Microstructures of SUS304 Stainless Steel after Long-Term Service in Vacuum Carburizing Quenching

Abstract

The microstructural changes and crack propagation in the carburized layer of SUS304 stainless steel due to repeated vacuum carburizing quenching as a heat treatment basket were investigated in this paper. As a result, beneath a graphite scale formed on the outermost surface, it was revealed that there were three types of carburized layers with different morphologies: M₇C₃ layer, M₇C₃/M₂₃C₆ mixed layer, and M₂₃C₆ layer (M = Cr, Mn, Fe). In addition, the surface was found to be uneven due to the occurrence of metal dusting. Besides, repeated heating and quenching caused the formation of voids in carbides and matrix in the carburized layer, and micro cracks appeared in the surrounding areas. Under the loading stress during vacuum carburizing quenching, these voids, micro cracks, and the uneven surface are considered to be the initial points for cracking.

 

This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 87 (2023) 211–218.

Fig. 3 EPMA quantitative mapping and line analysis result of carbon and chromium concentration in vacuum-carburized SUS304 stainless steel. (online color)

1. Introduction

In the heat treatment industry, both heat-resistant austenitic cast steels and stainless steels are used as materials for jigs that load heat-treated parts. Among these, SUS304, SUS309 and SUS310 steels exhibit excellent results. For example, baskets for carburizing and quenching are usually made by welding stainless steel round bars to form a frame followed by welding a stainless steel wire mesh inside the frame. On the other hand, the use of vacuum carburizing has been expanding in place of conventional gas carburizing because of the demand for improvement of mechanical properties such as wear resistance and fatigue properties of heat-treated parts, as well as the need to reduce energy consumption and running costs [1, 2]. In the vacuum carburizing process, the carburizing temperature is set relatively high, so damage to jigs (baskets, trays) such as external deformation and fracture due to carburization occurs early, resulting in a shortened life span of the jigs, i.e., a reduction in the number of uses. Many studies have been actively conducted on the vacuum carburizing mechanism of carbon steel and low alloy steel to clarify the difference in mechanical properties between gas carburizing and vacuum carburizing [3–5] but there are few reports on the material degradation of jigs used for long-term operation in a high-temperature carburizing vacuum. Although the relationship between the microstructure and creep properties of heat-resistant cast steel subjected to short thermal cycles of repeated heating and cooling has been clarified in our previous study [6], the factors that cause damage to heat-resistant cast steel and stainless steel jigs under actual operating environments have not yet been fully understood.

Therefore, in this study, microstructural analysis of carburized stainless steel was conducted using a SUS304 heat-treated basket that had been operated at high temperatures for a long-term in an industrial vacuum carburizing furnace. Based on the results obtained, the carburizing mechanism of SUS304 steel was elucidated, and the factors of crack propagation that damaged the basket were examined.

2. Experimental Methods

Figure 1 shows a photograph of a basket made of commercial SUS304 stainless steel round bars with a diameter of 22 mm and a wire mesh welded to the inside after two years of operation. This basket was used for vacuum carburizing and quenching of SCM steel and SNCM steel parts and high-temperature tempering of SKH steel and SKD steel parts. With regard to the carburizing conditions, since the target effective hardening depth as well as the materials of heat-treated parts were various, not only was the carburizing temperature of a heating chamber, which was reduced to 100 Pa, varied in the range from 1173 K to 1323 K, but also the carburizing time and diffusion time were controlled at the appropriate temperature. After carburizing, the basket with the heat-treated parts loaded was moved to a gas cooling chamber and subjected to a soaking treatment at 1123 K. The heat-treated parts and the basket were then simultaneously quenched in an oil bath at 443 K. High-temperature tempering was performed at 803 K in the air. In terms of heat treatment cycles of the basket, the basket was subjected to approximately 120 cycles of carburizing, quenching and tempering.

