- Email: [email protected]
Microstructure and mechanical response of single-crystalline high-manganese austenitic steels under high-pressure torsion: The effect of stacking-fault energy
Materials Science & Engineering A 604 (2014) 166–175
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Microstructure and mechanical response of single-crystalline high-manganese austenitic steels under high-pressure torsion: The effect of stacking-fault energy E.G. Astafurova a,n, M.S. Tukeeva a, G.G. Maier a, E.V. Melnikov a, H.J. Maier b a Laboratory of Physical Materials Science, Institute of Strength Physics and Materials Science, Siberian Branch of Russian Academy of Science, Akademichesky Prospect 2/4, 634021 Tomsk, Russia b Institut für Werkstoffkunde (Materials Science), Leibniz Universität Hannover, 30823 Garbsen, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 10 January 2014 Received in revised form 5 March 2014 Accepted 7 March 2014 Available online 17 March 2014
We investigate the kinetics of the structural deformation and hardening of single-crystalline austenitic Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C (in wt%) steels with different stacking-fault energies after cold high-pressure torsion. Independently of the stacking-fault energy, mechanical twinning was found to be the basic deformation mechanism responsible for the rapid generation of an ultrafine-grained microstructure with a high volume fraction of twin boundaries. Under high-pressure torsion, the spacing between twin boundaries increases, and the dislocation density and microhardness decrease as the stacking-fault energy increases. The formation of a twin net from the beginning of plastic flow in Fe–13Mn–1.3C steel provides a homogeneous distribution of microhardness values across the discs independent of strain under torsion. Lower hardness values in the disk centers compared to the periphery were observed for the two other steels, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C, with higher stacking-fault energies due to changes in the densities of the twin boundaries. An additional increase in the dislocation density for the Fe–13Mn–1.3C steel was detected compared with the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, which was a result of torsion in the temperature range of dynamic strain aging. The appearance of small fractions of ε and α0 phases in the structures of the Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C steels is discussed. & 2014 Elsevier B.V. All rights reserved.
Keywords: Austenite Steel Twinning High-pressure torsion Microstructure Stacking-fault energy
1. Introduction In recent decades, the physical studies of strength and plasticity have been directed at developing methods for achieving highstrength states in metals by refinement of their structures under severe plastic deformation (SPD) [1]. Progress in this area is not only associated with the creation of new deformation schemes but also, to a large extent, with the modification and combination of known methods, the optimization of the composition and structure of alloys prior to SPD and with grain-boundary engineering. High-pressure torsion (HPT) refers to the processing technique in which samples in the form of thin disks are subjected to simultaneous high compressive stress and torsion [1,2]. Using this method, nanocrystalline structural states are typically formed in various metals and alloys [2]. The obvious advantage of this SPD method is the ability to realize unlimited strains, which are often
n
Corresponding author. Tel.: þ 7 3822 286961; fax: þ 7 3822 491245. E-mail address: [email protected] (E.G. Astafurova). URL: http://www.ispms.ru (E.G. Astafurova).
http://dx.doi.org/10.1016/j.msea.2014.03.029 0921-5093/& 2014 Elsevier B.V. All rights reserved.
not achievable in conventional static tests, while maintaining the integrity of the workpiece and its shape (volume). The principles of the HPT method are described in detail in [2]. The HPT method has successfully been used to deform Armco iron [3,4] and steels [5–12]. HPT of twinning-assisted austenitic steels was not studied in detail [5,6,12]. However, mechanical twinning can be one of the factors that contribute to the rapid formation of ultrafine-grained structures with low-energy, highangle special boundaries Σ3n [13]. At temperatures that do not allow recrystallization, severe plastic deformation typically causes degradation of the dislocation structure by forming bands of localized deformation. The development of fine deformation twins in steels with a high concentration of interstitial atoms hinders this process because the deformation twins are more resistant to “disintegrating” compared with general boundaries. Some authors found deformation twins in nanocrystalline facecentered cubic (fcc) metals and alloys (with grain sizes of a few tens of nanometers), such as Cu, Al, Ni, and Cu–Zn, due to the emission of partial dislocations from the grain boundaries, overlapping of stacking-fault ribbons, grain boundary splitting, etc. [14,15]. The volume fraction of twin boundaries formed in
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
nanostructured materials is not considerable; it contributes but does not define the hardening of such materials. Forming a structural state with the maximum possible fraction of special Σ3n boundaries is not a trivial task because twinning in fcc alloys is largely complementary to the slip deformation mechanism and is generally observed in the low-temperature deformation regime [16]. In contrast, SPD is often conducted at elevated temperatures. Therefore, the deformation of alloys prone to high-temperature twinning offers the potential to create highstrength nanostructured materials with low-energy high-angle misorientations between the structural elements and that are resistant to degradation during annealing or subsequent deformation [17]. For AISI 304L stainless steel processed by equal channel angular pressing in the temperature range of 500–900 1C, Huang and coauthors [18] observed deformation twins in addition to deformation bands and subgrains (many bundles of deformation twins was observed at temperatures below 700 1C). Karaman et al. [19] reported the development of high-temperature twinning (up to 800 1C) for some difficult-to-work alloys, including 316L austenitic steel. They concluded that deformation twinning could be one of the main deformation modes in many metallic alloys over a wide temperature range when a high strength level is reached, irrespective of the way it is achieved [19]. As promising materials for the development of SPD methods, high-manganese austenitic steels with low stacking-fault energies (Hadfield steel-based) can be used. The high strength levels are reached in these steels due to solid solution hardening (high carbon content). Twinning in single- and polycrystalline Fe–(11–13)Mn–(1.0–1.3)C and Fe–13Mn–(1.3–2.7)Al–(1.3–1.75)C steels develops under uniaxial tension/compression at room temperature [20–25]. Under torsional deformation and high applied pressure, twinning had a significant contribution to the refinement and strengthening of the Hadfield steels Fe–13Mn–1.0C [6] and Fe–13Mn–1.2C [5]. The aim of this study is to characterize the microstructure and strength properties (microhardness) after the high-pressure torsion of single crystalline twinning-assisted Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C steels. The choice of such steel compositions allows one to change the stacking-fault energies by alloying Hadfield steel with manganese and aluminum while maintaining a high level of solid solution hardening. The use of single crystals allows for a detailed study of the strain hardening mechanisms and avoids the added complexity of the contribution from grain boundaries.
2. Experimental The high-Mn steels used in this study had the following chemical compositions: Fe–13Mn–1.0C, Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C (in wt%). Single crystals of the high-manganese austenitic steels were grown using the Bridgman technique under an inert gas atmosphere. All crystals were homogenized in an argon atmosphere at 1373 K for 24 h, solution-treated and water-quenched from 1373 K after 1 h. Electro-discharge machining was utilized to cut flat disks of 10 mm (diameter) 0.6 mm (height) for high-pressure torsion. The surfaces of the disks were cut parallel to the {001}-type planes for the Fe–13Mn–1.0C, Fe–13Mn–1.3C, and Fe–13Mn–2.7Al–1.3C steels and parallel to the {111}-type planes for the Fe–28Mn–2.7Al– 1.3C steel. The stacking-fault energies (SFE) for the investigated steels were assumed to be 25–30 mJ/m2 for Fe–13Mn–1.0C and Fe–13Mn–1.3C, 45 mJ/m2 for Fe–13Mn–2.7Al–1.3C, and 60 mJ/m2 for Fe–28Mn–2.7Al–1.3C according to [26]. The unconstrained high-pressure torsion was conducted at room temperature (p ¼5–6 GPa) for one to five revolutions at
167
a rotation speed of 1 rpm. The true logarithmic strain was calculated using the equation ε ¼ln(υr/h) [2], where r is the radius, υ is the rotation angle, and h is the thickness of the disk after highpressure torsion. Mechanical grinding and a final electrochemical polishing (50 g of CrO3 in 200 ml of H3PO4) were employed to remove all preparation-induced surface artifacts. The evolution of the microstructure was studied by repolishing and etching the specimens that were strained to different degrees. Chemical etching of the polished disk surfaces was conducted in a solution of 1 ml of HCl in 99 ml of H2O. Optical microscopy (OM) was performed using an Olympus GX-71 microscope. Electron-transparent foils were prepared by conventional electropolishing in a solution of 50 g CrO3 þ200 ml H3PO4 at an applied potential of 30 V. For microstructural analysis, transmission electron microscopy (TEM) was performed using a Philips CM200 transmission electron microscope. The foils for TEM were cut with the observation area located at half of the radius of the disks after high-pressure torsion. The average sizes of the microstructural elements (subgrains, twins) were determined using dark-field TEM images [27]. Selected area electron diffraction (SAED) patterns from areas of 0.5 and 12.6 μm2 were evaluated to identify the different phases. The density of twin boundaries was estimated as ρtw ¼ q/S, where q is the number of twin boundaries and S is the area of the TEM image. A Shimadzu XRD-6000 X-ray diffractometer with Cu-Kα radiation was utilized for the X-ray diffraction studies. The scalar dislocation density was determined by X-ray diffraction (XRD) using the method described in [28]. The microhardness was determined at room temperature using a Duramin 5 instrument with a load of 200 g.
