The crevice corrosion repassivation potential of austenitic UNS S30400 and UNS S31600 stainless steels was determined in 0.1 mol/L and 1 mol/L NaCl solutions and in 5 mol/L CaCl2 solution at temperatures between 0°C and 90°C. The repassivation potential of UNS S30400 decreased with increasing chloride concentrations and temperatures in the range from 0°C to 60°C, reaching a constant value of −0.430±0.015 VSCE in the range from 60°C to 90°C. The repassivation potential of UNS S31600 showed a continuous decrease with increasing chloride concentrations and temperatures. Crevice-corroded spots showed crystalline attack plus pitting corrosion. Crystalline attack prevailed in UNS S31600 and pitting corrosion prevailed in UNS S30400. A solution of 0.5 mol/L HCl was used to simulate the crevice-like environment. Results of repassivation potential were analyzed in the context of Galvele’s localized acidification model. The ohmic potential drop was the dominant contribution in determining the repassivation potential for both steels. The contribution of the polarization to sustain the critical crevice was nil. The main effect of the 2.5 wt% Mo addition in UNS S31600 compared to UNS S30400 on the crevice corrosion resistance was the increase of the ohmic potential drop within the corroded area. This effect was especially significant for the more dilute chloride solutions and for increasing temperatures.
Stainless steels are construction materials of equipment in almost every industry in applications where corrosion resistance is required. These alloys do not rust in environmental conditions where carbon or low-allow steels rust because of the spontaneous development of a chromium-rich passive film on their surface. A minimum 11 wt% to 12 wt% chromium is necessary on the alloy composition to provide passivation.1-2 Stainless steels may be classified according to their metallurgical structure into ferritic, austenitic, martensitic, and duplex. Ferritic stainless steels are body-centered cubic like pure iron, while austenitic stainless steels are face-centered cubic as a result of additions of austenite-stabilizing elements such as nickel, manganese, and nitrogen. Martensitic stainless steels are obtained by rapid quenching from the austenite temperature range and they are body-centered tetragonal or cubic. Duplex stainless steels contain similar amounts by volume of austenite and ferrite phases. Precipitation hardened stainless steels may be considered as a distinct alloy family. They are martensitic and/or austenitic stainless steels, which may be heat-treated to provide higher tensile strengths.1-3
Although stainless steels do not rust in unpolluted atmospheres, they are prone to localized corrosion in the forms of pitting corrosion, crevice corrosion, and stress corrosion cracking (SCC) when exposed to chloride-containing solutions or marine atmospheres.4 Molybdenum and nitrogen may be added to stainless steels to improve their resistance to pitting and crevice corrosion promoted by chlorides.5-6 Pitting resistant equivalent with nitrogen (PREN) is the parameter commonly used to rank the localized corrosion resistance of stainless steels. PREN is defined in Equation (1) as a function of the weight percentages of Cr, Mo, and N.1,7 The localized corrosion resistance of stainless steels is also affected by their metallurgical condition. The presence of delta ferrite, chromium carbides, chromium-depleted zones, sigma, chi and Laves phases, sulfide inclusions, and product form variations may affect the corrosion performance of stainless steels in certain environments.1 Bulk PREN does not account for the metallurgical conditions of the stainless steels; however, partitioning of alloying elements to each alloy phase and/or depleted zone (if large enough) may be considered as having a different chemical composition and consequently a different PREN. Attributing different PREN to different alloy phases has been used for studying the effect of thermal aging on crevice corrosion of duplex stainless steels.8
Austenitic stainless steels of the 300 series are the most popular corrosion resistant alloys.1 In the nuclear industry, they are used to manufacture canisters for radioactive intermediate-level waste and for dry cask storage of spent nuclear fuel after removal from the cooling pools. They are also promising materials as metal waste forms for high-level waste.9-11 Spent nuclear fuel may need to be dry cask stored on-site for very long periods of time before final disposal (more than 100 y). Crevice corrosion is a possible degradation mechanism for stainless steels in dry cask storage systems. Its occurrence may lead to the initiation of chloride-induced SCC that is the most likely mechanism to produce the penetration of the canisters walls used as the confinement boundary.11 UNS S30400(1) is the classic 18%Cr-8%Ni austenitic SS, while UNS S31600 benefits from the addition of 2.5 wt% Mo, which increases its localized corrosion resistance.1 The low-carbon versions of these alloys are UNS S30403 and UNS S31603 stainless steels, respectively, and the low-carbon plus nitrogen versions are UNS S30453 and UNS S31653 stainless steels, respectively.1 UNS S30400, UNS S30403, UNS S30453, UNS S31600, and UNS S31653 are currently licensed for use as the confinement boundary of dry cask storage systems canister for nuclear waste in the United States.11 These types of stainless steels are especially susceptible to crevice corrosion in the presence of chlorides even at low (ambient) temperatures.12-18 Crevice corrosion occurs as a result of the local acidification of the crevice or occluded region. It is widely accepted that pitting and crevice corrosion of stainless steels are the same electrochemical phenomena.12-13,19-20 Using Galvele’s interpretation, chloride-induced crevice corrosion of stainless steels stabilizes easier at lower potentials than pitting corrosion, as the diffusion length associated with a crevice is longer than that associated with a pit.4,19 Laycock, et al., showed that crevice corrosion initiation on austenitic stainless steels is a result of the stabilization of metastable pitting within the crevices. The frequency of metastable pitting events increases near the value of the critical crevice temperature (CCT).21
