This document summarizes a paper that critically reviews the current understanding of concrete deterioration due to corrosion of reinforcing steel. It discusses several key issues with the current understanding, including the mechanisms of steel passivation in concrete, chloride diffusion modeling, mechanisms of steel depassivation by chlorides, and the composition and protective effects of mineral scales. The paper aims to identify research needs to improve the fundamental understanding of deterioration mechanisms.
This document summarizes a paper that critically reviews the current understanding of concrete deterioration due to corrosion of reinforcing steel. It discusses several key issues with the current understanding, including the mechanisms of steel passivation in concrete, chloride diffusion modeling, mechanisms of steel depassivation by chlorides, and the composition and protective effects of mineral scales. The paper aims to identify research needs to improve the fundamental understanding of deterioration mechanisms.
This document summarizes a paper that critically reviews the current understanding of concrete deterioration due to corrosion of reinforcing steel. It discusses several key issues with the current understanding, including the mechanisms of steel passivation in concrete, chloride diffusion modeling, mechanisms of steel depassivation by chlorides, and the composition and protective effects of mineral scales. The paper aims to identify research needs to improve the fundamental understanding of deterioration mechanisms.
Concrete Due to Corrosion of Reinforcing Steel by Kimberly E. Kurtis and Kumar Mehta Synopsis: Throughout the world, sizable portions of national budgets must be appropriated for the repair and rehabilitation of concrete structures which have suffered damage due to corrosion of reinforcing steel. This paper presents a critical review of the current state-of-the-art on the mechanisms responsible for deterioration of concrete and reinforcing steel. The protective effect of the passive film and the depassivation of reinforcing steel by carbonation and chloride ingress are discussed. In addition to the passive film, some researchers believe that a mineral scale may contribute to the protection of embedded steel against corrosion. The following ambiguities in the current understanding of the deterioration of concrete due to corrosion of the reinforcing steel are discussed: mechanisms of passivation of steel in the concrete environment; stoichiometry of the passive film; applicability of Fick's second law for the prediction of chloride diffusion in concrete; mechanisms of depassivation of steel by chloride ions; threshold Cl-/OH ratio; the composition and mechanism of protection provided by mineral scales; mechanisms of expansion of the steel corrosion product in concrete; and the detrimental effects of aggressive ions on concrete properties. Keywords: Carbonation; chlorides; concretes; corrosion; durability; permeability; reinforced concrete; reinforcing steels. 535 536 Kurtis and Mehta Kimberly E. Kurtis is a graduate student researcher in the Civil Engineering Department at the University of California at Berkeley. She received her Master of Science degree in Structural Engineering and Materials at Berkeley and her Bachelor of Science in Civil Engineering at Tulane University. P. Kumar Mehta is Professor Emeritus in the Civil Engineering Department at the University of California at Berkeley. He is the author or co-author of numerous papers on cementitious materials and properties of concrete including a textbook on the subject. A Fellow of the American Concrete Institute, he has received several awards including the ACI's Wason Medal for materials research, CANMET/ACI award for outstanding contributions to research on performance of concrete in the marine environment, and the Mohan Malhotra Award for research on supplementary cementing materials. He has held the Roy Carlson Distinguished Professorship in Civil Engineering at Berkeley and has received the highest campus honor, the Berkeley Citation, for exceptional contribution to his field and .to the university. INTRODUCTION Steel corrosion and the resulting deterioration of reinforced concrete structures is one of the primary causes of increasing damage to infrastructure. Design of concrete structures based on material strength rather than durability, the growing use of deicing salts and chloride-containing admixtures, and construction in increasingly aggressive environments are some of the factors that have lead to the surge of concrete structures experiencing reinforcement corrosion in recent decades. The concrete durability crisis, particularly involving the corrosion of embedded steel, is extensive, and published literature is replete with case histories from all parts of the globe, such as those described below. United States The Federal Highway Administration (FHW A) recently reported that 226,000 U.S. reinforced concrete bridges, or 39% of the total, were deficient. In addition, 134,000 bridges, or 23% of the total, were classified as "structurally deficient", meaning that they can support only light traffic loads or that they require immediate repair [ 1]. In 1991, the United States Department of Transportation estimated rehabilitation costs for these damaged bridges at $90.9 billion [2]. Corrosion of the reinforcing steel is implicated in the majority of these damaged bridges [3]. Arabian Gulf In a survey of more than 100 reinforced concrete structures in the Arabian Gulf, Matta [4] found the typical service life to be 10-20 years, considerably less than the 50+ year design life. This reduction in service life resulted from the widespread corrosion of reinforcing steel that was caused by either chloride contamination of the concrete through mix ingredients or ingress, or by carbonation of concrete. Concrete Durability 537 Canada According to Khanna et al. [5], thermal cracking, frost action and corrosion of reinforcing steel caused severe damage to more than 400 concrete piles of the 7 year-old Rodney Terminal at Saint John harbor in Eastern Canada. Norway According to Gjorv [6], inspection of coastal bridges in Norway indicated that approximately 25 percent have experienced some reinforcing steel corrosion, mainly caused by chloride ingress. Many of these structures are less than 25 years old. Germany In May 1980, the roof of the Berlin Congress Hall, constructed in 1957, collapsed. Poor construction finishing at joints allowed moisture ingress, and hydrogen induced stress corrosion cracking in the prestressing tendons led to structural failure [7]. Mexico Castro et al. [8] assert that Mexico's infrastructure is corroding largely unchecked. This has resulted not only in economic losses but also the loss of human life. In the Yucatan Peninsula, where aggregates