On the mechanism of crevice corrosion of stainless Cr steels

On the mechanism of crevice corrosion of stainless Cr steels

Corrosion Science, 1971. Vol. 1I, pp. 499 tO 510. Pergamon Press. Printed in Great Britain ON THE MECHANISM OF CREVICE CORROSION OF STAINLESS Cr STEE...

605KB Sizes 0 Downloads 59 Views

Corrosion Science, 1971. Vol. 1I, pp. 499 tO 510. Pergamon Press. Printed in Great Britain

ON THE MECHANISM OF CREVICE CORROSION OF STAINLESS Cr STEELS* G/:iSTA KARLBERG a n d G6STA WRANGLEN Department of Applied Electrochemistry and Corrosion Science, Royal Institute of Technology, Stockholm 70, Sweden

Abstract--The mechanism of crevice corrosion of stainless Cr steels in neutral 3%NaC1 solution has been studied in a cell constructed for this purpose. Immediately after the start of an experiment, both pH and potential of the steel surface increase as a result of alkali formation (02 reduction) during corrosion of the steel in the passive state and thickening of the air-formed oxide film. After the oxygen available within the crevice has been consumed, the cathode process is concentrated to the outer surface whereas on the crevice surface only metal dissolution takes place. By hydrolysis of metal ions (Fe I+ and Cr s+) the crevice solution is now acidified. After having passed a maximum value of about 10, the oH in the crevice is lowered to a steady state value of about 4. The crevice surface is first activated locally (pitting) and finally in its entire extent and e~v. and e,~t. are both lowered. At the anode (crevice) surface, the Flade potential is raised whereas it is lowered on the outer surface where the pH is raised due to the cathode reaction. The crevice must therefore be active and the outer surface passive even if their potential difference is quite small (150 mV for 13%Cr, 50 mV for 17%Cr). R6sum6---Le m6canisme de la corrosion caverneuse d'acier inox au Cr dans une solution neutre de NaCI b. 3% a 6t6 6tudi6/t l'aide d'une cellule ad hoc. D6s le d6marrage d'un essai le pH et le potentiel de l'acier augmentent du fait de la r6duction de Os, lors de la corrosion de l'acier gt l'6tat passif et de l'6paississement de l'oxyde ~ l'air. Apr6s consommation de l'oxyg6ne dans la caverne, la r6action cathodique se concentre ~ la surface ext6rieure, tandis que dans la caverne se produit seule la dissolution du m6tal. Par hydrolyse des ions Fe 2+ et Cr a+, la solution de la caverne s'acidifie. Apr6s avoir atteint un maximum de 10, le pH s'y abaisse ~ une valeur stable voisine de 4. Dans la caverne le m6tal est d'abord activ6 localement (piqOres) et ensuite enti6rement; E~v et Ee~t sont tous deux abaiss6s. A l'anode (crevasse) le potentiel de Flade s'616ve alors qu'il s'abaisse ~ la surface ext6rieure, oO le pH monte du fait de la r6action cathodique. La crevasse doit de ce fait 6tre active et la surface externe passive, m~me si les potentiels y diff6rent peu (150 mV pour 14%Cr; 50 mV pour 17°oCr).

Zmammenfassung--Der Mechanismus der Spaltkorrosion an rostfreien Chromst/ihlen in 3%-iger NaCI-L6sung wurde in einer Versuchszelle studiert, die fiir die.sen Zweck speziell konstruiert wurde (Fig. 1). Unmittelbar nach Anfang eines Versuches steigt das pH sowie das Potential der Stahloberfl/iche, offenbar als Ergebnisse der Alkalibildung (Sauerstoffreduktion) w/ihrend der Korrosion des Stahles im passiven Zustand und der darans folgenden Verst/irkung des an der Luft gebildeten Oxydfilms. Nachdem der binnen dem Spalt zug'angliche Sauerstoff verbraucht worden ist, wird der Kathodenprozess zur /iusseren Stahloberfl/iche konzentriert w/ihrend an der Spaltoberfl/iche nur Metallaufl6sung stattfindet. Wegen Hydrolyse yon Metallionen (Fe ~+ und Cr 3+) wird die Spaltl6sung sauer. Nach dem Erreichen eines pH-Maximums yon etwa 10 stellt sich im Dauerzustand ein pH yon etwa 4 ein. Die Spaltoberfl/iche wird anf/inglich lokal (Lochfrass) aber schliesslich in ihrem ganzen Urnfang aktiviert und die Elektrodenpotentiale der Spaltoberfl/iche und der/iusseren Stahloberfl~.che werden beide erniedrigt. An der sauren Anoden-(d.h. Spalt-)-Oberfl~che wird das Fladepotential erhSht, w/i.hrend an der/i.usseren Oberfl/iche, wo Alkali durch die Kathodenreaktion entsteht, das Fladepotential ernledrigt wird. Die Spaltoberfl/iche kann deshalb aktiv und die ~ussere Oberfl/iche passiv sein auch wenn der Potentialunterschied nur klein ist (150 mV fiJr 13%Cr, 50 mV fiir 17%Cr). *Manuscript received 28 May 1970. 499

