olsson 2003, ea, review st.st

13
See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/223831935 Passive films on stainless steels—chemistry, structure and growth ARTICLE in ELECTROCHIMICA ACTA · APRIL 2003 Impact Factor: 4.5 · DOI: 10.1016/S0013-4686(02)00841-1 CITATIONS 242 DOWNLOADS 227 VIEWS 239 2 AUTHORS, INCLUDING: Dieter Landolt École Polytechnique Fédérale de Lausanne 227 PUBLICATIONS 5,325 CITATIONS SEE PROFILE Available from: Dieter Landolt Retrieved on: 08 August 2015

Upload: cherk

Post on 28-Aug-2015

226 views

Category:

Documents


5 download

DESCRIPTION

articulo

TRANSCRIPT

  • Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/223831935

    Passivefilmsonstainlesssteelschemistry,structureandgrowthARTICLEinELECTROCHIMICAACTAAPRIL2003ImpactFactor:4.5DOI:10.1016/S0013-4686(02)00841-1

    CITATIONS242

    DOWNLOADS227

    VIEWS239

    2AUTHORS,INCLUDING:

    DieterLandoltcolePolytechniqueFdraledeLausanne227PUBLICATIONS5,325CITATIONS

    SEEPROFILE

    Availablefrom:DieterLandoltRetrievedon:08August2015

  • Passive films on stainless steels*/chemistry, structure and growth

    C.-O.A. Olsson *, D. Landolt

    Departement des Materiaux, Laboratoire de Metallurgie Chimique, Ecole Polytechnique Federale de Lausanne, CH-1015 Lausanne EPFL, Switzerland

    Abstract

    The outstanding corrosion resistance of stainless steels results from the presence of a thin oxide*/passive*/film on the metalsurface, typically 1/3 nm thick. The characterisation of the composition and structure of such thin films and the study of theirinteraction with corrosive environments requires a combination of sophisticated experimental techniques. This paper reviews

    progress in the characterisation and understanding of passive films on stainless steels achieved over the past two decades. During

    this period, ex situ surface analysis methods have made substantial progress and new in situ methods for the study of passive films

    with atomic resolution have been introduced, giving real time information on film chemistry and growth. It has been found that

    whereas passive film growth occurs in seconds or minutes, long range film ordering is a considerably slower process that takes

    several hours. In situ investigations indicate that at short times, charge transfer at the metal/film or the film/solution interface limits

    the rate of film growth on stainless steels. In situ estimates of film composition confirm previous data obtained with ex situ

    techniques.

    # 2003 Elsevier Science Ltd. All rights reserved.

    Keywords: Passive films; Stainless steel; Passivity; X-ray photoelectron spectroscopy; Scanning tunnelling microscope; X-ray absorption near edge

    structure; Electrochemical quartz crystal microbalance

    1. Introduction

    Stainless steels were invented almost a century ago by

    Monnartz [1]. Today, their usage shows an average

    increase rate of about 5% per annum. They are divided

    into four main groups based on their microstructure:

    ferritic, austenitic, martensitic and austeno/ferritic(duplex). Among the ferritic steels, one finds a series

    of Fe/Cr alloys, e.g. Fe/17Cr (AISI 430), but alsosuperferritics containing high fractions of molybdenum.

    Austenitic stainless steels are usually alloyed with Ni, for

    example 18Cr/9Ni (AISI 304); superaustentic stainlesssteels containing high fractions of Mo and N can in

    certain environments surpass the corrosion resistance of

    Ni-base alloys. Stainless steels are subject to continuing

    development; some recent inventions are listed in Table

    1. Thermodynamic modelling has made it possible to

    develop new grades with high contents of molybdenum

    and nitrogen (654 SMO), that are extremely resistant to

    pitting attacks induced by halides[2]. Another example

    of the utility of thermodynamics is the super duplex

    stainless steel SAFUREX, that has been developed to

    withstand highly oxidative environments [3]. With

    thermodynamic models, it is possible to find composi-

    tions with high contents of Cr, Ni and N that give the

    desired phase balance, normally 50/50 austenite/ferrite.

    Extreme strength levels can be obtained by introducing

    high fractions of nitrogen and manganese. Further

    strength increase can, for austenitic materials, be

    obtained by cold deformation. One example is the

    high Mn, high N steel PANACEA [4]. Other develop-

    ments include surfaces with structures tailored to avoid

    bio fouling (UGICLEAN). Copper alloying in combi-

    nation with a heat treatment is used to produce a special

    line of anti-microbial stainless steels (NSS AM: 1/4).The literature shows that the corrosion resistance of

    stainless steels can be drastically improved by suitable

    alloying. This is possible because the composition and

    properties of the passive films depend on the alloy

    composition. Numerous investigations have contributed

    to our present understanding of the processes governing

    film growth and break down as a function of alloy

    composition and environment. Today, it is known that

    the passive film adapts to changes in potential or anion* Corresponding author. Fax: /41-21-693-3946.E-mail address: [email protected] (C.-O.A. Olsson).

    Electrochimica Acta 48 (2003) 1093/1104

    www.elsevier.com/locate/electacta

    0013-4686/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.doi:10.1016/S0013-4686(02)00841-1

  • concentration in the electrolyte. The dynamic properties

    of the passive film provide the key to the high resistance

    of stainless steels to corrosive attacks. The state of

    knowledge of passive films on stainless steels as of 20

    years ago was thoroughly reviewed by Fischmeister andRoll [5]. The aim of this paper is to present an overview

    of new methods and results reported in the past two

    decades. One could say that whereas the keyword of the

    70s and early 80s was ex situ surface analysis, recent

    research has emphasised the use of different in situ

    methods.

    2. Experimental methods for the study of passive films

    Passive films are formed during an exposure of the

    bare metal surface to an oxidising environment. Once a

    film is formed, the reaction rate between the metal andthe environment will be several orders of magnitude

    lower. The original theory of film formation goes back

    to Michael Faraday, who in the 19th century studied

    iron surfaces and found them altered. A review of the

    early days of passive film research has been written by

    Uhlig [6]. A general introduction to the theory of

    passivity has been published by Sato [7], whereas the

    electronic properties of passive films on differentmaterials have recently been reviewed by Schultze and

    Lohrengel [8].

    2.1. Ex situ methods

    Electron spectroscopy has been used since the early

    1970s for the study of the composition and thickness of

    the passive films, including the determination of oxida-

    tion states and of the depth distribution of the different

    constituting elements in the film. X-ray photoelectron

    spectroscopy (XPS, also known as ESCA, electron

    spectroscopy for chemical analysis) and Auger electron

    spectroscopy (AES) have been most widely used, butother methods such as secondary ion mass spectroscopy

    (SIMS) and ion scattering spectroscopy (ISS) have also

    provided important information.

    Based on surface analysis, a three-layer model has

    been suggested for passive films formed on austenitic

    stainless steels in acidic solutions: the outer part of the

    film consists of an hydroxide film on top of an oxide

    layer. The oxy-hydroxide film is formed on top of a

    nickel enriched layer in the metal, the origin of which is

    the selective oxidation of Fe and Cr during anodic

    polarisation. This picture is obtained essentially from

    XPS measurements on passive films, cf. Fig. 1 [9]. The

    film is enriched in chromium and the metal closest to the

    metal/film interface shows a strong enrichment in nickel.

