Effect of Tempering Temperature on the Microstructure, Deformation and Fracture Properties of an Ultrahigh Strength Medium‐Mn Steel Processed by Quenching and Tempering

Herein, a medium‐Mn steel containing 4 wt% Mn is designed to achieve ultrahigh tensile strength and good ductility by the quenching and tempering process. The effects of tempering temperature on the microstructure, tensile behavior, and fracture properties under different stress states are explored. It is observed that tempering temperature influences the fracture mechanisms of the ultrahigh‐strength steels. Both ductile fracture strain and cleavage fracture strength of the investigated medium‐Mn steel are affected by the tempering temperature. The impacts of tempering temperature on volume fraction, carbon content of retained austenite (RA), and tensile properties are quantitatively analyzed. Detailed microstructure analysis reveals that both volume fraction and carbon content of the RA are increased in the tempered steels, accompanied by improved tensile properties than in quenched conditions. The superior tensile properties are achieved after tempering at 250 °C, with the ultrahigh yield strength of 1590 MPa, ultrahigh tensile strength of 1963 MPa, and total elongation of 12%. The corresponding relationship between microstructure and mechanical properties is investigated, with a focus on the change of residual stress, the occurrence of carbon redistribution/carbide precipitation, the decrease of dislocation density of martensite, and the presence of transformation‐induced plasticity effect during deformation.


Introduction
Lightweight automotive engineering is an effective way to improve fuel efficiency, reduce fuel consumption, and decrease the emissions of CO 2 , sulfides, and nitrides. [1,2]The application of advanced high-strength steels in automotive components is an important method to achieve the lightweight engineering target. [2][5][6] Therefore, hot-stamped steel is an important type of material for the lightweight engineering of the body in white.At present, the hotstamped 22MnB5 steel, having the tensile strength of %1500 MPa and total elongation of %7%, is widely used in the structural parts of the automotive body. [7]owever, with the rapid development of lightweight automotive engineering, 22MnB5 steel will not meet the requirements of higher strength and better ductility for automobile steel in the future.Therefore, it is necessary to develop a new generation of hot-stamped steel with improved strength and ductility synergy.It is well known that the hot-stamped boron steel (22MnB5, 37MnB4, etc.) often has a relatively poor ductility because of its full martensitic microstructure. [8]To solve this problem, some studies have shown that the elongation of hot-stamping steel can Herein, a medium-Mn steel containing 4 wt% Mn is designed to achieve ultrahigh tensile strength and good ductility by the quenching and tempering process.The effects of tempering temperature on the microstructure, tensile behavior, and fracture properties under different stress states are explored.It is observed that tempering temperature influences the fracture mechanisms of the ultrahighstrength steels.Both ductile fracture strain and cleavage fracture strength of the investigated medium-Mn steel are affected by the tempering temperature.The impacts of tempering temperature on volume fraction, carbon content of retained austenite (RA), and tensile properties are quantitatively analyzed.Detailed microstructure analysis reveals that both volume fraction and carbon content of the RA are increased in the tempered steels, accompanied by improved tensile properties than in quenched conditions.The superior tensile properties are achieved after tempering at 250 °C, with the ultrahigh yield strength of 1590 MPa, ultrahigh tensile strength of 1963 MPa, and total elongation of 12%.The corresponding relationship between microstructure and mechanical properties is investigated, with a focus on the change of residual stress, the occurrence of carbon redistribution/ carbide precipitation, the decrease of dislocation density of martensite, and the presence of transformation-induced plasticity effect during deformation.be significantly improved by introducing retained austenite (RA) into the microstructure.The RA present in the martensite matrix can be achieved by hot stamping plus the quenching and partitioning (Q&P) process. [9,10]For example, Liu et al. reported that a duplex microstructure of RA and martensite was obtained in a 22MnB5-type steel through the combination of hot stamping with the Q&P process, resulting in an increase of the total elongation (14.8%) than the traditional hot-stamped and quenched condition (6.6%) without compromising strength. [9]However, it is difficult to control the isothermal partitioning process in the industrial hot-stamping line.Recently, Pan et al. confirmed that martensite together with film-like RA (<10 vol%) can be achieved in a 5.6% Mn warm stamped medium-Mn steels, leading to tensile strength of 1717 MPa and total elongation of 16%. [11]On the contrary, some studies showed that although the duplex microstructure (RA and martensite) was obtained, the tensile ductility of the hot-stamped medium-Mn steels without baking was relatively poor. [12,13]In this condition, an additional baking (tempering at 170 °C for 20 min) was needed to improve the ductility. [12,13]In addition, Du et al. reported that a high-uniform elongation (12%) combined with a high tensile strength (%1800 MPa) was achieved in a 7.5% Mn medium-Mn steel processed by quenching and tempering at 300 °C. [14]These previous studies confirm that additional tempering is a critical process for improving the mechanical properties of hot-stamped medium-Mn steel.The tempering temperature, which is a basic parameter for the tempering process, may have a significant influence on the mechanical properties of hot-stamped medium-Mn steel.However, studies of the effect of tempering temperature on the microstructure and mechanical properties of ultrahigh-strength hot-stamped medium-Mn steel are rare.
[20][21] However, only uniaxial tensile tests were used to evaluate the mechanical properties of different ultrahigh-strength medium-Mn steels in most of the studies in the literature, leading to the fact that fracture properties under other loading conditions were not fully considered.By performing tensile tests using different sample geometries with designed notch configurations, various stress states can be achieved in the critical region of the specimens.The failure strain typically decreases with increasing stress triaxiality.Cleavage fracture is also a critical factor for ultrahigh-strength body-centered cubic (bcc) steels, which is more prone to be triggered in high-triaxiality conditions.The activation of cleavage fracture is primarily dependent on the cleavage fracture strength and strain-hardening properties of the material. [22]Therefore, to understand the effects of heat-treatment parameters (tempering temperature) on fracture properties, different stress states should be considered.
In the present study, a medium-Mn steel containing 4 wt% Mn was designed and subjected to quenching and tempering process.The quenching process was used to roughly simulate the hot-stamping process.Ultrahigh strength and considerable ductility were achieved at a suitable tempering temperature.The effects of tempering temperature on the microstructure evolution and tensile properties were investigated.The relationship between microstructure and mechanical properties was also discussed.Moreover, the effects of tempering temperature on the fracture behaviors of ultrahigh-strength medium-Mn steel under different stress states were explored by a combination of experiments and finite-element simulations.