500

G~TA KARLBERGand GOSTA WRANGLI~N

INTRODUCTION SENSITIVITYtowards crevice corrosion is perhaps the greatest weakness of the stainless steels, and pitting of stainless steel is quite often connected with crevices or deposits. A survey of the literature x-x8 on crevice corrosion of stainless steels shows that relatively few experimental investigations on actual working crevices have been carried out. Conditions in really efficient, i.e. quite narrow crevices, such as threads, are not easy to specify or measure. On the other hand, wider crevices, in which measurements may be carried out, are not very effective and crevice corrosion may not occur. The mechanism of crevice corrosion on passivated metals is usually described as follows: On a free surface the diffusion-limited c.d. of Oo reduction will exceed the critical c.d. for passivation, and the metal will remain in the passive state with a corrosion potential above the Flade (or activation) potential. In a crevice, on the other hand, the diffusion-limited c.d. of the cathode reaction will soon be smaller than the critical c.d. for passivation and the corrosion potential, i.e. the point of intersection between the cathodic and the anodic polarization curve, will fall below the Flade potential. This means that the outer surface will be maintained in the passive state, whereas the crevice will be active. The p.d. between outer surface and crevice is assumed to be great and the Flade potential is tacitly assumed to be the same for outer surface and crevice. Studies have shown that the Flade potential is not changed with the 02 content of the solution. 2,8,4 On the other hand, the critical c.d. for passivation is increased in the presence of CI-. s, 1,t.15 The concept of quite a large p.d. between crevice and outer surface seems to be derived from potential measurements on steel specimens immersed in solutions of differing oxygen content. The potential of a stainless steel specimen, immersed in an aerated or a deaerated solution of the same p H may differ about 600 mV. In a study of crevice corrosion, Rosenfel'd and Marshakov 5 measured a potential difference of several hundred mV between a crevice and a free surface of stainless steel, which, however, were not in metallic contact with one another, whereas in an actual, working crevice, the latter is in direct galvanic contact with an outer surface. Since the two electrodes are mutually polarized, the potential difference in a working crevice corrosion cell would be expected to be much smaller. France and Greene x6 describe anodic protection of crevices in H2SO4 and measure potential differences of about 1000 mV between crevice and outer surface, which were both anodes with respect to a counter electrode as cathode. From the electroplating field it is well known how difficult it is to make the current penetrate into crevices and recesses. The great potential difference between crevice and outer surface in this case has no correspondence, therefore, in an actual crevice corrosion cell. More than 20 y ago, Uhlig ~ realized that crevice corrosion and pitting are, in principle, similar phenomena, and it has been stated 12 that pitting does not require a crevice--it creates its own ; it is also true that a crevice attack on stainless steels starts as small pits. The presence of 02 or some other oxidation agent and C1- seem to be necessary prerequisites for both pitting and crevice corrosion of stainless steels. The operation of an oxidant concentration cell between a small, active anode surface in contact with an acidified anolyte and a large, alkalized and passive cathode surface also seem to be common features. Rosenfel'd and Danilov 17 studied the potential

