Corrosion of Steel in Concrete Structures (Woodhead Publishing Series in Civil and Structural Engineering) [2 ed.] 0128218401, 9780128218402

Citation preview

Corrosion of Steel in Concrete Structures

This page intentionally left blank

Woodhead Publishing Series in Civil and Structural Engineering

Corrosion of Steel in Concrete Structures Second Edition Edited by

Amir Poursaee

Clemson University, Clemson, SC, United States

Woodhead Publishing is an imprint of Elsevier 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, OX5 1GB, United Kingdom Copyright © 2023 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. ISBN: 978-0-12-821840-2 For information on all Woodhead Publishing publications visit our website at https://www.elsevier.com/books-and-journals Publisher: Matthew Deans Acquisitions Editor: Gwen Jones Editorial Project Manager: Franchezca A. Cabural Production Project Manager: Anitha Sivaraj Cover Designer: Rogers Mark Typeset by TNQ Technologies

Contents

List of contributors 1

2

3

4

An introduction to corrosion of engineering materials Carolyn M. Hansson 1.1 Introduction e the ubiquitous nature of corrosion 1.2 Thermodynamics is on the side of corrosion 1.3 Kinetics are on the side of the metals 1.4 Forms of corrosion 1.5 A brief history of corrosion of reinforcing steel 1.6 The magnitude of the corrosion issue in general 1.7 Magnitude of corrosion issue specifically related to RC structures 1.8 Conclusion References Principles of corrosion of steel in concrete structures Amir Poursaee and Ueli M. Angst 2.1 Introduction References Assessing a concrete’s resistance to chloride ion ingress using the formation factor Robert Spragg, Chunyu Qiao, Timothy Barrett and Jason Weiss 3.1 Introduction 3.2 Background 3.3 Experimental techniques 3.4 Microstructural parameters 3.5 Specifying formation factor 3.6 Conclusions References Chloride corrosion threshold Carolyn M. Hansson 4.1 Background 4.2 Experimental laboratory tests 4.3 Field tests

xi 1 1 1 3 4 6 8 10 12 13 17 17 29

35 35 36 41 51 54 56 57 63 63 64 70

vi

Contents

4.4 4.5 4.6 4.7

5

6

7

The influence of concrete mix design and exposure conditions The CCT of corrosion-resistant grades of rebar The influence of the chloride cation Conclusions References

Corrosion of prestress and posttension reinforced concrete bridges Kingsley Lau, Samanbar Permeh and Ivan Lasa 5.1 Introduction 5.2 Materials 5.3 Overview of corrosion mechanisms 5.4 Overview of corrosion failures and recent corrosion problems 5.5 Cathodic protection 5.6 PT grout testing 5.7 Closing remarks References

70 71 73 75 76 81 81 82 87 93 96 97 100 100

Corrosion of stainless steel reinforcemnt in concrete Carolyn M. Hansson 6.1 Rationale for the use of stainless alloys as reinforcing bars in chloride-contaminated environments 6.2 Experience to date with stainless rebar in the field 6.3 Available grades of stainless rebar 6.4 Experimental corrosion tests for stainless rebar 6.5 Results to date on corrosion behavior of stainless rebar 6.6 Conclusions References

107

Corrosion of epoxy-coated steel in concrete David B. McDonald 7.1 Introduction 7.2 Use 7.3 Current specifications 7.4 Specification changes 7.5 Manufacture 7.6 Fabrication 7.7 Field handling 7.8 Corrosion research 7.9 Changes in agency specification of epoxy-coated reinforcing steel 7.10 Summary 7.11 Sources of further information References

137

107 109 111 112 120 131 131

137 137 139 140 141 142 143 143 155 156 157 157

Contents

8

9

10

11

Galvanized steel reinforcement: recent developments and future opportunities Stephen R. Yeomans 8.1 Introduction 8.2 Galvanized reinforcement 8.3 Laboratory studies 8.4 Field studies 8.5 Design and fabrication 8.6 Applications of galvanized reinforcement 8.7 Future opportunities 8.8 Summary 8.9 Further information References Influence of the microstructure of the carbon steel reinforcing bar on its corrosion in concrete Hamidreza Torbati-Sarraf and Amir Poursaee 9.1 Introduction 9.2 Influence of steel microstructure on its corrosion behavior 9.3 Influence of steel surface microstructure of its corrosion in concrete References Effect of different concrete materials on the corrosion of the embedded reinforcing steel Brett Holland, Prasanth Alapati, Kimberly E. Kurtis and Lawrence Kahn 10.1 Introduction 10.2 Chloride ingress resistance 10.3 Carbonation resistance 10.4 Future trends 10.5 Conclusions References Corrosion measurement and evaluation techniques of steel in concrete structures Amir Poursaee 11.1 Introduction 11.2 Half-cell potential technique 11.3 Linear polarization resistance (LPR) 11.4 Galvanostatic pulse technique 11.5 Electrochemical impedance spectroscopy (EIS) 11.6 Cyclic polarization 11.7 Cyclic voltammetry (CV) 11.8 Mott-Schottky technique 11.9 Corrosion sensors for field monitoring References

vii

161 161 162 165 172 175 177 181 182 183 183

187 187 189 191 193

199

199 200 207 210 215 215

219 219 219 221 224 226 230 232 234 237 239

viii

12

13

14

15

Contents

Monitoring corrosion of steel in concrete structures Ueli M. Angst 12.1 Introduction 12.2 Roles and concepts of monitoring 12.3 Parameters relevant for corrosion and types of sensors 12.4 Interpretation of monitoring data 12.5 Conclusions and outlook References Acoustic emission monitoring for corrosion damage detection and classification Paul Ziehl and Mohamed ElBatanouny 13.1 Overview of the acoustic emission technique 13.2 Mechanism of corrosion detection using acoustic emission 13.3 Case studies for corrosion detection using AE 13.4 Corrosion classification using AE 13.5 Small scale specimens 13.6 Special considerations and potential applications in field 13.7 Special considerations for wireless sensing 13.8 Summary Acknowledgments References

245 245 247 250 259 260 261

265 265 267 268 270 271 277 278 278 279 279

Practical field implementation of corrosion measurement methods Andrew Fahim and Pouria Ghods 14.1 Introduction 14.2 Half-cell potential 14.3 Linear polarization resistance 14.4 Galvanodynamic and potentiodynamic techniques 14.5 Galvanostatic and potentiostatic techniques 14.6 Coulostatic technique 14.7 Connectionless electrical pulse response analysis (CEPRA) 14.8 Practical aspects and case study References

283

Corrosion protection methods of steel in concrete Amir Poursaee 15.1 Cathodic protection 15.2 Electrochemical chloride extraction (ECE) 15.3 Inhibitors References

319

283 283 289 293 295 300 302 304 316

319 321 323 325

Contents

16

17

ix

Modeling corrosion of steel in concrete Burkan Isgor 16.1 Introduction 16.2 Kinetics of steel corrosion in concrete 16.3 Theoretical/numerical modeling of corrosion current density 16.4 Environmental conditions and material/chemical properties of concrete around steel reinforcement 16.5 Empirical models and other practical approaches for predicting corrosion current density 16.6 Atomistic/molecular modeling to develop fundamental understanding 16.7 Conclusions References

327

Future trends in research on reinforcement corrosion Carmen Andrade 17.1 Introduction 17.2 Processes leading to reinforcement corrosion 17.3 Corrosion onset. Chloride threshold 17.4 Corrosion propagation 17.5 Modeling of service life 17.6 Additional preventive measures 17.7 Repair techniques 17.8 Corrosion measurement techniques 17.9 Final comments Acknowledgments References

353

Index

327 328 331 337 339 344 345 346

353 353 362 363 366 368 369 369 370 370 371 375

This page intentionally left blank

List of contributors

Prasanth Alapati

Furno Materials, Atlanta, GA, United States

Carmen Andrade The International Centre for Numerical Methods in Engineering (CIMNE), Barcelona, Spain Ueli M. Angst

ETH Zurich, Zurich, Switzerland

Timothy Barrett

Federal Highway Administration, Washington DC, United States

Mohamed ElBatanouny Wiss, Janney, Elstner Associates, Inc., Northbrook, IL, United States Andrew Fahim Director of Research and Development, Giatec Scientific Inc., Ottawa, ON, Canada Pouria Ghods ON, Canada

Civil and Environmental Engineering, Carleton University, Ottawa,

Carolyn M. Hansson Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, ON, Canada Brett Holland Gumpertz & Heger, Waltham, MA, United States Burkan Isgor Oregon State University, School of Civil and Construction Engineering, Corvallis, OR, United States Lawrence Kahn

Georgia Institute of Technology, Atlanta, GA, United States

Kimberly E. Kurtis Georgia Institute of Technology, Atlanta, GA, United States Ivan Lasa

Lasa and Associates Corrosion Services, Gainesville, FL, United States

Kingsley Lau Department of Civil and Environmental Engineering, Florida International University, Miami, FL, United States David B. McDonald

Vinsi Partners, Sydney, Australia

Samanbar Permeh Department of Civil and Environmental Engineering, Florida International University, Miami, FL, United States Amir Poursaee Department of Civil Engineering, Department of Materials Science and Engineering, Clemson University, Clemson, SC, United States

xii

Chunyu Qiao Robert Spragg

List of contributors

University of Science and Technology Beijing, Beijing, China Federal Highway Administration, Washington DC, United States

Hamidreza Torbati-Sarraf

Lam Research, Fremont, CA, United States

Jason Weiss Oregon State University, School of Civil and Construction Engineering, Corvallis, OR, United States Stephen R. Yeomans University of New South Wales, Canberra, ACT, Australia Paul Ziehl Department of Civil and Environmental Engineering, University of South Carolina, Columbia, SC, United States

An introduction to corrosion of engineering materials

1

Carolyn M. Hansson Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, ON, Canada

1.1

Introduction e the ubiquitous nature of corrosion

With isolated exceptions, such as the native copper found in upstate Michigan, in the USA (Bornhorst and Barron, 2011), metals exist in the Earth’s crust as minerals, such as carbonates, sulfides, sulfates or oxides, formed over billions of years. These compounds are, therefore, considered to be in their equilibrium state, that is, their lowest energy state. Extracting the metals from these compounds requires a significant amount of energy in the form of heat, electrical and/or mechanical power. Once extracted, the metals and alloys are in a thermodynamically metastable state and, depending on the environment, will always attempt to revert to a lower energy compound, usually by corrosion or oxidation. Since the Earth’s atmosphere contains water and oxygen, and coatings and other barriers are inherently imperfect, it is impossible to completely prevent corrosion. Consequently, all the metals and alloys in use today are in metastable condition.

1.2

Thermodynamics is on the side of corrosion

The stability of the metal is usually described in terms of its electrochemical potential, a thermodynamic function which may be defined in the context of corrosion, as “the ease of ionizing an atom of the metal”. Potentials are dependent on the pH of the environment and factors such as oxygen availability. Potentials cannot be determined absolutely, and are, therefore, defined as the potential difference between the metal of interest and that of a reference electrode, so chosen to have a stable potential (Revie and Uhlig, 2008). The “standard hydrogen electrode” (SHE) is the reference electrode against which all other electrodes are measured and is arbitrarily given the electrochemical potential of 0.00 V. It is defined such that the reaction: 2Hþ þ 2e ¼ H2 is in equilibrium at pH 0 (i.e., 1 M acid solution), 1 atm hydrogen gas and 25 C. Because the hydrogen electrode is cumbersome to operate and transport, other electrodes, consisting of a metal in contact with a saturated solution of its ions, such as Cu/CuSO4 or Ag/AgCl, are used. The different common reference electrode potentials relative to that of the standard hydrogen electrode are given in Table 1.1. These electrodes are not generally suitable for embedding in concrete because they become unstable over a short period of time at high pH levels. However, a Mn/MnO2 electrode has been developed specifically for use in concrete (Arup et al., 1997). Corrosion of Steel in Concrete Structures. https://doi.org/10.1016/B978-0-12-821840-2.00001-8 Copyright © 2023 Elsevier Ltd. All rights reserved.

2

Corrosion of Steel in Concrete Structures

Table 1.1 Characteristics of common reference electrodes (Corrosion-doctors.org).

Electrode type

Equilibrium reactions Nernst equation

Conditions (activity)

Standard hydrogen electrode (SHE) Silver chloride

2Hþþ2e ¼ H2 E0 ¼ 0.059 pH

pH ¼ 0

AgCl þ e ¼ Ag þ Cl  0 E ¼ 0.059log10(aCl-)

Calomel

Hg2Cl2 þ2e ¼ 2Hg þ 2Cl E0 - 0.059log10(aCl-)

Mercurous sulfate Mercuric oxide

Hg2SO4 þ2e ¼ 2Hg þ SO24 E0e 0.0295log10(aSO2 4 ) HgO þ2e  þ 2Hþ ¼ Hg þ H2O E0 -0.059 pH Cu2þ þ 2e ¼ Cu (SO2 solution E0 þ0.0295log10(a2þ) Cu

aCl ¼ 1 0.1M KCl 1.0 M KCl Saturated KCL Seawater aCl ¼ 1 0.1M KCl 1.0 M KCl Saturated KCL

Copper sulfate

Potential V versus SHE

Temp. coefficient mV/8C

0.2224 0.2881 0.235 0.199

0.6 0.6 0.6 0.6

w0.250 0.268 0.3337 0.280 0.241

0.6 0.06 0.24 0.65

0.6151 0.926 aCu ¼ 1 Saturated

0.340 0.318

The ranges of electrochemical potential and pH in which the products of corrosion are either dissolved ions, e.g., Fe2þ or Cr3þ, or solid oxides or hydroxides are given in the “Atlas of Electrochemical Equilibria” (Pourbaix, 1974) for each metal in aqueous solutions. The solid products may protect the metal from further rapid corrosion and are then known as “passive films”. The dissolution, or oxidation, of a metal, e.g., M / M2þ þ 2e, constitutes the anodic half-cell reaction and the electrons released by this reaction are immediately consumed, usually by one of the two cathodic half-cell reactions: (a) 2Hþ þ 2e ¼ H2 (b) O2 þ 2H2O þ 4e ¼ 4(OH).

In low pH conditions and with little available oxygen, reaction (a), the evolution of hydrogen, is the most likely cathodic reaction whereas, with more oxygen and/or at neutral or high pH, reaction (b), the reduction of dissolved oxygen, is more likely. The limitations of these diagrams are that they consider only pure metals, only water and oxygen as the electrolyte and are calculated only for 25 C. Research is only fairly recently being conducted to calculate comparable electrochemical phase diagrams for alloys and for different environments using computational methods (Bale et al., 2009; Andersson et al., 2002; Ding et al., 2018) or by experiment, for example

An introduction to corrosion of engineering materials

3

(Beverskog and Puigdomenech, 1999a,b; Nishimura and Kodama, 2003; Gimenez et al., 1981).

1.3

Kinetics are on the side of the metals

–0.4

1.0 × 10–3

–0.5

0.75 × 10–3

–0.6

0.50 × 10–3

–0.7 Potential

Current (A)

Potential (V vs SCE)

First and foremost, only about one in 10 million of the atoms in a metal or alloy is on the surface and exposed to the environment and, thus, susceptible to corrosion. Second, although these surface atoms react, many of the reactions result in a solid oxide or hydroxide which protects the underlying metal from further rapid corrosion. The solid product is generally in the form of a very thin film, of the order of a few to a few hundred nanometers and may consist of multiple layers (Oranowska and Sklarska-Smialowska, 1981; Albani et al., 1986). It is described as a passive film rather than as a coating and is discussed in detail in Chapter 2. Because corrosion involves the exchange of electrons between the metal (the anode) and the cathodic species (e.g., hydrogen ions or dissolved oxygen), it is described as an “electrochemical process” and the current flowing between anode and cathode is a direct measure of the rate of dissolution or oxidation of the metal, known as the corrosion current. The passive film is not perfect and behaves as a leaky capacitor with some metal ions being transported through the film (Hakiki et al., 1998; Dean and Stimming, 1989). Thus, corrosion is not zero but is at a level which is considered insignificant relative to the expected service life of the part. Another major feature of natural passive films is their ability to recover after being damaged mechanically. This is illustrated in Fig. 1.1 showing the corrosion current and potential excursions when a sheet of grade UNS S2205 duplex stainless steel was immersed in a synthetic concrete pore solution (pH w13.5) and scratched with a knife. Note that the current recovered after w15 s but the potential drifted much more slowly to its original level. This is because it is

0.20 × 10–3

Current –0.8 200

0.00 400

600

800

Time (s)

Figure 1.1 Recovery time of passive film on 2205 duplex stainless steel when immersed in synthetic pore solution and scratched (Mammoliti and Hansson, 2007).

4

Corrosion of Steel in Concrete Structures

a “mixed potential” consisting partly of the potential of the scratched area and that of the undamaged area of the steel. Unfortunately, while metals and alloys may exhibit passive behavior resulting in loss of material of less than 1 mm/year, in practice that passive film can be breached, causing corrosion rates several orders of magnitude higher. A prime culprit in passive film failure is the chloride ion, which causes major problems in marine environments, in regions where deicing salts are used and in some industrial regions. Other, less wellunderstood culprits are microbes such as bacteria and molds (Stott, 1993; Lane, 2005; Sooknah et al., 2007), one of the major causes of corrosion and blockage in pipelines.

1.4

Forms of corrosion

There are many forms of corrosion (Revie and Uhlig, 2008; ASM, 1990; McCafferty, 2010) but only those most commonly encountered in concrete structures are discussed here.

1.4.1

General corrosion

Occurs in regions of potential and pH in which the corrosion products are either dissolved ions or nonprotective oxides/hydroxides. The rust produced on steel in the outdoor atmosphere is a typical example of a nonprotective hydroxide. In concrete, general corrosion can result from a decrease in pH to a range where a passive film is not stable and breaks down, leaving the steel to either dissolve (in very low pH solutions) or form a nonprotective oxide/hydroxide. This form of corrosion can result in a general thinning of the metal or alloy and, eventually, failure of the part.

1.4.2

Pitting corrosion

Is usually limited to those metals with passive films such as aluminum, titanium or zinc in neutral solutions, or steel at high pH. The most common cause is a very localized breakdown of the passive film by chlorides, typically from seawater and deicing salts but also from industry and even household salt use. There have been many studies on the breakdown of passive films but, because the films are so thin, the initiation of the pit cannot be observed directly. It is generally agreed that the first step is by adsorption of chloride or other halide ions. For the next step, theories diverge from the adsorptioninduced development of mechanical stresses in the film to a local thinning of the film, thereby increasing the potential gradient across the film and local “dielectric breakdown” (Sato, 1990; Szklarska-Smialowski, 2002; Macdonald, 1999). A recent model including the role of the oxide nanolayer is given in (Marcus et al., 2008). The semiconducting nature of the passive film is discussed in Chapter 2. Once pits are formed, however, there is a general consensus that they create a local “differential environmental cell” in which the electrolyte in the pit becomes depleted in

An introduction to corrosion of engineering materials

5

oxygen and more acidic than the solution adjacent to the nonpitted passive film. Moreover, the growth of the pit becomes autocatalytic because the chloride ions are not consumed in the process: they react with the metal but the metal chloride complex becomes unstable and reacts further, releasing the chloride ions to combine with other metal ions. The sequence of reactions is as follows: M/M nþ þ ne M nþ þ nCl /MCln MCln þ nH2 O/MðOHÞn þ nCl The electrons released in the first of these reactions travel through the metal, as shown in Fig. 1.2, and react with water and oxygen to create hydroxyl ions. In the case of chloride-induced corrosion of carbon steel in concrete, the attack is often described as localized corrosion, as shown in Fig. 1.3, because the attack extends laterally as well as in-depth. The third form of corrosion, which is rarely a concern for reinforcing bars, but can be a problem for prestressed or posttensioned steel, is environmentally assisted cracking which comprises stress corrosion cracking and/or hydrogen embrittlement. These phenomena are discussed in Chapter 5.

Cl–

Cl– Na+ OH–

O2 Na+

OH–

e–

e–

Na+

Na+ Cl–

O2 OH– e–

Na+

M+ M+ Cl– Cl– – + Cl – + Cl M M Cl– + M

Cl–

O2 Na+ OH– OH– e–

e–

Passive film

Figure 1.2 Schematic representation of chloride-induced pitting corrosion of a passivated metal (Hansson, 2011).

Figure 1.3 Chloride-induced localized corrosion of carbon steel rebar (Jaffer, 2007).

6

1.5

Corrosion of Steel in Concrete Structures

A brief history of corrosion of reinforcing steel

The problem of rebar corrosion has been recognized for over a century. In 1911, a report in the Proceedings of the American Institute of Electrical Engineers (Mahgnusson and Smith, 1911) based on laboratory studies of a few years earlier, stated: “While reinforced concrete was coming into general use as a structural material, much space in the technical press was given to discussions and reports on the durability of the encased iron. The results from a large number of experiments gave fairly conclusive evidence that under ordinary conditions the iron is protected, and that even if it was rusty when placed in the concrete, it will be free from the oxide after remaining in the concrete for some time. The time test on the durability of practical structures is, of course, the final arbiter, and for each year the increasing data bears out the assumption that properly constructed concrete-steel structures will stand indefinitely.” Nevertheless, in 1911, there were two reports of electrolytic corrosion (currently known as “stray current corrosion”). The first (Brown, 1911) involved damage to a large reinforced concrete packing house outside New York City. It was originally attributed to a stray current from a nearby electric railway but, was subsequently, found to be caused by a short between the D.C. electric light circuit and a water pipe, but “only where the concrete was damp and the pipe had an electric potential positive to ground.” At that time, “none thought it possible that electricity could leak to steel deeply buried in beams and girders and split the concrete apart” (Brown, 1911). The second report followed an inspection of four reinforced concrete structures in the harbor in Southampton, UK (Bamber et al., 1911). One of the structures, a small jetty, built in 1899 and believed to be the oldest reinforced concrete structure in the country, was found to be in excellent condition. Two other structures, a coal barge jetty and a quay, built in 1908, were showing significant deterioration: rusting of the steel and cracking and spalling of the concrete. This problem, too, was traced to the return wire of the electric lighting circuit being earthed (ground) to a metal plate on the deck of the jetty with a path to the sea through the concrete. Both of these studies used electrical potential difference measurements between the lighting wires, the rebar and the ground in order to identify the cause of deterioration. Much of the concern over the next few decades appears to have been focused on electrolytic, or stray current, corrosion. In 1949, a study of stray current corrosion of steel embedded in concrete and buried in soil, produced by an adjacent electrified railroad concluded that (1) a large (300 mm) concrete cover provided some but not complete protection from corrosion; (2) the electrochemical potentials produced attracted sulfate ions from the soil which deteriorated the concrete, (3) a.c. currents did not produce any degradation, (4) stainless steel bars were not corroded and (5) an asphalt coating on the concrete was effective in protecting the rebar (Magee, 1949). There was, however, an early realization of the destructive influence of chlorides.

An introduction to corrosion of engineering materials

7

In 1917, practically all the marine concrete structures in the USA were inspected (Wig and Ferguson, 1917). It was found that the majority of them, with ages ranging from 1 to 10 years, was showing some signs of deterioration or failure due to corrosion of the rebar above the waterline. For some of the structures, seawater had been used as mixing water but, for most, freshwater was used. After eliminating the possibility of “electrolysis” (stray current corrosion), it was concluded that the cause was the accumulation of salts in the concrete pores above the waterline by capillarity and evaporation e a process widely accepted today. It was also noted that, when steel corrodes, it occupies about twice its previous volume, develops stresses in the concrete and results in cracking and spalling. The role of preexisting cracks and those caused by corrosion was also discussed. It was concluded that preexisting cracks would be a large contributing factor but not a necessary one e a topic still under discussion today in Chapter 12. The recommended solutions were increasing the concrete cover and the use of galvanized steel e again, a current topic, Chapter 8. Recognition of concrete as an electrolyte was reported in Corrosion in 1957 (Stratfull, 1957). This paper concerned the deterioration of a seven mile-long causeway in California, USA, and is believed to be the first structure in which the areas of corrosion were identified by electrochemical corrosion mapping e a technique in routine use today by transport authorities around the world for condition monitoring. The map for the approaches to the bridge shown in Fig. 1.4 is shown in Fig. 1.5 (Poursaee, 2007). The use of this mapping technique for other structures was initiated in Europe in the 1970s after an apartment balcony in Copenhagen collapsed, killing several people. There were approximately 6000 balconies of similar design and age and potential mapping was used to identify those in need of immediate attention (Grønvold and Arup, 1979).

Figure 1.4 A view of the bridge at the University of Waterloo. (Poursaee 2007).

8

Corrosion of Steel in Concrete Structures

Figure 1.5 Contour map of the half-cell potential measurements of the approaches A and B of the bridge at the University of Waterloo; March 06, 2005; T ¼ 10 C, RH ¼ 72%, cloudy. (Poursaee 2007).

1.6

The magnitude of the corrosion issue in general

The cost of corrosion can be measured in social, environmental and financial terms (Hansson, 2011). There have been a number of projects aimed at estimating the financial costs, but there has been much less attention given to the social and environmental impacts. For example, a corrosion-induced oil pipeline failure, in which the oil contaminates the surrounding ground and water, is described in terms of the clean-up and repair costs, but not in terms of the impact on the neighboring population, flora and fauna and the time for these to return to their former state. Although few corrosion-induced failures result directly in human deaths, the environmental contamination can cause death to countless other species and future sickness and suffering for the human population. The worst example of this was the Bopal, India disaster of 1984, which resulted in the release of toxic gases and the immediate deaths of over 8000 people and the long-term health problems of half a million people (Brown, 2009; Peterson, 2009). Another, insidious form of corrosion is the long-term, slow release of metallic ions which can have unpredicted health impacts (Bertling et al., 2006a,b; Dietrich et al., 2004), particularly in the case of microbially-induced corrosion (Costerton et al., 1987). To quote George Hayes, Director General of World Corrosion Organization (Hays, 2010): The public hears about bridges and piers that collapse due to corrosion and environmental damage caused by pipeline failures, but does it ever hear about the loss

An introduction to corrosion of engineering materials

9

of potable water from water main corrosion or environmental damage caused by corroded sewer lines? The answer is no. Neither news media nor government agencies find such events sufficiently newsworthy to alert the public. Yet in many countries, the cost of water and wastewater system failures is far greater than any other single sector of the economy. Corrosion knows no national boundaries. Acid rain generated in one country pollutes the environment and causes corrosion damage far beyond that country’s borders and even beyond the borders of its neighbors. Toxic materials, released from corroded equipment in one area, pollute the air and water far beyond one country’s borders. And toxic material released into the world’s waterways poisons sea life, killing many species and making others toxic to humans.

Significant environmental damage is caused by the accident-prevention strategies in which an estimated 66 million tons of deicing salts are deposited on the roads globally every year (Houska, 2017), and by the atmospheric pollution by particulates and volatile gases produced maintenance of the highway infrastructure (Roadex). While little attention has been paid to the social and environmental costs, the financial costs of corrosion have received a huge amount of attention in recent years, largely in an attempt to draw the attention of decision-makers to the magnitude of the problem. The general conclusion is that corrosion is consuming approximately 3.4% of the world’s gross domestic product (GDP) i.e., about US$2.2 trillion in 2010 and US$2.5 trillion annually in 2013 (Hays, 2010; Koch et al., 2016). The estimates of the financial cost to several nations are given in Table 1.2, modified from (Biezma and San Crist obal, 2005). For the USA in 2002, the direct annual costs of corrosion were estimated at $137.9 B, broken down into sectors in Fig. 1.6. When the indirect costs were included, this total became $276 B e or 3.1% of the USA’s Gross Domestic Product. Similar data for the Chinese economy are shown in Fig. 1.7. Table 1.2 Estimated annual costs of corrosion in different countries (Biezma and San Crist obal, 2005).

Zhu (1986), Hou et al. (2017) Cherry and Skerry (1983) Al-Kharafi et al. (1995) Tullmin (2015) JSCE (2001) Tullmin (2015) Koch et al. (2002) (Sastri et al., 2003; Shipilov, 2009) Koch et al. (2016)

Year

Country

%GNP

Cost

1986 2017 1983 1995 1996 2001 2008 2002 2003

China

3.34

Australia Kuwait Switzerland Japan Australia USA Canada

1.5 5.2 3e5 0.77 2 3.1

India

4.5%

2127.8B RMB or wUS$325B US$2 B US$1 B 10e15 B Swiss Fr. 3938 B yen Aus.$1e5 B US$276 B CAD$32.8 B CAD$41 B US$48.4

10

Corrosion of Steel in Concrete Structures

Government $20.1 B

Production & manufacturing $17.6 B

Infrastructure $22.6 B Utilities $47.9 B Transportation $29.7 B

Figure 1.6 Direct annual costs of corrosion in USA (Koch et al., 2002).

Figure 1.7 The direct costs of corrosion of five major economic sectors in China in 2014 (Hou et al., 2017).

1.7

Magnitude of corrosion issue specifically related to RC structures

Concrete in the early part of the last century was relatively simple, as illustrated in Fig. 1.8A. Today’s concrete is a very sophisticated blend. Fig. 1.8B, shows how the proportions of the basic constituents can be tailored to the specific application and “seasoned” with a number of chemical and mineral admixtures to enhance performance, as described further in Chapter 10. Consequently, the diffusion of deleterious species through the today’s concrete pores is a very slow process (Sandberg et al., 1998; Stanish et al., 1997). Unfortunately, all structural concrete is cracked and ingress is significantly faster through the cracks than through the pores (Jacobsen et al., 1966;

An introduction to corrosion of engineering materials

(a)

Fuel

11

Aggregates Clinker Good old OPC for any concrete

Raw mix

Good old concrete for any application

Gypsum

28 days strength

Water

Color

(b) Anhydrite

Accelerator

Ordinary

Air entrainer

Reinforced

Plasticiser

SCC

Super plast

Structures

Retarder

Industrial

Aggregates

Roads

Workability

Tires Gypsum

Bleeding

Coke de petrol

Plaster

Fuel

Lime

Coal

Animal meal

Clinker A

Sulfo gypsum Water

Solvents

Clinker B

Waste water

Binder

Architectural

Concrete

Slabs

Slag

Fly ash Grinding aid

Alternative materials CKD

Aggressive env.

Limestone

Silica fume

Heat of hydration Early strength 2d, 28 d strength

Beams

Pozzolans Raw mix

Setting time

Blocks

Surface aspects

Durability

….

Figure 1.8 Concrete constituents (A) at the beginning of the last century and (B) today. (Courtesy: Y. Guitton-Fumet, Lafarge Central Research Laboratory).

Wang et al., 1997; Balakumaran et al., 2018; Otieno et al., 2016, Li and Li, 2019). The impact of the cracks on the time required to initiate corrosion of the rebar and on the subsequent corrosion rates remains a topic of discussion (Otieno et al., 2016). According to (Davis and Goldberg, 2013), most bridges in the USA were designed for a 50-year service life and their average age is now 45 years. The number of these that were described as structurally deficient bridges in 2012 was 66,405 and had an average age of 65 years. It is clear that this number is likely to increase significantly in the future raising the annual cost far beyond the 2002 estimate of $8.3 B. The age of the infrastructure in Canada, as a % of its useful life in 2007 is given in Fig. 1.9 showing that the average age of all sectors except the water supply systems, exceeds 50% (Gagnon et al., 2007). In Ontario, this translates to 2600 of the Provincial bridges in a 2000 survey, having reached an average of more than 80% of their planned service life. Although there are many factors contributing to the “end of life” for

12

Corrosion of Steel in Concrete Structures

Age as a % of useful life 65 60

Bridges and overpasses Wastewater treatment Highways and roads

55 50

Sewer sytems

45 40

Water supply systems 35 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 1.9 Age of Canadian infrastructure as a function of its useful life (Gagnon et al., 2007). Table 1.3 Cost of corrosion in the infrastructure section in the USA (Koch et al., 2002). Sector

Annual cost

Highway bridges Gas and liquid pipelines Waterways and ports Hazardous materials storage Gas distribution Drinking water and sewer

$8.3 B $7.0 B $0.3 B $7.0 B $5.0 B $36.6 B

reinforced concrete bridges, corrosion of the reinforcing steel is by far the most common. The total (direct plus indirect) annual cost of corrosion of the (largely reinforced concrete) infrastructure in the USA was estimated in 2002 as $63.6 B and was broken down into sectors as given in Table 1.3 (Koch et al., 2002). Note the huge cost to the drinking water and sewage sector with respect to the comments of Mr. Hayes, quoted above.

1.8

Conclusion

Corrosion is here to stay. The goal of the corrosion science and engineering community must be to slow the corrosion rate sufficiently so that the part or structure remains serviceable for its specified service life. This requires a thorough understanding of the nature of corrosion and the materials comprising the part or structure, to allow appropriate selection of both concrete and steel materials. The remaining chapters in this book are aimed at explaining the fundamentals, methods and materials to achieve this goal.

An introduction to corrosion of engineering materials

13

For most engineering applications, alloys should be selected which are passive in the environment in which they will operate in service or are provided with adequate protection for the expected useful life. Unfortunately, most engineers select materials for their functional properties, such as mechanical strength or electronic capabilities, without regard to the potential hazards of the service environment. To quote (Koch et al., 2002): “Corrosion research has, over the years, suffered from inadequate industry and government funding, especially given the cost and inconvenience associated with corrosion of water mains, bridge structures, automobiles, airplanes, and pipelines. In contrast, physicists and biologists have captured the attention of the public and Congress by describing the dynamics of high-energy physics and biotechnology. While not as noteworthy to the popular press as some other technologies, corrosion research has much to contribute to delivering social services more efficiently and more reliably while lowering the costs of many of the products and services purchased by the public.’” To conclude, there is one final objective: to educate the design and engineering community in order to break the “Corrosion Cycle” defined by J.C. Bowles, former President of NACE (Tullmin, 2015): • • •

•

Phase 1 e Neglect: Corrosion control is ignored; this may be “tempting” to (poor) management as corrosion problems may not show up immediately. It is easy to be lulled into a false sense of security. Phase 2 e Panic: The previously hidden corrosion danger becomes apparent, possibly with disastrous financial consequences and safety hazards. It is not easy to combat corrosion rationally and effectively in a state of panic. Phase 3 e Learning Curve: In dealing with the serious corrosion problems, effective corrosion control measures are eventually introduced reducing failure rates to manageable levels. Considerable effort (and time) may be required before effective solutions are found, qualified and implemented. Phase 4 e Unlearning Curve: Once the initial crisis is over, there is a risk that corrosion control will be neglected again and that hard lessons learnt in the past will be forgotten. This is when the corrosion cycle starts all over again, with the neglect stage re-establishing itself.

References Al-Kharafi, F., Al-Hashem, A., Matrouk, F., 1995. Economic Effects of Metallic Corrosion in the State of Kuwait. KISR Publications, Kuwait. Albani, O.A., Zerbino, J.O., Vilche, J.R., Arvia, A.J., 1986. A comparative elecrochemical and elliprometric study of iron electrodes in different alkaline electrolytes. Electrochimica Acta 31, 1403e1411. Andersson, J.-O., Helander, T., Höglund, L., Shi, P., Sundman, B., 2002. Thermo-calc & Dictra, computational tools for materials science. Calphad 26, 273e312. Arup, H., Klinghoffer, O., Mietz, J., 1997. Long term performance of MnO2-reference electrodes in concrete. Corrosion 97. NACE. ASM, 1990. Metals Handbook, Tenth ed. ASM International, Materials Park, OH. Corrosion. Balakumaran, S.S.G., Weyers, R.E., Brown, M.C., 2018. Linear Cracking in Bridge Decks. Virginia Transportation Research Council, Charlottesville, Virginia.

14

Corrosion of Steel in Concrete Structures

Bale, C.W., Bélisle, E., Chartrand, P., Decterov, S.A., Eriksson, G., Hack, K., Jung, I.-H., Kang, Y.-B., Melancon, J., Pelton, A.D., Robelin, C., Petersen, S., 2009. FactSage thermochemical software and databases e recent developments. Calphad 33, 295e311. Bamber, H.H.G., Marsh, C.F., Wentworth-Sheild, F.F., 1911. A British example of electrolytic corrosion of steel in a reinforced concrete structure. Engineering News-Record 66. Bertling, S., Wallinder, I.O., Berggren Kleja, D., Leygraf, C., 2006a. Long-term corrosioninduced copper runoff from natural and artificial patina and its environmental impact. Environmental Toxicology & Chemistry 25, 891e898. Bertling, S., Wallinder, I.O., Leygraf, C., Berggren Kleja, D., 2006b. Occurrence and fate of corrosion-induced zinc in runoff water from extermal sources. The Science of the Total Environment 367, 908e923. Beverskog, B., Puigdomenech, 1999a. Revised Pourbaix diagrams for Zn. Corrosion Science 39, 107e114. Beverskog, B., Puigdomenech, I., 1999b. Pourbaix diagrams for the ternary system of ironchromium-nickel. Corrosion 55, 1077e1087. Biezma, M.V., San Cristobal, J.R., 2005. Methhodology to study cost of corrosion. Corrosion Engineering, Science and Technology 40, 344e352. Bornhorst, T.J., Barron, R.J., 2011. Copper Deposits of the Western Upper Peninsular of Michigan. Geological Society of America, Washington, DC. Brown, H.H., 1911. Serious injury to a reinforced concrete building by electrolysis. Engineering News-Record 65. Brown, M., 2009. Bhopal Gas Disaster Legacy Lives on 25 Years Later [Online]. The Daily Telegraph. Available: http://www.telegraph.co.uk/news/worldnews/asia/india/5978266/ Bhopal-gas-disasters-legacy-lives-on-25-years-later.html (Accessed 18 July 2010). Cherry, B.W., Skerry, B.S., 1983. Corrosion in Australia. Australian National Centre of Corrosion Prevention and Control Feasilbility.Corrosion-Doctors.Org, Kingston, ON (Accessed). Costerton, J.W., Cheng, K.-J., Geesey, G.G., Ladd, T.I., Nickel, J.C., Dasgupta, M., Marrie, T.J., 1987. Bacterial boifilms in nature and disease. Annual Review of Microbiology 41, 435e464. Davis, S., Goldberg, D., 2013. The Fix We’re in for: The State of Our Nation’s Bridges 2013. Transportation for America, Washington, DC. Dean, M.H., Stimming, U., 1989. The electronic properties of disordered passive films. Corrosion Science 29, 199e211. Dietrich, A.M., Glindemann, D., Pizarro, F., Gidi, V., Olivares, M., Araya, M., Burlingame, G.A., Khiari, D., Edwards, M., 2004. Health and aesthetic impacts of copper corrosion on drinking water. Water Science and Technology 49, 55e62. Ding, R., Shang, J.-X., Wang, F.-H., Chen, Y., 2018. Electrochemical Pourbaix diagrams of NiTi alloys fron first principles calculations and experimental aqueous states. Computational Materials Science 143, 431e438. Gagnon, M., Gaudreault, V., Overton, D., 2007. Age of Public Innfrastructure: A Provincial Perspective. Statistics Canada, Ottawa. Gimenez, P.H., Rameau, J.J., Reboul, M.C., 1981. Experimental pH potential diagram for aluminium for Sea Water. Corrosion 37. Grønvold, F., Arup, H., 1979. Localization of Corroding Reinforcement by Electrochemical Potential Surveys. RILEM Symposium on Quality Control of Concrete Structures, Stockholm, pp. 251e258. Hakiki, N.E., Da Cunha Belo, M., Simoes, A.M.P., Ferreira, M.G.S., 1998. Semiconducting properties of passive films formed on stainless steels. Journal of the Electrochemical Society 145, 3821e3829.

