Structure and composition effects on pitting corrosion resistance of austenitic stainless steel after molybdenum ion implantation

Structure and composition effects on pitting corrosion resistance of austenitic stainless steel after molybdenum ion implantation

Surface & Coatings Technology 200 (2005) 2131 – 2136 www.elsevier.com/locate/surfcoat Structure and composition effects on pitting corrosion resistan...

364KB Sizes 0 Downloads 10 Views

Surface & Coatings Technology 200 (2005) 2131 – 2136 www.elsevier.com/locate/surfcoat

Structure and composition effects on pitting corrosion resistance of austenitic stainless steel after molybdenum ion implantation N. Mottu, M. Vayer*, J. Dudognon, R. Erre Centre de Recherche sur la Matie`re Divise´e, 1b rue de la Fe´rollerie, F45071 Orleans Cedex 2, France Received 20 April 2004; accepted in revised form 17 December 2004 Available online 8 February 2005

Abstract 316LVM, cold worked austenitic stainless steel, was implanted at 49 keV with molybdenum ions. Implantation doses varied between 11015 and 3.51016 ions cm 2. The structure of the implanted layer was examined by grazing incidence X-ray diffraction and the chemical composition was characterized by X-ray photoelectron spectroscopy combined with argon sputtering. Pitting corrosion studies were carried out on both unimplanted and implanted stainless steels in neutral chloride medium. The relationship between the pitting corrosion resistance, the structural and chemical modifications induced by Mo implantation was discussed. As a function of molybdenum ion dose, an expansion of fcc austenite was first observed, then above 81015 ions cm 2 a new bcc structure appeared and finally the implanted layer was partially amorphized. Electrochemical studies revealed that ion implantation enhances the pitting corrosion resistance. Increase in molybdenum implantation dose was beneficial up to 81015 ions cm 2 in improving the pitting corrosion resistance, beyond which it had a detrimental effect. D 2005 Elsevier B.V. All rights reserved. Keywords: Austenitic stainless steel; Ion implantation; Molybdenum; Pitting corrosion; SEM; XPS; GIXD

1. Introduction Molybdenum ion implantation could harden the surface of steel [1,2] and increase the pitting corrosion resistance of stainless steel [3,4]. Few corrosion studies have been performed to characterize Mo ion implantation effects on steels [3–5]. Some studies have been carried out to describe the passivation layer [6,7]. These studies show a relative Cr enrichment at the surface of Mo-implanted samples as compared with unimplanted samples. Other studies investigated the structural changes [8]. Ion implantation, by inducing defects, phase transformations, sputtering and amorphization introduces structural and chemical changes on the material surface. The induced modifications (structural changes, chemical changes, corrosion resistance) are highly dependant of the ion implantation conditions (ion energy, dose, ion current density).

* Corresponding author. Tel.: +33 238 494737; fax: +33 238 417043. E-mail address: [email protected] (M. Vayer). 0257-8972/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2004.12.017

In order to understand the beneficial effect of Mo for pitting corrosion resistance, it is fundamental to correlate the characteristics of the material surface and the corrosion behavior in the same study. Our study was thus focused on an AISI 316LVM austenitic stainless steel, used for medical applications. Its behavior in a medium representing the human body allowed its biocompatibility investigation. We investigated the implanted layer structure and composition, the surface topography and the passive layer composition and then we evaluated different aspects of the corrosion resistance in neutral chloride media. The structure of the implanted layer, the chemical composition of the passive layer, the topography of the surface and the electrochemical behavior were then correlated.

2. Experimental The chemical composition of the AISI 316LVM coldrolled austenitic stainless steel was (at.%) 18.71 Cr, 13.08 Ni, 1.71 Mn, 1.61 Mo, 1.03 Si, 0.29 S, 0.07 C, 0.04 P, and