On the mechanism of crevice corrosion of stainless Cr steels

501

distribution on a surface of stainless steel subjected to pitting and found a p.d. of about 100 mV between the interior of a pit and the surrounding, unattacked surface. Studies of crevice corrosion of Ti have also led to suggestions of a mechanism similar to that of pitting, ls,19 Fokin and co-workers 11 consider, however, that in the development of crevice corrosion, pH is of greater importance than CI-, which are more significant in the pitting of stainless steels. As a measure of the sensitivity of a stainless steel towards crevice attacks, they take the pH at which a free, passive electrode is activated. The mechanism of crevice corrosion of C-steels, which occurs primarily in inhibited solutions, seems to be substantially the same as that of stainless steels. Whereas the free surface is passivated by the inhibitor, a crevice may become active due to insufficient supply of the inhibitor. With reference to crevice corrosion of carbon steels, Skorcheletti and Golubeva "-° explained by means of polarization diagrams why the aerated electrode in a differential aeration cell corrodes less than the non-aerated electrode. Since the cathode reaction occurs more readily on the aerated electrode, the pH on its surface is raised and its Flade potential is lowered. For passive electrodes of Fe, Cr and Ni the Flade potential changes with pH according to the relation 2~ eFt = eFI. O -- 0"058 pH.

(1)

Even if the potential of aerated and non-aerated electrode is assumed to be the same, it is possible that this common corrosion potential lies within the passive range of the aerated electrode, but within the active range of the non-aerated electrode. Accordingly, these authors showed that buffering of the test solution prevented passivation of the aerated electrode, which then corroded more than the non-aerated electrode. It should be noted, on the other hand, that the mechanism of crevice corrosion seems to be substantially different in strongly acid solutions and when metal ion concentration cells (rather than differential aeration cells) are operative. In metal ion concentration cells, as with Cu, the attack may take place on the free surface close to the crevice and not within the crevice itself. 2z Finally, most cases of crevice corrosion and deposit attack in the atmosphere are apparently due to the simple fact that moisture and salts are more easily retained in crevices and under deposits than on free and clean surfaces. 23 EXPERIMENTAL The cell, which was built from polymethyl methacrylate in order to make direct observations of the metal surface possible (Fig. I), consisted of two separate vessels that could be screwed together to form a crevice. This construction enables perfect cleaning and degreasing of the walls of the crevice before each experiment, which is necessary in order to make the thin electrolyte film penetrate into the crevice completely and wet the surface of the thin steel sheet specimen. The width of the crevice was varied by placing a rubber gasket of desired thickness between the walls. A calomel electrode ending in the larger vessel, which is in contact with the atmosphere, was used to measure the potential of the outer surface; another calomel electrode whose capillary ended in the smaller vessel was a reference for pH and potential measurements in the crevice. For measuring the change of pH within the crevice

502

G(SSTA

KARLBE, RGand G(SSTAWRANGLI~N

l ~

Zno%° : g~

~

Perspexd/ ~ vessels~ 1 ~

g

a

contoct

s

k

Rubber t

e

Gl ass electrode FXG. 1. Top view of crevice corrosion cell 1 : 2.5.

solution, a glass electrode with a plane glass membrane penetrates through the smaller vessel and ends in the wall of the crevice, as shown by Fig. 1. The cell provided uniform aeration of the outer surface so that no differential aeration cells were created along the latter; the specimen was well below the solution surface along the whole of its length. N o r will corrosion products accumulate along the crevice by sinking towards the bottom as is the case with a vertical crevice. Further experimental details were as follows: Specimen: 0.2 m m steel strip, dimensions 30 x 1.8 cm, ground with paper No. 400, degreased with acetone and kept in a desiccator for 2 h. Crevice (between metal surface and cell wall): Dimensions 15 x 2 cm. Average thickness (a - - 0-2)/2 mm, where a = thickness of rubber gasket (0.5-2 mm). Electrolyte: Aerated, neutral 3%NaCI. Glass electrode: Radiometer, Type G242C. Moulding resin: Leukotherm E571, Hardener T3 (Bayer). Insulating lacquer: Struer edge protecting lacquer.

Specimen analyses Notation

C

Si

Mn

A (13% Cr) B (17%Cr)

0.10 0'07

0.30 0"22

0.29 0.29

P

0.016 0"016

S

Cr

Ni

Mo

0.004 0.010

13.6 17"3

0.22 0"14

-0.01

Attack first developed where the crevice was thinnest. In addition, attack almost always started in the slot forming the mouth of the crevice. By slightly bending the specimen attack could be provoked just opposite the capillary hole. In measuring the crevice potential accurately, a cell without glass electrode was used. In measuring the p H change of the crevice electrolyte it was more important for the attack to occur opposite the glass electrode, and in this case the potential measurement was only qualitative. The glass electrode was therefore mounted with its membrane 0.1 m m inside the wall of the crevice. Perfect sealing was required and the glass electrode was

