Effect of aging of 2507 super duplex stainless steel on sliding tribocorrosion in chloride solution

Author’s Accepted Manuscript Effect of Aging of 2507 Super Duplex Stainless Steel on Sliding Tribocorrosion in Chloride Solution J. Michael Shockley, Derek J. Horton, Kathryn J. Wahl www.elsevier.com/locate/wear

PII: DOI: Reference:

S0043-1648(16)30481-1 http://dx.doi.org/10.1016/j.wear.2017.03.019 WEA102122

To appear in: Wear Received date: 17 October 2016 Revised date: 7 March 2017 Accepted date: 9 March 2017 Cite this article as: J. Michael Shockley, Derek J. Horton and Kathryn J. Wahl, Effect of Aging of 2507 Super Duplex Stainless Steel on Sliding Tribocorrosion in Chloride Solution, Wear, http://dx.doi.org/10.1016/j.wear.2017.03.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of Aging of 2507 Super Duplex Stainless Steel on Sliding Tribocorrosion in Chloride Solution J. Michael Shockley3, Derek J. Horton1, Kathryn J. Wahl2 1

Center for Corrosion Science and Technology (Code 6134). Naval Research Laboratory, Washington, DC 20375 USA 2 Tribology and Molecular Interfaces Section (Code 6176). Naval Research Laboratory, Washington, DC 20375 USA 3 NRC Postdoctoral Research Associate sited in Chemistry Division, Naval Research Laboratory, Washington, DC 20375 USA

Abstract Grade 2507 super duplex stainless steel ordinarily achieves a balance of corrosion resistance and mechanical properties through its dual phase ferrite-austenite microstructure. However, heat treatment in the 600-900°C temperature range causes phase transformations to occur, developing complex microstructures with secondary phases including sigma phase, secondary austenite, chi phase, and chromium nitrides. It has already been demonstrated that in tribocorrosion experiments in 0.6 M NaCl in anodic potentiostatic conditions, the passivity of aged 2507 is eliminated due to mechanical wear, and pitting can occur in and near the wear track. However, the precise mechanism of this loss of passivity is not yet understood. In the present study, we explore the tribocorrosion behavior at selected phases and grain boundaries. Correlating the microstructural features to local surface topography changes and the current transient response revealed precise details of the wear-induced corrosion behavior.

1

Introduction

Page 1 of 29

Duplex stainless steels (DSS) have a good combination of corrosion resistance and mechanical properties due to their ferrite-austenite microstructures [1]. Corrosion resistance is accomplished through the formation of a chromium-rich passivation layer, and the dual phase microstructure is beneficial to both the corrosion resistance (particularly stress and crevice corrosion resistance from the ferrite phase) and mechanical properties (ductility and toughness from the austenite phase) [1]. SAF 2507 (UNS S32750) is noted for its exceptionally high pitting corrosion equivalency number (PREN) in excess of 40, placing it among the “superduplex” stainless steel (SDSS) grades [1]. When subject to heat treatments to temperatures in the 600-900°C range, 2507 and other grades of DSS form secondary phases such as sigma, secondary austenite, chi, and Cr2N phases [1, 2]. Such phases are often deleterious to both corrosion resistance and mechanical properties [2-5]. Among these phases, sigma phase forms in the highest volume fraction and is rich in chromium and molybdenum, resulting in depletion of passivating alloying elements in the surrounding phases [2, 4, 6, 7]. This process leads to a local relative increase of nickel concentration in the surrounding material, helping drive the formation of secondary austenite in a lamellar structure in the vicinity of the sigma phase. Secondary austenite can also form as Widmanstatten-like plates independent of the formation of sigma phase [8]. Although the formation of sigma and other phases is avoided as much as possible in normal processing conditions for engineering alloys, it can form during weld repairs and from other processes that cause local thermal cycling [1].

Page 2 of 29

Much of the previous characterization of the effect of sigma phase on the corrosion behavior of DSS has been studied with a wide range of conventional techniques such as potentiodynamic and potentiostatic polarization testing, measurements of material loss after various exposure times to corrosive media [7], and derivation of critical pitting temperatures (CPT) [2], as well as more advanced micro-techniques such as Scanning Kelvin Probe Microscopy [9]. However, 2507 and other DSS grades are often used in moving parts subject to mechanical contact, which can lead to removal or degradation of the passive film by wear processes, as well as subsurface plastic deformation and phase transformations [10]. Synergism between wear and corrosion can cause greater material losses than those from either process alone [11]. While the relationships between tribological stresses and corrosion are complex and challenging to unravel, the relative contributions of mechanical and chemical factors have been modeled with some success predicting experimental observations [12-18]. The influence of aging-induced secondary phases on the tribocorrosion behavior of 2507 has not been explored despite the widespread use of 2507 and its potential for sensitization in weld zones and similar areas. Given the well characterized microstructural changes associated with aging of 2507, this system presents a clear opportunity to induce microstructural changes and evaluate the interactions between different phases during tribocorrosion. In this work, 2507 SDSS was subjected to heat treatments to form varying quantities of sigma and chi phases in the microstructure, then subjected to sliding tribocorrosion tests in chloride electrolyte under anodic potentiostatic conditions. Friction and material loss rates were quantified and compared to the measured corrosion current.

