After a century of history of cathodic protection (CP) of iron and steel, this paper critically reviews the state of the art in the science and engineering and assesses the fitness of CP as an effective technology to tackle the challenges related to infrastructure corrosion. This paper focuses on CP of iron-based alloys embedded in porous media, such as soil or concrete, as these two major applications of CP technology share many similarities. First, the scientific understanding of CP is reviewed and different competing theories are discussed. There is wide agreement that corrosion protection of steel is achieved thanks to a combination of immediate activation polarization and the beneficial changes in electrolyte chemistry that are gradually occurring at the steel surface when a protection current is flowing toward a steel electrode. A major and well-documented technological advantage of these “chemical effects” is that the protective effect of CP is maintained during temporal loss of protection current, e.g., due to survey work related shut-offs or anodic interference. However, the relationships between these chemical concentration changes in the porous medium and the protection current are complex, and, as this review shows, cannot reliably be described with state-of-the-art approaches. Moreover, in this paper, different hypotheses for the mechanism of corrosion protection in heterogeneous situations (galvanic elements), as they are generally occurring in practice, are discussed. It is revealed that understanding the working mechanism of CP in heterogeneous conditions remains a critical scientific challenge. The longstanding debate concerns the question whether CP results mainly in a reduction of number and size of actively corroding areas, or in a reduction of the corrosion rate at the actively corroding sites. Additionally, the literature addressing the interrelation between microbiologically influenced corrosion and CP is here reviewed, and recent progress as well as limitations of the existing literature are highlighted. In a second part, engineering practice and CP protection criteria are reviewed. It is found that the approaches stipulated in international standard are unreliable. This can be traced back to the assessment criteria being empirical and incapable of adequately taking into account the complexity of the underlying processes. Finally, recommendations for future developments are made. Particular opportunities are seen in embracing the progress made in numerical modeling, such as reactive transport modeling in porous media, and considering the interdependence between the involved processes, namely the interdependence between transport processes, chemical reactions, and electrode kinetics.
INTRODUCTION
Cathodic protection (CP) is an electrochemical technique to control corrosion of metals. The technology is applied worldwide and across various industrial sectors to protect metallic structures and devices from corrosion, including underground structures, marine exposed structures, ship hulls, or heat exchangers.1-5 CP has a long history, dating back to Volta, Davy, and Faraday in the early 19th century.6 However, CP of iron or mild steel—as opposed to CP of other metals, such as copper—received research attention only in the beginning of the 20th century.2,7-9 This is important because the mechanism of corrosion protection of iron-based alloys is fundamentally different from CP applied to other metals.
Perhaps the first experiments investigating the current density needed to protect iron from corroding were performed by Bauer and Vogel and documented in 1918 in a comprehensive report of the Königliches Materialprüfungsamt in Berlin, Germany.7 The authors tested various galvanic elements involving different metals (such as zinc and iron) in a number of different aqueous electrolytes. By varying the geometry (area ratio and distance of the two electrodes) and the electrolyte conductivity, they seeked to determine the current density that was “just high enough to protect the iron electrode from corroding.” They concluded that this required protection current density was of the order of 0.1 A/m2. Similar protection current densities were later also reported in the dissertation by van Wüllen Scholten.8 Furthermore, extensive field tests with CP of iron were made in the late 1920s by Kuhn, when CP was applied to soil-buried cast iron pipe.9 A major result of these studies was the criterion for the potential to ensure protection (−0.85 VCSE ) that led to long-standing controversies until today.
While CP of iron and steel in soil has thus a track-record spanning over a century, the application of CP to reinforced concrete came later. First laboratory trials and subsequent application to the San Mateo-Hayward bridge in California are documented from the 1950s by Stratfull.10
Today, the need for effective corrosion control techniques is greater than ever, which is illustrated by the staggering socio-economic impact of corrosion.11-16 The worldwide direct costs related to corrosion of infrastructure—bridges, pipelines, etc.—is estimated at more than $1 trillion annually. Over the coming decades, the situation is expected to aggravate, because all industrialized countries face the problem of large stocks of aging infrastructure.14,17-18 The need for repair and replacement of civil infrastructure is expected to increase by a factor of 2 to 5 by 2050.19 The high societal relevance of infrastructure corrosion is the main motivation to—after 100 years of history of CP of steel—critically review the state of the art in the science and engineering in order to assess the fitness of CP as an effective technology to address the grand corrosion engineering challenges related to infrastructure aging.
The idea of CP is to polarize the metal structure that needs corrosion protection in cathodic direction by imposing an electrical current. This may be achieved by establishing a galvanic element with a sacrificial anode (that is, to connect the metal structure to a less noble metal) or by impressing a current with the help of a DC current source. Figure 1 schematically shows this for a soil buried structure and for reinforced concrete. In both cases, the metal is surrounded by a porous medium (soil or concrete).
