimental
All the chemicals used were analytical grade. Solutions were prepared with
“deionized” water (Bamstead, “Nanopure” 18 MQ cm). The absorbance vs. concentration behavior of chromate solutions is complicated by the equilibria between
HCr0 4 , Cr0 4 ^ , and CrzO?^, depending on both concentration and pH [165-169]. The combined concentration of these species in solution is indicated herein as [Cr'^‘], In order to determine the relationship between UV-Vis absorbance and [Cr'^'j, solutions of
KiCrjO? were prepared at various concentrations in the range of 1.0 x 10'^ M to
4 X 10"* M of Cr"^\ and adjusted to different pH values with HCIO 4 or NaOH and a pH meter. UV-Vis absorption spectra for solutions in 1-cm quartz cuvettes were collected using a Perkin Elmer “Lambda 20” spectrometer (see Figure 2-4). The spectra for one
Cr^ ' concentration are shown in Figure 4-1 for a pH range from 2.01 - 9.47. Since the absorption at 339 nm (A339 ) was independent o f pH, this wavelength was used to construct the pH-independent calibration line shown in Figure 4-2. Over the range of
112 339.0nm
<
(S II % 0 -
0 0 20300 350400 450 500250 A.(nm)
Figure 4-1. UV-Vis spectra of Cr^' solution at different pH but the same total concentration of 8.30 X lO'^M (the average is used as the second data point in Figure 4-2) = 0.0012+ 1475.8[Cr''‘] 0. 6 ” ‘339nm
Average from Figure 4-1
0. 2 -
0.0 -
0.0 5.0x10-5 10x104 1.5x10-4 2.0x10-4 2.5x10-4 3.0x10"* 3.5x10"* [Crv'l (M)
Figure 4-2. Calibration line of Ajj.,,,,,, vs. [Cr^' ] (pH range from 2.01 to 9.47) total [Cr^'] from 1 x 10'^ M to 4 x 10^ M and pH range 2.0 to 9.5, A 339 is linear with total
[Cr'"].
This calibration curve was used to determine [Cr'^*] from observed A 3 3 9 in
subsequent experiments. The molar absorptivity in terms of total Cr^' concentration is
1.49 X 10^ M'' cm ' at 339 nm over the pH range 2.0-9.5.
Cr'^^' concentrations in solution were determined spectrophotometrically as a
function of time for several starting conditions, including a CCC in water or salt solution,
Cr"'-Cr'^' mixed oxide in water, and Cr‘“ hydroxide immersed in dilute Cr^' solution.
CCC films were prepared by dipping polished AA2024-T3 alloy coupons into freshly prepared “Alodine” coating solution (7.5 g/1, pH = 1.3) at room temperature for 1 minute, washing with more than 100 ml flowing water and then drying in air for various period of
time (indicated in Table 4-1). Then the coupons were immersed in a known volume of deionized water or 0.1 M NaCl, in well-sealed glass containers. Cr'"-Cr^' mixed-oxide was prepared as described previously in Chapter 2. A known weight of freshly prepared
Cr’"-Cr'^' mixed-oxide was weighed and dried on a small piece of glass in air for known periods of time. The dried Cr^'-Cr'^' mixed-oxide powder samples on glass were immersed in known volumes of Nanopure water or 0.1 M NaCl solution and tightly covered. The containers were agitated occasionally over a period of several days.
Aliquots (2-3 ml) of each solution were withdrawn periodically for UV-Vis spectrophotometric measurement, and then returned to the corresponding container. The details of CCC surface/volume ratio and Cr‘"-Cr'^' mixed-oxide mass/volume ratio are listed in Table 4-1.
115 Material Aging time A/V or m/V (2024-T3 u n le s s (hrs) (cm^/ml) or (g/ml) [Cr'^'lcqu.i (M) indicated otherwise)
CCC 0 0.4 6.0 X 10 - CCC 0 0.8 1.2 X 10-* CCC 0 1.2 1.7 X 10"* CCC 0.25 4.5 1.34x10"*
CCC 22 4.5 1.54 X 10"*
CCC 96 4.5 1.32 X 10"* CCC 210 4.5 1.16x 10"* CCC 22 4.5 1.94 X 10“* CCC 2 1.67 1.0 X 10"*
CCC 2 5.0 3.6 X 10"*
Cr"'-Cr'^‘ mixed oxide 0.25 0.0032 2.11 X 10"*
Cr'"-Cr^' mixed oxide 22 0.0032 1.56 X 10"*
Cr'"-Cr^' mixed oxide 96 0.0032 1.33 X 10"*
Cr'"-Cr^' mixed oxide 210 0.0032 1.28 X 10"*
Cr‘“-Cr''* mixed oxide “ 0.25 0.0032 2.97 X 10"*
Cr'"-Cr^' mixed oxide “ 22 0.0032 2.92 X 10"*
Cr'"-Cr^' mixed oxide 0.25 0.00092 4.03 X 10"*
Cr"‘-Cr'^‘ mixed oxide 0.25 0.00092 4.48 X 10"*
Cr'"-Cr'^ ' mixed oxide 0.25 0.0013 2.70 X 10"*
Cr‘"-Cr'^' mixed oxide 0.25 0.0017 5.33 X 10"*
Cr‘"-Cr'' ' mixed oxide 0.25 0.0020 6.85 X 10"*
Cr'"-Cr'^' mixed oxide 0.25 0.0030 7.66 X 10"*
Cr"’-Cr'^‘ mixed oxide 0.25 0.0037 9.39 X 10"*
Cr'"-Cr^' mixed oxide 0.25 0.0042 9.81 X 10"*
SrCr04 N/A 0.020 4.65 X 10 ^ SrCrO;" N/A 0.020 7.81 X 10^ a. The medium was 0.1M NaCl solution. For others Noanpure water was used.
Table 4-1. Observed Equilibrium [Cr^'] in solution under various conditions
116 Saturation concentrations of SrCr 0 4 in deionized water or 0.1 M NaCl were
obtained by adding 2 g of SrCr 0 4 solid to 100 ml Nanopure water or 0.1 M NaCl, and
stirring for 6 days. After centrifugation, 1.00 ml of the supernatant was diluted to 50.00
ml with 2 M NaOH, and then a UV-Vis spectrum was obtained. The Cr 0 4 ^
concentration was determined from the molar absorptivity at 373 tun, 4884 M*' cm '.
The adsorption of aqueous Cr'^' by Cr"'-hydroxide was studied by adding a measured amount, (0.1- 1 g) of freshly prepared solid Cr(OH )3 to a 50.00 ml solution with known initial [Cr^'] and pH. The containers were sealed and shaken occasionally.
Aliquots of solution were withdrawn periodically, centrifuged, quantitatively diluted with
2 M NaOH, and analyzed spectrophotometrically at 373 nm. Details about the mass of
Cr"'-hydroxide and initial/final [Cr^'] are listed in Table 4-2.
The dynamic interaction between Cr”'-hydroxide and Cr^' was studied as follows:
A mixture of ~ 0.062 M (Cr(N 0 3 ) 3 • 9 H2O) and - 0.019 M Cr'^' (as KiCrzO?) was prepared. The total volume was 500 ml and the initial pH was below 3. In the first cycle, concentrated (-20 M) NaOH was added dropwise while agitating the solution. After each NaOH addition, the solution was stirred imtil the pH stabilized to ± 0.1 pH unit. 10 ml of the suspension was centrifuged for 5 minutes, then a 1.00 ml aliquot of supernatant solution was quantitatively diluted to 50.00 ml by 2 M NaOH and analyzed spectrophotometrically by using A 373nm- The remaining supernatant solution and solid were returned to the original 500 ml volume of solution. NaOH was added repeatedly such that UV-Vis. spectra were obtained at 0.5—1.0 pH unit increments. UV-Vis. spectra were obtained after aliquots were diluted with NaOH, since the NO 3 band (310 nm) overlaps with Cr'^’ bands at acidic or neutral pH. After reaching a pH of 12, the process
117 was reversed by incremental addition of concentrated HNO3 until the pH was less than 3.
A second complete cycle of NaOH and HNO3 addition was conducted, in order to
demonstrate reversibility. The numerical results of the experiments are shown in
Table 4-3.
Similar experiments were conducted by using 250 ml suspension o f—13 mM Cr^' plus - 0.75 g of Cr 2 0 3 , or a 500 ml mixtiue solution of — 66 mM A 1(N 0 3 ) 3 xHiO (x - 9) plus - 23 mM Cr'^‘. The corresponding results are shown in Table 4-4 and Table 4-5, respectively.
118 [Cr'^'U mass of Cr(OH )3 (g) initial [Cr'^'lcquil A [Cr'"] “ (mM) (in 50.00 ml solution) pH (mM) (mM )
10.17 1.051 7.19 8.939 1.231 10.17 0.1075 7.19 9.936 0.2340 10.12 1.309 2.76 6.558 3.562 10.12 0.1408 2.76 9.380 0.740 0.937 1.003 7.59 0.5108 0.4259 0.937 0.1563 7.59 0.8369 0.0998 0.947 1.051 2.67 0.0842 0.8595 0.947 0.1075 2.67 0.4621 0.4816 0.112 1.020 7.66 0.01997 0.09243 0.112 0.1321 7.66 0.011 0.1014 0.117 1.008 2.58 0.04097 0.07603 0.117 0.1076 2.58 0.04347 0.07353 10.14 0.2942 7.19 9.430 0.710 10.14 0.5169 7.19 9.182 0.958 10.14 0.8238 7.19 8.835 1.305 10.14 0.3216 2.67 8.531 1.609 10.14 0.5077 2.67 7.661 2.479 10.14 0.8146 2.67 6.324 3.816 10.14 2.8885 7.19 6.598 3.542 10.14 3.6067 7.19 5.002 5.138 10.14 6.0380 7.19 3.245 6.895 10.14 2.7110 2.67 2.412 7.728 10.14 3.9886 2.67 1.736 8.404 10.14 5.7103 2.67 0.5566 9.5834 5.91 1.0046 2.68 1.749 4.156 5.91 1.1340 7.18 4.189 1.716 3.01 1.0421 2.71 0.1383 2.8677 3.01 1.1022 7.18 1.701 1.305 - . Vît Vin ■ r ^ _ V I i -- ....
Table 4-2. Equilibrium Cr'^' concentration in solution after Cr‘“-hydroxideIII adsorption
119 cycle # 1 cycle #2
pH [Cr'"] (mM) [Cr'"] (mM) pH [Cr'"] (mM) [Cr'"] (mM)
2.39 19.27 63.00 4.60 0.6485 0.3548
3.81 19.60 62.80 5.50 1.212 0.2053
4.15 19.11 62.7 6.57 3.093 0.3964
5.93 0.500 0.3189 7.12 5.211 0.5254
7.03 2.066 0.2110 8.05 6.675 0.6947
7.78 4.186 0.4430 8.11 9.262 0.5558
8.29 6.385 0.1090 9.29 12.305 - 0 ^
9.25 10.20 0.0980 10.51 15.19 - 0 “
9.98 16.60 - 0 “ 12.79 16.49 - 0 “
8.44 9.985 0.0950 11.63 16.31 - 0 “
7.64 7.090 0.2850 10.13 14.12 -O'
6.78 3.703 0.1850 9.70 13.35 0.313
5.47 0.9415 0.0955 8.02 7.828 0.4166
3.13 0.5185 0.6135 7.19 4.501 0.3393
5.74 2.925 0.4466
2.78 0.6199 0.6915
a. The absorption is lower than detection limit - 0.002.
Table 4-3. Solution concentrations of Cr'" and Cr^' during Cr'"-Cr^' pH cycling experiment
120 pH [Cr'^'l (mM)
2.46 12.86
4.75 12.86
5.71 12.95
6.95 12.86
8.31 12.70
11.40 12.90
8.54 12.68
5.83 12.60
3.12 12.48
Table 4-4. Solution concentrations of during CrzOg-Cr'^' pH cycling experiment
121 cycle #1 cycle #2
PH [Cr^'] (mM) pH [Cr'"] (mM)
2.81 23.01 3.58 21.24 3.50 22.10 4.49 8.898 4.67 8.170 4.56 10.79 6.53 11.19 5.06 10.82 7.29 16.24 5.55 10.37 11.39 22.96 5.83 10.66 11.32 22.07 6.53 12.53 11.23 22.02 7.26 15.25 10.51 21.64 7.48 16.37 10.77 21.30 7.60 17.53 7.52 18.25 7.95 18.80 6.87 13.24 9.10 20.30 6.13 11.97 11.44 20.80 4.95 9.872 8.11 18.97 4.27 10.07 7.66 18.56 7.61 16.78 7.48 15.84 6.91 14.61 6.36 13.63 5.96 12.95 5.26 12.20 4.53 12.61 4.32 12.78
Table 4-5 Solution concentrations of Cr^' during Al“‘-Cr''‘III -VIpH cycling experiment
122 4.3 Results
When a CCC was immersed in water or salt solution, the Cr^' concentration in
solution increased with time, as shown in Figure 4-3. The Cr'^' is a mixture of CrO^" ,
CriO?' and HCrO^ , depending on concentration and pH, but the absorbance at 339 nm
permits assessment of total [Cr^'] in solution. All release curves of the type shown in
Figure 4-3 eventually reached a constant Cr^' concentration after several days, and these
“constant” concentrations, called [Cr'^*]equiu are listed in Table 4-1 for a variety of initial
conditions. Several observations deserve special note. First, a higher ratio of CCC area
to solution volume led to higher [Cr'^']equii in solution. Second, aging of the CCC before
exposure to water decreased the [Cr^']cquii, but not greatly. Third, the [Cr'^'jcquii observed
in 0.1 M NaCl was higher than that in deionized water, by approximately 26%, as shown
in Table 4-1.
Release of Cr’^* from Cr"'-Cr^' mixed oxide was studied by adding known
weights of synthetic mixed oxide to known volumes of water or salt solution, followed by
spectrophotometric monitoring. Release curves similar to those of Figures 4-3 were
observed, and the “constant” concentrations, [Cr'^'jequii, are listed in Table 4-1. As was
the case with the CCC, the final [Cr^'] after release from the synthetic mixed oxide was
in the range of 10“* to 10'^ M, decreased with aging time, and was slightly higher in 0.1
M NaCl than in water, as shown in Figure 4-4. Table 4-1 also lists the [Cr'^'jcquii
“released” from solid SrCrO^ into water and 0.1 M NaCl. For both the mixed oxide and
SrCr0 4 , solid remained after reaching the [Cr'^*]equii level. The ratio of the final [Cr'^'] observed in water to that in 0.1 M NaCl ranged fix>m 0.60 for SrCr 0 4 , 0.5-0.7 for the
CCC, to 0.79 for the mixed oxide. This ratio did vary somewhat with aging time and
123 1.6 X 10"» -
area/volume 1.2 X 10"» ■ - 0.4 cm^/ml I - 0.8 cm^/ml
U - 1.2 cm^/ml
0.0
0 50 100 150 200 250 300 350
Releasing time (hour)
Figure 4-3. “Release curves” of CCC in nanopure water under different area/volume ratio 3.5x10"* 1
3.0x10"* "NaCl O.IM
g 2.5x10"*
Ig 2,0x10-'• Nanopure water U o 1.5x10"* o
O 1.0x10"* UtK> 5.0x10-5 .j Air-dried for 10 minutes 0.0 Air-dried for 22 hours
—f— 0 50 100 150 200 250 300 350 time (hour)
Figure 4-4. “Release curves” of Cr'"-Cr''‘ mixed oxides in nanopure water or 0.1M NaCl solution area/volume ratio, but was consistently in the range of 0.6 to 0.8 for all three starting materials. The activity coefficient of the dominant Cr'^* species in solution (HCrOa ) in
0.1 M NaCl was calculated by using the extended-Debye-Huckle equation [165],
logy = ------O.SlZ^Æ2—p= Equation r- 4-1 , l + (a/305)vt where z is the charge of the ion, p is the ionic strength of the solution, and a is the effective hydrated radius, in pm, of the ion and its tightly bound sheath of water molecules. Under current conditions, for HCrOT,
z = 1
p = 0.1 M
a « 400 pm [165] therefore, y is 0.769. The experimental ratios are close to this calculated value, implying that the higher concentration observed in NaCl is mainly due to a reduced activity coefficient. Another possible effect of chloride on [Cr'^']equii is that chloride could compete with Cr'^' (aq) and form Cr'"-Cl, leaving more Cr^' in solution. However, chloride exchanges with -OH of Cr"'-hydroxide at a much slower rate than water [170].
Furthermore, [HiO] is much higher than Cl imder current experimental conditions, therefore the possible formation of Cr"‘-Cl does not play a significant role in this study.
If the constant [Cr^'] observed at long times in Figures 4-3 and 4-4 represents an equilibrium between solution and solid Cr^', then Cr'" oxide should absorb Cr^' from solution, according to Reaction 1.
126 f^atis Cr'"-OH (solid) + HCr 0 4 (aq) + FT - Cr^'-O-Cr'^'OsH (solid) + HiO (4-1 ) ^desorb
Reaction 1 is a simplified form of the reaction proposed in Chapter 2 to represent a dynamic equilibrium between the Cr'“-Cr'^' mixed oxide and aqueous Cr'^', and the remaining bonds to Cr*” are not shown to improve clarity. HCr 0 4 is used as an example of Cr^‘ (aq); similar equilibriums with Cr 0 4 ^' (aq) or CrzO?^ (aq) are also possible. As written, the reverse reaction describes release of Cr'^* into solution from the CCC or mixed oxide, while the forward reaction represents Cr'^' adsorption by Cr”’ hydroxide.
To test the forward reaction, Cr’” hydroxide was added to Cr^’ solution, and the Cr''' in solution was monitored spectrophotometrically. Figure 4-5 shows several plots of [Cr' ’] vs. time for two pH values and two weights of Cr’” hydroxide. The solid Cr(OH )3 does indeed adsorb Cr'^’ from solution, and the solution concentration decreases to an apparently constant [Cr''^’] level. Table 4-2 lists several final Cr'^’ concentrations
([Cr''^’]cquii) for different initial conditions; some are plotted in Figure 4-5 (C) to illustrate the trends. In general, adsorption is more pronounced at lower pH, at higher initial [Cr'^‘] and for higher weights of Cr’” hydroxide.