are rough and angular and the water/cement ratios are consequently high, there is much corrosion due to the high permeability of concrete. Pacific Rim In Hong Kong, the Kanmon Tunnels, which had been designed for 120 year life, leaked so profusely within 4 to 10 years of service that tunnels liners, rails, and equipment severely corroded. As a result, tunnel operation had to be interrupted to perform major repairs and reconstruction [9]. When internal steel reinforcement corrodes, the strength of the reinforced concrete member is undermined in several ways. Since corrosion products have a greater volume than the parent steel, internal tensile stresses will develop in the cement mortar at the steel/mortar interface. As a result, the surrounding concrete cracks and will eventually spall away as corrosion of the internal steel advances. In addition, under tensile stress developed during corrosion, existing fine cracks and microcracks in the surrounding concrete tend to enlarge and coalesce to form a network of interconnected cracks, providing increased ionic transport between the surface of the concrete and the surface of the reinforcing steel, effectively promoting the corrosion process. Crack growth decreases concrete stiffness and tensile strength, while the formation of a network of cracks increases concrete permeability. Thus, the structural integrity of the reinforced concrete member is increasingly compromised as cracking progresses. As the steel is progressively lost to corrosion, the reinforcing bar cross-section is reduced, causing a decrease the member's tensile strength. Furthermore, as corrosion advances, the bond between the steel and surrounding concrete is weakened, adversely affecting the load transfer between the two materials. To ensure that reinforced concrete members perform according to their design capacity and design service life, it is important to prevent or delay the occurrence of corrosion. 538 Kurtis and Mehta The growing incidence of reinforced concrete corrosion and the resulting repair costs have led the construction industry to acknowledge that changes are necessary in the design and construction for new reinforced concrete structures. Consequently, new products and techniques are being developed and made available for the construction of new structures. Techniques currently used to guard against corrosion include: sealants, epoxy-coated reinforcing bars, galvanized steel, fiber-reinforced plastic reinforcement, cathodic protection, and protective overlays and membranes for bridge decks. Many of these protective measures increase construction costs greatly, and some have met with mixed success. A more traditional, and typically less expensive, means of decreasing corrosion involves reducing the permeability of concrete. This can be accomplished by a combination of several techniques, such as providing adequate concrete cover thickness, using mineral admixtures and water-reducing agents in the concrete mixture, and proper curing. With sophisticated structures, a synergistic application of the traditional approach and newer methods and materials is required to ensure durability in aggressive environments. Despite the plenitude of publications on the subject of corrosion and its prevention, adequate fundamental information concerning the mechanisms of deterioration of reinforced concrete from corrosion is lacking. The purpose of this paper is to present an overview of the current state-of-the-art and to address deficiencies in our knowledge with the aim of providing a research impetus for better understanding of the underlying mechanisms. After a general review of the corrosion process, the nature of the passive film, the effects of carbonation and chloride ingress, and the role of mineral scale will be discussed. The electrochemical process of corrosion in concrete will be described, and a holistic damage model for reinforced concrete exposed to aggressive environmental conditions will be presented. THE CORROSION PROCESS In general, the corrosion process occurs when metals revert back to lower energy states. Independent of the metal involved, five specific conditions are required for corrosion to occur in an aerobic environment. These requirements are: 1) the presence of an anode to produce electrons, 2) the presence of a cathode to accept electrons, 3) the availability of oxygen at the cathode site, 4) the availability of water at the cathode site, and 5) an electrical connection between the anode and cathode sites to transfer electrons. If any of the above listed conditions is absent, corrosion will not occur. The anodic and cathodic regions may form on the metal embedded in the concrete. If dissimilar metals are present, one metal can act as a cathode and the other as an anode. For example, when aluminum conduits and steel reinforcing bars are embedded in concrete, the steel, being more noble than the aluminum, will be cathodic to the conduit. Even when only one type of metal is embedded, anodic and cathodic sites may form on the surface due to metal defects, uneven formation of films, presence of different microstructural phases, different grain orientations, or localized differences in the surrounding pore solution chemistry. Concrete Durability 539 Thus, requirements 1 and 2 are dependent on the characteristics of the metal or metals present. Requirements 3-5, listed above, are governed by the properties of the cement paste. To understand the effect of cement paste on corrosion, it is helpful to consider it as a two-phase material, composed of hydrated solid minerals and the pore fluid. The structure, size, distribution, and interconnection of the pores in the cement paste, in conjunction with the presence of cracks and microcracks, control the permeability of concrete. The more permeable the concrete, the greater the availability of oxygen and water at the cathode for electrochemical reaction. Gonzalez et al. have shown that higher concentrations of water and oxygen in the vicinity of the cathode will increase the rate of corrosion [ 10]. This is supported by the Nernst equationt, a fundamental concept used in the study of corrosion which demonstrates that the corrosion rate increases as the concentration of reactants, such as oxygen, increases. A more permeable concrete also allows greater ingress of aggressive chloride ions and carbon dioxide. Both of these substances contribute to the breakdown of an protective iron oxide film that is typically present on the steel surface when exposed to a highly alkaline environment. This film, termed a 'passive film', will be discussed further in the following sections. While the permeability of uncracked concrete is largely determined by the pore structure of the hydrated cement paste, it is the pore fluid and the hydrated paste that together act as the electrolyte needed for the