Page 3 of 29

Microanalysis after testing was used to correlate the material loss with the presence of sigma phase in and out of the wear track. 2 2.1

Experimental Sample preparation

Plates (75.4 x 75.4 x 6.35 mm) of rolled 2507 alloy of standard composition (ASTM A459 S32750, see Table 1) were heat treated in ambient air at 1250°C for 10 hours within the two phase austenite and ferrite phase field to promote grain growth and then water quenched. The plates were subsequently annealed at 800°C for 10, 30 and 60 minutes to promote the ferrite to sigma and chi phase transformation. After thermal processing, the surfaces were significantly ground to remove thermally induced oxides. Plates were sectioned into individual samples (10 x 25 x 6.35 mm) and then an insulated wire was spot welded to the back of each sample to permit electrical conductivity and the sample was then mounted in epoxy. All samples were polished to 1 µm diamond and the back of each mount was ground parallel to the sample surface. Electroplating tape was applied to minimize the exposed surface area and over all exposed epoxy-steel interfaces to prevent crevice corrosion. Table 1: Elemental composition of 2507 alloy. Fe Cr Ni Mo Mn Si Cu N C P S Wt.% 62.18 25.15 6.97 3.9 0.8 0.41 0.3 0.252 0.023 0.019 0.001

2.2

Microstructural characterization and hardness testing

Page 4 of 29

After metallographic polishing, samples were imaged in a scanning electron microscope (SEM) (JEOL 7100f) using backscattered electron (BSE) z-contrast. The relative elemental concentration of each phase was measured using energy-dispersive x-ray spectroscopy (EDS) using a 15 kV accelerating voltage. One sample was examined in cross-section after focused ion beam (FIB) milling with a gallium ion beam current of 6.5 nA. Hardness was measured using a nanoindenter (Hysitron UBI) equipped with a Performech digital controller and a high-load indentation head. A Berkovich indenter was used to make at least 120 indents per sample with 10 µm spacing between indents. Indents were performed with a 5 second loading time, 5 second hold time, and 5 second unloading time with a peak load of 50 mN. Hardness was calculated using the Oliver and Pharr method [19]. X-ray diffraction (XRD) was carried out using a Rigaku x-ray diffractometer. 2.3

Electrochemical testing

Potentiodynamic and potentiostatic polarization experiments were conducted using a traditional three-electrode cell with a platinum mesh counter electrode at ambient temperature and aeration. Potential measurements were made using a saturated calomel reference electrode (SCE). Scan rates during poteniodynamic polarization were 1 mV/s. Characteristic potentiostatic polarization experiments were conducted for at least 10,000 s in 0.6 M NaCl electrolyte solutions prepared using pure water (18.2 MΩ). Replicate experiments were conducted at least 3 times to ensure reproducibility. The following assumptions were used for volume loss calculations based on Faraday’s Law: the molar average equivalent valence state is assumed to be 2.28, the equivalent molar mass is 55.06 g/mol, and the density is 7.8 g/cm3. Page 5 of 29

2.4

Sliding tribocorrosion tests

A linear reciprocating tribometer was custom built using an open architecture platform in conjunction with a potentiostat (see Figure 1). Each mounted specimen was fixed to the bottom of an insulating polymer (polyvinyl chloride) well and encircled by a platinum mesh serving as the counter-electrode. An alumina ball (6.35 mm) was fixed on the end of an insulating ceramic tube using cyanoacrylate adhesive. The cantilever arm was dead loaded to 0.9 N above the pin and lateral force was measured using strain gauges in a Wheatstone bridge configuration. The calculated maximum Hertzian stress for the contact was 0.47 GPa. The cantilever arm was fixed on a brushless linear motor stage assembly allowing for reciprocating kinematics. The polymer well was filled with 0.6 M NaCl electrolyte prepared using deionized (DI) water (18.2 MΩ-cm) at ambient temperature and aeration.