The presence of a porous medium significantly influences the processes involved in CP. A major effect is that the porous medium limits transport processes, e.g., restricting convection and diffusion, and thereby controlling the electrolyte composition at the metal surface. The beneficial role of a porous medium adjacent to the steel surface is well known also from CP in seawater, where porous layer formation is fundamental for the CP to be effective.20 However, an important difference between CP of steel in a solution and CP of steel in soil and concrete is that in the latter cases, the porous medium is present initially and likely affects the corrosion situation before applying CP. For instance, different local properties at the metal/porous medium interface likely give rise to the formation of galvanic elements across the metal structure (Figure 1). Another commonality of CP in soil and reinforced concrete, distinguishing this fundamentally from CP of structures immersed in water, is that the porous medium may exhibit moisture states below saturation, which strongly influences ion transport processes within the porous medium and the availability of oxygen at the metal surface. In summary, CP of reinforced concrete and soil buried structures share significant similarities in terms of the underlying scientific mechanisms.
A major difference between CP in atmospherically exposed concrete and CP in soil, however, is that the distance between the impressed current (or sacrificial) anode and the protected steel is only a few centimeters in concrete, while this distance can easily be of the order of miles in CP of soil buried structures. This difference has major implications for the current distribution. Hence, anode positioning needs more careful attention in reinforced concrete compared to CP of metallic structures in soil.
This paper focuses on the working mechanisms of CP at the corroding (to be protected) steel structure. Aspects related more to technological questions such as sacrificial and impressed current anode materials and systems, including anode bedding and overlay systems, are not addressed here. Interested readers may find information in the literature.21-22
ADVANCES AND OPEN QUESTIONS IN THE SCIENCE OF CATHODIC PROTECTION
Electrochemical Kinetics
If a uniformly corroding metal electrode is by an external current polarized in cathodic direction, the rate of the anodic metal dissolution reaction will immediately be reduced. This is because a cathodic shift in potential kinetically reduces the anodic reaction rate according to the reaction kinetics that may be described with Tafel slopes or Butler-Volmer kinetics, as it is commonly presented in text books.5 This is schematically illustrated in Figure 2. The extent to which the anodic iron dissolution rate is slowed depends on the Tafel slope (ba) and the potential shift (ΔE). This shift of potential is generally termed a change in activation polarization. Note that according to this purely kinetic reasoning, as shown in Figure 2, in order to completely arrest corrosion, the potential would need to be shifted below the reversible potential of the iron dissolution reaction (that is, to the so-called immunity domain of iron23 ). However, as will be discussed later in this review, this is difficult to achieve due to the relatively high demand of protection current density in the corresponding potential range.
Another important aspect apparent from Figure 2 is that the applied protection current not only reduces the anodic reaction rate, but simultaneously leads to an increase in cathodic reaction rate. This enhanced cathodic reaction means an increase in the rate of generation of hydroxyl ions at the electrode surface, which almost inevitably leads to a change in the chemical composition of the electrolyte at the metal surface and thus affects both the thermodynamics and the electrochemical kinetics of the system subjected to CP. This means modifications of the current-potential relationships, deviating with time from those shown in Figure 2. These chemical effects and their impact are discussed in detail below. Later in this review, the impact of deviations from uniform corrosion conditions is also discussed. This is considered important as in engineering practice, uniformly corroding electrodes (such as the one assumed in Figure 2) are rare, which limits the applicability of considerations based on the assumption of uniform electrodes to conditions in practice.
The Recognized Importance of the pH and the Challenge of Predicting It
In recent years, various researchers came independently to the conclusion that it is the increase in pH, arising from the generation of OH− ions at the cathodically protected steel surface, which plays an important role in achieving corrosion protection of steel in soil,24 namely in promoting passivation.25-39 Thus, it is now recognized3 that effective corrosion protection with CP technology is largely achieved by favoring passivation, namely by changing the chemical environment of the steel, rather than by polarizing the steel into the immunity domain (i.e., the domain where iron is thermodynamically stable23 ). These chemical changes occurring upon CP include the removal of oxygen (due to oxygen reduction) and—if present—chlorides (due to migration) from the metal surface, as well as the generation and accumulation of OH− at the steel surface.
It is well known that concentration changes in reactants (or products) of electrochemical reactions at an electrode surface lead to so-called concentration polarization.12 Figure 3 shows an example of this, that is, the change in electrode potential (b) occurring upon changing the O2 concentration and the pH (a) of an electrolyte. It may be worth highlighting that in producing these results, the electrolyte was modified by gas purging and adding NaOH, and that no current was applied to the electrode throughout the experiment. The Pourbaix diagram (Figure 3[c]) illustrates how decreasing the O2 concentration shifted the potential toward the reversible potential of the hydrogen electrode (stages I and II), and how increasing the pH moved the electrode into the passive domain (stages II and III). In aerated conditions (stages III and IV), the potential increased, which can be explained by the potential of passive steel being dominated by the oxygen concentration. The figure thus illustrates how corrosion protection can be achieved without reaching the immunity domain, but thanks to beneficial changes of the electrolyte composition. In the example shown in Figure 3, these chemical changes were made by decreasing the oxygen concentration and by increasing the alkalinity of the electrolyte, in order to illustrate their impact on the corrosion situation. It is exactly these changes that occur during CP as a result of the electrochemical reactions (Figure 2) taking place. It may be of interest to note that the importance of the accumulation of hydroxide ions at the steel surface for corrosion protection was already postulated in 1928 by Kuhn,9 however, not yet in the context of passivation.
