Effect of Intercritical Annealing on Microstructure and Toughness of Medium-Mn Steel with Elongated Prior-austenite Grains Formed via Two-step Hot Rolling Process

Abstract

Fe-9 mass%Ni alloy is widely used as a cryogenic steel owing to its excellent low-temperature strength and toughness. However, Ni is an expensive element, with medium-Mn steel considered an inexpensive alternative. Considering the Fe-10%Mn-0.1%C alloy is brittle at low temperatures, the application of intercritical annealing with two-step hot rolling could lead to toughening. Herein, the effect of intercritical annealing on the toughness of a Fe-10%Mn-0.1%C alloy with elongated prior-austenite grains (PAGs) formed via a two-step hot-rolling process was investigated. Intercritical annealing was performed on the specimens with and without two-step hot rolling. For both specimens, intercritical annealing resulted in softening of α’-martensite and an increase in the amount of retained austenite. In the specimen not subjected to the two-step hot rolling process, the fracture morphology transitioned from ductile to intergranular with a decrease in the temperature. Intercritical annealing improved the toughness when ductile fracture occurred. In the case of intergranular fracture, the effect of intercritical annealing on the toughness was negligible. In the two-step hot-rolled specimen with elongated PAGs, the fracture morphology transitioned from ductile to separation fracture with ductile fracture, and intercritical annealing improved the toughness at all temperature ranges. The improvement in toughness during separation fracture is attributed to the expansion of the plastic zone owing to ductile crack progression and the formation of sub-cracks, which promote the strain-induced transformation of retained austenite and ε-martensite.

1. Introduction

Fe-9 mass%Ni alloy exhibits excellent strength and toughness even at low temperatures and hence is widely used as a cryogenic steel for liquefied natural gas storage tanks.^1,2,3) The excellent low-temperature toughness of the alloy can be attributed to the high dislocation mobility in the martensitic matrix owing to solid-solution Ni⁴⁾ and highly stable Ni-enriched retained austenite.⁵⁾ However, considering Ni is an expensive element, utilizing this alloy with large amounts of Ni leads to high equipment construction costs. Therefore, attempts have been made to replace Ni with Mn, a relatively inexpensive element.^6,7,8) Medium-Mn steel with 3–10% Mn has excellent hardenability, similar to the Fe-9%Ni alloy, and Mn behaves similarly to Ni as an austenite-stabilizing element in steel. Therefore, highly stable Mn-enriched retained austenite can be obtained by intercritical annealing (IA), improving its toughness in medium-Mn steel.^9,10) However, medium-Mn steels have fragile grain boundaries that cause intergranular fractures at low temperatures. Kwon et al.¹¹⁾ conducted Charpy impact tests on as-quenched Fe-9%Ni and Fe-8%Mn alloys at 77 K. Their results showed that intergranular fracture occurred in the Fe-8%Mn alloy, with the impact energy of the Fe-8%Mn alloy lower than that of the Fe-9%Ni alloy, indicating that the intergranular strength of the Fe-8%Mn alloys is lower than that of the Fe-9%Ni alloys. In addition, Yamanaka et al.¹²⁾ reported that the brittle fracture mode in Fe–Mn alloys shifts from cleavage to intergranular fracture with increasing Mn content.

While IA can marginally suppress intergranular fracture in medium-Mn steels,^13,14,15) completely suppressing intergranular fracture while achieving toughening at cryogenic temperatures is difficult. Murakami et al.¹³⁾ performed IA on Fe-6%Mn-0.05%C alloys and found a significant increase in toughness at temperatures higher than 173 K. However, intergranular fracture could not be suppressed at 77 K, with the effect of retained austenite formed by IA hardly noticeable at cryogenic temperatures. Therefore, intergranular fractures should first be suppressed to enhance the low-temperature toughness of medium-Mn steels with fragile grain boundaries.

The Fe-10%Mn-0.1%C alloy exhibits significant intergranular embrittlement at low temperatures. In our previous study on the suppression of intergranular fracture in the Fe-10%Mn-0.1%C alloy, the prior-austenite grains (PAGs) were elongated through a two-step hot rolling process—first in the recrystallization temperature range and then in a temperature range in which recrystallization is difficult.¹⁶⁾ Elongation of PAGs suppressed intergranular fracture and improved toughness at cryogenic temperatures via separation,^17,18,19) which suppressed the propagation of cracks by the relaxation of the triaxial stress condition, contributing to improving toughness. Furthermore, the formation of a fine-grained microstructure, obtained via the γ→ε→α′ two-step martensitic transformation, could effectively increase the absorbed energy at low temperatures. However, only the toughness of the as-quenched specimen without IA was investigated in our previous study.¹⁶⁾ The application of IA to the Fe-10%Mn-0.1%C alloy with two-step hot rolling is expected to lead to toughening owing to the elongated PAGs and retained austenite, which does not occur in conventional medium-Mn steels with intergranular fractures.

In this study, IA was performed on specimens with equiaxed and elongated PAGs. Microstructural evaluations and Charpy impact tests were conducted on these specimens to clarify the effect of IA on the toughness of medium-Mn steel with elongated PAGs. In addition, the toughening mechanisms of the intercritically annealed (IAed) medium-Mn steel with elongated PAGs are discussed in terms of the microstructure and fracture surface morphology.

