Low-Temperature Stress Corrosion Cracking of Stainless Steels in the Atmosphere in the Presence of Chloride Deposits

CORROSION SCIENCE SECTION

CORROSIONVol. 65, No. 2 105

Submitted for publication July 2008; in revised form, October 2008. Presented as paper no. 08484 at CORROSION/2008, March 2008, New Orleans, LA.

Corresponding author. E-mail: [email protected]. * Institut de la Corrosion/French Corrosion Institute, 220 rue

Pierre Rivoalon, 29200 Brest, France. ** Outokumpu Stainless AB, Avesta Research Center, PO Box 74,

SE-774 22 Avesta, Sweden. *** Swerea KIMAB AB, Drottning Kristinas vg 48, SE-114 28 Stock-

holm, Sweden.

Low-Temperature Stress Corrosion Cracking of Stainless Steels in the Atmosphere in the Presence of Chloride Deposits

T. Prosek,,* A. Iversen,** C. Taxn,*** and D. Thierry*

AbstrAct

Several cases of ceiling collapses and other failed elements have been reported in indoor swimming pool halls in the last two decades. The collapses were caused by stress corrosion cracking (SCC) of stainless steel fastening elements covered with chloride deposits at temperatures as low as room tem-perature. The goal of this study was to assess the applica-tion limits of different austenitic and austenitic-ferritic (duplex) stainless steels subject to tensile stress and contaminated with chloride deposits in atmospheric non-washing conditions as a function of temperature (20C to 50C), relative humidity (15% to 70% RH), and deposit composition. Austenitic stainless steels Type 304 (UNS S30400) and Type 316L (UNS S31603) were susceptible to SCC in the presence of magnesium and calcium chlorides at temperatures of 30C and higher and at low relative humidity. The tendency to SCC increased with increasing temperature and decreasing relative humidity. The corrosivity of chloride deposits under given exposure condi-tions decreased in the following order: calcium chloride (CaCl2) > magnesium chloride (MgCl2) > sodium chloride (NaCl). It was governed by the equilibrium chloride concentration in the sur-face electrolyte formed as a result of interaction of a given salt with water vapor in the air. Threshold values of the minimum chloride concentration and relative humidity intervals leading to SCC were established for Type 304 and Type 316L. Duplex stainless steels S32101 (UNS S32101), 2304 (UNS S32304),

2205 (UNS S32205), and 2507 (UNS S32750) were resistant to SCC but corroded selectively with the maximum depth of 200 m. Austenitic stainless steels Type 904L (UNS N08904) and Type S31254 (UNS S31254) showed no tendency to SCC.

Key woRDS: atmospheric corrosion, chloride stress corrosion cracking, duplex stainless steel, stainless steel

IntroductIon

Common austenitic stainless steel grades of 300 series have been widely applied in the building, chem-ical, engineering, and other industries for many decades. The corrosion properties of these materials are known and they serve well when properly used, since there is an understanding about the reactions and processes fundamental for material selection.

Austenitic stainless steels are also used in swim-ming pools for three main types of applications: the pool water conditioning plant, the equipment immersed in pool water, and the equipment exposed to swimming pool atmospheres. In general, there is no critical corrosion problem connected with the former two categories.1 However, problems occurred with the equipment exposed to swimming pool atmospheres, particularly with safety-relevant accessories subject to tensile stress. Several cases of ceiling collapses and other element failures were reported to have occurred in swimming pool halls.2-9 In all reported cases, failed parts, such as fasteners, wires, bars, strapping, bolts, hooks, hangers, and other fittings, were made of 300 series austenitic stainless steels. Acidic chloride deposits were found on the surface of damaged stain-less steels, with typical pH values of 3 to 4.6-8 Consid-

ISSN 0010-9312 (print), 1938-159X (online)09/000019/$5.00+$0.50/0 2009, NACE International


106 CORROSIONFEBRUARY 2009

ering cations, relatively large quantities of calcium, as well as magnesium, aluminum, and zinc, were often detected.7-11 The deposits were hygroscopic and formed water solutions even in the air with low rela-tive humidity. Metallographic analyses revealed that the collapses were caused by stress corrosion crack-ing (SCC). The cracks were transgranular and branched. In most cases, SCC was observed in con-junction with pitting.9,12 All failed parts were princi-pally in the cold-worked condition and/or were subject to an additional stress or work hardening on installation or in service.2,8

It should be noted that temperature in the respective swimming pool buildings was about 30C, which contradicts the general corrosion literature on SCC of austenitic stainless steels that considers this phenomenon to occur at temperatures exceeding 50C to 60C. Recent corrosion studies disclosed that there are two specific conditions leading to low-tem-perature SCC. First, solutions with chloride concen-trations from 0.5 M to 5 M and pH from 0.5 to 0.5 were found to cause SCC of austenitic stainless steels at temperatures below 60C.2,6,8,10,13-15 Since stainless steels under these conditions are in the active/pas-sive potential region,2,6,10,12,15-16 this mode of SCC was accompanied by considerable general corrosion and corrosion rates were typically a few mm per year. Sec-ond, SCC was observed in environments with high surface chloride concentrations greater than 20 wt%. Such solutions can be formed on a metal surface in the presence of salts with sufficient solubility, such as calcium or magnesium chlorides exposed to air at the relative humidity (RH) close to the point of deliques-cence.12,16-20 High chloride concentration is the only prerequisite for inducing this type of SCC, and acidic conditions are generally not necessary. Under these conditions, stainless steels are probably in a passive state.7,10,12,16 Results of examinations of failed stain-less steel segments indicate that the latter mechanism of SCC, i.e., the passive type, is more likely associ-ated with the conditions in swimming pool atmo-spheres.10,12,16 In addition, the former type of SCC was observed only under full immersion conditions and has never been simulated in the atmosphere.

