Alloy Design to Prevent Intergranular Corrosion of Low-Cr Ferritic Stainless Steel with Weak Carbide Formers

Table I lists the chemical composition of 11 experimental alloys in the first group. The specimen numbers starting with SCF means that the alloys are prepared with addition of the strong carbide former and those starting with WCF means that the alloys are prepared with addition of the weak carbide formers. The main purpose of this alloy group was to determine the range of amount of each alloying element which can prevent IGC, but does not impart any negative effect on the following welding and forming processes. SCF-1 and SCF-2 were prepared with addition of Ti as stabilizer with SR being 10.8 and 14.5, respectively. However, SCF-1 has higher C content than SCF-2. All the alloys prepared with addition of WCFs were designed to observe the effect of each element on IGC by fixing the amount of either one or two out of Mn, Mo and Si, and the levels of C and N content are controlled to have less than 100 ppm or 0.01 wt%.

Table II lists the chemical composition of 6 experimental alloys in the second group. SCF-11 and SCF-12 are prepared with addition of Ti and Nb as stabilizer with SR being 21 and 29, respectively. 4 experimental alloys with WCFs were prepared with different amount of Mo, Mn and Si to figure out adequate alloying composition to prevent IGC of low-Cr FSS and to investigate the IGC prevention mechanism by addition of WCF elements. From the preliminary study with the experimental alloys in the first group, the optimum composition of WCF-14 was to have 0.12 wt% Mo, 0.61 wt% Si and 0.37 wt% Mn. The alloy composition of WCF-14 meets the new alloy design concept mentioned earlier. On the other hand, other WCF alloys were prepared with insufficient amounts of one element out of Mo, Mn, and Si. WCF-11 had an insufficient amount of Mo, WCF-12 had an insufficient amount of Si, and WCF-13 had an insufficient amount of Mn.

The experimental alloys were produced by vacuum arc melting and hot rolled to 2.0 mm thick sheet in laboratory. The specimens were solution treated at 1300°C for 10 min, quenched by water, and then aged at 500°C for 2 h.

Intergranular corrosion resistance was examined by the chemical corrosion test using modified Strauss test solution (0.5% H₂SO₄ + 6% CuSO₄) introduced by Devine for low-Cr FSS (10–13 wt% Cr).^24,25 The concentration of H₂SO₄ in this solution was much lower than 16% of ASTM A262 and A763 to prevent heavy general corrosion of low-Cr FSS. Rectangular test coupons (12 × 12 × 1.2 mm) grinded with abrasive paper up to P2000 were immersed in the boiling solution of 300 ml and contacted electrochemically with copper balls in the solution to induce a galvanic effect on the specimen. After 20 h of immersion, the microscopic observation on the specimen surface was conducted for IGC attack.

The DL-EPR test is a well established method to evaluate IGC of the low-Cr FSS.^{2,4,13,26–28} The DL-EPR test was conducted in 0.5% H₂SO₄ solution with 0.0013 wt% KSCN at 30°C after grinding the surface of the specimens with abrasive paper P2000. The specimen was anodically polarized from the corrosion potential to 800 mV_SCE with the scan rate of 1.67 mV/s. In this polarization, the active dissolution occurs, and then a passive layer forms on the entire surface. At 800 mV_SCE, the polarization was reversed with the same scan rate, and the test was finished at the corrosion potential from which the polarization scan was initially started. The reverse scan induced the breakdown of the passive layer in the Cr depleted zone. The anodic loop was larger than the reverse loop because the entire surface of the specimen was activated just before reaching the primary passive potential. However, during the reverse scan, the small reverse loop was measured because most surface area was covered with the passive layer, except for the activated area of the Cr depleted zone around grain boundaries. The maximum activation current peak (I_a) and the maximum reactivation current peak (I_r) were determined from the polarization curves. The degree of sensitization (DOS) of the specimens was determined as the ratio of I_r/I_a. After the DL-EPR test, the microstructure of the specimens was examined to evaluate the susceptibility of IGC and grain boundary morphology was classified as either step (no sensitization), dual (a little sensitization), or ditch (significant sensitization).^26,29

Metallographic observation was performed with various techniques. Morphology of grain boundaries was observed with SEM (JEOL JSM-7401F). Analysis on the intermetallic compounds was carried out by using TEM (JEOL JEM-2200FS with Image Cs-corrector) with electron energy loss spectroscope (EELS). A TEM sample was prepared with a carbon extraction replica method. The nanoscale analysis on grain boundary was performed with Cameca LA-WATAPTM laser-assisted three-dimensional atom probe (3DAP), which was particularly useful for analyzing the elemental distribution with near-atomic resolution.³⁰ Samples for 3DAP analysis were prepared with a focused ion beam (FIB) milling process.