A variety of testing techniques and devices has been used for studying the crevice corrosion of austenitic stainless steels. Crevicing devices and materials have evolved and it is recognized that not all crevices are equally severe or demanding on promoting crevice corrosion. Variables such as crevice former materials, surface finish of the crevice former and of the testing specimen, and applied pressure/torque to the crevice former may influence the outcome of the tests.17,22-23 Early localized corrosion studies on stainless steels used metal-to-metal, polytetrafluoroethylene (PTFE)-to-metal, and/or rubber band-to-metal crevices exposed to oxidizing and acidic solutions.18,24 In the literature, many types of crevicing devices and materials were used to study the susceptibility of stainless steels to suffer localized corrosion. For example, Oldfield and Sutton used metal-to-metal crevices of UNS S31600 with an 800 grit finish surface. One metal specimen was just laid on top of the other (in contact), producing an average gap of 10 μm resulting from surface roughness.25 The authors stated that an induction period precedes the initiation of active corrosion within the crevice. Smialowska and Mankowski used a cylindrical rod of polymethylmethacrylate (PMMA) as a crevice former. The flat end of the rod was loaded with a weight, giving an applied pressure of 500 kPa onto the specimen. The obtained results depended on the surface condition of the tested material.12 Dawson and Ferreira used a square piece of an undetermined plastic material as a crevice former for potentiodynamic tests on UNS S31600. Moreover, they performed electrochemical impedance spectroscopy (EIS) tests using a special glass cell clamped to the stainless steel specimen. In the latter case, the crevice former was a rubber gasket placed between the cell and the specimen. The crevice was controlled by the applied torque on the fastening bolts, though the torque value was not informed.13 Jakobsen and Maahn studied the crevice corrosion of UNS S31600 at different temperatures using a special electrochemical cell designed for crevice corrosion testing.15 The crevice was formed by a ring of polyvinylidene fluoride (PVDF) that was ground to 1000 grit finish. The crevice former was pressed on the specimen surface by a screw mechanism, though the applied pressure/torque was not informed. A CCT value of 5±1°C was reported for UNS S31600 using this device in NaCl solutions.15 Alavi and Cottis studied a single macroscopic crevice on UNS S30400 with a length of 8 cm, a width of 2.5 cm, and a gap of 90±10 μm. This experimental setup was instrumented to obtain the potential and pH evolution during the crevice corrosion study. However such a macrocrevice may be significantly different from metal-to-metal and polymer-to-metal microcrevices.26
In order to homogenize the testing parameters, the ASTM Standard G48, methods D and F, and ASTM Standard G78 describe crevice testing assemblies in a specimen that consist of a central hole, along with nut, bolt, washers, and two serrated crevice formers.27-28 The crevice corrosion testing devices currently used for determining critical potentials are generally based on this design with varying crevice former materials, surface finishes, and applied torques.22,29-31 These specimens are usually referred to as multiple crevice assembly (MCA) or prism crevice assembly (PCA). The ASTM Standard G192 describes a variation of the MCA specimen design.32 Malik, et al., performed crevice corrosion exposure tests on MCA specimens of several stainless steels in seawater. The surface finishes investigated were (1) wheel ground, (2) 180 grit, and (3) as-received. The crevice formers were made of PTFE and torqued to 9.5 N·m. Malik, et al., reported that UNS S30403 and UNS S31603 were susceptible to crevice corrosion in seawater even at 25°C.14 Bäck and Singh performed polarization tests on several stainless steels (including UNS S31603) used for equipment in pulp bleaching plants.16 They used MCA specimens with a surface finish of 400 grit and PTFE crevice formers torqued to 0.28 N·m. The specimen was suspended with a titanium wire, although no galvanic effects were reported. Temperature and potential were the main studied variables affecting localized corrosion in this chlorine dioxide bleach environment.16 Cai, et al., studied the effects of the applied torque and crevicing material on the crevice corrosion of UNS S31603 in artificial seawater.17 They used cyclic potentiodynamic polarization (CPP) method (ASTM G6133 ) to study PTFE-to-metal, fluoroelastomeric-to-metal, and metal-to-metal crevices using modified MCA specimens. The tested specimens and the teeth of the serrated crevice formers were wet-polished to a surface finish of 800 grit. The metal-to-metal crevice was the most aggressive and also insensitive to changes in the applied torque.17 Anderko, et al., used MCA specimens with PTFE crevice formers torqued to 0.14 N·m to study the crevice corrosion of several alloys, including UNS S31603.29,34 They determined the crevice corrosion repassivation potential by the CPP and the potential-staircase methods. The collected data were used to feed a mechanistic model able to predict the repassivation potential for stainless steels in a range of environments wider than the range of testing.