Thus, non-washing conditions and the accumu-lation of highly soluble chloride deposits on the metal surface is necessary for initiation of the passive-type, low-temperature SCC. It has been postulated that vol-atile, chlorine-based products of water disinfectants such as chloramines play an important role in the transport of Cl to the metal surface. However, it has been proven by various laboratory experiments that the presence of chloramine or another oxidizer is not essential for initiation of this type of SCC12,16,21 and it has been observed, e.g., in plants operating in coastal areas.22-23 This study focused on the application limits of selected austenitic and austenitic-ferritic (duplex) grades of stainless steel submitted to stress under atmospheric non-washing conditions in terms of SCC and pitting. The corrosion resistance of U-bend speci-mens with deposited droplets of magnesium, calcium, and sodium chloride (NaCl) was evaluated as a func-tion of temperature and relative humidity at levels selected to represent typical and limiting application conditions in indoor environments.

ExpErImEntAl procEdurEs

MaterialsEight austenitic and duplex stainless steel grades

were included in the investigation. A list of tested stainless steels with a typical chemical composition and pitting resistance equivalent (PRE) values cal-culated using the formula PRE = %Cr + 3.3%Mo + 16%N is given in Table 1. U-bend specimens were pre-pared according to ASTM G30-94.24 Specimens were cut from pickled sheets longitudinally to the direction of rolling. The metal strip width was 13 mm and the strip thicknesses are listed in Table 1. Total strain on the outside of the bend was calculated according to an approximate equation = T/2R, where T is thick-ness and R is radius of the bend (10 mm). The speci-mens were used as-received and no other finish was applied.

Experimental SetupChloride deposits on stainless steel specimens

were formed according to a modified procedure devel-

tABlE 1Stainless Steel U-Bend Specimens

Microstructure thickness total AStM EN UNS (PRE) Cr Ni Mo N Other (mm) Strain 304 1.4301 S30400 Austenite (18) 18 8.4 0.1 0.04 2.9 0.15 316L 1.4404 S31603 Austenite (24) 18 11.4 2.1 0.04 2.9 0.15 904L 1.4539 N08904 Austenite (30) 20 24.7 4.5 0.04 1.5% Cu 1.9 0.10 S31254 1.4547 S31254 Austenite (43) 20 18.0 6.1 0.20 Cu 2.0 0.10 S32101(1) 1.4162 S32101 Duplex (26) 22 1.5 0.3 0.23 5% Mn 1.4 0.07 2304 1.4362 S32304 Duplex (26) 22 4.8 0.3 0.09 2.0 0.10 2205 1.4462 S32205 Duplex (35) 22 5.5 3.0 0.12 2.0 0.10 2507 1.4410 S32750 Duplex (43) 25 7.7 4.0 0.29 3.0 0.15(1) LDX 2101.

Stainless Steel Grade typical Chemical Composition (wt%)



oped by Shoji and Ohnaka,17 which was used later by other laboratories.8,18,20 Six droplets of saturated chlo-ride salt solution were deposited on the top of a U-bend specimen. The deposits were separated in six spots to allow evaluation of the probability of crack formation. Figure 1 shows a photograph of the U-bend specimen with freshly deposited droplets. The four droplets at the sides were smaller than the two central ones because the droplets on angling sides tended to drop down when their size surpassed a certain limit. All together, 662 L of the solution was applied on each specimen. The total covered area was approxi-mately 90 mm2. Details of the chloride surface con-centration are given in Table 2.

The specimens with deposited chlorides were exposed in climatic chambers at 20, 30, and 40C and 30, 50, and 70% RH for 10 weeks (1,680 h). Two series of specimens were exposed also for 4 weeks. The accuracy of humidity regulation in the cham-bers was 2%. The experiments started from the most detrimental conditions expected from the results of previous studies, i.e., at 40C and 30% RH. When it was clear from the acquired results that the follow-ing exposure conditions would be benign, the testing matrix was reduced. In addition, U-bend specimens contaminated with magnesium chloride (MgCl2) and calcium chloride (CaCl2) were tested at an elevated temperature of 50C over saturated solutions of the respective salt, keeping the relative humidity in the test chambers close to the deliquescence point. Rela-tive humidity was registered and it was 305% and 153%, respectively. Two sets of specimens were tested for 22 weeks and one set for 4 weeks.

After the exposure, photographs were taken (Fig-ure 2) and the specimens were cleaned in water and examined with an optical microscope at a magnifica-tion of 16X. Some of them were evaluated with a con-focal microscope to establish the depth of pits and etched areas. Selected specimens were cut in the lon-gitudinal direction, embedded into resin, polished, and examined with a metallographic microscope at the cross section. When needed, grain boundar-ies were highlighted by etching in 10% oxalic acid (H2C2O4) at 6 V for 5 s to 60 s.

rEsults And dIscussIon

Corrosion of Austenitic Stainless SteelsResults of all experiments are summarized in

Table 3, and the performance of the austenitic grades as a function of exposure conditions is presented in Figure 3. A detailed description of the degradation is given elsewhere.25-26 Only principal results and trends are presented in the following paragraphs.

Stress corrosion cracks were found in specimens of stainless steel grade Type 304 (UNS S30400)(1) exposed with MgCl2 and CaCl2 deposits at 50C/30% RH, 40C/30% RH, 40C/50% RH, and 30C/30% RH. Photographs of the Type 304 specimen prone to SCC before and after cleaning is shown in Figures 2 and 4. Similar observations were made for stainless steel grade Type 316L (UNS S31603), but the extent of SCC was lower and cracks were finer. In compari-son to Type 304, SCC did not initiate at 40C/50% RH and only calcium chloride was able to cause crack-ing at 40C/30% RH and 30C/30% RH. Morphol-ogy of SCC was studied with an optical microscope on selected cross-cut specimens. Photographs in Figure 5

(1) UNS numbers are listed in Metals and Alloys in the Unified Num-bering System, published by the Society of Automotive Engineers (SAE International) and cosponsored by ASTM International.

tABlE 2Surface Contamination with Droplets

of Saturated Chloride Solutions Chloride Concentration Applied Chloride in Saturated Chloride Concentration Salt Solution (g/l) (mg) (g/m2) MgCl2 350 23 260 CaCl2 400 26 290 NaCl 190 13 140

Metal Surface

FIGURE 1. U-bend specimen with chloride droplets.