The dynamic nature of the equilibrium between dissolved and adsorbed Cr' ' was examined further by cycling the pH of a Cr”’ + Cr'^’ mixture solution, as described in the experimental section. As concentrated NaOH was added dropwise, to soluble Cr”’ and
Cr' ' salts to form the brown Cr’”-Cr'^’ mixed oxide, and the suspension was allowed to equilibrate at a given pH for more than 2 hoiu-s. The total [Cr'^’j in the supernatant solution above the solid was determined spectrophotometrically and the pH was
127 0 .0 1 1 1 O.lg/pH-7 0.010 0.1g/pH~3 0.009 lg/pH~7 u 0.008 4 Ig/pH~3 0.007 '
0.006 0 50 100 150 200 250 Time (hours)
l.OOi 0.1g/pH~7 0.80
E 0.60 lg/pH~7
0.40
0.20' lg/pH~3 0.00 0 100 150 200 25050 Time (hours)
Figure 4-5. Cr^' adsorption by Cr'"-hydroxide.(A) [Cr'^]t=o«10mM, (B) [Cr'^’U » ImM
(Continued on next page)
128 (continued)
pH of initial solution • pH 2-3 0 pH 6-7
0.00 0.02 0.04 0.06 0.08 0.10 0.12 (M)
Figure 4-5 (C) Dependency of Cr^' adsorption on pH and [Cr^'),^Q (Mass ofCr(OH)j was kept at around 1 gram. A [Cr'^']=[Cr^'] ,=o-[Cr'''],.^„jn ) measured before adding another aliquot of NaOH and re-equilibration at a different pH.
After reaching pH 10-12, HNO 3 was added to decrease pH. Two complete pH excursions
were carried out in this fashion. A plot o f solution [Cr'^’jequii vs. pH is shown in Figure 4-
6 . Each point was taken after the addition of concentrated NaOH or HNO 3 and
equilibration at a particular pH. A similar experiment starting with solid Crz 0 3 and Cr'^’
solution exhibited no adsorption of Cr^' from solution, as shown in Figure 4-7. The
results obtained from Al‘“ (aq) + Cr^' are shown on Figure 4-8, overlaid with the Cr‘“ +
Cr^' results. The similarities and differences between those two systems, Al“‘ + Cr'^' and
Cr‘" + Cr^ are clear. Both can form hydroxide precipitates at certain pH range, and the
hydroxide adsorbed Cr'^‘ from solution. However, Al'"-Cr^' has a narrower stable pH
range than Cr'"-Cr^'. Total dissolution o f Al”'-Cr'^* solid was observed at pH < ~ 4 or pH
> -10, while Cr'^’-Cr'^' solid was still stable at pH - 2. Meanwhile, Al‘“-Cr'^‘ system has
lower Cr'’ ' capacity than Cr'“-Cr'^’, almost 50% less. To assure time for equilibration, the
results shown in Figure 4-8 were acquired over a period of 7 days, in a similar
experiment to that shown in Figure 4-6 and Figure 4-7.
Cr^' loading levels in Cr'"-Cr'^* or CCC film were measured as described previously in Chapter 2. CCC film was dissolved in basic solution, or the Cr'"-Cr^' mixed oxide was dissolved in concentrated HNO 3, then the pH of both was increased to >12 by adding NaOH. Spectrophotometric measurements showed that CCC film (Alodine, 1 minute on AA2024-T3) contains (1.03 ± 0.17) x 10'^ mol Cr'^Vcm^and Cr'"-Cr^' mixed- oxide contains (5.46 ± 0.25) x 1 0 "^ mol Cr^Vg. Both results are the average of three trials.
130 Starting point
0.025
Initial [CrV]
z 0.015-
5: 0.010 CydeWl Cycle #2 0.005-
.000 o •
0 2 4 6 8 10 12 14 pH
Figure 4-6. pH cycling of a mixture of ~ 62mM Cr(NOj)j 9H2O and ~ 9.6mM KjCrjOy mixture 0.016
0.014 Initial [Cr''']
0.012
0.010 I 0.008
0.006 wu> 0.004
0.002
0.000 2 4 6 8 10 12 pH
Figure 4-7. pH cycling of a mixture of 250ml ~ 6.5mM K^Cr^O^ and 0.75g CrjOj Around pH=4 Al"'-Cr^' starts to dissolve
25 Initial [Cr^'|~23mIVf 20 •...... *• ...... ^/lA...... Initial [Cr*''J ~ 19mM V J 15
AA A A A o A A^A • • ^ Ai'"-Cr''', cycle #1 A Al'"-Cr''', cycle #2 w u w 5 % • Cr^"-Cr^', cycle UI o • 0 O Cr"'-Cr^', cycle U2
—r~ T— —j— 0 4 6 8 70^ 12 14 pH
Figure 4-8. pH cycling of a mixture of ~ 66mM Al(NOj)j xH^O (x~9) and - 11.5mM K^Cr^O^ (Cr*" + Cr''* results are overlaid as a comparison) 4.4 Discussion
Before considering the quantitative implications of the observations, some useful conclusions are available about Cr'^' storage and release. A simple hypothesis might be that Cr^' is trapped in the CCC as a soluble salt, such as KzCrzO? In this case, the soluble Cr'^‘ would merely dissolve when the CCC is exposed to solution, and the final
Cr''‘ conentration, [Cr^'jcquii, in solution would increase linearly with the area/volume ratio. This case is plotted as the “depletion model” of Figure 4-9, for which the CCC (or mixed oxide) is merely a repository of a soluble Cr'^' salt. The x-axis in Figure 4-9 is the number of moles of Cr^' in the CCC or mixed oxide divided by the solution volume, and is proportional to either the CCC area/volume ratio or the Cr"'-Cr^' mixed oxide weight/volume ratio listed in Table 4-1. This depletion model is inconsistent with the observations for two major reasons. First, the observed [Cr'^’jequii is not linear with x in
Figure 4-9. Second, the uptake of Cr'^' by CrfOH)] indicates that Cr'^' is not completely soluble in the presence of the synthetic Cr'" hydroxide, or the mixed oxide present in the
CCC.
A second possibility is a “solubility model”, in which the Cr^' concentration over a CCC or the mixed oxide is controlled by a solubility product, analogous to the case for
SrCrOa. Ignoring reactions of Cr 0 4 ^' with or H 2O, SrCr04 should behave as a simple sparingly soluble salt, with a saturation concentration determined by its solubility product
(Ksp = 2.2 X I0'^[7l]). The behavior of the solubility model is also shown in Figure 4-9.
Once sufficient Cr^' is available to reach the solubility limit (~ 4.7 mM), the solution saturates and the addition of more SrCr 0 4 has no further effect on [Cr'^'jcquii in solution.
If the CCC or mixed oxide follows this behavior, we expect a saturation level of Cr^',
134 6.0 depletion 5.5
5.0 solubility 4.5
3.0
2.5
2.0 U) LA 1.5 Langmuir 1.0
0.5 CCC on 2024-T3 0.0 Cr'"-Cr'''-mixed-oxide
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 Nv/V (mM)
Figure 4-9. Comparison of three possible models which cannot be exceeded by adding more CCC (increased area/volume) or mixed oxide.
However, this is not the case, increasing area/volume always results in higher [Cr'^‘] in
solution. The solubility model is likely to apply to SrCrO^ contained in the primer, but
not to the CCC or the Cr**‘-Cr'^' mixed oxide.
It is clear that the release of Cr'"‘ from the CCC (or mixed oxide) is reversible and
pH dependent, and strongly dependent on the initial amoimt of Cr'"’ in the CCC or mixed
oxide relative to the solution volume. A mechanism that is consistent with the
observations is adsorption of Cr^' species to the insoluble Cr'" hydroxide, shown
schematically in Figure 4-10. We consider the insoluble Cr'" to have many surface
hydroxyl groups available for formation of Cr'" - O- Cr^' bonds. These hydroxyl groups
act as sites for covalent binding of Cr'^' according to Reaction 4-1. Cr'" - OH (solid) in
Reaction 4-1 represents insoluble Cr'" hydroxide, HCr'^'O^ (aq) is solution phase hexavalent chromium, Cr'" - O - Cr^' is the mixed oxide present in the CCC or prepared synthetically depending on pH and concentration, solution phase Cr^' may be CrOu",
HC1O 4 , or CrzO?" Reaction 4-1 is reversible, so the CCC may release Cr^', and the Cr'" hydroxide may adsorb Cr^' from solution. Release and adsorption of Cr^' are pH dependent, favoring the mixed oxide at low pH and soluble Cr^' at higher pH, as observed experimentally (Figure 4-5 ~ 4-8).
136 O ^ C r VI - o + -O Crv:----OH
O O - C r ^ + H+(aq) HO
C r ^ -G
+ H , 0
O - C r V i
■OH
Figure 4-10. Adsorption and release equilibrium
137 Such surface adsorption behavior can be represented mathematically similar to a
Langmuir isotherm [171]. If we consider the Cr”’ oxide matrix in the CCC or mixed oxide to have a finite surface area for adsorption with a saturation coverage of adsorbed
Cr'^’ to equal to Fs, then
= /3[Cr'" ][//* ] = Equation 4-2 Fs - ^yr where Fvi is the coverage of Cr'^’ on the Cr’” oxide in moles/cm", p is an equilibrium constant, and 6vi is the fractional coverage of Cr^’. Note that the area relevant to adsorption is the microscopic area of the Cr’” oxide matrix, not of the CCC itself.
Although this microscopic area is difficult to determine, we assume it to be proportional to the geometric CCC area or to the mass of synthetic mixed oxide. If we define Nvi as the moles of Cr^’ present in the CCC or mixed oxide solid, Nvi may be calculated after assuming a homogeneous distribution of Cr'^’ in the solid. For the CCC, Nvi equals the initial average Cr^’ concentration (moles/ geometric- cm^) times the CCC geometric area.
For the mixed oxide, Nvi equals the initial Cr^’ loading (moles/gram) times the weight of mixed oxide.
Under the postulate that Langmuir adsorption and Equation 4-2 control Cr''’ (aq) concentration of a solution in equilibrium with a CCC or mixed oxide, predictions of the behavior of [Cr'^’jcquii for several situations may be derived. When Cr'^’ (aq) is desorbed from an initially saturated CCC or mixed oxide, the [Cr^’jequii follows Equation 4-3, with
V representing solution volume (see section 4.6 of this chapter for derivatization).
- I + J l + 4>3[//"] [C/- U „,v=------2^0[/F] Equation 4-3
138 The observed pH at the end of release experiments was 4.6 ± 0.7. A fit of Equation 4-3
to the equilibrium values of [Cr'^‘] listed in Table 4-1 is shown in Figure 4-11. Results
for the CCC on AA2024, as well as for the synthetic Cr”‘-Cr'^' mixed oxide are shown,
along with curves predicted fi'om Equation 4-3 for three values o f the product P[H^]. As
noted earlier, the model assumes that Cr^' permeates the CCC and is able to be adsorbed
to CCC’s microscopic area. This permeation process certainly affects the release kinetics, but Equation 4-4 requires only that the entire CCC or mixed oxide has equilibrated with the solution. The release kinetics will be discussed in a separate publication [172].
The adequate fit of the experimental data to Equation 4-3 with P[H^] = 10^ supports the notion that a Langmuir adsorption process is the equilibrium that controls
[Cr'* ']equii- One cannot rigorously exclude other mechanisms, but at least the results are consistent with “Langmuir” adsorption, and not with the “solubility” or “depletion” models. Based on the solution pH observed for Cr'^' release from a CCC or mixed oxide
(4.6 ± 0.7), the fit shown in Figure 4-11 corresponds to a P value o f 10^ - 10^. Several observations about Equation 4-3 and Figure 4-11 deserve note. First, the agreement between the results from the CCC 2024-T3 alloy and the synthetic mixed oxide implies that the same chemistry controls the release process. As stated previously, we believe the
CCC consists of a porous Cr'”-Cr'^' mixed oxide that can slowly release Cr^' into solution. Second, there is a finite number of binding sites for Cr^' on the insoluble Cr'" oxide matrix of a CCC, and the sites can be saturated. The high Cr^' concentration and low pH present during CCC formation promotes Cr^' adsorption and favors saturation of the matrix. Third, Cr'^’ binding is strongly pH dependent, as is consistent with
139 Y = o CCC on 2024-T3 ▼ Cr'"-Cr'''-mixed-oxide
2.0 -
1.5
I ‘ ê ^ 0.5
0.0
— I------1------1------1------1------1— 0.0 0.5 1.0 1.5 2.0 2.5 Nv,/V (mM)
Figure 4-11. [Cr'''],-,^^, of releasing experiment and the fitting lines from Equation 4-3 Figure 4-6. This dependence is a simple and predictable consequence of Reaction 4-1, in which high [H^] favors Cr'^‘ binding. Fourth, the equilibrium [Cr’^'] has an unusual square-root dependence on Nvi (and therefore on the area/volume ratio of CCC). Unlike a solubility mechanism, in which the solution reaches a saturation level, the [Cr^'j predicted from Equation 4-3 increases with (area/volume)^ without boimd. In reality, an upper limit of [Cr^'] will occur at saturation when a CrzO?^ salt or CtO] precipitate, but this level is much higher than those observed during release experiments, such as > 0.1
M. In chemical terms, the square root dependence on area/volume stems from the fact that the activity of adsorbed Cr^' changes as release occurs, while the activity of a slightly soluble solid, such as SrCr 0 4 , remains constant. Unlike the case for SrCrOa solubility, release of Cr^' from the CCC affects both the Cr'^' activity in solution and adsorbed on the Cr”’ matrix.
Solution of Equation 4-2 for the case of an initial quantity of solid Cr”’ hydroxide exposed to Cr^’ solution yields Equation 4-4.
E q u a t i o n 4.4
z = l + p[H*] -P[H ^][C r where Nvi is the moles of Cr^’ adsorbed to the Cr’” hydroxide at saturation, and [Cr^ is the initial [Cr'^’] in solution. Under the pH-cycling experimental conditions, both
Equation 4-3 and Equation 4-4 take the form as shown below:
Equation 4-5
141 The experimental results of Figure 4-6 are consistent with Equation 4-5, and a P value of
10*’, as shown in Figure 4-12. Notice that, in order to fit the data to Equation 4-5,
[Cr''']t=o =l6mM, instead of 19mM has to be used. In other words, - 15% of the Cr'^‘ was
somehow "trapped" inside the Cr(OH) 3 , and could not be released even at high pH 13.
The possible reason for this will be discussed in Chapter 5. For the results listed in Table
4-2, there was significant variation in the equilibrium pH for different initial amounts of
Cr"' hydroxide. The fairly wide spread of final pH values (about 3 pH units) prevented a reliable fit of the results to Equation 4-4. However, the plots shown in Figure 4-5(C) demonstrated a Langmuir-like surface adsorption behavior, which is also affected by the solution pH.
Although it may appear to be different fi-om a conventional adsorption process on flat surfaces, such as Pt, the Cr^' adsorption-desorption equilibrium satisfies all the criteria for a Langmuir isotherm. First, the Langmuir model requires the total niunber of surface sites to be constant [171]. Cr"‘-hydroxide is not flat, but the adsorption on the surface of a porous solid is essentially the same as that on an ideally flat surface. Of course, the aging of Cr'"-hydroxide, which is essentially a series of polymerization reactions, will eventually change the number of sites. But the results from the pH-cycling experiments suggest that, under the current experimental conditions, aging does not affect the general quantitative results significantly. In other words, the total number of sites is approximately a constant imder the conditions described in this chapter. Another criterion says that Langmuir isotherm prevails only when the adsorbate molecules on the siuface do not interact with each other. Further studies, shown in Chapter 5, prove that there is only one type of siuTace Cr^' species adsorbed on Cr"'-hydroxide, regardless of the
142 0.0251
0.020 initial Cr''' concentration
0.015-
I 0.010
■u w b 0.005 Experimental data p=10’-'o 0.000
------1------r ■ I------1------r—"■ >------1 0 2 4 6 8 10 12 14 pH
Figure 4-12. pH cycling of Cr'" + Cr''' mixture and the fitting result from Equation 4-5 surface concentration of the Cr^'. This indicates that there is no adsorbate-adsorbate interaction on Cr"‘-Cr'^‘ mixed oxide surface. The only difference between the current model and the conventional Langmuir model is the involvement of in the adsorption reaction. At a certain pH, the mathematical relation between the amount of adsorbent and that of adsorbate is exactly the same as a classical Langmuir isotherm seen on a textbook
(see Equation 4-6).
The current research focuses on the Cr(OH) 3-Cr'^' system, but such adsorption- desorption behavior is not unique. Adsorption of Cr^' anions onto various hydroxide- oxides has been observed, including Fe“'-hydroxide [173-179], Fe"’-Bi"'-mixed- hydroxide[ 180-181], Ni(OH )2 [182-183], Mg(OH )2 [183], Fe'"-Cr"'-mixed- hydroxide[185], Sn02[185-186], Zn-Al-Cl layered double hydroxide [187], ZnO[188],
Nb(OH)$[189], Z 1O 2 [190], etc. Similar behavior was also observed on the surfaces of steel [191] and hydrous concrete particles [192], which is mainly hydrous iron-silicate. In spite of the differences in quantitative and structural details of those “adsorption” phenomena, the general trends are quite similar to what was observed here. Higher pH always favors the desorption [177, 186, 188, 190, 192], and higher [Cr'^‘]t=o results in more Cr^'(aq) being adsorbed onto the solid [176, 190, 192]. Although the authors of these research publications did not provide any mathematical models for their results, their results appear to be very similar to those shown in Figure 4-5, 4-6, 4-8, on a qualitative level.
144 4. S Implications to Corrosion Protection
The current results provide additional support for previous conclusions about
CCC chemistry, and add some new insights into chromate storage and release. The
“Langmuir” adsorption model is completely consistent with Cr^' storage in a CCC as a
Cr‘"-Cr'^' mixed oxide, according to Reaction 4-1. As noted earlier, the pH dependence of
Cr'* ' binding, apparent in Figure 4-6, favors Cr^' binding during CCC formation (low pH and high [Cr^']) and Cr^' release in a defect in the field (low [Cr'^’], neutral pH). The storage and release of Cr'^' is likely to be critical to the ability of chromate coatings to passivate defects, scratches, etc. While “self healing” may also result from Cr'^' released by the SrCrOa containing primer, the two processes are controlled by different equilibria.