    For an alkaline solution, the solubility is lower for Fe

    and higher for Cr; this affects the level of chromium

    enrichment in the film. With angular resolved XPS

    (ARXPS), it is also possible to obtain information on

    the depth distribution of different species and oxidation

    states, cf. Fig. 2, which illustrates the angular depen-

    dency of different oxidation states of molybdenum in

    the passive film after immersion in a ferric chloride

    solution [10]. Mo(VI) is enriched in the surface region,

    the metal peaks show a buried layer appearance, and the

    Mo(IV) oxide and Mo(IV) oxy-hydroxide show an

    angular distribution indicating a more homogeneous

    distribution through the passive film. The attribution of

    Table 1

    Examples of recent commercial developments in the field of stainless steel with high impact on the surface behaviour of the material

    Product Significance Producer

    654 SMO First 7% Mo austenitic stainless steel. Very high resistance to pitting corrosion at elevated temperatures Avesta Polarit

    SAFUREX First 29Cr duplex stainless steel. Very high resistance to oxidising acids and localised corrosion, in particular stress

    corrosion cracking.

    Sandvik Steel

    UGICLEAN A special surface structure modified to resist bacterial adhesion Ugine

    NSS AM1-4 Cu alloyed stainless steels with special heat treatment giving high resistance to bacterial adhesion Nisshin steel

    Zeron 100 Duplex stainless steel alloyed with W Weir Materi-

    als

    DP3-W W alloyed stainless steel Sumitomo

    Fig. 1. XPS results from a passive film formed in the high passive

    region in a 0.1 M HCl/0.4 M NaCl solution. The oxide film is foundto be strongly enriched in chromium, whereas the metal layers closest

    to the metal/film interface, i.e. the apparent metal concentration, are

    strongly enriched in nickel. The passive film itself shows close to no

    incorporation of nickel. After Olefjord and Elfstrom [9].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041094

  • oxidation states to binding energy has been detailed by

    Brox and Olefjord [11].

    Superior depth resolution can be obtained using ISS,

    which also gives quantitative results. By choosing the

    proper primary ions, a resolution sufficient to distin-

    guish between the Fe and Cr contributions to the signal

    can be obtained. With ISS, Calinski and Strehblow

    showed strong concentration gradients within the pas-

    sive film on an Fe/Cr alloy [12], cf. Fig. 3. The outerregion of the passive film consists of a layer enriched in

    iron, a feature that has also been observed with XPS

    sputter depth profiles [13]. In the centre of the passive

    film, a strong enrichment of chromium is found,

    whereas for the region closest to the metal/film interface,

    the composition corresponds to that of the bulk metal.

    This implies that the oxidation at the metal/film inter-face closely follows the composition of the bulk metal,

    and that the enrichment in chromium is caused by

    dissolution of iron. The investigation of passive layers

    on different metals with XPS and ISS has been reviewed

    by Strehblow [14].

    In 1984, Fischmeister pointed out that XPS is

    generally better suited for analysis of passive films on

    stainless steels than AES [5]. AES has a higher lateralresolution and a higher surface sensitivity for some

    elements, e.g. molybdenum. The difficulty in analysing

    Cr oxides with AES due to the peak overlap between the

    OKLL and CrLMM peaks can be resolved through target

    factor analysis (TFA) [15]. It is also possible to facilitate

    the analysis of Mo by using krypton as a sputtering gas

    instead of argon [15,16]. In spite of these improvements,

    Fischmeisters preference is still valid, since the state ofthe art resolution of XPS for spectral information has

    been improved to about 5 mm. The reproducibility ofAES and XPS measurements between different labora-

    tories has been investigated in a round-robin where two

    types of ferritic stainless steels were exposed to acidic

    sulphate solutions. In terms of raw data (binding energy

    shifts, relative intensities of oxide peaks etc.) the overall

    agreement between different laboratories was judgedsatisfactory [17].

    A very critical issue with ex situ surface analysis is the

    sample transfer from electrolyte to UHV. No matter

    how much precaution is taken, the surface will alter

    somewhat during the transfer. Different ingenious

    methods for sample transfer without contact with

    atmosphere have been developed; one example of such

    a design is given by Haupt et al. [18]. The level of utilityof an in vacuo transfer system is determined by the

    amount of sample oxidation that can be avoided. The

    amount of extra oxidation during transfer in air can be

    estimated by forming the passive film in a normal

    electrolyte, and making the transfer in an atmosphere

    enriched in 18O. The amount of 18O in the film will then

    be a measure of the amount of oxidation that has

    occured in atmosphere. The inverse experiment is alsopossible, i.e. a passive film formed in an electrolyte

    enriched in 18O and subsequent transfer in normal

    laboratory atmosphere. The different passive films are

    then compared using SIMS that can distinguish between

    the 16O and 18O isotopes. Graham et al. polarised FeCr

    alloys in water enriched in H218O [19/21], and found that

    for pure iron, there is a small effect of sample transfer on

    the passive film for polarisations in the passive region,whereas a more pronounced thickness change in the film

    was found for Fe/26Cr samples; it was suggested thatthis difference could be caused by the higher ability of

    Fe/Cr films to adapt to different environments. Similar

    Fig. 2. Concentration gradients in a passive film for molybdenum,

    recorded for a 6Mo superaustenitic stainless steel after immersion in

    ferric chloride using XPS. For angles close to grazing, there is a strong

    contribution of Mo(VI), whereas metallic Mo dominates for angles

    close to perpendicular. The IV-valued oxide and oxy-hydroxide states

    show less angular dependence. After Olsson and Hornstrom [10].

    Fig. 3. Enrichment and concentration variations for Cr in a passive

    film obtained by ISS. Even for low alloying amounts of Cr (5%), there

    is an enrichment of Cr in the film, with the exception of the outer part.

    After Calinski and Strehblow [12].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/1104 1095

  • experiments have been performed by Courty et al. [22].

    A direct estimate of the film change during transfer was

    obtained by comparing films formed in H218O and

    transfer in 16O2 with films formed in H216O and transfer

    in 18O2. For low passive and cathodic potentials, both

    laboratories found a significant effect of sample transfer.

    Graham et al. reported a transfer influence also for the

    high passive region. However, when Courty et al.

    transferred samples in 18O2 atmosphere, no uptake of18O in the film was seen, even though enrichment levels

    close to those of Grahams were observed for anodisa-

    tion in H218O and transfer in 16O2. Thus, the direct

    measurement showed an absence of oxygen uptake

    during transfer for samples polarised to the high passive

    region [22].

    The results indicate that caution is needed during the

    sample transfer. However, by ensuring that the transfer

    is made consistently for all samples, it is possible to draw

    valid conclusions from comparisons between series of

    measurements. Evidently, the only way to eliminateartefacts associated with sample transfer is to use in situ

    techniques.

    2.2. In situ techniques

    A technique new to the study of passive films is X-ray

    absorption near edge structure (XANES), which ac-quires in situ chemical information. It has been applied

    to the investigation of passive films on pure iron [23/27], iron/chromium alloys [28] and to the study ofsputter deposited iron/chromium oxides [29]. Thesputter deposited oxide films were used as model

    systems for passive films. XANES puts strong demands

    on the analysed material. The sample is a thin film of

    about 10 nm thickness deposited on top of a goldcontact layer. As substrate, a polymer carrier film that

    ensures X-ray transparency is used. Sample character-

    isation has been described by Stenberg et al. [30]. The

    study of reaction kinetics is somewhat limited, since

    acquisition times of the order of one hour are used [28].

    A short introduction to XANES applied to passivity

    related problems has been written by Schmuki and

    Virtanen [31].

    A XANES spectrum for Cr and Fe taken from an (Fe,

    Cr)2O3 sample is shown in Fig. 4. The height of the

    plateau at the high-energy side of the adsorption edge is

    proportional to the respective amount of the element in

    the sample, and the peak position reflects the oxidation

    state. Quantification of XANES spectra is less straight-

    forward than for AES and XPS. Davenport et al.

    introduced a formalism to separate the oxide and metal

    signals from an oxidised sample [26]. This method was

    further developed by Oblonsky et al. for the quantifica-

    tion of film composition on Fe/Cr alloys [28], and laterNi/Cr [32]. The measurements confirm XPS and AESresults, i.e. for a certain polarisation and bulk composi-

    tion, the amount of Cr in the passive film surpasses 50%

    of cations [33]. The XANES data also support the

    enrichment of Cr in the passive film by selective

    dissolution of iron. It has also been possible to

    demonstrate the presence of Cr(VI) in the film [34].

