- Email: [email protected]
Novel ferrite–austenite duplex lightweight steel with 77% ductility by transformation induced plasticity and twinning induced plasticity mechanisms
Available online at www.sciencedirect.com
ScienceDirect Acta Materialia 78 (2014) 181–189 www.elsevier.com/locate/actamat
Novel ferrite–austenite duplex lightweight steel with 77% ductility by transformation induced plasticity and twinning induced plasticity mechanisms Seok Su Sohn a, Kayoung Choi b, Jai-Hyun Kwak c, Nack J. Kim b, Sunghak Lee a,⇑ a
Center for Advanced Aerospace Materials, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea Graduate Institute of Ferrous Technology, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea c Sheet Products & Process Research Group, Technical Research Laboratories, POSCO, Kwangyang 545-090, Republic of Korea b
Received 11 March 2014; accepted 24 June 2014 Available online 19 July 2014
Abstract The need for lightweight materials has been an important issue in automotive industries to reduce greenhouse gas emission and to improve fuel efficiency. In addition, automotive steels require an excellent combination of strength and ductility to sustain automotive structures and to achieve complex shapes, but the traditional approach to obtain a reduction in weight from down-gauged steels with high strength has many limitations. Here, we present a new ferrite–austenite duplex lightweight steel containing a low-density element, Al; this steel exhibits tensile elongation up to 77% as well as high tensile strength (734 MPa). The enhanced properties are attributed to the simultaneous formation of deformation-induced martensites and deformation twins and the additional plasticity due to deformation twinning in austenite grains having optimal mechanical stability. The present work gives a promise for automotive applications requiring excellent properties as well as reduced specific weight. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Duplex lightweight steel; Annealing; Tensile properties; Transformation induced plasticity (TRIP); Twinning induced plasticity (TWIP)
1. Introduction Recently, there has been a focus on the need to reduce vehicle weight in order to reduce exhaust emissions and improve fuel efficiency [1,2]. The most efficient method is the use of materials lighter than conventional ones. Many efforts have been directed towards applying lightweight materials such as Al alloys or Mg alloys, despite their high costs [3,4]. However, these alloys show poor formability, which restricts the application [5,6]. The development of new advanced automotive steels, namely lightweight steels, ⇑ Corresponding author. Tel.: +82 54 279 2140; fax: +82 54 279 2399.
E-mail address: [email protected] (S. Lee).
is recognized as a more realistic measure [7,8]. As a part of this study, a considerable amount of Al has been added to automotive steels to obtain a lightweight material [9–11]. It has been shown that the addition of 1 wt.% Al leads to a 1.5% weight reduction in comparison with conventional steels. However, the addition of Al reduces the ductility, and thus the method utilizing microstructures such as a full austenitic phase strengthened with j-carbides, or a duplex phase of ferrite + austenite, has been suggested [9–19]. The full austenitic microstructure displays low work-hardening capacity, in spite of the excellent combination of strength and ductility, which results in the problem of poor formability [20]. When the austenite is utilized as a secondary phase in ferrite matrix, high-strength, high-ductility lightweight
http://dx.doi.org/10.1016/j.actamat.2014.06.059 1359-6454/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
182
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
steels having high work-hardening capacity can be developed by the austenite/martensite transformation during the deformation, namely the transformation induced plasticity (TRIP) mechanism [11–14]. According to research on an Fe–3.5Mn–5.9Al–0.4C steel [12] and an Fe–5.8Mn–3.1Al–0.12C–0.47Si steel [14], for example, the austenite was retained at room temperature when steels were annealed above 720 °C, and was transformed to a0 -martensite during the deformation, which led to excellent properties of tensile strength (990 MPa) and elongation (28%). This c/a0 transformation in (a + c) duplex structure mainly depends on the orientation, mechanical stability and stacking fault energy (SFE) of the retained austenite [11–13,21,22]. In order to achieve the more excellent combination of strength and ductility, the utilization of another powerful deformation mechanism of twinning induced plasticity (TWIP) by properly controlling the mechanical stability or SFE of austenite is essentially needed [23–27]. In a certain stability or SFE range of austenite, deformation twins prevent the movement of dislocations as they work for the grain refinement [26,27]. This variation in deformation mechanism might affect the strength and ductility by forming complex microstructures mixed with twins and martensite. Furthermore, such a microstructural evolution occurring during the final stage of deformation might play an important role in controlling deformation mechanisms. Thus, studies on deformation mechanisms are essential for the evaluation of alloy design, microstructural evolution and process control. In the present study, therefore, the (a + c) duplex lightweight steel showing the operation of both TRIP and TWIP mechanisms was developed by varying annealing conditions of an Fe–0.3C–8.5Mn–5.6Al steel, and tensile properties were evaluated. Detailed deformation mechanisms were investigated in relation to microstructural evolution by electron back-scatter diffraction (EBSD) and transmission electron microscopy (TEM) analyses, and the correlation between microstructural evolution process and tensile ductility was verified. 2. Experimental 2.1. Lightweight steels The lightweight steel used in this study was fabricated by a vacuum induction melting method, and its nominal composition is Fe–0.3C–8.5Mn–5.6Al–(<0.02)(P + S) (wt.%). After thick plates of 60 mm in thickness were homogenized at 1200 °C for 1 h, they were hot-rolled between 1100 and 900 °C. They were then cooled in a furnace from 650 °C after holding at this temperature for 1 h in order to simulate a coiling procedure. The hot-rolled steel sheets of 3 mm in thickness were rolled at room temperature to make 1 mm thick steel sheets. The sheets were annealed at 800 °C for 1 min or 30 min and at 900 °C for 30 min in
a continuous annealing simulator (model; CAS-AY-II, Ulvac-RIKO, Inc., Japan) to form a mixture of ferrite and austenite, and were cooled in the air. For convenience, the steel sheets annealed at 800 °C for 1 min, at 800 °C 30 min and at 900 °C for 30 min are referred to as “A81”, “A83” and “A93”, respectively. The effects of Al addition on weight reduction are attributed to the lattice expansion and the low atomic weight of substitutional solution [3]. The density of the present steel was measured to be 7.2 g cm3, by a densitometry (Mettler-Toledo XP205, Mettler-Toledo AG, Switzerland) on the basis of the Archimedes principle, which shows an apparent reduction of 8.5% in comparison to pure Fe. 2.2. Microstructural analysis Phases present in the specimens were identified by X-ray diffraction (XRD; Cu Ka radiation; scan rate, 2° min1; scan step size, 0.02°) and TEM. Their volume fractions were measured by the direct comparison method using XRD analysis [14]. Integrated intensities of (2 0 0)a and (2 1 1)a peaks and (2 2 0)c and (3 1 1)c peaks were used for this XRD method. For the TEM observation, specimens were mechanically polished to a thickness of 50 lm, punched to prepare disk specimens (diameter: 3 mm) by a disk cutter and then electro-polished by a twin-jet polisher (model; Tenupol-5, Struers, Denmark) in a solution of CH3COOH (90%) and HClO4 (10%) to prepare thin foil specimens. The thin foils were observed in a TEM (model 2100, JEOL, Japan) operated at an acceleration voltage of 200 kV. EBSD analysis (step size, 50 nm) was conducted by a field emission scanning electron microscope (FE-SEM, Quanta 3D FEG, FEI Company, USA). The data were then interpreted by orientation imaging microscopy (OIM) analysis software provided by TexSEM Laboratories, Inc. Electron probe microanalysis (EPMA) measurements employing wavelength-dispersive spectrometry (WDS) were also performed by an EPMA microprobe (model; JXA 8530F microprobe, JEOL, Japan) at an electron beam voltage of 15 keV. Since the precise measurement of C content was difficult by EPMA, the C content was measured by the XRD method using the following equation [28]: ac ¼ 3:578 þ 0:0330X C þ 0:0056X Al þ 0:00095X Mn ð1Þ ˚ , and XC, XMn where ac is austenite lattice parameter, in A and XAl are concentrations of C, Mn and Al, respectively, in wt.%. The austenite lattice parameter (ac) was determined from a d-spacing of (2 2 0)c position. 2.3. Tensile test Plate-type tensile specimens having gage length of 25 mm, gage width of 6 mm and gage thickness of 1 mm were prepared in the longitudinal direction. They were