Further indication that corrosion protection from CP can effectively be achieved by gradually changing the chemical environment of the protected electrode rather than by polarizing the steel to immunity is the fact that the immunity domain is located at potentials more negative than the reversible potential of the hydrogen evolution reaction23 (Figure 3[c]). Because iron has a small hydrogen evolution reaction overvoltage compared to other metals,40 relatively high current densities are needed to reach immunity. Accordingly, a protection current density of the order of 1 A/m2 would be needed to reach the immunity domain by pure activation polarization.41-43 Numerous studies, however, have demonstrated that 10 to 100 times lower protection current densities are generally sufficient to arrest corrosion.5,7-8,21,44-45 In this context, it may be worth drawing the attention to data published in Enos, et al.,46 as shown in Figure 4. In this study, hydrogen permeation was measured in mild steel coupons (self-contained Devanathan/Stachurski cells) cast in concrete. The coupons were located in differently exposed zones (submerged, splash, and atmospheric zone) and subjected to different levels of CP currents. Figure 4 shows that to obtain significant hydrogen evolution (and to reach immunity), current densities greater than 0.8 A/m2 were required.
The extent to which it is possible to increase the pH at the steel surface in soil by imposing a given protection current density, however, depends strongly on the environmental conditions including soil electrolyte chemistry,26 bedding conditions and soil microstructure,31,36 advection in moving water25,47 (stagnant vs. streaming water), and the influence of microorganisms.30 Thus, the empirical relationships between protection current density and achieved surface pH, as observed in various experimental studies and under different conditions related to CP in soil,26,30-31,33-34,36-37,48 diverge strongly (Figure 5). Additionally, the steepness of the gradient in pH from the protected steel surface into the soil also differs for different conditions. While the pH was found to drop to the level of the bulk solution within a few millimeters from the metal surface25,49 (in solution tests), other researchers reported the highly alkaline zone to extend from the steel surface into the soil in the order of centimeters or even decimeters.37,50-51 These differences can again be explained by influences such as the ionic strength or the pH buffer capacity of the electrolyte, advection, soil microstructure, but also the experimental time of the different studies.
As mentioned above, the working mechanism of CP of steel in concrete52 shares many similarities with that of CP in soil. There is wide agreement that CP applied to steel in concrete raises the pH at the steel surface and repels chloride ions.28,53-62 These beneficial effects are in the literature termed “secondary effects,” meaning that the “primary effect” offering corrosion protection is (activation) polarization, viz. shifting the potential in cathodic direction.63 However, a number of groups demonstrated that the instantaneous protective effects of CP are negligible (even at high CP current densities) and that time is needed to establish protection.55,62,64-66
This behavior is explained by the gradual changes in the environment (pH increase, chloride removal) that finally lead to corrosion protection. Further support for this hypothesis stems from theoretical considerations showing that usual CP current densities are far too low to lead to an immediate potential shift of 100 mV.67-68 However, the fact that 100 mV polarization can be achieved with these relatively low current densities over longer time spans indicates that the “secondary effects” are the main reason66 for the effectiveness of CP of steel in concrete. Moreover, a number of authors suggested that to re-passivate corroding sites, it is beneficial (or even needed) to initially deliver a high amount of charge; once passivity is restored thanks to the concentration changes at the steel surface, relatively low CP current densities are sufficient to maintain passivity.59,66,69-70
The importance of beneficially changing the chemistry in the porous medium adjacent to the steel surface during CP is also evident from studies that showed that corrosion protection remains active for prolonged times after the protection current is discontinued, both in soil71-73 and in concrete.61,64-65,74 It may be interesting to mention that this had already been known since the 1930s, as is apparent from the historical photograph shown in Figure 6. At that time, ensuring electrical power supply for CP was a challenge especially for locations remote from populated areas. Kuhn, who used a windmill to provide electrical power, wrote that despite the variability in wind, “the polarization and film formation on the cathodically protected pipes carries protection over into the (…) no wind periods.”71 Thus, Kuhn already acknowledged that upon temporal interruption of the protection current, it can take considerable time for the system to depolarize, which he explained by formation of an alkaline film at the steel surface. Upon switching off CP, the established gradients in electrolyte chemistry will start to gradually vanish.
The time needed to homogenize the electrolyte chemistry depends significantly on the volume of the porous medium in which the pH had been raised as well as on other influences including ion transport properties in the pore system, possible advection, and chemical reactions such as buffer effects (Figure 7[a]). It is well known that the associated time of depolarization, that is, the decay time of concentration polarization as apparent from potential measurements over time, may scatter over a large range, namely from minutes to days.2 The fact that this is much slower than the time needed for activation polarization to vanish, which according to Schwenk75 and von Baeckmann76 is a fraction of a millisecond, is further indication for the protective action to be related to changes in chemical environment rather than pure activation polarization. The relatively slow loss of corrosion protection upon current interruption is an important property of CP contributing to the generally positive track record of this technology, as temporal loss of protection current may occur during surveys, power outage, or due to anodic interference caused by telluric activity or time variant stray currents.