Experimental Section
The chemical composition of the investigated medium-Mn steel is shown in Table 1.The A e3 temperature of the steel calculated by Thermo-Calc software was 731 °C.The M s temperature of the steel measured by dilatometer equipment (DIL 805 A/D) was 210 °C (Figure 1) after heating at a rate of 10 °C s À1 to 800 °C and holding at this temperature for 5 min, followed by cooling to room temperature at a cooling rate of 50 °C s À1 .In particular, the M f temperature was lower than the room temperature.In addition, the result in Figure 1 indicates that the A c3 temperature was higher than 800 °C.The alloy raw materials were melted in a vacuum induction furnace and then produced a 90 kg ingot.The ingot was held at 1200 °C for 4 h, followed by forging into a slab with a section dimension of 70 Â 60 mm.Subsequently, the slab was held at 1200 °C for 2 h, then hot-rolled (HR) to %2.5 mm from 60 mm without reheating between the passes and aircooled to room temperature.The finish rolling temperature was %950 °C.Finally, the HR sheets were subjected to quenching and tempering process.The HR sheets were heat treated at 800 °C for 5 min in a tube-type resistance furnace, quenched in water, and then tempered at 250, 300, and 400 °C for 30 min in a constant temperature oven (resistance furnace with artificial air circulation by fans), respectively.For convenience, the HR sheets subjected to quenching and tempering at different temperatures were referred to as steels QU (without tempering), T250, T300, and T400, respectively.
The microstructures of the quenched sample and tempered samples were observed using JXA-8530F electron probe Table 1.Chemical composition of the experimental steel (mass fraction, %).311)γ peaks according to the following equation [23,24] V where I γ and I α are the average integral intensities of austenite and ferrite peaks, respectively.In addition, the average carbon content of RA (x C ) was calculated according to the following equation [25] a where a γ is the lattice parameter of RA in nm, x C , x Mn , x Cr , and x V are the content of C, Mn, Cr, and V in RA in wt%, respectively.The equation in reference [26] was used to calculate a γ based on (200)γ and (220)γ peaks.Vickers hardness was measured using an FM-700 Vickers hardness testing machine.The engineering stress-strain curves of steels QU, T250, T300, and T400 were obtained using smooth dog-bone (SDB) specimens.In addition, the notched-dog-bone (NDB) and shear (SH) specimens, with significantly different stress states, were manufactured to characterize the deformation and failure behavior of steels T250, T300, and T400 under different stress states.The detailed dimension of SDB, NDB, and SH specimens is shown in Figure 2. The longitudinal axis of all tensile specimens was parallel to the hot-rolling direction.A crosshead speed of 2 mm min À1 was applied during tensile tests at room temperature.The gauge length (GL) was 25 mm for the SDB specimen and 40 mm for the NDB and SH specimens.The stress-strain curves determined using the SDB specimens were used as inputs of finite-element simulations to predict the load-displacement results in the NDB and SH specimens.The slight difference in GL in these geometries does not affect the final results of the determined stress-strain curves.

