The precipitation kinetics of secondary phases in two austeno-ferritic lean duplex stainless steels (lean DSS) were examined after aging the materials at 800 °C. Owing to the instability of ferrite, all DSS are known to be sensitive to solid-state phase transformations in the critical temperature range 600-1,000 °C and different secondary phases may form, depending on composition and microstructure. The performed thermodynamic simulations revealed the proneness to the precipitation of such phases also have been done in lean DSS, but only information on the equilibrium microstructures were achieved. Therefore, the materials were aged at various times, in order to verify the simulations and determine the precipitation kinetics. The occurred structural modifications were observed and quantified by scanning electron microscope and X-ray diffraction measurements, determining phase type, composition and volumetric fraction. At 800 °C, grade 2101 was found to be only affected by Cr2N nitrides precipitation, whereas a significant amount of σ-phase was found to form in LDX 2404 for treatment longer than 1 h, almost totally replacing ferrite after 50 h. Up to now, the intermetallic σ-phase has been observed only in the high alloyed DSS, and the unexpected precipitation in grade 2404 highlighted that the increased content of molybdenum in this steel might be considered as determinant for the formation.
Duplex stainless steels (DSS) are biphasic austeno-ferritic stainless steels, in which equal volume fractions of the phases provide the best combination of mechanical and corrosion-resistance properties [1, 2]. After the forming operations, a solution treatment followed by water-quenching (WQ) is performed in all DSS. This treatment, the temperature of which is strongly composition dependent, is mandatory in order to achieve the characteristic duplex microstructure and to permit the re-dissolution of any precipitates that may form during the hot-working operations. DSS are indeed susceptible to secondary phases precipitated in a wide temperature range (600-1, 000 ° C), and their excellent characteristics may be negatively affected. Secondary phases (the intermetallics σ - and χ -phase, nitrides) may form during very short isothermal treatments [1, 2, 3, 4] or as a consequence of very slow cooling rates from the solubilization temperature [5, 6]. These phases are Cr-rich structural discontinuities that promote localized corrosion attacks and reduce the impact toughness resistance [1, 2]. Their presence must be avoided.
The formation of secondary phases in DSS as a result of continuous cooling and isothermal treatments has been extensively studied [1, 2, 3, 4, 5, 6, 7]. Thermodynamic simulations on multi-component systems have shown that σ -phase (the main intermetallic compound) is not a structural anomaly resulting from manufacturing errors, but belongs to the equilibrium microstructure of all DSS [6, 7]. However, previous studies have evidenced that the precipitation of intermetallics mainly concerns the high alloyed grades [1, 2, 3, 4, 5, 6, 7, 8, 9, 10], due to their high content of alloying elements (mainly Cr and Mo). On the other hand, lean DSS have been found to be affected only by nitrides formation . In any case, all DSS are sensitive to nitrides precipitation, as they contain significant percentages of nitrogen [7, 11]. Nitrides formation is strictly related to the variation of temperature of the ferrite solubility against nitrogen, which increases at high temperature, thus bringing ferrite in supersaturated solid-solution conditions. This metastable condition may cause nitrides precipitation during either the WQ operations or the isothermal aging of the steel.
In the present work, the microstructural modifications resulting from isothermal heat treatments in two lean DSS grades in the form of sheets were analyzed. Different treatments were performed within the critical temperature range, focusing on the effects of aging at 800 ° C, in order to verify the microstructural stability of the materials close to equilibrium conditions for relatively short treatment time.
The investigated LDX 2101 and LDX 2404 lean DSS, whose chemical compositions are listed in Table 1, were in the form of sheets (3 and 2 mm thick, respectively) obtained by hot and cold rolling, with a maximum thickness reduction of 50%. The cold rolling process was followed by the solution treatment [annealing and water-quenching (WQ)], to recover and recrystallize the deformed microstructure.
The steels were subjected to the same isothermal heat treatments in a muffle furnace at the temperature ranging from 600 to 1, 000 ° C for 5 to 6, 000 min (100 h). The aging temperatures were chosen by analyzing the equilibrium phase diagrams obtained by thermodynamic modeling, which provided information on equilibrium microstructure and temperature ranges favoring phases formation. Thermo-Calc software  based on the CALPHAD method  was employed for the simulation by using the TCFe6 database.
The metallographic observations of the as-received materials were performed by using a Leica DMRE optical microscope (OM) on Beraha’ s etched samples, whereas the microstructural analyses of the thermally treated materials were performed on unetched specimens, using a Leica Cambridge Stereoscan 440 scanning electron microscope (SEM) equipped with a Falcon FEI energy-dispersive spectrometer (EDS). SEM observation allows distinguishing the different microstructural constituents according to their average atomic number, exploiting the signal from the backscattered electrons (BSE). Ferrite appears darker than austenite; nitrides can be observed as small black particles, the shape and location of which depend on the precipitation process. The intermetallic compounds appear brighter than the other microstructural constituents. OM and SEM micrographs were edited on an image analysis software, to estimate the volume fraction of the phases.