2. Experimental Procedure

A 50-kg ingot of the Fe-10%Mn-0.1%C alloy was prepared via vacuum melting, with the chemical composition of the alloy listed in Table 1. The ingot was subjected to thermo-mechanical treatment as illustrated in Fig. 1. After austenitization at 1473 K for 7.2 ks, the ingot was subjected to two-step hot rolling: 50% rolling at temperatures above 1273 K and 75% rolling at 1017–1060 K, followed by water quenching (the resulting specimen is denoted RQ).¹⁶⁾ Furthermore, RQ was IAed at 873 K for 3.6 ks (denoted RQ-IA), where the maximum amount of retained austenite with high stability was obtained, even at 77 K. In addition, a reference specimen was prepared as follows. The specimen was austenitized at 1473 K for 7.2 ks, followed by water quenching without hot rolling (denoted AsQ), and IAed at 873 K for 3.6 ks (denoted Q-IA).

Table 1. Chemical compositions (mass%) of the Fe-10%Mn-0.1%C alloy.

C	Si	Mn	P	S	Al	Fe
0.097	0.019	10.03	<0.001	0.0029	0.012	bal.

Fig. 1. Schematic of the heat treatment process.

Microstructural observations were conducted via electron backscatter diffraction (EBSD) analysis using an orientation imaging microscope (OIM, TSL Solutions, Japan) equipped with a field-emission scanning electron microscope (FE-SEM, SIGMA 500, Zeiss, operated at 20 kV, Germany). The specimens used for microstructural observations were mechanically and chemically polished using colloidal silica. The volume fraction of each phase was measured using a time-of-flight neutron diffractometer (iMATERIA)²⁰⁾ at the Materials and Life Science Experimental Facility of the Japan Proton Accelerator Research Complex. The dimensions of the specimen were 8 × 6 × 65 mm, and the measurement time was 600 s. The size and power of the neutron beam were 20 × 20 mm and 520 kW, respectively. The neutron diffraction peak profiles were analyzed to estimate the volume fraction via Rietveld texture analysis using the Maud software.²¹⁾ In addition, the changes in the volume fraction of austenite during IA were measured via in situ neutron diffraction during heating. For specimens (70 × 10 × 2 mm) heated to 873 K, the neutron diffraction line profiles were obtained using a neutron beam with a beam power of 529 kW.

The hardness of each specimen was measured via the Vickers hardness test (load: 196 N, holding time: 10 s). Charpy impact tests were performed with full-size Charpy specimens (10 mm^t × 10 mm^w × 55 mm^l) with a 45° V-notch in the transverse direction (TD) plane in the temperature range of 77–373 K. The Charpy impact value was evaluated by dividing the absorbed energy by the cross-sectional area (80 mm²). The fracture surfaces were observed using a SIGMA500 (operated at 5 kV).

3. Results and Discussion

3.1. Microstructural Characterization

3.1.1. Microstructure of the as-quenched Fe-10Mn-0.1C Alloy (AsQ and RQ)

Figure 2 shows the crystallographic orientation maps and phase maps of AsQ and RQ. The phase map of the area enclosed by black squares in Fig. 2(a) is shown in Fig. 2(c). Figures 2(b) and 2(e) present the prior austenite maps of AsQ and RQ, respectively.²²⁾ The crystal orientation of the reconstructed prior austenite was calculated based on the Kurdjumov–Sachs (K–S) relationship ((111)_fcc//(011)_bcc, [101]_fcc//[111]_bcc). AsQ and RQ exhibit significantly different PAG morphologies. AsQ exhibits equiaxed PAGs with a grain size of approximately 147 μm. In contrast, as shown in Fig. 2(e), RQ exhibits elongated PAGs along the rolling direction, and large misorientations exist within these grains. This indicates that the recrystallization of hot-rolled austenite hardly occurs, and the austenite before quenching was work-hardened and had a high dislocation density. Both as-quenched specimens exhibit a complex microstructure comprising α’-martensite (bcc) matrix, ε-martensite (hcp), and retained austenite (fcc). The formation of such a multiphase fine microstructure occurs via a two-step martensitic transformation (γ→ε→α’) in medium-Mn steels with 10–15% Mn.^16,23) The addition of Mn decreases the stacking fault energy of austenite; hence, the plate-like ε-martensite forms within the austenite grain, followed by the nucleation of α’-martensite inside the ε-martensite. Therefore, the granular α’-martensite and ε-martensite with a small aspect ratio remained, and their morphologies differed from those of the general lath martensite with elongated blocks. The block widths of α’-martensite in AsQ and RQ are almost similar (0.4 ± 0.2 μm and 0.3 ± 0.2 μm, respectively), suggesting that the work-hardened austenite microstructure of RQ has minimal effect on the block width. Table 2 summarizes the Vickers hardness, PAG size, block width, and volume fraction of each phase. At room temperature (RT: approximately 300 K), the volume fraction of retained austenite in AsQ was 9 vol.%, whereas that in RQ was 19 vol.%, attributed to the improved mechanical stability of austenite in RQ owing to the dislocations introduced during hot rolling.²⁴⁾ The hardness values of AsQ and RQ are 471 ± 13 and 473 ± 11 HV, respectively. In general, the retained austenite decreases the hardness of the as-quenched specimen;^25,26) however, the hardness values of AsQ and RQ are close. The retained austenite in RQ exhibits a large number of dislocations introduced by hot rolling. Such dislocations are inherited into the α’-martensite (ausforming effect²⁷⁾), thereby increasing the hardness of the α’-martensite in RQ, which, in turn, suppresses the decrease in hardness of RQ with a large amount of retained austenite.

