Marcus P. Corrosion mechanisms in theory and practice

Подождите немного. Документ загружается.

228 Clayton and Olefjord

Figure 8 Measured intensity ratios, 100 × N/(O

2–

+ OH

–

), at the take-off angles 30°, 45° and 80° after

passivation of the Fe20Cr20Ni6Mo0.2N alloy at –75, 500 and 800 mV (SCE). The lower thick line and the

upper thin line represent the estimated intensity ratios from the N-distributions shown in the models to the

right of the figure. (From Ref. 15.)

segregation. The benefit of the technique was that, unlike the case of ion implantation

and plasma or thermal nitriding, a room temperature surface phase could be

formed without expected surface damage or sensitization. The advantage of studying

surface nitrides over the nitrogen-bearing alloys was reported to be that, whereas

anodic segregation is a continuously regenerating process, surface alloying provides

a finite quantity of nitrogen that can be monitored by surface analysis after anodic

dissolution under active and passive conditions.

As seen in Figure 10, the potentiodynamic behavior in deaerated 0.1 M HCl of

austenitic stainless steels is beneficially modified by the surface nitriding treatment, in

agreement with previous studies of the same steels alloyed with nitrogen [33]. Each of

the steels showed that the surface nitride stifles active dissolution and that this effect

Passivity of Austenitic Stainless Steels 229

Table 3 Chemical Composition of Main Elements of Steels (wt%)

Figure 9 Auger depth profile for alloy 30 after passivation for 24 h in 0.5 M HCl + 2 M

NaCl. Approximate etching rate was 0.01 nm/s. (From Ref. 30.)

is more pronounced as the addition of Mo is increased. In the lower alloyed steels,

304 and 317LX, the greatest influence of nitrogen was reported to be in raising the

pitting potential. As shown by Figure 10, the pitting resistance of surface-nitrided

317LX steel is markedly better than that of type 304 due to the higher Mo content. It

230 Clayton and Olefjord

Table 4 Chemical Composition of Steels (wt%)

Figure 10 Polarization curves for surface-nitrided and untreated austenitic stainless

steels in deaerated 0.1 M HCl where the specimens were permitted to float to the open-circuit

potential before polarization. Sweep rate: 1 mV/s. (From Ref. 33.)

is also apparent that the greatest anodic inhibition was achieved by the AL6X alloy,

which has the highest Mo content of the alloys studied and in particular is very

similar in composition to 904L except for an additional 1.8 wt % Mo. These

observations imply that N and Mo support the same mechanism of corrosion

inhibition.

The authors also noted that following potentiodynamic polarization from the

corrosion potential to 0 mV at a scan rate of 1 mV/s, XPS analysis was still able to

detect significant quantities of surface nitride. This is illustrated in Figure 11. The

most active of the alloys studied, type 304, was determined to have dissolved

approximately 20 monolayers. This suggested that the nitride may form a kinetic

barrier that is protected by the oxide passive film from rapid protonation to

ammonia and ammonium in the active range of potential. In the same study the

nitride phase formed on Ni had little effect on anodic behavior in 0.1 M HCl,

Passivity of Austenitic Stainless Steels 231

Figure 11 N Is spectra for austenitic stainless steels: (a) As nitrided; (b) follow polarization

from open circuit to 0 mV. Sweep rate: 1 mV/s. (From Ref. 33.)

whereas Fe became anodically activated by nitriding. Chromium exhibited complete

removal of the active nose following nitriding. This was attributed to suppression of

by direct chemical reaction of surface N and Cr forming a Cr

state as CrN, a

precursor of Cr

, which was identified by XPS. Molybdenum, however, showed

some ennoblement of the corrosion potential and transpassive potential. This effect

on Mo was later found to be even more pronounced following thermal nitriding [36].

It has been reported that nitrogen alloying and surface nitriding result in a

higher metal oxyanion content in the outer layers of the passive films of stainless

steels formed in acidic solutions. By inspection of the E-pH diagrams of Mo and W,

for instance, it is seen that MoO

2–

and WO

2–

are both more likely to form in the

middle to high pH range [37]. The presence of the metal oxyanions in the outer region

of passive films formed on stainless steels indicates that the products are developed

during the active stage of repassivation, precipitating as salt films. The well-known

pitting inhibition derived from metal oxyanions such as molybdate suggests that the

presence of such anions in the outer layers of the passive films would tend to reduce

the probability of pit initiation.

ALLOY SURFACE LAYERS

A stainless steel is normally used at potentials corresponding to the passive range.