    In situ analysis of binary alloys can also be performed

    by more conventional techniques. By solution analysis

    with ICP-AES/MS (inductively coupled plasma atomic

    emission spectrometry/mass spectrometry), it is possible

    to deduce composition changes in the passive film

    resulting from potential changes. This approach was

    successfully employed by Castle and Qiu [35] and Hamm

    et al. [36]. For an Fe/Cr alloy in the passive region, bothpapers concluded that the dissolved species consisted

    almost solely of iron. Similarly, rotating ring-disc

    electrode experiments with Fe/Cr alloys showed thatclose to no Cr was emitted from the sample during a

    potential change in the passive region [37], in good

    agreement with the ICP results. Annergren et al. used

    the rotating ring-disc electrode in combination with

    electrochemical impedance to study the dissolution and

    passivation behaviour of Fe/Cr alloys in sulphuric acid

    Fig. 4. XANES spectra (Fe and Cr K edges) of a mixed (Fe, Cr)2O3 sample containing 50% Cr2O3 during anodic potential steps in 0.1 M H2SO4. The

    potential was increased in 100 mV steps/15 min, which is illustrated with selected raw spectra. From Schmuki et al. [29].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041096

  • solution with additions of chloride [38/40]. Theycalculated that ferrous ions were leaving the electrode

    with a current efficiency even higher than unity for

    intermediate frequencies. The behaviour was explainedby postulating a concurrent chemical dissolution path of

    Fe(II) [39].

    2.3. Environmental influence on the film composition

    A passive film is in constant exchange of species with

    the electrolyte and consequently alters in thickness andcomposition with the environment. Among the factors

    that influence the passive film, one finds the potential,

    the presence of halides in the electrolyte, the pH, and the

    temperature.

    2.3.1. Potential

    The thickness of passive films on stainless steels grows

    linearly with the applied potential. This is illustrated in

    Fig. 5 for a Fe15Cr alloy in an acidic sulphate [41] as

    well as Fe10Cr and Fe20Cr alloys in a sodium hydroxide

    solution [42]. The basic solution gives considerably

    thicker films, since dissolution is less pronounced. The

    observed increase in film thickness with potential in an

    acidic solution is due mainly to the oxide part; thehydroxide part of the film is approximately independent

    of potential [43,44]. Development of the passive film

    with time has been investigated by, among others,

    Maurice et al. [45]. Between 20 min and 20 h of

    polarisation, they found only a small change in the

    total film thickness; the oxide part of the film was found

    to grow at the expense of the hydroxide part.

    The composition and chemistry of the film also vary

    with potential. For Fe/Cr alloys, enrichment of chro-mium occurs in the film in the low passive region, but in

    the high passive region the chromium content decreases,since the stability of iron surpasses that of chromium.

    Chromium, in turn, may oxidise to its soluble hexavalent

    state. At high passive potentials, Fe has a stabilising

    action on the trivalent Cr, making Fe/Cr alloys stableover a wider potential range than pure Cr. Higher

    oxidation numbers are also more dominant towards the

    higher end of the passive region, as confirmed by

    XANES [34].

    2.3.2. Chlorides and sulphates

    The presence of certain anions adsorbed on the

    surface or incorporated in the passive film can be

    detrimental to film stability and lead to pit initiation.

    For chloride, there is a vast literature attempting to

    explain its effects on pitting. Three different models are

    frequently quoted: adsorption leading to local film

    dissolution, penetration of anions in the film leadingto weakening of the oxide bonds, and film break down

    at defects such as cracks and dislocations [46]. Con-

    clusive analysis of the chloride content of passive films

    has proven difficult, since it is rather low; frequently

    quoted values range from 1 to 5%. Due to the very thin

    passive film and the possible presence of structural

    defects, it is not easy to distinguish adsorbed from

    incorporated anionic species. Using AES and XPS, ithas been shown that sulphates are present at or close to

    the film surface [47/50]. Chlorides are enriched at thepassive film surface, but the depth distribution is more

    homogeneous than for sulphates [51], especially at

    elevated temperatures [43]. Mischler et al. have shown

    that alloying with molybdenum reduces the surface

    enrichment of chlorides and sulphates [52]. Mitrovic-

    Scepanovic et al. [53] found a higher tendency forchlorides than sulphates to penetrate the passive film

    on Fe/Cr alloys. Chloride penetration into an intactpassive film can be investigated by forming the film in a

    non-chloride solution, followed by exposure to chlor-

    ides. Such a series of experiments has been performed by

    Hubschmid et al. for Fe/25Cr/X alloys. No incorpora-tion was found when chlorides were injected after an

    initial hour of film formation in a sulphuric acidsolution [54]. For films formed in chloride electrolytes,

    on the other hand, chlorides were found distributed

    through the film.

    2.3.3. pH

    The main effect of an increased pH is a lower

    dissolution rate. This leads to a thicker passive film,

    and a larger fraction of iron in the film, as iron oxidesare more stable in basic solutions. The pH effect is

    illustrated by the data of Schmutz and Landolt, who

    compared the response for two FeCr and FeCrMo

    Fig. 5. Oxide layer thickness on a stainless steel as a function of

    potential for a Fe15Cr alloy in 0.5 M H2SO4, and for Fe10Cr and

    Fe20Cr alloys in 1 M NaOH, estimated using XPS. The film growth

    region is considerably wider in the basic medium, which also gives

    thicker films. After Haupt and Strehblow [41] and Hoppe et al. [42].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/1104 1097

  • alloys to a potential change in basic and acidic solutions

    using the electrochemical quartz crystal microbalance

    (EQCM) [55]. Increasing the potential in the passive

    region lead to a mass loss in the acidic solution (0.1 MH2SO4/0.4 M Na2SO4) and to a mass increase in thebasic solution (0.1 M NaOH). A mass increase corre-

    sponds to a net film growth and a low dissolution rate.

    The effect of pH on the passive layer of Fe/Cr alloyshave also been studied by Strehblow et al. [41,42]. The

    films were found to be thicker in basic solutions. In

    addition, there was a marked increase of the amount of

    Fe(III) oxide.

    2.3.4. Temperature

    The effect of temperature on the passive film was

    studied by Jin and Atrens using XPS [56]. They found

    no differences in composition and thickness between

    passive films formed at room temperature and at 90 8Cin a 0.5 M NaCl solution. Mischler et al. quoted slightly

    thicker values for films formed on Fe/Cr/Mo alloys at65 8C when compared to room temperature [52]. Thetemperature effect on film thickness for a 6Mo stainless

    steel has been quantified by Wegrelius and Olefjord

    using XPS [43]. They compared film formation at 22 and

    65 8C in an acidic chloride solution. In spite of clearelectrochemical indications of pitting attacks during the

    exposure period at the higher temperature, the film was

    found to be only 2 A thicker for the higher temperature.

    2.4. Influence of microstructure and alloying elements on

    passive film composition

    2.4.1. Microstructure

    With classic metallurgical techniques, the composition

    range of the bulk material is limited by the precipitation

    of intermetallic phases. In addition, inclusions and

    segregation cause defects in the passive film. Fordetailed studies of passive films, the amount of inter-

    metallic phases and other imperfections can be mini-

    mised by utilising thin film materials produced with

    PVD (physical vapour deposition). Since PVD is a cold

    technique*/a typical deposition temperature is400 8C*/it is possible to obtain nanocrystalline oramorphous material that is virtually free from inter-

    metallic phases over a wide composition range.Inturi and Smialowska compared a nanocrystalline

    thin film alloy of the same composition with conven-

    tional 304 material [57]. The breakdown potential of the

    nanocrystalline material in a near neutral NaCl solution

    was found to be shifted anodically by about 850 mV, a

    behaviour that was attributed to the smaller grain size

    and the high purity of the deposited thin film material.