In the recent years, considerable research on crevice corrosion of Ni-Cr-Mo alloys has been performed, particularly for UNS N06022. The potentiodynamic-galvanostatic-potentiodynamic (PD-GS-PD)31 method, which is a modification of the Tsujikawa-Hisamatsu electrochemical (THE)32 method (ASTM G192), has led to the determination of reproducible and conservative repassivation potentials for UNS N06022. PD-GS-PD and THE methods avoid large current densities in the transpassive range of potentials and control the damage more efficiently that the CPP method (ASTM G61). The preferred crevicing devices were MCA/PCA specimens with a surface finish of 600 grit or smoother, along with PTFE-wrapped ceramic crevice formers torqued to no less than 3.4 N·m.22,32 The PTFE tape recommended for wrapping the crevice formers is 70-μm thick. The bolt, nut, and flat washer were made of a corrosion resistant material. The bolt was insulated from the specimen with a nonconductive sleeve.32 These testing conditions produce conservative and reproducible repassivation potentials for Ni-Cr-Mo alloys.22,30-31,35-36
In the present work, the authors assessed the crevice corrosion resistance of UNS S30400 and UNS S31600 austenitic stainless steels as a function of temperature and chloride concentration. The applied experimental techniques and devices were previously developed for the more corrosion resistant Ni-Cr-Mo alloys.22-23,35-37 A comparison is made between the current results and results from literature obtained by other testing techniques and crevicing devices.29 This work is meant to obtain quantitative parameters on the same basis for a proper comparison of the crevice corrosion resistance of stainless steels and Ni-Cr-Mo alloys. The results were analyzed within the context of Galvele’s localized acidification model to obtain insights on the crevice corrosion kinetics as previously done for Ni-Cr-Mo alloys.37 The fundamental equation of Galvele’s model can be adapted to crevice corrosion as indicated in Equation (2).37-38 The crevice corrosion repassivation potential (ER,CREV) is the addition of three terms: the corrosion potential of the alloy in the crevice-like solution (Ecorr*), a polarization needed to sustain the critical chemistry (η), and an ohmic potential drop (ΔΦ).19 The quantitative dependence of Ecorr*, η, and ΔΦ with the temperature for UNS S30400 and UNS S31600 was obtained from the present results. The effect of the 2.5 wt% Mo addition in UNS S31600 with respect to UNS S30400 is discussed in terms of Galvele’s model. The condition of a metal salt film to stabilize metastable pits within crevices is also discussed.
UNS S30400 and UNS S31600 specimens were prepared from wrought mill annealed plate stock. The specimens were heat-treated for 15 min at 1,050°C and then water-quenched to obtain full solubilization of any precipitates. The nominal chemical compositions of the alloys in weight percent are listed in Table 1. Prism crevice assemblies (PCA) specimens of approximate dimensions 19 mm × 19 mm × 9 mm were used. The specimens were fabricated based on ASTM Standard G192.32 The crevicing device contained 24 artificially creviced spots formed by two ceramic washers (crevice formers) wrapped with a 70-μm-thick PTFE tape. The applied torque was 5 N·m. The surface area exposed to the electrolyte solution was approximately 14 cm2. All of the specimens were ground using 600 grit abrasive paper and then degreased in ethanol and washed in distilled water. The surface preparation was performed 1 h prior to testing.