The finite solubility of SrCr04 (~ 5 mM) implies that SrCr 0 4 will dissolve until this limit is reached, perhaps leading to complete dissolution and depletion of SrCr 0 4 . The
“Langmuir” adsorption operative in the CCC does not result in a constant [Cr'^']equ,i, but rather one that decreases as the CCC film is depleted. Although the released [Cr^']equ,i decreases as the CCC is depleted, an equilibrium can be maintained at very low levels of both adsorbed and solution Cr^'. Depending upon the [Cr^'] level required for self- healing, and the relative amounts of Cr'^' in CCC and primer, the CCC may outlast a
SrCr0 4 primer, since the CCC will release less Cr'^‘ as it becomes depleted.
The effect of ionic strength on Cr^' release is attributable to an activity effect.
The activity coefficient calculated for HCr 0 4 in 0.1 M NaCl quantitatively accounts for the higher equilibrium [Cr'^’jcquii in NaCl, at least to the accuracy expected from the extended Debye-Huckel equation [165]. The current results do not take Cr'^’ spéciation into account, except for the spectophotometric analysis described in Figures 4-1 and 4-2.
145 The distribution of solution Cr'^' between HCr 0 4 , CrzO?^, Cr04 , and even H2Cr0 4 may
affect the equilibrium constant p. The dominant Cr'^‘ species in the conditions used here
is HCr04 , but the adsorption equilibriums may vary above pH 6.5 or below pH 2. The
value of P and the nature of the adsorbed Cr'^‘ may vary under these conditions, but
“Langmuir” adsorption is still expected to prevail.
An additional important issue CCC’s involves aging after CCC formation. Aging of both the CCC on AA 2024-T3 and the Cr'”-Cr''' mixed oxide for periods from 0.25 to
210 hours at room temperature decreased the equilibrium [Cr^']cquii released into water
(Table 4-1). However, the decrease in [Cr'^’jcquii release was not large; about 25% for the
CCC and 40% for Cr'"-Cr^' mixed oxide. A much greater decrease in [Cr^'jequii occurred after the CCC was heated at 50°C in air for 1 hour. These decreases could be the result of a reduction in CCC hydration, or of a structural rearrangement that tightly binds Cr^', and more discussion about this issue will be provided in Chapter 5.
In conclusion, the adsorption and release of Cr'^’ from a CCC is consistent with
“Langmuir” adsorption of Cr 0 4 ^’, HCrC>4 or CrzO?^ to a porous, insoluble Cr'" hydroxide matrix. The process is reversible and pH dependent, with higher pH favoring desorption of Cr' The low pH and relatively high [Cr'^'j in a CCC formation path favor adsorption o f Cr'^' to the Cr'" matrix formed by reduction of Cr'^' to the Cr'" by the alloy.
The observations are quantitatively consistent with a “Langmuir” adsorption-desorption equilibrium with a equilibrium constant o f 10^-10’ for the Cr'^' interaction with Cr'" hydroxide.
Another implication of the current results is that similar Cr'^' adsorption- desorption behavior also could happen on other metal or alloy surfaces. The
146 corresponding mathematical models and structure details of adsorbed species could vary according to the chemical nature of the metal, but they should share some common trends semi-quantitatively, such as the dependence on pH and [Cr'^']t=o-
4.6 Details of Adsorption-Reiease Model
4.6.1 Release of Cr'^' from CCC
Starting with Reaction 4-1 from the above text,
^ads Cr"'-OH (solid) + HCr 0 4 (aq) + Cr'"- 0 -Cr'^‘0 3 H (solid) + HzO Reaction 4-1 ^desorb
F.quil
C one 1-0VI [Cr'^'jequil. [H^lequil. 0VI 1
The equilibrium constant of this reaction, (3, can be expressed as Equation 4-6.
= P Equation 4-6 (1 - e,, )[Cr'" l^desort
k^ads. p = Ir desofb.
6 is fractional coverage, which can be calculated from equilibrium surface concentration , r vu (moles/cm"), and saturated surface concentration, Fs , (moles/cm").
r 0VI = —^ Equation 4-7 ^5 rVI equals the total moles of Cr^' per imit siuTace area. Substituting Equation 4-7 into
Equation 4-6, and rearranging resulting in Equation 4-2.
147 9 . p-frr- = Equation 4-2 1 5 1 1
Note that Fs is also the initial surface concentration of Cr'^' in the CCC, since the CCC
film is initially saturated with Cr'^’. Because there was no Cr'^* in solution initially, at
equilibrium the total moles of Cr^' in solution equals to the moles of Cr^' had been
released from the CCC:
Equation 4-8 which can be rearranged to.
1 _ (E j )S iequil. ~ jr Equation 4-9 and.
Equation 4-10
S is the microscopic siu*face area of porous CCC film, or synthetic powder samples. Since the released Cr'^* enters the solution as soluble Cr^'. Substitution of Fs and Fvi from
Equation 4-10 into Equation 4-2 yields:
: . [ C r Sr, rC r"l '+ - ______: ^ = 0 Equation 4-11
Application of the quadratic formula and assignment of FsS as Nvi, the initial moles of
Cr'^' in the whole CCC, yields Equation 4-3 in the main text.
148 iCf- ]ew. = ------Equation 4-3 iequil.
4.6.2 Release of Cr^' from mixed oxide:
Solution of Equation 4-2 for the Cr“‘-Cr''* mixed oxide also yields Equation 4-3.
Nvi has the same meaning, the total moles of Cr^' initially in the powder sample, and can be calculated as the weight of oxide times the initial Cr^' concentration in the oxide in terms of moles/gram.
4.6.3 Adsorption of Cr ' to synthetic Cr"' hydroxide.
The area available for adsorption on the Cr"' hydroxide matrix is unknown, but is assumed to be proportional to the mass of Cr'" hydroxide, m, by some constant, k, so that
S = km. Starting with an initial Cr^' in solution of [Cr'^']t=o, the equilibrium coverage of
Cr'" is:
Equation 4-12 mk
Substitution of Fvi from Equation 4-12 into Equation 4-2 yields:
The total moles of adsorbed Cr^', Nvi, is:
Ny, =Fgmk Equation 4-14
149 Substitution of Equation 4-13 into 4-14 and solving for [Cr ] yields Equation 4-4 in the main text.
iequil.
V y y
4.7 Conclusions
The reversible adsorption-desorption reactions between Cr'^’ and Cr‘"-hydroxide were studied quantitatively by using UV-Vis spectrophotometry. The adsorption is favored by low pH, high initial [Cr^'] concentration, or large amoimt of Cr"'-hydroxide.
A mathematical model, which is consistent with Cr"' -O-Cr^' structural model, was established based on the quantitative results. The model explains the dependence of
[Cr^']equii. on the experimental conditions (pH, [Cr'^‘]t=o, mass of Cr"‘-hydroxide), and fits experimental data adequately.
150 CHAPTERS
EVOLUTION OF STRUCTURE AND PROPERTIES
OF CCC-CHROMIUM(VI) SPECIES DURING AGING
5.1 Introduction
In previous chapters, the reversible interaction between Cr'^' and Cr“'-hydroxide was discussed. The spectroscopically invariant behavior of CCC-Cr^' species at different pH and dehydration conditions revealed the Cr‘”-0-Cr'^* bonding structure. However, the specific structure of the Cr"-O-Cr^' species is still not clear. Meanwhile, some spectral variations after thermal aging were also not well understood. Quantitative studies had indicated that the heated-CCC samples release much less Cr'^’ than air-aged ones. Since
Cr'^' release is essential for “self-healing”, the above property change appears to be undesirable for protection. An explanation for this would be important for designing non toxic coatings, such as coatings on metallic appliances under constant thermal radiation.
151 Another important issue related to “self-healing” is the possible structure-property variation during the Cr^' release process. Based on the previously described “Langmuir” model, the surface loading level o f Cr'^‘ in CCC film decreased when Cr'^' leached.
Therefore, the properties of Cr*"-hydroxide at various Cr^' loading levels become very important. Kendig et al. proposed that Cr'^' protects by adsorbing to Al"'-hydroxide surface, thus changing the surface charge and “discouraging” chloride adsorption and permeation [193]. The proposed charge effect could happen on both Cr'"-hydroxide and
Al‘"-hydroxide. In both cases, the knowledge of the specific structure of the surface Cr'" species is crucial in determining the surface charge distribution.
The goals of the studies described in the current chapter were to find out the structure variations of the CCC film in several different aging environments, including water-aging and thermal aging, and to test the protection properties of Cr‘“-hydroxide and
Al’"-hydroxide at different loading levels. In order to determine the specific structures of
Cr^’ species, either in regular CCC or in thermally treated CCC, isotopic substitution and ab initio calculations were carried out.
5.2 Experimental
5.2.1 Isotopic substitution
87.7% HzO'^(10% H iO ') was purchased from Isotec, Inc. (Miamiaburg, OH), and 99.9 % D 2O from Aldrich (Milwaukee, WI). (For simplicity, O'® will be referred to as O, and O'^ as O in this Chapter.) CriOi^' and Cr0 4 “' solutions were prepared by dissolving ~ 0.3 g Cr(N 0 3 )3«9H20, 8 mg K2Cr20?, or 8 mg K2Cr 0 4 in 1 ml H2O , respectively. Cr(OH )3 was prepared by dissolving I g Cr(N 0 3 )3«9H20 in 3 ml H2O, and
152 adding high concentration NaOH (—20 M) to pH - 6. The precipitate was collected and washed with —2 ml of regular nanopure water. CrfOD)] and aqueous Cr^' in D 2O (CriO?"
, Cr0 4 ^', and HCrO^l were prepared in similar ways by replacing H 2 O with D2O.
The Kaiser spectrometer, as described in Chapter 2, was used to collect the
Raman spectra of the above samples.
5.2.2 Synthesis of Cr'"-Cr'^' mixed-oxides and spectroscopic studies
In order to specify the structure of Cr'^'-species in the CGC film, the Cr'^' loading level on Cr"-hydroxide and aqueous Cr'^' equilibrium must be considered separately. The aqueous Cr^' equilibrium is controlled by both [Cr'^‘] and pH. The following experiments in section 5.2.2.1 were conducted to obtain Cr“'-Cr'^'-mixed-oxides with constant Cr^' loading level but under varied [Cr'^']t=o, which changes the solution Cr^' species distribution significantly. Section S.2.2.2 was aimed at synthesizing Cr'"-Cr^'-mixed- oxides with different loading levels.
5.2.2.1 Cr"'-Cr^' mixed-oxides at constant loading level, but different [Cr^']t=«
Cr(OH)3 was synthesized separately, by adding 20 M NaOH into 15 ml of 0.2 M
Cr(N0 3 )3.9H20 solution to pH -6. The precipitate was collected, washed, and used within one hour after filtration.
One fourth of the synthesized Cr(OH )3 was added to four different Cr^' solutions:
500 ml of 1.67 mM Cr^', 50 ml o f 16.7 mM Cr'^', 20 ml o f 42 mM Cr^ ', or 5 ml of 0.167 mM Cr^' solution. The pH of the Cr^^ solutions was adjusted to around 4.5 before adding the Cr(OH)3 and the Cr'” : Cr'^’ ratio in the suspension was approximately 3:1. The
153 suspensions were stirred and pH fine-tuned to ~ 4.5 before collecting the solid. The solids were stored in closed glass vials and their Raman spectra were collected wath the Kaiser spectrometer within 24 hours.
The synthesis method used here is different from that described in Chapter 2, due to the low initial concentration of Cr'". For example, when [Cr'^‘]t=o was 1.67 mM and
[Cr'^‘]t=o:[Cr‘"]t=o =1:3, the initial concentration of Cr"' in solution was 5 mM, which is too low to obtain reasonable amount of Cr'"-hydroxide at pH 4—6. Therefore, Cr'"- hydroxide was synthesized separately, from much higher [Cr'"]t=o.
S.2.2.2 Cr'"-Cr'^' mixed-oxides at various loading
Cr'"-Cr^' mixed-oxides with various Cr^' loading were prepared by adding NaOH to a Cr(N 0 3 )3 . 9 H2 0 -K 2Cr2 0 7 solution mixture imtil a significant amount of precipitate appeared, which occurred in the pH range 4—6. The initial molar ratio of Cr'":Cr^' was controlled at 90:1, 29:1, 20:1, 2.9:1, and 0.3:1, respectively. The Cr'" : Cr^' ratio in the resulting precipitate was measured by UV-VIS spectrophotometry, as described in chapter 2. Raman spectra of these powder samples were studied with both the Kaiser spectrometer and the Bruker FT-Raman system.
5.2.3 Aging treatment of CGC and Cr“'-Cr''' mixed-oxide
The CCC film and Cr'"-Cr^' mixed-oxide samples for aging studies were all prepared as described in Chapter 2. AA2024-T3 coupons were polished in water with a succession of 240, 400, 600, 800 and 1200 grit sandpaper. After drying in air for approximately 24 hours, the samples were immersed in room temperature Alodine 1200S
154 solution for 60 seconds, then rinsed thoroughly with more than 100 ml running nanopure water before any further treatment. C/"-Cr^' mixed-oxide was prepared by adding NaOH into Cr(N 0 3 )]-OHzO-KzCrzO? solution mixture to pH ~4, The precipitate was collected by filtration, and washed with nanopure water.
The aging processes of CCC film and Cr'"-Cr^' mixed-oxide in water were studied by immersing the samples in water. 112.5cm^ CCC coated AA 2024-T3 or 0.1 g
Cr'"-Cr'^' mixed-oxide was immersed in 150 ml nanopure water. The solution was changed with fresh nanopure water daily to speed up the depletion of Cr'^' from the samples. Raman spectra of the film or solid were collected during the whole experiment, which took about one week.
Heat treatment o f CCC film or Cr^'-Cr'"’ mixed-oxide was conducted as described in chapter 2, by heating in air in an oven at constant temperature for approximately 1 hour.
5.2.4 Protection from several ‘^artificial” films or treatments
Electrochemical studies were focused on Rp measurement of CCC coated or bare
AA2024-T3 after several types of treatments. All AA 2024-T3 samples used for electrochemical studies were Ixlcm^, and mounted in epoxy. “Depleted” CCC film was obtained after prolonged exposure (~ one week) of CCC film (Alodine 1 min) to water, with daily changing of the water. Cr^'-immersed AA2024-T3 was prepared by immersing
AA2024-T3 in ~ 50 ml of 0.5 mM Cr^' solution (pH ~5) in a covered container. “SOj"' - treated CCC” was prepared by covering a CCC film with around 0.5 ml of 0.1 M NazSO] solution, and the sample was stored in a tightly closed weighing bottle for about 24 hours.
155 The surface was then rinsed with nanopure water before any test or further treatment.
Cr'“-immersed AA2024-T3 was prepared by immersing AA2024-T3 in 0.1 mM Cr"' solution (pH ~7). The “artificial scratch” setup was used to test the “self-healing” property of the above samples by “sandwiching” them with untreated AA2024-T3 samples, with about 0.5 ml 0.1 M NaCl in between. All the Rp values were measured in
0.1M NaCl solutions.
5.3 Quantum Calculations
All the theoretical studies were conducted by using Gaussian-98 (Gaussian, Inc.,
Carnegie, PA). Hartree-Fock Self-Consistent Field (HF-SCF) method and 3-21G basis set were chosen for all the calculations shown in this chapter. For all the clusters, Cr04“',
CriO?", HCr0 4 , and the hypothetical clusters, three steps were followed to obtain calculated Raman vibration frequencies. First, the geometry parameter, such as bond- lengths and bond-angles, were optimized. Then the vibrational frequencies of the optimized clusters were calculated. Finally, scale factors were used to adjust the calculated frequencies to be closer to experimental values [194]. For example, the strongest experimental Raman band for Cr 0 4 ^* is at 846 cm ', while the calculated frequency for the same vibration of Cr 0 4 ^' was 1044 cm '. A scale factor of 0.81 has to be applied, by multiplying all the calculated frequencies of Cr 0 4 ^' by 0.81. The calculated frequencies, after correction by the scale factor, are shown in Table 5-1. For different clusters, the scale factors were slightly different, 0.810 for Cr 0 4 ^, 0.774 for HCr 0 4 , and
0.826 for CrzO?^ The average of the three factors, 0.803, were used to correct the calculated frequencies of all the hypothetical clusters shown in Table 5-2.
156 Experimental Raman Shift (aq)‘* Calculated Raman Shift Experimental Literature (cm ') (cm ') Raman Shift Raman Shift compounds - (cm ') (cm ') (aq) Q 1 6 q I6 q 18 o'" D2O (solid)*' [195-196]
Cl04^ 347 328 332 332 313 350 348 (Td) 373 373 358 341 389,396 368 846 796 840 846 798 853 847-848 885 852 878 865 833 869,880,906 884-886
246,273 235,260 HC1O4 310,344, 350 296, 327, 333 (Cs) - - 592,676 568,670 -- LA 898 [199] 898 853 931,940 896,907 3006 2996
215,239,269 205,229,256 220,234 217 364 349 373 314, 372,499 296,363,473 335,365,373, 367 CrzO?^ 387 (C2v) 556 537 628,658 604,637 565, 570 557 813 774 739,749, 892 776 902 860 899 904 853 911 904 936 906 939 981,999 938,957 925,945,952, 943-942 961 a. Scale factor = 0.810
Table 5-1. Comparison between experimental and calculated vibration frequencies of standard compounds (Ail bold numbers arc from the strongest bands) Calculated Experimental hypothetical Raman Shift for Chemical systems Raman Shift structures Cr^-O modes (cm ') (cm ') “
188, 282 379 Structure A 759 in Figure 5-3 888 365 (360 ~ 369) 978 Cr'"-Cr^' mixed- 449 oxide 535-527 235 858 Structure B 342, 465 945 in Figure 5-3 700 809 963 1008 Cr'"-Cr'’ ' mixed- 360 oxide prepared 444 350 in D2O 855 Structure C 460 923 in Figure 5-3 773 846 980
336 Structure D 493 in Figure 5-3 695 362 878 CCC 457 993 530 858 297 Structure E 943 442, 552 in Figure 5-3 753 835, 985 a. scale factor = 0.803
Table 5-2. Experimental frequencies of Cr'"^' species in Cr-mixed-oxide and CCC film. and the calculated frequencies of hypothetical Cr'"^' clusters (All bold numbers are from the strongest bands)
158 5.4 Results
5.4.1 Isotopic substitution
The replacement of O by O results in significant shift of the Cr-0 vibration frequencies toward lower energy. Figures 5-1 and 5-2 show the Raman spectra of various standard Cr'^'-, and Cr*”-compounds before and after O or D substitution. The
Cr^ '-O vibration of CrzO?^ solution shifted downward by more than 40 cm * due to O , and such replacement reaction took less than 10 minutes. However, dissolving K 2Cr0 4 in
H2O did not show any "isotope effect" initially. The 40 cm ' shift for Cr 0 4 ^’was observed after 6 days. For most of the Cr'^'-O vibration modes, there was no significant frequency shift when H was replaced by D, except for the slight shift of Cr20?' from 902cm ' (in
HiO) to 900cm ' (in D 2O), which might be an indication of HCr 0 4 [197-199]. Cr'"- species behave very similarly to Cr^' ones in terms of O or D substitution. Solid Cr'"- hydroxide showed one weak and broad peak at 527cm ', and the frequency shifted significantly after O-substitution, but not D-substitution.
