6.2 Electrochemical Nature of Molten Salt Corrosion

The loss of material is substantially attributed to the corrosion reaction, and therefore the material surface suffers from the homogeneous or heterogeneous attack of corrosive media. For the former case, Wagner and Traud (1938) described the fact that the occurrence of the corrosion reaction necessitates simultaneous dissolution of metals and reduction of oxidant. According to their theory, the location of the metal dissolution is not necessarily identified with that of oxidant reduction at the metal–solution interface. Therefore, the important factors in corrosion are not only impurities, defects, and other heterogeneities of the material but also the chemical nature of the oxidant involved in the liquid (i.e. impurities and by‐products due to dissociation). From this viewpoint, the corrosion reaction in molten salts is a fairly complex phenomenon compared with that in an aqueous medium where only a few chemicals such as oxygen and protons might be candidates for the oxidant of the corrosion reaction.

It should be noted that while in aqueous solutions metals are virtually insoluble in the electronated state, in molten or fused salts (which can be regarded as infinitely concentrated aqueous solutions), they can be appreciably soluble and corrosion occurring, in such cases, without the de‐electronation of the metal. In spite of this, the electrochemical approach for molten salt systems was put forward about 55 years ago and is finding general acceptance.

Based on the mixed potential theory, the corrosion reaction is expressed by the combination of metal dissolution, partial anodic reaction and reduction of oxidant, and partial cathodic reaction under the restriction of electric charge neutrality: thus, for the partial anodic reaction,

6.1

and for the partial cathodic reaction,

6.3

where M, Ox, and R represent the metal, oxidant, and reductant, respectively, and n is the number of electrons. Accordingly, in a molten salt, there are several oxidants (O₂, H⁺, H₂O, and OH⁻) common to the molten salts, and such oxidants are associated with the dissociation reaction of molten salts themselves.

At this stage and before moving further to thermodynamic considerations, it is important to briefly describe the meaning of the electrode potential, which is the central electrochemical parameter.

6.3 The Single Potential of an Electrode

In the most general sense of the word, an electrode is a system consisting of two phases in contact with each other, which can be the seat of an electrode reaction, i.e. a reaction in which certain constituents of the two phases participate and by which a transfer of charge takes place from the bulk of one phase to the bulk of the other. In a more restricted sense, an electrode is defined as a metal–electrolyte or metal–solution system of one or several electrolytes.

It has been shown over the last 100 years that at the interface of two phases, one finds an electrical double layer with a characteristic potential (Adam 1938; Butler 1940).

The origin and the characterization of the electrochemical double layer on a metal immersed in an aqueous electrolyte have been extensively studied. Similar studies in melts have been carried out only more recently. This section is not intended to deal with this aspect in detail, but the subject is mentioned because of its importance. Very good reviews on the subject have been published by Devanathan and Tilak (1965), Graves et al. (1966), and Ukshe et al. (1964).

Returning to the subject of this section, let the inner electric potential of the metal be Φ_M and that at a remote point in the molten electrolyte be Φ_S. Moreover, let ΔΦ represent the Galvani electric potential across the electrochemical double layer, i.e. ΔΦ = Φ_M − Φ_S. When a dynamic equilibrium across the double layer is set up, the electrochemical energy of 1 mol of metal in the surface will be equal to the electrochemical free energy of 1 mol of ions on the melt side of the double layer:

6.4

or

and

6.5

where μ_M − μ_S is the chemical energy balanced by the electrical energy liberated as Mⁿ⁺ ions traverse the double layer, F is Faraday’s constant, and n is the number of free electrons exchanged at the Mⁿ⁺/M interface. It is impossible to measure ΔΦ directly, but if the metal–melt electrode system is coupled “back to back” with a second arbitrarily chosen electrode system, it is possible to obtain a relative potential difference. If a zero potential and a convenient reference scale of potentials is defined, the combined system acts as a cell between those electrodes and there exists an electromotive force (e.m.f.) given by

6.6

where E_0.4 is called the single potential of the Mⁿ⁺/M electrode on the chosen standard reference scale. Considering now a more concise electrode reaction,

6.7

where M_i represents the constituents of the electrode taking part in the reaction and e⁻ represents the electron; the corresponding conditions of dynamic equilibrium can be expressed in the general form

6.8

Relating the chemical potentials, μ_i, to the standard chemical potential, , Eq. 6.8 can be put into the form

