Sensitization of an austenitic stainless steel due to the occurrence of δ-ferrite

1 Introduction

Austenitic stainless steels of American Iron and Steel Institute (AISI) type 304 are designed and processed to be free of δ-ferrite at room temperature, as this structure is unstable and transforms into austenite at temperatures <1200°C. However, δ-ferrite remains as a microstructural constituent in castings, welded pieces, and wrought heavy sections of austenitic stainless steels due to rapid cooling or insufficient hot working (Tseng et al., 1994). The occurrence of δ-ferrite in an AISI type 304 stainless steel has been considered to be responsible for unusually high values of the degree of sensitization (DOS) encountered in samples subjected to a solution annealing heat treatment at 1050°C for 1 h followed by water cooling; this phenomenon was detected by means of the electrochemical potentiokinetic reactivation single-loop (EPR-SL) and double-loop (EPR-DL) tests. Metallographic examination of the steel showed the presence of ribbons of δ-ferrite (Arganis-Juarez et al. 2007).

The decomposition kinetics of δ-ferrite in austenitic steels has been documented by means of the reduction in magnetic response in samples subjected to annealing treatments in the temperature range of 600–900°C for periods of time of 0.25–20 h, achieving a maximum rate of decomposition at 720°C (Tseng et al., 1994). The occurrence of δ-ferrite and austenite structures promotes the development of a profile of dissolution at the austenite/δ-ferrite interface, as has been reported by means of atomic force microscopy (AFM) in cast corrosion-resistant steel CF8 (equivalent to an AISI type 304 steel) (Yi & Shoji, 1996), and in type 2205 duplex stainless steel by electrochemical scanning tunneling microscopy (Femenia et al., 2001). The dissolution at the ferrite region can be attributed to the galvanic effect between ferrite and austenite, to the depletion of alloying elements (Cr, Mo, and Ni) in ferrite, or to a combination of these two causes (Yi & Shoji, 1996; Femenia et al., 2001).

The use of a technique such as EPR-DL, which was designed to detect depleted zones of Cr in stainless steels, and the fact that such detection occurred in an annealed austenitic 304 stainless steel treated at 1050°C, indicates that the chromium-depleted zones at the δ-ferrite/austenite interface are originated from the difference in chromium contents that arise from solidification. Decomposition of δ-ferrite at 720°C takes place by the eutectoid reaction δ→γ+M₂₃C₆ (Tseng et al., 1994); this reaction is confined to thin layers at the δ-ferrite/matrix austenite interface, while the remaining δ-ferrite will be prone to decompose into the σ phase when exposed to temperatures in the range of 600–900°C (Tseng et al., 1994). Formation of the σ phase adversely affects ductility, toughness, and corrosion resistance in austenitic stainless steels (Gray et al., 1978).

The aim of this work is to present the results of studies conducted to evaluate the DOS found in a solution annealed 304 steel. Scanning electron microscopy (SEM) and AFM techniques were used to evaluate the degree of dissolution at the δ-ferrite/austenite interface.

2 Materials and methods

An AISI type 304 austenitic stainless-steel plate with a thickness of 12.7 mm was used in this work; the chemical composition of the material is shown in Table 1. Small samples of 10×10×3 mm were machined from the plate and subjected to solution annealing heat treatments at 1050°C for 0.5, 1, 1.5, and 2 h followed by immersion in cold water. Solution-treated samples were subjected to an ageing treatment at 650°C for a time period of up to 12 h, to promote sensitization by precipitation of chromium carbides.

The magnetic response of δ-ferrite was measured with a magnetic gauge balance according to American Society for Testing and Materials (ASTM) standard A799/A799M-04 (ASTM A799, 2015). The DOS was obtained using the EPR-DL tests based on the methodology proposed by Majidi and Streicher (1984), in which a three-electrode electrochemical cell consisting of a calomel saturated electrode as reference and graphite and the steel samples as auxiliary and working electrodes, respectively, in a solution of 0.5 m H₂SO₄+0.01 m potasium sulphocyanide, was used. Data from the cell were captured in an ACM Instruments, model Gill AC potentiostat (Cark, Grange-over-Sands, UK). DOS was assessed using the ratio of the reactivation current I_r over the activation current I_a, i.e. I_r/I_a. The values for the DOS were reported as fractions and averaged from two independent measurements. The ferrite number is associated with the amount of ferrite, either α or δ, in the steel and was evaluated using a magnetic balance (McGowan et al., 2000).