Figure 2 presents the surface appearances of 11 experimental alloys in the first group after the modified Strauss IGC test. SCF-1 and SCF-2 are severely damaged even though they are stabilized with Ti. The degree of damage is more severe in SCF-1 than in SCF-2 since SCF-1 has more C content than SCF-2. The surface appearances of WCF-1, WCF-2 and WCF-3 indicate that addition of Mn, Si and Mo definitely improves IGC resistance. The surface appearances of WCF-4 and WCF-5 also suggest that the critical Mn content is important to secure the IGC prevention even though enough amount of Mo and Si is added. Through a careful examination on the IGC test results of the experimental alloys, a number of alloys such as WCF-6, WCF-7, WCF-8 and WCF-9 were prepared, and all of them show no sign of IGC attack. The IGC test on the first group of the experimental alloys suggests the minimum amount of WCFs for optimum alloy composition. For IGC prevention of low-Cr FSS, the critical amount of Mn, Mo and Si seems to be 0.35 wt%, 0.1 wt% and 0.6 wt%, respectively. On the basis of the IGC test result of the first experimental alloy group, the second alloy group was designed as listed in Table II to examine the IGC prevention mechanism by addition of WCFs.

Zoom In Zoom Out Reset image size

Figure 3 presents the grain boundary morphologies of the specimens after the modified Strauss IGC test. The evidence of IGC such as grain dropout (marked as A in Figure 3) and dissolution of grain boundary are observed in both SCF-11 and SCF-12. Both alloys stabilized with strong carbide formers are sensitized even though the SR value of SCF-11 (SR = 21) and SCF-12 (SR = 29) are higher than the recommended SR value (SR = 20).¹² However, WCF-14, which contains 0.12 wt% Mo, 0.37 wt% Mn and 0.61 wt% Si, is not sensitized at all, and it suggests that the co-addition of Mo, Mn and Si to low-Cr FSS is effective to suppress the sensitization and IGC. On the other hand, grain dropout was observed in WCF-11, which indicates that addition of only Mn and Si without Mo cannot prevent IGC. Though Mo was added more than 0.1 wt%, a lack of Si as in WCF-12 or a lack of Mn as in WCF-13 also induces IGC of low-Cr FSS. This result reveals that IGC of low-Cr FSS can be effectively prevented if the proper amounts of Mo, Mn, and Si are added.

Zoom In Zoom Out Reset image size

DL-EPR was conducted as an additional, quantitative measure of the sensitization level. Figure 4 compares the DL-EPR test result of the conventionally stabilized specimens, SCF-11 and SCF-12, with that of WCF-14 alloyed with weak carbide former. From the DL-EPR test results, the values of the peak reactivation current (I_r) and peak anodic current (I_a) were measured and DOS (I_r:I_a) were calculated as listed in Table III.^4,13,31 In the case of low-Cr FSS, previous studies suggest that IGC of low-Cr FSS can be prevented when DOS value is lower than 0.03.^13,19 It was revealed that DOS values of SCF-11 and SCF-12 were higher than this criterion, whereas DOS value of WCF-14 was lower than this criterion. As shown in Figure 5, the grain boundary morphology of SCF-11 and SCF-12 shows a ditch structure along the grain boundary after DL-EPR test, suggesting that the grain boundaries of both SCF-11 and SCF-12 were sensitized. On the other hand, the DOS value of WCF-14 was lower than 0.03, and step structure was clearly observed from the grain boundary after DL-EPR test. The results of DL-EPR are in good agreement with the previous result of the modified Strauss test and also suggests that the co-addition of Mo, Mn, and Si in WCF-14 was effective to prevent grain boundary sensitization.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 6 presents SEM characterization along grain boundary of WCF-14, and it indicates that most of grain boundary area is occupied by 200–300 nm size of intermetallic compounds. However, no Cr compound, including Cr-carbide or Cr-carbonitride, is observed near the grain boundary in WCF-14. The co-addition of Mo, Mn and Si prevents grain boundary sensitization and consequent IGC of WCF-14 by formation of blocky intermetallic compound along grain boundary. To investigate the chemical composition and crystal structure of the intermetallic compounds along grain boundaries, TEM characterization was performed on WCF-14 specimen. Figure 7 presents a bright field image with EELS mapping of the WCF-14 specimen prepared by the carbon replica method and its electron diffraction pattern analysis result. From EELS analysis, it is revealed that the intermetallic compound consisted of C, Mo, Mn, and Si as shown in Figure 7(a). Electron diffraction pattern analysis on this intermetallic compound identifies it as CMn₄MoSi, as shown in Figure 7(b).³² Formation of this CMn₄MoSi intermetallic compound requires the proper amount of Mo, Mn, and Si because WCF-11 (insufficient Mo content), WCF-12 (insufficient Si content), and WCF-13 (insufficient Mn content) do not form sufficient amount of CMn₄MoSi intermetallic compound to prevent IGC. Only WCF-14 has a sufficient amount of Mo, Mn, and Si and it promotes the formation of CMn₄MoSi intermetallic compound along grain boundaries to suppress effectively the sensitization of FSS.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The intermetallic compound CMn₄MoSi formed along grain boundary of FSS can block the diffusion of C and Cr toward the grain boundary during heat-treatment. Moreover, it is very interesting to observe from Figure 7(a) that Cr is segregated around the intermetallic compounds. It seems that the solubility of Cr in CMn₄MoSi is low enough to expel Cr from it. Then, Cr has little chance to react with C because C is already consumed to form CMn₄MoSi intermetallic compound. This means that formation of CMn₄MoSi intermetallic compound along grain boundaries helps to segregate the solute Cr around CMn₄MoSi intermetallic compound, so that Cr can be replenished along the grain boundary instead of Cr being depleted.