An introduction to corrosion of engineering materials

15

Hansson, C.M., 2011. The impact of corrosion on society. Metallurgical and Materials Transactions A 42A, 2952e2962. Hays, G.F., 2010. World Corrosion Organization. National Association of Corrosion Engineers Europe Board. Corrosia. Fall 2010 ed. Hou, B., Li, X., Ma, X., Du, C., Zhang, D., Zheng, M., Lu, D., Ma, F., 2017. The cost of corrosion in China. Npj Materials Degradation 1, 1e10. Houska, C., 2017. De-icing Salt e Recognizing the Corrosion Threat [Online]. International Molybdenum Association. Available: www.imoa.info (Accessed 2021). Jacobsen, S., Marchand, J., Boisvert, L., 1966. Effect of cracking and healing on chloride transport in OPC concrete. Cement and Concrete Research 26, 869e881. Jaffer, S.J., 2007. The Influence of Loading on the Corrosion of Steel in Cracked Ordinary Portland Cement and High Performance Concretes. University of Waterloo. Ph.D. Jsce, 2001. Committee on Corrosion Loss in Japan, vol. 50, pp. 490e512. Koch, G.H., Brongers, M.P.H., Thompson, N.G., Virmani, Y.P., Payer, J.H., 2002. Corrosion Cost and Prevention Strategies in the United States. Federal Highway Administration, McLean, VA. Koch, G.H., Varney, J., Thompson, N., Moghissi, O., Gould, M., Payer, J.H., 2016. In: Jacobson, G. (Ed.), International Measures of Prevention, Application and Economics of Corrosion Technologies Study. NACE International, Houston, TX. Lane, R.A., 2005. Under the microscope: understanding, detecting and preventing microbially influenced corrosion. Journal of Failure Analysis and Prevention 5 (5), 10e12. Li, K., Li, L., 2019. Crack-altering durability properties and performance of structural concretes. Cement and Concrete Research 124, 105811. Macdonald, D.D., 1999. Passivity e the key to our metals-based civilization. Pure and Applied Chemistry 71, 951e978. Magee, G.M., 1949. Electrolytic corrosion of steel in concrete. Corrosion 5, 378e382. Mahgnusson, C.E., Smith, G.H., 1911. Electrolytic corrosion of reinforced concrete. Proceedings of the American Institute of Electrical Engineers 30. Mammoliti, L., Hansson, C.M., 2007. Recovery of Mechanically Damaged Passivity of Stainless Steels in Alkaline Solutions. Unpublished work. Marcus, P., Maurice, V., Strehblow, H.-H., 2008. Localised corrosion (pitting): a model of passivity breakdown including the role of the oxide nanostructure. Corrosion Science 50, 2698e2704. Mccafferty, E., 2010. Introduction to Corrosion Science. Springer, New York, NY. Nishimura, T., Kodama, T., 2003. Clarification of chemical state for alloying elements in iron rust using a binary-phase potentialepH diagram and physical analyses. Corrosion Science 45, 1073e1084. Oranowska, H., Sklarska-Smialowska, Z., 1981. An electrochemical and ellipsometrric investigation of surface films grown on iron in saturated calcium hydroxide solution with or without chlorides. Corrosion Science 21, 735e747. Otieno, M., Beushausen, H., Alexander, M.G., 2016. Chloride-induced corrosion of steel in cracked concrete - Part II: corrosion rate prediction models. Cement and Concrete Research 79, 386e394. Peterson, M.J., 2009. Bhopal Plant Disaster. International Dimensions of Ethics Education in Science and Engineering. Univerity of Massachusetts, Amherst, MA [Online]. Available: www.umass.edu/sts/ethics (Accessed). Pourbaix, M., 1974. Atlas of Electrochemical Equilibria in Aqueous Solutions. National Association of Corrosion Engineers, Paris.

16

Corrosion of Steel in Concrete Structures

Poursaee, A., 2007. An Analysis of the Factors Influencing Electrochemical Measurements of the Condition of Reinforcing Steel in Concrete Structures. Ph.D., University of Waterloo. Revie, R.W., Uhlig, H.H., 2008. Corrosion and Corrosion Control. John Wiley & Sons, Hoboken, NJ. Roadex, N. Environmental Issues Related to Road Management [Online]. Roadex Network, Finnish transport Agency, Helsinki, Finland. Available: roadex.org. (Accessed 2021). Sandberg, P., Tang, L., Andersen, A., 1998. Recurrent studies of chloride Ingress in uncracked marine concrete at various exposure times and elevations. Cement and Concrete Research 28, 1489e1503. Sastri, V.S., Elboujdaini, M., Perumareddi, 2003. Economics of corrosion and wear in Canada. In: Lou, J., Elboujdaini, M., Shoesmith, D., Patnaik, P.C. (Eds.), International Conference on Environmental Degradation of Metals. Metallurgical Society of the Canadian Institute of Mining, Vancouver, BC. Sato, N., 1990. An overview of the passivity of metals. Corrosion Science 31, 1e19. Shipilov, S., 2009. What corrosion costs Canada e or can we afford to ignore corrosion. In: Wanjara, C.B.P. (Ed.), Materials Development and Performance of Sulphur Capture Plants. Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, QC. Sooknah, R., Papavinasam, S., Revie, R.W., 2007. Monitoring Microbially Influenced Corrosion: A Review of Techniques. Corrosion 2007. National Association of Corrosion Engineers, Nashville, TN. Stanish, K., Hooton, R.D., Thomas, M.D.A., 1997. Testing the Chloride Penetration Resistance of Concrete: A Literature Review. University of Toronto, Toronto, ON. Stott, J.F.D., 1993. What progress in understanding of microbially induced corrosion has been made in the last 25 Years? A personal viewpoint. Corrosion Science 35, 667e673. Stratfull, R.F., 1957. The corrosion of steel in a reinforced concrete bridge. Corrosion 13, 43e48. Szklarska-Smialowski, Z., 2002. Mechanism of pit nucleation by electrical breakdown of the passive film. Corrosion Science 44, 1143e1149. Tullmin, M., 2015. www.corrosio-club.com. (Accessed 2 August 2015). Wang, K., Jansen, D.C., Shah, S.P., 1997. Permeability study of cracked concrete. Cement and Concrete Research 27, 381e393. Wig, R.J., Ferguson, L.R., 1917. Reinforced concrete in sea water fails from corroded steel. Engineering News-Record 79, 689e693. Zhu, R., 1986. International Approaches to Reducing Corrosion Costs. Corrosion 86. National Association of Corrosion Engineers, Houston, TX.

Principles of corrosion of steel in concrete structures

2

Amir Poursaee 1 and Ueli M. Angst 2 1 Department of Civil Engineering, Department of Materials Science and Engineering, Clemson University, Clemson, SC, United States; 2ETH Zurich, Zurich, Switzerland

2.1

Introduction

The low cost of steel-reinforced concrete and the ready availability of raw materials from which it is formed make it the most widely used structural material available. The durability of structures made from steel-reinforced concrete is related to its ability to impede, or greatly reduce, the rate of moisture transport or the ingress of the aggressive ions. However, because many existing concrete structures are under constant exposure to aggressive environments, they suffer from durability issues, particularly, the corrosion of the reinforcing steel bars (rebar). This corrosion, in turn, has created a multi-billion-dollar infrastructure deficit in the US (Thompson, 2013). Currently, one in nine of America’s bridges are deemed structurally deficient (ASCE, 2013). Indeed, the Federal Highway Administration report on corrosion protection of concrete bridges estimated the average direct cost of corrosion for highway bridges in 2002 was $8.3 billion per year (Yunovich et al., 2001). Considering inflation and continued degradation, this cost is much higher in 2016. A recent (2019) study in Switzerland revealed that the direct costs caused by corrosion in road infrastructures amounts to 260e510 million Swiss Francs per year (Yilmaz and Angst, 2020) which related to the gross domestic product is several comparable order of magnitude as the figure revealed by the U.S. study 20 years ago. Given the current state of infrastructure in many parts of the world, there is a critical need to undertake corrosion research and education. The corrosion of the reinforcing steel in concrete is a serious problem from the perspective of both safety and economy and directly affects the sustainability of the infrastructure.

2.1.1

Passivation/depassivation

Concrete provides physical corrosion resistance to the steel reinforcement by acting as a barrier and chemical corrosion resistance due to its high pH. Concrete that is not exposed to any external influences usually exhibits a pH between 12.5 and 13.5 (Hansson, 1984). As shown in the Pourbaix diagram (Fig. 2.1), which defines the range of electrochemical potential and pH for the FeeH2O system in the alkaline environment, at potentials and pHs normally found for steel within the concrete, a protective passive layer forms on the surface of the steel. It is known that this layer is an ultrathin (5 nm (Mehta and Monteiro, 2006)), and their number and interconnectivity control the ingress of chloride ions, carbon dioxide, oxygen, and moisture into concrete (Jensen et al., 1999). Gel pores and interlayer spaces are believed to be too small and disconnected to significantly contribute to transport (Taylor, 1997). Two factors that significantly influence capillary porosity in concrete are the water to cementitious materials (w/cm) ratio (Kosmatka et al., 2002) and the use of supplementary cementing materials (Oh et al., 2002). Theoretically, a w/b ratio of 0.42 is required for the complete hydration of cement (Taylor, 1997). However, hydration is a gradual process and the unused mixing water is retained in the capillary pores (Aïtcin et al., 1997). Higher

26

Corrosion of Steel in Concrete Structures

w/b ratios, traditionally used to give a workable mixture, increase the amount and interconnectivity of capillary porosity in the cement paste allowing greater diffusion. With the advent of high range water reducing agents, much lower w/b ratios are now possible and significantly limit the penetration of chloride ions and carbon dioxide. The other factors that can affect the diffusion of chloride ions and carbon dioxide: • • •

• •

The inadequate cover provides a shorter path for diffusion and is regularly associated with areas of high corrosion risk due to both carbonation and chloride ingress. The age of concrete is another factor that affects the diffusion of aggressive ions (Verbeck, 1995). Over time, the curing process continues, and the diffusion becomes more difficult. Besides, diffusion is a function of time and its rate decrease with time. It is well known that the cement type or the use of supplementary cementitious materials can significantly influence the microstructure of the concrete, both in terms of the pore structure and the pore solution chemistry, and thus affect the corrosion of embedded steel. A complete review of these influencing factors is beyond the scope of this chapter. The reader is referred to dedicated literature for more details (Page and Page, 2007; Tang et al., 2012). Curing also affects the diffusion of chloride ions and carbon dioxide into the concrete (Balayssac et al., 1995; Lo and Lee, 2002). Better curing causes lower permeability, better hydration, more CH, and consequently, less carbonation and chloride diffusion. Temperature and relative humidity are the other factors that can affect the diffusion of aggressive species into concrete (Andrade et al., 1999; Jones et al., 1995; Khatib and Mangat, 2002; Woo-Yong et al., 2003). Diffusion is a function of temperature. For carbonation, there is a critical point that allows evaporation of the water, released by carbonation reactions, but does not dry out the concrete enough to stop the reaction (Mmusi et al., 2009). Relative humidity (Gonzalez and Andrade, 1982), wind and direction of sunlight, and the type of environment (e.g., pollutant and coastal regions) are the other effective factors.

2.1.6

Corrosion products

The volume of the corrosion products directly affects the proportion of concrete damage and consequently plays an important role in modeling the structural performance and service life prediction. As can be seen in Fig. 2.3 (Lide, 1999), different corrosion products of steel occupy different volumes. Suda et al. detected Fe3O4 (magnetite), a-FeOOH (goethite), and g-FeOOH (lepidocrocite) (Suda et al., 1993) while Jaffer and Hansson identified g-Fe2O3 (maghemite) (Jaffer and Hansson, 2009). The most expansive corrosion product detected to date in reinforced concrete is b-FeOOH (akaganeite), which is approximately 3.5 times the original volume of iron (Marcotte and Hansson, 1998, 2003). The reasons for such a variance in observations may be that the corrosion products were exposed to air and/ or allowed to dry out before analysis. Drying would convert the hydroxides to oxides. The lack of information on the mechanism of depassivation of the steel bars in the presence of chloride ions and the interaction between passive and active regions (anode to cathode ratio) at the nano/microscale may be one reason. One of the difficulties in determining the type of corrosion products, and consequently the volume of products, is that the corrosion product from the steel is unstable in air and changes almost immediately upon exposure to the atmosphere for laboratory analysis. An

Principles of corrosion of steel in concrete structures

27

Figure 2.3 The volume of corrosion products relative to iron.

α-Fe FeO Fe3O4 α-Fe2O3 γ-Fe2O3 δ-FeOOH α-FeOOH γ-FeOOH β-FeOOH Fe(OH)2 Fe(OH)3 Fe2O3.3H2O

0

1

2 3 4 Unit volume

5

6

7

in situ study of the corrosion products (e.g., with Raman spectroscopy and scanning electrochemical microscopy) would be advantageous in that it would provide more precise information on the volume and the distribution of the corrosion products.

2.1.7

Macrocell and microcell corrosion

Microcell corrosion is the term given to the situation where active dissolution and the corresponding cathodic half-cell reaction (the reduction of dissolved oxygen) take place at adjacent parts of the same bar, as illustrated in Fig. 2.4A. This process always occurs in practice and, in most cases, is the dominant corrosion process. Macrocell corrosion can occur when the actively corroding bar is coupled to another bar, which is passive, either because of its different composition or different environment. For example, the former situation might occur when black steel is in contact with stainless steel both in a corrosive environment (e.g., chloride contaminated concrete). The latter situation may occur when a top mat in a chloride-contaminated concrete bridge deck is coupled to a bottom mat in chloride-free and well-aerated concrete, as in Fig. 2.4B. Macrocells can also form on a single bar exposed to different environments within the concrete. The process is the same in all cases and, in all cases, the corrosive action of the macrocell is added to that of the microcells. The view of the “active steel becoming the anode and the passive steel becoming the cathode” is thus simplified and may not under all conditions be fully correct. In many situations, the anodic and cathodic reactions occur on both metal surfaces; when the two metals are coupled, the anodic corrosion of the active metal increases, and the anodic corrosion of the passive metal decreases. In situations such as localized chloride-induced corrosion, however, the anodic surfaces may indeed become largely deoxygenated so that the microcell component becomes negligible compared to the macrocell component of the corrosion process. While macrocell corrosion may be measured directly, for

28

Corrosion of Steel in Concrete Structures

Cl –

(a) O2

Cl –

O2

O2

Cl –

O2

OH– e–

1/2 O2 + H2O + 2e–

(b)

> 2OH–

Cl –

Cl –

Fe

> Fe2+ + 2e–

Cl –

Fe

> Fe2+ + 2e–

e– OH–

e–

O2

O2

1/2 O2 + H2O + 2e– O2

> 2OH– O2

Figure 2.4 Schematic illustration of (A) microcell corrosion and (B) macrocell corrosion (Hansson et al., 2006).

example, in dedicated experimental setups such as the configuration of an upper and a lower steel mat as shown in Fig. 2.4B, the same is not true of microcell corrosion. In the study by Hansson et al., 2006, it was concluded that, for carbon steel in OPC concrete with a relatively low electrical (ionic) resistance, the macrocell and microcell corrosion components were of the same order of magnitude and could be added together to provide the total corrosion rate on the actively corroding bar. In this case, the total corrosion rate was approximately three times that of the macrocell rate alone. In contrast, the macrocell corrosion of carbon steel in high-performance concrete with silica fume and either fly ash, or slag was negligible, and the corrosion was limited to microcells on the top bar by the ionic resistance of the concrete. Therefore, care must be taken in using the results of macrocell measurements. The absence of macrocell corrosion cannot be taken as an indicator that microcell corrosion is not occurring.

2.1.8

Corrosion under load

In their study of steel bars in simulated concrete pore solutions to determine passive behavior under loads and repassivation after loads were removed, Feng et al. noted

Principles of corrosion of steel in concrete structures

29

much more severe damage to passive layers under higher loads, which were hardly affected by the loading time change (Feng et al., 2011). In these applications of various loads, they also observed that the steel under lower loads did indeed repassivate after the load was removed, but did not behave similarly under higher loads, which resulted in plastic deformation. They did not, however, determine how compressive stress loads and chloride ions affect depassivation, and their experiment procedure was questionable. Specifically, this investigation was flawed in that the samples were immersed in a concrete pore solution for 24 h to form this passive layer which, based upon the previous experience, is insufficient for forming a protective layer (Poursaee and Hansson, 2007). Also, the cold air drying of samples rendered them vulnerable to a pH of less than 9, at which point the steel loses its passivity, meaning that the layer on the steel surface is a poor representation of the passive layer of steel within concrete. Finally, neither the presence of chloride ion nor its synergic effect on loading was considered when studying the depassivation of steel. Zhang and Poursaee also studied carbon steel passivation and depassivation from chloride attack in concrete simulated pore solution under different loading types (Zhang and Poursaee, 2015a, 2015b). They determined that in a chloride-free pore solution, the passive corrosion rate is less for specimens under tensile stresses than specimens under compressive stresses. In pore solutions contaminated with chloride ions, it was found that, despite the above observations, specimens (previously passivated in chloride-free pore solution) under tensile stresses corroded faster than those under either compressive stresses or no stress. These authors also noted that specimens exposed to chloride ions under compressive stresses exhibited superior performance compare to specimens under tensile or no loading types.

References ACI Committee 222, 1996. 222R-96: Corrosion of Metals in Concrete. Ahmed, S., 2003. Reinforcement corrosion in concrete structure, its monitoring and service life prediction-a review. Cement and Concrete Composite 30 (4e5), 459e471. Aïtcin, P.C., Neville, A.M., Acker, P., 1997. Integrated view of shrinkage deformation. Concrete International 19, 35e41. Allen, R.T.L., Forrester, J.A., 1983. In: Crane, A.P. (Ed.), The Investigation and Repair of Damaged Reinforced Concrete Structures. Society of Chemical Industry, London, pp. 193e222. Andrade, C., Sarria, J., Alonzo, C., 1999. Relative humidity in the interior of concrete exposed to natural and artificial weathering. Cement and Concrete Research 29 (8), 1249e1259. Angst, U., Elsener, B., 2017. The size effect in corrosion greatly influences the predicted life span of concrete infrastructures. Science Advances 3 (8), 1e8. Angst, U., Elsener, B., Larsen, C.K., Vennesland, Ø., 2009. Critical chloride content in reinforced concrete - a review. Cement and Concrete Research 39, 1122e1138. Angst, U., Rønnquist, A., Elsener, B., Larsen, C.K., Vennesland, Ø., 2011. Probabilistic considerations on the effect of specimen size on the critical chloride content in reinforced concrete. Corrosion Science 53, 177e187.

30

Corrosion of Steel in Concrete Structures

Angst, U., Geiker, M.R., Michel, A., Gehlen, C., Wong, H., Isgor, O.B., Elsener, B., Hansson, C.M., Francois, R., Hornbostel, K., Polder, R., Alonso, M.C., Sanchez, M., Correia, M.J., Criado, M., Sagues, A., Buenfeld, N., 2017. The steel-concrete interface. Materials and Structures 50 (20), 143. Angst, U., Geiker, M.R., Alonso, M.C., Polder, R., Elsener, B., Isgor, O.B., Wong, H., Michel, A., Hornbostel, K., Gehlen, C., François, R., Sanchez, M., Criado, M., Sørensen, H., Hansson, C.M., Pillai, R., Mundra, S., Gulikers, J., Raupach, M., Pacheco, J., Sag€ués, A., 2019. The effect of the steel-concrete interface on chloride-induced corrosion initiation in concrete e a critical review by RILEM TC 262-SCI. Materials and Structures 52, 88. Angst, U., Moro, F., Geiker, M., Kessler, S., Beushausen, H., Andrade, C., Lahdensivu, J., Köliö, A., Imamoto, K.-i., Greve-Dierfeld, S.v., Serdar, M., 2020. Corrosion of steel in carbonated concrete: mechanisms, practical experience, and research priorities - a critical review by RILEM TC 281-CCC. RILEM Technical Letters 5, 85e100. ASCE, 2013. 2013 Report Card for American’s Infrastructure: ASCE. ASTM, 2004. C856-04: Standard Practice for Petrographic Examination of Hardened Concrete (Vol. 04.02). Balayssac, J.P., Detriche, C.H., Grandet, J., 1995. Effect of curing upon carbonation of concrete. Construction and Building Materials 9 (2), 91e95. Bard, A.J., Mirkin, M.V. (Eds.), 2001. Scanning Electrochemical Microscopy. Marcel Dekker AG, Basel, Switzerlan. Belo, M.D.C., Hakiki, N.E., Ferreira, M.G.S., 1999. Semiconducting properties of passive films formed on nickelebase alloys type Alloy 600: influence of the alloying elements. Electrochimica Acta 44 (14), 2473e2481. Burleigh, T.D., Latanision, R.M., 1987. The use of photocurrents to characterize anodic films on Ti, Zr, Cu, and 304 stainless steel. Journal of Electroanalytical Society 134 (1), 135e141. Campbell, D.H., Sturm, R.D., Kosmatka, S.H., 1991. Detecting Carbonation. Concrete Technology Today, PL911, p. 12. http://www.portcement.org/pdf_files?PL911.pdf. Cao, Y., Gehlen, C., Angst, U., Wang, L., Wang, Z.D., Yao, Y., 2019. Critical chloride content in reinforced concrete - an updated review considering Chinese experience. Cement and Concrete Research 117, 58e68. Carnot, A., Frateur, I., Marcus, P., Tribollet, B., 2002. Corrosion mechanisms of steel concrete moulds in the presence of a demoulding agent. Journal of Applied Electrochemistry 32, 865e869. Enevoldsen, J.N., Hansson, C.M., Hope, B.B., 1994. The influence of relative humidity on the rate of chloride-induced corrosion of steel in concrete. Cement and Concrete Research 24, 1373e1382. Feng, X., Tang, Y., Zuo, Y., 2011. Influence of stress on passive behaviour of steel bars in concrete pore solution. Corrosion Science 53, 1304e1311. Germann Instruments Inc., 2006. Ghods, P., Isgor, O.B., McRae, G., Miller, T., 2009. The effect of concrete pore solution composition on the quality of passive oxide films on black steel reinforcement. Cement and Concrete Composites 31, 2e11. Glass, G.K., Reddy, B., Buenfeld, N.R., 2000. The participation of bound chloride in passive film breakdown on steel in concrete. Corrosion Science 42 (11), 2013e2021. Gonzalez, J.A., Andrade, C., 1982. Effect of carbonation, chlorides and relative ambient humidity on the corrosion of galvanized rebars embedded in concrete. British Corrosion Journal 17 (1).

Principles of corrosion of steel in concrete structures

31

Gunay, H.B., Ghods, P., Isgor, O.B., Carpenter, G.J.C., Wu, X., 2013. Characterization of atomic structure of oxide films on carbon steel in simulated concrete pore solutions using EELS. Applied Surface Science. https://doi.org/10.1016/J.APSUSC.2013.03.014. Hamada, M., 1968. Neutralization (carbonation) of concrete and corrosion of reinforcing steel. In: Paper Presented at the 5th International Symposium on Cement Chemistry, Tokyo, Japan. Hansson, C.M., 1984. Comments on electrochemical measurements of the rate of corrosion of steel in concrete. Cement and Concrete Research 14, 574e584. Hansson, C.M., Sørensen, B., 1990. The threshold concentration of chloride in concrete for the initiation of reinforcement corrosion. In: Paper Presented at the Corrosion Rates of Steel in Concrete, Baltimore, MD. Hansson, C.M., Poursaee, A., Laurent, A., 2006. Macrocell and microcell corrosion of steel in ordinary Portland cement and high performance concretes. Cement and Concrete Research 36, 2098e2102. Hansson, C.M., Poursaee, A., Jaffer, S.J., 2007. Corrosion of Reinforcing Bars in Concrete: Portland Cement Association (PCA). PCA R&D Serial No. 3013. Harrison, J.A., Williams, D.E., 1986. How does the electrochemical behavior of stainless steel reflect that of its constituent elements. Electrochimica Acta 31 (8), 1063e1072. Hausmann, D.A., 2007. Three myths about corrosion of steel in concrete. Materials Performance 46 (7), 70e73. Hong, K., Hooton, R.D., 1999. Effects of cyclic chloride exposure on penetration of concrete cover. Cement and Concrete Research 29, 1379e1386. Imamoto, K.-i., Kanematsu, M., Kiyohara, C., Hamasaki, H., Kinose, T., Teranishi, K., Noguchi, T., 2019. Field survey on rebar corrosion of carbonated existing concrete buildings in Japan. In: Paper Presented at the International Conference on Sustainable Materials, Systems and Structures (SMSS2019), Rovinj, Croatia. Jaffer, S.J., Hansson, C.M., 2009. Chloride-induced corrosion products of steel in crackedconcrete subjected to different loading conditions. Cement and Concrete Research 39, 116e125. Jensen, O.M., Hansen, P.F., Coats, A.M., Glasser, F.P., 1999. Chloride ingress in cement paste and mortar. Cement and Concrete Research 29 (9), 1497e1504. Jones, M.R., Dhir, R.K., Gill, J.P., 1995. Concrete surface treatment: effect of exposure temperature on chloride diffusion resistance. Cement and Concrete Research 25 (1), 197e208. Khatib, J.M., Mangat, P.S., 2002. Influence of high-temperature and low-humidity curing on chloride penetration in blended cement concrete. Cement and Concrete Research 32, 1743e1753. Kosmatka, S.H., Kerkhoff, B., Panarese, W.C., MacLeod, N.F., McGrath, R.J., 2002. Design and Control of Mixtures (Seventh Canadian Edition Ed. Vol. Engineering Bulletin 101). Cement Association of Canada, Ottawa, Ontario, Canada. Lahdensivu, J., 2012. Durability Properties and Actual Deterioration of Finnish Concrete Facades and Balconies. Tampere University of Technology, Tampere, Finland. Li, L., Sag€ués, A., 2004. Chloride corrosion threshold of reinforcing steel in alkaline solutions effect of specimen size. Corrosion 60 (20), 195e202. Lide, D.R., 1999. CRC Handbook of Chemistry and Physics. CRC Press, New York, NY. Lo, Y., Lee, H.M., 2002. Curing effects on carbonation of concrete using a phenolphthalein indicator and Fourier-transform infrared spectroscopy. Building and Environment 37, 507e514. Marcotte, T.D., Hansson, C.M., 1998. A comparison of the chloride-induced corrosion products from steel-reinforced industrial standard versus high performance concrete exposed to

32

Corrosion of Steel in Concrete Structures

simulated sea water. In: Paper Presented at the International Symposium on High Performance and Reactive Powder Concretes, Sherbrooke, Canada. Marcotte, T.D., Hansson, C.M., 2003. The influence of silica fume on the corrosion resistance of steel in high performance concrete exposed to simulated sea water. Journal of Materials Science 38, 4765e4776. Martin-Perez, B., Zibara, H., Hooton, R.D., Thomas, M.D.A., 2000. A study of the effect of chloride binding on service life predictions. Cement and Concrete Research 30 (8), 1215e1223. Mehta, P.K., Monteiro, P.J.M., 2006. Concrete: Microstructure, Properties and Materials. McGraw Hill, New York. Mmusi, M.O., Alexander, M.G., Beushausen, H.D., 2009. Determination of critical moisture content for carbonation of concrete. In: Paper Presented at the Concrete Repair, Rehabilitation and Retrofitting II: 2nd International Conference on Concrete Repair, Rehabilitation and Retrofitting,, Cape Town, South Africa. Montemor, M.F., Simoes, A.M.P., Ferreira, M.G.S., 1998. Analytical characterization of the passive film formed on steel in solutions simulating the concrete interstitial electrolyte. Corrosion 54 (5), 347e353. Morrison, S.R., 1981. Electrochemistry at Semiconductor and Oxidized Metal Electrodes. Plenum Press, New York. NORDTEST, 1989. NT BUILD 357: Concrete, Repairing Materials and Protective Coating: UDC 691.32:658.588 Carbonation Resistance. NORDTEST. Oh, B.H., Cha, S.W., Jang, B.S., Jang, S.Y., 2002. Development of high-performance concrete having high resistance to chloride penetration. Nuclear Engineering and Design 212, 221e231. Page, C.L., Page, M.M. (Eds.), 2007. Durability of Concrete and Cement Composites. Woodhead Publishing, Cambridge. Paola, A.D., 1989. Semiconducting properties of passive films on stainless steels. Electrochimica Acta 34 (2), 203e210. Paola, A.D., Shukla, D., 1991. Photoelectrochemical study of passive on stainless steel in neutral solutions. Electrochimica Acta 36 (2), 345e352. Paola, A.D., Quarto, F.D., Sunseri, C., 1986. A photoelectrochemical characterization of passive films on stainless steels. Corrosion Science 26 (11), 935e948. Pourbaix, M., 1974. Atlas of Electrochemical Equilibria in Aqueous Solutions. National Association of Corrosion Engineers (NACE), Houston, Texas. Poursaee, A., 2010. Determining the appropriate scan rate to perform cyclic polarization test on the steel bars in concrete. Electrochimica Acta 55 (3), 1200e1206. Poursaee, A., Hansson, C.M., 2007. Reinforcing steel passivation in mortar and pore solution. Cement and Concrete Research 37, 1127e1133. Powers, T.C., 1960. Properties of cement paste and concrete. In: Paper Presented at the Chemistry of Cement : Fourth International Symposium, Washington, D.C. USA. Revie, R.W., Uhlig, H.H., 2008. Corrosion and Corrosion Control : An Introduction to Corrosion Science and Engineering, fourth ed. Wiley. Rosenberg, A., Hansson, C.M., Andrade, C., 1989. Mechanisms of corrosion of steel in concrete. In: Skalny, J. (Ed.), The Materials Science of Concrete, vol. 1. The American Ceramic Society, pp. 285e313. Schmuki, P., Böhni, H., 1992. Metastable pitting and semiconductive properties of passive films. Journal of Electroanalytical Society 139 (7), 1908e1913. Simoes, A., Ferreira, M., Rondot, B., Belo, M., 1990. Study of passive films formed on AISI 304 stainless steel by impedance measurements and photoelectrochemistry. Journal of the Electrochemical Society 137 (1), 82e87.

Principles of corrosion of steel in concrete structures

33

Soroka, I., 1979. Portland Cement Paste and Concrete. The Macmillan Press Ltd, Surrey, England. Stefanoni, M., Angst, U., Elsener, B., 2018a. Corrosion rate of carbon steel in carbonated concrete - a critical review. Cement and Concrete Research 103, 35e48. Stefanoni, M., Angst, U., Elsener, B., 2018b. Electrochemistry and capillary condensation theory reveal the mechanism of corrosion in dense porous media. Scientific Reports 8 (1), 7407. Stefanoni, M., Angst, U., Elsener, B., 2019. Kinetics of electrochemical dissolution of metals in porous media. Nature Materials 18 (9), 942e947. Stefanoni, M., Angst, U., Elsener, B., 2020. The mechanism controlling corrosion of steel in carbonated cementitious materials in wetting and drying exposure. Cement and Concrete Composites 113, 103717. Su, K., Imamoto, K.-i., Noguchi, T., Kanematsu, M., Hamasaki, H., Teranishi, K., Kiyohara, C., Yamada, M., 2020. Post peak behavior of carbonated concrete structure - a case study of the former shime mining office vertical derrick in Japan. In: Paper Presented at the XV International Conference on Durability of Building Materials and Components, Barcelona, Spain. Suda, K., Misra, S., Motohashi, K., 1993. Corrosion products of reinforcing bars embedded in concrete. Corrosion Science 35 (5e8), 1543e1549. Sukhotin, A.M., Grilikhes, M.S., Lisovaya, E.V., 1989. The influence of passivation on the kinetics of the dissolution of iron - I. Outer layer of the passivating film as a heavy doped thin semiconductor and M-S equation. Electrochimica Acta 34 (2), 109e112. Sunseri, C., Piazza, S., Dipaola, A., Diquarto, F., 1987. A photocurrent spectroscopic investigation of passive films on ferritic stainless steels. Journal of the Electrochemical Society 134 (10), 2410e2416. Tang, L., Nilsson, L.O., Basheer, P.A.M., 2012. Resistance of Concrete to Chloride Ingress: Testing and Modelling. Spon Press, Abingdon. Taylor, H.F.W., 1997. Cement Chemistry, second ed. Thomas Telford Publishing, London. Thompson, D., 24 May 2013. The falling-bridge lesson: the U.S. infrastructure failure is still totally inexcusable. The Atlantic. Thuresson, T.I., 1996. Electrochemical Monitoring of the Influence of Concrete Quality on the Reinforcing Steel Corrosion in Industrial Effluent. MA. Sc.. Queen’s University. Tuutti, K., 1980. Service life of structures with regard to corrosion of embedded steel. In: Paper Presented at the International Conference on Performance of Concrete in Marine Environment, St. Andrews By-The-Sea, Canada. Tuutti, K., 1982. Corrosion of Steel in Concrete. CBI Research Report No. 4.82. Swedish Cement and Concrete Research Institute, Stockholm, Sweden. Verbeck, G.J., 1995. Carbonation of Hydrated Portland Cement. Research Department Bulletin RX087. Portland Cement Association. Woo-Yong, J., Young-Soo, Y., Young-Moo, S., 2003. Predicting the remaining service life of land concrete by steel corrosion. Cement and Concrete Research 33, 663e677. Yilmaz, D., Angst, U., 2020. Korrosionsbedingte Kosten an Ingenieurbauwerken im Schweizer Strassennetz. Beton- und Stahlbetonbau (in German) 115 (6), 448e458. Yunovich, M., Thompson, N.G., Balvanyos, T., Lave, L., 2001. Corrosion Costs and Preventive Strategies in the United States, Appendix D: Highway Bridges. Federal Highway Administration (FHWA), Office of Infrastructure Research and Development, Washington DC. Zakroczymski, T., Fan, C.-J., Szklarska-Smialowska, Z., 1985. Kinetics and mechanism of passive film formation on iron in 0.05M NaOH. Journal of the Electrochemical Society 132 (12), 2862e2867.

34

Corrosion of Steel in Concrete Structures

Zhang, J.Y., 2008. Corrosion of Reinforcing Steel in Concrete Structures: Understanding the Mechanisms. National Research Council (NRC)-Institute for Research in Construction, Ottawa. Zhang, Y., Poursaee, A., 2015a. Study the passivation and corrosion activity of carbon steel in concrete simulated pore solution under tensile and compressive stresses. ASCE Journal of Materials in Civil Engineering 27 (8), 1e9. Zhang, Y., Poursaee, A., 2015b. Study the semi-conductive behavior of the passive film on carbon steel in simulated concrete pore solution under stress. Anti-corrosion Methods & Materials 62 (6).

Assessing a concrete’s resistance to chloride ion ingress using the formation factor

3

Robert Spragg 1 , Chunyu Qiao 2 , Timothy Barrett 1 and Jason Weiss 3 1 Federal Highway Administration, Washington DC, United States; 2University of Science and Technology Beijing, Beijing, China; 3Oregon State University, School of Civil and Construction Engineering, Corvallis, OR, United States

3.1

Introduction

Assessing concrete quality, specifically to determine its resistance to ion ingress, has been gaining interest in recent decades. One particular method that has been gaining interest is the use of electrical resistivity or conductivity to assess the transport properties of concrete. Numerous studies have correlated resistivity to diffusion coefficients and other transport coefficients, among others are (Andrade, 1993; Tumidajski et al., 1996; Hall et al., 1997; McGrath and Hooton, 1999; McCarter et al., 2000; Castellote and Andrade, 2006; Nokken and Hooton, 2007; Radlinski et al., 2010; Rupnow and Icenogle, 2011; De la Varga et al., 2014; Bu and Weiss, 2014; Barrett, 2015). While these correlations have shown reasonable agreement, this chapter will highlight why, when assessing concrete quality, the formation factor may be preferred to resistivity alone. This chapter will begin by describing the background of electrical property measurement, that is, resistivity and conductivity, in cementitious materials and will define the formation factor. The formation factor, a microstructural material property that describes the pore network can then be related to a diffusion coefficient of different ionic species. This chapter will provide a review of the experimental techniques needed to determine the formation factor, specifically connectivity, pore volume, and pore solution resistivity and discusses the role of the degree of saturation. The procedure that has been used recently to determine the formation factor at different degrees of saturation will be presented. An illustration of the formation factor measured on Portland cement concretes saturated with different concentration of pore solutions will be presented, along with formation factor measured for variations in Portland cement concretes with moderate w/c. Lastly, the process of developing performance ranges will be discussed. An approximation utilizes surface concentration, reinforcement cover depth, and formation factor to approximate a timeto-corrosion initiation. The remainder of the chapter will discuss future research needs in advancing the use of formation factor in specifying durable concrete materials.

Corrosion of Steel in Concrete Structures. https://doi.org/10.1016/B978-0-12-821840-2.00025-0 Copyright © 2023 Elsevier Ltd. All rights reserved.