The main purpose of the above isotopic-substitution experiments is to provide experimental frequencies of standard compounds. These frequencies will be used to verifv the calculation method.
159 00 vD
in HjO
00 O y 30mM in HjO'» (pH < 3)
in D^O
in HjO
in HjO'* in D2O J 400 600 800 1000 1200 1400 Raman Shift (cm *) Figure 5-1. Raman spectra of Cr^' solutions in HjO or 87.7% H20'^ 9 CJ g Cr(OH) 3 solid Cr(0 ‘*H )3 solid Cr(OD)3 ^00 500 600 TOO 800 900 Raman Shift (cm ') Figure 5-2. Raman spectra of Cr*"-hydroxide prepared in HjO, 87.7% or DjO 161 5.4.2 Quantum calculation results The calculated Raman frequencies of several standard Cr'^'-containing compounds are listed in Table 5-1. For all the three Cr'^'-O standard compounds, the calculated results are consistent with the experimental values and the literature values. The correlation implies that the calculation can predict frequencies from new materials with reasonable accuracy. Therefore, the same basis set and methods were used in the following calculations. The vibrational frequencies of several hypothetical Cr'"-0-Cr^' type clusters were compared with those of CCC film or synthetic Cr*“-Cr'^‘ species. The optimized structures of these clusters are shown in Figure 5-3, and the calculated Raman spectra are shown in Figure 5-4. The calculated frequencies of these clusters are listed in Table 5-2. From experimental observations, the signals from Cr‘"-OH are much weaker than those of Cr"-O-Cr^'. For simplicity, and for better comparison with experimental spectra, only those vibration modes that involve Cr^-O and Cr'"-0-Cr^' are shown in Figure 5-4 and Table 5-2. All of the calculated spectra were plotted according to Lorenzian function [200], which is shown below. In the above equation, Avl is the full width at half maximum (FWHM), and v^ is the frequency of peak maximum. FWHM =100 was used in the plots shown in Figure 5-4. 1 6 2 B • Cr^i , Cr"' o O Figure 5-3 Optimized hypothetical Cr"'-0-Cr^' clusters (lines with open end represent O-H) 163 Raman Shift of synthetic _ Cr‘"-Cr''‘ mixed-oxide II A B C D Synthetic C r " '.C r V mixed-oxide CCC film 200 400 600 800 1000 1200 1400 Raman Shift (cm ') Figure 5-4. Comparison of calculated Raman spectra with experimental spectrum of CCC and Cr'^-Cr'^' mixed-oxide (A-E are the calculated Raman spectra of the five hypothetical structures showed in Figure 5-3) 164 In general, all the hypothetical Cr"^-Cr^' clusters showed their strongest Raman peaks at 800-1000cm’', which correspond to Cr"-O-Cr^' stretching, and weak peaks at 350-500cm ' from Cr"-O-Cr^' wagging vibrations. All these are semi-quantitatively consistent with the experimental spectra of CCC and Cr‘"-Cr'^' mixed-oxide. In addition, a general trend can be observed regarding the vibrational frequencies of adsorbed dichromate and chromate. The adsorbed dichromate has higher Raman shift than the adsorbed mono-chromates. 5.4.3 A stepwise study of the structural evolution during adsorption and release The CCC-Cr^' adsorption-release is a dynamic process. Based on the Langmuir equilibrium model proposed in Chapter 4, the surface Cr^' loading level, the solution [Cr'^'j, and the solution pH regulate each other until the whole system reaches the equilibrium. For example, when a CCC film starts to release Cr'^‘ into surrounding water, the surface loading decreases, the solution [Cr^'] increases, and pH of the solution decreases. Meanwhile, the Cr'^'(aq) can be adsorbed by the solid again. O f course, before reaching equilibrium, the released amount must be larger than the adsorbed amount. During the whole process, the [Cr'^'(aq)] and pH are constantly changing, as well as the aqueous Cr^' distribution in solution. In order to achieve a thorough understanding of “self-healing”, the possible structure and property evolution of CCC-Cr^' species and Cr'"-Cr^' mixed-oxide during such a dynamic process must be understood. In the current study, the possible effects of surface loading level, solution [Cr'^‘], and pH were studied separately. First, Cr"‘-Cr'^’ mixed-oxides with different Cr^' loading levels were prepared as described in Chapter 2, by keeping [Cr"'] at 0.2M, while varying [Cr'^'j from 0.0022M to 0.67M. The collection pH was approximately 4-5 for samples 165 with medium-high Cr^' loading levels (a, b in Figure 5-5), and 5-6 for the low Cr^' loading levels (c-e in Figure 5-5). The Raman spectra o f the resulting solids (see Figure 5-5) showed no significant spectral variation, except for the relative peak intensities. Ab initio calculations confirmed that the Raman bands at 858 cm ', 445-450 cm ' and -365 cm ' were all from the vibrations of Cr'^'-oxy ligands on Cr"‘-hydroxides. As the initial Cr'" : Cr^' ratio in solution was decreased, the 858 cm ', 445-450 cm ' and -365 cm ' Raman bands of the solids increased in magnitude imtil the Cr'" : Cr^' ratio in the solid was 3:1. For lower solution Cr'" : Cr'^' ratios, no further increase in band intensity from the solids was observed. These results imply that the solid became saturated with Cr^' when the solution ratio of Cr"' to Cr^' was 3 or less. Notice that none of the Cr'"-Cr'^' mixed-oxide samples showed bands in the 200-300cm ' region, where a band could be expected for a Cr^'-O-Cr^' bridging mode. The possible effect of pH on the structure of Cr'"-0-Cr^' species was discussed in Chapter 2. The results shown in Chapter 2 indicated that Cr'"-Cr^' mixed-oxides, prepared under different pH and a constant initial [Cr'"] : [Cr'^'j ratio of 3:1, had similar Raman bands at 858 cm '. The initial Cr^' concentration for those samples was [Cr'^']t=o = 68 mM and the pH ranged from 4 to 9. In the current study, a group of Cr'"-Cr'^' mixed-oxides were synthesized under different [Cr'^']t=o,but at saturation loading level and pH - 4.5. The [Cr'^']t=o varied from 1.67 mM to 0.167 M. The Raman spectra of the resulting solids (see Figure 5-6) are very similar to those obtained imder [Cr^']t=o = 68 mM and different pH. Therefore, the Cr'"-0-Cr^' species retained a particular structure under different pH, different initial [Cr^ '], and different surface Cr^' loading levels. 166 il»00 Initial [Cr"']:[Cr Cr'": Cr^' l» ^00 lO in solution in precipitate 0000 (average of three trials) 0.3:1 (2.2~4.0):l 0.2M: 0.67M 2.9:1 0.2M : 0.069M (2.5~3.5):l 20:1 0.2M: 0.0lOM (27-31): I 29:1 0.2M :0.0069M (34-44):! -|— I—I—r- 140 180 220260 ,------r-.----- ,------,------,------90:1 200 400 600 800 1000 1200 1400 0.2M : 0.0022M 81.7:1 Raman Shift (cm ') Figure 5-5. Raman spectra of Cr'"-Cr''' mixed-oxide at various loading level I All spectra were collected under similar conditions, and are shown on the same scale) oo m oo 400 600 800 lobo 1200 1400 1600 Raman Shift (cm ') Figure 5-6. Raman spectra of Cr"'-Cr^ mixed-oxides prepared under various [Cr^]t=Q, but same Cr^ loading level and pH (All spectra were collected under similar conditions, and are shown on the same scale) 168 A separate “structure evolution” study was conducted on regular Cr"'-Cr'^’ mixed- oxide and CCC film along the Cr^'-release process. The film or the powder was immersed in nanopure water and the water was changed daily. Several Raman spectra of the sample surface were collected every day. Deconvolution of the spectra revealed two Raman bands, one at 858 cm'*, the other at ~ 943 cm *. Plotting 858 cm * band intensity versus the intensity of 943 cm * band. Figures 5-7 and 5-8 are obtained. In general, the intensities of both peaks decreased upon Cr^* release, and regression of the data points yields a linear correlation coefficient of 0.96 for CCC, and 0.88 for Cr***-Cr^* mixed- oxide. In other words, intensities of the two peaks decreased colinearly when either CCC film or Cr*** -Cr'^* mixed-oxide were losing Cr'^‘. This implies that the two peaks are from the same species, and the structure of the species does not change during release. All the above results reveal that the Cr'^* species retains a particular structure despite the widely varied precipitation conditions. Some simple calculations based on the equilibrium constants firom reference 73 may be helpful in finding the common thread of those conditions. For a solution with total Cr'^* concentration of 68mM, at pH<5, 80% of the Cr^ * exists as CrzO?^ and 20% as HCr04 Transition to Cr 0 4 ^' happens between pH 5-7, and more than 80% of the Cr^* is Cr 0 4 ^ at pH 7. At pH 4.5, when the [Cr'^*]totai was increased from 0.00IM to 0.167M, the solution distribution coefficient of Cr^O-" increased from 19% to 77%, while that of HC 1O 4' dropped fi'om 78% to 13%. If we also consider the concentration and pH during a typical CCC release experiment, the predominant aqueous Cr'^* species is HC 1O 4 In addition, it has been reported that the stability of monochromate-ligands are better than their dichromate analogues [197]. 169 1 8 0 0 ’ Y = 76.9 + 8.96X R = 0.959 1400 i 10001 00 600 ' oo 200 i 0 20 40 60 80 100 120 140 160 180 943cm ' band intensity (AU) Figure 5-7. Intensity variation of 858cm ' and shoulder (-943cm ') Raman peaks from CCC film during Cr^' release 170 _ 3500 S Y =112 + 10.42 X < 3000 ■ R = 0.875 = 2500 ■ ^ 2000 ■ 1500 ■ OO 1000 OO 500 • 0 50100 150 200 250 300 945cm ' band intensity Figure 5-8. Intensity variation of 858cm ' and shoulder peaks (-945cm ') from Cr'"-Cr'^ mixed-oxide during Cr^' release 171 These results imply that the Cr^' on CCC film is monomeric for the pH and concentration range considered here. A likely subunit structure is Cr^'-O-Cr'^'OaH. Another observation fiom Figure 5-5 and Figure 5-6 is the weak Raman band near 450 cm The calculated spectra also showed Cr"-O-Cr^' vibrations at a similar wavenumber, and the spectnun calculated for Cr"' 2 0 2 (0 H2)?(0 4 Cr^'H)^ (structure E in Figure 5-3) is very similar to the experimental spectnun of Cr'"-Cr'^‘ mixed-oxide. In summary, considerable evidences indicated that the predominant Cr"'-0-Cr^ ' species in CCC is similar to structure E in Figure 5-3. First, the synthetic Cr"'-Cr^' mixed oxides showed similar Raman spectral featmes despite the widely varied preparation conditions, which include [Cr'^'jpzo , pH, and Cr^' loading levels. In addition, the Raman spectra of CCC film and Cr"‘-Cr'^' mixed oxide during the release experiments were also very similar to those before release, except for the lower band intensities. Furthermore, no Raman band was observed near 218 cm ', which is a signature of Cr^'-O-Cr^' structure. All of this suggests that the Cr^' species in the CCC film is monomeric, and it most likely has a subunit with the formula of Cr"'- 0 -Cr^'0 3 H. The calculated Raman spectrum of Cr'" 2 0 2 (0 H2)7(0 4 Cr'^'H)^ (structme E in Figure 5-4) shows strong resemblance to the experimental Raman spectra of CCC and Cr'"-Cr^' mixed oxide. 5.4.4 Structural evolution during heating Heat treatment of CCC and Cr'"-Cr'^' was revisited. Deconvolution of the spectra from heated CCC or heated Cr'"-Cr^' revealed Raman shift and intensity of each band. The deconvolution was conducted by using "ciu-ve-fitting" ftmction of Grams32 software ("curvefit.ab"). The 858 cm ' peak remained at a fixed fi'equency despite the dehydration 172 caused by heating. However, detailed spectral analysis showed changes in peak intensity, peak intensity ratios, and the emergence of new weak Raman bands, as shown in Figure 5-9, and 5-10. Since the spectral conditions (laser power and integration time) were reproducible and the typical sampling depth is longer than the film thickness, the 858cm ' peak intensities can be used to determine the relative amoimt of Cr'"-0-Cr^' species semi-quantitatively. Several spectra from each sample were collected and their 858 cm ' peak intensities were plotted in Figures 5-11 and 5- 12. For CCC film on AA2024-T3 alloy, generated by immersion in Alodine solution for 1 minute, heating resulted in lower 858 cm ' peak intensity, as shown in Figure 5-11. The higher the temperature, the lower the 858 cm ' peak intensity. A new peak near 780 cm ' emerged after heating, as well as the peak near 550 cm ', which is presumably from CrzOs. The spectral variations of the heated Cr'"-Cr'^' powder are, in general, similar to CCC film, and are shown in Figure 5- 10. The 858 cm ' peak intensity decreased after heating, as shown in Figure 5-12, and new bands near 818-785 cm ' also showed up after high temperature treatment. Due to the broad feature of the Raman bands under study, PCR (Principle Component Regression) was used to extract more detailed information. The “plsiq.ab” program in Grams 32 was used. The details of the data sets and calculation parameters are listed in Table 5-3. The number of factors represents the number of independent spectral variations among all the spectra inside a data-set, and their scores indicate the direction of such variations. 173 Air-dried W 4^ ^ (V, Vi 0\ ÛJ I—* ^ so U> o 00 o *! 300 400 500 600 ! 400 600 800 1000 1200 1400 1600 Raman Shift (cm ') Figure 5-9. Raman spectra of heated CCC films u> 4^ L/l00 v£>4^ 4^ u> 00 w vO vO O I Air-dried i i * . ' ...... '• 00 U> 4^ Lf) ^ 5 K) \ 0 Ui LA •••• ,V.. :i hit' ..... 400 600 800 1000 1200 1400 1600 300 400 500 600 Raman Shift (cm ') Figure 5-10. Raman spectra of heated Cr'"-Cr^' mixed-oxides 3500 ° individual spectrum 3000 - average < 2500 - 1.5 2000 - 1 § >*«»■ ON as 1000 - 00 »r> 500 - 00 0 20 40 60 80 100 120 140 160 180 Heating temperature ("C) Figure 5-11. 858cm ’ Raman band intensity of CCC film after heating 3000 individual spectrum 2500- I average I 2000 1500 I g 1000- 500 I 00 0 20 40 60 80 100 120 140 160 180 Heating temperature ("C) Figure 5-12. 858cm ' Raman band intensity of Cr'"-Cr''' mixed-oxide aOer heating sample CCC film Cr'"-Cr^' mixed-oxide Total number of files in the data-set 18 23 Number o f files from air-aged samples 4 6 Number o f files from 50°C heated samples 4 5 Number o f files from 113°C heated samples 5 7 Number o f files from 170°C heated samples 5 5 Diagnostic type Self-prediction Cross-validation Data preparation Mean center Mean center Path length correction Normalized Normalized Baseline automatic automatic Spectral region 297-1716 cm ' 297-1716 cm ' number of factors at best fit 3 3 Table 5-3. PCR parameters used in the analysis o f the spectra from heated CCC and Cr”'-Cr'^‘ mixed-oxide 178 For example, for CCC film, the data-set includes spectra of a CCC film before and after heating. The overall analysis revealed three independent spectral features, so- called “loading factors”, as shown in Figure 5-13. Loading factor 1 is mainly the 858cm’’ Raman band, and its amplitude is 10 times larger than the other factors, therefore it’s the dominant feature. Its score, as shown in Figure 5-14 (A), indicates that the amplitude of this Raman feature decreased under elevated temperature. The loading factor 2 includes Raman features fi’om CrzOs (547 cm ') and the thermally induced new surface species (760 cm ' and 898 cm '). The corresponding variation of its score, as shown in Figure 5- 14 (B), indicates that those three Raman bands increased simultaneously during the heating process. Loading factor 3 probably is the fluorescence of the surface, and its amplitude (Figure 5-14 C ) was approximately constant during the whole heating experiment. The PCR results fi-om heated Cr'"-Cr^' mixed-oxide are slightly different than that of CCC. The spectra of the loading factors and their scores are shown in Figure 5-15 and Figure 5-16, respectively. The loading factor I and its score variation are very similar to those of CCC. Loading factor 2 showed thermally induced new peaks at - 780 cm ', -900 cm ', and -968 cm ', with increasing scores at elevated temperatures. Loading factor 3 is spectroscopically similar to loading factor 1, but with slightly different wavenumber. When the heated CCC film or Cr'"-Cr^' mixed-oxide was immersed in nanopure water, heated-CCC did not release significantly Cr'^', while heated Cr'"-Cr^' mixed-oxide samples released more than regular ones. The solution [Cr'^'j values after one week immersion are listed in Table 5-4. 