- Email: [email protected]
Effect of sensitization on corrosion fatigue behavior of type 304 stainless steel annealed in nitrogen gas
Materials Science & Engineering A 640 (2015) 33–41
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of sensitization on corrosion fatigue behavior of type 304 stainless steel annealed in nitrogen gas Masayuki Akita a,n, Yoshihiko Uematsu b, Toshifumi Kakiuchi b, Masaki Nakajima c, Toshihiro Tsuchiyama d, Yu Bai e, Kenta Isono f a
Faculty of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan Department of Mechanical Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan c Department of Mechanical Engineering, National Institute of Technology, Toyota College, 2-1 Eisei-cho, Toyota 471-8525, Japan d Department of Materials Science and Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka Nishi-ku, Fukuoka 819-0395, Japan e Graduate Student, Materials Science and Engineering, Graduate School of Engineering, Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto 606-8501, Japan f Aisin Aw Co., Ltd., 10 Takane, Fujii-cho, Aichi, Anjo 444-1192, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 19 March 2015 Received in revised form 22 May 2015 Accepted 23 May 2015 Available online 28 May 2015
Effect of sensitization on fatigue behavior of type 304 stainless steel in laboratory air and in 3% NaCl solution was studied using specimens annealed in high temperature nitrogen gas. Annealing in high temperature nitrogen gas was performed at 1100 °C and at 1200 °C for 30 h, resulting in the solid solution of nitrogen and the precipitation of chromium nitride (CrN). The mechanical and fatigue properties were significantly improved by annealing. The improvement was due to the solid solution of nitrogen and the precipitation of CrN. Fatigue strengths of the untreated specimen in laboratory air and in 3% NaCl solution are nearly the same. However, fatigue properties of the annealed specimen in 3% NaCl solution change for the worse, because chromium (Cr) -depleted zones were formed along grain boundaries during the heat treatment, resulting in the remarkable sensitization. In order to prevent the sensitization, the re-solution treatment (RST) which enhanced the dissolution of CrN and the water quenching treatment (QT) which avoided the precipitation of CrN was performed. As a result, the fatigue properties of the RST and QT specimens in 3% NaCl solution were slightly improved, but were still lower than that of the untreated one. Since the oxalic etch tests proved the formation of Cr-depleted zone also in the RST and QT specimens, the influence of sensitization could not be fully eliminated by the both treatments. & 2015 Elsevier B.V. All rights reserved.
Keywords: Fatigue Austenitic stainless steel Annealed in nitrogen gas Corrosion Sensitization Microstructure
1. Introduction The excellent corrosion resistance of austenitic stainless steel type 304 (18Cr–8Ni) brings on the wide use in chemical, biomedical and mechanical engineering fields [1,2]. However, low strength and poor wear resistance of type 304 steel are the weak points when used as structural and mechanical components [3,4]. Some surface treatment techniques are known to improve the mechanical properties of type 304 steel. Recently, the annealing treatments in high temperature nitrogen gas are applied to stainless steels and Ti alloys to improve the mechanical properties and hardness by the solid solution of nitrogen into matrix [5–9]. This method is a kind of chemical heat treatment and is different from so-called nitriding [6]. Nitrogen is one of the austenite n
Corresponding author. Fax: þ 81 58 293 2378. E-mail address: [email protected] (M. Akita).
http://dx.doi.org/10.1016/j.msea.2015.05.080 0921-5093/& 2015 Elsevier B.V. All rights reserved.
(gamma o γ 4) stabilizing elements for austenitic stainless steels and can improve mechanical properties, fatigue strength [10,11], and corrosion resistance [11–14]. In the previous study [15], authors had reported that the fatigue and the mechanical properties of type 304 steel were highly improved by the annealing in high temperature nitrogen gas. It is well known that the γ-phase in type 304 steel is metastable as expected from its chemical composition, 18Cr–8Ni, taking place in the strain-induced martensitic transformation during fatigue loading [10,11]. However, the strain induced martensitic transformation during fatigue test was successfully suppressed by the annealing in high temperature nitrogen gas, that is, γ-phase was stabilized by an addition of nitrogen [15]. In general, the martensitic phase (α′-phase) has higher environmental sensitivity than γ-phase, thus the strain-induced martensitic transformation decreases the corrosion resistance of type 304 steel [16,17]. Since the γ-phase in type 304 steel was stabilized by the annealing in high temperature nitrogen gas [15],
34
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
it is expected that the corrosion fatigue properties could be also improved by this treatment. In the present study, the annealing in high temperature nitrogen gas was applied to type 304 stainless steel. Since stainless steels are generally required to be used in aggressive environments, the fatigue test should be performed in corrosive environment. But corrosion fatigue properties of type 304 steel annealed in high temperature nitrogen gas have not been fully studied yet. Therefore, in the present study, the fatigue tests in 3% NaCl solution were performed using type 304 stainless steel annealed in high temperature nitrogen gas, and the corrosion fatigue properties were discussed from the viewpoint of sensitization.
spectroscopy (EDX). The distribution of nitrogen concentration was measured using an electron probe microanalyser (EPMA). A measurement of hardness was carried out using a Vickers microhardness tester (test load 4.9 N, loading time 30 s). The residual stress was measured by an X-ray diffraction (XRD). A Cr-Kα X-ray radiation was used at 50 kV and 40 mA. The sin2ψ method was applied for the residual stress measurement [15].
2. Experimental procedures
3.1.1. . S–N diagrams in 3% NaCl solution Fig. 3 represents an S–N diagram in laboratory air and in 3% NaCl solution. The annealing treatments in nitrogen gas improved the fatigue properties in laboratory air (open symbols) in comparison with the untreated specimen, especially the annealing at 1100 °C highly improved the fatigue properties. The fatigue limit was 250 MPa, 380 MPa and 290 MPa for the untreated, 1100 °C and 1200 °C annealed specimens, respectively under a stress ratio of 1. In the previous study, it was indicated that the improved fatigue properties in the annealed specimens were due to the solid solution of nitrogen and the precipitation of chromium nitride (CrN) [15]. In 3% NaCl solution (solid symbols), the fatigue strength at 107 cycles was 240 MPa, 200 MPa and 160 MPa for the untreated, 1100 °C and 1200 °C annealed specimens, respectively. Based on this figure, the fatigue property of the untreated specimen in corrosive environment is nearly the same as that in laboratory air. On the other hand, the corrosion fatigue properties of the annealed specimens were significantly lower than those in laboratory air.
2.1. Materials and annealing treatment in nitrogen gas The material used is type 304 stainless steel (18Cr–8Ni), whose chemical compositions (wt%) are as follows; C: 0.05, Si: 0.29, Mn: 1.69, P: 0.038, S: 0.029, Ni: 8.19, Cr: 18.65, Fe: bal. [15]. Material was solution-treated at 1080 °C for 1 h followed by water quenching. After the solution treatment, material was machined to smooth round-bar fatigue specimens with a diameter of 5 mm and a gauge length of 5.04 mm as shown in Fig. 1. Prior to fatigue tests, the gauge section was mechanically polished using progressively finer grades of emery paper followed by buff-finishing. A vacuum furnace was filled with nitrogen gas at 0.1 MPa. The specimens were stored in high temperature nitrogen gas at 1100 °C or 1200 °C for 30 h [15]. Hereafter, the specimens heat treated at 1100 °C and 1200 °C are designated as the “1100 °C annealed specimen” and the “1200 °C annealed specimen”, respectively. In addition, the specimen without annealing treatment in nitrogen gas is referred to as “untreated specimen”. Fig. 2 shows specimen preparation and heat treatment procedures. The details of each treatment, RST (re-solution treatment) and QT (water quench treatment) in Fig. 2 will be mentioned later (see in Section 3.2.1). 2.2. Experimental procedures Fatigue tests were performed using an Ono-type four-point rotary bending fatigue testing machine with a capacity of 98 Nm operating at a stress ratio of 1 and at a frequency of 60 Hz. Testing conditions were in laboratory air and in 3% NaCl solution. The solution was trickled continuously on the specimen surface through the metering pump from a reserved tank. The microstructures and fracture surfaces of the specimens were examined in detail using a scanning electron microscope (SEM). The precipitates were analyzed by an energy dispersive X-ray
90 30
5
30
12
φ φ
5.04
4
R2 Fig. 1. Configuration of specimen for fatigue tests.
