Effects of Manganese on Microstructure and Work-hardening Behavior of Low-carbon Lath Martensitic Steel

Abstract

Microstructures of lath martensite have been studied intensively to understand their effect on the mechanical properties of steels. It is, however, said that the relation between microstructural factors and mechanical properties has not been clarified yet. The plastic deformation behavior of fully lath martensitic steels has become important because they are applied to automobile body structures such as bumper reinforcement. It is, therefore, important to understand the microstructural factors that control the work-hardening behavior of fully martensitic steels. Although we could not clarify differences in microstructural factors when manganese (Mn) concentrations of steels are altered, the work-hardening of 8 mass%Mn martensitic steel is much higher than that of 5 mass%Mn martensitic steel. It was found using the digital image correlation (DIC) method, that the strain concentration due to the in-lath-plane slip deformation is more developed in 5 mass%Mn martensitic steel than 8 mass%Mn martensitic steel. Transmission electron microscope (TEM) observations revealed the existence of two types of fine twins inside laths. Long twins that are parallel to the longitude of the lath are observed both in 5 mass%Mn and 8 mass%Mn martensitic steels. Short twins that partially cross the laths, on the other hand, can only be found in 8 mass%Mn martensitic steel. Since twin boundaries are high angle boundaries, the short twins are supposed to prevent the development of in-lath-plane slip deformation. This seems to be the mechanism of higher work-hardening behavior observed in 8 mass%Mn martensitic steel.

1. Introduction

Martensitic steel is the hardest steel with the same chemical composition.¹⁾ It is reported that ultra-low-carbon and low-carbon lath martensitic steels exhibit excellent strength and elongation ballance.²⁾ The high strength of lath martensitic steels is attributed to carbon atomic supersaturation of the solute into ferrous body-centered tetragonal or cubic structures, high-density dislocations introduced by martensitic transformation, crystal grain refinement by hierarchical microstructure dividing a prior austenite grain, and carbide precipitation.^1,3) However, the microstructural factors that determine elongation, particularly work-hardening during plastic deformation, of martensitic steels, are still unclear.

Tomota et al.²⁾ indicated that plastic working of martensitic steels is possible by reducing their carbon content of as-quenched low-carbon martensitic steels. Although carbon makes martensitic steels extremely strong, it reduces their elongation, toughness, and weldability. The addition of large amounts of carbon is undesirable. Thus, additions of other alloying elements are considered. Manganese can change the strength and elongation of martensitic steels. It is one of the least expensive substitutional alloying elements.⁴⁾ In this study, the authors discuss the relationship between microstructure and work hardening rate of several low carbon martensitic steels that have been altered in manganese concentration and quenched at different austenitizing temperatures.

Regarding the microstructure of low-carbon lath martensitic steels, Maki et al.⁵⁾ observed the microstructures of lath martensitic steels (Fe - 0.1 to 0.8 mass% (hereafter noted as%) C and Fe - 0.2%C to 1%X (X: Mn, Ni, Si, Cr, Mo)) in detail through an optical microscope. The martensitic steels containing 0.1 to 0.6% carbon had only lath martensite, and a martensitic steel containing 0.8% carbon had lath martensite and lenticular martensite. They also reported that martensitic packets and blocks could be clearly observed in martensitic steels containing 0.1 to 0.2% carbon, but their boundaries became unclear with increasing carbon concentration. Fujita et al.^6,7) investigated the prior austenite grain size of ultra-low-carbon steels containing 0.063 to 13.7% manganese based on JIS G 0551 (1956). They reported that manganese slightly decreased the prior austenite grain size, and this effect was saturated at approximately 1.5% manganese. Morito et al.⁸⁾ evaluated the sizes of individual hierarchical microstructure in ultra-low-carbon lath martensitic steels containing 0.02 to 3.14% manganese. Manganese reduces the size of the prior austenite grain by half and sizes of the packets, blocks, and laths by 10–30%. In addition, Morito et al.⁹⁾ studied the morphology and crystal orientation relationship of martensite in 0.0026 to 0.61%C martensitic steels. It was found that carbon reduces the size of the packets and blocks with increasing carbon concentration. The observed crystal orientation relationship between the prior austenite and martensite was close to the Kurdjumov - Sacks relationship. The Nishiyama - Wassermann relationship was also observed in some laths. They indicated the existence of sub-blocks in a block. The sub-blocks have a crystal misorientation of approximately 10° with another sub-block in a block.

Regarding the strength, especially yield strength of lath martensitic steels, it is known that the strength of lath martensitic steels increases with the decreasing size of the prior austenite grains.¹⁰⁾ However, Maki and Tamura¹¹⁾ argued that blocks are the fundamental unit in the Hall–Petch relationship based on crystallography. Packets determine the mechanical properties of low-carbon martensitic steels because their blocks are divided. Furthermore, the sizes of the packets and blocks were affected by the size of the prior austenite grains. Therefore, the yield strength can be expressed in terms of sizes of prior austenite grains, packets, or blocks.^1,8) The tensile strength of lath martensitic steels is 1.5 times more than the tensile strength of ferritic steels with the same chemical composition. However, their yield strengths are similar. Takaki et al.¹²⁾ formulated a relationship between the 0.2% proof strength and dislocation density of ultra-low-carbon steels containing nickel. They also concluded that the yield strength of the lath martensitic steels is low due to the high density of dislocations produced during martensitic transformation. Iwamura et al.¹³⁾ investigated the true and apparent dislocation density of 0.006 to 0.26%C as-quenched martensitic steels considering the c/a of the body-centered tetragonal structure by X-ray diffraction (XRD). The true dislocation density does not depend strongly on the carbon concentration. It saturates at 4.5×10¹⁵ m⁻² above 0.14%C.