179 o\00 o 0.0018 Loading factor 1 00VO 0.00022 00 ON Loading factor 2 00 o Loading factor 3 0.00022 T TT T 400 600 800 1000 1200 1400 1600 Raman Shift (cm ') Figure 5-13. PCR loading factors from Raman spectra of heated CCC 0.4 0.2 I 0.0 1 «2 8 -0.2 i GO -0.4 - 0.6 0 20 40 60 80 100 120 140 160 180 Temperature (°C) 0.6 0.4 0.2 o ë 0.0 - 0.2 CO! -0.4 - 0.6 0 20 40 60 80 100 120 140 160 180 Temperature (°C) Figure 5-14. Score variations of the loading factors for heated CCC (Dots are from PCR analysis, lines are the averages.) (continued on next page) 181 (continued) 0.6 0.4 0.2 w o 0.0 t s -0.2 c /3 -0.4 -0.6 0 20 40 60 80 100 120 140 160 180 Temperature (°C) Figure 5-14. Score variations of the loading factors for heated CCC (Dots are from PCR analysts, lines are the averages.) 182 LA00 00 0.0022 Loading factor 1 00 0.00016 u> 00 Loading factor 2 00 w00 0.00016 00 Loading factor 3 T TT TT T 400 600 800 1000 1200 1400 1600 Raman Shift (cm ') Figure 5-15. PCR loading factors from Raman spectra of heated Cr'"-Cr^'-mixed-oxides 0.6 0.4 1 y 0.2 1 I (% 0 0 - 0.2 ■ # . . 1 0 20 40 60 80 100 120 140 160 180 Temperature (°C) 0.4 0.21 o 0.0 2 - 0.6 0 20 40 60 80 100 120 140 160 180 Temperature (°C) Figure 5-16. Score variations of the loading factors for Cr*"-Cr'^-mixed-oxide (Dots are from PCR analysis, lines are the averages.) (continued on next page) 184 (continued) 0.6 0.4 m 0.2 0 0.0 1 ? - 0.2 c/3s -0.4 - 0.6 0 20 40 60 80 100 120 140 160 180 Temperature (°C) Figure 5-16. Score variations of the loading factors for Cr*“-Cr'^‘-mixed-oxide (Dots are from PCR analysis, lines are the averages.) 185 Sample Heating T (®C) [Cr'^'l (M) r 1 67cm^/ml 1.0 X 10"* Air-aging for „ 2 hours “ 5.0 cm^/ml 3.6 X 10"* CCC L (2.8cm^/ml) ^ 50 2.06 X 10"^ 113 7.18 X 10*^ L 170 2.75 X 10"* f Air-aging for 15 mins “ 2 .1 1 X 10"* 50 2.85 X 10"* Cr'"-Cr^' mixed-oxide (Wet weight 113 6.19 X 10 0.00328g /ml) 170 1.11 X 10'^ a. From Table 4-1 of Chapter 4 Table 5-4. [Cr'^] released by heated CCC film or Cr "-Cr^' mixed-oxide 5.4.5 Protection properties of several artificial **fUms'* and some treatments Since the Cr^' loading level in CCC film or in Cr"-Cr^' mixed-oxide changes during the release process, the protection property of CCC film for different loading levels might be an important issue. There are several ways to study this. The first is to allow the Cr^' in CCC film to be released, and then test the Rp of the residual film on the alloy. Another way is to use a reducing agent, such as NazSOs. to reduce the Cr^ in the film, thus generating a low-Cr^'-loading film. The protection properties of these films were tested and their Rp values are listed in Table 5-5 and Table 5-6. The surface images 186 of AA2024-T3 after the corresponding treatments are shown in Figure 5-17. All samples listed in Table 5-6 were treated as specified before measuring their Rp in 0.1 M NaCl solution. For some treated CCC samples, such as depleted CCC and Na 2S0 3 reduced CCC, "migration" experiments were conducted. The treated CCC was sandwiched with untreated AA2024-T3 with G.3-0.5 ml 0.1 M NaCl solution in between. After 48 hours, each sample was rinsed with nanopme water, and its Rp was measured individually in 0.1 M NaCl solution (see Table 5-5). Rp OCP Visual Treatment Sample (Q.cm^) (mV)* inspection Depleted -620 - “sandwich” depleted-CCC 4-6 X 10^ pit free with an untreated CCC -650 AA2024-T3 for 48 hours in O.lMNaCl (lcmVo.3ml) AA2024-T3 " 0.2-1 X 10"* ~-520 pitted NazSO] 5.46 X 10* -716 NazSO] on CCC for 24 reduced pit free hours, then “sandwiched” CCC* with an untreated 2024-T3 for 48 hrs AA2024-T3 ** 1.16 X 10"* -530 some pits a. Treated CCC samples b. Initially untreated AA2024-T3 c. Treated CCC films were sandwiched with imtreated AA2024-T3, and then Rp of each electrode was measured separately in 0.1 M NaCl solution. d. All potentials are vs. SCE. Table 5-5. Protection property of some artificial “films” 187 OCP Visual Sample Treatment Rp (Q.cm^) (m V ): inspection 0.1 M NaCl for 48 hrs AA2024-T3 (IcmVsO ml) 1-2 X IG^ --53G Pitted One piece of depleted im pact surface -686 pitted CCC was scratched in 9.67 X IG^ depleted CCC air, then immersed in 0.1M NaCl for 48 hrs inside scratch ~-52G pit free (Icm^/SOml) G.2-1 X IG^ Immersion in 0.02mM G.3G-1.61 -64G ~ Cr^'-O.IM NaCl for 6 AA 2024-T3 X IG^ -7G0 pit free days (IcmVsOml) Immersion in G.G2mM Cr'^'-G.lMNaCl for 48 2.69 X IG^ hrs, then immersed in AA 2024-T3 -526 pit free G.IM NaCl for 4 days (Icm^/SGml) Immersion in G.lmM Cr(N0 3 )3-G.lM NaCl G.84-1.55 .AA2024-T3 --68G very few pits for 48 hrs (lcm^/5Gml) X IG^ Kept in G.lmM CrC^- G.IM NaCl for 48 hrs G.G2-2.17 -528 ~ very few pits AA2024-T3 (lcm^/5Gml) X IG^ -688 a. All potentials are vs. SCE. Table 5-6. Protection property of Cr‘" and Cr'" VIon AA 2024-T3 188 plain AA2024-T3 after in O.IM NaCl for 48 hours M"- depleted CCC after “sandwich" Originally untreated AA2024-T3 with AA2024-T3 after “sandwich” with the in O.IM NaCl for 48 hours depleted CCC (shown in left) in O.IM NaCl for 48 hours Scratched depleted-CCC surface after in O.IM NaCl for 2 days Pit-ftee on intact surface Pitted inside a scratch 20pm Figure 5-17. Surface images of AA2024-T3 after various treatment (All images were collected by using x50 objective) (continued on next page) 189 (continued) Plain polished AA2024-T3 after immersion in 2xlO-5M Cr'^' -O.IM NaCl solution for 6 days Plain polished AA2024-T3 after immersion in 2xlO-*M Crvi -O.lMNaCl solution for 2 days, then in O.IM NaCl for 4 days AA 2024-T3 after immersion in AA 2024-T3 after immersion in 0.1 mM Cr(NO3)3-0.1 M NaCl 0.1 mM CrCl3-0.1 M NaCl solution for 48 hours solution for 48 hours 20^im Figure 5-17. Surface images of AA2024-T3 after various treatment (All images were collected by using x50 objective) 190 Cr^-depleted CCC film was obtained by immersing 1x1 cm^ CCC film (air-aged for 48 h) in 50 ml nanopure water for 7 days. Water was exchanged daily, until no Cr'^’ was detected by UV-Vis in the solution. Raman spectra of depleted films still show weak bands at 858 cm ', as shown in Figure 5-18. Further immersion in 0.1 M NaCl for two days did not cause any pitting on those samples (see Figure 5-17). But the depleted CCC can protect neither a nearby AA2024-T3 in an “artificial-scratch” set-up, nor a real scratch (see Figure 5-17). NaiSOj can reduce most of the Cr^' in CCC film, and generate a low-Cr^'-loaded CCC film. The spectra of a CCC before and after such treatment are shown in Figure 5-19. Since CCC can release Cr'^ when exposed to solution, a CCC covered A2024-T3 was immersed in 0.5 ml nanopure water for 3 days, the Raman spectrum of the resulting film is also shown in Figure 5-19. Similar to water-depleted- CCC film, the reduced film protects itself, but not scratches (see Table 5-6). One thing worth noticing is that both the depletion and the reduction processes involve prolonged exposure of the samples to diluted Cr'^' solution, which is unavoidable when putting CCC film in water. The observed protection could come fi’om the remaining Cr(OH) 3, with or without Cr^', or it could be the result of immersion in -lO ”^ M Cr^' solution. As listed in Table 5-5, immersion in Cr'^‘ provided protection, even when there was no Cr'^‘ in the solution surrounding the treated samples. In addition, -10"^ M Cr'" does show some protection, but far less effective than Cr'^'. This indicates that Cr'^' itself is responsible for most o f the protection; Cr'” is not the major factor here. The general results obtained firom the above observations are as follows; mobile Cr‘" provided some protection, but far less effective than Cr'^'. Immersion in diluted Cr^' provided protection for AA2024-T3, and such protection lasted when the treated 191 1000 800 Regular CCC 600 3 Depleted- CCC < i 400 u c 200 400 600 800 1000 1200 1400 1600 Raman shift (cm*') Figure 5-18. Raman spectra of regular CCC film and depleted-CCC film on AA2024-T3 (All spectra were collected under identical conditions, and are shown on the same scale.) 192 1000 800 (a) Regular CCC (b) Water immersed CCC 5 ^ 600 < I 400 (c) NajSOj immersed CCC 200 400 600 800 1000 1200 1400 1600 Raman shift (cm ') Figure 5-19. Raman spectra o f regular (a) CCC film, (b) CCC film after immersion in 0.5ml nanopure water for 3 days, and (c) the Na^SO^ reduced CCC film after 2 days immersion in 0.5ml O.IM NaCl (All spectra were collected under identical conditions, and are shown on the same scale.) 193 sample was exposed to 0.1 M NaCl solution. Both depleted CCC and reduced CCC contain small amount of Cr^' that was not mobile, and both protected themselves, but not a remote scratch. 5.5 Discussion In previous chapters, the Cr ' -O-Cr^ bonding model was established. In this chapter, the structural study results unraveled the specific details of the adsorbed Cr'^', through experimental observations and theoretical calculations. Although prepared under a wide variety of initial conditions, the synthetic Cr*" -Cr'^ mixed oxides showed very similar Raman spectra. A general analysis of the solution Cr^^ spéciation, as described in section 5.4.3, indicated that Cr"-O-Cr^'OjH is the most likely structural subimit of Cr^' species in CCC films and Cr'“-0-Cr'^’ mixed oxide. In order to provide further confirmation, an ab initio calculations were carried out. The calculation method was verified by comparing the experimental Raman spectra of some standard compoimds with those of calculated ones. Experimental Raman spectra of standard Cr^'-containing compoimds are shown in Figure 5-1*. The calculated Raman active vibrational frequencies are consistent with the experimental values, including those from o'* substituted Cr^'-oxy ions. The same method and basis set were used to calculate the Raman frequencies of the hypothetical clusters, whose structures are shown in Figure 5-4. The resulting Raman spectrum of structure-E, which contains a monodentate Cr‘"-0- Cr^'OjH subimit, was similar to the experimental spectrum of synthetic Cr-mixed-oxide. a. Notice that when dissolving KjCr^O? in HjO, the oxygen in dichronute ions exchanged with solvent O quickly, and the resulting Cr;Oi’‘ was revealed by a frequency shift of tnore than 40cm ' in less than 10 minutes. However, dissolving KjCrO* in H :0 did not show any frequency shift until six days later. The kinetics o f oxygen-exchange reaction between Cr'^-oxy ions and water has been studied previously [201-203]. It was reported that the rate of such oxygen exchange reaction is much slower at higher pH values [201 ]. The rate constant of reaction HCrO, + HjO was 2.4 x 10'’ s ', and that of CiO«’ + HjO was 3.2 x 10'’ s ' [201]. A H*-catalysis reaction mechanism was proposed [201]. For more details, please refer to the related references. 194 The possible Cr"-O-Cr^'OsH structure found here certainly helps in explaining the structural evolution during aging. For real-world airplanes, there are several different types of aging environments: in air at normal temperature, or under constant thermal radiation on a desert. Sometimes rain can repetitively wash away the solution layer on top of the coating. The studies discussed in the following section were all aimed at finding the possible structure-property variations during those aging processes. The Raman spectroscopic studies in the current chapter confirmed the structural details proposed in the “Langmuir” model (see Chapter 4). The aging in water is simply an adsorption-release process, and the remaining Cr^' surface species retains its structure despite the variations in its surface concentration, [Cr^'], or pH. The [Cr^'] and pH in a typical release experiment suggest that HCr 0 4 is the major aqueous Cr'^‘ species. An equilibrium between Cr ' -O-Cr^'OjH and HCr 0 4 (aq) fits the data in Chapter 2 and Chapter 4 quantitatively. The effect of air-aging on CCC and Cr'"-Cr^' mixed-oxide was discussed in Chapter 2 and Chapter 4. Spectroscopically, no significant change was observed except for some loss of water. Quantitatively, both CCC and Cr"'-Cr^' mixed-oxide release slightly less Cr^' after longtime aging in air. It is believed that further polymerization of Cr'"-hydroxide and gradual loss of water might be the major changes during air-aging [204-210]. Other structural rearrangement is also possible. In this chapter, thermal treatment was applied to accelerate the aging process and to mimic the conditions of high temperature environment. It has been reported that thermal treatment of Cr"'-hydroxide causes substantial structural changes. Cr'"-hydroxide does not crystallize after drying for a long time in 195 air[211], even under mild heating [212-214]. At a temperature lower than 150°C, it loses loosely bound water [212-214], and starts to form amorphous oxide at approximately 250°C [214]. At higher temperatures, ~ 400°C, it crystallizes and become CT2O3 [214]. The whole process is accompanied by a surface area decrease of - 50% [215]. Thermal and spectroscopic studies of a commercial Cr 2(Cr0 4 ) 3 6 H2O compound showed similar results [215]: a weight loss o f — 16% when heated to 170°C, mainly due to losing water, and a total weight loss of more than 30% when temperature was increased to 400 °C. Decomposition of Cr'^' was observed and total conversion to a- Cr 2 0 ] was achieved at - 400°C. In the current study, CCC samples were treated at 50°C, 113°C, or 170°C for one hour, and significant consequences were observed. First, the Cr"'-0-Cr^' species retained its center at 858 cm '(see loading factor 1 in Figure 5-13), and new Raman peaks at - 760 cm ', 898 cm ', and — 550 cm ' emerged after heating (see loading factor 2 in Figure 5-13). In addition, the average 858 cm"' peak intensity decreased steadily with increasing temperature (see Figure 5-14). Visually, the film turned darker after heating. A similar color change and slightly different spectral variations were observed for synthetic Cr'"-Cr^' mixed-oxide (see Figure 5-15 and Figure 5-16). However, the heated film and powder behaved quite differently when immersed in water. As listed in Table 5-4, the CCC film almost stopped releasing Cr'^' after heat treatment, while Cr'"-Cr^' mixed- oxide released more. It is understandable that heating can at least do three things to the hydrogel-solid film. First, structural changes of Cr'"-0-Cr^' species; second, the specific surface area of the hydroxide can be decreased due to water loss; third, thermally induced reactions between Cr^' and the aluminum matrix are possible. Oxidation of AI by O 2 is also 196 possible. For synthetic Cr‘"-Cr'^‘ mixed-oxide, only the former two consequences are applicable. Figures 5-20 and 5-21 show the possible structural changes on heated CCC film and Cr-mixed-oxide, respectively. The possible correlations between the thermal reaction products and the PRC results are also shown. For both CCC film and Cr"-O- Cr^' mixed oxide, the thermally induced structure change can take several possible routes, among them are decomposition and further anchoring [109, 216-218]. A simple comparison between CrfOH)] and Cr20] (see Figure 1-4 and Figure 1-6), reveals that thermal conversion from hydrogel to CrzO] requires not only further cross-linking and vvater-loss, but also dramatic structural rearrangement. When Cr'"-0-Cr^' species are involved, the competition between Cr"'-0-Cr"' and Cr”'-0-Cr'"' could result in the decomposition of Cr"-O-Cr^' species, generating Cr "-O-Cr " and “free” Cr'^'. The “free” Cr^' can be reduced to Cr'” by A1 when it is on the alloy surface. According to the PCR results of CCC film (see Figure 5-13 and Figure 5-14), such decomposition became more pronounced at higher temperatures, thus the 858 cm ' band intensity decreased and the 547 cm ' band intensity increased. Apparently, the conversion from Cr'^' to Cr"‘ results in a lower Nvi, which is the total moles of Cr^' in the film. According to Equation 4-3 in