Influence of N on precipitation behavior, associated corrosion and mechanical properties of super austenitic stainless steel S32654

1. Introduction

Super austenitic stainless steel (SASS) is an important developmental direction of high performance stainless steels. From the beginning of the previous century, a number of SASSs have been developed to satisfy the demands for high corrosion resistance and long life materials. N08904 (0.10 wt% N), S31254 (0.18-0.22 wt% N) and S32654 (0.45-0.55 wt% N) are three typical types of SASSs [1]. Recently, the most highly alloyed S32654 has become the focus of attention due to its remarkable corrosion resistance and outstanding comprehensive mechanical properties [[2], [3], [4]]. It has large amounts of Cr, Mo, and N and substantial potential for application in exceedingly harsh service environments, such as plate heat exchangers, flue gas desulphurization systems, desalination systems and chemical processing equipment [[5], [6], [7], [8]]. It is apparent that N plays an important role in the development of SASSs.

As well known, N has multiple effects on austenitic stainless steels and alloys [9]. Firstly, N is one of the most important elements that can improve corrosion resistance of steel. It can enhance the passive film stability, thereby increasing local corrosion resistance, particularly pitting and crevice corrosion resistance [[10], [11], [12], [13]]. Secondly, interstitial N atom has significant solid solution strengthening effect. The addition of N to austenitic stainless steels could result in a considerable increment in strength without significant reduction in plasticity and toughness [[14], [15], [16]]. Therefore, N-alloyed austenitic stainless steels generally exhibit better combinations of corrosion resistance and mechanical properties. Furthermore, as a strong austenite stabilizer, N can improve the stability of austenite structure as well as retard the precipitation of intermetallic phases [17,18]. However, excess N in turn promotes the formation of nitrides such as Cr₂N, AlN, and TiN [19,20].

In the past several decades, extensive researches have been carried out on the role of N in stainless steels. However, these have mainly focused on traditional stainless steels and seldom on SASSs. Only Lee et al. [21] investigated the effect of N on the precipitation characteristics of Fe-22Cr-21Ni-6Mo. They found that N could promote fine and uniform precipitation of secondary phases, but the related mechanism is still unclear. There appears to have been no report till now on the influence of N on S32654. As mentioned earlier, S32654 contains significant amounts of alloying elements. It is very sensitive to the precipitation of secondary phases, especially intermetallic compounds. For example, Koutsoukis et al. [22] found that four precipitates (σ, χ, and Laves phases and Cr₂N) formed in S32654. Moreover, they indicated that cold deformation could accelerate precipitation [23]. Heino et al. [24] discovered that in addition to σ, χ and Laves phases, R phase precipitated in S32654. Song et al. [25] pointed out that at least eight precipitates (σ, χ, μ, and Laves phases, M₃C, M₆C, M₂₃C₆, and Cr₂N) formed in 654SMO aged at 800-1100 °C. In our previous study [26], we firstly detected cellular precipitation of σ phase in S32654. We also observed two intermetallic phases (σ and R phases) and two nitrides (Cr₂N and π phase) in S32654 aged at its nose temperature [27,28]. It can be seen that the precipitation behavior of S32654 is highly complex and continues to be a matter of contention. Under such conditions, it is worth exploring what role N plays in so complicated precipitation process. Additionally, it is well acknowledged that the deleterious secondary phases degrade both the corrosion resistance and the mechanical properties of steel [[29], [30], [31], [32]]. Therefore, it is highly necessary to clarify the influence of N on the corrosion and mechanical properties of S32654 after precipitation.

In this work, the effect of N on the precipitation behavior of S32654 was first investigated at 1000 °C. The variations of driving force for precipitation and element activity with N content were calculated using Thermo-Calc software to reveal the influence mechanism of N on the precipitation. In addition, the evolutions of the intergranular corrosion (IGC) susceptibility and tensile properties with N content and aging time were also investigated. The corresponding corrosion and fracture failure mechanisms were proposed. These investigations will provide good guidance for the alloy composition optimization, hot working and heat treatment process formulation of SASSs.

2. Experimental procedure

2.1. Materials and heat treatments

The chemical compositions of three super austenitic stainless steels S32654 with various N contents were presented in Table 1. The steels were produced using a vacuum induction furnace. These steel ingots were first homogenizing annealed at 1280 °C for 10 h and then hot forged and hot rolled between 1150-1200 °C. Subsequently, the experimental samples were solution treated at 1200 °C for 30 min. Fig. 1 shows the typical microstructures of the samples after solution treatment. Obviously, all the three steels consist of full austenitic structure, and no secondary phase was observed. The average ASTM grain size of the solution-treated sample was 106 μm for the L-N steel, 94 μm for the M-N steel and 65 μm for the H-N steel. And they were essentially maintained throughout the aging process. All isothermal aging treatments were conducted in a tube furnace at 1000 °C for 30 min - 6 h, followed by water quenching.