Figure 1: Left: diagram of tribocorrosion test cell; right: photograph of actual laboratory setup.

Page 6 of 29

Sliding wear tests were conducted under potentiostatic conditions of either +100 mVSCE or +600 mVSCE, with samples being allowed to passivate for at least 30 minutes before sliding. Wear tests were conducted at 1 mm/s sliding velocity with an amplitude of 10 mm and run to 100 cycles or 3 cycles to determine the microstructural initiation sites of tribocorrosion induced degradation. At least three repeat measurements were conducted per heat treatment and cell potential. Tests were also run in deionized water (DI) to serve as a control environment with low ionic conductivity, no chloride and no applied potential. Wear track profiles were measured using a white light interferometer. Volumetric material loss rates were calculated by dividing the volume of material removed from the wear track by the applied load and the total sliding distance. Corrosion induced volume losses were calculated using Farday’s law. A weighted average of equivalent weight and valence number was assumed. Volume loss comparisons between profilometry and corrosion calculations assume that the corrosion volume is from within the wear track. 3 3.1

Results Microstructure and static corrosion testing

The microstructure after 10 minutes, 30 minutes, and 60 minutes at 800°C revealed more tertiary phases as heat treatment times increased (see Figure 2). After 10 minutes, the microstructure remains dual phase with only ferrite and austenite grains formed. After 30 minutes, chi phase has begun to form at the grain boundaries and shows high z-contrast because is particularly rich in molybdenum (see Figure 3); some sigma phase is visible as well growing into the ferrite phase. After 60 minutes, extensive Page 7 of 29

formation of sigma phase is evident, growing from the ferrite/austenite grain boundary into the ferrite grains (see Figure 3) [2]. Secondary austenite is visible in two places in the microstructure; one forms in the immediate vicinity of sigma phase lamellae, while independent secondary austenite forms in fragmented Widmanstatten-like plates in the middle of the ferrite phases. Both forms of secondary austenite were found to contain lower levels of chromium than the other phases (see Figure 3). The average hardness was found to increase with longer heat treatment times, from 4.2 ± 0.2 GPa after 10 minutes to 5.6 ± 1.3 GPa after 60 minutes (see Figure 2). XRD revealed no sigma phase present after the 10 minute heat treatment time and that it formed and increased in concentration with longer heat treatment times (see Figure 4). In the case of the 10 minute heat treatment the sample showed significant texture, with higher index austenite and ferrite peaks dominant. After longer heat treatments that texturing is lost as the secondary phases precipitate.

Figure 2: Sample microstructures after 10, 30, and 60 minutes at 800°C (left, center, and right, respectively), BSE z-contrast. The dark phase is δ-ferrite and the medium-shade grains γ-austenite. The light phase forming at grain boundaries is chi (χ) phase from which from which the light, feather-like sigma (σ) phase grows into the ferrite grains. The growth of sigma phase is more pronounced after 60 minutes than after 30 minutes. The measured average hardness plus and minus one standard deviation is shown in the bottom-right corner of each image.

Page 8 of 29

Figure 3: Left: high magnification BSE image showing δ (ferrite), γ (austenite), γ 2 (secondary austenite), γ2,indep. (independent secondary austenite), χ (chi), and σ (sigma) phases after a 60 minute heat treatment at 800°C. Right: relative mass concentrations of each phase as determined by EDS. See text for description of independent austenite.

Figure 4: XRD spectra of the tested samples, showing the formation of sigma phase with increasing heat treatment times.

Page 9 of 29

Electrochemical characterization tests revealed similar behavior at the three tested heat treatments, displaying passive current levels and negative hysteresis, indicative that no pitting took place. However, a notable difference between the three samples was that at above 800 mVSCE, the current was higher for samples exposed to longer 800°C heat treatment times (see Figure 6). Potentiostatic polarization at 100 and 600 mVSCE corroborates behavior expected from the cyclic polarization (Figure 6). At 600 mVSCE, exponential decay in the current is seen as the passive layer is formed, and that passive film is maintained over the course of the experiment. The active peak seen at an 800 mVSCE applied potential shows a clear increase in charge that correlates to the increase in annealing duration, yet no bulk phase dissolution was observed. Subsequent to this oxidation peak, the currents decay to similar passive current densities as those observed at 600 mVSCE.

Page 10 of 29

Figure 5: Potentiodynamic polarization curves; dashed horizontal lines at 600 and 100 mV represent the conditions at which tribocorrosion testing was carried out. The dashed box shows that above 800 mV, the current was higher with longer 800°C heat treatment times. The scan direction is indicated by arrows.