FIGURE 2. Deposits of CaCl2 on U-bend specimen of stainless steel Type 304 after exposure at 40C and 30% RH for 10 weeks.



make it well apparent that the cracks were branched and transgranular as typical for SCC of stainless steels in chloride environments.

There was practically no difference in the number of chloride spots with cracks on specimens exposed

for a shorter (4 weeks) and longer (10 or 22 weeks) period of time under identical exposure conditions. Indeed, the depth of the attack increased in time. It suggests that SCC initiated rapidly, probably within the first days of exposure. This is supported by visual

tABlE 3Corrosion Degradation of U-Bend Specimens Exposed with Chloride Deposits to Humid Air

Exposure Salt t (C) RH (weeks) 304 316l 904l S31254 S32101 2304 2205 2507 MgCl2 50 ~30 4 SCC 6 SCC 6 PIT PIT PIT PIT PIT PIT CaCl2 50 ~15 4 SCC 3 SCC 2 ET ET PIT ET ET ET MgCl2 50 ~30 22 SCC 6 SCC 6 PIT PIT PIT ET PIT PIT CaCl2 50 ~15 22 SCC 5 SCC 4 PIT PIT PIT PIT PIT PIT MgCl2 40 30 4 SCC 2 PIT NO NO PIT ET ET PIT + CaCl2 40 30 4 SCC 6 SCC 6 PIT PIT PIT ET ET PIT + MgCl2 40 30 10 SCC 2 PIT PIT NO ET ET ET NO CaCl2 40 30 10 SCC 6 SCC 6 PIT PIT ET ET ET PIT + NaCl 40 30 10 NO NO NO NO NO NO NO NO MgCl2 40 50 10 SCC 5 PIT PIT PIT ET PIT + PIT ET CaCl2 40 50 10 SCC 4 PIT PIT PIT ET PIT + PIT ET NaCl 40 50 10 NO NO NO NO NO NO NO NO MgCl2 40 70 10 ET NO NO NO ET NO NO NO CaCl2 40 70 10 ET NO NO NO ET NO NO NO NaCl 40 70 10 NO NO NO NO NO NO MgCl2 30 30 10 SCC 2 PIT PIT ET ET ET CaCl2 30 30 10 SCC 6 SCC 6 PIT ET PIT ET MgCl2 30 50 10 PIT NO NO PIT PIT ET CaCl2 30 50 10 PIT PIT ET PIT + PIT ET MgCl2 30 70 10 ET NO NO NO NO NO CaCl2 30 70 10 ET NO NO ET NO NO MgCl2 20 30 10 PIT PIT PIT ET PIT PIT CaCl2 20 30 10 PIT PIT PIT ET PIT PIT MgCl2 20 50 10 PIT PIT NO PIT + ET ET CaCl2 20 50 10 PIT PIT NO PIT + ET ET

SCC #: stress corrosion cracks found in a given number of six contaminated zonesPIT: pittingET: surface etchingNO: no corrosion: not tested+: tiny cracks

Exposure Conditions Stainless Steel Grade (UNS)

(a) (b)FIGURE 3. Summary of corrosion deterioration of austenitic stainless steels with chloride deposits exposed to air as a function of temperature, relative humidity, and grade; (C) 304, (#) 316L, () 904L, () S31254; (black) SCC, (gray) pitting corrosion, (empty symbol with thick border) etching, and (empty symbol with thin border) free of corrosion.



observations of specimens exposed at 50C where the first cracks were visible already after 7 days.

The systematic study at 20, 30, and 40C and 30, 50, and 70% RH with deposits of NaCl, MgCl2, and calcium chloride (CaCl2) clearly showed that the ten-dency of a given stainless steel grade to develop SCC depends on temperature. In the presence of MgCl2 deposits, stainless steel grade Type 316L was prone to SCC at 50C and no cracks developed at 40C and below. It cracked at 50, 40, and 30C in CaCl2. Lower alloyed austenitic stainless steel Type 304 was prone to SCC at 50, 40, and 30C in both environments and it was resistant at 20C.

The extent of corrosion diminished with increas-ing relative humidity. As seen in Figure 3, stainless steel Type 304 with CaCl2 and MgCl2 deposits exposed to air at 40C was prone to SCC at RH of 30% and 50%, whereas it was only superficially etched at 70% RH. At 30C, it was the RH of 30% that led to SCC, pits were found at 50% RH, and the specimens were only slightly etched when exposed to air at 70% RH. A similar trend was observed for Type 316L. In any case, no pitting or SCC was observed at the highest RH of 70%.

Considering the effect of chloride salts at iden-tical relative humidity and temperature, it is obvi-ous that deposits of CaCl2 were the most aggressive. The number and size of cracks and the extent of pit-ting was lower on surfaces contaminated with MgCl2. Specimens with NaCl spots were free of corrosion at all exposure conditions. The same observation was reported by Oshikawa, et al., who studied pitting and

SCC of Type 304 in the presence of different chlorides at RH from 33% to 75%.27

Cracks were found in all cases in the two center spots, while it was only pits with a typical depth of 10 m to 40 m that were observed in some of the side spots at less-aggressive conditions. Since the stress varies from zero at the ends to the maximum in the center of the U-bend specimen, the side areas rep-resent places with lower stress compared to the cen-tral zone. Thus, a higher stress level was needed for the SCC initiation under less-aggressive conditions.

It was also noticed that cracks in both Type 304 and Type 316L often came through pits under less-aggressive conditions. Although the pits were shal-low, they probably served as stress raisers facilitating the initiation of SCC. This was observed particularly

FIGURE 4. U-bend specimen of Type 304 after exposure with CaCl2 at 40C and at 30% RH for 10 weeks.

FIGURE 5. Morphology of SCC of Type 304 stainless steel exposed with MgCl2 at 50C and at 30% RH for 4 weeks; width 430 m (left) and 160 m (right).

