Marcus P. Corrosion mechanisms in theory and practice

Подождите немного. Документ загружается.

At the minimal c.d., Bp, typical diffusion control is observed consistent with

the local plateau of the current-voltage curve. A diffusion term is associated with a

transport limitation across the passive layer, whereas the inductive loop at the low-

frequency end visualizes the contribution of a decaying film protection at increasing

potentials. Application of a diffusion impedance equation to a finite-thickness layer

(

= 3 nm) given by

136 Keddam

Figure 18. Current-potential curves and schematic impedance diagrams of transpassive

Ni, 3 different single crystal orientations. 1 M H

. From ref [148].

led to a diffusivity D of the order of 10

–l6

–1

, This value is in agreement with

solid-state diffusion at room temperature, and a decrease of the film thickness by

a factor of 3 takes place between C

and E

An early paper [145] claimed on the basis of linear Tafel plots that the

dissolution in the transpassive range is controlled by a single charge transfer reaction.

It was then suggested that the electrochemical impedance behaves as a resonant

circuit, i.e., at least two relaxation times [146]. This prediction was directly verified

shortly after Ref. 147. Finally, a thorough investigation by impedance analysis at

various pH values and anion concentrations was at the origin of a more elaborate

model.

Figure 18 shows the current-voltage characteristics measured on three different

crystal plan orientations. As reported in Ref. 145, a Tafel law is observed over more

than two decades, but the impedance shapes display surprisingly intricate kinetics. A

very characteristic impedance feature is observed at point B (but not on the polar-

ization curves), where the impedance behaves like a perfect wave-trap circuit (the

impedance becomes infinite at a resonant frequency ≅ 0.4 Hz). Above this transition

point, e.g., C, the diagram is totally consistent with the electrochemical resonator

described in Ref. 146.

Figure 19(1) shows a detailed evolution of the impedance. The impedance

diagram turning counterclockwise when the frequency is decreased was related to the

reaction pattern shown below elaborated and compared with experiments by

numerical simulations [148].

Here Ni(II) stands for Ni cations pertaining to the lattice of the passive oxide

film and Ni

(II) is a cation in the film solubilized by chemical bonding with anions.

The likelihood of this species is also supported by the evidence of Ni=SO

band structures in the in situ Raman spectra of the Ni-H

interface in this

particular potential region. At higher potentials all the film cations are converted into

Ni(III) and Ni enters the secondary passivity region. Numerical simulations are

presented in Figure 19(2) and reproduce with a fair accuracy the main features of

experimental data, particularly the typical change of impedance trajectory between

C′ and D′.

Anodic Dissolution 137

Figure 19. Experimental and calculated impedance diagrams of transpassive nickel

(111) single crystal. 1M H

, 25°C. (1) Detailed evolution, (2) numerical simulations.

Measurement points are: A: 0.75 V/SSE; B: 0.8 V/SSE; C:0.8 V/SSE; D: 0.9 V/SSE; E: 0.95

V/SSE ; F: 1.1 V/SSE From ref [148].

Anodic dissolution of chromium in the transpassive range as well shows a

behavior generally attributed to the oxidation of the Cr(III) of the passive film to

soluble Cr(Vl) oxyanions [43,149,150].

Transpassive dissolution of chromium, molybdenum, and alloys containing

these elements has been extensively investigated by Bojinov et al. [151–153] using

impedance analysis and steady-state RRDE. Additional in situ information

concerning the conductance properties of the films is obtained by CER (contact

electric resistance). Many-parameter models were proposed combining reaction steps

138 Keddam

Figure 19 (Continued)

coupled by adsorbed species and bulk conduction in 3D layers controlled by the

mass balance of vacancies.

Anodic Dissolution Under Mass Transport Control

In corrosion practice, the rate of the heterogeneous anodic processes dealt with in

the foregoing sections is ideally low enough that mass transfer is never a limiting

factor. However, fast dissolution kinetics may occur in the following circumstances

of obvious concern to corrosion:

Active dissolution as a transient regime in the initial stages of metal passivity

Dissolution of locally depassivated metals following passivity breakdown (e.g.,

pits, crevices, grain boundaries)

Because of the irreversibility of the dissolution reactions at high overvoltages,

the coupling to mass transfer cannot be simply understood in the framework of

classical transfer-diffusion theory:

A high concentration of dissolved cations in the Nernst boundary layer is

incompatible with the supporting electrolyte approximation.