Fig. 2. Microstructures of (a)–(c) AsQ and (d)–(f) RQ: (a)(d) the crystallographic orientation maps of all phases, (b)(e) the reconstructed austenite maps of (a) and (d), respectively, and (c)(f) the phase maps of (a) and (e). (Field of view in (c) is the black-frame area in (a).). (Online version in color.)

Table 2. Phase fraction, hardness, PAGS, and block width of specimens quenched into RT and cooled into 77 K after quenching.

		Vickers hardness measured at RT	PAG size (μm)	Block width (μm)	Volume fraction (%)
					α’-martensite	ε-martensite	Retained γ
Specimens quenched into RT	AsQ	471 ± 13	147	0.4 ± 0.2	73	18	9
	RQ	473 ± 11	–	0.3 ± 0.2	61	20	19
	Q-IA	360 ± 8	166	0.4 ± 0.1	40	19	41
	RQ-IA	368 ± 8	–	0.5 ± 0.2	36	13	51
Specimens cooled into 77 K after quenching	AsQ	490 ± 13	168	0.4 ± 0.1	79	19	2
	RQ	478 ± 11	–	0.3 ± 0.1	75	19	6
	Q-IA	357 ± 12	188	0.4 ± 0.1	47	39	14
	RQ-IA	351 ± 7	–	0.3 ± 0.2	47	27	26

3.1.2. Microstructural Change via Intercritical Annealing

The reverse transformation behaviors of AsQ and RQ during heating to 873 K were investigated via in situ neutron diffraction analysis. Figure 3 illustrates the changes in the integrated intensity of the 200_α’, 200_γ, and 1011_ε diffraction peaks in AsQ²³⁾ and RQ. The heating rate was set at 10 K/s. Several diffraction peaks overlap because this alloy contains three different phases. Therefore, in Fig. 3, three independent diffraction peaks²³⁾ were utilized. In both specimens, the integrated intensity of the 1011_ε peak decreases, and that of the 200_γ peak begins to increase at approximately 535 K, indicating that the ε→γ reverse transformation occurs at 535 K. In our previous study, we found that the reversed austenite has high-density dislocations, and the ε→γ reverse transformation in this alloy is classified as a displacive transformation without partitioning of alloying elements.²³⁾ The finish temperatures of the ε→γ reverse transformation in AsQ and RQ are approximately 712 and 650 K, respectively. After the completion of ε→γ reverse transformation, the α’→γ reverse transformation starts at approximately 823 K, slightly lower than the IA temperature, with this process interpreted as a massive-type transformation without Mn partitioning.²³⁾ The volume fractions of massive austenite in AsQ and RQ, which transforms into α’-martensite during heating, were estimated from the changes in the integrated intensity (Fig. 3) and their respective values are approximately 7 and 3 vol.%. Therefore, the volume fractions of austenite immediately after reaching 873 K in AsQ and RQ are approximately 34 and 42 vol.%, respectively, and the remaining proportion corresponds to tempered α’-martensite.

Fig. 3. Changes in integrated intensity of 200_α’, 200_γ, and 1011_ε during heating to 873 K in (a) AsQ and (b) RQ. (Online version in color.)

Figure 4 shows the changes in volume fractions of austenite and full width at half maximum (FWHM) of the 200_γ peak during IA at 873 K in AsQ and RQ. The changes in FWHM at the 200_γ peak correspond to the change in the amount of strain in austenite. The figure also illustrates the austenite volume fractions just after reaching 873 K (〇). The α’→γ reverse transformation proceeds during IA at 873 K in both specimens. The volume fraction of austenite reaches the equilibrium volume fraction of austenite (69 vol.%) at 873 K and is saturated at approximately 3.6 ks. Reverse transformation during IA should proceed via the growth of the retained austenite present before heating and the austenite transformed from ε-martensite and α’-martensite during heating.²³⁾ The α’→γ reverse transformation proceeds by interface migration accompanied by Mn partitioning from α’-martensite to austenite because a local equilibrium is established at the γ/α’ interface during IA.²⁸⁾ In the reversed austenite region swept by the interface, the Mn concentration is expected to be close to the equilibrium composition (approximately 13 mass%). However, DICTRA simulations revealed that Mn could not be concentrated in the austenite interior at 873 K.²⁸⁾ In contrast, C, an interstitial solid solution element, can sufficiently diffuse into the austenite interior even at 873 K. Assuming that all Cs in α’ martensite are enriched into austenite, the C concentration in austenite becomes approximately 0.14 mass% during IA at 873 K.

Fig. 4. Changes in volume fraction of austenite and FWHM at the 200_γ peaks during intercritical annealing at 873 K in (a) AsQ and (b) RQ.

Furthermore, the FWHM of the 200_γ peak gradually decreases during IA in both AsQ and RQ, and the FWHM of RQ is always greater than that of AsQ, indicating that the dislocations introduced by the two-step hot rolling remain in the austenite of RQ even after IA at 873 K for 3.6 ks. However, the dislocations gradually decrease during IA.

