|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001966689
007 cr mnu|||uuuuu
008 081023s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002248
Corrosion assessment of mechanically formed aluminized steel
h [electronic resource] /
by Mersedeh Akhoondan.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Ribbed steel pipes made of Type 2 aluminized steel are commonly used for culvert pipes for highway drainage. Typically aluminized steel pipes have shown good durability and are expected to have long service life, e.g. 75 years; also, they are used in a wide variety of soil and water conditions. However, early corrosion of aluminized steel pipes has been recently observed in some inland locations. Initial observations showed severe corrosion in forms of pits, both along the ribs and at the nearby flat portions of the pipes. It is critical to determine the cause of early deterioration and establish methods of durability prediction. The possibility of unusual environmental conditions is being investigated elsewhere, but this research focuses on possible mechanical factors aggravating corrosion, since it is prevalent near pipe rib deformations. While forming the rib bends in the pipe, the outer bend surface is exposed to extreme tensile stresses which would cause small coating cracks (microfissures) exposing base metal. Those may lead to early corrosion as galvanic protection from the surrounding aluminum may not be sufficient under certain environments. Electrochemical impedance spectroscopy was used to measure corrosion rate of both formed and flat aluminized steel samples in simulated natural waters. Initial findings show that specimens formed by spherical indentation were susceptible to early corrosion development in moderately aggressive simulated natural water, but not in a more benign, precipitating simulated natural water solution.
Thesis (M.S.C.E.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 67 pages.
Alberto Sags, Ph.D.
Spiral ribbed pipes.
x Civil Engineeering
t USF Electronic Theses and Dissertations.
Corrosion Assessment of Mechanically Formed Aluminized Steel by Mersedeh Akhoondan A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Civil Engineering Department of Civil & Environmental Engineering College of Engineering University of South Florida Major Professor: Alberto Sags, Ph.D. Noreen Poor, Ph.D. Rajan Sen, Ph.D Date of Approval: November 1, 2007 Keywords: Spiral Ribbed Pipes, Corrugated, Mi crofissures, Scale-Forming Waters, Electrochemical Behavior Copyright 2007, Mersedeh Akhoondan
Dedication This work is dedicated to: My mother, Taraneh Bahary Your encouragement and comforting words are a tonic for my soul. My father, Ali Akhoondan Your unconditional love and sacrifices continue to enrich my life. My sister, Mandana Your friendship is a treasure beyond compare. Also, my Uncle Majid and the rest of my family who always supported me and provided me with the means to finish this project.
Acknowledgments First, I would like to sincerely appreci ate Dr. Alberto Sag s, my mentor and supervisor for his sage advice, insightful criticism, and patie nt encouragement. Also, I appreciate the help and guidance of Dr. Leonardo Casere s, the pioneer in this investigation. I gratefully acknowledge support from the Florida Department of Transportation and the Federal Highway Administ ration in funding this research. The opinions, findings, and conclusions expressed here are those of the author and not necessarily those of the supporting agency.
i Table of Contents List of Tables iii List of Figures iv Abstract vii Chapter One Introduction 1 1.1 Background on Aluminized Steel Type 2 1 1.2 Spiral Ribbed Pipe-Forming Process 3 1.3 Durability of Corrugated Culvert Pipe 5 1.3.1 Field Investigations 5 1.4 Review of Relevant Informati on on Characteristics and Corrosion Performance of Aluminized Steel 6 1.4.1 The Morphology of Intermetalli c Layer in Hot-dip Aluminized Steels 6 1.4.2 Formability of Metal-Coated Steel Sheets 7 1.4.3 Corrosion Mechanism of Aluminized Steel 8 1.4.4 Previous Studies Conducted by the University of South Florida 10 1.5 Objective of This Investigation 12 1.6 Approach 12 Chapter Two Preliminary Experiments and Results 14 2.1 Bending Process 14 2.2 Electrochemical Cell Set Up and Test Solution 15 2.2.1 Sponge-Cell System 15 2.2.2 Immersion-Cell System 18 2.3 Results and Discussion 20 2.3.1 Open Circuit Potential Trends 20 2.3.2 EIS Trends 22 2.3.3 Direct Observation of Corrosion 26 2.4 Conclusions Obtained From Preliminary Experiments 27 Chapter Three Deep Drawn Alumini zed Steel Experiments-Methodology 28 3.1 Dimpling Procedure 29 3.2 Immersion-Cell Experiment 31 3.3 Solution Compositions 33 3.4 Metallography 34 3.4.1 Metallography Process 35
ii Chapter Four Deep Drawn Aluminized Stee l Experiments-Results and Discussion 38 4.1 Metallographic Analysis 38 4.2 Direct Observation of Corrosion Performance 42 4.3 E OC Trends and Corrosion Behavior 44 4.4 Impedance Behavior 48 4.5 Long Term Trends and Continuation Studies 54 Chapter Five Conclusions of Deep Dr awn Aluminized Steel Experiments 56 References 57 Appendices 59 Appendix A: Sample Inventory 60 Appendix B: EIS Results from Dupli cate Dimpled Specimens in Solution C 61 Appendix C: EIS Results from Dupli cate Dimpled Specimens in Solution P 63 Appendix D: Metallographic Examination of Dimpled Samples. 65
iii List of Tables Table 2-1 Inventory of Specime ns Tested in the Sponge-Cell 16 Table 2-2 Inventory of Specimens Tested in the Immersion-Cell 19 Table 3-1 Solution Compositions and Properties 34 Table A-1 Number of Specimens Exposed in Each Category 60
iv List of Figures Figure 1-1 Aluminized Steel Type 2 Microstructure 3 Figure 1-2 Spiral Rib Pipe Manufacturing Proc ess (Dies Used to Form Rib (Left), Forming Spiral Pipe (Right)) 4 Figure 1-3 Minor Scratches During Fabrication Spiral Rib Pipe 4 Figure 1-4 Corroded Pipe from the Curl ew Rd. Site, Pinellas County, Fla. 6 Figure 2-1 Bending Process of Metal Strips 15 Figure 2-2 Sponge-Cell Confi guration-Bent Samples 17 Figure 2-3 Flat Samples in Sponge-Cell Set Up (Exposed Area 12 cm 2 ) 17 Figure 2-4 Photograph of a Sponge -Cell with a Bent Sample 18 Figure 2-5 Configuration of a Mild Bend Specimen for the Immersion-Cell System 19 Figure 2-6 Test-Cell of the Immersion-Cell System 20 Figure 2-7 E OC Trends with Exposure Time in Sponge-Cell System 21 Figure 2-8 E OC Trends with Exposure Time in Immersion-Cell System 22 Figure 2-9 EIS Behavior of Flat Specimens in Sponge-Cell System 23 Figure 2-10 EIS Behavior of Mild Be nt Specimens in Sponge-Cell System 23 Figure 2-11 EIS Behavior of Sharp Be nt Specimens in Sponge-Cell System 24 Figure 2-12 I Corr Trend of Specimens with Variou s Radii in Sponge-Cell System 24 Figure 2-13 EIS Behavior of Bent Specimens in Immersion-Cell System 25 Figure 2-14 I Corr Values of Mild-Bend and Two Sharp-Bend Specimens in the Immersion-Cell System 25
v Figure 2-15 Immediate Sign of Corrosion at Bend Region in Immersion-Cell Samples 26 Figure 3-1 Dimpling Specimens Using a Press-Sketch Shows Dimensions for the 1-in Ball Case 30 Figure 3-2 Forming Assembly Used to Dimple Specimens 31 Figure 3-3 Duplicate Aluminized Steel Spec imens (Left to Right) 1-in, -in, 9/16-in and Flat 31 Figure 3-4 Three-Electrode Cell Configuration 32 Figure 3-5 Simply Bent Metallographic Specimens 35 Figure 3-6 Example of Digitalized Im age Used for Finding the Radius of Curvature 36 Figure 3-7 Effect of Bending on the Aluminum Coating of a Sharp Bend Sample 37 Figure 4-1 Aluminized Steel Metallogra phy, Flat Specimen (left), Bent Specimen (right) 38 Figure 4-2 Strain and Radius Relationship 39 Figure 4-3 Outer Strain ( I per Eq(1), solid line) and M per Micrographic Analysis of Intermetallic Layer gaps, symbols) as Function of Radius of Curvature for Several Samples 41 Figure 4-4 Selected Microscopic Photos of Outer Layer of Specimens Introduced in Figure 4-3 42 Figure 4-5 Severe Corrosion at a 1-in Convex Dimple Exposed to Solution C 43 Figure 4-6 Light Corrosion at a 1-in Convex Dimple Exposed to Solution P 43 Figure 4-7 E OC Trends versus Exposure Time in Solution C 47 Figure 4-8 E OC Trends versus Exposure Time in Solution P 47 Figure 4-9 CPE Model Used to Fit the EIS Results 48 Figure 4-10 EIS Results of the Flat and 1-in Indentation Specimens Exposed to Solution C 49
vi Figure 4-11 EIS Results of the Flat and 1in Indentation Specimens Exposed to Solution P 50 Figure 4-12 Polarization Resistance Values for Different Bent Categories at Two Exposure Times in Solution C 51 Figure 4-13 Polarization Resistance Values for Different Bent Categories at Two Exposure Times in Solution P 52 Figure 4-14 I Corr Trend of Deep Drawn Specimens with Various Radii in Solution C 53 Figure 4-15 I Corr Trend of Deep Drawn Specimens with Various Radii in Solution P 53 Figure 4-16 i Corr (Nominal Current Density) Trend of Deep Drawn Specimens with Various Radii in Solution C 55 Figure B-1 EIS Results from Duplicate Dimp led Specimens in Solution C at 4 Days Bode (Top) and Phase Angle (Bottom) 61 Figure B-2 EIS Results from Duplicate Dimp led Specimens in Solution C at 10-Days Bode (Top) and Phase Angle (Bottom) 62 Figure C-1 EIS Results from Duplicate Dimp led Specimens in Solution P at 7-Days Bode (Top) and Phase Angle (Bottom) 63 Figure C-2 EIS Results from Duplicate Dimp led Specimens in Solution P at 21-Days Bode (Top) and Phase Angle (Bottom) 64 Figure D-1 Dimpled Specimen Cut with Hack Saw 65 Figure D-2 Side View of Dimpled Samples Cut at Indentation Region Perpendicular to Metal Rolling Direction 65 Figure D-3 Metallographic Dimpled Samples 65 Figure D-4 Low Magnification Microscopi c Photos of a Dimpled Specimen 66 Figure D-5 Radius Mapping of 9/16-in Dimpled Sample 66 Figure D-6 High Magnificati on Microscopic Photo of 9/16-in Dimpled Cell Before Exposure at Indentation Region 67 Figure D-7 Metallograph of Corroded 9/16in Dimpled Cell at Indentation Region 67
Corrosion Assessment of Mechanically Formed Aluminized Steel Mersedeh Akhoondan ABSTRACT Ribbed steel pipes made of Type 2 al uminized steel are commonly used for culvert pipes for highway drainage. Typically aluminized steel pipes have shown good durability and are expected to have long service life, e.g. 75 years; also, they are used in a wide variety of soil and wate r conditions. However, early corrosion of aluminized steel pipes has been recently observed in some inland locations. Initial observations showed severe corrosion in forms of pits, both along the ribs and at the nearby flat portions of the pipes. It is critical to determine the cause of early deterioration and establish methods of durability prediction. The possibility of unusual environmental conditions is being investigated elsewhere, but this research focuses on possible mechanical factors aggravating corrosion, since it is prevalent n ear pipe rib deformations. While forming the rib bends in the pipe, the outer bend surface is exposed to extreme tensile stresses which would cause small coating cracks (microfissure s) exposing base metal. Those may lead to early corrosion as galvanic protection fr om the surrounding aluminum may not be sufficient under certain environments. vii Electrochemical impedance spectroscopy was used to measure corrosion rate of both formed and flat aluminized steel samples in simulated natural waters. Initial findings vii
viii show that specimens formed by spherical inde ntation were susceptible to early corrosion development in moderately aggressive simulate d natural water, but not in a more benign, precipitating simulated natural water solution.