(a) (b)



on specimens of the more resistant grade Type 316L. A cross section of the Type 316L specimen exposed at 50C and 30% RH showing a shallow pit with a depth of 40 m and width of 200 m with cracks is shown in Figure 6. It indicates that pits as stress collectors enabled the initiation of SCC of more alloyed mate-rials or under less-aggressive conditions, whereas cracks could have been initiated on a free surface of lower alloyed stainless steel Type 304 exposed to higher temperature and lower relative humidity.

Superaustenitic stainless steel grades Type 904L (UNS N08904) and UNS S31254 were fully resis-tant to SCC. The specimens were usually pitted when exposed to lower relative humidity. The extent of cor-rosion deterioration increased with time of exposure and temperature. Metallographic examination and 3D profiling showed that pits were rather shallow with the depth ranging from 50 m to 100 m (Figure 7). Instead of pitting, small etched areas with the depth below 20 m were found under certain exposure con-ditions (Table 3, Figure 3). Etching was observed also along with pits. It is possible that pits would develop

there in time. No significant difference was seen in the performance of Type 904L and S31254.

The presented results agree well with the avail-able literature data. The experiments of Shoji and Ohnaka17 carried out at room temperature showed that grades Type 304 and Type 316L were prone to SCC when contaminated with MgCl2 and CaCl2 and exposed to RH of about 33% to 45% and 21% to 45%, respectively. Except for the results of Type 316L with MgCl2 deposits, it is in agreement with the results of this study, provided that the room temperature in the experiments of Shoji and Ohnaka was somewhat higher than 20C, e.g., 25C. Arlt, et al.,16 studied the SCC above saturated solutions of magnesium chlo-ride, i.e., at RH of about 30%. The results at 30, 40, and 50C for Types 304, 316 (UNS S31601), and 904L exposed for a comparable period of time correspond to those given in Figure 3. Fairweather, et al.,22 used a different experimental setup, i.e., C-ring specimens of stainless steel grade Type 304 contaminated with MgCl2 at about 2 g/cm

2 and 10 g/cm2 of chloride. The cracking frequency was 12% at 30C and 45%

(a) (b)FIGURE 6. SCC on Type 316L exposed with MgCl2 deposits at 50C and at 30% RH for 22 weeks; left: surface, width 4.3 mm; right: cross section, width 650 m.

(a) (b) (c)FIGURE 7. Pit morphology on UNS S31254 exposed with MgCl2 at 50C and at 30% RH for 4 weeks: (a) surface, width 800 m, (b) 3D depth profile of the same area, and (c) cross section, width 250 m.



RH. In this study, SCC was found at 30C and 30% RH but not at 30C and 50% RH. The data obtained by Fairweather, et al., at 45C are difficult to compare to our results because most experiments in this work were done at temperatures up to 40C.

Equilibrium Chloride Concentration as a Function of Temperature and Relative Humidity of Air

The results revealed that the tendency to SCC and pitting depends, beside temperature and stain-less steel grade, on relative humidity and the type of applied chloride salt. It is assumed that the aggres-siveness of chloride salt deposits is predominantly controlled by the activity of chloride ions in the sur-face solution formed in equilibrium with water vapor in air. Salt interacting with humid air tends to attract water from the atmosphere and a saturated salt solu-tion is formed at a certain level of relative humid-ity, the deliquescence point. Chloride concentration drops with increasing relative humidity as the solu-tion attracts more water from air and the volume of the solution increases. If the relative humidity of air is lower than the deliquescence point of a particular salt, the deposit is dry. Since the activity coefficient of chloride in solutions corresponding to the conditions of this study was not available, chloride concentra-tions were calculated and used for the estimation of corrosion properties of electrolytes on the metal sur-face instead of the activity.

Solubility of the chloride salts was calculated based on earlier literature data28 and it is plotted as a function of temperature in Figure 8. Solubility val-ues of NaCl and MgCl2 are very close within the whole temperature interval. However, chloride concentration in water solution is twice as high in the case of MgCl2 as a result of the presence of two chloride atoms in the molecule. Steep changes in solubility of CaCl2 cor-respond to the fact that different phases become sta-ble with increasing temperature, i.e., CaCl26H2O, CaCl24H2O (), and CaCl22H2O. Again, it should be noted that calcium chloride solution contains twice as many chloride ions as NaCl at the same salt con-centration.

To calculate the point of deliquescence and dependence of chloride concentration on relative humidity, the osmotic coefficient for each salt must be known as well. The osmotic coefficient varies with the concentration and temperature. The following equa-tion shows the relation between water activity, aW, osmotic coefficient, F, and concentration, m, in moles of salt per kilogram of water:

ln

,aw =

m MW1 000

(1)

In the equation, is the number of ions that the salt formally dissociates into, i.e., 2 for NaCl and 3 for CaCl2 and MgCl2. MW is molar weight of water. The

FIGURE 8. Solubility of chloride salts as a function of temperature.

FIGURE 9. Osmotic coefficient of water solutions; full curve: literature data, dashed curve: extrapolated data.

water activity has the same significance as the relative humidity.

Values of osmotic coefficients vary strongly between salts. Reliable data were obtained at 25C.29-30 Data at higher temperatures were found only for NaCl, where the temperature dependence of the osmotic coefficient is negligible. Therefore, the osmotic coeffi-cients were considered to be independent of tempera-ture for these calculations. Since the osmotic coeffi-cients were available only up to the concentrations of about 4.5, 6, and 10 mol/kg water for NaCl, MgCl2, and CaCl2, it was necessary to extrapolate the litera-ture data. Literature and extrapolated values of the osmotic coefficient are plotted in Figure 9 in solid and dashed curves, respectively.

The calculated equilibrium chloride concentra-tions in water solutions of NaCl, MgCl2, and CaCl2 as a function of the relative humidity of air are given in Figure 10 and Table 4. The calculated deliquescence points of concerned chloride salts at 20, 30, 40, and 50C and corresponding chloride concentrations are in Table 5.