To investigate atomic segregation behavior in the vicinity of grain boundaries, 3DAP characterization of WCF-13 and WCF-14 was performed and the results are compared with that of SCF-11. The concentration profile of alloying elements along the grain boundary of SCF-11 was investigated in our previous study and is shown in Figure 8.¹⁶ Figure 8(b) indicates that Cr atoms were extensively segregated up to 35.8 at.% on the grain boundary and consequent Cr depletion in the vicinity of the Cr segregation zone. The lowest Cr concentration of the depletion zone near the grain boundary of SCF-11 was only 5.3 at.%, which suggests that the grain boundaries of SCF-11 were sensitized. Figure 9 presents the concentration profile along the grain boundary of WCF-13, which contains sufficient amounts of Mo and Si, but an insufficient amount of Mn. The 3DAP elemental maps of WCF-13 shown in Figure 9(a) indicate that segregation of Cr, C, Mo, Mn, and Si occurs along grain boundaries. Figure 9(b) shows the concentration profile across the grain boundary of WCF-13, where Cr atoms were segregated up to 23.0 at.% and Cr depletion was down to 7.8 at.%. It is interesting that though Cr segregation occurred in both SCF-11 and WCF-13, the concentration of Cr (23.0 at.%) at the grain boundary of WCF-13 was much lower than that (35.8 at.%) of SCF-11.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

WCF-14, which contains a sufficient amount of all three elements of Mo, Si, and Mn, shows a totally different Cr profile near the grain boundary when compared with that of SCF-11. As shown in Figure 10, the highest amount of segregated Cr along the grain boundary of WCF-14 was only 18.2 at.%, which was much lower than those of SCF-11 and WCF-13, and there was virtually no Cr depletion near the grain boundaries because Cr content in the depleted area was only 9.9 at.%.

Zoom In Zoom Out Reset image size

The difference in the Cr profile along grain boundary of SCF-11, WCF-13 and WCF-14 is closely related with the C profile. During the heat-treatment and cooling or aging, solute C diffuses first to grain boundaries due to its much higher diffusivity than those of other elements. According to the conventional IGC mechanism, segregation of Cr is induced and fixed by the formation of Cr carbide. However, in case of the stabilized low-Cr FSS, Kim et al. have reported that, although Cr carbide is not formed, segregated C induces saturation of the solute Cr atoms along grain boundary and it consequently promotes sensitization.¹⁹ During aging, all the elements in the matrix tend to diffuse toward grain boundaries. Carbon diffuses first to the grain boundaries because of its much higher diffusivity than the other elements. Because of chemical affinity with C, both Cr and Ti diffuse to grain boundaries. Because Ti has stronger carbon affinity than Cr, Ti forms TiC preferentially. Cr cannot form Cr-carbide because no free C is available and thus solute Cr atoms are segregated near grain boundary. As the aging time increases, while precipitation of TiC or NbC progresses, back diffusion of these segregated Cr is insufficient, and thus grain boundaries of FSS is sensitized.¹⁶ Therefore, to avoid segregation of Cr and consequent IGC of low-Cr FSS, segregation of C along grain boundaries should be suppressed.

