TECHNICAL PAPERS

-

ELECTROCHEMICAL SCIENCE AND TECHNOLOGY

The Secondary Passive Film for Type 304

Stainless Steel in 0.5 M H 2SO4 A. Atrens The Department of Mining, Minerals, and Materials Engineering, The University of Queensland, Brisbane, Queensland, Australia 4072

B. Baroux* and M. Mantel Ugine Research Centre, F-73400 Ugine, France ABSTRACT An examination has been carried out of the secondary passive film on Type 304 stainless steel in 0.5 M HSO4. The characterization techniques used were electrochemical (potentiodynamic, potentiostatic, and film reduction experiments) and surface analytical. A bilayer model for the secondary passive film is proposed. It appears that next to the metal, there is a modified passive film which controls the electrochemical response; i.e., governs the current for any applied potential. On top of this modified passive film, the experimental data are consistent with a "porous" corrosion-product film which adds to the total film thickness but has little influence on the electrochemical response. The composition of the secondary passive film corresponds most probably to a mixed Fe/Cr oxide/hydroxide enriched in Cr3 +, with a composition similar to a primary passive film.

Introduction The aim of this study was to examine the secondary passive film on Type 304 stainless steel in 0.5 M HSO4 in the secondary passivity potential region, typically at potentials more positive than about 940 mVscE. The primary interest was in the characterization of the secondary passive film as it might be used in a technological application; that is after the film had been formed, dried, and exposed to air. Also of interest were the insights that could be gained concerning the secondary passive film by a study of film formation. The literature suggests three different possibilities for surface films in the secondary passive potential range: (i) a modified continuous passive film, (ii) a porous film, and (iii) a hydroxide gel. Possibility (i) derives from the passivity literature.'-" Possibilities (ii) and (iii) derive from the literature on colored stainless steels.35 43 The (primary) passive film on stainless steels has a complex structure'- 3 but can be thought of as a very thin (-1 to 3 nm) film composed of metal oxides/hydroxides and bound water. For stainless steels, the Cr enrichment in the bulk of the passive film is evidenced by a Cr°X/(FeOX + Cr °X) ratio significantly higher than the ratio Crm/(Fem + Crt) characteristic of the alloy 124'11 -13'16-22 There is further enrichment of Cr in the oxidized state, Cr° X, at the film/solution interface for films formed in acid conditions.l-.'3 The passive film grows on the metal surface by an electrochemical mechanism, covers the metal surface as a continuous layer, and inhibits corrosion by forming a barrier between the corroding metal and the environment. The film has a selfrepair property (under conditions such that the passive film is stable) because the film formation rate is highest on a film-free surface (as at a break in the film) and the film growth rate decreases rapidly with film thickness.4 '0 Thus, * Electrochemical Society Active Member.

the film can be thought of as having a steady-state thickness, although attainment of the steady-state thickness can take up to 100 h in acid solutions.4' 0 Both Clayton and Olefjord2 and Macdonald 4 2 2- 6 propose a bilayer structure for the primary passive film consisting of an inner barrier layer (consisting mostly of Cr2 O, for stainless steels') and an outer precipitate layer. This composite structure has been postulated by Sato 28 '29(see also Ref. 30, 31) to have ion selective properties which can be important in the formation stages of the passive film. Furthermore, ions from the solution can be incorporated in the outer precipitate layer 22 23' and provide increased protection against passivity breakdown induced by chloride ions and consequent pitting corrosion. Sugimoto and Matsudal s analyzed passive and secondary passive films using ellipsometry for films formed on FeCr, alloys in NaSO at pH 2.0 and 6.0. For Fe-Cr,0,, the passive film thickness increased linearly over the whole potential region from 0.4 to 1.3 VSCE; this behavior is similar to that for pure Fe measured by Sato et al. 4 For FeCr. (x = 0.15 and 0.20) the primary passive films increased linearly in thickness with increasing potential in the range 200 to 800 mVsce (consistent with the other passivity literature), there was then a transition region in which the film thickness was approximately constant, and then the secondary passive films increased linearly in thickness with increasing potential in the range 1000 to 1400 mVsc. The secondary passive films with average film thicknesses of 2.5 to 3.0 nm were somewhat thicker than the primary passive films (with average thicknesses 1.5 to 2.0 nm),15 and they had somewhat different optical constants, indicating that there were differences between the passive and secondary passive films. Similar trends were measured by Song et al.3 using ellipsometry for type 304 stainless steel in 0.5 M H2SO 4 at ambient temperature. Song et al." using x-ray photoelectron spectroscopy (XPS), found high levels