    Kraack et al. found a shift in the pitting potential tomore noble values when molybdenum was added to thin

    film Fe/Cr alloys in agreement with the well-knownbehaviour of bulk material [58]. A wide range of systems

    produced using PVD has been investigated by Hashi-

    moto et al. [59/64] who underlined the extremely highcorrosion resistance of amorphous alloys.

    A showcase for the importance of microsctructure isthe austeno/ferritic (duplex) stainless steels. Their two-phase structure makes them particularly sensitive to the

    formation of intermetallic phases. AES has been used to

    investigate the homogeneity of the passive film on the

    austeno/ferritic stainless steel 2205 (EN 1.4462) [65]. Atypical phase region for a commercial duplex stainless

    steel covers 10/20 mm. It has been demonstrated thatthe passive film composition shows a comparativelyhomogeneous composition, virtually independent of the

    underlying phase structure. This homogeneity continues

    some Angstroms down in the metal phase where a nickel

    and nitrogen enrichment is found for the austenitic and

    the ferritic phases [65]. This enrichment would be

    sufficient to produce a fully austenitic structure in a

    bulk material. Significant differences between phase

    regions were found by Hubschmid et al. who studied aFe/25Cr/11Nb model alloy with a considerably coarserdendritic microstructure than for commercial 2205 [66].

    It appears the lateral homogeneity of the passive film is

    linked to the lateral extension of different phases in the

    microstructure.

    2.4.2. Fe

    In acid solution, anodic polarisation of a Fe/Cr alloyin the passive potential region leads to selective dissolu-

    tion of iron, leaving chromium enriched in the passive

    film. This behaviour is governed by the difference in

    diffusion rates of iron and chromium in the passive film

    [13,67]. When performing sputter depth profiles with

    AES or XPS, there is normally an enrichment of iron in

    the outer region of the passive film. This is possibly a

    transfer artefact related to the higher diffusion rate ofiron through the film. At about 0.58 VSHE, iron changes

    valence from 2 to 3 [41] for an acidic solution; for a basic

    solution, the corresponding potential is about/0.3VSHE [18].

    2.4.3. Cr

    For moderate anodic polarisations in acidic solutions,the passive film consists essentially of chromium in its

    trivalent state. As the potential increases above the

    stability limit of chromium III, (about 0.6/0.8 VSHE),the passive film will start to change composition and the

    fraction of trivalent iron in the film will increase. For

    acidic solutions, the cation fraction of Cr in the passive

    film normally amounts to 50/70%. For basic solutions,the solubility of Cr increases, resulting in a higherfraction of iron. The surface chemistry of pure chro-

    mium has been studied by Bjornkvist and Olefjord

    [68,69]

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041098

  • 2.4.4. Ni

    Nickel is less readily oxidised than iron and chro-

    mium. Consequently, there is an enrichment of Ni in its

    metallic state in the metal closest to the oxide/metalinterface. This enrichment could assist in the formation

    of a nickel nitride [70]. Nickel could also bring down the

    overall dissolution rates of Fe and Cr. Recent scanning

    tunnelling microscope (STM) results indicate that the

    passive film formed on austenitic stainless steels shows a

    higher degree of long-range order within a shorter time

    span than the corresponding ferritic alloy [45].

    2.4.5. Mn

    Recently, addition of Mn to stainless steel has been

    used to increase the solubility of nitrogen and molybde-

    num, both of which have a strong beneficial influence onthe pitting resistance. Examples include the low nickel

    high strength austenitic material PANACEA [4] and the

    commercial superaustenitic stainless steel 654 SMO [2].

    2.4.6. Mo

    Molybdenum is an alloy element with a strong

    beneficial influence on the pitting resistance of a

    stainless steel. Pure molybdenum does not form a

    three-dimensional passive film [71]. It appears that the

    passive region of pure molybdenum is associated with a

    diffusion controlled oxidation of hydrogen released

    from the metal rather than with the formation of apassive oxide film [72]. On the other hand, when

    included as an alloying element in a stainless steel,

    molybdenum is incorporated into the passive film,

    showing a complex oxide chemistry with different states

    of oxidation. An example of the concentration gradients

    of molybdenum in a passive film on a stainless steel is

    given in Fig. 2. Hexavalent Mo is found to be enriched

    at the surface, whereas tetravalent states show a morehomogeneous distribution through the film [10]. This

    has been shown also for the two phases of a duplex

    stainless steel [73]. There are two possible hexavalent

    states: MoO3, which is soluble in acidic eletrolytes and

    MoO42 which shows a higher stability. Distinguishing

    between these two states is difficult even with a high

    resolution XPS spectrometer. Attempts have been made

    by Lu and Clayton, [70,74/78], who discussed theinteraction between nitrogen and molybdenum. The

    passivity of stainless steels was interpreted with a

    bipolar model. Another subject of discussion is the

    presence of a pentavalent state, cf. Kim et al. [79].

    Different molybdenum oxides have been thoroughly

    characterised by Brox et al. [11,71], who found the XPS

    peak of the MoO2 compound to be considerably less

    shifted than expected for an ionic bond. This effect wasnot seen for a tetravalent molybdenum oxy-hydroxide

    and was explained by core-level screening. The literature

    on the effects of Mo on the corrosion/pitting resistance

    has been reviewed by Jargelius-Pettersson and Pound

    [80].

    2.4.7. W

    Tungsten is fairly recent as a major alloying element

    in commercial stainless steels. It has been attributed

    properties similar to those of molybdenum [81]. The

    surface properties of ferritic alloys containing Mo and/

    or W have been thoroughly investigated by Landolt et

    al. [52,54,82/84]. An important difference betweentungsten and molybdenum oxides lies in the different

    stability of their oxides in acid solution. While hexava-lent molybdenum oxide dissolves at potentials well

    below oxygen evolution, the stability of hexavalent

    tungsten oxide extends to anodic potentials of several

    tens of volts [85].

    2.4.8. N

    Nitrogen is the element attributed the strongest

    beneficial influence on localised corrosion in the pitting

    resistance equivalent formula (PREN). Different typesof synergism with molybdenum have been proposed, as

    listed in [77]. As for molybdenum, nitrogen also shows

    strong concentration gradients in the passive film. It has

    been suggested that ammonia or ammonium ions react

    with free chlorine to form combined chlorine species

    that are less effective oxidants, thereby inhibiting

    chlorination enhanced localised corrosion [86]. Another

    possibility is the formation of a nitride at the metal/filminterface which brings down the dissolution rates for the

    individual elements in the alloy [70,87]. Nitride has also

    been put forward as a possible mechanism for synergy

    between nitrogen and molybdenum [70].

    2.4.9. Others

    Copper is added to highly corrosion resistant auste-

    nites (904L, 254SMO) to further boost the corrosionresistance. The influence of copper in Fe/Cr alloys hasbeen studied by Postrach et al. [88,89]. Copper additions

    are also used to give the surface anti-bacterial proper-

    ties, cf. Table 1. Another class of alloying elements with

    intriguing electronic properties are the rare-earth metals

    (REMs). They are used to enhanced the high tempera-

    ture corrosion resistance by improving the adhesion of

    oxide scales.