transfer of electrons during the corrosion process. As the resistivity of the pore solution is much less than that of the hydrated cement paste, the solution is the more conductive electrolyte. Therefore, the transfer of electrons during the corrosion process can be considered to be dependent on the size, distribution, and interconnection of pores and their degree of saturation. Typically, unsaturated concrete is a high resistance material, and, as a result, the corrosion rate of embedded steel is limited to some degree[l1]. Although the electrical resistivity of ordinary concrete is high (typically 10-20Qm), electron transfer from the cathode to the anode during the corrosion process does occur. However, according to Wolsiefer [12], when the electrical resistivity of concrete lies above 60Qm corrosion will not occur even in chloride contaminated concrete. From the above discussion, it is evident that corrosion requirements 1, 2, 4 and 5 - the presence of an anode, a cathode, water at the cathode, and an electrolyte - are readily met by most reinforced concrete members. The embedded steel contains anodic and cathodic sites, water present in the pores us available for the cathodic reaction, and both the surrounding pore fluid and cement paste provide a conductive medium for charge transfer. Requirement 3, the presence of oxygen at the cathodic site, is dependent on the permeability of the concrete. As will be explained further, the corrosion process cannot occur without the availability of oxygen at the cathodic site. However, it has been established that even very dense concrete is fairly permeable to oxygen [13, 14]. t E =Eo- (0.0592/n) log ([products]/[reactants]), where E represents the potential difference betweeen the metal and solution, E 0 is the equilibrium potential, n represents the number of equivalents exchanged, and [] denotes concentration in mol/L for a reduction reaction at room temperature. 540 Kurtis and Mehta Up to this point, the five general requirements for corrosion, independent of environment, have been examined. However, an important sixth requirement - a requirement specific to corrosion of steel embedded in portland cement concrete - has been neglected. This sixth requirement for corrosion of embedded steel in concrete is the penetration or removal of a thin oxide layer which forms a protective, passive film on the steel surface. When this film is intact, corrosion of the steel cannot occur because requirement 1 is not satisfied. The destruction of this film is dependent on both the permeability of the concrete and the chemistry of the cement paste as discussed in detail in the following sections. THE PASSIVE FILM Any metal or alloy is said to be in a "passive" state when it resists corrosion in an environment where corrosion is thermodynamically favorable [15]. Passivity is not a characteristic of a metal or alloy, but instead it is characteristic of a material with respect to a given environment. For steel embedded in concrete, the formation of a passive film is a function of environmental pH. Typically, the pH of the pore solution surrounding the steel lies between 12.5 and 13.5, depending on the concentration of alkalis in the cement. Sodium and potassium, derived from raw materials used in the production of cement clinker, are the source of much of the high pore solution alkalinity. Calcium hydroxide, one of the solid hydration products, acts as a buffer by maintaining the high pH of the pore solution [16]. As the Pourbaix diagram (Figure 1) for iron in an aqueous environment indicates, steel subjected to high pH environment remains passive over a wide range of potentials. Hausmann [18] demonstrated that a very limited supply of oxygen was sufficient to maintain potentials in this range. Many authors cite the Pourbaix diagram as a proof that a protective passive ferric or ferrous oxide film forms on reinforcing steel in concrete. Borgard et al. [ 11] caution that Pourbaix diagrams should not be used to predict whether a certain metal-oxide species will form a film that will provide corrosion protection. It is also important to note that these diagrams describe conditions of thermodynamic stability of metal species and do not consider the kinetics of the reactions implied. Even more, the Pourbaix diagram depicted in Figure I is applicable to the behavior of iron or steel in an aqueous solution. Yet, many researchers have wrongly used this diagram to explain the passivity of steel in concrete, assuming that pH is the only relevant parameter for passivation.* Because of these limitations, the Pourbaix diagram should not be used to predict the stoichiometry of the passive film or the amount of corrosion protection imparted by the film that forms on surface of steel embedded in concrete. Despite the confusion in the literature over the applicability of Pourbaix diagrams for predicting the behavior of steel in reinforced concrete, it is generally accepted that when steel is exposed to a highly alkaline environment a submicroscopic oxide film forms on the steel surface. The film is thought to form rapidly during the initial stages of cement hydration [ 19]. Adding to the uncertainties surrounding the understanding of this film, various stoichiometries * see "Mineral Scale" discussion for more details Concrete Durability 541 are reported in the literature. Kerkar et al. [20] state that the film consists of hydrated ferric oxide (FeOOH) or ferric oxide (Fe203), while others [21, 22] report that the film consists of an inner layer of ferrous oxide (Fe304) under an outer layer of some form of ferric oxide. Sagoe-Crentsil and Glasser [23] state that the "chemical and mineralogical compositions (of the passive film) are yet to be determined ... it is, however, feasible that the passive film consists of several phases, perhaps intimately mixed or zoned in response to changing oxygen gradients across the interface". In addition, the authors imply that calcium ions from the cement paste may be incorporated into the passive film protecting reinforcing steel embedded in concrete. Indeed, Devine** has suggested that the passive film may be composed of a single or a combination of previously unknown oxides, hydroxides, or oxyhydroxides. Because the passive film forms when the steel is subjected to the alkaline environment provided by concrete, any study of the film must take place within this environment. This constraint has posed a great challenge for the study and characterization of passive films. It has been determined, however, that the film is thinner than 2nm which is equivalent to a few layers of oxygen. As a result, the film is optically transparent. The strongly adherent passive film is thought to act as a barrier, inhibiting the anodic dissolution of iron, and, thus, protecting it from corrosion. While corrosion may not be completely