On the mechanism of crevice corrosion of stainless Cr steels

503

therefore surrounded by two concentric Perspex tubes, the outer of which was glued to the cell. The two cylindric interspaces created in this way were filled" with moulding resin. In order to compare pH values measured in the crevice with critical pH values for general activation of the same stainless steel, time-potential curves were recorded for steel specimens immersed in CI- solutions with varying pH and 02 content. The cell (Fig. 2) had a tightly fitting cover, and gas mixtures with varying oxygen content could be introduced from the bottom. The specimens were inserted in the walls of the cell. The surface of the specimen in contact with the cell was coated with lacquer in order to avoid crevice attacks. The specimens were ground with emery paper No. 400, degreased with acetone and then stored in a desiccator 24 h before mounting. Before each series of measurements, the actual gas mixture was passed through the electrolyte for a couple of days in order to reach a stationary state. The pH of the test electrolyte (3 %NaCI) was adjusted by means of a 0. I M acetate buffer and HCI. s

//~/////

SCE ELECTRODE DIPPINGINTOELECTROLYTE

~

/,//

i

r/ /

j,

~

GASOUT

'~ I/

{~---- HOLEFOR

t

EIG. 2.

GASINPROM ROTAMETER Cell for time-potential studies. Eight symmetrically placed steel specimens.

RESULTS 1. Crevice corrosion

Initially, the attack was very small. Quantitative values of the crevice potential (ecrev.) refer to experiments in which an attack was obtained just opposite the capillary hole. The glass electrode, on the other hand, had such a large area that it covered both the attacked and unattacked surfaces. Thus the pH within the attacked surface could not be evaluated with accuracy in short-term experiments. The following factors were found to be significant in crevice corrosion: a. Potentials o f outer and crevice surface and the p H in the crevice as a function of time

Figure 3 shows typical changes of potentials for the two steels. In the initiation period, the p.d. Ae = eout. ecrev"between the outer surface and the crevice was small. Initially, both potentials rose to a maximum and then fell, ecrev, falling more rapidly than eout.. After a few hours, both potentials attained an almost constant value. -

-

504

G~STA KARL.BERG a n d G(SSTA WR,ANOLI~N

e. v]

pH 9 8

7 6 5

Crevice pH Steel A

4 3 2 1

0 -100

10

20

30

40

I

I

i

i

~ ~l.. . . . . . . . . . . . . . . . . . . . . . . . .

-200

-_

_-..,......~steelB ......

tlme,hrs

eout ecre v

eout

S',eel' A -300

ecrev

FIO. 3. Potential a n d p H change for steel A (13%Cr) a n d potential change for steel B (17°/oCr) during crevice corrosion. W i d t h o f crevice = 0'15 m m . e ...... = potential of spec m e n within crevice, eout = potential o f outer surface.

The initial pH of the electrolyte was 6. It began to rise as soon as the electrolyte solution penetrated the crevice. After about an hour, the pH reached a maximum value of 9.5 and then started to decrease slowly. If localized attack started just opposite the glass electrode, the pH started to decrease earlier and did not attain such a high maximum value. In order to find out how far the pH in the crevice could sink, the cell was allowed to operate for a long time (2-14 d) with steel A. The crevice solution was collected and its pH was found to be 4-0. The pH, locally, was probably still lower, however.

b. Influence of experhnental variables The attack occurred more readily on steel A with 13%Cr than on steel B with 17%Cr. Furthermore, the potentials eout. and ecrev"finally assumed lower values with steel A than with steel B and their difference was greater (150 mV for A as compared to 50 mV for B). It proved difficult to provoke a crevice attack quickly, if the specimen was kept in the desiccator for a long time between grinding and insertion in the cell. This time was therefore limited to 2 h. The attack was effectively retarded by buffering the test solution with 0. I M acetate buffer, by bubbling air into the crevice through a capillary hole or by using 3 %NazSO4 instead of 3 %NaC1 as the electrolyte. The p.d. measured 8 cm apart from the crevice mouth was 150 mV compared with 90 mV immediately inside the mouth for steel A. If the crevice width (the distance between cell wall and specimen) was increased from 0.15 to 0.9 mm, the former p.d. was decreased from 150 mV to 75 mV.

