Fracture surface of Chloride Stress Corrosion Cracking
Figure 4 - Relative stresses and strains of the CERT experiments on AISI 304 in 0.001 M NaCI solution at 200 °C.
Figure 7 - Surface cracks resulting from strengthening during slow straining in inert gas at (a) 200 ° and (b) 300 °C (magnification: 250X).
|
|
Oxygen and Corrosion Potential Effects on Chloride Stress Corrosion Cracking W. M. M. Huijbregts * Corrosion Vol, 42, No 8, p 456-462, 1986. (paper 31) * NV KEMA, Utrechtseweg 310, 6812 AR Arnhem, The Netherlands
Abstract Constant extension rate test (CERT) experiments have been performed on AISI 304 at 200 °C in a 0.001 M NaCI solution. During straining, the specimens were kept at constant potentials in the range of -400 to +425 normal hydrogen electrode (NHE). Chloride stress corrosion crack-ing (SCC) was seen only at potentials above +150 mV NHE. At lower potentials, only small brittle surface cracks were formed because slow straining at high stress levels strengthens the steel. From corrosion potential measurements at 200 °C in oxygen-containing water, it follows that the critical potential value of + 150 mV NHE can be reached with oxygen contents above 10 ppb in nearly stagnant water. Introduction Small amounts of chloride in secondary pressurized water reactor (PWR) or primary boiling water reactor (BWR) water-steam circuits can cause stress corrosion cracking (SCC) of stainless steels (SSs). Important factors are the following:
Many results on this subject have been previously published (1-8). This paper focuses on the critical oxygen content related to chloride SCC in water-steam circuits. The corrosion potential of a passive SS is controlled primarily by the oxygen content in the water. Many electrochemical corrosion potential (ECP) measurements on SS have been performed under BWR water conditions, because in the piping and nozzles of BWRs, there were many in-service failures of intergranular stress-assisted cracking of sensitized SS. Figures 1 and 2 show the results of the measurements (1,2,9).
Figure 1 - Corrosion potentials of AISI 304 SS as a function of oxygen concentrations at 100, 150, 200, 250, and 288 C in high-purity water [measurements of Lee (2)].
Figure 2 - AISI 304 SS potential at high temperature: // // = 274 C (measurements of Indig (9); O = 288 C (measurements of Lee (2); and L = 288 C (measurements of Leibovitz (1)). These results vary considerably; thus, the present author decided to perform more research on this project, the results of which have been published recently (1). See Figure 3.
Figure 3 - Corrosion potential of AISI 304 SS in 200 C water. The water was ammoniated to pH 8.5: linear flow velocity in the autoclave, 0.001 m/s; pressure, 50 bar (the measurements at 274 °C of Indig (9) are also plotted). If the oxygen potential correlation is well known, exposure tests under controlled poten-tials will provide more knowledge on the influence of oxygen on corrosion processes. Thus, constant extension rate test (CERT) experiments were performed at controlled potentials. Experiments The experiments were conducted in a 2 L autoclave. The sample (2.5-mm diameter and 30-mm measuring length) was mounted in the straining bridge, by means of air-oxidized Zircaloy fittings, to prevent galvanic contact with the autoclave. The contact wire was spotwelded to one of the thread ends of the specimen. The electrode wires were pressure tightened with Conaxt fittings in the upper part of the autoclave. A 0.001 M NaCI solution was chosen as a non-flowing test environment. Nitrogen gas was purged through the solution during the heating time of the autoclave (2 h to reach 80 °C). The ECP of SS was then rather low [-350 mV NHE (normal hydrogen electrode)], indicating a very low oxygen content in the autoclave. Afterwards, the autoclave was closed and heated further to 200 °C. The sample was on the free corrosion potential until this temperature was reached. As soon as the straining was started, the sample was switched onto the desired potential by means of a potentiostat. A platinum counter electrode and Ag-AgCI (0.001 NaCI) reference electrode, mounted in a polytetrafluoroethylene (PTFE) capsule, were used. The calculated thermodynamic potential of this reference elec-trode at 200 °C amounts to +300 mV NHE (9). Extension rate experiments were performed both in inert gas environments (N2 and Ar) and in 0.001 M NaCI solution at 200 °C. The experiments in inert gas served as references for fractography. The AISI 304 steel was of the following chemical composition in weight percent (wt%): 0.06 C, 1.70 Mn, 0.44 Si, 0.017 S, 0.029 P, 18.3 Cr, 9.0 Ni, 0.24 Mo, and 0.16 Cu. The samples were machined from solution-annealed bar material. The deformation stresses from machining were not removed by annealing, except in the last stage of the research, in which two annealing procedures were used:
Each specimen had been wet polished with 600 mesh paper. Results Figure 4 shows the relative tensile strain and tensile stress values obtained from the experiments at controlled corrosion potentials. The relative strain and stress has been defined as Er = Ec/Ei* 100% at = ac/ai*100% in which Ei and ai = strain and stress in an inert environment (N2 or Ar), and Ec and ac = strain and stress in NaCI solution. The relative strain and stress values decrease at potentials above +150 mV NHE. Above 300 mV NHE, the relative strain increases again, but the relative stress continues to decrease. The number and depth of the stress corrosion cracks increase at greater potentials (see Figure 5). This results in a lower effective cross section of the specimens; therefore, the relative tensile strength decreases at high potentials. However, the many cracks above + 300 mV exhibit a strong increase in length, resulting in a minimum in the relative strain, as shown in Figure 4:
Figure 5 - CERT samples exposed in 0.001 M NaCI at different potential values. The sample diameter is 2.5 mm. The potentials are given in the photographs: extension rate, 7.5 10-5 mm/s More detailed microscopic examination revealed that in specimens exposed at lower potentials (< 150 mV), there were small surface cracks (Figure 6). These hemispherical cracks are clearly brittle, and the river pattern is perpendicular to the sample surface [Figures 6(a) and (b)]. This is contrary to the chloride-induced stress corrosion cracks at potentials above +150 mV NHE, where the well-known rosette-like crack surface is observed [Figure 6(c)]. It is remarkable that for specimens exposed at potentials above 150 mV, there were no characteristic surface cracks (also see Figure 5).