absent during this period of passivity, its rate is greatly limited by the presence of the film. Hansson [22] has shown that the corrosion rate of the passivated steel in concrete is about 0.1 J..Lrnlyear, an amount which can be considered negligible. The passive film has been studied for over two centuries, and various theories have been developed to explain its protective effect on metals. Yet, little research has been conducted on the mechanisms of steel passivation in concrete. Borgard et al. [ 11] state that this concept of protective passive film formation on reinforcing steel in concrete was "introduced (by Cornet et al. in 1968 [24]) with no experimental evidence to support the passivation mechanism". That is, the concept of passivity of embedded reinforcing steel has been embraced with little research to explain its protective mechanisms. Currently, the most plausible and accepted explanations for the passivation of steel in alkaline aqueous solutions are the adsorption theory and the film theory. The adsorption theory was first proposed by Faraday and presumes that the metal surface has such a great affinity for oxygen, that a thin, chemisorbed oxygen layer builds up on the metal surface over time. Because the adsorption of this layer is in some way "equivalent to oxidation" of the metal, the oxygen layer inhibits further chemical reaction of metallic iron with the surrounding electrolyte. The film theory is based on the presence of a hydrated metal oxide surface film as a result of environmental interaction. This film is assumed to be thicker than the oxygen layer adsorbed to the metal surface in the alternate theory presented here. By the film theory, it is postulated that impervious oxide films offer protection against corrosion by decreasing the diffusivity of metal ions into solution. Chawla and Gupta [15] state that these two theories may not be contradictory, but that instead they may complement one another. When the passive film is first forming and is thin, the adsorption theory may describe its protective behavior. But as the film thickens, the oxide theory may be more applicable. Again, these are theories only, and little experimental work has been **Personal communication, T.M. Devine, University of California, Berkeley, 1996. 542 Kurtis and Mehta done to determine the mechanisms of protection of steel in the concrete environment. It is known that the film formed at the metal surface is relatively unstable, and the passivity provided is temporary. For steel embedded in concrete, the two known means of depassivation are: a reduction in concrete alkalinity and attack by aggressive ions. It is reported that when the pH of the surrounding environment is higher than 11.5, the passive film will remain intact and will provide protection to the steel against corrosion [25]. Once the pH falls below this level, the film becomes unstable and the protection of the steel will be lost. Carbonation of concrete is the most common mechanism for the loss of alkalinity in concrete. In addition, aggressive ions, such as chloride, can dissolve or penetrate the passive film at any pH. Most often, this occurs when chlorides diffuse through the concrete and attack the passive barrier. The two most common means of depassivation, carbonation and chloride ion attack, are presented in the following sections. EFFECT OF CARBONATION Decaying organic matter often imparts significant concentrations of dissolved carbon dioxide (C02) into mineral, groundwater, and seawater [25]. The solution of this gas in water, carbonic acid (H2C03), reacts with the cement paste- a process termed "carbonation". Carbonation of calcium hydroxide (Ca(OH)2) present in the hydrated cement paste is represented by the reactions below: Ca(OH)2 + H2C03 -+ CaC03 + 2H20 CaC03 + C02 + H20 <-+ Ca(HC03)2 (1) (2) From these equations, it can been seen that calcium hydroxide in the hydrated cement paste is consumed through reaction with carbon dioxide or carbonic acid. This results in a gradual loss of buffering capacity which is essential for maintaining alkalinity in concrete. As stated previously, the passive film becomes unstable and no longer provides protection to iron at a pH below 11.5 which implies that carbonation can result in depassivation of reinforcing steel. The kinetics of the carbonation reaction in concrete have been widely studied, and a parabolic model is generally accepted to describe the depth of carbonation expected over a period of time. According to this model, the depth of carbonation can be predicted as follows: X= k t J/2 (3) In this equation, xis the thickness of the carbonation layer, tis exposure time, and k is a constant dependent on the permeability of the concrete cover. In good quality concrete, the rate of carbonation is typically on the order of lmrnlyear, indicating that carbonation should not present a problem during the design service life of a reinforced concrete structure which has 50mm or more concrete cover over the reinforcement [14]. Concrete Durability 543 Carbonation and the resulting loss in alkalinity typically begins at the concrete surface and progresses in the form of a "carbonation front" toward the center of the concrete section. Whereas the carbonated concrete will have pH of less than 8, the uncarbonated concrete maintains its high alkalinity. Thus, the uncarbonated concrete continues to provide a protective cover for the embedded steel. In permeable concrete, concrete with low cover depth, or in concrete containing microcracks connected to surface cracks, the carbonation front may reach the reinforcing steel and thus leads to the onset of corrosion [26]. Therefore, adequate cover depth and a concrete mixture with low permeability and low potential for microcracking is crucial for the protection of embedded steel from corrosion. Typically, corrosion induced by carbonation is a uniform attack. That is, the presence of very localized corrosion -or pitting - is notably absent. It is important to remember that depassivation of the steel by carbonation is not sufficient to cause corrosion. Water and oxygen must also be present at the cathode site. Because concrete structures are often subjected to cycles of drying and rewetting, an interesting interaction of carbonation and corrosion is found. Venuat [27] has shown that carbon dioxide diffusion is most rapid in concrete at intermediate humidity contents (50-80% relative humidity). When the concrete is completely dry, the carbonation reaction is very slow, and when the concrete is completely saturated, the diffusion of carbon dioxide is very slow. When the concrete is semi-dry, carbonation occurs. Upon wetting, the cathodic reaction of the corrosion process is enhanced. Thus, alternating cycles of drying and wetting provide the most aggressive environment for the carbonation-induced corrosion [16]. EFFECT OF CHLORIDE IONS As compared to corrosion resulting from concrete carbonation, corrosion of reinforcing steel caused by chloride ion ingress is much more prevalent. Chloride ions can be introduced into the concrete during its manufacture. This can occur when seawater or water with a high chloride concentration is used as mixing water, when chloride-contaminated coarse or fine aggregates are used, or when chloride-containing admixtures such as calcium chloride are used. Chloride ions present in the surrounding environment may penetrate the concrete. Deicing salts, seawater, and chloride-contaminated soils are the primary sources of external chloride. When concrete is dry, chlorides can penetrate several millimeters in a few hours by the capillary draw of salt water into the concrete [16]. When concrete is partly or fully saturated, chloride ions penetrate by diffusion through the pore solution. Typical diffusion rate for a fully saturated cement paste are on the order of w-8 cm2/s [28]. Different models exist to describe the ingress of fluids and ions, including chlorides, in concrete. For steady-state flow, Darcy's law relates the rate of flow dqldt to the hydraulic pressure gradient dhldl by: dqldt = K F (dhldl) (4) 544 Kurtis and Mehta with the permeability coefficient K and the cross section exposed to flow F: For the diffusion of gases, Fick's laws are typically applied. For the steady state, Pick's first law m = -D (fx:lox) (5) applies, and for the unsteady state, Pick's second law (6) applies, where c is concentration and D is the diffusion coefficient. Recent research by Chatterji [29] has shown that Fick's second law, in particular, is not necessarily applicable to concrete as it assumes that the material is non-ionic and completely saturated. Based upon field experience with concrete structures exposed to chloride environments such as seawater, Sandberg [30] asserted that difficulties in the application of Fick's laws may arise from a pore-blocking effect that results when hydrated cement paste interacts with seawater. Much research has been conducted in recent years to predict the time required for migration of chlorides through concrete. In most cases, some form of Pick's second law has been used to calculate the penetration depth over long periods of time, and, as discussed above, the predicted behavior may be unreliable for use in service life design of reinforced concrete structures. While it has been generally agreed that chloride ions act as catalysts for the loss of the protection offered by the passive film, the exact mechanisms of this process are not well understood. Two theories appear most often in the literature: the adsorption theory and the oxide film theory (note that the same nomenclature is used to the describe the mechanisms for passivation). According to the adsorption theory, the chloride ions replace oxygen atoms held within the passive film [16]. This causes differences in electrochemical potential across the film, and the film becomes unstable. The high reaction rate of steel and chloride, in areas where chloride ions have replaced oxygen, is thought to explain the occurrence of pitting on the steel surface which is typical of chloride-induced corrosion [23]. According to the oxide film theory, the passive film contains inherent defects and pores. By a selective dissolution of more reactive components of the passive film, chloride ions penetrate the film at these sites more readily than other anions present in the pore solution. The theory suggests that these localized attacks by chloride ions are responsible for pitting corrosion. Both theories suggest that attack of the passive film by chloride ions is a localized phenomenon. This form of attack causes microgalvanic cells to form on the reinforcing steel as shown in Figure 2. In regions where the depassivation has occurred, iron will be lost by oxidation. The areas that remain protected by the passive film will become cathodic and, thus, sites of oxygen reduction. These processes are described more fully in a subsequent section detailing the chemical reactions involved in the corrosion process. Because of the damaging effects of chloride in reinforced concrete, much research has been performed to determine a threshold chloride ion content below which corrosion will not occur in concrete. The ions exist in concrete in two forms -bound and free. Only the free chloride ions, those dissolved in the pore fluid, participate in the corrosion process, and as a result, it is the free chloride concentration, not the total chloride content, that is critical when determining Concrete Durability 545 threshold levels at the steel concrete interface. It is well-known that the chloride- binding capacity of a concrete depends on the cement composition. Since chloride ions react with calcium aluminate (C3A) present in the cement paste to form Friedel's salt (C3ACaCI215H20), concrete made from cement with a high C3A content will have a greater potential for binding chloride ions. Hussain et al. [31] demonstrated the beneficial effects of such cement. By raising the C3A content of cement from 2.43% to 14%, with all other factors remaining constant, the chloride threshold was increased by a factor of 2.85. Kayyali and Haque [32] found that the use of superplasticizing admixtures increased the levels of free chlorides in concrete, but that the addition of fly ash resulted in an increase in chloride binding capacity. Hussain et al. reported moderate increases in the chloride threshold value with higher levels of concrete alkalinity. Also, it was found that the presence of sulfates either moderately increased or decreased the threshold value depending on the cement composition. Hausmann [18] was the first to suggest a relationship between chloride content, concrete alkalinity, and the onset of corrosion. Subsequently, much research has been conducted to establish a CI-JOH threshold value for corrosion initiation. As shown in Table 1, a range of values are reported in the literature. For some time, the CI-JOH ratio of 0.6 reported by Hausmann was accepted by many, but it should be noted that the study that produced this value was conducted in an alkaline solution meant to simulate concrete pore fluid, not within. concrete itself. Recent research suggests that the source of chloride contamination has a significant influence on the threshold values. Diamond [33] reported a threshold of 0.3 when the chloride is introduced into the concrete as an admixture or in the mixing water. Lambert et al. [34] reported a much higher threshold ratio, 3.0, when the chloride is introduced from an external source. It is important to recognize that the threshold CI-/OH ratio is not easily defined and that no value has been accepted as a unique limit for all concrete mixes. Threshold values are highly dependent