Electrochemical tests were conducted in a 1 L vessel. The testing solutions were 0.1 mol/L and 1 mol/L NaCl and 5 mol/L CaCl2 (chloride concentration: [Cl−] = 0.1, 1, and 10 mol/L). Nitrogen (N2) was purged through the solution 1 h prior to testing and it was continued throughout the entire test. A water-cooled condenser was used to avoid evaporation of the solution. The temperature of the solution was controlled by immersing the cell in a water bath, which was kept at a constant temperature using a cryothermostat. The set point temperatures (T) were 0°C, 10°C, 30°C, 60°C, and 90°C. The tests were performed at atmospheric pressure. The reference electrode was a saturated calomel electrode (SCE). The reference electrode was connected to the solution through a water-cooled Luggin probe. A flag of platinum foil spot-welded to a platinum wire was used as the counter electrode. The total area of the counter electrode was 50 cm2. All of the potentials in this paper are reported in the SCE scale. After the tests, the specimens were examined in an optical microscope and some of them were observed in a scanning electron microscope (SEM).
The crevice corrosion repassivation potential was determined by the PD-GS-PD method.22,31,35 Before each PD-GS-PD test, the open-circuit potential was measured for 15 min and afterward a cathodic current of 50 μA was applied for 5 min (pretreatment). The PD-GS-PD method consists of three stages: (1) a potentiodynamic polarization (at a scan rate of 0.167 mV/s) in the anodic direction until reaching an anodic current of 300 μA, (2) the application of a constant anodic current of IGS = 300 μA (approximately iGS = 22 μA/cm2) for 2 h, and (3) a potentiodynamic polarization (at 0.167 mV/s) in the cathodic direction, from the previous potential at the end of Stage 2 continuing until reaching alloy repassivation. Three or more tests (repeatability) were performed for each testing condition (chloride concentration and temperature).
The open-circuit or corrosion potential (Ecorr) of noncreviced prismatic specimens of UNS S30400 and UNS S31600 was measured in HCl solutions aiming to simulate the solution within active crevices. Anodic polarization tests were performed in HCl solutions at 90°C using a scan rate of 0.167 mV/s. The surface finish of the specimens and experimental setup were identical to those previously described for the crevice corrosion tests.
Figures 1 and 2 show PD-GS-PD tests for UNS S30400 and UNS S31600 in 0.1 mol/L NaCl at 0°C. The two tested stainless steels showed forward and reverse scans (Stages 1 and 3, respectively) with similar features. Plots for UNS S31600 showed an anodic shift in Ecorr of ~100 mV compared to those of UNS S30400 (Figure 1). Both stainless steels showed a large potential drop during the galvanostatic stage (Stage 2). Most of this potential drop in Stage 2 occurred within the first 200 s of galvanostatic polarization (Figure 2). The repassivation potential was defined as the potential value in the reverse scan (Stage 3) at which the current density dropped below 1 μA/cm2 without any further increase. The repassivation potential defined in this way is usually called ER1 and is shown in Figure 1. Other researchers have also used this criterion for studying similar systems (ER,CREV = ER1).29,34
All of the tested environmental conditions led to crevice corrosion of UNS S30400 and UNS S31600. Consequently, their respective CCT must be below 0°C for the tested chloride concentrations. The crevice corrosion damage for both materials in 0.1 mol/L and 1 mol/L NaCl occurred below the crevice formers borderline (footprint), although in some tests the attack spread out from the physically creviced area toward the specimen surface exposed to the bulk solution. More corrosion products were observed as the temperature increased. In all of the tests performed in 5 mol/L CaCl2, the attack spread out from the crevice formers footprint. It may be argued that the localized attack always started below the crevice former borderline: in 0.1 mol/L and 1 mol/L chloride solutions, the attack generally grew deeper at these locations, while in 10 mol/L solutions, the attack spread out from the crevice toward the noncreviced surface. Figure 3 shows a typical example of a crevice corroded UNS S31600 specimen after a PD-GS-PD test in 1 mol/L NaCl at 60°C, where the attack is observed under every creviced tooth of the multiple crevice former.
At 30°C and 60°C, crevice corroded spots on UNS S30400 showed abundant pits (Figures 4[a] and 5[a]), while crevice corroded spots on UNS S31600 showed crystalline type of attack plus some pits (Figures 4[b] and 5[b]). Crystalline attack in crevice corrosion, where different crystal planes corrode at different rates, is typical of Ni-Cr-Mo alloys.37 The pit density in the creviced area increased with temperature for both alloys from 30°C to 60°C (Figures 4 and 5). However, there were no significant changes from 60°C to 90°C (not shown). Pit density was lower for UNS S31600 than for UNS S30400. Pit coalescence was observed very frequently for UNS S30400, while it was scarce for UNS S31600. The 2.5 wt% Mo content of UNS S31600 was responsible for its significantly lower pit density compared with that of UNS S30400 for all of the tested conditions. Pit nucleation in the crevice corroded spots was apparently random (not associated with any metallurgical features like inclusions or grain boundaries).