3. Results 3.1. Deterioration of corrosion fatigue properties in annealed specimens
3.1.2. Anodic polarization curves and surface morphologies The corrosion fatigue properties of the 1100 °C and 1200 °C annealed specimens significantly deteriorated as compared with the untreated specimen (Fig. 3). In order to evaluate the corrosion resistance of each specimen, anodic polarization behavior was measured by the potentiostatic approach. Conditions of anodic polarization measurement are listed in Table 1. Fig. 4 shows the anodic polarization curves in each material, where the potential was plotted as a function of current density. Anodic polarization curves were obtained using three samples for each specimen. As can be seen in this figure, the total differences among three curves are very small. The pitting potentials, Ep, are nearly the same, but current density is slightly higher in the untreated specimen, indicating that the annealed specimens should have slightly higher corrosion resistance. However, the corrosion fatigue properties of the annealed specimens were much lower than that of the untreated specimen (Fig. 3), and these lower fatigue properties of the annealed specimens in 3% NaCl solution could not be estimated from the anodic polarization behavior shown in Fig. 4. In order to investigate the deterioration of corrosion fatigue properties in the annealed specimens, the specimen surfaces were observed in detail after the anodic polarization tests. Fig. 5 reveals optical micrographs showing the surface morphologies after the polarization tests. Many corrosion pits (diameter o100 μm) within grains are uniformly seen in the untreated specimen (Fig. 5a). No selective attack against grain boundary was recognized. Corrosion pits are denser and larger than in the 1100 °C and 1200 °C annealed specimens (Fig. 5b and c). On the other hand, in the 1100 °C annealed specimen (Fig. 5b), the formation of network-like corrosion pits is seen along the grain boundaries (width 10–15 μm). In the 1200 °C annealed specimen (Fig. 5c),
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
35
Stress amplitude σa (MPa)
Fig. 2. Specimen preparation and heat treatment procedures.
600
Type304 Rotating bending
500
Untreated 1100°C annealed 1200°C annealed
400 300 200 Open : laboratry air Solid : 3%NaCl solution
100 103
6
104 105 10 107 108 Number of cycles to failure Nf
Fig. 3. S–N diagram in laboratory air and in 3% NaCl solution. Table 1 Conditions of anodic polarization measurement. Counter electrode
Pt
Reference electrode Solution Sweep velocity Coating material Measured range Amount of solution
SCE 3% NaCl 10 mV/s Epoxy resin 1.5 V–2 V 25 ml
Fig. 4. Anodic polarization curves.
largely deteriorated in corrosive environment compared with the result in laboratory air (Fig. 3). This behavior is due to Cr-depleted zones which resulted from the precipitation of CrN, namely sensitization. Thus following two heat treatments which aimed to prevent the CrN precipitation were examined.
corrosion pits (diameter o70 μm) are observed within grains similar to the untreated specimen (Fig. 5a), but the selective corrosion along grain boundaries (width 2–3 μm) is partly recognized as indicated by arrows in the figure. It is considered that the selective corrosion along grain boundaries in the annealed specimens (Fig. 5b and c) is due to the precipitation of CrN, resulting in the formation of chromium (Cr) -depleted zones along grain boundaries [18]. Consequently, this mechanism brings the reduction of corrosion fatigue endurance in the annealed specimens, that is the sensitization [19].
(1) The second solution treatment was applied after the annealing in high temperature nitrogen gas so as to dissolve the precipitated CrN. This post-heat treatment is designated as the resolution treatment (RST). (2) Since CrN precipitation predominantly occurred during the furnace cooling period of the annealing treatment in high temperature nitrogen gas, the water quenching treatment (QT) was performed. The samples were dropped into water bath directly from the furnace filled with high temperature nitrogen gas without furnace cooling.
3.2. Heat treatments for improvement
After these heat treatments, corrosion fatigue tests were again performed using the RST and QT specimens to compare with the 1100 °C and 1200 °C annealed specimens. The RST specimen was re-solution treated at 1200 °C for 1 h after the annealing in high temperature nitrogen gas (Fig. 2). On the other hand, the QT specimen was heat-treated in high
3.2.1. Preventive heat treatments of sensitization The anodic polarization test evaluation (Fig. 4) indicated similar corrosion resistances among the untreated and annealed specimens. However, the fatigue properties of the annealed specimens
36
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 5. Optical images of surface morphology after polarization test: (a) untreated specimen, (b) 1100 °C annealed specimen, (c) 1200 °C annealed specimen. Arrows indicate intergranular corrosion.
temperature nitrogen gas for 1200 °C for 24 h followed by water quenching (Fig. 2). 3.2.2. Microstructures of RST and QT specimens Fig. 6 shows the microstructures of RST and QT specimens, including that of 1200 °C annealed specimen [15] for comparison. The average grain size was 160 μm, 92 μm and 100 μm, respectively. It should be noted that QT specimen has the finest grains
Fig. 6. SEM micrographs showing microstructure near surface: (a) RST specimen, (b) QT specimen, (c) 1200 °C annealed specimen. Arrows indicate precipitates.
due to the elimination of furnace cooling period. In the RST and 1200 °C annealed (Fig. 6a and c) specimens, precipitates are recognized near specimen surface. In the previous study [15], the precipitate of 1200 °C annealed specimen was identified as a chromium nitride (CrN) by EDX. On the other hand, in the RST specimen, precipitates are coarser (diameter 4–6 μm) in comparison with CrN. The chemical composition of the precipitate was
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 7. SEM image for point analysis in the RST specimen. Table 2 The quantitative analysis result of precipitate in RST specimen. Element
Atomic concentration (%)
Cr Mn O
25.4 11.7 62.9
analyzed (see Fig. 7). From the quantitative analysis (see Table 2), Cr, Mn and O were detected, that is, the precipitate was identified as (Cr, Mn) oxide. The depths, where the precipitates were observed, are approximately 40 μm and 30 μm for the RST and 1200 °C specimens, respectively. On the other hand, precipitates were not recognized in QT specimen (Fig. 6b). 3.3. Fatigue properties of RST and QT specimens Fig. 8 shows an S–N diagram of the RST and QT specimens in laboratory air and in 3% NaCl solution, where the fatigue test in laboratory air was not performed for the QT specimen due to the limitation of specimen number. But it is assumed that the fatigue properties of the QT specimen might be similar to those of the RST and 1200 °C annealed specimens because of the similar hardness distributions among those heat-treated specimens (Fig. 10, Section
Fig. 8. S–N diagram after heat treatments to prevent sensitization.