    3. Atomic arrangement

    Besides chemistry and elemental composition of the

    passive films, their structure is also a vital parameter

    that influences the corrosion resistance. The first in-

    vestigation on the crystal structure of passive films onFe/Cr alloys was performed by McBee and Kruger in1972 [90] using electron diffraction in a transmission

    electron microscope. They found that the films formed

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/1104 1099

  • on alloys with 24% Cr were amorphous, whereas 0/10%Cr in the bulk resulted in a spinel-like structure. With

    the invention of the STM in the early 1980s, it became

    possible to study the crystal structure of the passive filmin situ with atomic resolution.

    3.1. Scanning tunnelling microscopy

    One of the first studies on the atomic arrangement of

    active and passive stainless steel surfaces was performed

    by Ryan et al. [91,92]. By studying iron in a borate

    buffer solution and Fe/Cr alloys in sulphuric acidsolutions, they confirmed the conclusion of McBee

    and Krueger [90] that the passive film is amorphousfor Cr concentrations above 12/19%. Marcus et al. havemade detailed investigations on a series of single crystal

    surfaces: Ni [93], Cr [94], Fe/Cr [48] and Fe/Cr/Ni[45]. The passive films were found to grow close to

    epitaxially (Cr (110)) in small (3 nm diam) crystalline

    regions on iron/chromium alloys. Aging under polar-isation led to growth of the crystallised regions, i.e. an

    enhanced long range order. The ordering process wasobserved to be slower for austenitics, which also showed

    a lower production rate of Cr2O3 on the surface. It was

    suggested that these effects were caused by the enrich-

    ment of nickel at the metal/film interface which limited

    the access of chromium. An example of an STM image

    is given in Fig. 6, which shows the atomic arrangement

    of a passive film on an austenitic Fe/Cr/Ni alloy [45].The STM can also be used to obtain depth informa-

    tion. This was demonstrated by Zuili et al. on Cr

    surfaces [94]. By varying the voltage of the scanning

    tip, it is possible to control the depth from which the

    major part of the response is acquired. The structure

    found was not a flat hydroxide layer, but resembled a

    buried island structure, much like the one suggested by

    Wegrelius and Olefjord based on XPS results [95]. STMstudies of passive films on different materials have

    recently been reviewed by Marcus [96].

    3.2. Geometrical considerations

    Conjectures based on the geometrical arrangement of

    atoms in or near the passive film have been developed as

    a possible explanation for some aspects of passivity. The

    simplest geometrical approach is to consider a bcc-lattice where every atom has eight neighbours. For

    iron/chromium, this means that if 1/8 (12.5%) of theatoms are Cr, at least 50% of the Cr atoms will have at

    least one Cr as their nearest neighbour. This corre-

    sponds to the onset of a continuous network of

    chromium atoms in the bulk material. The fraction 1/8

    also equals the lower limit of chromimum needed to

    make a steel stainless. Percolation theory is related tothe 1/8 concept. It has previously been used in materials

    science to treat different phenomena in amorphous

    materials, e.g. phase transformations. A general treat-

    ment of percolation has been presented by Zallen [97].

    Percolation is used to describe network connectivity.

    For a one-dimensional system (e.g. a string with node

    points), the percolation limit is trivial, since the first

    node cut will cause a break in connectivity between theend points. For a 2D structure, it is possible to find an

    exact percolation limit which will vary with the number

    of connections of each node (atom or molecule) in the

    network. For a 3D structure, the issue is more complex

    and percolation limits become less well defined, since the

    number of possible connections are much higher.

    The connectivity of Fe/Cr matrices was studied for aseries of different compositions by Newman et al. Fortheir Monte Carlo simulations, they initially considered

    a square two/dimensional lattice in which they ran-domly placed iron and chromium atoms [98,99]. By

    assuming a high dissolution rate for iron and zero

    dissolution for chromium, they derived percolation

    limits of 12% for onset and 17% for completed passivity.

    These percentages are in good agreement with practical

    experience.Later, the percolation model was modified to incor-

    porate pit initiation [100]. By assuming a low dissolution

    rate for Cr, it was shown that an intermittent dissolution

    of iron clusters at discrete times could be obtained. If

    large enough, these clusters could serve as initiation sites

    for pitting. Recently, the percolation results have been

    compared to experimental data for Fe/Cr PVD filmsexposed to acidic chloride media [101,102]. The percola-tion approach has also been modified by assuming

    higher dissolution rates at defect sites on the surface, e.g.

    kinks, or if one alloy component is restrained to the

    Fig. 6. Topographic STM image of a passive film on a (100) Fe/18Cr/13Ni substrate after passivation in 0.5M H2SO4 at 500 mVSHEfor 2 h. An almost hexagonal lattice is superimposed to show the

    directions of the close-packed rows. From Maurice et al. [45]

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041100

  • surface by thermodynamic limitations. These aspects

    have been reviewed by Heusler [103]. The percolation

    model can be used to investigate how different dissolu-

    tion rates of iron and chromium affect the surface

    roughness in the nano range [99].

    Newman et al. considered the surface to be passive

    when a hydroxide or oxide chain can bridge the gap and

    form a link between chromium atoms at the surface

    [104]. Another approach is to consider the connectivity

    between cations within the oxide film. This concept has

    been applied to Fe/Cr [105] and other binary Cr alloys[106] by McCafferty, who considered a series of binary

    systems following a formalism developed by Megirditch-

    ian [107] and Randic [108]. For a certain ratio of Fe/Cr

    atoms in the oxide matrix, a connectivity between the

    chromium atoms is obtained. Thus, for a certain

    amount of iron, the continuous chromium network

    through the oxide will be lost. The connectivity para-

    meter was successfully compared with parameters such

    as passivation currents, passivation potentials or experi-

    mental data on passive film compositions [105].

    4. Film growth kinetics

    Passive films are constantly changing and adapting to

    the environment. Thus, understanding film growth and

    dissolution is essential to the understanding of passive

    film stability in different and changing environments.

    High reaction rates and small dimensions of the film (1/3 nm thickness) make it difficult to obtain unique results

    for passive films on stainless steels. Some valve metals

    show thicker films and are thus more easily studied.

    Numerous attempts to model film growth based on

    different concepts have been made; a recent review has

    been written by Macdonald [109]. More details on

    earlier models for film growth dynamics have been

    given by Chao et al. [110].

    For a passive film covering the entire surface, two

    types of rate-limiting processes can be distinguished:

    high field assisted ionic transport through the film or

    charge transfer at either the metal/film or film/electro-

    lyte interface. These models will be referred to as high

    field (HFM) and interface models (IFM), respectively.

    Most models presented in the literature give growth

    equations that reduce to either the IFM or the HFM

    type.

    If the electric field is assumed to increase upon a

    change in potential, the film growth will be limited by

    high field ion conduction through the oxide (HFM). An

    early formalisation of this concept was made by Verwey

    [111]. The HFM is in widespread use for the modelling

    of passive currents, and can be respresented by an

    equation of the form

    @d

    @trgk

    hf1 e

    BU=d (1)

    where d is the film thickness, rg is the growth fraction,

    k1hf a high field model constant containing the specific

    volume and valence, B is an oxide-specific constant, and

    U is the applied potential. For the HFM, the thicknessof the passive film is found in the denominator of the

    exponential.

    The alternative to high field rate control is to consider

    a rate limiting reaction at either the metal/film or the

    film/electrolyte interface. A rate limiting reaction at the

    metal/film interface is the base for the point defect

    model (PDM) that was introduced in the early 80s. The

    PDM concept has since then been developed anddiscussed extensively in a series of publications by

    Macdonald et al. [110,112/117]. A growth rate limitingreaction at the film/electrolyte interface has been pro-

    posed by Vetter and Gorn [118]. It can be shown that

    the form of the rate limiting reaction is independent of

    which interface is considered [119]. Thus, the film

    growth equation for the interface type models (IFM:s)

    takes the form

    @d

    @tkif1 e

    k2 [DUE0 Dd] (2)

    where k1if and k2 are constants with different meanings

    depending on which model is represented [119]. For the

    IFM, the electric field E0 is usually assumed constantduring the applied external potential change. The film

    will show a thickness change Dd until the potentialapplied across the film is balanced. For the IFM, the

    thickness change is found in the nominator and the

    growth rate is not explicitly dependent on the absolute

    film thickness.