The treated samples were prepared for metallographic observation following conventional grinding and polishing procedures down to SiC paper of 600 grit and finishing with 0.5 μm alumina. Preparation for AFM measurements required the use of SiC paper of 4000 grit and then polishing with 0.5 μm alumina. Observations were made in an Asylum Research Microscope, model Cypher (Goleta, CA, USA); the tip was an AC240 type and the mode of operation was AC, with a spring constant of 2.2 N/m, amplitude of 70 nm, and frequency of 75 kHz. The microstructures were revealed by electrolytic etching with a 10% solution of oxalic acid following ASTM A-262 Practice A (ASTM A262, 2015); etching with a fresh solution of aqua regia (15 ml HCl, 5 ml HNO₃) was also carried out in selected samples.

Observation by SEM was conducted in a JEOL JSM model 5900LV (Peabody, MA, USA) coupled with an Oxford probe energy dispersive spectroscopy (EDS) detector. X-ray diffraction (XRD; Bruker Corporation, Billerica, MA, USA) was also carried out in a Siemens diffractometer, model D5000, with Cu Kα radiation (λ=1.540 Å). All etchants and reagents were provided by Sigma Aldrich, St. Louis, MO, USA.

3 Results and discussion

Samples of a 304 stainless steel showed the presence of δ-ferrite surrounded by austenite in its as-received condition after etching with aqua regia, as can be observed in Figure 1, which also includes X-ray maps for Ni, C, Si, Cr, and Fe. It can be appreciated that Ni and Fe are depleted in the region identified as δ, but enriched with C and Cr. The chemical compositions, obtained from EDS analyses, of the phases identified as δ and γ are shown in Table 2. The composition of δ-ferrite obtained is similar to that reported by Tseng et al. (1994). Metallographic examination of the material in its as-received condition did not put in evidence the precipitation of carbides at the grain boundaries; however, the value of DOS obtained fell within the range of lightly sensitized stainless steels.

Figure 1:

Scanning electron micrograph of the steel in its as-received condition showing the occurrence of δ-ferrite after etching in aqua regia. The presence of Ni, C, Si, Cr, and Fe is mapped.

Figure 2:

Atomic force micrograph of the steel in its as-received condition (A) and a line profile to show the localized dissolution at the interphase (B).

Figure 2 shows the AFM profile of a sample in the as-received condition etched with 10% solution of oxalic acid. The profile obtained was similar to that reported by Yi and Shoji (1996) and Femenia et al. (2001), and revealed the etching sensibility at the δ-ferrite/austenite interface. The oxalic acid attack is used in ASTM A-262 Practice A to reveal carbides and chromium-depleted zones (ASTM A799, 2015). Austenite, in the present case, was less sensitive to etching, followed by δ-ferrite, whereas a maximum dissolution depth of around 1150 nm was found to occur at the δ-ferrite/austenite interface. A similar profile was obtained when the material in the as-received condition is etched with aqua regia (Figure 3).

Figure 3:

Atomic force micrograph of the steel in its as-received condition after etching with aqua regia.

Figure 4 shows the microstructure of samples subjected to solution annealing treatments for 1 and 1.5 h at 1050°C, and cooled by immersion in water and etched with aqua regia. The microstructure shows the presence of ribbons of δ-ferrite within the austenite matrix. Higher magnification of the sample allows observing the presence of a ditch around the δ-ferrite that can be related to chromium depletion (Figure 4D). The absence of precipitated carbides in the sample treated for 1.5 h does not explain the value of 0.046 DOS obtained. Table 3 shows the chemical composition of the phases shown in Figure 4. It is appreciated that δ-ferrite is enriched in chromium and depleted in nickel, and the values agree with those reported by Tseng et al. (1994).