1 Chapter One Introduction Aluminum and aluminum alloys represen t an important category of materials due to their high technological va lue and wide range of industria l applications. Aluminum is also used as metallic coating for steel. It has been shown that aluminized steel is superior over galvanized steel sheets in corrosion resistance (Cerlane k and Powers, 1933) with up to 6 times longer service life. In galvanized st eel the zinc coating is subject to continuous corrosion to provide protection, while in alumin ized steel corrosion resistance is provided mainly by a stable thin film of aluminum oxi de. If this film is damaged or removed by abrasion, another layer of oxide is expected to form instantly to avoid further corrosion. Two types of aluminized steel are produced: Type 1 which have an aluminum coating with addition of 10% silicon to provide corr osion as well as heat resistance. Aluminized Type 1 is commonly used in hi gh temperature automobile i ndustry applications (muffler systems and heat exchangers).Type 2, however has aluminum coating and its typically applications are commercial roofing and culvert pipes. 1.1 Background on Aluminized Steel Type 2 Aluminized steel Type 2 is produced as a steel sheet, hot-dip coated on both sides with commercially pure aluminum (ASTM A929 and ASSHTO M274), which provides corrosion protection through low co rrosion rate of the aluminum when the aluminum is in passive condition, and also may confer galv anic protection to th e exposed underlying
2 steel under certain circumstances (Kimoto, 1999). For that reason, aluminized steel Type 2 is increasingly used for metallic drainage components in contact with natural waters. However, corrosion is an im portant durability limitation f actor in these components, which are often designed for very long servic e life (e.g. 75 yrs) (Cerlanek and Poweres, 1993). Microscopic examination of aluminized steel Type 2 in cross section shows a nearly pearlite-free ferrite low carbon steel substrate with regular grains, a partly columnar brittle inner alloy layer ~15 m th ick, and a softer outer aluminum-rich layer ~30 m thick. The inner alloy layer, commonly of composition Fe 2 Al 5 (An, Lui and Sun, 2001), (Li et al., 2003) is an essential ingredient of th e coating protection system, supplementing the outer aluminum-rich laye r and possibly providing a second line of defense against corrosion. The composition of the outer layer is predominantly nearly pure aluminum with Fe-rich intermetallic precipitates (6-11 wt% Fe) (Caseres and Sags, 2005). During manufacturing and/ or handling of the final material, discontinuities in the aluminized coating can extend to the substr ate steel, creating coating breaks. Those coating breaks exposing the steel base may result in the formation of galvanic macrocells. However, if the envi ronment is mild as t hose commonly found in Florida inland waters, sacrific ial protection to the exposed underlying steel may not be sufficient to prevent early corrosion as visually noted by the growth of rust projections at the aluminized coating breaks.
Figure 1-1. Aluminized Steel Type 2 Microstructure. (The oute r layer contains additional small amounts of aluminum-iron intermetallic precipitates). 1.2 Spiral Ribbed Pipe-Forming Process The ASTM 760 specification applies to the fabrication of ribbed pipes. The general process used to form the spiral ribs are as follows: The stock aluminized sheet is rolled over a series of press dies while it is lubricated with a soapy solution in order to decrease friction. Such construc tion creates open type ribs. In terlocking folds are formed in the opposite side of the sheet. As the pipe is rolled into the spiral the interlocking fold connects with a corresponding fo ld on the rib. These forme d, not-welded interlocks are called lock-seams and join the se gments of spiral pipes. 3
Figure 1-2. Spiral Rib Pipe Manufacturing Pr ocess (Dies Used to Form Rib (Left), Forming Spiral Pipe (Right )). [Photograph by the Author] During the fabrication of pipes, minor scratches and surface blemishes due to the rolling process and handling of the pipes are normal. As shown in Figure 1-3, aluminized stock rolls over against other metal component s during the forming process, and signs of minor scratching are readily observed. Figure 1-3. Minor Scratches During Fabrication Spiral Rib Pipe. [Photograph by the Author]. 4
5 1.3 Durability of Corrugated Culvert Pipe 1.3.1 Field Investigations Aluminized steel is commonly used for me tallic culvert pipes. This type of pipes is expected to have a long range service life from decades to beyond. The recommended condition for application of these pipes is pH rage of 5-9 and soil resistivity above 1500 ohm/cm (Kimoto, 1999). The minimum gage thickness for steel culverts and storm drains should be no less than 16-gage (Bednar, 1989). The gage thickness required for all performance modes should be determined fo r each installation base d on soil type, height of fill, and compaction of that soil (Cerla nek and Powers, 1993). The FDOT culvert pipe durability and materials selections guideline s (Cerlanek and Powers, 1993) project very long service life for alumini zed culvert pipe under thos e conditions. Earlier field investigations (Bednar, 1989) also indicate good prognosis for extended service life. However, recent field inspections conducte d by FDOT on four-year old spiral rib aluminized steel Type 2 culvert pipes expos ed to Florida inland waters have shown unexpected severe corrosion damage of aluminized pipe in the form of pits and generalized corrosion at the pipe ribs (where material form ing was more intense) and less so but still noticeable corrosion damage at the nearby flat portions of the pipe. The cause of that deterioration is under investigation, of which the present work is a part.
