Comparison of Variant Selection between Lenticular and Lath Martensite Transformed from Deformed Austenite

Abstract

A variant selection of ausformed lath martensite in an Fe–9Mn alloy (M_S = 602 K) and ausformed lenticular martensite in an Fe–25Ni–0.5C alloy (M_S = 220 K) are compared quantitatively using EBSD measurements combined with a method for the reconstruction of austenite orientation. Differences were found in the selected variants by ausforming between those two martensite morphologies. In ausformed lath martensite, variants with habit planes parallel to the active slip planes in austenite were preferentially selected. This variant selection in lath martensite may be attributed to the preferential nucleation on the microband structures in austenite and lack of growth inhibition from the microband boundaries. On the other hand, in ausformed lenticular martensite, variants with habit planes parallel to the compression plane were preferentially selected. The reason for this is that lenticular martensite whose habit plane is parallel to the compression plane can grow easily in austenite grains elongated along the compression plane. This means fewer disturbances for those variants to grow with a habit plane parallel to the compression plane than for other variants.

1. Introduction

Structural steel requires high strength and toughness. Because it provides better mechanical properties, martensite is increasingly the steel of choice, used in preference to ferrite and pearlite structures. Ausforming is a thermo-mechanical processes to transform work-hardened austenite (γ) into martensite. It is well-established that ausforming results in the strengthening of martensite without any deterioration in toughness. This explains the considerable attention being given to this process.¹⁾

Martensite has a specific orientation relationship close to the Kurdjumov-Sachs (K-S) relationship with respect to γ ((111)_γ//(011)_α, [ $\overline{1}01$ ]_γ//[ $\overline{1}\overline{1}1$ ]_α).²⁾ The K-S orientation relationship has an equivalent 24 variants. When martensite is transformed from work-hardened γ, the deformation texture in γ is inherited to the martensite. In addition, variant selection occurs where some specific variants are selected preferentially, leading to the development of a strong transformation texture in the martensite.³⁾ Since the transformation texture affects anisotropy, and therefore the mechanical properties, the variant selection of martensite has been investigated under various conditions.^4,5,6,7)

Abe et al.³⁾ warm-rolled Fe–25.7 mass%Ni and Fe–30.2 mass%Ni alloys (hereafter composition is noted by mass%) and investigated the texture of γ and martensite before and after subzero cooling. They found that the transformation texture differs between those alloys even though the rolling texture of γ is the same. Since the morphology of martensite changes from lath in the Fe–25.7%Ni alloy to lenticular in the Fe–30.2%Ni alloy,⁸⁾ the difference in transformation texture reported by Abe et al.³⁾ indicates that martensite morphology should affect variant selection. Details of this variant selection have yet to be investigated.

Recently, our research group developed a reconstruction program of γ orientation from the martensite orientation map obtained by electron backscatter diffraction (EBSD) analysis⁹⁾ and successfully quantified the fractions of 24 variants in ausformed lath martensite in an Fe–0.15%C–1.5%Mn–3%Ni alloy.¹⁰⁾ It was clarified that the variants formed with habit planes nearly parallel to the active slip planes in γ are dominant,¹⁰⁾ which is in good agreement with ausformed bainitic ferrite.¹¹⁾ When this method was applied to ausformed lenticular martensite in an Fe–25%Ni–0.5%C alloy,¹²⁾ a different variant selection rule appears where variants with habit planes nearly parallel to the compression plane are selected. In the present study, the variant selection of ausformed lath martensite is investigated again using an Fe–9%Mn alloy. Furthermore, the variant selection of lenticular martensite in an Fe–25%Ni–0.5%C alloy¹²⁾ is reviewed and compared with that in ausformed lath martensite to discuss the mechanism of variant selection in those two different martensite morphologies.

2. Experimental Procedure

In this study, an Fe–9%Mn alloy for lath martensite (M_S = 602 K)¹³⁾ and an Fe–25%Ni–0.5%C alloy (M_S = 220 K) for lenticular martensite were used. The specimens were homogenized at 1423 K for 345.6 ks. The Fe–9Mn specimens were austenitized at 1373 K for 0.6 ks while the Fe–25Ni–0.5C specimens were austenitized at 1323 K for 0.6 ks. After the austenitization treatment, the specimens were rapidly cooled to 973 K, compressed by either 30% or 50% at a strain rate of 1 s^–1 and immediately unloaded, followed by He gas quenching. Lenticular martensite was obtained by subzero cooling at 77 K or just below the M_S temperature. Details of sample preparations were reported in the previous paper.¹²⁾