6.9

Molten salt	Temperature (°C)	Equilibrium reaction	pO2−
KCl–LiCl, KCl	800
(Na, Li, K)2SO4	600
(Na, Li, K)2CO3	600
(Na, K)NO3	250

Changing from natural to Briggsian logarithms, the last relation can be rewritten in the form

6.10

in which has the following value in terms of the standard chemical potentials of the reactants:

6.11

Equation 6.10 is the Nernst equation. is the single potential of the electrode measured when the activity of every constituent is equal to unity and is known as the normal or standard potential. Once the normal electrode potential on the standard reference scale is known, the Nernst equation allows a calculation of what the equilibrium electrode potential will be when the system has any {M_i} ≠ 1. Extensive analytical treatments of this subject are given by Conway (1965), de Groot (1951), Grahame (1947), Guggenheim (1929), Parsons (1954), and Prigogine (1947). Experimental proof of the validity of the Nernst equation in molten media is given by Blander et al. (1959), Duke and Garfinkel (1961), Gordon (1899), Ranford and Flengas (1965), and many other workers.

6.4 Equilibrium Diagrams

Thermodynamic considerations can predict whether a metal is stable or whether it will corrode when it coexists with the oxidants common to molten salts. When the metal dissolution in Eq. 6.1 and the oxidant reduction in Eq. 6.3 are taken into account, the free energy of the global reaction associated with them can be negative, i.e. the potential difference ΔE(= E_0.3 − E_0.1) is positive, and therefore corrosion proceeds spontaneously. Since the metal reacts with chemical species such as oxide ions and oxyanions that constitute the molten salts, the equilibrium potential varies with the activity of the oxide ion in the melts. Hence, the diagram of chemical entities in the melt as a function of the electrode potential and the basicity is helpful in a comprehensive understanding of the corrosion behavior. The basicity of the melt is defined as pO²⁻ () on the basis of the Lux–Flood regulation (Flood and Förland 1947; Lux 1939) (see also Section 3.3.1). It is recognized that each molten salt has an appropriate acid–base equilibrium involving its anion in the case of the oxyanion system. For example, the basicity expressed as the logarithm , assuming an acid–base equilibrium in the alkali sulfate melt:

6.12

where K_T indicates the equilibrium constant of Eq. 6.12. Table 6.1 shows typical acid–base equilibria and the pO²⁻ equation for these molten salts.

The equilibrium diagrams E–pO²⁻ and its construction are fully described in Section 3.3.1 for the iron/sodium sulfate system at 1173 K. Similar diagrams for molten salts including single, binary, and ternary salt mixtures are also published in the open literature (Ingram and Janz 1965; Kunst and Duke 1963; Littlewood 1962; Marchiano and Arvia 1972; Rahmel 1968; Sequeira and Hocking 1977).

Thermodynamically calculated E–pO²⁻ diagrams have resulted in remarkable progress in understanding corrosion, but this method has several (qualitative) disadvantages for the realistic requirements of design engineers. In particular, the domains of thermodynamic stability considered in the diagrams give only a theoretical possibility of existence, not a certainty. Further on, as far as passivation is concerned, it must be pointed out that a proper passivity in molten media has not been obtained up to date, perhaps due to special properties of the film involved, mainly adherence, coherence, and deviation from stoichiometry. Therefore, elaborate kinetic techniques for obtaining precise corrosion rates are needed.

6.5 The Tafel Relationship

Considering now the dynamic side of the corrosion phenomenon, particular importance will be attached to electrochemical kinetic studies. A classic relationship precisely relating a characteristic parameter of the electrode, its electrode potential, and the current density traversing it is that found by Tafel in 1905. Therefore this semiempirical law enables us to establish a connection between the “thermodynamic electrode potential” and the current density that defines the kinetic character of the phenomenon.

The Tafel equation may be written as

6.14

where E is the relative electric potential of the electrode at a current density i (also called “reaction‐electrode potential”), a is a constant characteristic of the electrode, and b is the Tafel slope, which is one of the parameters indicating the mechanism of the electrode reaction.