Figure 5 shows the crystallographic orientation maps of Q-IA (IAed AsQ) and RQ-IA (IAed RQ). IA was performed at 873 K for 3.6 ks, followed by water cooling to RT. The data for ε-martensite and retained austenite were extracted from Figs. 5(a) and 5(d) and are shown in Figs. 5(b)(e) and 5(c)(f), respectively. Q-IA and RQ-IA respectively exhibited equiaxed and elongated PAGs along the rolling direction. There was no change in the morphology of the PAG owing to IA in either specimen. The Vickers hardness values of Q-IA and RQ-IA were 360 and 368 HV, respectively, and the hardness decreased significantly due to IA. Table 2 lists the volume fraction of each phase in Q-IA and RQ-IA measured by neutron diffraction. Figure 6 summarizes the changes in phase fraction (AsQ → Q-IA and RQ → RQ-IA) via IA (873 K–3.6 ks) and subsequent water cooling. The volume fractions of retained austenite significantly increased owing to IA and were estimated to be 41 and 51 vol.% in Q-IA and RQ-IA, respectively. In some cases, the crystallographic orientation maps are not consistent with the neutron diffraction results (e.g., the amount of ε-martensite in Q-IA), possibly because the retained austenite at the specimen surface is unstable and easily transformed to ε-martensite or α’ martensite. Figures 5(c) and 5(f) demonstrate that retained austenite with the same crystal orientation is present in the PAGs even after reverse transformation, indicating the occurrence of austenite memory.^29,30) Q-IA contains 26 vol.% tempered α’-martensite and 74 vol.% austenite just before water cooling. After water cooling post IA, 14 vol.% austenite transforms to fresh α’-martensite, and 19 vol.% austenite transforms to ε-martensite, resulting in 41 vol.% retained austenite. The volume fraction of retained austenite in RQ-IA was 10 vol.% larger than that in Q-IA. However, the Mn-enriched austenite in RQ-IA is expected to be less than that in Q-IA because the amount of reversed austenite formed during IA is smaller in RQ-IA than in Q-IA as shown in Fig. 4. Therefore, the reason why the amount of retained austenite is larger in RQ-IA is not because a large amount of Mn-enriched austenite is formed during IA, but because the austenite with high dislocation density remained even after IA and the mechanical stabilization by dislocations may contribute to the austenite stability in RQ-IA. The amount of retained austenite in RQ-IA can be explained by considering that the Mn-enriched austenite formed during IA (27 vol.%) and the retained austenite (with high dislocation density) present before IA (19 vol.%) remained even after water cooling.

Fig. 5. Crystallographic orientation maps of (a)–(c) Q-IA and (e)–(g) RQ-IA: ((a) and (e) all phases; (b) and (f) hcp; (c) and (g) fcc). (Online version in color.)

Fig. 6. Changes in volume fraction of each phase due to intercritical annealing at 873 K for 3.6 ks in (a) AsQ and (b) RQ. (Online version in color.)

3.2. Evaluation of Toughness and Fracture Behavior

3.2.1. Effect of Intercritical Annealing on the Toughness and Fracture Behavior of AsQ and Q-IA

Figure 7 illustrates the relationship between the Charpy impact value and temperature (ductile-to-brittle transition curve) for AsQ¹⁶⁾ and Q-IA. Figure 8 shows the SEM images of the fracture surfaces of the specimens subjected to Charpy impact tests at 373 K and 77 K. The Charpy impact value increased at test temperatures greater than 273 K, owing to IA. Ductile fracture surfaces characterized by dimples were observed in the AsQ and Q-IA specimens tested at 373 K (Figs. 8(a)–8(d)), and intergranular fractures were partially mixed. Q-IA has a lower hardness and contains more retained austenite than AsQ, as shown in Table 2. Therefore, the increase in toughness at temperatures above 273 K where the ductile fracture occurs would be because of the increased plastic deformation due to the softening of α’-martensite, plastic deformation of the retained austenite itself, and strain-induced transformation from ε-martensite and retained austenite to α’-martensite.^31,32,33)

Fig. 7. Charpy impact value vs. temperature curves of AsQ and Q-IA. (Online version in color.)

Fig. 8. (a), (b), (e), and (f) Fracture surfaces of AsQ; (c), (d), (g), and (h) Fracture surfaces of Q-IA: (a)–(d) tested at 373 K ((a) and (c) low magnification, (b) and (d) enlarged view of (a) and (c), respectively), and (e)–(h) 77 K ((e) and (g) low magnification, (f) and (h) enlarged view of (e) and (g), respectively). (Online version in color.)

However, the difference between the Charpy impact values of AsQ and Q-IA was small at test temperatures below 173 K, and the impact values of both specimens were close to zero at 77 K. The intergranular fracture surface can be observed in both specimens at 77 K (Figs. 8(e)–8(h)), and the fracture unit is approximately 150–200 μm, which corresponds to the PAG size, as presented in Figs. 2 and 5 and Table 2. These results indicate that the PAG boundary (PAGB) of this alloy was brittle at 77 K, even after IA. However, the intergranular fracture surface patterns differed between the two specimens. Figures 8(f) and 8(h) show the magnified images of the intergranular fracture surface. The intergranular fracture surface of AsQ (f) is flat, whereas that of Q-IA (h) has small facets. The microstructure of the AsQ and Q-IA specimens cooled to 77 K was examined via EBSD analysis to elucidate the differences in their intergranular fracture surfaces; the results are shown in Fig. 9 and Table 2. In AsQ, the volume fraction of retained austenite decreases owing to cryogenic cooling (9 → 2 vol.%), but the change in the volume fraction of ε-martensite is negligible (18 → 19 vol.%). By contrast, the decrease in retained austenite (41 → 14 vol.%) and the increase in ε-martensite (19 → 39 vol.%) were confirmed in Q-IA.

Fig. 9. Microstructures of (a) and (b) AsQ and (c) and (d) Q-IA after cooling at 77 K: ((a) and (c) crystallographic orientation map (b) and (d) phase map). (Online version in color.)

