Induction bending effects on mechanical properties and corrosion resistance of duplex stainless steel UNS S31803 pipes

Duplex stainless steels are well known austenitic-ferritic alloys which combines good mechanical resistance and excellent corrosion resistance. Due to this fact, this steel family is widely employed in offshore piping applications. However, these alloys require great care in thermomechanical processing, such as welding, since they might present serious decrease in corrosion resistance and toughness at the welded joint when compared to the base metal. Undeniably, welding is one of the major processes employed in pipe and industrial equipment assembly, thus requiring higher expertise and caution to achieve the desirable good combination of properties of duplex stainless steel alloys. In this context, induction bending process comes as an efficient alternative to reduce the number of weld joints in piping designs. In this work, induction bent tubes were characterized after hot bending in temperatures inside and outside the standard recommended range, that is, between 950 and 1150 °C. Microstructural characterization, mechanical testing, and corrosion tests were performed, indicating that the properties of curved materials remained comparable to the solubilized state. Therefore, the electromagnetic induction bending process proves to be not only a viable tool in thermomechanical processing of these materials, but also more productive.


Introduction
Duplex stainless steel (DSS) are corrosion resistant alloys with high mechanical properties when compared with conventional austenitic stainless steel grades, because of the higher Cr, Mo, and N content and a fine biphasic microstructure, which requires a roughly equal proportion of ferrite and austenite to obtain the best combination of mechanical properties and corrosion resistance [1]. Thus, this alloy family is employed mainly in petrochemical and offshore industries, where a continuous production operation at critical conditions is required [2], but can also be found in naval, chemical, paper-pulp, and nuclear industries [3,4].
Thus, DSS are materials frequently selected for pipes and fittings such as tees and elbows, pressure vessel, and other process equipment operating in harsh environments. Fabrication and assembly of such equipment usually requires hot forging, hot rolling, and welding. These thermomechanical processes, specially welding operations, when poorly selected might lead to deleterious phase precipitation [3,5,6], resulting in loss of toughness and corrosion resistance. Even when the processing parameters are considered satisfactory, it is still possible to note a slight decrease in toughness and corrosion resistance at the heat affected zone (HAZ) or weld metal (WM), when compared to the base metal (BM) [7].
Curves produced by hot induction bending of seamless pipes may be used in piping construction as an alternative to reduce the number of welded joints. The conventional method of spool construction uses forged curves which are welded to straight parts of tubes. Welding implies in several operations such as cutting, beveling and positioning with spot welds, and multipass welding. If the curves can be produced by hot bending, the productivity can be significatively enhanced. Also, the absence of a melted zone is an advantage over welding operations. It is worth noting that the construction with bended tubes must be designed from the beginning of the project because there are limitations in respect to the diameter/ external radius relation of the curves. Previous works [8][9][10][11][12] had already explored parameter selection and major causes of defects and instabilities in carbon steel hot pipe bending, such as wrinkling, cross-section distortion, and wall thinning. Many authors [13][14][15] also focused in microstructural features and eventual properties changes related to carbon steel hot pipe bending. In the case of duplex and superduplex spools, the thermal cycle must be controlled to avoid deleterious phases and to maintain the mechanical and corrosion resistance properties. One of the recommendations in literature is water quenching the pipe to reduce time at deleterious phase formation temperature [16]. Once the parameters are adjusted, the process is automated, and the reproducibility is very high.
In the present work, the microstructural features, the mechanical properties, and corrosion resistance of seamless pipe hot bended of duplex stainless steel UNS S32205 steel were investigated.