2.2. Microstructure characterization

The samples for metallographic observation were gradually ground from 240# to 2000# and polished with 1.5 μm diamond paste. Then these samples were electrolytically etched at 1 V for 3-20 s in 3 g oxalic acid +100 mL hydrochloric acid. The sample used for preliminary microstructure characterization was first observed by optical digital microscope (ODM), and at least 50 different micrographs were collected. Then, the OLYCIA m3 software [33] was applied to calculate the area fraction of secondary phases. Subsequently, the morphologies and compositions of the precipitates were further analyzed using a scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS). The identification of various precipitates was carried out in a Tecnai G² 20 transmission electron microscopy (TEM). TEM samples were first grounded to 50 μm thickness and then twin-jet electropolished in a solution of 8 vol.% perchloric acid +92 vol.% ethanol at 40 V and -20 °C. Additionally, the effect of N on the driving force for precipitation and the alloying element activity were calculated using Thermo-Calc software with TCFE9 database [34,35].

2.3. Intergranular corrosion tests

A modified double loop electrochemical potentiokinetic reactivation (DL-EPR) method [27] was employed to explore the effect of N on the intergranular corrosion susceptibility of S32654. The DL-EPR tests were performed using a potentiostat PARSTAT 2273 with a three-electrode cell: the reference electrode (a saturated calomel electrode, SCE), the counter electrode (a platinum foil), and the working electrode (aged samples). The samples were embedded in epoxy resin and gradually ground to 2000# and then polished with 1.5 μm diamond paste. The test medium was 2 M H₂SO₄ + 3 M HCl + 0.01 M KSCN solution at 25 °C. During the electrochemical tests, the potential was first swept from the open circuit potential (OCP) to +200 mV (vs. SCE) (forward scan) and held for 2 min and then reversed back to OCP (reverse scan) [36]. The scan rate was chose as 1.667 mV/s. During each scan direction, the peak current density was measured respectively, i.e. peak activation current density (I_a) for the forward scan and peak reactivation current density (I_r) for the reverse scan. The DL-EPR value (R) was calculated by R = (I_r/I_a) × 100, which represents the degree of sensitization (DOS) [37,38]. For each sample, the final R value is an average of at least five tests. After DL-EPR test, the microstructure was examined by ODM, and at least 50 different micrographs were recorded to calculate the average IGC attack area.

2.4. Tensile tests

To explore the influence of N on the mechanical properties of S32654 samples aged at 1000 °C, a series of tensile tests were conducted using a Shimadzu AG-X plus (200 kN) machine. According to ASTM E8/E8M-16a standard [39], flat dog-bone shaped samples with a cross-section of 6 mm × 5 mm and a gauge length of 25 mm were machined and tested at room temperature with a constant cross head speed of 1 mm/min. For each test, at least three samples were tested to get the mean value of the tensile properties. After tensile test, the microstructure near fracture was examined by SEM to reveal the fracture reason.

3. Results

3.1. Precipitation analysis

3.1.1. Microstructure observation

Fig. 2 shows the microstructures of the three S32654 steels aged at 1000 °C for 30 min - 6 h. The precipitation kinetics of the L-N steel appeared to be significantly faster. After 30 min of aging (Fig. 2(a)), almost all the grain boundaries were covered with σ phase. Meanwhile, a small amount of σ phase formed randomly in the austenite grain interior. As the aging time increased (Fig. 2(b) and (c)), the intergranular σ phase gradually grew, and more large σ phase precipitated within the austenite grains. In particular, numerous intragranular σ phases grew into large needles after 6 h of aging (Fig. 2(c)). As the N content increased, the precipitation behavior of S32654 presented several noteworthy variations (Fig. 2(d)-(i)). Firstly, increasing the N content led to grain refinement and an increment in grain boundaries. This phenomenon provides more favorable sites for the formation of intergranular precipitates. Thus, more σ phase as well as some nitrides (Cr₂N and π phase) precipitated along the grain boundaries in the M-N and H-N steels. In comparison, the two nitrides nucleated much earlier in the H-N steel. It should be noted that increasing the N content appears to limit the nucleation and growth of intergranular σ phase. Therefore, the number and size of the intergranular σ phase per unit grain boundary length in the H-N steel were marginally smaller than that in the L-N steel (Fig. 3). However, the total grain boundary length of H-N steel was much larger than that of L-N steel (Fig. 1, Fig. 2). Therefore, the total amount of intergranular σ phase still increased with the increment of N content. Moreover, the size and number of the intragranular σ phase gradually decreased with the increase of the N content (Fig. 2(d)-(i)). Fine block-like and rod-like σ phase rather than large needle-like σ phase formed in the H-N samples. All the above variations reflect the fact that increasing the N content promotes the formation of intergranular σ phase, Cr₂N, and π phase and inhibits the precipitation of intragranular σ phase.