36

Corrosion of Steel in Concrete Structures

3.2

Background

3.2.1

Electrical tests in porous materials

Electrical measurements in saturated porous materials with a nonconducting solid structure can be described using the modified parallel law (Forde et al., 1981; Garboczi, 1990; Torquato and Haslach, 2002) shown here in Eq. (3.1): 1 rT ¼ ro $ fb

(3.1)

where rT is the total resistivity (or the resistivity of the total bulk sample), ro is the resistivity of the pore solution which is a function of the ionic composition and concentration in solution, f is the porosity of the system that is fluid filled, and the b is the connectivity of the pores in the system (Dullien, 1991; Xie et al., 1992; Olson et al., 1995; Rajabipour, 2006). It is often difficult to separate the fluid filled pore volume and connectivity, so literature often presents the product of these two terms together, whose inverse has been termed the formation factor, shown as F in Eq. (3.2) (Snyder, 2001). fb ¼

1 F

(3.2)

The formation factor is a material property, which is independent of specimen size or shape, but is dependent on mixture characteristics such as the water-to-cementitious materials ratio, volume of paste, and degree of hydration (Forde et al., 1981; Gu et al., 1992; Christensen et al., 1994). They can change as the system hydrates (ages), often by several orders of magnitude in typical cementitious systems (Rajabipour, 2006).

3.2.2

NernsteEinstein Relationship

The formation factor is typically referred to as a transport property because it can be used to describe the diffusion of ionic species in porous materials. Diffusion is the process by which ions move through the pore solution due to a concentration gradient and is often considered a primary transport mechanism for chloride ions through cementbased materials. The formation factor is related to the diffusion coefficient by what is known as the NernsteEinstein Relationship (Snyder, 2001). This is presented in Eq. (3.3): rT Do ¼F ¼ ro D

(3.3)

where rT is the total resistivity, ro is the resistivity of the pore solution, F is the formation factor (this is classically defined at complete saturation), Do is the selfdiffusion coefficient which describes how different ionic species move through

Assessing concrete’s resistance to chloride ion ingress

37

Table 3.1 Values of selfdiffusion coefficient, from Yuan-Hui and Gregory (1974). Selfdiffusion coefficient, Do 10L10 m2/s

Ionic species

Naþ Kþ Mg2þ OH Cl SO2 4

08C

188C

258C

6.3 9.9 3.6 25.6 10.1 5.0

11.3 16.7 5.9 44.9 17.1 8.9

13.3 19.6 7.1 52.7 20.3 10.7

water, and D is the effective diffusion coefficient. The selfdiffusion coefficient, Do , has been determined for different ionic species, and shown in Table 3.1 for particular ions that might be of interest.

3.2.3

Electrical tests in cementitious materials

While electrical resistivity tests are easily conducted, the analysis of the test results and the implications of these values can prove challenging. Most notably the microstructure and pore solution change significantly over the lifetime of the material (Hammond and Robson, 1955; McCarter et al., 1981; Rajabipour, 2006; Sant et al., 2006). Curing conditions can also influence measurements. Sealed curing versus water curing can lead to a significantly different extent of reaction (degree of hydration), which will affect both the microstructure and the pore solution properties (Weiss et al., 2013; Spragg et al., 2013b). Additionally the storage conditions can lead to different levels of drying which impacts the volume of pore fluid, its connectivity, its pore solution resistivity and correspondingly the measured resistivity (Rajabipour and Weiss, 2007). For example, Presuel-Moreno et al. have shown that storage in lime water versus storage in a 100% relative humidity chamber can lead to differences in electrical measurements (Presuel-Moreno et al., 2009). There is a dependence of the measured electrical properties on the moisture content, which can be accounted for with a correction (Weiss et al. 1999, 2013; Schiessel et al., 2000; Barrett, 2015). Further, if samples are stored in a fluid there are additional complications. First, storage in a fluid doesn’t not mean samples will be completely saturated (Rajabipour, 2006; Bu et al., 2014). The concrete sample can also leach alkali species from the pore solution into the surrounding storage solution thereby changing the pore solution resistivity (Thomas et al., 2006; Spragg, 2013; Spragg et al., 2013b). Leaching of alkalis can influence the microstructure that forms (Bu and Weiss, 2014). Two separate correction factors have been used to account for temperature the first of which accounts for rate effects associated with hydration (e.g., an equivalent age) while the second accounts for changes in conduction with testing temperature (Sant

38

Corrosion of Steel in Concrete Structures

et al., 2008). However, a general equation has been proposed to account for all of these factors on electrical resistivity measurements at an equivalent age (Spragg et al., 2013a), shown in Eq. (3.4): r Tref ¼ ro $F$f ðSÞ$f ðTtest Þ$f ðLeachÞ

(3.4)

where rTref is the resistivity at a reference temperature at an equivalent age of teq , ro is the resistivity of the pore solution at saturation, F is the formation factor, f ðSÞ is a saturation function that has typically been used to compare to vacuum saturated concretes f ðTtest Þ describes the effect of testing temperature on resistivity measurements, f ðLeachÞ is a leaching function. The equivalent age teq is a function of the time and temperature history, that is, maturity or degree of hydration. Most importantly, it can be concluded that the formation factor is the microstructural property while resistivity is a related parameter that has many outside influences.

3.2.4

Differing levels of saturation

Power’s Gel Space Ratio Theory provides a conceptual idea of the relative volumetric proportions of a hydrating cement paste (Powers and Brownyard, 1946). The series of papers dealing with the model provided a quantification of the reaction between cement and water. The major tenets of Power’s model are •

• • •

Initially, the system is composed of cement and water. The free water is termed capillary water. The volume ratio of initial capillary water is given by the term initial porosity, p ¼ ðw =cÞ=ðw =c þrw =rc Þ. Where w=c is the water to cementitious mass ratio, rw is the density of water, and rc is the density of cement. The amount of reacted cement is termed the degree of hydration, a; where 0 means no cement is reacted and 1.0 means all of the cement has reacted. The reaction between cement and water results in a reduction of volume termed chemical shrinkage, which is approximately 6.4 mL/100 g reacted cement. Hydration products are classified as gel solids and gel water, with the later approximately 28% of the volume of the reacted products.

A graphical representation of Power’s model is shown in Fig. 3.1 for a 0.42 w/c paste, after (Jensen and Hansen, 2001). Initially, the volume fraction of capillary water is characterized by the term initial porosity described above, with the remainder of the volume representing cement. As the degree of hydration progress, the formation of gel solids and gel water take place, with a reduction in the volume of capillary water and unhydrated cement. It should be noted that while both capillary water and gel water represents porosity, that is, water-filled voids in the system, they represent very different size scales. Capillary pores are on the order of 10 mm to 10 nm (Mindess et al., 2003). Capillary water behaves as bulk water and pores of this large size and connectedness are a major contributor to the mass transport that occurs through the matrix. Gel pores are smaller, on the order of 10e2 nm and are a porosity inherent in the hydration process of cement

Assessing concrete’s resistance to chloride ion ingress

39

Figure 3.1 Graphical illustration of power’s model for a 0.42 w/c paste.

(Mindess et al., 2003). Since the gel pores are much smaller, they do not contribute much to the mass transport (Sant et al., 2011; Barrett, 2015). Additionally, due to reduction in volume between reactants and products, the formation of chemical shrinkage occurs. Chemical shrinkage is vapor filled voids in a sealed sample, contrasted with the gel and capillary pores which are typically viewed as fluid-filled porosity. As Power’s model describes cement paste, its application to concrete can be done by taking into account that concrete is typically only 25%e32% paste. As an example, Fig. 3.2 shows a 28% paste concrete with a 0.42 w/c. The paste portion is similar to that shown in Fig. 3.1; however, it only occupies 28% of the volume. Concrete typically also contains entrained air voids, usually on the order of 6%e8% by total volume, in this example, an air content of 6.5% is used. The remainder of the volume is composed of aggregates (Barrett, 2015; Todak, 2015). Air voids are typically air bubbles that are stabilized during the mixing process. As such, they are often vapor filled pores (and correspondingly nearly nonconductive) ranging in size from 0.05 to 1 mm (Lamond and Pielert, 2006; Lucero, 2015). Figure 3.2 Graphical illustration of power’s model for a 0.42 w/c concrete, with a 28% paste volume, 6.5% air content.

40

Corrosion of Steel in Concrete Structures

To assess paste properties with electrical tests and formation factor calculations, it is important to understand which of these pores are filled with fluid. One way to describe the pore space that is filled with fluid is to evaluate the degree of saturation. This is defined in Eq. (3.5). S¼

Fluid in specimen mat moisture state  mOD ¼ Fluid in specimen at Complete Saturation mSSD  mOD

(3.5)

Where mat moisture state is the specimen mass at the moisture state in question, mOD is the oven dry mass of the specimen, and mSSD is the saturated-surface dry mass of the specimen after vacuum saturation at an absolute pressure of 6 Torr (Bu et al., 2014). This parameter describes the fraction of the porosity that is filled with fluid at a given moisture state. The degree of saturation in a sealed sample can be calculated based upon Powers model (Barrett, 2015; Todak, 2015). In a paste system, the gel pores and the capillary pores remain fluid filled. The vapor filled space of chemical shrinkage represents the only reduction in degree of saturation (in a nonair entrained system). As such, the sealed degree of saturation can be calculated as simply the volume of gel and capillary pores, divided by the sum of the volume of gel pores, capillary pores, and chemical shrinkage. In the paste system, if the chemical shrinkage is filled, the degree of saturation would be 100%. Fig. 3.3a shows the reduction in degree of saturation as a function of degree for a sealed system or when the pores in the paste are saturated (Barrett, 2015; Todak, 2015). Concrete has three moisture states that are of the most interest when evaluating transport properties. These are a sealed condition; a condition where the porosity in the paste is fluid filled; and where all the porosity is fluid filled (Barrett, 2015). Calculation of the theoretical degree of saturation in each of these conditions can be accomplished by considering powers model and the volume of entrained air voids, noting that the air voids are not fluid filled for a sealed sample or a sample that is exposed to

Figure 3.3 Changes in the degree of saturation for a 0.42 (a) paste system or (b) 6.5% air content concrete system. The nick point represents when the paste porosity is filled, after Weiss et al. (2016).

Assessing concrete’s resistance to chloride ion ingress

41

drying. As such, the air content represents a significant reduction in the degree of saturation even before any chemical shrinkage develops in the paste. The first condition would be the sealed case, where no moisture is allowed to enter or leave the specimen. The degree of saturation in this case would be the gel and capillary pores divided by the gel and capillary pores, the chemical shrinkage volume, and the volume of air voids. The second moisture state would be the condition where the air voids remained empty but the porosity in the paste is filled, which has been termed the Nick Point as it corresponds to a kink in water absorption behavior (Todak et al., 2015). The degree of saturation in this case is calculated as the porosity in the paste (chemical shrinkage, gel, and capillary pores) divided by the paste porosity and the volume of air. Lastly, would be complete saturation. This does not just mean submerged in a tank, but vacuum saturated to allow for complete degassing and to allow the fluid to fill in all the voids in the concrete (Bu et al., 2014).

3.3

Experimental techniques

3.3.1

Concrete resistivity

The three most common methods for measuring the resistivity of concrete include: uniaxial, surface, and embedded geometries discussed in detail in this section. Resistivity is a material property that is independent of specimen size or shape and testing configuration. Values of resistivity are obtained from measurements of resistance, and related by Eq. (3.6), where r is the resistivity and k is the geometry factor, which depends on specimen size and shape, and testing configuration (Fig. 3.4). r ¼ R$k

(3.6)

Other configurations utilize embedded sensors. These have been used extensively in the literature, for example (Raupach and Schiessl, 1997; Weiss, 1999; Schiessel et al., 2000; Polder, 2001; Rajabipour et al., 2005; Rajabipour, 2006; Poursaee and Weiss, 2010; Castro et al., 2010; McCarter et al., 2015). These are often sensors that can be placed directly into the fresh concrete and cast-in-place. These have the benefit of being able to be placed into a specimen that is then sealed and testing would not necessitate removal from a sealed condition and risk extensive loss of moisture. This would be ideal for the monitoring of sealed specimens. As discussed previously, resistivity is material property that is independent of specimen geometry and electrode configuration. However, the resistivity is determined from a test of resistance. This resistance must be corrected for specimen size and electrode configuration to determine the resistivity. This correction has traditionally been termed the geometry factor, denoted using k in Eq. (3.6). Geometry factors can be determined experimentally by comparing to measurement of a solution of known resistivity or numerically (Morris et al., 1996; Rajabipour, 2006; Spragg et al., 2013b). The units of resistivity are typically ohm$length, where

42

Corrosion of Steel in Concrete Structures

Figure 3.4 Common testing geometries used on standard 100 mm  200 mm test cylinders in the (a) surface and the (b) uniaxial configurations.

the ohm comes from the measure of resistance and the measure of length comes from the geometry factor. When the resistance is multiplied by the correct geometry factor, the resistivity tests that are performed using samples of differing geometry will show consistent results between samples of different geometries (Spragg et al. 2012, 2013b). Also, resistivity measurements are temperature sensitive. The measurements conducted in this chapter were conducted at 23  1 C. If measurements are conducted at another temperature, they should be corrected for temperature back to 23 C reference temperature before they are used to compute the formation factor (Chrisp et al., 2001; Sant et al., 2008; Spragg et al., 2013b).

3.3.1.1

Surface configuration: Wenner Test

The Surface Resistivity test, often termed the Wenner Test, was initially developed for use in soil testing by Frank Wenner at the National Bureau of Standards in the 1910s (Wenner, 1916). A schematic of the four-point surface resistivity test applied to a concrete cylinder is shown in Fig. 3.5. In the development of the equations that are used to interpret the results from this test, an assumption of an infinite half-space was assumed, that is, the spacing of the electrodes is much smaller than the depth of the material being measured

Assessing concrete’s resistance to chloride ion ingress

43

Figure 3.5 Schematic of the surface resistivity test being conducted on a test cylinder, reproduced from Spragg et al. (2013b).

(a  d; L). If this criteria is satisfied, the geometry factor is given by b k 1 in Eq. (3.7): b k 1 ¼ 2pa

(3.7)

where a represents the spacing of the electrodes. Many commercial surface resistivity meters do this correction automatically. However, when testing is conducted on a cylindrical concrete test specimen, the assumption of infinite-half space is not completely satisfied because of the small testing geometry and the large probe tip spacing that is required to prevent interferences with coarse aggregates (Presuel-Moreno et al.; AASHTO TP95-11, 2011; Morris et al., 1996). An additional factor is often needed to account for constricted current flow in the material. Morris et al. (1996) has shown, using finite element simulations, the magnitude of additional corrections needed when using surface resistivity tests on various cylindrical geometries. These data have been digitized and can be described by Eq. (3.8) (Spragg et al., 2013b): 0:730 7:34 b k 2 ¼ 1:10    1 þ   2 d a d a

(3.8)

where a represents the spacing of the electrodes, d is the diameter of the test cylinder. It should be noted that this correction is only when d=a  6.0 and L=a 6.0, where L is

44

Corrosion of Steel in Concrete Structures

the length of the cylinder. For testing on a standard 100 mm  200 mm test cylinder, this value ranges from 1.8 to 1.9. Often times, b k 1 is applied automatically by the resistivity meter before a number is  displayed on the machine (Proceq, 2013). In that case, k ¼ 1 b k 2 should be applied. In the case that the resistivity meter simply shows the measured resistance, the geometry factor for a surface test is given by Eq. (3.9): k¼

b k1 b k2

(3.9)

It also worth noting that Eq. (3.9) is only valid for a (spatially) homogeneous system, and shows good agreement when the material is indeed homogeneous (Spragg et al., 2016a, 2016b). The surface resistivity test has gained popular usage by the Florida (FM 5-578, 2004; Vivas et al., 2007) and Louisiana (Rupnow and Icenogle, 2011) Departments of Transportation. Many sources in the literature advocate the use of this method because it can allow for measurements in situ, but a dependence on geometry, degree of saturation, leaching of alkalis and ingress of deicing salts applied from deicing operations, temperature, and location of reinforcing steel, means that in situ measurements, when assessing concrete quality, should be interpreted carefully (PresuelMoreno et al., 2009; Spragg et al., 2013a; Salehi et al., 2014).

3.3.1.2

Uniaxial configuration

The uniaxial resistivity test is another resistivity configuration for use on a concrete specimen (McCarter et al., 1981). A schematic of the uniaxial test is shown in Fig. 3.6. This configuration typically consists of a set of plates that are placed at the ends of the specimen with a conductive medium, typically a wet sponge or conductive gel, being used to create good electrical contact between the specimen and the electrodes. This configuration has the benefit of a more uniform current distribution throughout the sample and is not as dependent on surface characteristics. The geometry factor for this configuration is shown in Eq. (3.10): k¼

A L

(3.10)

where A is the cross-sectional area of the specimen and L is the length of the specimen. One of the important assumptions of this test is that there is a good electrical connection between the test specimen and the electrodes, typically accomplished through the use of a conductive medium (Newlands et al., 2007). This can be a conductive gel, a paper towel, or a sponge wetted with conductive solution (Shane, 2000; Thomas, 2008; Spragg et al., 2012). In some cases, these elements have an associated resistance. There has been a correction proposed to the measured resistance, to correct

Assessing concrete’s resistance to chloride ion ingress

45

Figure 3.6 Schematic of the uniaxial resistivity test being conducted on a test cylinder, reproduced from Spragg et al. (2013b).

for the resistance of the sponges, shown in Eq. (3.11) (for resistors in series) (Newlands et al., 2007): R ¼ Rmeasured  Rtop sponge  Rbottom sponge

(3.11)

where R is the resistance used in Eq. (3.6), Rmeasured is the measured resistance of the top sponge, specimen, and bottom sponge, and Rtop sponge and Rbottom sponge are the resistance of just the top and bottom sponge, respectively. When measuring the resistance of each sponge, it is important to use the pressure that each sponge experiences during the resistivity test. An example is shown in Fig. 3.7. Research has indicated that if a low enough resistance solution is used for the sponges and the concrete is of high resistivity, the correction tends to be negligible (Spragg et al., 2012).

3.3.1.3

Embedded configuration

The last configuration is the embedded configuration. While the literature has had many successful variants of the embedded sensor, one that has shown particular promise is a set of stainless steel rods in a standard 150 mm  300 mm test cylinder, shown here in Fig. 3.8. This particular configuration has been described elsewhere in the literature, for example (Castro et al., 2010; Spragg et al., 2013b). For this configuration, the geometry factor often needs to be determined numerically or experimentally. The experimental determination can be done using a

46

Corrosion of Steel in Concrete Structures

Figure 3.7 Configuration for the measurement of (a) Rtop sponge and (b) Rbottom sponge .

Figure 3.8 Schematic of the embedded configuration in a modified 150 mm  300 mm test cylinder, reproduced from Spragg et al. (2013b). HDPE, high density polyethylene.

Assessing concrete’s resistance to chloride ion ingress

47

solution of known conductivity, such as potassium chloride whose conductivity is presented in many chemistry or physics handbooks, for example (Dean and Lange, 1999). It is worth noting, that a solution of significantly low resistivity, that is, higher concentration, should be used, as high resistivity solutions can introduce spurious results.

3.3.2

Pore solution resistivity

The resistivity of the pore solution is an important factor when determining the formation factor. There are typically two methods of obtaining pore solution, experimental or theoretical.

3.3.2.1

Experimental pore solution expression

Experimental pore solution expression can be performed in accordance with a technique discussed by Barneyback and Diamond (1981). This technique using paste samples made from the cementitious materials of interest, using their relative proportions from the concrete of interest. The specimens are placed into a high pressure die system and specimens are squeezed, which expresses the pore solution. The resistivity of the expressed pore solution can then be measured by using a small pore solution cell. The high-pressure die and piston system is described by Barneyback and Diamond (1981) and shown schematically in Fig. 3.9. For this approach, approximately 50 mm diameter samples of cement pastes are made and then placed into the die. The apparatus is placed into a press or compression machine and loaded to express the solution from the pores. The fluid is collected in a syringe or small vial. Typically, only a few milliliters of solution are obtained. Once the solution is obtained, its electrical resistivity can be measured by testing in a pore solution resistivity cell, such as the one pictured in Fig. 3.10a (Castro et al., 2010). A schematic is shown in Fig. 3.10b, and is made from 3/800 diameter polycarbonate tubing approximately 100 long, end plates of stainless steel, with a small hole drilled one of the steel plates. The center of the plates is slightly milled such that the tube can fit in the end and no solution is lost. The solution is then filled from the hole, and the plates connected to an electrical resistivity meter, whose resistance is then measured. The resistivity of the pore solution can be determined using Eq. (3.6). The geometry factor for the type of cell shown in Fig. 3.10 is given by Eq. (3.10), that is, ratio of cross-sectional area to length. Another method of experimentally determining the pore solution resistivity is through the use of an embedded pore solution sensing system (Rajabipour et al., 2007; Rajabipour and Weiss, 2007). The system consists of two sensors and prior to use a calibration curve is developed to account for different relative humidities and solution resistivities. The system can then be embedded into fresh concrete, and the resistivities of the two sensors are measured. Using the resistivities of the sensors and the calibration curve, the pore solution resistivity can be determined (Rajabipour, 2006; Rajabipour et al., 2007).

48

Corrosion of Steel in Concrete Structures

Figure 3.9 High-pressure die and piston system for experimentally expressing pore solution from hardened cement pastes, after Barneyback and Diamond (1981).

Figure 3.10 Example pore solution cell shown in (a) photograph or (b) schematic.

Assessing concrete’s resistance to chloride ion ingress

3.3.2.2

49

Theoretical pore solution

An alternative method of approximating the pore solution resistivity is an estimation based upon the theoretical pore solution composition. The theoretical pore solution composition can be estimated from the chemistry of the cementitious materials, the mixture proportions, and the degree of hydration. This procedure was originally described by Taylor (1987). It has since been implemented in an online HTML/Javscript tool by Bentz (2007), available at http://concrete. nist.gov/poresolncalc.html. This method initially requires determination of the pore solution composition. For this, the moles (N) of Sodium (Na) and Potassium (K) are determined from the chemistry of the cementitious materials, specifically mass percentages of alkalis, which is information typically reported in the mill certificate. Also needed is the mass from the mixture design of each of the cementitious materials. This calculation is based upon the total amount of alkalis in the cementitious materials; however, not all of these materials are released into the pore solution. This concept has been incorporated into the online tool using the free ion factor (Bentz, 2007) and can be summarized as: •

• • •

75% of potentially available Na and K ions from the cement, silica fume, and fly are present in the pore solution after approximately 24 h. Prior to this time, sulfate reactions are occurring and can impact the pore solution chemistry and pore connectivity (De La Varga, 2013; Spragg et al., 2013b). The alkalis present in the slag cement are assumed to be incorporated into slag hydration products, and the alkalis in the slag will not be released into the pore solution. Alkalis are strongly absorbed by silica fume, and for mass fractions above 15%, the total amount of alkalis in the pore solution are only 45% of the total alkalis. Electroneutrality must be maintained, so the sum of the cation concentration, that is, Naþ and Kþ, are used to compute the concentration of OH in the pore solution.

For sealed specimens, Powers’ model can be used to estimate the volume of pore solution, in terms of L of solution per mass of cementitious materials, as simply the sum of capillary and gel water. The online tool also has an option for what it terms “saturated” curing, but this specifically means when the matrix is filled, which would mean the volume of pore solution is the sum of capillary, gel, and chemical shrinkage volumes (Bentz, 2007). This accounts for the dilution of the pore solution concentration by adding the volume of chemical shrinkage, but does not account for alkali leaching that can occur when moist curing (Spragg et al., 2016b). Previous research has shown good agreement between experimentally expressed, measured solutions and theatrical pore solution resistivity values (De La Varga, 2013; Spragg et al., 2013b; Bu and Weiss, 2014) for sealed cured, paste samples. The best agreement comes after about a 20% degree of hydration. It has been suggested that this is due aluminate reactions that can cause variations in the pore solution composition, specifically the presence of sulfate ions and a lower concentration of hydroxyl ions (Spragg et al., 2013b). Additionally, air entrained concretes can present challenges. Previous research has shown that changes in entrained air content, i.e., different dosages of air entraining

50

Corrosion of Steel in Concrete Structures

admixture, have shown to have little effect on sealed resistivity measurements (Castro et al., 2010). This makes sense, as the air entrained is typically thought as being nonfluid filled (and therefore nonconductive) in sealed systems. When the concretes are vacuum saturated, the pore solution concentration is diluted, as the fluid filled volume (in a sealed system) goes from the volume of gel and capillary pores in the matrix to the volume of gel, capillary, and chemical shrinkage of matrix and the entrained air volume (Barrett, 2015). The amount of dilution for a sealed system can be estimated by the air dilution factor shown in Eq. (3.12) and termed dAir and can be applied to the ionic concentration of each species, that is, Kþ, Naþ and OH. The premise of this factor is reducing the number of ions originally in the volume of gel and capillary pores (termed “water left” in the online Bentz tool) by the volume of gel and capillary pores, chemical shrinkage, and air content, which should be in units of L water/kg cement. The dilution parameter is shown in Eq. (3.12), and is termed dsealedsaturated to describe the dilution that takes place when going from a sealed specimen to a specimen that is vacuum saturated in lime water. The Bentz online tool (Bentz, 2007) has an option for a similar dilution, but only accounts for the volume of chemical shrinkage, not the volume of air (Barrett, 2015). The dilution parameter, dsealedsaturated , can be calculated according to Eq. (3.12). The terms A, B, and C are defined in Table 3.2. Where w=cm is the total water to cementitious materials ratio by mass, and a is the degree of hydration. The 0.23 and 0.06 coefficients that describe the reduction in porosity and chemical shrinkage, as discussed by Bentz, are generally accepted values but could be modified for a specific cement (Powers, 1966; Bentz, 2007). Parameter C describes the volume of air entrained voids, but converts from a volume percentage to consistent units with A and B. For this equation, the percentage of air should be used along with the total weight of cementitious materials per cubic yard. The factor of 16.878 is a factor that makes the conversion, when using % and lb of cementitious materials to L of water per kg of cementitious. dAir ¼

A AþBþC

(3.12)

Table 3.2 Description of the terms used in the calculation of the dilution parameter to account for dilution of the pore solution when vacuum saturating an air-entrained concrete. Letter

Description

Calculation

Units

A

Volume of gel and capillary porosity Volume of chemical shrinkage Volume of air

w=cm  0:23a

L/kg cementitious

0:06a

L/kg cementitious

B C

 3   Air ð%Þ ft L $27 $28:3 3 100 ft yd 3   kg Total Cementitious ðlbÞ$0:453 lb Or

Air ð%Þ Total Cementitious ðlbÞ$16:878

L/kg cementitious

Assessing concrete’s resistance to chloride ion ingress

51

The dilution factor, dAir , is applied to the ionic concentration given by the online tool. From there, the resistivity can be calculated as calculated as described by literature (Snyder et al., 2003; Bentz, 2007). This dilution of the pore solution due to filling in the air voids can be significant, and in the case of 6.5% air content and a 0.42 w/c, the dilution can represent a pore solution resistivity that is nearly double that of pore solution that is present when the air voids are not filled.

3.3.3

Determining the formation factor

The experiments described in Section 4 evaluate the resistivity of two conditions: sealed and vacuum saturated. As discussed in Section 2.4, for concrete materials, this evaluates the matrix pores (sealed condition) or the entire pore system (saturated condition), including the entrained air voids (Barrett, 2015). For sealed measurements, it is important to ensure test specimens are sufficiently sealed to prevent moisture loss from the hydrating concrete to the environment. To accomplish this, this work has utilized specimens that were left sealed in their original cylinder molds and placed in an additional 6 mil sealed plastic bag. When incorporated into a specification, in an effort to ensure that the specimens are adequately sealed, a mass change criteria between arrival at testing lab and time of testing could be specified (Spragg et al., 2013a; Barrett, 2015). The sealed formation factor Fsealed can be calculated by the ratio of concrete resistivity measured in a sealed condition, rsealed , to the resistivity of the pore solution in a sealed condition, rosealed , shown in Eq. (3.13). Fsealed ¼

rsealed rosealed

(3.13)

Specimens for saturated formation factor, Fsaturated , testing should first be oven dried then vacuum saturated using lime water at an absolute pressure of 6 Torr for a period of 3 h, adding deaired lime water for an additional 1 h at the 6 Torr vacuum level, and allowed to sit under lime water for an additional 18 h (Bu et al., 2014). The specimens can then be measured for their saturated resistivity, rsaturated . The pore solution at saturation, rosaturated , can be calculated as described in Section 3.2.2. From these two parameters, rsaturated and rosaturated ; the saturated formation can be calculated using Eq. (3.14). Fsaturated ¼

3.4

rsaturated rosaturated

(3.14)

Microstructural parameters

Formation Factor at Saturation for Concrete Materials. This study evaluated a Portland cement concrete at an age of 91 d, with the mixture proportions listed in Table 3.3. The samples were sealed cured, in their original cylinder molds and inside a 6 mil, sealed plastic bag.

52

Corrosion of Steel in Concrete Structures

Table 3.3 Mixture proportions of concrete evaluated for formation factor, in SSD condition. Constituent

Proportion (lb/yd3)

Cement Water Fine aggregate Coarse aggregate Air content w/c

658 267 1178 1780 6.3% 0.40

At an age of 91 d the specimens were demolded and prepared for testing. Specimens were oven dried, and vacuum saturated with two different solutions: (a) lime water and (b) artificial pore solution. The specimens were tested 24 h after vacuum saturation for resistivity. The pore solution was approximated for the lime water saturated specimens by the procedure described in Section 3.2.2, accounting for the dilution due to air content. The short amount of time under water was assumed to have negligible effects on the leaching of alkalis. The second set of specimens was saturated with artificial pore solution, with concentrations determined from the procedure determined in Section 3.2.2. The pore solution for this system would be twice the concentration, that is, the alkalis that were initially in solution that remained when the specimen was oven dried and the ions that were present in the artificial pore solution. This would result in a pore solution that is less than half the resistivity. This is summarized in Table 3.4. It is seen that while concrete specimens saturated with different solutions, that is, different pore solutions, had a drastically different resistivity, their formation factor

Table 3.4 Experimental data from concrete electrical measurements. Resistivity Specimen

Pore solution (mU$m)

Specimen (U$m)

Formation factor (unitless)

OPC-1a OPC-1b OPC-2a OPC-2b

95 95 40 40

21.6 22.8 8.6 9.4 Average St Dev

227 240 215 235 229.3 10.8

Assessing concrete’s resistance to chloride ion ingress

53

was relatively constant. A standard deviation was measured which corresponds to a coefficient of variation of 4.7%.

3.4.1

Influence of mixture proportions

The influence of mixture proportions on the formation factor was also investigated as a part of this work. Specifically, a concrete with a paste volume fraction of 28.2% was investigated using with the proportions given in Table 3.5. The cement used in this portion of the study was a type I OPC with 0.16% Na2O and 1.24% K2O content by mass. Specimens were sealed cured for a period of 1 year, demolded, and the resistivity measured in a sealed condition. The sealed pore solution resistivity was estimated using the online tool discussed by Bentz (2007). The specimens were oven dried then vacuum saturated in lime water, and the saturated pore solution concentration was estimated using the measured air contents and the sealed pore solution concentration. The resistivity was then calculated from the concentrations, as described by literature (Snyder et al., 2003; Bentz, 2007), and shown in Table 3.6. The resistivity and calculated formation factors for the three different w/c mixtures are shown in Fig. 3.11. The sealed measurements shown in Fig. 3.11a were conducted immediately following the sample being demolded from a sealed condition. Fig. 3.11b shows measurement that were conducted on the same specimens after being vacuum saturated in lime water. Testing was conducted at 23  1 C. First, it can be seen that both the values of resistivity and formation factor change quite significantly between sealed measurements and vacuum saturated

Table 3.5 Mixture proportions for different w/c mixtures with a constant paste content, in SSD lb/yd3. w/c

0.42

0.48

0.54

Fresh air content

5.0%

5.2%

6.5%

Cement Water Fine aggregate Coarse aggregate

645 271 1178 1780

596 286 1178 1780

554 300 1178 1780

Table 3.6 Estimated pore solution resistivities for sealed cured concretes at 90% degree of hydration, in mU m. w/c

0.42

0.48

0.54

Sealed pore solution Vacuum saturated

48 83

60 97

71 117

54

Corrosion of Steel in Concrete Structures

Figure 3.11 Plain concretes made with the same Type I OPC shown as (a) sealed measurements and (b) saturated factor at an age of 1 year.

measurements. This has been well documented and has been shown to follow an approximate power curve with degree of saturation between high values at low degrees of saturation and the saturated value (Schiessel et al., 2000; Weiss et al., 2013). Furthermore, recent research has shown that the fitting coefficient is a relatively constant value, except when considering the air entrained concrete (Barrett, 2015). The volume and spacing of the air entrained voids controls the amount of reduction in sealed resistivity. It is worth noting that for this study there is a clear trend with resistivity and w/c. The concretes in this study are made with the same Type I OPC, have similar air contents, and are made with a constant paste volume fraction. This trend in resistivity and w/c has often been observed in literature, for example (Henkensiefken et al., 2009), which makes resistivity good at seeing variations in water content of a concrete batch using consistent materials. Specifically, for use as a quality control test.

3.5 3.5.1

Specifying formation factor Why measure formation factor

Many studies have evaluated the relationship between concrete resistivity and diffusion properties and see that these relationships are often empirical and mixture specific (Berke and Hicks, 1992). One primary reason for this is because each binder system has its own pore solution properties. By normalizing the concrete resistivity to the pore solution resistivity, the formation factor can be determined. The formation factor, F described in Eq. (3.2), is the microstructural characterization parameters that describes the conductive volume and the connectivity of the volume. While simple resistivity tests describe pore characteristics and pore solution resistivity, normalizing by the pore solution gives a numerical quantification of the pore network that is independent of the pore solution resistivity. While this is important for OPC concretes, it becomes especially important for binder systems with supplementary materials. Research has shown that different

Assessing concrete’s resistance to chloride ion ingress

55

supplementary cements can have a large variation in pore solution properties (Feng et al., 2004). Furthermore, a test of formation factor can be accomplished relatively easily, with an electrical testing taking only a few minutes. Alternative tests, such as a standard diffusion test, can take upwards of months, for testing and data analysis.

3.5.2

Relating formation factor to life cycle

Recall that the formation factor is a microstructural material property, described in Section 2. The NernsteEinstein Relationship relates the formation factor to a measure of resistivity but also to an effective diffusion coefficient. This means that the formation factor can be used to obtain a target value for use in quality control testing, that is, resistivity, but also can be related to an estimated time to corrosion (Barrett, 2015; Weiss et al., 2016). One approach that would provide a conservative estimate of the relationship between formation factor and time to corrosion life cycle would be the use of the error function solution to Fick’s second law. This is termed conservative because the use of this equation does not account for the aging effects of the transport properties, effects of chloride binding or pore blocking reactions that would slow ionic ingress, the multimodal transport of chlorides, effects of temperature, or the gradual buildup of surface chlorides (Boddy et al., 1999; Bentz et al., 2014). The goal of the procedure, described below, is to allow for the quantitative comparison of microstructure quality. As it is a tool to measure the microstructure of the matrix phase of concrete, it is suggested that it be conducted not on saturated specimens, but on specimens that have been sealed from initial hydration or on specimens where only the matrix phase is filled with fluid (Bu et al., 2014; Todak et al., 2015; Barrett, 2015). The benefit of this method is that it would assess the durability of the matrix phase and allow for the air void system to be assessed separately. An estimated time to corrosion can be calculated using the error function solution to Fick’s second law, shown in Eq. (3.15).  x  Cðx; tÞ  Co ¼ 1  erf pffiffiffiffiffi Cs  Co 2 Dt

(3.15)

Where Cðx; tÞ is the concentration at a depth x at an exposure time t and typically taken as 0.05% by weight of concrete (a value typically recognized as corresponding to corrosion of reinforcing steel), Co is the background chloride content typically taken as 0.02% by weight of concrete, Cs is the surface concentration and D is the apparent chloride diffusion coefficient. The surface concentration Cs can be varied depending on the exposure class, that is, the level of chloride exposure at the surface, during inservice conditions. The apparent diffusion in Eq. (3.15) can be replaced by the ratio of the selfdiffusion coefficient of a chloride ion, 19.3  1010 m2/s at 23 C, to the formation factor. Specifically, in Eq. (3.15), D ¼ Do =F.

56

Corrosion of Steel in Concrete Structures

Figure 3.12 Influence of formation factor for a 50 mm cover on the estimated time to corrosion. Exposure classes of surface concentrations of 0.1%, 0.5%, and 1% represent a low, medium, and severe exposure class.

Fig. 3.12 shows the calculation for time to corrosion initiation for a 50 mm cover depth and a series of three different surface concentrations. These values were selected based upon surface concentrations from around the country, given by Life 365 (Life365, 2012). The values of 0.1%, 0.5%, and 1% by weight of concrete represent a low, moderate, and severe exposure category. This approach allows to quantify the relative improvement with an increase in formation factor for a mixture. The use of the formation factor to specify durable concrete shows promise because it can be consistently applied to a wide range of systems, primarily because many times the pore solution chemistry of these systems change. While the approach taken here provides a conservative estimation to the time to corrosion, a more detailed analysis could be done if needed that could account for variation in temperature, effects of chloride binding, and aging of the formation factor as shown by (Glasser et al., 2008; Barrett, 2015; Weiss et al., 2016).

3.6

Conclusions

Assessing a concrete’s resistance to ion penetration is primarily a function of the pore characteristics. While electrical tests provide some measure of the pore structure, electrical measurements are highly dependent on the pore solution. In cementitious systems, the pore solution can change quite appreciably depending on the mixture proportions, chemistry of the cementitious materials, and the degree of hydration. By normalizing the measured concrete resistivity by the resistivity of pore solution, the parameter of formation factor can be obtained. This is a numerical quantification of the microstructure. The use of formation factor will help to make a specification applicable to a wide range of cementitious systems that can have a variation in pore solution properties. Furthermore, the formation factor can be obtained in a few days of testing. This can be contrasted with other microstructural characterization tests, such as a diffusion test, which can take weeks or even months.

Assessing concrete’s resistance to chloride ion ingress

57

The NernsteEinstein relationship demonstrates how the formation factor can be used to determine the effective diffusion coefficient for a series of different ionic species. Using Fick’s second law, a conservative estimate of the time to corrosion initiation can be estimated for a series of different exposure classes and bar depths. Recent studies have shown that by properly accounting for ionic binding, aging, and temperature that the formation factor can serve as an input to life-cycle models that have accurately predicted chloride profiles of laboratory ponded-specimens. The formation factor can also be related to a simple measure of resistivity. This means that a formation factor specification can be used to establish target values for resistivity during concrete production by simply multiplying the specified formation factor by the pore solution resistivity (Barrett, 2015; Weiss et al., 2016). This would make a specification generic and would apply for a wide range of cementitious materials.