2132

N. Mottu et al. / Surface & Coatings Technology 200 (2005) 2131–2136

remainder Fe. Samples were cut from a bar of 14 mm in diameter. They were mechanically wet ground, polished and smoothed to get a mirror finish. Surface roughness (RMS), determined by AFM analysis, was about 2 nm. 98 Mo+ was implanted at room temperature with an energy of 49 keV corresponding to an implanted range of 30 nm with a maximum at 12 nm. Doses varied between 11015 and 3.5 1016 ions cm 2 and current density was 2 AA cm 2. Grazing incidence X-ray diffraction (GIXD) were performed on a PHILIPS parallel beam horizontal diffractometer using Cu-Ka source (1.54056 2). The information depth depends on the incidence angle. Use of a 0.48 incidence angle allows the structural investigation of the implanted layer (information depth of 13 nm (depth for 63% signal)). For 0.58 incidence angle the information depth is 27 nm. The peak positions were evaluated after correction of refraction effect due to grazing incidence. Angles were measured with a 0.038 precision; interplanar and lattice spacings were then determined with a 510 4 nm precision. The electrochemical behavior of the steel surface was studied in a doubled-walled glass cell, with a circulating water apparatus. The area of the steel working electrodes was fixed to 1 cm2. The counter-electrode was a platinum foil and the reference electrode was a saturated calomel electrode (SCE). Free potential measurements and galvanostatic curves were recorded at ambient temperature and in darkness in a NaCl 0.9% solution with an EGG 273A potentiostat driven by a software. The free potential was recorded for 90 min. It stabilized after 60 min except for unimplanted sample. Galvanostatic mode analysis consisted in applying a jump of current density (50 AA) after 60 min and following the evolution of the potential according to time (E=f(t)). Potentiodynamic curves of the implanted and unimplanted steel surfaces were recorded with a PGP 201 Radiometer Analytical potentiostat driven by a software. All the experiments were carried out in a NaCl 0.9% solution at 37 8C. Prior to measurements, samples were exposed to the chloride solution for 1 h so that the potential (free potential) could stabilize. Then the potential was increased in the rate of 8 mV/ min up to a potential corresponding to a current density of 500 AA cm 2. The potential was maintained to this value for 1 min and then decreased in the same rate. Current density was recorded versus potential (I=f(V)). The values of the characteristic potentials were given with an uncertainty of F20 mV gathering uncertainty on the reproducibility of the experiments. Implanted and unimplanted surfaces were examined before and after the corrosion attack with a Hitachi S4200 scanning electron microscope (SEM), using an acceleration voltage of 20 kV. Secondary electron mode was used to characterize the surface morphology and EDX mode (Energy Dispersive X-ray Analysis) to determine the elemental composition. X-ray photoelectron spectroscopy (XPS) measurements were performed on a VG ESCALAB 250 equipped with a

multidetection analyzer controlled by VG eclipse software, a 200W Al Ka source and a VG EXO 5 ion gun. The basic vacuum was 10 7 Pa. The reference energy was C1s at 284.6 eV (contamination carbon). All the samples were cooled down to liquid nitrogen temperature in order to limit diffusion process. They were alternatively sputtered by argon ion (4 keV, PAr=8.10 5 Pa, 2 AA cm 2) and analyzed by XPS in order to establish depth profile (concentrations and chemical species). We used a reference sample to estimate the sputter rate ((0.003F0.001) nm/s). A survey spectrum was first recorded to identify all elements present at the surface and then high-resolution spectra of the following regions were recorded: Fe2p, Mo3d, C1s, Cr2p, Ni2p, O1s. All the XPS recorded peaks were decomposed. For iron, three components were identified, namely Fe0 at 707.0 eV, Fe2+ at 708.2 eV and Fe3+ at 710.4 eV. For chromium Cr2p, we identified Cr0 at 574.0 eV, Cr in Cr2O3 at 576.4 eV, Cr in Cr(OH)3 at 577.5 eV, Cr in CrO3 at 578.1 eV. For oxygen O1s, 3 components were separated, namely O2 at 530.2 eV, OH at 531.7 eV, H2O at 533.0 eV. Molybdenum Mo3d had different states, Mo0 at 227.0 eV, Mo in MoO2 at 229.3 eV, Mo in Mo(OH)4 at 230.2 eV, Mo in MoO3 at 232.2 eV. Ni2p presented only one peak at 852.5 eV. All these binding energies were in agreement with the literature [9–17]. Concentration depth profiles were also corrected by taking into account the preferential sputtering of iron, chromium and nickel compared to molybdenum [18].