All the precipitates formed in S32654 were identified by TEM analyses. Fig. 4 shows the TEM bright-field micrographs and selected area electron diffraction (SAED) patterns of various precipitates. As shown in Fig. 4(a) and (b), the continuous and coarse intergranular precipitates as well as the typical needle-like intragranular precipitate were identified as σ phase. Their lattice parameters were determined as follows: intergranular σ phase (a = 0.882 nm and c = 0.453 nm) and intragranular σ phase (a = 0.883 nm and c = 0.455 nm). These are largely consistent with consistent with the reported values (a = 0.880 nm and c = 0.454 nm) [26,40]. In addition, the two nitrides were verified to be Cr₂N (a = 0.4796 nm and c = 0.4472 nm) and π phase (a = 0.637 nm), as shown in Fig. 4(c) and (d). The lattice parameters of the two nitrides are also essentially in accordance with the reported values [21,27,41].

3.1.2. Area fraction

In order to quantitatively evaluate the effect of N on the precipitation of S32654, the area fractions of intergranular, intragranular and total precipitates were measured using OLYCIA m3 software (Fig. 5). For the same steel, all the area fractions of intergranular, intragranular and total precipitates remarkably increased with the aging time prolonging. For different steels under same aging condition, these three groups of statistical data showed different variation regularities. As the N content increased, the area fraction of intergranular precipitates increased gradually, that of intragranular precipitates decreased gradually, and that of total precipitates first decreased and then increased. Additionally, the difference in the precipitate area fraction among the three steels was more evident during long-term aging. These variation regularities further indicate that increasing the N content accelerates the precipitation of intergranular precipitates and suppresses the formation of intragranular precipitates.

3.1.3. Chemical composition

The chemical compositions of σ, π phase and Cr₂N were measured by EDS, as presented in Table 2. The precipitate larger than 1 μm was selected for measurement. At least five particles were measured to obtain the average chemical composition of each phase. It is apparent that both σ and π phases are enriched in Cr and Mo, and Cr₂N is primarily rich in Cr. In addition, a certain amount of N was also detected in both π phase and Cr₂N. In fact, they are indeed the nitrides that generally precipitate in SASSs [23,41,42]. The elemental compositions of these precipitates are basically in agreement with those reported in the literature [26,27,41,42].

3.2. Intergranular corrosion tests

3.2.1. DL-EPR curves

Fig. 6 shows the DL-EPR curves obtained for the three S32654 steels after solution and aging treatments. A wide passivity range (0-200 mV) is observed in each DL-EPR curve. After solution treatment (ST), the samples did not present reactivation current peak during the reverse scan. This implies that the solution-treated samples did not suffer from intergranular sensitization. The aged samples of each steel exhibited apparent reactivation current peaks in the reverse curves. Furthermore, the reactivation current peak gradually increased with the prolonging of aging time, indicating a growing DOS.

3.2.2. IGC attack observation

Fig. 7 illustrates the ODM morphologies of the samples after the DL-EPR tests. The solution-treated samples presented no indications of IGC attack, whereas all the aged samples sustained varying degrees of IGC attack. It is evident that the regions around the intergranular and intragranular precipitates were severely corroded. In order to quantitatively examine the difference in the extent of IGC attack among the three steels, the intergranular, intragranular and total IGC attack areas of each sample were measured by OLYCIA m3 software (Fig. 8). As expected, the IGC attack areas exhibited similar variation regularities similar to those of the area fraction of precipitates (Fig. 5). That is, the intergranular and intragranular IGC attack areas gradually increased and decreased, respectively, as the N content increased. Moreover, the total IGC attack areas first decreased and then increased with the increase of the N content. This means that increasing the N content first weakens and then enhances the IGC susceptibility of S32654.

