Corrosion

Corrosion S Virtanen, University of Erlangen-Nuremberg, Erlangen, Germany & 2009 Elsevier B.V. All rights reserved. Introduction Corrosion can be generally defined as degradation of materials in a reaction between the material and its en-vironment. The nature of the reactions leading to deg-radation depends on the class of materials: for metals, corrosion is an electrochemical process, whereas ce-ramics can fail by purely chemical dissolution. This art-icle mainly discusses the corrosion processes of metallic materials, that is, electrochemical corrosion reactions. First, the thermodynamic and kinetic principles of electrochemical corrosion reactions are introduced, fol-lowed by a section on metals passivity. The most relevant modes of corrosion processes are briefly described. Physical–chemical principles of corrosion protection are described. At the end, selected examples of corrosion phenomena encountered in electrochemical power sources are briefly discussed. Basics Thermodynamics Corrosion reactions of metallic materials are of electro-chemical nature, that is, the reaction can be divided into an oxidation reaction (anodic reaction) and a reduction reaction (cathodic reaction). In the case of corrosion, metal oxidation reaction (eqn [I]) is coupled with the reduction of species in the environment/electrolyte – the two most typical reactions taking place in aqueous en-vironments are the reduction of protons (eqn [II]) or reduction of oxygen dissolved in water (eqn [III]). Me-Menþ þneÿ ½IŠ 2Hþ þ2eÿ-H2 ½IIŠ O2 þ2H2Oþ4eÿ-4OHÿ ½IIIŠ The anodic and the cathodic reactions can either take place statistically distributed all over the surface, leading to uniform corrosion, or be separated at specific cathodic and anodic sites of the surface, leading to localized cor-rosion. The latter phenomena (localized corrosion) take place when heterogeneities exist either on the material surface (e.g., different phases of a multiphase alloy) or in the environment (e.g., water-line corrosion of partially immersed objects owing to inhomogeneous oxygen dis-tribution in the depth). Owing to the electroneutrality requirement (i.e., all electrons produced in the anodic reaction must be con-sumed in the cathodic reaction), the oxidation and re-duction reactions must take place simultaneously and with an equal rate. To complete the circuit, the anodic and the cathodic sites must be electrically and electro-lytically connected. With the help of simple thermodynamic consider-ations as described in eqns [1] and [2], a first assessment of the risk of corrosion can be made by looking at the electrode potentials of the oxidation (anodic) and re-duction (cathodic) reactions DG ¼ ÿnFV ½1Š V ¼ EC ÿEA ½2Š where DG is the change in the Gibbs free energy, n is the number of electrons exchanged in the reaction, F is the Faraday constant, V is the cell voltage, EC is the potential of the cathodic reaction, and EA is the potential of the anodic reaction. The electrode potentials for the partial reactions are obtained from the standard potentialswith the help of the Nernst equation (eqn [3]), which takes into account the influence of temperature and activities of species par-ticipating in the reaction E ¼ E1 þRT lnPaox ½3Š red where E is the electrode potential, E1 is the standard electrode potential, R is the ideal gas constant, T is the temperature, n is the number of electrons exchanged in the reaction, F is the Faraday constant, a is the activity, and n is the stoichiometric factor. As an example, Table 1 lists the standard electrode potentials for some metal electrodes as well as for the most important oxidizing elements (these values were taken from the CRC Handbook of Chemistry and Physics). By comparing the electrode potentials of the metal elec-trodes with the potentials of the hydrogen and oxygen electrodes, it can be concluded that all metals with a standard potential o0V normal hydrogen electrode (NHE) are expected to corrode in the presence of water (and in the absence of oxygen), whereas metals with a standard potential 40V NHE corrode only in the presence of an oxidizing agent, the standard potential of which is higher than the standard potential of the metal. 56 Electrochemical Theory | Corrosion 57 Table 1 Standard electrode potentials for some Me/Menþ as well as O2/OHÿ and Hþ/H2 electrodes (all reactions written in the anodic direction) Reaction E1 (V vs SHE) Au-Au3þ þ3eÿ þ1.498 Ag-Agþ þeÿ þ0.7996 4OHÿ-O2 þ2H2Oþ4eÿ þ0.401 Cu-Cu2þ þ2eÿ þ0.3419 H2-2Hþ þ2eÿ 0.00 Pb-Pb2þ þ2eÿ ÿ0.1262 Ni-Ni2þ þ2eÿ ÿ0.257 Co-Co2þ þ2eÿ ÿ0.280 Fe-Fe2þ þ2eÿ ÿ0.447 Cr-Cr3þ þ3eÿ ÿ0.744 Zn-Zn2þ þ2eÿ ÿ0.7618 Al-Al3þ þ3eÿ ÿ1.662 Mg-Mg2þ þ2eÿ ÿ2.372 SHE, standard hydrogen electrode. 