Tendency to Stress Corrosion Cracking in View of the Equilibrium Chloride Concentration

Since the equilibrium chloride concentration on the surface of specimens contaminated with MgCl2 and CaCl2 was calculated for all relevant climatic con-ditions, it was possible to connect it to the occurrence of SCC and pitting corrosion. Charts in Figure 11 summarizing results for materials contaminated with MgCl2 and CaCl2 and exposed to 20, 30, and 40C and to 30, 50, and 70% RH demonstrate that the equilib-rium chloride concentration was crucial for the cor-rosiveness of the deposits in contact with air. CaCl2 solutions contained 20, 10, and 8% more chloride than MgCl2 solutions at the same relative humidity of 30, 50, and 70%, respectively, and the former depos-its were clearly more corrosive at temperatures from 20C to 40C (Table 3 and Figure 3).

Lower aggressiveness of CaCl2 in terms of a lower number of chloride spots with cracks (Table 3) in the experiments over saturated solutions of correspond-ing salts at 50C is somewhat surprising in light of these facts. Under these conditions, concentration of the solution formed in the presence of CaCl2 should be almost double compared to MgCl2 in an equilib-rium state according to data in Table 5. It is possible that the actual relative humidity in the chambers dif-fered from the equilibrium one, probably due to kinet-ics reasons. These data are not included in Figure 11.

Calculated deliquescence points listed in Table 5 show that the relative humidity of 30% was too low to dissolve MgCl2 and CaCl2 deposits at 20C and MgCl2 deposits at 30C. Despite that, pitting or even SCC developed under these conditions. It was shown by other authors31 that chloride salts are corrosive even at relative humidity slightly below the deliques-cence point. It is supposed that there is enough water absorbed in salts to provide electrochemical condi-tions close to those in saturated water solution. It implies that the effective chloride concentration in a deposit exposed to air at the relative humidity only slightly below the deliquescence point might be com-parable to that of the appropriate saturated solu-tion. Therefore, the saturated solution concentrations were used in charts presented in Figure 11 to describe these experimental conditions as well as in further considerations.

NaCl deposits should not dissolve at any of the tested relative humidity levels. The deliquescence point of NaCl is 75% RH in the whole range of tested temperatures. As discussed above, it can be assumed that at 70% RH there was enough water absorbed in NaCl deposits to make them corrosive. However, stainless steel specimens were free of corrosion. It may suggest that even the maximum chloride concen-tration in NaCl solution of 6.1 mol/kg to 6.2 mol/kg

FIGURE 10. Calculated equilibrium chloride concentration in water solutions of different salts as a function of the relative humidity of air; horizontal lines indicate solubility limits at 20, 30, 40, and 50C (from top).

tABlE 4Calculated Values of the Chloride Equilibrium Concentration in Salt Solutions in Contact with Air at Given Relative Humidity

and at Temperatures from 20C to 50C; in mol Cl/kg H2O Salt 10% 20% 30% 40% 50% 60% 70% 80% NaCl NA NA NA NA NA NA NA 5.2 MgCl2 NA NA 12.0+ 10.5 9.1 7.7 6.3 4.8 CaCl2 NA 18.8+ 14.4+ 11.8 10.0 8.4 6.8 5.1

NA: no solution is formed+: soluble only in a part of the temperature range

tABlE 5Calculated Deliquescence Points (DP) and Corresponding Chloride Concentrations for Salts in Equilibrium with Air

(mCl, in mol Cl/kg H2O)

Salt DP mCl DP mCl DP mCl DP mCl NaCl 75 6.1 75 6.2 75 6.2 75 6.4 MgCl2 34 11.4 32 11.7 30 12.0 28 12.4 CaCl2 33 13.6 22 17.5 17 20.8 13 23.8

20 C 30 C 40 C 50 C



H2O is too low to cause any serious corrosion dete-rioration to the tested stainless steels in atmospheric conditions at temperatures up to 40C.

It is apparent from Figure 11 that the threshold chloride concentrations defining conditions leading to SCC, pitting, and no corrosion can be defined for each material and temperature. For current experimental conditions, they are listed in Table 6.

Application Limits of Austenitic Stainless Steels in Atmospheric Conditions

The results allow us to define relative humidity intervals for each temperature and salt when SCC of stainless steels Type 304 and Type 316L might be ini-tiated. These ranges summarized in Table 7 can be understood as the application limits of stainless steel grades Type 304 and Type 316L in non-washing con-ditions where chloride deposits can be formed and some level of tensile stress is present. It was assumed that at a relative humidity more than 5% lower than the deliquescence point, there is no risk of SCC and the deposits are harmless. When relative humidity exceeds a certain limit, chloride concentration in the surface solution will be too low to initiate SCC. At 20C, the chloride concentration needed for the initia-tion of SCC is higher that the solubility of MgCl2 and CaCl2, and there is no danger of SCC under these deposits. At 30C, SCC can develop when the chloride concentration in water solution is above 10 mol/kg and 12 mol/kg H2O for Type 304 and Type 316L, respectively. Thus, stainless steel Type 304 can suffer from SCC when contaminated with MgCl2 and CaCl2 deposits and exposed to relative humidity below 44% and 49%. To form similarly corrosive environments for Type 316L, CaCl2 deposits must be present on the surface and the relative humidity must drop below

40%. Intervals of the relative humidity causing SCC of the austenitic stainless steel grades at 40C are logi-cally wider.

Mayuzumi, et al., showed that suppression of SCC of low-alloyed austenitic stainless steel grades Type 304 and Type 304L (UNS S30403) by reduc-ing the residual tensile stress of the structure might

FIGURE 11. The effect of the equilibrium chloride concentration in salt deposits on corrosion of austenitic stainless steels exposed to air at 20, 30, and 40C for 10 weeks. Results of experiments with CaCl2 and MgCl2 at 30, 50, and 70% RH are included: (C) Type 304, (#) Type 316L, () Type 904L, () S31254; (black) SCC, (gray) pitting corrosion, (empty symbol with thick border) etching, and (empty symbol with thin border) free of corrosion.

tABlE 6Threshold Values of Chloride Concentration for the Tendency to Stress Corrosion Cracking and Pitting Corrosion

of Austenitic Stainless Steels in Air at 20C to 40C (in mol Cl/kg H2O)