J. Electrochem. Soc., Vol. 144, No. 11, November 1997 © The Electrochemical Society, Inc.

3697

3698

J. Electrochem. Soc., Vol. 144, No. 11, November 1997 © The Electrochemical Society, Inc.

of Cr 3+ and some Cr6+ in secondary passive films formed in 0.5 M Na2 SO4 . Kirchheim et al.4 used XPS and Ar sputtering to measure concentration profiles for passive and secondary passive films formed on Fe-1OCr in 0.5 M H2 SO4 . The profiles of Cr°X/(Fe° x + Crox) were quite similar for both types of films with the secondary passive film (2.0 nm) being slightly thicker than the primary passive film (1.8 nm). Both were enriched in CrOx in the bulk of the film

-1 NIHq

-... ·

t

Tt

-I

ub-

.0 N! I

Experimental Type 304 stainless steel was used having the following chemical composition: 17.4% Cr, 8.4% Ni, 0.17% Mo, 1.4% Mn, 0.49% Si, 0.059% C, 0.15% Cu, 0.12% Co, 0.05% V, 0.013% Sn, 0.024% P, 0.0015% S, 0.05% N, 40 ppm 0. This was received from the mill in the bright annealed condition'(in-line continuous sheet anneal in a hydrogen atmosphere for about 5 min at 1000°C followed by rapid cooling in the same atmosphere) which ensures that all species are in solid solution and produce a bright stainless surface. The standard specimen preparation was a wet metallurgical polish to 1200 grit with silicon carbide, gentle air dry and exposure to laboratory air (of ambient humidity) for about a day (i.e., exposure to laboratory air for at least 8 h). The electrochemical apparatus' is shown in Fig. 1; this allowed the specimen to be introduced into the specimen cell and the specimen surface to be first deaerated by an equal mixture of H2 and N2, while this gas mixture was used to deaerate the electrolyte for 1 h in the deaeration a The electrochemical apparatus is one developed for studies of passivity and pitting corrosion. In studies of pitting corrosion, crevice corrosion at the specimen edges is always a concern and often a practical problem that masks the pitting corrosion that is being studied. This problem has been overcome4 in a number4 74of ways. One solution is that adopted by Baroux" and Atrens " where very careful attention is paid to deaeration of the sample surface and the deaeration of the test solution. This same approach was used in the present study. After the specimen is introduced into the specimen cell, the specimen is exposed to a flowing gas mixture of equal parts H and N2 . This gas flow removes oxygen from the specimen cell, removes (lightly bound) oxygen from the specimen surface and also removes oxygen from crevices at the specimen edges. Subsequently deaerated electrolyte is introduced into the specimen cell. For pitting studies, the procedure is useful in preventing crevice corrosion. If a less rigorous procedure is used, then there could be some oxygen trapped in crevices at the specimen edge and this could cause initiation of crevice corrosion. In our studies, removal of all oxygen was also a consideration to ensure low values of residual current in the reduction experiments. Our procedure was successful in producing low residual reduction values as indicated by the i values in Table III.

->N,fU, _

,I.i

Lab

'e aI i %

~1.i

I-I

.i

to similar levels with Cr°X/(Fe ° x + CroX) - 0.4. The thick-

ness of the CrOX region was somewhat greater in the secondary passive film. The primary passive film was further enriched in Crox at the very surface as discussed above. Coloring of stainless steel was developed 3 5 as a chemical anodization process. This was modeled 3 6-40 as due to the chemical precipitation of Fe(OH)3 and Cr2 (SO4 )3, 18H2 0 when their solubilities were exceeded locally at the metal surface. In contrast to the production of the passive film by electrochemical oxidation, these corrosion products were assumed to be produced by a chemical precipitation process. Porous corrosion products were assumed which did not influence the corrosion process, and the electrochemical response was modeled on the basis of the diffusion of ionic species in the liquid phase of the pores in the film. In this context it is worthwhile considering the properties of a hydroxide gel. Metal hydroxides can precipitate from solution to form a gel; there forms a skeletal threedimensional network of molecules throughout the medium giving the medium solidlike rigidity although these molecules (i.e., the solid) form a small fraction (-10 to 30%) of the total volume. Electrochemical corrosion is as for direct contact of the metal with liquid, as there remains a direct conduction path through a liquid phase of the gel. In the present study, secondary passive film formation was characterized by measuring potentiodynamic polarization curves and by carrying out potentiostatic experiments. Film characterization after film formation was carried out by electrochemical reduction and XPS.