Materials and equipment
This study was performed in the as received (AR) DSS UNS S31803 seamless pipe with 3½ in (88.9 mm) diameter and 19.0 mm wall thickness, after solution treatment condition, which chemical composition is presented at Table 1. This material is also referred, when delivered in the wrought condition, as AISI 2205, EN 1.4462, BS 318S13, and X2CrNiMoN22-5-3, depending on the standard. Some of the samples where then induction bent (solution treated and bent, identified as STB) and others remained at solution treated (ST) condition for comparison purposes. Temperatures of 950 °C, 1060 °C, and 1150 °C were selected to simulate the influence of an occasional fluctuation below and above the 1020 ~ 1100 °C range specified in ASTM A790 [17], and three different speeds were also evaluated: 0.2, 0.5, and 0.8 mm/s. To compare the hot bending and solution treated conditions, some samples were induction heated at the same solution treatment temperature for three different treatment times, which were based on the average duration necessary to perform the induction bending procedure for each speed.
The bent machine used was a high frequency bending machine Dai-ichi High Frequency Co. Ltd model HR-28 GB, with incorporated Yukon automation system. In order to achieve the selected temperatures, the frequency, tension, and current applied varied between 1.5 and 1.8 kHz, 400 and 700 V, and 6.8 and 8.5 kA, respectively. The induction coil, presented in Fig. 1, had 28 mm of width with 41 mm of lift off. The temperature was monitored by an infrared pyrometer and the cooling was carried out by water spray at room temperature with a pressure of 0.1 kgf/cm 2 . Figure 2 shows the DSS tube curved to 90° with a 4D radius after the induction bending process.

Testing conditions
Three selected bending speeds of 0.2, 0.5, and 0.8 mm/s were analyzed at 950 °C, 1060 °C, and 1150 °C. These bending speeds correspond to a strain rate (mm·s −1 /mm) of 0.001,

Specimen removal
The test samples were removed from the cross-section area, as shown in Fig. 9, for the solubilized condition in AR and ST conditions. For induction bent specimens (STB), the samples were removed from the extrados of the curved section to perform microstructural characterization.

Metallographic characterization
Metallographic characterizations by Beraha, KOH, and oxalic acid were performed to reveal phase proportions

Mechanical property evaluation
Brinell hardness tests were performed at the pipe surfaces with a King Portable Brinell according to ASTM E10 [18]. After proper preparation of the surface by sanding, a load of 3000 kgf was applied with a 10 mm diameter ball. Three Brinell hardness measurements were made to calculate an average value. For conditions AR and ST, random regions were considered, but for STB, the measurements were taken by the neutral line of the curved section. The Charpy impact test was conducted at − 46 °C according to ASTM E23-18 [19], using an impact machine model JBW-300. The specimens with dimensions 55 mm × 10 mm × 10 mm with V-notch were removed in the transverse direction, with the opening of the notch in the radial direction and positioned on the external surface of the straight tubes. The external surface of the extrados was employed for bent tubes.

Critical pitting temperature
The critical pitting temperature (CPT) measurements were based in ASTM G150 requirements [20], using a solution of 1 mol of NaCl, without previous deaeration. The solution was heated from 10 °C at a rate of 6 °C/min. Temperature was verified manually at intervals of 15 s, using a digital thermometer. The Omnimetra® PG-3901 galvanostat potentiostat was used to apply a potential of 700 mV SCE to the sample. For each sample studied, three tests were performed and the average of the CPT values was calculated. If the CPT values differed more than 5 °C from each other for a single given sample, a retest was carried out to ensure a better result.  Figure 10 shows the micrographs for the DSS along the tube's length. It is possible to verify the sporadic presence of large austenite islands, as well as large ferrite islands. Therefore, it can be said that the phase distribution along the length of the straight tube in the as received condition is not perfectly homogeneous. No intermetallic phases were found.

