Deleterious Phases Precipitation on Superduplex Stainless Steel UNS S 32750 : Characterization by Light Optical and Scanning Electron Microscopy

Deleterious phases precipitation in superduplex stainless steels is the main concern in fabrication by welding and hot forming of this class of material. Sigma, chi and secondary austenite phases are considered deleterious phases because they produce negative effects on corrosion resistance. Besides, sigma and chi phases also promote strong decrease of toughness. In the present work, the precipitations of sigma, chi and secondary austenite under aging in the 800-950 °C interval were studied in two UNS S32750 steels with different grain sizes. The deleterious phases could be quantified by light optical microscopy, with no distinction between them. Scanning electron microscopy was used to distinguish the individual phases in various aging conditions. The results elucidate the influence of the aging temperature and grain size on the kinetics precipitation and morphology of deleterious phases. The kinetics of deleterious phases is higher in the fine grained material in the initial stage of aging, but the maximum amount of deleterious phases is higher in the coarse grained steel.


Introduction
Superduplex stainless steels (SDSS) are high strength and corrosion resistant materials used in the chemical and petrochemical industries.UNS S32750 steel has been applied in off-shore equipments used to store and process aggressive fluids 1 .
It is well known that the optimum mechanical and anti-corrosion properties of SDSS are obtained with equal proportions of ferrite (δ) and austenite (γ) phases 2 .The high pitting corrosion resistance is provided by high amounts of Cr, Mo and N contents, which must be dissolved in the ferrite and austenite phases and not precipitated as deleterious compounds.
Sigma and chi phases precipitate during re-heating in the 800-950 °C interval.However, χ is found to be formed in the first stages of aging.According to Michalska et al. 8 and Chen et al. 9 , χ phase is metastable and undergoes the transformation χ→σ under prolonged aging.
Sigma is the deleterious phase which may precipitate in higher amounts, generally assuming a lamellar morphology as a result of the ferritic eutectoid decomposition according to the reaction δ → σ + γ 2 .However, at higher temperatures σ may assume other morphologies 9 .
Secondary austenite (γ 2 ) phase may also be independently formed by diffusion at high temperatures.One of the characteristics of this phase is the low nitrogen content, which leads to a poor corrosion resistance when compared to the austenite (γ) previously formed in the solution heat treatment 10 .Chen et al. 9 observed plates of σ phase associated with γ 2 in the ferrite boundaries.However, γ 2 may also grow from the primary γ toward the ferritic domains, or by heterogeneous nucleation in the intergranular chromium nitride (Cr 2 N) particles 11 .
In duplex and superduplex stainless steels intense Cr 2 N precipitation occurs when the material is ferritized at high temperatures and rapidly cooled.This is typically found in heat affected zones of welded joints 12 .
Thus, the present work aims at studying the kinetics precipitation of the deleterious phases σ, χ, and γ 2 which occurs at almost the same temperature range.Scanning electron microscopy was used to distinguish the three phases and to better understand the evolution of each phase.The study was conducted in two SDSS of the same grade but different grain sizes.

Experimental
Two superduplex steels, named as "SD-A" and "SD-B" in this work, with chemical compositions shown in Table 1, were studied.Both materials were previously solution treated by the purchaser.However, steel SD-A presents much finer ferritic and austenitic grain sizes than steel SD-B, which will be shown later.Samples of 15 × 10 × 5 mm were cut and heat treated at 800, 850, 900 and 950 °C by 5, 15, 30, 45, 60, 90, 120 and 240 minutes, followed by water cooling.Microstructures analysis were carried out in a light optical microscope (LOM) NEOPHOT 32 and in a scanning electron microscopy (SEM) JEOL 6460LV with microanalysis by energy dispersive spectroscopy (EDS).Table 2 shows the list of metallographic etches used and the objective of each one.Some samples were observed without etch in the SEM, using the backscattered mode.Electron beam energy used in the EDS analysis was 20 kV.
Deleterious phases revealed by electrochemical etching with KOH solution were quantified by image analysis using Image Tools software 13 , analyzing 20 fields per condition with 240×, 475× and 950×.