1.6 1.2 (b) 0.8 Fe3+ 0.4 Passivity 0.0 Corrosion II Fe2O3 (a) −0.4 Fe2+ −0.8 I −1.2 Fe ‘immunity’ Metallic Fe is stable −1.6 Fe3O4 HFeO2 1 3 5 7 9 11 13 It should be considered, however, that the standard potentials of course are valid only for standard con-ditions. Hence, for instance, the presence of metal cation complexing species, or anions leading to precipitation of salts of the metal cation and the anion, changes the ac-tivities of the dissolved cations and hence the potential of the metal electrode. Moreover, for many metals, solid oxide layers can form on the surface under specific conditions, and the formation of these oxide films strongly influences the corrosion behavior. In the best case, the oxide films form a very efficient kinetic barrier against metal dissolution – in this case, the oxide films are called passive films. In the case of passivity, the metal dissolution rate is many orders of magnitude lower than active dissolution. Therefore, to assess the risk of cor-rosion, it is not only sufficient to know whether a metal oxidation reaction will take place or not, but also infor-mation on the final state of the metal cation (dissolved or bound in a solid layer on the surface) is required. The thermodynamic considerations for different possible reactions for a metal/H2O system can be sum-marized in potential–pH diagrams, the so-called Pour-baix diagrams (such diagrams were compiled by Marcel Pourbaix in 1963). Figure 1 shows as an example the Pourbaix diagram for iron. The diagram gives regions of stability for different dissolved or solid species as a function of pH and potential. The dotted lines represent the equilibria for the hydrogen and oxygen electrodes. The region between these two lines hence indicates the stability region of water. The solid lines represent the calculated equilibria for the different relevant electro-chemical and chemical reactions considered; the regions between these lines represent the stability ranges for iron and its corrosion products. Such diagrams therefore en-able the prediction of whether the metal will be inert (no reaction: thermodynamic stability region of the metal), activelydissolve (stability region of Menþ), or be covered pH Figure 1 pH–potential diagram for Fe/H2O. The dotted lines (a) and (b) show the pH dependence of the oxygen electrode and the hydrogen electrode. Redrawn based on the data given in Pourbaix M (1963) Atlas d’equilibres electrochimiques. Paris: Gautier-Villars & Cie. by surface oxides (stability region of Me oxides). As can be seen from the iron diagram, iron is not inert in aqueous solutions, but passivation by iron oxides becomes possible with a suitable combination of potentials and pH. Such pH–potential diagrams can be used for the first assessment of the (thermodynamic) stability of a system. Since these calculations are based purely on thermo-dynamic considerations, this approach gives no infor-mation on the rate (kinetics) of the possible corrosion reactions. Even if a reaction is thermodynamically pos-sible, its rate can be so negligible that it lacks practical relevance. Moreover, in the case of real practical systems, effects such as temperature variations or presence of different aggressive or inhibitive species in the environ-ment lead to a corrosion behavior, which hardly can be predicted by the simple pH–potential calculations for Me/H2O systems. Kinetics The purely thermodynamic considerations give the first idea of the risk of corrosion. For the practical case, of course, information on the kinetics of the corrosion re-actions is required. The kinetics of electrochemical re-actions can be described and studied bycurrent–potential curves (the so-called polarization curves). For electro-chemical corrosion reactions, different rate-controlling factors can be dominant: activation energy control (also called charge-transfer control), mass-transport control, or passivation. 58 Electrochemical Theory | Corrosion Activation energy control (charge-transfer control) stems from the fact that ions or electrons must be transferred through the electrochemical double layer across the metal/electrolyte interface. Active metal dis-solution and hydrogen evolution reaction at low pH values are typical electrochemical reactions, which are under activation control. As the activation energy de-pends on the applied potential, the current–potential relationship follows the so-called Butler–Volmer equa-tion, showing an exponential i/E relationship. Alternatively, mass transport in the solution can be-come rate determining. The reaction is then diffusion (or diffusionþmigration) controlled. In this case, the charge-transfer process is fast compared with the dif-fusion of the reacting species to the surface or of the dissolving cations away from the surface. In aqueous solutions, diffusion control is frequently encountered for the cathodic oxygen reduction reaction: the reaction rate depends on the supply of oxygen (gas) to the electrode surface. In this case, a limiting current density, which is independent of the applied potential, is observed in the i/E curves. The limiting current density depends