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
tested at room temperature at a strain rate of 103 s1 by a universal testing machine (model; 8801, Instron, Canton, MA, USA) of 100 kN capacity. 3. Results 3.1. Microstructure EBSD phase maps of the A81, A83 and A93 steels are shown Fig. 1a–c. The annealed steels have duplex microstructures of ferrite and austenite. The overall volume fraction and grain size of austenite in the A81 steel is 36% and 0.93 lm, respectively (Fig. 1a). As the annealing time at 800 °C increases, the austenite grain size increases up to 1.56 lm, while its volume fraction remains (37%) (Fig. 1b). In the A93 steel, both the volume fraction and size of austenite grain increase to 42% and 2.82 lm, respectively (Fig. 1c). Many austenite grains contain annealing twins. 3.2. Tensile properties Fig. 2a shows room-temperature engineering stress– strain curves. As the annealing time at 800 °C increases from 1 to 30 min, the yield and tensile strengths decrease,
(a)
whereas the elongation increases. The A81 and A83 steels show a common tensile flow behavior of yielding, necking and failure. When the annealing temperature increases from 800 to 900 °C, the elongation dramatically increases up to 77%, although the yield and tensile strengths are slightly reduced. This increased elongation is quite outstanding, and has not been reported in previous studies on duplex lightweight steels. It is also noted that the tensile behavior shows the steady stress flow after the rapidly increased strain-hardening. Fig. 2b shows the true stress–strain curve, strain-hardening rate (dr/de) curve and volume fraction of austenite measured by the direct comparison XRD method, and are indicated by the black line, red dotted line and blue circular symbols, respectively, for the A93 steel. The true stress– strain curve shows a continuously increasing strain-hardening behavior, which is different from the steady stress flow behavior of the engineering stress–strain curve, and the tensile strength reaches 1280 MPa. The strain-hardening rate curve displays a multiple stage strain-hardening behavior. In the initial deformation stage at the true strain of 0–0.25, the strain-hardening rate continuously decreases with increasing plastic strain, while the austenite volume fraction is not varied much in the range of 38–42%. Beyond the strain of 0.25, the strain-hardening rate increases and
Dγ : 1.56 μ m Vγ (XRD) : 37 %
Dγ : 0.93 μ m Vγ (XRD) : 36 % A81
(b)
183
Dγ : 2.82 μ m Vγ (XRD) : 42% A83
(c)
BCC FCC
A93
TD 5 μm
5 μm
5 μm
RD
Fig. 1. EBSD phase maps of the (a) A81, (b) A83, and (c) A93 steels, showing duplex microstructures of ferrite and austenite.
184
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
(b) 2500
A83
800
TrueStress, dσ/dε (MPa)
A81 A93
600 400 - A93 Steel YS : 502 MPa TS : 734 MPa El. : 77 %
200
A93 Steel True σ - ε Strain Hardening Rate ( dσ / dε )
2000
50 40 30
1000
20 500
Stage 1
Stage 2
Stage 3
10 0 0.7
0
10
20
30
40
50
60
70
60
1500
0 0
70
0.0
80
0.1
0.2
0.3
0.4
0.5
Austenite Fraction (%)
Engineering Stress (MPa)
(a) 1000
0.6
True Strain
Engineering Strain (%)
Fig. 2. (a) Room temperature tensile engineering stress–strain curves of the A81, A83 and A93 steels, and (b) true stress–strain curve, strain-hardening rate (dr/de) curve and volume fraction of austenite, which are indicated by black line, red dotted line and blue circular symbols, respectively, for the A93 steel. Two peaks are marked by black arrows on the strain-hardening rate (dr/de) curve. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
decreases twice, indicating two up–down peaks as marked by black arrows. In the strain range of 0.25–0.45 where the strain-hardening rate shows the first peak, the austenite volume fraction rapidly drops to 22%, and then slightly decreases to 20% in the strain range of 0.45–0.55 (fracture ε = 0.15
(a)
BF
strain) where the strain-hardening rate shows the second peak. For convenience, the true strain ranges of 0–0.25, 0.25–0.45 and 0.45–0.55 are referred to as Stages 1, 2 and 3, respectively, as shown in Fig. 2b. Considering the rapid drop of austenite volume fraction, Stage 2 might be related BF
(b)
b
(-11-1) (-200)
(11-1)
(000)
0.5 μm
(c)
BF
ε = 0.35
DF
(d)
d
BF
(e)
Deformaon Twin
e
α’-Martensite
(-11-1) (-200) (11-1) (000) (1-21)
0.5 μm
0.5 μm
(f)
ε = 0.55
0.5 μm
BF
(2-11)
(000) (110)
DF
(g) Deformaon Twin
(-200) (-11-1) (000)
0.5 μm
0.5 μm
(11-1)
Fig. 3. (a, b) TEM bright field images of an austenite grain in Stage 1 (true strain: 0.15), showing the dislocation pile-up at the annealing twin boundary. (c–e) TEM bright and dark field images of an austenite grain in Stage 2 (true strain: 0.35), showing deformation twins and a0 -martensites formed in the left and right sides, respectively, of an annealing twin. (f, g) TEM bright and dark field images of an austenite grain in Stage 3 (true strain: 0.55), showing thick deformation twins.
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
to the TRIP mechanism [29]. In order to elucidate causes of extremely high elongation of 77% in the A93 steel, more detailed analyses including TEM data are essentially needed.
185
The transition in the strain-hardening behavior (Fig. 2b) is closely related with deformation mechanisms. TEM analyses were conducted in Stages 1–3, and the results are shown in Fig. 3. In Stage 1 at the true strain of 0.15, austenite grains are deformed by the dislocation slip mechanism (Fig. 3a). The inset in Fig. 3b shows the selected-area diffraction (SAD) pattern taken along the [0 1 1] zone axis of an austenite grain, which further confirms that the austenite grain contains an annealing twin. The density of dislocations is higher at the annealing twin boundary than inside (Fig. 3b). At the strain of 0.35 in Stage 2, deformation twins are observed in the left side of an annealing twin (Fig. 3c), as verified by the dark-field image and SAD pattern taken along the [0 1 1] zone axis of the austenite (Fig. 3d). In the right side of the annealing twin, a0 -martensite is found without e-martensite (Fig. 3e), as confirmed by a SAD pattern taken along the [0 1 1] zone axis of the body centered cubic (bcc) phase. It is noted from these TEM results that both deformation twins and a0 -martensite can be formed inside one austenite grain. At the strain of 0.55 in Stage 3, a number of thick deformation twins are well formed inside the whole austenite grain (Fig. 3f and g). These twins are not developed into secondary twins, unlike in conventional TWIP steels [26]. EBSD inverse pole figure maps of the same area after the interrupted tensile test at the true strains of 0.35, 0.4 and 0.45 in Stage 2 are shown in Fig. 4. At the strain of 0.35, deformation twins and a0 -martensite are independently formed inside austenite grains whose boundaries are surrounded by white dotted lines, as marked by blue and white arrows, respectively (Fig. 4a and d). Deformation twins and a0 -martensite are formed from the opposite boundary inside a grain. At the strain of 0.4, a0 -martensite initiated from the boundary are grown into the grain interior, and deformation twins are thickened or newly formed at the grain boundary (Fig. 4b and e). In the grain marked by a yellow arrow, deformation twins and a0 -martensite are independently formed from the opposite boundary. At the strain of 0.45, a number of deformation twins are developed, while a0 -martensite is grown further (Fig. 4c and f). Some a0 -martensite makes inroads into deformation twins.