Figure 1-4. Corroded Pipe from the Curlew Rd. Site, Pinellas County, Fla. (2007) [Photograph by the Author]. 1.4 Review of Relevant Information on Ch aracteristics and Corrosion Performance of Aluminized Steel. 1.4.1 The Morphology of Intermetallic La yer in Hot-dip Aluminized Steels An important factor which directly a ffects the mechanical characteristics of aluminized steel is the thickness of intermet allic layer, in between steel substrate and outer aluminum layer. The in termetallic layer, wh ich is and alloy of aluminum and iron, is produced during the immersion of the steel base in molten aluminum. This layer is brittle in nature; therefore, it is desired to keep the thickness of this layer to a minimum (Awan and Faiz, 2006). In hot-dip aluminizin g with pure aluminum, such as the stock used for this investigation, the interlayer is thick and exhibits finger-like columnar crystal growth into the steel (Kimoto, 1999). SEM investigati ons illustrated that these finger-like features are composed of Fe 2 Al 5 Studies showed that the presence of silicon in the molten aluminum decreases the thickness of the intermetallic layer and promotes the formation of a cubic Fe 3 Al phase. The thickness of intermetallic layer may be also 6
7 associated with the carbon content of the steel substrate; as the carbon content increases the thickness of this layer decreases. Since aluminized steel Type 2 is produced by low carbon steel and nearly pure aluminum coating with no silicon, the in termetallic layer has an appreciable thickness, in order of 15 m. As shown later, fractur e of this brittle layer is a conspicuous manifestation of forming stra in in the aluminized steel evaluated here. 1.4.2 Formability of Metal-Coated Steel Sheets In order to study the e ffect of bending on alumini zed steel in the present investigation, flat metal sheet stock was hemispherically indented to various radii (Chapter Four). The deformation and strain analysis of stressed regions under that complex biaxial regime is in progress and detailed analysis will be presented in the future. That analysis will be based in part on prior work on deformation of coated steel. For example, comparable studies have been done by Gupta and Kumar (2005) on formability of galvanized inte rstitial-free steel. Galvanized steel is produced by hotdipping a steel substrate in molten zinc (450 C 460 C) to form an intermetallic bond between the zinc coating and the steel. As indicated by the aut hors, interstitial-free steel is produced by addition of titanium and/or niob ium to an extra low carbon grade steel to precipitate interstitial carbon and nitrogen atoms (Gupta and Kumar, 2006). Such materials are commonly used in the automobile industry due to their formability and high temperature corrosion resistance. In this i nvestigation, Heckers simplified technique which includes three steps: grid marking th e sheet samples, punch stretching the grid marked samples to failure or onset of localized necking, and measurement of strain(Gupta and Kumar, 2006). Grid marki ng of samples was done by printing a circle on the surface of the sheet before punch stretching. Samples were punched using
8 hemisphere dies. After stretching the sheet the grid circles turned into ellipses. To obtain the strain distribution of formed samples, long itudinal strain of ellipses were measured. It was found that the ellipse which is located right under the punch suffers from a minimal strain. The result illustrated that the strain s of uncoated sheets are higher than coated sheets. This is due to presence of a brittl e intermetallic Fe-Zn layer. A similar method will be applied in future work to analyze the strain of dimpled aluminized steel specimens. Appendix D illustrates the work in progress to establish the relationship between the bending radius and strain in dimpled cells. 1.4.3 Corrosion Mechanism of Aluminized Steel A recent review of the corrosion behavior of aluminized steel has been conducted by Caseres (Caseres, 2007). Hi ghlights of that work are ex cerpted and summarized in the next indented section, para phrasing as appropriate: Legault and Pearson (1978) conducted a fi ve-year investigati on on atmospheric corrosion of aluminized steel Type 2. In th at investigation aluminized steel test panels with uncoated cut edges (exposi ng the base steel) were exposed to industrial and marine environment. Small (~0.2 m/yr) and moderate (~0.45 m/yr) corrosion rates were observed in industrial and marine environments, respectively. The cut edges were free of corrosion in marine environments and illustrated rust formation in industrial environments due to insufficient galvanic protection to the exposed st eel (Legault and Pearson, 1978). Similar specimens were tested by Townsend and Borzillo (1987) in severe marine, moderate marine, rural, an d industrial environments for 13 years.
9 The authors concluded that in aggressive environments, the aluminized coating is anodic to the exposed steel where chloride ions impair the passivity of aluminum. However, in industrial and rural atmosphere s the aluminized coa ting passivated so that little to none galvan ic protection to the underlyi ng steel was noted (Townsend and Borzillo, 1987). Johnsson and Nordhag (1984) also perf ormed a four-year investigation to compare the sacrificial corrosion performan ce of several metallic coatings on steel exposed to atmospheric environments a nd seawater. Weight loss measurement was used to determine the corrosion rates of uncoated cut edges aluminized steel specimens with and without scribe marks, exposing underlying steel. The corrosion of aluminized steel even after one year was mostly in form of pitting, especially in a marine atmosphere. The investigation showed that except in marine environments, the aluminized steel without scribe marks outperformed the galvanized steel. The scribed specime ns demonstrated the poor galvanic protection of the aluminized coating to the exposed steel in all environments and seawater, in the form of heavy red rust formation along the cut edges and at the scribe mark (Johnsson and Nordhag, 1984). The majority of the studies on galvan ic corrosion of aluminized steel were conducted by atmospheric exposure or by im mersion in highly aggressive solutions. Limited information exists on the galvanic be havior of aluminized steel with coating breaks exposed to fresh waters of varying scaling tendencies with moderate chloride contents, where galvanic protection may not take place at all (Caseres, 2007).
10 Therefore, further research was conducted by the University of South Florida to address these issues, in the form of the inve stigation summarized in section 1.4.4. Another recent investigation for methods to increase the corro sion durability of aluminized steel was conducted by Hong et al., (2002) .The i nvestigation concluded that the addition of 1% Cr to aluminum coating increases the durability of aluminized steel Type 2. In fact, addition of Cr in the coating produces Al 13 Cr 2 intermetallic layer which replaces Al 5 Fe 4 layer. In this experiment samples we re exposed to 3.5% NaCl solutions at room temperature. It was found that in alumin ized steel with no addition of Cr, the outer aluminum coating initially corroded, follo wed then by the intermetallic layer (Al 5 Fe 4 ). In samples with 1% Cr coating the aluminum layer also started to corrode first; however, corrosion stopped when it re ached the intermetallic Al 13 Cr 2 (Hong, 2002). Therefore, addition of Cr has a positive impact on the dur ability of aluminized steel materials. 1.4.4 Previous Studies Conducted by the University of South Florida An investigation just completed at the University of South Florida (Caseres, 2007) addressed the issue of corrosion behavior of aluminized steel Type 2 with coating breaks (made by milling the aluminized steel sheet ) of significant size (2 mm to 2 cm in diameter), compared with that of an unbl emished aluminized steel surface. In that investigation samples were expos ed to the same solutions used in the present work. The main findings are summarized in the following: An extremely low corrosion rate was det ected for aluminized steel stock with no coating breaks exposed to solutions with m oderate Cl content a nd high alkalinity/high hardness with consequently high scaling te ndency. In solutions with moderate Cl content and low alkalinity, early pitting of the outer aluminized layer and strong
11 surface discoloration were observed. Howe ver, discoloration resulted from a momentary pH increase of the solution early on. In conditions where the pH remained near neutral, discoloration was delayed. In solutions with high scaling tendency and moderate chloride content, intense early steel corrosion at both small and large coa ting breaks was observed. The solution was benign and did not promote passivity loss of aluminum, so no cathodic protection of the exposed steel took place. However, in several instances there was delayed activation of the aluminum after long exposure times (e.g. 2,000 hrs), after which cathodic protection of the expos ed steel took pla ce. Little or no st eel corrosion was observed in solutions aggressive enough to cause early passivity breakdown of aluminum. No clear pattern was established between corrosion rates and the size of the coating break. The major objective of previous studies wa s to predict the corro sion behavior of flat stock aluminized sheet as a function of the scaling tendency in simulated natural waters, whereas, this investigation seeks to predict the corrosion performance of formed aluminized steel exposed to various bending stresses. As noted earlier, formed samples have micro breaks on the surface due to lack of ductility of brittle intermetallic interface. Formed samples are representative of conditio ns in corrugated and ribbed pipes which are also exposed to bending stresses during manufacturing. In previous studies it was shown that la rge coating breaks, intentionally created on the surface of metal, woul d dominate the corrosion perf ormance of the system. These breaks are significantly larger than those mi crofissures created due to bending stresses;
12 therefore, the corrosion trends predic ted for large coating breaks may not be representative of the behavior of bent samples with small mi cro cracks. That issue will be investigated here. 1.5 Objective of This Investigation This work aims at determining the co rrosion behavior of mechanically formed aluminized steel Type 2. The microstructura l aspects of mechanic al distress upon bending as a precursor to corrosion development are characterized. Specimens with various bending radii of curvature are electrochemica lly assessed in waters of varying scaling tendencies with moderate chlo ride content, resembling co mpositions typically found in inland Florida waters. This investigation is in progress and initial findings are presented here, to be amplified in the future as additional information is developed. 1.6 Approach To isolate the effect of bending, a set of aluminized steel samples are bent to various radii and are examined under enviro nments representative of field conditions. The corrosion rate of these samples is mon itored by visual inspection and electrochemical impedance spectroscopy. Also, in future wor k, the corrosion of existing field and factory samples will be analyzed and characterized. The investigation was conducted in two stages. A series of preliminary experiments were made using simply bent sp ecimens and sponges saturated with the test solution or epoxy-cast stoppers to delimit the zone of the bent region exposed to the test solution. Those experiments serv ed as ranging test s to obtain initial indication of the
13 effect of bending on corrosion performance. However, the preliminary tests also indicated that it was very difficult to delim it the portion of the specimen being exposed, with consequent uncertainty due to formation of crevice corrosion in regions of the specimen not related to the bent condition. Consequently, specimen design was changed to use flat sheet stock dimpled at the center to a controlled deformation degree. That design avoided unwanted crevice corrosion and proved to be a reliable and si mple procedure for determining effects of forming on corrosion performance. The main corrosion characterization technique used was Electrochemical Impedance Spectroscopy (EIS). Findings from the preliminary and the final specimen designs are presented in separate chapters.