37
4.1). The results of 1200 °C annealed and untreated specimens are included in the figure for comparison. In laboratory air (open symbols), the fatigue strength of RST specimen was higher than that of the untreated specimen. Fatigue limit was 270 MPa, 290 MPa and 250 MPa for the RST, 1200 °C annealed and untreated specimens, respectively. The improved fatigue properties in the RST specimen, as well as in the 1200 °C annealed specimen, were due to the solubility of nitrogen (see Section 4.2). In 3% NaCl solution (solid symbols), the corrosion fatigue properties of the 1200 °C annealed, RST and QT specimens were significantly lower than those in laboratory air. The fatigue strength at 107 cycles in 3% NaCl solution was 240 MPa, 160 MPa, 180 MPa and 200 MPa for the untreated, 1200 °C annealed, RST and QT specimens, respectively. It should be noted that the fatigue life of heat-treated specimen in 3% NaCl solution was the longest in the QT specimen and followed by the RST and 1200 °C annealed specimens in decreasing order. Fig. 9 indicates the typical examples of the fracture surface near the crack initiation site in 3% NaCl solution. In all specimens, corrosion pits were not observed, and fatigue crack initiated by cyclic slip deformation. In the 1200 °C annealed specimen, CrNs was distributed as shown in Fig. 6c. But it was confirmed in the previous work that CrN was irrelevant to the fatigue crack initiation mechanism [15]. The magnified view of the crack initiation site in Fig. 9c (RST specimen with (Cr, Mn) oxides) is shown in Fig. 10. It should be noted that the specimen was titled about 45° to observe fracture and specimen surfaces simultaneously. Corrosion pits were not found on the specimen surface side in accordance with Fig. 9. Furthermore, (Cr, Mn) oxide was not recognized at the crack initiation site. It indicates that fatigue crack in the RST specimen initiated regardless of (Cr, Mn) oxides similar to the 1200 °C annealed specimen with CrNs.
4. Discussion 4.1. Distribution of nitrogen concentration and mechanical properties The fatigue properties of annealed specimens in laboratory air were enhanced compared to the untreated one. This improvement is based on the increase in surface hardness caused by the solid solution of nitrogen and the precipitation of nitrides. The detailed discussion about the improvement was made elsewhere [15]. On the other hand, in 3% NaCl solution, the fatigue properties of annealed specimens were deteriorated by the sensitization due to the formation of nitrides. In this section, the reasons of this behavior are discussed. Fig. 11 represents Vickers hardness profiles and the distributions of nitrogen concentration. Vickers hardness of the untreated specimen are shown by the dotted line (HV 168). The hardness at surface are 245 HV, 235 HV and 234 HV for the RST, QT and 1200 °C annealed specimens, respectively. In addition, each hardness distribution profile shows a similar tendency. This increase in surface hardness brings an improvement of fatigue properties in laboratory air. In order to clarify the relationship between the hardness profile and the nitrogen concentration, the distribution of nitrogen concentration was measured by EPMA. In all heat-treated specimens, the nitrogen concentration decreased with increasing depth from surface and then reached a constant value. It should be noted that the hardness profiles correspond well with the distribution of nitrogen concentration as seen in Fig. 11. Table 3 gives the nitrogen contents (mass%) of heat-treated specimens. In all specimens, the nitrogen concentrations at the
38
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 9. SEM micrographs showing fracture surfaces near crack initiation site in 3% NaCl sol.: (a) untreated specimen (sa ¼ 250 MPa, Nf ¼ 6.9 104), (b) 1200 °C annealed specimen (sa ¼190 MPa, Nf ¼ 1.5 106), (c) RST specimen (sa ¼ 240 MPa, Nf ¼1.8 106), (d) QT specimen(sa ¼ 220 MPa, Nf ¼7.5 106). Arrows indicate crack initiation cites.
Fig. 10. SEM micrograph showing crack initiation site of RST specimen in 3%NaCl sol. (sa ¼ 240 MPa, Nf ¼1.8 106). The magnified view of Fig. 9c. Arrow indicates crack initiation cite.
Table 3 Nitrogen contents in specimens after each heat treatment (mass%). Specimen
Surface
Center
1200 °C annealed RST QT
0.55 0.37 0.61
0.23 0.10 0.18
Table 4 Mechanical properties. Specimen
Proof stress s0.2 (MPa)
Tensile strength sB (MPa)
Elongation δ (%)
Reduction of area φ (%)
Untreated 1200 °C annealed RST
242 352
586 726
65 56
70 47
317
716
53
70
Fig. 11. Distribution of Vickers hardness profiles and nitrogen concentration.
surface are higher than those near the center. While the precipitation of nitrides took place in the RST (Fig. 6 a) and 1200 °C annealed (Fig. 6c) specimens near the surfaces, the hardness
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Residual stress σR (MPa)
100 0
Type304 Scatter band Average value
–100 –200 –300 Untreated 1200°C annealed
RST
Fig. 12. Residual stress distribution.
profile and the nitrogen concentration were similar to the QT one (Fig. 6b) without precipitates. Consequently, it is suggested that the higher surface hardness and nitrogen concentration in all specimens are mainly attributed to the solid solution of nitrogen. The mechanical properties were obtained as an average of three tensile specimens for each specimen. These mechanical properties are summarized in Table 4, but QT specimen has no tensile data due to the limited number of samples. Both RST and 1200 °C annealed specimens exhibit higher 0.2% proof stress, s0.2, and tensile strength, sB, than the untreated one. s0.2 was 317 MPa, 352 MPa and 242 MPa for the RST, 1200 °C annealed and untreated specimens, respectively. The elongation, δ, of both heat-treated specimens decreased in comparison with that of the untreated specimen. The higher s0.2 and sB of the RST and 1200 °C annealed
39
specimens could be attributed to the solubility of nitrogen. Consequently, the addition of nitrogen is effective for the improvement of mechanical properties. 4.2. Effects of residual stress and grain size on fatigue properties In the present study, type 304 steel was heat treated, such as annealing, RST and QT, to improve the fatigue properties in 3% NaCl solution. Due to these heat treatments, compressive residual stress could have been generated near specimen surface. Fig. 12 shows the results of residual stress measurements on the specimen surfaces in the untreated, 1200 °C annealed and RST specimens. Residual stress of QT specimen was not measured due to the limitation of specimen number. Average residual stresses are 107 MPa, 142 MPa and 160 MPa for the untreated, 1200 °C annealed and RST specimens, respectively. The compressive residual stress in the untreated specimen was due to the machining of specimen and following polishing process of specimen surface. The effect of residual stress on the fatigue properties seems to be small because the difference of residual stresses among three specimens was small. In addition, as described in the previous Section 3.2, the average grain sizes of RST and 1200 °C annealed specimens are 160 μm and 100 μm, respectively, whereas that of the untreated one is 60 μm [15]. It is well known that the grain coarsening has a detrimental effect on fatigue limit. But in laboratory air, both heattreated specimens exhibit higher fatigue limits than the untreated one, indicating that beneficial effects of nitrogen addition on fatigue properties overcame the effect of grain coarsening. 4.3. Microscopic corrosion behavior Anodic polarization curves in Fig. 4 could not evaluate the enhanced sensitivity to corrosive environment, that is, the sensitization of the annealed specimens shown in Fig. 5. The anodic polarization curves indicate the macroscopic corrosion behavior on the sample surface, while the sensitization is dependent on the highly localized Cr-depleted zones along grain boundaries. As regards the fatigue properties in 3% NaCl solution, the fatigue lives of
Fig. 13. Optical micrographs showing surface morphology after 10% oxalic etch tests: (a) untreated specimen, (b) 1200 °C annealed specimen, (c) RST specimen, (d) QT specimen.