3.2.3. DL-EPR values

Fig. 9 shows the DL-EPR values (R) of the three S32654 steels aged at 1000 °C for various time periods. Similarly, the R value also presented variation regularities similar to those of the area fraction of precipitates (Fig. 5). That is, the R value of the same steel increased significantly with the prolongation of aging time, indicating a gradually increasing DOS. For different steels, the R value first decreased and then increased with the increment of N content. This further reveals that increasing the N content first decreases and then enhances the IGC susceptibility of S32654.

3.3. Tensile tests

3.3.1. Strength and elongation

Fig. 10 depicts the ultimate tensile strength (UTS), yield strength (YS), and elongation of the three S32654 steels as a function of aging time. The solution-treated samples exhibited high levels of UTS ranging from 815 to 880 MPa. Moreover, their UTS increased gradually with the increment of N content. After 30 min of aging, there was a marginal decrease in the UTS of all these steels. During the subsequent aging process, the UTS of these steels decreased sharply as the aging time increased. However, compared with the UTS, the YS presented contrary variation trends. That is, the YS of these steels increased significantly with the prolongation of the aging time. A special phenomenon is that the YS of the H-N steel exhibited a remarkable decrease at the initial stage of aging. The reason for this will be discussed in Section 4.3. It is also noteworthy that both the UTS and YS are always higher in the order of L-N, M-N and H-N steels throughout the solution and aging processes. This demonstrates that increasing the N content exerts a significant strengthening effect on both the solution-treated and aged steels. The elongation exhibited variation characteristics significantly different from those of the strength. Firstly, the elongation was much more sensitive to the aging treatment than the strength was. After only 30 min of aging, the elongations of all the steels decreased dramatically from 73 % to 75 % (the solution-treated samples) to 48 %-51 % (the samples aged for 30 min). Upon continued aging time to 2 h, the elongations reduced further to rather low levels (18 %-24 %). When aged for 6 h, the elongations of the samples decreased to less than 15 %. However, the variation in the elongation with the N content was not as apparent as that of the strength. The elongations of the three solution-treated samples were largely identical. The elongations of the aged samples differed only marginally during long-term aging. That is, the elongation of the M-N steel was marginally higher than that of the H-N steel and then that of the L-N steel. This indicates that the influence of N on the elongation of S32654 is not significant.

3.3.2. Microstructure near fracture surface

In order to identify the reason for fracture, the samples were cut open along the tensile direction after the tensile tests. Fig. 11 shows the cross-section microstructure near the fracture surface of S32654 steels aged at 1000 °C for 30 min - 6 h. The aged steels with different N contents exhibited an identical fracture mode: intergranular fracture. For the samples aged for 30 min (Fig. 11(a), (d), and (g)), although the intergranular precipitates were few and small, they were adequate for originating and propagating cracks along grain boundaries. Meanwhile, the austenitic grains were elongated apparently, indicating that the austenitic matrix still exhibited a certain amount of plasticity. As the aging time increased to 2 and 6 h, the intergranular precipitates coarsened and grew up, resulting in more apparent intergranular fracture. Meanwhile, more and larger precipitates formed within the austenitic grains. Under the combined actions of intergranular and intragranular precipitates, the deformation degree of the austenitic grains decreased, revealing a gradual decrease in plasticity. This phenomenon corresponds to the variations in elongations with the aging time (Fig. 10(b)).

Fig. 12 illustrates the magnified SEM images of the cross-section microstructure near the fracture surface. Several more clear characteristics are observed. Firstly, neither the intergranular nor intragranular precipitates underwent shape-alteration and rotation after the tensile tests. The cracks were initiated from the grain boundaries and propagated across the precipitates. As a result, most intergranular precipitates broke or cracked along the tensile direction, as indicated by the yellow arrows in Fig. 12. Furthermore, a few intragranular precipitates were also pulled apart into several segments along the tensile direction, as indicated by the green arrows in Fig. 12. Moreover, σ phase was mainly responsible for the intergranular fracture owing to its larger size and number. Apart from the σ phase, the marginal amounts of π phase and Cr₂N could also have caused the cracking (Fig. 12(b) and (c)).

4. Discussion

4.1. Influence of N on precipitation behavior

The above experimental results (Fig. 2, Fig. 5) reveal that increasing the N content provides more favorable sites for the formation of intergranular σ phase, Cr₂N and π phase and significantly inhibits the precipitation of intragranular σ phase. Therefore, the area fraction of total precipitates first decreased and then increased with N content increasing. The influence of N on the precipitation behavior could be explained as follows.