Thus, a definition of overpotential for the reaction can be introduced by the expression

6.15

where E₀ is the reversible or equilibrium electric potential possessed by the electrode at electrochemical equilibrium (i.e. at zero imposed current) or, more generally, the equilibrium potential of the electrochemical reaction of the type (Eq. 6.7). Equation 6.15 defines the overpotential, and considering Eq. 6.14, it is seen that the Tafel equation is itself closely related to the overpotential (η). These problems of electrochemical kinetics are discussed in more detail later.

The brief discussion given above is significant enough for the present purpose of showing that the Tafel relationship leads to the experimental polarization curves. Figures 6.1 and 6.2 represent the anodic and cathodic potential–current curves for two typical electrochemical reactions, with either an arithmetic scale for the reaction current (Figures 6.1a and 6.2a) or with a logarithmic one (Figures 6.1b and 6.2b). The exchange current density (i₀), which is defined as the exchange rate per unit area of the potential‐determining electrode process at equilibrium, is represented in Figures 6.1b and 6.2b. The aspect of the polarization curves is very important because it gives an account of the degree of irreversibility of the reactions. In fact, in the case of a reversible reaction (Figure 6.1), the application of very small polarizing potentials is sufficient for producing significant current densities. In the case of irreversibility (Figure 6.2), on the contrary, the existence of high overpotentials is a measure of the extent of irreversibility. The reversible or fast electrode processes exhibit high exchange current densities; on the other hand, the irreversible or slow ones exhibit low exchange current densities. Thus, if an electrode is more likely to behave reversibly, the higher is its intrinsic exchange current density.

The Tafel curves giving E as the function of log i also provide interesting results about the kinetics of the phenomena. In fact, the plot of E against log i is not usually a straight line, i.e. the kinetics of the reaction does not obey the Tafel relation. The experimental curves obtained can be considered as built up of rectilinear portions or, on the contrary, as showing systematic departures from the linear Tafel relation for certain current densities reached. It is well established in electrochemistry that the Tafel relation does not apply precisely for small and large current densities. One reason for this is that the Tafel equation is derived purely by considering the overpotential due to polarization caused by charge transfer requirements, which assumes that a cation moves away from the metal‐fused electrolyte interface as soon as it is dissolved. This is generally not the case. The removal of the anodic product does not increase in the same proportion as the current density and the concentration of these products will increase and cause a back e.m.f. This effect is particularly relevant in the case of fast electrode processes as observation of Figures 6.1 and 6.2 shows: the difficulty in obtaining a pure Tafel slope, i.e. unaffected by double layer and mass transfer effects, increases with kinetics of the electrode processes (i.e. with high i₀) due to increasing overlapping of the diffusion zones over the Tafel one. Polarization measurements also include ohmic overpotential that arises from an IR drop through a portion of the electrolyte and between the test electrode and the reference one, but this term is not significant in molten media as compared with an aqueous system. The validity of the Tafel equation has been verified by Laitinen and Gaur (1957), Randles and White (1955), and others for many fused systems.

The facts mentioned above are very important especially due to their repercussion on the interpretation of electrochemical phenomena. So, it is seen, by this simple enumeration, that it is possible to connect thermodynamic considerations to kinetic ones.

6.6 Corrosion Potential–pO²⁻ Relationship

In the section above, where some concepts of electrode kinetics were discussed, a quantity of crucial importance to electrodics was defined: the overpotential (η). In electrochemical kinetics, the only reaction directly affected by the potential is the charge transfer reaction, i.e. the reaction in which charge carriers are transferred across the electrochemical double layer at a phase boundary. The rate of the charge transfer reaction determines the charge transfer or activation overpotential. This kind of overpotential has been treated in detail by Audubert (1942), Bard and Faulkner (1998), Bockris (1954), Butler (1924), Conway (1965), Erdey‐Gruz and Volmer (1930), Gerischer (1960, 1961), and others.

Prior to discussion, it is noted that only the basic electrodic equations will be considered.