Chapter Two Preliminary Experiments and Results The following describes the procedur es and findings of the preliminary investigation of the effect of simple bending on corrosion of aluminized steel. Additional details have been presented elsewhere (Akhoo ndan and Sags, 2007). It is noted that these experiments were semiquantitative in nature, and generally aimed at developing an adequate test set up for relia ble testing. The final method, us ing dimpled or "deep drawn" specimens, is detailed in the next chapters. 2.1 Bending Process In the preliminary experiments, 16-gage (1.58 mm thick) aluminized steel sheet was cut to rectangular strips, ty pically, 2-in x 6-in, 2-in x 3-in and 1-in x 6-in. Next, 2-in x 6-in, and 1-in x 6-in samples were bent to different radii using a bending jig. For sharper bends, after specimens were pre-bent as above, U shaped strips were pressed in a vise to obtain a sharper radius. Th e 2-in x 3-in strips were not bent, but used instead as flat specimens to compare the corr osion behavior of bent and unbent materials. The specimens were classified depending on the bending radius as sharp-bend (r=1.5 to 3 mm), mid-bend.(r= 3 to 5 mm), large (r=5 to 15 mm) and flat (r > 50 mm). 14
Figure 2-1. Bending Process of Metal Strips 2.2 Electrochemical Cell Set Up and Test Solution Two types of electrochemical cell systems were designed to conduct the preliminary experiments. The electrolyte c hosen for both cells was a normally protective solution (designated "P") used in the prev ious work by Caseres (Caseres and Sags, 2005). This solution tends to precipitate a prot ective carbonate film on the metal, but it also contains ~ 372 ppm chloride ions. The solution P is made from de-ionized water of resistivity > 10 6 -cm to which NaHCO 3 HCl, and Ca(OH) 2 are added as indicated in Table 3-1. The pH stayed almost constant in the range of 8 to 8.5 during the experiment. The solution in the cells was naturally aerated and the temperature was 21-24 o C. 2.2.1 Sponge-Cell System The sponge-cell system re presents conditions that might be encountered in the outer (soil side) surface of alumin ized steel culvert pipes. In this case, the sponge acts as a porous medium which absorbs, transfers and holds the solution to the surface of metal. 15
16 Sponge-cell samples were assembled as in Figure 2-2. In order to prevent exposing the cut sheet edges, which because of the exposed steel are not representative of the main surfaces, they were covered with a two component (EP-308, Thermoset Plastics Inc., Indianapolis, IN) industr ial epoxy resin and allowed to set for 24 hrs. Then the samples were degreased with etha nol and acetone, and stored in a desiccator before use. A low impedance activated titanium reference electrode (Castro et al., 1996) was firmly embedded in the sponge 1 cm away from the bend, and periodically calibrated against a saturated calomel reference electrode (SCE). The tip of the titaniu m reference electrode used for the EIS tests is embedded in the sponge firmly 1 cm away from the bend. The counter electrode is a stainless steel wire which is in contact with the electrolyte. Open circuit potential monitoring as well as periodic Electrochemical Impedance Spectroscopy (EIS) tests (f requency range from 100 kHz to 1 mHz using sinusoidal signals of 10 mV rms amplitude) were conducted by Gamry PCI4-300 potentiostat on this system to measure the corrosion rate and compare the results from both flat and bent samples. The type and number of samples test ed in the sponge-cell experiments are listed in Table 2-1. Table 2-1 Inventory of Specime ns Tested in the Sponge-Cell Flat SharpBend MildBend Exposure Time Total No. of Specimens Tested 4 3 9 Specimens not showing crevice effects (results reported for these only) 2 2 3 288 840 hrs
Ref. Aluminized Steel Type 2 Sponge Epoxy wicking Sponge Epoxy Working Electrode wicking Electrode Sponge Epoxy P Solution Counter Electrode (Stainless Steel) Wicking 200 cc Solution Figure 2-2. Sponge-Cell Configuration-Bent Samples Aluminized Steel Type 2 Epoxy Sponge Wicking Figure 2-3. Flat Samples in Sponge-Cell Set Up (Exposed Area 12 cm 2 ) 17
Figure 2-4. Photograph of a Sponge -Cell with a Bent Sample 2.2.2 Immersion-Cell System The immersion-cell system was designed to represent conditions that might exist in the inner surfaces of pipes, which are expos ed to flowing or standing water. The same bending process described earlier was used for these samples as well. However, the edges of specimens were completely covered wi th a highly adherent metallographic mount epoxy (Buehler Epoxicure, Buehler, USA), in stead of the EP-308 industrial epoxy which tended to lose adherence early during the tests. Figure 2-5 illu strates the configuration of the immersion-cell specimens. An activated titanium refere nce electrode as described earlier was placed in the solution 1 cm away from the specimen and periodically 18
calibrated against an SCE. The electrolyte was the same as used in the sponge-cell system. The type and number of samples tested in the sponge-cell experiments are listed in Table 2-2. Table 2-2 Inventory of Specimens Tested in the Immersion-Cell Mild Sharp Exposure Time No. of Specimens 1 3 Specimens not showing crevice effects (results reported for these only) 1 3 408 1512 hrs Metallographic Aluminized Steel Type 2 Metallographic Epoxy Aluminized Steel Type 2 Figure 2-5. Configuration of a Mild Bend Specimen for the Immersion-Cell System 19
Figure 2-6. Test-Cell of th e Immersion-Cell System 2.3 Results and Discussion Many of the specimens in the sponge-cell systems developed crevice corrosion due to poor performance of the Ep-308 e poxy which allowed disbondment and moisture intrusion at the epoxy-metal interface. Specimens showing those symptoms were discarded. Therefore, even though 15 sample s were exposed in the sponge-cells, the results of only six samples were taken into account. 2.3.1 Open Circuit Potential Trends 20 The Open Circuit potential (E OC ) graphs illustrated that in both sponge-cell and immersion-cell systems, flat specimens tended to develop the lower potentials (approaching E OC values typical of aluminum). In contrast, bent specimens tended to develop higher E OC values, thus reflecting a mixe d potential approaching the E OC value of steel, a sign of preferentia l steel corrosion in a steel-al uminum couple (Caseres and Sags, 2005). Typically, the E OC of pure aluminum in neutral aerated water is in the order of -1 V (SCE), while E OC of actively corroding steel in aerated conditions is in the order of -0.6 V (Caseres and Sags, 2006). The polarizability of passive aluminum is
very high (Caseres and Sags, 2006), so in an Al-steel couple even very small amounts of exposed steel are likely to result in a mixed potential a pproaching that of only steel (Caseres and Sags, 2006). -1 -0.95 -0.9 -0.85 -0.8 -0.75 -0.7 -0.65 -0.6 -0.55 -0.5 0 100 200 300 400 TIME/ HRS SHARP BEND 1 SHARP BEND 2 MILD BEND 2 MILD BEND 3 FLAT 1 FLAT 2Eoc / V (vs SCE ) Figure 2-7. E OC Trends with Exposure Time in Sponge-Cell System 21
-1 -0.95 -0.9 -0.85 -0.8 -0.75 -0.7 -0.65 -0.6 -0.55 -0.5 0 100 200 300 400 TIME/ HRS SHARP ( NO CORR) SHARP (CORR) MILD BEND FLATEoc / V (vs SCE) Figure 2-8. E OC Trends with Exposure Time in Immersion-Cell System 2.3.2 EIS Trends The EIS results for the sponge-cell system are shown in Figure 2-9 to 2-11. Also, EIS results are illustrated in terms of the corrosion current I Corr in Figure 2-12. Apparent Rp values (Rpa) were assumed to be equa l to the impedance modulus |Z|@ 0.001 Hz minus the value of the solution resistance. The solution resistance Rs was assumed to be equal to |Z| @ 10 5 Hz. The underlying principles are described in standard literature sources and course summaries (Sags, 2006). The corrosion current I Corr was calculated from I Corr =B/(Rpa Rs) Eq(1) 22 where B = 26 mV (Typical value of the Stearn -Geary constant for corroding steel (Jones, 1996).