Figure 6 - Scanning electron and optical micrographs of CERT samples exposed to 0.001 M NaCI at [(a) and (b)] -175 mV and [(c) and (d)] +425 mV. At potentials above +150 mV NHE, typical chloride stress corrosion cracks [(c) and (d)] were formed. At lower potentials, small brittle surface cracks [(a) and (b)] were found [450X for (a) and (c) and 250X for (b) and (d)]. Figure 6(d) shows the branching of the chloride stress corrosion cracks. In specimens exposed at potentials above +300 mV, slight pitting was noted. This pitting became more apparent at greater potentials. The pitting potential of 300 mV in 0.001 M NaCI was also found by means of polarization curve measurements, with a scan rate of 60 mV/min. To study the cracks formed at the low potentials in more detail, many experiments were performed in an inert gas environment at various temperatures (11). From Figure 7, it can be concluded that small transgranular surface cracks are formed. Because the surface cracks could result from surface deformation during machining of the samples, some samples were annealed. Two annealing procedures were chosen:
Figure 8 shows the sample surfaces after the CERT.
FIGURE 8 - [(a) and (b)] Stress-relieved and [(c) and (d)] solution-annealed samples exposed in an inert gas environment [(a) and (c)] and in 0.001 M NaCI at -400 mV NHE [(b) and (d)]: extension rate, 7.5 10-5 mm/s [100X for (b) through (d) and 1600X for (a)]. In both of the annealed specimens, small surface cracks are formed. However, the number of cracks appeared to be smaller than that in the non-annealed material. Cracks in the inert gas environment samples do not exhibit clear river patterns. The river-line pattern is more pronounced in the NaCI samples. Discussion Much attention has been paid to the formation of the small brittle surface cracks in inert gas environments (11). These cracks are formed in both annealed and non-annealed samples: The cracks seem to have been formed during the plastic deformation period of the slow extension rate test. Because of the low deformation rate, strengthening of the steel will occur, and at imperfect places, small cracks will be formed. This effect will be greater in non-annealed steel samples than in the solution- or stress relief-annealed conditions. In the case of NaCI solution (0.001 M) under less corrosive conditions (< +150 mV NHE), the cracks resulting from strengthening appeared to become larger and more brittle. However, this corrosion process occurs at stresses and deformation grades, which can be considered to be unrealistic under most conditions of practical interest. Thus, the occur-rence of small surface cracks under these test conditions should not be taken as a criterion for SCC. These surface cracks are not present in the samples exposed at high potentials (in which real chloride SCC occurred), since the tensile stresses in these samples are too low. In the CERT experiments in pure water, transgranular surface cracks were also observed by Agrawal and Staehle (3) in sensitized as well as solution-annealed AISI 304 at 288 °C. Hirano concluded from CERT experiments at 290 °C on sensitized AISI 304 SS in pure water that transgranular stress corrosion cracking (TGSCC) occurred at oxygen contents below 100 ppb, and intergranular stress corrosion cracking (IGSCC) occurred above 1000 ppb oxygen (12). Agrawal and Hirano did not report results from experiments in inert gas environments. Indig (9) conducted CERT experiments in a 0.01 M sodiumsulphate solution at controlled potential at 274 °C. At low potent6am (-400 and -100 mV NHE), small transgranular cracks were formed that sometimes exhibited fatigue characteristics on the fracture surface. Indig also concluded that the brittle surface cracks are of no practical importance because they are formed at unrealistically high stress levels. These crack are not attributed to TGSCG. Much research has been conducted on sensitized AISI 304 SS because of the many SCC failures in the nozzles and pipes of BWRs. The critical potentials and/or oxygen contents for IGSCC were determined with CERT. This technique was applied less frequently for studies on chloride-induced SCC in high-temperature, high-pressure conditions. Table 1 summarizes the critical potentials mentioned in the literature. Indig, Rosborg, and Rosengren found a critical potertial for IGSCC of ca -325 mV NHE in a BWR environment (13, 14 and 15). The sensitization of the SS was very severe in these experiments Andresen (6), Indig (9), Cragnolino (10), and Macdonald (17) studied the influence of chloride on IGSCC. The chloride did not influence the critical potential. Poznansky (8) performed CERT experiments in sulfate/chloride environments; sulfate ions shifted the potential to a 100 mV higher