2025 Volume 65 Issue 1 Pages 20-25
From the viewpoint of expanding the allowable P limit, the effect of warm tempforming on delayed fracture resistance was evaluated for 0.09% P-doped 0.4%C–1%Cr–0.7%Mn–0.2%Mo steel (mass%). The P-doped steel was warm tempformed at 500°C with a caliber-rolling reduction of 78% and annealed at 550°C for 1 h. This thermomechanical treatment created an ultrafine elongated grain structure with a strong <110>//rolling direction fiber texture, in which P definitely cosegregated with Mn and Mo at grain boundaries. The slow-strain-rate-test and immersion test demonstrated that warm tempforming markedly enhanced the delayed fracture resistance of the P-doped steel at a tensile strength of 1100 MPa level, in contrast to conventional quenching and tempering treatment.
To achieve carbon neutrality, steel scrap recycling is attracting growing attention. However, repeated recycling of steel scrap increases the contents of impurity elements such as P, S, Cu and Sn, which negatively affect the properties of steels.1) Therefore, in addition to reducing the impurity elements, a major challenge in steel scrap recycling is to develop manufacturing technologies that minimize the negative effects of impurity elements and further expand their allowable limits.
Although P induces intergranular embrittlement in steels, its addition offers advantages such as enhanced strength1) and weatherability.2) Hence, attempts have been made to utilize P as an alloying element through microstructure control of steels. For example, Kawakubo et al. reported that 0.1% C weathering steel with a P content of 0.3% (in mass%), which cannot be joined by fusion welding, could be joined through friction stir welding (FSW) process.2) In the stir zone, a significant decrease in ductile-to-brittle transition temperature was observed as the average ferrite grain size decreased from 23 to 2.5 μm. Min et al. focused on the use of P embrittlement for delamination toughening in warm tempformed 0.4%C–1%Cr–0.7%Mn–0.2%Mo steels with P contents up to 0.09%.3) Warm tempforming involves plastic deformation of quenched and tempered (QT) steel at elevated temperatures, and an ultrafine elongated grain (UFEG) structure with a strong <110>//rolling direction (RD) fiber texture can be obtained by warm multi-pass caliber-rolling.4,5,6,7) The result showed that 0.09% P-doped steel with a UFEG structure was especially toughened by delamination and exhibited superior Charpy impact property compared with QT steel with a P content of 0.001% even at subzero temperatures.
Regarding the practical application of high-strength steels, high resistance to delayed fracture is also required. Warm tempformed (TF) steels with a tensile strength, σB, of 1400–1800 MPa have been shown to exhibit superior delayed fracture properties compared to QT steels.4,5,6,7) However, the P contents of the TF steels were within the allowable limit of P (0.03% or less) of JIS low-alloy steels for machine structural use, such as SCM440 steel (JIS G4053: 2023). It is well known that in QT steels, P facilitates hydrogen embrittlement.8,9) Therefore, from the viewpoint of expanding the allowable P limit, we studied the effect of warm tempforming on delayed fracture resistance for 0.09% P-doped steel.
0.4%C–1%Cr–0.7%Mn–0.2%Mo steels containing 0.001 and 0.09% P, which are similar in basic chemical composition to JIS-SCM440 steel, were prepared by 100 kg vacuum induction melting and casting. Table 1 lists the chemical compositions of 0.001% P (LP) and 0.09% P (HP) steels. The HP steel was oil-quenched after austenitizing at 920°C for 1 h, warm tempformed at 500°C using caliber-rolling with a cumulative reduction of 78% and annealed at 550°C for 1 h (HPTF). For reference, normalized bars were austenitized at 920°C for 1 h, oil-quenched, and tempered at 550°C for 1 h. The average prior-austenite grain size of the QT samples was approximately 30 μm, regardless of the P content. Details of the material preparations were described elsewhere.3) As shown in Table 2, the σB of LPQT, HPQT and HPTF is comparable at approximately 1100 MPa.