on the diffusion constant, the bulk concentration of the diffusing species, and the hydrodynamic conditions. In the corrosion process, the anodic and cathodic re-actions occur simultaneously and with the same rate (owing to the law of conservation of charge). The cou-pled situation of both redox equilibria is described by the so-called ‘mixed potential theory’. The mixed oxidation (e.g., Fe/Fe2þ) and reduction (e.g., OHÿ/O2) systems will equilibrate to zero net current. The resulting potential that lies between EFe2þ=Fe and EO2=OHÿ is called the corrosion potential (Ecorr). Figure 2 schematically illus-trates the situation for Fe in an oxygen-containing neu- tral solution, assuming active dissolution of iron. The anodic reaction hence is under activation control, and an exponential potential dependence is observed for the current. The oxygen reduction reaction is diffusion-controlled showing over a wide potential region a po-tential-independent limited current density. (For further details on the kinetics of electrochemical reactions, the i reader is referred to many books on electrochemistry or physical chemistry.) As illustrated in Figure 2, at the corrosion potential the sum current is zero, and hence the corrosion current density cannot be directly determined from the polar-ization curves. However, plotting log(i) versus E enables the determination of dicorr by extrapolating the linear parts of the curves to the corrosion potential. The cor-rosion current density can be used for the calculation of, for instance, mass loss, by using the Faraday’s law. Passivity As already mentioned, the rate of the metal dissolution reaction is very strongly retarded if a passive film is formed on the surface. Figure 3 shows a schematic po-larization curve for a metal showing an active/passive transition. The critical current density (see Figure 3) is a measure of the passivation ability: the lower the icrit, the easier the passivation. In other words, alloys with a low critical current density for passivation require less oxi-dizers (e.g., oxygen) present in the environment than alloys with a high critical current density. An example of alloying leading to a strongly enhanced passivation ability is Fe–Cr alloys: addition of chromium to iron strongly decreases the critical current density, and this is one of the reasons for the superior corrosion resistance of high-chromium-containing iron base alloys (e.g., stainless steels). The passive current density (ipass), on the con-trary, is a measure of the protective quality of the passive film. Typically for alloys with highly protective passive films, passive current densities o1mAcmÿ2 are observed. The beneficial effect of, chromium in Fe–Cr alloys is further demonstrated by a decrease of the passive current density by increasing, the chromium content. It should be mentioned that the anodic polarization behavior of any metal or alloy of course depends on the environment, as corrosion is a property of the material/environment system, not a materials property. Therefore, not only alloying can lead to enhanced passivity, but also changes in the environment can enable passivation, and this may – for some systems – be an efficient way of corrosion protection (use of inhibitors leading to spontaneous passivation; an example is alkalization of the environment Fe icorr Ecorr O2 + 2 H2O + 4e− Fe2+ + 2e− i icrit E 4 OH− ipass Epass E Figure 2 Schematic polarization curves for Fe (actively corroding) in neutral solutions, with a dominant cathodic reaction Figure 3 Schematic polarization curve for a metal showing an of oxygen reduction. active/passive transition. Electrochemical Theory | Corrosion 59 enabling passivation of iron). In this context, it should also be mentioned that the role of dissolved oxygen is very different for active metals and for metals with an active/passive transition. For actively dissolving metals, an increase in the oxygen concentration increases the corrosion rate, but for metals with a possibility of pas-sivation, a sufficient amount of oxygen (or another oxi-dizing agent) leads to spontaneous passivation. Corrosion of a passivated metal is strongly influenced by the ionic and electronic properties of the passive film. This is illustrated by the schematic representation of a metal surface covered bya passive film in Figure 4. On the one hand, dissolution in the passive state can take place by cation transport through the passive film. The barrier properties of the film of course depend on the chemical composition, structure, and defectiveness of the oxide film. On the other hand, dissolution of the oxide film in the electrolyte can take place combinedwith a re-formation of the oxide film at the metal/oxide interface. Passivity hence is a dynamic equilibrium between film dissolution and re-formation. Highly protective passive films therefore should have a low solubility in a variety of environments. The