Stainless Steel 20C 30C 40C 20C 30C 40C 304 13.6 10.0 to 11.7 6.8 to 9.1 316L 13.6 11.7 to 14.4 12.0 to 14.4 904L 6.8 to 9.1 >13.6 >14.4 >14.4 S31254 10.0 to 11.4 10.0 to 11.7 6.8 to 9.1 >13.6 >14.4 >14.4

: Not tested

Stress Corrosion CrackingPitting Corrosion

tABlE 7Approximate Ranges of Relative Humidity and Chloride Concentrations that Can Lead to SCC

of Stainless Steel Grades 304 and 316L in the Presence of Salt Deposits at Temperatures 20, 30, and 40C

Salt Deposits 20C 30C 40C 20C 30C 40C MgCl2 27 to 44 25 to 67 CaCl2 17 to 49 12 to 69 17 to 40 12 to 39 Chloride concentration (mol Cl/kg H2O) >14 >10 >7 >14 >12 >12: No danger of SCC

316l304



be very difficult.32 The authors found that the thresh-old stress of SCC for specimens with deposited drop-lets of synthetic seawater and exposed to air at 30C and 35% RH was as low as one-half of the 0.2% proof stress for solution-annealed materials, and smaller than one-fourth of the 0.2% proof stress for sensitized Type 304. Thus, if the deposit formation and access to air with relative humidity within given intervals are possible, application of such materials must be avoided.

The SCC cannot initiate on austenitic stainless steel grades Type 904L and UNS S31254 in the pres-ence of any tested chloride salt and at temperatures up to 50C within the whole range of relative humid-ity. It can be assumed that the threshold value of the chloride concentration leading to SCC is higher than the solubility of the salts. There is probably no dan-ger in application of these grades even at conditions where salt deposits can be formed.

Indeed, there are other salts than those studied in this work with a solubility high enough to cause SCC, e.g., lithium chloride (LiCl), zinc chloride (ZnCl2), or nickel chloride (NiCl2). Currently, it is also not clear if salts affect the tendency to SCC only due to differ-ent chloride concentrations in the surface electrolyte. It is possible that pH of the solution is also relevant. Due to hydrolysis, the pH of CaCl2 and MgCl2 solu-tions is lower than that of the NaCl water solution. However, the effect of pH cannot explain the difference in the aggressiveness of calcium and magnesium chlorides observed in this study. Calcium chloride solutions are less acidic than solutions of magnesium chloride, and it was the former salt that caused a higher degree of SCC in most experimental conditions. For example, the pH of NaCl, MgCl2, and CaCl2 water solutions at the chloride concentration of 5 mol/kg water are 6.3, 5.4, and 5.6 at 20C.33 Also, the free

corrosion potential of stainless steels in solutions of the chloride salts differs. Although it was shown by Arlt and coworkers that the effect of the potential on the studied mode of SCC is rather low, it cannot be fully neglected.12,16,21 Moreover, the salt cation might play a direct role in the degradation mechanism.34-35 These concerns will be addressed by further research.

The conclusions are based on experiments with pure chloride deposits. In real applications, mixed deposits are formed. Since there are no data on the effect of other salts on the tendency to SCC in avail-able literature, further experiments are planned to evaluate this aspect. Finally, it should be noted that the data for this discussion were obtained in 10-week (1,680 h) exposures. The possibility that a prolonged time of exposure would lead to SCC at lower chloride concentrations and temperature cannot be excluded. However, results of exposures for 4, 10, and 22 weeks indicate that the initiation process is rather rapid. No fundamental difference was observed on specimens exposed for shorter and longer periods. Also, data of Shoji and Ohnaka17 show relatively small differences in the presence of cracks on specimens exposed for 3 months and 24 months. Therefore, the authors believe that a 10-week exposure under very severe conditions in terms of both stress and aggressiveness of the environment are sufficiently representative for long-term applications of stainless steel members in non-washing conditions.

Austenitic-Ferritic Stainless SteelsCorrosion of all tested specimens is summarized

in Table 3. The surface of duplex stainless steels was typically etched (Figure 12). Etched zones were usu-ally large, covering almost the entire contaminated surface of lower alloyed materials at more aggres-sive conditions. The zones were often elongated in the

(a) (b)FIGURE 12. Etched surface of duplex stainless steels; left: 2205 exposed with CaCl2 to 50C and to 15% RH for 4 weeks, width 1.2 mm; right: 2507 exposed with MgCl2 to 50C and to 30% RH for 22 weeks, width 4.0 mm.



direction perpendicular to the longer side of the metal strip, i.e., in the same direction as cracks in austen-itic stainless steels. Under certain conditions, the metal deterioration appeared to have been closer to pitting corrosion (Figure 13). Also, the pits were often perpendicularly elongated. It is supposed that there was no principal disparity in the two corrosion fea-tures and that etching and pitting might differ only in the depth and/or density of the attack.

Tiny cracks were seen on the surface of some duplex stainless steel specimens. As shown in Figure 14, these cracks usually passed through small pits. Observations of the superficial cracks are also sum-marized in Table 3. It is obvious that the cracks were found only on lower-alloyed duplex stainless steels at less-aggressive conditions; on specimens of more alloyed materials, they were present only at higher temperature and lower relative humidity. It might indicate that these cracks were, in fact, formed in more conditions but they were rapidly etched off as a result of general corrosion (etching). Consequently, no cracks were found on lower-alloyed duplex stain-less steels after 10 weeks in more aggressive environ-ments, since elongated etched zones developed on the surface.

In total, 19 duplex stainless steel specimens exposed to different conditions were investigated in cross section. Zones of selective corrosion were found under all surface cracks. An example of a large selec-tively corroded area is given in Figure 15. Selected specimens without observable cracks were examined as well. The corrosion morphology varied between rather shallow pitting (Figure 13) and selective corro-

(a) (b)FIGURE 13. Morphology of pitting on UNS S32101 exposed with CaCl2 at 50C and at 15% RH for 4 weeks; left: surface, width 1.5 mm; right: cross section, width 160 m.

FIGURE 15. Selective corrosion of duplex stainless steel UNS S32101 exposed with CaCl2 deposits at 30C and 50% RH for 10 weeks; an arrow indicates location of a tiny crack; width 400 m.