The proportion of retained austenite and ε-martensite in Q-IA is higher than that in AsQ after cooling (Figs. 9(b)(d)). The retained austenite and ε-martensite in Q-IA are uniformly distributed and present at the PAGB and within the PAG. Therefore, the small facets on the intergranular fracture surface in Q-IA are considered to be the traces of ε-martensite and retained austenite at the PAGB. Previous studies have reported similar fracture surfaces on the Fe-9%Ni alloy and Fe-6%Mn alloys containing retained austenite.^13,34) However, the Charpy impact value hardly increases because of IA. Thus, when cracks propagate along the PAGBs and intergranular fracture occurs, the effects of ε-martensite and retained austenite at the PAGB and the softening of the matrix by IA on the toughness are minimal because the PAGBs in Fe-10Mn steel are extremely brittle.

3.2.2. Effect of Intercritical Annealing on the Toughness and Fracture Behavior of RQ and RQ-IA

Figure 10 shows the relationship between the Charpy impact value and the temperature for RQ¹⁶⁾ and RQ-IA. Figure 11 shows the fracture surfaces at 373 K and 77 K. At most test temperatures, the impact values of RQ were higher than those of AsQ, as shown in Fig. 7. In addition, the Charpy impact value increased significantly due to IA at all temperatures, including 77 K.

Fig. 10. Charpy impact value vs. temperature curves of RQ and RQ-IA. (Online version in color.)

Fig. 11. Fracture surfaces of (a), (b), (e), (f), (i), and (j) RQ and (c), (d), (g), (h), (k), and (l) RQT: tested at (a)–(d) 373 K ((a) and (c) low magnification; (b) and (d) enlarged view of (a) and (c), respectively); (e)–(l) 77 K ((e) and (g) low magnification; (f) and (h) enlarged view of (e) and (g), respectively); (i) and (k) tilted specimen; (j) and (l) enlarged view of (i) and (k), respectively). (Online version in color.)

RQ and RQ-IA mainly exhibited ductile fracture at 373 K. In the low-temperature range, the Charpy impact values of RQ and RQ-IA were significantly different from those of AsQ and Q-IA (Fig. 7). The Charpy impact values for RQ were higher than those for AsQ. As shown in Fig. 11(e), multiple sub-cracks parallel to TD (coinciding with strike direction (SD)) are generated at intervals of 100–200 μm in RQ tested at 77 K. Figures 11(i) and 11(j) show the side surface of a sub-crack observed from a different direction. Brittle, smooth fracture surfaces were observed at the sides of the sub-cracks. The main fracture surfaces perpendicular to the RD between the sub-cracks were ductile fracture surfaces characterized by dimples even at 77 K, as shown in Fig. 11(f). This unique fracture with sub-cracks in the RQ is called a separation fracture.¹⁶⁾ The separation fracture in RQ contributes to toughening by relieving triaxial stress,^17,18,19) attributed to the propagation of sub-cracks along the brittle PAGB, which extends along the direction perpendicular to the SD. In addition, the ultra-fine microstructure within PAG, resulting from the two-step γ→ε→α’ transformation, contributes to toughening.

Next, the effects of IA on the toughness and fracture behavior of RQ are discussed. As shown in Fig. 10, the Charpy impact values of RQ-IA were higher than those of RQ, indicating that IA significantly increased the Charpy impact value even at 77 K, in contrast to the behavior observed for AsQ and Q-IA. The fracture surface of RQ-IA at 77 K shows a sub-crack and a ductile fracture surface, as shown in Figs. 11(g), 11(h), 11(k), and 11(l), and this fracture surface morphology is similar to that in RQ. Figure 12 shows the crystal orientation maps of the cross-section (perpendicular to SD) of RQ and RQ-IA fractured at 77 K. The reconstructed austenite maps (c) and (f) indicate that the sub-cracks propagated along the PAGBs in RQ-IA and RQ. The lengths of the sub-cracks affect the relaxation of the triaxial stress in the separation fracture.¹⁹⁾ Hence, the average lengths of the sub-cracks in RQ and RQ-IA were measured. Figure 13 shows the SEM images of the cross-section (perpendicular to SD) of RQ and RQ-IA fractured at 77 K. Multiple sub-cracks propagate along the RD, and the measured average lengths of the sub-cracks were 1.25 ± 0.75 mm and 1.65 ± 1.19 mm in RQ and RQ-IA, respectively. The lengths of the sub-cracks tended to be slightly larger in RQ-IA than in RQ; however, the difference was too small to affect the Charpy impact values. In addition, Figs. 11(e) and 11(g) indicate that the difference between the spacing of the sub-cracks in RQ and RQ-IA was negligible. These results suggest that the increase in the Charpy impact value by IA at low temperatures is not related to the changes in the morphology of the separation fracture.

Fig. 12. Crystallographic orientation maps at cross-section of (a)–(c) RQ and (d)–(f) RQ-IA Charpy impact tested at 77 K ((b) and (e) enlarged view of the area marked by a red-frame in (a) and (d), respectively; (c) and (f) reconstructed austenite map of (b) and (e), respectively). (Online version in color.)

Fig. 13. Cross-section of (a) RQ and (b) RQ-IA Charpy impact tested at 77 K. (Online version in color.)

