Intergranular corrosion behavior and mechanism of the stabilized ultra-pure 430LX ferritic stainless steel

1. Introduction

Ferritic stainless steels (FSSs) have been widely used in modern energy and manufacturing industries owing to their good mechanical properties, corrosion resistance and moderate price [[1], [2], [3]]. However, a major obstacle to the use of FSSs is their susceptibility to intergranular corrosion (IGC) after welding or improper heat treatment [4,5]. A great number of environments lead to damaging IGC of sensitized FSSs.

According to the previous researches [[6], [7], [8], [9]], IGC mechanism of FSS is similar to that of austenitic stainless steel (ASS). The well-known IGC mechanism is that Cr-rich compounds such as carbides (M₂₃C₆) and nitrides (M₂N) precipitate at the grain boundaries during heat treatment. The formation of Cr-rich compounds leads to the consumption of chromium atoms and the formation of Cr-depleted zones. The Cr-depleted zones adjacent to intergranular precipitates contain lower level of chromium than the matrix and are therefore prone to corrosion. However, in virtue of the lower solubility and more rapid diffusion speed of carbon and nitrogen in ferrite than austenite [7], Cr-rich compounds precipitate more easily in FSS compared with that in ASS [4], and thus FSS is more susceptible to IGC than ASS.

To prevent the precipitation of Cr-rich compounds at the grain boundaries and improve the IGC resistance of FSS, the amount of carbon and nitrogen is reduced as low as possible. Meanwhile, various stabilizing elements such as titanium and niobium are also added to react preferentially with carbon and nitrogen [[10], [11], [12], [13], [14], [15]]. Although the IGC resistance of FSS has been greatly improved, some failures induced by IGC still widely emerge. Huang et al. [13] reported that IGC occurred in the Ti-Nb-stabilized 430 FSS heat-treated above 1050 °C. For the stabilized ultra-pure FSS, the susceptibility to IGC is induced by giving a two-step heat treatment, which consists of an initial high temperature solution treatment and a subsequent low temperature aging treatment [16,17]. Kim et al. [18] reported that IGC occurred in the Ti-stabilized ultra-pure 409 L FSS aged at 400-600 °C after solution treatment at 1300 °C for 10 min. Li et al. [19] also reported that IGC occurred in the Ti-Nb-stabilized ultra-pure 429 FSS aged at 500-700 °C after solution treatment at 1200 °C for 1 h.

Although Cr-depleted zones have been generally considered as the main cause of IGC occurring in the stabilized ultra-pure FSS, there are still some disagreements on what induces Cr-depleted zones. Niekerk et al. [20,21] reported that M₂₃C₆ precipitates formed at the grain boundaries in the welding HAZ of the Ti-stabilized ultra-pure AISI 409 FSS. Qiang et al. [22] also observed M₂₃C₆ carbides along the grain boundaries in the Ti-stabilized ultra-pure 409 L FSS aged at 600 °C after solution treatment. They supported that the precipitation of Cr-rich compounds induced Cr-depleted zones. On the other hand, Devine et al. [16] pointed out that (Ti,Cr)(C,N) precipitates formed at the grain boundaries in the 18Cr-2Mo-Ti stabilized FSS and the Cr/Ti ratio increased with the growth of (Ti,Cr)(C,N) precipitates, and thus proposed that the increase of Cr/Ti ratio in (Ti,Cr)(C,N) precipitates caused the formation of Cr-depleted zones during aging treatment. Kuzucu et al. [11] and Suzuki et al. [23] proposed that chromium atoms might segregate around MC-type precipitates at the grain boundaries, such as TiC, NbC and (Ti,Nb)C. Kim et al. [18,[24], [25], [26], [27]] also reported the chromium segregation phenomenon at the grain boundaries and proposed that Cr-depleted zones were induced by the segregation of chromium atoms near MC precipitates at the grain boundaries. Accordingly, the IGC mechanism of the stabilized ultra-pure FSS still needs to be further investigated.

In order to avoid improper heat treatment and guide the welding process, the IGC behavior of the Ti-Nb-stabilized ultra-pure 430LX FSS was investigated. A double loop electrochemical potentiokinetic reactivation (DL-EPR) test and an oxalic acid etch test were conducted to examine the susceptibility of specimens given a two-step heat treatment. In addition, to determine the cause of inducing Cr-depleted zones and propose the IGC mechanism occurring in the stabilized ultra-pure 430LX FSS, the precipitates at the grain boundaries were identified using a transmission electron microscope (TEM) with an energy dispersive spectroscope (EDS).