As is established, the nucleation site, driving force and element activity play important roles in the precipitation of secondary phases [8,43,44]. Moreover, these three factors are remarkably affected by the chemical composition of the steel. In the present study, it is apparent that the N content significantly affects the nucleation sites (grain boundaries) for precipitates (Fig. 1, Fig. 2). However, the influences of N on the driving force and element activity are still unclear, because it is challenging to verify them experimentally. Hence, effective calculations were carried out using Thermo-Calc software to clarify this issue, as shown in Fig. 13. It is apparent that both the driving force for σ phase and Cr activity linearly decrease with the N content. In contrast, increasing the N content leads to linear increments in both the driving force for Cr₂N and Mo activity. Taking the above results into consideration, the influence of N on the intergranular and intragranular precipitates will be detailedly discussed, respectively.

The influence of N on intergranular precipitates is mainly attributed to its effects on the nucleation site, driving force and element activity. Firstly, the austenitic grains were apparently refined as the N content increased. This implies that the number of grain boundaries in the steel would increase significantly. It is well established that grain boundaries have high energy and could provide a rapid diffusion channel for the precipitate forming elements (Cr, Mo, N, etc.) [41,45,46]. Thus, they are generally considered as favorable nucleation sites for precipitates. That is, nucleation sites exhibit a positive association with the number of grain boundaries. Therefore, it is easy to understand that the number of nucleation sites would also increase with N content increasing. As a corresponding result, more intergranular precipitates formed in the M-N and H-N steels (Fig. 2, Fig. 5). As observed in Fig. 3, increasing the N content limited the nucleation and growth of intergranular σ phase, marginally decreasing the number and size of intergranular σ phase per unit grain boundary length. This is attributed to the significant reductions in both the driving force for σ phase and the Cr activity (Fig. 13). Although the number and size of intergranular σ phase per unit grain boundary length decrease marginally, the total amount of grain boundaries increases considerably with the increment of the N content (Fig. 1, Fig. 2). Thus, the total amount of intergranular σ phase would still increase as the N content increases. Apart from that, the increase in the N content also promoted the formation of two nitrides: Cr₂N and π phase. The reasons for this are discussed as follows. Firstly, the driving force for Cr₂N was substantially enhanced with the N content increasing (Fig. 13(a)). This is conducive to the precipitation of Cr₂N. As an important forming element of Cr₂N, N is more sufficient in the M-N and H-N steels. This is also beneficial for the formation of Cr₂N. However, as another essential forming element of Cr₂N, Cr is limited owing to its lower activity at a higher level of N content (Fig. 13(b)). It is challenging for Cr₂N to grow to a large size and amount. Thus, under the combined effects of the above aspects, a few small Cr₂N precipitated in the M-N and H-N steels (Fig. 2). The π phase is generally reported as a metastable phase in SASS [25]. Therefore, the driving force for π phase was not calculated in this work. Table 2 illustrates that Cr, Mo and N are three key elements for the formation of π phase. Increasing the N content not only provides sufficient N element but also improves the Mo activity (Fig. 13b). Both the features are favorable for the precipitation of π phase. As mentioned above, the Cr activity linearly decreased with the N content increasing (Fig. 13(b)). This in turn limited the formation of π phase. Thus, only a few π phases were observed in the M-N and H-N steels (Fig. 2).

The precipitation behavior of intragranular precipitates was more apparently affected by the N content. Both the number and size of intragranular σ phase gradually decreased with the N content increasing. Certain fine block-like and rod-like σ phase rather than large needle-like σ phase formed in the H-N steel (Fig. 2). All these variations are mainly related to the N solubility in σ phase and the variations in the driving force and element activity with N content. Several previous studies have indicated that N exhibits a very low or zero solubility in σ phase [21,41,47]. The chemical composition in Table 2 also verifies this fact. From this perspective, a low-N condition is conducive to the precipitation of σ phase. Furthermore, the driving force for σ phase is considerably high in the L-N steel (Fig. 13(a)). This could also provide a favorable condition for the nucleation of σ phase. Although the Mo activity is lower under a low-N condition (Fig. 13b), the diffusion of Mo is faster than that of the other elements (Fe, Cr, etc.) [42,48]. Moreover, the Cr activity is high enough in the L-N steel. Therefore, these conditions could also satisfy the nucleation and growth of σ phase. Consequently, numerous σ phases formed within the austenitic grains in the L-N steel and finally grew into large needles (Fig. 2(a)-(c)). With the N content increasing, both the driving force for σ phase and the Cr activity decreased to rather low levels (Fig. 13). Under such conditions, the nucleation and growth of σ phase were severely constrained. Thus, the number and size of intragranular σ phase progressively decreased with the N content increasing. The morphology of intragranular σ phase also altered from needle-like to fine block-like and rod-like. In our previous study [26], we observed that the nucleation and growth of needle-like σ phase follows the following process: once the intragranular σ phase is nucleated, a coherent or semi-coherent interface and an incoherent interface forms around it. These interfaces must migrate to satisfy the σ phase growth. Because σ phase (BCT) has a different crystal structure from austenite phase (FCC), the mobility of the coherent or semi-coherent interface is rather low, whereas that of the incoherent interface is very high. Thus, the intragranular σ phase is more likely to grow into needles. In the present study, the Cr activity linearly decreased with the N content increasing. The diffusion of Cr used for σ phase growth slows down, which may reduce the mobility of the incoherent interface. In this case, the intragranular σ phase tends to grow into blocks and rods rather than needles. Finally, only some fine block-like and rod-like σ phases formed in the austenite grain interior of the H-N steel (Fig. 2(i)).