Experiments combining electron backscatter diffraction (EBSD) and digital image correlation (DIC) have been reported. These studies have elucidated the influence of microstructures on the plastic deformation of lath martensitic steels. Morsdorf et al.¹⁴⁾ compared the microstructure of a low carbon steel containing nickel before and after deformation and analyzed the strain by DIC. According to their report, strain concentration occurs along the longitude of the block and tilts by 45° against the macroscopic tensile direction at the early stage of deformation. Subsequently, the strains expanded to other blocks. The retained austenite thin films sandwiched between the laths were transformed into martensite before the average strain reached 0.04. Therefore, the influence of the retained austenite on the overall plastic deformation behavior is limited. Ishimoto et al.¹⁵⁾ observed the microscopic strain distribution up to the average strain of 0.1 using multi-layered steels with three layers constructed from two austenitic stainless steels and a low-carbon martensitic steel containing manganese, chromium, and molybdenum. During the early deformation stage (average strain < 0.02), the strain was uniform. Subsequently, the strain was concentrated in particular blocks with increasing the average strain. The slip systems were classified as in-lath and out-of-lath plane slips based on an analysis of the crystal orientation. The in-lath plane slip was observed in blocks with high Schmid factors and consequently, the strain was concentrated within the block. Ryou et al.¹⁶⁾ studied low- and medium-carbon martensitic steels. In narrow-width blocks, the habit plane slip was more active than in-lath plane slip with high Schmid factor. Ungár et al.¹⁷⁾ observed the deformation of low-carbon steel containing manganese using high-resolution neutron diffraction. According to their observations, the growth direction of the lath relative to the tensile direction determines the difficulty of deformation. In each packet, the habit planes of the laths included were identical. Therefore, the authors concluded that the deformation difficulties of each packet differed. They named the hard to be deformed and easy to be deformed packets as “hard orientation (HO)” and “soft orientation (SO),” respectively. In HO, the out-of-lath plane slips were activated, and the characteristic of the dislocation changed from screw to edge with increasing dislocation density by deformation. Conversely, in the SO, the in-lath-plane slips were activated, and the characteristic of the dislocation changed from edge to screw with decreasing dislocation density by deformation. Sakaguchi et al.¹⁸⁾ introduced the size of the microstructure in the calculation of the true stress-true strain curves based on the continuous composite approach (CCA) model. They estimated the average distance of dislocation movement from the size of the microstructure of a low-carbon martensitic steel containing nickel and the size of the strain distribution observed by DIC with the iso-work assumption.

Regarding the work-hardening of lath martensitic steels, Niino et al.¹⁹⁾ confirmed the increase in the work-hardening rate and the increasing rate of the dislocation density with carbon content of steels conducting tensile tests of multi-layered steels with three layers constructed from two austenitic stainless steels and low-carbon (0.001 to 0.32% carbon) martensitic steels. They attempted to explain the increase in the work-hardening rate using the decline in the rate of decrease in mobile dislocation density. They also indicated that the suppression of the construction of dislocation cells by solute carbon appears to promote an increase in the work-hardening rate. Harjo et al.^20,21) claimed that HO is important for the work-hardening of lath martensitic steels based on a study by Ungár et al.¹⁷⁾ Regarding the effect of manganese addition, Hanamura et al.¹⁰⁾ reported a large elongation of low-carbon steel containing 5% manganese. Maeda et al.²²⁾ revealed that manganese addition increases both the tensile strength and uniform elongation. The cause was explained the difference of the work hardening by increasing the dislocation density in the high-strain region, because generations of the dislocation cells were suppressed, by in-situ observation using synchrotron.

Arlazarov et al.²³⁾ investigated the effect of manganese on true stress-true strain carves comparing medium-carbon steels and low-carbon steels containing 5% manganese, and simulated the true stress-true strain curves based on the CCA model.

The effects of alloying elements on the microstructure and strength, particularly the yield strength, of low-carbon lath martensitic steels have been discussed. It has been established that manganese additions increase the work-hardening rate, therefore, both the elongation and tensile strength. This phenomenon has been discussed based on the movement of dislocations and heterogeneity of the microstructure. Nevertheless, the relationship between the effects of manganese concentrations more than 5% on work-hardening and the hierarchical microstructure of martensitic steels remains uncertain. In this study, the authors investigate the microstructure that dominates the work hardening of lath martensitic steels containing manganese to ascertain the relationship between the microstructure and microscopic heterogeneity of deformation using tensile test with EBSD and DIC.

2. Experimental Procedure

The samples were low-carbon steels containing 3, 5, and 8% manganese, as listed in Table 1. The samples were hot- and cold-rolled after being cast in a laboratory. The plates were austenitized at 950, 1050, and 1100°C in argon gas and quenched in water. Each sample is referred as C_MnMn steel or C_MnMn-TQ steel from the manganese concentration C_Mn% and the austenitizing temperature T°C. After quenching, the samples were processed into tensile test specimens, as shown in Fig. 1.