3. Results 3.1. Composition depth profiles The composition depth profile obtained by XPS coupled with argon sputtering of unimplanted sample (Fig. 1a) could be divided into three layers: (i) the contamination layer (0.3 nm thick) containing only carbon species, (ii) the passive layer (1.5 nm thick) characterized by the presence of an oxygen peak, (iii) the bulk containing the nominal concentration of the stainless steel where there was no oxygen. The passive layer contained only Fe and Cr. These species were in oxide and hydroxide forms. Ni, in metallic state, presented a high concentration at the interface between the passive layer and the bulk (20 at.%). Ni in metallic state segregated at this interface. The passive layer could be divided into two parts: the outer one contained oxygen into OH and H2O state and the inner part contained mainly oxygen in oxide form. For the implanted samples (Fig. 1b,c,d,e), the concentration depth profile could be divided into four layers: (i) the contamination layer (0.3 nm thick) containing only carbon species, (ii) the passive layer (2.5 nm thick) characterized by the presence of oxygen species, (iii) the implanted layer about 30 nm thick characterized by an excess of molybde-

N. Mottu et al. / Surface & Coatings Technology 200 (2005) 2131–2136

b 70 60 50 40 30 20 10 0

Ni Cr O

0

1

2

3

Fe C Mo

4

5

composition (at.%)

composition (at.%)

a

70 60 50 40 30 20 10 0

Ni Cr O

0

1

depth (nm)

2

3

Fe C Mo

4

5

depth (nm)

d

70 60 50 40 30 20 10 0

Ni Cr O

0

10

20

composition (at.%)

c composition (at.%)

2133

Fe C Mo

30

60 50 Ni Cr O

40 30 20 10 0 0

40

1

composition (at.%)

2

3

4

5

depth (nm)

depth (nm)

e

Fe C Mo

70 60 50

Ni Cr O

40 30

Fe C Mo

20 10 0

0

10

20

30

40

depth (nm) Fig. 1. (a) Raw XPS atomic profile of an unimplanted sample. Depth range 0–5 nm. (b) Raw XPS atomic profile of a 81015 Mo+/cm2 implanted sample. Depth range 0–5 nm. (c) Raw XPS atomic profile of a 81015 Mo+/cm2 implanted sample. Depth range 0–40 nm. (d) Raw XPS atomic profile of a 2.51016 Mo+/cm2 implanted sample. Depth range 0–5 nm. (e) Raw XPS atomic profile of a 2.51016 Mo+/cm2 implanted sample. Depth range 0–40 nm.

num compared to the bulk (see Fig. 1c and e) and (iv) the bulk. Molybdenum concentration increased logically with the implantation dose in the implanted layer (comparison between Fig. 1c and e). The ratios between the Fe and Ni concentrations in this layer were comparable for unimplanted and implanted sample. However, the ratio between Fe and Cr concentrations was slightly higher (20%) for implanted samples than for unimplanted samples. Mo implantation leaded to a relative depression of Cr in the implanted layer. The passive layer formed on Mo-implanted samples contained Mo, Fe and Cr. All these species were in oxide or hydroxide forms. No metallic species was present in the passive layer. We also observed Ni segregation at the interface passive layer/implanted layer. We thus considered that the passive layer was constituted of oxidized chromium (oxide or hydroxide), oxidized molybdenum (oxide or hydroxide) and oxidized iron (oxide or hydroxide). We

determined the concentrations of these components in the passive layer. As displayed in Table 1, Mo implantation led to an enrichment in oxidized Mo and Fe combined with a depletion of oxidized Cr. As the doses increased, oxidized Fe decreased in the same proportion than oxidized Cr and oxidized Mo increased. 3.2. Structural characterization Unimplanted samples (Fig. 2) presented an face cubic centered g austenitic structure in the outermost layer as in the bulk with a lattice parameter of 3.602 2. Mo ion implantation modified only the near surface layer since only the diffraction spectrum for 0.48 and 0.58 incidence angles were modified, the 1.58 diffraction spectrum remained unchanged [19]. An expanded austenite was observed with a lattice expansion, up to 1% (Fig. 2). From 1.51016 ions cm 2, a new a ferritic structure was

Table 1 Composition of the passive layer of unimplanted and Mo implanted samples Implantation dose in ions cm Oxidized Fe [%] Oxidized Cr [%] Oxidized Mo [%]

2

Unimplanted

81015

1.51016

2.01016

3.01016

3.51016

68 32 0

80 16 4

75 13 12

68 13 19

65 12 23

54 10 35

2134

N. Mottu et al. / Surface & Coatings Technology 200 (2005) 2131–2136 0.25

α(200) Mo 1.5

1016

potential (V)

intensity (au)

α(110) 2

ion/cm

Mo 8.1015 ion/cm2

38

γ(200) 48

unimplanted l

58

γ(220)