| Steels | C | Si | Mn | Cr | Mo | P | S |
|---|---|---|---|---|---|---|---|
| LP1 | 0.40 | 0.24 | 0.73 | 1.04 | 0.22 | 0.001 | <0.001 |
| LP2 | 0.41 | 0.25 | 0.69 | 1.02 | 0.20 | 0.001 | <0.001 |
| HP | 0.41 | 0.24 | 0.71 | 1.00 | 0.20 | 0.093 | 0.0005 |
| Samples | σ0.2 (MPa) | σB (MPa) | UEL (%) | TEL (%) | RA (%) |
|---|---|---|---|---|---|
| LP1QT | 999 | 1122 | 5.4 | 15.6 | 60 |
| LP2QT | 1002 | 1112 | 5.5 | 14.5 | 57 |
| HPQT | 993 | 1137 | 6.1 | 13.2 | 46 |
| HPTF | 1059 | 1133 | 7.9 | 16.6 | 52 |
Microstructures and fracture surfaces were observed by field emission scanning electron microscopy (FE-SEM). Crystallographic orientation was analyzed using an electron backscatter diffraction (EBSD) detector installed in a FE-SEM, with a step size of 50 nm. Elemental analysis was performed using energy-dispersive X-ray spectrometer (EDS) in scanning transmission electron microscope (STEM) (JEM-ARM300F). The thin film sample for STEM was prepared with a focused ion beam (FIB).
Slow-strain-rate testing (SSRT) was performed for notched bar specimens with a stress concentration factor of 4.9 (see Fig. S1(a)).4,5,6,7) Hydrogen was introduced homogeneously into the specimens by cathodic charging with a 0.1 N NaOH aqueous solution or a 3%NaCl + 0.3%NH4SCN aqueous solution at current densities of 0.2–10 A/m2 for 168 h. The hydrogen-charged specimens were tensioned at a crosshead speed of 1 mm/min to a nominal stress of 230 or 920 MPa at the notch and tensioned at a crosshead speed of 0.005 mm/min to failure. The purpose of tensile loading before SSRT was to shorten the testing time and minimize hydrogen release during SSRT. Notch tensile strength, σNB, was calculated by dividing the maximum load by the cross-sectional area at the notch. In the QT sample of LP2 steel with a σB of 1290 MPa, the relationship between σNB and diffusible hydrogen (HD) content in this SSRT condition was almost the same as that in SSRT for Cd-plated specimens (see Fig. S1(b)). An immersion test was conducted for cylindrical specimens (7×20 mm) at 30°C with a 0.5 mol/L NaCl + 0.01 mol/L HCl aqueous solution (pH=2), simulating the atmospheric corrosive environment in Okinawa.10) Thermal desorption spectrometry (TDS) analysis was conducted to measure the HD content; hydrogen that desorbed up to 300°C during TDS was defined as HD. Apparent activation energy for hydrogen desorption, Ea, was evaluated by Kissinger plots of the hydrogen desorption peak temperature, TC, in TDS at different heating rates, Φ, of 100–300°C11) for the plate specimens (1×5×15 mm); hydrogen charging was performed using a 0.1 N NaOH aqueous solution at a current density of 25 A/m2 for 48 h to saturate the specimens with hydrogen.6,7)
Figure 1 shows a UFEG structure with a strong <110>//RD fiber texture in HPTF. The UFEG structure consisted of ribbon- and rod-shaped grains, which inherited the hierarchical heterogeneity of the tempered martensitic structure consisting of blocks and packets within prior-austenite grains. The average transverse intercept length was measured to be 0.5 μm for high angle grain boundaries (HAGBs) with a misorientation angle of 15° or more. Cementite particles were aligned in the RD and more spherical compared to those in LPQT and HPQT.3) Figure 2 presents the results of the STEM-EDS analysis in the UFEG structure. As expected, Cr, Mn, and Mo partitioned into cementite particles. Elemental mapping further revealed that P and S, as well as Mo and Mn, segregated at the HAGB. Depletion zones of Cr and Mn were also observed around the HAGB. The elemental line profile indicated that the element concentration at the grain boundaries (GBs) was in the order of P<Cr<Mn<Mo; however, S concentration could not be quantified because S-Kα peak (2.31 keV) overlapped with Mo-Lα peak (2.29 keV). Regarding the cosegregation of P and these elements, Morita12) showed by a first-principles calculations on Fe Σ3(111) GB that Cr, Mn, and Mo exhibited repulsive interactions with P at GB and reduce the GB segregation of P. The first-principles calculations by Ito et al.13) also showed that Mo segregation at Fe Σ3(111) GB significantly reduced the GB segregation of H due to repulsive interactions between H and Mo. Furthermore, Mo is a strong cohesion enhancer.14) Hence, Mo may suppress intergranular embrittlement in the present steel.