2. Experimental procedures

2.1. Material and heat treatment

The experimental material was the Ti-Nb-stabilized ultra-pure 430LX FSS provided by Baosteel. The chemical composition (in wt%) is given in Table 1. The experimental specimens were cut into 11 mm × 11 mm plates with a thickness of 0.7 mm. The specimens were solution treated at 1200 °C for 1 h in N₂ flow, then quenched in water [19,22,25]. After that, the specimens were aged at various temperatures (550, 600, 650, 700 or 750 °C, below the stabilizing treatment temperature) for various times (2 min, 10 min, 30 min, 1 h, 2 h or 8 h) respectively.

2.2. Electrochemical measurements

Electrochemical measurements were carried out with a PARSTAT MC electrochemical system, using a conventional three-electrode cell. The heat treated specimens were embedded in epoxy resin with an exposure area of 0.85 cm × 0.85 cm as working electrodes. Prior to electrochemical tests, the specimens were ground mechanically from 600 to 2000 grit SiC papers with water as lubricant, then polished with 2.5 μm diamond paste, rinsed with ethanol and dried in hot air. A platinum sheet was used as a counter electrode, and a saturated calomel electrode (SCE) as a reference electrode. All potentials presenting in this work refer to SCE.

DL-EPR test is a fast and non-destructive method to evaluate the susceptibility to IGC [5,[28], [29], [30]]. The tests were conducted in the solution of 0.5 mol/L H₂SO₄ with an addition of 0.0001 mol/L KSCN at 30 °C. This condition was well used to examine the susceptibility to IGC of 429 FSS in previous research [19]. The working electrodes were cathodically polarized at -0.7 V for 2 min, then stabilized at open circuit potential for 10 min. After that, the specimens were anodically polarized from -0.6 V to 0.3 V and then cathodically polarized to -0.6 V from 0.3 V at a scan rate of 0.1 V/min. To ensure reproducibility, DL-EPR tests for each specimen were repeated three time. The maximum current densities during the forward and reverse scans are referred to as the activation peak current density (I_a) and reactivation peak current density (I_r) respectively. The susceptibility to IGC is estimated as [19,22,31]

R_a = (I_r / I_a) × 100%.

To quickly and qualitatively identify the IGC attack, oxalic acid etch test was performed according to ASTM A763-93 Practice W standard. The specimens were etched at a current density of 1 A/cm² for 90 s in a 10% oxalic acid solution (100 g oxalic acid crystal dissolving in 900 mL deionized water). During the etch test, the specimens worked as the anode, and a stainless steel sheet was used as the cathode. Following etching, the specimens were rinsed in deionized water and ethanol, and then dried in hot air. The etched structures are classified as step, dual and ditch.

2.3. Microstructural characterization

The specimens' surface micrographs after IGC tests were characterized via optical microscopy (OM) and scanning electron microscopy (SEM). The precipitates at the grain boundaries in aged specimens were characterized using TEM. The TEM specimens were prepared in two steps. First, a thick slice was pre-thinned to 60 μm by grinding on successively finer SiC papers of 600-2000 grit. From such 60 μm thin sheet, 3 mm discs were punched out. The pre-thinned 3 mm discs of 60 μm thick specimens were then twin-jet polished in a mixture of 5% perchloric acid and 95% ethanol at -20 °C and 50 V potential to create a small hole approximately in the center of the disk. The chemical compositions of precipitates were analyzed using TEM-EDS. Selected area electron diffraction (SAED) patterns were acquired from the precipitates to identify them.

3. Results and discussion

3.1. Effect of heat treatment on the susceptibility to IGC

Fig. 1 shows the DL-EPR curves of the as-received and solution treated specimens. From Fig. 1, typical DL-EPR curves are seen with significant activation current density peaks and small reactivation current density peaks. During the anodic scan, the entire surface is activated to form the activation current density peak. Then the current density decreases to 1 × 10^-4 A/cm² within a wide passivity range from -0.15 V to 0.3 V, in which a thin passive film mainly composed with chromium and iron oxides can form on the specimen's surface. The reactivation current density peak in the reverse scan is attributed to the Cr-depleted zones. In addition the peak current densities of activation and reactivation, I_a and I_r, do not occur at the same potential and the potential of I_a is about 0.25 V higher than that of I_r, attributed to the ohmic resistance drop. The difference between the potentials of I_a and I_r was also reported in previous studies [29,30,32]. The value of R_a = I_r/I_a can be defined as the degree of sensitization [33,34]. The R_a value of as-received specimen is 4.46%, and the R_a value of solution treated specimen increases a little to 6.02%. The increase of R_a values indicates that solution treatment partly increases the specimens' susceptibility to IGC.