3. Results 3.1. Microstructural observations High-pressure torsion leads to the fragmentation of single crystals of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C, and Fe–28Mn– 2.7Al–1.3C mainly due to the formation of a high dislocation density, twins and localized deformation bands, which were detected by optical metallography (Fig. 1) and under TEM studies (Figs. 2–4). Analysis of the crystal surface after repolishing and etching shows that there are several twinning systems and that the twins and shear bands are internally twinned (Fig. 1). The tendency for plastic flow to localize increases as the stacking-fault energy increases (Fig. 1). In the Fe–13Mn–1.3C steel, the localization processes are suppressed because, from the beginning of deformation, a twin net forms with a spacing between the boundaries of a few nanometers. In single crystals of the Fe– 28Mn–2.7Al–1.3C steel, localized shear bands were formed prior to or simultaneously with the development of deformation twinning (Fig. 1c). In the Fe–13Mn–1.3C steel at N ¼1 revolutions, there is a formation of twin packs with a micron width and shear bands (SB) (Fig. 1a). Using TEM, we observed twins in which the thicknesses of the plates were tens of nanometers (Fig. 2a, b). The effective size of the structural fragments is determined from the thickness of the deformation twins and the distance between the twin boundaries, which is 5–15 nm. With an increase in strain (N ¼3) in the Fe–13Mn–1.3C steel, the twin net remains; however, in the TEM and optical images, deformation and a large number of microbands were detected (Figs. 1d and 2c, d). Twins and shear bands with the fragment widths of 0.5 mm are filled with deformation twins with a thickness of 5–10 nm (Fig. 2c). Fragments within such a twin net have a high density of slip
168
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 1. Optical images of the etched surfaces of single crystals of Fe–13Mn–1.3C (a, d, g), Fe–13Mn–2.7Al–1.3C (b, e, h), and Fe–28Mn–2.7Al–1.3C (c, f, i) after HPT to different strains. The number of revolutions is shown in the figures.
dislocations ρsl 1015 m 2. When N ¼ 5, the twin net in the Fe– 13Mn–1.3C steel is significantly destroyed. On the etched surfaces of the disks, distorted twin traces are observed (Fig. 1g). The average size of the structural elements observed in the dark-field TEM images is approximately 100 nm (Fig. 2e). Within these structural elements, fragments of twin boundaries are observed, and characteristic twin reflections are observed in the SAED patterns (Fig. 2f). TEM observations of the Fe–13Mn–1.3C steel after high-pressure torsion at room temperature revealed a small fraction of ε-martensite (diffuse reflections observed in the SAED patterns at N ¼3–5; see Fig. 2f). High pressures applied during the given experimental technique (6 GPa) lead to the development of mechanical twinning in single-crystalline Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels with medium to high SFE. The presence of twinning was confirmed both metallographically and under TEM studies (Figs. 1, 3 and 4). Consequently, due to the high compressive stress under high-pressure torsion, twinning stresses in these steels are reached at room temperature, which often cannot be realized in experiments on static tension and compression. The proportion of twinned volumes after high-pressure torsion (N ¼ 1) is approximately 40%, as determined from the images of the surfaces of single crystals after repolishing and etching, and it does not change in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels compared to the Fe–13Mn–1.3C steel. However, the morphological features of the twins, the localization and strain hardening differ from those of the Hadfield steel. The bright- and dark-field TEM images for the Fe–13Mn–2.7Al– 1.3C and Fe–28Mn–2.7Al–1.3C steels after high-pressure torsion indicate that for N ¼1–3, the width of the twins and the size of “cells” limited by twin lamellae are higher than for the Hadfield
steel: 5–500 nm for Fe–13Mn–1.3C; 100–700 nm for Fe–13Mn– 2.7Al–1.3C; and 50–350 nm for Fe–28Mn–2.7Al–1.3C (Figs. 3a and 4a–c). When N¼3–5, the twins are fragmented and twin boundaries are observed over the entire volume of the samples (Figs. 3b–e and 4c–e). The reflections in the SAED patterns for the Fe–13Mn–1.3C, Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels are significantly diffused but not confined to the rings (Figs. 3 and 4), which is often observed after torsional strain [2]. One can distinguish a bright textural reflection in the SAED patterns, which also indicates the formation of deformation texture under HPT. 3.2. X-ray studies High-pressure torsion leads to line broadening in the X-ray diffraction patterns of the steels under investigation, but regardless of the number of revolutions, this method revealed a sharp deformation texture (Fig. 5a–c). Measurement of the coherent scattering regions provides the values of the structural parameters after high-pressure torsion of o20 nm (Fig. 5d). In the Fe–28Mn– 2.7Al–1.3C steel, the coherent scattering regions vary slightly slower with the strain, and the dislocation density is 2–3 times lower compared to the other two steels (Fig. 5d, e). Of the three studied steels, a maximal growth in the crystal lattice microstrain and dislocation density with increasing number of revolutions was observed in single crystals of Fe–13Mn–1.3C (Fig. 5d, e). 3.3. The microhardness tests High-pressure torsion caused a high strain hardening in the austenitic steels. The formation of highly nonuniform states in
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
169
Fig. 2. TEM images of twins in single crystals of Fe–13Mn–1.3C steel after HPT: (a, b) Ν¼ 1, bright-field image, SAED pattern, and dark-field image in twin reflection; (c, d) Ν¼ 3, bright-field image, SAED pattern, and dark-field image in matrix reflection; (e, f) Ν¼ 5, bright-field image, dark-field image using a twin reflection, and SAED pattern. SAED patterns were obtained from an area of 0.5 μm2.
steels under high-pressure torsion is accompanied by a significant increase in the microhardness from 2.5 GPa in the initial states up to 6.4–7.8 GPa after torsion for five revolutions (Fig. 5f). As shown in Fig. 5f, the microhardness of the Fe–13Mn–1.3C and Fe–13Mn– 2.7Al–1.3C steels after N ¼3 reaches a maximum value and then changes slowly as the strain increases. Work hardening in the Fe–13Mn–1.3C steel is higher than that in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels; thus, the Hadfield steel has the highest microhardness values after torsion (Fig. 5f). The microstructure formed under HPT in the Fe–13Mn–1.3C steel with a low SFE is very homogeneous independent of strain, which provides constant microhardness values through the diameter of the disk. Detailed Vickers microhardness measurements revealed lower hardness values in the disk centers for the Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels with a higher SFE (Fig. 6). 3.4. The influence of carbon content in Hadfield steel on phase composition and hardening TEM observations of the Fe–13Mn–1.3C steel (as two other steels) after high-pressure torsion at room temperature revealed a small fraction of ε-martensite (blurry reflections observed in the SAED patterns at N ¼3–5; see Fig. 2f). In single crystals of the Hadfield steel, which has a lower carbon content of Fe–13Mn–1.0C, the reflections corresponding to ε-martensite are clearly visible in SAED patterns after upset (pressing without rotation of a plunger) [6]. With an increase in strain, the proportion of martensite increases in the Fe–13Mn–1.0C steel, and after high-pressure
torsion for 2–3 revolutions, X-ray reflections corresponding to the ε-phase appear (Fig. 7). In single crystals of Fe–13Mn–1.3C, even after torsion for 5 revolutions, ε and α0 phases were not detected during the X-ray studies. The α0 phase volume fraction estimated using magnetization measurements [29] was not greater than 0.8% independent of the steel composition and strain.