Considering the conversion of M⁺ ions into metallic M, the number of moles of positive ions reacting per second by crossing unit area of the melt–metal interface is proportional to the electronation current density i and as a first approximation (Erdey‐Gruz and Volmer 1930):

6.16

where is the activation energy required for the electronation process at the potential E, K₋ and are constants, {M⁺} is the activity of cations M⁺ on the metal side of the interface, is the chemical activation free energy barrier when there is zero electric field acting on the ion, and α is the transfer coefficient for the de‐electronation reaction introduced by Erdey‐Gruz and Volmer (1930) that has values ranging from zero to unity (0 < α < 1) (Vetter 1952). K₋ is dependent on the reference potential chosen since . Moreover, positive ions can move from the melt to the electrode and also in the opposite direction. Thus, there is an electronation reaction

6.17

and also a de‐electronation reaction

6.18

In this case, the total current density passing through the double layer is

6.19

where {M} is the activity of the metallic atoms immediately at the electrode surface (unity). If the metal–melt interface reaches its equilibrium state, there is no net current (i = 0), but there is an exchange current. The electronation and de‐electronation reactions continue to occur but at the same rate. The current densities corresponding to these individual reactions become the exchange current density i₀ (i₀ = i₊ = |i₋|). Hence,

6.20

where E₀ is the potential difference across the interface at equilibrium. On dividing Eq. 6.19 by the corresponding expressions in Eq. 6.20 and taking into consideration the definition for overpotential η, the following important relationship is obtained (Erdey‐Gruz and Volmer 1930):

6.21

For higher charge transfer overpotentials, |η| > RT/F, the first or second term on the right‐hand side (r.h.s.) of Eq. 6.21 can be neglected depending on the sign of the current so that if this Eq. 6.21 is put into a logarithmic form, a linear relationship results between η and log|i| similar to the Tafel equation:

6.22

Here, the subscripts “+” and “−” to the overpotential denote anodic and cathodic polarization, respectively.

Let us now consider the following reaction of a metal M in a molten sulfate (Burrows and Hills 1966):

6.23

This overall reaction can be considered as the sum of the two half‐reactions:

6.24

6.25

For these two processes to take place simultaneously, it is necessary and sufficient that the potential difference across the interface be more positive than the equilibrium potential for the reaction 6.24

6.26

and more negative than the equilibrium potential of the electronation reaction 6.25 involving electron acceptors (such as SO₃) contained in the melt:

6.27

The behavior for the potential–current curve E(I) that can be envisaged for the process Eq. 6.23 is illustrated in Figure 6.3 by the continuous thick line. (I) E is composed additively of the metal dissolution current I₂₄ (E) and the electronation current I₂₅ (E), and the two are independent of each other at the same potential value. The principle of additive combination of all partial processes at an electrode surface to obtain the total potential–current curve was first formulated by Wagner and Traud (1938). The potential E at zero current is neither E_0.25 nor E_0.24, but a so‐called mixed potential (E_M) for which the total potential–current curve gives I (E_M) = 0. The total metal dissolution current I₂₄ and electronation current I₂₅ are then equal in magnitude but opposite in sign, i.e.

6.28

The rate of corrosion of the metal is obviously given directly by the rate of metal dissolution; hence, the free corrosion current I_c (i.e. the rate at which the metal destroys itself) is given by

6.29

Considering the electrode kinetics, we can thus write

6.30

6.31

where A₂₄ is the sink area and A₂₅ is the electron‐source area (at the metal–electrolyte interface). Assuming that overpotentials are sufficiently large, Eqs. 6.30 and 6.31 become

6.32

The free corrosion potential, E_c, i.e. the uniform potential difference over the surface of the corroding metal, can be evaluated from the expressions above. It follows that

6.33

It is clear from this that the free corrosion potential and the mixed potential are identical.

In accordance with the theory of Lux (1939) and Flood et al. (1952) who assign a definite acid–base equilibrium to all oxyanionic melts, the acid–base equilibria for sulfates can be expressed by the Eq. 6.12 and by the equilibrium constant

6.34

Assuming , it follows that

6.35

where pO²⁻ =  − log {O²⁻} is a nonprotonic function that measures the acidity of the melt, as already reported. The expression 6.35 for the active species SO₃ can be introduced into Eq. 6.33, giving

6.36

Assuming that the ratio of A₂₅ and A₂₄ as well as α₂₄ and α₂₅ is constant during the overall corrosion reaction (which is not free from error), Eq. 6.36 becomes

6.37

with a and b being two constants. Thus, we attain a linear relationship between the corrosion potential and the function pO²⁻, which should be valid for all oxyanionic melts.

What has been presented above is a very elementary account of electrodics of corrosion under quite ideal conditions. The details of the complex corrosion phenomena in real systems are out of the scope of this chapter.