Page 11 of 29

Figure 6: Representative potentiostatic current transients at 600 mVSCE (left) and 800 mVSCE (right) in the absence of wear. The current transient at 100 mVSCE (not pictured) was similar to that at 600 mVSCE.

3.2

Tribocorrosion tests

3.2.1 Corrosion current during and after sliding The amount of current flowing through the electrochemical cell during sliding varied according to the applied potential and the heat treatment time at 800°C (see Figure 7). For each sample at 100 mVSCE, once sliding started after the initial passivation, the corrosion current was more or less stable and was between 2 and 3.5 µA throughout sliding. Once sliding stopped, the samples quickly repassivated to corrosion currents close to those from the initial passivation. At 600 mV, the samples behaved quite differently. For the sample heat treated to 10 minutes at 800°C, the corrosionrepassivation current trend was similar to 100 mVSCE, showing rapid repassivation. However for the 30 minute sample the current trended upward while fluctuating significantly, and by the end of sliding was roughly 3 times that of the 10 minute sample. After sliding stopped, the current immediately dropped slightly, but then began to rise

Page 12 of 29

continually at a similar slope to its rise during sliding. The 60 minute sample behaved similarly to the 30 minute sample except that the current rose at a much higher rate, and by the end of testing the current was orders of magnitude higher than either the 30 or 10 minute samples. When tribocorrosion tests were run to only 3 cycles and stopped (not pictured), the samples exhibited the same repassivation behavior as after 100 cycles, where the behavior depended on heat treatment time and applied potential.

Page 13 of 29

Page 14 of 29

Figure 7: Representative corrosion current transients from tribocorrosion tests conducted in potentiostatic conditions of (a) 100 mVSCE and (b,c) 600 mVSCE.

Friction forces monitored during sliding showed that the friction was generally similar between 100 and 600 mV conditions, with an initial spike to around 0.5 and then settling between 0.38 and 0.45 (see Figure 8, left). In DI, the friction was considerably higher and more unstable, remaining between 0.5 and roughly 0.6. The high friction levels for the 100 and 600 mV conditions did not develop until roughly the third cycle; before that, the COF of the first cycle was low (~0.15) and increased with each cycle (see Figure 8, right), with the same trend visible in the corrosion current.

Figure 8: Left: Average coefficient of friction (COF) per cycle at the 600 mV, 100 mV, and in DI; right: current (blue) and COF per cycle (black) in the opening cycles of sliding.

3.2.2 Wear track surface analysis To observe tribocorrosion-induced microstructural changes, wear track surfaces were observed using electron microscopy (see Figure 9). For tests conducted at 100 mV, the wear tracks resembled one another in that small voids (<5 um) were visible on the Page 15 of 29

surfaces. Grain boundaries, which had previously been well defined, became blurry and jagged. For tests conducted at 600 mVSCE, the wear tracks differed from one another considerably. For the sample heat treated for 10 minutes, the wear track resembled those tested at 100 mVSCE. For the sample heat treated for 30 minutes, the same was largely true, except that occasional clusters of small pits (~2 µm in diameter) were visible in regions of sigma phase. The sample treated for 60 minutes showed extensive damage to the surface in the form of clusters of small pits, as well as some larger pits (~5 µm in diameter).

Figure 9: Representative micrographs of wear track surface analysis: 100 mV (top) and 600 mV (bottom), backscattered electron z-contrast. Dark features are pits.

The regions of the 30 and 60 minute heat treated samples that exhibited the most pitting tended to be associated with sigma phase (see Figure 10). Pits were most extensive at the σ/γ2 grain boundaries, while pits were also occasionally visible at the δ/γ 2and also at Page 16 of 29

the γ/χ grain boundaries. Heavily pitted regions were visible within the wear track, but also outside of them, up to 100 µm away from the edge of the wear track.

Figure 10: Representative micrographs of wear track surface analysis: 100 mV (top) and 600 mV (bottom), backscattered electron z-contrast. Dark features are corrosion products and/or wear debris sitting on the surface of the sample.

It was noted that when heavily pitted samples were left out for several days and exposed to room temperature air and ~60% relative humidity, debris and even wormlike structures formed on the surface (see Figure 11). The debris formed in the vicinity of heavily pitted regions, although not every heavily pitted region formed these structures. One region with surface debris from a 60 minute heat treated sample subjected to sliding under 600 mV conditions, and then sat at potential after sliding stopped for several minutes, was cross-sectioned using FIB (see Figure 11, top). Underneath the surface debris, a porous (~2 µm diameter) microstructure was visible, similar to the surface pitting in the vicinity of the debris. A larger void (~5 µm) was also

Page 17 of 29

present. Debris from a similar region was analyzed using EDS (see Figure 11, bottom) revealing elevated levels of chlorine and oxygen, with low levels of iron and molybdenum compared to the other alloying elements.