Another aspect worth highlighting is the fact that the pH needed for corrosion protection (passivation) may be different from case to case. This is because of the influence of other chemical species in the electrolyte, such as the presence of chlorides, sulfides, or carbonates. Changes in ferrous ion concentrations, such as due to complexation with carbonates, extend the domain of active corrosion in the Pourbaix diagram,77-78 which gives rise to an active dissolution domain above pH 9. This may lead to the paradox situation that polarizing cathodically can increase the corrosion rate, as was observed by Schwenk79 in carbonate buffers. On the other hand, the precipitation of siderite may contribute to the formation of protecting or at least advection limiting layers, the latter promoting an additional increase in pH. This depends on numerous factors such as the soil chemistry and the bedding conditions, and is in present approaches of ensuring and assessing the effectiveness of CP not explicitly taken into account.
In summary, the concentration changes—particularly related to OH− ions—occurring at the steel surface and in the pore system adjacent to it when a protection current is flowing toward a steel electrode play a crucial role in achieving corrosion protection and retaining it for a certain time during temporal interruption of the protection current. However, the relationships between these chemical concentration changes in the porous medium and the protection current are complex and depend on many factors, including ion transport properties, the local pore structure of the porous medium (soil, sand), chemical reactions (e.g., pH buffering), advection in moving water, or the influence of microorganisms. Thus, it is impossible to predict these beneficial concentration changes based on simple engineering approaches, e.g., merely based on current density (Figure 5). At the same time, there is a lack of mechanistic, quantitative models to predict the concentration changes at the steel surface based on the relevant physical and chemical parameters.
One of the main reasons for this lack of fundamental models can be found in the complexity that arises from the interdependence of the involved processes. Figure 7 schematically illustrates the different processes and how they are coupled. The next section, Advances in Numerical Modeling of Cathodic Protection, discusses the extent to which numerical models have attempted to consider these processes. Nevertheless, the present lack of fundamental models hampers making recommendations about which protection current density is needed—or which protection criteria in terms of potentials have to be satisfied—in order to generate a chemical environment at the steel surface of structures in their actual exposure conditions to ensure corrosion protection.
Advances in Numerical Modeling of Cathodic Protection
Numerical modeling is used to predict current and potential distributions on electrode surfaces and in the electrolyte (soil, concrete) to design and study CP. Both finite element modeling (FEM) and boundary element modeling80 (BEM) are well-established (even used in combination, e.g., Brichau and Deconinck81 and Liu, et al.82 ) and integrated in commercial software.83-84 The conceptual approaches fall in different classes, as summarized in Figure 8. Early approaches typically applied the so-called potential theory,85 based on the Laplace equation and different boundary conditions. For electrode boundaries, either uniform current/potential distributions (yielding so-called “primary current distributions”),81 or linear/non-linear boundary (static) electrode kinetics in the form of current-potential relationships such as Tafel law or Butler-Volmer expressions (yielding “secondary current distributions”) were used.69,82,86-97 Recently, more refined approaches that also consider ion transport have been used.98-101 This permits calculating “tertiary current distributions.”102-103 A number of authors have used reactive transport models to take into account precipitation reactions occurring in the electrolyte, typically within crevices at disbonded coatings.104-105 Some authors focused on modeling transport in crevices or in soil, however, did not explicitly model corrosion kinetics at the electrode, but assume constant source or sink terms for the considered species.50,106
In recognition of the time-dependent changes at the electrodes, such as formation of calcareous deposits in CP protected structures in seawater and the related changes in electrode kinetics, suggestions were also made to take these effects into account.107 Nisancioglu20,108 used experimentally determined decays of current density with time (in seawater) to describe the nonstatic electrode kinetics with so-called dynamic polarization curves (for the cathodic reactions). This conceptual approach of empirically considering the changes in potential-current density (E-j) relationship of the electrode as a function of time (but not explicitly depending on physical or chemical conditions) was later also adopted for soil leachates109 and other aqueous solutions.110-111 For CP in concrete, similar (empirical) approaches were also—although rarely—used to take into account the changes of the polarization behavior of sacrificial zinc anodes112-113 and reinforcing steel114 as a function of time or charge passed. Regarding cathodic protection in soil, a model using dynamic electrode kinetics on a fundamental basis was proposed by Büchler.115-116 This model explicitly considers the electrode kinetics of both the anodic and cathodic reactions as a function of pH and oxygen concentration; however, the transport in the soil is treated in an empirical manner and, moreover, does not take into account chemical reactions.
