Effect of partial replacement of carbon by nitrogen on intergranular corrosion behavior of high nitrogen martensitic stainless steels

1. Introduction

As a key component of machinery, bearings require high standard of both mechanical properties and corrosion resistance [1]. The martensitic stainless steels (MSSs) containing high level of carbon (C) and chromium (Cr) are widely used in bearings because of their superior hardness, strength and wear resistance [2,3]. Nevertheless, the precipitation of coarse eutectic carbides would deteriorate the fatigue life and corrosion resistance [4,5]. Previous studies revealed that nitrogen (N) existed in MSSs in the form of nitrides and solid solution state [[6], [7], [8]], and the N-induced short range atomic ordering could prevent Cr-clustering and delay the precipitation of Cr-rich carbides [9,10]. Thus, the addition of N in MSSs could increase hardness and strength [11,12], improve corrosion resistance [6,[13], [14], [15]] and fatigue life [3].

It is worth noting that the MSSs alloyed with both C and N exhibit more homogeneous distribution of Cr than those of the C or N alloyed MSSs [9,10], which provide an efficient solution for solving the contradiction between good mechanical properties and high corrosion resistance of bearing steels. Cronidur 30 with 0.3 wt.% C and 0.4 wt.% N shows superiority in microstructure, mechanical properties and corrosion resistance, and has been widely applied in fuel pump bearing of spacecraft, main bearing of engine, etc. [3,7]. However, excess N is also detrimental to corrosion resistance due to the massive precipitation of Cr-rich nitrides [6]. Besides, much higher pressure is needed in the manufacturing process to obtain higher N content in high nitrogen MSSs, which would increase the costs and risks of production. Therefore, it is necessary to reasonably match the C and N contents in MSSs. Berns et al. [16] reported that up to 0.7 wt.% C + N could be dissolved into austenite with 15 wt.% Cr at a austenitizing temperature of 1050 ℃. Therefore, three high nitrogen MSSs with C + N ≈ 0.7 wt.%, i.e. 0.50C-0.16 N, 0.35C-0.37 N and 0.20C-0.54 N, were manufactured in order to tailor the precipitation and avoid excess precipitation of carbides or nitrides. According to our previous work, the mechanical properties of high nitrogen MSSs with different C and N contents are listed in Table 1. It shows that the partial substitution of C by N first reduces the strength and improves impact toughness of 0.35C-0.37 N steel, and then increases the strength and deteriorates impact toughness of 0.20C-0.54 N steel. The 0.35C-0.37 N steel containing medium C and N contents provides good combination of hardness, strength and toughness.

Intergranular corrosion (IGC) is a potential corrosion form of the MSSs, which always work in the tempered condition [17]. The Cr-rich precipitates formed in the tempering process together with undissolved precipitates in the austenitizing process distribute along prior austenite grain boundaries and martensitic laths [17,18]. Simultaneously, the Cr-depleted zones emerge adjacent to these precipitates [[19], [20], [21]], which are preferential sites for IGC. The addition of N induces the formation of Cr-rich nitrides, which also have an impact on IGC, yet there are few studies focusing on the effect of partial replacing C by N on IGC resistance of MSSs.

The present work aims at investigating the effect of partial replacement of C by N on microstructure and IGC behavior of high nitrogen MSSs, thus optimizing the chemical composition and obtaining good combination of mechanical properties and IGC resistance. The microstructure evolution was analyzed using scanning electron microscope (SEM), X-ray diffraction (XRD), transmission electron microscope (TEM) and Thermo-Calc calculation. And nitric acid tests and double-loop electrochemical potentiokinetic reactivation (DL-EPR) tests were applied to evaluate the IGC resistance.

2. Experimental procedure

2.1. Materials preparation

In order to explore the effect of partial replacement of C by N on IGC behavior of high nitrogen MSSs, three types of steels with different C and N contents were smelted by a 25 kg pressurized induction furnace under different nitrogen pressures [22,23]. The chemical compositions of the steels are listed in Table 1, and the steels are named according to the C and N contents as 0.50C-0.16 N, 0.35C-0.37 N and 0.20C-0.54 N, respectively. To eliminate the dendritic segregation, the steels were subjected to diffusion annealing at 1260 ℃ for 15 h, and then cooled to room temperature in the furnace. Afterwards, they were hot forged into round bars with diameter of 70 mm in the temperature range of 1000-1190 ℃. Then the bars were annealed at 875 ℃ for 5 h, furnace cooled to 700 ℃ and kept for 3 h, followed by cooling to 600 ℃ in furnace at the speed of 1 ℃/min, finally cooled to room temperature in the furnace. Subsequently, the annealed steels were subjected to austenitizing treatment at 1040 ℃ for 30 min. After air-cooling to room temperature, the cryogenic treatment was performed in a cooling chamber, with the as-quenched specimens cooling to -80 ℃ at 1-2 ℃/min and then keeping for 2 h. Finally, tempering with air-cooling at 200 ℃ for 2 h was conducted twice.