Microstructural observation was performed by optical microscopy (OM) and EBSD. The observation area for EBSD analysis contains several γ grains in both specimens. The orientation relationship was determined first in the non-deformed specimen in accordance with the method described in a previous study.¹⁴⁾ The deviation angles between the close-packed plane and the close-packed direction for the orientation relationships determined were 1.06° and 3.12° for the Fe–25Ni–0.5C alloy, and 2.09° and 2.27° for the Fe–9Mn alloy, respectively. Then reconstruction of the γ orientation map was made and area fractions of the 24 variants were quantified for non-deformed and ausformed specimens.¹⁰⁾

3. Results

Figures 1(a) and 1(b) show the OM images of non-ausformed and 50% ausformed lath martensite in the Fe–9Mn alloy, respectively. As shown in Fig. 1(a), typical lath martensite structures containing block and packet structures were observed. In the 50%-ausformed lath martensite (Fig. 1(b)), lath blocks frequently grow along the direction inclined by 40–60° degrees from the compression axis (CA) direction. In addition, a portion of the block structure is substantially refined by ausforming. Figures 1(c) and 1(d) show the α orientation map of the 50%-ausformed lath martensite and the γ orientation map reconstructed from Fig. 1(c) by using the reconstruction program.⁹⁾ In ausformed lath martensite or upper bainite structures,^10,11) it was reported that habit planes of favorably formed variants tend to be parallel to the active slip planes of γ, i.e., {111}_γ the planes belonging to the slip systems with the highest or second highest Schmidt factors for compression. Therefore, the active slip planes in γ are calculated from reconstructed γ orientation and traces of them are shown in Figs. 1(c) and 1(d) to compare directions of martensite lath and active slip planes. Here, {111}_γ planes for the highest and the second highest Schmidt factors are indexed as (111)_γ and ( $\overline{1}11$ )_γ, respectively. It seems that lath blocks tend to grow along the active slip planes, which is in good agreement with ausformed low-carbon martensite¹⁰⁾ and upper bainite.¹¹⁾

Fig. 1.

Optical microscopy images of lath martensite in the specimens (a) without ausforming, (b) after 50%-ausforming, (c) α-orientation map of 50%-ausformed lath martensite and (d) corresponding γ-orientation map reconstructed from (c). Black/white dashed lines and white lines represent prior austenite grain boundaries (PAGB) and traces of active slip planes in γ.

Figure 2(a) shows an OM image of lenticular martensite in non-deformed Fe–25Ni–0.5C alloy cooled to the M_S temperature.¹²⁾ Typical lenticular martensite, which forms autocatalytically in a zig-zag manner, was observed. The growth of lenticular martensite stops at the γ grain boundary as indicated by the dashed lines and other martensite plates. In a 50%-ausformed specimen (Fig. 2(b)), the γ/lenticular martensite interface became irregular, and although the martensite plates tended to be fragmented, only small changes were made to the width of the large martensite plate by 50%-ausforming. Figure 2(c) shows an α orientation map of 50%-ausformed lenticular martensite after subzero cooling to 77 K. It is clear that the martensite plates tended to grow perpendicularly to the CA direction, i.e. the variants with habit planes nearly parallel to the compression plane became dominant. Figure 2(d) shows a γ orientation map reconstructed from Fig. 2(c). The γ grains were elongated parallel to the compression plane. In Fig. 2(c), there was no clear parallel relationship between the direction of the martensite plate and the active slip planes in γ.

Fig. 2.

Optical microscopy images of lenticular martensite in (a) non-ausformed specimen, (b) 50%-ausformed specimen, subzero-cooled to M_S temperature (c) α-orientation map of 50%-ausformed specimen subzero cooled to 77 K and (d) corresponding γ-orientation map reconstructed from (c). Black/white dashed lines and white lines represent prior austenite grain boundaries and traces of active slip planes in γ.