?tfirWfoAg1A& swa 1_-C@n 4 W d(t)

1~- -

ki J-I%

l.-

,, ssamce

! ve

fK

Fig. 1. Electrochemical cell. cell. All solutions were made using analytical grade reactants and deionized water. The deaerated electrolyte was then introduced into the specimen cell without exposure to air, and electrochemical measurements were started a period of 2 to 4 min after electrolyte contact was established with the specimen. Prior experiments have determined that the oxygen concentration in the solution is around 10 ppb under these conditions. The counterelectrode was a platinum sheet. The reference electrode was Ag/AgCl. The potentials are reported with respect to a saturated calomel electrode to allow easier comparison with the literature. The temperature was maintained at 23°C by the circulation of the thermostated water through the outer water jacket of both the specimen chamber and the deaeration cell. This electrochemical apparatus was used for (i) characterization of film formation by measuring potentiodynamic polarization curves and by carrying out potentiostatic experiments, and (ii) film characterization after electrochemical film formation by electrochemical reduction. A rapid overview of the electrochemical behavior relevant to film formation was obtained by measuring the potentiodynamic polarization curve, in 0.5 M H2 SO4 . For comparison polarization curves were also measured in 5 M and 6 M H2 SO4 ; a borate buffer, 0.15 M sodium borate (Na2 B4 O, · 10 HO) + 0.15 M boric acid (H3BO3 ) at pH 8.4; and in a phosphate solution (4.7% Zn3(PO,) 2 + 9.3% H3PO 4 which had a pH of 1.47). All polarization curves were measured starting from a cathodic potential (-800 mVscE) at the rate of 10 mV/min. Details of the electrochemical behavior during film formation were obtained from potentiostatic experiments. In these experiments, a specimen was prepared in the standard manner. Its surface was deaerated. It was exposed for a few minutes to the deaerated 0.5 M H2SO 4 at the free corrosion potential (and passive film was formed on the specimen surface as shown by the potentiodynamic experiments). Subsequently, the potential was "instantaneously" applied at the value of interest, and the current recorded as a function of time. The secondary passive film after electrochemical film formation was characterized by electrochemical reduction at -350 mVscE in fresh deaerated 0.5 M H 2SO,. The reduction conditions were chosen to minimize contributions from other reduction reactions, including reduction of species in solution and also hydrogen evolution. b The potential value of -350 mVs,, was chosen because of the low current densities measured in the polarization curves in 0.5 M H2SO4 . For the reduction experiments reported in this paper a secondary passive film was first formed by holding a new specimen prepared in the standard manner at constant potential in deaerated 0.5 M H2SO4 while the current was recorded. The solution was drained from the specimen cell. The specimen, in the specimen holder, was removed from the specimen cell, washed in distilled water, dried in a jet of clean air, and reintroduced into the specimen cell where the surface of the specimen was deaerated b In a series of preliminary experiments, reduction was carried out immediately after film formation in the same solution. The experimental results were similar to those reported herein, although the background current densities were somewhat higher than those measured in the experiments reported in this paper.