Figure 4:

Scanning electron micrograph of the specimen after annealing at 1050°C for 1 h (A and B) and 1.5 h (C and D), with aqua regia as etchant.

Figure 5A shows the profile obtained by AFM of the sample subjected for 1 h at 1050°C and with oxalic acid; a ditch at the δ-ferrite/austenite interface is appreciated. Figure 5B shows a similar profile in a sample solution treated for 1.5 h at 1050°C, but etched with aqua regia. The localized attack at the δ-ferrite/austenite interface is evident. Figure 6A corresponds to an AFM micrograph for a sample annealed for 1 h at 1050°C and etched with aqua regia. Figure 6B presents the profile along the line indicated in Figure 6A; the deeper attack at the δ-ferrite/austenite interface suggests the depletion of Cr in such a region.

Figure 5:

Atomic force micrographs of the steel annealed at 1050°C for 1 h (oxalic acid etching, A) and 1.5 h (aqua regia etching, B).

Figure 6:

Atomic force micrograph (A) of the steel after annealing for 1 h at 1050°C, aqua regia etching, and a profile indicating the depth of dissolution (B).

The EPR-DL test has been established to evaluate the DOS in stainless steel (Majidi & Streicher, 1984). The test consists of conducting a couple of scans in a sample within the device and solution described in Section 2. The first, forward or activation, scan is carried out from the open circuit potential to the passive range. The second, reversed, scan is conducted from the passive to the initial corrosion potential, which leads to the preferential breakdown of the passive film in a sensitized material. The ratio of the maximum current generated in the two loops, I_a/I_r, is used to evaluate the DOS. Figure 7 shows the loops that were registered in two samples from the 304 type steel under study; the first one was in its as-received condition, whereas the second one was treated at 650°C for 3 h to enhance sensitization. Table 4 presents the results obtained by this test in various samples.

Figure 7:

Potential vs. current curves obtained by the EPR-DL test conducted in samples of type 304 stainless steel in as-received and sensitized conditions.

Figure 8A documents the effect of solution annealing at 1050°C on the ferrite number; the decrease of this parameter is due to the decomposition of ferrite. The dotted line in this figure is fit to

Figure 8:

Dependence of the ferrite number in the steel as a function of the annealing time at 1050°C, and that of the DOS as a function of the ferrite number.

where FN is the ferrite number and t is the time in hours. Figure 8B relates the DOS as a function of the ferrite number in the solution treatment. It is found that the steel exhibits a value of DOS above the threshold for a non-sensitized steel of 1×10⁻³, which is associated with Cr-depleted zones, in its as-received condition. Annealing for at least 1 h results in the increase of DOS up to values associated with lightly sensitized steels (in the range of 1×10⁻³–5×10⁻²) (Majidi & Streicher, 1984); further annealing reduces the sensitization below the threshold. Such behavior may be explained in terms of the presence of chromium-depleted zones at the δ-ferrite/austenite interface, which may arise during solidification of the steel (Fredriksson, 1972; Suutala et al., 1980; Kim et al., 2003). The decomposition of δ-ferrite into austenite and chromium carbides, as proposed by Tseng et al. (1994), does not take place at the solution temperature of 1050°C, but chromium would exhibit high mobility and enhanced diffusion; therefore, it would be expected for δ-ferrite to transform into austenite, reducing with it the DOS and the ferrite number.

Some of the samples that were solution annealed at 1050°C were subjected to a sensitization treatment for up to 12 h at 650°C to promote sensitization of the microstructure. It was found that chromium carbides precipitated at the grain boundaries and at the δ-ferrite/austenite interface. Figure 9 shows a series of images obtained by SEM and AFM of a sample sensitized for 1 h at 650°C and etched with the aqua regia solution. Close examination at the δ-ferrite/austenite interface (Figure 9B and C) shows the dissolution of ferrite into a mixture of carbides and austenite, similar to that reported elsewhere (Tseng et al., 1994). The AFM micrograph (Figure 9D) indicates the resistance of the carbides to aqua regia etching, whereas the remnants of ferrite are dissolved. The chemical compositions of the various microstructural constituents, obtained by EDS, are shown in Table 5. The content of 25% of chromium in δ-ferrite agrees with published results (Tseng et al., 1994). The analyses conducted by EDS suggest that the composition of the chromium carbides ranged from MC to M₂C; however, it is considered that further research should be conducted to define their real stoichiometry.