40
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 14. Optical micrographs showing surface morphology after 10% oxalic etch tests: (a) 1200 °C annealed specimen, (b) RST specimen, (c) QT specimen. Arrows indicate intergranular corrosion.
the RST and QT specimens were slightly improved compared with 1200 °C annealed one. The difference in corrosion behavior between 1200 °C annealed, RST and QT specimens was examined. To confirm the corrosion behavior in the RST and QT specimens, the 10% oxalic acid tests were performed on the cross sections of the samples in accordance with ASTM A262 [20]. Fig. 13 reveals the surface morphologies after 10% oxalic acid test in each specimen. In the untreated specimen (Fig. 13a), corrosion pits were uniformly observed on the specimen surface. The corrosion pit size was small, and the network-like corrosion pits along grain boundaries as seen in Fig. 5b were not recognized. In the 1200 °C annealed specimen (Fig. 13b), corrosion pits were uniformly distributed similar to the untreated specimen, while the size and the number of corrosion pits were larger than the other specimens. The surface morphologies of the RST and QT specimens as seen in Fig. 13c and d are nearly the same as compared to the untreated specimen. The corrosion pit sizes were small, and the network-like corrosion pits were not observed at the grain boundaries. These micrographs imply that the corrosion properties of the RST and QT specimens (Fig. 13c and d) were improved by the re-solution or water-quenching treatment compared with the 1200 °C annealed specimen (Fig. 13b). 4.4. Decrease in corrosion fatigue strength by sensitization As seen in the surface morphologies after 10% oxalic etch tests (Fig. 13), corrosion properties of the RST and QT specimens seem to be similar to that of the untreated one. However, the fatigue properties in 3% NaCl solution of the RST and QT specimens were much poorer compared with the untreated one (Fig. 7). The results of the 10% oxalic etch tests (Fig. 13) were obtained on the cross sections of the samples. But the gradient of nitrogen concentration exists on the cross section in these samples as shown in Fig. 11, that is, the nitrogen concentration at the surface was the highest. Therefore, it is expected that the corrosion behavior is different between the surface and center of specimen. Since the fatigue tests were conducted under rotating bending condition, it was considered that the corrosion fatigue properties were strongly dominated by the corrosion property on the specimen surface, not on the cross section. Thus the 10% oxalic etch tests were performed on the surfaces of round bar specimens. Fig. 14 represents the specimen surfaces after the 10% oxalic etch tests, where the differences were apparent. In the 1200 °C annealed (Fig. 14a) and RST specimens (Fig. 14b), grain boundaries are heavily corroded, resulting in the network-like corrosion pits along grain boundaries. In the RST specimen, the width and depth of network-like corrosion pits were smaller than the 1200 °C
annealed specimen. However, the sensitization also took place for the RST specimen. On the other hand, in the QT specimen (Fig. 14c), the network-like corrosion pits are only partially observed along the grain boundaries as shown by the arrows in the figure, but the size of pits is much smaller. As a result, the Crdepleted zone was fully formed along all grain boundaries in the 1200 °C annealed and RST specimens, while partially formed in the QT one. As mentioned in Section 3.2.2, coarse CrN did not precipitate significantly in the RST and QT specimens (Fig. 6a and b). However, it is anticipated that very fine CrN precipitated along grain boundaries near the surface, resulting in the sensitization. Consequently, the deterioration in corrosion fatigue properties of the RST and QT specimens could be attributed primarily to the precipitation of fine CrN and the formation of Cr-depleted zones. In 3% NaCl solution, the fatigue property of the QT specimen was improved as compared with those of the RST and 1200 °C annealed specimens (Fig. 8). It should be noted that the QT treatment could partially suppress the sensitization as seen in S–N diagram (Fig. 8) and the etch test (Fig. 14), but not fully eliminate the precipitation of CrN. The mechanism of corrosion fatigue is highly localized phenomenon, therefore the localized CrN precipitation of the QT specimens had detrimental effect on the corrosion fatigue property.
5. Summary In the present study, the effect of sensitization on corrosion fatigue behavior of type 304 stainless steel annealed in nitrogen gas was investigated. Two heat treatments, namely re-solution treatment (RST) and water quenching treatment (QT) were tried aiming at the prevention of CrN precipitation. Using those specimens, rotary bending fatigue tests were performed to investigate the effect of sensitization on fatigue properties in 3% NaCl solution. The results obtained are summarized as follows: 1. The grains of specimens annealed in nitrogen gas were coarsened. The average grain size of the untreated, 1200 °C annealed, RST and QT specimens was 60 μm, 100 μm, 160 mm and 92 μm, respectively. 2. Annealing in nitrogen gas enhanced the hardness and mechanical properties due to the solubility of nitrogen. The surface hardness of the annealed specimens is significantly higher than the core area. 3. In laboratory air, the fatigue properties of the annealed specimens were enhanced compared with the untreated one. Higher
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
fatigue strength could be attributed to the high surface hardness resulting from the solubility of nitrogen and CrN precipitation. 4. In 3% NaCl solution, the fatigue properties of the annealed specimens were significantly deteriorated in comparison with the untreated one. However, the fatigue strengths of the RST and QT specimens in 3% NaCl solution were slightly improved compared with 1200 °C annealed one in the increasing order. 5. Although the precipitation of CrN was suppressed by RST and QT, the influence of sensitization in the RST and QT specimens was not fully eliminated by both treatments. But QT was still better than RST in the view point of suppression of sensitization.
References [1] [2] [3] [4] [5]
K.H. Lo, C.H. Shek, J.K.L. Lai, Mater. Sci. Eng. R 65 (2009) 39–104. M. Talha, C.K. Behera, O.P. Sinha, Mater. Sci. Eng. C. 33 (2013) 3563–3575. K. Tokaji, K. Kohyama, M. Akita, Int. J. Fatigue 26 (2004) 543–551. M. Akita, K. Tokaji, Surf. Coat. Technol. 200 (2006) 6073–6078. T. Tsuchiyama, H. Ito, K. Kataoka, S. Takaki, Metall. Mater. Trans. A 34A (2003)
41
2591–2599. [6] T. Tsuchiyama, T. Takaki, Denkiseiko 77 (2006) 163–169 , in Japanese. [7] T. Tsuchiyama, Y. Fujii, Y. Terazawa, K. Nakashima, T. Ando, S. Takaki, ISIJ Int. 48 (2008) 861–867. [8] T. Nakanishi, T. Tsuchiyama, H. Mitsuyasu, Y. Iwamoto, S. Takaki, Mater. Sci. Eng. A 460–461 (2007) 186–194. [9] T. Ando, K. Nakashima, T. Tsuchiyama, S. Takaki, Mater. Sci. Eng. A 486 (2008) 228–234. [10] M. Akita, M. Nakajima, Y. Uematsu, K. Tokaji, J.-W. Jung, Fatigue Fract. Eng. Mater. Struct. 35 (2012) 1095–1104. [11] M. Nakajima, Y. Uematsu, K. Tokaji, J.-W. Jung, J. Soc. Mater. Sci. 58 (2009) 956–961 , in Japanese. [12] Y.S. Lim, J.S. Kim, S.J. Ahn, H.S. Kwon, Y. Katada, Corros. Sci. 43 (2001) 53–68. [13] M. Sumita, T. Hanawa, S.H. Teoh, Mater. Sci Eng. C 24 (2004) 753–760. [14] J.W. Simmons, Mater. Sci. Eng. A 207 (1996) 159–169. [15] Y. Bai, M. Akita, Y. Uematsu, T. Kakiuchi, Y. Nakamura, M. Nakajima, Mater. Sci. Eng. A 607 (2014) 578–588. [16] A.S. Hamada, L.P. Karjalainen, M.C. Somani, Mater. Sci. Eng. A 431 (2006) 211–217. [17] M. Nakajima, Y. Uematsu, T. Kakiuchi, M. Akita, K. Tokaji, Procedia Eng. 10 (2011) 299–304. [18] W. Liang, Appl. Surf. Sci. 211 (2003) 308–314. [19] T. Sourmail, C.H. Too, H.K.D.H. Bhadeshia, ISIJ Int. 43 (2003) 1814–1820. [20] ASTM Standard A262-91, Standard Practice for Detecting Susceptibility to Intergranular Attack on Austenitic Stainless Steels, ASTM, Philadelphia, 1993 pp. 1–18.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of sensitization on corrosion fatigue behavior of type 304 stainless steel annealed in nitrogen gas Masayuki Akita a,n, Yoshihiko Uematsu b, Toshifumi Kakiuchi b, Masaki Nakajima c, Toshihiro Tsuchiyama d, Yu Bai e, Kenta Isono f a
Faculty of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan Department of Mechanical Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan c Department of Mechanical Engineering, National Institute of Technology, Toyota College, 2-1 Eisei-cho, Toyota 471-8525, Japan d Department of Materials Science and Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka Nishi-ku, Fukuoka 819-0395, Japan e Graduate Student, Materials Science and Engineering, Graduate School of Engineering, Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto 606-8501, Japan f Aisin Aw Co., Ltd., 10 Takane, Fujii-cho, Aichi, Anjo 444-1192, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 19 March 2015 Received in revised form 22 May 2015 Accepted 23 May 2015 Available online 28 May 2015
Effect of sensitization on fatigue behavior of type 304 stainless steel in laboratory air and in 3% NaCl solution was studied using specimens annealed in high temperature nitrogen gas. Annealing in high temperature nitrogen gas was performed at 1100 °C and at 1200 °C for 30 h, resulting in the solid solution of nitrogen and the precipitation of chromium nitride (CrN). The mechanical and fatigue properties were significantly improved by annealing. The improvement was due to the solid solution of nitrogen and the precipitation of CrN. Fatigue strengths of the untreated specimen in laboratory air and in 3% NaCl solution are nearly the same. However, fatigue properties of the annealed specimen in 3% NaCl solution change for the worse, because chromium (Cr) -depleted zones were formed along grain boundaries during the heat treatment, resulting in the remarkable sensitization. In order to prevent the sensitization, the re-solution treatment (RST) which enhanced the dissolution of CrN and the water quenching treatment (QT) which avoided the precipitation of CrN was performed. As a result, the fatigue properties of the RST and QT specimens in 3% NaCl solution were slightly improved, but were still lower than that of the untreated one. Since the oxalic etch tests proved the formation of Cr-depleted zone also in the RST and QT specimens, the influence of sensitization could not be fully eliminated by the both treatments. & 2015 Elsevier B.V. All rights reserved.