The inverse pole figures (IPF) for the CA direction in α and γ in Fig. 3 compare transformation texture in α and deformed texture in γ between ausformed lath and lenticular martensite. In 50%-ausformed lath martensite, the transformation texture with a peak around <111>_α′//CA and deformation texture with peak around <011>_γ//CA were developed, as shown in Figs. 3(a) and 3(b), respectively. The IPF for the CA direction in martensite shown in Fig. 3(c) indicates the expected orientations of the 24 variants and their variant numbers transformed from a textured γ. The 24 variants were numbered according to Morito et al.¹⁵⁾ As an orientation of γ, [ $\overline{1}$ , 10, 11]_γ that is misoriented by 4.7° from [011]_γ is selected to avoid the overlapping of variant orientations in the IPF in order to identify each variant. Figure 3(c) indicates that the 24 variants can be divided into two groups according to their orientations: one is a group of variants close to <111>_α′//CA, and the other is a variant group close to <001>_α′//CA. Hence, the strong transformation texture around < $\overline{1}11$ >_α′ indicates the variants belonging to the <111>_α group in Fig. 3(c) are favored more than the variants of the <001>_α′ group. Figures 3(d)–3(f) show the results of the same analyses for ausformed lenticular martensite. The transformation texture of ausformed lenticular martensite in Fig. 3(d) has a peak close to < $\overline{1}12$ >_α′//CA, which is slightly different from that in ausformed lath martensite, yet the deformation texture in γ is nearly the same (Figs. 3(b), 3(e)). These results are similar to that reported by Abe et al.³⁾ The small difference in the orientation relationships between lath and lenticular martensite explains the closeness of the orientations of 24 variants for lenticular martensite transformed from a textured γ shown in Fig. 3(f) to those for lath martensite. The closeness of the orientations makes it difficult to investigate the difference between the variant selection between lath and lenticular martensite on the basis of texture analysis.

Fig. 3.

Inverse pole figures of (a) α, (b) γ in 50%-ausformed lath martensite, (c) expected orientations of 24 variants transformed from $\overline{1}$ 10 11_γ. The inverse pole figures of (d), (e) and (f) correspond to 50%-ausformed lenticular martensite.

Based on the orientation data of martensite and γ at various ausforming strains, the fraction of 24 variants were quantified using a method previously used by Miyamoto et al.¹⁰⁾ The 24 variants can be classified into four variant groups according to the parallel relation in the close-packed planes (CP) ({111}_γ//{011}_α′). In this method, the γ orientation, indexed as the CA direction, is contained in a [001]_γ-[011]_γ-[ $\overline{1}11$ ]_γ standard stereographic triangle. According to this rule, the primary and secondary slip systems for compression correspond to (111) [ $\overline{1}01$ ]_γ and ( $\overline{1}11$ ) [101]_γ, respectively. Figure 4 shows the variant fractions of lath martensite in the Fe–9Mn alloy and lenticular martensite in the Fe–25Ni–0.5C alloy with and without ausforming. Variant fractions obtained from ausformed carbon martensite¹⁰⁾ are also shown in Fig. 4(d) for comparison purposes. The dashed lines represent a fraction when the 24 variants are formed equally. In the non-ausformed martensite in Figs. 4(a) and 4(e), the 24 variants are equally formed as expected. The fractions of variants belonging to CP1 and CP3 group are increased in lath martensite slightly by 30% ausforming (Fig. 4(b)). In 50%-ausformed lath martensite (Fig. 4(c)), V3, V6, V14 and V17 are preferentially formed. Those variants belong to CP1 and CP3 groups. Those tendencies are quite similar to the variant selection in ausformed martensite in Fe–3Ni–1.5Mn–0.15C alloy (Fig. 4(d)).¹⁰⁾

Fig. 4.

The variant fractions of (a) non-ausformed, (b) 30%-ausformed, (c) 50%-ausformed lath martensite and (d) 50%-ausformed Fe–3Ni–0.15C–1.5Mn(mass%) alloy¹⁰⁾ and (e) non-ausformed, (f) 30%-ausformed and (g) 50%-ausformed lenticular martensite after subzero cooling to 77 K. Dashed lines represent a fraction when the 24 variants are formed equivalently.

On the other hand, in ausformed lenticular martensite, V1, V6, V16 and V17 are preferentially formed by compression of 30% (Fig. 4(f)) or 50% (Fig. 4(g)).

Figures 5(a) and 5(b) show a (001)_γ standard stereographic projection with the habit plane normal of 24 variants for lath martensite {5 7 5}_γ and lenticular {3 15 10}_γ, respectively. The habit planes of the observed variants in ausformed lath martensite (V3, V6, V14 and V17) are nearly parallel to the primary (111)_γ and secondary ( $\overline{1}11$ )_γ slip planes. In contrast, the habit planes of the observed variants in lenticular martensite (V1, V6, V16 and V17) are nearly parallel to the compression plane.

Fig. 5.