2.2. Microstructure analysis

After the heat treatment, the metallographic specimens were ground, polished and etched for 4-5 s in the solution of 1 g picric acid, 5 g ferric chloride, 15 mL hydrochloric acid and 50 mL alcohol to observe the size, morphology and distribution of the precipitates using Hitachi SU8000 SEM. The distribution of Cr, C and N was analyzed using the energy dispersive spectrometer (EDS) attached to SEM. To reveal the evolution of precipitation types, the precipitates were extracted by electrolytical dissolution process [6] and then characterized by XRD using D8 advance with Cu Kα radiation at 40 kV, 40 mA and 0.04°/s from 20° to 90°. Furthermore, the thin foils were prepared [22,24] and JEM-2100 F TEM at 200 kV was used to identify the type of precipitates.

2.3. Thermodynamic calculations

The variation of precipitation with austenitizing temperature in the range of 700-1150 ℃ in the equilibrium state was calculated by Thermo-Calc with TCFE7 database.

2.4. Nitric acid tests

To investigate the IGC behavior of high nitrogen MSSs with different C and N contents, the nitric acid tests were performed in an Erlenmeyer flask equipped with a cold finger-type condenser according to ASTM A262-13 [25]. The test solution (65 wt.% nitric acid solution) was prepared by mixing 1 L of reagent grade nitric acid with 108 mL of deionized water. The specimens with the size of 25 mm × 27 mm × 3 mm were immersed in the boiling nitric acid solution for 12 h. Afterwards, to remove the corrosion products on the surface, the corroded specimens were immersed in the solution of 70 g citric acid and 397 mL deionized water in an ultrasonic vibrator for 2 h according to ISO 4807-2009 [26]. The weight of the specimens before and after the tests was measured using an electronic balance with an accuracy of 0.0001 g, and the weight loss was calculated. The surface of the specimen was observed using Carl-Zeiss Ultra Plus SEM, and then the cross section of the specimen was observed by DSX 510 OM after being ground and polished to measure the percolation depth.

2.5. Double-loop electrochemical potentiokinetic reactivation tests

To further evaluate the IGC behavior of the specimens, the DL-EPR tests were performed on Gamry Reference 600 potentiostat equipped with standard three-electrode electrochemical flat cell (a platinum electrode as counter electrode, a saturated calomel electrode (SCE) as reference electrode and a specimen as working electrode) [27]. Since there is no standard electrolyte for 15 wt.% Cr MSSs, the test solution was referred to a recent research [17], which reported that the DL-EPR values conducted in 0.03 M H₂SO₄ fitted well with the results of nitric acid tests. Therefore, the DL-EPR tests were conducted in 0.03 M H₂SO₄ solution at room temperature. Before the tests, the specimens were wet ground to 800-grit on emery papers, then subjected to passivation treatment in 25 wt.% HNO₃ solution at 50 ℃ for 1 h and embedded in the epoxy resin with the exposing area of 1 cm² to the electrolyte. After the solidification of resin, the embedded specimens were further ground to 2000-grit, and then stored in clean and dry sample box after being rinsed and dried. Prior to the tests, the working electrode was immersed in the electrolyte for 20 min to reach steady state. The DL-EPR tests were performed at the scan rate of 1 mV/s from 0.1 V below OCP to 0.5 V/SCE and then reversed to the initial potential. After the DL-EPR tests, the surface morphologies were observed using SEM.

3. Results and discussion

3.1. Microstructure characterization

Fig. 1 shows the SEM micrographs of the investigated high nitrogen MSSs with different C and N contents. There are massive rod-like and spheroidal precipitates along the grain boundaries and in the grain interior of 0.50C-0.16 N steel (Fig. 1(a)), and the size of precipitates was up to $\widetilde{3}$.5 μm. But in 0.35C-0.37 N steel, the content of precipitates is drastically reduced, and the granular precipitates with smaller size (0.2-0.7 μm) in the grain interior are observed (Fig. 1(b)). Similar precipitation variation was also reported by Wang et al. [28], in which Cr-rich precipitates existed along the grain boundaries in Fe-12.8Cr-0.19C steel, while replacing C with N eliminated the precipitates along grain boundaries in Fe-13.22Cr-0.095C-0.07 N steel. However, with further replacing C by N, the content and size (0.2-1.5 μm) of precipitates are inclined to rise again, and a few rod-like precipitates reappears (Fig. 1(c)). Our recent work [6,29] revealed that excess N induced the massive precipitation of Cr-rich M₂N due to the N content exceeding the N solubility in solid solution. Therefore, the partial replacement of C by N first reduces and then increases the content and size of precipitates.