Fig. 1

Appearance of SUS304 heat treatment basket after being in service. (online color)

Visual inspection of the entire basket revealed that the surface layer of all SUS304 round bars (hereafter referred to as “experimental steel”) was coated with black scale. The thick black scale formed on the surface of the experimental steel was easily detached, resulting in localized surface irregularities. Additionally, as indicated by the arrows, the cracks on the surface were generated along the longitudinal direction. Here, carbon was analyzed in the non-carburized center region using the combustion method, and other elements were analyzed using optical emission spectroscopy. Table 1 shows the chemical compositions of SUS304 before carburizing.

Table 1 Chemical compositions of SUS304 heat treatment basket prior to carburizing (mass%).

Hardness was measured using a micro Vickers hardness tester (Shimadzu, HMV-G21FA) from the surface to the inside with a load of 2.94 N (0.3 kgf). Each measurement value of the hardness profile was the average of five measurements.

For the microstructural analysis, the cut surfaces of the samples were etched with Villela’s reagent (HCl: 5 ml, picric acid: 1 g, ethanol: 100 ml) for a few seconds, and the cross-sectional microstructures were observed using an optical microscope and a scanning electron microscope equipped with EDS (SEM-EDS, JEOL, IT-100LA). The backscattered electron (BE) mode was used, which allows clear observation of any carbides present. The area fraction occupied by each carbide formed by the carburization was calculated using image analysis software based on the BE images. An electron probe microanalyzer (EPMA, JEOL, JXA8230) was used to measure the concentration distribution of alloying elements in the surface layer. EPMA analysis was carried out in an area 6 mm deep and 0.6 mm wide from the surface layer.

Furthermore, the experimental steel was cut vertically at depths of 0.05, 1, 2, and 3 mm from the surface, and the cross sections at these positions were subjected to X-ray diffraction (XRD, Malvern Panalytical, Empyrean, target: Cu) to identify carbides. For reference, the non-carburized region in the center of the SUS304 round bar was also investigated by XRD. Changes in carbides at these positions were also examined using SEM-EDS.

On the other hand, the thermal expansion coefficient, which is used to evaluate a physical property, was measured using a thermomechanical analyzer (TMA, Rigaku, Thermo plus EVO2) from room temperature to 1323 K on test pieces with a diameter of 2.5 mm and a height of 15 mm cut by wire cutting from two locations: at a distance of 0.05 mm from the surface and the center of the experimental steel. The measurement conditions were an argon gas atmosphere, a heating rate of 0.167 K/sec, and a load of 0.098 N.

3. Results and Discussion

3.1 Carburizing behavior of SUS304 steel

A low-magnification photograph of the cross-section of the experimental steel etched with Villela’s reagent is displayed in Fig. 2(a). The carburized surface layer and the center, which appear to be the non-carburized region, can be distinguished by their color tone. The carburized surface layer discolored much faster than the center even with the same etching time. As a result, a ring-shaped carburized layer with a thickness of about 2.8 mm appears on the surface of the experimental steel. Moreover, there was a color variation in the ring-shaped carburized layer. From its color tone, it was seen that it consisted of three layers. On the other hand, as indicated by the arrows in the figure, all cracks propagated from the surface. When the crack tip opening displacement exceeded 0.15 mm, the depth of the carburized layer increased. The hardness distribution in the cross-section was measured for the crack-free area. From the hardness profile shown in Fig. 2(b), it can be seen that the maximum hardness just below the surface is 750 HV and the hardness gradually decreases from the surface. In the area more than 4 mm away from the surface, i.e., the non-carburized layer, the hardness is almost constant, about 200 HV.

Fig. 2

Cross-sectional image of vacuum-carburized SUS304 stainless steel (a); Hardness distribution on a round bar cross-section (b). (online color)