Samples were characterized by X-ray diffraction (XRD) analysis on a Siemens D500 diffractometer (CrK α radiation, step of 0.05° and 3 s of acquisition time) in the angular range 2θ = 55° -145° .
The microstructures of the as-received steels are shown in Fig. 1. Grain fragmentation due to the preceding rolling process was evident. Comparing nominally equal DSS grades with the same chemical composition, the role of the microstructure cannot be neglected when diffusion mechanism is involved. In this regard, forming processes induced microstructural modifications will alter the precipitation kinetics, especially in the first stages of precipitation [14, 15]. Image analysis confirmed that the as-received microstructures of both steels were well balanced with similar phase fractions (about 52 vol% of ferrite and 48 vol% of austenite).
The calculated phase diagrams in Fig. 2 show the equilibrium constituents of the steels as a function of temperature. The diagrams highlighted the presence of σ -phase at equilibrium in both steels, with a stability field slightly enlarged toward higher temperatures in 2404, probably due to the higher content of Cr and Mo, which promote its formation. On the contrary, the χ -phase was not foreseen, as instead occurs in the high alloyed DSS , probably due to the lower Mo content of the lean grades. Besides the σ -phase, the equilibrium diagrams also predict a significant quantity of Cr2N, in both cases close to 2-4 vol% of the volume fraction, and the presence of a very small quantity of M23C6 (< 1 vol%).
The calculations confirmed that between 500 and 600 ° C, σ -phase is completely formed, reaching approximately 30 vol% at the equilibrium. However, after 100-h treatment at 600 ° C, only small dark precipitates were detected in both steels (Fig. 3). These particles were preferentially located at ferritic grain boundaries and can be probably identified as Cr2N. Despite the thermodynamic calculations, no σ -phase was detected at this temperature. As a matter of fact, if one considers the kinetics governing the diffusion processes, this temperature is rather low and the time required to reach the equilibrium conditions may be very long. On the contrary, at 800 ° C, diffusion kinetics are more favoured and conditions close to the thermodynamic equilibrium can be reached in shorter time. Nevertheless, for grade 2404, this temperature falls in the σ -phase stability field, while for 2101, it is outside that range.
No phase precipitation was observed after 5-min treatment, whereas after 10-min treatment, small dark particles, identified as Cr2N, were detected. These phases were preferentially arranged at ferrite/ferrite grain boundaries and rarely observed at the ferrite/austenite interfaces (Fig. 4a). After 30-min treatment, only a moderate increment in nitrides volume fractions was noted.
Compared to other lean DSS grades [7, 9], a completely different microstructural evolution was observed after longer heat treatment times, and a new phase was formed after 1-h treatment (Fig. 4b). The EDS analysis confirmed the presence of σ -phase (Table 2), significantly enriched in Cr and Mo respect to ferrite and austenite. After 2-h treatment, the fraction of σ increased, but not significantly enough to be detected by XRD measurements.
A progressive growth of σ -phase with the increasing in soaking time was revealed. After 50-h treatment, ferrite was almost totally decomposed (Fig. 4c-f). It is demonstrated that such transformation occurred very slowly if compared to the high alloyed DSS. The EDS analysis and the XRD patterns (Fig. 5) confirmed that such phase was an intermetallic compound. In addition, the very long-time treatment caused the coalescence of the Cr2N particles, which no longer appeared as continuous chains located at grain boundaries. It is observed that less interconnected distribution of nitrides arose from the formation of σ .
The very slow ferrite decomposition in LDX 2404 can be mainly attributed to compositional factors. Compared to the high alloyed DSS grades [3, 7, 16], the lower content of alloying elements might have caused the observed delay in the precipitation kinetics. Grade LDX 2404 has a higher Mo content than other lean DSS grades, and this can be considered as the discriminating factor for σ -phase formation, as it is believed that Mo controls the diffusion processes and the precipitation kinetics involved in the genesis of the intermetallic phase [8, 16]. Mo, in fact, possesses the highest value of the diffusivity coefficient at any temperature within the critical range , and its contribution is essential in increasing the driving force for the formation of σ . As expected, no χ -phase was detected. In this case, the Mo content in the steel was not sufficient to allow the formation of the other intermetallic compound.
For this grade, 10-min treatment solely caused the precipitation of Cr2N. Particle size of Cr2N was slightly smaller than that observed in 2404, probably due to the lower Cr content, but which were present in a comparable fraction. The number of precipitates increased with the increasing in treatment time which resulted the coarsening of particles, but still below EDS and XRD detection limits. After 2-h treatment, nitrides are still the only precipitates, nearly maintaining similar dimensions but with a greater fraction (Fig. 6a).