Based on previous studies on crevice corrosion of Ni-Cr-Mo alloys37 and published research for pitting corrosion of stainless steels,39-40 0.5 mol/L and 1 mol/L HCl solutions were selected for testing as crevice-like solutions for UNS S30400 and UNS S31600. Then, the 1 mol/L HCl solution was discarded for giving very high Ecorr values and the 0.5 mol/L HCl was selected as representative of the crevice-like solution. Ecorr in simulated pit or crevice solutions is a strong function of pH and a weak function of chloride concentration.41 Ecorr of UNS S30400 and UNS S31600 was monitored for 2 h in deaerated HCl solutions at different temperatures. The values of potential in the last 10 min were averaged, obtaining a standard deviation lower than 2 mV. These average values of corrosion potential are represented as a function of temperature in Figure 6. Ecorr for UNS S30400 and UNS S31600 showed a slight increase (~30 mV) from 0°C to 90°C. Most of this increase occurred from 60°C to 90°C for UNS S30400, and from 0°C to 30°C for UNS S31600. Ecorr of UNS S31600 was higher than the Ecorr of UNS S30400 by 20 mV to 50 mV, depending on the temperature.
Ecorr* (the potential inside the localized corroded area) is an important parameter of Galvele’s localized acidification model, which has been found to be relatively insensitive to temperature changes.37 An average value of Ecorr* was calculated for UNS S30400 and UNS S31600. Table 2 shows the crevice-like solutions and the Ecorr* values of the stainless steels tested in the present work and the previously tested Ni-Cr-Mo alloys.37 These Ni-Cr-Mo alloys are much more corrosion-resistant than UNS S30400 and UNS S31600 as their respective PREN indicates. Consequently, the crevice-like solutions of UNS S30400 and UNS S31600 are more dilute and the Ecorr* values are hundreds of mV lower than those of Ni-Cr-Mo alloys. Newman and Ajjawi report Ecorr* = −0.400 VSCE for UNS S30400 in the saturated pit solution.42 This value is in agreement with the present results. Pits are expected to be slightly more acidic than crevices because of their higher current densities leading to higher acidification and, consequently, higher Ecorr* values.43 A value of 1 mol/L HCl is generally used to simulate pit-like solutions of 300 series stainless steels, while 0.5 mol/L was the selected crevice-like solution in the present work. Galvele, et al., reported Ecorr* = −0.550 VSCE for 18%Cr ferritic stainless steels and Ecorr* = −0.480 VSCE for 18%Cr-2%Mo ferritic stainless steels.39 Surface enrichment of nickel in austenitic stainless steels was likely to lead to the higher Ecorr* reported in the present paper compared to the corresponding ferritic stainless steels, though effects of minor alloying elements cannot be discarded. Molybdenum addition in UNS S31600 produced a further Ecorr* ennoblement with respect to UNS S30400 (Table 2).
Figure 7 shows SEM images of noncreviced UNS S30400 and UNS S31600 after 2-h immersions in 1 mol/L HCl at 60°C, in open-circuit conditions. The appearance of the alloys specimens’ bulk surfaces after testing in 0.5 mol/L and 1 mol/L HCl solutions was the same. Both stainless steels showed crystalline type of attack plus pitting. UNS S30400 showed more pits than UNS S31600. This type of attack resembles that observed in the crevice corroded spots (in NaCl solutions) of these materials. It is worth noting that the noncreviced specimens immersed in HCl solutions showed pits in spite of having suffered corrosion in the active state (no passive film). Crevice corroded spots (Figures 4 and 5) showed more localized attack than specimens tested in HCl solutions (Figure 7).
Figure 8 shows the anodic polarization curves of UNS S30400 and UNS S31600 in 0.5 mol/L HCl, at different temperatures. UNS S31600 showed lower current densities than UNS S30400 in all of the tested conditions reaching a passive or pseudo-passive state after an anodic peak. The current density of the anodic peak increased with the temperature. Galvele, et al., polarized molybdenum-containing ferritic stainless steels in HCl solutions.39 They stated that above −0.2 VSCE, localized corrosion entered into competition with pseudo-passivity. Their explanation may be also suitable for UNS S31600 (Figure 8). All of the tested specimens were severely pitted after polarization in HCl solutions. Current densities up to 100 mA/cm2 were observed in these tests. The molybdenum addition of UNS S31600 with respect to UNS S30400 not only improved the corrosion resistance of the material in the active range (low potentials), but also allowed for some degree of passivation at higher potentials.