Figure 3 shows the results of the immersion test to determine the maximum absorbed HD content, HE, in an atmospheric corrosive environment. Note that the rust layers for LPQT were removed by sandblasting because they could not be removed by ultrasonic cleaning as for HPQT and HPTF. In all samples, the HD content reached nearly a constant value after the immersion test for 120 h or longer. HE values were thus determined to be 0.06 mass ppm for HPQT, 0.09 mass ppm for HPTF, and 0.13 mass ppm for LPQT. Although the HE of LPQT was higher than that of HPQT and HPTF, it was as low as that for QT steels without effective hydrogen trapping sites4,5,6,7,11) such as nanometer-sized V-carbides15) and Mo-rich precipitates.5,7) Table 3 lists the Ea values. From the TDS analysis, Ea was estimated to be 15.3 kJ/mol for HPQT, 13.3 kJ/mol for HPTF, and 18.9 kJ/mol for LPQT (see Fig. S2). The comparison of these Ea values with those for other steels4,5,6,7,11) suggests that hydrogen might be reversibly trapped in αFe grain boundaries, αFe/Fe3C interfaces and dislocation strain fields in these samples, as previously discussed in 0.6%C-2%Si-1%Cr QT and TF steels.4) The hydrogen trapping ability of HPQT and HPTF was slightly weaker than that of LPQT.
Figure 4 shows the change in σNB as a function of HD content. SSRT from 920 MPa was performed under the hydrogen-precharged condition where σNB exceeded 1000 MPa in SSRT from 230 MPa. There was no significant difference in hydrogen degradation between SSRT from 230 and that from 920 MPa, indicating that hydrogen release during SSRT from 230 MPa was not significant. In the hydrogen-uncharged state, σNB of the QT samples deteriorated from 1840 to 1690 MPa by the addition of 0.09% P but improved to 1890 MPa by warm tempforming. On the other hand, hydrogen charging degraded σNB in all samples. The σNB decreased significantly in the low- HD content range of ~1 mass ppm, followed by a gradual decrease. However, when compared at the same HD content, the decrement in σNB was in the order of HPTF<LPQT<HPQT, indicating that hydrogen embrittlement susceptibility was in the order of HPTF<LPQT<HPQT. It should be noted that HPTF retained a high σNB above 1000 MPa even with a high- HD content of around 3 mass ppm, where the σNB for LPQT sample dropped to around 500 MPa.
Delayed fracture resistance under atmospheric corrosive environments has been discussed in terms of the balance between HE and the critical HD content, HC, for hydrogen embrittlement.4,5,6,7,15,16) Here, the relationship between σNB and HD content can be approximated by the power-law relationship shown in Fig. 4, to determine HC at a given applied stress, σa. Table 4 lists the HC and HE values. When σa was 0.9σB, HC was calculated to be 0.73 mass ppm for LPQT, 0.16 mass ppm for HPQT and 3.7 mass ppm for HPTF. The HC/HE value was thus calculated to be 5.6 for LPQT, 2.7 for HPQT, and 41 for HPTF. HC/HE of HPQT was between that of 0.2%C–1%Mn–0.002%B (=3.6) and JIS-SCM435 (=2.1) steels,16) which were QT steels with respective σB of 1305 and 1320 MPa; 1300 MPa-grade bolts of these steels suffered delayed fracture during outdoor exposure tests in Tsukuba and Okinawa. On the other hand, HC/HE of LPQT was close to that of 0.2%C–1%Mn–0.002%B QT steel (=6.7) with a σB of 1050 MPa,16) where no delayed fracture occurred in the outdoor exposure tests of the bolts. As for HPTF, its HC/HE was much larger than that of LPQT and was 7.0 even at a σa of 1500 MPa. Therefore, it can be concluded that the delayed fracture resistance of HP steel is markedly improved by warm tempforming, in addition to Charpy impact property.3)
Figure 5 shows fractographs from the SSRT specimens for LPQT and HPQT. Almost 100% of the fracture surfaces of HPQT consisted of intergranular fracture along prior-austenite GBs, regardless of the HD content. However, in the hydrogen-charged specimen, an uneven fracture surface suggested that intergranular fracture occurred more frequently (Fig. 5(e)). From a first-principles calculations Yamaguchi et al.17) indicated that the combined mobile and immobile effects of hydrogen largely reduced GB cohesive energy (70–80% reduction), compared to the sole immobile effect of hydrogen (10–20% reduction). As shown in Fig. 4, the decrement in σNB is much larger in hydrogen than in phosphorus. This may be due to the contribution of the mobile effect of hydrogen on intergranular decohesion. Although not shown in Fig. 5, the uncharged specimen for LPQT failed in a completely ductile manner. As the HD content increased, the predominant fracture mode changed from microvoid coalescence to quasicleavage (Figs. 5(c), 5(d)). Furthermore, intergranular fracture was observed in the high- HD content range where its σNB dropped to around 500 MPa (Figs. 5(g), 5(h)). These findings demonstrate that P facilitates hydrogen-induced intergranular fracture.