View original graphic| Download| PPT slide

Fig. 1.   DL-EPR curves of as-received and solution treated specimens.

Fig. 2 presents the corresponding specimens' surface micrographs after DL-EPR and oxalic acid etch tests. From the surface micrographs after DL-EPR test in Fig. 2(a) and (c), no obvious IGC attack is found in as-received and solution treated specimens. But the grain size increases from dozens of micrometers to hundreds of micrometers due to recrystallization during solution treatment [35,36]. According to the microstructural results of oxalic acid etch test in Fig. 2(b) and (d), there is also no ditch at grain boundaries, indicating that the both specimens exhibit good resistance to IGC.

View original graphic| Download| PPT slide

Fig. 2.   Specimens' surface micrographs after IGC tests: (a) as-received specimen after DL-EPR test, (b) as-received specimen after oxalic acid etch test, (c) solution-treated specimen after DL-EPR test, and (d) solution-treated specimen after oxalic acid etch test.

In order to further investigate the effect of aging time and aging temperature on IGC, the specimens were aged at 550-750 °C for various times. Fig. 3 shows the DL-EPR curves of aged specimens. All aged specimens also present a wide passivity range from -0.15 V to 0.3 V, with anodic current density close to 1 × 10^-4 A/cm². It should be noted that the maximum current density in the anodic scan is almost independent of the susceptibility to IGC, which is in good agreement with the results proposed by Reichert and Stoner [37]. However, the magnitude of the reactivation peak current density increases to values as high as 0.017 A/cm² for severely sensitized specimens from 0.0042 A/cm² for as-received specimen, originating from attack on the Cr-depleted zones and not on the Cr-rich precipitates. Fig. 4 shows the variation of R_a values calculated from DL-EPR curves with aging time at various aging temperatures. As seen in Fig. 4, it can be found that R_a values of the specimens aged at all temperatures (550, 600, 650, 700 or 750 °C) firstly increase, and then decrease with the increase of aging times, indicating that the susceptibility to IGC firstly increases and then decreases. However, the aging time at which the highest R_a values occur at different temperatures are varying in Fig. 4. At low temperatures, the specimens need a long aging time to reach the maximum of R_a values, for example, 1 h at 550 °C, 30 min at 600 and 650 °C. At the higher temperatures, such as 700 and 750 °C, the highest R_a values occurs at just 10 min.

View original graphic| Download| PPT slide

Fig. 3.   DL-EPR curves of the specimens aged at various temperatures for various times: (a) 550 °C, (b) 600 °C, (c) 650 °C, (d) 700 °C, and (e) 750 °C.

View original graphic| Download| PPT slide

Fig. 4.   Variation of Ra values calculated from DL-EPR curves with aging time at various aging temperatures.

Fig. 5 shows R_a values and the corresponding surface micrographs of the specimens aged at 650 °C for various times after DL-EPR and oxalic acid etch tests. For the specimens aged at 650 °C, when aging time increases from 10 min to 30 min, R_a values increase from 16.52% to 24.94% and evident attacks occur at the grain boundaries in the surface micrographs. As the aging time increases further to1 h, 2 h and 8 h, R_a values decrease from 13.03% to 7.83% and only very light attacks at the grain boundaries are observed. Fig. 6 shows R_a values and the corresponding surface micrographs of the specimens aged at various temperatures for 30 min after DL-EPR and oxalic acid etch tests. For the specimens aged for 30 min, when aging temperature rises from 550 to 600 °C, R_a values increase rapidly from 12.60% to 26.22% and wider IGC attacks occur at the grain boundaries. However, when aging temperature rises further from 650 to 750 °C, R_a values decrease from 24.94% to 10.70% and the IGC attacks at the grain boundaries gradually narrow. It can be seen that the change of R_a values is in good accordance with that of aged specimens' surface micrographs.

View original graphic| Download| PPT slide

Fig. 5.   Comparison of R_a values and surface micrographs of the specimens aged at 650 °C for various times after DL-EPR and oxalic acid etch tests.

View original graphic| Download| PPT slide

Fig. 6.   Comparison of R_a values and surface micrographs of the specimens aged at different temperatures for 30 min after DL-EPR and oxalic acid etch tests.