4. Discussion 4.1. Fragmentation mechanisms for high-Mn steels under high-pressure torsion: the influence of stacking-fault energy Cold severe plastic deformation of single-crystalline Fe–13Mn– 1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels is associated with the formation of boundaries of a common and special type – twin – and their interaction and degradation with an increase in the number of revolutions. The tendency for plastic flow to localize is apparent in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels compared to the Hadfield steel and is a result of the SFE increase and micro-band plasticity, as it appeared in high-Mn–Al–C austenitic steels with high SFE (E85 mJ/m2) [30]. The development of mechanical twinning in single crystals of the Fe–13Mn–1.3C and Fe–13Mn–2.7Al–1.3C steels under highpressure torsion is consistent with data obtained in [20,21]. Under tensile deformation at room temperature, twinning develops in Fe–13Mn–1.3C single crystals from the early stages of deformation
170
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 3. Multiple twinning in the single crystals of Fe–13Mn–2.7Al–1.3C steel after HPT: (a) Ν¼ 1, bright-field image, SAED pattern obtained from the area of 12.6 μm2; (b) Ν¼ 5, bright-field image; (c) Ν¼ 5, dark-field image in combined (111)γ and (120)ε reflections; (d) Ν¼5, dark-field image in combined (200)γ and (1 2 1)ε reflections; (e) Ν ¼5, dark-field image in twin (111)γ reflection; (f) Ν ¼5, SAED pattern for (b–e). SAED patterns were obtained from an area of 12.6 μm2.
(ε 40.5%, depending on the initial orientation of the single crystal) and in o111 4-, o011 4-oriented specimens of the Fe–13Mn– 2.7Al–1.3C steel (ε 4 10%) [20,21]. The initially single-crystalline Hadfield steel specimen under stress transforms to a polycrystalline specimen with continuous misorientation filled with slip dislocations and twins of 5–15 nm (Table 1). Therefore, the structural elements are restricted by twin boundaries – mainly high-angle Σ3n boundaries of a special type. The increase in the twin fraction under plastic flow is mainly due to the nucleation of new thin twins. This is characteristic for the deformation twinning of high-alloyed substitution and interstitial solid solutions [16]. The multiplicity of twinning is attributed to a quasihydrostatic stress state under HPT. An increase in SFE leads to increased twinning stresses [16]; however, under HPT for the investigated steels, they are clearly reached in the stress process for the Fe–13Mn–1.3C and Fe–13Mn– 2.7Al–1.3C steels or very beginning of rotation under HPT for the Fe–28Mn–2.7Al–1.3C steel. The effective distance between the twin boundaries (twin thickness t and distance between twins e) increases with SFE during the alloying of Hadfield steel with Al and Mn (in steels Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C, see Table 1) and simultaneously reduces the effectiveness of reinforcement and strain-hardening (Fig. 5f). Consequently, the shear accommodation and energy dissipation processes may play a significant role under increasing SFE of high-strength steels. Wang and coauthors [15] demonstrated that the twins and stacking faults (SF) act as effective sites for blocking slipping and storing dislocations in a Cu–10 wt%Zn alloy with a low stacking-
fault energy. They determined that the dislocation density in grains with twins/SFs was 4–10 times greater than that for grains without any twins/SF. This result suggests that Fe–13Mn–1.3C could contain a higher dislocation density in comparison with the other two steels as the effective distance between twin boundaries increases with decreasing SFE. This is confirmed by X-ray diffraction studies (Fig. 5e) considering that measured dislocation density values are not very precise. Because processes of dislocation dissociation, recombination and cross-slip becomes difficult in materials with low SFE, these processes impede the recovery process and lead to a smaller minimum grain size under SPD [31,32]. The high-pressure torsion of the investigated steels at room temperature did not produce a microstructure with well-developed high-angle boundaries of a common type. Severe plastic deformation resulted in the fragmentation of the initial single crystal structure with microand nano-twins and shear bands. Nevertheless, this structure complies with the definition of an ultrafine-grained structure as it is fully homogeneous and an equiaxed structure, the average grain size (distance between twin boundaries in our case) is less than 1 μm, and the majority of grain boundaries have high-angles of misorientation (Σ3n). 4.2. The effect of dynamic strain aging on strengthening of high-Mn steels under HPT Strain aging is peculiar to steels that contain interstitial elements in solid solution, which segregate to dislocations and thus induce dislocation pinning. The effect of dynamic strain aging
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
171
Fig. 4. TEM images of twins in single crystals of Fe–28Mn–2.7Al–1.3C steel after HPT: (a, b) Ν¼ 1, bright-field image, SAED pattern and dark-field image in twin reflection; (c) Ν¼ 3, bright-field image and SAED pattern; (d, e) Ν¼ 5, bright-field image, SAED pattern, and dark-field image in twin reflection. SAED patterns were obtained from an area of 12.6 μm2.
(DSA) is one of the major factors in the hardening of Hadfield steel, which also contributes to the accumulation of a high density of dislocations under static deformation at room temperature [23,25,33,34]. Alloying with aluminum suppresses the DSA in polycrystalline high-Mn steels near room temperature [25]. For the investigated single crystals, the DSA regions were determined in tensile tests at temperatures where the strain-rate sensitivity β ¼ Δs=Δ ln ε_ (Δs is the change in stress when the strain rate varies) had zero or negative values as the strain rate ε_ was changed from 4 10 4 s–1 to 4 10–3 s 1 at a strain of o1% (Fig. 8). The interval of the negative SRS for the Fe–13Mn–1.3C steel is rather wide across room temperature; therefore, even assuming some local heating of the specimen under torsional processing, plastic strain is realized in the DSA region (Fig. 8a). This effect is clearly not appreciable for the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels under HPT at room temperature (Fig. 8b, c). The strain rate under HPT depends on the distance from the disk radius and can be calculated pffiffiffi based on the relationship for true equivalent strain ε ¼ 2π rN= 3h (N – number of revolutions, r – distance from the disk center, and h – disk thickness) [2] and rotation speed (1 rpm). For the given experimental conditions, the strain rate is 5 10 3 s 1 next to the disk center (r ¼ 0.05 mm) and is 5 10 1 s 1 on the disk periphery (r ¼5 mm). Therefore, the strain rate under HPT is appropriate for DSA. Based on these data, we assume that the excess dislocation density in the Fe–13Mn–1.3C steel compared to the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels also contributed to the DSA effect in addition to the SFE effect described in Section 4.1. It should also be
noted that the phenomenon of strain aging is described in the literature only for the case of dislocation glide, and the features of this process in the case of collective rotational deformation mechanisms are not clear and require a separate study. 4.3. Strain-induced γ–ε, γ–α0 transformations in steels Fe–13Mn–1.0C, Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C under HPT After high-pressure torsion, strain-induced martensite was found by X-ray analysis in the Fe–13Mn–1.0C steel, which possesses a lower carbon content compared to the other steels with 1.3 wt% of interstitials. The amount of ε-martensite after the first highpressure torsion revolution was not detected in the X-ray diffraction patterns. Three revolutions led to the appearance of the α0 phase and to an increase in the ε-phase fraction, as revealed by XRD in the Fe–13Mn–1.0C steel (Fig. 7). In steels with a higher carbon concentration (1.3 wt%), the martensite onset temperature is lower, and this suppresses the martensitic transformation in the Fe–13Mn–1.3C, Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels at room temperature. The (111) XRD lines are symmetric independent of strain for the Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, and this result confirms that there are no appreciable γ–α0 , γ–ε-phase transformations (Fig. 5a–c). Nevertheless, weak reflections corresponding to the ε phase are detected in the SAED patterns for steels with a high carbon content (1.3%) (Figs. 2f, 3f, and 4c). Teplov et al. [5] reported that in polycrystalline Fe–13Mn–1.2C deformed by cold HPT (p¼ 6–15 GPa, N¼1–36, T¼300 K), the ε phase was not observed during TEM and X-ray diffraction
172
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 5. X-ray diffraction patterns (a–c) and microstrain of the crystal lattice (Δd/d), coherently scattering regions (CSR) (d), dislocation density (e) versus strain in steels under investigation. Microhardness of steels (in the center of the disks radius) after high-pressure torsion (f).
