Microstructural Evolution Due to One Thermal Cycle in a Superduplex Stainless Steel ASTM A890/A890M-Grade 6A in the As-Weld and Post-Weld Heat Treatment Conditions

During welding of the superduplex stainless steels are exposed to thermal cycles, allowing precipitation of deleterious phases as well as unbalance as the ferrite and austenite phases. This work analyzed the microstructural changes taking place on a plate ASTM A890/A890M grade 6A steel submitted to arc welding one welding of pass conducted on a flat surface using coated electrode Zeron®100, in the as welded and post-welded heat treatment conditions. Microstructural characterization in the regions base metal, heat affected zone, and fusion zone were conducted by optical microscopy, scanning electron microscopy and X-ray diffraction. The post-weld heat treatment in the fusion zone changed the ferrite/austenite volumetric fraction from 59/41 to 52/48. Cr, Mo and Ni distributions in the fusion zone were altered after the post-weld heat treatment, with the concentration ratio ferrite/ austenite being around 1.0; 1.1 and 0.9 to 1.1; 1.4 and 0.7, respectively. In addition, hardness reduced from 300 to 270 HV0.3 at the fusion zone.


Introduction
The superduplex stainless steels (SDSS) depict a microstructure composed of austenite (γ) and ferrite (δ), normally with a 1:1 ratio 1-3 . The formation of these two phases is determined by the preferential partitioning of the alloying elements, ferrite, and austenite phases, especially the Cr, Ni and Mo 4 . Cr and Mo enrich ferrite, while Ni and N are concentrated in austenite 5,6 . These materials show a compelling combination of excellent corrosion resistance and high mechanical properties 1,7 . For example, SDSS steels belonging to the Fe-Cr-Ni system present outstanding corrosion resistance, with pitting resistance equivalent number (PREN) above 40 1,3 , with Cr and Mo contents around 25-27% and 3-4.5 %, respectively 6 .
High temperature changes the morphology of the microstructure of SDSS steels, as well as the partitioning of alloying elements, which affects the mechanical and corrosion properties of the heat-affected zone (HAZ) 8 and fusion zone (FZ). The greater the presence of alloying elements in these steels, the greater the probability of forming deleterious phases, such as sigma, when the material is exposed to thermal treatments or welding thermal cycles 9 .
Some authors have reported that in the HAZ may precipitate Cr 2 N, χ, and σ at different typical temperatures 10,11 . In these regions, high or very low heat input may be unfavorable for phase adjustment, under high heat input conditions may result in precipitation of intermetallic phases, and when low heat input conditions are presented, may raise the ferrite content 11 . A heat input of 0.5 to 2.5 kJ/mm is recommended for 25% Cr 12 .
Partitioning of alloying element and partition coefficient are dependent on cooling rates, which in slower cooling rates can diffuse completely elements that stabilizes austenite (N, Ni) of the ferrite matrix to austenite precipitates 6,13 . The degree of partitioning of the alloying elements between the phases δ and γ calculated by the partition coefficient greatly affects the corrosion resistance of the SDSS 2,5,8 . Several authors have reported that post-welded heat treatment (PWHT) can improve corrosion resistance properties [14][15][16][17] , due to the homogenization of phases ratio. The PREN is affected by the partition of the alloying elements, the ferrite tends to increase with temperature and the Cr and Mo gets more diluted in this phase [17][18][19] .
The objective of this work is to evaluate the effect of one welding pass conducted on a flat surface of superduplex stainless steel and to verify the effect on the microstructure of the welded joints. The steel analyzed in this research was obtained in sand casting, it is regulated by ASTM A890/ A890M with grade 6A specification. The joints were evaluated in the as-welded and PWHT conditions. The analyzes were performed in the regions base metal (BM), heat-affected zone (HAZ) and fusion zone (FZ

Experimental Procedure
Plates with dimensions of 360 mm x 110 mm x 40 mm of a ASTM A890/A890M grade 6A 20 steel were studied. The steel was prepared in a 60 Hz Vacuum Induction Melting (VIM) furnace and poured in plate shaped sand molds bounded with phenolic-urethane resin. After solidification and demolding, the plates were then heated for 4 hours at a temperature of 1140 °C to adjust ferrite/austenite balance recommended by the ASTM A890/A890M standard 20 .
The welding parameters of the company that supplied the material. Bead-on-plate pass was conducted using shielded metal arc welding (SMAW) process, with voltage 20-25 V and current 100-130 A, heat input 1.3 kJ/mm and speed 1.23 mm/s. The coated electrode was the Zeron®100, whose chemical composition is similar to the steel used in this work, as shown in Table 1. After welding, the plate was sectioned in two parts, with one of them subjected to a PWHT at 1150 °C for 2 h 45 min and subsequent cooling in water. This procedure aimed to approximate the amount of ferrite and austenite to a 1:1 ratio.
For microstructural characterization, samples of the area of interest, such as BM, HAZ and FZ, transversal to the weld, were grinded using sandpaper with grit size of 180 to 1200, and polished with diamond paste 6, 3 and 1 µm. To revealing the microstructure of the samples, electrolyte etching was conducted with 3 volts for 15 seconds in a 20% aqueous solution of NaOH.
The images were obtained by light optical microscope (LOM) and by secondary electron detector coupled to a scanning electron microscope (SEM). The chemical analysis was performed by Energy Dispersion X-Ray Spectroscopy (EDS). The partition coefficients (K) was calculated by Equation 1, where C δ is the concentration of ferrite and C γ is the concentration of austenite 21 .