6.7 Electrochemical Polarization and Monitoring

In this section, an account is given of several available methods for obtaining the potential–current curves, the free corrosion rates, and the electrochemical monitoring of corrosion. It is customary to classify the electrochemical methods for obtaining such potential–current curves into potential‐sweep, current‐sweep, potential‐step, and current‐step methods. The former method, sometimes called potentiodynamic (or potentiokinetic) polarization, is accomplished by continuously changing the applied electrode potential (with feedback control) at a constant rate and simultaneously recording current. In the current‐sweep (intensiokinetic) method, the electrolysis current (with feedback control) is varied with time according to a given program, and the electrode potential is recorded as a function of current. In the potential‐step method, electrode potential is rapidly changed over a finite increment, and current measured after a predetermined time interval, and this process is repeated. In the last (above) method, a current step is applied, and the resulting variation of potential with time is recorded. These two‐step methods are also referred to as potentiostatic and galvanostatic (or intensiostatic), respectively, to indicate that at any instant E (or I) is held constant, by means of feedback control.

All the polarization methods described above are well established in the ambient temperature range (Edeleanu 1958; Greene 1962; Greene and Leonard 1964; Pourbaix and Vandervelden 1965; Prazak 1963) for electrochemical studies of metals, the best or most accurate of them depending, of course, on its intended use: if the purpose is to study the kinetics phenomena, the methods in which the electrode potential is controlled are the best; when the permanent physicochemical modifications, due to applied current, are the objective, the methods in which the electrolysis current is controlled are preferable. Generally speaking, however, potential‐step and potential‐sweep methods will be more useful for the study of corrosion phenomena than current‐step and current‐sweep methods because they give more precise indications concerning the quality of the passivation and concerning the conditions of activation. The investigation of corrosion processes in melts by these standard electrochemical techniques was pioneered by a reasonable number of researchers (Arvia et al. 1971, 1972a, b; Baudo et al. 1970; Davis and Kinnibrugh 1970; Kazantsev et al. 1968; Sequeira and Hocking 1978a, b, 1981), and, in general, it can be said that almost all workers in the field of high temperature agree that the polarization methods appear to be quite suitable in corrosion studies at high temperature in molten electrolytes.

The polarization curves have several shapes, but near the potential axis their character is similar to that shown by the curves plotted in Figures 6.1 and 6.2.

A potentiodynamic polarization curve provides a graphic summary of the corrosion characteristics of tested material so that its plot and evolution can be used as a very rapid and practically nondestructive corrosion test. However, the corrosion characteristics obtained are accurate only for the medium in which the curve was obtained. Therefore, the importance of the method does not lie in any final evaluation of the rates of corrosion in a particular medium, but rather in the rapid evaluation of fundamental corrosion properties with the possibility of sensitive relative comparison (Prazak 1963). Another feature of the method is that analysis of such potential–current curves, together with E − pO²⁻ diagrams, enables us to predetermine those reactions theoretically possible and impossible for each electrode potential value.

One of the most important electrochemical values for the determination of the kinetics and mechanisms of primary corrosion processes is the value of the free corrosion current. When the primary corrosion process is slow (absence of concentration polarization), the free corrosion current can be obtained by potentiodynamically polarizing the alloy electrode at a very slow sweep rate in the neighborhood of the free corrosion potential. The alloy dissolution rate is obtained by plotting potential versus log current, and extrapolating the linear anodic and cathodic branches is the free corrosion potential, the intercept giving the free corrosion current. This method can also be applied when some (relatively small) concentration polarization is present; the latter can be taken into account by using the approximate expression

6.38

where I_meas. is the measured current in the presence of concentration polarization and I_d is the diffusion current (of course this correction implies the exact knowledge of the diffusion current). When the primary corrosion processes are sufficiently fast, other methods, namely, relaxation methods, must be used. The objective of these methods is the study of the electrode processes before they become diffusion controlled. The experimental techniques associated with these methods are very complex, but they are the most important ways allowing accurate determination of fast charge transfer corrosion rates. Several relaxation methods for these measurements have been considerably revised by Graves and Inman (1970), and useful exposition of the “classical” relaxation methods is that of Damaskin (1967). The selection of the method to measure the free corrosion currents, depending precisely on rate phenomena, is obviously difficult (because the values of these corrosion rates are not known a priori).

In general, the high temperatures of fused salt systems cause most corrosion processes to be rapid (reversible with respect to the non‐relaxation methods). Generally speaking, therefore, relaxation methods will be the most appropriate for the determination of corrosion rates in melts.