on such parameters as the water/cement ratio, cement content, cement surface area, cement composition, ambient temperature, and alkalinity. As a result, a well-defined threshold relationship universally applicable to all concrete mixtures remains elusive. Consequently, service life predictions based on certain assumed threshold values are not reliable. ROLE OF MINERAL SCALES While it is generally accepted that the presence of a passive film on the steel surface provides some corrosion protection, some researchers, believe that an additional phenomenon may offer further protection. Many authors consider the pore fluid to be the environment of exposure in concrete. As a result of this assumption, much reinforced concrete corrosion research has been conducted in aqueous solutions that are meant to simulate the concrete pore fluid. Consequently, the corrosion behavior of steel in concrete has been extrapolated from its behavior in aqueous solutions of similar pH. This approach neglects the effect of the interfacial steel-cement paste zone which consists of hydrated cement paste minerals. 546 Kurtis and Mehta Borgard et al. [11] postulate that mineral scales, consisting of precipitates of cementitious compounds, form on metal surfaces in concrete and act in conjunction with the metal oxide passive film to protect reinforcing steel. The authors suggest that this phenomenon has been ignored in much of the literature as a consequence of the use of pore solution models, which do not account for the presence of the hydrated cement paste minerals, in corrosion studies. Page [35] has also proposed that passivation may result from a lime-rich layer of hydration products at the interface between steel and concrete. Borgard et al. state that calcium carbonate scales are used for protection of oil well production tubing even in the presence of chloride concentrations much greater than in seawater. They assert that the formation of high calcium scales on steel embedded in concrete is likely since portland cement has a high calcium content. The authors reason that corrosion in concrete may be prevented by a similar mechanism and that corrosion will occur only after the calcium scales have been removed. It should be recognized that chemical attack on the calcium hydroxide scale by carbon dioxide, chloride, or sulfate ions can result in eventual removal of the scale. However, the process of removal can provide additional protection to the passive film on the steel surface because of the dissociation of the calcium hydroxide by reaction with such species as chloride ions and carbon dioxide. Such reactions produce free hydroxyl ions which act to buffer the alkalinity of the pore solution. If this reaction takes place in the vicinity of the steel, passivity can be maintained by ensuring that the pH remains above 11.5. However, once the calcium hydroxide scale has been depleted, further ingress of chloride ions or carbon dioxide damage the passive film. CORROSION CHEMISTRY As stated previously, corrosion is an electrochemical process involving the transfer of ions. In reinforced concrete, electrons are produced at the anodic site by the oxidation of the steel. That is, metal ions from the parent metal go into solution at the anode, and the loss of metal indicates that this is the site of corrosion. The rate of steel corrosion in concrete is dependent on the rate of the anodic reaction: Fe ---+ Fe++ + 2e- (7) Electrons released at the anodic site are consumed at the cathodic site. Typically, at the cathode, either oxygen or hydrogen is reduced. The thermodynamic driving force for the reduction of oxygen is much greater than that for hydrogen. As a result, the oxygen reaction is more likely to occur, and the reduction of hydrogen in concrete corrosion is rarely a factor. The cathodic reactions are: (8) (9) Concrete Durability 547 Ferrous ions (Fe++) produced in the anodic reaction (eq. 7) combine with the hydroxyl ions produced in the cathodic reaction (eq. 8) to form the corrosion product ferrous hydroxide (Fe(OH)2). The corrosion rate can be increased by the presence of other ions in the concrete. According to Hime and Erlin [36], the reaction of ferrous ions and hydroxyl ions in the presence of chloride ions can also result in the formation of the corrosion product Fe(OH)2 (eqs. 10-12). The reaction represented in equation 16 demonstrates that chloride ions are produced by this process, increasing the corrosion rate of the steel reinforcing. Fe+++ 6Cl- +-> FeC16-4 (10) Fe+3 + 6Cl- +-> FeCl6-3 (11) FeC16-3 + 20H- +-> Fe(OH)2 + 6Cl- (12) The process of corrosion for steel reinforcement bars embedded in concrete is illustrated in Figure 2. The diagram shows that the electrons are released at the anodic site and travel through the steel to the cathodic site, where they are consumed. The reduction of oxygen at the cathode produces hydroxyl ions (OH-). The production of hydrogen gas, caused by the reduction of hydrogen, is less likely, but possible. Equations 7 and 8 demonstrate that the corrosion process may be limited by the rate of diffusion of the hydroxyl ions, as well as by the availability of oxygen (02), water (H20), or other aggressive ions (Cl- in Fig. 2 and eq. 10-12). Glasser and Sagoe-Crensil [37] have examined a 27 year old reinforced concrete post to determine the morphology and characteristics of the corrosion product (i.e. rust) present. In the specimen studied, the corrosion product was found to be composed of two layers: an inner, highly crystalline and dense region of Fe304 and y-Fe203 and an outer more porous region of primarily o:-FeOOH (goethite). The inner layer was found to have preferential orientation related to that of the substrate metal microstructure that resulted in epitaxial growth of the inner layer corrosion product. Wang and Monteiro [39] confirmed that the corrosion product tends to form a continuous band poorly bound to the reinforcing steel and that this band is composed of multiple layers. It has been suggested that this structure results from variations in oxygen availability and pore solution conductivity across the region of corrosion. MECHANISMS OF EXPANSION AND CRACKING OF CONCRETE The solid-state transformation of iron to the ferrous hydroxide corrosion product resulting in a volumetric expansion is generally assumed to be the mechanism by which concrete expands and cracks by the corrosion of embedded steel. However, the corrosion chemistry shows that the corrosion of iron to ferrous hydroxide occurs as a result of a through-solution process. That is, iron present in steel must oxidize to soluble ferrous ions. The ferrous ions combine, in solution, with hydroxyl ions generated at the cathode to produce ferrous hydroxide. Consequently, Mehta [39] and Figg [40] postulate that poorly 548 Kurtis and Mehta crystalline or gel-like ferrous