4.2. Influence of N on intergranular corrosion susceptibility

The IGC attack observation (Fig. 7) and the DL-EPR results (Fig. 8, Fig. 9) reveal that increasing the N content first weakens and then enhances the IGC susceptibility of S32654. Moreover, the IGC susceptibility exhibits variation regularity similar to that of the area fraction of precipitates (Fig. 5, Fig. 9). In order to clarify the influence mechanism of N on the IGC susceptibility, the cause of the IGC of S32654 must be first understood.

In general, the IGC of stainless steels is mainly attributed to the Cr depletion surrounding the precipitates [45,[49], [50], [51], [52], [53], [54]]. In our previous studies [27,28], we have detected apparent Cr- and Mo-depleted zones adjacent to both σ and π phase. We also observed a Cr-depleted zone in the vicinity of Cr₂N. Thus, it could be concluded that the IGC of S32654 was mainly caused by the Cr- and Mo-depleted zones around the precipitates. In the present study, noticeable IGC attack regions surrounding the precipitates were observed (Fig. 7). This further verifies that the Cr and Mo depletions around the precipitates cause the IGC of S32654. The IGC process in the DL-EPR test has been revealed as follows [27]: during the forward scan, a full passivation film was generated on the surface of the sample. However, the passivation film over the depleted zones was easily to be damaged in the reverse scan owing to its low stability. This resulted in a severe IGC attack.

It is well known that the extent of depleted zones depends heavily on the overall amount of precipitates, further determining the IGC susceptibility of the steel [[55], [56], [57]]. In this study, the overall amount of precipitates consists of two parts: intergranular and intragranular precipitates (Fig. 5). Both the parts exert non-negligible influence on the IGC susceptibility. Fig. 7 reveals two facts: firstly, both the regions around each intergranular and intragranular precipitate were severely corroded; secondly, the wider the intergranular precipitate or the larger the intragranular precipitate, the more extensive the IGC attack areas surrounding it. In general, the IGC attack areas are governed by the extent of depleted zones. Therefore, these two facts indicate that the IGC susceptibility of S32654 is related to the extent of depleted zones, which is considerably dependent on the total amount (number and size) of precipitates. In summary, the influence of N on the IGC susceptibility of S32654 is mainly determined by its effect on the precipitation. As shown in Fig. 2, Fig. 5, the area fraction of total precipitates first decreased and then increased as the N content increased. This implies that the extent of depleted zones in the steel would also first exhibit a decrease and then an increase with the increment of the N content. As mentioned earlier, the IGC susceptibility relies directly on the extent of depleted zones. Thus, it is easy to understand why increasing the N content first weakens and then enhances the IGC susceptibility of S32654 (Fig. 9). In fact, the actual total IGC attack areas also first decreased and then increased with the increase of N content (Fig. 7, Fig. 8). It is noteworthy that during long-term aging, the intragranular and intergranular precipitates mainly contribute to the IGC susceptibility of L-N and H-N steel, respectively. The intragranular and intergranular IGC attack areas occupy the dominant role in the total IGC attack areas of L-N and H-N steel, respectively. These results further verify the accuracy of the above statements.

4.3. Influence of N on tensile properties

As observed from the tensile test results (Fig. 10, Fig. 11, Fig. 12), N substantially influences on the tensile properties of both the solution-treated and aged S32654 samples. The UTS sharply decreased as the aging time increased. Meanwhile, the YS exhibited contrary changing trends. Increasing the N content enhanced both the UTS and YS significantly throughout the solution and aging processes. Compared with the strength, the elongation was much more sensitive to aging treatment. However, it varies marginally with the N content. Additionally, all the aged steels with various N contents basically exhibited identical fracture mode: intergranular fracture. The above phenomena are discussed in detail below.

