Mitigating the susceptibility to intergranular corrosion of alloy 625 by friction-stir welding

In this work, friction-stir welding (FSW) was employed to alloy 625 grade I (soft annealed) sheets. Therefore, solid-state based welding was undertaken with a tool rotational speed of 200 rpm and a welding speed of 1 mm/s. Microstructural features were analyzed by light optical and scanning electron microscopy (LOM and SEM). Moreover, microhardness measurements were performed. The susceptibility to intergranular corrosion was verified by the double-loop electrochemical potentiokinetic reactivation (DL-EPR) test. Complementary, intergranular corrosion was also evaluated by the ASTM G28 Method A. FSW promoted grain refinement, increased microhardness, and reduction in the degree of sensitization. Finally, the mean corrosion rate observed in the ASTM G28 Method A test was 0.4406 mm/year, which suggests a good weld quality.


Scientific Reports
| (2022) 12:3482 | https://doi.org/10.1038/s41598-022-07473-0 www.nature.com/scientificreports/ shown FSW in nickel-based alloys 1,2,9,[12][13][14][15][16]23 , where pcBN tool wear is often reported. It remains unknown in investigating the intergranular corrosion properties of friction-stir-welded alloy 625, as indicated in 1 . Intergranular corrosion of various alloys with a passive surface layer has been analyzed over time by distinct corrosion tests 17 . The first electrochemical tests were named after their inventors (Streicher, Strauss and Huey), as mentioned in 18 . Overall, if the investigated condition is sensitized, grain dropping would occur 19 . Aiming at substituting the long-term assays, the double-loop electrochemical potentiokinetic reactivation (DL-EPR) test was gratefully developed and accepted 18,20,21 . The main advantage of the DL-EPR test is to obtain fast results for the degree of sensitization (DOS). In this sense, nickel-based alloys have also been evaluated by DL-EPR, with some improvements in the test 17,22 . However, results of the DL-EPR tests in friction-stir-welded alloy 625 have not been found. Still, the intergranular corrosion using the ASTM G28 Method A remains to be explored in friction-stir-welded alloy 625, as proposed in the current investigation.
As intergranular corrosion is one of the most important degradation mechanisms of alloy 625 5 ; it must also be evaluated in friction-stir welds of this nickel-based alloy. Therefore, the current work focuses on mitigating the susceptibility to intergranular corrosion of alloy 625 (grade I soft annealed) through the FSW process.

Materials and methods
Friction-stir-welded alloy 625 sheets (300 mm × 80 mm × 3.2 mm) were investigated in this work, and their chemical composition is given in Table 1 (data from the supplier). Moreover, alloy 625 was in the soft annealed condition (grade I), which has exceptional corrosion resistance in varied corrosive media 23 . Also, this nickelbased alloy is often chosen for cladding rigid pipes due to its outstanding corrosion properties.
Alloy 625 sheets were prepared by cleaning the surfaces by abrasive sanding and acetone, aiming at the removal of grease, oxides, and impurities. This step was recommended to increase the surface quality of the joint 2 . Afterwards, the sheets were fixed and aligned on the welding table. Welded joints were processed in the butt joint configuration on a rigid gantry machine, which has servomotors and automated control systems. A polycrystalline cubic Boron Nitride (pcBN) Q70 tool, with W-Re (binder phase), a shoulder of 25-mm diameter, and a probe of 3-mm long was used. The tool was tilted 1.5° to the vertical axis to obtain the welded joint. Also, welding was performed with an argon atmosphere to reduce oxidation. FSW was carried out with the best process condition (tool rotational speed of 200 rpm and welding speed of 1 mm/s), as done elsewhere 2,9,12,24 . These parameters can minimize wear of the pcBN tool. Details on pcBN tool wear in FSW of alloy 625, recently investigated by an experimental and numerical approach, can be found in 24 .
For microstructural analysis, the samples were cut by electrical-discharge machining (EDM), prepared by conventional metallography practices (sanding and polishing), and etched with glyceregia (10 ml HCl, 10 ml HNO 3 , and 0.5 ml glycerin). Thus, the microstructural features were observed by light optical microscopy (LOM) and scanning electron microscopy (SEM).
Local mechanical properties were evaluated in terms of microhardness. The Vickers microhardness profile measured on the top surface was obtained with a 500 gf load and distance between indentations of 0.2 mm.
The investigation of the degree of sensitization was based on the DL-EPR test, which is schematically shown in Fig. 1. The parameters of this electrochemical test were taken from 17 , where the regions chosen were firstly anodically polarized to the metal passivation zone and then scanned in the reverse direction. In other words, when a region is anodically polarized, it is theoretically covered by a passive layer. Next, when the scan is reversed, www.nature.com/scientificreports/ the passive film breaks in the metal-depleted areas. These metal-depleted areas (diminished Cr or Mo content) tend to be preferentially attacked due to anodic polarization effects that can create uneven surfaces 25 or even a thinner and less protective oxide layer 26 , so that these act as preferential sites for the depassivation. Hence, the selected regions investigated were the base material (BM) and the stir zone (SZ). The susceptibility to intergranular corrosion through weight loss was carried out in accordance with ASTM G28 Method A 27 . Two samples (12.5 × 50 × 3.2 mm) were machined by EDM, sanded and polished. The weight change was verified after the tests. A solution of 25 g FeSO 4 , 236 ml H 2 SO 4 , and 400 ml deionized water was used. The duration of the tests was 120 h (starting from the time that the solution reached the boiling point). Hence, the corrosion rate was calculated via Eq. (1), as follows: where the constant was 86.760 and the density of alloy 625 was 8.44 g/cm 35,27 . Furthermore, the samples tested were observed by LOM.
A schematic sketch (Fig. 2) is therefore detailing the samples for both corrosion tests (DL-EPR and ASTM G28 Method A).