Equation 1
Phase quantification in the regions of the BM, HAZ and FZ under the as-welded and PWHT conditions was performed by image analysis according to ASTM E1245 standard 22 . The phase identification was performed by X-ray diffraction (XRD) technique and a cobalt (CoKα1) source with a wavelength length of 1,789 Ǻ, 40 kV, 40 mA and step of 0.02 °/min. The Vickers (HV) hardness measurements conducted with a load of 2.94 N (0.300 kgf) and a loading time of 15 seconds using the ASTM E384 standard 23 . Figure 1a show the cross-section and the transition between the BM and FZ weld bead produced on SDSS plate. From images, three distinct regions are depicted, the fusion zone (FZ), the heat-affected zone (HAZ) and the base metal (BM). A clear difference in the weld bead as-welded and PWHT can be observed in the Figure 1d-e. The Figure 2 shows the volumetric fraction of the ferrite and austenite phases in the BM and FZ, in the as-welded and PWHT conditions.

Results
In the microstructural analysis, it was observed that BM depicted austenite islands precipitated in ferrite matrix, as shown in Figure 1b, with a ferrite/austenite fraction of 47/53 shown in Figure 2. The post-weld heat treatment showed little change in the structure of BM as shown in Figure 1c, with a ferrite/austenite ratio of 49/51. The change was not significant, since the BM had already been heat-treated after the plate was made, to balance the volumetric fraction of ferrite and austenite. In the transition region between the BM and the FZ a transition region, called the HAZ, the precipitation of ferrite occurs within the pre-existing austenite islands as shown in Figure 1d. In multipass welds the austenite dissolves during heating to form the ferrite and during cooling the austenite precipitates, especially, in the contours of ferrite grains. In HAZ, however, austenite does not dissolve completely, so it is observed that the morphology at the interface between ferrite and austenite is partially altered 24 .
After welding, FZ presented a ferritic matrix with acicularized austenitic structure (Widsmanstätten type) 25 and precipitated in the contour and inside of the ferritic grain, see Figure 1f. This acicularized structure is typical in regions that melt during welding and which solidify rapidly. Rapid solidification inhibited the growth of austenite from the ferrite. In the solidification of this alloy the first phase to be formed from the liquid is the ferrite, as can be observed in 70Fe/25Cr/5Ni steel 26 , whose basic chemical composition is close to the steel used in this work.
The post-welded heat treatment in the FZ favored austenite growth and the ferrite/austenite volumetric fraction ratio approached the balance, from 59/41 to 52/48 1,2 as shown in     Figure 2. The growth of austenite with PWHT also reduced its acicular shape somewhat according Figure 1f. As shown in Table 2 and Figure 3, Cr, Mo and Ni distributions in BM were not significantly altered after PWHT, with the ferrite/austenite concentration ratio being around 1.1, 1.6 and 0.7 respectively. These data are consistent with literature, which indicate a ferrite/austenite ratio greater than 1 for ferrite stabilizing elements such as Cr and Mo; and less than 1 for stabilized austenite elements, such as Ni 2,5,13,21,27 . However, in the FZ, which was cooled rapidly, with PWHT there was diffusion of Cr and Mo from austenite to ferrite and from Ni in the opposite direction.
The PREN (%Cr+3.3%Mo+16%N) of the ferrite and austenite phases is shown in Table 2, as well as the overall value ((PREN δ + PREN γ )/2), the nitrogen value (% wt) of 0.29 for the BM and 0.26 for the FZ, values that are available in Table 1, since this chemical element is light to be detected by the EDS. It was observed that austenite presents lower values of PREN in BM, for both conditions. However, in FZ in the as-welded condition the values were equivalent, but in the condition with PWHT obtained smaller value. The PREN in FZ after PWHT was equivalent to BM in both conditions. In thermally treated materials PREN tends to increase in the ferrite phase and decrease in the austenite phase 28 . This condition tends to improve the resistance to pitting corrosion since the ferrite is more susceptible to corrosion and in this state, this phase presents the best conditions 13 .
however, as the fast solidification acicularized the austenite and raised the volumetric fraction of the ferrite, which is a phase of greater hardness than the austenite, the hardness of the region increased to 300 HV 0.3 . The PWHT in the FZ softened the acicularization of austenite and increased its volumetric fraction from 41 to 48%, at the expense of the ferrite consumption, facts that resulted in a reduction of the hardness of the steel to 270 HV 0.3 . The austenite peaks became more evident in the condition after PWHT, it is noticed that in the annealed BM the present peaks of both phases are balanced indicating that there is volumetric balance of the phases. It was also observed that in the FZ as-welded ferrite peaks prevailed, this result is coherent with the volumetric fractions of the present phases, because in this condition it is common that these steels have higher ferrite contents. The hardness in the BM remained at 250 HV 0.3 after the PWHT, see Figure 5, this is justified because the heat treatment did not significantly alter the volumetric fractions of the phases and nor the microstructure thereof. In the FZ, The hardness reduction occurs because the austenite has a lower hardness than the ferrite with values of 266 ± 9 HV 0.3 and 301 ± 13 HV 0.3 , respectively, results obtained in BM in annealed condition. The higher presence of Cr and Mo in the ferrite phase increases the hardness of the same. In addition, ferrite has less slip systems than austenite at room temperature 29 .