In general, the strengthening mechanism of austenitic stainless steels mainly includes three types: solid solution strengthening, grain refinement strengthening and precipitation strengthening. It is well accepted that N is the strongest solid solution strengthening element in austenitic stainless steels. The strengthening effect of N is mainly achieved through two means: firstly, the N atoms interact with dislocations to form N-dislocation complexes, thus pinning the dislocations and hindering their movement [14,58]; secondly, the interaction between interstitial N atoms and other displacement atoms alters the distribution of the alloying elements and promotes the formation of short-range order. This increases the resistance to slip deformation [59,60]. Both the means can markedly improve the strength of the steel. In addition to solid solution strengthening, N can also refine austenitic grains. This results in significant grain refinement strengthening. The principle of grain refinement strengthening is that the grain boundaries block dislocation slip during deformation. This results in dislocation pile-up at grain boundaries, thus improving the steel’s strength [61]. The main cause of precipitation strengthening is that the hard and brittle precipitate could function as a barrier to the dislocation movement, thus increasing the strength of the steel [62,63]. In this work, all the above three strengthening mechanisms should be present and play different roles with the variations in the aging time and N content.

For the solution-treated samples, no secondary phase was observed, and the austenitic grains were noticeably refined with the N content increasing (Fig. 1). Hence, the solid solution and grain refinement strengthening should be mainly responsible for the increments of both the UTS and YS of the M-N and H-N steels (Fig. 10(a)). In particular, the YS of the H-N steel is much higher than that of the other two steels. This indicates that the solid solution and grain refinement strengthening effects are much more apparent under high-N condition. For the aged samples, an increasing amount of precipitates formed in the steel during the aging process (Fig. 2), and the UTS and YS exhibit contrary changing trends as the aging time prolonged (Fig. 10(a)). It could be inferred that these precipitates exert opposite effects on the UTS and YS. The related reasons are discussed separately as follows.

For the UTS, these precipitates play a deteriorating rather than a strengthening role. Thus, the UTS of these steels sharply decreased with the aging time increasing (Fig. 10(a)). The deterioration effect of precipitates on the UTS arises from two aspects. On the one hand, the precipitation of σ phase, Cr₂N, and π phase consumed an exceptionally amount of solid solution atoms (Cr, Mo, N, etc.), reducing the solid solution strengthening. On the other hand, the intermetallic σ phase and the nitrides Cr₂N and π phase are hard and brittle precipitates [64]. It is challenging for these precipitates to absorb energy by their own deformation or rotation during the deformation process. On the contrary, they are easily cracked and separated from the matrix, as observed in Fig. 12. Both the adverse aspects significantly weaken the UTS of the steel. In addition, the deteriorating effect is associated with the size and amount of precipitates [62,65]. During long-term aging, more and larger precipitates formed in the steels, thereby resulting in more noticeable deterioration of the UTS. Furthermore, the total amount of precipitates in the M-N steel is apparently the lowest among those in the three steels (Fig. 5). Therefore, the UTS of the M-N steel decreased more gradually over the long period of aging (Fig. 10(a)).

For the YS, the numerous precipitates have an apparent precipitation strengthening effect on it. Therefore, the YS of these steels increased significantly with the prolongation of aging time (Fig. 10(a)). However, there is an abnormal phenomenon wherein the YS of the H-N steel decreased markedly from the solution state to the initial aging state (30 min). This could be explained as follows. As mentioned earlier, the high YS of the solution-treated sample was obtained mainly by the combination of the solid solution and grain refinement strengthening. After 30 min of aging, in addition to σ phase, some Cr₂N also precipitated along the grain boundaries (Fig. 2(g)). Hence, large amounts of Cr and Mo and some N were consumed from the matrix, resulting in a decrease in solid solution strengthening. In this case, the precipitation strengthening may be not capable of compensating the loss of solid solution strengthening. Under the competitive influence of the two strengthening mechanisms, the H-N samples aged for 30 min exhibited a decrease in the YS. During the subsequent aging process, the YS of all the three steels kept increasing with the prolongation of the aging time. This is because the growth and coarsening of the precipitates make their effect of hindering dislocation movement more apparent. This would promote the increment in precipitation strengthening to surpass the reduction in solid solution strengthening, thus increasing the YS. The precipitation strengthening also depends heavily on the size and amount of the precipitates. Thus, these steels exhibit very high YS after long-term aging. Compared with the L-N and H-N steels, the M-N steel also exhibited a slower increasing rate of YS due to its lower total amount of precipitates. It should also be noted that the total amount of precipitates first decreased and then increased (Fig. 5), whereas both the UTS and YS always increased with the N content increasing (Fig. 10(a)). This is because the combined action of the solid solution and grain refinement strengthening has always been the dominant influence on the strength of steel.