Based on the aged specimens' microstructural results of oxalic acid etch test in Fig. 5, Fig. 6, the specimens aged at 650 °C for 10, 30 min and 600 °C for 30 min exhibit a ditch structure. The grain boundaries are corroded severely, and all grains are completely surrounded by ditches. The specimens aged at 650 °C for 1 h, 550 and 700 °C for 30 min exhibit a dual structure, mixed with step and ditch structures. The specimens aged at 650 °C for 2 h, 8 h and 750 °C for 30 min show a step structure. The grain boundaries only show some steps, and no ditches is observed. It can be seen that the specimens exhibit step structure with the values of R_a below 12%. The specimens exhibit ditch structure with the values of R_a above 15%. The specimens exhibit dual structure with the values of R_a between 12% and 15%. Therefore, the critical value of the susceptibility to IGC for aged specimens is approximately 15% under the experimental condition used in this work. This method, defining the critical value of the susceptibility to IGC, was widely used in the previous researches [28,38].

To further reveal the correlation among aging time, aging temperature and the susceptibility to IGC, time-temperature-sensitization (TTS) diagram is plotted in Fig. 7, derived from the critical value of R_a = 15%. The sensitive region of IGC was indicated as dark region in Fig. 7. The sensitive region exhibits some features of the stabilized ultra-pure 430LX FSS's IGC behavior. Firstly, IGC occurs in the specimens aged at 600-750 °C for a short time. Secondly, when aging treatment exceeds a certain period of time, the specimens regain the resistance to IGC. Thirdly, the higher the aging temperature, the less time it takes to induce the susceptibility to IGC, but the earlier the specimens regain the resistance to IGC. Finally, it can be speculated that IGC might also occur within the temperature range of stabilization treatment if the time is not long enough.

View original graphic| Download| PPT slide

Fig. 7.   TTS curve of the stabilized ultra-pure 430LX FSS.

3.2. Characterization of precipitates

Fig. 8 shows SEM micrographs of the specimens aged at 650 °C for 30 min and 8 h after oxalic acid etch test. The grains in the specimen aged for 30 min are completely surrounded by ditches due to the selective dissolution of chromium carbides, indicating the presence of large quantities of chromium carbides at the grain boundaries [28,38,39]. However, no ditch is observed at grain boundaries in the specimens aged for 8 h, indicating good resistance to chromium carbide-type intergranular attack. In addition, some precipitates with a size of a few micrometers are observed in the grains of the specimens aged for both 30 min and 8 h. The precipitates in the grains were identified as (Ti,Nb)N, mainly TiN, which were also reported in the previous researches [18,[24], [25], [26]]. They precipitated during the steel production processes such as hot rolling and annealing and still remained stable even during high temperature solution treatment [18,25,40].

View original graphic| Download| PPT slide

Fig. 8.   SEM micrographs and EDS analysis results of the specimens aged at 650 °C for 30 min and 8 h after oxalic acid etch test: (a) SEM micrograph of the specimen aged for 30 min, (b) EDS analysis result of 'A' marked in (a), (c) SEM micrograph of the specimen aged for 8 h, and (d) EDS analysis result of 'B' marked in (c).

The intergranular precipitates were further analyzed using TEM. Fig. 9 shows the TEM micrographs, EDS analysis results and SAED patterns of the precipitates along the grain boundaries in the specimens aged at 650 °C for 30 min and 8 h. The sites 'A' and 'B' in Fig. 9(a) and (d) represent the intergranular precipitates, respectively. Rod-liked precipitates (site 'A') with a length of approximately 200 nm are observed in the specimens aged at 650 °C for 30 min. As aging time increases to 8 h, the precipitates become nanoscale spheroidal particles (site 'B'), with a diameter of approximately 20 nm. The EDS result in site 'A' shows significantly high contents of chromium, iron and carbon, but low contents of titanium and niobium. As such, these precipitates should be the compounds of chromium, iron and carbon. On the basis of the SAED pattern analysis, the precipitates were identified as M₂₃C₆ with the zone axis of [2-2-3]. This result is similar to the TEM result in the previous research by Qiang et al. [22], in which M₂₃C₆ precipitates were observed along the grain boundaries in the Ti-stabilized 409 L FSS. On the other hand, the EDS result in site 'B' shows high contents of carbon, nitrogen and niobium. On the basis of the SAED pattern analysis, the precipitates were identified as Nb(C, N) with the zone axis of [0 -11]. The MC precipitates along the grain boundaries were also reported in the previous researches on the Ti-stabilized and Ti-Nb-stabilized FSSs after a long time sensitization [18,[24], [25], [26], [27]].