View original graphic| Download| PPT slide

Fig. 1.   SEM morphologies and elemental mapping of (a)(d) 0.50C-0.16 N, (b)(e) 0.35C-0.37 N and (c)(f) 0.20C-0.54 N steels. The area fractions of precipitates are about 4.84%, 1.41% and 2.56%, respectively.

EDS mapping was employed to analyze the elemental distribution in the steels (Fig. 1(d)(e)(f)). It shows that the precipitates in all the steels are rich in Cr. The enrichment of C and N are obvious in 0.50C-0.16 N and 0.20C-0.54 N steels, respectively, indicating the precipitates might be mainly carbides and nitrides, respectively. It is worth noting that Cr, C and N distribute more uniformly in 0.35C-0.37 N steel, which could reduce the detrimental effects of precipitates on corrosion resistance [22,30]. Thus, the partial replacement of C by N first alleviates and then increases elemental segregation, which would significantly impact the IGC resistance.

Fig. 2(a) shows the XRD analysis of electrolytically extracted precipitates to identify the types of precipitation. The dominant precipitates in 0.50C-0.16 N and 0.20C-0.54 N steels are M₂₃C₆ (Cr_15.58Fe_7.42C₆) and M₂N (Cr₂N), respectively, while both of them are present in the 0.35C-0.37 N steel. It is noteworthy that the absence of M₂₃C₆ or M₂N peaks does not mean the absence of these precipitates in the specimens, since there exists an detection limit (5%) by the XRD analysis [31]. Furthermore, TEM analysis was conducted to confirm the types of precipitation, and the TEM images and selected area electron diffraction (SAD) patterns of the precipitates are shown in Fig. 2(b)-(d). The granular and rod-like M₂₃C₆ are observed in 0.50C-0.16 N steel, and granular M₂N is observed in 0.20C-0.54 N steel. Therefore, the partial replacement of C by N in MSSs converts the dominant precipitates from M₂₃C₆ to M₂N, which is consistent with the variation in types of precipitation in high nitrogen MSSs with different N contents [6] and high nitrogen austenitic stainless steels with different C and N contents [32]. Kim et al. suggested that the driving force for M₂₃C₆ was much higher than that for M₂N in Fe-15Cr-15Mn-4Ni austenitic steels [32], which could explain the higher content of M₂₃C₆ in 0.50C-0.16 N steel than that of M₂N in 0.20C-0.54 N steel in the present study.

View original graphic| Download| PPT slide

Fig. 2.   (a) XRD patterns, and TEM images and SAD patterns of (b)(c) 0.50C-0.16 N and (d) 0.20C-0.54 N steels.

The variation of precipitates was calculated using Thermo-Calc software, as shown in Fig. 3. It shows that for 0.50C-0.16 N steel, M₂₃C₆ and M₂N would completely dissolve into the matrix at the austenitizing temperature of 1040 ℃, while the non-equilibrium state in the heat treatment induces the existence of M₂₃C₆. Besides, the absence of M₇C₃ in the XRD pattern might be attributed to the content lower than the detection limit. The partial replacement of C by N eliminates the precipitation of M₇C₃, and reduces the content and dissolution temperature of M₂₃C₆, while increases those of M₂N. The variation in contents and types of precipitates is consistent with the EDS mapping and XRD results (Fig. 1, Fig. 2(a)). This might be attributed to the inhibiting effect of N on the carbide precipitation [13,33]. According to the previous researches [22,33], the precipitates emerged in the annealing process, and nitrides are more coherent with matrix than carbides, which contribute to the smaller size, more homogeneous distribution and slower growth speed of nitrides than those of carbides [3,22,33]. Gavriljuk et al. [9] investigated the atomic distribution in Fe-15Cr-1Mo MSSs alloyed with 0.6C, 0.62 N or 0.29C+0.35 N using Mössbauer spectroscopy. They found that the N-induced short range atomic ordering contributed to the homogeneity of Cr atoms distribution sorted as 0.29C+0.35 N, 0.62 N and 0.6C, which could further delay precipitation and decrease the size of Cr-rich precipitates. However, further replacing C by N substantially enhances the stability of M₂N, which induces the massive precipitation of M₂N in 0.20C-0.54 N steel (Fig. 1(c)).