Effect of Tempering Temperature on Microstructure Evolution
Figure 3 shows secondary electron images of the HR sheet together with steels QU, T250, T300, and T400.Martensite was present in the HR sheet.Martensite together with a small quantity of globular carbides was observed in steel QU.Typical tempered martensite was observed in steels T250, T300, and T400.The martensite laths in steels T250, T300, and T400 were blurred because of the interdiffusion of atoms. [27]n addition, the XRD result in Figure 4 confirms that RA exists in the martensitic matrix of the HR sheet and steel QU.The volume fraction of RA in steel QU was calculated to be 13.6% from the XRD results.The RA remaining after quenching to room temperature was related to the fact that the M f temperature was lower than room temperature.Compared with the steel QU, the volume fraction of RA and the C content of RA in the tempered steels was increased by the tempering process, as shown in Figure 4. Similar results were found in a previously studied medium-Mn steel [14] and a hot-stamping-bake toughening steel, [28] which were related to the fact that carbon partitioned across RA/martensite interfaces and the RA content increased via interface migration during tempering. [14,29][31] In addition, it was noted that the RA content in steels T250, T300, and T400 decreased with the increase in tempering temperature.The reason might be that higher fractions of carbide formed in the martensite matrix consume C at higher tempering temperatures, which is not beneficial to the formation of RA.Moreover, the C content of RA in steels T250, T300, and T400 increased with the increase of tempering temperature, which was related to the higher diffusion rate of C from the martensite matrix to RA at higher tempering temperature.
Figure 5a,b shows bright-field TEM images of the steel QU.Both lath martensite and twinned martensite were observed.The formation of twinned martensite was related to the high C content. Figure 5d shows the high angle angular dark field scanning TEM (HAADF-STEM) image of the precipitates in steel QU.Two different size ranges were observed, with diameters of %35-80 and %7-15 nm, respectively.Two typical precipitates were analyzed by the EDS installed in TEM, as shown in    The precipitate with the size of %10 nm was enriched with V, indicating that the precipitate with a small size (%7-15 nm) was vanadium carbide (VC), as shown in Figure 6a-d.In contrast, Cr was enriched in the precipitate with the size of %50 nm, as shown in Figure 6e-h.Therefore, the precipitate with a large size (%35-80 nm) was Cr-enriched carbide.These globular Cr-enriched carbides might be M 7 C 3 or M 23 C 6 according to refs.[6,27].
Figure 7a,b shows bright-field TEM images of steel T250.Unlike the sample of steel QU, in addition to globular Cr-enriched carbides and globular VC, acicular carbides were observed in steel T250. Figure 7c shows the selected area electron diffraction (SAED) pattern of acicular carbide.The interplanar spacings were precisely measured to determine the carbide species because cementite (θ), η carbide, and ε carbide had similar diffraction patterns and interplanar spacings.The results in Table 2 showed that the carbide interplanar spacings (d 0110 and d 0001 ) of measured values for acicular carbide in the present study were close to that of ε carbide, which indicated that the acicular carbide was ε carbide.In addition, it is observed in Figure 7a that twinned martensite did not disappear after the tempering process.Both lath RA and blocky RA were observed in steel T250, as shown in Figure 7d-f.Figure 7g-l shows the bright-field TEM images of steel T300 and steel T400, respectively.Globular Cr-enriched carbides, globular VC, and acicular ε carbides were observed in steels T300 and T400.Twinned martensite was also found in steels T300 and T400.
The average dislocation density in the martensite matrix of steels QU, T250, T300, and T400 could be measured by XRD experiments and calculated using the following equation [32] ρ ¼ where ρ is dislocation density, β is the full width at half maximum of (211)α peak, and b is Burgers vector.The value of b was 2.48 Â 10 À10 m. [33] The measured dislocation density in the martensite matrix is shown in Figure 8.The result showed that there was a continuous decrease in the dislocation density in the martensite matrix with an increase in tempering temperature because of the recovery of martensite matrix.