Effects of Temperature and Chloride Concentration
Figures 9 and 10 show ER1 of UNS S30400 and UNS S31600 in chloride solutions as a function of temperature, respectively. ER1 decreased linearly with the increase of temperature in the range from the lowest tested temperature to 60°C for UNS S30400, and in the entire temperature range for UNS S31600. For UNS S30400, ER1 reached a constant value in the range from 60°C to 90°C regardless the chloride concentration (Figure 9). This minimum value is the corrosion potential in the crevice-like solution (ER1 = Ecorr* = −0.430±0.015 VSCE). ER1 for UNS S31600 is expected to remain flat at Ecorr* = −0.400±0.017 VSCE at temperatures above 90°C (Figure 10). Equation (3) was fitted to the data of both materials at the three tested chloride concentrations. Table 3 shows the fit parameters, where A and B are constants that depend on the tested stainless steel and chloride concentration.
ER1 for UNS S30400 decreased as temperature increased from 0°C to 60°C with slopes of −4.4 mV/K and −3.5 mV/K in [Cl−] = 0.1 mol/L and [Cl−] = 1 mol/L solutions, respectively. In [Cl−] = 10 mol/L solutions, the temperature dependence of ER1 was significantly lower (almost constant). ER1 was constant from 60°C to 90°C for the three tested chloride concentrations (Figure 9). ER1 for UNS S31600 decreased as temperature increased with slopes of −3.0 mV/K ([Cl−] = 0.1 mol/L), −2.5 mV/K ([Cl−] = 1 mol/L), and −0.7 mV/K ([Cl−] = 10 mol/L). The absolute value of the slopes (B in Table 3) decreased as the chloride solutions became more concentrated. This behavior is in agreement with that of Ni-Cr-Mo alloys in the same testing conditions.37 However, Ni-Cr-Mo alloys have higher temperature dependencies (4 mV/K to 9 mV/K) than the tested stainless steels in 0.1 mol/L and 1 mol/L NaCl solutions.37 Laycock and Newman reported that the pitting potential (EP) of UNS S30400 and UNS S31600 in 1 mol/L NaCl decreased linearly with increasing temperatures with slopes of −2.9 mV/K and −3.8 mV/K, respectively.44 In the present work, the slopes of −3.5 mV/K (UNS S30400) and −2.5 mV/K (UNS S31600) were obtained. The temperature dependence of critical potentials for pitting and crevice corrosion is not expected to be the same, as pitting corrosion involves much higher current densities, and consequently higher values of η and generally higher values of Δϕ than for crevice corrosion.
Effects of Testing Techniques and Crevice Formers
Figure 11 shows ER1 vs. [Cl−] fit and typical dispersion for UNS S31600 from the present work, obtained by the PD-GS-PD method with PTFE-wrapped ceramic crevice formers, and those of Anderko, et al., for UNS S31603, obtained by the CPP/potential-staircase methods with PTFE crevice formers.29 Other testing conditions between the present work and that of Anderko, et al., are the same. The crevice corrosion resistances of UNS S31600 and UNS S31603 are similar in the solution annealed condition. Present results are more conservative, especially for the higher temperatures and chloride concentrations. Chloride concentration dependence of ER1 is also affected by the testing method and crevicing material.37 More conservative repassivation potentials and a lower dependence of chloride concentration is also observed when comparing data for Ni-Cr-Mo alloys from PD-GS-PD method with PTFE-wrapped ceramic crevice formers vs. CPP/potential-staircase methods with PTFE crevice formers. In the case of Ni-Cr-Mo alloys, the quantitative differences among repassivation potentials and slopes are larger than those of the considered stainless steels.29,37 Open-circuit potentials of the tested stainless steels in seawater are typically higher than the ER1 values determined in the present work and, consequently, they are susceptible to crevice corrosion in this environment.45
Crevice Corrosion Attack and Salt Film Precipitation
Several features of the localized corrosion in crevice corroded spots (Figures 4 and 5) were similar to the corrosion in HCl solutions (Figure 7) for UNS S30400 and UNS S31600. The localized corrosion may be described as crystalline attack plus pitting corrosion. Crystalline attack prevailed in UNS S31600, while pitting corrosion prevailed in UNS S30400. Crevice corroded spots in UNS S30400 only showed some crystalline attack at the lowest tested temperature (not shown). The stabilization of the metastable pitting resulting from the crevice geometry is the process that leads to crevice corrosion initiation.21,46 According to Frankel, et al., metastable pitting of UNS S30200 occurs at potentials as low as −0.210 VSCE at room temperature.47 However, the pits in the crevice corroded spots of UNS S30400 and UNS S31600 are not necessarily the outcome of metastable pit growth. They may be etch pits, as they are also observed in acidic chloride solutions where the steels are not passive. Each pit occurring on an open passive surface has a cover of a porous lace-like structure7,47-48 which has not been observed in the HCl tests. The absence of this lace-like structure over the pits observed in this work does not rule out metastable pitting stabilization as the origin of crevice corrosion. It is likely that metastable