In this condition only a fraction of the oxidized atoms remain in the passive film;

most are dissolved into the solution. The oxidation rate of the metal corresponds

to the passive current density. As established earlier, the alloying elements Cr and

Mo are enriched in the passive film. The other two main alloying elements, Fe and

Ni, could in principle be found in the solution due to selective dissolution of these

elements. XPS spectra [1,38–40] recorded from stainless steels after polarization

to potentials in the active and passive regions show that the measured Fe content

in the metal phase is lower than the composition of the bulk alloy. Fe must therefore

be selectively dissolved and can thereby be found enriched in the solution. Ni and

Mo, on the other hand, are enriched in the metal phase underneath the oxide. It has

been proposed [1,38–40] that during active dissolution and passivation of stainless

steel a thin layer of an intermetallic compound is formed in the outermost layers

of the metal phase. Inhibition of the active dissolution of stainless steel by elemental

Mo has been considered [41]. It was proposed that Mo serves to decrease active

dissolution by binding to active surface sites such as kinks and thus increasing the

coordination of more active species. More recently, it has been shown [42] by cyclic

polarization and pit propagation rate tests on a series of Fe-Ni-Cr-Mo alloys having

Mo contents of 3, 6, and 9 wt % for three Fe/ Ni ratios that active dissolution was

governed by an Ni-Mo surface complex. These observations led us to the

conclusion that in describing the corrosion properties of stainless steel it is not enough

to describe the reactions leading to formation of the barrier layer; it is also necessary

to describe the reactions taking place in the outermost layers of the metallic phase.

The influence of the alloy composition on the polarization diagrams of

austenitic stainless steels polarized in 0.1 M HCl + 0.4 M NaCl is shown in Figure 12.

The curves represent the following steels: curve 1, 20Cr-18Ni-6.1Mo-0.2N (wt %);

curve 2, AISI 316, 18Cr-13Ni-2.7Mo; curve 3, AISI 304, 18Cr-9Ni. The polarization

curves show that the high-alloyed steel is passive within a broad potential range. The

232 Clayton and Olefjord

other two alloys are sensitive to pitting corrosion in this solution. The critical current

density is about the same for the two Mo-containing alloys, and it is an order of

magnitude larger for the non-Mo-containing steel. The conclusion from the polar-

ization diagrams is that the composition of the metal influences markedly the passi-

vation and the pitting behavior of the steels. It will be shown in the following that syn-

ergistic effects exist between the alloying elements.

The spectra in Figure 13 were recorded from a high-alloyed austenitic

stainless steel (16.7Cr-15Ni-4.3Mo) after polarization in 0.1 M HCl + 0.4 M

NaCl at –320 mV (SCE). The potential represents the active dissolution potential

slightly above the corrosion potential. The result is from the same study as

Figure 2. It appears from the spectra of Figure 13 that the intensities of the

metallic states are much higher after polarization to the active potential

compared with the passive potentials of Figure 2. This is, of course, due to the

fact that the oxide film formed on the sample polarized at the active potential is

much thinner than on the passivated samples. One should in principle expect no

oxide or at most a very small amount of oxidized species on the alloy polarized

to the active region. However, during the rinsing and the transfer of the sample

from the electrochemical cell to the XPS analyzer, the surface is slightly oxidized.

The thickness of the oxide formed during handling of the sample polarized to

the active region is about 0.5 nm.

The composition of the metal phase can be estimated from the recorded

intensities. The dots in Figure 14a [1,40] show the apparent compositions of the metal

phases of the alloy after polarization to the active potential. The composition is

calculated from the measured peak intensities in Figure 13, taking into account the

Passivity of Austenitic Stainless Steels 233

Figure 12 Polarization diagrams obtained during exposure to 0.1 M HC1 + 0.4 M NaCl

of the following alloy steels: (1) Fe-200Cr-18Ni-6.1Mo-0.2N; (2) Fe-18Cr-13Ni-2.7Mo;

(3) Fe-18Cr-9Ni. (From Ref. 32.)

234 Clayton and Olefjord

Figure 13 XPS spectra of an Fe-20Cr-18Ni-6.1Mo-0.2N (wt %) alloy recorded after polarization to the active potential

–320 mV(SCR). (From Ref. 1.)

yields and the attenuation lengths of the photoelectrons. The horizontal lines in the up

and down triangles in Figure 14 show the composition of the alloy. It appears from

Figure 14a that Ni and Mo are enriched in the metal phase, while the elements Fe and

Cr are depleted. The depth to which the composition changes from the bulk

composition can be only a few atomic planes because of the very low diffusion rate

at room temperature. The depth of analysis, which is about three times the

attenuation length, is larger than the enriched zone. Hence, the bulk of the alloy

contributes to the recorded spectral intensity. Thus, the composition in Figure 14a is

an apparent composition. The actual surface composition can be found only if the

distribution of the elements in the surface region is known. Unfortunately, this is not

the case. It is therefore necessary to assume a realistic distribution function of the

elements in order to be able to calculate the surface composition of the metal phase.

The simplest distribution function describes enrichment of the elements Ni

and Mo in only one atomic plane. If this model is applied to the spectra shown in

Figure 13, it becomes apparent that it is not possible to obtain mass balance.

However, most of the oxide present on the surface during the analysis of the sample

polarized to the active potential is formed during transfer of the sample from the

cell. Therefore, to assess the composition of the surface during anodic dissolution,

the cations detected have to be converted to their metallic states and added to the

contribution from the metallic spectra. By assuming a model for the distribution of

the elements, the surface composition can be calculated. The details of the procedure

Passivity of Austenitic Stainless Steels 235

Figure 14 Composition of the metal phase during dissolution at an active potential.