Table 1. Chemical compositions of the initial specimens.

Specimen name	C	Si	Mn	P	S	Fe
	mass%	mass%	mass%	mass%	mass%
3Mn steel	0.10	<0.01	3.0	<0.003	0.001	Bal.
5Mn steel	0.09	<0.01	5.0	<0.002	0.002	Bal.
8Mn steel	0.07	<0.01	8.1	<0.002	0.003	Bal.

Fig. 1. Dimensions of macroscopic and microscopic DIC tensile test specimens.

The rolled surfaces of the tensile test specimens were ground and polished with waterproof sandpaper, aluminum abrasive grains, and colloidal silica grains for observation and analysis using scanning electron microscopy (SEM, FEI Scios), EBSD (Oxford Instrument Symmetry S3), and XRD (Rigaku SmartLab).

TEM samples were prepared from the grip part of the tensile test specimens using two methods for observation and analysis through transmission electron microscopy (TEM, JEOL JEM-ARM-200F) and scanning precession electron diffraction (SPED) (NanoMegas ASTAR). In the first method, a TEM sample (<100 nm thickness) was prepared using a focused ion beam (SEM-FIB) (FEI Versa 3D) from the grip part of the tensile test specimens ground with waterproof sandpaper. In the second method, the sample was prepared using an electrolytic polishing instrument (Struers TenuPol-5) after punching 3 mm diameter disc from the grip part of the tensile test specimens ground to a thickness of 60–80 μm with waterproof sandpaper. A mixture of ethanol and perchloric acid was used as the electrolyte.

Tensile tests were performed using an Instron-type tensile test machine (Shimadzu AG-10TA). The load was measured using the load cell of the tensile test machine and the strain was measured using digital image correlation (DIC) method. The movement of the random pattern (spray paint) was photographed using a digital camera (AnMo Electronics Dino-LitePremier500M) and the images were analyzed using the DIC software (Correlated Solutions VIC-2D). The analysis conditions were as follows: image resolution, 2592 × 1944 pixels; subset size, 29 pixels; step size, 7 pixels. One pixel corresponds to 5.7 μm. This DIC method is called macroscale DIC.

To clarify the relationship between the microstructure and plastic deformation, the before and after deformations of the 5Mn-1050Q and 8Mn-1050Q steels were observed using SEM (Zeiss Ultra55). Tensile tests were applied using an Instron-type tensile test machine (Shimadzu AG-10TA). The random pattern used for DIC was silver nanoparticles. This DIC method is called microscale DIC. Both ends of the tensile test specimen were fixed with a jig during transfer from the tensile test machine to the SEM instrument not to change the strain. Therefore, the total strain, including the elastic strain, could be observed without stress relaxation. Secondary electron images were analyzed using the DIC software (Correlated Solutions VIC-2D) to calculate strain. The analysis conditions had an image resolution of 3072 × 2304 pixels, subset size of 41 pixels, and step size of 14 pixels. One pixel corresponds to 37 nm. In addition, the crystal orientation of the same view as that of the DIC was measured using an EBSD instrument (TSL DVC5).

3. Results

3.1. True Stress-true Strain Curves

Figure 2 shows the true stress-true strain curves of the 5Mn-950Q, 5Mn-1050Q, and 5Mn-1100Q steels austenitized at different temperatures, and Fig. 3 shows the true stress-true strain curves of the 3Mn-1050Q, 5Mn-1050Q, and 8Mn-1050Q steels austenitized at 1050°C. The influence of the austenitizing temperature on the true stress-true strain curves was less prominent than that of the chemical composition. This trend was also observed for the 3Mn and 8Mn steels. The work-hardening rate of the 8Mn-1050Q steel was greater than those of the 3Mn-1050Q and 5Mn-1050Q steels. Figure 4 shows the effect of the manganese content on the yield strength, tensile strength, and magnitude of work-hardening of the 3Mn-1050Q, 5Mn-1050Q, and 8Mn-1050Q steels. The yield strength is defined as the elastic limit. Tensile strength is defined as the stress at which the true stress, σ_t, coincides with the work-hardening rate (dσ_t)/(dε_t). The magnitude of work-hardening is defined as the difference between the yield and tensile strengths. The change in yield strength with manganese concentration was comparatively small. The tensile strength markedly increased with increasing manganese concentration. And magnitude of the work-hardening increased with increasing manganese concentration. The changes in the tensile strength and magnitude of work-hardening from 5%Mn steel to 8%Mn steel were distinctive.

Fig. 2. True stress-true strain curves of Fe-0.1C-5Mn steels as-quenched at 950°C, 1050°C, and 1100°C. (Online version in color.)

Fig. 3. True stress-true strain curves of Fe-0.1C-3Mn, Fe-0.1C-5Mn, and Fe-0.1C-8Mn steels as-quenched at 1050°C. (Online version in color.)

Fig. 4. Effect of Mn concentration on the yield strength, tensile strength, and work-hardening of Fe-0.1C steels as-quenched at 1050°C. The yield strength is defined as the stress of the true strain-true stress curve deviating from a straight line with a slope of the Young’s modulus. The tensile strength is defined as the intersection of the true strain-true stress curve and work-hardening rate. Work hardening is defined as the difference between the tensile strength and yield strength. (Online version in color.)