View original graphic| Download| PPT slide

Fig. 3.   Variation of precipitates with temperature in high nitrogen MSSs calculated by Thermo-Calc.

3.2. Intergranular corrosion behavior

Fig. 4(a) shows the weight loss of the investigated high nitrogen MSSs after the nitric acid tests. It indicates that the partial replacement of C by N first decreases and then increases the weight loss, and the 0.20C-0.54 N steel exhibits higher IGC resistance than 0.50C-0.16 N steel. Fig. 5(a)(b)(c) shows the surface SEM morphologies of the specimens after nitric acid tests. In 0.50C-0.16 N steel, the severe grain drooping occurs, while the partial replacement of C by N alleviates the detachment of grains, and just the corroded grain boundaries are observed in 0.35C-0.37 N steel. However, the further replacing C by N aggravates the grain drooping again in 0.20C-0.54 N steel. Furthermore, the optical micrographs of the cross section in Fig. 5(d) confirm the severe detachment of grains in 0.50C-0.16 N steel. The 0.20C-0.54 N steel also exhibits obvious preferential attack along grain boundaries as well as grain drooping (Fig. 5(f)), and the IGC attack is reduced in 0.35C-0.37 N steel (Fig. 5(e)). The average IGC depths in Fig. 4(b) reveal that the percolation depth follows the similar trend as that of the weight loss. It is noteworthy that the severe grain dropping makes it difficult to find out the position of the original surface in 0.50C-0.16 N steel, thus the actual percolation depth might be higher than the measured value.

View original graphic| Download| PPT slide

Fig. 4.   (a) Weight loss and (b) percolation depths of the investigated high nitrogen MSSs after nitric acid tests.

View original graphic| Download| PPT slide

Fig. 5.   Surface morphologies by SEM and cross section micrographs by OM of (a)(d) 0.50C-0.16 N, (b)(e) 0.35C-0.37 N and (c)(f) 0.20C-0.54 N steels after nitric acid tests.

The DL-EPR curves of the high nitrogen MSSs with different C and N contents are shown in Fig. 6(a)-(c). According to the previous literatures [[34], [35], [36]], the ratio of the maximum current density or the whole charge generated during the reverse scan (i_r or Q_r) and those during the forward scan (i_a or Q_a) can be defined as degree of sensitization (DOS). In the forward scan, the surface of the specimen suffers first active dissolution and then passivation. Subsequently, in the reverse scan, the passive film on the non-sensitized steels would be mostly intact, thus the value of i_r might be stabilized or slightly decreased. In contrast, on the sensitized steels, the active dissolution of passive film on Cr-depleted zones would increase the value of i_r, which results in a high DOS value [37]. Therefore, the lower DOS value indicates the higher IGC resistance of the steel, and the values of i_a, i_r and DOS are listed in Table 2 based on the DL-EPR tests.

View original graphic| Download| PPT slide

Fig. 6.   DL-EPR results and corresponding surface SEM micrographs of (a)(d) 0.50C-0.16 N, (b)(e) 0.35C-0.37 N and (c)(f) 0.20C-0.54 N steels. The arrows indicate the scanning direction.

As can be seen, the 0.50C-0.16 N steel has the highest DOS value, indicating the poorest IGC resistance. Meanwhile, the surface observation reveals IGC attack along prior austenite grain boundaries, precipitates and martensitic lath interfaces (Fig. 6(d)). It is worth noting that the DOS value for 0.50C-0.16 N steel exceeds 100%, as reported by Alonso-Falleiros et al. on Fe-12.5Cr-0.13C MSS tempered at 550 ℃ (DOS value about 300%), which was ascribed to an increase in effective surface area by the increased asperity of etched surface [38]. The partial replacement of C by N significantly reduces the value of DOS for 0.35C-0.37 N steel, which means that the IGC susceptibility is drastically reduced. Meanwhile, the surface just exhibits etching along prior austenite grain boundaries and precipitates, and the scratches could be clearly observed (Fig. 6(e)). The further replacing C by N leads to a rise in DOS value in 0.20C-0.54 N steel, indicating the reduction of IGC resistance. Correspondingly, the etching along prior austenite grain boundaries and precipitates is also observed (Fig. 6(f)), but the extent is much severer than those of 0.35C-0.37 N steel.

