Marcus P. Corrosion mechanisms in theory and practice

Подождите немного. Документ загружается.

of this dissolution process will finally result in breakdown of the passive layer and

exposure of bare metal surface to the electrolyte. For F

–

this occurs after intermediate

breakdown and repair events at the total surface. For the other halides, local

effects and the formation of pits are found corresponding to their less pronounced

complexing tendency. They will form complexes and cause their effect only at

special active sites of the oxide surface. It should also be mentioned that fluoride

causes a reduced local attack in more alkaline solutions such as phthalate buffer

pH 5 [42], which finally leads to the formation of corrosion pits and not a general

attack as observed in strongly acidic electrolytes, The thinning of the oxide layer

is also much less pronounced or hardly detected for pH>5. For a higher pH the

passive layer gets to or close to its dissolution equilibrium in the sense of the

thermodynamically deduced potential-pH diagrams. In this situation, the tendency

for the formation and transfer of soluble fluoride complexes will be reduced so

that again only local effects are observed.

In this regard, the special situation of Cr should be discussed. This metal does

not show pitting and the passive layer is not attacked in the presence of chloride. The

stability constants of CrX

complexes are smaller than 1. Besides that, it is well

known that the exchange of Cl

–

ligands and H

O between the inner and the outer

sphere is an extremely slow process with a half-time of several hours [49]. This is a

consequence of the large ligand field stabilization of the Cr(III) complexes with an

octahedral coordination shell with three electrons in the lower t

and none in the e

level. This situation causes “insolubility of CrCl

and Cr

in cold water” as

reported in literature [50]—however, it is rather the slow dissolution rate of the Cr

salt or oxide than their low solubility. The exchange of a ligand of the inner with one

of the outer coordination shell requires a large activation energy, which makes the

Mechanisms of Pitting Corrosion 259

Figure 8 Decrease of oxide thickness with time of halide exposure at 0.5 V after 60 min

prepassivation at 1.20 V for different halides, deduced from XPS measurements. (From

Ref. 48.)

dissolution process even via complex formation extremely slow. Therefore one

has to conclude that the presence of a CrCl

complex at the surface will not

increase the dissolution rate because it will form and dissolve very slowly by

itself. In contrast to this situation, the exchange is rapid for Fe

complexes. Thus,

a chemical change of Cr

ions from a part of the oxide matrix to a CrCl

complex

will not increase the dissolution rate. Besides these circumstances, the smaller stability

constants of the Cr

complexes are also in favor of the stability of the passive

layer. In consequence, the tendency for Cr

-halide complexes to form is negligibly

small, and once they have formed their dissolution rate is not increased relative to

within an oxide matrix. Therefore the halides will not attack the passive layer

of chromium and pitting cannot occur, in agreement with the experimental findings.

For similar reasons, the dissolution rate of Cr(III) oxide is extremely slow in the

passive state. Additions of Cr therefore stabilize the passive behavior of Fe-Cr alloys

and stainless steel. Fe-Cr alloys are more resistant to pitting in chloride-containing

electrolytes with more positive pitting potentials compared with pure Fe. The Cr

concentration is increased within the passive layer relative to the composition of

the bulk metal. Thus Fe-Cr alloys are more protected against the attack of aggressive

anions and pitting by the beneficial effect of Cr.

Comparison of the Different Nucleation Mechanisms

The discussion of the different nucleation mechanisms on the basis of experimental

results for iron and nickel leads to the conclusion that the film breaking and

adsorption mechanisms are very effective. As usual in kinetics, the fastest reaction

path is dominating. This, however, depends on the experimental or environmental

conditions. For a stationary state of the passive layer the adsorption mechanism

seems to be most effective, as demonstrated for Fe in weakly acidic electrolytes. If a

nonstationary state is attained by a fast change of the potential, film breaking is most

probable. Of course, other nucleation mechanisms may contribute as well, such as

mechanical damage of the surface, dissolution of inclusions, and last but not least the

penetration mechanism. Penetration is believed to be the leading mechanism for

pitting of Ni in Cl

–

-containing electrolytes. A conclusive critical experiment to

determine whether penetration of Cl

–

is an initial step for breakdown of Ni passivity

is still missing. The role of inclusions is the subject of another chapter in this book

and will not be discussed in detail here. It was the aim of this chapter to discuss effects

related to pure or at least single-phase metals. In the technical world, however, these

other effects are very important. As chemistry plays a decisive role in the pitting

process, one should discuss the tendency of the different cations to form complexes

with halides. Thus, one has to include the properties of the aggressive anions and of

the specific metals under study as well. This idea is often neglected but seems to be

a key question for breakdown of passivity and the stable growth of corrosion pits, and

it will be discussed again in the next section.