References AASHTO TP95-11, 2011. Standard Method of Test for Surface Resistivity Indication of Concrete’s Ability to Resist Chloride Ion Penetration. American Association of State Highway and Transportation Officials, Washington, D.C. Andrade, C., 1993. Calculation of chloride diffusion coefficients in concrete from ionic migration measurements. Cement and Concrete Researc 23, 724e742. https://doi.org/ 10.1016/0008-8846(93)90023-3. Barneyback, R.S., Diamond, S., 1981. Expression and analysis of pore fluids from hardened cement pastes and mortars. Cement and Concrete Research 11, 279e285. https://doi.org/ 10.1016/0008-8846(81)90069-7. Barrett, T., 2015. Improving Service Life of Concrete Structures Through the Use of Internal Curing: Impact on Practice. PhD. Purdue University. Bentz, D.P., 2007. A virtual rapid chloride permeability test. Cement and Concrete Composites 29, 723e731. https://doi.org/10.1016/j.cemconcomp.2007.06.006. Bentz, D.P., Guthrie, W.S., Jones, S.Z., Martys, N.S., 2014. Predicting service life of steelreinforced concrete exposed to chlorides. Concrete International 55e64. Berke, N.S., Hicks, M.C., 1992. Estimating the life cycle of reinforced concrete decks and marine piles using laboratory diffusion and corrosion data. In: Chaker, V. (Ed.), Corrosion Forms and Control for Infrastructure, ASTM STP 1137. ASTM International, Philadelphia, Pennsylvania, pp. 207e231. Boddy, A., Bentz, E., Thomas, M., Hooton, R., 1999. Overview and sensitivity study of a multimechanistic chloride transport model. Cement and Concrete Research 29, 827e837. https://doi.org/10.1016/S0008-8846(99)00045-9. Bu, Y., Spragg, R., Weiss, J., 2014. Comparison of the pore volume in concrete as determined using ASTM C642 and vacuum saturation. Advances in Civil Engineering Materials. https://doi.org/10.1520/ACEM20130090. Bu, Y., Weiss, J., 2014. The influence of alkali content on the electrical resistivity and transport properties of cementitious materials. Cement and Concrete Composites 51, 49e58. https:// doi.org/10.1016/j.cemconcomp.2014.02.008.

58

Corrosion of Steel in Concrete Structures

Castellote, M., Andrade, C., 2006. Round-Robin Test on methods for determining chloride transport parameters in concrete. Materials and Structures 39, 955e990. https://doi.org/ 10.1617/s11527-006-9193-x. Castro, J., Spragg, R., Kompare, P., Weiss, J., 2010. Portland Cement Concrete Pavement Permeability Performance. West Lafayette, Indiana. Chrisp, T.M., Starrs, G., McCarter, W.J., et al., 2001. Temperature-conductivity relationships for concrete: an activation energy approach. Journal of Materials Science Letters 20, 1085e1087. https://doi.org/10.1023/A:1010926426753. Christensen, B.J., Coverdale, T., Olson, R.A., et al., 1994. Impedance spectroscopy of hydrating cement-based materials: measurement, interpretation, and application. Journal of the American Ceramic Society 77, 2789e2804. https://doi.org/10.1111/j.1151-2916.1994. tb04507.x. De La Varga, I., 2013. Increased Fly Ash Volume and Internal Curing in Concrete Structures and Pavements. PhD. Purdue University. De la Varga, I., Spragg, R.P., Di Bella, C., et al., 2014. Fluid transport in high volume fly ash mixtures with and without internal curing. Cement and Concrete Composites 45, 102e110. https://doi.org/10.1016/j.cemconcomp.2013.09.017. Dean, J., Lange, N., 1999. Lange’s Handbook of Chemistry. McGraw-Hill, New York, NY. Dullien, F.A.L., 1991. Porous Media: Fluid Transport and Pore Structure, second ed. Academic Press, San Diego. Feng, X., Garboczi, E., Bullard, J., et al., 2004. Expanding a Tool for Predicting Chloride Diffusivity in Concrete So It Can Be Used by Manufacturers to Evaluate the Durability of Concrete Made with Blended Cements. Part I: Characterizing Blended Cement Materials (Gaithersburg, MD). FM 5-578, 2004. Florida Method of Test for Concrete Resistivity as an Electrical Indicator of its Permeability (Tallahassee, FL). Forde, M.C., McCarter, J., Whittington, H.W., 1981. The conduction of electricity through concrete. Magazine of Concrete Research 33, 48e60. https://doi.org/10.1680/macr.1981. 33.114.48. Garboczi, E.J., 1990. Permeability, diffusivity, and microstructural parameters: a critical review. Cement and Concrete Research 20, 591e601. https://doi.org/10.1016/0008-8846(90) 90101-3. Glasser, F.P., Marchand, J., Samson, E., 2008. Durability of concreteddegradation phenomena involving detrimental chemical reactions. Cement and Concrete Research 38, 226e246. https://doi.org/10.1016/j.cemconres.2007.09.015. Gu, P., Xie, P., Beaudoin, J.J., Brousseau, R., 1992. A.C. impedance spectroscopy (I): a new equivalent circuit model for hydrated Portland cement paste. Cement and Concrete Research 22, 833e840. https://doi.org/10.1016/0008-8846(92)90107-7. Hall, C., Marchand, J., Gerard, G., Sosoro, M., 1997. Transport of fluids in homogenous isotropic cementitious composites. In: Reinhardt, H. (Ed.), Penetration and Permeability of Concrete, RILEM Report 16, pp. 5e79. Hammond, E., Robson, T., 1955. Comparison of electrical properties of various cements and concretes. I-II. The Engineer 199, 78e80. Henkensiefken, R., Castro, J., Bentz, D.P., et al., 2009. Water absorption in internally cured mortar made with water-filled lightweight aggregate. Cement and Concrete Research 39, 883e892. https://doi.org/10.1016/j.cemconres.2009.06.009. Jensen, O.M., Hansen, P.F., 2001. Water-entrained cement-based materialsdI. Principles and theoretical background. Cement and Concrete Research 31, 647e654. https://doi.org/ 10.1016/S0008-8846(01)00463-X.

Assessing concrete’s resistance to chloride ion ingress

59

Lamond, J., Pielert, J. (Eds.), 2006. Significance of Tests and Properties of Concrete and Concrete-Making Materials, STP 169D, STP 169D. ASTM International, West Conshohocken, PA. Life-365, 2012. Life-365 User Manual, pp. 1e80. Lucero, C.L., 2015. Quantifying Moisture Transport in Cementitious Materials. MSCE, Purdue University. McCarter, W.J., Chrisp, T.M., Starrs, G., et al., 2015. A durability performance-index for concrete: developments in a novel test method. International Journal of Structural Engineering 6, 2. https://doi.org/10.1504/IJSTRUCTE.2015.067966. McCarter, W.J., Forde, M.C., Whittington, H.W., 1981. Resistivity characteristics of concrete. ICE Proceedings 71, 107e117. https://doi.org/10.1680/iicep.1982.1993. McCarter, W.J., Starrs, G., Chrisp, T.M., 2000. Electrical conductivity, diffusion, and permeability of Portland cement-based mortars. Cement and Concrete Research 30, 1395e1400. https://doi.org/10.1016/S0008-8846(00)00281-7. McGrath, P.F., Hooton, R.D., 1999. Re-evaluation of the AASHTO T259 90-day salt ponding test. Cement and Concrete Research 29, 1239e1248. https://doi.org/10.1016/S00088846(99)00058-7. Mindess, S., Young, J.F., Darwin, D., 2003. Concrete, second ed. Prentice Hall, Upper Saddle River, NJ. Morris, W., Moreno, E.I.I., Sag€ués, A.A., 1996. Practical evaluation of resistivity of concrete in test cylinders using a Wenner array probe. Cement and Concrete Research 26, 1779e1787. https://doi.org/10.1016/S0008-8846(96)00175-5. Newlands, M., Jones, R., Kandasami, S., Harrison, T., 2007. Sensitivity of electrode contact solutions and contact pressure in assessing electrical resistivity of concrete. Materials and Structures 41, 621e632. https://doi.org/10.1617/s11527-007-9257-6. Nokken, M.R., Hooton, R.D., 2007. Using pore parameters to estimate permeability or conductivity of concrete. Materials and Structures 41, 1e16. https://doi.org/10.1617/s11527006-9212-y. Olson, R.A.A., Christensen, B.J.J., Coverdale, R.T.T., et al., 1995. Interpretation of the impedance spectroscopy of cement paste via computer modellingdPart III Microstructural analysis of frozen cement paste. Journal of Materials Science 30, 5078e5086. https:// doi.org/10.1007/BF00356052. Polder, R.B., 2001. Test methods for on site measurement of resistivity of concreteda RILEM TC-154 technical recommendation. Construction and Building Materials 15, 125e131. https://doi.org/10.1016/S0950-0618(00)00061-1. Poursaee, A., Weiss, J., 2010. An automated electrical monitoring system (AEMS) to assess property development in concrete. Automation in Construction 19, 485e490. https:// doi.org/10.1016/j.autcon.2009.12.016. Powers, T.C., 1966. The Nature of Concrete. Portland Cement Association, Skokie, IL. Powers, T.C., Brownyard, T.L., 1946. Studies of the physical properties of hardened portland cement paste. Journal of the American Concrete Institute. https://doi.org/10.14359/15301. Presuel-Moreno, F., Liu, Y., Paredes, M., 2009. Understanding the Effect of Rebar Presence and/or Multilayered concrete resistivity on the apparent surface resistivity measured via the four-point wenner method. In: Corrosion 2009. NACE International, Houston, TX. Presuel-Moreno F, Suares A, Liu Y Characterization of New and Old Concrete Structures Using Surface Resistivity Measurements. Florida Atlantic UniversitydSea Tech Campus, Dania Beach, FL. Proceq, 2013. Resipod family operating instructions. Schwerzenbach.

60

Corrosion of Steel in Concrete Structures

Radlinski, M., Olek, J., Nantung, T., 2010. Development and application of maturity method for prediction of concrete’s resistance to chloride ion penetration. Transportation Research Record Journal of the Transportation Research Board 2164, 105e112. https://doi.org/ 10.3141/2164-14. Rajabipour, F., 2006. Insitu Electrical Sensing and Material Health Monitoring in Concrete Structures. Purdue University. PhD. Rajabipour, F., Sant, G., Weiss, J., 2007. Development of electrical conductivity-based sensors for health monitoring of concrete materials development of electrical conductivity-based sensors for health monitoring of concrete materials. In: Proceedings of the Transportation Research Board, pp. 1e16. Washington, DC. Rajabipour, F., Weiss, J., 2007. Electrical conductivity of drying cement paste. Materials and Structures 40, 1143e1160. https://doi.org/10.1617/s11527-006-9211-z. Rajabipour, F., Weiss, J., Shane, J.D., et al., 2005. Procedure to interpret electrical conductivity measurements in cover concrete during rewetting. Journal of Materials in Civil Engineering 17, 586e594. https://doi.org/10.1061/(ASCE)0899-1561(2005)17:5(586). Raupach, M., Schiessl, P., 1997. Monitoring system for the penetration of chlorides, carbonation and the corrosion risk for the reinforcement. Construction and Building Materials 11, 207e214. https://doi.org/10.1016/S0950-0618(97)00039-1. Rupnow, T.T., Icenogle, P., 2011. Evaluation of Surface Resistivity Measurements as an Alternative to the Rapid Chloride Permeability Test for Quality Assurance and Acceptance. Louisiana Department of Transportation, Baton Rouge, LA. Salehi, M., Ghods, P., Burkan Isgor, O., 2014. Numerical investigation of the role of embedded reinforcement mesh on electrical resistivity measurements of concrete using the Wenner probe technique. Materials and Structures. https://doi.org/10.1617/s11527014-0498-x. Sant, G., Bentz, D.P., Weiss, J., 2011. Capillary porosity depercolation in cement-based materials: measurement techniques and factors which influence their interpretation. Cement and Concrete Research 41, 854e864. https://doi.org/10.1016/j.cemconres.2011.04.006. Sant, G., Rajabipour, F., Fishman, P., et al., 2006. Electrical conductivity measurements in cement paste at early ages: a discussion of the contribution of pore solution conductivity, volume, and connectivity to the overall electrical response. In: RILEM Workshop on Advanced Testing of Fresh Cementitious Materials, pp. 213e222. Stuttgart, Germany. Sant, G., Rajabipour, F., Weiss, W.J., 2008. The influence of temperature on electrical conductivity measurements and maturity predictions in cementitious materials during hydration. Indian Concrete Journal 82, 7e16. Schiessel, A., Weiss, W.J., Shane, J.D., et al., 2000. Assessing the moisture profile of drying concrete using impedance spectroscopy. Concrete Science and Engineering 2, 106e116. Shane, J., 2000. Electrical Conductivity and Transport Properties of Cement-Based Materials Measured by Impedance Spectroscopy. PhD. Northwestern University. Snyder, K., Feng, X., Keen, B., Mason, T., 2003. Estimating the electrical conductivity of cement paste pore solutions from OH, Kþ and Naþ concentrations. Cement and Concrete Research 33, 793e798. Snyder, K.A., 2001. The relationship between the formation factor and the diffusion coefficient of porous materials saturated with concentrated electrolytes: theoretical and experimental considerations. Concrete Science and Engineering 3, 216e224. Spragg, R., Bu, Y., Snyder, K.A., et al., 2013a. Electrical Testing of Cement-Based Materials: Role of Testing Techniques, Sample Conditioning, and Accelerated Curing. Purdue University, West Lafayette, Indiana.

Assessing concrete’s resistance to chloride ion ingress

61

Spragg, R., Villani, C., Snyder, K., et al., 2013b. Factors that influence electrical resistivity measurements in cementitious systems. Transportation Research Record Journal of the Transportation Research Board 2342, 90e98. https://doi.org/10.3141/2342-11. Spragg, R., Villani, C., Weiss, J., 2016a. Electrical properties of cementitious systems: formation factor determination and the influence of conditioning procedures. Advances in Civil Engineering Materials. https://doi.org/10.1520/ACEM20150035. Spragg, R.P., 2013. The Rapid Assessment of Transport Properties of Cementitious Materials Using Electrical Methods. MSCE, Purdue University. Spragg, R.P., Castro, J., Nantung, T., et al., 2012. Variability analysis of the bulk resistivity measured using concrete cylinders. Advances in Civil Engineering Materials 1, 104596. https://doi.org/10.1520/ACEM104596. Spragg, R.P., Jones, S.Z., Bu, Y., et al., 2016b. Alkali-Leaching in Cementitious Systems: Implications to Measurements of Electrical Resistivity. Taylor, H.F.W., 1987. A method for predicting alkali ion concentrations in cement pore solutions. Advances in Cement Research 1, 5e17. https://doi.org/10.1680/adcr.1987.1.1.5. Thomas, M., 2008. The Use of Conductive Gel. Thomas, M., Fournier, B., Folliard, K., et al., 2006. Test methods for evaluating preventive measures for controlling expansion due to alkali-silica reaction in concrete. Cement and Concrete Research 36, 1842e1856. https://doi.org/10.1016/j.cemconres.2006.01.014. Todak, H., 2015. Durability Assessments of Concrete Using Electrical Properties and Acoustic Emissions. MSCE, Purdue University. Todak, H., Lucero, C., Weiss, W.J., 2015. Why is the air there? Thinking about freeze-thaw in terms of saturation. Concrete inFocusdSpring 27e33. Torquato, S., Haslach, H., 2002. Random Heterogeneous Materials: Microstructure and Macroscopic Properties. Springer, New York, NY. Tumidajski, P.J., Schumacher, a. S., Perron, S., et al., 1996. On the relationship between porosity and electrical resistivity in cementitious systems. Cement and Concrete Research 26, 539e544. https://doi.org/10.1016/0008-8846(96)00017-8. Vivas, E., Hamilton, H., Boyd, A., 2007. Permeability of ConcretedComparison of Conductivity and Diffusion Methods (Gainesville, FL). Weiss, J., Barrett, T., Qiao, C., Todak, H., 2016. Toward a specification for transport properties of concrete based on the formation factor of a sealed specimen. In: Transportation Research Board 95th Annual Meeting (Washington, DC. Weiss, J., Shane, J., Mieses, A., et al., 1999. Aspects of monitoring moisture changes using electrical impedance spectroscopy. In: Persson, B., Fagerlund, G. (Eds.), Self Desiccation and its Importance in Concrete Technology. Lund Institute of Technology, Lund, Sweden, pp. 31e48. Weiss, J., Snyder, K., Bullard, J., Bentz, D.P., 2013. Using a saturation function to interpret the electrical properties of partially saturated concrete. Journal of Materials in Civil Engineering 25, 1097e1106. https://doi.org/10.1061/(ASCE)MT.1943-5533.0000549. Weiss, W.J., 1999. Prediction of Early-Age Shrinkage Cracking in Concrete Elements. PhD. Northwestern University. Wenner, F., 1916. A method of measuring earth resistivity. Bull Bur Stand 12, 469. https:// doi.org/10.6028/bulletin.282. Xie, P., Gu, P., Zu, A., Beaudoin, J.J., 1992. A rationalized A.C. Impedance model for the microstructural characterization of hydrating cement systems. Cement and Concrete Research 23, 359e367. Yuan-Hui, L., Gregory, S., 1974. Diffusion of ions in sea water and in deep-sea sediments. Geochimica et Cosmochimica Acta 38, 703e714. https://doi.org/10.1016/0016-7037(74) 90145-8.

This page intentionally left blank

Chloride corrosion threshold

4

Carolyn M. Hansson Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, ON, Canada

4.1 4.1.1

Background The concept

The concept that a chloride corrosion threshold value, CCT, exists, below which corrosion of rebar would not occur and above which active corrosion would be initiated was introduced more than 5 decades ago (Hausmann, 1964; Richartz, 1969). From subsequent attempts at measuring this concentration, two aspects of this concept soon became clear. The first is that there is no unique value; instead, it is a function of a large number of variables/factors. The second is that the concept of such a value is a laboratory construct, in that a value can be determined in the laboratory for a specific combination of these factors, but that it cannot be determined for actual structures in the field. As explained in earlier chapters, chloride-induced corrosion is not uniform and, because it is not feasible to continuously monitor all parts of the reinforcement in a structure, it is impossible for the chloride concentration of a particular location to be determined as soon as it starts to actively corrode. Nevertheless, knowledge of the chloride threshold value is of major interest to the practicing engineer who may need to use water or aggregates that are contaminated with chlorides, and to those involved in inspection, repair and maintenance of structures exposed to marine environments or deicing agents. It is also critical for the development of service life models for reinforced concrete exposed to these environments. Most such models are patterned after the hypothesis proposed by Tuutti (1982a,b) that describes an incubation period during which the steel remains passive until chlorides from the environment reach the steel at the critical concentration level. Thereafter, the rate of active corrosion determines the remaining life of the structure, as illustrated in Fig. 4.1. Since the publication of this thesis, the concepts of initiation period and critical chloride concentration have been considered indispensable for service life models. Unfortunately, the third factor, the rate of corrosion subsequent to active corrosion initiation, has often been ignored on the assumption that it contributes very little to the overall service life (ACI365.1R-17, 2017).

4.1.2

The variability of CCT values

There have been several extensive reviews of the available CCT data in the literature over the last 2 decades, all of which confirm the very large variability in reported

Corrosion of Steel in Concrete Structures. https://doi.org/10.1016/B978-0-12-821840-2.00002-X Copyright © 2023 Elsevier Ltd. All rights reserved.

64

Corrosion of Steel in Concrete Structures

Figure 4.1 “If only the rate of corrosion after initiation is taken into consideration, the life of a concrete structure cannot be assessed. The initiation time can be completely decisive for the life of the structure.”(Tuutti 1982a,b).

values (Glass and Buenfeld, 1997; Ann and Song, 2007; Alonso and Sanchez, 2009; Angst et al., 2009; Cao et al., 2019). These are attributed to different experimental test procedures, different combinations of concrete mixtures and reinforcing alloys and different exposure conditions Angst et al. (2022) provides a detailed analysis of the data and recommendations of the specific details to be reported in future publications. The experimental techniques employed have included half-cell potential monitoring, macrocell current measurements, cyclic polarization curve determination, linear polarization resistance measurements, potentiostatic tests, and electrochemical impedance spectroscopy, which are described in Chapter 12. The immediate environment around the rebar has included concrete, mortar, and synthetic pore solution. In addition, the type and condition of rebar, the exposure conditions, for example, outdoor, laboratory, continuous immersion or wet/dry cycles, and the type and concentration of the salts have all been found to influence the chloride threshold value. This chapter will provide some details of each of these factors as well as examples of the various testing methods that have been used.

4.2 4.2.1

Experimental laboratory tests Tests in synthetic concrete pore solution

As described in Chapter 2, corrosion initiation and subsequent active corrosion rates are controlled by the availability and composition of the ionic solution in the pores of the concrete. In noncontaminated concrete, this is a highly alkaline solution (generally with pH in excess of 13), containing KOH and NaOH with some Ca(OH)2 and minor amounts of sulfates and other cations (Barneyback and Diamond, 1981, Byfors

Chloride corrosion threshold

65

et al., 1986; Tritthart, 1990, Ghods et al., 2009). The specific composition of this solution is a function of cement type; any supplementary cementitious materials and/or chemical admixtures (Byfors et al., 1986, Arya et al., 1990, Luo et al., 2003, Scott and Alexander, 2016, Vollpracht et al., 2016, Van Niejenhuis et al., 2020); the water-tocementitious materials ratio (w/cm) (Buckley et al., 2007; Van Niejenhuis et al., 2020; Arya et al., 1990); and the age of the concrete (Arya et al., 1990; Buckley et al., 2007). However, chlorides can alter the pore solution composition quite significantly, depending on the specific salt (Hansson and Sørensen, 1990; De Weerdt et al., 2015) and to a degree dependent on whether they were present in the concrete mix or penetrated the hardened concrete from the environment (Arya et al., 1990; Enevoldsen et al., 1994). Some of the chlorides chemically react with components of the cement, particularly with ettringite and monosulphate, replacing some of the sulfate component of these phases and leaving less Cl and more SO24 in the pore solution (Van Niejenhuis et al., 2020). One of the “shortcuts” used to study corrosion of reinforcement and determine the critical chloride threshold is to use a synthetic version of this ionic solution rather than embed the bars in concrete or mortar. The advantages of this include a more rapid test, control of the chloride content and easier determination of active corrosion initiation, as well as the location and extent of the corrosion. Various compositions of the “synthetic pore solution” have been used, from a simple, saturated Ca(OH)2 solution with pH 12.6 at 25 C (Li and Sag€ ués, 2002; Hurley and Scully, 2006; Yang et al., 2014), to the compositions of analyzed pore solution from cements with different (admixed) chloride contents, with pH levels typically >13 (Li and Sag€ués, 2002; Randström et al., 2010; Scott and Thomas, 2013; Yang et al., 2014). It should be noted, however, that tests in pore solution inevitably give a higher value of CCT than those in concrete because the chloride concentration is uniform over the rebar surface and, so there is less possibility of any galvanic contribution between local anodic and cathodic sites as there is in rebar embedded in concrete. Moreover, there is negligible solution resistance in the system. Because a ribbed bar with mill scale takes some time in concrete to form its equilibrium passive film (Poursaee and Hansson, 2007), it is generally recommended for pore solution tests that the rebar be exposed to the pore solution and allowed to passivate for a few days prior to any measurements and/or additions of salt to the solution. After the initial equilibration period, a chloride salt can be added incrementally to the solution and the corrosion behavior monitored by one of the techniques mention above at each incremental Cl concentration. The chloride threshold is designated as that concentration at which the potentials, Ecorr, becomes significantly more negative and the current density, icorr, increases sharply, as illustrated in Fig. 4.2, determined by EIS (Jiang et al., 2012). Unfortunately, observations of a sharp change in Ecorr and icorr are dependent on the test procedure and so the threshold values are not always clear. Instead, the chlorides can produce defects in the passive film which increase in density with increasing Cl concentration, resulting in a gradual increase in leakage current, that is, passive current density, determined by linear polarization resistance for three grades of rebar

66

Corrosion of Steel in Concrete Structures

Figure 4.2 (A) Variations in the Ecorr and (B) the icorr of polished steel bars in saturated Ca(OH)2 solutions with concentrations of chloride ions derived from NaCl, KCl, MgCl2, and CaCl2 determined by EIS (Jiang et al., 2012).

as shown in Fig. 4.3. Similar behavior has been observed for rebar in mortar and was attributed to a general increase on the total area of active corrosion with the increase of chlorides (Oh et al., 2003). An alternative technique, which has been shown to be applicable to both carbon steel and stainless steel rebar, involves determining anodic cyclic polarization curves

Figure 4.3 Average values of corrosion current density determined by LPR for the three reinforcing steels, black steel, 304 stainless steel and LDX 2101, duplex stainless steel, tested in synthetic pore solution with different chloride additions (Jaffer and Hansson, 2010).

Chloride corrosion threshold

67

for bars in synthetic pore solution with different chloride contents and monitoring both the half-cell potential, ECORR and the pitting potential, EPIT, Fig. 4.4A. These values are then plotted versus chloride content and the two sets of data extrapolated until they intersect, as shown in Fig. 4.4B. The assumption is made that the intersection represents the chloride content of the pore solution at which corrosion will initiate without any applied potential, thus representing the threshold value(Ogunsanya and Hansson, 2018; Van Niejenhuis et al., 2020). Of course, for all evaluations of the threshold value using pore solution tests, the data must then be translated into the corresponding chloride content of cement or concrete. This can be accomplished for a specific cementitious mixture if the synthetic pore solution is based on the pore solution composition of that mixture. It should be noted that the pore solution compositions used in the above CCT tests were based on those expressed from cement pastes with admixed NaCl (Ogunsanya and Hansson, 2018). Different values are obtained using this CCT test but with different salts (Van Niejenhuis et al., 2020). A third pore solution test method is the measurement of the macrocell current, which is described in the ASTM A955 Procedure 1 “rapid macrocell test” (ASTM, 2018). This test was specifically designed for stainless steel but there is no reason why it should not also be used for other rebar grades. It consists of 2 cells, one containing two lengths of rebar immersed in synthetic pore solution, which act as cathodes. The second cell contains one length of rebar, the anode, immersed in the same synthetic pore solution, but also containing NaCl. The 2 cells are connected via a salt bridge and the 3 bars are connected across a standard resister. The potential across the resister is monitored and, applying Ohm’s Law, the galvanic current between the anode and cathode bars can be calculated. By varying the chloride content of the anode cell, a measure of the CCT can be obtained.

Figure 4.4 (A) Anodic cyclic polarization curves for triplicate specimens of black rebar tested in synthetic pore solution with different additions of NaCl; (B) the open circuit potentials and pitting potentials from (A) plotted versus chloride content of the solution (Ogunsanya and Hansson, 2018).

68

4.2.2

Corrosion of Steel in Concrete Structures

Tests in concrete and mortar

Exposing reinforced concrete to chloride solution and monitoring the corrosion potential is the traditional method of determining the initiation time for corrosion and, if followed by removing and analyzing the concrete cover can also provide the chloride threshold value. While this is the most realistic method, it can take many years for results. The following test is designed to provide more rapid results. Potentiostatic polarization tests have been used to measure the CCT of rebar embedded in concrete and immersed in a chloride solution. A small anodic overpotential (100e400 mV) is applied to the rebar which has two effects: it serves to drive the chloride into the mortar toward the rebar more rapidly than diffusion and it allows the corrosion current to be monitored. When the current starts to increase (and this is can be by as much as three orders of magnitude), the specimens can be autopsied and the chloride content of the cover adjacent to the rebar can be determined by titration or by X-ray fluorescence or by mass spectroscopy. This technique has been used by a number of researchers including (Hansson and Sørensen, 1990; Hurley and Scully, 2006). An increasingly common method of determining the CCT of rebar in concrete or mortar is a macrocell corrosion, or galvanic corrosion, method. The attraction of this test is its low cost: it requires only a high impedance voltmeter and a standard resistor. Also, the results do not require sophisticated analysis. One macrocell corrosion test is outlined in the ASTM G-109 (ASTM, 2007), in which a concrete prism, such as that in Fig. 4.3, is cast with a single rebar (the anode) embedded 25 mm below a ponding well and connected via a standard resistor to two couple bars (the cathodes) at the bottom of the prism. The potential across the resister is monitored and, applying Ohm’s Law, the galvanic current between the anode and cathodes can be calculated, Fig. 4.5. The ponding well is filled with a chloride solution for 2 weeks and kept dry for 2 weeks and cycled every 4 weeks until at least half the samples exhibit an

Figure 4.5 G-109 concrete prism.

Chloride corrosion threshold

69

integrated macrocell current of at least 15 Coulombs. The bars are then autopsied and the chloride content of the concrete adjacent to the top bar is then measured to determine the CCT value. There have been a number of variations of the macrocell corrosion test, including one recently recommended by the ACI 222 Committee on Corrosion of Metals in Concrete (Trejo et al., 2021) In this case, the anode and cathode bars are cast into separate mortar specimens. The anode bars are embedded into individual “dog-bone” specimens, and attached to a prism containing several longer bars, Fig. 4.6. Because of the very small (5 mm) cover depth on the anode bars, this test can be completed in a much shorter time than the ASTM G109 test.

Figure 4.6 ACI Committee 222 recommended macrocell test: top: front view; bottom: top view showing the set-up for 10 replicate specimens. Courtesy D. Trejo, Oregon State University.

70

Corrosion of Steel in Concrete Structures

Similarly, to determine the effect of different anode/cathode area ratios, Chalhoubet al. (2019) used a 6 mm length of rebar cast into a small mortar cylinder attached to a 16 mm long rebar cast into a second large mortar cylinder. The “anode” cylinders were dried at 45 C and 25% RH until they reached a constant weight and then immersed in a chloride solution. After reaching an equilibrium weight, both anode and cathode cylinders were immersed in a tank of NaOH solution, connected via a resistor, and the galvanic current monitored for 180 h. Thereafter, the anode cylinders were autopsied and the corroded areas of the bars measured to determine the actual anode/cathode area ratio. This procedure was carried out for steels in four different surface condition: as received, with mill scale; pickled to remove the mill scale; pickled then oxidized at 400 C; and pickled and rusted in a humid atmosphere. The results presented in Fig. 4.7 show that despite the huge difference in the cathode/anode area ratio, the corrosion initiation occurred at approximately the same free chloride content. Moreover, the CCT values for the as-received, oxidized, and corroded bars were very similar, 0.6% wt. cement, but that for the pickled bar was approximately 2.0% wt. cement.

4.3

Field tests

Because of the logistical difficulties and the costs, as well as the heterogeneities of the materials and the inconsistencies in exposure conditions of different parts of structures, there have been far fewer attempts to determine critical chloride thresholds in existing structures. The techniques and devices employed for this purpose are the subjects of Chapters 13e15 and will not be discussed further here.

4.4

The influence of concrete mix design and exposure conditions

It is highly likely that chlorides penetrating into the hardened cement or concrete would give very different CCT values from those obtained in concrete with admixed

Figure 4.7 Macrocell current measured for different free chloride contents for apparent cathode/ anode area ratios of 16 and 2950 (Chalhoubet al., 2019).

Chloride corrosion threshold

71

chlorides and, thus, than those in synthetic pore solution based on expressed solutions from the latter. For example, an investigation of (i) mortar with 2% admixed chloride (as NaCl) soaked in Ca(OH)2 solution; (ii) chloride-free mortar soaked in in Ca(OH)2 with 3.5% Cl; and (iii) mortar with admixed chlorides and also soaked in the Ca(OH)2 with 3.5% Cl solution (Enevoldsen et al., 1994). After 1 year in solution, the mortars were analyzed, by DSC, for the presence of chloroaluminate phases and, by titration, for the amount of chloride in solution. Chloroaluminate was observed in all specimens but the water-soluble chloride content of the specimens with admixed chlorides was more than twice that of the initially chloride free specimens immersed in Cl solution. However, the bound chloride (i.e., the acid-soluble minus the water-soluble chloride) was w0.04% by mass of cement in both cases. Moreover, Alonzo and Sanchez (Alonso and Sanchez, 2009) have reported that “the chloride thresholds determined from natural tests, adding chlorides or penetrating did not show significant differences.” There is a large body of research covering the influence of concrete mix design on the diffusion and migration of chlorides into concrete as discussed in Chapter 3. Mineral admixtures, such as ground granulated blast furnace slag, fly ash, or silica fume, produce such a significantly beneficial reduction in the penetration rates of chlorides However, much less attention has been given to their effect once the chlorides reach the rebar. Many service life models assume the propagation period to be much shorter than the initiation time and therefore, simply add a few years to the initiation time or assume they are equivalent (ACI365.1R-17, 2017). Nevertheless, the mineral admixtures do have an influence on the CCT and it is generally detrimental. Examples include the decrease in CCT value observed in field exposure tests with increasing cement replacement by fly ash (Thomas, 1996). Similarly, Manera et al. (2008) observed a lower CCT for bars in selfcompacting concrete with 10% silica fume than bars in Portland cement selfcompacting concrete. Although Cao et al. (1994) did not measure the CCT, they found the pitting potential of OPC to be significantly more positive than those of blended cements with fly ash or slag. In contrast, Cao et al. (2019) report that, in laboratory tests in mortar, Li et al. (2014) found slightly higher CCT values for the blended cement concretes than for OPC concrete.

4.5

The CCT of corrosion-resistant grades of rebar

Rebar grades are often categorized as carbon steel (black) rebar, corrosion resistant rebar, or stainless steel rebar. The middle category includes galvanized bar, described in detail in Chapter 8) and ferritic low chromium steels: the approximately 9% Cr (ASTM A1035), and 3Cr12, containing 11%e12% Cr (S41003). The most commonly available stainless steels in North America are the austenitic grades 304L (S30403) and 316LN (S31653) and duplex grades 318 (S32205) and S2304. Several additional stainless steel grades are available in Europe. The stainless steels are discussed in more detail in Chapter 6.

72

Corrosion of Steel in Concrete Structures

Many researchers investigating the corrosion behavior of different rebar grades have presented their chloride threshold data relative to that of carbon steel, rather than provide specific values. Thus, for example, Lysogorskiet al. (2005) report the CCT of several alloys tested in synthetic pore solution in terms of rank as follows: 316 z 304 z 2205 z SSC1 > SSC2   > LDX2201    z 9% Cr > 3Cr12 > carbon steel *316 Stainless steel clad from Stelex **316 Stainless steel clad from SMI Corp ***Proprietary duplex stainless steel from Outokumpu

Similarly, Hurley and Scully (2006) gave the following ranking for resistance to corrosion in potentiostatic tests in synthetic pore solution 316LN > 2102 > 9% Cr > carbon steel

4.5.1

Galvanized steel

Unlike iron and steel, zinc is not naturally passivated at the high pH of concrete. However, after a short period in the concrete, galvanized steel develops a protective coating of calcium hydroxyzincate, which limits the corrosion rate to that similar to passivated carbon steel (Tan and Hansson, 2008). Therefore, when embedded galvanized steel is exposed to chlorides, the chlorides must breakdown this protective layer to activate corrosion of the zinc layer. When the galvanized layer is breeched locally, the zinc then acts as a sacrificial anode and corrodes more rapidly, while the exposed steel corrodes much more slowly. Maldonado (2009) found the CCT of galvanized steel rebar to be two to three times that of carbon steel rebar embedded in concrete cylinders and exposed on the coast of the Yucatan Peninsula for several years. Similarly, Darwin et al. (2009) found the threshold value of galvanized to be almost twice that of the carbon steel bar but with greater variability.

4.5.2

Low chromium steels (w9% Cr & 3Cr12)

As mentioned above, two lower chromium grades of steel are available as reinforcement and have shown to be more corrosion resistant to chlorides than carbon steel but do not have the performance of the stainless steels. The approximately 9% Cr steel, also known as microcomposite steel, has an ultrafine martensitic microstructure with fine laths of austenite which gives the steel a much higher strength (almost 700 MPa) than carbon steel while also providing toughness. The chloride threshold values quoted in the literature vary from twice (Scott and Thomas, 2013) to more than 9 times ( (Pillai and Trejo, 2005) that of carbon steel. The w12% Cr alloy (3Cr12, S41003) has received very little attention although it was shown to be significantly more corrosion resistant than carbon steel in seawater (Callaghan, 1993) and to have a CCT twice that of carbon steel but somewhat less than that of the 9% microcomposite steel (Lysogorskiet al., 2005).

Chloride corrosion threshold

4.5.3

73

Stainless steels

Stainless steels used for rebar are typically austenitic with 17%e18% Cr and 8%e10% Ni. Because stainless steels can withstand significantly higher levels of chlorides than the other rebar types, they require even more accelerated tests in order to obtain results in a reasonable amount of time. Consequently, the tests in concrete have involved admixed chloride (Schönning and Randström, 2011; Van Niejenhuis et al., 2016), potentiostatic tests (Scully and Hurley, 2007), cracked concrete (Islam et al., 2013; Van Niejenhuis et al., 2016), or combinations of these approaches. Again, the aim was not in obtaining CCT values per se, but in ranking the steels’ behavior. Table 4.1 gives the ranking by different investigators and shows, clearly that there is little consensus. However, there is general agreement that all the CCT for the stainless steels is between five and 10 time that of carbon steel. The values vary considerably because of the different concrete or mortar mixes in which they were tested. Pore solution tests of five different stainless grades (Ogunsanya and Hansson, 2020) actually indicated that the CCT values for all the rebar, obtained by the test shown in Fig. 4.4, were greater than 6% by mass of cement, which is greater than the solubility limit for chloride (as NaCl) in that particular mixture. It should be noted, however, that the pore solution compositions were obtained from pastes with admixed chlorides which, as discussed in Section 4.2.1, have a greater degree of chemical binding than chlorides penetrating into the hardened cement or concrete. Thus, a higher proportion of the chlorides penetrating the concrete from the environment would likely remain in solution.