Keywords: Fatigue Austenitic stainless steel Annealed in nitrogen gas Corrosion Sensitization Microstructure
1. Introduction The excellent corrosion resistance of austenitic stainless steel type 304 (18Cr–8Ni) brings on the wide use in chemical, biomedical and mechanical engineering fields [1,2]. However, low strength and poor wear resistance of type 304 steel are the weak points when used as structural and mechanical components [3,4]. Some surface treatment techniques are known to improve the mechanical properties of type 304 steel. Recently, the annealing treatments in high temperature nitrogen gas are applied to stainless steels and Ti alloys to improve the mechanical properties and hardness by the solid solution of nitrogen into matrix [5–9]. This method is a kind of chemical heat treatment and is different from so-called nitriding [6]. Nitrogen is one of the austenite n
Corresponding author. Fax: þ 81 58 293 2378. E-mail address: [email protected] (M. Akita).
http://dx.doi.org/10.1016/j.msea.2015.05.080 0921-5093/& 2015 Elsevier B.V. All rights reserved.
(gamma o γ 4) stabilizing elements for austenitic stainless steels and can improve mechanical properties, fatigue strength [10,11], and corrosion resistance [11–14]. In the previous study [15], authors had reported that the fatigue and the mechanical properties of type 304 steel were highly improved by the annealing in high temperature nitrogen gas. It is well known that the γ-phase in type 304 steel is metastable as expected from its chemical composition, 18Cr–8Ni, taking place in the strain-induced martensitic transformation during fatigue loading [10,11]. However, the strain induced martensitic transformation during fatigue test was successfully suppressed by the annealing in high temperature nitrogen gas, that is, γ-phase was stabilized by an addition of nitrogen [15]. In general, the martensitic phase (α′-phase) has higher environmental sensitivity than γ-phase, thus the strain-induced martensitic transformation decreases the corrosion resistance of type 304 steel [16,17]. Since the γ-phase in type 304 steel was stabilized by the annealing in high temperature nitrogen gas [15],
34
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
it is expected that the corrosion fatigue properties could be also improved by this treatment. In the present study, the annealing in high temperature nitrogen gas was applied to type 304 stainless steel. Since stainless steels are generally required to be used in aggressive environments, the fatigue test should be performed in corrosive environment. But corrosion fatigue properties of type 304 steel annealed in high temperature nitrogen gas have not been fully studied yet. Therefore, in the present study, the fatigue tests in 3% NaCl solution were performed using type 304 stainless steel annealed in high temperature nitrogen gas, and the corrosion fatigue properties were discussed from the viewpoint of sensitization.
spectroscopy (EDX). The distribution of nitrogen concentration was measured using an electron probe microanalyser (EPMA). A measurement of hardness was carried out using a Vickers microhardness tester (test load 4.9 N, loading time 30 s). The residual stress was measured by an X-ray diffraction (XRD). A Cr-Kα X-ray radiation was used at 50 kV and 40 mA. The sin2ψ method was applied for the residual stress measurement [15].
2. Experimental procedures
3.1.1. . S–N diagrams in 3% NaCl solution Fig. 3 represents an S–N diagram in laboratory air and in 3% NaCl solution. The annealing treatments in nitrogen gas improved the fatigue properties in laboratory air (open symbols) in comparison with the untreated specimen, especially the annealing at 1100 °C highly improved the fatigue properties. The fatigue limit was 250 MPa, 380 MPa and 290 MPa for the untreated, 1100 °C and 1200 °C annealed specimens, respectively under a stress ratio of 1. In the previous study, it was indicated that the improved fatigue properties in the annealed specimens were due to the solid solution of nitrogen and the precipitation of chromium nitride (CrN) [15]. In 3% NaCl solution (solid symbols), the fatigue strength at 107 cycles was 240 MPa, 200 MPa and 160 MPa for the untreated, 1100 °C and 1200 °C annealed specimens, respectively. Based on this figure, the fatigue property of the untreated specimen in corrosive environment is nearly the same as that in laboratory air. On the other hand, the corrosion fatigue properties of the annealed specimens were significantly lower than those in laboratory air.
2.1. Materials and annealing treatment in nitrogen gas The material used is type 304 stainless steel (18Cr–8Ni), whose chemical compositions (wt%) are as follows; C: 0.05, Si: 0.29, Mn: 1.69, P: 0.038, S: 0.029, Ni: 8.19, Cr: 18.65, Fe: bal. [15]. Material was solution-treated at 1080 °C for 1 h followed by water quenching. After the solution treatment, material was machined to smooth round-bar fatigue specimens with a diameter of 5 mm and a gauge length of 5.04 mm as shown in Fig. 1. Prior to fatigue tests, the gauge section was mechanically polished using progressively finer grades of emery paper followed by buff-finishing. A vacuum furnace was filled with nitrogen gas at 0.1 MPa. The specimens were stored in high temperature nitrogen gas at 1100 °C or 1200 °C for 30 h [15]. Hereafter, the specimens heat treated at 1100 °C and 1200 °C are designated as the “1100 °C annealed specimen” and the “1200 °C annealed specimen”, respectively. In addition, the specimen without annealing treatment in nitrogen gas is referred to as “untreated specimen”. Fig. 2 shows specimen preparation and heat treatment procedures. The details of each treatment, RST (re-solution treatment) and QT (water quench treatment) in Fig. 2 will be mentioned later (see in Section 3.2.1). 2.2. Experimental procedures Fatigue tests were performed using an Ono-type four-point rotary bending fatigue testing machine with a capacity of 98 Nm operating at a stress ratio of 1 and at a frequency of 60 Hz. Testing conditions were in laboratory air and in 3% NaCl solution. The solution was trickled continuously on the specimen surface through the metering pump from a reserved tank. The microstructures and fracture surfaces of the specimens were examined in detail using a scanning electron microscope (SEM). The precipitates were analyzed by an energy dispersive X-ray
90 30
5
30
12
φ φ
5.04
4
R2 Fig. 1. Configuration of specimen for fatigue tests.