pits developed at an initial stage, and once the acidified solution spreads from them, etch pits are thus formed. Metastable pits are stabilized by a porous pit cover which acts as a diffusion barrier.48 This cover is necessary to maintain a concentrated local chemistry inside the pit. When this pit cover collapses, the precipitation of a metal salt film is necessary for the pit survival in an open (noncreviced) surface.47,49 However, the very occlusion of the crevice geometry may be enough to stabilize pits at low current densities without metal salt precipitation.50 Present tests did not show corrosion products in the crevice corroded spots (Figures 4 and 5). A small amount of corrosion products was observed outside crevices, at 90°C. This observation may result from the current density limitation of the PD-GS-PD method. Absence of corrosion products was also reported for Ni-Fe-Cr alloys N06600, N06690, and N08800.51 Laycock and Newman reported artificial pit experiments in 50-μm-diameter wires of UNS S30200 stainless steels.44,49 They estimated the potential for the salt film precipitation (ET) from polarization tests in artificial pits as a function of the temperature.44 ET is defined as the potential at which the anodic current is equal to the diffusion-limited current but the surface is salt-free. ET is a transition potential between activation/ohmic control and diffusion control with a metal salt film and it has a good correlation with the pitting potential. Although the aspect ratio of stabilized pits within crevices in this work is unknown, it is useful to compare ET with the highest anodic potential in the current crevice corrosion tests. Figure 12 shows the maximum potential reached in the PD-GS-PD tests (EMAX) as a function of the temperature for UNS S30400 and UNS S31600. The determination of EMAX is shown in Figure 1. Potentials within crevices can be lower than EMAX as a result of ohmic drop effects, or equal to EMAX. In all of the tests performed in 0.1 mol/L NaCl solutions and in the tests performed in 1 mol/L NaCl solutions at 0°C, EMAX was higher than ET, although the potential within the crevices may be lower. In the tests performed in 1 mol/L NaCl solutions above 0°C, EMAX was in the same range as ET. Potential within crevices are expected to be below ET. In all of the tests in 5 mol/L CaCl2 solutions, EMAX was 0.1 V to 0.2 V below ET (Figure 12). These results suggest that a salt film precipitation may not be necessary for stabilizing single pits within crevices.
Galvele’s Localized Acidification Model and Crevice Corrosion Kinetics
Ecorr* of UNS S30400 and UNS S31600 was obtained from measurements in crevice-like solutions (Table 2). A simplifying assumption is necessary to obtain the parameters η and ΔΦ as a function of temperature. In a previous work on Ni-Cr-Mo alloys, the authors found that it is reasonable to assume ΔΦ = 0 for [Cl−] = 10 mol/L.37 Consequently, Equation (2) becomes ER1 = Ecorr* + η, and η can be easily obtained as the difference between ER1 for [Cl−] = 10 mol/L and Ecorr* (Equation ). η is independent of the chloride concentration of the bulk solution. ΔΦ can be calculated for [Cl−] = 1 mol/L and [Cl−] = 0.1 mol/L according to Equations (5) and (6), respectively. ΔΦ depends on the crevice geometry and present calculations are only valid for the type of crevice used in the present tests. Calculations of the term ΔΦ of Galvele’s equation based on the potential ohmic drop on a 1D crevice can be found elsewhere.52
Figure 13 shows η as a function of temperature for UNS S30400, UNS S31600, and selected Ni-Cr-Mo alloys. The continuous variation of η with temperature was obtained by combining Equations (4) and (3). The contribution of η to the repassivation potential for the tested stainless steels was low when compared to that of Ni-Cr-Mo alloys. For temperatures higher than 50°C, the contribution was nil considering the assumptions and errors of measurement involved for determining ER1 and Ecorr*. Figure 14 shows Δϕ for crevice corrosion tests in chloride solutions of [Cl−] = 1 mol/L as a function of temperature for UNS S30400, UNS S31600, and selected Ni-Cr-Mo alloys. The continuous variation of Δϕ with temperature was obtained by combining Equations (5) and (3). The contribution of Δϕ to the repassivation potential for the stainless steels was large at low temperatures, but it dropped steeply as temperature increased. Δϕ decreased more steeply for UNS S30400 than for UNS S31600. At some temperatures, Δϕ for UNS S31600 was comparable to Δϕ for UNS N06625 and UNS N06022, though all Ni-Cr-Mo alloys showed an overall better performance (higher ER1). Laycock and Newman made similar analyses, but for the pitting corrosion of austenitic stainless steels as a function of temperature.44 They report that Δϕ is the main contribution to the pitting potential but, unlike crevice corrosion, η is significant. This difference is reasonable as crevice corrosion occurs at lower potentials and with lower current densities than pitting corrosion. Consequently, the effect of η on crevice corrosion is low, as the critical current for sustaining the local acidic solution is reached at lower anodic polarizations.