Discussion
The partition of the elements between austenite and ferrite is driven by the phenomenon of diffusion, i.e., the partition coefficient depends on the cooling rate. In slow cooling there is an excellent partition of the elements between the two phases, based on the thermodynamic characteristics of each element. When the cooling is fast, the diffusion is inhibited, producing austenite and ferrite of homogeneous composition, leading to partition coefficients with very high values close to 1.
At high temperatures the Ni tends to be concentrated also in the ferrite phase, the behavior observed in the FZ 1 . The partition coefficient of the Cr alloy element for all conditions was around 1, slightly stabilizing the ferrite phase 2 . The alloying element Mo preferably stabilizes the ferrite phase, but its concentration is decreased in this phase in the FZ of the as-welded condition, due to the cooling conditions, not having time for diffusion in the ferrite phase and being part retained austenite. High cooling rates induced by the welding process leads to significant variation in the grain size, orientation and shape in the FZ microstructures when compared to BM, as also reported in previous studies 12 . The PWHT enabled the partitioning of the alloying elements and changed/modified 1.0; 1.1 and 0.9 to 1.1; 1.4 and 0.7 for chemical elements: Cr, Mo and Ni, respectively.
Hardness in the FZ was the highest and even after PWHT was higher than BM, since the welding can cause the tension induced hardening during the solidification of the weld, fast thermal heating and cooling cycle, residual tension and higher concentrations of elements of alloys trapped in the crystalline lattice of the phases 30 , which ends increasing hardness in the matrix.
In these steels the ferrite presents a higher hardness than austenite because it contains higher Cr and Mo contents in its composition 12 and because its volumetric fraction was higher in the FZ before the PWHT and there was still a higher Ni absorption, the hardness tends to be higher in these regions. Moreover, BCC crystal (ferrite) structures have less slip systems than the crystalline FCC structure at room temperature 29 . Austenite has much more deformation modes, such as dislocation slips, shear banding and phase transformations, while the ferrite undergoes deformations by dislocation slips on many operating slip systems 30-34 .

Conclusions
Based on the materials and results of the experimental techniques used, it is possible to conclude that: • The PWHT in the fusion zone changed the ferrite/ austenite volumetric fraction from 59/41 to 52/48, The partition coefficient ratio ferrite/austenite of the chemical elements Cr, Mo and Ni being around 1.0; 1.1 and 0.9 to 1.1; 1.4 and 0.7, respectively after PWHT. This result confirms that Cr and Mo tend to stabilize the ferrite phase and Ni the austenite phase, since PWHT allowed the diffusion of these alloying elements. • PREN in FZ after PWHT was equivalent to BM in both conditions. In thermally treated materials PREN tends to increase in the ferrite phase and decrease in the austenite phase; • The PWHT balanced the phases proportions produced after welding. Likewise, alloying elements partition and microhardness equilibrium was favored by the PWHT. • Austenite and ferrite were the only phases found by the XRD analysis, however the intensity or the appearance of the peaks showed that there is a certain tendency in the formation of the phases, because in the conditions with PWHT the peaks are equivalent and on the other hand, in the as-welded FZ the peaks of ferrite are more evident. • The fusion zone presented a refined austenite grains, while BM depicts austenite plates. This characteristic was observed in the two conditions, although the adjustment of the phases after the PWHT occurred to a more proportion close to 1:1 ratio.