Temperature and time effect of thermal aging treatment on microstructure and corrosion resistance of UNS S31803 duplex stainless steel

Duplex stainless steels correspond to a class of steel in which the microstructure is composed basically by the phases ferrite and austenite. Given the metastable character of ferrite, associated phase transformations can occur during thermal treatments, which can lead to a reduction in corrosion resistance. Therefore, the aim of this work was to evaluate the influence of the solution-treated at 1100°C for 30min and aging at 500 and 600°C for 1, 3 and 12h, on the microstructure and corrosion resistance of DSS UNS S31803. The steels were characterized by SEM, EDS, XRD, hardness, microhardness and thermal analysis. The steels aged at 500°C for 1h presented the phase alpha line dispersed in the ferrite, and the steels aged at 600°C for 12h presented the phases alpha line, sigma and chi. In the assays of double cycle potentiodynamic reactivation in solution of 0.5M H2SO4 and 0.01M KSNC, it was observed that no sensitization occurred. Furthermore, in the tests of cyclic potentiodynamic polarization, in 3.5% NaCl solution, it was verified that the precipitation of the alpha line phase did not alter corrosion resistance; nevertheless, the sigma phase reduced corrosion resistance.


Introduction
Duplex stainless steels (DSS) present low carbon content because of the process of decarburization by oxygen blowing associated to the argon used in its production, which, associated to a biphasic microstructure, composed of ferrite (α) and austenite (γ), promote high corrosion resistance, good weldability, besides high mechanical resistance. Nevertheless, the typical microstructure can be modified by the thermal treatments of solubilization and aging (Örnek, Burke, Hashimoto, Lim & Engelberg, 2017, Sicupira, Cardoso Junior, Bracarense, Frankel & Lins, 2016.
Phase α' is a precipitate rich in chromium and it occurs inside α, it presents a body-centered cubic (BCC) structure, coherent with the ferritic matrix, and since it has low energy of interface with the ferritic matrix, it presents high resistance to coalescence (Chen & Limin, 2015, Cronemberger et al., 2015, Hilders, Zambrano & Ochoa, 2018, Jinlong, Tongxiang, Lacerda, Cândido, & Godefroid, 2015. The precipitation of phase α' is associated to the diffusion of chromium and, therefore, it promotes the formation of regions poor in chromium around it. Thus, DSS becomes more susceptible to the localized corrosion, for being sensitive to active dissolution and contributing to the degradation of passive film stability (Hilders et al., 2018, Jinlong et al., 2015, Lacerda et al., 2015, Li, Lo, Kwok, Sun & Lai, 2018, Mandal, 2016, Mohammed, Shrikrishna & Sathiya, 2016, Verma & Taiwade, 2017.
Phase α' can be formed by two mechanisms, nucleation and growth, and spinodal decomposition of ferrite, depending on the chemical composition of the ferrite phase and the temperature of the treatment. Both mechanisms occur to reduce the Gibbs free energy of the system and, consequently, promote a thermodynamically more stable state (Appleton, 2019, Llorca-Isern et al., 2016, Marques et al., 2020, Mohammed et al., 2016, Paulraj & Garg, 2015, Rezende et al. 2018, Rivolta et al., 2018, Verma & Taiwade, 2017.
The mechanism of spinodal decomposition of ferrite occurs when the chromium content is superior to 23% (in mass) and the thermal treatment of aging is performed at 280-475°C (Hilders et al., 2018, Jinlong et al., 2015, Lacerda et al., 2015, Li et al., 2018. The condition for spinodal decomposition to occur is that the alloy composition be the one in which the curve of free energy presents a negative curvature, in other words: This condition is observed for systems at low temperatures and where mixing enthalpy is positive, in other words, ΔHmix > 0. This last factor occurs because of a characteristic of atoms of having more favorable ligations among equal species than different species. Thus, when chemical composition and temperature are located inside the chemical spinodal, phase α is unstable, there is no nucleation barrier, and decomposition occurs in diffusion of the "up-hill" type, nucleating phase α', by spinodal decomposition.
Nonetheless, in case the composition is inside the miscibility GAP and outside the chemical spinodal, phase α is metastable; thus, there might occur the formation of phase α', by nucleation and growth, and the diffusion is of the "down-hill" type.
This reaction is started around the grains α/γ or α/α, and the consumption of the ferrite phase grows. Its precipitation occurs by the diffusion of the elements chromium, molybdenum and silicon, and with this, there is the formation of regions poor in these elements, leaving DSS with lower corrosion resistance, besides reducing ductility and toughness (Deng & Zheng, 2019, Dille et al., 2017, Matias, Tavares, Pardal & Ribeiro, 2017, Mehta et al., 2019, Nithin et al., 2019, Ogawa & Osuki, 2020, Xiang, Liu, Silva et al., 2016. Phase chi is thermodynamically unstable and is usually precipitated around the grains α/γ or α/α (Örnek et al., 2017).