Microstructural characterization
In Fig. 11, it shows the solution treated pipe in three different temperatures for three holding times, totaling nine solubilizing combinations. It is possible to note the presence of large random austenite islands at temperatures of 950 and 1060 °C. In solubilized samples at 1150 °C, the presence of the darker phase, ferrite, is more evident, which indicates a Fig. 10 Microstructures of AR condition over pipe length. Letter "F" indicates the ferritic darker phase and "A" the austenite Fig. 11 Micrographs of the solution treated pipe by induction heating bigger concentration of ferrite at higher temperatures A similar microstructure observation was verified by Jiang et al. [21], which submitted the same DSS in solution treatments between 1000 and 1200 °C during 24 h. No intermetallic phases were detected with Beraha solution. Figure 12 shows the micrographs for the induction bent pipes at three temperatures and three different speeds, totaling nine combinations of bending. The micrographs of curved samples indicated in Fig. 12 show to be very similar to each other, as noted by [22]. It was shown that the austenitic islands found themselves slightly smaller in curved samples with speed of 0.8 mm/s at the three tested temperatures, corroborating the statements of [23] that high heating rates prevent the grain growth and induce the formation of an ultra-fine microstructure after subsequent rapid cooling. Relatively large islands and random austenite were observed in curved samples, as well as in solubilized samples. The STB samples did not present a higher ferrite content at 1150 °C, oppositely to ST ones.
The results for the chromium nitride analysis are presented in Fig. 13, for both the solution treated and induction bent specimens at 1150 °C. It is possible to observe It is important to note that the higher the solution treatment time at 1150 °C, the higher the concentration of chromium nitrides. This is explained by the lower concentrations of austenite, which favors the segregation of nitrogen in the ferrite phase and the precipitation of Cr 2 N when cooling. Similar results were also found by [16,24,25]. The bent tubes (STB) showed a lower amount of chromium nitrides precisely for this reason, as can be seen comparing Figs. 11 and 12, where there was a more prominent presence of the ferrite at 1150 °C for the ST samples. Figure 14 shows the ST sample at 950 °C for 24 min and attacked with KOH for sigma phase detection. As can be noted at the referred figure, there were no signs of sigma phase. This result was corroborated by [16,26], where only treatments for periods longer than 10 h started to show appreciable sigma phase amounts. Figure 15 shows the results for ferrite concentration (vol.%), CPT, and toughness (absorbed energy, J) of the solution treated samples at 950 °C for 3, 6, and 24 min. The dotted line, labeled AR in Fig. 15, identifies the as received condition. Error bars represent standard deviation. Figure 15 indicates that there were no significant variations in the ferrite concentrations and toughness in the 950 °C solubilized samples at all three holding times, when compared to the as received condition. The most obvious variation occurred in the sample that was solution treated for 24 min, where it is possible to note a slight decrease in the ferrite concentration and, proportionally, a slight increase in toughness. Regarding the CPT, the solution treated samples presented higher corrosion resistance when compared to the as received condition. Many studies [26,27] have demonstrated an inversion relation between the solution treatment time and CPT or impact energy for long periods. However, this has not been observed for durations of less than 1 h.