martensite and deformation twins in Stage 2 and the additional plasticity due to deformation twins in Stage 3. Relatively fine austenite grains 0.9–2.8 lm in size are formed after the annealing (Fig. 1a–c). The austenite grain refinement raises the thermal stability of austenite, and reduces the martensite start temperature (Ms) below the room temperature [30]. Deformation mechanisms of the austenite existed even at room temperature are varied with mechanical stability or SFE [9,21,25]. Though the SFE is not the absolute standard for classifying deformation mechanisms, many studies on the measurement or calculation of SFE in austenitic steels have been conducted [25,31,32]. However, the experimental SFE data of the austenite in (a + c) duplex steels have been rarely reported [33] because their deformation behavior is determined by reactions with the adjacent ferrite as well as SFE. In general, deformation mechanisms related with the mechanical stability of austenite are varied mainly with the content of Mn [9]. In high-Mn (20 wt.% or higher) duplex lightweight steels, the austenite is deformed mainly by mechanisms of microbanding and wavy glide because of very high SFE [17,21,23]. In low-Mn (3–4 wt.%) (a + c) duplex lightweight steels, the only TRIP mechanism, i.e., deformation-induced martensitic transformation, is working [12,14]. Here, the optimal mechanical stability of austenite importantly works for excellent mechanical properties, and the deformation mechanism cannot be escaped from the TRIP regime even in the cases of the increased stability or SFE [34]. The Fe–12Mn–1Al–0.7C steel, one of medium-Mn austenitic steels, whose austenite grain size was 6 lm, was deformed by the only TWIP mechanism [35]. In the present Fe–8.5Mn–5.6Al–0.3C steel, one of medium-Mn (a + c) duplex steels, both the TRIP and TWIP mechanisms can work as major deformation mechanisms at an optimal mechanical stability of austenite, which has not been reported previously. The mechanical stability against the formation of strain-induced martensite and the SFE which is related with the kind of deformation mechanism are influenced by the grain size and composition of austenite. The increase in austenite grain size readily provides nucleation sites for stacking faults and shear-band intersections, and activates the formation of martensite due to the decrease in elastic strain energy required for lattice displacement and in strain energy required for shear deformation [36–38]. Effects of alloy composition and grain size on Ms can be explained by the following equations suggested by Jimenez-Melero et al. and Mahieu et al. [30,39]:
4. Discussion
M s0 ð CÞ ¼ 539 30:4Mn 7:5Si þ 30Al
3.3. Tensile deformation behavior
The most unique characteristic of the present lightweight steel is the persistent elongation up to 77% in the A93 steel, as shown in Fig. 2. This is mainly attributed to the simultaneous formation of deformation-induced
M s ð CÞ ¼ M s0 AX C BðV
1=3 Þ c
273
ð2Þ ð3Þ
where Ms0 shows the effect of alloying composition except C on Ms. XC, A, Vc and B are C content in austenite, constant (423 K wt.%1), average volume of austenite grain
186
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
ε
ε
0.35
(a)
ε
0.4 (b)
0.45 (c)
Deformaon Twin
5 μm
5 μm
5 μm
(d)
(e)
(f)
α’-Martensite
TD 5 μm
5 μm
5 μm
RD // Tensile Direcon
Fig. 4. EBSD IPF maps of (a–c) face centered cubic (fcc) and (d–f) bcc in the same area after the interrupted tensile test at the true strains of 0.35, 0.4 and 0.45 in Stage 2, showing deformation twins (blue arrow marks) and a0 -martensites (white arrow marks) independently formed inside austenite grains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 C, Mn and Al contents, austenite grain size, Ms and SFE of the A81, A83 and A93 steels. Steel
Ca (wt.%)
Mnb (wt.%)
Alb (wt.%)
Austenite grain size (lm)
Ms0c (°C)
Msc (°C)
SFEc (mJ m2)
A81 A83 A93
0.58 0.58 0.55
11.1 11.3 11.5
4.58 4.62 4.74
0.93 1.56 2.82
339 334 332
417 215 69
51.6 51.8 51.1
a b c
Content of C measured from the XRD analysis. Contents of Mn and Al measured from the EPMA analysis. Ms and SFE calculated from Eqs. (2–4).
and geometry coefficient of austenite grain (475 lm K), respectively. The SFE (C is expressed by the following equation [31]: C ¼ 2qDGc!e þ 2rc=e þ 2qDGex ; DGex ¼ 170:06 expðd=18:55Þ c!e
c/e
ð4Þ
where q, DG , r , and DGex show the molar surface density along {1 1 1} planes, free energy for c ! e phase transformation, c/e interfacial energy and excess part of the SFE due to the grain size (d) effect, respectively. The DGc!e is related to chemical compositions. According to Eq. (4), the increase in austenite grain size reduces the SFE. However, the SFE has only an indirect effect on the twinning stress, and thus a homogeneous slip length is the most relevant microstructural variable directly
influencing the twinning stress [40]. The increase in homogeneous slip length reduces the critical stress required for twin nucleation. C, Mn and Al contents, austenite grain size, Ms and SFE of the A81, A83 and A93 steels are shown in Table 1. Here, C, Mn and Al contents are measured from the XRD and EPMA analyses, and Ms and SFE are calculated from Eqs. 2–4. Contents of Mn and Al are not varied much by the annealing conditions, and hardly influence the Ms and SFE. The C content, which can affect the Ms slightly, decreases in the A93 steel having the highest volume fraction of austenite, but its variation is not large. Thus, the Ms0 are almost same in the three steels. The Ms of the A93 steel is much higher (about six times) than that of the A81 steel because the austenite grain size is considered in the calculation of Ms. This implies that the deformation
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
amount required for martensite formation is relatively small. In view of the SFE, the SFE is hardly varied in the range of 51–52 mJ m2, although the austenite grain size is considered in its calculation. This indicates that the SFE of the austenite is not a main parameter affecting deformation mechanisms in the duplex microstructure. It is also well known that the martensitic transformation rarely occurs in this range of SFE [41–43]. Rather than these parameters such as alloying contents and SFE, the decrease in critical stress required for twin nucleation due to the increase in austenite grain size can critically affect the activation of twinning. The A81 and A83 steels, whose stability of austenite is quite high because of grain refinement below 1.6 lm in size, show a common tensile flow behavior (Fig. 2a). In the A93 steel whose austenite grain size increases to 2.8 lm, on the other hand, the TRIP and TWIP mechanisms are working simultaneously beyond the true strain of 30% according to the relatively low stability of austenite, which results in the extremely high elongation of 77%. This unusual deformation behavior related with the extremely high elongation in the A93 steel is determined by the crystallographic preferential orientation of austenite grain. EBSD IPF maps of the same area after the interrupted tensile test at the true strains of 0.35 and 0.45 in Stage 2 are shown in Fig. 5. When an austenite grain deformed by the dislocation slip without deformation twins or a0 -martensite at the strain of 0.35 (marked by white dotted line in Fig. 5a) is further deformed (strain:
ε = 0.35
FCC IPF
(a)
187
0.45), deformation twins and a0 -martensite are independently formed inside the grain (Fig. 5c and d). According to the Schmid factor map of the austenite grain of Fig. 5a (Fig. 5b), the local orientation variation appears inside the grain during the strain up to 0.35. Here, the “A” region whose Schmid factor is low at 0.25–0.3 and the “B” region whose Schmid factor is high at 0.3–0.4 correspond to nucleation sites for deformation twins and a0 -martensite, respectively (Fig. 5c and d). It directly shows the orientation dependence of the TRIP and TWIP mechanisms. Considering that the martensitic transformation is triggered by the supply of strain energy as a consequence of dislocation pile-ups at strong barriers such as grain boundaries [23,39], the martensitic transformation preferentially occurs in regions having high resolved shear stresses, i.e. high Schmid factors. It is also possible to trigger deformation twins in regions having relatively low resolved shear stresses because the A93 steel has considerably higher mechanical stability or SFE than low-Mn duplex steel [9,12]. The present duplex lightweight steel (the A93 steel) shows the excellent elongation of 77% together with the high tensile strength, which are far superior to conventional lightweight steels [11–14,17,34]. The enhanced tensile elongation has been achieved by the operation of multideformation mechanisms such as (1) the formation of deformation-induced martensite together with twinning and (2) deformation twinning in austenite grains having optimal mechanical stability. The present work also shows