4.6

The influence of the chloride cation

The vast majority of CCT investigations have used NaCl as the salt because it is the major component of seawater and of rock salt. Increasingly in recent years, however, brines of MgCl2 and CaCl2 as well as NaCl and mixtures of these three salts, are being used as antiicing agents, for three reasons: first, unlike the crystalline rock-salt, they adhere to the pavement and are not blown into the verge by passing traffic; second, they have lower eutectic temperatures and so are effective to lower ambient temperatures (Brady, 2009) and, third, they are less environmentally destructive than NaCl (Oswald et al., 2019). So, while they are better at maintaining ice-free conditions, they are not so kind to the reinforced concrete. CaCl2 reacts with the Ca(OH)2 in cement paste to form an expansive phase, calcium hydroxychloride, which can cause cracking (JulioBetancourt and Hooton, 2009). MgCl2 also reacts with Ca(OH)2 to form MgCO3 which is highly insoluble and precipitates, filling the pores and, thereby, limiting further ingress of the chlorides, but reducing the pH to w9, a level at which the steel is no longer passive. The MgCl2/Ca(OH)2 reaction also produces CaCl2 which can then form the expansive oxychlorides. Whereas NaCl raises the pH by reacting with the Ca(OH)2 producing more NaOH, both CaCl2 and MgCl2 lower the pH by removing the calcium hydroxide from solution (Yuan et al., 2009; Jiang et al.,

74

Table 4.1 Ranking of stainless steel rebar embedded in concrete, except where noted.

UNS/AISI designation

Schönning and Randström (2011)

Van Niejenhuis, Bandura et al. (2016)

Admixed ClL D potentiostatic control 1

Van Niejenhuis et al., 2016

Islam et al. (2013)

Admixed ClL D LPR

Cracked concrete D LPR

Macrocell

1

1

3

3 2

1 5 2 4

2

5 1 2 3 4

Bautista et al. (2006)

Ogunsanya and Hansson (2020)

Pore solution

1

3

2

1

2

3 1 4

2 3

5

3 2 1 4

Corrosion of Steel in Concrete Structures

S30403/ 304L S32653/ 316LN S32205/318 S32304/ S32101/ S24100 16Cr/0.2Ni/ 16Cr/1.5Ni/ 2Mo 316Ti 204Cu

GarciaAlonzo et al. (2007)

Chloride corrosion threshold

75

2012). It is known that the passive film on steel is more protective at higher pH values (Ghodset al., 2013), so the CCT for NaCl attack should be higher than that for the other salts as observed. However, the chloride binding increases with decreasing pH (Tritthart, 1989) so there will be more chlorides from NaCl available in solution to attack the rebar than from the other salts. The net result from these conflicting factors is very interesting, as shown in Table 4.2. The results of the test shown in Fig. 4.4 with salts with different cations indicated the CCT in terms of chloride concentration of the pore solution was significantly higher for NaCl than for CaCl2 or MgCl2 but when “translated” into chloride concentration as % by mass of cement, the values were very similar, because of the much higher proportion of bound chlorides in the latter two mixes (Ogunsanyaet al., 2020).

4.7

Conclusions

It is clear from the above that there is no unique critical chloride threshold for rebar corrosion, even for carbon steel rebar in ordinary Portland cement. Consequently, those who need to use the concept of a threshold concentration for active corrosion in order to predict the remaining practical service life of existing structures, or to design structures to meet the required specified service life, need to measure, or define, the parameters controlling the CCT. These factors are not confined to the mixture design, rebar type, and structural design, such as cover depth, but should include the climate as well as the extent and type of chloride exposure. For example, the local climate (temperature, relative humidity, amount and form of precipitation, the exposure to winds) will determine the frequency and extent of wetting and drying and, hence, the moisture content. This, in turn, will determine the composition and availability of the pore solution adjacent to the rebar. This author believes that the CCT is best defined as total chloride as a percentage of the dry cement in the concrete, in agreement with the conclusions of (Glass and Buenfeld, 1997). Defining it as the water-soluble chloride, may underestimate the content because the amount of bound chloride can change with time. Using the Cl /OH ratio, is recommended by Angst et al. (2009) because it takes into account the pH and buffer capacity of the concrete and, therefore, considers the effects of chloride binding. Nevertheless, measuring the hydroxide content of concrete is not always practical, especially for field structures.

Table 4.2 Critical chloride content of synthetic pore solutions based on the composition of expressed solution from pastes with different admixed salts (Ogunsanyaet al., 2020). Chloride content

NaCl

CaCl2

MgCl2

CCT % of pore solution CCT % cement

3.05 1.67

1.12 1.49

1.00 1.63

76

Corrosion of Steel in Concrete Structures

References ACI365.1R-17, 2017. Report on Service Life Prediction American Concrete Institute, p. 21. Alonso, C., Sanchez, M., 2009. Analysis of the variability of chloride threshold values in the literature. Materials and Corrosion 60 (8), 631e637. Angst, U., Elsener, B., Larsen, C.K., Vennesland, Ø., 2009. Critical chloride content in reinforced concrete - a review. Cement and Concrete Research 39, 1122e1138. Angst, U., Isgor, B., Hansson, C., Sag€ués, A., Geiker, M., 2022. Beyond the chloride threshold concept for predicting corrosion of steel in concrete. Applied Physics Reviews 9, 011321. https://doi.org/10.1063/5.0076320. Ann, K.Y., Song, H.-W., 2007. Chloride threshold level for corrosion of steel in concrete. Corrosion Science 49, 4113e4133. Arya, C., Buenfeld, N.R., Newman, J.B., 1990. Factors influencing chloride-binding in concrete. Cement and Concrete Research 20, 291e300. ASTM, 2007. G109-07 Standard Test Method for Determining Effects of Chemical Admixtures on Corrosion of Embedded Steel Reinforcement in Concrete Exposed to Chloride Environments West Conshocken. American Society for Testing and Materials, PA. ASTM, 2018. A955/A955M-18 Standard Specification for Deformed and Plain Stainless Steel Bars for Concrete Reinforcement [Metric]. American Society for Testing and Materials, Philadelphia, PA. Barneyback, R., Diamond, S., 1981. Expression and analysis of pore fluids from hardened cement pastes and mortars. Cement and Concrete Research 11 (2), 279. Bautista, A., Blanco, G., Valasco, F., 2006. Corrosion behaviour of low nickel austenitiv stainless steel reinforcements: a comparative study in simulated pore solutions. Cement and Concrete Research 36 (10), 1922e1930. Brady, J.B., 2009. Magma in a beaker: analog experiments with water and various salts or sugar for teaching igneous petrology. The Canadian Mineralogist 47, 457e471. Buckley, L.J., Carter, M.A., Wilson, M.A., Scantlebury, J.D., 2007. Methods of obtaining pore solution from cement pastes and mortars for chloride analysis. Cement and Concrete Research 37, 1544e1550. Byfors, K., Hansson, C.M., Tritthart, J., 1986. Pore solution expression as a method to determine the influence of mineral additives to cement on chloride binding. Cement and Concrete Research 16, 761. Callaghan, B.G., 1993. The performance of a 12% chromium steel in concrete in severe marine environments. Corrosion Science 35 (5e8), 1535e1541. Cao, H.T., Bucea, L., Sirivivatnanon, V., 1994. Influence of binder type on anodic dissolution o steel embedded in cement pastes. Cement and Concrete Research 24 (2), 203e213. Cao, Y., Gehlen, C., Angst, U., Wang, L., Wang, Z., Yao, Y., 2019. Critical chloride content in reinforced concrete - an updated review considering Chinese experience. Cement and Concrete Research 117 (1), 58e68. Chalhoub, C., François, R., Carcasses, M., 2019. Determination of chloride threshold initiating corrosion: a new set-up taking the localized aspect of corrosion into account. Cement and Concrete Research 124, 105825e105839. Darwin, D., Browning, J., O’Reilly, M., Xing, L., Ji, J., 2009. Critical chloride corrosion threshold of galvanized reinforcing bars. ACI Materials Journal 106 (3), 176e183. De Weerdt, K., Colombo, A., Coppola, L., Justnes, H., Geiker, M.R., 2015. Impact of the associated cation on chloride binding of Portland cement paste. Cement and Concrete Research 68, 196e202.

Chloride corrosion threshold

77

Enevoldsen, J.N., Hansson, C.M., Hope, B.B., 1994. Binding of chlorides in mortar containing admixed or penetrated chlorides. Cement and Concrete Research 24, 1534e1548. Garcia-Alonzo, M.C., Escudero, M.L., Miranda, J.M., Vega, M.I., Capilla, F., Correira, M.J., Salta, M., Bennani, A., Gonzalez, J.A., 2007. Corrosion behaviour of new stainless steel reinforcing bars embedded in concrete. Cement and Concrete Research 37 (10), 1463e1471. Ghods, P., Isgor, O.B., Carpenter, G.J.C., Li, J., McRae, G.A., Gu, G.P., 2013. Nano-scale study of passive films and chloride-depassivation of carbon steel rebar in simulates concrete pore solutions using FIB/TEM. Cement and Concrete Research 47, 55e68. Ghods, P., Isgor, O.B., McRae, G., Miller, T., 2009. The effect of concrete pore solution on the quality of passive oxide films on black steel reinforcement. Cement and Concrete Composites 31 (1), 2e11. Glass, G.K., Buenfeld, N.R., 1997. The presentation of the chloride threshold level for corrosion of steel in concrete. Corrosion Science 39 (5), 1001e1013. Hansson, C.M., Sørensen, B., 1990. The Threshold Concentration of Chloride in Concrete for the Initiation of Reinforcement Corrosion. Corrosion Rates of Steel in Concrete. ASTM STP, Baltimore, MD, p. 1065. Hausmann, D.A., 1964. Electrochemical behaviour of steel in concrete. Journal of the American Concrete Institute 61, 171e187. Hurley, M.F., Scully, J.R., 2006. Threshold chloride concentrations of selected corrosionresistant rebar materials compared to carbon steels. Corrosion 62 (10), 892e904. Islam, M.A., Bergsma, B.P., Hansson, C.M., 2013. The chloride-induced corrosion behaviour of stainless steel and carbon steel reinforcing bars in sound and cracked concrete. Corrosion 69 (3), 303e313. Jaffer, S.J., Hansson, C.M., 2010. Corrosion Behaviour of LDX2101. University of Waterloo. Jiang, L., Huang, G., Xu, J., Zhu, Y., Mo, L., 2012. Influence of chloride salt type on threshold level of reinforcement corrosion in simulated concrete pore solutions. Construction and Building Materials 30, 516e521. Julio-Betancourt, G.A., Hooton, R.D., 2009. Calcium and magnesium chloride attack on cement-based materials: formation, stability and effects of oxychlorides. In: 2nd International RILEM Workshop on Concrete Durability and Service Life Planning (ConcreteLife ’09). RILEM, Haifa, Israel. Li, L., Sag€ués, A.A., 2002. Chloride corrosion threshold of reinforcing steel in alkaline solutions - cyclic polarization behavior. Corrosion 58 (4), 305e376. Li, Y., Zhu, Y., Fang, J., 2014. Experimental study of chloride ion’s critical content causing reinforcement corrodio in concrete. Hydro-Science and Engineering 2014 (1), 24e28. Luo, R., Cai, Y., Wang, C., Huang, X., 2003. Study of chloride binding and diffusion in GGBS concrete. Cement and Concrete Research 33 (1), 1e7. Lysogorski, D.K., Cros, P., Hartt, W.H., 2005. Performance of Corrosion Resistant Reinforcement as Assessd by Accelerated Testing and Long Term Exposure in Chloride Contaminated Concrete. Corrosion 2005. National Association of Corrosion Engineers. Paper #15258. Maldonado, L., 2009. Chloride threshold for corrosion of galvanized reinforcement in concrete exposed in the Mexican Caribbean. Materials and Corrosion 60 (7), 536e539. Manera, M., Vennesland, Ø., Bertolini, L., 2008. Chloride threshold for rebar corrosion in concrete with addition of silica fume. Corrosion Science 50 (2), 554e560. Ogunsanya, I.G., Hansson, C.M., 2018. Detection of the critical chloride threshold of carbon steel rebar in synthetic concrete pore solutions. RILEM Technical Letters 3, 75e83.

78

Corrosion of Steel in Concrete Structures

Ogunsanya, I.G., Hansson, C.M., 2020. The critical chloride threshold of austenitic and duplex stainless steel reinforcing bars of stainless steel rebar. Metallurgical and Materials Transactions A 51, 4685e4694. Ogunsanya, I.G., Kristufek, L., L’e Meur, P., Hansson, C.M., 2020. Influence of Chloride Cations on Pore Solution Chloride and Critical Chloride Threshold of Carbon Steel Rebar. ACI Materials Journal Submitted for Publication. Oh, B.H., Jang, S.Y., Shin, Y.S., 2003. Experimental investigation of the threshold chloride concentration for corrosion initiation in reinforced concrete structures. Magazine of Concrete Research 55, 117e124. Oswald, C.J., Giberson, G., Nicholls, E., Wellen, C., O’ni, S.R., M., 2019. Spatial distribution and extent of urban land cover control watershed-scal chloride retention. The Science of the Total Environment 652, 278e288. Pillai, R.G., Trejo, D., 2005. Surface condition effects of critical chloride threshold of steel reinforcement. ACI Materials Journal 102 (2), 103e109. Poursaee, A., Hansson, C.M., 2007. Reinforcing steel passivation in mortar and pore solution. Cement and Concrete Rsearch 37 (7), 1127e1133. Randström, S., Almén, M., Petersson, R., Adair, M., 2010. Reproducibility of Critical Chloride Threshold Levels for Stainless Steel Reinforcement. Structural Faults and Repair, Edinburgh, UK. Richartz, W., 1969. Die bindung von Chlorid bei der Zementerh€artung. Zement-Kalk-Gips 22, 447e456. Schönning, M., Randström, S., 2011. Adaptation of EN 480-14:2006 as a Test Ethod for Determining a Critical Chloride Threshold Level for Stainless Steel Rebar. Eurocorr 2011. Scott, A., Alexander, M.G., 2016. Effect of supplementary cementitious materials (binder type) on th epore solution chemistry and the corrosion of stel in alkaline environments. Cement and Concrete Research 89, 45e55. Scott, A.N., Thomas, M.D.A., 2013. Chloride resistance of 9% Cr steel in a simulated pore solution. Corrosion 69 (11), 1073e1080. Scully, J.R., Hurley, M.F., 2007. Investigation of the Corrosion Propagation Characteristics of New Metallic Reinforcing Bars. Virginia Transportation Research Council, p. 58. Tan, Z.Q., Hansson, C.M., 2008. Effect of surface condition on the initial corrosion of galvanized steel embedded in concrete. Corrosion Science 50, 2512e2522. Thomas, M.D.A., 1996. Chloride thresholds in marine concrete. Cement and Concrete Research 26 (4), 513e519. Trejo, D., Vaddey, N.P., Halmen, C., 2021. Standardizing test to quantify chloride threshold of steel in concrete. ACI Materials Journal 118 (1). Tritthart, J., 1989. Chloride binding in cement, part I. Cement and Concrete Research 19, 683e691. Tritthart, J., 1990. Pore solution composition and other factors influencing the corrosion risk of reinforcement in concrete. In: Third International Symposium on Corrosion of Reinforcement in Concrete Construction, Belfry Hotel. Elsevier Applied Science (for Society of Chemical Industry), Wishaw, Warwickshire, United Kingdom. Tuutti, K., 1982a. Corrosion of Steel in Concrete. Cement and Concrete Research. Tuutti, K., 1982b. Corrosion of Steel in Concrete PhD. Swedish Cement and Concrete Research Institute. Van Niejenhuis, C.B., Bandura, T.W., Hansson, C.M., 2016. Evaluation of th eproposed European test procedure for ranking stainless steel rebar. Corrosion 72 (6), 834e842.

Chloride corrosion threshold

79

Van Niejenhuis, C.B., Ogunsanya, I.G., Hansson, C.M., 2020. Analysis of the pore solution of different cement pastes with and without admixed chlorides. ACI Materials Journal 117 (3), 21e28. Van Niejenhuis, C.B., Walbridge, S., Hansson, C.M., 2016. The performance of austenitic and duplex stainless steels in cracked concrete exposed to concentrated chloride brine. Journal of Materials Science 50 (1), 362e374. Vollpracht, A., Lothenbach, B., Snellings, R., Haufe, J., 2016. The pore solution of blended cements: a review. Materials and Structures 49 (8), 3341e3367. Yang, L., Chiang, K.-T., Pabalan, R.T., Dasgupta, B., Ibarra, L., 2014. Threshold chloride levels for localized carbon steel corrosion in simulated concrete pore solutions using coupled multielectrode array sensors. Corrosion 79 (8), 850e857. Yuan, Q., Shi, C., De Schutter, G., Audenaert, K., Deng, D., 2009. Chloride binding of cementbased materials subjected to external chloride environment e a review. Construction and Building Materials 23 (1), 1e13.

This page intentionally left blank

Corrosion of prestress and posttension reinforced concrete bridges

5

Kingsley Lau 1 , Samanbar Permeh 1 and Ivan Lasa 2 1 Department of Civil and Environmental Engineering, Florida International University, Miami, FL, United States; 2Lasa and Associates Corrosion Services, Gainesville, FL, United States

5.1

Introduction

Implementation of prestressed concrete elements in highway bridges is a continued engineering practice from the late twentieth century and into the present. The application of prestress techniques to concrete has allowed for greater structural design considerations to be made by bridge engineers (Naaman, 2004; Hewson, 2003). Light and slender prestressed structural elements provide advantages for long bridge spans. The integration of precast concrete elements in the prestress techniques provides good quality control during element casting as well as erection. Furthermore, prestressed concrete provides nice aesthetics for public use and appreciation of highway bridges. Several decades of experience with the techniques with continued improvements and innovations in its application has further helped garner acceptance of prestressed concrete by bridge engineers and owners. In contrast to the advantages and general good use of prestressed concrete for over 60 years, documented durability problems and component failures in arguably isolated cases continue to illustrate difficulties in bridge applications. Unfortunately, corrosion concerns of the steel components of prestress concrete bridges have persisted. Prestressed bridge systems are complicated and many structural, construction, material, and environmental factors can be involved in corrosion development. The use of prestressing techniques for concrete bridges is diverse and has been used for bridge superstructure and substructure elements. Examples of prestressed concrete elements are shown in Fig. 5.1. Extensive information on the current practice and structural applications of prestressed concrete for highway bridges can be found in guidelines, handbooks, and other publications from state transportation departments, the Federal Highway Administration (FHWA), trade organizations promoting the construction, and from materials providers (Corven and Morenton, 2013; PCI; PTI M55.1-19; ACI-318R; ACI-222.2). For brevity and in context to case studies of corrosion development, prestress applications can be broadly categorized by two prestressing methods: pretensioning and posttensioning. Pretensioning of high strength steel wires

Corrosion of Steel in Concrete Structures. https://doi.org/10.1016/B978-0-12-821840-2.00013-4 Copyright © 2023 Elsevier Ltd. All rights reserved.

82

Corrosion of Steel in Concrete Structures

(a)

(b)

Figure 5.1 Examples of prestressed concrete elements. Left: (A) Sunshine skyway bridge, FL (prestressed segmental box girders and columns). Right: (B) St. George Island bridge, FL (prestressed concrete cylinder piles and girders).

prior to encapsulation in concrete and release after concrete hardening is commonly associated with prefabricated prestressed beams and prestressed piles for bridge application. Posttensioning of high strength steel, as its namesake implies, occurs after the concrete element (often precast) has been fabricated. The concrete element typically contains embedded duct conduits where the prestressing steel is subsequently placed and stressed. The assembly of the prestressed steel and other relevant hardware components is called a tendon. Posttensioned tendons are differentiated as being bonded or unbonded where in the former, the prestress steel is permanently set in an injected cementitious grout to prevent movement relative to the concrete. Prestressing forces can be developed along the length of the steel in bonded tendon systems to transfer the forces onto the surrounding concrete. Posttensioned tendons are further identified as being an internal or external tendon where the former is cast within the concrete of the structural element and the latter have free spans external to the structural concrete. Posttensioned tendons have been commonly associated with bridge superstructure box girders but other bridge components such as columns and caps have been designed as well.

5.2

Materials

The variety of materials and hardware components used in prestressed concrete is wide as is the many applications for bridge components. Specification changes, technology innovations, and economic conditions continue to allow for the introduction of new components and materials. Constructability and bridge durability concerns, too, provide opportunities for new materials and designs. For brevity, the reader is referred to the literature available from FHWA, trade organizations, and materials providers for a detailed description of important hardware components. This section highlights select materials and components that are relevant to later discussion on corrosion development in select prestressed concrete systems.

Corrosion of prestress and posttension reinforced concrete bridges

5.2.1

83

Prestressing steel

Effective prestressing in practice requires steel with high strength and high elongation. Other important steel characteristics include elastic behavior and ductility at stress levels at loading, low relaxation of tension, resistance to stress corrosion cracking, and resistance to hydrogen embrittlement. Wire and strand are typically high carbon steel with the application of cold-work and thermal treatment to attain desired material characteristics. Standard specifications for prestress wire, strand, and bar can found in ASTM A421, A416, and A722. An example of corrosion of a wire from a posttensioned strand compared to rebar corrosion is shown in Fig. 5.2. Vertical and horizontal pitting corrosion on the wire is seen in Fig. 5.2A and corrosion progression into the microstructure of the rebar steel is shown in Fig. 5.2B. Prestressed concrete, similar to conventional reinforced concrete, can be susceptible to chloride and carbonation induced corrosion in aggressive exposure environments. Like conventional reinforced concrete, alternative materials to the prestressing steel have been presented. Epoxy coated strand is specified in ASTM A882. Work on application of austenitic and duplex stainless steel (Mullins et al., 2014; Moser), carbon fiber composite cables (ACI 440.4R; Roddenberry et al., 2014) (CFCC), and galvanized strands have been recently reported. Stainless steels are expected to provide greater resistance to chloride-induced corrosion but issues on stress corrosion cracking, mechanical properties, and costs require further deliberation. Mechanical performance, detailing, and costs are issues that are to be resolved for CFCC wire and strand but extensive work is being done on this material.

5.2.2

Prestressing components and materials

Innovations in tendon detailing and design including strand configuration to address tendon corrosion durability have also been presented. Electrically isolated tendons among others have been presented as useful designs for health monitoring of the system. More generally, appropriate material selection and design of the other prestress system components are important to mitigate corrosion development. Examples of bonded posttensioned tendon components are shown in Fig. 5.3. The tendon duct

(a)

(b)

Figure 5.2 Micrographs of corrosion. Left: (A) Wire from post-tensioned strand (unetched). Right: (B) Rebar (etched with 2% nital).

84

Corrosion of Steel in Concrete Structures

(a)

(b)

(c)

(d)

(e)

(f)

Figure 5.3 Examples of bonded post-tensioning components and materials with corrosion related degradation and deficiency. (A) Damaged external tendon. (B) Corroded steel duct. (C) Corroded strand in tendon with sand contamination. (D) Segregated grout in anchor cap. (E) Corroded strand, wedges, and wedge plate. (F) Surface corrosion on anchor in pourback.

for posttensioned concrete systems not only provides the form for the continuous longitudinal space for the prestressing steel but it serves as an outer barrier to environmental contaminants including moisture and salt. The duct can be made out of galvanized steel, plastic, or high-density polyethylene pipe (HDPE) depending on its application. Corrugated galvanized steel and plastic (polyethylene or polypropylene) pipe are often used for internal tendons. Galvanized pipe can be used for tendon portions through deviators and diaphragms. HDPE pipe is common for external tendons. Tendon systems contain valve inlets and outlets to allow for proper grout injection. Those same locations should contain appropriate plugs during and after construction to minimize exposure to the external environment including moisture, salt,

Corrosion of prestress and posttension reinforced concrete bridges

85

and atmospheric carbon dioxide. Posttensioned concrete steel anchors also require appropriate design and material selection consideration to account for possible tendon deficiencies especially for bonded tendons where proper grouting is vital to mitigate corrosion development. The anchorage system in general typically consists of the trumpet, bearing plate, wedge plate, individual strand wedges, and an outer permanent protective cap. Materials can vary depending on the design and make for the requirements of the technology as well as conforming to specifications; different metallic components in the prestress anchor may introduce aggravating galvanic interactions. The anchorage trumpet can be of ductile iron, galvanized steel, or plastic. Appropriate design details and materials selection of the prestress concrete elements, the joints, and pour-backs also are important to avoid premature corrosion. For example, direct moisture and chloride penetration may occur at the joints between concrete segments. Alignment of the tendons at girder end sections should be ensured to allow proper coupling and sealing at pourback sections. Also, as indicated by Sprinkel and Balakumaran (2017), detailing of the pourback should provide spacing to allow adequate concrete pour and consolidation. To minimize the potential for corrosion of the internal tendons in segmental box bridge construction, epoxy-resin joints have been generally used to good effect (Trejo et al., 2009a,b) compared to dry joint construction (although a record of flawed sealing has also been documented in the former). Similarly, shrinkage cracks and other degradation of cementitious materials can occur at the concrete pour-back, possibly allowing intrusion of moisture and salts. Epoxy grout encapsulating material and concrete seal coatings have been used in practice over pour-back areas to act as a barrier coating. The use of plastic ducts at joints and permanent grout caps further prevents exposure to external contaminants. For posttensioned tendons, materials for duct couplings and fittings should be compatible with the duct material and duct filler, be resilient to degradation and mechanical damage, provide unobstructed internal space, and provide protection for leaks and exposure to external contaminants. Various duct couplers and connection techniques are prescribed based on design and use. Use of heat fusion welding techniques for plastic pipes, polymer boots (ethylene propylene diene monomer), shrink sleeves, and duct couplers are used in practice. Poor construction practices of using adhesive tape and cloth should be avoided.

5.2.3

Concrete, grouts and filler material

The concrete for the prestressed structural elements especially for locations that embed or cover prestressing steel provides some level of protection from external environmental exposure. For external tendons, the concrete sections can minimize direct exposure of the tendons from salt spray and deicing salts in applications where the tendons are located inside the structure (segmental bridges). For pretensioned wire and posttensioned internal tendons, the concrete clear cover provides a barrier for aggressive chemicals such as chloride ions and atmospheric carbon dioxide to diffuse in the same manner as that for conventionally reinforced concrete elements. High performance concretes with low permeability provide good corrosion durability of the embedded steel elements.

86

Corrosion of Steel in Concrete Structures

The cementitious grout used in bonded posttensioned tendons has received significant attention in the past 25 years due to its association with severe corrosion in a number of bridges in Florida and elsewhere despite its intended function to provide corrosion protection of the prestressing steel. The grout in current practice is essentially Portland cement and can include supplemental cementitious materials (such as slag, silica fume, and fly ash), admixtures (for set-control, air-entraining, antibleeding, and corrosion inhibition), and fine-sized aggregates. In earlier practice, expansioncausing admixtures were used but have subsequently been shelved after revisions to grout specifications, particularly due to durability concerns. Prior to specification changes in the early 2000s, posttension grouts were neat grouts and were documented to have challenges due to the excess formation of bleed water and resulting in grout void formation and severe corrosion of tendons in a number of bridges. In response to that corrosion development in the late 1990s, posttensioning systems in recent practice have used prepackaged thixotropic grouts with low-bleed water formationd unfortunately, recent developments since 2011 have shown that challenges in corrosion mitigation persist for these grouts as well. In specifications, the allowable grouts are grouped into four categories depending on tendon exposure conditions, thixotropic properties of the grout, and prepackaging. Construction of bonded posttensioned tendons in aggressive environments in the past 20 years has typically used prepackaged thixotropic grouts to incorporate qualified grout products with expected good quality control and low-bleeding characteristics. Low bleed characteristics are typically defined by bleed tests such as the wick-induced bleed test and Schupak pressure bleed test. Other tests prescribed in Europe include the inclined-tube tests (VSL, 2002). As alluded to earlier, qualified grouts used in construction and conforming to the revised grout specifications had a few documented incidences of severe deficiency (nominally called soft grout) and in some cases directly associated with fast and severe corrosion of the prestressing steel in the absence of chloride contamination and grout pore water carbonation (Lau et al., 2013a,b,c). The deficient grout was characterized as having poor cohesive bulk properties, high moisture content, and elevated sulfate ion concentrations in the pore water. Another recent development of quality control issues not directly related to soft grout development of some prepackaged grout used in construction included the presence of enhanced chloride presence in the cement component (U.S. Department of Transportation, 2011). Many questions concerning these cases were posed and concerns about grout material quality control (grout contamination), product packaging quality control (underweight bags), grout robustness (segregation), construction practices (excess moisture), and corrosion development (corrosion initiation in chloride-free alkaline environments) were being addressed. Other grout material parameters that have received recent attention included chloride threshold concentrations in prepackaged grouts and segregation testing (Lee and Zielske, 2014). Some construction-related issues related to grout quality that has received recent attention included mandate for rigorous personnel qualifications, grout shelf-life expiration, and use of field mockup tests. The Virginia Department of Transportation (VDOT) as part of the construction of a recent posttensioned concrete bridge required full-scaled test mockups to qualify grout material.

Corrosion of prestress and posttension reinforced concrete bridges

87

Filler materials such as waxes and greases have been used for the protection of cables for utilities and in the application for posttensioned concrete elements in nuclear facilities. Adoption of flexible fillers for posttensioned concrete bridges has been made in Europe due to corrosion associated with bleeding and void formation in cementitious grouts. In current practice in Germany and France, filler wax material was thought to have good stability at high temperatures (Parsons, 2012). The wax filler heated above its congealing point is injected into the tendon with the intent to completely encapsulate the prestressing steel. Acceptable wax materials (bitumastic petroleum-based waxes) and oil-based greases specified in the European Organization for Technical Approval ETAG013 (ETAG 013) have minimum material required criteria for material stability, separation, penetration, aggressive ion content, and corrosion protection. The wax and grease material is intended to provide corrosion protection by encapsulating the prestressing steel and acting as a physical barrier. European standards require minimum corrosion resistance in spray tests including salt spray. Unlike cementitious grout (where the normally high pH pore water solution provides conditions for steel passivation), the grease and wax material itself is not necessarily intended to provide corrosion mitigation but rather provide a barrier against the elements. The European standards have stipulations to specify the corrosivity of the wax by copper-strip corrosion tests at 100
C and of the grease by Emcor tests. In the US, Florida has recently begun taking steps for the implementation of flexible fillers for some posttensioned bridge applications to provide alternative means of corrosion protection of the prestressing steel (Robertson, 2015). Concerns with issues in structural design and detailing of nonbonded tendons in bridges, construction and injection procedures with flexible fillers, and possible durability concerns for highway bridge applications are still to be resolved before wide adoption of the technique and its materials.

5.3

Overview of corrosion mechanisms

The corrosion mechanisms of prestressing steel in highway bridges can vary by many complicating factors involving the structural component, prestressing technique, construction practices, deficiencies in the building material, and the natural and service environments. As such, only select topics that have been of recent interest in bridge durability are presented here. Corrosion of prestressing steel encapsulated in cementitious grout and concrete as in pretensioned concrete and bonded posttensioned concrete is in general respects similar to that of conventional reinforced concrete; however, the structural implications of even minor cross section loss of prestressing wire may be more consequential than that for reinforced concrete. Other factors such as the use of high strength steel, multiwire strand, the presence of a prestress force, the presence of segment joints, use of cementitious grout, and the presence of other building materials further complicate the system. Possible hydrogen embrittlement and stress-corrosion cracking with the use of high strength steels and the presence of the prestress force can be a concern in certain environments.

88

5.3.1

Corrosion of Steel in Concrete Structures

Pretensioned steel concrete

Prestressing steel in pretensioned concrete bridge girders and piles are exposed to the environment similar to that of conventional reinforced concrete. Bridge structural elements can be exposed to deicing salts, marine splash, or exposed to soils. The prestressing steels in these systems are not surrounded by ducts as in posttensioned concrete components and may be susceptible to corrosion by exposure to aggressive chemical species through the concrete cover. Transport of the aggressive chemical species such as chloride ions can occur through the concrete cover at a rate depending on the quality of the concrete and environmental conditions. Like conventional reinforced concrete, chloride or carbonation induced corrosion can occur. A detailed description of chloride- and carbonation-induced corrosion can be found in Chapter 2 and elsewhere (Bentur et al., 1997; Bertolini et al., 2004). Various changes in the cement pore water can allow for depassivation of the steel to occur but corrosion has commonly been attributed to the elevated presence of chloride ions and pore water carbonation. Maximum chloride limits for prestressed concrete prescribed by ACI 318 and ACI 222 are 0.06% by weight of cement (water-soluble chloride ion) and 0.08% by weight of cement (acid-soluble chloride ion) (ACI318R; ACI-222.2). Carbonation of the cementitious material at the prestressing steel interface causing a drop in the pore water pH below 12 will allow for depassivation of the steel. Larger concrete cover and the use of less permeable concrete should provide long bridge service times. Enhanced transport of moisture, chlorides, and atmospheric carbon dioxide can occur at localized cracks that may be present in the prestressed concrete components depending on its exposure to moisture, salty humid air, water, and soil. Cracks in prestressed concrete components in bridge substructures such as piles (Fig. 5.4) can occur during installation. Work by Lau and Sag€ués (2009) indicated that isolated cracking in good quality conventionally reinforced concrete should create topical corrosion damage but adverse crack conditions could cause increased damage.

Figure 5.4 Longitudinal crack in prestressed concrete cylinder pile.