Limitation of the reaction rate by mass transfer must imply either back diffusion

of an acceptor of metal cations (complexing or solvating species) toward the

electrode surface or growth of a new phase on the surface with limitation of

the reaction rate by ohmic drop or space charge overvoltage.

Little is understood at a mechanistic level because of an almost total lack of

knowledge of the local chemistry within the boundary diffusion layer or solid films.

The problem is essentially investigated in the framework of pitting, crevices, and

stress corrosion cracking [154] by modeling and experiments on occluded cells

(artificial pit designing).

Mass Transport Control and Corrosion

It is generally accepted that transport effects (diffusion and electromigration) are at

the origin of a buildup of metal cation concentration concomitant with a depletion of

and a overconcentration of anions (for electroneutrality reasons). Homogeneous

chemical reaction can add their own contribution, e.g., hydrolysis of Cr(III) resulting

in a local acidification of neutral media [155].

The most relevant consequence of the change of interfacial chemistry for

corrosion is undoubtly the modification of the active-passive transition and the

possible generation of self-sustained large-amplitude oscillations between active

and passive states [156,157]. These phenomena were tightly associated with the

manifestations of passivity from the earliest research, particularly in the case of

iron in acid solutions. Theoretical investigations imply difficult nonlinear

mathematics. A significant renewal of the field was observed in relation to the

increasing interest in nonlinear phenomena and such concepts as bifurcation theory

and chaos in chemistry [158–160].

In spite of this new sophistication, the basic model is still that due to Franck and

FitzHugh [156] in which the following reactions are supposed to participate in a

cyclic sequence.

An oversimplified form of dissolution and passivation reaction is considered:

Anodic Dissolution 139

Reaction (46) at high dissolution rates tends to decrease the pH by H

electromigration and to shift the passivation equilibrium reaction (47) to the right.

At the same constant potential, the iron surface turns into the passive state, and

back diffusion of H

takes place and restores the active dissolution by displacing the

passivation equilibrium to the left. The same sequence restarts, generating a periodic

regime. It should be observed that this model ignores the likely participation of salt

layers (FeSO

,7H

O) and of ohmic drop and is unable to explain the coexistence

at a same electrode potential of side-by-side active and passive areas, i.e., of stable

localized corrosion coexisting with a passive sample. The latter self-stabilizing

process of dissolution is very generally ascribed to the depassivating role of Cl

–

enriched by electromigration toward the active areas. However, similar

behavior can be observed in chloride-free solutions [33,74].

Active Dissolution of Iron under Mass Tranport Control

The anodic dissolution of iron in sulfuric acid medium exhibits, with well-defined

convection on a rotating disk electrode, a current plateau as shown in Figure 20,

attributed to the limiting rate of transport between the dissolving surface and the

solution bulk [161]. The curves shown in Figure 20 suggest that some critical

phenomenon is associated with the common branching point from which all the

curves merge toward their plateau region, whereas they overlap perfectly below.

The nature of the process controlling the heterogeneous steps was investigated by

a detailed analysis of the EHD impedance.

140 Keddam

Figure 20. Current-potential curves for iron at different rotation speeds. Johnson-Matthey

iron. 1.8 M H

, pH = 0, 25°C. From ref [161].

EHD impedances are interpreted in terms of the existence of a viscosity

gradient in the boundary layer estimated to be four- to fivefold between solution

bulk and electrode surface. On this basis, the profiles of species participating in

the mass and charge transport have been computed and are plotted in Figure 21.

The interfacial concentration of Fe

is in good agreement with solubility

data. The increase of pH by H

migration is reflected in the profile of SO

2–

It is concluded that no barrier layer is involved in the mass transport control of

iron dissolution. The heterogeneous reaction rate is reported as not dependent

on the HSO

–

concentration; therefore, the only acceptor species likely to limit

the dissolution rate is water [74]. This is compatible with the modified form (29)

of the initial step of dissolution in which one water molecule is dissociated and

the depletion of free water at the electrode surface by Fe(II) hydration and

electromigration of hydrated Fe(II) away from the surface.