hydroxide tends to swell by water adsorption and that the hydraulic pressure generated is responsible for expansion and cracking. Recently, Wang and Monteiro [38] studied the mechanisms by which corrosion of embedded steel reinforcement undermines the strength of concrete structures. Once the depassivation of the steel occurs and corrosion begins, they found that oxidation products form as a band around the steel. Loss of reinforcing steel by corrosion leads directly to loss in the strength of the reinforced concrete member. While the presence of water is necessary for the corrosion of reinforcing steel, Wang and Monteiro found that cycles of relative wet and dry conditions at the concrete and steel interface lead to weakening of the corroding reinforced concrete member. When wet, adsorbed water will force corrosion product particles further apart, resulting in expansion and cracking of the surrounding concrete. Upon drying, the corrosion product will tend to shrink and crack. In this way, cycles of wetting and drying will cause the steel and concrete to debond at the interface. In addition, Wang and Monteiro found that corrosion products diffuse away from the interface into the microcracks and pores of the surrounding concrete. The researchers reason that 'plugging' of existing voids in the concrete by diffused corrosion product will decrease the probability of crack arrest, effectively embrittling the concrete and weakening the reinforced member. Loss of steel, debonding at the steel/concrete interface, and cracking and embrittlement of surrounding concrete is detrimental to the strength and integrity of reinforced concrete. Mehta [41] has proposed a holistic model, Figure 3, which illustrates all the physico-chemical changes occurring in the reinforcing steel and concrete in response to all environmental effects including weathering and ingress of aggressive chloride ions and carbon dioxide. Most descriptions of the effects of corrosion of steel embedded in concrete center on the volume expansion associated with the corrosion products and the resulting cracking and debonding of the concrete. As discussed previously, corrosion of reinforcing steel damages the surrounding concrete as well as the steel. The mechanisms depicted in Figure 3 are holistic. That is, no part of the system is overlooked. The mechanisms that lead to the corrosion of steel also affect the surrounding concrete, and these relationships are considered in this model. For instance, since C-S-H is the primary source of strength in the cement paste and since the stability of C-S-H is dependent on the concentration of hydroxyl ions in the pore fluid, a decrease in alkalinity has a detrimental effect on the strength and elastic modulus [41]. Free hydroxyl ions in the pore fluid can combine with other species, such as carbonate, sulfate, and chloride. This type of reaction, which results in a decrease of the pore fluid alkalinity, causes loss of strength and elastic modulus in the concrete. The mechanism of expansion and cracking of concrete shown here takes into consideration the combined effect of weakened microstructure of the cement paste and the development of hydraulic pressure in the pores of the water-saturated system. Thus, Figure 3 demonstrates an integration of all environmental effects on both components of the reinforced concrete system. Concrete Durability 549 CONCLUSIONS Upon critical review of the state-of-the-art understanding of expansion and cracking of reinforced concrete due to corrosion of the embedded steel, several areas of uncertainty or ambiguity emerge. These are identified as: Mechanisms of steel passivation in the concrete environment. The stoichiometry of the passive film on steel and the mechanisms of its formation. The relative significance of the passive film and mineral scales in providing corrosion protection to steel. The composition of the mineral scales formed in the vicinity of the steel and the mechanism of corrosion protection. Applicability of Fick's second law for the prediction of chloride diffusion in concrete. Mechanisms of depassivation of reinforcing steel by chloride ions and carbonation. The threshold Cl-JOH ratio for corrosion to occur. Composition, morphology, and mechanisms of expansion of the steel corrosion products. Detrimental effects of aggressive species on the material properties of concrete itself. The study of corrosion is limited to some extent by the inherent difficulties associated with observing the corrosion of steel within concrete. However, some of the gaps in our understanding of the steel corrosion process in concrete seem to stem from the reductionist or fragmentary approach taken by researchers on this subject. Corrosion engineers and material scientists tend to study the corrosion of steel in alkaline aqueous solutions and, as a result, neglect the possible effects of the cement paste microstructure on the corrosion process and changes in the microstructure during corrosion. For instance, the effect of chloride attack and carbonation on the surface chemistry of steel is considered, whereas their detrimental effect on the strength and elastic modulus of the concrete is often completely overlooked in our corrosion damage models. Widespread misapplication of Pourbaix diagrams for prediction of passivity of steel embedded in concrete is an example of the narrow approach applied to the study of corrosion of steel embedded in concrete. Only when researchers take a holistic approach for the study of combined response of both the reinforcing steel and the surrounding concrete to all environmental effects will these ambiguities be resolved. To address the deficiencies and advance the state-of-the-art in concrete science, we obviously need a paradigm shift in our approach to corrosion research. REFERENCES [1] Federal Highway Administration, 1991 Report to Congress, June, 1991. [2] Federal Highway Administration. "1991 Status of the Nation's Highways and Bridges: Conditions, Performance, and Capital Investment Requirements." July 2, 1991. 550 Kurtis and Mehta [3] ELTECH Research Corp. , "Cathodic Protection of Reinforced Concrete Bridge Elements: A State-of-the-Art Report.", SHRP-S-337, 1993. [4] Matta, Z.G., "More Deterioration of Reinforced Concrete in the Arabian Gulf", Concrete International, Nov. 1993, p. 50-51. [5] Khanna, J.; Seabrook, P.; Gerwick, B.C. ; and Bickley, J. "Investigation of distress in precast concrete piles at Rodney Tenninal", Performance of Concrete in Marine Environment ed. V. Malhotra, ACI SP-109, 1988, p. 277-320. [6] Gjorv, O.E. "Steel Corrosion in Concrete Structures Exposed to Norwegian Environment", Concrete International, April, 1994, p. 35-39. [7] Isecke, B. "Collapse of the Berlin Congress Hall Prestressed Concrete Roof", Materials Performance, Dec. 1982, p. 36-39.