Corrosion of prestress and posttension reinforced concrete bridges

5.3.2

89

Posttensioned steel concrete

Corrosion damage to the prestressing steel of posttensioned concrete components can occur by different mechanisms depending on the prestressing configuration and environmental exposure. The use of a visual guide for acceptance of the corroded strand was proposed by Sason (1992) particularly for corrosion that formed prior to construction. The steel strand may be exposed to moisture, sea spray, and other deleterious environmental contamination during transport and storage. A significant presence of corrosion pitting can reduce the structural capacity of the strand and may result in failure during stressing or in service (Trejo et al., 2009a,b; Sason, 1992). After construction, unbonded posttensioned tendons rely on fillers such as grease and wax to provide encapsulation of the steel, providing a barrier to environmental exposure. The authors were not aware at the time of writing of any cases of strand corrosion in the contemporary filler materials used in Europe for unbonded bridge tendons. Corrosion of steel strand prior to wax injection is expected to be similar to that of bonded tendons prior to group pumping. During the construction sequence of bonded posttensioned tendons, the strand is installed through the ducts of aligned concrete segments and then stressed prior to grouting. The time until grouting may be delayed due to complications in the construction. The time limit for grouting in humid and marine environments is 7 days (PTI M55.1-19) (20e40 days in dry environments) but may be extended with the use of additional means for corrosion protection upon the approval by the construction engineer. Runoff water may become entrapped in the ducts if they are not properly sealed during the construction delay. Corrosion activity in near neutral pH or even slightly elevated pH can be further exacerbated with the presence of chloride ions (and other aggressive ions) and possible carbonation in the latter case. Internal ducts were left ungrouted for up to 15 months during the construction of the San FranciscoeOakland Bay Bridge in 2006 and runoff water from the deck entered the duct through unsealed grout tubes (Reis, 2006, 2007a,b). Strand with moderate corrosion with an indication of shallow pitting was observed and some failed to meet specified tensile strength and ductility requirements. Transverse cracks and fractures in some wires in strand touching a duct hard-point (metal duct defect caused during construction) may have been associated with environmentally assisted cracking. Recent research in 2011 and 2014 (Sag€ ués et al., 2011; Fernandez et al., 2014) indicated that no significant corrosion of steel strand in sheltered marine exposure was observed and only minor pitting in closed systems containing some water occurred in stressed tendon mockups exposed for 4 weeks and unstressed tendons exposed for up to 9 months. Superficial deposition of salt resulted in early enhanced corrosion in the closed system containing water (exposed for 4 weeks). In all those cases, no loss of tensile strength was measured. Greater hydrogen buildup was measured in the closed system containing water than in the open exposed and dry closed systems. Due to a variety of material and construction-related issues, portions of bonded posttensioned tendons may be inadvertently left ungrouted, therefore creating large void spaces that can expose the steel strand for long periods of time. Also, movement

90

Corrosion of Steel in Concrete Structures

of air and water during grouting can leave a bleed trail resulting in similar improper encapsulation of the strand in the cementitious material. Barring other chemical deficiencies of the grout and other interaction with free moisture availability (discussed later), the void space itself may or may not create a region where the steel can experience severe corrosion but are still considered a deficiency. With the simplifying assumption that no other chemical interactions are important, possible corrosion of the exposed steel in the void space can be generally viewed as being exposed to atmospheric conditions. In that scenario, the major factors involved include the time of wetness, relative humidity, temperature, and presence of contaminants such as salts (Roberge, 2011; Revie and Uhlig, 2008; Jones, 1996). Sufficient moisture film on the steel surface is needed to provide the electrolyte for the electrochemical reactions. The stability of that moisture film depends on the humidity level, temperature, and level of pollution (i.e., sulfur dioxide) in the duct. Significant corrosion occurs above a critical relative humidity level (reported to be >75%(Matsushima, 2011)) that in turn is dependent on the extent of surface contaminants on the metal surface. Hygroscopic salts such as chlorides can promote water condensation from the air at lower relative humidity than in noncontaminated conditions and thus increase the time of wetness and corrosion development. Also, surface contamination (i.e., chlorides and sulfates) can result in greater corrosion development due to the formation of nonprotective corrosion product films. In practice, inspection of voided tendons with dry internal conditions often reveal minor to no corrosion development. In other documented cases, voids have been associated with severe corrosion but other tendon or material deficiencies were often also related (Powers, 1999; Corven Engineering, 2001; Hartt and Venugopalan, 2002). Corrosion of strand in the void spaces created by the formation of grout bleed water has been widely discussed in posttensioned tendon failures in the past 25 years. Prior to the use of low-bleed prepackaged thixotropic grouts, plain cement grouts with expansive agents were commonly used. Case studies revealed the tendency of those materials to form bleed water thus leading to subsequent changes in grout material specifications and testing methodologies for bleed water formation. Capillary action through the narrow interstitial spaces of the strand and vertical deviation of draped tendons allowed for the accumulation of the bleed water at high points such as at the anchorage assemblies. The movement of air and bleed water followed by grout subsidence/segregation allow for voids to expose the strand in a mixed environment of grout, bleed water, and air. Severe corrosion of strand exposed to the bleed water may still occur even though the pH of the bleed water itself can be alkaline (pH > 12). It has been proposed that corrosion can occur due to the carbonation of the bleed water and grout by the presence of CO2 or intrusion of CO2 at the grout-to-air interface where pH can drop to 8 (Hartt and Venugopalan, 2002). At these pH levels, active corrosion of the steel at the carbonated grout can develop. After the formation of the void, the level of bleed water can diminish either by reentering in the grout pore space or by drying and evaporative processes. The corrosion rate may diminish in dry conditions; however, corrosion of the strand in the void space and in the grout-to-air interface may continue if moisture becomes available in periods when the relative

Corrosion of prestress and posttension reinforced concrete bridges

91

humidity is high and condensation of moisture occurs (Hartt and Venugopalan, 2002) or generally if external moisture seeps through improperly sealed or defective posttensioning components (Wang et al., 2005). Furthermore, the pore water residuals and corrosion products may be hygroscopic allowing moisture condensation at lower relative humidity levels. The large movement of water to accumulate bleed water would likely aggregate chloride and other aggressive ionic chemical species. Indeed even the vestigial amount of chlorides in the raw grout materials meeting specification limits may allow for sufficient aggregation of chloride ions at the defect location (Wang et al., 2005). The enhanced chloride content in the grout at the void location in conjunction with carbonation of the grout would create situations where the [Cl]/ [OH] threshold is exceeded and steel passivity breakdown can occur. The level of corrosion is most severe at the grout-to-air interface not only due to the active corrosion caused by grout carbonation and chloride enhancement but also due to aggravating macrocell development between the anodic regions at the grout-to-air interface and the rest of the passive strand in otherwise normal grout. Furthermore, adverse macrocell development can occur between the strand and the rest of the anchorage assembly (Wang et al., 2005). Since 2011, isolated cases of severe corrosion of steel strand in deficient grout not directly associated with voids, bleed water, carbonation, and chlorides were reported in Europe and in Florida and Virginia (Sprinkel and Balakumaran, 2017; Lau et al., 2013a,b,c; Bertolini and Carsana, 2011; Carsana and Bertolini, 2015). In those accounts, the grout was described as having poor cohesive bulk properties, high moisture content, high alkali content, and high sulfate ion concentrations. In the European case, failure of a bridge prestressing cable occurred in less than 2 years of service. In Florida, two posttensioned bridges with less than 8 years in service had severe steel corrosion in grouts with these characteristics but unexpectedly occurred in the newer low-bleed prepackaged thixotropic grouts. In two Virginia bridges built in 2006 and 2007, soft grout developed from a prepackaged grout product was associated with corrosion of steel in continuous posttensioned tee beams after w8 years (Sprinkel and Balakumaran, 2017). The grout anomalies from the prepackaged thixotropic grouts were also documented in Texas (Merrill, 2010). Research was conducted to identify the cause of the grout deficiency in the prepackaged grout products (Hamilton, 2014). Early findings attributed the failure in part to excess water and materials intolerant to excess water. The lack of supporting evidence of significant chloride presence and carbonation and inconsistent association of severe corrosion to void spaces would indicate different corrosion mechanisms than those conventionally associated with corrosion of prestressed steel (Bertolini and Carsana, 2011; Blactot et al., 2007). Bertolini and Carsana (2011) proposed that low oxidizing conditions due to limited oxygen supply in crevice regions and highly alkaline pore water conditions with pH greater than 14 would allow for stability of the HFeO-2 ion and allow for active corrosion conditions. Coincidently, high sulfate ions were consistently associated with segregated grout (Lau et al., 2013a,b,c; Carsana and Bertolini, 2015; Merrill, 2010). Hydrolysis reactions involving sulfate ions such as Fe2þ þ 2H2O þ 2SO2 4 / Fe (OH)2 þ 2H2SO4 in crevice environments may allow for localized corrosion (Jones,

92

Corrosion of Steel in Concrete Structures

1996). Information from the technical literature also indicated that there could be conditions where enhanced sulfate ion concentrations may lead to corrosion development but a uniform description of the corrosion activation processes of steel in relevant grout environments is lacking. Gouda (1970) showed that breakdown of steel passivity in saturated calcium hydroxide solution (pHw12.1) can occur at sodium sulfate concentrations greater than 0.2%. Acha et al. (1990) also indicated the role of potassium sulfate in inhibiting the passive layer formation of steel in saturated calcium hydroxide solution. Al-Tayyib et al. (1988) also indicated increased corrosion current of steel in calcium hydroxide solution with sodium sulfate presence. Gouda and Halaka (1970) furthermore investigated the role of sulfates (up to 8%) in the corrosion of steel in concrete. Reported test results showed no breakdown of steel passivity with the presence of sulfates and the authors posed that alumina-containing cement phases react with the sulfate to form insoluble compounds. Al-Tayyib and Khan (1991) reported only a modest increase in the corrosion rate of steel in concrete with up to 1.8 kg/ m3 sulfate by weight of concrete. Research on the role of sulfates in the corrosion of steel in deficient grout was conducted by (Permeh et al., 2016, 2018, 2021a,b; Krishna Vigneshwaran et al., 2018). Results of electrochemical testing of steel in alkaline sodium sulfate solution indicated that at pH 12.6, the early presence of sulfate ions (>4000 ppm sodium sulfate) in solution could be aggressive by impairing the initial passive film development. Testing of the grout material showed that mixing of grout with extra mix water (15% excess of recommended levels) promoted the development of grout deficiency and that accumulation of sulfate ions in the deficient grout can form without external sulfate sources. This increases the concerns of water being present in ducts when grouting. A threshold concentration of 0.7 mgsulfate/ggrout was proposed for free sulfate ions in the grout (Permeh et al., 2019). Concerns of corrosion development of steel strand after repair of void spaces with dissimilar grout materials may be relevant particularly for tendon systems containing deficient grout materials. Corrosion failure of an external tendon in the Varina-Enon Bridge in Virginia in 2007 has often been cited as the tendon failure occurred less than 4 years after the repair of a void space (Hansen, 2007). In the absence of grout degradation and other tendon deficiencies, complete repair of void space with dissimilar materials can be expected not to produce any significant corrosion development if both materials yield high pH environments where the steel should maintain its passive behavior. Moderate levels of polarization in the anodic region due to coupling of the steel in the dissimilar grouts should not cause significantly enhanced corrosion rates if the repair is made to prevent moisture accumulation and external exposure. Voids created by bleed water and grout segregation could lead to localized grout deficiencies such as carbonation and enrichment of aggressive chemical species where steel depassivation can occur, and the condition could be aggravated by macrocell development. Laboratory testing of repaired void spaces filled with repair grout in field extracted external tendons containing some form of grout segregation indicated corrosion activity after placement of the repair material (Lau et al., 2013a,b,c; O’Reilly et al., 2012). Removal of deficient grout prior to repair remains a challenge.

Corrosion of prestress and posttension reinforced concrete bridges

5.4

93

Overview of corrosion failures and recent corrosion problems

The corrosion failures of early posttensioned concrete bridges in the United Kingdom in 1967 and 1985 have often been referenced in written work regarding the corrosion durability of posttensioned concrete bridge systems (Hewson, 2003). The catastrophic failure of the Ynys-y-Gwas Bridge in 1985 occurred without warning due to the ineffective corrosion protection of the prestressing steel at mortar-filled segment joints from moisture and deicing salts. In the United Kingdom, a moratorium on internal posttensioned tendons was placed in 1992 over concern on the corrosion durability of the prestressing steel located at joints (Hewson, 2003). Lessons learned from this bridge failure included the need for multilevel strand protection and detailing to provide corrosion durability. In the United States, corrosion of external grouted posttensioned tendons occurred in several bridges in Florida (Niles Channel, Mid-Bay, Seven-Mile, and Sunshine Skyway) (Powers, 1999; Corven Engineering, 2001; Hartt and Venugopalan, 2002) and Virginia (Varina-Enon) (Hansen, 2007) in the late 1990s to the mid-2000s. The design of these bridges incorporated materials thought to provide corrosion durability such as the use of HDPE duct and sealing of anchors with coating materials. Although specific details vary from case to case, corrosion of the prestressing steel in these cases was largely attributed to the void formation in the tendon due to bleed water formation in the neat grout commonly used at the time. The void spaces generally formed at high point tendon anchors. In some of the cases, recharge of moisture occurred from inadequately sealed anchor heads. Cases of cracked HDPE ducts were observed in MidBay and Sunshine Skyway but were not directly attributed to corrosion initiation (Hartt and Venugopalan, 2002). As recently as 2018, corrosion failures of external posttensioned tendons constructed in the I-526 Wando River Bridge in South Carolina caused for emergency closure after 29 years of service (Adcox, 2018). Following these events in the late 1990s, strategies to improve corrosion durability of posttensioned systems were implemented. These included the use of prebagged thixotropic grout materials to improve quality control to prevent bleed-water formation, use of enhanced posttensioning system components, detailing of multiple levels of corrosion protection for the duct and anchors, and sealing of joints and anchors (Corven Engineering, 2002). The Ringling Causeway Bridge is a posttensioned segmental bridge with internal and external tendons spanning across an intracoastal waterway in Florida. In 2011, corrosion failures of two longitudinal external posttensioned tendons occurred after only approximately 8 years of service (Lau et al., 2013a,b,c). Further inspection showed that over 10% of the external posttensioned tendons (17 tendons) had to be replaced due to corrosion. Not only were the number and severity of corroded tendons alarming but also the fact that corrosion occurred in tendons containing a prepackaged thixotropic grout product conforming to updated specifications introduced to minimize corrosion. The severe corrosion of the strand was not consistently associated with grout voids as described for earlier cases but occurred at locations with deficient grout. In the most severe manifestation of the deficient grout, the grout remained unhardened

94

Corrosion of Steel in Concrete Structures

Figure 5.5 Corrosion of strand embedded in deficient grout. Pink color shows the pH indicator (phenolphthalein) sprayed on grout surface.

(Fig. 5.5) with a putty-like consistency and becomes friable upon drying. In nearby regions, the deficient grout had a white color and had a chalky consistency. The deficient grout material contained as much as 80% moisture. Pore-water pH, determined by an ex situ leaching method, was generally greater than 12 and the pH in the soft grout with nearby corrosion was sometimes greater than 13. No consistent indication of carbonation being the primary corrosion mechanism was observed. The total chloride content was low to moderate but may be elevated in the deficient grout due to capillary transport during the grout segregation process. There may be some instances where the chloride and pH conditions may exceed critical [Cl]/[OH] concentrations, but in general, these conditions may not have always been consistent in locations with severe corrosion. Severe corrosion and deficient grout were associated with high free sulfate concentrations (as high as w10,000 ppm). As seen in Fig. 5.6, a variation of grout consistency and chemical makeup was observed in stratified grout layers at a low point anchor cap. Corrosion of the strand and wedge plate was predominant in regions of severe grout deficiency that had enhanced moisture and sulfate ion presence. As of 2014, five other bridges in Florida built with low-bleed water formation grouts had been identified to have analogous deficient grout, but the amount and

100 Ca 2+ K+ Na + Cl NO3 2SO4 2Moisture Content

9000

7000 6000

80 70 60

20

1000

10

0

V

30

2000

NEW

3000

III

40

IV

50

4000

I

5000

II

Concentration (ppm)

8000

I

90 II Moisture Content %

10000

III

IV

V

0

Figure 5.6 Visual and chemical makeup of stratified deficient grout at an anchor cap.

Corrosion

Corrosion of prestress and posttension reinforced concrete bridges

95

severity varied significantly between bridges (Lau et al., 2014). One of those five bridges had significant corrosion of the metal duct from internal tendons where the thickness of the duct was fully penetrated due to corrosion (Fig. 5.7). Fourteen tendons had deficient grout. Similar to the deficient grout found in the Ringling Causeway, the extracted deficient grout there had high sulfate ion concentrations (ranging from 1000 to 12,000 ppm) relative to the extracted hardened grout (10e200 ppm). Studies on the corrosion behavior of steel in deficient grout were conducted to identify root cause (Permeh et al., 2016, 2018, 2019, 2021a, b, c; Krishna Vigneshwaran et al., 2018). As further exacerbation by the recent corrosion failures, the prepackaged grout product sometimes contained elevated chloride levels over 5% by weight of cement (U.S. Department of Transportation, 2011). Bridges identified to contain batches of the contaminated grout have recently been inspected to measure chloride content. Laboratory studies further investigating the critical chloride threshold concentrations in prepackaged thixotropic grout materials were conducted. In an FHWA report (Lee and Zielske, 2014), corrosion risk was described to be low if the total chloride content is less than 0.4% and high if the total chloride content is greater than 0.8%. These values were not applicable to grouts that have any other form of deficiencies. Permeh et al. (2019) showed that elevated free sulfate ion concentrations in deficient grout had synergistic effects with the low chloride ion concentrations (below conventional chloride threshold values) to reduce corrosion resistance. They concluded that the assessment of corrosion susceptibility in deficient grout by chloride values alone was insufficient as sulfate ion presence and grout characteristics were also important. Methodologies to inspect and repair tendons with deficient grout can be difficult due to the complexities associated with the geometry of internal and external draped tendons. Following the reports of corrosion of strand in deficient grout in Florida bridges, a guideline for sampling and assessment of deficient grout was published by FHWA (Theryo et al., 2013). Where tendon sections cannot be completely replaced, removal of the deficient grout may be needed. Hamilton et al. 2020, evaluated hydrodemolition methodologies to remove deficient grout and indicated that hydrodemolition did not completely remove grout especially in the case of soft grout. Also, water must not be

Figure 5.7 Corrosion of metal duct with deficient grout. Photos courtesy of FDOT.

96

Corrosion of Steel in Concrete Structures

entrapped within the tendon after the procedure. Methodologies to rehabilitate tendons with deficient grout by internal drying and injection of inhibitors were tested in Florida bridges (Whitmore and Tore Arnesen, 2020). Nondestructive testing to identify corrosion of strand and deficient grouts in tendons is needed. Positive results were obtained in recent demonstrations of techniques such as Magnetic Main Flux and Magnetic Flux Leakage for detection of steel metal loss within external tendons and sonic/ultrasonic impact echo testing for detection of anomalies in the internal and external tendons of posttensioned girders (Rehmat et al., 2019). Also, acoustic emission monitoring devices have been implemented to identify wire breaks. The reader is referred to the technical literature for a detailed description of nondestructive testing and evaluation techniques (Permeh and Lau, 2022; Lee et al., 2014; DMJM Harris, 2003; Azizinamini and Gull, 2012). Studies to identify suitable nondestructive tests for posttensioned systems have been made, including NCHRP 14e28. Due to the recent corrosion failures of posttensioned tendons that incorporated a prepackaged thixotropic grout product and the associated difficulties in its inspection and repair, there has been momentum in Florida for the specification of nonbonded posttensioned systems using flexible fillers such as waxes for continuity tendons, IBeams, hammerhead piers, and straddle piers (Robertson, 2015; Hamilton and Rice, 2015). The arguments for the change are that the wax materials should remain stable and inert and good performance has been observed in the nuclear industry and in Europe; if corrosion were to occur, failure would be progressive and the tendons can be replaced. Disadvantages for implementation of the technology in the US include the need to develop material and structural specifications, increased material costs, specialized construction methods, limited material suppliers, and limited expertise in the workforce in the technology. Specifications have been developed for flexible fillers by the Florida Department of Transportation.

5.5

Cathodic protection

Cathodic protection (CP) systems have been implemented for prestressed concrete systems (notingly for prestressed concrete cylinder pipes but ongoing experience for prestressed bridge piles and beams continue). CP systems and its materials are diverse but can be generally grouped as galvanic systems or impressed current systems. Application in atmospheric, buried, or submerged environments can further diversify CP systems and its materials. A NACE International state-of-the art-report on CP of prestressed concrete structures is available (NACE Int., 2002). Limitations of the use of CP systems for prestressed concrete include possible adverse effects due to the enhanced formation of products from the cathodic reaction. Alkalinization of the pore water at the steel surface due to cathodic polarization may contribute to the reduction of the bond strength at the steel/cement interface (Eyre, 1998). At potentials more negative to the half-cell potential of the reaction H2O þ e/Hads þ OH, excessive atomic hydrogen dissolution into the steel atomic lattice can cause hydrogen embrittlement. The pore water chemistry and environment

Corrosion of prestress and posttension reinforced concrete bridges

97

can affect the condition and rate to which hydrogen can form. Steel material makeup, properties, and surface conditions can also affect its susceptibility to hydrogen embrittlement, further complicating practical CP criteria (NACE Int., 2002; Eyre, 1998; Bennett and Schue, 1998). Generally, CP systems are maintained to prevent hydrogen formation. Other practical complications for CP of prestressed concrete include the need to consider appropriate or inappropriate continuity of strand and other steel components, requirement for uniform current distribution despite adverse geometry and mixed environments, and current shielding.

5.6

PT grout testing

Institutions such as the Post-Tensioning Institute (PTI), the American Concrete Institute (ACI), the Prestressing Concrete Institute (PCI) as well as the FHWA and state transportation departments have developed guidelines and specifications relating to the mitigation of corrosion of the steel strand in bonded posttensioned bridge elements. The specifications prescribe material, design, and construction requirements to improve corrosion durability. The prescribed material testing of the grout delineates nominal characteristics for grout workability as well as service and corrosion durability. The specification limits the content of aggressive chemical compounds (such as chlorides) that can initiate corrosion and the level of acceptable material performance and stability (such as grout bleed). Examples of grout specifications relating to corrosion durability are shown in Table 5.1. The thixotropic grouts, developed in the late 1990s, were developed to meet specifications to minimize grout bleed water formation and development of voids. Standardized tests to measure the extent of grout bleeding include the wick-induced bleed test (ASTM C940 (ASTM C940-16, 2016)), Schupak pressure bleed test (ASTM C1741 (ASTM C1741-18, 2018)), and the incline-tube test. The PTI M55195 Specification for Grouting of PT Structures specifies a wick bleed of 0.0% volume after 3 h and a range of maximum bleed water depending on the design vertical rise of tendon deviation. Tendons with a vertical rise greater than 6 ft should have maximum bleed water of 0% volume for the prescribed test pressures (50e100 psi). Texas and Florida DOT have a more general specification where the maximum bleed at 100 psi was defined as 0.0% volume for any vertical tendon applications (Section 938, 2020; DMS-4670, 2019). The inclined tube test (specified in Euronorm standards, EN 445.2) (CEN/TC 104, 2007) was developed as a large scale bleed test and also for evaluating the tendon configuration, grout materials, mixing guidelines, and mixing equipment. PTI and FDOT adopted the test in 2019 and specified that the grout bleeding shall not exceed 0.3% of the initial volume of the grout after 3 h kept at rest. In regards to grout segregation, the modified inclined-tube (MIT) test was adopted in recent research by Hamilton et al. (2014), Hamilton (2014), Permeh et al. (2016) to assess grout segregation. In those research programs, in part to identify grout robustness, the grout was mixed in excess of the manufacturer’s recommended upper water limit. Research by Permeh et al. (2018) indicated that the MIT tests allowed for better

98

Corrosion of Steel in Concrete Structures

Table 5.1 Grout specifications relating to corrosion durability. Test value

Property

PTI (PTI M55.1-19) [1]

FDOT (Section 938, 2020) [1]

TxDOT (DMS4670, 2019) [2]

VDOT (IIMS&B-91, 2016) [1]

Bleeding @ 3 ha

0.0%

Pressure Induced bleeding (vertical)a Permeability test @ 28 day

6e20 ft,b 0.0% @ 50 psi, >20 ft, 0.0% @ 100 psi 2500 C @ 30 V for 6 h.

0.0% @ 100psi

0.0% @ 100psi

0.0% @ 100psi

-

2500 C @ 30 V for 6 h.

-

Surface resistivity @ 28 day Inclined tube testa

-

16 kUcm

-

16 kUcm

15 15

Polarization resistance

>5

Potentiostatic: þ200mV SCE

2.8e10.6

Polarization resistance

>5

Potentiostatic: þ200mV SCE

5e10 >10

Potentiostatic: þ200mV SCE

6.5e10

10 years outdoor exposure 10 years of outdoor exposure

3.2

Pore solution

Mortar, admixed Cl

Pore solution

Mortar

Sand blasted surface Admixed

Ca(OH)2 pH 12.6 NaOH pH 13.9 Ca(OH)2

302

Concrete

Admixed

315

Concrete

Admixed

6e7

>3.2

Reference Sørensen et al. (1990) McDonald et al. (1995a) Kouril et al. (2010) Ogunsanya and Hansson (2019a) GarciaAlonzo et al. (2007b) (Randström et al., 2010)] GarciaAlonzo et al. (2007b) Bertolini et al. (1996) Bertolini and Gastaldi (2009) Treadaway et al. (1989) Treadaway et al. (1989)

Corrosion of stainless steel reinforcemnt in concrete

125

Table 6.3 Continued CCT % by wt. cement or solution

Alloy

Environment

Salt application

Exposure or test

316 316LN

Concrete

Admixed

Potentiostatic: þ200mV SCE

>8%

Concrete

Admixed

10 years of outdoor exposure Pickled With mill scale Polarization resistance

>3.2

Potentiostatic: þ200mV SCE

5.7e6.8

Potentiostatic: þ200mV SCE

>10 >10

Potentiostatic: þ200mV SCE

>10

Pore solution Mortar,

Admixed

Pore solution

Sand blasted surface

Ca(OH)2 pH 12.6 NaOH pH 13.9 Ca(OH)2

24100

>15 15 >5

Ca(OH)2

Pickled

Potentiostatic: þ200mV SCE

2.8e11.3

Pore solution

Pickled

Cyclic polarization

6e7

Mortar,

Admixed

Polarization resistance

>5

Pore solution

Pickled

Cyclic polarization

5.7e6

Reference Sørensen et al. (1990) Treadaway et al. (1989) Kouril et al. (2010) GarciaAlonzo et al. (2007b) Randström et al. (2010) Bertolini et al. (1996) Bertolini and Gastaldi (2009) Hurley and Scully (2006) Ogunsanya and Hansson (2019a) GarciaAlonzo et al. (2007b) Ogunsanya and Hansson (2019a) Continued

126

Corrosion of Steel in Concrete Structures

Table 6.3 Continued

Alloy

Environment

2205

Pore solution

2304

Salt application

Pore solution

Pickled

Concrete

Ponding

Pore solution

Sand blasted

Exposure or test Pickled With mill scale Cyclic polarization

Potentiostatic: þ200mV SCE

4.6e10.6

7.9

Potentiostatic: þ200mV SCE

>10 >10

Cyclic polarization

6.0e6.7

Ca(OH)2

Potentiostatic: þ200mV SCE

7.5e8

Pore solution pH 12.8e13.3 Ca(OH)2

Potentiostatic: þ200mV SCE

5.7e10.6

Potentiostatic: þ200mV SCE

3.5e6

Potentiostatic: þ200mV SCE

1.1e1.4

Ca(OH)2 pH 12.5 NaOH pH 13.9 Pore solution

2102

>15 >15 7e8

3.5

Pore solution

2101

CCT % by wt. cement or solution

Pickled

Ca(OH)2

Pickled in lab

Concrete

Ponding

2.9

Reference Kouril et al. (2010) Ogunsanya and Hansson (2019a) Ji et al. (2005) Schönning et al. (2013) Mahmoud et al. (2015) Bertolini et al. (1996) Ogunsanya and Hansson (2019a) Bertolini and Gastaldi (2009) Schönning et al. (2013) Bertolini and Gastaldi (2009) Hurley and Scully (2006) Ji et al. (2005)

Corrosion of stainless steel reinforcemnt in concrete

127

Figure 6.16 SEM micrographs of pits in rebar after polarization in chloride-containing synthetic pore solution. (A) austenitic 304L; (B) duplex 2001 (Alvarez et al., 2011). (iv) The ability of the corrosion products to protect the underlying steel from very rapid corrosion rates. When corrosion is initiated, the corrosion current density is very high but decreases over time. This may be attributed to a lack of oxygen for the cathodic half-cell reaction or to the solid corrosion products creating a partial barrier to the transfer of ions. Cui and Sag€ués (2003) measured corrosion activity at the defect sites in a stainless clad bar and found that corrosion moderated with time, which they attributed to progressively developing corrosion products. (v) The ability of the corrosion products to block, or partially block, any preexisting cracks. Corrosion of stainless rebar has been observed at the intersection between structural cracks and the embedded rebar when the concrete has been exposed to antiicing brine containing 21% Cl, one of the brines currently in use in Canada (Van Niejenhuis et al., 2015). Once the passive film has broken down, there is a rapid surge in corrosion current density, followed by a slow decline, as illustrated for the austenitic S24100 (XM28) and the duplex S32304 (2304) in Fig. 6.17. The curves in these figures represent the average current densities for five replicate specimens each in sound (uncracked) concrete, in concrete specimens with a loading crack transverse to the bars and concrete specimens with a loading crack longitudinal to the bar. Autopsying the specimens revealed corrosion only at the location of the cracks and there was an indication that the corrosion products were deposited within the crack.

In another study of the influence of loading cracks, N€urnberger and Beul (1999) and N€ urnberger (2005) did not detect any “serious corrosion” of 316LN or 2205 in concrete with cracks > RU, the resistance that the LPR measures is close enough to the polarization resistance that can be used as the actual value. However, in some environments, such as concrete, the RU is significant and should be considered (Jones, 1995). According to some researchers, corrosion current densities over w1 mA/cm2 are identified as a level of high corrosion risk, and corrosion current densities below 0.1 mA/cm2 are characterized as passive corrosion in the system (Gonzalez et al., 1995; Alonso et al., 2000; Polder and Peelen, 2002). However, it seems that the equipment used by these researchers generally gives lower values than the other commercial equipment (Gepraegs, 2002). Therefore, applying such definitions may over or underestimate the corrosion rate and cause errors in evaluations and life predictions. Interpreting the corrosion current density values of embedded steel bars in concrete obtained from the LPR technique is difficult in large part because determining the actual corroding area of steel is almost impossible and usually causes underestimation of the actual corrosion current density in the areas of active corrosion. The LPR has some advantages over the other measurement techniques, which make it popular in the evaluation of the corrosion rate in reinforced concrete: it is a nondestructive technique; it is a simple method, and it usually needs only a few minutes for corrosion rate determination. Because of its rapidity, it can effectively be used in kinetic studies of corrosion monitoring (Jones, 1995).

11.3.1

Potentiostatic LPR (potentiostatic transient technique)

In the potentiostatic LPR technique, a constant potential signal (usually 10 mV or, for steel embedded in concrete, 20 mV) in the form of a square wave between the working electrode (steel bar in concrete) and reference electrode and the response current is measured. The potential application period is determined by the time the current reaches steady state; using Eq. (11.1), the RP and hence, the corrosion current density and corrosion rate can be calculated. Fig. 11.3 schematically shows the applied potential and current response during potentiostatic transient measurement. This technique can also be used to directly determine the corrosion current density of steel rebars without using the Stern-Geary constant (B value) (Poursaee, 2010).

Corrosion measurement and evaluation techniques of steel in concrete structures

223

Figure 11.3 Applied potential and current response during potentiostatic LPR measurement.

When steel is polarized, the charge consumed from time t ¼ 0 to t can be written as follows: Zt qtotal ¼

Idt

(11.5)

0

where qtotal is the consumed charge in coulomb, t is the time (s), and I is current (A). Therefore, the area under the currentetime curves (Sa þ Sc in Fig. 11.3) can be used to determine the consumed charge during the polarization process. It is imperative to note that this is the total consumed charge in the interfacial corrosion process and the double-layer capacitance. Therefore, the effect of capacitance should be considered in the calculations. The charge used by the double layer capacitance can be determined using Eq. (11.8): qdl ¼ Cdl  V

(11.6)

where Cdl is the double layer capacitance (F), and V is the potential (V). Electrochemical impedance spectroscopy or galvanostatic pulse technique can determine the capacitance value. By deducting qdl from qtotal, the consumed charged during the corrosion processes can be calculated: qcorr ¼ Dqtotal
qdl

(11.7)

qcorr represents both charges obtained from anodic and cathodic polarization, considering the effect of double-layer capacitance. The mass loss during the polarization time can then be calculated by applying Faraday’s law as following: m¼

qcorr  M nF

(11.8)

224

Corrosion of Steel in Concrete Structures

where m (g) is the mass loss during the polarization time, M is the atomic weight of steel (w55.845 g/mol for iron), n is the number of equivalents exchanged electrons (in this case considered 2), and F is Faraday’s constant (96,500 C/mol). The amount of mass loss can then be determined per year, and the penetration depth can be calculated as: d¼

my nF

(11.9)

where d is the penetration depth (mm/year), my is the mass loss per year (g), and A is the corroding area (cm2). For iron, M ¼ 55.845 g/mol, r ¼ 7.875 g/cm3, and n ¼ 2, therefore:   Corrosion rateðmm=yearÞ icorr A = m2 ¼ 0:116

11.4

(11.10)

Galvanostatic pulse technique

The galvanostatic pulse technique was introduced for field application in 1988 (Newton and Sykes, 1988). This method is a rapid nondestructive polarization technique. A short-time anodic current pulse is applied galvanostatically between a counterelectrode placed on the concrete surface and the rebar. The applied current is usually in the range of 10e100 mA, and the typical pulse duration is between 5 and 30 s. The reinforcement is anodically polarized, and the resulting change in the electrochemical potential of the reinforcement is measured with respect to a reference electrode and recorded as a function of polarization time (Klinghoffer, 1995; Elsener et al., 1997). A typical potential response for a corroding reinforcement is shown in Fig. 11.4.

Figure 11.4 Schematic illustration of galvanostatic pulse results.

Corrosion measurement and evaluation techniques of steel in concrete structures

225

Figure 11.5 Schematic illustration of Eq. (11.12).

When the constant current, Iapp, is applied to the rebar, the polarization of the rebar, ht, at a given time t can be expressed as (Jones and Greene, 1966): 2

2

6 6 ht ¼ Iapp  4Rp  41
e



3


t Rp Cdl

3

7 7 5 þ RU 5

(11.11)

where: Rp ¼ polarization resistance, Cdl ¼ double-layer capacitance, and RU ¼ ohmic resistance of the concrete cover. By transferring Eq. (11.11) to logarithmic form, the values of Rp and Cdl, can be calculated as following (Newton and Sykes, 1988):   lnðhmax
hÞt ¼ ln Iapp  Rp




t Rp Cdl

 (11.12)

where hmax is the final steady-state potential value. Fig. 11.5 plots Eq. (11.12). If the line through the data points is extrapolated to t ¼ 0, it will give an intercept of Iapp  Rp and the slope of the line in 1/RpCdl. The remaining overpotential corresponds to Iapp  RU, which is the ohmic voltage drop across the concrete cover. After determining the polarization resistance (Rp) using the above method, the corrosion current Icorr can be calculated from the Stern-Geary equation (Stern and Geary, 1957; Newton and Sykes, 1988), Eq. (11.2).

11.4.1 Equipment with the guard ring An auxiliary counterelectrode, referred to as a guard ring electrode can confine the polarization to a known length of reinforcing bar. The counterelectrode and the guard ring are typically arranged as annular metal rings with a reference electrode in the center. The potential or current applied from the guard ring is intended to repel the signals from the central counterelectrode and confine them to an area of the structure located approximately under the counterelectrode, as schematically shown in Fig. 11.6.

226

Corrosion of Steel in Concrete Structures

Reference electrode

Measurement computer

Counter electrode Guard ring Sponge

Confined polarized area

Figure 11.6 Schematic plan of the GalvaPulse with guarding to limit the polarized area while performing the corrosion measurement.

There are two well-known commercially available instruments based on the galvanostatic pulse technique: (i) The GECOR is considered a galvanostatic pulse technique instrument, although it is often referred to as an LPR device (Broomfield, 1996; Feliu et al., 1996; Newhouse and Weyers, 1996). The electrode assembly has three reference electrodes, one located in the center and two between the counterelectrode and guard ring. The latter is used to adjust the guard electrode current to maintain the potential difference between the two reference electrodes constant during the polarization procedure. (ii) The GalvaPulse is a galvanostatic pulse instrument developed by the FORCE Institute in Denmark. Signal confinement is over a 70 mm length of the bar, and measurements can be made with or without the guard ring electrode. The measuring cell has an Ag/AgCl reference electrode at the center, a zinc counterelectrode, and a zinc guard ring.

11.5

Electrochemical impedance spectroscopy (EIS)

The EIS studies the system response to the application of a small amplitude alternating potential or current signal at different frequencies. The popularity of the EIS methods for reinforced concrete has increased remarkably in recent years because analysis of the system response provides information about the double-layer capacitance, interface, structure, reactions which are taking place, corrosion rate, and electrolyte (environment) resistance (Silverman, 1990; Jones, 1995; Lasia, 1999). An electrochemical process can be considered an electrical circuit with basic elements such as resistors, capacitors, and inductors. Therefore, in interpreting the response to an AC, the AC circuit theory can be used successfully to demonstrate a corrosion process, and also it may be used to understand the corrosion behavior and prediction of the corrosion rates.

Corrosion measurement and evaluation techniques of steel in concrete structures

227

In direct current, ohm’s law is given as: V ¼ IR

(11.13)

where V ¼ Potential, I ¼ Direct current, and R ¼ Actual resistor. In the AC condition, ohm’s law becomes: V ¼ IZ

(11.14)

where V ¼ Potential, I ¼ Amplitude of the alternating current, and Z ¼ Impedance. Direct current can be viewed as alternating current at zero frequency. In this case, the resistance comprises only one or more actual resistors. When the frequency is not zero, all circuit elements affecting the current flow, for example, resistors, capacitors, and inductors, cause the impedance. The created resistance by capacitors and inductors depends on frequency, while that created by a resistor is not dependent on frequency (Silverman, 1986). A sinusoidal current or voltage can be represented as a rotating vector, as shown in Fig. 11.7. In the right-hand diagram of Fig. 11.7, the X component shows the observed current, so it becomes the real component of the rotating vector, while the Y component is a contribution that is not observed; therefore, it is named the imaginary component of the rotating vector. The mathematical descriptions of the two components are as follows: Real current ¼ Ix ¼ jIjcosðutÞ

(11.15)

Imaginary current ¼ Iy ¼ jIjsinðutÞ

(11.16)

where t ¼ time and u ¼ frequency in radians per second ¼ 2pf (f ¼ frequency in Hertz).

Sinusoidal representaon

Imaginary

Ireal=Isin(ωt)

ωt Real

Time

Figure 11.7 Relationship between sinusoidal AC and rotating vector representation.

228

Corrosion of Steel in Concrete Structures

To separate the real (x) and imaginary (y) components, the magnitude of the imagpffiffiffiffiffiffiffi inary part should be multiplied by j ¼
11, and then the real and imaginary values can be reported separately. The equations for AC impedance become: Etotal ¼ Ereal þ Eimaginary ¼ E0 þ jE00

(11.17)

Itotal ¼ Ireal þ Iimaginary ¼ I0 þ jI00

(11.18)

Ztotal ¼ Z0 þ Z00 ¼

E0 þ jE00 I0 þ jI00

(11.19)

Absolute amplitude, jZj, of the impedance (that is, the length of the vector) and the phase angle q, are defined by: jZj ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z0 2 þ Z00 2

tanq ¼

(11.20)

Z00 Z0

(11.21)

AC impedance aims to measure the impedance Z as Z0 and Z00 and then model the response using an equivalent simple circuit (Silverman, 1986).

11.5.1

Data presentation

There are different ways to illustrate the response of an electrochemical system to an applied AC potential or current. The most common plots are the Nyquist plot and Bode plots. If at each excitation frequency, the real part is plotted on the x-axis, and the imaginary part is plotted on the y-axis of a chart, a “Nyquist plot” is formed. A simple corroding system can be assumed as solution resistance, in series, with a combination of a resistor and a capacitor, which represent the polarization resistance and doublelayer capacitance, respectively. This simple representation is called a Randles circuit and is shown in Fig. 11.8. Polarization resistance (Rp) Solution resistance (RΩ)

Double layer capacitance

Figure 11.8 The equivalent circuit for a simple electrochemical system. 1

Mathematicians use i to stand for current.

pffiffiffiffiffiffiffi
1, but electrochemists use j to avoid confusion with i, the symbol for

Corrosion measurement and evaluation techniques of steel in concrete structures

229

Fig. 11.9 schematically illustrates the Nyquist plot for a simple electrochemical system corresponding to the analog circuit in Fig. 11.8. It should be noted that each point on the Nyquist plot is the impedance at one frequency. On the Nyquist plot, the impedance can be represented as a vector of length jZj, and the angle between this vector and the x-axis is the phase angle “q.” At high frequencies, at the leftmost end of the semicircle, where the semicircle intersects the x-axis, the impedance of the Randles cell is entirely produced by the ohmic resistance, RU. The frequency reaches its low limit at the rightmost end of the semicircle. At this frequency, the Randles cell also approximates a pure resistance, but now the value is (RU þ Rp). The Nyquist plot has some limitations: (i) The frequency is not clearly shown on the plot, and it is not possible to determine, for a specific point, the frequency used to record that point; (ii) The ohmic and polarization resistances can be directly determined from the plot, but the electrode capacitance can only be calculated if the frequency information is known, using Eq. (11.22):

C¼

1 umax  Rp

(11.22)

(iii) If there are high and low impedance components in the circuit, the larger impedance controls plot scaling, and distinguishing the low impedance semicircle would probably be impossible.

Bode plots are another popular presentation method for impedance data. In the Bode plot, the data are plotted with the log of frequency on the abscissa and the log of the absolute value of the impedance (jZj) and phase-shift (q) on the ordinate. Fig. 11.10 shows a typical Bode Plot for the same system as shown in Fig. 11.8 (Silverman, 1986). Since the frequency appears as one of the axes in the Bode plot, it is easy

Figure 11.9 Nyquist plot for a simple electrochemical system.

Corrosion of Steel in Concrete Structures

4 fmax

Log|Z|

3

Contribution from resistor-capacitor is fully charged (only corrosion process)

2

90 75 60 45 30 15 0

f2

1

Contribution from Capacitor (surface capacitance) Contribution from resistor and capacitor

0 -2

-1

0

1 2 Log ω

3

4

Phase angle, q degree

230

5

Figure 11.10 Bode plots of the frequency response (dotted line) and phase angle (solid line) for an electrochemical system with RU ¼ 10 U, Rp ¼ 1000 U, and Cdl ¼ 93.8 mF.

to understand the dependence of impedance on the frequency from the plot. The log jZj versus log u curve can be used to determine the values of Rp and RU. jZj becomes independent of frequency at very high and low frequencies. The ohmic resistance controls the impedance at the highest frequencies, and the log(RU) can be read from the high-frequency horizontal level. On the other hand, at the lowest frequencies, log(Rp þ RU) can be read from the low-frequency horizontal portion. The Bode format is advantageous when data scatter prevents satisfactory fitting of the Nyquist semicircle. In general, the Bode plot provides a more understandable description of the frequency-dependent behavior of the electrochemical system than does the Nyquist plot, in which frequency values are not clear.