68

88

3.3. Surface reactivity results Surface reactivity was characterized by studies of the electrochemical behavior. We successively examined the evolution of the free potential in solution, the evolution of the current density versus potential in potentiodynamic mode (measurement of pitting and repassivation potentials), the evolution of the potential versus time for an imposed current density (50 AA) in galvanostatic mode (measurement of the electrochemical attack resistance). Fig. 3 shows the influence of the ion implantation dose on the value of the free potential. This free potential and its evolution characterize the stability of the passive layer. Unimplanted samples had a free potential which decreased with time. The passive layer became more and more unstable with time. Implanted samples presented metastable pits (pitting followed by an immediate repassivation) at the beginning of their immersion in solution. The creation of defects and the increase of the surface roughness, induced Table 2 Structural characteristics of unimplanted and Mo implanted samples

0.3602 0.3616 0.3618 0.3622 0.3638 0.3630 0.3626

0

Mo 3.5 1016 ion/cm2

unimplanted

TA

60 54

Ferrite lattice parameter (nm)

0.2914 0.2917

500

1000

1500

2000

2500

3000

3500

time (seconds)

formed. With high doses from 2.01016 ions cm 2, the diffracting structures were destroyed and this leaded to a partial amorphization. These modifications are presented in Table 2.

Unimplanted 11015 21015 41015 81015 1.51016 2.01016

0.05

0

Fig. 2. Diffraction spectrum (incidence angle=0.48) of an unimplanted sample and for 21015 Mo+/cm2, 81015 Mo+/cm2, 1.51016 Mo+/cm2 and 3.51016 Mo+/cm2 implanted samples.

Maximum austenite lattice parameter (nm)

0.1

-0.1

γ(311) 78

2θ ((°))

Mo+/cm2 doses

0.15

-0.05

Mo 2.1015 ion/cm2

γ(111)

Mo 2.1015ion/cm2

Mo 8.1015 ion/cm2

0.2

Mo 3.5 1016 ion/cm2

R

19 10

TA—ratio between intensities of amorphized diffracting structure (base of (111) peak) and of the well-defined diffracting structure (area under (111) peak). R—ratio between the ferrite and the austenitic peaks area in arbitrary units.

Fig. 3. Free potential curves in NaCl 0.9% solution at 25 8C of an unimplanted sample and of 21015 Mo+/cm2, 81015 Mo+/cm2 and 3.51016 Mo+/cm2 implanted samples.

by the collision effect of ion implantation, generated metastable pits formation. This observation agrees well with roughness measurements obtained by AFM which showed that all the implanted samples had a higher roughness than the unimplanted samples (cf. 3.4). After 2000s, molybdenum implanted samples had free potential values higher than those of the unimplanted samples. Molybdenum ion implantation thus improved the stability of the oxide layer for implanted samples compared to unimplanted samples. However, free potential values depended on the molybdenum implantation dose. Indeed, the samples implanted with low doses (lower than 1.51016 ions cm 2) had an oxidation layer which was more stable than those implanted with high doses. In potentiodynamic mode the evolution of the current density was recorded versus potential. The pitting potential was evaluated either by the first current density jump (noted Epit) or by a current density value of 100 AA/cm2 (noted Epit*). Scanning decreasing potential after pitting gave access to the value of repassivation potential (noted Erp), potential from which samples were not sensitive any more to pitting. The shape of the curve of polarization of unimplanted samples (Fig. 4) indicates that: – pits were initiated very quickly ( 50 mV) but developed very slowly in solution (current did not increase significantly between 50 mV and 710 mV), – the surface was attacked (many corrosion pits and/or very deep pits) and the corrosion pits were not repassived quickly (broad loop of the curve return). The implantation dose influenced highly the polarization curves form (Fig. 4). Corrosion pits were formed very quickly in the case of the unimplanted samples (very low values of Epit) contrary to the implanted samples (Epit shifted towards greater values). On the other hand, corrosion pits development was faster for the implanted samples compared to the unimplanted samples (lower values of Epit*). Implantation enhanced the repassivation capacity.

N. Mottu et al. / Surface & Coatings Technology 200 (2005) 2131–2136 3.E-03

1.5 unimplanted l

Mo 3.5 1016 ion/cm2

15

Mo 8.10

1.4 1.3

2

ion/cm

potential (V)

3.E-03

2135

2.E-03 2.E-03

Mo 1.1015 ion/cm2

1.E-03

Mo 1.5 1015

1.2

Mo 8.1015 ion/cm2

1.1 Mo 3.1016 ion/cm2

1 0.9

unimplanted

0.8

5.E-04

0.7 0.E+00 -0.10

0 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

potential (V) Fig. 4. Polarization curves in NaCl 0.9% solution at 37 8C of an unimplanted sample and 11015 Mo+/cm2 , 81015 Mo+/cm2 and 3.51016 Mo+/cm2 implanted samples.