3.2. Microstructures

Figure 5 shows the crystal orientation maps measured by the EBSD analysis of the 5Mn-950Q, 5Mn-1050Q, 5Mn-1100Q, 8Mn-950Q, 8Mn-1050Q, and 8Mn-1100Q steels. The microstructures of all specimens were deemed typical lath martensite by observation. Furthermore, ε martensite, carbide, and retained austenite were not observed in the SEM and EBSD analyses. Figure 6(a) shows the XRD patterns of the 5Mn-1050Q and 8Mn-1050Q steels before and after deformation and Fig. 6(b) presents an enlarged view of the 8Mn-1050Q steel before deformation within the 50≤2θ≤60 range. A 2 0 0 peak of retained austenite was observed only in the 8Mn-1050Q steel before deformation. No other peaks were observed for the retained austenite. Consequently, a quantitative analysis is difficult. Figure 7 shows that the average grain size of the prior austenite estimated by intercept method. The average grain size of the prior austenite increased with the increasing austenitizing temperature. This trend is consistent with that reported by Fujita et al.^6,7) They reported that the refining effect on prior austenite of manganese saturated above 1.5% manganese. For simplicity, the average width of blocks was estimated by measuring appearance width of many blocks, rather than calculating true width based on appearance width and crystal direction. Figure 8 shows the average block widths. The changes in block width with austenitizing temperature and manganese concentration were small.

Fig. 5. Crystal orientation maps measured by the EBSD analysis obtained for Fe-0.1C-5Mn steel as-quenched at 950°C (a), 1050°C (b), and 1100°C (c), and Fe-0.1C-8Mn steel as-quenched at 950°C (d), 1050°C (e), and 1100°C (f).

Fig. 6. X-Ray diffraction patterns of Fe-0.1C-5Mn and Fe-0.1C-8Mn steels as-quenched at 1050°C. (a) shows a comparison between the before and after deformation for each steel. (b) shows an enlarged view of Fe-0.1C-8Mn steel as-quenched at 1050°C before deformation.

Fig. 7. Average grain sizes of prior austenite as-quenched Fe-0.1C-5Mn and Fe-0.1C-8Mn steels. Error bars represent the standard deviation.

Fig. 8. Average appearance widths of the martensitic blocks of as-quenched Fe-0.1C-5Mn and Fe-0.1C-8Mn steels. Error bars represent the standard deviation.

3.3. Microscale DIC

Figure 9 shows the crystal orientation maps measured by the EBSD and the strain distributions obtained by the DIC. Figures 9(a) and 9(b) are maps of 5Mn-1050Q (a) steel, and Figs. 9(c) and 9(d) are maps of 8Mn-1050Q steel. The crystal orientation maps in Figs. 9(a) and 9(c) and the strain distributions in Figs. 9(b) and 9(d) are from the same view field, respectively. The macroscale tensile direction was parallel to the x-axis (horizontal direction of paper). The average strains of the 5Mn-1050Q steel were 0.014 and 0.017, as measured by the macroscale and microscale DICs, respectively. The average strains of the 8Mn-1050Q steel were 0.015 and 0.020, as measured by macroscale and microscale DICs, respectively. The differences between the nominal stress and true stress and between the nominal strain and true strain were considered small because the strains were small. It is considered that all the specimens underwent uniform macroscopic deformation based on the true stress-true strain curves shown in Fig. 3. In both specimens, the strain concentrations (local strains) parallel to the blocks represented by the H ellipses in Fig. 9 were observed. However, the local strains in the 8Mn-1050Q steel were not more clearly recognized than those in the 5Mn-1050Q steel. The strain distribution in the 8Mn-1050Q steel was more uniform than that in the 5Mn-1050Q steel.

Fig. 9. Crystal orientation maps measured by EBSD analysis obtained for Fe-0.1C-5Mn steel (a) and Fe-0.1C-8Mn steel (c) as-quenched at 1050°C before tensile deformation, and corresponding engineering total strain distributions parallel to tensile direction measured by DIC analysis obtained at an average microscopic engineering strain of 0.017 (b) for Fe-0.1C-5Mn steel as-quenched at 1050°C, and at an average microscopic engineering strain of 0.020 (d) for Fe-0.1C-8Mn steel as-quenched at 1050°C.

4. Discussion

The effect of manganese on the work-hardening behavior of lath martensitic steels is discussed. In this study, the magnitude of work-hardening was defined as the difference between the yield and tensile strengths. First, the effects of solute strengthening are evaluated. The amount of change in the tensile strength Δσ_n,TS MPa is defined by Eq. (1) based on the carbon effect on tensile strength reported by Takaki et al.¹²⁾ and the manganese effect on friction stress reported by Allain et al.²⁴⁾ Using true strain ε_t,TS corresponding to the tensile stress on true stress obtained by experiment, Eq. (1) is rewritten as Eq. (2).