dissolution of the oxide film can be purely chemical (e.g., aluminium oxide (Al2O3) in acidic or alkaline solutions) or electrochemical in nature (e.g., dissolution of chromium oxide (Cr2O3) under sufficiently oxidizing conditions owing to the formation of soluble CrO 2ÿ, or reductive dissolution of iron oxide (Fe O ) owing to the formation of soluble Fe2þ). Furthermore, for the corrosion behavior in the passive state, the electronic properties of the passive film are of relevance, as they determine whether the cathodic counterreaction requiring electron transfer through the metal/electrolyte interface can take place on the passive surface.Inaddition,upon anodicpolarization,a significant difference is seen between metals with well-conducting passive films and those with insulating passive films: if the passive film enables electron flow in the anodic direction, at potentials above the anodic water de-composition potential strongoxygen evolution takes place, limiting the use of high voltages for anodization (to achieve thicker oxide layers on the surface). In the case of the so-called valve metals (e.g., aluminum, tantalium, hafnium, niobium, titanium, tungsten), the oxide layer is Dissolution of n+ the passive film O2 OH– Electrolyte e– Passive film (Me-oxide) Re-formation of the passive film Figure 4 Schematic illustration of reactions taking place on a metal surface covered by a passivating oxide layer. blocking under anodic bias, and therefore anodization can be carried out up to high voltages until dielectric break-down of the oxide film – this opens up many possibilities for the modification of the metal surface. Moreover, the electronic properties of passive films become highly rele-vant when a passive metal is used in applications where electronic conductivity plays a role. Consequently, the chemistry, structure, and the elec-tronic properties of passive films on many technically important metals and alloys have been widely studied. As the nature of the passive films depends not only on the alloy itself but also on all passivation parameters such as potential, time, temperature, and chemistry of passiv-ation, it is out of the scope of this article to give details of the nature of the passive films. For this, the reader is referred to review articles by J. W. Schultze and M. M. Lohrengel (2000) and P. Schmuki (2002). It should be mentioned, however, that a detailed studyof the nature of passive films formed under ambient conditions (low temperatures and low potentials) is challenging owing to the low thickness of the films; typically do10nm. Hence, highly surface-sensitive techniques are required for the study of the properties of passive films. Nowadays, there is relativily good agreement on the general nature of the passive films on the technically most important materials such as iron, stainless steels, and aluminium. As an example, for stainless steels (Fe–Cr alloys), it is very well known that a strong Cr enrichment takes place in the passive film (for instance, for an alloy containing 18% chromium, over 80% of chromium oxide can be found in the passive film). Details of the chemistry of the passive film are more difficult to summarize, as the Cr/Fe ratio and depth distribution in the passive film, as well as the thickness of the film, vary for different pas-sivation conditions (e.g., pH of the solution, time of passivation, potential). Regarding the crystal structure of the passive films, the use of sophisticated techniques such as electrochemical scanning tunneling microscope (STM) has revealed interesting insights into the structure of passive films on some metals and alloys. The electronic properties of passive films have also been explored by manygroups. Passive films on many metals and alloys are of semiconductive nature. However, in many cases, pas-sive films have been found to show a very high doping density, owing to the high defect density in the films. As the exact chemistry of the passive film on an alloy de-pends not only on the alloy composition, but also on the environment, also the electronic conductivity of a passive film formed on a certain alloy can vary in a certain range. It should be mentioned that apart from the low rate of passive dissolution taking place across the passive film as schematically illustrated in Figure 4, also specific corrosion processes owing to localized breakdown of passivity can take place in the passive region. Electro-chemically, the onset of localized dissolution phenomena 60 Electrochemical Theory | Corrosion can be observed by a sudden increase of the anodic current density at a specific potential in the anodic po-larization curve (the so-called breakdown potential). The value of the breakdown potential again depends on the material and on the environment. Some specific corrosion modes are described in brief in the next section. Different Corrosion Phenomena Galvanic Corrosion In practical