Phase quantification, CPT, and toughness measurements
In the case of Fig. 16, it shows the outcomes for toughness, ferrite concentration, and CPT for the bent samples at 950 °C in 0.2, 0.5, and 0.8 mm/s speeds. The dotted line, labeled as AR in Fig. 16, represents the pipe in the as received condition.
It is evident from the results of Fig. 16, except for the toughness in the bent sample at the speed of 0.5 mm/s, that the values are consistent with each other. The discrepancy in the stated measurement can be explained by the results presented in Fig. 10, where it can be clearly seen that the ferrite concentration varies in the as received condition. In the case of specimen bent in 0.5 mm/s speed, the percentage of ferrite could have varied greatly, causing the discrepant measurement when compared to the other samples. Nevertheless, there was no significant variation between the values of toughness from the other bent samples when compared to the sample as received. The same statement is true for the ferrite content. Finally, the curved samples showed slight increase in pitting corrosion resistance as it increased the bending speed. This result can be explained by the lower time spent at the temperature of 950 °C, preventing the precipitation of sigma phase. Figure 17 shows the results for toughness, ferrite concentration, and CPT of the solution treated samples at 1060 °C for 3, 6, and 24 min.
It is shown in Fig. 17 that the sample solution treated at 1060 °C for 6 min presented a CPT superior to the other measured, when it was expected to fall due to the reduced absorbed energy at impact test. However, the maximum difference among the CPTs of solubilized samples to 1060 °C was 10.9 °C which, in practical terms, is not significant. In the case of Fig. 18, it shows the results for toughness, ferrite concentration, and CPT for the bent samples at 1060 °C in 0.2, 0.5, and 0.8 mm/s speeds.
Through Fig. 18, it can be noted a small variation between the ferrite concentrations and toughness from the bent samples and the sample as received. However, it is possible to note a slight raise in pitting corrosion resistance with increasing bending speed. This is an advantage of induction bending as manufacture process, when compared with welding, as the latter usually reduces CPT, even when mechanical properties values were acceptable [7,28]. Moreover, the analysis of Figs. 18 and 19 shows that induction bending at 1060 °C tends to enhance pitting corrosion resistance, toughness and decrease ferrite content when comparing to solution treatments at the same temperature. Figure 19 shows the results for toughness, ferrite concentration, and CPT of the solution treated samples at 1150 °C for 3, 6, and 24 min.  Figure 19 shows that solubilized samples at 1150 °C presented higher values than the sample as received, which is justifiable due to the higher temperature. Both the absorbed impact energy and the CPT showed lower values than the sample as received, due to the presence of chromium nitrides in the ferrite phase at these samples, as presented in Fig. 13. Similar behavior was also verified by [24,25,29]. Summing up, when the same treatment time was considered, an increase in temperature resulted in higher ferrite content with a consequent CPT and toughness reduction.
Finally, Fig. 20 shows the results for toughness, ferrite concentration, and CPT for the bent samples at 1150 °C in 0.2, 0.5, and 0.8 mm/s speeds.
Through Fig. 20, it can be shown that the bent samples at 1150 °C showed CPT values higher than the sample as received, despite evidence of precipitated chromium nitride in the ferritic phase. When comparing the induction heating  Fig. 20, it comes clear that ST led to rise in ferrite content, which resulted in a reduced absorbed energy and CPT, while induction bending enhances these properties. Some works [15,16] has displayed that induction bending process could be deleterious to mechanical properties and corrosion resistance, being necessary a solution treatment at 1120 °C for at least 15 min to restore them. However, this work shows that solution treatments at higher temperatures during long times can also result in deleterious effects due to ferrite content increase and consequently chromium nitride precipitation. Furthermore, it was detected that induction bending improves CPT, toughness, and austenite content when compared with induction heating treatment.

Hardness measurements
The values of hardness measured for the ST and STB specimens are shown in Figs. 21 and 22, respectively. By comparing both graphs, it is possible to note that the hardness of the induction bent tubes is relatively smaller than that of the ST condition, regardless of temperature, bending speed, and solution time. This is because the specified bending rates suggest a very rapid time of exposure to the three temperatures studied, providing a short time for the samples to reach phase equilibrium at each temperature. This resulted in similar hardness values, all lower than AR, for all curved tubes, regardless of temperature. On the other hand, Collie et al. [15,16] have reported a hardness increase after induction bending, related majorly to deleterious phase precipitation. The values were higher than the ones allowed by standard, but a solution treatment after bending was sufficient to achieve acceptable values. All measurements were lower than the 290 HB, specified by ASTM A790 [17].

Conclusions
This work concludes that bending process in a duplex stainless steel tube can be employed satisfactorily in piping assembly during pre-fabrication stages. This statement is true because the hardness, CPT, phases balance, and toughness evaluated in 950-1150 °C temperature range did not change significantly when compared to the as received condition, and its values are all within the range specified by normative entities [17,30]. Based on the results presented, it is recommended a temperature of 1060 °C to perform the bending process. The increase in hot bending speed did not lead to higher hardness, and instead resulted in better corrosion resistance and a more balanced microstructure.
Comparing the results of this work with the welding process as employed in [7], the hot induction bending process resulted in higher toughness and corrosion resistance (CPT), making it a good alternative to be used in DSS piping and tubing assembly.
Author contribution All authors contributed to the study conception and design. Material preparation and data collection were performed by Larissa Ribeiro de Souza, Antônio Marcelo Meireles, and Maria Cristina Lopez Areiza. The data analysis was performed by Pedro Soucasaux Pires Garcia, Juan Manuel Pardal, and Sergio Souto Maior Tavares. The first draft of the manuscript was written by Pedro Soucasaux Pires Garcia and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES)-Finance Code 001, by CNPq and FAPERJ.

Availability of data and material Not applicable.
Code availability Not applicable.