1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10 1.0E-031.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05|Z| / ohmf / Hz P4-13Days P5-14Days FLAT Figure 2-9. EIS Behavior of Flat Specimens in Sponge-Cell System 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+101.0E-031.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05|Z| / ohmf / Hz p1-27 Days p2-24 Days p6-12 Days MILD BEND Figure 2-10. EIS Behavior of Mild Be nt Specimens in Sponge-Cell System 23
1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10 1.0E-031.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05|Z| / ohmf / Hz BP1-30 Days BP2-30 Days SHARP BEND Figure 2-11. EIS Behavior of Sharp Be nt Specimens in Sponge-Cell System 0.01 0.1 1 10 UNBENT MILD SHARPI CORR ( A) Sponge-Cell Figure 2-12. I Corr Trend of Specimens with Variou s Radii in Sponge-Cell System 24
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E-03 1.0E-011.0E+011.0E+03|Z| / ohmf / Hz SP1-1Day SP1-3 Days SP2-1 Day SP3-1 Days SP2-2 Days SP1 : Mild Bend ( No Sign of Corr.) SP2: Sharp Bend ( No Sign of Corr.) SP3: Sharp Bend (Corr. At Bend) 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E-03 1.0E-011.0E+011.0E+03|Z| / ohmf / Hz SP1-1Day SP1-3 Days SP2-1 Day SP3-1 Days SP2-2 Days SP1 : Mild Bend ( No Sign of Corr.) SP2: Sharp Bend ( No Sign of Corr.) SP3: Sharp Bend (Corr. At Bend) Figure 2-13. EIS Behavior of Bent Specimens in Immersion-Cell System 0.01 0.1 1 10Mild Bend (No Sign of Corr. ) Sharp Bend (No Sign of Corr.)Sharp Bend (Corr. at Bend)I CORR ( A) Immersion-Cell Figure 2-14. I Corr Values of Mild-Bend and Two Sharp-Bend Specimens in the Immersion-Cell System 25
In both systems unbent specimens had th e greatest polarization resistance (Rp) values. Mild bend specimens had smaller Rp values compared to those of the flat samples. Finally, sharp-bend specimens had the smallest Rp compared to mild bend and unbent specimens. Sharp-bend spec imens have the highest I Corr (33-43 A) values. As the severity of bending increased the appare nt corrosion current also increased. Results correlated with the E OC trends. 2.3.3 Direct Observation of Corrosion The bend region was covered in the sponge -cell specimens so it was not amenable for direct examination early in the test. s ponge-cell specimens were not opened until late in the test exposure when extensive corrosion had taken place, so those results will not be considered. In the immersion-cell system, no crevice corrosion was observed in any of the samples. A sharp-bend specimen clearly showed corrosion signs (re d rust) at the bend region soon after exposure began. 26 Figure 2-15. Immediate Sign of Corrosion at Bend Region in Immersion-Cell Samples
2.4 Conclusions Obtained From Preliminary Experiments Preliminary findings indicated preferential deterioration in bent samples and damage associated with mechanical distress. E OC measurements in general showed more positive potentials for sharper bends, indicative of exposing steel at coating breaks. Sharp-bend specimens had highest corrosion rates compared to mild bend and flat specimens. Exposed steel (at microfissures due to the bending process) in a sharp-bend aluminized steel specimen experienced intense corrosion early in the test. The above findings were affected by significant uncertainty due to artifacts that led to the development of the next generation test-cell. 27
28 Chapter Three Deep Drawn Aluminized St eel Experiments-Methodology The following describes the procedures fo r the main phase of this investigation, which resulted in the development of an im proved dimpled specimen configuration. This work evolved from the preliminary study of th e effect of simple bending on corrosion of aluminized steel described in the previous chapter. The preliminary experiments showed that crevice corrosion occurred frequently at bent specimen edges, where the deformed surf ace needed to be delimited. That limitation was overcome in the present design by using flat specimens with a central controlleddeformation dimple. Surface delimitation took place at the flat por tion of the specimen, where crevice formation was less common due to use of a gasket assembly perfected earlier (Caseres, 2007). The work also seeks to quantify the effects of forming on microstructural damage, for later relating that damage to extent of resulting corrosion. E xperiments to date examine the effect of simple (uniaxial) bending and methodology, and results of that phase of the work are presented here. The effects of biaxial bending, as in dimpling, have only been examined in exploratory tests (see Appendix D) and will be explored in detail in the future. Additional details of this phase of th e work have been accepted for publication elsewhere (Akhoondan, Sags and Caseres, 2008).
29 3.1 Dimpling Procedure Circular 16-gage aluminized steel specimens with exposed area of 12.56 in 2 (82 cm 2 ), and with minimal as-received surface distress were hemispherically indented at the center using stainless steel ball-bearing balls with diameters of 1 in (2.54 cm) 3/4 in (1.90 cm), or 9/16 in (1.43 cm) as shown in Figure 3-1. The indentation was made by pressing the bearing ball, socketed in a steel plate into the initially flat specimen. An indented Teflon plate was used to prot ect the convex face of sample until full hemispheric penetration was achieved. The conv ex face was the one later exposed to the test solution. Afterwards, the specimens we re cleaned with ethanol and acetone, and stored in a desiccator prior to immersion. Fi gure 3-3 shows the appe arance of the dimpled specimens. Control flat specimens without surface indentation were also used for comparison. Multiple specimens were prepared for duplicate testing.
Figure 3-1. Dimpling Specimens Using a PressSketch Shows Dimensions for the 1-in Ball Case. 30
Figure 3-2. Forming Assembly Used to Dimple Specimens. (Picture shows the 9/16-in case). Figure 3-3. Duplicate Aluminized Steel Spec imens (Left to Right) 1-in, -in, 9/16-in, and Flat 3.2 Immersion-Cell Experiment A three-electrode test-cell (Figure 3-4) was designed for exposing horizontally the convex side of the specimen, where distress is expected to be worst because of the tensile 31
stresses. Corrosion of the concave side will be assessed in future experiments. A metalmetal oxide activated titanium mesh placed pa rallel ~6 cm from the specimen surface was used as a counter electrode A low impedance activated titanium pseudo reference electrode 0.3 cm diameter and 5 cm long (C astro et al., 1996) was placed ~1.5 cm above the specimens indentation and periodically calibrated against an SCE. All potentials reported here are in the SCE scale. Each test -cell was filled with 500 mL of solution, not replenished during the test for reason s explained below and in section 2.2.2. Figure 3-4. Three-Electr ode Cell Configuration The immersion tests were conducted in duplicate for each test solution, at 22 C. Solution pH was verified, and E OC was monitored The E OC measurements were taken 32
33 nearly daily and plotted as function of expos ure time to reveal transitions indicative of corrosion condition. EIS measurements were obtained at the E OC with a Gamry PCI4300 potentiostat in the frequency range from 100 kHz to 1 mHz using sinusoidal signals of 10 mV rms amplitude. The results for the first ~600 hrs and ~1,000 hrs of exposure are presented here for solutions C and P respectiv ely as well as limited initial EIS data. Exposure continues and long-term performa nce will be presented in a subsequent publication. 3.3 Solution Compositions Two test solutions (Table 3-1) were prepared from decarbonated de-ionized water of resistivity > 10 6 -cm combining reagent grade NaCl and NaOH (solution C of negligible carbonate precip itating tendency) and NaHCO 3 HCl, and Ca(OH) 2 (solution P of high carbonate prec ipitating tendency). These solutions are identical to those used by Caseres (Caseres, 2007), who showed that the Langelier Index was -5.9 for solution C and +1.50 for solution P, consiste nt with the observation of a powdery precipitate layer of CaCO 3 uniformly distributed on the specimen surf ace shortly after initiation of exposure. The test solutions in the test-cells were quiescent and naturally aerated through a small opening. As in previous work by Caseres (Caseres, 2007) and noted elsewhere (Akhoondan, Sags and Caseres, 2008), the rela tively small electrolyte volume/total specimen area ratio was intended to be repres entative of, for instance, worst-case culvert pipe conditions with stagnant water on the pi pe invert, or of occluded conditions for pore water on the soil side of a pipe.
34 Table 3-1. Solution Compositions and Properties Solution TA TH FC BI pH Ca +2 mg/L Cl mg/L C (Control) 6 2 0 8 ~7.4 0 P (Precipitating) 184 52 13 223 ~7.4 200 372 Preparation (Per 1 L of Solution) 0.05 mL of 0.01 M Na(OH) C 5.25 mL of 2 M NaCl 100 mL of 0.1 M NaHCO 3 0.37 g of Ca(OH) 2 P 105 mL of 0.1 M HCl Legend: TA: total alkalinity (mg/L) of CaCO 3, TH: total hardness (mg/L) of CaCO 3, FC: free CO 2 : solution conductivity 3.4 Metallography Microstructural assessme nt of mechanical distre ss due to bending has been performed to date with 16-gage (1.58 mm thick) Type 2 aluminized steel sheets simply bent to various extents. The bent region was mounted metallogra phically and examined with an optical metallographic microscope.