Chapter 4, the [Cr^'jgquii decreases with the square-root of Nvi, therefore the intensities from 858 cm ' band, and increased intensities from 968 cm ' band at elevated temperatures. The thermally generated free Cr^' is very likely to form polychromate species, and the broad band at 968 cm ' of loading factor 2 in Figure 5-15 is very close to the Cr^ '-O vibration of crystalline CrOa at 977 cm ' (see Figure 2-10). This implies that this Raman peak might come from the “free” Cr^', and its amoimt increased when more 197 nicrm al Reditcliof^ on AI x - Loading factor 1 III— 0 decomposition ♦ ( 'O r A ) ► L ose w a te r + Free Cr''* Loading factor 2 ; (probably in polychromate form) (547 cm') Thermally induced Cr'^' anchorinj Loading factor 2 VO 00 I (760,898 cm-') L o se w a ter Oxidalion \Cagefonnalion LosewMT * Figure 5-20. Possible structural changes in CCC film during thermal treatment o Loading factor 1 Thermal decomposition Loading factor 2 L o se w a ter 968 cm' —■ Cr'"—O (probably in Z polychromate form) I _ Loading factor 2 Thermally (-780,-896 cm' induced g Cr'^' anchorini I L o se w a te r —OH Cr'^'OjH Cage formation ‘ ------p . L o se w a ter Figure 5-21. Possible structural changes in Cr-mixed-oxide during thennal treatment Cr'"-0-Cr'^' bonds were broken at elevated temperatures. The loading factor 3 of Cr"‘-0- Cr^' mixed oxide, as shown in Figure 5-15, appear to be similar to the Cr"‘-0-Cr'^' heated-film releases less Cr'^* than the air-aged ones. However, such reduction by A1 only exists when Cr^' is in close contact with alloy during the heating. The Cr"'-Cr^' mixed- oxide samples were held in glass vials when undergoing thermal treatment, therefore the heated-Cr"'-Cr^' mixed-oxide should contain significant amounts of “free” Cr'^', which can be dissolved totally when exposed to water. The PCR results of Cr'"-0-Cr^' mixed oxide (see Figure 5-15, 5-16) support the thermal decomposition by showing decreased species in CCC, but its relative intensity was constant with temperature. The Cr^' species corresponding to this loading factor might be the thermal reaction product of Cr'"- hydroxide and residual Cr^' from mixed-oxide synthesis. The “residual Cr^'” refers to the small amount of Cr^'(aq) that could be “trapped” inside the porous mixed-oxide during synthesis. It is worth mentioning that the relative intensity variation of the 898,760, 547cm ' bands from CCC track each other. Similarly, the relative intensities of 968, 896 and -783cm ' bands from Cr ' -O-Cr^' mixed oxide increased simultaneously with elevated temperature. These imply that the thermal reaction products, CrzO] and bidentate Cr'"-0-Cr^' species for CCC, or “free” Cr'^' and bidentate Cr'"-0-Cr'^‘ species for Cr'"- O-Cr^' mixed oxide, were generated together. In addition to decomposition, some surface Cr^% such as Cr^'-O-Cr'^'OjH, can also under go further anchoring and form bidentate Cr^-O ligand, as shown in Figures 5- 20 and 5-21. The tight multiple bonds result in the “fixation” of Cr^' in the film or the powder. This provides the second reason to explain CCC film’s low [Cr'^'jequii. after 200 heating. The new Raman peaks at 780 cm ' and 898-896 cm ' were observed from both heated CCC film and heated Cr"'-0-Cr'^' mixed oxide, and might be from the multiple- bonded Cr^' species, such as the bidentate Cr'^' species shown in the center of Figures 5- 20 and 5-21. For both CCC film and Cr'"-0-Cr'^' mixed oxide, it is possible that the polymerization of Cr(OH ) 3 can generate some random sized closed-cage structures, which might encapsulate some Cr"'-0-Cr'^' species. This process can happen in any aging environment, including in air, in water, and in thermal treatment. The resulting closed- cages can hinder the encapsulated Cr'^' from being accessed by surrounding water, thus “trapping” the Cr^' permanently. The long-range structure of amorphous Cr'"-hydroxide is not accessible by conventional characterization techniques, such as X-ray Diffraction. But crystalline Cr'"-cage compoimds, as well as a wide variety of polynuclear transition metal cage complexes, have been synthesized and characterized. Several transiton metals, including Co, Ni, Cr, Mn, V, W, can form cages with OH, H 2O or O spacers [219-223]. For example, vanadium, very close to Cr in periodical table, forms a mixed-valent cage compound [V 2 2 0 5 4 (Mo 0 4 )]®' with the Mo 0 4 ^‘ encapsulated inside the closed cage [223]. Some Cr"'-cages contain O, OH and OH 2 as “bridges” between adjacent Cr"' centers, with each Cr'" surrounded by six oxygen atoms in octahedral coordination [219-220]. In order to stabilize the cage and form a crystal, some organic ligands were inserted [219- 220]. Other than those organic ligands, the structural features in Cr"'-cages are very similar to that of Cr'"-hydroxide. An example of one of the Cr'"-cage compounds is shown in Figure 5-22[219]. The possible formation of such cage-like structures could contribute to the release behavior of heated CCC, and it might also be the origin of the 201 so-called “trapped” Cr'^’ in regular CCC film or Cr’^-Cr'^* mixed-oxide. As we mentioned in the Cr"'-Cr^' titration experiment in Chapter 4, Cr"'-hydroxide apparently “trapped” - 10-20% Cr^' and these Cr^' could not be released even at pH 13. Furthermore, Raman of CCC film after prolonged exposure to water showed small portion of residual Cr^'. This Cr'"' species is spectroscopically similar to the regular 858 cm*' Cr"'-Cr^' species, but seems to be “un-releasable”. The possible formation of cage-like structures could also be the reason for these observations. Another important issue is the possible ftmction of Cr^' in protection. Current results revealed that a “depleted-CCC”, with “trapped” Cr^', can protect itself from aggressive chloride ions, as long as that there is no scratch caused by external forces (see Table 5-5). In addition, AA2024-T3 that had been immersed in Cr^' also protects itself, even when there is no Cr^' in solution (see Table 5-5). In both cases, the protection could come from either Cr'"^', or its reduction product, Cr'". Further studies, also shown in Table 5-5, indicated that Cr'" alone does not accoimt for the high Rp of Cr'^'-immersion treatment or the depleted-CCC film. Therefore, it is clear that mobile Cr^' must be present in order to achieve effective protection. One interesting observation is the Cr'^'-loading level on Cr'" -hydroxide. The solid became saturated when Cr'":Cr^' ratio reaches ~ 3:1. As shown in Figure 1-8, trimer is the most abundant solution species at pH 4-5 over a wide range of [Cr"']totai- A trimer can "attract" one HCr 0 4 *. When the solution Cr'":Cr^' ratio approaches 3:1 or lower, most of the trimers could be "attached" by HCr04, thus a Cr'":Cr'^' ratio 3:1 in the solid. Such trimer-HCr0 4 interaction could exist in solution, before the precipitation. When the solution Cr'":Cr'^' ratio is much higher than 3:1, not enough Cr^' is available for 2 0 2 C r" i Figure 5-22. Crystal structure of a Cr'”-cage compound (adapted from Figure 4 of reference [219].) 203 all the trimer clusters. But "bare" trimers can cross link with trimer-HCr 0 4 , therefore the Cr'“ -O-Cr^' mixed-oxides appear to have a continuously variable Cr"‘:Cr'^' ratio from 3 to higher. 5.6 Conclusions Several important conclusions can be drawn from current results. First, the 858cm ' species retained its structure imder the controlled, but widely varied conditions. Based on the solution spéciations imder those conditions, a Cr"'- 0 -Cr'^'0 3 H (see structure E in Figure 5-3) was proposed. This structure was supported by the ab initio calculation results and is consistent with the Langmuir model (Chapter 4) quantitatively. Second, aging in water was found to be a simple Cr'^‘ adsorption-release process. The structure of Cr^' surface species remained to be the same despite the varied Cr^' loading level on the solid, or the changing [Cr'^‘] and pH in the surrounding solution. Third, heating decreases the [Cr' ‘icquii released from CGC film, but increases that from Cr‘“-Cr'^' mixed-oxide. Decreasing of 858cm ' band intensity and generation of new species were observed in both cases. Possible Cr"'-0-Cr^' degradation, reaction with Al, and formation of “closed- cage” structure by Cr'"-oxy-hydroxide were proposed and shown in Figure 5-20 and Figure 5-21. 204 CHAPTER 6 SOME EFFECTS OF COPPER ON ALUMINUM 6.1 Introduction The study of inter-metallic compounds (IMC's) in Cu-rich aluminum alloys has been an active area for corrosion scientists for a long time[149-164]. Extensive research has been done and well established models are available to explain possible roles of Cu- rich IMC’s in CCC formation and corrosion [149-164]. In this chapter, some of the effects of copper on Al will be discussed, mainly because of their close correlation with the results shown in previous chapters. The Open- Circuit-Potentials (OCP's) of pure Al and pure Cu in various combinations of Alodine ingredients were tested. In addition to pure Al or Cu, a model for 2024-T3 alloy was made by electrically connecting piue Al to pure Cu, making a “Al-Cu pair” electrode. OCP's of such Al-Cu pair electrodes were also studied. 6.2 Experimental Al-Cu pair electrodes were made by coupling electrodes made from 99.999% Al and 99.999% Cu. The pure metal plates were cut into Ixlcm^ coupons and mounted in epoxy after attaching conductive wires through Ag-epoxy-resin. All the electrodes were 205 polished as described in Chapter 2. Rp values of pure Al, pure Cu, and Al-Cu pair electrodes were measured in O.IM NaCl solution by using the Gamry CMS 100/105/PC3- 300 system (Gamry Instruments, Inc., Warminster, PA). The film growth rates on pure Al, pure Cu, or the Al-Cu pair electrodes were measured as described in Chapter 3. AA2024-T3 electrodes were prepared as described in Chapter 2. Several synthetic coating solutions were prepared by using reagent grade compounds and Nanopure water. The default solution Q contains 0.1 M NaCl, 0.0144 M NaF, 0.0146 M KBF4, and 0.002 M KzZrF^. The pH was adjusted to 2.1 by adding HNO 3. The solution-Q was prepared such that it can be used to mimic the Alodine solution, but allow the CCC film formation to proceed step by step. Open-Circuit-Potentials (OCP’s) of AA2024-T3 electrodes were monitored in solution-Q, and then Cr'^' was injected to reach a final concentration of approximately 40 mM. When the OCP reached a stable value, Fe(CN) 6 ^‘ solution was injected to a concentration of approximately 3 mM. The solution was stirred manually after each injection. Similar experiments were conducted by adding Fe(CN)ô^'solution first, then Cr'^'. 6.3 Results The Rp values of pure Al, pure Cu, and Al-Cu pair electrodes, before and after Alodine treatment are shown in Table 6-1. The Alodine treated AA2024-T3 and pure Al showed similar Rp values of ~10^ Q»cm^, and OCP’s of 700-800mV. It is clear that coupling Al with Cu greatly decreased the overall Rp and changed the OCP value, which is actually close to the OCP of bare AA2024-T3. Alodine treatment of the Al-Cu pair results in CCC film formation on the Al side and a high Rp of the Al electrode, but no 206 significant spectroscopic or electrochemical changes were observed on the Cu side. As mentioned in Chapter 3, the CCC film growth rate on Al is slightly higher than that on 2024-T3. The results shown in Figure 6-1 clearly indicate that the CCC film growth rate on the pure Al electrode decreased significantly when electrically connected with Cu. There was no film growth on the Cu side, whether connected to Al or not, even after 5 minutes in Alodine solution. Since the surface area ratio of Al:Cu here is around 1:1, which is much lower than the ratio in a real 2024-T3 electrode, the growth rate here is only semi-quantitatively comparable with that of AA2024-T3. Rp OCP Rp (Q cm^) OCP (mV) " Metal (Qcra^) (mV): after Alodine Imin After Alodine Imin 2024-T3 -lO f --5 0 0 3.357 X 10* -853 99.999% Al 1.004 X 10^ -882 1.058 X 10* -745 99.999% Cu 6.829 X 10^ -162 4.280 X 10" -140 Al-Cu pair “ 1.574 X 10^ -671 N/A N/A 99.999% Al ^ 1.406 X 10* -735 99.999% Cu ” 8.771 X 10^ -130 a. Rp of the Al-Cu pair was measured before Alodine treatment. b. Al-Cu pair electrode was treated by Alodine, then the two electrodes were separated, and their Rp values measured individually. c. vs. Ag/AgCl reference electrode Table 6-1. Rp of pure Al, pure Cu, Al-Cu pair electrodes 207 1600 ^ 1200 I 00 00 0 800 i 1S 400 Al in alodine AA in Alodine Al(-Cu) in Alodine 0 T T 1 2 3 4 Coating time (min) Figure 6-1. CCC film growth rate on pure Al, AA2024-T3, and the aluminum side of the Al-Cu pair electrodes in Alodine solution The OCP’s of pure Al, 2024-T3, and Al connected with Cu in synthetic coating solutions revealed more differences among them. The OCP of AA2024-T3 in aerated default solution-Q is -570 ~ -550 mV, as shown in Figure 6-2. The addition o f — 40 mM of Cr'^‘ changed the OCP to —370 mV, and the following addition of Fe(CN)&,^' resulted in an OCP shift back to about -600 mV. Visual observation provides some hints about what was happening on the surface. The AA 2024-T3 surface was clear in solution-Q, very similar to the appearance of AA2024-T3 in air, but with some tiny bubbles coming out of the interface between AA2024-T3 and the solution. The addition of Cr^' changed the surface to a yellowish color, and a layer of porous gel-like brown film formed quickly following the addition of Fe(CN)6^‘. The gel-like film was so fragile that it fell off the surface when there was any slight shaking or vibration of the solution. The above experiments were also conducted on pure Al or pure Cu electrodes and the results are also shown in Figure 6-2. The OCP of Cu was ~ -90 mV in solution-Q, and it was changed to -20 mV after adding - 40 mM of Cr^', but there was no further change when -3 mM of Fe(CN)ô^ was also present. In the case of pure Al, the OCP Increased significantly after adding Cr^', but did not change significantly following the addition of Fe(CN)6^’. Similar experiments, but with different sequences, were conducted on AA2024- T3. The stabilized OCP of electrodes in default solution-Q was obtained first, then Fe(CN)e^ was injected into the solution to reach a concentration o f —3 mM. After the potential become stabilized, Cr^' was added and the net [Cr^'] was — 40 mM. The resulting OCP vs. time plots are shown in Figure 6-3. The OCP dropped to — 670 mV following the addition of Fe(CN)6^*, and the surface showed a very light blue color at this 209 add CrVI addFe(CN).^ -20mV / -22mV 0 -88mV \ r - 0.2 add Fe(CN)^^ addCr''* -370mV -0.4 0 -570~-550mV \ -590mV - 0.6 -750niV f-.... -720mV < - 0.8 N» cu a d d F e ( Œ ) ^ 3- O - 1. 0 - o addCr^' AA-2024-T3 - 1.2 -1350mV \ pure Al -1.4 ——— pure Cu 0 200 400 600 800 1000 1200 1400 1600 1800 Time (second) Figure 6-2. Open Circuit Potential of pure Al, pure Cu, and AA2024-T3 pair electrodes in synthetic coating solutions add Cr''' - 0.2 ^ -0.3 % < -0.4 a d d F e ( Œ ) -450mV wi -0.5 > -570mV CU y -0-6 -670mV -0.7 0 1000 2000 3000 4000 Time (second) Figure 6-3. Open Circuit Potential of AA2024-T3 electrode in synthetic coating solutions point. The OCP increased to - 250 — 240mV after adding Cr'^’, and quickly stabilized at -470 ~ -450 mV. During this time, the surface was quickly covered by a thick layer of gel-like brown film. “Oscillations” of OCP were observed after the visual formation of the brown film. The above experiments were repeated using de-aerated solutions. Solution-Q, Cr^' solution, and Fe(CN)6^' solution were all de-aerated by Ar for at least 30 mins beforehand. The containers for OCP measurement were tightly sealed with Seal view (Norton Performance Plastics Corp., Akron, OH). Both Cr'^' + Fe(CN)6^‘, or Fe(CN)6'* + Cr^' sequences were tested, and the results are shown in Figure 6-4. The general trends were very similar to those in aerated solutions, except for the final OCP values. The interesting thing was that the Fe(CN)6^’ + Cr^' sequence in de-aerated solution resulted in the same OCP as Cr^' + Fe(CN)6^' in aerated solution, and Fe(CN)6^‘ + Cr^' sequence in aerated solution resulted in the same OCP of Cr^' + FefCN)^^' in de-aerated solution. 