Figure 3 shows the concentration distribution of carbon and chromium from the carburized surface layer to the center of the experimental steel, measured by EPMA. The upper part and lower part of the figure show the mapping of C Kα and Cr Kα in the analyzed area, respectively. The corresponding analysis for each mapping is also shown in the same figure. As can be seen from the color changes shown in the mapping of C Kα, the cross-section of the experimental steel appears to be divided into three regions. Detailed observation shows that the cross-section consists of a region from the surface of the carburized layer to about 3 mm, a region above 4 mm toward the center of the non-carburized part, and the intermediate layer where the carburized layer and the non-carburized part are mixed. Furthermore, the carbon distribution curve shows that the carbon concentration at the surface was a maximum of 5.0 mass%, decreasing toward the center. However, at a distance of 2 mm from the surface, the decrease in carbon concentration is more gradual. The average carbon concentration in the matrix-only area is 0.043 mass%, which is almost the same as the carbon content of the experimental steel as shown in Table 1. On the other hand, it was seen that the Cr concentration decreased in the surface layer and increased toward the center, showing the opposite trend to the change in C concentration. A close-up of the change in Cr concentration from the non-carburized area to the surface shows that the measured value for the non-carburized area is 19.0 mass%, while the value just below the surface is about 17.8 mass%. In between, particularly in the carburized layer, the Cr concentration decreases in a zigzag pattern within the range of 18 to 19 mass%. From these results, it is assumed that not only does an increase in C concentration result in a relative decrease in Cr concentration, but also that a deficiency of Cr concentration occurs due to the decomposition and exfoliation of carbides caused by the metal dusting reaction described below.

Fig. 3

EPMA quantitative mapping and line analysis result of carbon and chromium concentration in vacuum-carburized SUS304 stainless steel. (online color)

To identify the carbides formed by the vacuum carburizing and quenching, cross sections at depths of 0.05, 1, 2, and 3 mm from the surface and the center of the non-carburized region were examined by XRD in the range of 35 to 95°. The obtained XRD patterns are displayed in Fig. 4. For reference, the black scale on the surface of the experimental steel was also examined by XRD at a low angle range. As a result, it was seen that the austenite peaks were dominant in the non-carburized region. The closer to the surface, the more prominent carbide peaks appeared. At a depth of 3 mm near the intermediate layer, the carbides (marked with △) formed were identified as M₂₃C₆. Meanwhile, at a depth of 2 mm, M₇C₃ (marked with ▲) and M₂₃C₆ coexisted. Only M₇C₃-type carbides were present at depths of 1 mm and 0.05 mm from the surface. At any depth, only carbides were detected, and the presence of the σ phase that causes high-temperature embrittlement was not observed. In the case of the black scale on the outermost surface of the experimental steel, a low-angle XRD result revealed the presence of graphite in the scale.

Fig. 4

XRD patterns of outermost surface, cross-sections of carburized layer at 0.05, 1, 2, and 3 mm away from surface and non-carburized region.

Figure 5 shows backscattered electron images of the cross-sectional microstructure of the carburized layer at depths of 0.05, 1, 2, and 3 mm from the surface. For comparison, the microstructure of the non-carburized center part is also shown in the same figure. Figure 6 shows the compositional maps by EDS corresponding to these depths. Based on the contrast of BE images, it was observed that the color of carbides formed by the carburization reaction is dark gray from a depth of 0.05 mm to 2 mm, and the color of carbides changes to a light gray from a depth of 2 mm. At a depth of 3 mm, the grain boundaries as well as the matrix consist only of light gray carbides. From these results and the XRD patterns, it is considered that the dark gray carbides are M₇C₃, whereas the light gray carbides are M₂₃C₆. The area fraction of each carbide was determined by the image analysis method to evaluate the quantitative changes in these carbides. The obtained result is shown in Fig. 7. The average area fraction of M₇C₃ was about 80% at a depth of 0.05 mm. It decreases linearly from the surface to the center. At a depth of 2 mm, where the two types of carbides coexist, the gradient change in the area fraction of M₂₃C₆ was found to be different from that of M₇C₃. This result is in good agreement with the gradient change in carbon concentration shown in Fig. 3. Furthermore, from the EDS compositional maps, it was revealed that at a depth of 1 mm, massive carbides precipitated at the grain boundaries, whereas they precipitated as fine particles within the matrix. In addition, a carbide-free zone (white layer) was observed around the massive carbide. At a depth of 2 mm, carbides that formed at both grain boundaries and in the matrix became finer. In this region, the carbides coexisting in the matrix have two shapes: needle-like M₇C₃ and slightly larger cuboid-like M₂₃C₆. In the center, i.e., the non-carburized region, there was almost no carbide, and the microstructure was similar to that of SUS304 after solution heat treatment.