3. Results 3.1. Deterioration of corrosion fatigue properties in annealed specimens
3.1.2. Anodic polarization curves and surface morphologies The corrosion fatigue properties of the 1100 °C and 1200 °C annealed specimens significantly deteriorated as compared with the untreated specimen (Fig. 3). In order to evaluate the corrosion resistance of each specimen, anodic polarization behavior was measured by the potentiostatic approach. Conditions of anodic polarization measurement are listed in Table 1. Fig. 4 shows the anodic polarization curves in each material, where the potential was plotted as a function of current density. Anodic polarization curves were obtained using three samples for each specimen. As can be seen in this figure, the total differences among three curves are very small. The pitting potentials, Ep, are nearly the same, but current density is slightly higher in the untreated specimen, indicating that the annealed specimens should have slightly higher corrosion resistance. However, the corrosion fatigue properties of the annealed specimens were much lower than that of the untreated specimen (Fig. 3), and these lower fatigue properties of the annealed specimens in 3% NaCl solution could not be estimated from the anodic polarization behavior shown in Fig. 4. In order to investigate the deterioration of corrosion fatigue properties in the annealed specimens, the specimen surfaces were observed in detail after the anodic polarization tests. Fig. 5 reveals optical micrographs showing the surface morphologies after the polarization tests. Many corrosion pits (diameter o100 μm) within grains are uniformly seen in the untreated specimen (Fig. 5a). No selective attack against grain boundary was recognized. Corrosion pits are denser and larger than in the 1100 °C and 1200 °C annealed specimens (Fig. 5b and c). On the other hand, in the 1100 °C annealed specimen (Fig. 5b), the formation of network-like corrosion pits is seen along the grain boundaries (width 10–15 μm). In the 1200 °C annealed specimen (Fig. 5c),
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
35
Stress amplitude σa (MPa)
Fig. 2. Specimen preparation and heat treatment procedures.
600
Type304 Rotating bending
500
Untreated 1100°C annealed 1200°C annealed
400 300 200 Open : laboratry air Solid : 3%NaCl solution
100 103
6
104 105 10 107 108 Number of cycles to failure Nf
Fig. 3. S–N diagram in laboratory air and in 3% NaCl solution. Table 1 Conditions of anodic polarization measurement. Counter electrode
Pt
Reference electrode Solution Sweep velocity Coating material Measured range Amount of solution
SCE 3% NaCl 10 mV/s Epoxy resin 1.5 V–2 V 25 ml
Fig. 4. Anodic polarization curves.
largely deteriorated in corrosive environment compared with the result in laboratory air (Fig. 3). This behavior is due to Cr-depleted zones which resulted from the precipitation of CrN, namely sensitization. Thus following two heat treatments which aimed to prevent the CrN precipitation were examined.
corrosion pits (diameter o70 μm) are observed within grains similar to the untreated specimen (Fig. 5a), but the selective corrosion along grain boundaries (width 2–3 μm) is partly recognized as indicated by arrows in the figure. It is considered that the selective corrosion along grain boundaries in the annealed specimens (Fig. 5b and c) is due to the precipitation of CrN, resulting in the formation of chromium (Cr) -depleted zones along grain boundaries [18]. Consequently, this mechanism brings the reduction of corrosion fatigue endurance in the annealed specimens, that is the sensitization [19].
(1) The second solution treatment was applied after the annealing in high temperature nitrogen gas so as to dissolve the precipitated CrN. This post-heat treatment is designated as the resolution treatment (RST). (2) Since CrN precipitation predominantly occurred during the furnace cooling period of the annealing treatment in high temperature nitrogen gas, the water quenching treatment (QT) was performed. The samples were dropped into water bath directly from the furnace filled with high temperature nitrogen gas without furnace cooling.
3.2. Heat treatments for improvement
After these heat treatments, corrosion fatigue tests were again performed using the RST and QT specimens to compare with the 1100 °C and 1200 °C annealed specimens. The RST specimen was re-solution treated at 1200 °C for 1 h after the annealing in high temperature nitrogen gas (Fig. 2). On the other hand, the QT specimen was heat-treated in high
3.2.1. Preventive heat treatments of sensitization The anodic polarization test evaluation (Fig. 4) indicated similar corrosion resistances among the untreated and annealed specimens. However, the fatigue properties of the annealed specimens
36
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 5. Optical images of surface morphology after polarization test: (a) untreated specimen, (b) 1100 °C annealed specimen, (c) 1200 °C annealed specimen. Arrows indicate intergranular corrosion.
temperature nitrogen gas for 1200 °C for 24 h followed by water quenching (Fig. 2). 3.2.2. Microstructures of RST and QT specimens Fig. 6 shows the microstructures of RST and QT specimens, including that of 1200 °C annealed specimen [15] for comparison. The average grain size was 160 μm, 92 μm and 100 μm, respectively. It should be noted that QT specimen has the finest grains
Fig. 6. SEM micrographs showing microstructure near surface: (a) RST specimen, (b) QT specimen, (c) 1200 °C annealed specimen. Arrows indicate precipitates.
due to the elimination of furnace cooling period. In the RST and 1200 °C annealed (Fig. 6a and c) specimens, precipitates are recognized near specimen surface. In the previous study [15], the precipitate of 1200 °C annealed specimen was identified as a chromium nitride (CrN) by EDX. On the other hand, in the RST specimen, precipitates are coarser (diameter 4–6 μm) in comparison with CrN. The chemical composition of the precipitate was
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 7. SEM image for point analysis in the RST specimen. Table 2 The quantitative analysis result of precipitate in RST specimen. Element
Atomic concentration (%)
Cr Mn O
25.4 11.7 62.9
analyzed (see Fig. 7). From the quantitative analysis (see Table 2), Cr, Mn and O were detected, that is, the precipitate was identified as (Cr, Mn) oxide. The depths, where the precipitates were observed, are approximately 40 μm and 30 μm for the RST and 1200 °C specimens, respectively. On the other hand, precipitates were not recognized in QT specimen (Fig. 6b). 3.3. Fatigue properties of RST and QT specimens Fig. 8 shows an S–N diagram of the RST and QT specimens in laboratory air and in 3% NaCl solution, where the fatigue test in laboratory air was not performed for the QT specimen due to the limitation of specimen number. But it is assumed that the fatigue properties of the QT specimen might be similar to those of the RST and 1200 °C annealed specimens because of the similar hardness distributions among those heat-treated specimens (Fig. 10, Section
Fig. 8. S–N diagram after heat treatments to prevent sensitization.