  

$$\mathrm{\Delta }{\sigma }_{n,TS}=2\mathrm{\hspace{0.17em} }560\times C_C{}^{\frac{2}{3}}+33\times C_{Mn}$$(1)

  

$$\begin{array}{l}\mathrm{\Delta }{\sigma }_{t,TS}=\mathrm{\Delta }{\sigma }_{n,TS}\cdot exp\left({\epsilon }_{t,TS}\right)\\ \hspace{1em}\hspace{1em}\hspace{1em}=\left(2\mathrm{\hspace{0.17em} }560\times C_C{}^{\frac{2}{3}}+33\times C_{Mn}\right)exp\left({\epsilon }_{t,TS}\right)\end{array}$$(2)

Here C_C and C_Mn are the carbon and manganese concentrations (mass%), respectively. Equation (3) was derived from Eq. (2) when 3Mn steel was selected as the standard.

  

$$\mathrm{\Delta }{\sigma }_{t,TS}^{\prime }=\left(2\mathrm{\hspace{0.17em} }560\times C_C{}^{\frac{2}{3}}+33\times C_{Mn}\right)exp\left({\epsilon }_{t,TS}\right)-678$$(3)

The calculated values and results of the experiments of the Δσ_n,TS for the 3Mn, 5Mn and, 8Mn steels are listed in Table 2. The experimental values are the results for 3Mn-1050Q, 5Mn-1050Q, and 8Mn-1050Q steels. The calculated and experimentally observed values for the 5Mn steel were similar. In contrast, the 8Mn steel exhibited a large difference between the experimental value and the calculated value. Therefore, the increase in tensile strength from 5Mn steel to 8Mn steel could not be explained using solute strengthening with carbon and manganese.

Table 2. Comparison of the tensile strength (true stress) changes in the calculation and experiment.

Specimen name	Tensile strength change (Calculation) $\mathrm{\Delta }{\sigma }_{t,TS}^{\prime }$	Specimen name	Tensile strength change (Experiment)
3Mn steel	Reference	3Mn-1050Q steel	Reference
5Mn steel	+31	5Mn-1050Q steel	+35
8Mn steel	+98	8Mn-1050Q steel	+390

On the other hand, the yield strength or 0.2% proof stress of the martensitic steels containing retained austenite has been reported to be lower, and their work-hardening rate and elongation are greater than those of the fully martensitic steels because the retained austenite transformed to the martensite by deformation. Lath martensitic steels contain a small volume percent of very hard, film-like retained austenite. This retained austenite affects the macroscopic mechanical properties at an early stage of deformation.^{4,16,25,26,27)} The 8Mn-1050Q steel contained a small amount of retained austenite, as shown in Fig. 6. However, it is considered that this retained austenite does not affect the plastic deformation because it is transformed into martensite in the early stage of deformation, as reported by Morsdorf et al.¹⁴⁾

Therefore, investigating the elementary steps of deformation is necessary to discuss the effect of manganese on work-hardening behavior. The authors analyzed the relationship between the Schmid factor and local strains parallel to the macroscopic tensile direction of 5Mn-1050Q and 8Mn-1050Q steels. Here, the stress was assumed to be uniaxial in microscopic and the slip systems {0 1 1}⟨1 1 1⟩ and {1 1 2}⟨1 1 1⟩ were selected. The maximum value of the Schmid factor, characteristics (slip system, in-lath plane slip or out-of-lath plane slip), and local strain measured by microscale DIC of the blocks are listed in Table 3. Habit plane slip, a type of in-lath plane slip, was also considered. The habit plane was assumed to correspond to the {0 1 1} planes. The local strain was expressed as the total strain.

Table 3. Relationship between the Schmid factor and engineering total strain parallel to the tensile direction of Fe-0.1C-5Mn steel as-quenched at 1050°C at an average microscopic engineering strain of 0.017 and Fe-0.1C-8Mn steel as-quenched at 1050°C at an average microscopic engineering strain of 0.020.

Specimen name	Block	Maximum Schmid factor	Character of maximum Schmid factor	Total engineering strain parallel to the tensile direction
5Mn-1050Q steel	1	0.341	$(1\mathrm{\hspace{0.17em} }\overline{2}\mathrm{\hspace{0.17em} }1)[\overline{1}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ In-lath	0.009
	2	0.469	$(0\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1})[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1]$ In-lath (Habit)	0.018
	3	0.486	$(\overline{2}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1)[1\overline{1}\mathrm{\hspace{0.17em} }1]$ Out-of-lath	0.009
	4	0.472	$(1\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ In-lath (Habit)	0.021
	5	0.480	$(\overline{1}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{2})[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.012
	6	0.466	$(\overline{1}\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1]$ In-lath (Habit)	0.022
	7	0.480	$(\overline{2}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1]$ Out-of-lath	0.017
	8	0.470	$(1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }0)[1\mathrm{\hspace{0.17em} }\overline{1}1]$ In-lath (Habit)	0.028
	9	0.473	$(\overline{2}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }1]$ Out-of-lath	0.014
	10	0.365	$(1\mathrm{\hspace{0.17em} }\overline{2}\mathrm{\hspace{0.17em} }1)[\overline{1}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ In-lath	0.016
8Mn-1050Q steel	1	0.491	$(1\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.028
	2	0.491	$(1\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.020
	3	0.477	$(1\mathrm{\hspace{0.17em} }\overline{2}\mathrm{\hspace{0.17em} }1)[\overline{1}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ In-lath	0.027
	4	0.491	$(1\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.018
	5	0.479	$(0\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1})[\overline{1}\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.016
	6	0.479	$(0\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1})[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.024
	7	0.491	$(1\mathrm{\hspace{0.17em} }0\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.030
	8	0.492	$(0\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1})[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.015
	9	0.475	$(\overline{1}\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }1)[1\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.015
	10	0.484	$(0\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\overline{1})[1\mathrm{\hspace{0.17em} }\overline{1}\mathrm{\hspace{0.17em} }\overline{1}]$ Out-of-lath	0.027