Thus, while over the last 20 years significant advances were made regarding the numerical modeling of various aspects of CP, the field has been reluctant to develop models that integrate all relevant physical, chemical, and electrochemical processes—and especially their interdependence (Figure 7). Taking into account that electrode kinetics, release/consumption of species, and reactive transport are coupled is, however, essential for reliable predictions. This is because upon transition from activation controlled anodic dissolution to a passive behavior, the electrode kinetics change dramatically. It is increasingly being recognized that considering the interdependence of the processes (Figure 7) presents a major opportunity in the development of future numerical CP models. Various authors have highlighted the need to consider the changes in electrode kinetics that result from CP-induced changes in electrolyte chemistry.69,94,99,114,117
The Question of Heterogeneous Corrosion Conditions
Exposed steel surfaces in both soil and concrete rarely show uniform corrosion. It is well documented that the corrosion attack is typically localized because of the formation of local galvanic elements.118-120 This can be explained by the heterogeneity in local microstructure and environment at the steel/soil and steel/concrete interface.121-123 Particularly in aerated soils where steel surfaces are only partially wetted or where the thickness of the contacting electrolyte layer is variable,124 there is a high probability for establishing differential aeration cells within short distances. As was early illustrated in the famous Evans droplet experiment and later summarized by Pourbaix,125 this differential aeration “produces an increase in the corrosion rate in the non-aerated regions, and a decrease in the corrosion rate in the aerated regions.” Galvanic elements within coating defects were confirmed, e.g., by the scanning vibrating electrode technique.126 Another possible cause for the formation of galvanic elements in soil is the presence of microorganisms at the steel surface.127 For steel in concrete, particularly for chloride-induced corrosion, the corrosion is typically of localized nature with confined anodic zones and larger zones of cathodic reinforcing steel.119-120 Due to the macroscopic nature of the arising galvanic elements, that may in concrete easily span over several tens of centimeters, the term macro-cell corrosion is typically used.10,119
The question of how CP acts on heterogeneous electrodes and the actual mechanism potentially leading to corrosion protection in the presence of local galvanic elements has been controversially discussed for a long time. In the 1930s, Mears and Brown47 used galvanic elements formed with zinc and copper as a model system. They concluded that “in order to obtain complete cathodic protection, it is necessary to polarize the cathodes in the corrosion cell to the open circuit potential of the local anodes.” Mears and Brown used both the terms “open-circuit potential of the anode” and “unpolarized potential of the anode” to describe the corrosion potential of the anodic metal (zinc) in the actual electrolyte when not galvanically coupled to the copper. It is important to recognize that this “unpolarized” or “open-circuit potential” (OCP) referred to by Mears and Brown is different from the reversible potential of the zinc (the potential at which the metal dissolution and the metal deposition reactions are in equilibrium). In fact, the OCP is anodic to the reversible potential. This implies that the anodic dissolution rate of the anode is not zero at the OCP, although it may be low. Thus, in stating that “it is necessary to polarize the cathodes (…) to the [OCP] of the local anodes,” Mears and Brown essentially proposed to suppress the galvanic element. Figures 9(a) and (b) illustrate that the corrosion rate of the anodes will be strongly decreased as the anodic current leaving the metal at the anodic sites is to a large extent offset by the impressed CP current, but not entirely suppressed. The extent to which these changes in current density at local anodes and cathodes occur was suggested by Mears and Brown to depend on the Tafel slopes of the anodic and cathodic partial reactions (Figure 9[b]).
It is worthwhile mentioning that this hypothesis considers CP as a protection process that would immediately be active once the impressed current is switched on and that would immediately vanish once this current is switched off (which disagrees with early practical experience, see Figure 6). Any time-dependent changes, such as due to modifications of the electrolyte chemistry, are neglected in the Mears and Brown theory. In this regard, it may be interesting to mention that Evans, in a comment to the contribution by Mears and Brown,47 noted that in case the protection current “produces an inhibitive cathodic product” such as an alkaline solution, the situation is expected to be different.
An alternative hypothesis considering time-dependent changes was postulated in 1965 by LaQue and May (their original publication from 1965 was reprinted in 1982128 ). The key element of this theory is that the effect of CP applied to heterogeneous electrodes is to progressively increase the area of the cathodes at the expense of the local anodic areas (Figure 9[c]). This theoretical proposal was later experimentally confirmed by Dexter, et al.,129 who showed that the release of OH− at the cathodic areas leads to lateral spread of alkalinity and as a result, the cathodic areas grow. An important implication of this is that small remnants of anodically behaving surfaces may exhibit large corrosion current densities, which may mean high rates of localized corrosion. In other words, if the level of CP is insufficient to completely eliminate the anodic sites, the applied protection current may create a situation in which the corrosion attack is even enhanced locally.
The second of the above hypotheses is underpinned by observations made by various researchers. In field tests in aerated soils, Funk, et al.,130 observed that coupons exhibited very non-uniform corrosion rates, and that local corrosion rates significantly exceeded the average rates and amounted up to 0.03 mm/y despite common protection criteria for aerated soil being satisfied. Similarly, Barlo44 found increased local corrosion rates in a field test with uncoated steel pipe, where even at IR-free potentials as negative as –1.1 VCSE the corrosion rate was locally reported to be 0.1 mm/y. Funk, et al.,130 noted that the formation of galvanic elements is more difficult the smaller the coating defect. Other authors have also shown that with increasing exposed steel surface area, both the susceptibility for corrosion and the variability in corrosion behavior increase.73,131-133 Based on this and on the references cited above, a size effect is expected, namely that the risk for non-uniform conditions increases with increasing coating defect size.