TRANSITION FROM PIT NUCLEATION TO PIT GROWTH,

MICROSCOPIC OBSERVATIONS

In recent literature the transition of nucleation to a stable growth of corrosion pits

is examined with the scanning tunneling microscope (STM) [51,52]. These studies

260 Strehblow

try to follow the development of corrosion pits from a diameter of some few

nanometers to several micrometers using besides the STM the scanning electron

microscope (SEM). The results thus give an extension from previous investigations

with the SEM and the light microscope [6] from the μm to the nm range. STM

studies may trace corrosion pits down to a size of some few nm. In principle, these

effects have been followed in two different ways. Maurice et al. [51] passivated

Ni(111) single crystals in ~ 0.1 M sulfate solution of pH 3 for 30 min at E = 0.85

and exposed them finally to this solution after the addition of 0.05 V NaCl at the

same potential (pitting potential 0.75 V). Numerous small pits (5–9 × 10

–2

)

with a preferential round or some with a triangular shape aligned at steps of

terraces of the (111) plane could be found (Fig. 9). They have an average lateral

size of 20 nm and a depth of 2 nm, which is in the thickness range of the passive

layer of ~ 1 nm. A comparison of the total area of these pits (number and size)

leads to a charge of metal dissolution of ~ 0.7 mC cm

–2

, which is about one order

of magnitude less than the directly measured charge of 1.6 to 8 mC cm

–2

. This

result leads to the conclusion that increased metal dissolution in the passive state

occurs, which is expected according to the adsorption mechanism for pit nucleation.

These small pits are interepreted as pit nuclei that apparently did not grow further.

About 1% of these pit nuclei grow larger (70 nm) but with the same depth and a

triangular shape oriented still parallel to the step edges at the Ni surface.

Mechanisms of Pitting Corrosion 261

Figure 9 STM topographic image of a pit formed on Ni(111) passivated for 1800 s at

0.85 V (SHE) in 0.05 M Na

, pH 3 and subjected to pitting in the same solution with

additions of 0.05 M NaCl for 4500 s at the same potential. (From Ref. 51.)

They presumably are repassivated pits. About 0.1% of the pits grow to a size of

some few 100 nm with an elongated irregular shape parallel to the steps. These

pits are seen as repassivated, i.e., metastable pits and grow when current transients

are observed in the potentiostatic current-time curves.

Another approach is the growth of corrosion pits at potenials several hundred

mV more positive than the pitting potential for short times in the ms range. Kunze

and Strehblow prepassivated Ni(111) surfaces by changing the potential from –0.1 V

to values below the pitting potential (E = 0.39 V) for a few seconds in a 0.2 M

NaCl solution of pH 5.6 to prepassivate the surface [52]. Then the electrode was

pulsed to potentials in the range 1.0 to 1.5 V for some few ms to grow pits and

finally stepped back to –0.1 V to stop pit growth. These pulse measurements

facilitated pit nucleation and thus caused a large density of very small corrosion

pits [30,53] that could be studied in situ. The sulfate-free NaCl solution enhanced

pit nucleation and hindered the early change of their shape from a polygon to a

hemisphere [6]. Figure 10 presents a sequence of typical corrosion pits in different

stages of development [52]. Small pits of less than 10 nm lateral size and 1.5 nm

depth have an irregular slightly elongated shape. They change to triangular pits

during further growth. This shape is easily explained by (111) crystal planes

forming the pit surface and their intersection with the (111) Ni surface. They are

the most densely packed crystal planes of face-centered cubic (fcc) Ni and thus

have the largest resistance to metal dissolution. As a consequence, these planes

remain during the dissolution of an actively corroding surface site and thus form

the inner pit surface. They still grow further as shown on the SEM micrograph of

262 Strehblow

Figure 10 Sequence of in situ STM topographic images of growing pits formed on

Ni(111) by potentiostatic pulses in 0.2 M NaCl. pH 5.6 during 50 ms at 1.1 V (SHE) after

3 s prepassivation at 0.39 V. (From Ref. 52.)

Mechanisms of Pitting Corrosion 263

264 Strehblow

Figure 11 SEM images of pits formed on polycrystallinc Ni electrodes: (a) formed in

0.2 M NaCl, pH 5.6 for 5 ms at 1.5 V (SHE) after 10 s prepassivation at 0.5 V (from Ref.

52); (b) formed in 0.05 M phthalate buffer, pH 5.0 with addition of 0.1 M KCl for 200 ms

at 1.00 V (SHE) after preactivation at –0.5 V and 1 s prepassivation at 0.50 V (from

Ref. 58).

Figure 11a for a similar experiment on polycrytalline Ni [52]. Depending on the

orientation of the crystallite surfaces, other shapes of the pit orifice also appear

such as squares (Fig. 11b) [58] or triangles with cut-off edges approaching a hexagon.

Figure 12 presents examples of the combination of (111) and (100) planes for

pits on crystallites of body-centered cubic (bcc) Fe grown with the potentiostatic

pulse technique described earlier [6]. These polygonal pits change their shape to

hemispheres when the increasing concentration of corrosion products leads to the

preciptation of a salt film. This situation causes electropolishing of the pit surface by

diffusion-limited metal dissolution within the pit. It starts at the pit bottom, where the

electrolyte is concentrated first, and finally includes the whole pit surface, changing the

polygonal form to a hemisphere [55]. These electropolished hemispheres keep their

form when crossing a grain boundary, whereas the polygons may change their shape

because of the different orientation of the adjacent crytallite (Fig. 13a and b) [6]. The

change of the pit surface from a polygon to a hemisphere will be discussed later in

detail together with the calculation of the electrolyte composition for later stages of

pit growth.