3.2.3. Toughening Mechanism in RQ via Intercritical Annealing

Both RQ and RQ-IA exhibited ductile fractures at high temperatures. Therefore, the increase in the Charpy impact values of RQ by IA at these temperatures can be attributed to the softening of the matrix and the contribution from retained austenite, as in the case of Q-IA described above. However, further investigation is required because the fracture mode of RQ-IA (separation fracture) differs from that of Q-IA (intergranular fracture) at low temperatures. To understand the high toughness of RQ-IA at low temperatures, the effects of microstructural changes due to IA on the toughness in the case of a separation fracture must be clarified. As previously mentioned, the effect of IA on the morphologies of the separation fracture was negligible, and ductile fractures occurred at the main fracture surface in RQ and RQ-IA. Thus, the microstructural changes caused by IA may increase the energy required for the propagation of ductile cracks.

Figure 14 shows the crystallographic orientation maps and phase maps of RQ and RQ-IA cooled to 77 K. The volume fraction of ε-martensite and retained austenite in RQ-IA is larger than that in RQ, even at 77 K (Table 2). Therefore, the plastic deformation and phase transformation of ε-martensite and retained austenite can affect the crack propagation behavior. Figure 15(a) shows an SEM image of the cross-section (perpendicular to SD) of RQ-IA fractured at 77 K. Figures 15(b)–15(e) show the crystallographic orientation maps and phase maps of the area marked by a black frame in the SEM image. The left side of the SEM image shows the main fracture surface. In the region far from the main fracture surface (Fig. 15(b)), a large amount of ε-martensite and retained austenite is observed. The amounts of ε-martensite and retained austenite in Fig. 15(b) are similar to those in Fig. 14(d). However, the proportion of these phases decreases as the observed area approaches the vicinity of the main fracture surface, and they are hardly observed near the main fracture surface (Fig. 15(e)) owing to the transformation from ε-martensite and retained austenite to α’-martensite induced by the plastic deformation around the main crack tip. The same observation was also performed for Q-IA fractured at 77 K, and the obtained SEM image, crystallographic orientation maps, and phase maps are shown in Fig. 16. The amounts of ε-martensite and retained austenite in the region close to the main fracture surface (Fig. 16(e)) are lower than those in the region away from the main fracture surface. However, a large amount of ε-martensite is observed in all regions. Figure 17 shows the relationship between the residual ratio of (ε-martensite + retained austenite) obtained from the phase maps and the distance from the fracture surface for Q-IA and RQ-IA. The residual ratio of (ε-martensite + retained austenite) is calculated by dividing the volume fraction of (ε-martensite + retained austenite) in each observed area by the volume fraction of those phases before testing (Figs. 9(d) and 14(d)). The residual fraction of (ε-martensite + retained austenite) increases with an increase in the distance from the main fracture surface and reaches a value of 1 for both Q-IA and RQ-IA. However, the distance at which the residual ratio was saturated differs for the two specimens. In Q-IA, the residual ratio increased rapidly near the fracture surface, and the distance at which the residual ratio saturated was approximately 0.35 mm. In contrast, the corresponding distance is approximately 4 mm in RQ-IA, suggesting that the region where the transformation of ε-martensite and retained austenite into α’-martensite is larger in RQ-IA than in Q-IA. Ductile fracture occurred at the main fracture surface in RQ-IA, whereas intergranular fracture occurred in Q-IA; therefore, the plastic zone at the crack tip in RQ-IA was considered larger than that in Q-IA. In addition, the formation of sub-cracks in RQ-IA causes the plane stress condition where no stress exists in the thickness direction, and the plastic region at the crack tip was expanded.¹⁸⁾

Fig. 14. Microstructures of (a) and (b) RQ; (c) and (d) RQ-IA after cooling at 77 K: ((a) and (c) crystallographic orientation map and (b) and (d) phase map). (Online version in color.)

Fig. 15. Distribution of each phase in RQ-IA Charpy impact tested at 77 K ((a) SEM image of the cross-section, (b)–(e) crystal orientation maps and phase maps of the area marked by a black-frame in (a)). (Online version in color.)

Fig. 16. Distribution of each phase in Q-IA Charpy impact tested at 77 K ((a) SEM image of the cross-section, (b)–(e) crystal orientation maps and phase maps of the area marked by a black frame in (a)). (Online version in color.)

Fig. 17. Relationship between residual ratio of (γ + ε) and distance from fracture surface of Q-IA and RQ-IA. (Online version in color.)

The impact absorption energy of RQ-IA at 77 K includes the energy required for slip deformation of each phase and α’-martensitic transformation of ε-martensite and retained austenite,³⁵⁾ the energy consumed for heat generation, and the surface energy from the newborn surface. Quantifying all these energies is challenging; however, the importance of strain-induced transformations in impact absorption energy has been reported in previous studies. For example, Zou et al.¹⁰⁾ investigated the effect of the strain-induced transformation of retained austenite on crack initiation and propagation using two Fe-5%Mn-0.01%C alloys with different retained austenite stabilities. The results revealed that the higher the amount of γ→α’ transformation, the higher the energy required for crack initiation and propagation, possibly owing to the relaxation of the stress concentration at the crack tip, which increases the impact absorption value. Tomota et al.³³⁾ reported that the ε→α’ transformation during tensile deformation reduced the work hardening rate due to local stress relaxation in high Mn steels with 16–20%Mn. Therefore, the strain-induced transformation of ε-martensite and retained austenite in RQ-IA contributes to the toughening by energy absorption owing to the transformation and stress relaxation around the crack tip.

Here, the effect of the IA on toughness was amplified by separation fractures at low temperatures. Separation fracture is a unique phenomenon in this alloy, and microstructural control to induce this fracture mode is critical for improving its low-temperature toughness.