ε = 0.35
FCC Schmid Factor
(b) A
B B 2 μm
ε = 0.45
FCC IPF
(c)
2 μm
ε = 0.45
BCC IPF
(d)
α’ Martensite Deformaon Twin
TD
2 μm
2 μm
RD // Tensile Direcon
Fig. 5. EBSD (a) IPF map of fcc, (b) Schmid factor map of fcc, (c) IPF map of fcc, and (d) IPF map of bcc in the same area after the interrupted tensile test at the true strains of 0.35 and 0.45 in Stage 2, showing deformation twins (blue arrow marks) and a0 -martensites (white arrow marks) independently formed inside an austenite grain. In (b), the “A” region, whose Schmid factor is low at 0.25–0.3, and the “B” region, whose Schmid factor is high at 0.3– 0.4, correspond to nucleation sites for deformation twins and a0 -martensites, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
188
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
that the unique deformation behavior is dependent on the crystallographic preferential orientation. Since this steel has outstanding properties of strength and ductility as well as reduced density, it gives a promise for automotive applications requiring excellent properties. In order to further enhance microstructures and properties of duplex lightweight steels, more intensive studies to design new lightweight steels, to establish optimized annealing conditions and to clarify mechanisms involved in improved strength and ductility should be continued. 5. Conclusions In the present study, the (a + c) duplex lightweight steel showing the operation of both TRIP and TWIP mechanisms was developed by varying annealing conditions of an Fe–0.3C–8.5Mn–5.6Al (wt.%) steel, and tensile properties were analyzed in relation with microstructural evolution and deformation mechanisms. (1) The annealed steels showed duplex microstructures of ferrite and austenite. As the annealing time at 800 °C increased, the austenite grain size increased up to 1.56 lm, while its volume fraction remained constant (37%). In the A93 steel (annealed at 900 °C for 30 min), both the volume fraction and size of austenite grain increased to 42% and 2.82 lm, respectively. (2) The A81 and A83 steels (annealed at 800 °C for 1 min and 30 min, respectively) showed a common tensile flow behavior of yielding, necking and failure with elongation of 35%. In the A93 steel, however, the elongation dramatically increased up to 77%, and the tensile behavior showed the steady stress flow after the rapidly increased strain-hardening. The strain-hardening rate curve of the A93 steel displayed a multiple stage strain-hardening behavior. In the initial deformation stage, the strain-hardening rate continuously decreased with increasing plastic strain. Beyond the strain of 0.25, the strain-hardening rate increased and decreased twice, indicating two up– down peaks. (3) The detailed multi-deformation mechanisms of the A93 steel were analyzed by TEM and EBSD in each deformation stage. In Stage 1, austenite grains were deformed by the dislocation slip. In Stage 2, deformation twins and a0 -martensites were independently formed inside one austenite grain. Finally in Stage 3, a number of thick deformation twins were well formed inside the whole austenite grain. (4) The enhanced tensile elongation and operation of multi-deformation mechanisms were attributed to the optimal mechanical stability of austenite. The C, Mn and Al contents were not varied much with annealing conditions, and had little effect on mechanical stability. The austenite grain size acted as a factor determining the activation of martensitic transformation and twinning. When the stability of austenite
was quite high because of the grain refinement below 1.6 lm, like in the A81 or A83 steel, the tensile flow showed the common behavior. In the A93 steel whose austenite grain size increased to 2.82 lm, however, the TRIP and TWIP mechanisms were working simultaneously beyond the true strain of 30% according to the relatively low stability of austenite. (5) This unusual deformation behavior related to the extremely high elongation in the A93 steel was determined by the crystallographic preferential orientation of austenite grain. The martensitic transformation preferentially occurred in regions having high resolved shear stresses, i.e., high Schmid factors.
Acknowledgment This work was supported by the Ministry of Knowledge Economy under a Grant No. 10031723-2011-21. References [1] Mayyasa A, Qattawia A, Omara M, Shana D. Renew Sustainable Energy Rev 2012;16:1845. [2] Kuziak R, Kawalla R, Waengler S. Arch Civ Mech Eng 2008;8:103. [3] Kulekci MK. Int J Adv Manuf Technol 2008;39:851. [4] Miller WS, Zhuang L, Bottema J, Wittebrood AJ, De Smet P, Haszler A, et al. Mater Sci Eng 2000;A280:37. [5] Hirsch J. Acta Mater 2013;61:818. [6] Lang L, Danckert J, Nielsen KB. J Mater Process Technol 2004;148:119. [7] Matlock DK, Speer JG. In: Haldar A, Suwas S, Bhattacharjee SD, editors. Microstructure and texture in steels. London: Springer; 2009. [8] Ma M, Yi H. In: Weng Y, Dong H, Gan Y, editors. Advanced steels. Heidelberg: Springer; 2011. [9] Frommeyer G, Bru¨x U. Steel Res Int 2006;77:627. [10] Rana R, Liu C, Ray RK. Scr Mater 2013;68:354. [11] Hwang SW, Ji JH, Lee EG, Park KT. Mater Sci Eng 2011;A528:5196. [12] Seo CH, Kwon KH, Choi K, Kim KH, Kwak JH, Lee S, et al. Scr Mater 2012;66:519. [13] Park SJ, Hwang B, Lee KH, Lee TH, Suh DW, Han HN. Scr Mater 2012;68:365. [14] Suh DW, Park SJ, Lee TH, Oh CS, Kim SJ. Metall Mater Trans A 2010;41A:397. [15] Frommeyer G, Drewes EJ, Engl B. Rev Metall Cah Inf Technol 2000;97:1245. [16] Falat L, Schneider A, Sauthoff G, Frommeyer G. Intermetallics 2005;13:1256. [17] Sutou Y, Kamiya N, Umino R, Ohnuma I, Ishida K. ISIJ Int 2010;50:893. [18] Choi K, Seo C-H, Lee H, Kim SK, Kwak JH, Chin KG, et al. Scr Mater 2010;63:1028. [19] Park KT, Jin KG, Han SH, Hwang SW, Choi K, Lee CS. Mater Sci Eng 2010;A527:3651. [20] Chin KG, Kang CY, Shin SY, Hong S, Lee S, Kim HS, et al. Mater Sci Eng 2011;A528:2922. [21] Gra¨ssel O, Kru¨ger L, Frommeyer G, Meyer LW. Int J Plast 2000;16:1391. [22] Ryu JH, Kim DI, Kim HS, Bhadeshia HKDK, Suh DW. Scr Mater 2010;63:297. [23] Remy L, Pineau A. Mater Sci Eng 1977;A28:99. [24] Zhang HW, Hei ZK, Liu G, Lu J, Lu K. Acta Mater 2003;51:1871. [25] Dumaya A, Chateau JP, Allain S, Migot S, Bouaziz O. Mater Sci Eng 2008;A483–484:184.
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189 [26] Bouaziz O, Allain S, Scott CP, Cugy P, Barbier D. Curr Opin Solid State Mater Sci 2011;15:141. [27] Sahraouia T, Hadji M, Yahi M. Mater Sci Eng 2009;A523:271. [28] Babu SS, Specht ED, David SA, Karapetrova E, Zschack P, Peet M, et al. Metall Mater Trans A 2005;36A:3281. [29] De Cooman BC. Curr Opin Solid State Mater Sci 2004;8:285. [30] Jimenez-Melero E, van Dijk NH, Zhao L, Sietsma J, Offerman SE, Wright JP, et al. Scr Mater 2007;56:421. [31] Saeed-Akbari A, Imlau J, Prahl U, Bleck W. Metall Mater Trans A 2009;40A:3076. [32] Yoo JD, Park KT. Mater Sci Eng 2008;A496:417. [33] Reick W, Pohl M, Padilha AF. Steel Res 1996;67:253. [34] Sohn SS, Lee B-J, Lee S, Kim NJ, Kwak J-H. Acta Mater 2013;61:5050.
189
[35] Lee S, Kim J, Lee SJ, De Cooman BC. Scr Mater 2011;65:1073. [36] van Dijk NH, Butt AM, Zhao L, Sietsma J, Offerman SE, Wright JP, et al. Acta Mater 2005;53:5439. [37] Berrahmoune MR, Berveiller S, Inal K, Moulin A, Patoor E. Mater Sci Eng A 2004;378:304. [38] Takaki S, Fukunaga K, Syarif J, Tsuchiyama T. Mater Trans 2004;45:2245. [39] Mahieu J, Maki J, De Cooman BC, Claessens S. Metall Mater Trans A 2002;33A:2573. [40] El-Danaf E, Kalidindi SR, Doherty RD. Metall Mater Trans A 1999;30A:1223. [41] Schumann H. Neue Hu¨tte 1972;10:605. [42] Allain S, Chateau JP, Bouaziz O. Mater Sci Eng A 2004;143:387–9. [43] Frommeyer G, Bru¨x U, Neumann P. ISIJ Int 2003;43:438.