212 a d d F e ( C N ) - 0.2 - -300mV > -470mV Ic/l -0.5 - > -590mV K» & W - 0.6 - -570mV o -600mV V. -0.7 - -700mV add Fe(CN)/ add Cr'^' - 0.8 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Time (second) Figure 6-4. Open Circuit Potential of AA2024-T3 electrodes in de-aerated synthetic coating solutions 6.4 Discussion The OCP’s of the electrodes in the various solutions provide more hints about the possible electrochemical changes during CCC formation. It is clear that the slow reaction between Al and Cr'^' in the absence of Fe(CN)6^* helps the OCP to stabilize at a very high value, and prevents obvious changes tothe AA2024-T3 surface. The subsequent addition of Fe(CN)6^’accelerates the reaction, thus resulting in a thick layer of hydrated gel film. The OCP dropped back to more negative values after the surface was covered by the film. However, the final OCP on pure Al is almost the same as before adding Fe(CN)ô^ • Based on the results in Chapter 3, there was significant film growth on pure Al without mediation. Therefore the surface was covered quickly by a thin layer of CCC film after being exposed to Cr^', and the following addition of Fe(CN)6^' only changed the film thickness, but not OCP. Adding Fe(CN)g^' first into the highly acidic and aggressive solution Q changed the surface significantly, based on both the OCP and the blue color of the AA2024-T3 surfaces. E° o f Fe(CN)6^’^'‘' is +0.12 V(vs. Ag/AgCl), but the OCP of AA2024-T3 in solution Q plus FefCN)^^' was — 0.7V (vs. Ag/AgCl), which is more negative than the OCP of AA2024-T3 in solution Q (>~0.6V vs. Ag/AgCl). There are some possible reasons for this. Since the OCP of AA2024-T3 in solution is the mixed potential of Al oxidation and O 2 reduction, which presumably happens on Cu-rich sites, anything that prevents O 2 reduction can lower the OCP potential, as shown in Figure 6-5. McGovern et al. [164] observed CN-species on Cu-rich IMC’s, and proposed that such CN-complex prevents CCC film deposition since the IMC’s were “deactivated” by such CN-coverage. It is possible that the IMC’s can also be covered by certain type of adsorbed CN-species when the AA2024-T3 was immersed in solution-Q with Fe(CN)6 \ 214 such coverage could hinder the oxygen reduction, thus a lower OCP was observed. It could also be that Fe(CN)6^'promotes anodic activity of Al matrix, as shown in Figure 6- 5. The corrosion current at OCP would be helpful in differentiate the two possibilities. A puzzling observation is the final OCP of AA2024-T3 in aerated or de-aerated solutions after different Cr^% Fe(CN)6^ sequences (see Figure 6-2,6-3, 6-4). Apparently, adding the two chemicals in different sequences results in different film properties. They could be different in terms of morphology or they could even be different chemically. The limited time prevents the author from pursuing this topic any further. But any effort related to these observations might be a crucial step in explaining CCC film’s protection mechanism. The slowing down of CCC formation on Al after being connected with Cu implies that the mediation efficiency by Fe(CN)6^" was impaired slightly. The results in Chapter 3 indicated that Fe(CN)6^' mediation is mainly controlled by kinetics, but did not rule out the possibility that electrode potential may also play some roles in the formation kinetics. The whole mediation cycle involves at least two reactions, Al + Fe(CN)6'” and Cr^' + Fe(CN)6^'. The large potential difference between pure Al and the Al connected with Cu, might have some effects on the reaction rate of Al + Fe(CN)6^', thus the whole mediation cycle. 215 E«(V) VO OCP in solution-Q W ON OCP in solution-Q with Fe(CN)/ »I(A) Oxygen reduction impaired by ferricyanide or its product Accelerated Al dissolution on AA2024-T3 Figure 6-5. Possible changes when AA2024-T3 was immersed in solution-Q with or without Fe(CN)^'' SUMMARY The results discussed in previous six Chapters are closely related to each other, although describing different aspects of the CCC formation and function. A summary will be helpful in connecting all the observations and conclusions. The overall formation and storage-release processes are schematically shown in the figure at the end of this summary. The formation of CCC film on AA 2024-T3 is initiated by acid and fluoride, through the formation of soluble AlFe^'. Fresh Al therefore is exposed to the coating solution. Although the redox reactions between Al° and Cr^' are thermodynamically favorable, the slow kinetics hinders effective film formation. Fe(CN)6^ accelerates the Al° + Cr'^* reaction. The acceleration involves two redox cross reactions, in particular Fe(CN)6^* + Al and Fe(CN)6^ + Cr^'. Such a mediation process is essential in providing necessary amoimt of Cr'" on the alloy surface in a short time span; then Cr'" hydrolysis can proceed and generate polymeric Cr"'-hydroxide. During the formation of CCC film, the deposition of nascent Cr"'-hydroxide was accompanied by the chemisorption of Cr^\ through formation of Cr"'-0-Cr^' bonds. Simultaneously, some Fe(CN)6^', and its degradation product Berlin Green, are physisorbed into the film. 217 As the result of the above reactions, a thin layer of Cr*"-Cr'^’ hydro gel is deposited on the surface of the alloy. The CCC film is exposed to moisture, or air, or heat, depending on the environment of the coated aircraft. One of the most common environmental conditions is air with some moisture. The condensation of moisture on the coating surface results in a thin layer of solution, or at least hydrated film. The Cr"-O-Cr^' can break up and release Cr^ ' into the solution, and the released Cr'^‘(aq) can also come back to the coating film by re-forming Cr"'-0-Cr^' bond with Cr’"-hydroxide. Such an equilibrium is controlled by pH, [Cr^'(aq)], and Cr^^-surface loading level in the CCC film. In general, the Cr"' - Cr^' equilibrium follows a mathematical model that is very similar to a Langmuir isotherm. Both experimental results and the model indicate that the formation of Cr"'-0-Cr^ ' species is favored at lower pH, higher [Cr^'(aq)], and larger amounts of Cr'"-hydroxide. The opposite of those conditions favors Cr^' release. These conclusions shed some light onto the formulation strategy of Alodine coating, and also that of possible alternatives. The Alodine coating conditions combine high [Cr'^'(aq)], low pH, high fluoride concentration, and Fe(CN)6^‘ to generate a high local concentration of Cr"', which polymerizes to form Cr'"-hydroxide bonded to Cr^'. As the result, the Cr'"-hydroxide in CCC was saturated by Cr^', with a Cr'":Cr'^' mole ratio of 3:1. The neutral (or even basic) conditions in the field favor Cr^' release, tlius permitting "self-healing". In addition, the effect of aging under elevated temperature was also studied. The lack of releasable Cr'^' fi-om CCC was correlated with Cr'"' thermal reduction on Al alloy and possible formation of multiple bonded Cr^' species. 218 H + F AlF^^'(soluble) AlOx film > i n i t i a t i o n Expose fresh J Al to coating solution -y - Fe(CN)^"--^yFe(CN)i^- A|3+ > m e d i a t i o n Fe(CN)^^ Al® < - deprotonation Cr(OH)2+ H O (HjO)^Cr Cr(OHj)^ H HjO H O . / (HjO)^Cr ; ) C r ^ " H HjO h y d r o l y s i s J rearrange HjO ^ \ ' ^ O H HjO I OH I OH " “ - i r - ” » HjO OHj y Figure 7-1 Schematic illustration of the overall CCC formation and release ( continued on next page) 219 (continued from previous page) i Further polymerizati^^^;. C r f " ^ hydrolysis c < : N a s c e n t Cr^i'-hvdroxide Cr'^' (aq) B i n d i n g ofCr^J } i n C C C f i l m Cr"'-Cr‘" in CCC film ( continued on next page) 220 In-field storage-release equilibrium DB + SO + fS SB release I X o 'storage' SB O O SB I SB I s ■I I FUTURE WORK In closing, some possibilities for future research on aluminum alloy corrosion deserve note. The first one is the effect of intermetallic particles on Al alloy. The detailed chemistry of these particles and their impact on surrounding Al could be an interesting topic. Several hypothetical corrosion mechanisms involve the role of IMC inclusions. For example, it is believed that oxygen reduction happens preferably on Cu-rich sites. It could be proved if a dye can be found to indicate such localized reactions. A pH- indicating-dye is a possibility, as is an H 2O2 indicating fluorescence probe. Alternatively, a molecule whose Raman spectrum changes with pH could also be used to map the local pH on a corroding surface. Of course, “smart paints” could evolve from such studies, if the formulation is practical. More could be done to explore the protection mechanism of Cr^'. A difficulty in Cr^' protection mechanism study is how to isolate Al from other alloying elements. Any electrochemical result from an alloy is the overall outcome of possibly several parallel reactions on different sites of the same surface. On the other hand, physically coupled bulk pure Al and pure Cu might not represent the alloy precisely. It might be possible to couple Al and Cu on a microscopic scale, and still allow the conditions (such as potential) of Al and Cu to be controlled or monitored separately. 222 Since the replacement of the CCC by other coatings appears unavoidable, more of my thoughts are orientated toward designing new coating systems. Considerations of stainless steel might help in explaining the idea. Fe can form an insoluble compound with P04^\ thus cover the steel chemically with an adsorbed film. Al"' has extremely rich organometallic chemistry and many organic chemicals form stable complexes with Al'". However, most of these complexes are water soluble. If some “fixation” mechanism could be found, such as polymerization of the ligands of these Al'"-complexes, an alternative might be formulated. The same idea can be applied to the intermetallic particles. Taking advantage of the enormous amount of possibilities provided by organometallic chemistry, some chemicals might be available or could be made to chemically “cover” the Al and the “active” inter metallic particles. Of course, some free ligands should be stored in the coating, along with the chemicals that can trigger the polymerization. Whenever there is a defect, the fi-ee ligands might form stable complex with the exposed Al, followed by the polymerization of such Al-complex, thus covering the defect. Such ideas are just speculations at present, but may provide some guidance in formulating future research approaches. 223 References 1. M. G. Fontana, Corrosion Engineering, International Edition, McGraw-Hill, 1987 2. D. A. Jones, Principles and Prevention of Corrosion, 2"“^ Edition, Prentice Hall, New Jersey, 1996 3. M. G. Fontana, and R. W. Staehle, Advances in Corrosion Science and Technology, Vol. 2, Plenum Press, New York - London, 1972 4. F. A. Cotton, G. Wilkinson, F. R. S., Advanced Inorganic Chemistry, Edition, Interscience Publishers, 1972 5. L. L. Shreir, R. A. Jarman, G T. Burstein, Corrosion, 3"^ Edition, Vol. 1, Butterworth-Heinemann Ltd., 1994 6. L. L. Shreir, R. A. Jarman, G. T. Burstein, Corrosion, Edition, Vol. 2, Butterworth-Heinemann Ltd., 1994 7. Technical Process Bulletin, No. 981, Parker Amchem, 1993 8. G. W. Critchlow, D. M. Brewis, J. Adhes., Vol. 61, No. 1-4, 213 (1997) 9. A. T. Evans, J. D. Scantlebury, L. M. Callow, Proc.-Electrochem. Soc., Vol. 95- 13 {Advances in corrosion Protection by Organic Coatings II), 266 (1995) 10. X.-H. Jin, J. D. Scantlebury, Cailiao Baohu, Vol. 25, No. 4, 4 (1992) 11. A. E. Yaniv, N. Fin, H. Dodiuk, 1. E. Klein, Appl. Surf. Sci., Vol. 20, No. 4, 538 (1985) 12. S. Maeda, M. Yamamoto, Progress in Organic Coatings, Vol. 33, 83 (1998) 13. J. S. Wainwright, O. J. Murphy, and M. R. Antonio, Corr. Sci., Vol. 33, 281 (1992) 14. M. W. Kendig, A. J. Davenport, and H. S. Isaacs, Corr. Sci., Vol. 34, 41 (1993) 15. J. Zhao, G. Frankel, and R. Mccreery, J. Electrochem. Soc., Vol. 145, 2258 (1998) 16. W. Mertz, Physiol. Rev., Vol. 49, 163 (1969) 17. R. A. Anderson, Trace Elements in Health and Disease, Norstedt Stockholm, 1985 224 18. R. A. Anderson, M. M. Polansky, N. A. Bryden, and J. J. Canary, Am. J. Clin. Nutr., Vol. 54, 909 (1991) 19. R. A. Anderson, Clin. Physiol. Biochem., Vol. 4, 31 (1986) 20. A. Elbetieha, M. H. Al-Haraood, Toxicology, Vol. 116, 39 (1997) 21. B. L. Finley, B. D. Kerger, M. W. Katona, M. L. Gargas, G. C. Corbett, and D. J. Paustenbach, Toxicology and Applied Pharmacology,VoX. 142, 151 (1997) 22. B. L. Finley, B. D. Kerger, M. W. Katona, M. L. Gargas, G. C. Corbett, and D. J. Paustenbach, Toxicology and Applied Pharmacology, Vol. 142, 151 (1997) 23. C. T. Dillon, P. A. Lay, A. M. onin, M. Cholewa, G. J. F. Legge, T. J. Collins, and K. L. Kostka, Chem. Res. Toxicol., Vol. 11, 119 (1998) 24. D. M. Stearns and K. E. Wetterhahn, Chem. Res. Toxicol., Vol. 10, 271 (1997) 25. L. Zhang, and P. A. Lay, Inorg. Chem., Vol. 37, 1729 (1998) 26. L. Zhang, and P. A. Lay, J. Am. Chem. Soc., Vol. 118, 12624 (1996) 27. P. A. Layer, and A. Levina, J. Am. Chem. Soc., Vol. 120, 6704 (1998) 28. R. P. Farrell, P. A. Lay, A. Levina, I. A. Maxwell, R. Bramley, S. Brumby, and J.- Y. Ji, Inorg. Chem., Vol. 37, 3159 (1998) 29. T.-C. Tsou, R.-J. Lin, and J.-L. Yang, Chem. Res. Toxicol., Vol. 10, 962 (1997) 30. X. Shi, M. Ding, J. Ye, S. Wang, S. Leonard, L. Lunyi, V. Castranova, V. Vallyathan, A. Chiu, N. Dalal, K. Liu, J. Inorg. Biochem., Vol. 75, No. 1, 37 (1999) 31. J. K. Speetjens, R. A. Collins, J. B. Vincent, and S. A. Woski, Chem. Res. Toxicol., Vol. 12, 483 (1999) 32. J. Singh, and E. T. Snow, Biochemistry, Vol. 37, 9371 (1998) 33. Development Document for Proposed Effluent Limitations Guidelines and Standards fo r the Centralized Waste Treatment Industry, Vol. 1, United States Environmental Protection Agency, December, 1998 34. Report by the Air Force Blue Ribbon Advisory Panel on Aircraft Coating, November, 1995 225 35. B. R. W. Hinton, Metal Finishing, September, 55 (1991) 36. E. Eichinger, Metal Finishing, March, 36 (1997) 37. A. E. Hughes, and R. J. Taylor, and B. R. W. Hinton, Surf. Interface Anal., Vol. 25,223 (1997) 38. A. J. Davenport, and H. S. Isaacs, Corr. Sci., Vol. 31, 105 (1990) 39. A. Kh. Bairamow, S. Zakipour, and C. Legraf, Corr. Sci., Vol. 25, N o.l, 69 (1985) 40. J. A. Treverton, and M. P. Amor, Trans. Inst. Metal Finishing, Vol. 60, No. 3, 92 (1982) 41. J. A. Treverton, and N. C. Davies, Surf. Interface Anal., Vol. 3, No. 5, 194 (1981) 42. K. Asami, M. Oki, G. E. Thompson, G. C. Wood, and V. Ashworth, Electrochimica Acta, Vol. 32, No. 2, 337 (1987) 43. F. W. Lytle, R. B. Greegor, G. L. Bibbins, K. Y. Blohowiak, R. E. Smith, and G. D. Tuss, Corr. Sci., Vol. 37, No. 3, 349 (1995) 44. H. Habazaki, K. Shimizu, P. Skeldon, G. E. Thompson, X. Zhou, J. De Laet, and G. C. Wood, Corr. Sci., Vol. 39, No. 4,719 (1997) 45. A. G. Schrott, G. S. Frankel, A. J. Davenport, H. S. Isaacs, and C. V. Jahnes, Surf. 5c/., Vol. 250, 139(1991) 46. J. K. Hawkins, H. S. Isaacs, S. M. Heald, J. Tranquada, G. E. Thompson, and G. C. Wood, Corr. Sci., Vol. 27, No. 4, 391 (1987) 47. G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson, and G. C. Wood, Corr. Sci., Vol. 33, No. 9, 1371 (1992) 48. G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson, and G. C. Wood, Corr. Sci., Vol. 34, No. 7, 1045 (1993) 49. G. M. Brown, K. Shimizu, K. Kobayashi, G. E. Thompson, and G. C. Wood, Corr. Sci., Vol. 35, Nos. 1-4, 253 (1993) 50. M. F. A. Rabbo, J. A. Richardson, and G. C. Wood, Corr. Sci., Vol. 18, 117 (1978) 226 51. M. F. A. Rabbo, J. A. Richardson, and G. C. Wood, Corr. Sci., Vol. 16, 667 (1976) 52. P. Schmutz, and G. S. Frankel, J. Electrochem. Soc., Vol. 145, No. 7, 2285 (1998) 53. P. Schmutz, and G. S. Frankel, J. Electrochem. Soc., Vol. 145, No. 7, 2295 (1998) 54. K. Schmidt, and M. Stratmann, Corr. Sci., Vol. 40, No. 8, 1441 (1998) 55. J. R. Waldrop, and M. W. Kendig, J. Electrochem. Soc., Vol. 145, No. 1, L11 (1998) 56. P. Schmutz, and G. S. Frankel, J. Electrochem. Soc., Vol. 146, No. 12, 4461 (1999) 57. A. M. Ahem, P. R. Schwartz, and L. A. Shaffer, Appl. Spectro., Vol. 46, No. 9, 1412(1992) 58. M. Lenglet, F. Petit, and J. Y. Malvault, G. E. Thompson, and G. C. Wood, Phvs. Status Solid. A, Vol. 143, 361 (1992) 59. J. D. Ramsey, R. L. McCreery, J. Electrochem. Soc., Vol. 146, No. 11, 4076 (1999) 60. G. W. Ewing, Analytical Instrumentation Handbook, 2"** Edition, Marcel Dekker, Inc., 1997 61. C. Klauber, Microbeam Analysis, Vol. 4, 341 (1995) 62. B. G. Yacobi, D. B. Holt, and L. L. Kazmerski, Microanalysis o f Solids, Plenum Press, New York, 1994 63. Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer Corporation 64. A. G. Schrott, G. S. Frankel, A. J. Davenport, H. S. Isaacs, C. V. Jahnes. Surf. Sci., Vol. 250, 139 (1991) 65. M. Kendig, R. Addison, and S. Jeanjaquet, J. Electrochem. Soc., Vol. 146, No. 12,4419(1999) 66. H. J. Humecki, Practical Guide to Infrared Microscopy, Marcel Dekker, Inc., New York-Basel-Hong Kong, 1995 67. W. Suetaka, Surface Infrared and Raman Spectroscopy (Methods and Application), Plenum Press, New York-London, 1995 227 68. J. J. Lasema, Modem Techniques in Raman Spectroscopy, John Wiley & Sons, 1996 69. D. Lin-Vien , et al. The Handbook o f infrared and Raman Characteristic Frequencies o f Organic Molecules, Academic Press, Boston, 1991 70. K. Nakamoto, Infrared and Raman Spectra o f Inorganic and Coordination Compounds, Part A and B, S'**. Edition, John Wiley & Sons, Inc., 1997 71. B. J. Clark, T. Frost, M. A. Russell, UVSpectroscopy Techniques, Instrumentation, Data Handling, Chapman & Hall, 1993 72. G. R. Van Hecke, K. K. Karukstis, A Guide to Laser Chemistry, Jones and Barlett Publishers, 1998 73. C. L. Rollison, The Chemistry o f Chromium, Molybdenum and Tungsten, Pergamon Press, 1973 74. P. A. Cox, Transition Metal Oxides, An introduction to their electronic structure and properties, Clarenton Press, Oxford, 1992 75. L. Spiccia, and W. Marty, Inorg. Chem., Vol. 25, 266 (1986) 76. L. Spiccia, H. Stoeckli-Evans, W. Marty, and R. Giovanoli, Inorg. Chem., Vol. 26, 474(1987) 77. M. J. Avena, C. E. Giacomelli, and C. P. De Pauli, J. Colloid Interface Sci., Vol. 180, 428(1996) 78. P. P. Rotzinger, H. Stunzi, and W. Marty, Inorg. Chem., Vol. 25, 489 (1986) 79. A. Drljaca, and L. Spiccia, Polyhedron, Vol. 14, No. 12, 1653 (1995) 80. A. Drljaca, and L. Spiccia, Polyhedron, Vol. 15, No. 17, 2875 (1996) 81. M. R. Grace, and L. Spiccia, Polyhedron, Vol. 10, No.20-21, 2389 (1991) 82. L. Spiccia, and W. Marty, Polyhedron, Vol. 10, No. 6, 619 (1991) 83. M. Thompson, and R. E. Connick, Inorg. Chem., Vol. 20, 2279 (1981) 84. M. Ardon, and A. Linenberg, J. Phys. Chem., Vol. 65, 1443 (1961) 228 85. H. U. Gudel, U. Hauser, and A Furrer, Inorg. Chem., Vol. 18, No. 10, 2730 (1979) 86. S. J. Crimp, L. Spiccia, H. R. Krouse, and T. W. Swaddle, Inorg. Chem., Vol. 33, 465(1994) 87. J. E. Finholt, M. E. Thompson, and R. E. Connick, Inorg. Chem., Vol. 20, 4151 (1981) 88. K. Wieghardt, W. Schmidt, H. Endres, and C R. Wolfe, Chem. Ber., Vol. 112, 2837 (1979) 89. H. Stunzi, and W. Marty, Inorg. Chem., Vol. 22, 2145 (1983) 90. H. Stunzi, F. P. Rotzinger, and W. Marty, Inorg. Chem., Vol. 23, 2160 (1984) 91. L. Monsted, O. Monsted, and J. Springborg, Inorg. Chem., Vol. 24, 3496 (1985) 92. M. T. Flood, R. E. Marsh, H. B. Gray, J. Am. Chem. Soc., Vol. 91, 193 (1969) 93. T. Damhus, and E. Pedersen, Inorg. Chem., Vol. 23, 695 (1984) 94. H. Stunzi, L. Spiccia, F. P. Roztinger, and W. Marty, Inorg. Chem., Vol. 28, 66 (1989) 95. L. Spiccia, Polyhedron, No. 16, 1865 (1991) 96. A. Drljaca, and L. Spiccia, Polyhedron, Vol. 15, No. 24, 4373 (1996) 97. D. F. Shriver, P. W. Atkins, and C. H. Langford, Inorganic Chemistry, 6‘*'. Edition, W. H. Freeman and Company, New York, 2000 98. M. R. Grace, and L. Spiccia, Inorganica Chimica Acta, Vol. 213, 103 (1993) 99. D. T. Richens, The Chemistry o f Aqua Ions, John Wiley & Sons, 1997 100. P. Lofgren, Acta Cryst., B29, 2141 (1973) 101. P. Norby, A. N. Christensen, H. Fjellvag, and . Nielsen, J. Solid State Chem., Vol. 94, 281 (1991) 102. K. A. Wilhemi,/Ic/a Chem. Scand., Vol. 12, 1965 (1958) 103. O. Jonsson, Acta Chem. Scand., Vol. 24, 3627 (1970) 229 104. L. Suchow, I. Fankuchen, and R. Ward, J. Phys. Chem., Vol. 74, 1678 (1951) 105. K. A. Wilhemi, Acta Chem. Scand., Vol. 22, 2565 (1968) 106. K. A. W ilhemi,/Ic/û Chem. Scand., Vol. 19, 165 (1965) 107. Material Safety Data Sheet for Alodine ^ / 200S, Endura Manufacturing Co. Ltd., Edmonton Alberta, Canada, 1991 (http://www.ehs.ohio- state.edu/Script2.acgi?66005&., reviewed 01/13/2000) 108. P. L. Hagans, and C. M. Hass, Surface and Interface Analysis, Vol. 21, 65 (1994) 109. B. M. Weckhuysen. I. E. Wachs, and R. A. Schoonheydt, Chem. Rev., Vol. 96, 3327(1996) 110. F. W. Eppensteiner, and M. R. Jenkins, Metal Finishing, Vol. 97, Issue 1, 494 (1999) 111. M. H. Abbas, Metal Finishing, December, 8 (1995) 112. D. B. Brown, D. F. Shriver, and L. H. Schwartz, Inorg. Chem., Vol. 7, 77 (1968) 113. M. W. Fuller, K.-M. F. Le Brocq, E. Leslie, and I. R. Wilson, Aust. J. Chem., Vol. 39, 1144(1986) 114. M. G. De Oliveira, G. J. Langley, and A. J. Rest, J. Chem. Soc., Dalton Trans., 2013 (1995) 115. S. E. Ronco, and P. J. Aymonino, Transition Met. Chem., Vol. 12, 174 (1987) 116. S. J. Robuck, and R. G. Luthy, Wat. Sci. Tech., Vol. 21, 547 (1989) 117. A. F. Wells, Structural Inorganic Chemistry, 4^. Edition.,Clarendon Press, Oxford, 1975 118. D. A. Dzombak, C. L. Dobbs, C. L. Culleiton, J. R. Smith, D. Krause, Proc - WEFTEC'96, Annu., Conf. Expo., 69'*'., Vol. 3, 107 (1996) 119. V. D. Neff, J. Electrochem. Soc., Vol. 125, No. 6, 886 (1978) 120. J. Suo, Y. Chen, R. Li, Faming Zhuanli Shenqing Gongkai Shuomingshu, 8pp (1994) 121. W. Zhang, Y. Li, S. Dong, Chin. Sci. Bull., Vol. 37, No. 4, 347 (1992) 230 122. Ninomiya, Kagaku to Kyoiku, Vol. 40, No. 1, 50 (1992) 123. G. Socrates, Infrared Characteristic Group Frequencies, John Wiley & Sons, New York (1994) 124. S. Hoste, L. Verdonck, and G. P. van der Kelen, Bull. Soc. Chim. Belg., Vol. 7, 91 (1982) 125. D. M. Drew, Anal. Chem., Vol. 45, 2423 (1973) 126. P. K. Dutta, and M. Puri, J. Phys. Chem., Vol. 93, 376 (1989) 127. B. I. Swanson, and J. J. Rafalko, Inorg. Chem., Vol. 15, 249 (1976) 128. N. R. de Tacconi, N. Myung, and K. Rajeshwar, J. Phys. Chem., Vol. 99, 6105 (1995) 129. K. Kunimatsu, y. Electroanal. Chem., Vol. 262, 195(1989) 130. T. Ozeki, and D. E. Irish, J. Electroanal. Chem., Vol. 280, 451 (1990) 131. W. R. McGovern, P. Schmutz, R. G. Buchheit, R. L. McCreery, in preparation 132. A. Amine, J.-M. Kauffinann, and G. J. Patriarche, Talanta, Vol. 40, No. 8, 1157 (1993) 133. S. Basak, and K. Rajeshwar, J. Electroanal. Chem., Vol. 259, 403 (1990) 134. P. Jeroschewski, K. Haase, A. Trommer, and P. Grundler, Electroanalysis, Vol. 6, 769(1994) 135. S. Uchiyama, V. Kurokawa, Y. Hasebe, and S. Suzuki, Electroanalysis, Vol. 6, 63 (1994) 136. F. Torabi, K. Ramanathan, P.-O. Larsson, L. Gorton, K. Svanberg, Y. Okamoto, B. Danielsson, M. Khayyami, Talanta, Vol. 50, No.4, 787 (1999) 137. A. S. Bassi, E. Lee, J.-X. Zhu, Food Res. Int., Vol. 31, No.2, 119 (1998) 138. P. C. B. Page, M. J. McKenzie, D. R. Buckle, Tetrahedron, Vol. 54, No. 48, 14581 (1998) 139. E. Kilinc, A. Erdem, L. Gokgunnec, T. Dalbasti, M. Karaoglan, M. Ozsoz, Electroanalysis, Vol. 10, No.4, 273 (1998) 231 140. X. Dou, Y. Ozaki, Appl. Spectrosc., Vol. 52, No. 6, 815 (1998) 141. F. Ge, X.-E. Zhang, Z.-P. Zhang, X.-M. Zhang, Anal. Lett., Vol. 31, No. 3, 383 (1998) 142. Y. Liu, L. C. Seefeldt, V. D. Parker, Anal. Biochem., Vol. 250, No. 2, 196 (1997) 143. A. Noll, V. Rudolf, E. W. Grabner, Electrochim. Acta, Vol 44, No. 2-3, 415 (1998) 144. M. A. Watzky, J. F. Endicott, X. Song, Y. Lei, and A. Macatangay, Inorg. Chem., Vol. 35, 3463 (1996) 145. A. Boffi, and E. Chiancone, Biochem., Vol. 36, 4505 (1997) 146. G. Eaton, A. s. Dena-Nunez, and M. C. R. Symons, J. Chem., Soc., Faradav Trans. , /, Vol. 84,2181 (1988) 147. A. Simplerler, and W. Mikenda, Monatsh. Chem., Vol. 128 (1997) 148. P. S. Braterman, C. Tan, and J. Zhao, Mater. Res. Bull., Vol. 29, 1217 (1994) 149. T. J. R. Leclere, A. J. Davenport, J. Deakin,. Raikes, R. C. Newman, Proc.- Electrochem. Soc., Vol. 98-17 {Critical factors in Localized Corrosion III), 130 (1999) 150. T. Suter, R. C. Alkire, Proc.-Electrochem. Soc., Vol. 98-17 {Criticalfactors in Localized Corrosion III), 118 (1999) 151. E. A. DeBartolo, B. M. Hillberry, Int. J. Fatigue, Vol. 20, No. 10, 727 (1998) 152. P. Schmutz, G. S. Frankel, J. Electrochem. Soc., Vol. 145, No. 7, 2295 (1998) 153. P. Schmutz, G. S. Frankel, J. Electrochem. Soc., Vol. 145, No. 7, 2285 (1998) 154. C.-M. Liao, J. M. Olive, M. Gao, R. P. Wei, Corrosion (Houston), Vol. 54, No. 6, 451 (1998) 155. J. E. Wittig, J. O. Milewski, Proc. Int. Conf. Trends Weld. Res., 3"*, 199 (1993) 156. J. R. Scully, T. O. Knight, R. G. Buchheit, D. E. Peebles, Corr. Sci., Vol. 35, Nos. 1-4 {Advances in Corrosion and Protection, Ptl), 185 (1993) 157. R. G. Buchheit, J. Electrochem. Soc., Vol. 142, No. 11, 3994 (1995) 232 158. R. G. Buchheit, R. P. Grant, P. F. Hlava, B. Mckenzie, G. L. Zender, J. Electrochem. Soc., Vol. 144, No. 8, 2621 (1997) 159. R. G. Buchheit, J. Appl. Electrochem., Vol. 28, No. 5, 503 (1998) 160. F. D. Wall, K.-A. Son, N. A. Missert, J. C. Barbour, M. A. Martinez, K. R. Zavidil, J. P. Sullivan, R. G. Copeland, W. R. Cieslak, R. G. Buchheit, H. S. Isaacs, Proc.-Electrochem. Soc., Vol. 98-17 {Critical Factors in Localized Corrosion III), 701 (1999) 161. R. G. Buchheit, L. P. Montes, M. A. Martinez, J. Michael, P. F. Hlava, J. Electrochem. Soc., Vol. 146, No. 12, 4424 (1999) 162. G. P. Halada, C. R. Clayton, M. J. Vasquez, J. R. Keams, M. W. Kendig, S. L. Jeanjaquet, G. G. Peterson, G. S. McCarthy, G. L. Carr, Proc.-Electrochem. Soc., Vol. 98-17 {Critical factors in Localized Corrosion III), 139 (1999) 163. K. Shimizu, G. M. Brown, K. Kobayashi, P. Skeldon, G. E. Thompson, G. C. Wood, Corr. Sci., Vol. 40, No.7, 1049 (1998) 164. W. R. McGovern, Chromium Deposition on Copper Containing Aluminum Alloys Studied via Raman, Profilometry, Auger, and Electrochemical Methods, M.S. thesis. The Ohio State University, 1999 165. D C. Harris, Quantitative Chemical Analysis, 4th Edition, pp 186-188, W.H. Freeman (1995). 166. V. G. Poulopoullou, E. Vrachnou, S. Koinis and D. Katakis, Polyhedron, Vol. 16, 521 (1997) 167. J. J. Cruywagen, J. B. B. Heyns, and E. A. Rohwer, Polyhedron, Vol. 17, 1741 (1998) 168. R. K. Tandon, P. T. Crisp and J. Ellis, Talanta, Vol. 31, 227 (1984) 169. T. Shen-Yang and L. Ke-An, Talanta, Vol. 33, 775 (1986) 170. F.-C. Xu, H. R. Krouse, and T. W. Swaddle, Inorg. Chem., Vol. 24, No. 3, 267 (1985) 171. A. J. Bard and L. R. Faulkner, Electrochemical Methods, Wiley, New York, 1980, p516ff 172. E. Akiyama, L. Xia, R. McCreery, A. Markworth, G. S. Frankel, in praparation 233 173. V. F. Komarov,B- Nazirmadov, I. V. Melikhov, A. I. Novikov, Zh. Neorg. Khim. Vol. 31, No. 2, 434(1986) 174. K. G. Stollenwerk, D. B. Grove, J. Environ. Quai., Vol. 14, No. 1, 150 (1985) 175. T. Aoki, M. Munemori, Water Res., Vol. 16, No. 6, 793 (1982) 176. M. M. Benjamin, Environ. Sci. Technol., Vol. 17, No.l 1, 686 (1983) 177. H. Abdel-Samad, P. R. Watson, Appl. Surf. Sci., Vol. 108, 371 (1997) 178. D. Baron, C. D. Palmer, and J. T. Stanley, Environ. Sci. Technol., Vol. 30, 964 (1996) 179. M. A. Olazabal, N. P. nikolaidis, S. A. Suib, J. M. Mudariage, Environ. Sci. Technol., Vol. 31, 2898 (1997) 180. F. Ishino, M. Munemori, Nippon Kagaku Kaishi, Vol. 9, 1404 (1984) 181. F. Ishino, M. Munemori, Nippon Kagaku Kaishi, Vol 4, 569 (1984) 182. M. Balasubramanian, C. Melendres, J. Synchrotron Radiat., Vol. 6, No. 3, 594 (1999) 183. M. M. Sokolova,V. V. VolTchin, Zh. Neorg. Khim., Vol. 42, No. 7, 1091 (1997) 184. V. M. Rogov, T. L. Shvetsova, Khim. Tekhnol. Vody, Vol. 8, No. 3, 22 (1986) 185. M. M. Bhutani, A. K. Mitra, R. Yiuman, Adsorpt. Sci. Technol., Vol. 8, No. 1, 44 (1991) 186. J. Iqbal, M. L. Mirza, and S. Malik, y. Chem. Soc. P ak, Vol. 17, No. 1, 11 (1995) 187. B. Houri, A. Legrouri, A. Barroug, C. Forano, J.-P. Besse, Collect. Czech. Chem. Commun., Vol. 63, No. 5, 732 (1998) 188. E. Selli, A. De Giorgi, G. Bidoglio, Environ. Sci. Technol., Vol. 30, 598 (1996) 189. Y. 1. Sukharev, Y. V. Egorov, Neorg. Mater., Vol. 4, No. 6, 996 (1968) 190. M. Valigi, D. Gazzoli, and A. Cimino, J. Phys. Chem., B, Vol. 103, 11318 (1999) 191. D. Prasad, S. Sanyal, J. Indian Chem. Soc., Vol. 74, No. 8, 637 (1997) 234 192. C. H. Wang, C. P. Huang, H. E. Allen, P. B. Leavens, P. F. Sanders, Environ. Sci. Technol., Vol. 30, 371 (1996) 193. M. Kendig, R. Addison, S. Jeanjaquet,, J. Electrochem. Soc., Vol. 146, No. 12, 4419(1999) 194. Gaussian 98 User's Reference (Revision A I), M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Toraasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Gui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle, and J. A. Pople, Gaussina, Inc., Pittsburgh PA, 1998 195. P. Gili, P. A. Lorenzo-Luis, Coordination Chemistry Reviews, Vol. 193-195, 747 (1999) 196. N. Spanos, S. Slavov, Ch. Kordulis, and A. Lycourghiotis, Langmuir, Vol. 10, 3134(1994) 197. J. D. Ramsey, R. L. McCreery, in preparation 198. J. D. Ingle, Jr., S. R. Crouch, Spectrochemical Analysis, Prentice Hall, New Jersey, 1988 199. D. H. Cho, S. D. Yim, G. H. Cha, J. S. Lee, Y. G. Kim, J. S. Chung, and l.-S. Nam, J. Phys. Chem. A, Vol. 102, 7913 (1998) 200. P. G. Harrison, N. C Lloyd, W. Daniell, C. Bailey, and W. Azelee, Chem. Mater.. Vol. 11, 896(1996) 201. A. Okumura, M. Kitani, Y. Toyomi, and N. Okazaki, Bull. Chem. Soc. Jpn., Vol. 53,3143 (1980) 202. A. Okumura, N. Yamamoto, and N. Okazaki, Bull. Chem. Soc. Jpn., Vol. 46, 3633 (1973) 203. A. Okumura, S. Watanabe, M. Sakaue, and N. Okazaki, Bull. Chem. Soc. Jpn., Vol. 52, 2783(1979) 235 204. A. I. Mun, L. G. Gurvich, N. M. Kazantseva, A. V. Zasypalov, Miner. Syr va. Vol. 3, 83 (1983) 205. A. I. Mun, G. V. Zheludkova, N. M. Kazantseva, D. I. Zul’ftigarly, Azerb. Khim. Zh., Vol. 3, 81 (1981) 206. A. I. Mun, G. V. Zheludkova, N. M. Kazantseva, Miner. Syr'ya, Vol. 6, 32 (1981 ) 207. R. A. Abdurakhmanov, E. A. Aripov, S. Z Muminov, R. M. Rustamova, Zh. Neorg. Khim, Vol. 22 (10), 2634 (1977) 208. K. M. Akhmetov, D. K. Arginbaev, M. Z. Ugorets, A. Z. Beilina, Deposited Doc., (VINITI 933-74), 38 (1974) 209. S. 1. Smyshlyaev, L. A. Simonova, Zh. Fiz. Khim, Vol. 48, No. 2, 487 (1974) 210. M. S. Kovel, Tr. Ural. Politech. Inst., No. 152, 5 (1966) 211. S. 1. Pechenyuk, L. P. Kuz’mich, S. I. Matveenko, E. V. Kalinkina, Colloid and Surfaces, A., Vol. 144, 43 (1998) 212. K. K. Singh, P. R. Sarode, P. Ganguly, J. Chem. Soc. Dalton Trans., 1895 (1983) 213. C. Volzone, A. M. Cesio, Materials Chemistry and Physics, Vol. 48, 216 (1997) 214. P. Ratnasamy, and A. J. Leonard,/. Phys. Chem., Vol. 76, No. 13, 1838 (1972) 215. R. Toranzo, M. A. Vicente, and M. A. Banares-Munoz, Chem. Mater., Vol. 9, 1829(1997) 216. B. M. Weckhutsen, and I. E. Wachs, J. Phys. Chem., B, Vol. 101, 2793 (1997) 217. P. G. Harrison, N. C. Lloyd, W. Daniel, C. Bailey, and W. Azelee, Chem. Mater., Vol. 11, 896(1999) 218. B. M. Weckhuysen, B. Schools, and R. A. Schoonheydt, / Chem. Soc., Faraday Trans., Vol. 93, No. 11, 2117(1997) 219. E. K. Brechin, A. Graham, P. E. Y. Milne, M. Murrie, S. Parsons, and R. E. P. Winpenny, Phil. Trans. R. Soc. Lond., A, Vol. 357, 3119 (1999) 220. F. E. Mabbs, E. J. L. Mclnnes, M. Murrie, S. Parsons, G. M. Smith, C. C. Wilson, and R. E. P. Winpenny, Chem. Commum., 643 (1999) 236 221. R. P. Winpenny, Comments Inorg. Chem., Vol. 20, Nos. 4-6, 233 (1999) 222. J. S loan, J. L. Hutchison, R. Tenne, Y. Feldman, T. Tsirlina, and M. Homyonfer, J. Solid State Chem., Vol. 144, 100(1999) 223. T. Yamase, L. Yang, R. Suzuki, J. Molecular Catalysis A., Vol. 147, 179 (1999) 237