37
4.1). The results of 1200 °C annealed and untreated specimens are included in the figure for comparison. In laboratory air (open symbols), the fatigue strength of RST specimen was higher than that of the untreated specimen. Fatigue limit was 270 MPa, 290 MPa and 250 MPa for the RST, 1200 °C annealed and untreated specimens, respectively. The improved fatigue properties in the RST specimen, as well as in the 1200 °C annealed specimen, were due to the solubility of nitrogen (see Section 4.2). In 3% NaCl solution (solid symbols), the corrosion fatigue properties of the 1200 °C annealed, RST and QT specimens were significantly lower than those in laboratory air. The fatigue strength at 107 cycles in 3% NaCl solution was 240 MPa, 160 MPa, 180 MPa and 200 MPa for the untreated, 1200 °C annealed, RST and QT specimens, respectively. It should be noted that the fatigue life of heat-treated specimen in 3% NaCl solution was the longest in the QT specimen and followed by the RST and 1200 °C annealed specimens in decreasing order. Fig. 9 indicates the typical examples of the fracture surface near the crack initiation site in 3% NaCl solution. In all specimens, corrosion pits were not observed, and fatigue crack initiated by cyclic slip deformation. In the 1200 °C annealed specimen, CrNs was distributed as shown in Fig. 6c. But it was confirmed in the previous work that CrN was irrelevant to the fatigue crack initiation mechanism [15]. The magnified view of the crack initiation site in Fig. 9c (RST specimen with (Cr, Mn) oxides) is shown in Fig. 10. It should be noted that the specimen was titled about 45° to observe fracture and specimen surfaces simultaneously. Corrosion pits were not found on the specimen surface side in accordance with Fig. 9. Furthermore, (Cr, Mn) oxide was not recognized at the crack initiation site. It indicates that fatigue crack in the RST specimen initiated regardless of (Cr, Mn) oxides similar to the 1200 °C annealed specimen with CrNs.
4. Discussion 4.1. Distribution of nitrogen concentration and mechanical properties The fatigue properties of annealed specimens in laboratory air were enhanced compared to the untreated one. This improvement is based on the increase in surface hardness caused by the solid solution of nitrogen and the precipitation of nitrides. The detailed discussion about the improvement was made elsewhere [15]. On the other hand, in 3% NaCl solution, the fatigue properties of annealed specimens were deteriorated by the sensitization due to the formation of nitrides. In this section, the reasons of this behavior are discussed. Fig. 11 represents Vickers hardness profiles and the distributions of nitrogen concentration. Vickers hardness of the untreated specimen are shown by the dotted line (HV 168). The hardness at surface are 245 HV, 235 HV and 234 HV for the RST, QT and 1200 °C annealed specimens, respectively. In addition, each hardness distribution profile shows a similar tendency. This increase in surface hardness brings an improvement of fatigue properties in laboratory air. In order to clarify the relationship between the hardness profile and the nitrogen concentration, the distribution of nitrogen concentration was measured by EPMA. In all heat-treated specimens, the nitrogen concentration decreased with increasing depth from surface and then reached a constant value. It should be noted that the hardness profiles correspond well with the distribution of nitrogen concentration as seen in Fig. 11. Table 3 gives the nitrogen contents (mass%) of heat-treated specimens. In all specimens, the nitrogen concentrations at the
38
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 9. SEM micrographs showing fracture surfaces near crack initiation site in 3% NaCl sol.: (a) untreated specimen (sa ¼ 250 MPa, Nf ¼ 6.9 104), (b) 1200 °C annealed specimen (sa ¼190 MPa, Nf ¼ 1.5 106), (c) RST specimen (sa ¼ 240 MPa, Nf ¼1.8 106), (d) QT specimen(sa ¼ 220 MPa, Nf ¼7.5 106). Arrows indicate crack initiation cites.
Fig. 10. SEM micrograph showing crack initiation site of RST specimen in 3%NaCl sol. (sa ¼ 240 MPa, Nf ¼1.8 106). The magnified view of Fig. 9c. Arrow indicates crack initiation cite.
Table 3 Nitrogen contents in specimens after each heat treatment (mass%). Specimen
Surface
Center
1200 °C annealed RST QT
0.55 0.37 0.61
0.23 0.10 0.18
Table 4 Mechanical properties. Specimen
Proof stress s0.2 (MPa)
Tensile strength sB (MPa)
Elongation δ (%)
Reduction of area φ (%)
Untreated 1200 °C annealed RST
242 352
586 726
65 56
70 47
317
716
53
70
Fig. 11. Distribution of Vickers hardness profiles and nitrogen concentration.
surface are higher than those near the center. While the precipitation of nitrides took place in the RST (Fig. 6 a) and 1200 °C annealed (Fig. 6c) specimens near the surfaces, the hardness
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Residual stress σR (MPa)
100 0
Type304 Scatter band Average value
–100 –200 –300 Untreated 1200°C annealed
RST
Fig. 12. Residual stress distribution.
profile and the nitrogen concentration were similar to the QT one (Fig. 6b) without precipitates. Consequently, it is suggested that the higher surface hardness and nitrogen concentration in all specimens are mainly attributed to the solid solution of nitrogen. The mechanical properties were obtained as an average of three tensile specimens for each specimen. These mechanical properties are summarized in Table 4, but QT specimen has no tensile data due to the limited number of samples. Both RST and 1200 °C annealed specimens exhibit higher 0.2% proof stress, s0.2, and tensile strength, sB, than the untreated one. s0.2 was 317 MPa, 352 MPa and 242 MPa for the RST, 1200 °C annealed and untreated specimens, respectively. The elongation, δ, of both heat-treated specimens decreased in comparison with that of the untreated specimen. The higher s0.2 and sB of the RST and 1200 °C annealed
39
specimens could be attributed to the solubility of nitrogen. Consequently, the addition of nitrogen is effective for the improvement of mechanical properties. 4.2. Effects of residual stress and grain size on fatigue properties In the present study, type 304 steel was heat treated, such as annealing, RST and QT, to improve the fatigue properties in 3% NaCl solution. Due to these heat treatments, compressive residual stress could have been generated near specimen surface. Fig. 12 shows the results of residual stress measurements on the specimen surfaces in the untreated, 1200 °C annealed and RST specimens. Residual stress of QT specimen was not measured due to the limitation of specimen number. Average residual stresses are 107 MPa, 142 MPa and 160 MPa for the untreated, 1200 °C annealed and RST specimens, respectively. The compressive residual stress in the untreated specimen was due to the machining of specimen and following polishing process of specimen surface. The effect of residual stress on the fatigue properties seems to be small because the difference of residual stresses among three specimens was small. In addition, as described in the previous Section 3.2, the average grain sizes of RST and 1200 °C annealed specimens are 160 μm and 100 μm, respectively, whereas that of the untreated one is 60 μm [15]. It is well known that the grain coarsening has a detrimental effect on fatigue limit. But in laboratory air, both heattreated specimens exhibit higher fatigue limits than the untreated one, indicating that beneficial effects of nitrogen addition on fatigue properties overcame the effect of grain coarsening. 4.3. Microscopic corrosion behavior Anodic polarization curves in Fig. 4 could not evaluate the enhanced sensitivity to corrosive environment, that is, the sensitization of the annealed specimens shown in Fig. 5. The anodic polarization curves indicate the macroscopic corrosion behavior on the sample surface, while the sensitization is dependent on the highly localized Cr-depleted zones along grain boundaries. As regards the fatigue properties in 3% NaCl solution, the fatigue lives of
Fig. 13. Optical micrographs showing surface morphology after 10% oxalic etch tests: (a) untreated specimen, (b) 1200 °C annealed specimen, (c) RST specimen, (d) QT specimen.