    An attempt to unify the IFM and HFM conjectures

    was made by Kirchheim in 1987 [120], focussing ongalvanostatic experiments on iron in acid environments.

    Kirchheim divided the current into a growth and a

    corrosion part: itot/icorr/igrowth. The growth currentwas assumed to be controlled by an interface equili-

    brium, as previously suggested by Vetter and Gorn

    [118,121], and the total current was assumed to be

    limited by the HFM. The corrosion part can be

    estimated by solution analysis. Kirchheim showed thatneither the interface equilibrium nor the high field

    limitation for the total current on its own could

    successfully explain all experimental observations [120].

    To understand steady-state dissolution in the passive

    region, Heusler introduced different diffusion rates for

    the iron and chromium cations in the film. He found an

    agreement with experimental data for diffusion rates of

    iron eight times higher than for those of chromium [67].A similar approach was argued by Kirchheim et al., who

    adopted different transport rates for cations in combi-

    nation with the concepts developed for iron [120].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/1104 1101

  • Measurements of the passive current [122] as well as

    estimates of the amount of dissolved material [123] were

    both interpreted as a lower mobility for chromium in the

    passive film. Further investigations also included ellip-

    sometry [124], that was used to estimate the film growth

    current igrowth. Another possibility is to use solution

    analysis to measure the dissolution current icorr. This

    approach was adopted successfully by Castle and Qiu

    [35], who showed that the dissolved material was heavily

    dominated by iron.

    A combination of the above ideas was introduced in

    an attempt to create a general theory for passive film

    growth-the mixed conduction model (MCM), which has

    been applied to pure Cr [125] as well as Fe/Cr [126/128], Ni/Cr [129], Fe/Mo [130], and Fe/Cr/Mo alloys[131,132]. The MCM includes a rate limiting step at an

    interface as well as transport through the film. It starts

    out from a diffusion equation, which for potential

    sweeps in the anodic direction essentially reduces to

    the HFM. The MCM has been successful in fitting

    impedance as well as passive film conductivity data.

    Recently, the electrochemical quartz crystal micro-

    balance (EQCM) has been used to investigate passivefilm growth [55,133,134]. It gives in situ information

    with a time resolution sufficient to provide real time

    growth curves of the passive film. This has been

    illustrated by Olsson et al., who used the EQCM to

    study passive films on Cr [119], Fe/Cr [135], and Fe/Cr/Ni/(Mo/W) alloys [136]. The film growth during apotential change in the anodic direction for a Fe/Cr/Ni/Mo stainless steel PVD alloy is illustrated in Fig. 7a[136]. The fit parameters for the IFM and HFM were

    fixed for a 304L type PVD material and it was assumed

    that the addition of Mo would essentially affect the

    ability of the film to withstand an increased electric field,

    a hypothesis supported by the difference in thickness

    change between the alloys. As can be seen, the IFM

    produces a satisfactory fit to the growth curve, whereas

    less correspondence is found for the HFM. This experi-ment was repeated for alloys with additions of Cr and

    W. The electric fields in the film derived from the HFM

    and IFM fits were then compared to that calculated

    directly from dividing the applied potential change by

    the experimental EQCM film thickness change, cf. Fig.

    7b. The IFM fit values of the electric field gave a good

    correlation with the direct estimates, whereas an oppo-

    site trend was found for the HFM. In addition, it wasfound that alloying with Mo and W gave considerably

    higher mass losses than alloying with Cr. This indicates

    that their presence renders the film more permeable to

    ion transport, which would allow the film to reach an

    equilibrium composition faster.

    5. Concluding remarks

    There is no such thing as a static passive film. The

    passive film on a given stainless steel changes with the

    environment. It can grow or dissolve, and may adsorb

    or incorporate anions. A parameter critical to the

    stability of a passive film is the time necessary to

    respond to an environmental change. With the intro-

    duction of in situ techniques, different time dependent

    parameters have become accessible.The passive film thickness changes within a couple of

    seconds in response to a potential change. Structural

    ordering, on the other hand, follows considerably slower

    kinetics. Both these parameters are important factors in

    the growth and break down of passive films.

    Acknowledgements

    We are indebted to Mats Liljas at Avesta Polarit

    R&D, and Marie-Gilles Verge for valuable comments to

    Fig. 7. (a) Experimental EQCM growth curve for a passive film on a

    Fe/Cr/Ni/Mo alloy undergoing a potential change in the anodicdirection within the passive region. This curve is compared to growth

    curves simulated using Eqs. (1) and (2) for the HFM and IFM. (b)

    Values of the electric field yielding a best fit to experimental data is

    shown for different alloys for the IFM and HFM:s. The IFM fit values

    compare better with the electric fields calculated directly from the

    EQCM thickness change. From Olsson et al. [136].

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041102

  • the script. Financial support was provided by Fonds

    Nationale Suisse.

    References

    [1] P. Monnartz, Metallurgie 8 (1911) 161.

    [2] B. Wallen, M. Liljas, P. Stenvall, in: P.J. Tunturi (Ed.),

    Eurocorr, 1992, Corrosion Society of Finland, Esbo, 1992, p. 81.

    [3] T. Thorvaldsson, Sci. J. Met. 26 (1997) 71.

    [4] P.J. Uggowitzer, R. Magdowski, M.O. Speidel, ISIJ 36 (1996)

    901.

    [5] H. Fischmeister, U. Roll, Fresenius Z. Anal. Chem. 319 (1984)

    639.

    [6] H.H. Uhlig, Corrosion Sci. 19 (1979) 777.

    [7] N. Sato, Corrosion Sci. 31 (1990) 1.

    [8] J.W. Schultze, M.M. Lohrengel, Electrochim. Acta 45 (2000)

    2499.

    [9] I. Olefjord, B.-O. Elfstrom, Corrosion 38 (1982) 46.

    [10] C.-O.A. Olsson, S.E. Hornstrom, Corrosion Sci. 36 (1994) 141.

    [11] B. Brox, I. Olefjord, Surf. Interface Anal. 13 (1988) 3.

    [12] C. Calinski, H.-H. Strehblow, J. Electrochem. Soc. 136 (1989)

    1328.

    [13] R. Kirchheim, B. Heine, S. Hofmann, H. Hofsass, Corrosion Sci.

    31 (1990) 573.

    [14] H.-H. Strehblow, Corrosion and Prevention 97, Brisbane, Paper

    no. 5, Australasian Corrosion Association, Melbourne, 1997.

    [15] C.-O.A. Olsson, Surf. Interface Anal. 21 (1994) 846.

    [16] R. Goetz, D. Landolt, Electrochim. Acta 27 (1982) 1061.

    [17] P. Marcus, I. Olefjord, Corrosion Sci. 28 (1988) 589.

    [18] S. Haupt, C. Calinski, U. Collisi, H.-W. Hoppe, H.-D. Speck-

    mann, H.-H. Strehblow, Surf. Interface Anal. 9 (1986) 357.

    [19] J.A. Bardwell, G.I. Sproule, M.J. Graham, J. Electrochem. Soc.

    140 (1993) 50.

    [20] M.J. Graham, Corrosion Sci. 37 (1995) 1377.

    [21] M.J. Graham, J.A. Bardwell, G.I. Sproule, D.F. Mitchell, B.R.

    Macdougall, Corrosion Sci. 35 (1993) 13.