As expected, after 50-h treatment, it is exhibited that there was no precipitation of any intermetallic compound, the microstructure was found to be composed of ferrite, austenite and nitrides (Fig. 6b). The long-time aging caused an increasing in Cr2N fraction, making it possible to be identified by means of XRD (Fig. 6c). Also, the 50-h treatment caused nitrides coalescence, and the precipitation at the grain boundaries was more fragmented than that from shorter time treatment.
As previously reported, only LDX 2404 was affected by σ -phase precipitation, and its fraction as a function of the treatment time is presented in Fig. 7a. Owing to the lower content of alloying elements, ferrite decomposition was very slow, and if compared to what is usually observed in high alloyed DSS , the formation of σ -phase was retarded. After 50-h treatment, the equilibrium conditions were not totally reached, since ferrite was not completely transformed.
At all temperatures and soaking times, Cr2N precipitation took place in both steels. The corresponding precipitation kinetics at 800 ° C is presented in Fig. 7b. In the early stages of precipitation (up to 10 min), the amount of nitrides can be considered comparable, whereas for longer time, the higher Cr content in 2404 affected the precipitation by favoring the formation of a greater fraction of nitrides. However, after 50 h, the quantity of precipitates settled on close values. As it is reported, the 800 ° C treatment primarily caused nitrides formation in 2404, and the precipitation of σ -phase occurred when nitrides were already present. In this steel, the observed saturation in nitrides fraction can be ascribed to the approaching of the equilibrium conditions, and the appearance of σ seemed to affect the nitrides displacement only at the boundaries, rather than the precipitated amount, favoring the coalescence of the particles.
Isothermal treatments were also performed in the whole critical temperature range (600-1, 000 ° C) for 10 min to determine the temperature at which nitrides formation may be favoured, on the basis of the observed precipitated amount. The results confirmed that the precipitation peaks were in the range of 800-850 ° C for LDX 2404 and 750-850 ° C for LDX 2101 (Fig. 8).
In a previous study , the kinetics of nitrides precipitation in a rod-shaped SAF 2101 DSS has been investigated. This steel had nearly the same chemical composition as that of LDX 2101, except for lower Mo (0.07 wt%) and N (0.16 wt%) contents. The manufacturing cycle was different, even though the final solubilization may be probably performed at about the same temperature. Respect to the SAF grade, an anticipated precipitation of nitrides was observed in the LDX under study, although the estimated critical temperature range (750-850 ° C) was found to be the same. This difference can be explained by considering the influence of compositional effects and microstructural differences. Besides Mo content, the higher N amount in LDX 2101 makes the steel more susceptible to nitrides formation, owing to an increased saturation level of ferrite. Moreover, in the LDX grade, the grains were rather fragmented, owing to cold rolling, and the solubilization had not completely restored the duplex matrix in terms of grains size, lattice strains and dislocation density. Therefore, the higher propensity to nitrides precipitation can be also explained by the presence of an increased number of triple points, grain boundaries, dislocations and residual stresses, which enhance the precipitation kinetics by shifting the formation of nitrides toward lower treatment time. This is supported by recent studies [14, 15] that reported an accelerated kinetics of phase precipitation by increasing cold working in SAF 2205 and 2507 DSS. However, in-depth studies are still required in order to assess the singles contributions to nitrides formation.
In the present work, the isothermal precipitation kinetics of secondary phases in two lean duplex stainless steels (LDX 2101 and LDX 2404 DSS) in form of sheets were examined. At any temperature within the range 600-1, 000 ° C and for treatment time longer than 5 min, intergranular chromium nitrides (Cr2N type) were detected, as usually occur in this class of steels. However, depending on the considered grade, the precipitation occurred with different kinetics, depending upon composition and microstructure.
A more detailed investigation on the effect of treating the materials at 800 ° C was performed. At this temperature, nitrides precipitation was found to be faster in 2404 than 2101, even if a comparable volume fraction of precipitates was revealed after 50-h treatment. The higher Cr content in 2404, in combination with the treatment temperature, allowed for an enhancement of nitrides precipitation in the (relatively) early treatment stages, making the microstructure more prone to their formation.
Conversely to what generally happens in lean DSS grades, the 800 ° C aging treatment of grade LDX 2404 caused the formation of σ -phase after 1 h. The harmful effects of this intermetallic phase are well known, and its precipitation generally affects the high alloyed DSS grades, such as SAF 2205 and 2507 DSS. In current study, the formation of σ -phase was anyhow slower than in higher alloyed DSS, which can be ascribed to the increased Mo content in LDX 2404 respect to other lean DSS. Molybdenum, together with chromium, is one of the main constituents of the intermetallic compound, and its content in grade 2404 is 5-10 times higher than what is generally found in lean DSS, thus making the material more sensitive to σ precipitation. Therefore, the Mo content may be considered as the crucial factor for the σ -phase appearance, as it controls the diffusion kinetics and contributes to the increase in the driving force of the genesis of intermetallic compound.
The authors have declared that no competing interests exist.