40
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
Fig. 14. Optical micrographs showing surface morphology after 10% oxalic etch tests: (a) 1200 °C annealed specimen, (b) RST specimen, (c) QT specimen. Arrows indicate intergranular corrosion.
the RST and QT specimens were slightly improved compared with 1200 °C annealed one. The difference in corrosion behavior between 1200 °C annealed, RST and QT specimens was examined. To confirm the corrosion behavior in the RST and QT specimens, the 10% oxalic acid tests were performed on the cross sections of the samples in accordance with ASTM A262 [20]. Fig. 13 reveals the surface morphologies after 10% oxalic acid test in each specimen. In the untreated specimen (Fig. 13a), corrosion pits were uniformly observed on the specimen surface. The corrosion pit size was small, and the network-like corrosion pits along grain boundaries as seen in Fig. 5b were not recognized. In the 1200 °C annealed specimen (Fig. 13b), corrosion pits were uniformly distributed similar to the untreated specimen, while the size and the number of corrosion pits were larger than the other specimens. The surface morphologies of the RST and QT specimens as seen in Fig. 13c and d are nearly the same as compared to the untreated specimen. The corrosion pit sizes were small, and the network-like corrosion pits were not observed at the grain boundaries. These micrographs imply that the corrosion properties of the RST and QT specimens (Fig. 13c and d) were improved by the re-solution or water-quenching treatment compared with the 1200 °C annealed specimen (Fig. 13b). 4.4. Decrease in corrosion fatigue strength by sensitization As seen in the surface morphologies after 10% oxalic etch tests (Fig. 13), corrosion properties of the RST and QT specimens seem to be similar to that of the untreated one. However, the fatigue properties in 3% NaCl solution of the RST and QT specimens were much poorer compared with the untreated one (Fig. 7). The results of the 10% oxalic etch tests (Fig. 13) were obtained on the cross sections of the samples. But the gradient of nitrogen concentration exists on the cross section in these samples as shown in Fig. 11, that is, the nitrogen concentration at the surface was the highest. Therefore, it is expected that the corrosion behavior is different between the surface and center of specimen. Since the fatigue tests were conducted under rotating bending condition, it was considered that the corrosion fatigue properties were strongly dominated by the corrosion property on the specimen surface, not on the cross section. Thus the 10% oxalic etch tests were performed on the surfaces of round bar specimens. Fig. 14 represents the specimen surfaces after the 10% oxalic etch tests, where the differences were apparent. In the 1200 °C annealed (Fig. 14a) and RST specimens (Fig. 14b), grain boundaries are heavily corroded, resulting in the network-like corrosion pits along grain boundaries. In the RST specimen, the width and depth of network-like corrosion pits were smaller than the 1200 °C
annealed specimen. However, the sensitization also took place for the RST specimen. On the other hand, in the QT specimen (Fig. 14c), the network-like corrosion pits are only partially observed along the grain boundaries as shown by the arrows in the figure, but the size of pits is much smaller. As a result, the Crdepleted zone was fully formed along all grain boundaries in the 1200 °C annealed and RST specimens, while partially formed in the QT one. As mentioned in Section 3.2.2, coarse CrN did not precipitate significantly in the RST and QT specimens (Fig. 6a and b). However, it is anticipated that very fine CrN precipitated along grain boundaries near the surface, resulting in the sensitization. Consequently, the deterioration in corrosion fatigue properties of the RST and QT specimens could be attributed primarily to the precipitation of fine CrN and the formation of Cr-depleted zones. In 3% NaCl solution, the fatigue property of the QT specimen was improved as compared with those of the RST and 1200 °C annealed specimens (Fig. 8). It should be noted that the QT treatment could partially suppress the sensitization as seen in S–N diagram (Fig. 8) and the etch test (Fig. 14), but not fully eliminate the precipitation of CrN. The mechanism of corrosion fatigue is highly localized phenomenon, therefore the localized CrN precipitation of the QT specimens had detrimental effect on the corrosion fatigue property.
5. Summary In the present study, the effect of sensitization on corrosion fatigue behavior of type 304 stainless steel annealed in nitrogen gas was investigated. Two heat treatments, namely re-solution treatment (RST) and water quenching treatment (QT) were tried aiming at the prevention of CrN precipitation. Using those specimens, rotary bending fatigue tests were performed to investigate the effect of sensitization on fatigue properties in 3% NaCl solution. The results obtained are summarized as follows: 1. The grains of specimens annealed in nitrogen gas were coarsened. The average grain size of the untreated, 1200 °C annealed, RST and QT specimens was 60 μm, 100 μm, 160 mm and 92 μm, respectively. 2. Annealing in nitrogen gas enhanced the hardness and mechanical properties due to the solubility of nitrogen. The surface hardness of the annealed specimens is significantly higher than the core area. 3. In laboratory air, the fatigue properties of the annealed specimens were enhanced compared with the untreated one. Higher
M. Akita et al. / Materials Science & Engineering A 640 (2015) 33–41
fatigue strength could be attributed to the high surface hardness resulting from the solubility of nitrogen and CrN precipitation. 4. In 3% NaCl solution, the fatigue properties of the annealed specimens were significantly deteriorated in comparison with the untreated one. However, the fatigue strengths of the RST and QT specimens in 3% NaCl solution were slightly improved compared with 1200 °C annealed one in the increasing order. 5. Although the precipitation of CrN was suppressed by RST and QT, the influence of sensitization in the RST and QT specimens was not fully eliminated by both treatments. But QT was still better than RST in the view point of suppression of sensitization.
References [1] [2] [3] [4] [5]
K.H. Lo, C.H. Shek, J.K.L. Lai, Mater. Sci. Eng. R 65 (2009) 39–104. M. Talha, C.K. Behera, O.P. Sinha, Mater. Sci. Eng. C. 33 (2013) 3563–3575. K. Tokaji, K. Kohyama, M. Akita, Int. J. Fatigue 26 (2004) 543–551. M. Akita, K. Tokaji, Surf. Coat. Technol. 200 (2006) 6073–6078. T. Tsuchiyama, H. Ito, K. Kataoka, S. Takaki, Metall. Mater. Trans. A 34A (2003)
41
2591–2599. [6] T. Tsuchiyama, T. Takaki, Denkiseiko 77 (2006) 163–169 , in Japanese. [7] T. Tsuchiyama, Y. Fujii, Y. Terazawa, K. Nakashima, T. Ando, S. Takaki, ISIJ Int. 48 (2008) 861–867. [8] T. Nakanishi, T. Tsuchiyama, H. Mitsuyasu, Y. Iwamoto, S. Takaki, Mater. Sci. Eng. A 460–461 (2007) 186–194. [9] T. Ando, K. Nakashima, T. Tsuchiyama, S. Takaki, Mater. Sci. Eng. A 486 (2008) 228–234. [10] M. Akita, M. Nakajima, Y. Uematsu, K. Tokaji, J.-W. Jung, Fatigue Fract. Eng. Mater. Struct. 35 (2012) 1095–1104. [11] M. Nakajima, Y. Uematsu, K. Tokaji, J.-W. Jung, J. Soc. Mater. Sci. 58 (2009) 956–961 , in Japanese. [12] Y.S. Lim, J.S. Kim, S.J. Ahn, H.S. Kwon, Y. Katada, Corros. Sci. 43 (2001) 53–68. [13] M. Sumita, T. Hanawa, S.H. Teoh, Mater. Sci Eng. C 24 (2004) 753–760. [14] J.W. Simmons, Mater. Sci. Eng. A 207 (1996) 159–169. [15] Y. Bai, M. Akita, Y. Uematsu, T. Kakiuchi, Y. Nakamura, M. Nakajima, Mater. Sci. Eng. A 607 (2014) 578–588. [16] A.S. Hamada, L.P. Karjalainen, M.C. Somani, Mater. Sci. Eng. A 431 (2006) 211–217. [17] M. Nakajima, Y. Uematsu, T. Kakiuchi, M. Akita, K. Tokaji, Procedia Eng. 10 (2011) 299–304. [18] W. Liang, Appl. Surf. Sci. 211 (2003) 308–314. [19] T. Sourmail, C.H. Too, H.K.D.H. Bhadeshia, ISIJ Int. 43 (2003) 1814–1820. [20] ASTM Standard A262-91, Standard Practice for Detecting Susceptibility to Intergranular Attack on Austenitic Stainless Steels, ASTM, Philadelphia, 1993 pp. 1–18.