The change of the pit shape with its size and thus its age gives support to the

film breaking mechanism. A rapid change of the electrode potential causes small

cracks within the passive layer with a size similar to the thickness of the passive

layer of 1 to 2 nm. This leads to small elongated pits of some few nm, which then

change to polygonal forms during their growth to some 10 nm (Fig. 10a–c). If the

potential is high and thus the local current density at the pit surface is large

enough, the increasing concentration of corrosion products finally leads to a round

and electropolished hemisphere. Low potentials with smaller pit current densities

do not reach this stage at all or after a longer time so that large polygonal pits may

be observed for these conditions. In the vicinity of the pitting potential, pit growth

becomes instable. As a consequence, a metastable pit growth with a stop and go

sequence will lead to the observed irregular shape.

The presence of large amounts of nonaggressive anions in the bulk electrolyte

also causes their accumulation within the pits. This competition leads to smaller

halide concentrations with a less stable condition for pit growth. The smaller

dissolution current in the presence of these nonaggressive anions (see Fig. 15) further

supports unstable growth and possible repassivation. As a consequence,

discontinuous growth of pits with irregular shapes is observed. Figure 14 gives an

example of a pit grown for 10 s on Ni in 0.05 M phthalate buffer, pH 5 with

0.1 M KCl and 0.1 M K

. For these conditions pits grow into the metal leaving

a metal cover that is perforated from their inner surface. Apparently, this irregular

shape compensates for the lower current density. The covered pit keeps the

accumulated halide inside and also causes a large ohmic drop that stabilizes its

growth under less favorable conditions. At the inner surface of this larger pit a

small triangular pit is observed, which supports the explanation of continuous

passivation and nucleation for irregular growth. The formation of small corrosion

tunnels with (100) surfaces during pitting of Al in the vicinity of its critical pitting

potential further supports the dicontinuous irregular growth at microscopic and

submicroscopic sites [54]. At more positive potentials, repassivation no longer

occurs and pits with a hemispherical and electropolished surface are obtained.

Mechanisms of Pitting Corrosion 265

266 Strehblow

Figure 12 SEM images of pits on polycrystalline Fe formed in phthalate buffer, pH 5.0

with addition of 0.01 M KCl for 3 s at 1.10 V on crystallites with (111) orientation. (a)

(110) planes at pit surface; (b) combination of (110) and (100) planes at pit surface. (From

Ref. 6.)

PIT GROWTH

As visualized in the previous section, different stages of pit growth have to be

distinguished, which are closely related to the question of the stability of localized

corrosion. A pit will run through these different stages, each of which has its own

Mechanisms of Pitting Corrosion 267

Figure 13 Microscopic images of pits on polycrystalline Fe formed in phthalate buffer

pH 5.0. (a) With 0.01 M KCl and 0.1 M K

for 30s at 1.10 V (SHE). (b) with 0.01 M

KCl only, for 3 s at 1.10 V. (From Ref. 6.)

characteristic conditions of stability. Once a pit has nucleated, its local current

density refers to the applied electrode potential. For very positive potentials well

within the passive range, extremely high dissolution current densities of several

tens of A/cm

up to more than 100 A/cm

have been deduced from the growth of

the corrosion pits using Faraday’s law and taking into account their size and form

(Fig. 15) [53,55]. Equivalent current densities are measured directly on small wire

electrodes of <1 mm diameter embedded in resin when the passivation of the

whole electrode surface is prevented in solutions of high chloride content (>1 M)

starting with a preactivated oxide-free surface [53,56]. Extrapolating the dissolution

kinetics of free metal corrosion to potentials in the passive range of the polarization

curve leads to extremely high local current densities of 10

to 10

A/cm

[57].

These large current densities would cause precipitation of a salt film within 10

–4

to 10

–8

s. Current densities of more than 10

A/cm

are possible in principle but

only up to 120 A/cm

has been observed for a free corroding metal surface [53,55].

The local current density increases with the chloride concentration, whereas

the critical pitting potential decreases with increasing chloride concentration. If,

however, i

c,p

is presented relative to E

, the data fit a single curve. Apparently an

increasing chloride concentration shifts the whole i

c,p

/(E – E

) curve in the negative

direction [58, p. 55]. The addition of nonaggressive anions such as sulfate or nitrate

shifts the exponential current increase to positive potentials with a large potential-

independent plateau at low potentials [53,58, p. 57]. These plateau values are in the

range of a few A/cm

, i.e., 5 A/cm

for Ni in 0.1 M KCl + 0.1 M K

and 2 A/cm

268 Strehblow

Figure 14 Microscopic image of a metal-covered pit with irregular shape on

polycrystalline Ni formed in 0.05 M phthalate buffer, pH 5.0 with 0.1 M K

and 0.1 M

KCl for 10 s at 0.90 V (SHE), Small triangular pit at inner pit surface. (From Ref. 58.)