Effect of Tempering Temperature on Mechanical Properties and Transformation of RA
The Vickers hardness values of steels QU, T250, T300, and T400 are shown in Figure 9a.The Vickers hardness of steel QU was extremely high with a value of %643 HV, while it decreased significantly after tempering.The Vickers hardness of steels T250, T300, and T400 was %569, %557, and %547 HV, respectively.Figure 9b shows engineering stress-strain curves of steels QU, T250, T300, and T400.A poor ductility due to brittle fracture without plastic deformation was observed in steel QU.The engineering stress-strain curves presented a continuous yielding phenomenon in steels T250, T300, and T400, which could be related to a large number of mobile dislocations in the microstructure. [27]Compared with steel QU, the strength and ductility of steel T250 were increased after tempering.The yield strength, tensile strength, and total elongation of steel T250 were 1590, 1963 MPa, and 12%, respectively, leading to a high product of tensile strength and total elongation (PSE) of 23.56 GPa%.Moreover, it was found that the tensile strength and elongation of steels T250, T300, and T400 were decreased with the increase of tempering temperature.Tensile strength of 1923 MPa and total elongation of 7.6% were obtained in steel T300.When the tempering temperature was increased to 400 °C, the tensile ductility of steel T400 was significantly reduced.The tensile strength and total elongation of steel T400 were 1831 MPa and 2%, respectively.
Figure 10a shows a comparison of the volume fraction of RA in deformed and undeformed samples.The volume fraction of RA in the deformed samples increased with the increase of tempering temperature.Consequently, the transformation ratio of RA decreased from 63.4% to 10.6% when the tempering temperature increased from 250 to 400 °C, as shown in Figure 10b.The mechanical stability of RA during tensile deformation mainly depended on the C content of RA.The RA with higher C content possessed higher mechanical stability. [34,35]Therefore, lower transformation ratio of RA in steel T300 than in steel T250 was related to the higher C content of RA in steel T300 than in steel T250, which led to higher mechanical stability of RA in the former than the latter.Moreover, the transformation ratio of RA in steel T400 was obviously lower than in steel T300.The reason was that the fracture during tensile test occurred at a low engineering strain of 2% in steel T400.Therefore, only the RA with low mechanical stability transformed to martensite at low strain, while the RA with high mechanical stability did not have the opportunity to transform to martensite at higher engineering strain than 2%.