    [22] C. Courty, H.J. Mathieu, D. Landolt, Electrochim. Acta 36

    (1991) 1623.

    [23] A.J. Davenport, J.A. Bardwell, C.M. Vitus, J. Electrochem. Soc.

    142 (1995) 721.

    [24] A.J. Davenport, J.A. Bardwell, C.M. Vitus, J. Electrochem. Soc.

    142 (1995) 721.

    [25] L.J. Oblonsky, A.J. Davenport, M.P. Ryan, H.S. Isaacs, R.C.

    Newman, J. Electrochem. Soc. 144 (1997) 2398.

    [26] A.J. Davenport, M. Sansone, J. Electrochem. Soc. 142 (1995)

    725.

    [27] S. Virtanen, P. Schmuki, M. Buchler, H.S. Isaacs, J. Electro-

    chem. Soc. 146 (1999) 4087.

    [28] L.J. Oblonsky, M.P. Ryan, H.S. Isaacs, J. Electrochem. Soc. 145

    (1922) 1998.

    [29] P. Schmuki, S. Virtanen, H.S. Isaacs, M.P. Ryan, A.J. Daven-

    port, H. Bohni, T. Stenberg, J. Electrochem. Soc. 145 (1998) 791.

    [30] T. Stenberg, J. Keranen, P. Vuoristo, T. Mantyla, S. Virtanen, P.

    Schmuki, M. Buchler, H. Bohni, Vacuum 52 (1999) 477.

    [31] P. Schmuki, S. Virtanen, Electrochem. Soc. Interface 6 (1997) 38.

    [32] L.J. Oblonsky, M.P. Ryan, J. Electrochem. Soc. 148 (2001)

    B405.

    [33] K. Asami, K. Hashimoto, S. Shimodaira, Corrosion Sci. 18

    (1978) 151.

    [34] J.A. Bardwell, G.I. Sproule, B.R. Macdougall, M.J. Graham,

    A.J. Davenport, H.S. Isaacs, J. Electrochem. Soc. 139 (1992)

    371.

    [35] J.E. Castle, J.H. Qiu, Corrosion Sci. 29 (1989) 591.

    [36] D. Hamm, K. Ogle, C.-O.A. Olsson, S. Weber, D. Landolt,

    Corrosion Sci. 44 (2002) 1443.

    [37] S. Haupt, H.-H. Strehblow, J. Electroanal. Chem. 216 (1987)

    229.

    [38] I. Annergren, M. Keddam, H. Takenouti, D. Thierry, Electro-

    chim. Acta 38 (1993) 763.

    [39] I. Annergren, M. Keddam, H. Takenouti, D. Thierry, Electro-

    chim. Acta 41 (1996) 1121.

    [40] I. Annergren, M. Keddam, H. Takenouti, D. Thierry, Electro-

    chim. Acta 42 (1997) 1595.

    [41] S. Haupt, H.-H. Strehblow, Corrosion Sci. 37 (1995) 43.

    [42] H.-W. Hoppe, S. Haupt, H.-H. Strehblow, Surf. Interface Anal.

    21 (1994) 514.

    [43] L. Wegrelius, I. Olefjord, 12th International Corrosion Con-

    gress, Houston, TX, NACE, Houston TX, 1993, p. 3887.

    [44] C.-O.A. Olsson, D. Hamm, D. Landolt, J. Electrochem. Soc. 147

    (2000) 2563.

    [45] V. Maurice, W.P. Yang, P. Marcus, J. Electrochem. Soc. 145

    (1998) 909.

    [46] H.-H. Strehblow, Werkst. Korr. 35 (1984) 437.

    [47] P. Marcus, J.M. Grimal, Corrosion Sci. 33 (1992) 805.

    [48] V. Maurice, W.P. Yang, P. Marcus, J. Electrochem. Soc. 143

    (1996) 1182.

    [49] I. Olefjord, B. Brox, U. Jelvestam, J. Electrochem. Soc. 132

    (1985) 2854.

    [50] B. Brox, I. Olefjord, Stainless Steel 1984, Gothenburg, Sweden,

    The Institute of Metals, London, 1984, p. 134.

    [51] I. Olefjord, L. Wegrelius, Corrosion Sci. 31 (1990) 89.

    [52] S. Mischler, A. Vogel, H.J. Mathieu, D. Landolt, Corrosion Sci.

    32 (1991) 925.

    [53] V. Mitrovic-Scepanovic, B.R. Macdougall, M.J. Graham, Cor-

    rosion Sci. 27 (1987) 239.

    [54] C. Hubschmid, D. Landolt, H.J. Mathieu, Fresenius J. Anal.

    Chem. 353 (1995) 234.

    [55] P. Schmutz, D. Landolt, Corrosion Sci. 41 (1999) 2143.

    [56] S. Jin, A. Atrens, Appl. Phys. A 45 (1988) 83.

    [57] R.B. Inturi, Z. Szklarska-Smialowska, Corrosion 48 (1992) 398.

    [58] M. Kraack, H. Bohni, W. Muster, J. Patscheider, Surf. Coat.

    Technol. 68/69 (1994) 541.

    [59] M.-W. Tan, E. Akiyama, A. Kawashima, K. Asami, K.

    Hashimoto, Corrosion Sci. 38 (1996) 1495.

    [60] P.Y. Park, E. Akiyama, H. Habazaki, A. Kawashima, K. Asami,

    K. Hashimoto, Corrosion Sci. 38 (1996) 1649.

    [61] X.Y. Li, E. Akiyama, H. Habazaki, A. Kawashima, K. Asami,

    K. Hashimoto, Corrosion Sci. 39 (1997) 1365.

    [62] J. Bhattrai, E. Akiyama, H. Habazaki, A. Kawashima, K.

    Asami, K. Hashimoto, Corrosion Sci. 40 (1998) 155.

    [63] J. Bhattrai, E. Akiyama, H. Habazaki, A. Kawashima, K.

    Asami, K. Hashimoto, Corrosion Sci. 39 (1997) 355.

    [64] M.-W. Tan, E. Akiyama, H. Habazaki, A. Kawashima, K.

    Asami, K. Hashimoto, Corrosion Sci. 38 (1996) 2137.

    [65] C.-O.A. Olsson, S.E. Hornstrom, Duplex Stainless Steel IV,

    Glasgow UK, Paper no. 68, Abington Publishing, Cambridge,

    UK, 1994.

    [66] C. Hubschmid, H.J. Mathieu, D. Landolt, Surf. Interface Anal.

    20 (1993) 755.

    [67] K.E. Heusler, Corrosion Sci. 31 (1990) 597.

    [68] L. Bjornkvist, I. Olefjord, Corrosion Sci. 32 (1991) 231.

    [69] L. Bjornkvist, I. Olefjord, Proc. Eurocorr 1987, Dechema,

    Karlsruhe, 1987, p. 325.

    [70] R.D. Willenbruch, C.R. Clayton, M. Oversluizen, D. Kim, Y.

    Lu, Corrosion Sci. 31 (1990) 179.

    [71] B. Brox, W. Yi-Hua, I. Olefjord, J. Electrochem. Soc. 135 (1988)

    2184.

    [72] F. Falkenberg, V.S. Raja, E. Ahlberg, J. Electrochem. Soc. 148

    (2001) B132.

    [73] C.-O.A. Olsson, Corrosion Sci. 37 (1995) 467.

    [74] C.R. Clayton, Y.C. Lu, J. Electrochem. Soc. 133 (1986) 2465.

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/1104 1103

  • [75] Y.C. Lu, C.R. Clayton, A.R. Brooks, Corrosion Sci. 29 (1989)

    863.

    [76] Y.C. Lu, C.R. Clayton, Corrosion Sci. 29 (1989) 927.