ScienceDirect Acta Materialia 78 (2014) 181–189 www.elsevier.com/locate/actamat
Novel ferrite–austenite duplex lightweight steel with 77% ductility by transformation induced plasticity and twinning induced plasticity mechanisms Seok Su Sohn a, Kayoung Choi b, Jai-Hyun Kwak c, Nack J. Kim b, Sunghak Lee a,⇑ a
Center for Advanced Aerospace Materials, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea Graduate Institute of Ferrous Technology, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea c Sheet Products & Process Research Group, Technical Research Laboratories, POSCO, Kwangyang 545-090, Republic of Korea b
Received 11 March 2014; accepted 24 June 2014 Available online 19 July 2014
Abstract The need for lightweight materials has been an important issue in automotive industries to reduce greenhouse gas emission and to improve fuel efficiency. In addition, automotive steels require an excellent combination of strength and ductility to sustain automotive structures and to achieve complex shapes, but the traditional approach to obtain a reduction in weight from down-gauged steels with high strength has many limitations. Here, we present a new ferrite–austenite duplex lightweight steel containing a low-density element, Al; this steel exhibits tensile elongation up to 77% as well as high tensile strength (734 MPa). The enhanced properties are attributed to the simultaneous formation of deformation-induced martensites and deformation twins and the additional plasticity due to deformation twinning in austenite grains having optimal mechanical stability. The present work gives a promise for automotive applications requiring excellent properties as well as reduced specific weight. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Duplex lightweight steel; Annealing; Tensile properties; Transformation induced plasticity (TRIP); Twinning induced plasticity (TWIP)
1. Introduction Recently, there has been a focus on the need to reduce vehicle weight in order to reduce exhaust emissions and improve fuel efficiency [1,2]. The most efficient method is the use of materials lighter than conventional ones. Many efforts have been directed towards applying lightweight materials such as Al alloys or Mg alloys, despite their high costs [3,4]. However, these alloys show poor formability, which restricts the application [5,6]. The development of new advanced automotive steels, namely lightweight steels, ⇑ Corresponding author. Tel.: +82 54 279 2140; fax: +82 54 279 2399.
E-mail address: [email protected] (S. Lee).
is recognized as a more realistic measure [7,8]. As a part of this study, a considerable amount of Al has been added to automotive steels to obtain a lightweight material [9–11]. It has been shown that the addition of 1 wt.% Al leads to a 1.5% weight reduction in comparison with conventional steels. However, the addition of Al reduces the ductility, and thus the method utilizing microstructures such as a full austenitic phase strengthened with j-carbides, or a duplex phase of ferrite + austenite, has been suggested [9–19]. The full austenitic microstructure displays low work-hardening capacity, in spite of the excellent combination of strength and ductility, which results in the problem of poor formability [20]. When the austenite is utilized as a secondary phase in ferrite matrix, high-strength, high-ductility lightweight
http://dx.doi.org/10.1016/j.actamat.2014.06.059 1359-6454/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
182
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
steels having high work-hardening capacity can be developed by the austenite/martensite transformation during the deformation, namely the transformation induced plasticity (TRIP) mechanism [11–14]. According to research on an Fe–3.5Mn–5.9Al–0.4C steel [12] and an Fe–5.8Mn–3.1Al–0.12C–0.47Si steel [14], for example, the austenite was retained at room temperature when steels were annealed above 720 °C, and was transformed to a0 -martensite during the deformation, which led to excellent properties of tensile strength (990 MPa) and elongation (28%). This c/a0 transformation in (a + c) duplex structure mainly depends on the orientation, mechanical stability and stacking fault energy (SFE) of the retained austenite [11–13,21,22]. In order to achieve the more excellent combination of strength and ductility, the utilization of another powerful deformation mechanism of twinning induced plasticity (TWIP) by properly controlling the mechanical stability or SFE of austenite is essentially needed [23–27]. In a certain stability or SFE range of austenite, deformation twins prevent the movement of dislocations as they work for the grain refinement [26,27]. This variation in deformation mechanism might affect the strength and ductility by forming complex microstructures mixed with twins and martensite. Furthermore, such a microstructural evolution occurring during the final stage of deformation might play an important role in controlling deformation mechanisms. Thus, studies on deformation mechanisms are essential for the evaluation of alloy design, microstructural evolution and process control. In the present study, therefore, the (a + c) duplex lightweight steel showing the operation of both TRIP and TWIP mechanisms was developed by varying annealing conditions of an Fe–0.3C–8.5Mn–5.6Al steel, and tensile properties were evaluated. Detailed deformation mechanisms were investigated in relation to microstructural evolution by electron back-scatter diffraction (EBSD) and transmission electron microscopy (TEM) analyses, and the correlation between microstructural evolution process and tensile ductility was verified. 2. Experimental 2.1. Lightweight steels The lightweight steel used in this study was fabricated by a vacuum induction melting method, and its nominal composition is Fe–0.3C–8.5Mn–5.6Al–(<0.02)(P + S) (wt.%). After thick plates of 60 mm in thickness were homogenized at 1200 °C for 1 h, they were hot-rolled between 1100 and 900 °C. They were then cooled in a furnace from 650 °C after holding at this temperature for 1 h in order to simulate a coiling procedure. The hot-rolled steel sheets of 3 mm in thickness were rolled at room temperature to make 1 mm thick steel sheets. The sheets were annealed at 800 °C for 1 min or 30 min and at 900 °C for 30 min in
a continuous annealing simulator (model; CAS-AY-II, Ulvac-RIKO, Inc., Japan) to form a mixture of ferrite and austenite, and were cooled in the air. For convenience, the steel sheets annealed at 800 °C for 1 min, at 800 °C 30 min and at 900 °C for 30 min are referred to as “A81”, “A83” and “A93”, respectively. The effects of Al addition on weight reduction are attributed to the lattice expansion and the low atomic weight of substitutional solution [3]. The density of the present steel was measured to be 7.2 g cm3, by a densitometry (Mettler-Toledo XP205, Mettler-Toledo AG, Switzerland) on the basis of the Archimedes principle, which shows an apparent reduction of 8.5% in comparison to pure Fe. 2.2. Microstructural analysis Phases present in the specimens were identified by X-ray diffraction (XRD; Cu Ka radiation; scan rate, 2° min1; scan step size, 0.02°) and TEM. Their volume fractions were measured by the direct comparison method using XRD analysis [14]. Integrated intensities of (2 0 0)a and (2 1 1)a peaks and (2 2 0)c and (3 1 1)c peaks were used for this XRD method. For the TEM observation, specimens were mechanically polished to a thickness of 50 lm, punched to prepare disk specimens (diameter: 3 mm) by a disk cutter and then electro-polished by a twin-jet polisher (model; Tenupol-5, Struers, Denmark) in a solution of CH3COOH (90%) and HClO4 (10%) to prepare thin foil specimens. The thin foils were observed in a TEM (model 2100, JEOL, Japan) operated at an acceleration voltage of 200 kV. EBSD analysis (step size, 50 nm) was conducted by a field emission scanning electron microscope (FE-SEM, Quanta 3D FEG, FEI Company, USA). The data were then interpreted by orientation imaging microscopy (OIM) analysis software provided by TexSEM Laboratories, Inc. Electron probe microanalysis (EPMA) measurements employing wavelength-dispersive spectrometry (WDS) were also performed by an EPMA microprobe (model; JXA 8530F microprobe, JEOL, Japan) at an electron beam voltage of 15 keV. Since the precise measurement of C content was difficult by EPMA, the C content was measured by the XRD method using the following equation [28]: ac ¼ 3:578 þ 0:0330X C þ 0:0056X Al þ 0:00095X Mn ð1Þ ˚ , and XC, XMn where ac is austenite lattice parameter, in A and XAl are concentrations of C, Mn and Al, respectively, in wt.%. The austenite lattice parameter (ac) was determined from a d-spacing of (2 2 0)c position. 2.3. Tensile test Plate-type tensile specimens having gage length of 25 mm, gage width of 6 mm and gage thickness of 1 mm were prepared in the longitudinal direction. They were
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
tested at room temperature at a strain rate of 103 s1 by a universal testing machine (model; 8801, Instron, Canton, MA, USA) of 100 kN capacity. 3. Results 3.1. Microstructure EBSD phase maps of the A81, A83 and A93 steels are shown Fig. 1a–c. The annealed steels have duplex microstructures of ferrite and austenite. The overall volume fraction and grain size of austenite in the A81 steel is 36% and 0.93 lm, respectively (Fig. 1a). As the annealing time at 800 °C increases, the austenite grain size increases up to 1.56 lm, while its volume fraction remains (37%) (Fig. 1b). In the A93 steel, both the volume fraction and size of austenite grain increase to 42% and 2.82 lm, respectively (Fig. 1c). Many austenite grains contain annealing twins. 3.2. Tensile properties Fig. 2a shows room-temperature engineering stress– strain curves. As the annealing time at 800 °C increases from 1 to 30 min, the yield and tensile strengths decrease,
(a)
whereas the elongation increases. The A81 and A83 steels show a common tensile flow behavior of yielding, necking and failure. When the annealing temperature increases from 800 to 900 °C, the elongation dramatically increases up to 77%, although the yield and tensile strengths are slightly reduced. This increased elongation is quite outstanding, and has not been reported in previous studies on duplex lightweight steels. It is also noted that the tensile behavior shows the steady stress flow after the rapidly increased strain-hardening. Fig. 2b shows the true stress–strain curve, strain-hardening rate (dr/de) curve and volume fraction of austenite measured by the direct comparison XRD method, and are indicated by the black line, red dotted line and blue circular symbols, respectively, for the A93 steel. The true stress– strain curve shows a continuously increasing strain-hardening behavior, which is different from the steady stress flow behavior of the engineering stress–strain curve, and the tensile strength reaches 1280 MPa. The strain-hardening rate curve displays a multiple stage strain-hardening behavior. In the initial deformation stage at the true strain of 0–0.25, the strain-hardening rate continuously decreases with increasing plastic strain, while the austenite volume fraction is not varied much in the range of 38–42%. Beyond the strain of 0.25, the strain-hardening rate increases and
Dγ : 1.56 μ m Vγ (XRD) : 37 %
Dγ : 0.93 μ m Vγ (XRD) : 36 % A81
(b)
183
Dγ : 2.82 μ m Vγ (XRD) : 42% A83
(c)
BCC FCC
A93
TD 5 μm
5 μm
5 μm
RD
Fig. 1. EBSD phase maps of the (a) A81, (b) A83, and (c) A93 steels, showing duplex microstructures of ferrite and austenite.