FIGURE 14. Tiny crack on the surface of duplex stainless steel 2304 exposed with MgCl2 at 40C and at 50% RH for 10 weeks; width 1.4 mm.



sion (Figure 16). With energy-dispersive x-ray (EDX) analysis, it was found in all cases that the corroded phase contained more chromium and it was therefore identified as the ferritic phase. The depth of the selec-tively corroded zones depended systematically neither on the exposure conditions nor on the grade. Maxi-mum depths of the selective attack were 200, 90, 140, and 190 m for UNS S32101, 2304, 2205, and 2507, respectively. No principal difference in the aggres-siveness of magnesium and calcium chloride was observed. NaCl deposits did not cause any corrosion.

Although elongated pits and etched zones and tiny cracks were often observed on the surface, the metallographic examination did not confirm the pres-ence of cracks in the structure. It is generally known that duplex stainless steels have better SCC resis-tance than the corresponding austenitic stainless steels. Newman36 attributes these properties to the differences in stress intensity factor of ferritic and austenitic phases. Very likely, one phase assists in blocking the crack propagation.36 It is possible that the microcracks initiated in the austenitic phase that is more susceptible to SCC and were subsequently blocked in the ferritic phase.

It is often observed that one phase in duplex stainless steel corrodes preferentially. The more vul-nerable phase can be either austenite or ferrite, depending on the exact material composition and exposure conditions. It was reported by Lothongkum, et al., that the concentration of nitrogen in highly alloyed duplexes was essential for the phase sensi-tivity to pitting corrosion in aerated NaCl solution.37 The corroded structure in the tested stainless steels without nitrogen alloying was the austenite phase, whereas it was the ferritic phase in materials alloyed with nitrogen. Batista and Kuri showed that due to different concentrations of alloying elements in the

ferritic and austenitic structure, PRE of the phases can vary significantly.38 Based on this work and the results of the EDX analyses, it can be estimated that a PRE ratio of ferritic and austenitic phases was about 0.8, 1.0, 1.1, and 0.9 for grades UNS S32101, 2304, 2205, and 2507, respectively. The ferritic phase of duplex stainless steels with a lower content of nitrogen and higher content of chromium and molyb-denum tend to have higher PRE than the austenitic phase and vice-versa. Thus, it might be expected that stainless steels UNS S32101 and 2507 would be cor-roded preferentially in the ferritic phase and the oth-ers in the austenitic phase. However, experiments showed that the ferritic phase was preferentially cor-roded in all duplex materials.

Since rather large zones of selective corrosion of the ferritic phase, reaching depths up to 14% of the cross section, developed in duplex stainless steels under chloride deposits in atmospheric condi-tions and no systematic effect of climatic parameters was observed, no application limits can be set based on this work. Further measurements are planned to assess the impact of the selective corrosion on mechanical properties of these materials.

conclusIons

v Austenitic stainless steel grades Type 304 and Type 316L were susceptible to the atmospheric SCC in the presence of magnesium and calcium chloride deposits at temperatures from 30C and at low relative humid-ity. Cracks were branched and transgranular.v The corrosivity of salt deposits decreased in the fol-lowing order: CaCl2 > MgCl2 >> NaCl. It increased with temperature and decreased with relative humidity.v The corrosivity of salt deposits was controlled by equilibrium chloride concentration in the surface elec-

(a) (b)FIGURE 16. Selective corrosion of duplex stainless steel 2304 exposed with MgCl2 at 50C and at 30% RH for 4 weeks; width 100 m (left) and 60 m (right).



trolyte formed by absorption of water from air. CaCl2 was more corrosive than MgCl2 because it forms more concentrated solutions at a given relative humidity.v Intervals of relative humidity giving rise to SCC of Type 304 and Type 316L at 30C and 40C were defined together with threshold chloride concentra-tions. They limit application of these grades when highly soluble chloride deposits can be formed and tensile stress is present.v The tendency of austenitic stainless steels to SCC and pitting decreased with increasing PRE. Superaustenitic stainless steels Type 904L and UNS S31254 showed no tendency to SCC. It can be assumed that the threshold value of chloride concen-tration leading to SCC of these materials is beyond the solubility of the tested salts at temperatures up to 50C.v Duplex stainless steels were resistant to SCC. The dominant mode of degradation was selective corrosion of the ferritic phase. Although the corrosion attack was often only superficial, under certain circum-stances the materials were corroded up to 200 m in depth, i.e., 14% of the material cross section. No sys-tematic effect of the climatic parameters and PRE on the degradation was observed. Further research is needed to clarify their application limits in non-wash-ing atmospheric conditions.v Application of austenitic stainless steels with the molybdenum content from 4% such as grades Type 904L or UNS S31254 is recommended in non-washing conditions where it is impossible to eliminate depo-sition of highly soluble chloride salts, temperatures above 20C, relative humidity below 70%, and the presence of tensile stress.

AcknowlEdgmEnts

We gratefully acknowledge financial support of Outokumpu Stainless AB. The authors wish to thank E. Johansson of Outokumpu Stainless AB for valu-able discussions on the results.

rEfErEncEs

1. J.W. Oldfield, B. Todd, Mater. Perform. 29, 12 (1990): p. 57. 2. J.W. Fielder, B.V. Lee, D. Dulieu, J. Wilkinson, The Corrosion

of Stainless Steels in Swimming Pools, Applications of Stainless Steel 92, vol. 2 (Stockholm, Sweden: Jernkontoret [The Swedish Steel Producers Association], 1992), p. 762.