Light Optical Microscope (LOM) analysis
Figures 1a and b show the microstructures in the as received condition of materials SD-A and SD-B, respectively.These images were obtained with Beraha's etching which provides a very good contrast between ferrite and austenite phases, but does not reveal grain boundaries.However, it is clear that material SD-A presents a much finer microstructure than material SD-B.The best measurements of the grain sizes of ferrite and austenite phases were achieved by electron backscattered scanning diffraction (EBSD) analysis, whose details are explained in another work 16 .Table 3 summarizes the grain sizes and phase amounts of each material in the as received condition.
Figures 2a-f show the microstructural evolution of images obtained by LOM with KOH etch in samples of both materials aged at 800 °C.The dark phases were quantified and computed as "deleterious phases".These dark regions include σ and χ phases and the γ 2 formed by eutectoid decomposition.Figures 3  and 4 show the evolution of the deleterious phases with time for the four investigated temperatures.The kinetics precipitation in the fine grained material SD-A is clearly faster than in material SD-B at the initial stages, which means that the grain refinement promotes a shorter incubation time and higher nucleation rate of deleterious phases 16 .
The precipitation curves of Figures 3 and 4 show a saturation value for the amount of deleterious phases.This saturation value decreases with aging temperature as shown in Figure 5.It is interesting to note that, despite the faster initial kinetics of precipitation, the fine grain material "SD-A" present lower saturation values than SD-B in all temperatures.Figures 6a and b show the microstructures of specimens treated at 850 °C by 240 minutes.At this heat treatment condition, material SD-A show many γ 2 large particles precipitated inside the ferrite domains, while SD-B shows a higher proportion of fine deleterious phases originated from the δ eutectoid decomposition, as delimitated by a circle in Figure 6b.The higher amount and size of independent γ 2 phase in material SD-A can be attributed to its higher grain boundaries areas available to the δ→γ 2 transformation.Although the γ 2 particles precipitated at the ferrite grain boundaries can create Cr and Mo-rich regions and accelerate the χ and σ phases formation in the beginning, in a further step they act as barriers, just like an impingement effect.This explains the lower saturation values achieved in material SD-A.Figures 7a and b and Figures 8a and b show this barrier effect of γ 2 in specimens aged at 900 and 950 °C for 240 minutes, respectively.

Scanning Electron Microscopy (SEM) analysis
Figure 9 shows the deleterious phases observed in detail by backscattered electron image in the specimen treated at 900 °C by 240 minutes.This image was obtained in a specimen polished but not etched.Points 1, 2 and 3 were analyzed by EDS and identified as χ, σ       and γ 2 phases, respectively.Table 4 shows the chemical compositions of each phase.Although these results are rather qualitative than quantitative, due to the small size of the particles analyzed, they are important to identify the different phases.Phase χ has the highest Mo content and, as a consequence, corresponds to the lightest regions in the image.Figure 10a and b show the SEM images of samples in the early stages of aging (850 °C/15 minutes), in which χ phase in the γ/δ and δ/δ boundaries were mainly observed.In Figure 9, a very small amount of phase χ is present, which is in accordance with its metastable feature, as observed by Michalska et al. 8 and Chen et al. 9 .Figure 9 also shows σ and γ 2 phases.At this condition only small remaining ferrite is present, since this phase was almost completely transformed.EDS results show that γ 2 particles very small Mo and Cr contents.According to Lippold Kotecki 10 , these γ 2 particles formed at low temperatures are also nitrogen depleted.These chemical composition features makes γ 2 particles very susceptible to pitting corrosion and to wear by grinding and polishing.The small black points mainly concentrated in the γ 2 observed in Figure 9, may be attributed to pitting and/or pulling by wear during the sample preparation.
Figure 11a and b show the images of the specimens of materials SD-A and SD-B aged at 900 °C for 60 minutes, for comparison.Table 5 shows the EDS analysis from points 1, 2, 3, 4 and 5 of Figure 11a.The following features must be highlighted based on these figures: • Typical eutectoid γ 2 + σ structure are observed in Figure 11b from SD-B.• A third form of γ 2 is observed, that results from the growth of pre-existing γ phase.This is more evident in Figure 11b, where small pits and/or chromium carbides are also observed in the γ 2 layer adjacent to the γ island.The nature of these small black points could not be clarified by SEM.The hypothesis of carbide    5) contains more Mo and Cr than the γ 2 formed by eutectoid decomposition observed in Figure 9 (Table 4).• σ and γ 2 phases observed in SD-B are much finer than in SD-A.The fact that by reducing grain size the σ particles become coarser may be explained by the higher solute diffusivity in the fine grained material.Figures 12a and b from material SD-B aged for 240 minutes at 800 and 900 °C show that the increase of aging temperature also promotes the coarsening of σ particles.