In summary, the principles of ensuring corrosion protection in heterogeneous situations remain unclear. This applies particularly to the case of local galvanic elements, where it is difficult to guide the protection current by positioning the anode to certain areas of the galvanic couple, because the anode system will always be on a relatively remote position with respect to the two electrodes involved in the galvanic element, at least for CP of steel structures in soil. The longstanding debate concerns the question of whether CP results mainly in a reduction of number and size of actively corroding areas, or in a reduction of the corrosion rate at the actively corroding sites (not affecting their size).28,128-129,134-135 Understanding this remains a fundamental challenge that needs resolving to establish criteria for the effectiveness of CP.
Microbiologically Influenced Corrosion in Soil
Sulfate reducing bacteria (SRB) are anaerobes, thriving in pH values up to 10. SRB reduce sulfate to sulfide by utilizing various electron donors, namely molecular hydrogen, organic compounds (e.g., lactic acid), or—as found recently—directly metallic iron.136-138 The first hypothesis of SRB influenced corrosion, suggested in 1934, became known as the “cathodic depolarization theory.”139 The authors proposed that SRB use hydrogen as the sole electron donor to reduce sulfates to sulfides, and that the microbial consumption of H2 enhances the corrosion rate. Today, it is widely accepted that this theory does not adequately describe the mechanism by which SRB influence corrosion, and that the microbial consumption of hydrogen, although it may occur, plays no more than a secondary role.138,140-141 More modern theories take also into account the role of biogenic sulfide, i.e., H2S generated by SRB, which combines with ferrous ions to form iron sulfide films (FeS). Under some conditions, these films were found to offer some degree of corrosion protection,142 but deposition of iron sulfide was also found to accelerate corrosion,143-144 e.g., by formation of galvanic elements between the base metal and the iron sulfide film.140 FeS enhances the cathodic reaction,144 either because it catalyzes the reduction of protons or, perhaps more likely, because it enlarges the surface area available for the cathodic reaction.137,141 In recent years, this view considering the influence of H2S as major component in SRB influenced corrosion, which was termed “chemical microbially influenced corrosion” (CMIC),141 was further expanded by taking into account an alternative mechanism, in which certain strains of SRB are able to derive electrons directly from metallic iron, rather than through oxidizing organic matter or molecular hydrogen.136 This mechanism was termed “electrically microbially influenced corrosion” (EMIC).141 Under some conditions, EMIC can lead to much higher corrosion rates than CMIC.141 However, in the presence of organic matter such as lactate, it appears that the few strains capable of promoting EMIC become outcompeted by the other (organotrophic) SRB, which then gives rise to CMIC only. It was thus suggested that further research should distinguish more clearly between organotrophic and lithotrophic cultivation.141
Thus, the mechanisms of microbiologically influenced corrosion (MIC) of steel in soil and the conditions under which MIC occurs still wait to be completely understood. The situation becomes even more complex if the interrelation of MIC and CP is considered. While there is general agreement that a more negative protection potential is needed in the presence of MIC than in their absence, there is debate about the actual threshold value.36 Numerous international standards145-147 specify a protection criterion of –0.95 VCSE (IR-free potential) in the presence of SRB, which can be traced to theoretical (thermodynamic) considerations148 and to experimental studies30,149-150 (all published in the last century). In recent years, however, several authors experimentally observed that even at IR-free potentials in the range of –0.95 VCSE to –1.10 VCSE, corrosion could not be stopped in the presence of SRB,151-154 or that CP promoted the growth of SRB.155 This was explained by the CP-driven evolution of hydrogen, which “feeds” the SRB,152,155 promoting their growth and the related microbial back conversion of H2 to H+, thereby limiting the increase in pH.30,152 In 2016, Guan, et al.,154 hypothesized, by referring to the recent findings regarding direct electron uptake by SRB from metallic iron, that a cathodically polarized steel surface provides an additional source of electrons for SRB. Kajiyama30 also suggested that within an intermediate range of pH, reached upon CP, corrosion protection was achieved by formation of a dense FeS film. Thus, CP may have both beneficial (film formation) and adverse (promoting bacterial growth) effects, but there is a lack of thorough understanding of the conditions under which one of these mechanisms becomes dominant.
In summary, while literature results are contractionary, an increasing number of studies indicate that common international CP standards may lead to insufficient corrosion protection in the presence of MIC. This imposes an urgent need for clarification, which was also highlighted in a recent panel discussion.156 The disagreement among different studies can be traced to the different approaches (thermodynamic/theoretical vs. various experimental methods), and to the fact that common studies made no explicit distinction between the fundamentally different mechanisms of SRB influenced corrosion that can arise from different SRB strains and cultivation conditions. A further shortcoming of traditional experimental studies on CP and MIC is the use of laboratory reactors containing solutions rather than soil, which, due to the pronounced differences in terms of transport properties, is hardly representative of field conditions. Finally, it may be mentioned that understanding the interrelation of CP and MIC is also impeded by the many open questions related to CP in the absence of MIC.