Figure 3-5. Simply Bent Metallographic Specimens 3.4.1 Metallography Process To prepare the metallographic samp les the following steps were executed: First, the specimens were cut out from th e sheet stock using a manual hack saw and carefully avoiding marring of the surfaces ; then the specimens were degreased with ethanol. Next, they were molded in (Buehler Epoxicure, Buehler, USA) metallographic epoxy. After curing for 24 hrs, the specimens were demolded and were ready for grinding and polishing. Abrasive papers (SiC 80, 120, 600, 1200 grit, Struers Inc, USA) were used to grind the samples in a Buehler Ecomet III turntable. During grinding tap water was used as a lubricant. Then specim ens were polished using (SiC 2400, 4000 grit, Struers Inc, USA) papers and a lubricant (D-P Lubricant, Blue, Struers Inc, USA) which is non-aqueous. Final polishing steps were conducted using a cloth (MD-Dac, Struers Inc, Denmark), with Dp Blue Lubr icant and, and 1m diamond pa ste (Dp-paste, P, Struers Inc, Denmark). An even force of about 20 N on a medium speed wheel was applied to accomplish a better result. A demonstration samp le preparation was also made by Struers 35
(Struers Inc, USA) using automated equipm ent. After the samples were polished and cleaned, the cross section of the outer bent was examined at low magnification and digitized to determine local radius of curvat ure (Figure 3-6). The same regions were then examined at high magnification to determ ine the type and extent of distress, as exemplified in Figure 3-7. Figure 3-6. Example of Digita lized Image Used for Finding the Radius of Curvature As it is explained in the next chapter, the calculated strain for each simply bent sample was obtained by assuming a constant neut ral axis located at the center of metal. Future experiments and calculations will address the deformation of samples formed by spherical indentation as those explained in section 3-1. The analysis of such deformation is more complex, since these dimpled sample s were exposed to biaxial stress. A more complete analysis is being inve stigated in continuation work. 36
Figure 3-7. Effect of Bending on the Alum inum Coating of a Sharp Bend Sample. (Analysis emphasis was on the convex portion of the specimen; inner portion is shown for illustration of other forms of damage to be examined in the future). 37
Chapter Four Deep Drawn Aluminized Steel E xperiments-Results and Discussion 4.1 Metallographic Analysis Figure 4-1 (left) shows a metallographic cross section of the as-received (flat) material showing the coating microstructure described in the Introduction. Figure 4-1 (right) shows the coating condi tion of a sharply unaxially be nt sample (radius ~2 mm). The most notable feature is nearly completely br ittle fracture of the intermetallic layer, as manifested by its many elongation gaps. The outer coating was much more ductile, with a few instances of ductile fracture such as that illustrated on the left side of Figure4-1, apparently originating at one of the larger brittle gaps of the inner layer. Figure 4-1. Aluminized Steel Metallography, Flat Specimen (l eft), Bent Specimen (right). Special Automated Sample Prepara tion Courtesy of Struers Inc. The number and severity of fractures in the inner layer incr eased as the bending stress increased. For completely brittle fract ure, the integrated gap length measured on a 38
line parallel to the perimeter, divided by the in itial length (that is, the sum of the lengths of the remaining visible interm etallic segments) should equal th e total strain at the outer fiber. To test that hypothesis, the strain M calculated from measurements of the ratio of inner layer gaps to visible in termetallic length was compared with the strain at the outer fiber expected from the local macroscopic be nding radius and the sheet thickness. For a simply bent sheet in which the neutral axis of metal sheet stays constant, the theoretical relationship between the strain of outer fiber and radius of curvature (Dieter, 1986) is: I = T 2 R T 2 Eq(2) where: I is the ideal strain T is the thickness of the sheet (16-gage = 1.58 mm) R is the outer bending radius The derivation of Eq(2) is illustrated in Figure 4-2 and the following text. Figure 4-2. Strain and Radius Relationship 39
40 R external = r 0 + T/2 r 0 = R external T/2 external = ( L 0 + L 0 ) / L 0 = ( R external r 0 )/ r 0 = (r 0 + T/2 r 0 )/ r 0 = T/2 / r 0 = T / [2 ( R external T / 2)] which is the same as Eq(2). Figure 4-2 compares the values of ideal outer fiber strain computed per Eq(1) from the radius of curvature (determined for various specimens from the low magnification micrographs) and the sheet thickness, and M from the micrographic analysis of the intermetallic la yer (Figure 4-5). Flat samples were assigned a radius > 50 mm in the graph. The results show good agr eement between measured and expected strain, supporting the hypothesis of nearly completely brit tle behavior for the inner intermetallic layer. Tests in progress address behavior of specimens bent by spherical indenting and ductile fracture of the outer la yer, which will be reported in the future.
0.01 0.10 1.00 10.00 11 0External StrainLog (Radius) / mm 1 0 0 Figure 4-3. Outer Strain ( I per Eq(1), solid line) and M per Micrographic Analysis of Intermetallic Layer gaps, symbols) as Func tion of Radius of Curvature for Several Samples 41
Figure 4-4. Selected Microscopic Photos of Outer Layer of Specimens Introduced in Figure 4-3 4.2 Direct Observation of Corrosion Performance Visual examination of the specimen surfaces in solution P (high total alkalinity and hardness, with consequent car bonate precipitation tendency) showed early light yellowish discoloration uniformly distri buted over the convex indentation surface in all dimpled specimens tested. The discolor ation was indicative of ongoi ng, but mild, corrosion there. 42
Symptoms were more severe in solution C (low alkalinity and hardness, with no precipitating tendency), where strong reddish discoloration was first noted early on the convex indentation of all dimpled specimens and gradually propagating covering the entire indentation su rface (Figure 4-5). Figure 4-5. Severe Corrosion at a 1-in Convex Dimple Exposed to Solution C. (Picture shows immersed sample surface at 19 days exposure). 43 Figure 4-6. Light Corrosion at a 1-in Convex Dimple Exposed to Solution P. (Picture shows immersed sample surface at 21 days exposure).
44 In either solution corrosion appeared to be more intense for the 1-in diameter indentation specimens. However, it is noted that the deformed area was greatest for those specimens even though maximum local strain may have been less severe than for the smaller radius indentation specimens. In both solutions, corrosion of the flat aluminized surface around the dimple and associated deformation was minimal, suggesting preservation of passive be havior there. Correspondingly, no sign of corrosion was observed either in any of the flat surface sp ecimens exposed to both solutions, suggesting good corrosion performance in those cases at le ast during the time period investigated to date. Clearly, any galvanic protection from the aluminized portion was insufficient to avoid initiation of active corros ion of the exposed steel especi ally in the so lution C, in agreement with the results reported elsewh ere for specimens with machined coating breaks (Caseres, 2007). It is noted that in the work by Caseres (Caseres 2007) activation of flat aluminized surfaces tended to take place in some circumstances after longer exposure times (e.g. 2,000 hrs or more). Future testing will examine that possibility. 4.3 E OC Trends and Corrosion Behavior The E OC tended to fluctuate upon laboratory temperature ch anges. In order to compensate for that temperature effect to better evaluate E OC trends with time, all potentials were converted to estimated values at 23 C o The conversion was made by first plotting the change in potenti al between consecutive day readings as function of the corresponding change in temperature. The data followed an approximately linear trend in that relatively narrow interval, with a trend slope of 0.0204 V/ o C. That result was thus approximated by:
(T+ T) = E (T) T. (0.0204) V/ o C Eq(3) (23 C 0 + T) = E (23 o C) (T m -23C 0 ). (0.0204) V/ o C Eq(4) E (23 o C) = E(T m ) + (T m -23C 0 ). (0.0204) V/ o C Eq(5) Where T m is the measured temperature and E(23 o C) is the estimated value that E OC would have adopted if the test temperature was 23 o C. This correction is an empirical adjustment to assist in revealing longer time potential trends, and it is not assigned mechanistic significance at present. However, future work will assess if the corresponding activation energy that can be derived from this analysis may provide an additional descri ptor of the kinetics of reactions relevant to the corrosion process. Figure 4-7 exemplifies the average E OC trends versus exposure time for the different indentation radii expos ed to solution C. It is no ted that in previous work (Caseres and Sags, 2006) with these solutions the E OC of plain steel was found to be in the order of ~-600 to -700 mV, while an unblemished aluminized surface tended to develop potentials typically > 100 mV more negative. In th e present experiments, a few hours after immersion the E OC values were ~-650 mV for all specimens. The potentials then decayed to reach values of < ~-800 mV for the flat specimens, and for the dimpled specimens of 9/16-in and -in indentation radii. The E OC value for the dimpled specimens with 1-in indentation radius was ~100 mV more positive compared to all the other specimens. That observation showed that the severity of the indentation (as pointed out above, apparently responding in th ese tests more to the total area affected than to how strong it is locally) played an im portant role in dicta ting the overall system potential. This potential elevation likely refl ected the presence of a substantial number of 45
46 microscopic aluminized coating breaks exposing the underlying steel base in the 1-in indentation specimens, which dominated the over all potential trends in those cases. The general declining trends of both flat and formed specimens, however, suggest that the potentials are approaching that of corroding aluminum as exposure time increases. In the more aggressive solution C, the aluminum coating may tend to depassivate and provide galvanic protection for steel; however, this pr otection if present was not sufficient during the test period examined so far to prevent the corrosion of bare steel exposed to the solution. Indeed, appreciable co rrosion of steel manifested as brownish discoloration was observed in the dimpled region, which is expect ed to have numerous microfissures. In solution P there was much scatter in the data but the flat samples had in average the most negative potentials compared to formed samples. There was an increasing E OC trend with time for all specimens in solution P, most notably for the formed specimens, suggesting that the E OC in those tended to appr oach the potential of corroding steel. In this benign environment the al uminum coating is likely to be in a well established passive condition, thus not providing any important galvanic protection for steel. Therefore the corrosion of unprotected exposed steel predominantly polarized the system and caused a elevating trend of potentials in the system. However, it is noted that the specimens in solution P exhibited only light discoloration at the indentation, so that the E OC trend toward more positive values may be eventually arre sted if the above interpretation is correct.