Figure 11: Surface debris formed after several days of exposure to ambient air (~60% relative humidity) on a 60 minute 800°C sample after sliding under 600 mV conditions. Top left: secondary electron image; top right: the same region observed after a rough cut FIB cross-section showing corrosion pit and sigma phase microstructure; bottom: a different set of corrosion products and its corresponding EDS mapping.

Page 18 of 29

3.2.3 Material loss rates The measured volume of material lost during each tribocorrosion test condition is shown in Figure 12. After sliding in DI, loss rates were the lowest of all sliding conditions and were slightly lower with increasing heat treatment time at 800°C. Tests at 100 mV had substantially higher wear rates but maintained the same trend of slightly decreased material loss rates with annealing time. After sliding at 600 mV, however, loss rates were higher overall than at 100 mV and were also considerably higher with increasing heat treatment time.

Figure 12: Material loss rates for 10, 30, and 60 minute heat treated samples after sliding in DI and 0.6 M NaCl under 100 and 600 mVSCE. Error bars represent plus and minus the standard error of the mean.

Page 19 of 29

4 4.1

Discussion Connecting the present observations

The experimental results show the tribocorrosion behavior of the samples tested was dependent on the microstructure as well as the applied cell potential. The heat treatements at 800°C produced the expected tertiary microstructural components including sigma, chi, and secondary austenite. The relative concentrations of chromium, molybdenum, and nickel in each phase (Figure 3) were in agreement with previous literature, with higher nickel concentrations in the austenite phases, the highest molybdenum content in the chi phase, and the secondary austenite phases having lower levels of chromium than the other phases. Also, the sigma phase XRD spectra show good agreement in relative intensity and position to those published [20, 21]. However, the sigma phase formed more slowly than that typically reported; the lack of significant changes after 10 minutes at 800°C contrasts with the changes observed after just a few minutes by Berecz et al. [22]. Time-temperature-transformation diagrams also tend to predict faster kinetics than those observed in the present results [23]. This may be related to this study’s initial heat treatment at 1250°C to increase grain size over the as-received condition, which would increase the diffusion distances and decrease the total grain boundary area. Nevertheless, the development of sigma and other phases in the microstructure were accompanied by an increase in hardness, as previously reported [2, 24]. The microstructural changes did not affect the room temperature corrosion resistance in the absence of wear at +100 mVSCE and +600 mVSCE.

Page 20 of 29

In tribocorrosion tests conducted in DI and in 0.6 M NaCl at 100 mVSCE, the longer heat treatment time at 800°C resulted in slightly decreased material loss rates. This trend can be explained by the increased hardness as a function of longer heat treatment time. The current transients at 100 mV show that the wear tracks repassivated as soon as sliding stopped for each heat treatment condition. The higher overall material loss rates at 100 mVSCE compared to sliding in DI indicate that the anodic potential and presence of chloride ions dissolved some material at the the surface exposed by sliding, but the presence of sigma phase did not significantly impact the ability of the surface to repassivate under 100 mV potential. However, the current transients at the more strongly anodic 600 mV condition showed different behavior depending on heat treatment time. The heat treatment of 10 minutes at 800°C, where no sigma phase was observed in the microstructure, resulted in similar sliding repassivation behavior at 600 mV to all samples tested at 100 mV. However, after 30 minutes and 60 minutes at 800°C, the samples did not repassivate, with a steadily increasing background current that continued even once sliding was stopped. The fact that this behavior became more pronounced with greater quantities of sigma phase, and that it correlated to higher material losses, suggests that the presence of sigma and other tertiary phases resulted in a wear-induced loss of passivity of secondary austenite within, and as far as 100 µm away from, the wear track. This is evidently a wear initiated process, given that during the potentiostatic and potentiodynamic electrochemical characterization of these samples (Figure 5 and Figure 6), no bulk phase dissolution was observed. This was the case in spite of the

Page 21 of 29

oxidation peaks at 800mV, driving voltages higher than the 100 and 600 mVSCE applied during tribocorrosion. The surface debris that formed after several days in ambient air resembles corrosion products in that it contains high levels of oxygen and chlorine, suggesting that it may arise from salt residues in the pits reacting with remaining water and/or atmospheric humidity, then protruding from the wear track upon reaction. 4.2