measurements. According to their data, decarburization under highpressure torsion is approximately 0.3 wt%, and after torsion of the Fe–13Mn–1.2C steel, the high residual carbon content of approximately 0.9 wt% imparts austenite with stability against γ–α, γ–εmartensitic transformations [5]. Therefore, in the case of the lower carbon content Fe–13Mn–1.0C steel we used in our investigation, the carbon content could be less than 0.9 wt% after HPT and provides a strain-induced phase transformation. However, the loss of 0.3 wt% carbon in our case appears ambiguous because lower pressure and strain were used compared to [5], and no evidence of precipitates was found. According to XRD data, a change in the crystal lattice parameter for the steels with 1.3 wt% C under HPT for 5 revolutions at room temperature corresponds to the release of 0.2 wt% of carbon from the solid solution. The small fractions of ε and α0 phases could be the result of severe plastic deformation by HPT and some local inhomogeneity in the chemical composition in the single crystals, which is difficult to
remove by homogenization alone. Moreover, the strain-induced –ε-martensitic transformation becomes more likely with a decrease in SFE as both the grain size and carbon content decrease [26]. Lastly, as both twins and ε-martensite are results of the motion of a/6o2114 partial dislocations on {111}-planes in austenite [16,35], the violation in the sequence of partial dislocations under twin growth could lead to the occurrence of thin plates of ε-martensite. Notably, the proportions of the ε and α0 phases in the structure of the steel under investigation are low and have no noticeable effect on the hardening of the single crystals under HPT. For this reason, a strain-induced phase transformation could not also produce a difference in strain hardening for steels, as shown in Fig. 5f. 4.4. Variation in homogeneity across the disks after HPT A well-known limitation of the HPT process is that the strain and consequent strain hardening vary across the disk diameter [2].
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
173
Table 1 Influence of high-pressure torsion on twin fraction (f), twin thickness (t), distance between twins (e), and density of twin boundaries (ρtw) in the investigated highmanganese steels.
Fig. 6. Microhardness versus distance from the specimen center in the investigated steels: (a) Fe–13Mn–1.3C; (b) in Fe–13Mn–2.7Al–1.3C; (c) in Fe–28Mn–2.7Al–1.3C.
Fig. 7. X-ray diffraction patterns of Fe–13Mn–1.0C steel depending on the number of revolutions under high-pressure torsion.
e (nm)
ρtw, (m 2) (TEM)
Steel composition
N
f (%) (OM)
f (%) t (nm) (TEM)
Fe–13Mn–1.3C
1 3 5
40
27 21 30
7 32 30
14 20 25
5.2 1014 1.5 1014 2.5 1014
Fe–13Mn–2.7Al–1.3C
1 3 5
20
21 20 23
194 72 70
700 221 210
6.2 1012 2.3 1013 2.7 1013
Fe–28Mn–2.7Al–1.3C
1 3 5
20
33 18 19
50 57 22
350 130 60
1.1 1013 2.5 1013 8.0 1013
As a rule, under low stress and strain, there is a non-uniform distribution of microhardness values across the diameter of a sample, with significantly lower values in the center of the disks for metals and alloys with slow recovery and low SFEs. The relationship is inverse for HPT materials with a rapid recovery rate and high stacking-fault energy [2]. Furthermore, this difference decreases as the pressure and number of revolutions under HPT increase [2]. The results available to date suggest that complete homogeneity across the HPT fcc pure Ni and Al samples required torsional straining through more than 5 revolutions [2,36]. For austenitic/ferritic duplex stainless steel (C 0.017, Si 0.3, Mn 0.5, P 0.015, S 0.001, Ni 7.0, Cr 25, Mo 3.3, W 2.0, and N 0.28, wt%) processed by HPT (4–8 GPa, 1–20 rev.), Cao [37] found that after 5 revolutions of HPT, the hardness in the disk center is appreciably lower (6 GPa) than at the periphery (8 GPa) and that after 6 revolutions, the average hardness values at locations E0.5 mm away from the disk center become saturated. For austenitic TWIP-steel Fe– 24Mn–3%Al–2%Si–1%Ni–0.06%C, Matoso [12] reported that under HPT (6 GPa, RT) there are variations in the hardness distribution in the through-thickness direction of the disks up to 10 revolutions. According to their data, pronounced twinning was found in the TWIP-steel in the early stages of deformation, and twin boundaries cover almost the entire surface after processing for 1/4 revolution (with a minimal twin boundary spacing of E400 nm). Thereafter, the twin saturation was reached at a strain of 1 revolution, and the fraction of twined volumes decreased with strain [12]. In contrast to previous data, the Hadfield steel exhibits a highly homogeneous hardness distribution in the given HPT strains from 1 to 5 revolutions (Fig. 6a). This is a result of pronounced mechanical twinning and reduced twin thickness at a given strain. Differences in hardness homogeneity for the Hadfield steel and TWIP-steel [12] arise from the different twinning morphologies. The structure of the Hadfield steel contains a high concentration of carbon atoms, which results in a reduced twin thickness (5–15 nm) compared to TWIP-steel ( E400 nm [12]). For the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, it is reasonable to conclude that the slight inhomogeneity in hardness values around the disks (Fig. 6b, c) is primarily due to the increase in twin boundary spacing. It is reasonable to assume that the increase in strain and stress under torsion (N 45, P4 6 GPa) could destroy the twinning grid in the investigated steels while increasing the number of HPT revolutions. However, because the destruction of twin boundaries is accompanied with the formation of a nanocrystalline structure with a grain size of 10–15 nm [5], it is difficult to expect further inhomogeneity in the mechanical properties because, despite the differences in the mechanisms of fragment formation, the structural parameters (size of fragments) remain constant.
174
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
homogeneous and an equiaxed one, the majority of grain boundaries have high-angles of misorientation (Σ3n), and the average distance between twin boundaries is less than 1 μm. All three steels showed pronounced twinning across the disks, and the density of twin boundaries, twin thickness and mean distance between twins increased when the Fe–13Mn–1.3C steel was alloyed with aluminum and manganese. The microstructure formed under HPT in the Fe–13Mn–1.3C steel with a low stacking-fault energy is very homogeneous independent of the strain that provides a constant microhardness values through the disk diameter. A lower hardness value in the disk centers for the two Fe–13Mn–2.7Al–1.3C and Fe–28Mn– 2.7Al–1.3C steels with higher stacking-fault energies was revealed, which is due to changes in twin morphology. An additional increase in the dislocation density was found in the Fe–13Mn–1.3C steel compared with the two other steels, which was a result of torsion in the temperature range of dynamic strain aging and small values of the distance between twin boundaries. Small fractions of ε and α0 phases were found in single crystals of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C high-carbon austenitic steels as a result of severe plastic deformation by HPT (release of 0.2 wt% of carbon from the solid solution) and some initial local inhomogeneity of the single crystals.