Usually, phase chi is consumed by the phases sigma and austenite, and this tends to weaken and reduce the localized corrosion resistance of DSS (Alvarez-Armas & Degallaix-Moreuil, 2009, Martins & Hara, 2004, Morais & Magnabosco, 2017, Santos & Magnabosco, 2015, Warren et al., 2015. Therefore, the aim of this work was to evaluate the influence of the thermal treatments of solubilization and aging on phase precipitation and electrochemical corrosion behavior, employing the techniques of cyclic potentiodynamic polarization and by the double cycle potentiodynamic reactivation (DL-EPR).

Methodology
The research methodology applied in this study is experimental/laboratory, relying on quantitative and qualitative approaches, as it is aimed at testing hypotheses in numerical and observed forms about the phase transformation resulting from Research, Society andDevelopment, v. 10, n. 8, e24910817369, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i8.17369 4 the aging heat treatment (Ludke & Andre, 2013, Koche, 2011, Pereira, Shitsuka, Parreira & Shitsuka, 2018. Thus, the set of analyzes used to measure the influence of the phase transformation phenomenon occurring in duplex stainless steel about the modification of its corrosion resistance is described below. The nominal chemical composition of the steel studied was determined by inductively coupled plasma atomic emission spectroscopy and is presented in Table 1. Source: Authors (2021).
The samples were solution-treated at 1100ºC for 30 min and cooled in water. After that, aging treatments were performed at 500 and 600ºC for 1, 3 and 12h, followed by cooling in water.
The samples were analyzed by scanning electron microscopy (SEM) and the semi-quantitative microanalysis of the phases was obtained by energy-dispersive spectroscopy (EDS), after chemical attack with Behara (2.02M HCl and 0.045M K2S2O5).
X-ray diffraction (XRD) analysis was performed with Cu Kα radiation, in a range of 2θ from 40º to 100º, with a step of 0.02º for 2 s/step.
Thermal analysis was performed using differential scanning calorimetry (DSC) in the equipment Netzsch Júpiter STA 449F3, the samples were submitted to heating from 30ºC to 1200ºC, at a heating rate of 5ºC/min, under an argon and nitrogen atmosphere.
The corrosion behavior was evaluated by the cyclic potentiodynamic polarization method using a potentiostat Metrohm model Autolab/PGSTART 302 connected to a typical electrochemical cell with a saturated calomel electrode (SCE) used as a reference electrode, a platinum plate employed as counter-electrode and the working electrode made from the steel studied. Electrochemical measurements were performed in triplicate for each condition. The electrolyte used was 3.5 % sodium chloride solution at 25ºC. The potentiodynamic curves were measured at a potential scan rate of 1mV/s in the range from -1.5V, with reversion of the direction when the anodic current density of 10 -3 A/cm 2 was reached.
The corrosion tests by double cycle potentiodynamic reactivation (DL-EPR) were performed in the solutions 0.5M H2SO4 and 0.01M KSNC. The DL-EPR curves were obtained by scanning the potential from 0.5 to 0.3V, the scan was reversed and swept from positive to negative potentials at -0.5V at a constant rate of 2mV/s.
The corrosion tests were performed in the samples in which there was precipitation of the alpha line phase, to evaluate its effect in relation to the samples under the conditions as-received and solubilized.
The microhardness (HV) test was performed with a load of 0.1 kgf and printing time of 15 seconds, and hardness (HRC) was also determined with a conical penetrator of 120° and load of 150kgf. Ferritoscopy assay was also performed for the DSS samples under the conditions studied.

Results and Discussion
The micrograph of the DSS UNS S31803 as received are shown in Figure 1a, and in the regions indicated by A and B, the EDS analyses were performed. The austenite phase was observed in the form of elongated islands, and in alternate regions of refined and acicular structures, dispersed in the ferritic matrix. Figure 1b shows a micrograph of UNS S31803 solution- Research, Society andDevelopment, v. 10, n. 8, e24910817369, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i8.17369 5 treated at the temperature of 1100°C for 30 min and cooled with water. respectively. It is observed that the thermal treatment of aging performed at 500°C for 1h ( Figure 2a) was able to precipitate phase α', inside the ferrite, by the mechanism of nucleation and growth. The morphology of phase α' is a consequence of the differences between the thermal expansion of ferrite and of α', besides the proximity between the network parameters between the matrix and the precipitate which led to the formation of coherent interfaces and a spherical appearance with diameters of approximately 100nm. Furthermore, it is noted by Figures 2b and 2c that there was a growth of the austenite grains because of the diffusion of the alloy elements, analogously to the observed after the solution-treated, Figure 1b The micrographs obtained by scanning electron microscopy for UNS S31803 aged at 600°C for 1, 3 and 12h are shown in Figure 3. An alteration in austenite morphology was observed in the three times of aging treatment, changing from elongated islands to a massive structure, deriving from the growth of the austenite grains followed by their union, promoting a heterogeneous microstructure. The treatment of aging at 600°C for 12h promoted the precipitation of phase α', by the method of nucleation and growth, with morphology similar to the treatment of aging at 500 o C for 1h (Figure 2a).