On the mechanism of crevice corrosion of stainless Cr steels

505

2. Critical pH valuefor activation Figures 4 and 5 illustrate how steels A and B behave in 0-1M acetate buffered 3 % N a C I solution of different p H and Oz content. According to Fig. 4, steel A does not stand p H 3.5 in Ar-saturated electrolyte but corrodes under Hz evolution. It does stand p H 4 in the same solution, however. It is also evident that steel A stands pH 3.5 with 02 bubbling. The difference between electrode potentials of steel A in

eh [mV] 400I

3OO[ 02, pH 3.5

Potential fluctuations

200 100

70 I J 10 ~ . ~

8,o

9o

loo _i

time, hrs

Air.pH 4.0

-100 I

-200[ j .

Ar, pH4.0

-300-~

Ar, pH 3.s

bright

._

dark coloured hydrogen evolution

F1o. 4. Free electrode of steel A (13%Cr) in acetate-buffered, 3%NaCI solution.

%rmvU 02,PH 3.o Air, pH 3.0

0'--

I

2,0

3o

4o

5o,

6o,

10

70

80

90

I

I

I

100 I tlme~hr5

-100 -200

Ar, pH 3.5

-300

At, pH 3.0

bright

dark coloured '--

hydrogenevolution

FIG. 5. Free electrode of steel B (17%Cr) in acetate-buffered, 3%NaCI solution.

506

G(SSTA K.ARLBERGand G(SSTA WRANGLI~N

O o-saturated and O2-free solution is as high as 700 mV. As regards steel B (Fig. 5), it stands pH 3.5 but not 3.0 in Ar-saturated solution. The p.d. for steel B in aerated and deaerated solution is 600-700 mV at pH 3-0.

DISCUSSION The results show that the initiation of a crevice attack does not involve the development of a great p.d. between aerated outer surface and air-free crevice. In the beginning, the two potentials are practically the same and the whole of the steel specimen is passive. Its small general corrosion in the passive state consumes 02 and produces OH-, however. The pH on the steel surface rises rapidly during the first part of the initiation period, obviously corresponding to the growth of a thicker oxide film than that formed in air. The potential of both the free electrodes exposed in the cell according to Fig. 2 and of the specimen in the crevice cell rises to begin with. If the oxide film is allowed to thicken for a long time in air before exposure to the salt solution, the initial change in pH and potential is smaller, however, and a crevice attack takes much longer time to develop. After the initial rise of ecrev,and eout. both potentials fall when crevice attack starts. Whereas the potential of the anode (crevice) surface is lowered due to activation, the potential of the cathode (outer) surface is lowered due to cathodic polarization. During the propagation of a crevice attack there develops between anode and cathode a certain p.d., whose magnitude is a measure of the rate of attack in accordance with Ohm's law. As the active surface within the crevice is increased, the corrosion current increases and so does the ohmic potential drop and the cathodic polarization of the outer surface. That steel A (13%Cr) gives a larger p.d. than steel B (17%Cr) is apparently due to the fact that the corrosion rate of stainless steel in its active state also is reduced by an increased Cr content. It is natural, furthermore, that an attack occurs where the crevice film is thinnest since both the Oz available and the buffer capacity is smallest in a thin film. The influence of the crevice width on the p.d. between anode and cathode is explained by the larger resistance and higher ohmic potential drop along a thin solution film. By the same reason, the potential measured is lower far into the crevice if the attack occurs at its inner end. That the pH of the crevice solution is increased rapidly immediately after filling the cell with electrolyte is obviously due to the fact that initially Oz is consumed and O H - is produced also within the crevice according to the cathode reaction: 02 + 2H20 + 4e- -+ 40H-.

(2)

After all 02 within the crevice is consumed, the cathode reaction (2) occurs entirel2~ on the outer surface whereas on the crevice surface mainly the anode reactions:

Fe ~ Fe 2÷ + 2e-

(3)

Cr ~ Cr 3+ + 3e-

(4)

and

On the mechanism of crevice corrosion of stainless Cr steels

507

occur. When the Fe ~+ and Cr 3+ concentrations reach certain values, hvdrolysis takes place: Cr 3+ + 3H20 -+ Cr(OH)3 + 3H +

(5)

Fe ~+ + 2HoO -+ Fe(OH)2 + 2H +.