Indeed, the loop carried out by the curve of return in potential was less broad and indicated the formation of pits less deep and/or fewer (Table 3). Curves obtained in galvanostatic mode for the unimplanted and the molybdenum implanted samples are presented in Fig. 5. Pitting corrosion resistance was improved for molybdenum implanted samples. Samples implanted with high doses (2.01016 to 3.51016 ions cm 2) had an electrochemical behavior similar to a sample implanted with low dose (21015 ions cm 2). They resisted to the application of the current for 15 to 20 min then their potential decreased. The implanted sample with a dose of 81015 ions cm 2 was the sample which resisted the best to the electrochemical attack because its potential fell only after 50 min. The increase in the implantation dose improved the resistance of steels up to an implantation dose of 81015 ions cm 2. There was an optimal dose for which the corrosion resistance in galvanostatic mode was improved. Whatever the implantation dose, molybdenum implanted samples behavior was always better than that of the unimplanted samples. 3.4. Surface topography The surface roughness was evaluated by means of AFM. The RMS roughness of the unimplanted samples was 2.5 nm. Ion implantation increased the RMS roughness. The Table 3 Characteristics of the potentiokinetic curves of the implanted and unimplanted samples Mo+/cm2 doses Unimplanted 11015 21015 41015 81015 1.51016 2.01016 3.01016

Ecorr (mV) 180 140 130 100 180 190 220 240

Epit (mV)

EpitT (mV)

55 350 310 300 290 260 230 90

715 465 330 360 340 330 300 250

Erp (mV) 95 100 120 60 130 130 100 100

500

1000

1500

2000

2500

3000

3500

time(seconds) Fig. 5. Curves in galvanostatic mode with a current of 50 AA in NaCl 0.9% solution at 37 8C for an unimplanted sample and 11015 Mo+/cm2, 81015 Mo+/cm2 and 3.01016 Mo+/cm2 implanted samples.

dose seemed to have no influence on the roughness. The RMS roughness was 2.9 nm. Fig. 6 presents pits morphologies obtained by SEM for unimplanted and implanted samples. For implantation doses up to 151015 ions cm 2, a fine film covered pits which had a round shape. EDX analyses indicated that this film was enriched in molybdenum. This film was no longer visible for the implantation with high doses (starting from 151015 ions cm 2) and the pit morphology was highly modified (many gathered pits). Optical microscopy allowed us to determine the repartition in the size of the corrosion pits and to evaluate the proportion of the surface area covered by pits. Unimplanted samples and implanted samples with a low doses (up to 1.01016 ions cm 2) had similar pits size. The surfaces area covered by the pits were also similar (near 0.3% of the total surface area).The pits sizes had a diameter comprised between 5 and 40 Am with a mean diameter of 10 Am. For the implantations with high doses (from doses from 1.51016 Mo ions cm 2), the proportion of the surface area covered by corrosion pits highly increased (3%). The mean pits size also increased (25 Am) with some very large pits.

4. Discussion and conclusion As we saw previously, the samples had a complex behavior with respect to corrosion resistance and different aspects of electrochemical behavior of the samples have to be considered. The samples can be then divided into two groups, the transition dose is 81015 ions cm 2:! up to a dose of 810 15 ions cm 2 . Mo implantation induces: –an improvement of the stability of the oxidation layer; –an improvement of pitting corrosion resistance; –an improvement of the repassivation of pits created by electrochemical attack. The size of the pits and the proportion of surface area covered by pits are the same as those observed for unimplanted samples.

2136

N. Mottu et al. / Surface & Coatings Technology 200 (2005) 2131–2136

Fig. 6. Corrosion pits morphology imaged by SEM after electrochemical attack of (a) an unimplanted sample (white scale length=10 Am), (b) an 81015 Mo+/ cm2 implanted sample (white scale length=2 Am), (c) an 3.01016 Mo+/cm2 implanted sample (white scale length=50 Am).