    [77] Y.C. Lu, M.B. Ives, C.R. Clayton, Corrosion Sci. 35 (1993) 89.

    [78] C.R. Clayton, Y.C. Lu, Corrosion Sci. 29 (1989) 881.

    [79] D. Kim, S.V. Kagwade, C.R. Clayton, Surf. Interface Anal. 26

    (1998) 155.

    [80] R. Jargelius-Pettersson, B.G. Pound, J. Electrochem. Soc. 145

    (1998) 1462.

    [81] G.P. Halada, C.R. Clayton, J. Vac. Sci. Technol. A 11 (1993)

    2342.

    [82] R. Goetz, D. Landolt, Electrochim. Acta 29 (1984) 667.

    [83] D. Landolt, S. Mischler, A. Vogel, H.J. Mathieu, Corrosion Sci.

    31 (1990) 431.

    [84] H.J. Mathieu, D. Landolt, Surf. Interface Anal. 14 (1989) 744.

    [85] A. DiPaola, F. DiQuarto, G. Serravalle, J. Less-Common Met.

    42 (1975) 315.

    [86] M.B. Ives, Y.C. Lu, J.L. Luo, Corrosion Sci. 32 (1991) 31.

    [87] G.P. Halada, D. Kim, C.R. Clayton, Corrosion 52 (1996) 36.

    [88] B. Postrach, I. Garz, H.-H. Strehblow, Werkst. Korr. 45 (1994)

    508.

    [89] B. Postrach, H.-H. Strehblow, I. Garz, Werkst. Korr. 45 (1994)

    544.

    [90] C.L. McBee, J. Kruger, Electrochim. Acta 17 (1972) 1337.

    [91] M.P. Ryan, R.C. Newman, G.E. Thompson, J. Electrochem.

    Soc. 142 (1995) L177.

    [92] M.P. Ryan, R.C. Newman, G.E. Thompson, J. Electrochem.

    Soc. 141 (1994) L164.

    [93] D. Zuili, V. Maurice, P. Marcus, J. Electrochem. Soc. 147 (2000)

    1393.

    [94] D. Zuili, V. Maurice, P. Marcus, J. Phys. Chem. B 103 (1999)

    7896.

    [95] L. Wegrelius, F. Falkenberg, I. Olefjord, J. Electrochem. Soc.

    146 (1999) 1397.

    [96] P. Marcus, Electrochim. Acta 43 (1998) 109.

    [97] R. Zallen, The Physics of Amorphous Solids, Wiley, New York,

    1983.

    [98] S. Qian, R.C. Newman, R.A. Cottis, J. Electrochem. Soc. 137

    (1990) 435.

    [99] S. Qian, R.C. Newman, R.A. Cottis, K. Sieradzki, Corrosion

    Sci. 31 (1990) 621.

    [100] D.E. Williams, R.C. Newman, Q. Song, R.G. Kelly, Nature 350

    (1991) 216.

    [101] M.P. Ryan, N.J. Laycock, R.C. Newman, H.S. Isaacs, J.

    Electrochem. Soc. 145 (1998) 1566.

    [102] M.P. Ryan, N.J. Laycock, H.S. Isaacs, R.C. Newman, J.

    Electrochem. Soc. 146 (1999) 91.

    [103] K.E. Heusler, Corrosion Sci. 39 (1997) 1177.

    [104] R.C. Newman, F.T. Meng, K. Sieradzki, Corrosion Sci. 28

    (1988) 523.

    [105] E. McCafferty, Corrosion Sci. 42 (1993) 2000.

    [106] E. McCafferty, Corrosion Sci. 44 (2002) 1393.

    [107] J.J. Meghirditchian, J. Am. Chem. Soc. 113 (1991) 395.

    [108] M. Randic, J. Am. Chem. Soc. 97 (1975) 6609.

    [109] D.D. Macdonald, Pure Appl. Chem. 71 (1999) 951.

    [110] C.Y. Chao, L.F. Lin, D.D. Macdonald, J. Electrochem. Soc. 128

    (1981) 1187.

    [111] E.J.W. Verwey, Physica 2 (1935) 1059.

    [112] L.F. Lin, C.Y. Chao, D.D. Macdonald, J. Electrochem. Soc. 128

    (1981) 1194.

    [113] D.D. Macdonald, L. Zhang, E. Sikora, J. Sikora, in: P.M.

    Natishan, H.S. Isaacs, M. Janik-Czachor, V.A. Macagno, P.

    Marcus, M. Seo (Eds.), Passivity and its Breakdown, Paris, The

    Electrochemical Society, Pennington NJ, 1997, p. 411.

    [114] D.D. Macdonald, M. Urquidi-Macdonald, J. Electrochem. Soc.

    137 (1990) 2395.

    [115] D.D. Macdonald, J. Electrochem. Soc. 139 (1992) 3434.

    [116] L. Zhang, D.D. Macdonald, E. Sikora, J. Sikora, J. Electro-

    chem. Soc. 145 (1998) 898.

    [117] D.D. Macdonald, M. al-Rifaie, G.R. Engelhardt, J. Electro-

    chem. Soc. 148 (2001) B343.

    [118] K.J. Vetter, F. Gorn, Electrochim. Acta 18 (1973) 321.

    [119] C.-O.A. Olsson, D. Hamm, D. Landolt, J. Electrochem. Soc. 147

    (2000) 4093.

    [120] R. Kirchheim, Electrochim. Acta 32 (1987) 1619.

    [121] K.J. Vetter, Electrochim. Acta 16 (1923) 1971.

    [122] R. Kirchheim, H. Fischmeister, S. Hofmann, H. Knote, U. Stolz,

    Corrosion Sci. 29 (1989) 899.

    [123] B. Heine, R. Kirchheim, Corrosion Sci. 31 (1990) 533.

    [124] J. Hafele, B. Heine, R. Kirchheim, Z. Metallkd. 83 (1992) 395.

    [125] M. Bojinov, G. Fabricius, T. Laitinen, T. Saario, G. Sundholm,

    Electrochim. Acta 44 (1998) 247.

    [126] M. Bojinov, G. Fabricius, T. Laitinen, K. Makela, T. Saario, G.

    Sundholm, Electrochim. Acta 45 (2000) 2029.

    [127] M. Bojinov, G. Fabricius, T. Laitinen, K. Makela, T. Saario, G.

    Sundholm, J. Electrochem. Soc. 146 (1999) 3238.

    [128] M. Bojinov, I. Betova, G. Fabricius, T. Laitinen, R. Raicheff, T.

    Saario, Corrosion Sci. 41 (1999) 1557.

    [129] M. Bojinov, G. Fabricius, P. Kinnunen, T. Laitinen, K. Makela,

    T. Saario, G. Sundholm, Electrochim. Acta 45 (2000) 2791.

    [130] M. Bojinov, I. Betova, R. Raicheff, Electrochim. Acta 44 (1998)

    721.

    [131] M. Bojinov, G. Fabricius, T. Laitinen, T. Saario, Electrochim.

    Acta 44 (1999) 4331.

    [132] M. Bojinov, G. Fabricius, T. Laitinen, K. Makela, T. Saario, G.

    Sundholm, Electrochim. Acta 46 (2001) 1339.

    [133] P. Schmutz, D. Landolt, Electrochim. Acta 45 (1999) 899.

    [134] C. Gabrielli, M. Keddam, F. Minouflet, H. Perrot, Electrochim.

    Acta 41 (1996) 1217.

    [135] D. Hamm, C.-O.A. Olsson, D. Landolt, Corrosion Sci. 44 (2002)

    1009.

    [136] C.-O.A. Olsson, D. Landolt, J. Electrochem. Soc. 148 (2001)

    B438.

    C.-O.A. Olsson, D. Landolt / Electrochimica Acta 48 (2003) 1093/11041104