Synergistic model of material loss

One of the earliest and widely applied tribocorrosion models is the synergistic model [16] of material loss, which has been adapted and contributed to by numerous researchers [10, 15, 25] and standardized [26]. The total volume lost, 𝑉

is broken into

individual components: 𝑉

=𝑉 +𝑉 +𝑉

+𝑉

(1)

where 𝑉 is the purely mechanical wear loss, 𝑉 is the background corrosion current, and 𝑉

and 𝑉

represent the synergistic terms of wear-induced corrosion and corrosion-

induced wear, respectively. The 𝑉 and 𝑉

terms are both calculated through

Faraday’s Law: (2)

𝑉 = 𝑉

=

(3)

Page 22 of 29

where Q represents the total charge transfer calculated by integrating the current transient, 𝑀 the molar mass, 𝑛 the mole averaged oxidation number (see Section 2.3), F the Faraday constant, and 𝜌 the mass density of the alloy. Generally the alloy is assumed to be passivating, with the background current dropping to a low level and only a small region of the wear track active as it re-passivates [10, 27]. Table 2: Material loss from wear induced corrosion (𝑉 ) calculated from Faraday’s law after subtracting the background current, expressed as a fraction of the total measured volume loss (𝑉 ) during tribocorrosion

10 minutes 800°C

30 minutes 800°C

60 minutes 800°C

100 mV

87%

94%

86%

600 mV

92%

187%

2,630%

The wear-induced corrosion was quantified using Equation 3 and expressed as a percentage of the total material loss (see Table 2). When expressed this way, the chemical contributions range from 87% of the total material loss, to greater than 100% for the 30 minute sample at 600 mV, to greater than 2,500% for the 60 minute sample at 600 mV. This exceeds any physical possibility if only material in the wear track is considered. However, these very high values, particularly for the cases of the 30 and 60 minute samples at 600 mV, can be reconciled by considering several factors of the experiment. First, Figure 11 shows that the material loss penetrates below the surface of the wear track. The lamellar network formed by the sigma phase and secondary austenite creates pathways for electrolyte to penetrate under the surface as one phase or grain boundary is preferentially corroded. Thus, a systematic underestimate of real material Page 23 of 29

losses is already an issue given that white light interferometry cannot penetrate the surface. This cannot be easily reconciled by weighing the change in sample mass because even the highest material loss rates correlate to only a few tens of µg, compared to the tens of grams of the samples themselves. This would introduce considerable uncertainty in the material loss rate measurements, as has already been established [28]. Furthermore, it is evident in Figure 9 that corrosion pits extend outside the wear track, appearing to follow contiguous grains and/or grain boundaries. This suggests, as does the penetration of pitting below the surface, that the current flow is not simply associated with re-passivation. Instead a total loss of passivity in the vicinity of the wear track is taking place due to the highly anodizing potentials and the presence of deleterious phases induced by heat treatment. This process likely starts through local depassivation in the wear track in a region more susceptible to corrosion, such as the Cr-depleted secondary austenite. Once initiated, the pitting is driven by purely corrosive pit growth phenomena such as local pH increases. This behavior makes the synergistic model an inappropriate model to apply to this system, as it was largely developed for passivating alloys. 4.3

Comparison to scratch corrosion tests

The pitting potential is defined by the potential at which passive film breakdown occurs and undergoes rapid local dissolution as the pit forms. This potential is measured experimentally by scanning potential at a fixed rate through the passive region until a rapid increase in anodic current is observed [29]. In situ re-passivation transient scratch

Page 24 of 29

tests, typically performed with a diamond scribe, have been used fairly extensively to further characterize the oxide at applied potentials below an expected pitting potential. If the surface no longer re-passivates after the scratch test, and instead the current climbs continually, this is an indication of susceptibility to pitting corrosion [30-32]. Interestingly, this approach has been applied once to 2205 duplex stainless steel heat treated to form various quantities of sigma phase, then subject to scratching while immersed in 0.6 M NaCl at various potentials [33]. In those experiments, the samples, initially not susceptible to localized breakdown and pitting, exhibited pitting potentials as low as 300 mVSCE as the amount of sigma phase increased through extended heat treatment time at 850°C. This compares favorably with the results in the present study, where 100 mVSCE was too weakly anodic to induce pitting under any conditions regardless of heat treatment, while 600 mVSCE was sufficient to induce pitting when sigma, chi, and secondary austenite phases were present and the surface was abraded. As the pitting initiation behavior was observed after only 3 sliding cycles, this confirmed that mechanical surface damage from the hard alumina ball was sufficient to break down the oxide layer and allow pitting to initiate when sigma, chi, and secondary austenite phases were present. In the scratch test re-passivation experiments, the rise in pitting susceptibility was attributable to either an increase in occluded surface sites or by removal of the oxide allowing a critical amount of ionic interaction with the bare surface thereby detrimentally influencing passivity [30, 31]. Reciprocating sliding is similar in character to the scratch re-passivation tests. The surface is deformed during sliding which leads to an increase in the number of occluded active sites, likely sigma and secondary austenite phases, Page 25 of 29

decreasing the ability to form a contiguous passive film. Reciprocating sliding also causes periodic removal of the passive film, diminishing any protective effect from thickening of the passive film. The underlying cause initiating passivity breakdown, such as composition or microstructural changes accompanying sliding, is not yet known. 5