Acknowledgments The authors wish to thank Professor Y. Chumlyakov for providing the single crystals and for fruitful discussions. This research was partially supported by the Russian Ministry of Education and Science (Contract no. 8749, 01.10.2012) and Russian President Scholarship (SP-4384.2013.1). References
Fig. 8. Variations of strain-rate sensitivity of flow stress (a–c) with temperature in steels under investigation. Data for polycrystalline steels (red symbols) are taken from [23,25]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
5. Conclusions Disks of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al– 1.3C high-carbon austenitic steels with different SFEs (30, 45 and 60 mJ/m2, respectively) were processed by HPT for 1–5 revolutions under an applied pressure of 5–6 GPa. The evolution of the deformation structure and hardening as a function of strain was investigated. High-pressure torsion did not produce a true ultrafine-grained microstructure with high-angle boundaries of a common type. It was found that HPT deformation results in the formation of an ultrafinegrained microstructure that includes a high fraction of twin boundaries and a high dislocation density. This microstructure is fully
[1] R.Z. Valiev, Y. Estrin, Z. Horita, T.G. Langdon, M.J. Zehetbauer, J. Mater. 58 (4) (2006) 33–39. [2] A.P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53 (2008) 893–979. [3] R.Z. Valiev, Yu V. Ivanisenko, E.F. Rauch, B. Baudelet, Acta Mater. 44 (12) (1996) 4705–4712. [4] Yu Ivanisenko, R.Z. Valiev, H.-J. Fecht, Mater. Sci. Eng. A 390 (2005) 159–165. [5] V.A. Teplov, L.G. Korshunov, V.A. Shabashov, R.I. Kuznetsov, V.P. Pilyugin, D.I. Tupitsa, Phys. Met. Metall. 66 (1988) 563–571 (in Russian). [6] E.G. Astafurova, M.S. Tukeeva, G.G. Zakharova, E.V. Melnikov, H.J. Maier, Mater. Charact. 62 (2011) 588–592. [7] O.V. Rybal’chenko, S.V. Dobatkin, L.M. Kaputkina, G.I. Raab, N.A. Krasilnikov, Mater. Sci. Eng. A 387–389 (2004) 244–248. [8] A. Etienne, B. Radiguet, C. Genevois, J.-M. Le Breton, R. Valiev, P Pareoge, Mater Sci. Eng. A 527 (2010) 5805–5810. [9] H. Wang, I. Shuro, M. Umemoto, H.-H. Kuo, Y. Todaka, Mater. Sci. Eng. A 556 (2012) 906–910. [10] Yu. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.-J. Fecht, Acta Mater. 51 (2003) 5555–5570. [11] E.G. Astafurova, S.V. Dobatkin, E.V. Naydenkin, S.V. Shagalina, G.G. Zakharova, Yu.F. Ivanov, Nanotechnol. Russ. 4 (1–2) (2009) 109–120. [12] M. Matoso, R.B. Figueiredo, M. Kawasaki, D.B. Santos, T.G. Langdon, Scr. Mater. 67 (2012) 649–652. [13] N.R. Tao, K. Lu, Scr. Mater. 60 (2009) 1039–1043. [14] Y.T. Zhu, X.Z. Liao, X.L Wu, Prog. Mater. Sci. 57 (2012) 1–62. [15] Z.W. Wang, Y.B. Wang, X.Z. Liao, Y.H. Zhao, E.J. Lavernia, Y.T. Zhu, Z. Horita, T.G. Langdon, Scr. Mater. 60 (2009) 52–55. [16] J.W. Christian, S. Mahajan, Prog. Mater. Sci. 39 (1995) 1–157. [17] S.M. Schlegel, S. Hopkins, M. Frary, Scr. Mater. 61 (2009) 88–91. [18] C.X. Huang, G. Yang, Y.L. Gao, S.D. Wu, Z.F. Zhang, Mater. Sci. Eng. A 485 (2008) 643–650. [19] I. Karaman, G.G. Yapici, Y.I. Chumlyakov, I.V. Kireeva, Mater. Sci. Eng. A 410–411 (2005) 243–247. [20] E.G. Astafurova, I.V. Kireeva, Y.I. Chumlyakov, H.J. Maier, H. Sehitoglu, Int. J. Mater. Res. 98 (2007) 144–149. [21] I. Karaman, H. Sehitoglu, K. Gall, Y.I. Chumlyakov, H.J. Maier, Acta Mater. 48 (2000) 1345–1359. [22] K.S. Raghavan, A.S. Sastri, M.J. Marcinkowski, Trans. Metall. Soc. AIME 245 (1969) 1569–1575. [23] Y.N. Dastur, W.C. Leslie, Metall. Trans. A 12 (1981) 749–759. [24] P.H. Adler, G.B. Olson, W.S. Owen, Metall. Trans. A 17 (1986) 1725–1737.
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
[25] B.K. Zuidema, D.K. Subramanyam, W.C. Leslie, Metall. Trans. A 18 (1987) 1629–1639. [26] A. Saeed-Akbari, J. Imlau, U. Prahl, W. Bleck, Metall. Mater. Trans. A 40 (2009) 3076–3090. [27] D.B. Williams, C.B. Carter, Transmission Electron Microscopy, Springer, USA (1996) 2009. [28] G.K. Williamson, R.E. Smallman, Philos. Mag. 1 (1956) 34–46. [29] K. Mumtaz, S. Sakahashi, J. Echigova, Y. Kamada, L.F. Zhang, H. Kikuchi, K. Ara, M. Sato, J. Mater. Sci. 39 (2004) 85–97. [30] J.D. Yoo, K.T. Park, Mater. Sci. Eng. A 496 (2008) 417–424.
175
[31] L. Balogh, T. Ungár, Y. Zhao, Y.T. Zhu, Z. Horita, C. Xu, T.G. Langdon, Acta Mater. 56 (2008) 809–820. [32] F.A. Mohamed, S.S. Dheda, Mater. Sci. Eng. A 558 (2012) 59–63. [33] W.S. Owen, M. Grujicic, Acta Mater. 47 (1999) 111–126. [34] D. Canadinc, C. Efstathiou, H. Sehitoglu, Scr. Mater. 59 (2008) 1103–1106. [35] K. Otsuka, C.M. Wayman (Eds.), Cambridge University Press, 1998. [36] M. Kawasaki, R.B. Figueiredo, T.G. Langdon, Acta Mater. 59 (2011) 308–316. [37] Y. Cao, Y.B. Wang, R.B. Figueiredo, L. Chang, X.Z. Liao, M. Kawasaki, W.L. Zheng, S.P. Ringer, T.G. Langdon, Y.T. Zhu, Acta Mater. 59 (2011) 3903–3914.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Microstructure and mechanical response of single-crystalline high-manganese austenitic steels under high-pressure torsion: The effect of stacking-fault energy E.G. Astafurova a,n, M.S. Tukeeva a, G.G. Maier a, E.V. Melnikov a, H.J. Maier b a Laboratory of Physical Materials Science, Institute of Strength Physics and Materials Science, Siberian Branch of Russian Academy of Science, Akademichesky Prospect 2/4, 634021 Tomsk, Russia b Institut für Werkstoffkunde (Materials Science), Leibniz Universität Hannover, 30823 Garbsen, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 10 January 2014 Received in revised form 5 March 2014 Accepted 7 March 2014 Available online 17 March 2014
We investigate the kinetics of the structural deformation and hardening of single-crystalline austenitic Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C (in wt%) steels with different stacking-fault energies after cold high-pressure torsion. Independently of the stacking-fault energy, mechanical twinning was found to be the basic deformation mechanism responsible for the rapid generation of an ultrafine-grained microstructure with a high volume fraction of twin boundaries. Under high-pressure torsion, the spacing between twin boundaries increases, and the dislocation density and microhardness decrease as the stacking-fault energy increases. The formation of a twin net from the beginning of plastic flow in Fe–13Mn–1.3C steel provides a homogeneous distribution of microhardness values across the discs independent of strain under torsion. Lower hardness values in the disk centers compared to the periphery were observed for the two other steels, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C, with higher stacking-fault energies due to changes in the densities of the twin boundaries. An additional increase in the dislocation density for the Fe–13Mn–1.3C steel was detected compared with the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, which was a result of torsion in the temperature range of dynamic strain aging. The appearance of small fractions of ε and α0 phases in the structures of the Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C steels is discussed. & 2014 Elsevier B.V. All rights reserved.