Results and discussion
Base material. The base material microstructure is shown in Fig. 3. Therefore, alloy 625 grade I (soft annealed) has an amount of M(C, N) carbonitrides (Fig. 3a), M 6 C (Fig. 3b), and M 23 C 6 carbides (Fig. 3c). In this context, M(C, N) carbonitrides were mainly located in the matrix, with the metal being Nb or Ti. In addition, M 6 C and M 23 C 6 carbides, typically rich in Mo and Cr, were observed at isolated grain boundaries, as prior reported in 2,28 . In general, when etching with glyceregia, the matrix around M 6 C carbide is less etched than the matrix around M 23 C 6 carbide, as shown in Fig. 3b,c. Figure 3c shows an M 23 C 6 carbide (0.25 μm) at the grain boundary where some grooves around it can be observed, and thus neighboring metal-depleted zones may be more susceptible to intergranular corrosion. As these carbides are very small, the EDS result can be influenced by the interaction of the electron beam with alloy 625 matrix due to the beam diameter, which is larger than the carbide size. Furthermore 17 , also verified microstructural features as MC, M 6 C, and M 23 C 6 , suggesting that they evaluated a similar alloy condition.
In 17,19,28 , it has been indicated that Mo-depleted zones strongly influence the corrosion resistance, which can be understood by the pitting resistance equivalent number (PREN) calculated using Eq. (2). Thus, Mo-rich phases could result in lower local PREN in their depleted adjacent zones. In this work, as M 6 C and M 23 C 6 carbides were observed in the soft annealed condition, and, since the latter group was mostly at grain boundaries, it seemed to have the main effect on sensitization and further intergranular corrosion.
Friction-stir-welded alloy 625. The macrostructure on the top surface of friction-stir-welded alloy 625 is shown in Fig. 4a, where small-sized grains are verified through a darker microstructure (next detailed in Fig. 4b). Therefore, an MC type carbonitride (metal being Ti) was noted as well as a remarkable grain refinement of the austenite microstructure, which is typically accomplished in FSW of alloy 625 2,9 . In addition, no significant quantity of M 6 C and M 23 C 6 carbides were observed by SEM imaging. Thus, it seems that most of the M 6 C and M 23 C 6 carbides at grain boundaries (or even all) were fragmented and dissolved due to the stirring action of this severe deformation process (SDP). Furthermore, EDS analysis indicated tungsten (W) content that is not an alloying element of the alloy 625 base material and hence it came from the pcBN tool wear. In our recent work 24 , it was demonstrated that the average %W in the SZ (on the top surface) of this joint was 0.27%wt (lowest value among the FSW conditions evaluated), which represents minimal tool wear. www.nature.com/scientificreports/ Figure 5 presents the time-temperature-sensitization diagram of alloy 625 grade I (soft annealed) related to corrosion loss (mm/year) when tested according to ASTM G28 Method A 23 , along with the thermal cycles measured in the current investigation (temperatures at 15 mm distance from the weld centerline). With the soft annealed condition, sensitization of alloy 625, regarding the 50 µm criterion, could occur at around 750 °C after approximately 3 h 23 . In this work, the maximum temperature verified on the advancing side (AS) (513 °C) was higher than that on the RS (retreating side) (436 °C) 28 . Moreover, it is interesting to note that the FSW process would not lead to considerable sensitization in alloy 625 (temperatures reached were outside of the "C curve" for substantial sensitization). Otherwise, FSW tends to promote cleaner microstructure compared to the alloy 625 (grade I) base material, thus diminishing the susceptibility to sensitization. The cleaner microstructure in the SZ concerning the carbides at grain boundaries is a consequence of the recrystallization during FSW. Thus, the movement of recrystallization interfaces promoted the dissolution of M 23 C 6 and M 6 C carbides. www.nature.com/scientificreports/ The Vickers microhardness profile measured on the top surface can be seen in Fig. 6. Therefore, alloy 625 grade I (base material) had microhardness values ranging from 250 to around 280 HV. In the stir zone (SZ), microhardness values up to approximately 340 HV were achieved. Thus, the increased microhardness in the weld zone is commonly explained by finer grains obtained through FSW (clearly seen in Fig. 4). Recently, we suggested that an achievement of small-sized grains due to the FSW process can be related to the dynamic recrystallisation coming from the interaction between two variables (plastic flow and heat generated), which are sufficient to thus drive the recrystallization process 28 .