View original graphic| Download| PPT slide

Fig. 9.   TEM micrographs, EDS analysis results and SAED patterns of the precipitates along the grain boundaries in the specimens aged at 650 °C for 30 min and 8 h: (a) TEM micrograph of the intergranular precipitates in the specimen aged for 30 min, (b) EDS analysis result of 'A' marked in (a), (c) SAED pattern of 'A' marked in (a), (d) TEM micrograph of the intergranular precipitates in the specimen aged for 8 h, (e) EDS analysis result of 'B' marked in (d), and (f) SAED pattern of 'B' marked in (d).

3.3. IGC mechanism of the stabilized ultra-pure 430LX FSS

For the stabilized ultra-pure 430LX FSS, M₂₃C₆ precipitates firstly form at the grain boundaries in a short time, and the susceptibility to IGC increases. However, with the increase of aging time, intergranular precipitates translate into Nb(C, N), and the susceptibility to IGC decreases simultaneously with the change of intergranular precipitates. A possible IGC mechanism of the stabilized ultra-pure 430LX FSS can be concluded as follows. At first, carbon atoms are released from carbides and disperse in the matrix during solution treatment [17,25] as illustrated in Fig. 10(a). During aging treatment, carbon atoms first diffuse to the grain boundaries because of their high diffusion speed, as shown in Fig. 10(b). These two processes were proposed in the previous research by Kim et al. [25]. Dissimilarly, most carbon atoms react with chromium atoms, not the atoms of stabilizing elements because the amount of titanium and niobium atoms around the grain boundaries is very small at the early of aging treatment, although carbon atoms preferentially react with them. M₂₃C₆ precipitates form at grain boundaries and lead to the consumption of chromium atoms and the appearance of Cr-depleted zones. Fig. 10(c) illustrates the development of Cr-depletioned zones. Thermodynamically, the atoms of stabilizing elements have much stronger affinity for carbon atoms than chromium atoms [14,41]. As aging time is further prolonged, the atoms of stabilizing elements diffuse to grain boundaries from the crystals. They substitute chromium atoms, combine with carbon atoms and finally form MC precipitates. At that time, the chromium atoms replaced by the atoms of stabilizing elements and dispersing in the crystals diffuse to Cr-depleted zones, improve the level of chromium in Cr-depleted zones and thus decrease the susceptibility to IGC. Therefore MC precipitates remain near the grain boundaries, as shown in Fig. 10(d).

View original graphic| Download| PPT slide

Fig. 10.   Schematic diagram illustrating the IGC mechanism of the stabilized ultra-pure 430LX FSS.

To date, commonly accepted IGC mechanism for the stabilized ultra-pure FSS is mainly divided into two types: the Cr-rich compound theory and the chromium segregation theory. The Cr-rich compound theory is still based on the precipitation of Cr-rich compounds along the grain boundaries. The precipitation of Cr-rich compounds induces Cr-depleted zones, and thus leads to IGC [[20], [21], [22]]. On the other hand, the chromium segregation theory suggests that with the addition of stabilizing elements, carbon atoms react with their atoms at the grain boundaries since stabilizing elements have much stronger affinity for carbon than chromium. At that moment, chromium atoms have a tendency to segregate at the MC/matrix interface to relieve supersaturation. Consequently, the chromium segregation develops Cr-depleted zones and leads to IGC [18,[24], [25], [26], [27]].The results of this study support the Cr-rich compound theory rather than the chromium segregation theory, and even suggest that the formation of MC precipitates along the grain boundaries decreases the susceptibility to IGC.

4. Conclusions

The IGC behavior of the Ti-Nb-stabilized ultra-pure 430LX FSS was investigated by using DL-EPR and oxalic acid etch tests. The precipitates at the grain boundaries were identified using TEM with EDS. The following conclusions are made from the present study.

(1) The results of DL-EPR and oxalic acid etch tests indicate that IGC occurs in the specimens with Ra value over 15% and the results from two different IGC tests are consistent with each other.

(2) IGC occurs in the specimens aged at the temperature range of 600-750 °C for a short time. The aging time that is required to cause IGC decreases with the increase of aging temperature. Long-time aging treatment reduces the susceptibility to IGC.

(3) The M₂₃C₆ precipitates form along grain boundaries, causing Cr-depleted zones. The presence of Cr-depleted zones results in the susceptibility to IGC. However, after long-time aging treatment, stabilizing elements replace chromium in the intergranular precipitates to form MC precipitates, resulting in the chromium replenishment of Cr-depleted zones and the reduction of the susceptibility to IGC.

Acknowledgments

The authors acknowledge financial support from the National Key Research and Development Program of China (No. 2018YFB0704400) and the National Natural Science Foundation of China (Nos. 51501041, 51871061 and 51671059).

The authors have declared that no competing interests exist.