IMPLICATIONS FOR ENGINEERING
International Standards and Cathodic Protection Assessment Criteria
In many applications, e.g., in the oil and gas industry or in concrete structures, insufficient corrosion protection may have catastrophic consequences. Engineering methods thus need to be capable of reliably assessing the effectiveness of the corrosion protection method. Thus, protection criteria are defined in international standards.145-147,157
Most of the protection criteria used today allow one to assess the effectiveness of CP only with a low level of confidence. This is illustrated in Figure 10, which summarizes an extensive evaluation of field tests,44,158 following the methodology of evaluation described in Joos and Büchler.159 The reviewed field studies reported data related to the protection criteria for buried coupons ranging in size from 1 cm2 to 18 cm2, as well as the corresponding corrosion rates, based on weight loss measurements upon excavation or electrical resistance (ER) probes. Here, the assessment of the protection criteria is based on a threshold corrosion rate of 0.01 mm/y, which is in standards145,147 defined as the threshold below which the corrosion rate of a structure is claimed to be so low that corrosion can be considered “arrested.” According to the standards,145,147 this should be the case if the protection criteria are satisfied. However, referring to Figure 10, it is particularly unsatisfactory that in a substantial portion of the situations, state-of-the-art CP criteria suggest corrosion protection (corrosion rate <0.01 mm/y), while this is not the case (actual corrosion rate >0.01 mm/y). These cases, termed “wrong ‘dangerous’” in Figure 10, expose the public to risks. In an even higher portion of the situations, the CP criteria suggested critical conditions, although this was not the case (“wrong ‘conservative’”). Conservatism is not only uneconomic, but may also lead to overprotection and the related adverse effects of CP, such as coating disbondment or hydrogen embrittlement and stress corrosion cracking.46,160 It may be noted that the popular 100 mV criterion1,98,161 is the least reliable of the evaluated protection criteria, as only in approx. 20% of the situations was this criterion able to correctly indicate the corrosion state of the coupons.
A general problem with engineering approaches, both for CP in soil and concrete, is that they are largely based on so-called IR-free potentials.147,162 As will be discussed in the next section, A Critical Scientific Analysis of Cathodic Protection Assessment Criteria, this is a parameter that can hardly be measured—especially on structures (rather than coupons)—for a number of different reasons.36 Another example of problematic engineering is the fact that the different requirements stipulated in the different standards for CP in soil are contradictory, even within the same family of standards (e.g., European standards147,163-164 ). In soil-buried structures, more negative on-potentials are needed to achieve the required IR-free potentials145-147 and to ensure protection in the presence of stray currents.164 However, less negative on-potentials are required for protection against AC corrosion163 and prevention of overprotection (hydrogen embrittlement and coating disbonding).145 In many cases, these requirements leave only a very narrow window of CP operation, or even are contradictory, thus, the relevant standards may not simultaneously be satisfied. Thus, it may for the practitioner be impossible to comply with the codes of practice.
A Critical Scientific Analysis of Cathodic Protection Assessment Criteria
For a detailed historic evolution of CP criteria in standards, one is referred to the literature.1,165-166 It becomes apparent that protection criteria are largely based on empirical observations, such as those previously reported,9,24,130 rather than on scientifically rigorous concepts. Accordingly, protection criteria and their interpretation have been subject to long-standing controversies.1,36,38,48,159,165-168 Much of the controversy can be traced to poor understanding of the working mechanism of CP, essentially the incomplete distinction between activation polarization (Figure 2) and polarization arising from concentration changes (Figure 3), the lack of acknowledging passivation (and the resulting dramatic changes in electrode kinetics), not adequately considering galvanic elements, and overlooking the impact of diffusion potentials.
Significant discussion relates to the question of whether or not potentials should be free of IR drops.169-172 A frequently held view is that IR drops—arising from current flow through soil or concrete of limited conductivity—affect the interpretation of potential measurements and thus need to be corrected for, before they can be compared with protection criteria such as the –850 mVCSE criterion.145-147,170 Other authors suggested that the effect of IR drops on assessing the effectiveness of CP may in many situations of practical relevance be overrated,171-172 especially in the presence of galvanic elements on the structure, because the galvanic corrosion current (as illustrated in Figure 1) partially or completely offsets the IR drop arising from the protection current.173
Potentials that are corrected for the IR drop are termed IR-free potentials145,147 (or polarized potentials145-146 ). Determining IR-free potentials in practice, however, is not straightforward. A common approach is to utilize instant-off potentials in order to “closely approximate the potential without IR drop.”146 This means that a potential reading is taken after interruption of the protection current. General recommendations are to measure the potential between 0.1 s and 1 s after interruption the CP current.145,174 This, however, is somewhat surprising when considering that the time needed for both activation polarization and ohmic drops to vanish is shorter than 0.1 s. The time constants of activation polarization and ohmic effects have been reported to be of the order of 10−4 s…10−1 s, and 10−7 s, respectively.75-76,175 Thus, by the time one takes instant-off potential readings, activation polarization has also most likely decayed to considerable extent, while significant—however, hardly quantifiable—concentration polarization is still present.175-176 It may thus be concluded that instant-off potentials are essentially a measure of some residual concentration polarization, which may explain why instant-off potentials allow in some cases to assess the corrosion state, but that the conceptual understanding of instant-off potential measurements as an adequate approach to remove the IR drop while maintaining activation polarization may be questioned. In agreement with this, various authors indicated that instant-off potentials hardly provide reasonable estimates of IR-free potentials (polarized potentials),36,175-177 and that the instant-off technique may not be an adequate approach to remove IR drops.98,178
Additional phenomena that complicate the interpretation of both instant-off potentials and criteria based on depolarization (such as the so-called 100 mV criterion1,98,161 ) are related to galvanic elements. First, it is well known that the potential (or potential decay) that is recorded in the presence of galvanic elements is highly sensitive to the location of the reference electrode with respect to the location of anodic and cathodic sites on the protected structure.98,178 In fact, this geometrical dependency is the basis for a well-established technique to detect corrosion, e.g., in reinforced concrete structures.179-180 This, however, implies that it is challenging to generalize a depolarization criterion, such as the 100 mV criterion, without considering the location of the reference electrode with respect to the anodic and cathodic sites on the protected structure. Second, equalizing currents, arising upon interrupting the protection current applied to galvanic elements, are well-known to disturb instant-off potentials.31,72,181 These problems are correctly acknowledged in standard EN 13509162 (valid for soil) that excludes the use of instant-off potentials for non-uniformly polarized electrodes. However, standard EN 12696157 (valid for concrete) strongly relies on instant-off measurements in exactly this case—which is rather stunning.