value. Poznansky attributed this increase to inhibition of the anodic dissolutio-process at the crack tip by the sulfate ions. An ECP value of -325 mV NHE at 288 °C is achieved on sensitized SS in pure water with an oxygen content of ca 100 ppb, according to the measurements of Lee (2) and a value of 15 a 40 ppb is achieved, according to the measurements of Leibovitz (1). From CERT experiments in the BWR Dresden-2, a critical oxygen content for IGSCC at 270 °C was determined to be 20 to 40 ppb (18). This value agrees with the critical ECP of -325 mV NHE and the oxygen ECP measurements of Leibovitz. Rosengren (15) measured a critical potential of -300 mV at 250 °C, which corresponded to 10 ppb oxygen in the water of those experiments. Such low oxygen contents in a BWR can be achieved by dosing hydrohen (18,19). However, some reservations still exist concerning the application of this alternate water treatment because of the formation of radioactive ammonia: N2 + 3H2 - 2NH3 (2) Without H2 dosing water, soluble nitrates are formed in the BWR water. The formation of radioactive NH3 increases the radiation level at the turbine, if adequate countermeasures are not taken. From oxygen ECP measurements at 200 °C, it appears that less oxygen can be tolerated at this temperature if low ECP values should be attained. In chloride-containing water (350 ppm). Macdonald measured a critical potential of -240 mV NHE. Assuming that chloride does not influence the value of critical potential for IGSCC at 200 °C, only a rather low oxygen content of 1 to 5 ppb can be tolerated. See Figure 3. From the measurements at 250 to 290 °C, it appeared that the critical potential was not significantly influenced by the chloride con-tent. Regarding the TGSCC of solution-annealed SS, relatively few studies on CERT experiments in high-temperature, high-pressure water have been publisheds (8). See Table 1.
The critical potential for chloride-induced TGSCC is obviously greater than that of the IGSCC of sensitized SS. Andresen measured -20 mV in a 100 ppm chloride solution at 290 °C. Macdonald found IGSCC on sensitized SS at low potentials, but also found TGSCC above a potential of 0 mV in 350 ppm chloride solution. In sulfate-chloride environments, the critical potentials shift to much greater values (+160 to +810 mV). These measurements of Poznansky are also cited in Table 1. It is concluded from Figure 3 that the O2 content, causing an ECP value equal to + 150 mV, ranges from 10 to 1000 ppb. Speidel summarized the results of SCC long-term (> 1000 h) exposure tests mentioned in the literature. See Figure 9. In this figure, the areas of no cracking, IGSCC, and TGSCC are drawn. The figure shows that the critical oxygen content for IGSCC is 200 ppb, and that TGSCC at this oxygen value initiates at a chloride content of 5 ppb. The mentioned data of Table 1 are plotted in a similar diagram. See Figure 10.
Figure 10 - SCC areas from the CERT data. A critical potential of -325 mV at 290 °C is attained at an oxygen content of 10 to 40 ppb, which is an obvious difference with the 200 ppb value from the long-term exposure tests. In the long-term experiments, oxygen in the water system can be consumed by the oxidation of the autoclave wall and specimen. This can possibly explain the measured difference in the critical oxygen content (10 to 40 ppb in the CERT experiments vs 200 ppb in the long-term exposure tests). The greater chloride content of 10,000 ppb, which was determined from extrapolation of the ductile-IGSCC and IGSCC-TGSCC lines in Figure 10, is much greater than the value of 5 ppb chloride given by Speidel. In the long-term experiments, a relatively high oxygen content was applied (> 100 ppb). Under these conditions, the ECP value will be rather high and certainly above the pitting potential. Pitting corrosion will occur, and TGSCC will initiate from these sites. This can explain the much lower chloride content, at which TGSCC occurs in the long-term experiments. In the present experiments, the pitting potential at 200 °C was - 150 mV above the critical potential for TGSCC. This was concluded from the CERT experiments as well as from the polarization curves measured in chloride solutions at a scan rate of 60 mV/min. These measured pitting potentials are also plotted in Figure 10. Macdonald found that the critical potential for TGSCC coincides with the pitting potential in a 350 ppm chloride solution at 250 and 290 °C. Considering that the measured potential values are accurate, it must be concluded that at higher temperatures, the pitting potential decreases more than the critical potential for TGSCC. Conclusions
References
|
||
Figure 9 - SCC areas from data collected from Speidel (20) long-term (> 1000 h) exposures. |
||||
|
|
||||