Anodic Dissolution Controlled by Transport in the Presence

of Solid Layers

In many practical cases of corrosion, the anodic dissolution takes place on a surface

partially or totally covered by thick (order of micrometers) solid layers grown by

Anodic Dissolution 141

Figure 21. Concentration field calculated under the conditions of Figure 20. Top :

viscosity and potential profiles. From ref [161].

several possible mechanisms (heterogeneous nucleation and growth, dissolution-

reprecipitation) and usually regarded as porous. Until the past decade the mechanism

by which this layer interferes, with the anodic dissolution, and particularly the

transport phenomena, remained poorly understood. The last developments deal

with iron and copper in HCl solutions and take advantage of steady-state RDE and

RRDE associated with AC impedance (EIS and EHD) techniques. Mass transport

and solution chemistry [162–163] let to a dissolution model [164] in which the

electrochemical monoelectronic surface step:

142 Keddam

is followed by a series of complexation reactions, the more likely being

More advanced RDE studies were interpreted by the formation of a porous 3D

layer of CuCl

ads

[156] in a step such as

Because of the reversibility of reactions (48) and (49), both Cl

–

and CuCl

–

diffusing

in opposite directions may control the dissolution rate. In the plateau region it was

shown [164] by RRDE that compact coverage is attained. Therefore in the model

elaborated in Ref. 165 the contribution of the layer to the rate of dissolution is

twofold and constitutes the central ingredient of the model: the CuCl film is

considered at the same time to dissolve at its outer boundary with the electrolyte:

CuCl

film

+ Cl

–

→ CuCl

–

and to limit the rate of Cl

–

diffusion to the Cu surface to form the film by reaction

(48). This form of the model is sketched in Figure 22.

Figure 22. Model of dissolution of copper through a solid layer growing by a Tafel

kinetics at the metal interface and dissolving by a chemical kinetics at the solution

interface. Adapted from ref [165].

It is noteworthy that this model is formally similar to the point defect model

(transport of a metal acceptor toward the metal surface, film growth at its inner

interface, dissolution at its outer one) shown in Figure 15. EIS and EHD responses

associated with this model have been derived [165] taking into account the diffusion

of both Cu

–

and CuCl

–

in the solution (diffusion layer δ) and of Cl

–

through the

film (diffusion layer λ) and allowing for the modulation of the layer thickness by

the AC potential, an effect discarded in Ref. 164. Comparison with the experiments

substantiated the model. The mass transport by diffusion and migration and the

potential drop at soluble anodes covered by salt films have been dealt with [166,167]

by assuming compact (saturated) or porous (supersaturated) coverages (wet) on iron.

Even more intricate situations on nickel have been tentatively approached [168].

ANODIC BEHAVIOR OF ALLOYS

A detailed understanding of the electrochemistry of alloys is still far from being

achieved on the basis of the rather crude theory of pure metals. The complexity is

expected to increase more rapidly than the number of components (cross-

interactions) and it is doubful whether a unique mechanism can be at work in view

of the extreme variety of compositions. As far as mechanistic information is

concerned, only homogeneous (single phase) and binary alloys are considered in

this section.

The Basic Concepts: A Survey

Thermodynamics and Rate Constant Approaches

It was thought very early that the change of rest potential and of dissolution rate with

alloy composition cannot be understood without assuming surface modifications

induced by the anodic processes themselves [169]. The nature of these modifications

has been the aim of many electrochemical and surface analysis studies but is still

not fully elucidated.

In an ideal case, in the absence of any free energy of mixing, the free energy

of a metal in an alloy with respect to that of the pure metal is given by

ΔG = RTln x

where x is the molar fraction of the metal in the alloy, and the corresponding shift of

electrode potential is

Anodic Dissolution 143

This equation is obeyed only by a few alloys and dilute amalgams [10]. The

individual dissolution rates of Cu and Ni in Ni-Cu were investigated as a function

of alloy composition [169,170]. In the framework of the activated state theory, a

standard rate constant of the solution of the ith component of an alloy, k°ia, is

calculated from the exchange current density at the standard potential:

where γ

= activity of the element in the alloy

= activity of the element cation in the solution

= transfer coefficient of the metal i–cation equilibrium

The k°ia is found independent of the alloy composition up to 60% Cu, above

which a new phase is invoked.

This purely formal kinetic treatment is of very little utility because it is quite

unable to dissociate the more elementary contributions to the alloy electrochemistry,

namely from metal solid-state physics to surface physical chemistry:

1. Change of elementary rates of charge transfer by electronic interaction in

alloy structure (alloying effect)

2. Change of the surface composition of alloy by selective dissolution of the less

noble element

3. Growth of 2D or 3D layers resulting from chemical or electrochemical

reactions of the alloy components with the solution

Following Rambert and Landolt [172], dissolution of single-phase alloys can be

classified in two categories:

Simultaneous dissolution, where at steady state the alloy elements go into the solution

at a rate proportional to their atomic concentration in the alloy. Well-known

examples pertain to the iron-base stainless steels and especially ferritic Fe-Cr,

on which this type of dissolution is repeatedly found [173,174].