4. Conclusions

The effects of two-step hot rolling and intercritical annealing on the microstructure and toughness of Fe-10 mass%Mn-0.1 mass%C alloys can be summarized as follows:

(1) Following the intercritical annealing of AsQ (as-quenched specimen) with equiaxed PAGs and RQ (two-step hot-rolled specimen) with elongated PAGs at 873 K, the morphologies of the PAGs remained largely unaffected. However, a noticeable reduction in hardness and an increase in the volume fraction of retained austenite were observed for both samples. The increase in the retained austenite volume fraction is attributed to the enhanced stability of austenite, facilitated by the concentrations of C and Mn in the retained austenite and the dislocations induced by the two-step hot-rolling process before quenching.

(2) Ductile fractures were observed at 373 K, whereas at 77 K, AsQ and Q-IA (intercritically annealed AsQ) displayed intergranular fractures. Although intercritical annealing enhanced the Charpy impact values for ductile fractures (at temperatures above 273 K), it had minimal influence on the values for intergranular fractures. The matrix softening and prominent presence of ε-martensite and retained austenite due to intercritical annealing predominantly contribute to toughness in the case of ductile fracture. Conversely, their effects are negligible in cases of intergranular fractures.

(3) In RQ and RQ-IA (intercritically annealed RQ), toughening from intercritical annealing was evident at all test temperatures. For both RQ and RQ-IA specimens, ductile fracture was observed at 373 K. However, at 77 K, both specimens displayed a separation fracture and ductile characteristics.

Statement for Conflict of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment(s)

The neutron experiments were performed at the Materials and Life Science Experimental Facility of Japan Proton Accelerator Research Complex (J-PARC) using a user program [Proposal no. 2021BM0008]. We thank Dr. Yusuke Onuki (Tokyo Denki University) for his technical assistance and valuable advice regarding the neutron experiments. This study was supported by the JST and the establishment of university fellowships for the creation of science and technology innovation [grant number JPMJFS2132].