It is widely accepted that there exists Cr-depleted zones in the vicinity of Cr-rich precipitates (Cr₂₃C₆, Cr₂N, σ, etc.), which act as the anodic zones during corrosion and are preferential initiation sites for IGC [[39], [40], [41], [42], [43]]. The change of IGC resistance of high nitrogen MSSs with the partial replacement of C by N is mainly attributed to the variation in content and type of precipitation. On the one hand, the partial replacement of C by N first reduces and then increases the content of precipitates. The massive precipitation in 0.50C-0.16 N steel induces the high area fraction of Cr-depleted zones, which further results in the severe IGC sensitivity. The partial replacing C by N decreases the precipitation content, and hence alleviates IGC sensitivity. However, the further replacement increases precipitation content, which would provide more initiation sites for IGC [17], thus aggravating IGC sensitivity. The variation trend of precipitation content is consistent with that of the IGC sensitivity, which is the main reason for the change in IGC resistance. On the other hand, the partial replacement of C by N gradually converts the dominant precipitates from M₂₃C₆ to M₂N. Our previous studies and the work by Berns et al. reported that M₂N had smaller size and induced slighter Cr-depletion than those of M₂₃C₆ [22,33,44], and the metastable pitting corrosion preferentially initiated around M₂₃C₆ [22,44]. Similarly, the Cr-depleted zones around M₂₃C₆ would suffer severer attack in IGC tests. Therefore, the massive precipitation of M₂₃C₆ in the grain interior and along grain boundaries of 0.50C-0.16 N steel induced the severest IGC sensitivity. The combined effect of content and type of precipitation induces the initial increase and the following deterioration of IGC resistance.

The effect of partial replacement of C by N on microstructure, IGC resistance and mechanical properties is schematically illustrated in Fig. 7. Specifically, Fig. 7(a) shows the influence of C/N content on carbide/nitride content and Cr-depleted zones. The decrease of C content reduces the content of Cr-rich M₂₃C₆ and the area of Cr-depleted zones in Fe-15Cr-1Mo MSS, thus alleviating the Cr-depletion and reducing the IGC sensibility. The increase of N content increases the content of Cr-rich M₂N and the area of Cr-depleted zones, thereby enhancing the IGC sensibility. Taking into account the influence of both C and N on precipitation and Cr-depleted zones, the variation of M₂₃C₆ and M₂N contents and area of Cr-depleted zones is shown in Fig. 7(b). Due to the lower content, smaller size, more homogeneous distribution and slighter Cr-depletion of M₂N than those of M₂₃C₆, the partial substitution of C by N first alleviates and then aggravates Cr-depletion, which contributes to the least percolation depth and IGC susceptibility of 0.35C-0.37 N steel (Fig. 7(c)). Besides, both C and N are prone to higher strength through solid solution and precipitation strengthening [12]. The partial substitution of C by N slightly influences the strength of the steels, while substantially enhances the toughness of 0.35C-0.37 N steel, which would be discussed in detail in our further study. Therefore, the high nitrogen MSS containing medium C and N contents possesses good combination of mechanical properties and IGC resistance.

View original graphic| Download| PPT slide

Fig. 7.   Schematic of partial replacement of C by N on microstructure, IGC resistance and mechanical properties of high nitrogen MSSs: (a) influence of C/N content on carbide/nitride content and Cr-depleted zone, (b) variation of M₂₃C₆ and M₂N contents and area of Cr-depleted zones, (c) tensile strength and percolation depth.

4. Conclusions

The effect of partial replacement of C by N on microstructure and IGC resistance of high nitrogen MSSs was investigated, and the main conclusions are obtained as follows:

(1) There exists massive large size Cr-rich M₂₃C₆ in 0.50C-0.16 N steel along prior austenite grain boundaries and in grain interior. The partial replacement of C by N first significantly reduces and then increases the size and content of precipitates in 0.35C-0.37 N and 0.20C-0.54 N steel, respectively, and converts the dominant precipitates from M₂₃C₆ to M₂N.

(2) The nitric acid and DL-EPR tests reveal that the partial replacement of C by N first enhances and then deteriorates the IGC resistance, which could be ascribe to the variation in content and type of precipitation. The high nitrogen MSS containing medium C and N contents possesses good combination of mechanical properties and IGC resistance.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Acknowledgements

This work was supported by the National Natural Science Foundation of China [grant numbers 51434004, U1435205, 51774074], Fundamental Research Funds for the Central Universities [N172512033, N172507002] and Transformation Project of Major Scientific and Technological Achievements in Shenyang [grant number Z17-5-003].