3. J.J.M. Heselmans, J.W.H.van Duijn, Stainless Steel world 12 (2001): p. 21.

4. C.P. Dillon, Mater. Perform. 29, 12 (1990): p. 66. 5. C.L. Page, R.D. Anchor, The Structural engineer 66, 24 (1988):

p. 416. 6. G. Herbsleb, F. Theiler, Mater. Corros. 40 (1989): p. 467. 7. M. Faller, P. Richner, Mater. Corros. 54, 5 (2003): p. 331. 8. J.W. Oldfield, B. Todd, Br. Corros. J. 26, 3 (1991): p. 173. 9. U. Nrnberger, Stahl eisen 110, 6 (1990): p. 141. 10. D. Hirschfeld, H. Busch, I. Stellfeld, N. Arlt, E. Michel, D. Grimme,

G. Steinbeck, Steel Res. (Arch. eisenhuttenwes.) 64, 8-9 (1993): p. 461.

11. G.R. Lobley, Prakt. Metallogr. 27, 9 (1990): p. 472. 12. N. Arlt, E. Michel, D. Hirschfeld, I. Stellfeld, Corrosion Behaviour

of Stainless Steels in the Atmosphere of Indoor Swimming Pools, Process and Materials: Innovation Stainless Steel, vol. 3 (Florence, Italy: Associazione Italiana di Metallurgia, 1999), p. 99.

13. D.S. Min, L. Hong, Stress Corrosion Cracking of Austenitic Stainless Steel 0Cr18Ni9 in Hydrochloric Acid Plus Sodium Chlo-ride Media at Ambient Temperature, 9th Int. Cong. on Metallic Corrosion, vol. 2 (Toronto, Canada: National Research Council of Canada, 1984), p. 216.

14. H.K. Juang, C. Altstetter, Corrosion 46, 11 (1990): p. 881. 15. S. Torchio, Corros. Sci. 20 (1980): p. 555. 16. N. Arlt, E. Michel, D. Hirschfeld, H. Busch, I. Stellfeld, D. Grimme,

G. Steinbeck, Steel Res. (Arch. eisenhuttenwes.) 64, 10 (1993): p. 526.

17. S. Shoji, N. Ohnaka, Boshoku Gijutsu 38 (1989): p. 92. 18. N. Arnold, P. Gmpel, T. Heitz, P. Pscheidl, Mater. Corros. 48, 10

(1997): p. 679. 19. N. Arnold, P. Gmpel, T.W. Heitz, Mater. Corros. 49, 7 (1998): p.

482. 20. J. Mietz, J. Baszynski, J. Fischer, Mater. Corros. 50, 5 (1999): p.

273. 21. N. Arnold, N. Arlt, P. Gmpel, W. Heimann, Stainless Steels for

Fasteners for Architectural Application in Highly Corrosive Envi-ronments, Proc. of Stainless Steels 96 (Dsseldorf/Neuss, Ger-many: Verein Deutscher Eisenhttenleute, 1996), p. 138.

22. N.D. Fairweather, N. Platts, D.R. Tice, Stress Corrosion Crack Initiation of Type 304 Stainless Steel in Atmospheric Environ-ments Containing Chloride: Influence of Surface Condition, Relative Humidity, Temperature and Thermal Sensitization, CORROSION/2008, paper no. 08485 (Houston, TX: NACE Inter-national, 2008).

23. H. Hayashibara, M. Mayuzumi, Y. Mizutani, J. Tani, Effects of Temperature and Humidity on Atmospheric Stress Corrosion Cracking of 304 Stainless Steel, CORROSION/2008, paper no. 08492 (Houston, TX: NACE, 2008).

24. ASTM G30-94, Standard Practice for Making and Using U-bend Stress Corrosion Test Specimens (West Conshohocken, PA: ASTM International, 1994).

25. A. Iversen, T. Prosek, Atmospheric Stress Corrosion Cracking of Austenitic Stainless Steels in Conditions Modelling Swimming Pool Halls, Eurocorr 2007, paper no. 1142 (Freiburg im Breisgau, Germany: EFC, 2007).

26. T. Prosek, A. Iversen, C. Taxen, Low-Temperature Stress Corro-sion Cracking of Stainless Steels in the Atmosphere in Presence of Chloride Deposits, CORROSION/2008, paper no. 08484 (Houston, TX: NACE, 2008).

27. W. Oshikawa, S. Itomura, T. Shinohara, S. Tsujikawa, Boshoku Gijutsu 49 (2000): p. 1,075.

28. M. Broul, J. Nyvlt, O. Shnel, Solubility in Inorganic Two-Com-ponent Systems, Physical Sciences Data, vol. 6 (Amsterdam, The Netherlands: Elsevier, 1981).

29. R.A. Stokes, R.H. Robinson, electrolyte Solutions, 2nd ed. (Lon-don, U.K.: Butterworths, 1959).

30. V.M.M. Lobo, Handbook of Electrolyte Solutions, in Physical Sciences Data, vol. 41 (Amsterdam, The Netherlands: Elsevier, 1989).

31. T. Ohligschlger, P. Pohjanne, T. Taulavuori, R. Karppi, Low-Temperature Stress Corrosion Cracking Under Salt Deposits of Austenitic Stainless Steels in Annealed and Hard Cold-Rolled Conditions, 5th Stainless Steel Science and Market Congress (Sevilla, Spain: Instituto de Recursos Naturales y Agrobiologa [CSIC], 2005).

32. M. Mayuzumi, T. Arai, K. Hide, Boshoku Gijutsu 52 (2003): p. 227. 33. T. Prosek, D. Thierry, C. Taxn, J. Maixner, Corros. Sci. 49

(2007): p. 2,676. 34. T. Shibata, Corros. Sci. 49 (2007): p. 20. 35. Z. Shenghan, T. Shibata, T. Haruna, Corros. Sci. 47 (2005): p.

1,049.36. R.C. Newman, Corrosion Mechanisms in Theory and Practice,

eds. P. Marcus, J. Oudar (New York, NY: Marcel Dekker, 1995), p. 311.

37. G. Lothongkum, P. Wongpanya, S. Morito, T. Furuhara, T. Maki, Corros. Sci. 48 (2006): p. 137.

38. S.R.F. Batista, S.E. Kuri, Anti-Corros. Method Mater. 51 (2004): p. 205.

Low-Temperature Stress Corrosion Cracking of Stainless Steels in the Atmosphere in the Presence of Chloride Deposits

Documents

atmospheric corrosion

outokumpu stainless

uns n08904

uns s30400

uns s32304

uns s31603

uns s32750

uns s32205