(6)

These reactions generate H +, which means that pH is lowered again after having passed a maximum value. Since the glass electrode covers both anode and cathode surfaces to begin with, it does not respond adequately to this lowering of pH. Collection of the crevice solution and pH measurement showed pH 4 after a few days. It seems reasonable to assume that the effect of this acidification due to hydrolysis within the crevice is the same as that postulated in the theory of pitting. That buffering of the test solution prevented crevice attack is also an indication that the lowering of pH is a condition for activation. The relation between the critical pH for activation of a stainless steel and its tendency to crevice corrosion, reported by Fokin et al. 11 points in the same direction. It is interesting to note that the final pH value, measured in the crevice solution, coincides with the critical pH required for activation of the steels concerned in deaerated NaC1 solution. Below this pH value H, evolution takes place and it seems reasonable to assume that the corresponding consumption of H + ~revents the crevice pH from sinking to a still lower value. It should be noted, furthermore, that H + ions migrate out of the crevice to an increasing extent as their concentration is raised. The mechanism of crevice corrosion of stainless steels in neutral aerated C1solutions is illustrated by the schematic polarization diagrams for crevice and outer surface shown in Fig. 6. i,. denotes cathodic polarization curves for O0. reduction, and i, anodic polarization curves for stainless steel, which in the passive range are vertical lines. Potential values refer to the time-potential curves for steel A in Fig. 3. (a) In the beginning, the polarization diagrams for crevice and outer surface are identical. All the steel surface is passive, the air-formed oxide film allowing a fairly high anodic current density, however. The redox potential for the cathode reaction (2) above is given by RT Po~ e =- eo + zF- In ]OI_i_14

(7)

which shows that a change of pH causes a potential change 4 times that catlsed by a corresponding change of 02 concentration. Although both a lowering of 02 content and a rise of pH are factors which cause a lowering of the potential, the measured potentials are initially changed in positive direction, ill accordance with Figs. 3-5 and apparently due to a strengthening of the air-formed oxide film. (b) The anode curves are moved towards lower currents, both in the crevice and on the outer surface, due to a thickening of the oxide film and corresponding to more noble values of measured potentials. From the crevice towards the outer surface there flows in the solution a small current of the size (i,,)..... -- (ic)crev. = (ia)out. -- (ic)o~t., which means that there is an ohmic potential drop in the same direction. This is so

508

GOSTA KARLBERG and GOSTA WRANGLEN

CREVICE

OUTER

SURFACE

e ia

I00i

O.

0

......

_ec2r!_

-

-100 -200 -30(3

log i

I

100 b.

0 ~

-

~og~

e __~,_ (io)°ut

(iQlcrev I ...........

e c-°tr - "

-

0

-100 -200 I 2

-300

I

(ic)crev

log ~'

hQlcrev ............

IO0

('c)out

Io; i

(ialout ec_or_r_

e. o~i-.~," -'°°II I"'
o~ I t ~

I I

i~ I

'

I

L

m I

~

(Ic)crev

"

J

i

I

log i

hc)ou t

log i

!

-.,c

(ialcrev

100 d. 0 -f00 -200 -300

/

I (Ic)crev

100 e

I

I



log a

(ic)ou t

log i

ha)crev

0

-100 -200 -300

ecrev

__. I

log (ia)crev



e~,

/! i

log i ( iclout

FIG. 6. Mechanism of crevice corrosion initiation stage, i, and i, denote cathodic and anodic polarization curves. The subscripts ~ . and out refer to crevice and outer surface respectively. D i a g r a m e shows propagation stage. R. = internal (electrolyte) resistance in crevice corrosion cell. er~ = Flade potential.

On the mechanism of crevice corrosion of stainless Cr steels

509

small, however, that crevice a n d outer surface still have practically the same potential. (c) The difference in O,. content a n d c a t h o d i c c.d. between crevice a n d outer surface is now large. In the crevice, the a n o d e reaction d o m i n a t e s over the c a t h o d i c reaction, which means that the a m o u n t o f O H - f o r m e d in the crevice is negligible in c o m p a r i s o n with that o f the Fe z+ generated. A t all times we have the relation

(ia)crev. "~ (/a)ouL :

(ic) ..... "~ (/c)out."