It is observed that the alloy elements chromium and molybdenum are majoritarian in the ferrite phase, because of the ferritizing character of these elements, whereas the alloy element nickel is majoritarian in austenite, given the austenitizing character. It was also observed that there were no significant variations in the contents of chromium, molybdenum and nickel under the studied conditions, even with the precipitation of phase α' inside ferrite. Figure 4 presents the X-ray diffraction spectra of DSS UNS S31803 under the studied conditions. It was observed that the solution-treated the intermetallic phases σ and χ present under the condition as-received. Furthermore, the precipitation of phase σ was observed, in the thermal treatments of aging performed at 500°C for 1h and at 600°C at the times of 1, 3 and 12 h, and the precipitation of phase χ, aging at 600°C for 1 hour.   Table 4 shows the hardness values obtained for DSS under the studied conditions, and a decrease in hardness was observed after the solution-treat, deriving from the dissolution of the intermetallic phases σ and γ, present in the as-received steel. In the steel aged at 500°C for 1h, there was the precipitation of phases α' and σ, which present weakening and hardening characteristics, promoting higher hardness, in relation to the aging times of 3 and 12h. The treatment of aging at 500 o C for 1h ( Figure 2a) and 600°C for 12h (Figure 3c) promoted the precipitation of phase α', which presents hardening characteristics, which led to an increase in hardness in relation to the other times.

DSS UNS S31803 Hardness (HRC)
As-received 24 ± 1 Solution-treated at 1100ºC 15 ± 2 Aged at 500 o C for 1h 23 ± 1 Aged at 500 o C for 3h 21 ± 1 Aged at 500 o C for 12h 22 ± 1 Aged at 600 o C for 1h 22 ± 1 Aged at 600 o C for 3h 22 ± 1 Aged at 600 o C for 12h 23 ± 1 Source: Authors (2021).    Source: Authors (2021). Research, Society andDevelopment, v. 10, n. 8, e24910817369, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i8.17369 11 Assays of double cycle potentiodynamic reactivation (DL-EPR) and cyclic potentiodynamic polarization (CPP) were performed in a solution of 0.5M H2SO4 and 0.01M KSN, in the samples in which there was precipitation of phase α' (aged at 500°C for 1h and 600°C for 12h) and under the conditions as-received and solubilized, as shown in Figure 6. The values of corrosion potentials (Ecorr) and degree of sensitization are presented in Table 6. It was observed that under the conditions as-received and solubilized, the values of the potentials increased with time, deriving from the formation of a stable passive layer that adheres to the steel surface. Nevertheless, under the conditions of aging at 500°C for 1h and 600°C for 12h, the values of the potentials decreased over time, followed by a stabilization at the times of 250 seconds and 50 seconds, respectively, evidencing a generalized corrosion followed by the formation of a passive layer. Table 6: Open circuit corrosion potential and degree of sensitization by double cycle potentiodynamic reactivation, of DSS UNS S31803 in solution of 0.5M H2SO4 and 0.01M KSNC.
It was observed that the degree of sensitization was maintained below 0.5, for all conditions, indicating that no sensitization occurred (Chail & Kangas, 2016). This fact evidences that the passive layer was not disrupted and, therefore, the thermal treatments of aging and solubilization did not alter the resistance to intergranular corrosion of steel. Thus, the presence of phases α', σ and χ did not have significant effect on the degree of sensitization, because of the low precipitation of these deleterious phases, which promoted regions poor in chromium and molybdenum. Additionally, in Figure 6b, two peaks are observed in the activation current, the more noble potential corresponding to the austenite phase, and the less noble corresponding to the ferrite phase.  From Figure 7a, the values of the corrosion potentials (Ecorr) were obtained, as shown in Table 7. It was observed that under the studied conditions, the values of the potentials increased over time, characterizing the formulation of a passive, stable and adherent layer on the surface of the steel. Table 7: Open circuit corrosion potential of DSS UNS S31803 in a 3.5% sodium chloride solution.
From the polarization curves, Figure 7b, the electrochemical parameters were obtained, by the Tafel's extrapolation method, as shown in Table 8.