Fig. 5

BE images showing non-carburized region and cross-sections of carburized layer at 0.05, 1, 2, and 3 mm away from surface.

Fig. 6

EDS compositional maps showing non-carburized region and cross-sections of carburized layer at 0.05, 1, 2, and 3 mm away from surface. (online color)

Fig. 7

Area fraction of M₂₃C₆ and M₇C₃ formed due to carburizing.

On the other hand, from Fig. 6, apart from the carbides formed by carburization, many voids were observed at a depth of 0.05 mm. Therefore, high-magnification SEM observation was performed near the surface of the area where no cracks propagate in the cross-section. The results are shown in Fig. 8. According to the secondary electron image and backscattered electron one, there was the presence of voids just below the surface and cracks in the M₇C₃ carbide and austenitic matrix. These cracks were not associated with the fracture zone but were prominent along the outer diameter of the experimental steel. Traces of metal particles were also detected in the graphite scale. Figure 9 shows the EDS compositional maps of the metal particles (area 1) and matrix (area 2) observed at high magnification using SEM. In these metal particles, major elements such as Mn, Cr, and Ni were oxidized. The carbides consisted mainly of Cr and Mn, whereas Mn, Cr, Ni, and enriched Si were detected in the matrix. Table 2 summarizes the EDS quantitative analysis results for the metal particle, carbide, and matrix. It was seen that the metal particles contain trace amounts of Ni, unlike the matrix. The appearance of such metal particles indicates the occurrence of the “metal dusting” phenomenon in the experimental steel. This phenomenon is the transformation of metals and alloys into metal particles and graphite dust in a strong carburizing atmosphere where the activity of carbon, a_C, is >1. Regarding the gas carburizing atmosphere, based on the composition of the carburized material, three mechanisms of metal dusting have been proposed. Type 1: the decomposition of metastable carbides (M₃C) in iron or low-alloy steel [7], Type 2: the formation of graphite in thermodynamically stable carbides (M₂₃C₆ and M₇C₃) in austenitic Fe-Ni alloys or Ni [8], and Type 3: due to the activity of oxygen and carbon in the gas carburizing atmosphere [9]. In the case of gas-carburized SUS304, metal dusting was reported to occur as a Type 2 or Type 3 mechanism [9]. In the present study, the mechanism of metal dusting is currently unknown and needs to be investigated further.

Fig. 8

SE and BE images showing cross-sectional microstructure near-surface.

Fig. 9

EDS compositional maps showing metal particle in graphite scale (Region 1), carbide and matrix in carburized layer near surface (Region 2). (online color)

Table 2 EDS analysis results of metal particle in graphite scale, carbide and matrix near surface (mass%).

Based on the equilibrium phase diagram shown in Fig. 10, the microstructural changes of SUS304 with the progress of carburization are discussed here. As described in the experimental method, various steel parts loaded into the SUS304 basket were carburized at temperatures ranging from 1173 K to 1323 K. This carburizing temperature range is indicated by the dashed line in the figure. The two white circle marks indicated the non-carburized state (as received state) and the early stage of carburization for the experimental steel at 1173 K and 1323 K. Since the experimental steel contains a trace amount of carbon, the thermodynamically stable carbide is M₂₃C₆ at the early stage of carburization at both temperatures. The SEM image of the non-carburized region (the center part) shown in Fig. 5 corresponds well to the equilibrium phase diagram. As carburization progresses, M₂₃C₆ transitions to the more stable M₇C₃ carbide. When the carbon concentration is above 5.1% at 1173 K and above 5.7% at 1323 K, graphite (Gr) becomes the stable phase and coexists with austenite (γ) and M₇C₃. As for the carbide transition during carburization, the phase identification results using XRD and SEM observation are in good agreement with the equilibrium phase diagram. However, the measured carbon concentration just below the surface was 5.0 mass%, lower than the reading from the equilibrium phase diagram (gray circle mark). Therefore, the difference in carbon concentration may be due to the fact that the microstructure just below the surface has not fully reached thermodynamic equilibrium.