Keywords: Austenite Steel Twinning High-pressure torsion Microstructure Stacking-fault energy
1. Introduction In recent decades, the physical studies of strength and plasticity have been directed at developing methods for achieving highstrength states in metals by refinement of their structures under severe plastic deformation (SPD) [1]. Progress in this area is not only associated with the creation of new deformation schemes but also, to a large extent, with the modification and combination of known methods, the optimization of the composition and structure of alloys prior to SPD and with grain-boundary engineering. High-pressure torsion (HPT) refers to the processing technique in which samples in the form of thin disks are subjected to simultaneous high compressive stress and torsion [1,2]. Using this method, nanocrystalline structural states are typically formed in various metals and alloys [2]. The obvious advantage of this SPD method is the ability to realize unlimited strains, which are often
n
Corresponding author. Tel.: þ 7 3822 286961; fax: þ 7 3822 491245. E-mail address: [email protected] (E.G. Astafurova). URL: http://www.ispms.ru (E.G. Astafurova).
http://dx.doi.org/10.1016/j.msea.2014.03.029 0921-5093/& 2014 Elsevier B.V. All rights reserved.
not achievable in conventional static tests, while maintaining the integrity of the workpiece and its shape (volume). The principles of the HPT method are described in detail in [2]. The HPT method has successfully been used to deform Armco iron [3,4] and steels [5–12]. HPT of twinning-assisted austenitic steels was not studied in detail [5,6,12]. However, mechanical twinning can be one of the factors that contribute to the rapid formation of ultrafine-grained structures with low-energy, highangle special boundaries Σ3n [13]. At temperatures that do not allow recrystallization, severe plastic deformation typically causes degradation of the dislocation structure by forming bands of localized deformation. The development of fine deformation twins in steels with a high concentration of interstitial atoms hinders this process because the deformation twins are more resistant to “disintegrating” compared with general boundaries. Some authors found deformation twins in nanocrystalline facecentered cubic (fcc) metals and alloys (with grain sizes of a few tens of nanometers), such as Cu, Al, Ni, and Cu–Zn, due to the emission of partial dislocations from the grain boundaries, overlapping of stacking-fault ribbons, grain boundary splitting, etc. [14,15]. The volume fraction of twin boundaries formed in
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
nanostructured materials is not considerable; it contributes but does not define the hardening of such materials. Forming a structural state with the maximum possible fraction of special Σ3n boundaries is not a trivial task because twinning in fcc alloys is largely complementary to the slip deformation mechanism and is generally observed in the low-temperature deformation regime [16]. In contrast, SPD is often conducted at elevated temperatures. Therefore, the deformation of alloys prone to high-temperature twinning offers the potential to create highstrength nanostructured materials with low-energy high-angle misorientations between the structural elements and that are resistant to degradation during annealing or subsequent deformation [17]. For AISI 304L stainless steel processed by equal channel angular pressing in the temperature range of 500–900 1C, Huang and coauthors [18] observed deformation twins in addition to deformation bands and subgrains (many bundles of deformation twins was observed at temperatures below 700 1C). Karaman et al. [19] reported the development of high-temperature twinning (up to 800 1C) for some difficult-to-work alloys, including 316L austenitic steel. They concluded that deformation twinning could be one of the main deformation modes in many metallic alloys over a wide temperature range when a high strength level is reached, irrespective of the way it is achieved [19]. As promising materials for the development of SPD methods, high-manganese austenitic steels with low stacking-fault energies (Hadfield steel-based) can be used. The high strength levels are reached in these steels due to solid solution hardening (high carbon content). Twinning in single- and polycrystalline Fe–(11–13)Mn–(1.0–1.3)C and Fe–13Mn–(1.3–2.7)Al–(1.3–1.75)C steels develops under uniaxial tension/compression at room temperature [20–25]. Under torsional deformation and high applied pressure, twinning had a significant contribution to the refinement and strengthening of the Hadfield steels Fe–13Mn–1.0C [6] and Fe–13Mn–1.2C [5]. The aim of this study is to characterize the microstructure and strength properties (microhardness) after the high-pressure torsion of single crystalline twinning-assisted Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C steels. The choice of such steel compositions allows one to change the stacking-fault energies by alloying Hadfield steel with manganese and aluminum while maintaining a high level of solid solution hardening. The use of single crystals allows for a detailed study of the strain hardening mechanisms and avoids the added complexity of the contribution from grain boundaries.
2. Experimental The high-Mn steels used in this study had the following chemical compositions: Fe–13Mn–1.0C, Fe–13Mn–1.3C (Hadfield steel), Fe–13Mn–2.7Al–1.3C, and Fe–28Mn–2.7Al–1.3C (in wt%). Single crystals of the high-manganese austenitic steels were grown using the Bridgman technique under an inert gas atmosphere. All crystals were homogenized in an argon atmosphere at 1373 K for 24 h, solution-treated and water-quenched from 1373 K after 1 h. Electro-discharge machining was utilized to cut flat disks of 10 mm (diameter) 0.6 mm (height) for high-pressure torsion. The surfaces of the disks were cut parallel to the {001}-type planes for the Fe–13Mn–1.0C, Fe–13Mn–1.3C, and Fe–13Mn–2.7Al–1.3C steels and parallel to the {111}-type planes for the Fe–28Mn–2.7Al– 1.3C steel. The stacking-fault energies (SFE) for the investigated steels were assumed to be 25–30 mJ/m2 for Fe–13Mn–1.0C and Fe–13Mn–1.3C, 45 mJ/m2 for Fe–13Mn–2.7Al–1.3C, and 60 mJ/m2 for Fe–28Mn–2.7Al–1.3C according to [26]. The unconstrained high-pressure torsion was conducted at room temperature (p ¼5–6 GPa) for one to five revolutions at
167
a rotation speed of 1 rpm. The true logarithmic strain was calculated using the equation ε ¼ln(υr/h) [2], where r is the radius, υ is the rotation angle, and h is the thickness of the disk after highpressure torsion. Mechanical grinding and a final electrochemical polishing (50 g of CrO3 in 200 ml of H3PO4) were employed to remove all preparation-induced surface artifacts. The evolution of the microstructure was studied by repolishing and etching the specimens that were strained to different degrees. Chemical etching of the polished disk surfaces was conducted in a solution of 1 ml of HCl in 99 ml of H2O. Optical microscopy (OM) was performed using an Olympus GX-71 microscope. Electron-transparent foils were prepared by conventional electropolishing in a solution of 50 g CrO3 þ200 ml H3PO4 at an applied potential of 30 V. For microstructural analysis, transmission electron microscopy (TEM) was performed using a Philips CM200 transmission electron microscope. The foils for TEM were cut with the observation area located at half of the radius of the disks after high-pressure torsion. The average sizes of the microstructural elements (subgrains, twins) were determined using dark-field TEM images [27]. Selected area electron diffraction (SAED) patterns from areas of 0.5 and 12.6 μm2 were evaluated to identify the different phases. The density of twin boundaries was estimated as ρtw ¼ q/S, where q is the number of twin boundaries and S is the area of the TEM image. A Shimadzu XRD-6000 X-ray diffractometer with Cu-Kα radiation was utilized for the X-ray diffraction studies. The scalar dislocation density was determined by X-ray diffraction (XRD) using the method described in [28]. The microhardness was determined at room temperature using a Duramin 5 instrument with a load of 200 g.
3. Results 3.1. Microstructural observations High-pressure torsion leads to the fragmentation of single crystals of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C, and Fe–28Mn– 2.7Al–1.3C mainly due to the formation of a high dislocation density, twins and localized deformation bands, which were detected by optical metallography (Fig. 1) and under TEM studies (Figs. 2–4). Analysis of the crystal surface after repolishing and etching shows that there are several twinning systems and that the twins and shear bands are internally twinned (Fig. 1). The tendency for plastic flow to localize increases as the stacking-fault energy increases (Fig. 1). In the Fe–13Mn–1.3C steel, the localization processes are suppressed because, from the beginning of deformation, a twin net forms with a spacing between the boundaries of a few nanometers. In single crystals of the Fe– 28Mn–2.7Al–1.3C steel, localized shear bands were formed prior to or simultaneously with the development of deformation twinning (Fig. 1c). In the Fe–13Mn–1.3C steel at N ¼1 revolutions, there is a formation of twin packs with a micron width and shear bands (SB) (Fig. 1a). Using TEM, we observed twins in which the thicknesses of the plates were tens of nanometers (Fig. 2a, b). The effective size of the structural fragments is determined from the thickness of the deformation twins and the distance between the twin boundaries, which is 5–15 nm. With an increase in strain (N ¼3) in the Fe–13Mn–1.3C steel, the twin net remains; however, in the TEM and optical images, deformation and a large number of microbands were detected (Figs. 1d and 2c, d). Twins and shear bands with the fragment widths of 0.5 mm are filled with deformation twins with a thickness of 5–10 nm (Fig. 2c). Fragments within such a twin net have a high density of slip
168
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 1. Optical images of the etched surfaces of single crystals of Fe–13Mn–1.3C (a, d, g), Fe–13Mn–2.7Al–1.3C (b, e, h), and Fe–28Mn–2.7Al–1.3C (c, f, i) after HPT to different strains. The number of revolutions is shown in the figures.