(8)

(d) Hydrolysis o f Fe 2~- a n d Cr a÷ result in a lowering o f the p H o f the crevice solution. W i t h the successively increasing cell current C I - migrate into the crevice. T h e result is an acid solution o f F e 2+ and Cr s+ chlorides. Both H ÷ a n d C1- increase the rate o f the a n o d e reaction, H + by dissolving the oxide film, C I - by peptizing it. The acidification o f the crevice solution raises eFl a c c o r d i n g to e q u a t i o n (I) above. (e) eFl o f the crevice surface has now been raised so m u c h that the crevice surface is active whereas the outer surface is still passive. The p r o p a g a t i o n stage o f the crevice a t t a c k has been reached. The c o r r o s i o n current a n d hence also the p.d. (Ae) between crevice a n d outer surface is increased. Ae reaches a steady state value. I n the slot (crevice m o u t h ) Fe(OH)a accumulates. It is f o r m e d u n d e r O.z c o n s u m p t i o n a n d also renders Oz diffusion into the crevice m o r e difficult. Crevice c o r r o s i o n is self-prop a g a t i n g (autocatalytic) in the same way as pitting. F o r steel A, e ..... in the steady state is below the potential o f the h y d r o g e n electrode at p H 4 (i.e. - - 235 mV). As a final conclusion, it m a y be stated that several c o - o p e r a t i n g factors are necessary for crevice c o r r o s i o n o f stainless steels to occur. These are 1. The presence o f an o x i d a t i o n agent, usually oxygen, f o r m i n g a concentration cell between crevice a n d outer surface. 2. The presence o f activating anions, usually C1- ions. 3. A low e n o u g h buffer capacity to allow a substantial lowering o f p H within the crevice. Acknowledgement--This study was promoted in several ways by the Sandvik Steel Co., Sweden. The authors are indebted to Mr. J. Chr. Carl6n, Chief Engineer, and to Dr. Lars Troselius of this company for fruitful discussions.

REFERENCES I. H. H. UHUG, Corrosion Handbook, p. 172. John Wiley, New York (1940). 2. N. D. GREENE,J. electrochem. Soc. 107, 457 (1960). 3. G. J. SCHAFER,J. R. GABRIELand P. K. FOSTER,J. electrochen~. Soc. 107, 1002 (1960). 4. U. R. EVANS,J. electrochem. Soc. 108, 613 (1961). 5. [. L. ROSENFELDand I. K. MARSHAKOV,Corrosion 20, l15t (1964). 6. Yu. M. KOROVINand I. B. ULANOVSKII,Corrosion 22, 16 (1966). 7. P. M. SrRoccm, B. VICENTINIand D. SINIGAGLIA,Electrochim. Met. 1, 239 (1966). 8. I. ]3. ULANOVSKII,J. appl. Chem. U.S.S.R. 39, 768 (1966). 9. G. ~BOMBARA,D. SINIGAG-LIAand G. TACCANI,Electrochim. Met. 3, 81 (1968). 10. T. P. HOAR, Galvanotechnik u. Oberfli~chenschutz 9, 18 (1968). l 1. M. N. FOKIN, M. M. KURTEPOV, V. K. ZHURAVLEVand V. I. ORESHICtN,Proc. 1st Inst. Syrup. Water Desalination, Washington D.C., 1965, p. 2-531. U.S. Department of the Interior, Washington 0967). 12. M. G. FONTANAand N. D. GREENE, Corrosion Engineering, p. 54. McGraw-Hill, New York (1967). 13. U. R. EVANS, The Corrosion alul Oxidation o f Metals: First Supplementary Volume, p. 51. Edward Arnold, London 0968).

510 14. 15. 16. 17. 18. 19. 20. 21.

G0STA KARLaEgG and GOSTA WRANGL~I

O. L. RaGGS, Corrosion 19, 180t (1963). K. NOBE and R.. F. TOBtAS, Corrosion 20, 263t (1964). W. D. FRANCE and N. D. GREENE, Corrosion 24, 16 (1968). L L. R.OSENFELD and I. S. DANILOV, Corros. Sci. 7, 129 (1967). J. D. JACKSON and W. K. BOYD, A S T M Spec. Tech. Pnbl. 432, 219 (1968). J. C. GRIESS, Corrosion 24, 96 (1968). V. V. SKORCHELETTI and N. K. GOLUBEVA, Corro& Sci..5, 203 (1965). K. J. VETTER,Electrochemical Kinetics. Theoretical and Experimental Aspects, p. 752. Academic Press, New York (1967). 22. H. R.. CoPSON in LAQUE and COPSON (Editors), Corrosion Resistance o f Metals andAIloys, p. 10. Reinhold Publ. Corp., New York (1963). 23. J. KAESCHE, Werkstoffe Korros. 15, 379 (1964).