Fig. 10

Equilibrium phase diagram of SUS304 stainless steel calculated by Thermo-calc software. (Database: Fe-DATA Ver. 4)

From the above results, it is clear that the carbides formed in SUS304 during carburizing in the temperature range from 1173 K to 1323 K change in the order of M₂₃C₆ $ \Rightarrow $ M₇C₃ $ \Rightarrow $ Graphite.

3.2 Crack propagation in SUS304 during carburizing and quenching

Figure 11 shows a representative example of the crack propagation in the carburized layer observed by SEM. It can be seen that the propagation path of the crack is jagged. Cracks propagated to the vicinity of the intermediate layer adjacent to the carburized zone, but no crack was recognized in the non-carburized region. The enlarged views of one area of crack propagation and the crack tip are shown in the lower part of the same figure. Microstructural observation at high magnification revealed that the carbides formed and coarsened at grain boundaries tended to crack more easily than those in the matrix. Since the carbides were densely distributed near the surface, it was difficult to identify the crack propagation path. However, observation of the crack tip indicates that the crack propagated along the carbides that were fractured at the grain boundaries.

Fig. 11

BE images showing the region of crack propagation in a cross-section of vacuum-carburized SUS304 stainless steel.

Generally, in an environment where heating and cooling are continuously repeated, not only does deformation due to thermal expansion and contraction occur but also thermal stress associated with temperature gradients is induced, so the thermal expansion coefficient and thermal conductivity are important physical properties [10]. In this study, TMA was used to determine whether there is a difference in the thermal expansion coefficient between the surface layer (carburized layer), where M₂₃C₆ and M₇C₃-type carbides were formed and grown by carburizing, and the center (non-carburized region), where there was almost no carbide. Considering the thickness of the carburized layer, TMA tests were conducted on cylindrical test pieces with a diameter of 2.5 mm and a height of 15 mm cut from the carburized layer just below the surface and from the center. The measurement results are shown in Fig. 12. The thermal expansion coefficient of the surface layer and the non-carburized region increases with increasing temperature, up to 1100 K. Above a temperature of 1323 K, which corresponds to the maximum carburizing temperature during carburizing and quenching, the thermal expansion coefficient in the non-carburized region was almost the same, while it tends to decrease slightly in the surface layer where carbides exist. At 1323 K, the measured thermal expansion coefficients of the carburized layer and non-carburized region were 9.3 × 10⁻⁶ and 19.0 × 10⁻⁶ (1/K), respectively. Since there is a large difference in the thermal expansion coefficient between the two regions, it is presumed that as the carburization progresses, the formation, the growth and the transition of carbides result in a marked difference in expansion and contraction relative to the austenite matrix.

Fig. 12

Thermal expansion coefficient of carburized layer and non-carburized region in vacuum-carburized SUS304 stainless steel.

In our previous papers, cracking in austenitic heat-resistant cast steel after several hundred cycles of repeated heating and quenching under no load has been investigated [11, 12]. As a result, it was clear that micro voids and fine cracks generated at the boundary between the primary carbides and the matrix, and inside the carbide by thermal shock were the initial points of cracking. The results obtained in the present study were confirmed to be similar. Observation of the entire surface of the experimental steel revealed that the volume of defects such as cracks in the carbides, cracks in the austenite matrix and voids in the carburized layer, especially just below the surface, was higher than in the non-carburized region. From these results, it is assumed that the high applied stress during carburizing and quenching causes thermal stress concentration to be more likely to occur in areas where voids or fine cracks exist. In addition, it was found that the formation and exfoliation of carbide-free metal particles proceeded from the outermost surface under a carburizing atmosphere, and this was considered to be a metal dusting phenomenon. Therefore, the carburized SUS304 steel surface no longer has the same surface smoothness as when the steel was manufactured. The surface has become uneven, and stress concentration is more likely to occur.