dislocations ρsl 1015 m 2. When N ¼ 5, the twin net in the Fe– 13Mn–1.3C steel is significantly destroyed. On the etched surfaces of the disks, distorted twin traces are observed (Fig. 1g). The average size of the structural elements observed in the dark-field TEM images is approximately 100 nm (Fig. 2e). Within these structural elements, fragments of twin boundaries are observed, and characteristic twin reflections are observed in the SAED patterns (Fig. 2f). TEM observations of the Fe–13Mn–1.3C steel after high-pressure torsion at room temperature revealed a small fraction of ε-martensite (diffuse reflections observed in the SAED patterns at N ¼3–5; see Fig. 2f). High pressures applied during the given experimental technique (6 GPa) lead to the development of mechanical twinning in single-crystalline Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels with medium to high SFE. The presence of twinning was confirmed both metallographically and under TEM studies (Figs. 1, 3 and 4). Consequently, due to the high compressive stress under high-pressure torsion, twinning stresses in these steels are reached at room temperature, which often cannot be realized in experiments on static tension and compression. The proportion of twinned volumes after high-pressure torsion (N ¼ 1) is approximately 40%, as determined from the images of the surfaces of single crystals after repolishing and etching, and it does not change in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels compared to the Fe–13Mn–1.3C steel. However, the morphological features of the twins, the localization and strain hardening differ from those of the Hadfield steel. The bright- and dark-field TEM images for the Fe–13Mn–2.7Al– 1.3C and Fe–28Mn–2.7Al–1.3C steels after high-pressure torsion indicate that for N ¼1–3, the width of the twins and the size of “cells” limited by twin lamellae are higher than for the Hadfield
steel: 5–500 nm for Fe–13Mn–1.3C; 100–700 nm for Fe–13Mn– 2.7Al–1.3C; and 50–350 nm for Fe–28Mn–2.7Al–1.3C (Figs. 3a and 4a–c). When N¼3–5, the twins are fragmented and twin boundaries are observed over the entire volume of the samples (Figs. 3b–e and 4c–e). The reflections in the SAED patterns for the Fe–13Mn–1.3C, Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels are significantly diffused but not confined to the rings (Figs. 3 and 4), which is often observed after torsional strain [2]. One can distinguish a bright textural reflection in the SAED patterns, which also indicates the formation of deformation texture under HPT. 3.2. X-ray studies High-pressure torsion leads to line broadening in the X-ray diffraction patterns of the steels under investigation, but regardless of the number of revolutions, this method revealed a sharp deformation texture (Fig. 5a–c). Measurement of the coherent scattering regions provides the values of the structural parameters after high-pressure torsion of o20 nm (Fig. 5d). In the Fe–28Mn– 2.7Al–1.3C steel, the coherent scattering regions vary slightly slower with the strain, and the dislocation density is 2–3 times lower compared to the other two steels (Fig. 5d, e). Of the three studied steels, a maximal growth in the crystal lattice microstrain and dislocation density with increasing number of revolutions was observed in single crystals of Fe–13Mn–1.3C (Fig. 5d, e). 3.3. The microhardness tests High-pressure torsion caused a high strain hardening in the austenitic steels. The formation of highly nonuniform states in
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
169
Fig. 2. TEM images of twins in single crystals of Fe–13Mn–1.3C steel after HPT: (a, b) Ν¼ 1, bright-field image, SAED pattern, and dark-field image in twin reflection; (c, d) Ν¼ 3, bright-field image, SAED pattern, and dark-field image in matrix reflection; (e, f) Ν¼ 5, bright-field image, dark-field image using a twin reflection, and SAED pattern. SAED patterns were obtained from an area of 0.5 μm2.
steels under high-pressure torsion is accompanied by a significant increase in the microhardness from 2.5 GPa in the initial states up to 6.4–7.8 GPa after torsion for five revolutions (Fig. 5f). As shown in Fig. 5f, the microhardness of the Fe–13Mn–1.3C and Fe–13Mn– 2.7Al–1.3C steels after N ¼3 reaches a maximum value and then changes slowly as the strain increases. Work hardening in the Fe–13Mn–1.3C steel is higher than that in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels; thus, the Hadfield steel has the highest microhardness values after torsion (Fig. 5f). The microstructure formed under HPT in the Fe–13Mn–1.3C steel with a low SFE is very homogeneous independent of strain, which provides constant microhardness values through the diameter of the disk. Detailed Vickers microhardness measurements revealed lower hardness values in the disk centers for the Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels with a higher SFE (Fig. 6). 3.4. The influence of carbon content in Hadfield steel on phase composition and hardening TEM observations of the Fe–13Mn–1.3C steel (as two other steels) after high-pressure torsion at room temperature revealed a small fraction of ε-martensite (blurry reflections observed in the SAED patterns at N ¼3–5; see Fig. 2f). In single crystals of the Hadfield steel, which has a lower carbon content of Fe–13Mn–1.0C, the reflections corresponding to ε-martensite are clearly visible in SAED patterns after upset (pressing without rotation of a plunger) [6]. With an increase in strain, the proportion of martensite increases in the Fe–13Mn–1.0C steel, and after high-pressure
torsion for 2–3 revolutions, X-ray reflections corresponding to the ε-phase appear (Fig. 7). In single crystals of Fe–13Mn–1.3C, even after torsion for 5 revolutions, ε and α0 phases were not detected during the X-ray studies. The α0 phase volume fraction estimated using magnetization measurements [29] was not greater than 0.8% independent of the steel composition and strain.
4. Discussion 4.1. Fragmentation mechanisms for high-Mn steels under high-pressure torsion: the influence of stacking-fault energy Cold severe plastic deformation of single-crystalline Fe–13Mn– 1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels is associated with the formation of boundaries of a common and special type – twin – and their interaction and degradation with an increase in the number of revolutions. The tendency for plastic flow to localize is apparent in the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels compared to the Hadfield steel and is a result of the SFE increase and micro-band plasticity, as it appeared in high-Mn–Al–C austenitic steels with high SFE (E85 mJ/m2) [30]. The development of mechanical twinning in single crystals of the Fe–13Mn–1.3C and Fe–13Mn–2.7Al–1.3C steels under highpressure torsion is consistent with data obtained in [20,21]. Under tensile deformation at room temperature, twinning develops in Fe–13Mn–1.3C single crystals from the early stages of deformation
170
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 3. Multiple twinning in the single crystals of Fe–13Mn–2.7Al–1.3C steel after HPT: (a) Ν¼ 1, bright-field image, SAED pattern obtained from the area of 12.6 μm2; (b) Ν¼ 5, bright-field image; (c) Ν¼ 5, dark-field image in combined (111)γ and (120)ε reflections; (d) Ν¼5, dark-field image in combined (200)γ and (1 2 1)ε reflections; (e) Ν ¼5, dark-field image in twin (111)γ reflection; (f) Ν ¼5, SAED pattern for (b–e). SAED patterns were obtained from an area of 12.6 μm2.