-0.850 -0.800 -0.750 -0.700 -0.650 -0.600 012345678910111213141516171819202122232425(EW-SCE)est 23 C/ VTime / Day Average 1-in Average 3/4-in Average 1/2-in Average Flat Figure 4-7. E OC Trends versus Exposure Time in Solution C -0.850 -0.800 -0.750 -0.700 -0.650 -0.600 -0.550 -0.500 024681012141618202224262830323436384042444648505254(EW-SCE)est 23 C/ VTime / Day Average 1-in Average 3/4-in Average 1/2-in Average Flat Figure 4-8. E OC Trends versus Exposure Time in Solution P 47
4.4 Impedance Behavior The analog circuit used to fit the EIS data is shown in Figu re 4-9. This analog 1 was chosen mainly for overall simplicity and having provided a reasonable account of the observed impedance spectra. A more complicated circuit is expected to be needed for later stages of the system e volution (Caseres, 2007). Rs represents the effective solution resistance. The analog circuit consists of a polarization admittance (Rp -1 ) reflecting the combined activation polarization of the anodi c and cathodic reactions, in parallel with a Constant Phase Element (CPE) representing th e interfacial charge storage at the metal surface. The proposed analog circuit yielded good best-fit simulations of the EIS responses in both solutions. Figure 4-9. CPE Model Used to Fit the EIS Results Figure 4-10 illustrates the EIS results of the flat and 1-in indentation specimens exposed to solution C at 4 and 10 days afte r exposure. The 1 mHz impedance moduli for the flat specimens showed incr easing trends attaining ~11 k after 10 days, consistent with generally passive behavior and the absen ce of visual evidence of active corrosion. In contrast, the 1 mHz impedance moduli for the 1 in indentation specimen in C were ~ < 2 48 1 In general, the analog circuit chosen to represent the impedance respon se for the present cases may not be unique; alternative analog circuits may explain equally well the observed impedance behavior.
k and roughly constant with exposure time. Comparison of results computed by fitting the data per the Figure 4-9 analog circuit fo r all the specimens are shown in Figure 4-12. The comparison shows that in solution C the dimpled specimens had distinctly lower Rp values (and consequently stronger indication of corrosion) than the flat specimens. These results are consistent with the visual obser vations and provided further evidence that forming distress, at the levels explored here, promotes corrosion of the aluminized composite in the non-scaling, more aggressive solution tested. -7000 -6000 -5000 -4000 -3000 -2000 -1000 0 0 2000 4000 6000 8000 10000 12000Z' / ohmZ" / ohm 1" Dimpled Sample-10 Days of Exposure 1" Dimpled Sample -4 Days of Exposure Flat Sample-10 Days of Exposure Flat Sample-4 Days of Exposure Figure 4-10. EIS Results of the Flat and 1-in Indentation Specimens Exposed to Solution C. Bode diagram is presented in Appendix B. 49
-12000 -10000 -8000 -6000 -4000 -2000 0 0 2000 4000 6000 8000 10000 12000Z' / ohmZ" / ohm 1-in Dimpled Sample-7 Days of Exposure 1-in Dimpled Sample-21 Days of Exposure Flat Sample -7 Days of Exposure Flat Sample -21 Days of Exposure Figure 4-11. EIS Results of the Flat and 1-in Indentation Specimens Exposed to Solution P. Bode Diagram Is Presented in Appendix C. Impedance diagrams of specimens in the more protective Solution P showed much larger low frequency impedance moduli th an in Solution C, and less differentiation between flat and formed specimens. The relative Rp values and variabili ty for the various dimple sizes in both solutions and at two exposure times is furt her illustrated in Fi gures 4-12 and 4-13. Figures 4-14 and 4-15 show the nominal cal culated corrosion currents assuming a SternGeary constant of 26 mV. Clearly the solution P is less aggressive overall and in it there is greater scatter and little differentiation between the flat and formed specimens. The 50
polarization resistance on average tended to increase with time (with matching decreasing time for the apparent corrosion current) attes ting to the protective character of this precipitating solution. In the more aggressive solution C, Rp values are about one order of magnitude lower (or I Corr values one order of magn itude higher) than in the solution P, and there is notably more apparent corrosion current in th e formed than in the flat specimens. There seems to be a trend of increasing Rp (decreasing I Corr ) with time but not consistent for all test conditions. 1 10 100 1000 0.56-in .75-in 1-in Flat Specimen 1 Age :4 Days Specimen 1 Age :10 Days Specimen 2 Age: 4 Days Specimen 2 Age: 10 Days Duplicates in C SolutionLog Rp/( Kohm ) Figure 4-12. Polarization Resistance Values for Different Bent Categories at Two Exposure Times in Solution C 51
1 10 100 1000 0.56-in .75-in 1-in Flat Duplicates in P Solution Specimen 1 Age :7 Days Specimen 1 Age :21 Days Specimen 2 Age: 7 Days Specimen 2 Age: 21 DaysLog Rp/(Kohm) Figure 4-13. Polarization Resistance Values for Different Bent Categories at Two Exposure Times in Solution P At this time, no clear differentiation in behavior has been established between specimens formed with different radii indentat ions, in part because of the early stage of the tests and also because the relative area of the mechani cally distressed zones has not been quantitatively determined. Long-term moni toring of specimens, and development of ways to account for surface area normalization of corrosion at the different indentation sizes, is in progress. An initial attemp t at normalization is presented next. 52
0.01 0.1 1 10 100 0.56-in .75-in 1-in FlatI CORR ( A) Specimen 1 Age :4 Days Specimen 1 Age :10 Days Specimen 2 Age: 4 Days Specimen 2 Age: 10 Days Duplicates in C (Dimpled Samples) Figure 4-14. I Corr Trend of Deep Drawn Specimens with Various Radii in Solution C 0.01 0.1 1 10 0.56-in .75-in 1-in FlatI CORR ( A) Specimen 1 Age :7 Days Specimen 1 Age :21 Days Specimen 2 Age: 7 Days Specimen 2 Age: 21 Days Duplicates in P (Dimpled Figure 4-15. I Corr Trend of Deep Drawn Specimens with Various Radii in Solution P 53
54 Although a precise evaluation of the area specific corrosion rate cannot be done at this time, on first approximation the expected corroding area can be assumed to be the same as the area of a circle with a diameter equal to the diameter of the indenting steel ball. Therefore, the corrosion current valu e obtained based on polarization resistance (Figures 4-14 and 4-15) can be divided by those nominal corroding areas to obtain a nominal corrosion current density i Corr In order to account for the apparent corrosion rate of the flat portion, the average corrosion cu rrent of control flat samples was first subtracted from the average corrosion current of the formed samples. The analysis was only performed for the specimens exposed to the solution C where flat-formed differentiation was apparent. The resulting nomin al current density of the formed region is shown in Figure 4-16. These results show greater nominal corrosion current densities, (by about one order of magnitude ) in the two lower indentation diameter cases than in the 1-in diameter case. This finding agrees with the expectation that more sharply formed material may attain a higher corrosion ra te. The highest nominal corrosion current densities are in the order of a few A/cm 2 which correspond to average metal penetration rates in the order of a few tens of micrometer s per year. Such rates, if sustained over the long term could be of concern in significantl y reducing the time for local perforation of a pipe wall. 4.5 Long Term Trends and Continuation Studies This investigation has been predominantly of exploratory character and served to set the methodology and preliminary trends toward long term evaluation of the effect of forming on corrosion performance of formed alum inized steel. The experimental set ups
created in this work will undergo continuing testing ove r extended periods (e.g. > 3,000 hrs) to establish long term ef fects of forming on corrosion. Of particular interest will be to assess whether activation of the aluminum takes place later on as seen elsewhere (Caseres 2007), which could have an important protective effect on the formed regions seen to undergo early corrosion in the tests performed to da te. Analysis will also be extended to better characterize the extent of mechanical damage for the dimpled geometry and that used in actual production ribbed culvert pipes. 0.01 0.1 1 100.56-in .75-in 1-inI CORR ( A)/cm2 Specimen 1 Age :4 Days Specimen 1 Age :10 Days Specimen 2 Age: 4 Days Specimen 2 Age: 10 Days Duplicates in C (Dimpled Samples) Figure 4-16. i Corr (Nominal Current Density) Trend of Deep Drawn Specimens with Various Radii in Solution C 55