In blocks 4, 6, and 8 of 5Mn-1050Q steel, the Schmid factors of the habit plane slip systems {0 1 1}⟨1 1 1⟩ were maximum in the slip systems, and local strains that were approximately 1.5 times the average strain parallel to the tensile direction in the view were observed. In block 3 of 8Mn-1050Q steel, the Schmid factor of the in-lath plane slip systems {1 1 2}⟨1 1 1⟩ was maximum in the slip systems and local strains that were approximately 1.5 times the average strain parallel to the tensile direction in the view were observed. In blocks 1, 2, and 10 of the 5Mn-1050Q steel, where the Schmid factors for the in-lath plane slip were the maximum in the slip systems, the local strains were smaller than the average strain parallel to the tensile direction. Therefore, strain concentration does not always occur in the block where the Schmid factor of the in-lath plane is the largest in the slip systems, regardless of the slip systems.

The local strains in blocks 3, 5, 7, and 9 of 5Mn-1050Q were equal to or half the average strain parallel to the tensile direction. The Schmid factors of these blocks were maximized in the out-of-lath plane. The local strains in blocks 2, 4, 5, 8, and 9 of 8Mn-1050Q which Schmid factors were maximized in the out-of-lath plane, were equal to or half the average strain parallel to the tensile direction, regardless of the slip systems. However, there was no extremely small local strain in the 8Mn-1050Q steel compared with that in the 5Mn-1050Q steel. In the 8Mn-1050Q steel, local strains that were over the average strain parallel to the tensile direction were observed in blocks 1, 6, 7, and 10. The maximum values of the local strain were approximately twice the average strain parallel to the tensile direction. The Schmid factors of these blocks were maximum in the out-of-lath slip {0 1 1}⟨1 1 1⟩. Therefore, in 8Mn-1050Q steel, large local strains also occurred in the blocks where the Schmid factor of the out-of-lath plane was high.

Comparing the local strains between each block, in the 5Mn-1050Q steel, the blocks exhibited a large local strain, and the blocks with a small local strain existed alternately. Consequently, the heterogeneity of the local strain was high. The change in the local strain of the 8Mn-1050Q steel was relatively gradual and uniform compared with the 5Mn-1050Q steel.

Figure 10 shows the histograms of the local strains of the 5Mn-1050Q and 8Mn-1050Q steels. The average strains parallel to the tensile direction in the view measured by microscale DIC were 0.017 and 0.020 in the 5Mn-1050Q and 8Mn-1050Q steels, respectively. The deviation of the strain distribution of the 8Mn-1050Q steel was smaller than that of the 5Mn-1050Q steel. In the 5Mn-1050Q steel, the local strains are concentrated in the blocks where the Schmid factor of the in-lath plane is high. It is difficult for the local strains to concentrate in blocks where the Schmid factor of the out-of-lath plane is high. Therefore, the deviation in the local strain is large. In contrast, in the 8Mn-1050Q steel, strains occurred in blocks where the Schmid factor of the in-lath plane was high, as well as in blocks where the Schmid factor of the out-of-lath plane was high. Therefore, the deformation of 8Mn-1050Q steel was relatively uniform. As a result, it is considered that the deviation in the local strain was small. Harjo et al.^20,21) defined HO which is not easy to deform and SO which is easy to deform based on the relationship between the tensile direction. They considered that HO was more closely related to the work-hardening of lath martensitic steels. In 8Mn steels, deformation occurs in HO, which requires a large stress to deform, that is, in the out-of-lath plane. As a result, it can be concluded that the work-hardening rates of the 8Mn steels were higher than those of the 5Mn steels. The authors thought that there was an unknown microstructure factor that prevented only in-lath plane slips and promoted out-of-lath slips in 8Mn steels. As observed through SEM and EBSD, the size of the prior austenite, width of the block, and retained austenite did not clearly change with manganese concentration. Therefore, the authors observed the interiors of the laths using TEM and SPED. Figure 11 shows the TEM bright field images of the 5Mn-1050Q and 8Mn-1050Q steels sampled by FIB. Fine twins in laths were observed in both samples. Fine twins in laths were first observed by Das and Thomas²⁸⁾ in the 1960s. These amounts increase with increasing concentrations of carbon, manganese, and chromium.¹¹⁾ Ping et al.^29,30) and Liu et al.³¹⁾ observed fine twins in ultra-low-carbon steel and low-carbon steel using TEM. They distinguished the fine twins to short twin dividing the longitudinal direction of the lath and the long twin dividing the transverse direction of the lath. Sugiyama et al.³²⁾ observed the fine twins in the low-carbon steel containing 1% manganese by ultra-high voltage electron microscopy, SEM, and EBSD. They indicated that SEM and EBSD observation of fine twins in laths was possible. In addition, they conducted a 2-dimensional analysis of fine twins using FIB. However, the relationship between the mechanical properties and fine twins was not clear. In our observation, the long twins indicated by arrows A and B in Fig. 11 were observed in both 5Mn steel and 8Mn steels. The short twins indicated by arrow C were observed only in the 8Mn steel. TEM samples of the 5Mn-950Q and 8Mn-950Q steels prepared by electrolytic polishing were also observed to confirm that the FIB did not affect the generation of fine twins. Similar to the observation of the FIB samples, long twins were observed in the 5Mn-950Q and 8Mn-950Q steels, and short twins were observed only in the 8Mn-950Q steel. The crystal directions of the fine twins in 8Mn steel were measured using SPED. Figure 12 shows the crystal orientation maps of the short twins in the 8Mn-1050Q steel. The twin boundary between the lath, as the parent phase, and the twin was a high-angle grain boundary. The twin boundaries were considered to prevent the dislocation movement. Therefore, high stress was needed for deformation. And the out-of-lath plane slip, which was generally difficult to active, became active. As a result, it is considered that the work-hardening rate increased. From these observations, it appears that the large increase in the tensile strength and work-hardening rate from 5% to 8% manganese concentration was caused by the fine twins in the lath. In future, it will be necessary to clarify the mechanism of fine twin generation and the interaction between fine twins and slip deformation.