4. Conclusion

Microstructural changes in the carburized layer and crack propagation in a SUS304 heat-treated basket used for a long period in a vacuum carburizing and quenching environment were investigated. The results are as follows:

1. (1)    During carburizing, as the carbon concentration increases in SUS304 steel, the carbide reaction sequence is: M₂₃C₆ $ \Rightarrow $ M₇C₃ $ \Rightarrow $ Graphite
2. (2)    The thermal expansion coefficient of the surface layer is lower than that of the non-carburized region due to the formation of carbides by carburizing. In addition, repeated heating and cooling causes cracking of the carbides in the surface layer and the formation of voids at the boundary between the carbide and the austenite matrix due to thermal shock.
3. (3)    The outermost surface becomes uneven due to the formation and exfoliation of carbide-free metal particles.
4. (4)    Under the load during the vacuum carburizing and quenching, the areas with voids and micro-cracks, as well as an uneven surface, are prone to thermal stress concentration, and these areas are considered to be the initial points for cracking.

Acknowledgment

The authors would like to thank Emeritus Professor Masaru Nakayama of the Department of Materials Science and Engineering at National Institute of Technology, Kurume College, for suggesting useful discussions in this study.

REFERENCES

• 1)   N.  Okumura and  A.  Iwase: Vacuum Carburizing Using Acetylene Gas, Denso Technical Review 5(1) (2000) 87–93.
• 2)   T.  Tanaka: Advantages and Future Vision of Vacuum Carburizing Furnace, Denkiseiko 84 (2013) 123–128.
• 3)   K.  Okamoto,  M.  Tominaga,  H.  Ono and  Y.  Nishiyama: Effects of Microstructure in the Carburized Layer on the Fatigue Strength of Vacuum Carburized Cr-Mo Steel (SCM 420), J. Jpn. Inst. Metals 46 (1982) 827–834. doi:10.2320/jinstmet1952.46.8_827
• 4)   T.  Morita,  Y.  Matsumura and  M.  Umemoto: Effect of Excess Carburizing at Edge on Fatigue Strength of Vacuum-Carburizing Steels, Denkiseikou 79 (2008) 15–23. doi:10.4262/denkiseiko.79.15
• 5)   K.  Inoue: Recent Development Trends of Carburizing Steels, Denkiseikou 90 (2019) 79–86.
• 6)   N.H.K.  Luan,  T.  Okuyama,  M.  Nakayama,  K.  Mizuno and  K.  Koreaki: Influence of Nb Addition on Microstructure and Creep Property of Heat-Resistant Cast Steel in Vacuum Carburizing and Quenching, Mater. Trans. 63 (2022) 1557–1566. doi:10.2320/matertrans.F-M2022810
• 7)   H.J.  Grabke,  E.M.  Muller-Lorenz and  A.  Schneider: Carburization and Metal Dusting on Iron, ISIJ Int. 41(Supplement) (2001) S1–S8. doi:10.2355/isijinternational.41.Suppl_S1
• 8)   H.J.  Grabke: CARBURIZATION, CARBIDE FORMATION, METAL DUSTING, COKING, Mater. Technol. 36 (2002) 297–305.
• 9)   P.  Szakalos,  R.  Pettersson and  S.  Hertzman: An active corrosion mechanism for metal dusting on 304L stainless steel, Corros. Sci. 44 (2002) 2253–2270. doi:10.1016/S0010-938X(02)00053-7
• 10)   T.  Nagashima and  T.  Iikubo: Thermal Fatigue Property of Ferritic Stainless Cast Steels, Deniseiko 64 (1993) 99–106. doi:10.4262/denkiseiko.64.99
• 11)   N.H.K.  Luan,  K.  Koizumi,  K.  Sawada,  K.  Mizuno and  T.  Okuyama: Influence of Pouring Temperature on Microstructure and Elevated Temperature Properties of Heat Resistant Cast Steel, J. JFS 92 (2020) 351–359. doi:10.11279/jfes.92.351
• 12)   N.H.K.  Luan,  K.  Koizumi,  K.  Mizuno,  Y.  Yamada,  T.  Okuyama and  M.  Nakayama: Influence of Pouring Temperature on Microstructure and High Temperature Mechanical Properties of Nb-added Austenitic Heat Resistant Cast Steel, J. JFS 93 (2021) 185–194. doi:10.11279/jfes.93.185