Figure 6 shows fractographs from the SSRT specimens for HPTF. Macroscopic fracture modes of HPTF were characterized by delamination in the RD and shear fracture, and the frequency of delamination was more pronounced with hydrogen charging. The shear fracture progressed by microvoid coalescence even with a high- HD content of around 3 mass ppm (Fig. 6(h)). By contrast, the predominant mode of delamination in the uncharged specimen was quasicleavage (Fig. 6 (b)), which appeared to change to intergranular fracture along UFEG boundaries with hydrogen charging (Fig. 6(d)). Figure 7 shows delamination cracks that preferentially propagated along UFEG boundaries. Hence, the σNB for HPTF is considered to largely depend on the occurrence of delamination, as discussed for other TF steels with UFEG structures.4,5,6,7)
Possible mechanisms by which hydrogen embrittlement resistance is enhanced by warm tempforming are as follows. Firstly, intergranular fracture along prior-austenite GBs is suppressed in the transverse direction.4,5,6,7) Secondly, the UFEG structure with a strong <110>//RD fiber texture provides more ductile and tougher planes normal to the RD whereas weaker GBs and planes along the RD. This further suppresses transverse brittle cracking and creates the conditions for delamination under a triaxial tension state that generates near the notch root.7) Although delamination cracking lowers the σNB, it relaxes the triaxial tension state and mitigates hydrogen accumulation. Thirdly, multiple microcracks along the RD near the notch root or crack tip might reduce the driving force for the main crack propagation through the stress-shielding effect associated with the interference of multiple microcracks.7) As shown in Fig. 2, the GB segregation of P further facilitates the hydrogen-assisted delamination cracking and microcracks along the RD. The synergistic effect of these mechanisms is thus considered to suppress brittle main crack propagation and maintain a high σNB above 1000 MPa or over even at a high- HD content of 3 mass ppm in HPTF. Similar hydrogen embrittlement behavior has been observed in 0.37%C–0.9%Mn–0.002%B TF steels (σB =1400 MPa),6) and cold caliber-rolled pearlitic 0.6%C–2%Si–1%Cr steel (σB=1670 MPa).18) On the other hand, the σNB of 0.6%C–2%Si–1%Cr TF steel (σB=1540 MPa) dropped to 600 MPa at high- HD content of 3 mass ppm,4) even though it has a hydrogen trapping ability similar to HPTF. This may be explained in terms of differences in the occurrence of hydrogen-assisted delamination fracture in the UFEG structures with various GB characters. Clarification of this requires a more detailed investigation considering the GB characters of the UFEG structure.
Warm tempforming effect on delayed fracture resistance was evaluated for 0.09% P-doped steel at a σB of 1100 MPa level. Warm tempforming combined with annealing at 550°C for 1 h evolved a UFEG structure with a strong <110>//RD fiber texture, in which P definitely cosegregated with Mn, and Mo at GBs. In quenching and tempering at 550°C for 1 h, the addition of 0.09% P significantly degraded hydrogen embrittlement resistance for the steel by facilitating intergranular fracture along prior-austenite GBs. By contrast, warm tempforming markedly enhanced the hydrogen embrittlement resistance of the P-doped steel due to the UFEG structure. The TF steel was shown to have a significantly high delayed fracture resistance when considering the balance between the critical HD content for hydrogen embrittlement and the maximum absorbed HD content in an atmospheric corrosive environment. This fact indicates that an impurity element of P can be utilized as an effective element against hydrogen embrittlement if we correctly chose the microstructure.
[Figs. S1–S2] This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2024-189.
The authors declare no conflicts of interest.
The authors thank Mr. Iida, Mr. Kobayashi and Mr. Hibaru for materials processing with caliber rolling, and Ms. Seki for her help with TDS analysis and FE-SEM observation. A part of this study was supported by the Electron Microscopy Unit, NIMS. This study was partly supported by grants from the JSPS KAKENHI Grant Number 24K01204, and MEXT Program: Data Creation and Utilization Type Material Research and Development Project Grant Number JPMXP1122684766.