Conclusions

The present study sought to explore the influence of microstructural changes in aged 2507 SDSS on its tribocorrosion behavior. To test this, isothermal heat treatments of different lengths were used to drive varying degrees of microstructural transformations, and then the samples were subject to conventional corrosion testing (in the absence of sliding contact) as well as sliding tribocorrosion experiments at low and high anodic electrochemical potentials. It was found that the formation of sigma, chi, and other tertiary phases through heat treatment at 800°C produced no notable differences in the corrosion behavior of 2507 at low and high anodic potential (100 and 600 mV, respectively) measured in the absence of sliding contact. Sliding tribocorrosion experiments at low anodic potential (100 mVSCE) revealed that all samples repassivated at the end of sliding, regardless of heat treatement time and thus regardless of the presence of secondary phases. Tribocorrosion experiments at high anodic potential (600 mVSCE) revealed that the presence of sigma and other secondary phases influenced the re-passivation behavior, with longer heat treatment times resulting in a steadily increasing background current that continued to climb even after sliding stopped. This indicated that at high anodic potential, the sliding disrupted the passivity in microstructural regions containing low chromium content, as this correlated to pitting

Page 26 of 29

corrosion at and below the surface at the secondary austenite phase, both within and in regions near to the wear track.

Acknowledgements This work was supported by the Basic Research Program of the Naval Research Laboratory (NRL). JMS was supported through a National Research Council Postdoctoral Research Associateship.

References 1. 2.

3. 4. 5.

6. 7. 8. 9.

10.

Nilsson, J.O., Super duplex stainless steels. Materials Science and Technology, 1992. 8(8): p. 685700. Nilsson, J.O. and A. Wilson, Influence of isothermal phase transformations on toughness and pitting corrosion of super duplex stainless steel SAF 2507. Materials Science and Technology, 1993. 9: p. 545-554. Potgieter, J.H., Influence of sigma phase on general and pitting corrosion resistance of SAF 2205 duplex stainless steel. British Corrosion Journal, 1992. 27(3): p. 219. Lopez, N., M. Cid, and M. Puiggali, Influence of sigma phase on mechanical properties and corrosion resistance of duplex stainless steels. Corrosion Science, 1999. 1999(41): p. 1615. Nilsson, J.O., et al., Mechanical properties, microstructural stability and kinetics of phase formation in 29Cr-6Ni-2Mo-0,38N superduplex stainless steel. Metallurgical and Materials Transactions A, 2000. 2000(31A): p. 35. Hsieh, C.-C. and W. Wu, Overview of intermetallic sigma phase precipitation in stainless steels. ISRN Metallurgy, 2012. 2012(732471): p. 16. Ravindranath, K. and S.N. Malhotra, Influence of Aging on Intergranular Corrosion of a 25% Chromium-5% Nickel Duplex Stainless Steel. Corrosion, 1994. 50(4): p. 318-328. Junior, A.M.J., G.S. Reis, and O. Balancin, Influence of the microstructure on the plastic behaviour of duplex stainless steels. Materials Science and Engineering A, 2011. 528(6): p. 2259-2264. Sathirachinda, N., et al., Scanning Kelvin probe force microscopy study of chromium nitrides in 2507 super duplex stainless steel—Implications and limitations. Electrochimica Acta, 2011(56): p. 1792-1798. Mischler, S., S. Debaud, and D. Landolt, Wear-accelerated corrosion of passive metals in tribocorrosion systems. Journal of the Electrochemical Society, 1998. 145(3): p. 750-758.

Page 27 of 29

11. 12. 13. 14. 15. 16. 17. 18.

19.

20.

21.