The electrochemical behavior of these implanted samples is better than this of unimplanted samples. This has to be connected to all the characteristics of the samples. Mo implantation leads to an increase of molybdenum concentration in the implanted layer and consequently in the passive layer. The beneficial effect of molybdenum for corrosion resistance has been already demonstrated by many authors [11,20,21] when molybdenum is introduced as alloying element. They have demonstrated that the presence of molybdenum in the passive layer plays a fundamental role. The surface topography can play a major role in pitting corrosion [22] since holes, strays and surface defects are preferential sites where pitting corrosion takes place. We evaluated the surface topography by determining the RMS roughness. The implanted samples are rougher than the unimplanted sample. Consequently if we consider only the roughness the electrochemical behavior must be worse and consequently another parameter with a higher influence must explain the better electrochemical behavior of these implanted samples. Implanted samples have the same structure (austenite) as the unimplanted samples. The lattice parameter of austenite only increases.! starting from a dose of 810 15 ions/cm 2 . Implantation dose increase induces: –a reduction in the stability of the oxidation layer, –a reduction in the pitting corrosion resistance, –an improvement of the repassivation of the pits created by electrochemical attack, –an increase in the size and density of the pits. In the same time, the increase of molybdenum implantation dose increases the molybdenum concentration in the implanted layer and consequently in the passive layer. This must be have a beneficial effect of corrosion resistance, as it has been previously discussed. The increase of implantation dose has a high influence on the observed structure of the implanted layer. Indeed, in this group, ferritic structure appears and then amorphization takes place. Austenitic structure has a better corrosion resistance than the ferritic structure. Moreover, the amorphization process is not favorable to better corrosion resistance since it creates defects and is responsible for degradation of

the oxide layer stability. Size of pits and proportion of surface area covered by pits increase. Improvement of the repassivation of the passive layer observed in this group has to be related to the increase of Mo concentration in the passive layer.

References [1] Yu.P. Sharkeev, B.P. Gritsenko, S.V. Fortunab, A.J. Perry, Vacuum 52 (1999) 247. [2] Yu.P. Sharkeev, E.V. Legostaeva, S.V. Panin, B.P. Gritsenko, Surf. Coat. Technol. 158–159 (2002) 674. [3] X. de Buchere, P. Andreazza, C. Andreazza-Vig nolle, C. Clinard, R. Erre, Surf. Coat. Technol. 80 (1996) 49. [4] H.S. Isaacs, S.M. Huangs, J. Electrochem. Soc. 143 (1996) 277. [5] P.D. Hicks, F.P.A. Robinson, Corros. Sci. 24 (1984) 885. [6] M.F. Lopez, A. Gutie´rrez, F.J. Pe´rez, M.P. Hierro, F. Pedraza, Corros. Sci. 45 (2003) 2043. [7] A. Gutie´rrez, M.F. Lopez, F.J. Pe´rezTrujillo, M.P. Hierro, F. Pedraza, Surf. Interface Anal. 30 (2000) 130. [8] N. Mottu, M. Vayer, P. Andreazza, Th. Sauvage, G. Blondiaux, R. Erre, Surf. Coat. Technol. 151–152 (2002) 47. [9] V. Di Castro, S. Ciampi, Surf. Sci. 331–333 (1995) 294. [10] Y. Okazaki, T. Tateishi, Y. Ito, Mater. Trans., JIM 38 (1997) 78. [11] G. Latha, N. Rajendran, S. Rajeswari, J. Mater. Eng. Perform. 6 (1997) 743. [12] D. Briggs, M.P. Seah, Practical surface analysis by Auger and XPS, John Willey and Sons, Chichester, 1990. [13] A.M. Beccaria, G. PoggiI, G. Castello, Br. Corros. J. 30 (1995) 283. [14] I. Olefjord, B. Brox, U. Jelvestam, J. Electrochem. Soc. 132 (1985) 2854. [15] B. Stypula, J. Stoch, Corros. Sci. 36 (1994) 2159. [16] N. Moncoffre, G. Hollinger, H. Jaffrezic, G. Marest, J. Tousset, Nucl. Instrum. Methods Phys. Res., B Beam Interact. Mater. Atoms 7/8 (1985) 177. [17] D. Rats, L. Vandenbulcke, R. Herbin, R. Benoit, R. Erre, V. Serin, J. Sevely, Thin Solid Films 270 (1995) 177. [18] N. Mottu, M. Vayer, R. Benoit, R. Erre, Surf. Interface Anal. 34 (2002) 276. [19] N. Mottu, M. Vayer, R. Erre, Surf. Coat. Technol. 183 (2004) 165. [20] H.C. Brookes, J.W. Bayles, F.J. Graham, J. Appl. Electrochem. 20 (1990) 223. [21] K. Sugimoto, Y. Sawada, Corros. Nace 32 (1976) 347. [22] M. Pourbaix, L. Klimzack-Mathieu, Ch. Mertens, J. Meunier, J. Laureys, L. Neelemans, M. Warzee, Corros. Sci. 28 (1963) 135.