184
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
(b) 2500
A83
800
TrueStress, dσ/dε (MPa)
A81 A93
600 400 - A93 Steel YS : 502 MPa TS : 734 MPa El. : 77 %
200
A93 Steel True σ - ε Strain Hardening Rate ( dσ / dε )
2000
50 40 30
1000
20 500
Stage 1
Stage 2
Stage 3
10 0 0.7
0
10
20
30
40
50
60
70
60
1500
0 0
70
0.0
80
0.1
0.2
0.3
0.4
0.5
Austenite Fraction (%)
Engineering Stress (MPa)
(a) 1000
0.6
True Strain
Engineering Strain (%)
Fig. 2. (a) Room temperature tensile engineering stress–strain curves of the A81, A83 and A93 steels, and (b) true stress–strain curve, strain-hardening rate (dr/de) curve and volume fraction of austenite, which are indicated by black line, red dotted line and blue circular symbols, respectively, for the A93 steel. Two peaks are marked by black arrows on the strain-hardening rate (dr/de) curve. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
decreases twice, indicating two up–down peaks as marked by black arrows. In the strain range of 0.25–0.45 where the strain-hardening rate shows the first peak, the austenite volume fraction rapidly drops to 22%, and then slightly decreases to 20% in the strain range of 0.45–0.55 (fracture ε = 0.15
(a)
BF
strain) where the strain-hardening rate shows the second peak. For convenience, the true strain ranges of 0–0.25, 0.25–0.45 and 0.45–0.55 are referred to as Stages 1, 2 and 3, respectively, as shown in Fig. 2b. Considering the rapid drop of austenite volume fraction, Stage 2 might be related BF
(b)
b
(-11-1) (-200)
(11-1)
(000)
0.5 μm
(c)
BF
ε = 0.35
DF
(d)
d
BF
(e)
Deformaon Twin
e
α’-Martensite
(-11-1) (-200) (11-1) (000) (1-21)
0.5 μm
0.5 μm
(f)
ε = 0.55
0.5 μm
BF
(2-11)
(000) (110)
DF
(g) Deformaon Twin
(-200) (-11-1) (000)
0.5 μm
0.5 μm
(11-1)
Fig. 3. (a, b) TEM bright field images of an austenite grain in Stage 1 (true strain: 0.15), showing the dislocation pile-up at the annealing twin boundary. (c–e) TEM bright and dark field images of an austenite grain in Stage 2 (true strain: 0.35), showing deformation twins and a0 -martensites formed in the left and right sides, respectively, of an annealing twin. (f, g) TEM bright and dark field images of an austenite grain in Stage 3 (true strain: 0.55), showing thick deformation twins.
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
to the TRIP mechanism [29]. In order to elucidate causes of extremely high elongation of 77% in the A93 steel, more detailed analyses including TEM data are essentially needed.
185
The transition in the strain-hardening behavior (Fig. 2b) is closely related with deformation mechanisms. TEM analyses were conducted in Stages 1–3, and the results are shown in Fig. 3. In Stage 1 at the true strain of 0.15, austenite grains are deformed by the dislocation slip mechanism (Fig. 3a). The inset in Fig. 3b shows the selected-area diffraction (SAD) pattern taken along the [0 1 1] zone axis of an austenite grain, which further confirms that the austenite grain contains an annealing twin. The density of dislocations is higher at the annealing twin boundary than inside (Fig. 3b). At the strain of 0.35 in Stage 2, deformation twins are observed in the left side of an annealing twin (Fig. 3c), as verified by the dark-field image and SAD pattern taken along the [0 1 1] zone axis of the austenite (Fig. 3d). In the right side of the annealing twin, a0 -martensite is found without e-martensite (Fig. 3e), as confirmed by a SAD pattern taken along the [0 1 1] zone axis of the body centered cubic (bcc) phase. It is noted from these TEM results that both deformation twins and a0 -martensite can be formed inside one austenite grain. At the strain of 0.55 in Stage 3, a number of thick deformation twins are well formed inside the whole austenite grain (Fig. 3f and g). These twins are not developed into secondary twins, unlike in conventional TWIP steels [26]. EBSD inverse pole figure maps of the same area after the interrupted tensile test at the true strains of 0.35, 0.4 and 0.45 in Stage 2 are shown in Fig. 4. At the strain of 0.35, deformation twins and a0 -martensite are independently formed inside austenite grains whose boundaries are surrounded by white dotted lines, as marked by blue and white arrows, respectively (Fig. 4a and d). Deformation twins and a0 -martensite are formed from the opposite boundary inside a grain. At the strain of 0.4, a0 -martensite initiated from the boundary are grown into the grain interior, and deformation twins are thickened or newly formed at the grain boundary (Fig. 4b and e). In the grain marked by a yellow arrow, deformation twins and a0 -martensite are independently formed from the opposite boundary. At the strain of 0.45, a number of deformation twins are developed, while a0 -martensite is grown further (Fig. 4c and f). Some a0 -martensite makes inroads into deformation twins.