(a) Apparent metal content; (b) estimated metal content in the outermost atomic layer [1,40].

have been published [1,40]. Figure 14b shows the quantitative analysis of the metal

composition during anodic dissolution. The distribution function used was an error

function where the composition varies over three atomic planes. The composition

of the second plane is 50% of the difference between the bulk composition and the

composition of the surface. The values shown with the coarse solid line represent

the composition of the outermost atomic layer during anodic dissolution. As

already pointed out, it was assumed that most of the oxide observed on the surface

was due to the transfer of the sample. It appears from Figure 14 that Ni and Mo

are markedly enriched on the surface of the metal during anodic dissolution. Even

the Cr concentration in the outermost layer is significantly higher than the bulk

concentration. Because of the enrichment of Cr, it is easy to understand that the

oxide formed during handling of the sample was mainly Cr oxide. Furthermore,

Figure 14b shows that the Fe content in the outer surface layer is markedly lower

than the alloy composition.

It was shown before that the Ni content of the passive film is low. One can

therefore expect that the Ni content at the metal-oxide interface is still high after

passivation. Figure 15 [1] illustrates the measured apparent metal content of the

metal phase after polarization to the active potential, –320 mV (SCE), and the

passive potentials, –100 mV and 500 mV (SCE). It appears from the figure that Ni

largely remains in the alloy phase after passivation.

It has been pointed out that the passive currents for all three alloys whose

polarization diagrams were shown in Figure 12 are about the same. Even the passive

current for pure Cr (Fig. 1) is about the same as for the alloys. The noticeable

difference in electrochemical behavior between the alloys is in the passivation current

and the pitting potential. As shown earlier, the alloying elements are enriched on the

236 Clayton and Olefjord

Figure 15 Apparent metal content of the alloy vs. the potentials (SCE). (From Ref. 1.)

surface in their metallic states during active dissolution. It has been suggested

[1,40] that an intermetallic surface phase is formed during dissolution.

The formation of an intermetallic surface phase can be understood from the

Engel-Brewer [43] valence bond theory of metallic bonding by considering the ground

state electronic configuration of the elements and the nature of the possible bonding

processes. The model predicts intermetallic bonding between “hyper” and “hypo”

d-electron transition metals, resulting in very strong bonding. Such bonding, in

principle, should result in low dissolution rates due to a higher activation energy for

anodic dissolution. Such systems are formed between transition metals from the left of

the periodic table having more vacant d electrons and those to the right having fewer

d-electron vacancies. The intermetallic bonding between elements suitablyseparated in

the transition row may result from penetration of an electron pair from the hyper

d electron into the d orbital of the hypo d-electron metal. XPS analysis [44] of thin

Ni-Mo intermetallic layers has confirmed the charge transfer. Compared with those of

pure metal lattices, the binding energies of Ni and Mo from the intermetallic layer are

shifted higher and lower, respectively. It has been pointed out that the strength of the

d-electron bonding increases from 3d to 4d, indicating that thestability of the inter-

metallic bond between Ni and Cr should be lower than that between Ni and Mo [33].

Therefore, it has been suggested that the overall latticeenergy increases during dissolu-

tion of Mo-alloyed austenitic stainless steel in the metal layers close to the passive film

due to formation of intermetallic bonds between Ni and Mo atoms [45]. Consequently,

the activation energy increases for anodicdissolution and the dissolution rate decreases.

According to this model, bonding between Mo and Fe is predicted to be weaker than

between Mo and Ni and Mo and Cr, and therefore Fe is selectively dissolved.

The Engel-Brewer model [43] of intermetallic bonding would explain how

Ni can lower the critical current density and elevate the pitting potential of

austenitic alloys, while playing no direct role in the construction of the passive

film. It is evident that Mo not only plays several direct roles in the formation and

stability of the passive film but also enhances Ni’s role by further enhancing the

anodic segregation of Ni. Therefore, it is proposed that through sluggish dissolution

kinetics alone, Ni, when bonded with Cr and more strongly with Mo, will lower

the rate of metal dissolution during the pitting process and thereby reduce the

maximum metal chloride concentration in the pit solution [45].

In the N-bearing stainless steels, N has a pronounced influence on the pitting

corrosion properties. It has already been pointed out [15,26,31] that the positive

effect of N is obtained for the Mo-alloyed stainless steels. The lower dissolution rate

of the Mo-containing alloys due to formation of intermetallic layer during active

dissolution provides a model for the synergistic effect between Mo and N. Figure 16

shows the proposed mechanism. Three alloys with and without Mo and N are

assumed to be exposed to an acid chloride–containing solution. Pits are initiated on

all three alloys. The pits formed on the Mo-containing alloys become smaller than

the pits on the Mo-free alloy because the Mo lowers the dissolution rate of the alloy

by formation of an intermetallic surface layer. If the acidity and/or the chloride

content of the solution is high enough, the pits formed on the Mo-free alloy will

become critical and grow. The dissolved ions are hydrated and the hydrolysis causes

increased acidity in the pits. The pH value in the small pits formed on the

Mo-containing alloys becomes low due to the outlet diffusion of H

ions. The

Passivity of Austenitic Stainless Steels 237