The values of corrosion potential obtained by open circuit were higher than those obtained by the cyclic potentiodynamic polarization curves, Tables 7 and 8, respectively, because the partial removal of the passive layer in the polarization scan started in more negative potentials than the corrosion potential.
The polarization curves show the presence of two passive regions in all conditions tested. This occurred because of the difference in chromium and molybdenum contents between the ferrite and austenite phases, elements that passivate steel.
Thus, the ferrite phase that presents higher contents of these elements passivates before the austenite phase.
The values of the electrochemical parameters: corrosion potential (Ecorr), pitting potential (Epit), protection potential (Eprot), corrosion current density (Icorr) and passivation current density (Ipass), under the conditions as-received, solutiontreated and aged at 500°C for 1h did not present significant variations, indicating that the precipitation of phase α', present in these conditions, did not affect steel corrosion properties.
Nevertheless, in the treatment of aging at 600°C for 12h, a reduction in the values of corrosion, pitting and protection potential was observed, because of the precipitation of phase sigma (σ) and promoting the formation of regions poor in chromium, and consequently, reducing corrosion resistance. Furthermore, in this condition, higher passivation current density was observed, indicating that the passive layer was formed more slowly than in the other conditions.
Figures 8a and 8b show the micrographs obtained by SEM of DSS UNS S31803, aged at 500ºC for 1h and aged at 600ºC for 12h, respectively, after the assays of corrosion by polarization, in a 3.5% sodium chloride solution. In the two conditions, the localized corrosion by pits was observed, being more intense under the condition aged at 600°C for 12h. The pits were classified as crystallographic (Pit-C), hemispherical (Pit-H) and irregular (Pit-I) and it is noted that pit nucleation occurred preferentially around grain α/α and α/γ and inside the austenite grains, characterizing the selective corrosion. (a) Aged at 500ºC for 1h; (b) Aged at 600ºC for 12h.
Source: Authors (2021). Table 9 shows the pitting resistance equivalent number (PREN) of the ferrite and austenite phases, calculated from the chemical composition of the main alloy elements obtained by EDS (Table 2). It can be observed that a preferential selective corrosion of one of the phases may occur due to the significant difference of the PREN values between the phases. This difference is caused by the non-homogeneous distribution of the passivating elements; consequently, the passive layer can present a different composition between the phases, increasing the probability of occurring defects.
In Figure 8, it is noted that the pits are nucleated around the grains and they grow in ferrite, even with this phase presenting a higher PREN value than austenite. This derives from the fact that the alloy elements Cr and Mo, besides passivators, are stabilizers in the ferrite phase, whereas in the austenite phase they act only as passivators. Therefore, for a corrosion resistance by pits similar to the austenite phase, the ferrite phase should contain higher contents of the elements Cr and Mo.

Conclusion
The steel in the condition as-received presented a ferritic matrix with elongated austenite islands and the presence of phases σ and χ. The thermal treatment of solution promoted higher homogenization of the phases and growth of the austenite grains, besides promoting the dissolution of the intermetallic phases σ and χ.
The thermal treatment of aging performed at 500°C for 1h promoted the precipitation of phase α' dispersed in the ferritic matrix, whereas the aging performed at 600°C for 12h, besides the precipitation of phase α', promoted the precipitation of phases σ and χ.
Additionally, the thermal treatments of aging reduced the volumetric fraction of the ferrite phase and/or promoted the precipitation of phases α' and σ, increasing steel hardness.
The ferrite phase of the steels thermally treated by aging at 500°C for 1h and 600°C for 12h presented higher values of microhardness because of α' precipitation. The austenite phase did not undergo significant variation in microhardness, indicating that phase G did not precipitate inside it.
By thermal analysis, the ferromagnetic-paramagnetic transformation of ferrite was observed at 428.9 and 520.4°C, besides ferrite consumption by paramagnetic phases at 842.2 and 860.6°C.
In the tests of double cycle potentiodynamic reactivation, it was observed that the steel did not suffer sensitization under the studied conditions. The open circuit potential assays revealed the formation of a passive and stable film, adherent to the steel surface. By the tests of cyclic potentiodynamic polarization, it was observed that the precipitation of phase α' did not affect corrosion resistance; nevertheless, phase σ reduced corrosion resistance.
In the steels aged at 500°C for 1h and 600°C for 12h, the pits were nucleated around the grains and in the ferrite phase. The ferrite phase presented higher PREN values in relation to austenite, for presenting higher contents of Cr and Mo, which are passivating and alphagenic elements.
In addition, it is important to mention that future studies can be developed in terms of investigation of stress corrosion crack on the conditions evaluated, this topic is little explored in duplex stainless steels. It is also suggested an investigation of reactions resulting from corrosion by electrochemical impedance tests.