Selective dissolution, where the less noble element dissolves selectively leaving

behind a “porous” metal phase enriched in the more noble constituent,

resulting in dealloying. The more common case is the selective dissolution

of Zn (dezincification) from α-brass (Cu-Zn). Selective dissolution is in

some instances associated with circumstances of stress corrosion cracking

(SCC) [175].

Simultaneous Dissolution and Associated Formalism

The basic formulation was elaborated sometime ago [175,176]. A linear relationship

of the form

144 Keddam

was suggested for representing the dissolution current of an alloy AB as a function

of the elementary current densities of A and B, γ

and (1 – γ

) being the atomic

proportions of A and B [commonly denoted A – (1 – γ

)B].

It was simultaneously assumed [178] that, as a result of the alloy dissolution,

its surface composition, (γ, 1 – γ), can deviate from the bulk one. Consequently,

the proportionality of the atomic fluxes of A and B going into the solution to their

alloy content is expressed by

where Z

and Z

are the dissolution electrovalences.

In principle, Eq. (53) yields a γ

value matching simultaneous dissolution at

any potential, I

and I

being known from the individual curves of A and B in the

same electrolyte. In general, there is no particular reason why this γ

would restitute

the correct value of I

. In the case of Fe-Cr alloys, the essential features of the

current-voltage characteristics in the transition range (7 to 12%) are not satisfactorily

reproduced [178] and nonlinear interactions must be introduced [105,106].

Application of EIS to this problem in the case of Fe-Cr alloys in the active and

passivation ranges is illustrated in the following. In the full passive domain,

compositional changes in the passive film, determined by XPS, play a role similar

to the adjustement of the surface composition of the alloy phase to achieve the

simultaneous dissolution [179]. According to Ref. 180, percolation phenomena,

apparently relevant to selective dissolution, could also explain some features of the

dissolution of Fe-Cr in the incompletely passivated state.

Selective Dissolution

Different mechanisms have been proposed to explain the selective dissolution of

alloys and the formation of a porous dealloyed layer. A dissolution-redeposition

mechanism [180] was proposed for α-brass dealloying. A larger group of models

requires the description of atomistic processes of restructuring of the more noble

atoms A in order to allow the dissolution of the more soluble element B to proceed

across an A-enriched porous layer. A critical review of the mechanisms likely to

participate in these surface phenomena can be found in Ref. 32. It was suggested

[182] that the rate-determining step is a solid-state diffusion of the less noble

atoms via divacancies. Surface diffusion can also be taken into account. Roughening

by a mechanism of “negative” aggregation known to generate fractal interfaces

[161] constitutes a fruitful approach. In the 1980s, a series of papers [175,180,184]

underlined the existence of critical compositions and of a threshold concentration

of the less noble constituent below which no dealloying appears. This sharp

dealloying threshold is not consistent with any of the diffusion-based models of

selective dissolution, which are supposed to produce essentially continuous

behaviors. A model based on percolation phenomena was proposed and extensively

worked out. Its main background and developments will be exposed.

Simultaneous Dissolution: Fe-Cr Alloys

In contrast to selective dissolution, evenly dissolving alloys can be dealt with, up

to a sophisticated level including non-steady-state responses, by macroscopic, i.e.,

kinetic descriptions. As shown in Figure 23, Olivier [1851] pointed out that the

steady polarization curves of Fe-Cr alloys in a 0.5 M sulfuric solution display the

decay of active and passive currents with increasing Cr content and the emergence

of the transpassive dissolution of Cr to the hexavalent state.

The studies have focused on two aspects of the behavior of Fe-Cr: the

modifications brought about by the addition of chromium to iron in the active

dissolution and prepassive ranges on the one hand and in the passive state on the other

hand. Electronic interaction by filling up of the d level of Fe [186] was put forward.

Most of the subsequent contributions concluded in some progressive change from

an iron-to a chromium-like behavior without a further detailed mechanism in the

absence of a satisfactory model for pure iron on its own. A particular ability of Cr to

enhance the passive state of iron, even with low surface coverage, was repeatedly

reported [187,188]. Regarding the passive state, advances of in situ surface analysis

[Auger electron spectroscopy (AES) and XPS] associated with electrochemical

Anodic Dissolution 145