Figure 9:

Scanning electron (A–C) and atomic force (D) micrographs of the steel after solution annealing at 1050°C for 1 h followed by aging at 650°C for 1 h, aqua regia etching.

Figure 10 shows a higher-magnification electron micrograph of the δ-ferrite/austenite interface of the sensitized sample together with the mappings for Cr, Fe, and Ni, in which the enrichment of Cr at the interface is due to carbide precipitation; depletion of iron at the interface and of nickel in δ-ferrite are appreciated. Precipitation of chromium carbides results in increase in sensitization (Bruemmer, 1986; Trillo et al., 1995) and reduction in the ferrite number, as shown in Figure 11. The increase of DOS in samples aged at 650°C (Figure 11) can be attributed to precipitation of chromium carbides; this treatment results in the reduction of ferrite (right hand side of Figure 11), although values obtained by the magnetic balance are within the range of the detection error. Figure 12 shows the X-ray diffraction spectra of samples that were solution treated for 1 h at 1050°C and aged for 2, 4, and 12 h at 650°C, in which the presence of austenite and ferrite is indicated; it is worth noticing the presence of ferrite at 2θ of 64.5° and 82.2° in the sample treated for 12 h, which did not appear for shorter times. This ferrite correspond to 80.1% Fe and 11.7% Cr (JCPDS 034-0396).

Figure 10:

Scanning electron micrograph of the steel in the conditions shown in Figure 8, together with mapping for Cr, Fe, and Ni.

Figure 11:

DOS and ferrite number as a function of time of aging at 650°C; values corresponding to the solution treatment are shown at zero time.

Figure 12:

X-ray diffraction spectra of samples solubilized for 1 h at 1050°C and aged at 650°C for 2, 4, and 12 h.

The decomposition of ferrite during the aging treatment agrees with the results from Tseng et al. (1994), in which the eutectoid reaction δ→γ+M₂₃C₆ is proposed to occur in thin layers at the δ-ferrite/austenite interface. They also proposed that the remaining δ-ferrite should decompose into σ and austenite in the temperature range of 600–900°C. They reported chromium contents in the range of 31–34% and a maximum of 4% Ni (Tseng et al., 1994), which were not detected in the present study, as the time required to precipitate the σ phase in the present steel will be >20 h (Hsieh & Wu, 2012).

The chromium and nickel content (Table 1) should assure fully austenitic steel after the solution treatment at 1050°C, rather than the mixed phase observed in Figures 1–5. Solidification of stainless steel of the 18/8 type proceeds by the precipitation of primary dendrites of δ-ferrite that may be able to transform into austenite during cooling; δ-ferrite is found at room temperature in castings or welds of these steels, as it has a higher concentration of Cr and lower concentration in Ni (Fredriksson, 1972; Suutala et al., 1980; Kim et al., 2003). Wrought stainless steels, such as the 304 type, are hot worked at temperatures around 1200°C after reheating them for long cycles to decompose δ-ferrite (Rezayat et al. 2016); the ribbon shape of δ-ferrite in the steel under study (Figure 4) indicates that δ-ferrite did not fully decompose while hot worked, but remained in the steel.

4 Conclusions

The AISI type 304 steel under study exhibited a certain DOS after being solution annealed at 1050°C. The sensitization is associated with the presence of δ-ferrite ribbons that have higher concentrations of Cr and lower Ni. These ribbons likely originate during solidification, do not decompose during hot working, and are retained at room temperature. The highest DOS was observed in samples that were solution treated for 1 h; longer periods of time contributed to reduced sensitization. This phenomenon is attributed to the decomposition of δ-ferrite, evidenced by the reduction in the ferrite number. Precipitation of chromium carbides was observed in samples subjected to aging at 650°C. These samples exhibited an increase in DOS as a function of time due to the reduction of chromium in solid solution.