martensite and deformation twins in Stage 2 and the additional plasticity due to deformation twins in Stage 3. Relatively fine austenite grains 0.9–2.8 lm in size are formed after the annealing (Fig. 1a–c). The austenite grain refinement raises the thermal stability of austenite, and reduces the martensite start temperature (Ms) below the room temperature [30]. Deformation mechanisms of the austenite existed even at room temperature are varied with mechanical stability or SFE [9,21,25]. Though the SFE is not the absolute standard for classifying deformation mechanisms, many studies on the measurement or calculation of SFE in austenitic steels have been conducted [25,31,32]. However, the experimental SFE data of the austenite in (a + c) duplex steels have been rarely reported [33] because their deformation behavior is determined by reactions with the adjacent ferrite as well as SFE. In general, deformation mechanisms related with the mechanical stability of austenite are varied mainly with the content of Mn [9]. In high-Mn (20 wt.% or higher) duplex lightweight steels, the austenite is deformed mainly by mechanisms of microbanding and wavy glide because of very high SFE [17,21,23]. In low-Mn (3–4 wt.%) (a + c) duplex lightweight steels, the only TRIP mechanism, i.e., deformation-induced martensitic transformation, is working [12,14]. Here, the optimal mechanical stability of austenite importantly works for excellent mechanical properties, and the deformation mechanism cannot be escaped from the TRIP regime even in the cases of the increased stability or SFE [34]. The Fe–12Mn–1Al–0.7C steel, one of medium-Mn austenitic steels, whose austenite grain size was 6 lm, was deformed by the only TWIP mechanism [35]. In the present Fe–8.5Mn–5.6Al–0.3C steel, one of medium-Mn (a + c) duplex steels, both the TRIP and TWIP mechanisms can work as major deformation mechanisms at an optimal mechanical stability of austenite, which has not been reported previously. The mechanical stability against the formation of strain-induced martensite and the SFE which is related with the kind of deformation mechanism are influenced by the grain size and composition of austenite. The increase in austenite grain size readily provides nucleation sites for stacking faults and shear-band intersections, and activates the formation of martensite due to the decrease in elastic strain energy required for lattice displacement and in strain energy required for shear deformation [36–38]. Effects of alloy composition and grain size on Ms can be explained by the following equations suggested by Jimenez-Melero et al. and Mahieu et al. [30,39]:
4. Discussion
M s0 ð CÞ ¼ 539 30:4Mn 7:5Si þ 30Al
3.3. Tensile deformation behavior
The most unique characteristic of the present lightweight steel is the persistent elongation up to 77% in the A93 steel, as shown in Fig. 2. This is mainly attributed to the simultaneous formation of deformation-induced
M s ð CÞ ¼ M s0 AX C BðV
1=3 Þ c
273
ð2Þ ð3Þ
where Ms0 shows the effect of alloying composition except C on Ms. XC, A, Vc and B are C content in austenite, constant (423 K wt.%1), average volume of austenite grain
186
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
ε
ε
0.35
(a)
ε
0.4 (b)
0.45 (c)
Deformaon Twin
5 μm
5 μm
5 μm
(d)
(e)
(f)
α’-Martensite
TD 5 μm
5 μm
5 μm
RD // Tensile Direcon
Fig. 4. EBSD IPF maps of (a–c) face centered cubic (fcc) and (d–f) bcc in the same area after the interrupted tensile test at the true strains of 0.35, 0.4 and 0.45 in Stage 2, showing deformation twins (blue arrow marks) and a0 -martensites (white arrow marks) independently formed inside austenite grains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 C, Mn and Al contents, austenite grain size, Ms and SFE of the A81, A83 and A93 steels. Steel
Ca (wt.%)
Mnb (wt.%)
Alb (wt.%)
Austenite grain size (lm)
Ms0c (°C)
Msc (°C)
SFEc (mJ m2)
A81 A83 A93
0.58 0.58 0.55
11.1 11.3 11.5
4.58 4.62 4.74
0.93 1.56 2.82
339 334 332
417 215 69
51.6 51.8 51.1
a b c
Content of C measured from the XRD analysis. Contents of Mn and Al measured from the EPMA analysis. Ms and SFE calculated from Eqs. (2–4).
and geometry coefficient of austenite grain (475 lm K), respectively. The SFE (C is expressed by the following equation [31]: C ¼ 2qDGc!e þ 2rc=e þ 2qDGex ; DGex ¼ 170:06 expðd=18:55Þ c!e
c/e
ð4Þ
where q, DG , r , and DGex show the molar surface density along {1 1 1} planes, free energy for c ! e phase transformation, c/e interfacial energy and excess part of the SFE due to the grain size (d) effect, respectively. The DGc!e is related to chemical compositions. According to Eq. (4), the increase in austenite grain size reduces the SFE. However, the SFE has only an indirect effect on the twinning stress, and thus a homogeneous slip length is the most relevant microstructural variable directly
influencing the twinning stress [40]. The increase in homogeneous slip length reduces the critical stress required for twin nucleation. C, Mn and Al contents, austenite grain size, Ms and SFE of the A81, A83 and A93 steels are shown in Table 1. Here, C, Mn and Al contents are measured from the XRD and EPMA analyses, and Ms and SFE are calculated from Eqs. 2–4. Contents of Mn and Al are not varied much by the annealing conditions, and hardly influence the Ms and SFE. The C content, which can affect the Ms slightly, decreases in the A93 steel having the highest volume fraction of austenite, but its variation is not large. Thus, the Ms0 are almost same in the three steels. The Ms of the A93 steel is much higher (about six times) than that of the A81 steel because the austenite grain size is considered in the calculation of Ms. This implies that the deformation
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
amount required for martensite formation is relatively small. In view of the SFE, the SFE is hardly varied in the range of 51–52 mJ m2, although the austenite grain size is considered in its calculation. This indicates that the SFE of the austenite is not a main parameter affecting deformation mechanisms in the duplex microstructure. It is also well known that the martensitic transformation rarely occurs in this range of SFE [41–43]. Rather than these parameters such as alloying contents and SFE, the decrease in critical stress required for twin nucleation due to the increase in austenite grain size can critically affect the activation of twinning. The A81 and A83 steels, whose stability of austenite is quite high because of grain refinement below 1.6 lm in size, show a common tensile flow behavior (Fig. 2a). In the A93 steel whose austenite grain size increases to 2.8 lm, on the other hand, the TRIP and TWIP mechanisms are working simultaneously beyond the true strain of 30% according to the relatively low stability of austenite, which results in the extremely high elongation of 77%. This unusual deformation behavior related with the extremely high elongation in the A93 steel is determined by the crystallographic preferential orientation of austenite grain. EBSD IPF maps of the same area after the interrupted tensile test at the true strains of 0.35 and 0.45 in Stage 2 are shown in Fig. 5. When an austenite grain deformed by the dislocation slip without deformation twins or a0 -martensite at the strain of 0.35 (marked by white dotted line in Fig. 5a) is further deformed (strain:
ε = 0.35
FCC IPF
(a)
187
0.45), deformation twins and a0 -martensite are independently formed inside the grain (Fig. 5c and d). According to the Schmid factor map of the austenite grain of Fig. 5a (Fig. 5b), the local orientation variation appears inside the grain during the strain up to 0.35. Here, the “A” region whose Schmid factor is low at 0.25–0.3 and the “B” region whose Schmid factor is high at 0.3–0.4 correspond to nucleation sites for deformation twins and a0 -martensite, respectively (Fig. 5c and d). It directly shows the orientation dependence of the TRIP and TWIP mechanisms. Considering that the martensitic transformation is triggered by the supply of strain energy as a consequence of dislocation pile-ups at strong barriers such as grain boundaries [23,39], the martensitic transformation preferentially occurs in regions having high resolved shear stresses, i.e. high Schmid factors. It is also possible to trigger deformation twins in regions having relatively low resolved shear stresses because the A93 steel has considerably higher mechanical stability or SFE than low-Mn duplex steel [9,12]. The present duplex lightweight steel (the A93 steel) shows the excellent elongation of 77% together with the high tensile strength, which are far superior to conventional lightweight steels [11–14,17,34]. The enhanced tensile elongation has been achieved by the operation of multideformation mechanisms such as (1) the formation of deformation-induced martensite together with twinning and (2) deformation twinning in austenite grains having optimal mechanical stability. The present work also shows