Relationship Between Microstructure and Mechanical Properties
In the present study, the applied tempering processes had a significant influence on the mechanical properties of the investigated steels (Figure 9), which was related to the different microstructures and the transformation-induced plasticity (TRIP) effect during tensile deformation in steels QU, T250, T300, and T400.In steel QU, the martensitic matrix possessed a high hardness because of the combination of high defect density, interstitial solution-hardening, and high residual stress.The high residual stress resulted from martensitic transformation and temperature variation during the quenching process.Therefore, the ductility of steel QU was extremely poor, leading to brittle fracture occurring at a relatively low engineering strain of 0.56%.
Both strength and ductility of steel T250 were significantly improved than steel QU.The increase in tensile ductility of steel T250 was due to the following factors.First, the residual stress caused by the quenching process was partially eliminated after tempering at 250 °C.Second, it is well known that the ductility of the carbon-based martensitic steels could be improved by tempering through carbon redistribution and carbide precipitation. [12,36]The result in Figure 4 confirmed carbon diffused from martensite to RA, and the TEM image in Figure 7 proved the precipitation of acicular ε carbide during tempering in steel T250.Both carbon redistribution and carbide precipitation occurred in steel T250, reducing the carbon content in the martensite matrix of steel T250.Previous studies also reported that the formation of ε carbide via low-temperature tempering was beneficial in improving the ductility of quenched martensitic steels. [36,37]Third, the dislocation density in the martensite matrix of steel T250 was significantly decreased (Figure 8) because of the recovery of martensite during the tempering process.Consequently, the hardness of martensite matrix was decreased, while the ductility of the steel T250 was improved.Fourth, RA contributed to the mechanical properties of steel T250. Figure 11 shows strain-hardening exponent-true strain ε [44] η [45] θ [46] Present study   curves of steels T250, T300, and T400.The strain-hardening exponent (n) was calculated using the following equation [38,39] n ¼ dðlnσÞ=dðlnεÞ where σ is true stress and ε is true strain.The strain-hardening exponent-true strain curve of steel T250 could be divided into 3 stages.The strain-hardening exponent decreased rapidly at stage 1, followed by a slow decrease at stage 2 and a fast decrease at stage 3.The slow decrease of strainhardening exponent at stage 2 was attributed to the TRIP effect, which enhanced the strengthening capacity and delayed the necking. [40]Therefore, the RA in steel T250 was beneficial to the mechanical properties of steel T250 due to TRIP effects.
For steel T300, the dislocation density in martensite matrix was lower than in steel T250, contributing to the lower tensile strength of steel T300 than that of steel T250.In addition, steel T300 possessed a lower volume fraction of RA than steel T250, and the transformation ratio of RA in steel T300 was significantly lower than in steel T250 because of the different mechanical stability of RA in the two steels.The mechanical stability of RA could be compared using the following equation [41,42] f γε ¼ f γ0 expðÀkεÞ (5)   where k is the mechanical stability coefficient of RA, f γ0 is the initial RA fraction before tensile deformation, and f γε is the RA fraction in a fractured sample at true strain ε.A lower k value indicated higher mechanical stability of RA.The k values calculated using Equation ( 5) in steels T250 and T300 were 8.86 and 4.41, respectively, confirming that the mechanical stability of RA in steel T300 was higher than that in steel T250.In this condition, the TRIP effect occurred in a smaller strain range in steel T300 than that in steel T250, as confirmed in Figure 11.Consequently, weaker TRIP effect occurred in steel T300 than that in steel T250, which also led to lower tensile strength and lower elongation of the former than that of the latter.Moreover, it's worth noting that the ductility of steel T400 was so poor that fracture occurred at a low engineering strain of 2%.This phenomenon might be related to the reason that the carbon content of RA in steel T400 was the highest among the investigated conditions.The high-carbon martensite formed by the blocky RA containing high carbon content during the early stage of tensile deformation, which possessed a high hardness, was harmful to the plasticity of steel T400.