J. Electrochem. Soc., Vol. 144, No. 11, November 1997 © The Electrochemical Society, Inc. by the N2 /H, gas mixture for about 1 h. During this hour, a new 0.5 M H2SO, solution was deaerated in the deaeration cell. After the new deaerated solution had been introduced into the specimen cell by means of the deaeration gas pressure, the potential of -350 mVscE was applied and the current recorded as a function of time. For each reduction experiment, there was a new specimen and a new deaerated solution. Some experiments were carried out to study the stability of the secondary passive film. These had the same procedure except that, after film formation, the specimen was removed from the specimen holder and subjected to an ultrasonic cleaning bath in an acetone/alcohol mixture. Subsequently, the specimen was washed in distilled water, reintroduced into the specimen cell, its surface was deaerated, and the surface film was characterized by reduction in a new deaerated 0.5 M HSO 2 4 solution. Film characterization after film formation was also carried out using XPS. XPS measurements were carried out to characterize films formed for (i) 200 s and (ii) 30 min at 1200 mVscE in deaerated 0.5 M HSO 4 (followed in each case by washing in distilled water and drying); these were compared with (iii) a passive film on a specimen after the standard specimen preparation (wet abrasion with 1200 grit, dried in air, and at least 8 h exposure to laboratory air), and (iv) the residual film on a specimen with secondary passive film formed for 30 min at 1200 mVsc in deaerated 0.5 M HSO,, reduced in a new deaerated 0.5 M HSO 2 4 solution at -350 mVscE, washed in distilled water, and dried in air. Because major emphasis was on gaining an understanding of the secondary passive film as it might be relevant in a technological application of the stainless steel, all specimens were exposed to the atmosphere during transfer to the XPS apparatus. This transfer arrangement might cause the loss of fragile surface species like OH and adsorbed water. Furthermore, the fine structure of a modified passive film may revert to that more characteristic of a passive film. Nevertheless, the XPS measurements should provide valid measurements of chemical species and film thickness. It is also worthwhile mentioning that the passive film has been shown to be quite robust and to change surprisingly little with air exposures. 1 XPS analysis was performed using a VG XR3E2 spectrometer employing a Mg K0 (1253.6 eV) monochromatic x-ray source operated at 15 kV with a power of 300 W. Typical operating pressures were -30 nPa (2 10 0 Torr). The spectrometer was calibrated using the following photopeaks: Au 4 f7/2 at 83.8 eV, Cu 3p at 75.2 eV, and Cu 2 p 2/3 at 932.7 eV All binding energies were referenced to the carbon C-H photopeak at 285.0 eV. The area analyzed was 10x 4 mm. Survey (wide) scan spectra were measured with a pass energy of 60 eV for all samples to determine what elements were present in the top 5 nm of the surface. Mutiplex (narrow) scan spectra were obtained with a pass energy of 30 eV for all major photopeaks. Angle dependent analysis at 30 and 90 was done by varying the take-off angle between the surface and the direction of electron detection. Detailed scans were carried out for Fe, Cr, 0, C, S, and Si with take-off angles of 300 and 900. Particular attention was devoted to the S measurements, because there was an expectation of sulfate incorporation in the secondary passive film from the literature on colored stainless steels 3-4 3

Potentiodynamic Polarization The polarization curves are presented in Fig. 2. Low currents were measured in the passive range (-200 to -950 mVsCE) in all cases. In 5 M H2S0 4 there was active corrosion followed by passivation. Specimens were selfpassivating in the borate buffer solution, in 0.5 M H2 SO, and in the phosphate solution. This is of significance for the interpretation of the subsequent potentiostatic experiments; these experiments show that there was a passive film on the specimen surface after the standard specimen preparation and exposure for a few minutes to

16.0- -

3699

I

-a- 6MH2S04 --

5M 142S04 borate buffer -- - 0.5MH2S04 -- phosphate

12.0

a

r

8.0

0

b

.0

4.4.0 W

/ .0~

[i

Si--

/10

0.0

;ry' _

7R- .7

No7, -, ! 4

lr

b

I

I

I

I

'-

-0.6

-0.2

0.2

0.6

1.0

1.4

I

1.8

potential, VsCE

Fig. 2. Potentiodynamic polarization curves to characterize film formation conditions. the borate buffer-solution, to the phosphate solution, and to 0.5 M HSO4 . For 0.5 M HSO 2 4, within the secondary passivity range, the current increased to a plateau of about 4.6 mA/cm 2 at 1400 mV. In the more concentrated solutions, i.e., in5 and 6 M HSO the current densities were in general much 2 4 higher in the secondary passivity region. For the phosphate solution, the current density increased to a plateau of about 2 mA/cm 2 at 1250 mV The curve measured in the borate buffer was different from the other curves in that the current density increased nearly linearly with increasing potential; the "noise" on this curve of potentials above about 1175 mV might be indicative of oxygen evolution which is expected at these potentials in the higher pH values of the borate buffer. The potential E,, at which there was a significant current density increase indicative of the onset of secondary passivity, was similar for all the solutions used, Table I, with an average value of 940 mVscE, and soE, was approximately independent of pH over the range 0 to 8.4.