Fig. 10. Comparison of engineering total strain distributions parallel to tensile direction measured by DIC analysis obtained at an average microscopic engineering strain of 0.017 for Fe-0.1C-5Mn steel as-quenched at 1050°C, and at an average microscopic engineering strain of 0.020 for Fe-0.1C-8Mn steel as-quenched at 1050°C. (Online version in color.)

Fig. 11. Bright field TEM images Fe-0.1C-5Mn (a) and Fe-0.1C-8Mn (b) steels as-quenched at 1050°C. The arrows A in (a) and B in (b) indicate long twins. Arrow C in (b) indicates short twins.

Fig. 12. Bright field image (a) obtained by TEM and corresponding crystal orientation map (b) measured by SPED analysis obtained for Fe-0.1C-8Mn steel as-quenched at 1050°C. The arrow A in (b) indicates a short twin. The blue lines in (b) represent the twin boundaries.

5. Conclusions

The following conclusions were drawn from the tensile tests, observations of the microstructure, and the strain distribution with the DIC method for low-carbon lath martensitic steels containing manganese austenitized at various temperatures and quenched.

(1) The effect of the austenitizing temperature on the true stress-true strain curve is smaller than the effect of the composition changes in the steel. The work-hardening rate of lath martensitic steels containing 8% manganese was greater than those containing 3% and 5% manganese.

(2) The microscopic strain concentration was caused by in-lath plane slip near the block boundaries in lath martensitic steels containing 5% manganese at an average strain of 0.02. However, such strain concentration was not evident in lath martensitic steels containing 8% manganese.

(3) The short twins dividing the longitude of the lath were only observed inside laths of lath martensitic steels containing 8% manganese.

(4) Short twins appear to inhibit the dislocation motion, prevent the development of in-lath plane slip, and promote out-of-lath plane slip, which requires a higher stress. Consequently, short twins may increase the work-hardening rate.

Acknowledgments

This research was conducted performed in collaboration with Nippon Steel Corporation and supported by JST SPRING, Grant Number JPMJSP2136. The authors used the equipment of the Ultramicroscopy Research Center, Kyushu University. The authors would like to thank Prof. S. Hata and Mr. Y. Zhao for their help with the TEM observations.