References

• 1)   S.-H.  Kim,  C.-Y.  Kang and  K.-S.  Bang: J. Mater. Sci., 36 (2001), 1197. https://doi.org/10.1023/A:1004890027527
• 2)   B.  Fultz,  J.I.  Kim,  Y.H.  Kim,  H.J.  Kim,  G.O.  Fior and  J.W.  Morris, Jr.: Metall. Trans. A., 16 (1985), 2237. https://doi.org/10.1007/BF02670423
• 3)   S.  Hany,  B.  Duponchel,  C.  Poupin,  A.A.  Kais,  D.  Dewael,  J.B.  Vogt,  J.  Bouquerel,  H.  Kacem,  A.  Mouftiez,  S.  Hariri,  M.  Milochova,  E.  Bychkov and  E.A.  Aad: Adv. Mater. Res., 936 (2014), 1953. https://doi.org/10.4028/www.scientific.net/AMR.936.1953
• 4)   K.  Maeno,  M.  Tanaka,  N.  Yoshimura,  H.  Shirahata,  K.  Ushioda and  K.  Higashida: Tetsu-to-Hagané, 98 (2012), 667 (in Japanese). https://doi.org/10.2355/tetsutohagane.98.667
• 5)   C.K.  Syn,  B.  Fultz and  J.W.  Morris: Metall. Trans. A, 9 (1978), 1635. https://doi.org/10.1007/BF02661946
• 6)   Z.  Liao,  Y.  Dong,  Y.  Du,  X.  Wang,  M.  Qi,  H.  Wu,  X.  Gao and  L.  Du: J. Mater. Res. Technol., 23 (2023), 1471. https://doi.org/10.1016/j.jmrt.2023.01.058
• 7)   J.  Han,  A.K.  da Silva,  D.  Ponge,  D.  Raabe,  S.-M.  Lee,  Y.-K.  Lee,  S.-I.  Lee and  B.  Hwang: Acta Mater., 122 (2017), 199. https://doi.org/10.1016/j.actamat.2016.09.048
• 8)   S.K.  Hwang and  J.W.  Morris, Jr.: Metall. Trans. A, 10 (1979), 545. https://doi.org/10.1007/BF02658317
• 9)   M.  Kuzmina,  D.  Ponge and  D.  Raabe: Acta Mater., 86 (2015), 182. https://doi.org/10.1016/j.actamat.2014.12.021
• 10)   Y.  Zou,  Y.B.  Xu,  Z.P.  Hu,  X.L.  Gu,  F.  Peng,  X.D.  Tan,  S.Q.  Chen,  D.T.  Han,  R.D.K.  Misra and  G.D.  Wang: Mater. Sci. Eng. A, 675 (2016), 153. https://doi.org/10.1016/j.msea.2016.07.104
• 11)   K.H.  Kwon,  I.-C.  Yi,  Y.  Ha,  K.-K.  Um,  J.-K.  Choi,  K.  Hono,  K.  Oh-ishi and  N.J.  Kim: Scr. Mater., 69 (2013), 420. https://doi.org/10.1016/j.scriptamat.2013.05.042
• 12)   K.  Yamanaka and  M.  Kowaka: J. Japan Inst. Metals., 43 (1979), 1151 (in Japanese). https://doi.org/10.2320/jinstmet1952.43.12_1151
• 13)   M.  Murakami,  K.  Shibata and  T.  Fujita: Tetsu-to-Hagané, 72 (1986), 241 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.72.2_241
• 14)   S.-H.  Shin,  Y.-C.  Yoon,  S.-I.  Lee and  B.  Hwang: J. Mater. Res. Technol., 26 (2023), 3558. https://doi.org/10.1016/j.jmrt.2023.08.154
• 15)   J.  Hu,  L.-X.  Du,  W.  Xu,  J.-H.  Zhai,  Y.  Dong,  Y.-J.  Liu and  R.D.K.  Misra: Mater. Charact., 136 (2018), 20. https://doi.org/10.1016/j.matchar.2017.11.058
• 16)   T.  Maeda,  S.  Okuhata,  K.  Matsuda,  T.  Masumura,  T.  Tsuchiyama,  H.  Shirahata,  Y.  Kawamoto,  M.  Fujioka and  R.  Uemori: Mater. Sci. Eng. A, 812 (2021), 141058. https://doi.org/10.1016/j.msea.2021.141058
• 17)   Y.  Kimura,  T.  Inoue,  F.  Yin and  K.  Tsuzaki: Science, 320 (2008), 1057. https://doi.org/10.1126/science.1156084
• 18)   L.  Liu,  Q.  Yu,  Z.  Wang,  J.  Ell,  M.X.  Huang and  R.O.  Ritchie: Science, 368 (2020), 1347. https://doi.org/10.1126/science.aba9413
• 19)   T.  Inoue,  F.  Yin,  Y.  Kimura,  K.  Tsuzaki and  S.  Ochiai: Metall. Mater. Trans. A, 41 (2010), 341. https://doi.org/10.1007/s11661-009-0093-x
• 20)   Y.  Onuki,  T.  Hirano,  A.  Hoshikawa,  S.  Sato and  T.  Tomida: Metall. Mater. Trans. A, 50 (2019), 4977. https://doi.org/10.1007/s11661-019-05415-6
• 21)   A.I.  Saville,  A.  Creuziger,  E.B.  Mitchell,  S.C.  Vogel,  J.T.  Benzing,  J.  Klemm-Toole,  K.D.  Clarke and  A.J.  Clarke: Integr. Mater. Manuf. Innov., 10 (2021), 461. https://doi.org/10.1007/s40192-021-00224-5
• 22)   G.  Miyamoto,  N.  Iwata,  N.  Takayama and  T.  Furuhara: Acta Mater., 58 (2010), 6393. https://doi.org/10.1016/j.actamat.2010.08.001
• 23)   K.  Matsuda,  T.  Masumura,  T.  Tsuchiyama,  Y.  Onuki,  M.  Takanashi,  T.  Maeda,  Y.  Kawamoto,  H.  Shirahata and  R.  Uemori: ISIJ Int., 64 (2023), 486. https://doi.org/10.2355/isijinternational.ISIJINT-2023-053
• 24)   K.  Tsuzaki,  S.  Fukasaku,  Y.  Tomota and  T.  Maki: Mater. Trans. JIM, 32 (1991), 222. https://doi.org/10.2320/matertrans1989.32.222
• 25)   S.  Inthidech,  K.  Boonmak,  P.  Sricharoenchai,  N.  Sasaguri and  Y.  Matsubara: Mater. Trans., 51 (2010), 1264. https://doi.org/10.2320/matertrans.M2010018
• 26)   B.A.  Akhavan Tabatabae,  F.  Ashrafizadeh and  A.M.  Hassanli: ISIJ Int., 51 (2011), 471. https://doi.org/10.2355/isijinternational.51.471
• 27)   S.-W.  Seo,  G.-S.  Jung,  J.S.  Lee,  C.M.  Bae,  H.K.D.H.  Bhadeshia and  D.-W.  Suh: Mater. Sci. Technol., 31 (2015), 436. https://doi.org/10.1179/1743284714Y.0000000574
• 28)   K.  Kaneko,  T.  Maeda,  Y.  Kawahara,  K.  Ichino,  T.  Masumura,  T.  Tsuchiyama,  H.  Shirahata and  R.  Uemori: Scr. Mater., 193 (2021), 112. https://doi.org/10.1016/j.scriptamat.2020.10.044
• 29)   S.  Matsuda and  Y.  Okamura: Trans. Iron Steel Inst. Jpn., 14 (1974), 363. https://doi.org/10.2355/isijinternational1966.14.363
• 30)   L.  Liu,  Z.-G.  Yang,  C.  Zhang: J. Alloys Compd., 577 (2013), S654. https://doi.org/10.1016/j.jallcom.2012.04.021
• 31)   S.B.  Zhou,  F.  Hu,  W.  Zhou,  L.  Cheng,  C.Y.  Hu and  K.M.  Wu: J. Mater. Res. Technol., 14 (2021), 1021. https://doi.org/10.1016/j.jmrt.2021.07.011
• 32)   K.  Sugimoto,  H.  Tanino and  J.  Kobayashi: Steel Res. Int., 86 (2015), 1151. https://doi.org/10.1002/srin.201400585
• 33)   Y.  Tomota,  M.  Strum and  J.W.  Morris, Jr.: Metall. Trans. A, 17 (1986), 537. https://doi.org/10.1007/BF02643961
• 34)   S.  Yano,  H.  Sakurai,  H.  Mimura,  N.  Wakita,  T.  Ozawa and  K.  Aoki: Tetsu-to-Hagané, 59 (1973), 752 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.59.6_752
• 35)   J.R.  Patel and  M.  Cohen: Acta Metall., 1 (1953), 531. https://doi.org/10.1016/0001-6160(53)90083-2