Potentiostatic Experiments Figure 3 presents the current as a function of time measured at 500, 1000, and 1300 mVscE in the potentiostatic experiments. In the passive region (at 500 mVsc), the current decreased steadily with time, whereas in the secondary-passivity region, the current became independent of time after an initial transient; this is consistent with the literature.'," Figure 4 presents the initial shapes of the transients in the secondary-passivity potential range forpotentiostatic experiments at 1100, 1200, and 1300 mVscE. One curve at 1200 mVcE was recorded for 200 s, whereas all the other experiments were of 30 min duration, although Fig. 4 illustrates only the initial parts of the transient where there was significant change. The repeat experiments show excellent reproducibility. The transient at 1100 mVsCE displayed a single minimum similar to that at 1000 mVsCE (see Fig. 3). The transients at 1200 and 1300 mVscE showed a minimum followed by a maximum, which was more pronounced for the curve at 1300 mVscF. The steady-state value of the current, i, was very similar to the current density, i, measured in the potentiodynamic experiments. This is recorded Table . Potential, E, Solution

marking the onset of secondary passivity in different solutions. Esp, mVsCE

6 M HSO4 5 M HSO

970

0.5 M H2 SO 4 Phosphate Borate

910 940 960

970

J. Electrochem. Soc., Vol. 144, No. 11, November 1997 © The Electrochemical Society, Inc.

3700

Table II.Steady-state current density, i,, measured in potentiostatic experiments compared with the current density, id,measured in the potetiodynamic experiments.

10 -

j

B---

-`^I'----1

Potential E (mV)

'..

Current density (mA/cm') id i

C)4

[1 1

* 0.1

©3 r::

ooW

00

I 00

a

001

"0 t)

+ 12 0,001 :-

10

i

* 500 mV 0.0001

time, s Fig. 3. Potentiostatic curves to characterize film formation for type 304 stainless steel in deaerated 0.5 M H2SO 4 at 500, 1000, and 1300 mVscE.

in Table II. The speed of the transient response indicated that there is a fairly rapid transformation of the passive film into the secondary passive film (which controls the electrochemical response) with most of the film formation completed within 400 s.

Reduction Experiments Typical reduction curves are presented in Fig. 5 for secondary passive films formed for different periods at 1200 mVscE. For secondary passive films, formed under the normal conditions, the reduction curves could be interpreted as being composed of three reduction waves. Table III presents the charge associated with each reduction wave, Q1, Q2, and Q, as defined in Fig. 6. Table III also records values of the film thickness, d,, which has been estimated from" d = Q V,,ox/3F

1400

4.6

5.6

1300 1200

4.0 3.2, 3.3

1100 1000

1.8, 1.8 0.18, 0.16

4.0, 4.1 3.3, 3.4, 3.3, 3.3, 3.1, 3.3 2.0, 1.9 0.32

where 100% current efficiency is assumed during reduction, it is assumed that all the reduction charge Q is associated with M3+; VOX, the oxide molecular volume is taken as Vox = 10 cm3/mol which is the average value for iron oxides.' ° The charge associated with each reduction wave increased with increasing time of film formation, although there was considerable data scatter, for example, as shown by the differences in the two reduction curves for films formed for 30 min at 1200 mV. This data showed that these secondary passive films could withstand drying. However, there was very little charge (0.43 mC/cm2 ) associated with the reduction curve for the film subjected to the acetone/alcohol ultrasonic bath indicating that most of the secondary passive film had not survived this treatment. Low values of residual current, i3 (see Table III) indicated success in choosing reduction conditions that minimized reduction reactions other than film reduction. Measurement of the residual current, i3 due to other reduction reactions, allowed evaluation of the reduction charge associated only with film reduction. A most interesting observation is that the film thickness, as measured by reduction, continued to increase for film formation times greater than 400 s, Fig. 7. In this time period, there was little change in the current transients measured potentiostatically, see Fig. 4, indicating little change in the electrochemical response. It is worth stressing that the film thickness measurements using cathodic reduction are to be considered to underestimate the true film thickness for two reasons. First, it is widely believed that Cr,O, is not reduced under these conditions.' 44'45 This can place considerable limitations on this cathodic reduction technique as a means of measuring the film thickness. Second, the reduction experiments were carried out after the specimen had been washed in distilled water and dried. This specimen handling procedure could remove

[1]

o.-

I --

-I ;

-10.0

-I

_-

-

45 4.0

' -20.0

3.5

-

30 min

t -30.0

'-x-

3.0 2.5 = 2.i 05 2.0

-40.030 I

0.0

Is