11.6

Cyclic polarization

The cyclic potentiodynamic polarization technique is a relatively nondestructive measurement that can provide information about the corrosion rate, corrosion potential, susceptibility to pitting corrosion of the metal, and concentration limitation of the electrolyte in the system. The technique is built on the idea that predictions of the behavior of a metal in an environment can be made by forcing the material from its steady-state condition and monitoring how it responds to an increasing force and as the force is removed at a constant rate and the system is reversed to its original steady-state condition. An applied potential is the force and is increased at a continuous, often slow, rate by using a potentiostat. This rate is called the polarization scan rate and is an experimental parameter. As the specimen’s potential changes continuously, the resulting current is monitored, and then the applied potential is plotted versus the logarithm of the resulting current density. The conductivity of the electrolyte (environment) is a significant factor that should be considered in all electrochemical experiments, especially in the cyclic polarization technique. The electrolyte resistance causes a potential drop between the working electrode and the reference electrode and can cause errors. This effect has an important impact on the interpretation and should be compensated.

Corrosion measurement and evaluation techniques of steel in concrete structures

231

11.6.1 Scan rate An appropriate cyclic polarization scan rate for the specific system under study must be used; otherwise, the results do not accurately reflect the corrosion behavior. To ascertain that the current/voltage relationship reflects only the interfacial corrosion process at every potential of the polarization scan, the effect of capacitance should be minimized. For this purpose, the capacitor should remain fully charged; otherwise, some of the current generated would reflect the charging of the surface capacitance in addition to the polarization resistance, and the measured current would then be greater than the current produced by the corrosion reactions. To reduce the effect of the capacitance, the scan rate should be slow enough to keep the capacitance fully charged during the experiment. In this case, the current-potential relationship reflects the polarization resistance (interfacial corrosion process) at every potential of the polarization scan (Silverman, 1998). Because the capacitance and resistance are functions of the material, environment, and applied potential, choosing the appropriate scan rate is not easy. An outline of an approach to determine the maximum scan rate is given by Mansfeld and Kendig (1981). The principle of the method is based on the Bode plot, Fig. 11.10, represented by the Randles circuit (Fig. 11.8). At low frequencies, jZj ¼ RU þ Rp; therefore, to determine the polarization resistance accurately, the frequency characterization of the scan rate should be less than the fmax in the low-frequency portion of the Bode plot. u ¼ 2pf, where f is the frequency of the applied sine wave, and E can be calculated by using Eq. (11.23) E ¼ Eo sinðutÞ ¼ 1=2DEej2pft

(11.23)

DE is the peak-to-peak amplitude and is equal to 2Eo. DE is usually taken w10 mV. This assures the linear response of the system to the applied potential (Stern and Geary, 1957). The rate of change of potential can be calculated by taking the derivative of Eq. (11.23):   dE=dt ¼ 1=2DE  ðj2pftÞ  ej2pft

(11.24)

and the maximum rate of change would be: DEmax ¼ pfDE

(11.25)

This maximum rate of change corresponds to scan rate, S, and by using fmax in Eq. (11.18), the maximum scan rate, Smax, can be calculated as following: Smax ¼ p,DE,f max
600e >800e >1000e >1200e

50e52 52e54 54e56 56e58 58e60

>1500e >2000e >2500e >3000e >3500e

60e65 65e70 70e75 75e80 90e100

>4500 >6500 >7500 >9000 >10,000

Acoustic emission monitoring for corrosion damage detection and classification

13.5

271

Small scale specimens

A total of 20 specimens were tested under an accelerated corrosion test setup. The specimens were reinforced with an embedded 1/2 inch diameter prestressing strand; the dimensions of each specimen were 4.5 in  4.5 in  20 in. (114  114  1270 mm). The test matrix included 11 precracked specimens and nine pristine specimens to evaluate the effect of cracks on AE attenuation. Specimens were immersed in a tank filled with a 3% NaCl solution at room temperature to a level 0.25 inches (7 mm) below the reinforcing strand. A copper plate with the same length as the specimens was placed below each specimen to serve as the cathode. Accelerating the corrosion process was established by forming a galvanic cell using a rectifier to impress a direct external current to the specimens. The current that flows between the dissimilar metals controls the degree of corrosion activity in the cell. The rectifier was connected between the copper plate (cathode) and the prestressing strand (anode). Fig. 13.5 shows a diagram of the corrosion cell for a single cracked specimen placed in the test vessel (Mangual et al., 2013a). The results showed that stresses induced by volume expansion resulted in dense AE that was correlated to the onset of corrosion and nucleation of cracks as shown in Fig. 13.6. Comparing acoustic emission data with electrochemical HCP measurements showed that the cumulative signal strength relates well with potential variations as shown in Fig. 13.7. It also allowed discrimination between different corrosion stages. The magnitude of the slope in the cumulative signal strength (CSS) versus the time curve was able to portray the depassivation process and onset of corrosion as corroborated by HCP. This shows that AE is capable of detecting and discriminating between early corrosion stages while mimicking the behavior of resistivity changes in the concrete.

Digital voltmeter

Cu/CuSO4 Reference electrode

Power supply

– +

AE sensor

Resistor 3% NaCl water solution AE/parametric data acquisition

Figure 13.5 Schematic of the accelerated corrosion setup (Mangual et al., 2013a).

272

Corrosion of Steel in Concrete Structures 100

100

Amplitude (dB)

Onset

Cracking

90

90

80

80

70

70

60 50

C9 0

2000

1000

3000

4000

0

Duration (µ µs)

1000

2000

3000

60 50 4000

Duration (µ µs)

Figure 13.6 Amplitude (dB) versus duration (ms) plot (Mangual et al., 2013a). 600

Corrosion onset

300 0

1.0 –350 mV

–300 –600

0.0

0

50

100

Half-cell potential (mV)

CSS (mVs)

2.0

–900

Time (hr)

Figure 13.7 Cumulative signal strength and half-cell potential versus time.

Source location based on AE data enabled the accurate detection of events as a result of passivity breakdown along with the reinforcement and debonding (Mangual et al., 2013a). Fig. 13.8 shows the source location results for a precracked specimen. As seen in the figure, most of the AE activity is concentrated near the crack where corrosion is predicted to take place as a result of chlorides accessibility in this region (no AE activity was detected at the exact crack location as it is already formed). At the conclusion of the test, the specimens were taken apart and the strand was removed for visual inspection of corrosion. The strand was then cleaned and reweighed to measure the steel mass loss as detailed in Mangual et al., (2013a). As shown in Fig. 13.9, heavy corrosion damage, in terms of the formation of longitudinal and tangential cracks, was observed in some of the specimens. ElBatanouny et al. (2011) proposed that AE Intensity Analysis can be used to classify corrosion damage. The method was first proposed by Fowler et al. (1989) to quantify damage in FRP vessels and tanks by calculating two parameters: historic index, H(t), and severity, Sr. Historic index is a form of trend analysis that measured

Y position (in.)

4.5

No. of events

Acoustic emission monitoring for corrosion damage detection and classification

20

ch. 1

273

ch. 3

3.0 1.5 ch. 2

0 0

5

ch. 4 10 X position (in.)

ch. 1,2

15

20

15

20

ch. 3,4

10 0 0

5

10 X position (in.)

Figure 13.8 Source location and number of events for a precracked specimen.

Steel strand

Midsection crack Corrosion crack

Steel strand

Corrosion cracks 1 in.

Diffused oxides

1 in.

Tangential crack 0.1 in.

Figure 13.9 Corrosion damage in steel and concrete.

the rate of change in the cumulative signal strength while severity is the average of a certain number of hits (50 hits) with the highest signal strength. The increase in these parameters can be related to the accumulation of damage. Historic index and severity can be calculated using Eqs. (13.4) and (13.5) where: N is the number of hits up to a time (t), Soi is the signal strength of the i-th event, and K is empirically derived factor that varies with the number of hits. In this study, the value of K was selected to be: a) N/A if N  50, b) K¼N-30 if 51  N  200, c) K ¼ 0.85N if 201  N  500, and d) K¼Ne75 if N  501. N P

N HðtÞ ¼ NK

Soi

i¼Kþ1 N P

(13.4)

Soi

i¼1

Sr ¼

¼50 1 iX Soi 50 i¼1

(13.5)

274

Corrosion of Steel in Concrete Structures

Intensity Analysis results were correlated with HCP measurements and measured sectional mass loss yielding two AE based corrosion classification charts for precracked and uncracked specimens as shown in Fig. 13.10. For precracked specimens the chart divides the corrosion damage into four categories as illustrated in Mangual et al. (2013a) as follows: A) No damage: at this level the steel is still in the passive condition and no corrosion damage occurred, b) Depassivation: at this level, corrosion has just initiated with sectional mass loss less than 15%, c) Cracking: refers to the level at which cracks due to corrosion started to form and the sectional mass loss is less than 21%, and d) Severe damage: more cracks form and the sectional mass loss exceeds 21%. For pristine specimens, specimens in which depassivation was absent laid in region A of the intensity analysis grading chart, whereas depassivated specimens lay in the B region (Mangual et al., 2013b). These charts enable early detection of corrosion and classification of corrosion damage. Such charts also could extend the use of AE to personnel that are less acquainted with the method to perform damage evaluation without the need to send the collected data to AE specialists. It is noted that these figures include uncertainties associated with AE monitoring which should be studied and quantified in future studies.

(a) 107

Severity

106

Specimen-section loss Cl–19% C2–15% C3–11%

D C

105

C4–17% C5–20% C6–24% C7–26%

104 A 1

C8–28% C9–21% C10–23% C11–13%

B

Historic index

10

(b) 107

Specimen A1

Severity

A2 A3

106

A4

B

A5

105

104

A6 B1 B2

A 1

Historic index

10

B3

Figure 13.10 AE based corrosion classification charts for: (A) precracked specimens, and (B) pristine specimens (Mangual et al., 2013a, 2013b).

Acoustic emission monitoring for corrosion damage detection and classification

275

13.5.1 Medium scale specimens Long term corrosion tests were performed on medium scale PC beams to provide a better representation of actual environments for corrosion. The corrosion was accelerated using 3 days wet/4 day dry cycles with a 3% NaCl solution as shown in Fig. 13.11. Each beam was reinforced with two 1/2 in. (12.7 mm) prestressing strands located in the compression zone and measured 16 ft. Four in. in length (4.98 m). The beams were T-shaped with a total height of 15 in. (380 mm), a web thickness of 6 in. (150 mm), and a flange width and depth of 24 in. (610 mm) and 3 in. (75 mm), respectively. The test included three specimens where two were precracked to 0.016 in. (0.4 mm), specimen CC-0.4, and 0.032 in. (0.8 mm), specimen CC-0.8, while the last specimen was pristine. The specimens were continuously monitored by AE for 140 days. No corrosion damage was detected in the pristine specimen from any method due to the limited ability of chlorides to penetrate through concrete subjected to prestressing force. The results of the precracked specimens showed that the AE parameter CSS is able to detect corrosion with a higher sensitivity than HCP. As shown in Fig. 13.12, HCP for specimen CC-0.8 showed that corrosion initiated within the first week of testing, which agreed with CSS results which showed a high rate of corrosion activity. For specimen CC-0.4, HCP results only approached the corrosion threshold toward the end of the test while AE was showing a steady increase in the AE activity illustrating that corrosion is occurring. To further investigate the corrosion activity, linear polarization resistance (LPR) measurements were also taken to calculate the corrosion rate. Based on Andrade et al., (1990) classification, LPR results showed that CC-0.4 had a moderate corrosion rate while CC-0.8 had a high corrosion rate. This agrees with the AE results where the rate of AE activity for CC-0.8 was higher than that of CC-0.4. Visual inspection of damaged prestressing strands showed clear evidence of pitting (localized) corrosion in both specimens as shown in Fig. 13.13. This corrosion type leads to a significant reduction in the residual capacity of the specimens. Crack width is a significant factor in the formation and intensity of pitting in terms of pit depth. Load testing of the beams at the conclusion of the test showed a reduction in the capacity of the beams where corrosion was detected by AE as compared to the pristine specimen where no corrosion was detected. This indicated that AE has the ability to detect corrosion before a reduction in the strength of the structure occurs. Therefore,

Prestressing strand

AE sensor 76

1220

Acrylic dike Copper plate 3% NaCl solution

25 430

406

430

4980

Figure 13.11 Corrosion test setup and AE sensor layout (in mm; 1 in. ¼ 25.4 mm) (ElBatanouny et al., 2014c).

276

Corrosion of Steel in Concrete Structures

(a) 1.00

–200 Corrosion threshold

0.50

HCP

0.25 0.00

0

(b) 1.00

–600

50

100 Time (days)

–800 150 0

CSS HCP

0.75 CSS (mV-s)

–400

CSS

–200 Corrosion threshold

0.50

–400

0.25

–600

0.00 0

50

100 Time (days)

Potential (mV)

CSS (mV-s)

0.75

Potential (mV)

0

–800 150

Figure 13.12 Cumulative signal strength and half-cell potential versus time. (A) Specimen precracked crack width (A) 0.016 in. (0.4 mm), specimen CC-0.4 and (B) 0.032 in. (0.8 mm), specimen CC-0.8 (ElBatanouny et al., 2014c).

(a)

(b)

(c)

(d) 200 µm

200 µm

Figure 13.13 Photographs showing pitting corrosion: (A and B) specimens CC-0.4 and CC-0.8, respectively, and (C and D) SEM micrographs specimens CC-0.4 and CC-0.8, respectively. (1 in. ¼ 25.4 mm) (ElBatanouny et al., 2014c).

Acoustic emission monitoring for corrosion damage detection and classification

277

107 D

CC-0.8

Severe damage

Severity (pV-s)

CC-0.4 C 106

105

Cracking

1

B

Depassivation

A

No damage

10 Historic index

100

Figure 13.14 Corrosion intensity analysis results for cracked specimens (ElBatanouny et al., 2014c).

adopting this method for real-time corrosion monitoring can reduce, if not eliminate, the risk of sudden failure as a result of corrosion damage (ElBatanouny et al., 2014c). Intensity Analysis was performed on the cracked specimens using the limits set in the small scale specimens study as shown in Fig. 13.14. The results showed that AE Intensity Analysis can enable the detection and classification of corrosion damage. This is true for small and medium scale specimens illustrating that the method may be independent of specimen size and duration of exposure. These results were also compared with LPR results and had a good agreement.

13.6

Special considerations and potential applications in field

AE is a non-intrusive method that has the ability to detect damaged areas through source triangulation. The high sensitivity of the method enables it to perform a global assessment of the structure. However, effective filtering protocols are needed for field applications to reject noise related to environmental containments such as rain, hail and wind with debris. The method is currently deployable in elements such as piles and foundations. For superstructures, a method to separate AE data from corrosion and that from other sources such as service loading should be developed. More studies should be conducted on the applicability of the developed charts for field applications. Uncertainties related to noise rejection, AE wave speed, and signal attenuation in the field should also be investigated (ElBatanouny et al., 2014d). Note that subsequent studies by others

278

Corrosion of Steel in Concrete Structures

showed that the proposed intensity analysis method presented in this chapter is also sensitive for corrosion damage detection in small and full-scale reinforced concrete samples (Abousussien, 2017; Abouhussien and Hassan, 2018)

13.7

Special considerations for wireless sensing

AE systems have been steadily developing over the past decades, to become affordable and deployable. Currently, a self-powered wireless AE system is commercially available in the United States through Mistras Group, Inc. This makes AE suitable for remote/real-time/rapid inspection of massive structures using a minimal number of sensors. The authors of this chapter successfully implemented AE wireless systems in a previous project with the Savannah River National Laboratory where AE sensors were embedded in a Meso-Scale testbed to monitor cracks in grout used for In-Situ Decommissioning (ISD) of nuclear structures (Ziehl et al., 2012; ElBatanouny et al., 2012). The wireless AE system “smart node” reduces the data transferred wirelessly by sending the waveform parameters without including the waveforms. The waveforms are saved on an SD card that is attached to the node for more detailed signal processing. This data reduction technique enables the system to process and send data wirelessly in real-time and reduce the possibility of data loss due to insufficient buffer. However, the capacity of data collection of the wireless system is less than that of wired systems. Therefore, future efforts should focus on comparing AE data collected from wireless AE systems to wired AE systems. Since all damage evaluation techniques proposed are based on wired systems, a probabilistic analysis should be conducted to evaluate and modify damage assessment limits for wireless systems.

13.8

Summary

The feasibility of acoustic emission to detect corrosion was evaluated by comparing the AE results to electrochemical methods, sectional mass loss, and visual evidence of corrosion damage. AE parameters, such as cumulated events and signal strength, were found to detect the initiation of corrosion prior to electrochemical measurements. The method can be effectively used in places where there is no provision for electrochemical measurements. AE based Intensity Analysis charts can enable the detection and classification of corrosion damage using empirical limits for corrosion levels. This chart classifies corrosion damage in specimens with different sizes and exposure times showing that it may be independent of size and duration. Unlike some electrochemical techniques, the proposed AE corrosion classification charts have the ability to detect and quantify corrosion damage at early stages. This enables the development of acoustic emission into a damage quantification tool for maintenance prioritization because significant damage (such as macro-cracking and

Acoustic emission monitoring for corrosion damage detection and classification

279

spalling) is not required for detection. The proposed chart can also be used to estimate safe remaining service life as it is linked to cross-sectional mass loss results. However, the uncertainties associated with the relation between AE and mass loss should be quantified prior to full implementation.

Acknowledgments Portions of this work were published in the Precast/Prestressed Concrete Institute ConventionNational Bridge Conference Proceedings (PCI-NBC), 2014 held in Washington, D.C. in a paper titled “Review of Acoustic Emission Corrosion Monitoring of Prestressed Concrete Bridges”.

References Abdelrahman, M., 2013. Assessment of Damage in Concrete Structures Using Acoustic Emission (Master thesis). University of South Carolina, p. 146. Abdelrahman, M., ElBatanouny, M.K., Ziehl, P., 2014. Acoustic emission based damage assessment method for prestressed concrete structures: modified index of damage. Engineering Structures 60, 258e264. Abouhussien, A.A., 2017. Acoustic Emission Based Structural Health Monitoring of Corroded/ Un-Corroded Concrete Structures (Doctoral dissertation). Memorial University of Newfoundland, p. 200. Abouhussien, A.A., Hassan, A.A., 2018. Acoustic emission monitoring of corrosion damage propagation in large-scale reinforced concrete beams. Journal of Performance of Constructed Facilities 32 (2), 04017133. Appalla, A., ElBatanouny, M.K., Velez, W., Ziehl, P., 2016. Assessing corrosion damage in posttensioned concrete structures using acoustic emission. Journal of Materials in Civil Engineering 28 (2), 04015128, pp. 10. Aki, K., Richards, P.G., 1980. Quantitative Seismology, Theory and Methods, vol. 1. W. H. Freeman and Company, New York, N.Y. Andrade, C., Alonso, M.C., Gonzalez, J.A., 1990. An Initial Effort to Use the Corrosion Rate Measurements for Estimating Rebar Durability. Corrosion rates of steel in concrete, ASTM STP, 1065, pp. 29e37. ASTM E1316-13c, 2013. Standard Terminology for Nondestructive Examinations. American Standard for Testing and Materials, pp. 1e38. Bhat, C., Bhat, M.R., Murthy, C.R.L., 2003. Acoustic emission characterization of failure modes in composites with ANN. Composite Structures 61, 213e220. ElBatanouny, M., Mangual, J., Ziehl, P., Matta, F., 2011. Corrosion intensity classification in prestressed concrete using acoustic emission technique. In: Proc. American Society for Nondestructive Testing (ASNT) Fall Conference and Quality Testing Show 2011, Palm Springs, CA, p. 8. ElBatanouny, M., Larosche, A., Ziehl, P., Yu, L., 2012. Wireless monitoring of in situ decommissioning of nuclear structures using acoustic emission. In: 8th International Conference on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies 2012 (NPIC & HMIT), San Diego, CA, July 22e26.

280

Corrosion of Steel in Concrete Structures

ElBatanouny, M., Ziehl, P., Larosche, A., Mangual, J., Matta, F., Nanni, A., 2014a. Acoustic emission monitoring for assessment of prestressed concrete beams. Construction and Building Materials 58, 46e53. ElBatanouny, M., Larosche, A., Mazzoleni, P., Ziehl, P., Matta, F., Zappa, M., 2014b. Identification of cracking mechanisms in scaled FRP reinforced concrete beams using acoustic emission. Experimental Mechanics 54 (1), 69e82. ElBatanouny, M., Mangual, J., Ziehl, P., Matta, F., 2014c. Early corrosion detection in prestressed concrete girders using acoustic emission. Journal of Materials in Civil Engineering 26 (3), 504e511. ElBatanouny, M., Abdelrahman, M., Ziehl, P., 2014d. Review of acoustic emission corrosion monitoring of prestressed concrete bridges. In: PCI Convention and National Bridge Conference, Washington, D.C., September 6e9. Enoki, M., Kishi, T., 1988. Theory and analysis of deformation moment tensor due to microcracking. International Journal of Fracture 38 (4), 295e310. Farid Uddin, A.K.M., Numata, K., Shimasaki, J., Shigeishi, M., Ohtsu, M., 2004. Mechanisms of crack propagation due to corrosion of reinforcement in concrete by AE-SiGMA and BEM. Construction and Building Materials 18 (3), 181e188. Fowler, T., Blessing, J., Conlisk, P., Swanson, T.L., 1989. The MONPAC system. Journal of Acoustic Emission 8 (3), 1e8. Grosse, C.U., Ohtsu, M., 2008. Acoustic Emission Testing. Springer, Berlin, Germany. Gutkin, R., Green, C.J., Vangrattanachai, S., Pinho, S.T., Robinson, P., Curtis, P.T., 2011. On acoustic emission for failure investigation in CFRP: pattern recognition and peak frequency analyses. Mechanical Systems and Signal Processing 25 (4), 1393e1407. Hossain, M., Ziehl, P., Yu, J., 2012. Source characterization of acoustic emission during fatigue crack growth in steel bridge material. In: NDE/NDT for Highways and Bridges: Structural Materials Technology (SMT) Topical Conference, Aug 21e24, NY. Idrissi, H., Limam, A., 2003. Study and characterization by acoustic emission and electrochemical measurements of concrete deterioration caused by reinforcement steel corrosion. NDT & E International 36 (8), 563e569. Jaffer, S., Hansson, C., 2009. Chloride-induced corrosion products of steel in crackedconcrete subjected to different loading conditions. Cement and Concrete Research 39, 116e125. Kim, K.Y., Sachse, W., 1984. Characteristics of AE signals from indentation cracks in glass. In: Progress in Acoustic Emission II (JSNDI), Proceedings of the 7th International AE Symposium, Zao, Japan, pp. 163e172. Li, J., Du, G., Jiang, C., Jin, S., 2012. The classification of acoustic emission signals of 304 stainless steel during stress corrosion process based on K-means clustering. Anti-corrosion Methods & Materials 59 (2), 76e80. Li, Z., Li, F., Zdunek, A., Landis, E., Shah, S.P., 1998. Application of acoustic emission technique to detection of reinforcing steel corrosion in concrete. ACI Materials Journal 95 (1), 68e76. Maji, A.K., Ouyang, C., Shah, S.P., 1990. Fracture mechanism of quasibrittle materials based on acoustic emission. Materials Research 5 (1), 206e217. Mangual, J., ElBatanouny, M.K., Ziehl, P., Matta, F., 2013a. Acoustic-emission-based characterization of corrosion damage in cracked concrete with prestressing strand. ACI Materials Journal 110 (1), 89e98. Mangual, J., ElBatanouny, M.K., Ziehl, P., Matta, F., 2013b. Corrosion damage quantification of prestressing strands using acoustic emission. Journal of Materials in Civil Engineering 25 (9), 1326e1334.

Acoustic emission monitoring for corrosion damage detection and classification

281

Marec, A., Thomas, J.-H., El Guerjouma, R., 2008. Damage characterization of polymer-based composite materials: multivariable analysis and wavelet transform for clustering acoustic emission data. Mechanical Systems and Signal Processing 22 (6), 1441e1464. Mejia, F., 2012. Advanced Computing Methods for Knowledge Discovery and Prognosis in Acoustic Emission Monitoring. Dissertation. University of Miami, Coral Gables, FL. Miller, R.K., Mclntire, P., 1987. Non-destructive Testing Handbook, vol. 5. Acoustic Emission Testing, American Society for Nondestructive Testing, Columbus, OH. Moser, R.D., Singh, P.M., Kahn, L.F., Kurtis, K.E., 2011. Chloride-induced corrosion of prestressing steels considering crevice effects and surface imperfections. Corrosion 67 (6), 065001-1e065001-14. Murakami, Y., Yuyama, S., Shimizu, T., Kouyama, H., Matsushima, M., 1993. Deformation behavior and AE characteristics for anchorage pulling tests on the foundations of power transmission pylons (part II, moment tensor analysis). In: Proceedings of the 9th National AE Conference (JSNDI), Okinawa, Japan, pp. 143e150. Ohno, K., Ohtsu, M., 2010. Crack classification in concrete based on acoustic emission. Construction and Building Materials 24 (12), 2339e2346. Ohtsu, M., 1987. Acoustic emission characteristics in concrete and diagnostic applications. Journal of Acoustic Emission 6 (2), 99e108. Ohtsu, M., 1991. Simplified moment tensor analysis and unified decomposition of acoustic emission source: application to in-situ hydrofracturing test. Journal of Geophysics Research 96 (B4), 6211e6221. Ohtsu, M., 1995. Acoustic emission theory for moment tensor analysis. Research in Nondestructive Evaluation 6 (3), 169e184. Ohtsu, M., Ono, K., 1984. A generalized theory of acoustic emission and green’s functions in a half space. Journal of Acoustic Emission 3 (1), 27e40. Ohtsu, M., Tomoda, Y., 2008. Phenomenological model of corrosion process in reinforced concrete identified by acoustic emission. ACI Materials Journal 105 (2), 194e199. Oliveira, R.d., Marques, A.T., 2008. Health monitoring of FRP using acoustic emission and artificial neural networks. Computers and Structures 86 (3e5), 367e373. Omkar, S.N., Karanth, R., 2008. Rule extraction for classification of acoustic emission signals using ant colony optimisation. Engineering Applications of Artificial Intelligence 21 (8), 1381e1388. Perrin, M., Gaillet, L., Tessier, C., Idrissi, H., 2010. Hydrogen embrittlement of prestressing cables. Corrosion Science 52, 1915e1926. Pollock, A.A., 1986. Classical wave theory in practical AE testing. In: Progress in AE III, Proceedings of the 8th International AE Symposium, Japanese Society for Nondestructive Testing, pp. 708e721. Prateepasen, A., Jirarungsatian, C., 2011. Implementation of acoustic emission source recognition for corrosion severity prediction. Corrosion 67 (5), 11. Qian, W., Xie, L., Huang, D., Yin, X., 2009. Pattern recognition of fatigue damage acoustic emission signal. In: IEEE International Conference: Mechatronics and Automation (IEEEICMA), Changchun, China, pp. 4371e4375. Sause, M.G.R., Gribov, A., Unwin, A.R., Horn, S., 2012. Pattern recognition approach to identify natural clusters of acoustic emission signals. Pattern Recognition Letters 33 (1), 17e23. Scruby, C.B., Baldwin, G.R., Stacey, K.A., 1985. Characterization of fatigue crack extension by quantitative acoustic emission. International Journal of Fracture 28 (4), 201e222. Shigeishi, M., Ohtsu, M., 1992. A sigma analysis of the 2-dimensional PMMA model. In: Progress in Acoustic Emission VI (JSNDI), Proceedings of the 11th International AE Symposium, Fukuoka, Japan, pp. 211e217.

282

Corrosion of Steel in Concrete Structures

Shokri, T., Nanni, A., 2013. Crack source location by acoustic emission monitoring method in RC strips during in-situ load test. Smart Structures and Systems (in press). Wadley, H.N.G., Scrubby, D.B., 1983. Elastic wave radiation from cleavage crack extension. International Journal of Fracture 23 (2), 111e128. Weng, M.S., Dunn, S.E., Hartt, W.H., Brown, R.P., 1982. Application of acoustic emission to detection of reinforcing steel corrosion in concrete. Corrosion 38 (1), 9e14. Yoon, D., Weiss, W., Shah, S., 2000. Assessing damage in corroded reinforced concrete using acoustic emission. Journal of Engineering Mechanics 126 (3), 273e283. Yuyama, S., 2005. Acoustic emission for fracture studies using moment tensor analysis. The Journal of Strain Analysis for Engineering Design 40 (1), 33e44. Zdunek, A.D., Prine, D.W., Li, Z., Landis, E., Shah, S., 1995. Early detection of steel rebar corrosion by acoustic emission monitoring. In: Paper No. 547 Presented at CORROSION95, the NACE International Annual Conference and Corrosion, p. 9. Ziehl, P., 2008. Applications of acoustic emission evaluation for civil infrastructure. In: SPIE Proc., SPIE Smart Structures NDE, San Diego, CA, p. 9. Ziehl, P., Yu, L., ElBatanouny, M., Larosche, A., 2012. Sensor system phase two test report: insitu decommissioning sensor network, meso-scale test bed. In: Prepared for DOE/SRNL, Final Report e Phase II, 25 pp.

Practical field implementation of corrosion measurement methods

14

Andrew Fahim 1 and Pouria Ghods 2 1 Director of Research and Development, Giatec Scientific Inc., Ottawa, ON, Canada; 2Civil and Environmental Engineering, Carleton University, Ottawa, ON, Canada

14.1

Introduction

It has been previously illustrated (Chapter 12) that corrosion of reinforcing steel is an electrochemical process that can be detected, analyzed, and understood by utilizing measurements of half-cell potential (Ecorr) and corrosion current density (icorr). Measuring these parameters have been essential components in virtually every modern corrosion inspection. These parameters provide a reliable electrochemical method to accurately detect corrosion onset as well as quantify the corrosion propagation rate. This chapter will deal with methods and devices that are typically used in the field to measure these parameters and how they can be interpreted to detect corrosion. The focus of the chapter is on the practical implementation of these methods and best-practices to ensure a successful field utilization. Some of the theoretical background of the method is presented herein as understanding the theory behind these methods is essential in successfully implementing the methods and interpreting the collected results.

14.2

Half-cell potential

The half-cell potential monitoring technique is the most widely used qualitative technique to detect corrosion of reinforcing steel. The method was standardized by ASTM in 1980 as C876 “Standard Test Method for Half-cell Potentials of Uncoated Reinforcing Steel in Concrete” and is to this date the only standardized method for corrosion detection in the field. The method relies on measuring the electrochemical potential of the steel with respect to a standard reference electrode and using these potentials to understand corrosion activity. When an electrode (steel rebar or any type of embedded metal in our case) is immersed in an electrolyte (the concrete pore solution in our case), a potential is set up across the electrode/electrolyte interface that is correlated with the corrosion activity of the electrode. Measuring this potential therefore allows a qualitative estimation of the corrosion activity. Simply put, there is no direct way of measuring this potential difference between the electrode and the electrolyte; since attempting to connect one terminal of a voltmeter to the electrolyte and another to the electrode will lead to the development of another half-cell potential caused by the contact of the voltmeter’s Corrosion of Steel in Concrete Structures. https://doi.org/10.1016/B978-0-12-821840-2.00006-7 Copyright © 2023 Elsevier Ltd. All rights reserved.

284

Corrosion of Steel in Concrete Structures

terminal with the electrolyte. Therefore, it is necessary to introduce another “standard” half-cell with a known potential in order to determine the half-cell potential of the system under consideration. If two such cells are connected, the potential of the electrode/ electrolyte system under consideration can be calculated by measuring the potential difference between these two half-cells using a voltmeter. This half-cell potential can be used to understand the corrosion activity of the system under consideration. Since numerous half-cells can be used for this purpose, the introduction of a halfcell that can act as a datum/reference to measure all potentials becomes necessary. This cell was universally chosen to be the standard hydrogen electrode (SHE). This cell is composed of an inert platinum electrode in a molar solution of hydrogen ions at the unit activity (i.e., 1.0 mol Hþ ions in 1.0 L of water) saturated with hydrogen gas (H2) at a temperature of 25 C (Bertolini, 2013). The SHE merely acts as an electrode of reference potential and is not used for physical measurements for systems such as reinforced concrete. Two frequently used half-cells in concrete research and practice are silver-silver chloride (SSC), and copper-copper sulfate (CCS). These half-cells have an experimentally known potential against the SHE, have an excellent ruggedness and size for field applications, have good stability, and come at a reasonable cost. When connecting these half-cells to a reinforcement, through a hole in the concrete, and observing the potential difference, through a voltmeter, the corrosion potential of the reinforcement can be found. Fig. 14.1 shows a typical half-cell potential measurement system, which uses a reference electrode and a voltmeter, with its ends connected to both the steel reinforcement and the reference electrode.

Figure 14.1 A schematic of a half-cell potential measurement.

Practical field implementation of corrosion measurement methods

285

The observed potential difference will be a function of the metal’s tendency to dissolve, among other factors such as oxygen availability. By convention, the positive terminal of the voltmeter is connected to the steel and the negative terminal to the halfcell. This gives an increasingly negative reading for an increasing corrosion activity. If the cell is moved along the steel, different potentials will be observed because the reinforcement is in different environments. At the anode, it can readily dissolve, therefore more negative potentials are observed. While at the cathode, steel resists dissolution, due to the condition of passivation, providing more positive potential. ASTM C876 presents an empirical way to evaluate and classify the corrosion activity within a system using the obtained half-cell potential measurements versus the copper/copper sulfate half-cell. This is shown in Table 14.1. Since other half-cells have a direct offset from the copper/copper sulfate half-cell, these half-cells can be used along with the C876 recommendation directly. The guidelines for other common half-cells are shown in Table 14.1. Despite its common use, various limitations exist for the half-cell potential method. First and foremost, it is a qualitative, not a quantitative, method. Although the technique is useful in determining if the metal is undergoing active corrosion or is maintaining passivity, there is no general relationship that can be used to extract corrosion rates from half-cell potential measurements. This is since both parameters respond differently to variables such as moisture, oxygen availability, temperature, and concrete resistivity (Andrade and Alonso, 2004). In theory, the difference in half-cell potential between anodic and cathodic sites can be used to calculate the corrosion rate. However, this is not attainable in field conditions since it is difficult to isolate potentials measured from anodic and cathodic sites, especially for cases with microcell corrosion. One of the most important factors influencing the half-cell potential is the availability of oxygen. When concrete is near saturation, the results obtained tend to be more negative than the real potential. This might lead to wrongful assessment where corrosion is being estimated to be present while in reality the reinforcement is passive but has low oxygen availability in its vicinity. It should be also noted that the absence of Table 14.1 Corrosion condition evaluation criteria based on the measured half-cell potential. Copper/ Copper sulfate

Silver/ Silver chloride

Mercury/Mercury chloride (calomel)

Standard hydrogen electrode

> 200

> 106

> 126

> þ116

200 to 350

106 to 256

126 to 276

þ116 to 34

350 to 500 < 500

256 to 406 < 406

276 to 426

34 to 184

< 426

< 184

Corrosion condition Low (90% risk of corrosion) Severe corrosion

286

Corrosion of Steel in Concrete Structures

Table 14.2 The effect of concrete condition on the measured half-cell potentials. Condition of concrete

Typical ranges of potentials

Water-saturated concrete without oxygen Wet, chloride-contaminated concrete Humid, chloride-free concrete Humid, carbonated concrete Dry, carbonated concrete Dry concrete

900 400 þ100 þ100 þ200 þ200

to 1000 mV to 600 mV to 200 mV to 400 mV to 0 mV to 0 mV

Adapted from Elsener et al. (2003).

oxygen, in saturated conditions, tends to slow down the actual corrosion rate to negligible levels, which opposes what is seen through half-cell potential (Broomfield, 1997). This can be seen in Table 14.2, which shows experimental potentials obtained for a large number of structures at different conditions (Elsener et al., 2003). It can be clearly seen that passivity and activity can cover a wide range of overlapping potentials depending on the oxygen content and the resistivity of the concrete, which opposes the values outlined in C876. It should also be noted that such a method cannot detect localized corrosion and that the potentials measured at the surface of the concrete are not the potentials at the steel’s surface directly below the point where the half-cell is located. There is a dilution of potential that is highly affected by the cover depth, resistivity and anode-to-cathode ratio. Other factors influencing the overall electrical resistance between the half-cell and the rebar will influence this dilution such as cases where coatings and sealers are used. Results by Pour-Ghaz (2007), which were obtained through a finite element model in which potential maps were analyzed at the concrete surface and compared to that at the steel surface, are shown in Figs. 14.2 and 14.3. The effect of resistivity on the measured half-cell potentials, shown in Fig. 14.2, indicates that anodic and cathodic surfaces can only be differentiated in cases of resistivity much higher than those associated with chloride-laden concrete. For low resistivities, the potential measured at the concrete surface tends to become similar to the average overall potential of the zone surrounding the measurement. The effect of cover depth, shown in Fig. 14.3, indicates this distinction is also not possible for large cover depths, which supports the recommendations in ASTM C876 on the influence of cover depth. At such high cover depths, the measurement is an average of potentials over an area much larger than the point under consideration. The half-cell potential technique is generally not suitable for conducting measurements on epoxy-coated or galvanized reinforcements. Epoxy-coated steel (in cases with no significant damage to the epoxy) is electrically isolated from other bars and the concrete; subsequently, isolated from the reference electrode and does not allow a direct half-cell potential measurement. Stable measurements can only be taken in cases with significant epoxy-coating defects or damage (significant bare-metal

Practical field implementation of corrosion measurement methods

287

Figure 14.2 Effect of resistivity (A: 50 U m; B: 5000 U m; C: 10,000 U m) on half-cell potentials measured at the concrete surface and the steel-concrete interface (Pour-Ghaz, 2007).

exposure), provided that an electrical connection to the bare metal is done. For galvanized rebar, the galvanic metal (commonly zinc) protects the steel reinforcement. In these cases, the measured half-cell potential reading is not truly the potential of the steel reinforcement under consideration but is the mixed potential of steel and zinc (which is significantly more negative than the values for steel rebar). Therefore, interpreting these values in the context provided by C876 is irrelevant. Measurement of potential for galvanized rebar has been used extensively for research purposes on the efficacy of galvanized rebar. Different criteria need to be applied to these reinforcements based on a knowledge of the galvanized rebar characteristics (Gu and Beaudoin, 1998). In cases where cathodic protection systems are in use, traditional half-cell measurements do not provide useful information when the system is active/cathodically polarized. These measurements only quantify the extent to which the rebar is cathodically polarized by the protection system and not the corrosion activity. In these cases, halfcell potential measurements are typically done after the cathodic protection system is shut down and at least a 24-hour period of de-polarization is allowed. Corrosion inhibitors also have a complex influence on the half-cell potential since some of these chemicals modify the corrosion potential of the steel reinforcement

288

Corrosion of Steel in Concrete Structures

Figure 14.3 Effect of cover depth on measured half-cell potential for a cover depth of (A) 20 mm and (B) 140 mm (Pour-Ghaz, 2007).

positively or negatively, as a means of cathodic or anodic protection. An anodic inhibitor is simply a strong oxidizing agent, with calcium nitrite being by far the most common type of anodic inhibitor. Calcium nitrite use will shift the corrosion potential to a more positive value and reduce the corrosion rate. A half-cell potential reading obtained from a reinforced concrete structure containing calcium nitrite would normally be more positive than one without it. A cathodic inhibitor acts in the opposite manner. The effects of the corrosion inhibitor must be considered when interpreting the halfcell potential readings and the inhibitor supplier should be asked to clarify how the mixed corrosion inhibitor affects the corrosion potential of the steel before any interpretation of the results is done (Gu and Beaudoin, 1998). Most concrete coatings have extremely high electrical resistivity. In most cases, these coatings can be considered as an isolating material. As a result, the half-cell potential technique cannot provide meaningful data when the concrete surface is coated with an intact coating. Even though reading can be made in cases with significant coating damage, or when the coating is removed, the coating may have had an effect on the oxygen concentration in concrete and around the reinforcement; therefore, the readings can be unrealistically negative similar to what is shown in Table 14.2 (Gu and Beaudoin 1998).