In Fig. 7, the differences between the peak activation current density (Ia) and the peak reactivation current density (Ir) indicate a distinct electrochemical behavior among the passivation and depassivation for each region evaluated. Moreover, a significant variation in the Ir/Ia ratio between the base material (BM) and the stir zone (SZ) was verified. Thus, it was noted that alloy 625 grade I soft annealed achieved a degree of sensitization of 1.23. On the other hand, FSW promoted a clean resulting microstructure concerning the carbides that would influence sensitization, a fact that caused a lower degree of sensitization (0.66) in SZ. This has good agreement with the findings shown in Figs. 4 and 5. Other studies reported similar outcomes, also indicating that FSW can  www.nature.com/scientificreports/ be beneficial to mitigate the susceptibility to sensitization of distinct metals (AA 5083 aluminum alloy 30 and 304 stainless steel 31 ). The DL-EPR results indicate that although the grain refinement may reduce the intergranular corrosion resistance; the presence of M 23 C 6 and M 6 C carbides is suggested to be more deleterious for the intergranular corrosion, resulting in an SZ with improved intergranular corrosion resistance. The corrosion rate evaluated by the ASTM G28 Method A test is presented in Table 2, where both samples showed similar weight losses. Thus, the average corrosion rate was found to be of 0.4406 mm/year. According to 5 , the error of this test is ± 0.05 mm/year. In addition, some authors mentioned that the threshold for sensitization in the industry is 1 mm/year 13 , thus an acceptable friction-stir-welded alloy 625 was obtained. It is worth noting that authors 32 showed that Inconel 625, produced by wire-arc additive manufacturing (WAAM), reached an average corrosion rate of 0.609 mm/year (slightly higher than that of the current work), which was then suggested to have an excellent resistance to intergranular corrosion. Furthermore, for Inconel 686, a nickel-based alloy that was processed by fusion welding processes such as gas metal-arc welding (GMAW) and gas tungsten-arc    33 . Overall, based on the above observations, a suitable welded joint of alloy 625 was produced by solid-state welding (FSW process). Figure 8 shows the micrographs after using ASTM G28 Method A. As this assessment comprises the weight loss, and the BM had a certain amount of M 6 C and M 23 C 6 at some grain boundaries, and these carbides are the most relevant to intergranular corrosion, it is assumed that the BM lost more weight than the SZ. Thus, it was noted that the BM (Fig. 8a) achieved a worse behavior than the SZ (Fig. 8b) regarding intergranular corrosion, which agrees well with the outcomes from DL-EPR test (Fig. 7). Therefore, the BM had the highest degree of sensitization and weight loss. In addition, in this test, grain dropping would take place if the alloy condition has a certain level of sensitization 19 , as here seen in the BM. In contrast, the SZ showed improved resistance to intergranular corrosion. Moreover, intergranular boundaries were not deeply corroded in the SZ compared with that of the BM. In other words, FSW thus mitigated the susceptibility to intergranular corrosion of alloy 625 grade I soft annealed. In a similar way, for FSW in AISI 304 34 , it was also revealed that the SZ had better corrosion properties than the BM.

Conclusions
This work presented friction stir welding (FSW) for mitigating the susceptibility to intergranular corrosion of alloy 625 grade I (soft annealed). The results of the present investigation can be summarized as follows: 1. FSW led to the formation of small-sized grains and increased microhardness in the stir zone. 2. Double-loop electrochemical potentiokinetic reactivation (DL-EPR) tests showed that FSW decreased the degree of sensitization of alloy 625, which is mainly related to the absence of carbides at grain boundaries (verified by SEM) that would influence sensitization. 3. The mean corrosion rate verified with the intergranular corrosion ASTM G28 Method A was 0.4406 mm/ year, which agrees with DL-EPR findings and suggests a good weld quality. 4. The presence of M 23 C 6 and M 6 C carbides at grain boundaries is more deleterious for intergranular corrosion resistance. Therefore, the FSW process diminished the susceptibility to intergranular corrosion of alloy 625 grade I soft annealed. Furthermore, if a nickel-based alloy achieves a high level of sensitization due to fusion welding, manufacturing processes, or even service time-temperature conditions, it can be mitigated by using FSW.