The elongation was considerably sensitive to the aging treatment, whereas it exhibited marginal variation with the N content (Fig. 10(b)). After solution treatment, the three steels with different N contents exhibited essentially identical elongations. It is well accepted that solid solution strengthening of N deteriorates the plasticity of steel, whereas grain refinement can significantly improve it [66]. As observed in Fig. 1, the increment of the N content refined the austenitic grains of S32654 remarkably. Thus, the enhancement of plasticity caused by the grain refinement could compensate for the decrement of plasticity induced by solid solution strengthening. Finally, under the combined influence of the two, the elongations of the solution-treated steels varied negligibly with the N content. In general, the plasticity of steel decreases significantly when precipitation strengthening occurs. Accordingly, the elongations of the three steels exhibited a sharp reduction in the initial stage of aging (Fig. 10(b)). This could be attributed to the intergranular failure and separation caused by the coverage of precipitates on the grain boundaries, as shown in Fig. 11. Because both σ phase and Cr₂N are hard and brittle, the formation of these precipitates significantly suppresses the grain boundary deformation, thereby deteriorating the plasticity of the steel. It should be recognized that the reduction of the elongation is also heavily dependent on the size, distribution and amount of the precipitates. With the aging time prolonging to 2 and 6 h, the intergranular precipitates grew and coarsened. Moreover, extensively large precipitates formed within the austenitic grains (Fig. 2). The deformation of both grain boundaries and grains became more difficult (Fig. 11), and the intergranular cracking and separation became more straightforward (Fig. 12), resulting in further degradation of the elongations. Additionally, there was no significant variation in elongations with the N content increasing during the whole aging process (Fig. 10(b)). This arises from the fact that the difference of the joint influence of solid solution strengthening, grain refinement, and precipitation on the plasticity is not large between the different steels. Only after long-term aging, did the M-N steel exhibit slightly higher elongation than the L-N and H-N steels. This is also because the total amount of the precipitates in the M-N steel is the lowest among those in the three steels (Fig. 5). Therefore, it exerts a smaller deterioration effect on the plasticity.

5. Conclusions

In this study, the influence of N on the precipitation behavior, associated corrosion, and mechanical properties of S32654 was investigated. The main conclusions are summarized as follows:

(1) Increasing the N content results in several alterations in the steel: (i) grain refinement, which promotes the formation of intergranular precipitates; (ii) a linear increase in the driving force for Cr₂N and Mo activity, which accelerates the precipitation of intergranular Cr₂N and π phase, respectively; (iii) a linear decrease in the driving force for σ phase and Cr activity, which suppresses the formation of intragranular σ phase. As a result, the total amount of precipitates first decreases and then increases with the N content increasing.

(2) The IGC susceptibility depends heavily on the extent of Cr- and Mo-depleted zones, which is mainly decided by the total amount (number and size) of precipitates. Thus, increasing the N content also leads to a first decrease and then an increase in the IGC susceptibility.

(3) Aging precipitation causes a considerable decrement in the UTS and a remarkable increment in the YS. The deterioration of the UTS mainly arises from two reasons: firstly, the reduction in the solid solution strengthening effect caused by the consumption of massive atoms from solid solution to precipitation; secondly, the easy cracking and separation of hard and brittle precipitates from the matrix. The enhancement in the YS is mainly attributed to the precipitation strengthening of these precipitates. Additionally, both the UTS and YS always increase with the N content increasing throughout the solution and aging processes. This is because the combined action of solid solution and grain refinement strengthening has always been the dominant influence on the strength of steel.

(4) The elongation is considerably sensitive to the aging treatment, because the hard and brittle precipitates suppress grain boundary deformation and cause intergranular failure and separation of steel. Furthermore, the elongation exhibits marginal variation with the increment of N content. This is because the joint influence of solid solution strengthening, grain refinement and precipitation on the plasticity does not vary significantly with the N content increasing.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time due to technical limitations.

Acknowledgements

This work was financially supported by National Natural Science Foundation of China (No. U1860204), the Fundamental Research Funds for the Central Universities (No. N172507002) and the Transformation Project of Major Scientific and Technological Achievements in Shenyang (No. Z17-5-003).

---