References

• 1)   G.  Krauss: Mater. Sci. Eng. A, 273–275 (1999), 40. https://doi.org/10.1016/S0921-5093(99)00288-9
• 2)   Y.  Tomota,  M.  Aoyama,  M.  Takenaka,  H.  Kuratomi and  I.  Tanimoto: Trans. Jpn. Soc. Mech. Eng. A, 62 (1996), 1605 (in Japanese). https://doi.org/10.1299/kikaia.62.1605
• 3)   G. R.  Speich: Trans. Metall. Soc. AIME, 245 (1969), 2553.
• 4)   K.  Ushioda,  M.  Yoshimura,  H.  Kaidoh and  K.  Kimura: Tetsu-to-Hagané, 100 (2014), 716 (in Japanese). https://doi.org/10.2355/tetsutohagane.100.716
• 5)   T.  Maki,  K.  Tsuzaki and  I.  Tamura: Tetsu-to-Hagané, 65 (1979), 515 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.65.5_515
• 6)   K.  Fujita,  T.  Mori,  M.  Ohnishi,  S.  Hashimoto and  M.  Yada: Tetsu-to-Hagané, 59 (1973), 764 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.59.6_764
• 7)   T.  Tsumura,  Y.  Kamada,  S.  Tanoue and  H.  Ohtani: Tetsu-to-Hagané, 70 (1984), 1993 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.70.15_1993
• 8)   S.  Morito,  X.  Huang,  T.  Furuhara,  T.  Maki and  N.  Hansen: Acta Mater., 54 (2006), 5323. https://doi.org/10.1016/j.actamat.2006.07.009
• 9)   S.  Morito,  H.  Tanaka,  R.  Konishi,  T.  Furuhara and  T.  Maki: Acta Mater., 51 (2003), 1789. https://doi.org/10.1016/S1359-6454(02)00577-3
• 10)   T.  Hanamura,  S.  Torizuka,  S.  Tamura,  S.  Enokida and  H.  Takechi: ISIJ Int., 53 (2013), 2218. https://doi.org/10.2355/isijinternational.53.2218
• 11)   T.  Maki and  I.  Tamura: Tetsu-to-Hagané, 67 (1981), 852 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.67.7_852
• 12)   S.  Takaki,  K.-L.  Ngo-Huynh,  N.  Nakada and  T.  Tsuchiyama: ISIJ Int., 52 (2012), 710. https://doi.org/10.2355/isijinternational.52.710
• 13)   M.  Iwamura,  M.  Tsukahara,  O.  Idohara,  Y.  Misaka and  S.  Takaki: Tetsu-to-Hagané, 107 (2021), 853 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2021-068
• 14)   L.  Morsdorf,  O.  Jeannin,  D.  Barbier,  M.  Mitsuhara,  D.  Raabe and  C. C.  Tasan: Acta Mater., 121 (2016), 202. https://doi.org/10.1016/j.actamat.2016.09.006
• 15)   Y.  Ishimoto,  M.  Michiuchi,  S.  Nambu,  K.  Asakura,  J.  Inoue and  T.  Koseki: J. Jpn. Inst. Met., 73 (2009), 720. https://doi.org/10.2320/jinstmet.73.720
• 16)   K. H.  Ryou,  S.  Nambu and  T.  Koseki: Mater. Sci. Eng. A, 777 (2020), 139090. https://doi.org/10.1016/j.msea.2020.139090
• 17)   T.  Ungár,  S.  Harjo,  T.  Kawasaki,  Y.  Tomota,  G.  Ribárik and  Z.  Shi: Metall. Mater. Trans. A, 48 (2017), 159. https://doi.org/10.1007/s11661-016-3845-4
• 18)   K.  Sakaguchi,  S.  Yamasaki,  H.  Kawata,  K.  Hayashi and  M.  Takahashi: ISIJ Int., 62 (2022), 2008. https://doi.org/10.2355/isijinternational.ISIJINT-2022-114
• 19)   T.  Niino,  J.  Inoue,  M.  Ojima,  S.  Nambu and  T.  Koseki: Tetsu-to-Hagané, 106 (2020), 488 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2019-109
• 20)   S.  Harjo,  T.  Kawasaki,  Y.  Tomota,  W.  Gong,  K.  Aizawa,  G.  Tichy,  Z.  Shi and  T.  Ungár: Metall. Mater. Trans. A, 48 (2017), 4080. https://doi.org/10.1007/s11661-017-4172-0
• 21)   S.  Harjo,  W.  Gong,  T.  Kawasaki,  S.  Morooka and  T.  Yamashita: ISIJ Int., 62 (2022), 1990. https://doi.org/10.2355/isijinternational.ISIJINT-2022-207
• 22)   A.  Maeda,  S.  Torizuka and  H.  Adachi: Trans. Soc. Automot. Eng. Jpn., 49 (2018), 856 (in Japanese). https://doi.org/10.11351/jsaeronbun.49.856
• 23)   A.  Arlazarov,  O.  Bouaziz,  A.  Hazotte,  M.  Gouné and  S.  Allain: ISIJ Int., 53 (2013), 1076. https://doi.org/10.2355/isijinternational.53.1076
• 24)   S.  Allain,  O.  Bouaziz and  M.  Takahashi: ISIJ Int., 52 (2012), 717. https://doi.org/10.2355/isijinternational.52.717
• 25)   S.  Morito,  K.  Oh-ishi,  K.  Hono and  T.  Ohba: ISIJ Int., 51 (2011), 1200. https://doi.org/10.2355/isijinternational.51.1200
• 26)   S.  Morito,  T.  Ohba,  A. K.  Das,  T.  Hayashi and  M.  Yoshida: Tetsu-to-Hagané, 98 (2012), 245 (in Japanese). https://doi.org/10.2355/tetsutohagane.98.245
• 27)   H.  Natsumeda,  A.  Kitahara and  S.  Hashimoto: Tetsu-to-Hagané, 102 (2016), 525 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2015-104
• 28)   S. K.  Das and  G.  Thomas: Metall. Trans., 1 (1970), 325. https://doi.org/10.1007/BF02819291
• 29)   D.  Ping,  T.  Liu,  M.  Ohnuma,  T.  Ohmura,  T.  Abe and  H.  Onodera: ISIJ Int., 57 (2017), 1233. https://doi.org/10.2355/isijinternational.ISIJINT-2016-624
• 30)   D. H.  Ping,  S. Q.  Guo,  M.  Imura,  X.  Liu,  T.  Ohmura,  M.  Ohnuma,  X.  Liu,  T.  Abe and  H.  Onodera: Sci. Rep., 8 (2018), 14264. https://doi.org/10.1038/s41598-018-32679-6
• 31)   X.  Liu,  D. H.  Ping,  H. P.  Xiang,  X.  Lu and  J.  Shen: J. Appl. Phys., 123 (2018), 205111. https://doi.org/10.1063/1.5034146
• 32)   M.  Sugiyama,  M.  Takei,  S.  Sekida and  N.  Maruyama: IOP Conf. Ser.–Mater. Sci. Eng., 1249 (2022), 012020. https://doi.org/10.1088/1757-899X/1249/1/012020