Chapter Five Conclusions of Deep Drawn Aluminized Steel Experiments Forming by bending results in br ittle fracture of the inner (intermetallic) aluminized coating layer, in a manner consistent with predictions of simplified bending mechanics of the system. Failure of the outer (mostly aluminum) layer by ductile deformation was much less common. The convex portion of specimens formed by s pherical indentation was susceptible to early corrosion development in moderately ag gressive simulated natural water. Not formed (flat) aluminized control surfaces showed much less deterioration during the test interval (from visual a nd electrochemical evidence). Both formed and control specimens showed little corrosion damage when exposed to a more benign, precipitating simula ted natural water solution. Open Circuit Potentials in precipitating solutions with high precipitating tendency suggested a increasing trend in which poten tials are approaching the potentials of corroding steel. In more aggressive so lution with low precipitating tendency, potentials illustrated a declining trend approaching the poten tial of corroding aluminum. The above are preliminary findings from test s in progress. Confirmation as well as revealing the evolution of other modes of deterioration is pending on longer term evaluation with the test assemblies created in the present work. Additional Future work will focus on analyz ing the strain of deep drawn specimens, and establishing a relationship between th e strain and corrosion rate of exposed specimen. 56
References AK Steel, Aluminized Steel Type 2 Product Features, (2007) http://www.aksteel.com/pdf/markets _products/carbon/T2_Data_Bulletin.pdf Akhoondan M., Sags A. (2007), Corrosion of Mechanically Formed Aluminized Steel, Student Poster /2007, NACE Intern ational, Nashville, TN, 2007 (Archived 2 nd Place of Herro Award for Best Student Poster). Akhoondan M, Sags A., Cseres L. (2008), Corrosion Assessment of Mechanically Formed Aluminized Steel, submitted for publication as Corrosion/2008, Paper No. (1462), NAC E International, Houston, TX, 2008. Amercian Society for Metals, 1972, 8 th Ed., Vol. 7, Ohio. Amit Kumar Gupta, D. Ravi Kumar (2006), Fo rmability of galvan ized interstitialfree steel sheets, Journal of Mate rials Processing Technology, 172, pp. 225-237. An J., Liu Y.B., Sun D.R. (2001), Mechan ism of Bonding of Al Pb Alloy Strip and Hot-Dip Aluminised Steel Sheet by Hot Rolling, Materials Science Technology 17, pp. 451-454. Armstrong R.D., Braham V.J. (1996), the Mechanism of Alumin um Corrosion in Alkaline Solutions, Corrosion Science 38, pp. 1463-1471. Bednar L. (1989), Galvanized Steel Draina ge Pipe Durability Estimation with a Modified California Chart, Paper No. 88-0341, 68th Annual Meeting, Transportation Research Board, Washington, D.C. Caseres L. (2007), Electroche mical Behavior of Aluminized Steel Type 2 in ScaleForming Waters, Ph.D. Dissertation, University of South Florida. Caseres L. and Sags A. (2006), Galvanic Be havior of Type 2 Aluminized Steel in Simulated Natural Waters", L. Cancun, Mexico, "Corrosion of Infrastructure", The Electrochemical Society, Pennington, NJ, 2007, ECS Transactions Vol. 13, Issue 13, 210th ECS Meeting October 29-November 3, 2006, pp. 147-157. Caseres L., Sags A. (2005), Corrosion of Aluminized Steel in Scale Forming Waters, Paper No. 05348, 13 pp. Corrosion/2005, NACE International, Houston, 2005. 57
58 Castro P., Sags A ., Moreno E.I. Maldonado L., Genesca J. (1996), Characterization of Activated Titanium Solid Reference Electrodes for Corrosion Testing of Steel in Concrete, Corrosion 52, p. 609. Cerlanek W.D., Powers R.G. (1993), Drainage Culvert Service Life Performance and Estimation, State of Florida Departme nt of Transportation Report No. 93-4A. Creus J., Mazille H., Idrissi H. (2000), Po rosity Evaluation of Protective Coatings Electrodeposited on Steel, Surface and Coatings Technology 130, pp. 224-232 Dieter G., (1986), Mechanical Meta llurgy, McGraw-Hill, New York, pp. 659-662. Gul Hameed Awan, Faiz ul Hasan (Unive rsity of Engineering and Technology, Lahore), (2006), The morphology of coa ting/substrate inte rface in hot-dip aluminized steels, Materials Scien ce and Engineering, Vol. 390, pp. 437-443. Hong, Lim, Chung, Choi (2002), Texture and Corrosion Mechanisms of Aluminized Steel Sheets, Material Science Forum, 408-412, pp. 1031-1036. Hwang, Song, Kim (2004), Effects of car bon content of carbon steel on its dissolution into a molten aluminum alloy, Materials Science and Engineering 390, pp. 437-443. Johnsson T., Nordhag L. (1984), Corrosion Resistance of Coatings of Aluminum, Zinc and their Alloys: Results of Four Years Exposure, Interfinish. Jones. D.A. (1996), Principles and Prev ention of Corrosion, Prentice Hall, NJ, pp. 75-113. Kimoto H. (1999), Corrosion Engineering 48, p. 579. Legault R.A., Pearson V.P. (1978), Kinetics of the Atmospheric Corrosion of Aluminized Steel, Corrosion 34, pp. 344-348. Li W., Liu S., Huang Q., Gu M. (2003), Ho tDipped Aluminising (HDA) of a Low Carbon Steel Wire, Materials Sc ience and Technology 19, pp. 1025-102. Mohanty (2000), Developments in autom obile steel grades-e xperience at TATA steel, Iron Steel Rev. Pp. 19-27. Sags A. (2006), Electrochemical Im pedance Spectroscopy Course Notes, University of South Florida. Townsend H.E., Borzillo A.R. (1987), Twen ty-Year Atmospheric Corrosion Tests of Hot-dip Coated Sheet Steel, Materials Performance 26, pp. 37-41.
Appendix A: Sample Inventory Table A-1 Number of Specimens Exposed in Each Category Specimens Formed Category (Indentation Size) Solution P Solution C =1-in PDIM (1) PDIM (2) CDIM (7) CDIM (8) =3/4-in PDIM (3) PDIM (4) CDIM (10) CDIM (11) =9/16-in PDIM (5) PDIM (6) CDIM (13) CDIM (16) Specimens Exposed Flat PF (1) PF (2) CF(1) CF (2) 60
Appendix B: EIS Results from Dupli cate Dimpled Specimens in Solution C 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+04f / Hz|Z| / ohm CDIM(7)-9/16-in CDIM(8)-9/16-in CDIM(10)-3/4-in CDIM(11)-3/4-in CDIM(13)-1-in CDIM(15)-1-in CF(1) CF(2) Age of ~ 4 Days -9.0E+01 -8.0E+01 -7.0E+01 -6.0E+01 -5.0E+01 -4.0E+01 -3.0E+01 -2.0E+01 -1.0E+01 0.0E+00 1.0E+01 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05f / Hz 61 Zphz (o) CDIM(7)-9/16-in CDIM(8)-9/16-in CDIM(10)-3/4-in CDIM(11)-3/4-in CDIM(13)-1-in CDIM(15)-1-in CF(1) CF(2) Age of ~ 4 Days Phase Angle / Deg Figure B-1. EIS Results from Duplicate Dimp led Specimens in Solution C at 4 Days Bode (Top) and Phase Angle (Bottom)
Appendix B (Continued) 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+04f / Hz|Z| / ohm CDIM(7)-9/16-in CDIM(8)-9/16-in CDIM(10)-3/4-in CDIM(11)-3/4-in CDIM(13)-1-in CDIM(15)-1-in CF(1) CF(2) Age of ~ 10 Days -9.0E+01 -8.0E+01 -7.0E+01 -6.0E+01 -5.0E+01 -4.0E+01 -3.0E+01 -2.0E+01 -1.0E+01 0.0E+00 1.0E+01 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05f / Hz Age of ~ 10 Days CDIM(7)-9/16-in 62 Zphz (o) CDIM(8)-9/16-in CDIM(10)-3/4-in Phase Angle / Deg CDIM(11)-3/4-in CDIM(13)-1-in CDIM(15)-1-in CF(1) CF(2) Figure B-2. EIS Results from Duplicate Dimp led Specimens in Solution C at 10-Days Bode (Top) and Phase Angle (Bottom)
Appendix C: EIS Results from Dupli cate Dimpled Specimens in Solution P 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+04f / Hz|Z| / ohm PDIM(1)-1-in PDIM(2)-1in PDIM(3)-3/4-in PDIM(4)-3/4-in PDIM(5)-9/16-in PDIM(6)-9/16-in PF(1) PF(2) A ge of ~ 7 Days 63 -9.0E+01 -8.0E+01 -7.0E+01 -6.0E+01 -5.0E+01 -4.0E+01 -3.0E+01 -2.0E+01 -1.0E+01 0.0E+00 1.0E+01 1.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05f / HzZphz ( o) PDIM(1)-1-in PDIM(2)-1in PDIM(3)-3/4-in PDIM(4)-3/4-in PDIM(5)-9/16-in PDIM(6)-9/16-in PF(1) PF(2) A ge of ~ 7 Days Phase Angle / Deg Figure C-1. EIS Results from Duplicate Dimp led Specimens in Solution P at 7-Days Bode (Top) and Phase Angle (Bottom)
Appendix C (Continued) 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E-031.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+04f / Hz|Z| / ohm PDIM(1)-1-in PDIM(2)-1in PDIM(3)-3/4-in PDIM(4)-3/4-in PDIM(5)-9/16-in PDIM(6)-9/16-in PF(1) PF(2) A ge of ~ 21 Days -9.0E+01 -8.0E+01 -7.0E+01 -6.0E+01 -5.0E+01 -4.0E+01 -3.0E+01 -2.0E+01 -1.0E+01 0.0E+00 1.0E+01 1.0E-031.0E-021.0E-011.0E+001.0E+011.0E+021.0E+031.0E+041.0E+05f / Hz 64 Zphz (o) PDIM(1)-1-in PDIM(2)-1in PDIM(3)-3/4-in PDIM(4)-3/4-in PDIM(5)-9/16-in PDIM(6)-9/16-in PF(1) PF(2) A ge of ~ 21 Days Phase Angle / Deg Figure C-2. EIS Results from Duplicate Dimp led Specimens in Solution P at 21-Days Bode (Top) and Phase Angle (Bottom)
Appendix D: Metallographic Examination of Dimpled Samples. (Illustration of Initial Work and Methodology Under Development). Figure D-1. Dimpled Specimen Cut with Hack Saw Figure D-2. Side View of Dimpled Samples Cu t at Indentation Region Perpendicular to Metal Rolling Direction Figure D-3. Metallographic Dimpled Samples 65
Appendix D (Continued) Figure D-4 Low Magnification Mi croscopic Photos of a Dimpled Specimen. (As It Is Seen in These Photos, the Thickness of Metal Sheet Changes Strongly Due to the Severity of the Bending Stress ; Therefore, the Location of Neutral Axis Is Markedly Displaced from the Center of the Sheet). Figure D-5. Radius Mapping of 9/16-in Dimpled Sample 66
Appendix D (Continued) Following illustrates the initial metallogra phic examination of unexposed and exposed (trial test, solution P, >500 hrs) dimpled specimens. Figure D-6. High Magnification Microscopic Photo of 9/16-in Dimpled Cell Before Exposure at Indentation Region 20 m Figure D-7. Metallograph of Co rroded 9/16-in Dimpled Cell at Indentation Region. (The Disappearance of Intermetallic Layer in Several Regions is Observed. This Disappearance Can Be Explained by the Corrosion of Intermetallic Layer and the Base Steel). This Issue Will Be Examined in Continuation Work). 67