Galvele’s Localized Acidification Model and the Effect of Alloy Composition
The main difference between UNS S31600 and UNS S30400 is the 2.5 wt% Mo addition of UNS S31600 with respect to UNS S30400. UNS S31600 showed higher ER1 values than UNS S30400 in all of the tested conditions (Figures 9 and 10). It is of interest to assess which parameters of Galvele’s model (Ecorr*, η, and/or Δϕ) contribute to the enhanced crevice corrosion resistance of UNS S31600. Figure 15 shows the difference between the corresponding parameters of Galvele’s model for UNS S31600 and UNS S30400 as a function of temperature. As stated earlier, an average positive difference of 30 mV in Ecorr* is observed for UNS S31600. The term η was not significant for either of the two materials (Figure 13), although it is slightly higher for UNS S31600 (Figure 15). η depends on the Tafel slope of the alloy in the crevice-like solution.19 Artificial pit studies in austenitic stainless steels suggest a slight increase in the anodic Tafel slope resulting from Mo additions.40 The main clear difference for UNS S31600 was observed for Δϕ. This difference increases for increasing temperatures and it is more pronounced for dilute chloride solutions (Figure 15). Alloyed molybdenum and molybdates are reported to be more beneficial to pitting corrosion resistance as the temperature increases.53 A comprehensive discussion of the effect of molybdenum on localized corrosion is out of the scope of this paper. However, the present analysis clearly shows that the main effect of the 2.5 wt% Mo addition of UNS S31600 with regard to UNS S30400 on crevice corrosion was the increase of the term Δϕ of Galvele’s model, which accounts for the ohmic potential drop. Molybdenum oxides are highly insoluble in the acidic solutions typical of localized corrosion processes; therefore, these oxides may contribute to stifling the propagation to localized corrosion as a physical barrier inside the corroded areas.
SUMMARY AND CONCLUSIONS
The resistance of UNS S30400 and UNS S31600 to chloride-induced crevice corrosion was assessed in a wide range of temperatures and chloride concentrations. Results were analyzed in the context of Galvele’s localized acidification model.
The repassivation potential of UNS S30400 decreased with increasing chloride concentrations and temperatures in the range from 0°C to 60°C, and it reached a constant value in the range from 60°C to 90°C regardless of chloride concentration. This minimum and constant repassivation potential value for UNS S30400 was the corrosion potential of UNS S30400 in the crevice-like solution (−0.430±0.015 VSCE). The repassivation potential of UNS S31600 decreased with increasing chloride concentrations and temperatures in the entire tested temperature range (0°C to 90°C). The repassivation potential of UNS S31600 is expected to level at −0.400±0.017 VSCE when it reaches the corrosion potential of UNS S31600 in the crevice-like solution.
Crevice-corroded areas on UNS S30400 and UNS S31600 showed crystalline attack plus pitting corrosion. Crystalline attack prevailed in UNS S31600, while pitting corrosion prevailed in UNS S30400. The appearance of UNS S30400 and UNS S31600 noncreviced specimens tested in HCl solutions was similar to the crevice corroded material in near-neutral salt solutions. As pits developed in the specimens exposed to HCl solutions where they suffered active corrosion (not in a passive state), it may be argued that they were etch pits in both cases. Crevice corrosion was able to initiate in both UNS S30400 and UNS S31600 in conditions where the potential associated with salt film precipitation in artificial pits was not exceeded. This observation suggests that the occluded crevice geometry itself is sufficient to stabilize metastable pitting.
The ohmic potential drop was the main contribution controlling the repassivation potential for UNS S30400 and UNS S31600. The contribution of the polarization to sustain the critical chemistry was nil. The main effect of the 2.5 wt% Mo addition in UNS S31600 compared to UNS S30400 was the increase in the ohmic potential drop inside the corroded area. This effect was most significant for dilute chloride solutions and for increasing temperatures. After 40 years of the appearance of the seminal paper of the localized acidification model, it continues to give insights on localized corrosion and rationalizing new knowledge.19
UNS numbers are listed in Metals and Alloys in the Unified Numbering System, published by the Society of Automotive Engineers (SAE International) and cosponsored by ASTM International.
Financial support from the Argentine Agencia Nacional de Promoción Científica y Tecnológica of the Ministerio de Ciencia, Tecnología e Innovación Productiva is acknowledged.