22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

Landolt, D. and S. Mischler, Tribocorrosion of Passive Metals and Coatings. Series on Metals and Surface Engineering. 2011: Woodhead Publishing. Mischler, S., Triboelectrochemical techniques and interpretation methods in tribocorrosion: a comparative evaluation. Tribology International, 2008. 41(7): p. 573-583. Jemmely, P., S. Mischler, and D. Landolt, Electrochemical modeling of passivation phenomena in tribocorrosion. Wear, 2000. 237(1): p. 63-76. Diomidis, N., et al., A methodology for the assessment of the tribocorrosion of passivating metallic materials. Lubrication Science, 2009. 21(2): p. 53-67. Jiang, J., M.M. Stack, and A. Neville, Modelling the tribo-crrosion interaction in aqueous sliding conditions. Tribology International, 2002. 35(10): p. 669-679. Watson, S.W., et al., Methods of measuring wear-corrosion synergism. Wear, 1995. 181-183(2): p. 476-484. Maldonado, S.G., et al., Mechanical and chemical mechanisms in the tribocorrosion of a Stellite type alloy. Weaer, 2013. 308(1-2): p. 213-221. Cao, S. and S. Mischler, Assessment of a recent tribocorrosion model for wear of metal-on-metal hip joints: Comparison between model predictions and simulator results. Wear, 2016. 362-363: p. 170-178. Oliver, W.C. and G.M. Pharr, An improved technique for determining hardness and eleastic modulus using load and displacement sensing indentation experiments. Journal of Materials Research, 1992. 7(6): p. 1564-1583. Elmer, J.W., T.A. Palmer, and E.D. Specht, Direct Observations of Sigma Phase Formation in Duplex Stainless Steels using In Situ Synchrotron X-Ray Diffraction. Metallurgical and Materials Transactions A, 2007. 38(3): p. 464-475. Chen, T.H. and J.R. Yang, Effects of solution treatment and continuous cooling on sigma phase precipitation in a 2205 duplex stainless steel. Materials Science and Engineering A, 2001. A311: p. 28-41. Berecz, T., et al., Decomposition kinetics of ferrite in isothermally aged SAF 2507-type duplex stainless steel. Journal of Materials Engineering and Performance, 2015. 24(11): p. 4777-4788. Li, X., A.P. Miodownik, and N. Saunders, Modelling of materials properties in duplex stainless steels. Materials Science and Technology, 2002. 18: p. 861-868. Angelini, E., B.D. Benedetti, and F. Rosalbino, Microstructural evolution and localized corrosion resistance of an aged superduplex stainless steel. Corrosion Science, 2004. 46(6): p. 1351-1367. Diomidis, N., et al., Tribocorrosion of stainless steel in sulfuric acid: Identification of corrosion– wear components and effect of contact area. Wear, 2010. 269: p. 93-103. ASTM G119-09(2016), Standard Guide for Determining Synergism Between Wear and Corrosion. 2016: ASTM International, West Conshohocken, PA. Garcia, I., D. Drees, and J.-P. Celis, Corrosion-wear of passivating materials in sliding contacts based on a concept of active wear track area. Wear, 2001. 249(5-6): p. 452-460. Schmitz, T.L., et al., Wear-rate uncertainy analysis. Journal of Tribology, 2004. 126: p. 802-808. Stoyanov, P., et al., Combining in situ and online approaches to monitor interfacial processes in lubricated sliding contacts. MRS Communications, 2016. 6(3): p. 301-308. Zhang, W., A.G. Carcea, and R.C. Newman, Pitting of steam-generator tubing alloys in solutions containing thiosulfate and sulfate or chloride. Faraday Discussions, 2015. 180: p. 233-243. Newman, R.C., Pitting of Stainless Alloys in Sulfate Solutions Containing Thiosulfate Ions. Corrosion, 1985. 41(8): p. 450-453. Newman, R.C., H.S. Isaacs, and B. Alman, Effects of sulfur compounds on the pitting behavior of type 304 stainless steel in near-neutral chloride solutions. Corrosion, 1982. 38(5): p. 261-265.

Page 28 of 29

33.

Elshawesh, F., N. Elahresh, and A. Elhoud, Effect of sigma phase on pitting corrosion of 22-5 duplex stainless steel. British Corrosion Journal, 1998. 33(4): p. 285-287.

Highlights



The tribocorrosion behavior of 2507 super duplex stainless steel in 0.6M NaCl solution was studied under reciprocating sliding conditions.



Heat treatments were used to enlarge the grains and precipitate undesirable secondary phases such as those found in weld repairs. These microstructural changes did not affect the resistance of the material to electrochemical attack under potentiostatic and potentiodynamic conditions.



By contrast, sliding wear initiated current transients and resulted in an inability to repassivate the surface. This effect was correlated with pitting in and near wear tracks, was associated with secondary austenite, and negated the effect that secondary-phase induced hardening had on reducing wear rates.

Page 29 of 29