The measurement of free corrosion rates in molten salts was pioneered by Baudo et al. (1970) and Cutler (1971), among others. The first workers determined the corrosion rates using the conventional method, extensively and successfully applied at low temperatures in aqueous systems; they presented the results in the form of Tafel plots, but they do not give information on the current–potential measured values from which the plots of E versus log i were obtained. It is rather surprising in view of the relatively high corrosion rates quoted that it had been possible to obtain pure Tafel zones. It is certain that this method enables the measurement of rough values (more of qualitative than quantitative value) of the free corrosion rates, but it is advisable to put into practice LaQue’s opinion that new and more accurate tests should be introduced for a better understanding of the corrosion mechanisms (LaQue 1969). A cyclic potential‐sweep method is used by Cutler (1971) to determine the free corrosion rates that seems scientifically more appropriate than the conventional one.

The search for accurate methods for the determination of the kinetics and mechanisms of corrosion processes in molten salt systems has led to the study of newer electrochemical methods suitable for corrosion studies in aqueous systems. As a result, DC polarization curves, linear polarization resistance methods, AC impedance, and electrochemical noise/galvanic coupling are continually increasing their application to molten salts, and it can be concluded that apart from being useful additions to investigate high temperature corrosion processes in molten salts, they have demonstrated feasibility for use in online monitoring and for estimating corrosion under realistic operational conditions in boilers, gas turbines, and other high temperature equipment (Mu et al. 1995; Pardo et al. 1992; Rapp and Zhang 1994; Ratzer‐Scheibe 1991; Sequeira 1998). It should be noted that the techniques widely used for the electrochemical monitoring of corrosion in aqueous solutions have not been applied to molten salt corrosion very often because of the limitations of the experimental technique under high temperature conditions. Nishikata and coworkers have investigated the possibility of electrochemical monitoring of molten salt corrosion by the linear polarization resistance and AC impedance techniques (Nishikata and Haruyama 1986; Nishikata et al. 1981; Numata et al. 1983).

The correlation between the polarization resistance and corrosion rate was first derived by Stern and Geary (1957), assuming that both the anodic and cathodic partial reactions obey the Tafel relation. The polarization resistance R_p is written as

6.39

where I_c indicates the corrosion rate and K is a constant that depends on the temperature and corrosion mechanisms. The K value could be determined theoretically from the Tafel slopes of the partial anodic and cathodic polarization curves. However, it is very difficult to measure the Tafel slopes, especially in molten salts because of extremely fast charge transfer rates (Laitinen et al. 1960; Nishikata et al. 1983).

According to the generalized derivation presented by Ohno and Haruyama (1981), the value of K is expressed as

6.40

The terms in parentheses in Eq. 6.40 are the slopes of the partial anodic and cathodic polarization curves on a semilogarithmic scale at a corrosion potential. According to Eqs. 6.39 and 6.40, the polarization conductance I/R_p is proportional to I_c even if a distinct Tafel region is not observed on the polarization curves. The K value can be determined practically by plots of log(I/R_p) measured from the electrochemical method versus logI_c calculated from weight loss measurements. The polarization resistance R_p is determined from the slope of the straight portion of the polarization curve in the vicinity of the corrosion potential as shown in Figure 6.3.

In general, K values differ significantly, and K depends on the natures of the metal and molten salt. For example, in molten chloride, the K value in the absence of oxygen is approximately half that in the presence of oxygen. This is caused by the difference between the cathodic polarization curves (Nishikata et al. 1984).

As a result, if the K value in the molten salt of interest is measured in advance, the instantaneous corrosion rates can be obtained by polarization resistance measurements. For example, as the polarization resistance is possible to measure continuously and to record using an AC corrosion monitor (Mansfeld and Bertocci 1981), the corrosion rates can be precisely determined by the Stern–Geary equation. By this method, the effects of alloying metals and other elements on the corrosion rates have been investigated in diverse molten salts.