A 001_γ standard stereographic projection showing the (a) {5 5 7}_γ habit plane normal and (b) {3 15 10}_γ habit plane for the 24 variants. The [001]_γ-[011]_γ-[ $\overline{1}11$ ]_γ triangle highlighted in gray shows the location of the CA direction. The solid circles indicate the habit plane of variants formed preferentially.

4. Discussion

The quantification of variant fractions in ausformed martensite in this study clearly indicates a difference between the variant selection of ausformed lath and lenticular martensite. This difference leads to the difference in the transformation texture because the orientations of V1, V6, V16 and V17 in lenticular martensite (Fig. 3(f)) are closer to ( $\overline{1}12$ )_α than those for V3, V6, V14 and V17 in lath martensite (Fig. 3(c)).

Miyamoto et al.¹⁰⁾ reported that the variants of lath martensite with habit planes parallel to the active slip planes are selected by ausforming in the Fe–0.15%C–1.5Mn–3Ni alloy. They supposed that microband structures are formed in γ along active slip planes by uniaxial compression and that martensite variants with habit planes nearly parallel to the active slip plane in γ nucleate preferentially on the microband boundaries. They also suggested that those variants can grow more easily than the other variants if the microband boundaries act as obstacles for growth.¹⁰⁾ A similar variant selection rule observed in lath martensite in the Fe–9%Mn alloy indicates that the microband structures in γ may also be a reason for the variant selection of lath martensite in this alloy. Since the variant fractions are quite similar between Fe–9Mn and Fe–3Ni–1.5Mn–0.15C alloys, the composition dependence of variant selection in ausformed lath martensite is not large.

In the Fe–25%Ni–0.5%C alloy, the deformation structure in γ can be observed directly since the M_S temperature is lower than the ambient temperature. Thus, the deformation structure was investigated using an electron channeling contrast imaging (ECCI) technique incorporate with and EBSD analysis on the same field of view. The ECCI is a scanning electron microscopy (SEM) imaging technique with sufficient sensitivity to clearly visualize the local crystallographic orientation and the dislocations near the specimen surface.^16,17) Figure 6(a) shows the ECCI micrograph of 50%-deformed austenite in the Fe–25Ni–0.5C alloy. In most of the γ grains, the microband structures are formed along either the primary or secondary slip planes in γ for uniaxial compression. Figure 6(b) shows an ECCI micrograph of the same specimen after subzero cooling to just below the M_S temperature. The EBSD analysis clarified that lenticular martensite was formed in the center of this field of view where martensite/γ interfaces are described by the white dashed lines. Lenticular martensite grew across the microbands and the microband structures were clearly inherited by martensite. This fact strongly indicates that the microband structures in γ do not interrupt the growth of lenticular martensite, as can be seen in Fig. 7(b). This behavior is in contrast to the case of ausformed lath martensite (Fig. 7(a)). As seen in the Fig. 2, the growth of lenticular martensite is stopped at γ grain boundary and at other martensite plates. Therefore, when lenticular martensite is formed in deformed γ, which is elongated along the compression plane, the lack of inhibition from the γ grain boundary allows variants with habit planes parallel to the compression plane to grow more easily than other variants. This also indicates that lenticular martensite interacts more weakly with the deformation structure in γ than lath martensite, as supposed previously.¹⁸⁾ Since the habit plane of V17 is the closest to the compression plane among selected V1, V6, V16 and V17, as shown in Fig. 5(b), the highest fraction of V17, shown in Fig. 4(g), also supports the preferential growth model explained above. According to Bokros and Parker¹⁹⁾ and Okamoto et al.,²⁰⁾ V1, V6, V16 and V17 form together in a variant group even in non-ausformed plate-type martensite due to self-accommodation. Therefore, it is also possible that V17 is formed first, with V1, V6 and V16 then induced by self-accommodation.

Fig. 6.

ECCI micrographs of the Fe–25Ni–0.5C alloy after 50% ausforming (a) before subzero cooling to M_S temperature and (b) after subzero cooling to M_S temperature Black lines and white lines represent traces of active slip planes and M/γ interfaces, respectively.

Fig. 7.

Schematic illustrations of variant selection model of ausformed (a) lath and (b) lenticular martensite.

Besides the deformation structure and γ grain boundary, the residual stress that can be assumed to exist in γ after deformation is capable of causing variant selection. According to Patel and Cohen,¹⁹⁾ stress can induce the formation of the specific variant if work done by the stress against the shape deformation of the variant is large. When the direction of the residual stress is assumed parallel to the compression axis, resolved stresses on the habit planes of V1, V6, V16 and V17 can be expected to be smaller than the other variants because their habit planes are almost parallel to the compression plane. Therefore, it can be concluded that stress-induced martensite transformation by residual stress cannot explain variant selection in ausformed lenticular martensite.