[13] Gjorv, O.E.; Vennesland, 0. and El Busaidy, A.H.S., Proc. Corrosion, 76, NACE, Houston, 1976. [8] Castro, P.; Maldonado; and deCoss, R. "Study of Chloride Diffusion as a Corrosive Agent in Reinforced Concrete for a Tropical Marine Environment", Corrosion Science, vol. 35, no.5-8, 1993, p. 1557-1562. [9] Gerwick, B.C. "Pressing Needs and Future Opportunities in Durability of Concrete in the Marine Environment", International Experience with Durability of Concrete in the Marine Environment, Ed. P. K. Mehta, Jan. 16-17, 1989, p. 1-5. [10] Gonzalez, J.A.; Lopez, W. ;and Rodriguez, P. "Effects of Moisture Availability on Corrosion Kinetics of Steel Embedded in Concrete", Corrosion, December 1993, p. 1004-1010. [11] Borgard, B.; Warren, C.; Somayaji, S.; and Heidersbach, R. "Mechanisms of Corrosion of Steel in Concrete", Corrosion Rates of Steel in Concrete, STP1065, ASTM, 1990. [12] Wolsiefer, J.T. "Silica Fume: A Solution to Steel Reinforcement Corrosion in Concrete", ACI SP-126, 1991. [13] Gjorv, O.E.; Vennesland, 0.; and El Busaidy, A.H.S., Proc. Corrosion, 76, NACE Houston, 1976. [14] Page, C.L. and Treadaway, K.W.J., "Aspects of the Electrochemisty of Steel in Concrete", Nature, 297, May 1982, p. 109-115. [15] Chawla, S.L. and Gupta, R.K. Materials Selection for Corrosion Control. ASM International, 1993. [16] Rosenberg, A.; Hansson, C.M.; and Andrade, C. "Mechanisms of Corrosion of Steel in Concrete", Materials Science of Corrosion I, The American Ceramic Society, 1989. [17] Pourbaix, M. Atlas of Electrochemical Equilibrium in Aqueous Solutions. Pergamom, London, 1976. Concrete Durability 551 [18] Hausmann, D.A. "Steel Corrosion in Concrete: How Does it Occur?", Materials Protection, November 1967, p.19-23. [19] ACI Committee 222. "Corrosion of Metals in Concrete", ACI Journal, Jan./Feb. 1985, p. 3-32. [20] Kerkar, J.; Robinson, A.J.; and Forty, A.J. Faraday Discussion, Chemical Society, 89, 1990, p.31. [21] Sato, N., Passivity of Metals, Electrochemical Society, Princeton, NJ, 1978, p. 29. [22] Hansson, C.M., "Comments on the Electrochemical Measurements of the Rate of Corrosion of Steel in Concrete", Cement and Concrete Research, Vol. 14, 1984, p. 574-584, Pergamon Press Ltd. [23] Sagoe-Crentsil, K.K. and Glasser, F.P. "Steel in Concrete: Part I. A Review of the Electrochemical and Thermodynamic Aspects", Magazine of Concrete Research, Dec. 1989, p. 205-212. [24] Cornet, I.; Ishikawa, T.; and Bresler, B. Materials Protection, Vol. 7, No.3, March 1968, p.44-47. [25] Mehta, P.K. and Monteiro, P.J.M., Concrete: Structure. Properties. and Materials. Second Edition, Prentice Hall, 1993. [26] Parrott, L.J. "A Study of Carbonation Induced Corrosion", Magazine of Concrete Research, vol. 46, no. 166, 1994, p. 23-29. [27] Venuat, M. "Relationship between Concrete Carbonation and the Corrosion of Reinforcement", Recentres CEFRACOR-77, JTBTP, Oct. 1977. [28] Ngala, V.T.; Page, C.L.; Parrott, L.J.; and Yu, S.W. "Diffusion in Cementious Materials: Further Investigations of Chloride and Oxygen Diffusion in Well-Cured OPC and OPC/30%PFA Pastes", Cement and Concrete Research, Vo1.24, No.4, 1995, p.819-826. [29] Chatterji, S. "On the Applicability of Pick's Second Law to Chloride Ion Migration through Portland Cement Concrete", Cement and Concrete Research, Vol. 25, No.2, 1995, p. 299-303. [30] Sandberg, P. "Critical Evaluation of Factors Affecting Chloride Initiated Reinforcement Corrosion in Concrete", Report TVBM-3068, University of Lund, Sweden, 1995. [31] Hussain, S.E.; Rasheeduzzafar; Al-M.usallam, A.; and Al-Gahtani, A.S. "Factors Affecting Threshold Chloride for Reinforcement Corrosion in Concrete", Cement and Concrete Research, Vol. 25, No.7, 1995, p. 1543- 1555. 552 Kurtis and Mehta [32] Kayyali, O.A. and Haque, M.N. "The Cl-/OH Ratio in Chloride- Contaminated Concrete- a Most Important Criterion", Magazine of Cement and Concrete Research, 47, No. 172, Sept. 1995, p. 235-242. [33] Diamond, S. "Chloride Concentrations in Concrete Pore Solutions Resulting from Calcium and Sodium Chloride Admixtures", Cement, Concrete, and Aggregates, Vol. 8, No.2, Winter 1986, p. 97-102. [34] Lambert; Page, C.L.; and Vassie, P.R.W. "Investigations of Reinforcement Corrosion 2: Electrochemical Monitoring of Steel in Chloride- Contaminated Concrete", Structures and Materials, 24, 1991, p. 351-358. [35] Page, C.L. "Mechanisms of Corrosion Protection in Reinforced Concrete Marine Structures", Nature, Dec. 1975, p. 514-515. [36] Hime, W. and Erlin, B., "Some Chemical and Physical Aspects of Phenomena Associated with Chloride-Induced Corrosion". Corrosion. Concrete. and Chlorides, ACI SP-102, p. 1-12, Ed. F.W. Gibson, 1987. [37] Glasser, F.P. and Sagoe-Crentsil, K.K. "Steel in Concrete: Part II. Electron Microscopy Analysis", Magazine of Concrete Research, Dec. 1989, p.213- 220. [38] Wang, K. and Monteiro, P.J.M. "Corrosion Products of Reinforcing Steel and Their Effects on the Concrete Deterioration", currently in press. [39] Mehta, P.K. "A Disscussion of the Paper on 'Osmotic Pressure and Swelling of Gels' by L.S. Dent Glasser", Cement and Concrete Research, vol.lO, n.l, Jan. 1980, p. 123. [40] Figg, J. "Salt, Sulfate, and Other Chemical Effects", Intemational Experience with Durability of Concrete in the Marine Environment, Ed. P. K. Mehta, Jan. 16-17, 1989, p.49-69. [41] Mehta, P.K. "Concrete Technology at the Crossroads- Problems and Opportunities", Concrete Technology: Past, Present, and Future. ACI SP- 144, 1994, p. 1-30. TABLE 1-RANGE OF REPORTED C 1'/0H THRESHOLD VALUES Researcher Experimental Method CI-/OH Ratio Hausmann [18] Steel bars in chloride-contaminated alkaline solution 0.6 Diamond [33] Steel bars embedded in concrete subjected to internal chloride attack. (made from Type I slightly greater cement (9.1% C3A);w/c=0.50; with 0.5% by than 0.3 weight CaC!z admixture) Lambert, Page Steel bars embedded in concrete subjected to and Vassie [34] external chloride attack; no details on mix approximately proportions reported 3 Concrete Durability 553 i.li' ::r: Cl) "' --- ;> 2: Fe20J (;l 0 -G-- Fe++ 0 '+=I -== -- 2 corrosic;'r1------ _ 0 0.... immunity Fe 2 2 0 4 8 12 16 pH Fig. 1-Pourbaix diagram (potential vs. pH) for iron in an aqueous solution [171 Fig. 2-Schematic illustration of the corrosion process of reinforcing steel in concrete (adapted from [22]) 554 Kurtis and Mehta Reinforced Concrete I Weathering and Loading Effects l Increase in the Permeability of Cover I
Penetration of Water, J Oz, COz, and Cl-
Loss of OH- Ions from I Cement Paste 1 A: (i.) Depassivation of the Steel Reinforcement (ii.) Formation of a Colloidal Rust B. Gradual Loss of Adhesion by C-S-H A: Expansion of Rust Increases the Hydraulic Pressure in Pores B: Reduction in Concrete Strength and Stiffness - Expansion and Cracking of Concrete I Fig. 3-A model of reinforced concrete damage from exposure to aggressive environments [42]