Effect of Tempering Temperature on Fracture Behavior Under Different Stress States
The fracture surface morphology in the SDB samples of steels QU, T250, T300, and T400 is shown in Figure 12a-e.
The fracture surface morphology in the NDB samples of steels T250, T300, and T400 is shown in Figure 12f-h.An extremely small SH lip zone was observed in the SDB samples of steel QU, and cleavage fracture appeared in the radial zone, corresponding to the poor ductility of steel QU.In contrast, the SH lip zone on the fracture surface of both SDB and NDB samples was large in steels T250 and T300.Dimples were observed in the radial zones on the fracture surface of both SDB and NDB samples in steels T250 and T300, indicating the occurrence of ductile fracture.In steel T400, cleavage fracture was the dominant failure mechanism, while the SH lip zone of SDB samples in steel T400 was larger than in the steel QU.The fracture surface morphology in the SH samples of steels T250 and T300 is shown in Figure 12i,j.Due to the lower stress triaxiality, the SH dominant ductile fracture was the primary failure mechanism in the SH samples of steels T250 and T300.Surprisingly, fracture did not occur in the SH region in the SH sample of the brittle steel T400, as shown in Figure 12k.Cleavage fracture was the observed failure mechanism in the SH sample of steel T400, which was attributed to the fact that the cleavage fracture strength in the tension region was reached before the critical ductile failure strain was reached in the SH region, as shown in Figure 12l.
The fracture behavior in tensile tests of NDB and SH specimens was simulated by finite-element methods.A fine mesh (0.1 Â 0.1 Â 0.1 mm 3 ) was assigned in the critical region of fracture specimens during finite-element simulations using the ABAQUS/Explicit software.To increase the computational efficiency, half-thickness geometrical models were used for all specimens and the 3D brick elements (C3D8) from the ABAQUS library were selected.The strain-hardening properties of steels T250, T300, and T400 were described using Swift-hardening law, and the hardening parameters were determined based on the uniaxial tensile properties, as shown in Figure 13a.The classical Johnson-Cook fracture criterion, [43] without considering strain rate, temperature, and Lode angle effects, was used to simulate the ductile fracture in NDB and SH specimens of steels T250 and T300.More comprehensive experimental characterization needed to be performed to reveal the Lode angle effects on fracture properties as suggested in recent studies, [18][19][20][21] which was not considered in this study for simplicity.The failure strain as a function of stress triaxiality is shown in Figure 13b.When the ductile fracture criterion was applied in the simulations, the fracture properties of steels T250 and T300 were accurately predicted.The fracture strength in uniaxial tensile tests was taken as the cleavage fracture strength of steels QU and T400, which was shown in Figure 13c.Given the hardening properties and cleavage fracture strength, the cleavage fracture in NDB specimens of steel T400 was accurately predicted.Cleavage fracture was not triggered in tensile tests of NDB specimens in steels T250 and T300 because the cleavage fracture strength was not reached.Therefore, the largest value of maximum principal stress observed in NDB specimens was taken as the lower boundary of cleavage fracture strength of steels T250 and T300.
The determined parameters of hardening law and fracture criterion are summarized in Table 3 for different materials.As summarized in Table 3 and Figure 13, the tempering temperature significantly affected the ductile fracture strain and the cleavage fracture strength of the investigated medium-Mn steels.

Conclusions
1) A relatively low M s temperature of 210 °C and M f temperature lower than room temperature were achieved in the medium-Mn steel designed in this study after holding at 800 °C for 5 min.The martensite matrix with 13.6 vol% RA, Cr-enriched carbide, and VC was obtained in the steel under quenching conditions (QU).2) Compared with the quenched medium-Mn steel QU, both the volume fraction of RA and C content of RA could be increased via tempering treatment to achieve improved tensile properties.
3) The volume fraction of RA and C content of RA in the tempered steels were affected by the tempering temperature.The volume fraction of RA in the tempered steels was decreased with increasing tempering temperature, while the C content of RA was increased.4) Tensile strength and ductility of tempered steel were significantly affected by tempering temperature.The designed medium-Mn steel tempered at 250 °C possessed ultrahigh yield strength of 1590 MPa, ultrahigh tensile strength of 1963 MPa, and total elongation of 12%.5) Fracture mechanisms were affected by the tempering temperature.Both the ductile fracture strain and cleavage fracture strength of the investigated medium-Mn steel show a clear dependence on the tempering temperature.