Practical field implementation of corrosion measurement methods

289

Finally, a very common use case for the half-cell potential method is analyzing patch repairs. Patch repairs are known to cause macrocells (known as the anodic ring/halo effect) due to the differences in the environment within the patch and outside the patch and correspondingly having different potentials (Elsener, 2002, Pruckner and Gjrv, 2002). Therefore, in patch repairs, generally, the macrocell corrosion can be considered as a governing mechanism of reinforcement corrosion. The measurement of half-cell potential in cases of macrocell corrosion significantly depends on the location of the probe with respect to the interface of new and old concrete as well as the concrete resistivity (Elsener et al., 2003). Measurements, when patch repairs are used, should be interpreted putting into consideration the influence of the patch repair on half-cell potentials both within and in the periphery of the patch.

14.3

Linear polarization resistance

In Chapter 12, the Stern and Geary equation has been introduced as a method to determine the corrosion rate electrochemically. This section discusses the theoretical background of this equation, how it is derived, the physical meaning of its parameters, and how it has been used for decades to find corrosion rates in the field. As electrons flow from an anodic to a cathodic site, due to the difference in their equilibrium potentials (anodes having a more negative potential than cathodes), this flow induces a current that is directly proportional to the electrons being produced. The amount of metal that is corroding is directly proportional to the anodic current that is being passed by this reaction, and therefore the number of electrons that are released. These electrons are consumed at the cathodic site, during the reaction with oxygen and water to form hydroxyl ions. This means that the cathodic and anodic currents must equal each other, leading to an overall net current of zero. This current is the main interest of a practitioner when trying to classify the corrosion state of reinforcement as it is an indicator of the rate of flow of electrons which itself is an indication of the rate of the corrosion of the reinforcement. As the overall observed current in this system is zero, there is no direct way of measuring these currents, especially since the anodic and cathodic sites are not usually clearly defined. The linear polarization resistance (LPR) method overcomes this limitation by applying a change in potential and measuring a change in current to find the corrosion rate. The relationship between the change in potential from the equilibrium (half-cell) potential and the corresponding current is described by the Butler-Volmer equation, which is shown in Eq. (14.1) (Frankel, 2016).
i ¼ i0

   nF nF h  exp  ð1  bÞ h exp b RT RT 

(14.1)

Where io is the exchange current density (a constant for the system that is a factor of the chemical activation energy, the temperature, the concentration of oxidizing and reducing agents, and the material), h is the overpotential (the change in potential from

290

Corrosion of Steel in Concrete Structures

the equilibrium half-cell potential), b is the symmetry coefficient (which is usually taken as 0.5, indicating that the change in potential influences the oxidation and reduction reactions equally), n is the valency of the metal, F is Faraday’s constant (96,485 C/mol), R is the gas constant (8.314 J/molK), and T is the absolute tempernF h ] represents the ature (in K). It should be noted that the positive term [exp b RT    nF h anodic portion of the reaction, while the negative term, exp  ½1 b RT rep-

resents the cathodic portion (Roberge, 2008). The goal here is to measure io by measuring both i and h after a polarization and fitting the observed data using the Butler-Volmer equation to determine io (it will be shown below that this equation can be simplified significantly). When plotting the Butler-Volmer equation of overpotential versus current, it produces a curve such as that shown in Fig. 14.4. Where in the left portion, the cathodic current increases with the increase in negative potential (decrease in potential), while the anodic current increases with the increase in potential and in the middle portion a linear relationship can be seen. It should be noted that the two sides of the figure plotted are not the anodic and cathodic reactions of the corrosion process, but the anodic (e.g., Fe / Feþ2 þ 2e) and cathodic (e.g., 2e þ Fe2þ / Fe) portions of one idealized reaction; such as the metal’s dissolution/deposition. The same curve can also be drawn for oxygen reduction. In order to represent this in a simpler manner, the constants in the equations, such as Faraday’s number, the gas constant, the valency, the temperature, and the b coefficient are all gathered producing two constants ba, and bc, the so-called anodic and cathodic Tafel constants, shown in Eqs. (14.2) and (14.3).

Figure 14.4 Typical Butler-Volmer curve.

Practical field implementation of corrosion measurement methods

291

ba ¼

2:303RT bnF

(14.2)

bc ¼

2:303RT ð1  bÞnF

(14.3)

The use of these constants leads to a more convenient representation of the ButlerVolmer equation, as shown in Eq. (14.4).  h h i ¼ io 10ba  10bc

(14.4)

Provided that h is small in this equation (in the range of 10e30 mV), an approximation can be done to this equation as shown in Eq. (14.5) if io is renamed to the corrosion current (Icorr) and i is renamed to the measured current during polarization (Note h that 10 b ¼ 1  2:3 hb ). Icorr ¼

imeas ba bc h 2:3ðba þ bc Þ

(14.5)

If all of the constants are grouped into the constant B, termed the proportionality ba bc constant, where B ¼ 2:3ðb and the rate of change of voltage to change in current a þbc Þ

h is termed Rp, the polarization resistance, where Rp ¼ imeas ¼ DV DI , the Stern and Geary equation, Eq. (14.6), is derived (Stern and Geary, 1957).

Icorr ¼

B Rp

(14.6)

Devices that use the linear polarization resistance technique rely on the fact that by polarizing the electrode, through applying a certain polarizing current, a change in potential will results, which is used to determine the polarization resistance that is inputted in the Stern and Geary equation. Such techniques typically use what is known as a three-electrode system, including a reference electrode, a counter electrode, and a working electrode. The reference electrode is similar to the half-cells discussed previously. The counter electrode is the electrode that applies the potential or current perturbation necessary for polarizing the reinforcement. The working electrode is the steel reinforcement. The final equation used by all of the devices that rely on LPR is shown in Eq. (14.7), where the Stern and Geary equation is normalized by the area polarized by the polarizing current (A) in order to present a corrosion current density (note than lower case icorr is used to denote the corrosion current density which is the areanormalized value of Icorr, the corrosion current). A schematic of this system is shown in Fig. 14.5. icorr ¼

B A Rp

(14.7)

292

Corrosion of Steel in Concrete Structures

Figure 14.5 A schematic of modern LPR measurements showing the three primary components of the measurement, connected using a hand-held device along with the polarized area.

A value for B of 26 mV is typically used in the case of an actively corroding system. If the steel is considered to be passive, a Stern-Geary coefficient of 52 mV is used. This is based on the findings of Andrade and Gonzalez (1978) in which they found that satisfactory results were obtained with these coefficients. These values also have theoretical merit (Eqs. 14.2 and 14.3 will result in a value of ba ¼ bc ¼ 0.12 mV for actively corroding systems and ba ¼N, due to passivation control, and bc ¼ 0.12 mV for passive reinforcements, which deduces the same proportionality constant values) and can be calculated from the theoretical values of the inputs. The same coefficients have been used for decades to accurately measure corrosion rate. Since it is not always known whether the steel is in a passive or active state of corrosion, an accepted way is to assume a coefficient of 26 mV for all cases, since assuming a lower B coefficient for a passive case in which icorr is already small, will not affect the classification of the reinforcement, and will still maintain a correct classification (Millard and Broomfield, 2002). The obtained corrosion current density, icorr, is not only an indication of corrosion activity but is also a direct way to calculate the loss of reinforcement section due to corrosion. This is done using Faraday’s law as shown in Eq. (14.8). m¼

MIT zF

(14.8)

where m is the mass of steel consumed by corrosion (g), I is the corrosion current (A), t is time (s), F is Faraday’s constant (96,500 A s) z is the ionic charge (2 for the case of

Practical field implementation of corrosion measurement methods

293

steel) and M is the molar mass of the metal (56g for Fe). Table 2.2 shows the classification of corrosion activity according to the corrosion current density. Different corrosion classifications were used in the past by different authors, but they generally fall in the same range shown.

14.4

Galvanodynamic and potentiodynamic techniques

The galvanodynamic and potentiodynamic techniques are the most direct methods to measure the polarization resistance. The potentiodynamic method was standardized by ASTM in 1997 as G57 “Standard Test Method for Conducting Potentiodynamic Polarization Resistance Measurements” and is more commonly used for reinforcing steel evaluation than the galvanodynamic method. In this method, a small potential scan DE is applied to the reinforcement, with respect to the half-cell potential (Ecorr); such that DE ¼ E  Ecorr, and the resultant current is recorded. The polarization resistance is therefore defined according to Eq. (14.9).  Rp ¼

vDE vi

 (14.9) i¼0;

dE /0 dt

Typically, this test method is performed using a potentiostat capable of varying potential at a constant scan rate and measuring the current. After the three-electrode system is connected to the reinforcement, the half-cell potential is recorded for a period of time in order to ensure stability (typically less than 0.5 mV of change in 5 s) before the sweep is performed. After ensuring stability, a potential that is 10e20 mV more negative than the half-cell potential is applied and then the potential is swept to more negative values with a certain sweep rate that ensures linearity of the voltage-current relationship (sweep rates discussed later), which is typically in the range of 0.04e0.167 mV/s; the latter being the recommendation of ASTM G59 until reaching 10e20 mV more positive than the half-cell potential. The obtained voltage and current are fitted linearly, and the deduced slope is the polarization resistance. Fig. 14.6 shows an example of data obtained from a potentiodynamic sweep showing a polarization resistance of 268.48 U. It is worth mentioning that Eq. (14.9) only applies for systems where the polarization resistance is the dominant resistance present, or in other words, the resistance of the electrolyte (in our case the concrete cover) is low/negligible compared to that of the polarization resistance. In reality, reinforced concrete has a very high resistance between the working electrode and the reference/counter electrode caused by the concrete cover. In these cases, it was found that the reinforced concrete system can be electrically represented using the Randles circuit shown in Fig. 14.7, where RO represents the ohmic resistance (concrete cover resistance in our case), Cdl is the doublelayer capacitance at the rebar/concrete interface and Rp represents the polarization resistance. This circuit is used by most electrochemical techniques to represent the electrical response of the system when undergoing polarization.

294

Corrosion of Steel in Concrete Structures

Figure 14.6 Example of a potentiodynamic test result.

Figure 14.7 Randles circuit.

Finding that the reinforced concrete system is represented as a Randles circuit has two primary consequences. The first one is that the change in voltage versus the change in current does not represent only the polarization resistance but represents the summation of both polarization resistance and ohmic resistance. Therefore, Eq. (14.9) is modified as shown in Eq. (14.10). 

 v DE Rp ¼  RU vi

(14.10)

This means that the ohmic resistance needs to be determined and used to correct the slope of the potential-current relationship. This can be done in many ways; the three most common methods are: (a) Current-interrupt methods: where the instantaneous change in voltage caused by interrupting the polarizing current for short periods of time (in the millisecond range) is found. In these interrupt periods, the potential drop due to ohmic resistance disappears immediately

Practical field implementation of corrosion measurement methods

295

while the electrode polarization only decreases slowly due to the (charged) double-layer capacitance, allowing a calculation of the ohmic resistance through this instantaneous change in voltage. (b) determining the resistance of the system at high frequencies (usually in excess of 10 KHz) where the double-layer capacitance acts as a short-circuiting element (impedance of a capacitor is inversely proportional to frequency) and the overall measured impedance corresponds to the magnitude of the ohmic resistance. (c) Positive feedback method: where the current applied to the working electrode is measured by a current follower with an output potential directly proportional to the applied current; this can be a simple shunt-resistor over which the potential drop is measured. This potential is fed back to the input of the potentiostat and added to the applied voltage to compensate for the ohmic resistance

Devices utilizing the potentiodynamic method must apply one of these three methods to correct the measured resistance to the electrolyte resistance. If these methods are not applied, and the polarization resistance is determined directly using the change in the potential to change in the current ratio, significant errors can arise. As shown in Eq. (14.10), these errors increase as the concrete resistivity or the cover depth increase (both increase RO). For instance, in dry/high-resistivity concrete, the polarization resistance can be overestimated if RO is not accounted for. This is especially true for actively corroding systems where Rp is low and could be within the same order of magnitude of RO. Discussion of these methods and their advantages and limitation can be found in (Elsener and Bohni, 1983; Andrade et al., 1984; Mansfeld et al., 1990; Oelssner et al., 2006; Lambie et al., 2007). The other primary consequence of representing the system as a Randles’ circuit is the availability of the double-layer capacitance. This means that the system has a specific time constant (time taken for the capacitance to charge/discharge) which governs the time for the voltage to reach a steady-state at every point of the sweep. Due to this, the potential sweep rate needs to be defined and needs to be low enough to ensure a steady-state is achieved (in other words the capacitive effects need to be exhausted in order to determine the potential shift solely related to the Faradaic processes of polarization). Sweep rates in the range of 0.04e0.167 mV/s were found to be acceptable for these purposes. A complete discussion of this and the mathematical derivation of the effect of the sweep rate can be found in (Gonzalez et al., 1985). The galvanodynamic method is very similar to the potentiodynamic method, although, as the name entails, in the galvanodynamic method the current is controlled and not the potential shift. The primary reason behind this method being used less frequently than the potentiodynamic method is related to the difficulty of controlling the current change to ensure that small potential shifts are achieved. As mentioned earlier, ensuring small potential shifts is essential in ensuring the linearity of the potential-current relationship as necessary in the derivation of the Stern-Geary equation.

14.5

Galvanostatic and potentiostatic techniques

The galvanostatic pulse technique was introduced for field application in 1988. It is a rapid technique used mainly for on-site monitoring. In this technique, a current pulse is

296

Corrosion of Steel in Concrete Structures

imposed on the reinforcement from a counter electrode placed on the concrete surface. The applied current is usually in the range of 10e200 mA and the typical pulse duration is in the range of 10e100 s. The reinforcement is typically polarized in an anodic (more positive) direction compared to its free corrosion potential. The resulting change in the electrochemical potential of the reinforcement is recorded via a reference electrode mounted on top of the concrete cover over the reinforcement under consideration. Using the recorded voltage change and the known current applied, the polarization resistance could be calculated. A typical potential transient response obtained after the application of a galvanostatic pulse is shown in Fig. 14.8. In the galvanostatic technique, it is again assumed that a simple Randles circuit describes the potential response, Vt, of the steel-concrete system as a function of time when a galvanostatic current, Iapp, is applied. Under this assumption, the potential response, Vt, as a function of the polarisation time, tp, can be expressed by Eq. (14.11) (Elsener, 2005), which is the response of the idealized Randles circuit.
VðtÞ ¼ Iapp





t Rp 1  exp Rp Cdl



 þ RU

(14.11)

where Iapp is the applied polarization current, t is the time from the application of current, Cdl is the capacitance of the steel-concrete interface and RO is the Ohmic system resistance previously described. Fitting the observed potential transient to Eq. (14.11), using an exponential curve fitting procedure, Cdl, Rp and RO can be found and Rp is used in the Stern and Geary equation to determine the corrosion rate. Fig. 14.9 shows a schematic of the GalvaPulse device, developed by FORCE technology and Germann Instruments, which utilizes the galvanostatic method. The device uses an Ag/AgCl reference electrode mounted at the center of the counter electrode. Two circular zinc electrodes are used in the device electrode assembly: a counterelectrode and an auxiliary guard ring with outer/inner diameters of 60/30 mm and

Figure 14.8 Typical galvanostatic polarization plot.

Practical field implementation of corrosion measurement methods

297

Figure 14.9 Galvapulse device schematic.

100/86 mm, respectively. The current applied from the guard ring is hypothesized to repel the lines of current from the central counter electrode, confining them to an area of the structure located approximately under the counter electrode (Germann, 2004). This is deemed essential in order to control and accurately determine the polarized area used in the Stern-Geary equation. In this setup, a reinforcement length of 70 mm is assumed to be the polarized length. A pulse time of 10 s and a current of 20e100 mA are typically used. Three primary issues affecting accuracy have been extensively discussed in the literature related to the use of this device. First, since the potential shift is not controlled, in many cases the application of a pre-determined empirical polarizing current leads to potential shifts that are far beyond the range of linearity typically thought of as 10e30 mV. This is especially true for passive reinforcements, with high polarization resistance, since the potential shift is the product of polarization resistance and the applied current. This can make the use of the Stern and Geary equation, which is based on the assumption of small potential changes, questionable. When this device is used, it is recommended that the user attempts a polarization at low values of applied current and monitors the total potential shift to evaluate if the potential shift is within the range of 10e30 mV. If higher shifts are found, a lower applied current should be used. Based on these trials, the applied current can be determined. If this method is used, it is essential to wait a sufficient time between tests to ensure de-polarization. Secondly, short waiting times (less than 30 s) can cause significant errors such that the polarization resistance can be significantly underestimated, and the corrosion rate is overestimated, when using short wait times. The waiting times recommended for this method are typically in the range of 30 s for corroding rebar to 100 s for passive rebars; considering that passive rebars have a much larger time constant (the time constant is the product of polarization resistance and double-layer capacitance; refer to Eq. 14.11). This device has typically been found to overestimate corrosion rates for passive reinforcements significantly (Martinez et al., 2010; Fahim et al., 2018) when short waiting times are used.

298

Corrosion of Steel in Concrete Structures

Finally, perhaps the most discussed topic is the use of guard-ring electrodes utilized to confine the polarizing current. Outside of the laboratory, in actual structural elements, the counter electrode is much smaller in size than the working electrode (virtually infinite in field conditions). Therefore, ensuring that the polarizing current is confined directly below the counter electrode is important to ensure the accuracy of the polarized area value used. In the confinement method used by this device, the current applied by the guard ring is controlled through an initial measurement of the potential of the counter-electrode and guard ring versus the reinforcement. The current applied from the guard ring is then estimated in order to ensure that the potential difference between the counter and guard electrodes remains the same as that before the polarization. This confinement method has been found to perform poorly, especially for passive reinforcements and localized corrosion cases (Nygaard et al., 2009), with the true polarized areas being very different from those assumed by the device. For passive reinforcements, Nygaard et al. (2009) reported a polarized area upwards of 560 mm, which is far from the assumed 70 mm. For localized corrosion, a phenomenon termed self-confinement has been reported, where the current flows to the path of least resistance (the corroded areas) even in cases where the guard ring is distant from these areas (Nygaard et al., 2009; Andrade and Alonso, 2004) leading to a measurement that is not representative of the area under consideration. Fig. 14.10 shows the GECOR device, developed by Geocisa, which also incorporates the galvanostatic method. The device’s measurement probe consists of stainless-steel electrode rings with a counter-electrode and a guard ring with outer/inner diameters of 70/11 mm and 180/140 mm, respectively. A Cu/CuSO4 reference electrode is used in the center of the counter electrode. The device also uses the guard ring method where two auxiliary Cu/CuSO4 reference electrodes are used for controlling the guard ring are positioned between the counter-electrode and the guard ring at a distance of 45 and 60 mm from the center of the probe. This device overcomes two of the major issues associated with the GalvaPulse. Instead of applying a constant pre-set polarizing current, this device starts by applying a short current pulse from the counter electrode and measures the potential response. From the measured response, the instrument calculates an assumed “optimal” polarizing current, to maintain the potential shift in the range of

Figure 14.10 Photograph of the GECOR device components (James Instruments, 2020).

Practical field implementation of corrosion measurement methods

299

linearity of the polarization. Although this is a more theoretically sound method to control the potential shift, work by Nygaard et al. (2009) have indicated that potential shifts in excess of 100 mV were recorded during testing. The other differentiation is that the device uses much longer polarization periods (100 s as default) to ensure a steady-state is achieved thereby accurately determining the potential shift associated with the polarization resistance, even for passive reinforcements. The confinement method used by the device is termed the modulated guard-ring. This method uses twin reference electrodes, S1 and S2, located between the counter electrode and the guard ring in order to achieve the required counterbalancing electrical field. A schematic representation of this method is shown in Fig. 14.11. The potential difference between S1 and S2 is continuously measured to permanently control the electrical field between these two electrodes to equilibrate internal and external currents. Therefore, the current applied by the guard ring changes during the measurement depending on the potential measured between the two ring controllers in order to ensure the potential difference between S1 and S2 remains near zero during the measurement. This method has been found to perform more successfully in terms of confinement for passive reinforcements when compared to the non-modulated method but has been found to have the same limitation of self-confinement for localized corrosion (Nygaard et al., 2009). Both methods have been extensively compared by Nygaard et al. (2009) showing the associated differences and comparative confinement success. The potentiostatic technique is similar to the galvanostatic technique, except that a constant potential perturbation of 10e20 mV is applied instead of the polarizing current. In this setup, the measured current corresponds to the Randles circuit, which is fitted to the current-time relationship corresponding to Eq. (14.12)

Figure 14.11 Theoretical confinement of the applied current using the modulated confinement technique (Modified from Andrade and Alonso, 2004).

300

Corrosion of Steel in Concrete Structures

2 It ¼

RU



0

13

6 B t C7 DE C7 6 RU þ Rp expB 4 @Cdl RU Rp A5 RU þ Rp RU Rp

(14.12)

where it has a steady-state component: IN ¼

DE  RU þ Rp

(14.13)

and a transitionary component: 0 Itr ¼

1

B t C DERp C

 expB @Cdl RU Rp A RU RU þ Rp RU Rp

(14.14)

with a time constant: s¼

Cdl RU Rp RU Rp

(14.15)

If the ohmic resistance is assumed to be lower than the polarization resistance, which is typically the case for passive reinforcements and for some cases with active corrosion for chloride-laden concrete, the time constant reduces to Eq. (14.16). s ¼ Cdl RU

(14.16)

Eq. (14.16) indicates the primary advantage of the potentiostatic method over the galvanostatic one. Clearly, the time constant for potentiostatic measurements can be lower than that for galvanostatic measurements (CdlRp). This has been also proven experimentally and values in the range of 15e60 s have been found to be sufficient for this method for active and passive reinforcements, respectively (Gonzalez et al., 1985).

14.6

Coulostatic technique

The coulostatic technique involves the application of a known amount of charge, and the observation of the potential discharge following the charge application. The potential transient obtained after the application of a quantity of charge follows that shown in Eq. (14.17) (Hassanein et al., 1998) ht ¼ ho exp

  t sc

(14.17)

Practical field implementation of corrosion measurement methods

301

Figure 14.12 Typical coulostatic transient for the case of passive and active rebars.

where ht is the potential shift at any time t, ho is the initial potential shift and sc is the coulostatic time constant (sc ¼ CdlRp). An example of these transients obtained for the passive and active conditions is shown in Fig. 14.12. The theoretical transient may be fitted to the measured data by adjusting ho and sc to minimize the squared differences between the predicted and measured data. This step, however, does not result in the determination of the double-layer capacitance or the polarization resistance explicitly, but their multiplication product (sc). When an amount of charge is applied to the Randles circuit, the double-layer capacitance becomes instantaneously charged. In this case, the potential shift measured at time zero corresponds to the potential shift caused by the charging of the capacitor. Therefore, the double-layer capacitance can be directly found from Eq. (14.16), using the initial potential shift, ho, which is obtained from curve fitting as shown in Eq. (14.15); if it’s assumed that the electrolyte capacitance is negligible. Finding this capacitance then allows the determination of the polarization resistance (Rp ¼ sc/Cdl) through the time constant obtained from Eq. (14.18) Cdl ¼

qs ho

(14.18)

where qs is the applied quantity of charge. A marked difference between coulostatic transient and the galvanostatic one is that the transient response equation for the coulostatic technique does not include an ohmic resistance term. This is due to the transient being an open-circuit discharge which is not expected to be influenced by any ohmic contribution. This is since the Randles circuit, which is assumed to model the steel-concrete system, assumes that no capacitive effects are imposed by the electrolyte, and therefore assumes that the electrolyte cannot be charged by the application of such a quantity of charge. A prototype device using this method has shown good accuracy compared to gravimetric measurements in Fahim et al. (2018).

302

14.7

Corrosion of Steel in Concrete Structures

Connectionless electrical pulse response analysis (CEPRA)

The techniques illustrated previously all require a direct connection between the counter electrode and the working electrode in order to determine the change in half-cell potential and subsequently determine the polarization resistance. The CEPRA technique is a recently developed technique that overcomes this limitation using a 4point (Wenner) probe setup, commonly used for surface resistivity measurements, to determine the corrosion rate. A schematic of this setup is shown in Fig. 14.13. As noted in Chapter 12 when discussing EIS, the dependence of the impedance caused by the double-layer capacitance on the applied current’s frequency can be used to determine the system’s electrochemical characteristics through performing frequency sweeps, where at high frequencies, the short-circuiting effect caused by the double-layer capacitance can directly yield the ohmic resistance and at the lowfrequency limit, the double-layer capacitance acts as an open circuit and the summation of the ohmic resistance and polarization resistance can be found. The same concept can be seen in the time-domain in the galvanostatic method for instance, where instantaneously after polarization (t ¼ 0), the ohmic resistance can be found directly as the double-layer capacitance acts as a short circuit and at a steady-state (t ¼N) the summation of polarization resistance and ohmic resistance can be found as the double-layer capacitance is fully charged and acts as an open circuit element. This simple representation of the steel-concrete system stems from the simple Randles circuit which has been used frequently to obtain corrosion rates for the steelconcrete system. In the connectionless technique, which utilizes the 4-pt probe setup, there is higher system complexity. If a current pulse or a step voltage is applied from one of the two outer probes of a 4-pt probe device, this current has two primary flow paths. One path is normal to the metallic electrode, which causes the charging of the

Figure 14.13 A wenner probe setup used on top of the rebar.

Practical field implementation of corrosion measurement methods

303

double-layer capacitance or the polarization of the electrode (depending on the frequency of the applied current) and another path that is parallel to the metallic electrode, in which the current applied by one electrode is consumed by the other similar to resistivity measurements. The portion of the current flowing in each of these paths is highly dependent on the applied current’s frequency, the concrete cover characteristics (cover depth and resistivity), the polarization resistance value and the double-layer capacitance. These are all interrelated factors that affect the current flow path and the obtained results. This system can be represented schematically using a more complex circuit model shown in Fig. 14.14 (Fahim et al., 2019). In this case, Rc1 represents the probes’ contact resistance and all of the current is faced by this resistance. This circuit model clearly identifies the two major current flow paths in the concrete medium through Rc2 and Rc3, where the magnitude of current passing by each of these resistors is dependent on: (1) the magnitude of Rc2 and Rc3 resistance; (2) the impedance caused by the capacitance or the extent of charging of this capacitance; (3) the magnitude of the polarization resistance; (4) the concrete cover depth and reinforcement diameter. If the applied current is swept from very high to very low values, this circuit can be solved using iterative methods. This is, however, a very time-consuming process similar to impedance spectroscopy and could require several hours. Alternatively, a narrow current pulse can be used similar to the galvanostatic method and the resulting data can be converted to the frequency domain, or solved in the time domain, to extract the same information. If a narrow current pulse or a step voltage is applied for a short period of time, the measured voltage response as a factor of time is similar to that of a charging RC circuit, shown in Eq. (14.19), assuming that the electrolyte capacitance is negligible (same assumption as that in all of the other monitoring techniques). 

Vin ðtÞ ¼ Vex : A  BeDt

(14.19)

where Vex is the constant voltage applied through the external electrodes and Vin is the potential difference between the two inner electrodes. The model shown in Fig. 14.14 was solved in order to determine the variables A, B and D. It was found that these

Figure 14.14 Circuit model used to represent the CEPRA technique.

304

Corrosion of Steel in Concrete Structures

variables follow functions shown in Eqs. (14.20), (14.21) and (14.22). By measuring the voltage response over time as shown in Eq. 14.19, A, B and D constants can be calculated by fitting an exponential curve to the measured data. These factors can then be used to calculate the circuit components as shown in Eqs. (14.20), (14.21) and (14.22). This solution approach is rather complicated compared to the Randles circuit used by all of the other techniques. However, such a circuit can be solved using more complicated iterative solution procedures. A ¼ f ðRc1 ; Rc2 ; Rc3 ; Rc4 Þ

(14.20)

B ¼ gðRc1 ; Rc2 ; Rc3 ; Rc4 Þ

(14.21)

D ¼ hðRc1 ; Rc2 ; Rc3 ; Rc4 ; Cdl Þ

(14.22)

The device used to implement this technique uses a four-point array similar to the conventional Wenner-probe, with an electrode spacing of 5 cm. The outer probes apply a narrow DC step voltage for a short period of time (3e30 s) and simultaneously record the voltage of the system with a relatively high sampling rate. This typically leads to applied currents in the range of 0.5e2 mA (this current is not the polarizing current since a large portion of the applied current flows explicitly between the two probes in the Rc2 path). The obtained transient is then fitted to Eq. (14.19) to yield the constants A, B and D, which can be used to calculate the system components shown in Fig. 14.14. Fig. 14.15 shows a photograph from a corrosion inspection done using the CEPRA technique without requiring a connection to the reinforcing steel. This technique has been compared to gravimetric measurements of corrosion rate (Fahim et al., 2019) and has been found to show an accuracy similar to that shown by other well-established techniques such as the galvanostatic method.

14.8

Practical aspects and case study

Knowledge of the electrochemical methods does not on its own guarantee the success of field inspections. In many cases, it is important to keep in mind the practical aspects

Figure 14.15 A photograph of the hand-held corrosion device utilizing the CEPRA method as well as an inspection being done using the device without requiring a rebar connection.

Practical field implementation of corrosion measurement methods

305

of these measurements. This section outlines the best practices related to the field implementation of these methods and devices as well as two case studies where these guidelines are used. This guidance is adapted from the recommendation provided by RILEM TC154-EMC: “Electrochemical Techniques for Measuring Metallic Corrosion” which has been used successfully for field applications for many years. The reader can refer to (Andrade and Alonso, 2004) for a more detailed guide.

14.8.1 Devices used Numerous commercial devices are available in the market to implement the techniques illustrated in this chapter. When the guidance provided in this chapter is taken into consideration (e.g., ohmic resistance correction, measurement time, polarized area considerations, sweep rates, etc.), all of these devices should show similar, if not identical, corrosion rates (Fahim et al., 2018). When these commercial devices are not used, any potentiostat or galvanostat can be used as a means of controlling and measuring potential and current. It is recommended that the potential measuring circuit of the instrument is able to maintain a voltage measurement within 1 mV of the true values over the range of currents expected. The current circuit should have a sensitivity such that the icorr could be determined at least of the order of 0.05 mA/cm2. The potential or the current is applied in one of two modes: (1) a potentiodynamic or galvanodynamic sweep and (2) a stepwise polarization using a single potential, current or charge steps, while ensuring that the potential shifts are small enough to maintain linearity of the current-potential relationship. Potentiodynamic or galvanodynamic measurements allow for the calculation of the slope DE/DI, which has to be multiplied by the polarized rebar area and used in the Stern-Geary equation. In these methods, it is important to use sweep rates that are in the range of 0.04e0.167 mV/s, with the lower range of values being preferable for passive rebars (Gonzalez et al., 1985). Potentiostatic, galvanostatic, coulostatic and CEPRA measurements allow a determination of transient current or potential which are fitted into the Randles circuit idealized equation to obtain the polarization resistance (or more complex circuits for EIS or CEPRA methods). For these methods, the optimum polarization time is of the utmost importance. In galvanostatic measurements, this is typically from 30 to 100 s for passive and corroding rebar. respectively. For potentiostatic measurements, a period of 15e60 s is typically recommended. The device used has to be equipped with a probe containing both the counter and references electrodes. The counter electrode is typically made of stainless steel as it is a cost-effective material for corrosive environments enabling easy maintainability; although it could be made of any other metal able to provide the polarizing current. The reference electrodes can be any type that provides the required stability, measurement range and ruggedness required for field application; with the primary ones being: silver/silver chloride, copper/copper sulfate or calomel. The reference electrodes have to be checked periodically and before inspections in order to detect any issues caused by drying of the porous plug or leakage of the solution. This is best done by comparing the measured corrosion potential at a specific location to those obtained using other

306

Corrosion of Steel in Concrete Structures

reference electrodes available at the same location. The manufacturer’s instructions for storage, calibration, and maintenance must be followed. Electrodes should not be allowed to dry out or become contaminated. The porous plug has to be covered when not in use for long periods to ensure that it does not dry out. The device used has to implement a method for the ohmic resistance calculation. For potentiodynamic or galvanodynamic methods this could be one of: current interrupt, positive feedback or high-frequency impedance measurements. Measurements that do not make use of ohmic resistance can cause an overestimation of polarization resistance (since the DV/DI will correspond to the summation of Rp and RO not solely Rp) and an underestimation of the corrosion rate. This is especially important for corroding systems with low Rp. For galvanostatic and potentiostatic measurements, the ohmic resistance is determined by fitting the obtained transient to the Randles circuit solution, although a high sampling rate is required around t ¼ 0 in order to accurately determine this value.

14.8.2

Preparation

The procedure of performing field inspections to detect corrosion includes several steps. Almost every such inspection will start by visually inspecting the area/structure and recording evidence such as rust staining, cracking, spalling, etc. A map showing the location, type and severity of such observed defects is an essential component of these inspections and will affect how electrochemical methods are utilized. More information on conducting a visual inspection can be found in ACI 201.1R “Guide for Conducting a Visual Inspection of Concrete in Service.” The environment of the structure, as well as the concrete condition (saturation level, exposure to marine environment or de-icing salt, humidity of the environment and temperature) need to also be recorded. Based on the findings from the visual inspection, the number of measurements and location of measurements need to be determined. The number of points will depend on the time available, access, size of the structure and observed defects, among other factors. Other previously taken measurements or collected evidence can also influence these data points. For example, if elements of varying structural importance are faced, the ones with the highest determined structural criticality are usually prioritized. Typically, the time required per measurement (as a rule of thumb, 5 min per measurement; but this will vary based on the device used) is used to determine the density of measurements taken. Following this, the determination of rebar characteristics is essential. An inspector must determine rebar location, cover depth, rebar diameter and rebar spacing. Once this is determined, the coordinate system and grid system are detailed, with an appropriate grid spacing. Usually, a grid of 0.2e1.5 m is used. Similar to the number of points, the grid spacing will be based on severity, available time, previous observations, structure size, etc. This grid is typically painted/labeled in the area of inspection. Once the grid is decided upon, the area needs to be prepared for measurements. It is essential to check for electrical continuity between the device used and the rebar under consideration. This is especially important if the connection is not done through the

Practical field implementation of corrosion measurement methods

307

rebar under consideration but using another rebar that is deemed to be electrically connected to the rebar being measured. This is most accurately done by exposing two or more rebars across the structure and measuring the potential difference between them with a high impedance voltmeter. Differences of less than 1 mV indicate connectivity. Alternatively, resistance can be measured. The resistance should be less than 1 U in all cases. If discontinuities are found, then it may be necessary to make regular rebar connections instead of relying on one or two connections to the rebar mesh. In ASTM C876, it is recommended that the electrical connection to the reinforcing steel is done by means of a compression-type ground clamp, by brazing or welding a protruding rod, or by using a self-tapping screw in a hole drilled in the bar. In many cases, the connection can have poor stability due to connecting to a rusted rebar surface layer. In these cases, the rebar needs to be thoroughly cleaned (grinding, sanding, scraping or brushing) before connecting. The stability of the measurement is highly influenced by the saturation of the concrete and its resistivity. In most cases, the concrete surface needs to be thoroughly pre-wetted to decrease the electrical resistance. It is recommended to first connect a reference electrode to the rebar similar to performing a regular measurement and observing the voltage fluctuation. A fluctuation of less than 20 mV within 5 min means that pre-wetting is not required. If this stability is not achieved, pre-wetting is required. This can be achieved by spraying or wetting the entire concrete surface, or the measurement points until stability is achieved. In cases where stability is not achieved after spraying or wetting, saturating sponges with a conductive contact solution or water and placing them on the concrete surface is recommended. The sponges should be left in place for the period necessary to obtain stable measurements. When measurements are performed, it is essential to ensure that no free surface water remains between grid points. During all LPR measurements, a wet sponge or any other conductive material or cloth has to be used below the counter electrode and reference electrode. In order to ensure the stability of LPR measurements, perform two or more measurements at the same location, while allowing for several minutes between measurements (to ensure complete de-polarization) until the half-cell potential returns to the same range as before polarization. These measurements should differ by less than 0.05 mA/cm2. If repeatability is not achieved, the same guidelines as that provided for half-cell potential measurements should be followed. On coated or hydrophobically treated surfaces, the ohmic resistance (previously described) has to be measured to ensure it is not significantly higher (more than an order of magnitude) than the polarization resistance measured or anticipated. The resistivity of the concrete can also be checked using a Wenner probe to ensure a reasonable range is found (