6.8 Electrochemical Nature of Metal Oxidation

In the sections above, the basic fundamentals of molten salt electrochemistry are described, and it is shown how high temperature electrochemistry can be applied to understand and mitigate corrosion in molten media. The reader should also give attention to Chapters and , where further electrochemical descriptions relevant to molten salt corrosion are presented. Nowadays, high temperature electrochemistry, and its unique electrolyte features to process materials in nonaqueous environments, provides many opportunities including concentrated solar power systems, nuclear fuel reprocessing, tritium recovery in fusion energy technologies, light metal production, and others. Although high temperature solid‐state electrochemistry and its applications are less well known, namely, in the area of corrosion, it is a fact that the evidence of ionic or electrolytic conductivity in many corrosion products found in high temperature oxidation systems led to the conclusion that corrosion in these dry (or wet) oxide systems is also an electrochemical phenomenon that is important to understand and mitigate by electrochemical means. One specific area of high temperature solid‐state electrochemistry that has been increasing in importance is concerned with the electrochemical behavior of ionic or quasi‐ionic solid oxides. Here, we will begin to show that the metal–oxide systems can be viewed as systems where oxidation occurs by electrochemical processes. Then, particular attention will be given to the electrochemistry of ionic solid oxides or oxide solid electrolytes and the relationships that govern their properties, namely, ionic and electronic conductivities, chemical diffusion coefficients, transport numbers, mobilities, stoichiometry, and thermodynamic factor, among others. More specifically the equations that allow the characterization of the total conduction process in an oxide electrolyte, over a wide range of oxygen partial pressures in both electronic and ionic conduction regimes, are to be examined. These equations are applied in several electrochemical situations, namely, that where current is supplied to reversible electrodes at the interfaces of an oxide conductor and that of the ionic conductor with ohmic contacts but no current supply. Mathematics are presented that will allow one to calculate the electron and hole conductivities, as well as ionic transport numbers, directly from polarization measurements.

To easily demonstrate the electrochemical nature of the metal–oxide systems, let us recall from Chapter that the chemical reaction between a metal and oxygen is called oxidation, which takes place at low and high temperatures. This reaction is heterogeneous, as it involves different phases – solid (metal) and a gas – that take place depending on different factors at different interfaces. The result of heterogeneous reactions is usually the formation of a new phase as they occur between two immiscible phases.

If no natural oxide can be found on the surface, oxidation starts with the adsorption of oxygen (O₂) on the surface, followed by O₂ splitting into O atoms, and – as the reaction proceeds – oxygen is dissolved in the metal, and the oxide is formed.

The chemical reaction can be written as

6.41

Equation 6.41 can be determined by the second law of thermodynamics, which can be written in terms of Gibbs’ free enthalpy ΔG° (Eq. (3.2)) at high temperature, as the temperature and pressure are constant. The reaction is spontaneous if ΔG° < 0, or in equilibrium if ΔG° = 0, and thermodynamically impossible if ΔG° > 0.

It should also be noted that the oxide will also be thermodynamically formed only if the ambient oxygen pressure is greater than the dissociation pressure of the oxide in equilibrium with the metal. If we consider the formation of MO according to Eq. 6.41 for simplicity, it can be assumed that the oxide is simply divalent and ΔG° (MO) as the standard free energy of the reaction at a certain temperature, the oxidation of M is only possible if the following inequality holds:

6.42

The different dissociation pressures of the oxide and the corresponding standard free energies of formation of some oxides are summarized in the well‐known Ellingham diagrams, which do not take the kinetics into account. In Section 3.2.1, the Ellingham diagrams are deeply analyzed.

After the scale becomes continuous, the metal is separated from the gas, and the further reaction is carried on through diffusional transport of the reactants through the oxide scale. The difference between a thin (or anodically) and a thick grown oxide is in the driving force, which is known to be the electric field for the thin film and the chemical potential for the second category.

The resistance of materials at high temperature is defined by the formation of a continuous, adherent, slow growing, and thermodynamically stable oxide. The reaction takes place at both metal–oxide and oxide–oxygen interfaces as shown in Figure 6.4. There are several reactions occurring at the different interfaces:

• Metal–oxide interface:
  
  6.43
  
  

  which represents the two possible reactions at this interface – the oxidation of the metal and the reaction between the metal and the ionized oxygen, respectively.

  
  
• Oxide–gas interface:
  
  6.44
  Only gold members can continue reading. Log In or Register to continue

  Share this:

  • Click to share on Twitter (Opens in new window)
  • Click to share on Facebook (Opens in new window)

  Related

  Related posts:

  1. Nitridation
  2. Protective Coatings
  3. Diffusion in Solid‐State Systems
  4. Oxidation