5. Conclusions

The variant selection of ausformed lath martensite in Fe–9Mn alloy and lenticular martensite in Fe–25Ni–0.5C alloy was compared quantitatively. The following results were obtained.

(1) In ausformed lath martensite, variants with habit planes parallel to the active slip planes are selected preferentially. This is possibly due to preferential nucleation on microband structures in austenite and low levels of growth inhibition from the microband boundaries. Those selected variants are quite similar to those in ausformed Fe–3Ni–1.5Mn–0.1C alloy¹⁰⁾ even though the alloy compositions differ considerably, indicating that variant selection for lath martensite is not dependent on composition.

(2) In ausformed lenticular martensite, variants with habit planes parallel to the compression plane are selected preferentially. SEM/ECCI observations reveal that lenticular martensite can grow by crossing microband boundaries in austenite, indicating the interaction between lenticular martensite and the microbands is not significant. The variant selection of ausformed lenticular martensite is explained by their preferential growth along elongated austenite grains since they encounter few disturbances at the austenite grain boundaries.

(3) The difference in the variant selection can be attributed to the difference in the interaction between martensite and the deformation structure in austenite between lath and lenticular martensite.

Acknowledgment

The authors gratefully acknowledge financial support provided by the Ministry of Education, Culture, Sports, Science and Technology through a Grant-in-Aid for Young Scientists (A) No. 23686098 (2011–2013), and the Iron and Steel Institute of Japan through a Research Promotion Grant (2012–2013).

References

• 1)   I.  Tamura: Trans. Iron Steel Inst. Jpn., 27 (1987), 763.
• 2)   G.  Kurdjumov and  G.  Sachs: Ztsch. Phy., 64 (1930), 325.
• 3)   H.  Abe,  K.  Ito and  R.  Fukumoto: J. Jap. Inst. Met., 31 (1966), 37.
• 4)   H.  Inagaki: Proc. of the 6th Int. Conf. on Textures of Materials, ISIJ, Tokyo, (1981), 149.
• 5)   R. K.  Ray and  J. J.  Jonas: Int. Mater. Rev., 35 (1990), 1.
• 6)   W. P.  Liu and  H. J.  Bunge: Mater. Lett., 10 (1991), 336.
• 7)   N. J.  Wittridge,  J. J.  Jonas and  J. H.  Root: Metall. Trans. A., 32A (2001), 889.
• 8)   M.  Umemoto,  E.  Yoshitake and  I.  Tamura: J. Mater. Sci., 18 (1983), 2893.
• 9)   G.  Miyamoto,  N.  Iwata,  N.  Takayama and  T.  Furuhara: Acta Mater., 58 (2010), 6393.
• 10)   G.  Miyamoto,  N.  Iwata,  N.  Takayama and  T.  Furuhara: Acta Mater., 60 (2012), 1139.
• 11)   W.  Gong,  Y.  Tomota,  M.  Koo and  Y.  Adachi: Scr. Mater., 63 (2010), 819.
• 12)   T.  Chiba,  G.  Miyamoto and  T.  Furuhara: Scr. Mater., 67 (2012), 324.
• 13)   K.  Tsuzaki,  S.  Fukasaku,  Y.  Tomota and  T.  Maki: Mater. Trans. JIM, 32 (1991), 222.
• 14)   G.  Miyamoto,  N.  Takayama and  T.  Furuhara: Scr. Mater., 60 (2009), 1113.
• 15)   S.  Morito,  H.  Tanaka,  R.  Konishi,  T.  Furuhara and  T.  Maki: Acta Mater., 51 (2003),1789.
• 16)   A. J.  Wilkinson and  P. B.  Hirsch: Micron, 28 (1997), 279.
• 17)   I.  Gutierrez-Urrutia,  S.  Zaefferer and  D.  Raabe: Scr. Mater., 61 (2009), 737.
• 18)   J. R.  Strife,  M. J.  Carr and  G. S.  Ansell: Metall. Trans. A, 8A (1977), 1471.
• 19)   J. R.  Patel and  M.  Cohen: Acta Metall., 1 (1953), 531.
• 20)   H.  Okamoto,  M.  Oka and  I.  Tamura: Trans. JIM, 19 (1978), 674.