applications, corrosion processes can be significantly more complex than either active dissolution of the surface or passivity. First, when different materials are combined in a device or in a construction, these materials (when electrically and electrolytically con-nected) can form a galvanic cell. The galvanic coupling accelerates the dissolution of the less noble metal, whereas the dissolution rate of the more noble metal is retarded (or even completely stopped). The driving force for galvanic corrosion is the difference in the electrode potentials of the materials. Factors limiting the rate of galvanic corrosion are the kinetics of the anodic and cathodic reactions on the materials surface and the conductivity of the electrolyte. If the kinetics of the re-actions are strongly hampered, even a large driving force (DE¼V) will not lead to strong galvanic corrosion ef-fects. The conductivity of the electrode determines the potential and current distribution, and hence the effec-tive distance of galvanic coupling. A further important factor influencing galvanic corrosion is the area ratio of the anode and the cathode, a small anode coupled with a large cathode being the most deleterious case. Pitting Corrosion From the mechanistic point of view, manydifferent types of localized corrosion phenomena are interesting. Mostly, these types of attack are encountered with passive metals and alloys, which do not fail by uniform dissolution of the surface. Pitting corrosion shows an attack morphology of small distinct pits on the surface, surrounded by passive, unattacked surface. Pitting corrosion takes place in the presence of specific aggressive anions in the solution, often halide ions – for most metals and alloys, chlorides lead to highest susceptibility to pitting attack. Some of the technically important materials prone to pitting corrosion are aluminium and aluminium alloys, iron (when in passive state), and stainless steels. However, the susceptibility to pitting corrosion strongly depends not only on the alloy group, but also on the exact alloy composition. For instance, in stainless steels, increased concentrations of chromium and molybdenum increase the resistance against pitting corrosion, whereas impur-ities – especially sulfur – stronglydecrease the resistance. In technical alloys, typically surface heterogeneities such as inclusions or second-phase particles act as initiation sites for pitting corrosion. A large body of information on pitting corrosion of metals has been compiled by Z. Szklarska-Smialowska (1986). Crevice Corrosion The mechanism of crevice corrosion is similar to pitting corrosion. Crevices can be present on the metal surface because of the construction, but a crevice type of situ-ation can also evolve in practice, for instance if deposits of another material form on the metal surface. Alloys that are susceptible to pitting corrosion typically also fail by crevice corrosion, but the crevice exacerbates the situ-ation so that less critical environments (e.g., lower Clÿ content) are required for the crevice corrosion to take place. This is due to the fact that in narrow crevices a local aggressive environment can easily form, involving oxygen depletion, chloride-ion enrichment, and acidifi-cation – all this hampers repassivation reactions and ac-celerates dissolution. Similar local chemistries evolve in growing pits, but in the case of crevices no pit initiation step is required, as the crevice inherently acts as an ‘occluded site’ hindering exchange of solution with the environment. Intergranular Corrosion Intergranular corrosion describes preferential dissolution along the grain boundaries and is related to precipitation of second-phase particles at the grain boundaries and solute depletion in the vicinity of the grain boundaries. A very well-known case is intergranular corrosion of stainless steels, owing to precipitation of chromium-rich carbides in the temperature range of E450–8501C (as can be encountered for instance in the heat-affected zones of weld joints) at the grain boundaries, and as a consequence the formation of chromium-depleted zones in the vicinity of the grain boundaries. As chromium in solid solution is the alloying element crucial for the passivation ability of Fe–Cr alloys, the chromium-de-pleted zones will activate and hence preferentially dis-solve. The remaining passive surface of the matrix acts as a large cathode to drive the dissolution of the grain boundary zones. Other alloys prone to intergranular corrosion owing to metallurgical effects are certain alu-minium alloys in specific heat treatment conditions. Mechano-Chemical Attack In addition, corrosion can take place in a conjoint action of mechanical stresses and chemical attack, or wear and corrosion. The reader is referred to text books on cor-rosion for more information on stress corrosion cracking, corrosion fatigue, and tribocorrosion. ... - tailieumienphi.vn