A final comment should be made to the influence of diffusion and streaming potentials. While these disturbing effects are well known in other fields,182-187 they are generally ignored in CP. Diffusion potentials arise in the presence of concentration gradients in an electrolyte and from differences in ionic mobility.188 Streaming potentials are caused by the movement of water and thus the movement of charge, combined with its interaction with charged surfaces. Both diffusion and streaming potentials, when present between the position of the working electrode and reference electrode, arithmetically add to the potential at the steel/porous medium interface, see Figure 11. Diffusion potentials are often dominated by differences in pH. As became apparent in the first part of this review, CP likely leads to marked differences in pH from the protected steel surface into the bulk, particularly in soil. Thus, diffusion potentials that may easily exceed 120 mV182-184 are expected. Without doubt, this is an order of magnitude that has to be considered relevant for the interpretation of potential measurements in CP.189 In fact, diffusion potentials may often exceed the IR drop.173 This means that considering the contribution of diffusion potentials in the interpretation of measured potentials is essential for a scientifically sound understanding of the working mechanism of CP. Ignoring diffusion potentials renders the validity of theoretical reasoning—e.g., thermodynamic considerations based on measured potentials, that are not corrected for the shift caused by diffusion potentials—questionable.
CONCLUSIONS
This review presents a critical assessment of the scientific and engineering state-of-the-art of cathodic protection of steel in porous media such as soil and concrete. The following major conclusions are drawn:
- •
Applying a cathodic current to a corroding steel surface has a twofold effect: (i) activation polarization, leading to an immediate kinetic reduction in the iron dissolution rate, and (ii) gradual beneficial changes in electrolyte chemistry, namely the generation of corrosion inhibiting species (hydroxyl ions) and the removal of corrosive species (chlorides, oxygen) at the metal surface. There is broad evidence in the literature that it is these (time-dependent) chemical effects, which are generally responsible for achieving corrosion protection of iron and steel, and not necessarily the polarization into the immunity domain.
- •
A major and well-documented technological advantage of these chemical effects is that corrosion protection is maintained also during temporal loss of cathodic protection current, e.g., due to maintenance or survey work related shut-offs, power outage, or anodic interference caused by telluric activity or stray currents.
- •
The relationship between protection current density and resulting changes in chemical composition of the electrolyte at the metal surface and thus corrosion protection is complex. This is because of the numerous influencing factors, including the properties of the porous medium (microstructure, the presence of reactive phases), the electrolyte chemistry, water movement (stagnant vs. convection), and the possible presence of microorganisms.
- •
Understanding the mechanism of corrosion protection in heterogeneous situations (that is, galvanic elements, which commonly occur in practice) remains a scientific challenge that needs to be resolved in order to devise reliable protection criteria.
- •
An assessment of state-of-the-art CP assessment criteria revealed that these engineering approaches are unreliable. In a substantial fraction of cases, they either suggest corrosion protection while this is not the case, or, if too conservative, they may lead to overprotection and the related adverse side effects. This can be traced back to the assessment criteria being empirical and incapable of adequately taking into account the complexity of the underlying processes.
- •
Major opportunities for future developments are seen in embracing recent scientific advances in numerical modeling, such as reactive transport modeling in porous media as well as devising novel models that are capable of considering the interdependence between the involved processes, namely the interdependence between transport processes, chemical reactions, and electrode kinetics (as schematically sketched in Figure 7). Fundamentally rigorous models will pave the way toward new engineering tools that will allow us to abandon empiricism and thus increase the effectiveness and reliability of CP as a technology to mitigate corrosion of both soil buried and reinforced concrete structures.
ACKNOWLEDGMENTS
The author is grateful to the Swiss National Science Foundation for the financial support (SNSF, project no. PP00P2_163675) and to the European Research Council (ERC) for the financial support provided under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 848794). Furthermore, the author wishes to thank Dr. Markus Büchler and David Joos from the Swiss Society for Corrosion Protection for the valuable discussions. Finally, the author wishes to thank Federico Martinelli Orlando for his support in carrying out the measurements shown in Figure 3.