Conclusions
The precipitation of deleterious phases σ, χ and γ 2 during isothermal aging at 800, 850, 900 and 950 °C were investigated by light optical microscopy (LOM) and scanning electron microscopy (SEM).Based on the results, the following conclusions may be highlighted: 1. Three mechanisms of γ 2 precipitation were observed: eutectoid decomposition (δ → γ + σ), independent precipitation and growth of γ 2 plates, and growing of pre-existing γ phase islands.2. The initial stage of aging at 850 °C is characterized by the precipitation of Mo-rich phase at the γ/δ and δ/ δ boundaries.The increase of aging time promotes the replacement of the metastable χ phase by σ. 3. The deleterious phases σ, χ and γ 2 formed by eutectoid reaction were quantified by LOM in samples prepared by electrolytic etching in KOH solution (3V, 12 seconds).Curves of deleterious phases against aging time were presented.These curves present a saturation value of the deleterious phase amounts for long aging times.The saturation value is found to be lower in the fine grained steel, which is attributed to the higher grain boundaries areas available to the δ→γ 2 transformation.4. Since deleterious phase precipitation starts in the grain boundaries, the fine grained material presents a faster kinetic precipitation in the initial stages.However, the independent γ 2 particles precipitation is more pronounced in the fine grained material.These γ 2 particles act as barriers to σ precipitation in advanced stages of aging.As consequence, the saturation value is found to be higher in the coarse grained steel.The increase of aging temperature also intensifies the independent γ 2 precipitation and then decreases the saturation value.5. γ 2 precipitated by eutectoid reaction has very low pitting resistance because of its low Cr and Mo contents.6.The σ and γ 2 particles are coarser in the fine grained than in the coarse grained steel.The increase of aging temperature also promotes the coarsening of these particles due to the increase of solute atoms diffusivities.

Figure 2 .
Figure 2. Evolution microestructural images obtained by LOM in aged conditions at 800 °C for a SD-A and SD-B, respectively.

Figure 5 .
Figure 5. Saturation value of deleterious phases as function of aging temperature in SD-A and SD-B.

Figure 6 .
Figure 6.LOM images from samples treated at 850 °C for 240 minutes: a) SD-A; and b) SD-B.

Figure 7 .
Figure 7. LOM images from samples treated at 900 °C for 240 minutes: a) SD-A; and b) SD-B.

Figure 8 .
Figure 8. LOM images from samples treated at 950 °C for 240 minutes: a) SD-A and b) SD-B.

Figure 9 .
Figure 9. SEM image from sample treated at 900 °C for 240 minutes in a SD-B.

Figure 10 .
Figure 10.SEM images from samples treated at 850 °C for 15 minutes: a) SD-A; and b) SD-B.

Figure 11 .
Figure 11.SEM images from samples treated at 900 °C for 60 minutes: a) SD-A; and b) SD-B.

Figure 12 .
Figure 12.SEM images from SD-B samples treated at 800 °C for 240 minutes and 900 °C for 240 minutes, respectively.

Table 1 .
Chemical compositions of the materials studied.

Table 2 .
Reagents employed to reveal microstructural characteristics in a SDSS.

Table 3 .
Percentages of phases and grain sizes in materials SDA and SDB as received.
Figure 3. Percentage of deleterious phases precipitates against aging time at 800, 850, 900 and 950 °C for SD-A.

Table 4 .
EDS analysis of deleterious phases obtained in Figure9.

Table 5 .
EDS analysis of deleterious phases obtained in Figure11a.