ε = 0.35
FCC Schmid Factor
(b) A
B B 2 μm
ε = 0.45
FCC IPF
(c)
2 μm
ε = 0.45
BCC IPF
(d)
α’ Martensite Deformaon Twin
TD
2 μm
2 μm
RD // Tensile Direcon
Fig. 5. EBSD (a) IPF map of fcc, (b) Schmid factor map of fcc, (c) IPF map of fcc, and (d) IPF map of bcc in the same area after the interrupted tensile test at the true strains of 0.35 and 0.45 in Stage 2, showing deformation twins (blue arrow marks) and a0 -martensites (white arrow marks) independently formed inside an austenite grain. In (b), the “A” region, whose Schmid factor is low at 0.25–0.3, and the “B” region, whose Schmid factor is high at 0.3– 0.4, correspond to nucleation sites for deformation twins and a0 -martensites, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
188
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189
that the unique deformation behavior is dependent on the crystallographic preferential orientation. Since this steel has outstanding properties of strength and ductility as well as reduced density, it gives a promise for automotive applications requiring excellent properties. In order to further enhance microstructures and properties of duplex lightweight steels, more intensive studies to design new lightweight steels, to establish optimized annealing conditions and to clarify mechanisms involved in improved strength and ductility should be continued. 5. Conclusions In the present study, the (a + c) duplex lightweight steel showing the operation of both TRIP and TWIP mechanisms was developed by varying annealing conditions of an Fe–0.3C–8.5Mn–5.6Al (wt.%) steel, and tensile properties were analyzed in relation with microstructural evolution and deformation mechanisms. (1) The annealed steels showed duplex microstructures of ferrite and austenite. As the annealing time at 800 °C increased, the austenite grain size increased up to 1.56 lm, while its volume fraction remained constant (37%). In the A93 steel (annealed at 900 °C for 30 min), both the volume fraction and size of austenite grain increased to 42% and 2.82 lm, respectively. (2) The A81 and A83 steels (annealed at 800 °C for 1 min and 30 min, respectively) showed a common tensile flow behavior of yielding, necking and failure with elongation of 35%. In the A93 steel, however, the elongation dramatically increased up to 77%, and the tensile behavior showed the steady stress flow after the rapidly increased strain-hardening. The strain-hardening rate curve of the A93 steel displayed a multiple stage strain-hardening behavior. In the initial deformation stage, the strain-hardening rate continuously decreased with increasing plastic strain. Beyond the strain of 0.25, the strain-hardening rate increased and decreased twice, indicating two up– down peaks. (3) The detailed multi-deformation mechanisms of the A93 steel were analyzed by TEM and EBSD in each deformation stage. In Stage 1, austenite grains were deformed by the dislocation slip. In Stage 2, deformation twins and a0 -martensites were independently formed inside one austenite grain. Finally in Stage 3, a number of thick deformation twins were well formed inside the whole austenite grain. (4) The enhanced tensile elongation and operation of multi-deformation mechanisms were attributed to the optimal mechanical stability of austenite. The C, Mn and Al contents were not varied much with annealing conditions, and had little effect on mechanical stability. The austenite grain size acted as a factor determining the activation of martensitic transformation and twinning. When the stability of austenite
was quite high because of the grain refinement below 1.6 lm, like in the A81 or A83 steel, the tensile flow showed the common behavior. In the A93 steel whose austenite grain size increased to 2.82 lm, however, the TRIP and TWIP mechanisms were working simultaneously beyond the true strain of 30% according to the relatively low stability of austenite. (5) This unusual deformation behavior related to the extremely high elongation in the A93 steel was determined by the crystallographic preferential orientation of austenite grain. The martensitic transformation preferentially occurred in regions having high resolved shear stresses, i.e., high Schmid factors.
Acknowledgment This work was supported by the Ministry of Knowledge Economy under a Grant No. 10031723-2011-21. References [1] Mayyasa A, Qattawia A, Omara M, Shana D. Renew Sustainable Energy Rev 2012;16:1845. [2] Kuziak R, Kawalla R, Waengler S. Arch Civ Mech Eng 2008;8:103. [3] Kulekci MK. Int J Adv Manuf Technol 2008;39:851. [4] Miller WS, Zhuang L, Bottema J, Wittebrood AJ, De Smet P, Haszler A, et al. Mater Sci Eng 2000;A280:37. [5] Hirsch J. Acta Mater 2013;61:818. [6] Lang L, Danckert J, Nielsen KB. J Mater Process Technol 2004;148:119. [7] Matlock DK, Speer JG. In: Haldar A, Suwas S, Bhattacharjee SD, editors. Microstructure and texture in steels. London: Springer; 2009. [8] Ma M, Yi H. In: Weng Y, Dong H, Gan Y, editors. Advanced steels. Heidelberg: Springer; 2011. [9] Frommeyer G, Bru¨x U. Steel Res Int 2006;77:627. [10] Rana R, Liu C, Ray RK. Scr Mater 2013;68:354. [11] Hwang SW, Ji JH, Lee EG, Park KT. Mater Sci Eng 2011;A528:5196. [12] Seo CH, Kwon KH, Choi K, Kim KH, Kwak JH, Lee S, et al. Scr Mater 2012;66:519. [13] Park SJ, Hwang B, Lee KH, Lee TH, Suh DW, Han HN. Scr Mater 2012;68:365. [14] Suh DW, Park SJ, Lee TH, Oh CS, Kim SJ. Metall Mater Trans A 2010;41A:397. [15] Frommeyer G, Drewes EJ, Engl B. Rev Metall Cah Inf Technol 2000;97:1245. [16] Falat L, Schneider A, Sauthoff G, Frommeyer G. Intermetallics 2005;13:1256. [17] Sutou Y, Kamiya N, Umino R, Ohnuma I, Ishida K. ISIJ Int 2010;50:893. [18] Choi K, Seo C-H, Lee H, Kim SK, Kwak JH, Chin KG, et al. Scr Mater 2010;63:1028. [19] Park KT, Jin KG, Han SH, Hwang SW, Choi K, Lee CS. Mater Sci Eng 2010;A527:3651. [20] Chin KG, Kang CY, Shin SY, Hong S, Lee S, Kim HS, et al. Mater Sci Eng 2011;A528:2922. [21] Gra¨ssel O, Kru¨ger L, Frommeyer G, Meyer LW. Int J Plast 2000;16:1391. [22] Ryu JH, Kim DI, Kim HS, Bhadeshia HKDK, Suh DW. Scr Mater 2010;63:297. [23] Remy L, Pineau A. Mater Sci Eng 1977;A28:99. [24] Zhang HW, Hei ZK, Liu G, Lu J, Lu K. Acta Mater 2003;51:1871. [25] Dumaya A, Chateau JP, Allain S, Migot S, Bouaziz O. Mater Sci Eng 2008;A483–484:184.
S.S. Sohn et al. / Acta Materialia 78 (2014) 181–189 [26] Bouaziz O, Allain S, Scott CP, Cugy P, Barbier D. Curr Opin Solid State Mater Sci 2011;15:141. [27] Sahraouia T, Hadji M, Yahi M. Mater Sci Eng 2009;A523:271. [28] Babu SS, Specht ED, David SA, Karapetrova E, Zschack P, Peet M, et al. Metall Mater Trans A 2005;36A:3281. [29] De Cooman BC. Curr Opin Solid State Mater Sci 2004;8:285. [30] Jimenez-Melero E, van Dijk NH, Zhao L, Sietsma J, Offerman SE, Wright JP, et al. Scr Mater 2007;56:421. [31] Saeed-Akbari A, Imlau J, Prahl U, Bleck W. Metall Mater Trans A 2009;40A:3076. [32] Yoo JD, Park KT. Mater Sci Eng 2008;A496:417. [33] Reick W, Pohl M, Padilha AF. Steel Res 1996;67:253. [34] Sohn SS, Lee B-J, Lee S, Kim NJ, Kwak J-H. Acta Mater 2013;61:5050.
189
[35] Lee S, Kim J, Lee SJ, De Cooman BC. Scr Mater 2011;65:1073. [36] van Dijk NH, Butt AM, Zhao L, Sietsma J, Offerman SE, Wright JP, et al. Acta Mater 2005;53:5439. [37] Berrahmoune MR, Berveiller S, Inal K, Moulin A, Patoor E. Mater Sci Eng A 2004;378:304. [38] Takaki S, Fukunaga K, Syarif J, Tsuchiyama T. Mater Trans 2004;45:2245. [39] Mahieu J, Maki J, De Cooman BC, Claessens S. Metall Mater Trans A 2002;33A:2573. [40] El-Danaf E, Kalidindi SR, Doherty RD. Metall Mater Trans A 1999;30A:1223. [41] Schumann H. Neue Hu¨tte 1972;10:605. [42] Allain S, Chateau JP, Bouaziz O. Mater Sci Eng A 2004;143:387–9. [43] Frommeyer G, Bru¨x U, Neumann P. ISIJ Int 2003;43:438.