Figure 4 .
Figure 4.The volume fraction of retained austenite (RA) and C content of RA in steels HR, QU, T250, T300 and T400.HR is hot-rolled sheet.

Figure 5 .
Figure 5. a,b) Bright-field TEM images of the steel QU, c) SAED pattern of the area denoted by a white circle in (b), and d) HAADF-STEM image of the steel QU.SAED is selected area electron diffraction; LM and TM are lath martensite and twinned martensite, respectively; HAADF-STEM is high angle angular dark field-scanning TEM.

Figure 6 .
Figure6.The precipitate with the size of %10 nm was enriched with V, indicating that the precipitate with a small size (%7-15 nm) was vanadium carbide (VC), as shown in Figure6a-d.In contrast, Cr was enriched in the precipitate with the size of %50 nm, as shown in Figure6e-h.Therefore, the precipitate with a large size (%35-80 nm) was Cr-enriched carbide.These globular Cr-enriched carbides might be M 7 C 3 or M 23 C 6 according to refs.[6,27].Figure7a,b shows bright-field TEM images of steel T250.Unlike the sample of steel QU, in addition to globular Cr-enriched carbides and globular VC, acicular carbides were observed in steel T250.Figure7cshows the selected area electron diffraction (SAED) pattern of acicular carbide.The interplanar spacings were precisely measured to determine the carbide species because cementite (θ), η carbide, and ε carbide had similar diffraction patterns and interplanar spacings.The results in Table2showed that the carbide interplanar spacings (d 0110 and d 0001 ) of measured values for acicular carbide in the present study were close to that of ε carbide, which indicated that the acicular carbide was ε carbide.In addition, it is observed in Figure7athat twinned martensite did not disappear after the tempering process.Both lath RA and blocky RA were observed in steel T250, as shown in Figure7d-f.Figure7g-l shows the bright-field TEM images of steel T300 and steel T400, respectively.Globular Cr-enriched carbides, globular VC, and acicular ε carbides were observed in steels T300 and T400.Twinned martensite was also found in steels T300 and T400.The average dislocation density in the martensite matrix of steels QU, T250, T300, and T400 could be measured by XRD experiments and calculated using the following equation[32]

Figure 6 .
Figure 6.a,e) HAADF-STEM images of precipitates in the steel QU; b-d) elemental mappings of the corresponding region in (a); f-h) elemental mappings of the corresponding region in (e).

Figure 7 .
Figure 7. Bright-field TEM images of steels: a,b,d) T250, g-i) T300, and j-l) T400; c) SAED pattern of acicular ε carbide in steel T250; e) dark-field TEM image of RA in (d); f ) corresponding SAED pattern of the RA in (d).

Figure 10 .
Figure 10.a) Comparison of the volume fraction of RA in deformed/undeformed samples and b) transformation ratio of RA.

Figure 12 .
Figure 12.Fracture surface morphology in the a-e) SDB, f-h) NDB, and i,j,l) SH samples of steels: (a,b) QU; (c,f,i) T250; (d,g,j) T300; and (e,h,l) T400.k) The photo of the SH sample of steel T400 after tensile deformation.The inserted images in (a,c-h) show the fracture surface morphology in the radial zone.

Figure 13 .
Figure 13.Fracture properties of steels T250, T300, and T400 in different stress states.a) Strain hardening curves, b) fracture strain, c) cleavage fracture strength, and d-f ) experimental and simulated force-displacement curves of tensile tests.

Table 2 .
Comparison of carbide interplanar spacings (Å) between measured values in the present study and reported values of other carbides in the literature.

Table 3 .
Determined materials parameters of hardening law and fracture criterion for different tempering conditions.