From the 3DAP results, the relationship between IGC and atomic distribution of Cr and C along grain boundary can be obtained and the results are listed in Table IV. After the aging treatment, C atoms in SCF-11 are segregated along grain boundaries up to 11.2 at.%, and they induce segregation of Cr up to 35.8 at.%. In contrast, when WCF elements of Mo, Mn, and Si are added to low-Cr FSS, the concentration of C along grain boundaries significantly decreases. In the case of WCF-13 in which even though an insufficient amount of Mn was added, C atoms were segregated along grain boundaries to 5.5 at.% and it was lower than that of SCF-11. As the segregation of C along grain boundaries was reduced, that of Cr was also decreased to 23.0 at.%, and concentration of Cr in the Cr depleted zone was higher in WCF-13 (7.8 at.% Cr) than in SCF-11 (5.3 at.% Cr). In the case of WCF-14, when a sufficient amount of Mo, Mn, and Si was added to FSS, C atoms were segregated only to 0.26 at.%, which was much lower than that in both SCF-11 and WCF-13. Because 3DAP analysis was conducted in the nano-scale range, for proper analysis, extreme care and expertise was required for the sample preparation. Therefore, the quantitative analysis on C distribution along the grain boundaries should be precisely investigated with different sets of specimens in further research. However, both WCF-13 and WCF-14 with co-addition of Mo, Mn, and Si show a similar tendency to suppress C segregation along the grain boundaries, and this phenomenon is promoted as the concentration of Mo, Mn, and Si increases (SCF-11 < WCF-13 < WCF-14). Therefore, comparison of 3DAP analysis data of alloys with strong carbide former (SCF-11) and with weak carbide former (WCF-13 and WCF-14) suggests that the co-addition of weak carbide formers (Mo, Mn, and Si) is effective to suppress segregation of C along the grain boundaries in low-Cr ferritic stainless steel and seems to be related to the formation of CMn₄MoSi intermetallic compound. Because the C concentration profile by 3DAP was measured within a limited area (∼40 nm) near the grain boundaries, the scan area of 3DAP for WCF-14 shown in Figure 10 is significantly smaller than the size of the intermetallic compound (∼250 nm as shown in Fig. 6) formed along the grain boundary. Therefore, the C profile near the grain boundaries of WCF-14 does not represent C concentration in the matrix, but it was influenced by formation of the intermetallic compound. Because CMn₄MoSi intermetallic compounds along the grain boundaries consume C atoms, they can effectively reduce the concentration of C along grain boundaries when compared to C concentration along grain boundaries of SCF-11 where no CMn₄MoSi intermetallic compound can be formed. Therefore, because of low C concentration along grain boundaries in WCF-14, both the formation of Cr carbide and the segregation of Cr atoms were suppressed and no significant depletion of Cr was observed near the grain boundaries. Although the solubility of carbon in CMn₄MoSi intermetallic compound has not yet been investigated, it has been reported that intermetallic compounds formed in stainless steel such as Laves phase has a certain amount of carbon solubility.^33,34 Moreover, as shown in Figure 7, the EELS intensity of carbon in CMn₄MoSi intermetallic compounds was significantly higher than that of the back ground carbon intensity of the replica. It is clear that the high C solubility and blocky size of the intermetallic compound formed in WCF-14 have three effects to prevent IGC, as schematically presented in Figure 11. One is to act as an effective C trap site, the second is to act as the effective diffusion barrier against the solute Cr diffusion toward grain boundary, and the third is to replenish Cr back into the Cr depleted zone due to its low solubility of Cr in CMn₄MoSi intermetallic compound.

Zoom In Zoom Out Reset image size

This study proves that co-addition of weak carbide formers to low-Cr ferritic stainless steel can prevent IGC by formation of intermetallic compound along the grain boundary. In general, precipitation of an intermetallic compound may decrease toughness and mechanical strength depending on the size and distribution of the precipitate. The low-Cr ferritic stainless steel is not usually recommended for an engineering component requiring high toughness. Therefore, in this study, we did not specifically focus on the effect of precipitation of the intermetallic compound on the toughness and mechanical strength. For proper engineering applications, however, it is recommended to validate the mechanical property of the ferritic stainless steel when alloyed with co-addition of weak carbide formers.