(ε 40.5%, depending on the initial orientation of the single crystal) and in o111 4-, o011 4-oriented specimens of the Fe–13Mn– 2.7Al–1.3C steel (ε 4 10%) [20,21]. The initially single-crystalline Hadfield steel specimen under stress transforms to a polycrystalline specimen with continuous misorientation filled with slip dislocations and twins of 5–15 nm (Table 1). Therefore, the structural elements are restricted by twin boundaries – mainly high-angle Σ3n boundaries of a special type. The increase in the twin fraction under plastic flow is mainly due to the nucleation of new thin twins. This is characteristic for the deformation twinning of high-alloyed substitution and interstitial solid solutions [16]. The multiplicity of twinning is attributed to a quasihydrostatic stress state under HPT. An increase in SFE leads to increased twinning stresses [16]; however, under HPT for the investigated steels, they are clearly reached in the stress process for the Fe–13Mn–1.3C and Fe–13Mn– 2.7Al–1.3C steels or very beginning of rotation under HPT for the Fe–28Mn–2.7Al–1.3C steel. The effective distance between the twin boundaries (twin thickness t and distance between twins e) increases with SFE during the alloying of Hadfield steel with Al and Mn (in steels Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C, see Table 1) and simultaneously reduces the effectiveness of reinforcement and strain-hardening (Fig. 5f). Consequently, the shear accommodation and energy dissipation processes may play a significant role under increasing SFE of high-strength steels. Wang and coauthors [15] demonstrated that the twins and stacking faults (SF) act as effective sites for blocking slipping and storing dislocations in a Cu–10 wt%Zn alloy with a low stacking-
fault energy. They determined that the dislocation density in grains with twins/SFs was 4–10 times greater than that for grains without any twins/SF. This result suggests that Fe–13Mn–1.3C could contain a higher dislocation density in comparison with the other two steels as the effective distance between twin boundaries increases with decreasing SFE. This is confirmed by X-ray diffraction studies (Fig. 5e) considering that measured dislocation density values are not very precise. Because processes of dislocation dissociation, recombination and cross-slip becomes difficult in materials with low SFE, these processes impede the recovery process and lead to a smaller minimum grain size under SPD [31,32]. The high-pressure torsion of the investigated steels at room temperature did not produce a microstructure with well-developed high-angle boundaries of a common type. Severe plastic deformation resulted in the fragmentation of the initial single crystal structure with microand nano-twins and shear bands. Nevertheless, this structure complies with the definition of an ultrafine-grained structure as it is fully homogeneous and an equiaxed structure, the average grain size (distance between twin boundaries in our case) is less than 1 μm, and the majority of grain boundaries have high-angles of misorientation (Σ3n). 4.2. The effect of dynamic strain aging on strengthening of high-Mn steels under HPT Strain aging is peculiar to steels that contain interstitial elements in solid solution, which segregate to dislocations and thus induce dislocation pinning. The effect of dynamic strain aging
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
171
Fig. 4. TEM images of twins in single crystals of Fe–28Mn–2.7Al–1.3C steel after HPT: (a, b) Ν¼ 1, bright-field image, SAED pattern and dark-field image in twin reflection; (c) Ν¼ 3, bright-field image and SAED pattern; (d, e) Ν¼ 5, bright-field image, SAED pattern, and dark-field image in twin reflection. SAED patterns were obtained from an area of 12.6 μm2.
(DSA) is one of the major factors in the hardening of Hadfield steel, which also contributes to the accumulation of a high density of dislocations under static deformation at room temperature [23,25,33,34]. Alloying with aluminum suppresses the DSA in polycrystalline high-Mn steels near room temperature [25]. For the investigated single crystals, the DSA regions were determined in tensile tests at temperatures where the strain-rate sensitivity β ¼ Δs=Δ ln ε_ (Δs is the change in stress when the strain rate varies) had zero or negative values as the strain rate ε_ was changed from 4 10 4 s–1 to 4 10–3 s 1 at a strain of o1% (Fig. 8). The interval of the negative SRS for the Fe–13Mn–1.3C steel is rather wide across room temperature; therefore, even assuming some local heating of the specimen under torsional processing, plastic strain is realized in the DSA region (Fig. 8a). This effect is clearly not appreciable for the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels under HPT at room temperature (Fig. 8b, c). The strain rate under HPT depends on the distance from the disk radius and can be calculated pffiffiffi based on the relationship for true equivalent strain ε ¼ 2π rN= 3h (N – number of revolutions, r – distance from the disk center, and h – disk thickness) [2] and rotation speed (1 rpm). For the given experimental conditions, the strain rate is 5 10 3 s 1 next to the disk center (r ¼ 0.05 mm) and is 5 10 1 s 1 on the disk periphery (r ¼5 mm). Therefore, the strain rate under HPT is appropriate for DSA. Based on these data, we assume that the excess dislocation density in the Fe–13Mn–1.3C steel compared to the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels also contributed to the DSA effect in addition to the SFE effect described in Section 4.1. It should also be
noted that the phenomenon of strain aging is described in the literature only for the case of dislocation glide, and the features of this process in the case of collective rotational deformation mechanisms are not clear and require a separate study. 4.3. Strain-induced γ–ε, γ–α0 transformations in steels Fe–13Mn–1.0C, Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C under HPT After high-pressure torsion, strain-induced martensite was found by X-ray analysis in the Fe–13Mn–1.0C steel, which possesses a lower carbon content compared to the other steels with 1.3 wt% of interstitials. The amount of ε-martensite after the first highpressure torsion revolution was not detected in the X-ray diffraction patterns. Three revolutions led to the appearance of the α0 phase and to an increase in the ε-phase fraction, as revealed by XRD in the Fe–13Mn–1.0C steel (Fig. 7). In steels with a higher carbon concentration (1.3 wt%), the martensite onset temperature is lower, and this suppresses the martensitic transformation in the Fe–13Mn–1.3C, Fe– 13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels at room temperature. The (111) XRD lines are symmetric independent of strain for the Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, and this result confirms that there are no appreciable γ–α0 , γ–ε-phase transformations (Fig. 5a–c). Nevertheless, weak reflections corresponding to the ε phase are detected in the SAED patterns for steels with a high carbon content (1.3%) (Figs. 2f, 3f, and 4c). Teplov et al. [5] reported that in polycrystalline Fe–13Mn–1.2C deformed by cold HPT (p¼ 6–15 GPa, N¼1–36, T¼300 K), the ε phase was not observed during TEM and X-ray diffraction
172
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
Fig. 5. X-ray diffraction patterns (a–c) and microstrain of the crystal lattice (Δd/d), coherently scattering regions (CSR) (d), dislocation density (e) versus strain in steels under investigation. Microhardness of steels (in the center of the disks radius) after high-pressure torsion (f).
measurements. According to their data, decarburization under highpressure torsion is approximately 0.3 wt%, and after torsion of the Fe–13Mn–1.2C steel, the high residual carbon content of approximately 0.9 wt% imparts austenite with stability against γ–α, γ–εmartensitic transformations [5]. Therefore, in the case of the lower carbon content Fe–13Mn–1.0C steel we used in our investigation, the carbon content could be less than 0.9 wt% after HPT and provides a strain-induced phase transformation. However, the loss of 0.3 wt% carbon in our case appears ambiguous because lower pressure and strain were used compared to [5], and no evidence of precipitates was found. According to XRD data, a change in the crystal lattice parameter for the steels with 1.3 wt% C under HPT for 5 revolutions at room temperature corresponds to the release of 0.2 wt% of carbon from the solid solution. The small fractions of ε and α0 phases could be the result of severe plastic deformation by HPT and some local inhomogeneity in the chemical composition in the single crystals, which is difficult to
remove by homogenization alone. Moreover, the strain-induced –ε-martensitic transformation becomes more likely with a decrease in SFE as both the grain size and carbon content decrease [26]. Lastly, as both twins and ε-martensite are results of the motion of a/6o2114 partial dislocations on {111}-planes in austenite [16,35], the violation in the sequence of partial dislocations under twin growth could lead to the occurrence of thin plates of ε-martensite. Notably, the proportions of the ε and α0 phases in the structure of the steel under investigation are low and have no noticeable effect on the hardening of the single crystals under HPT. For this reason, a strain-induced phase transformation could not also produce a difference in strain hardening for steels, as shown in Fig. 5f. 4.4. Variation in homogeneity across the disks after HPT A well-known limitation of the HPT process is that the strain and consequent strain hardening vary across the disk diameter [2].
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
173
Table 1 Influence of high-pressure torsion on twin fraction (f), twin thickness (t), distance between twins (e), and density of twin boundaries (ρtw) in the investigated highmanganese steels.
Fig. 6. Microhardness versus distance from the specimen center in the investigated steels: (a) Fe–13Mn–1.3C; (b) in Fe–13Mn–2.7Al–1.3C; (c) in Fe–28Mn–2.7Al–1.3C.
Fig. 7. X-ray diffraction patterns of Fe–13Mn–1.0C steel depending on the number of revolutions under high-pressure torsion.
e (nm)
ρtw, (m 2) (TEM)
Steel composition
N
f (%) (OM)
f (%) t (nm) (TEM)
Fe–13Mn–1.3C
1 3 5
40
27 21 30
7 32 30
14 20 25
5.2 1014 1.5 1014 2.5 1014
Fe–13Mn–2.7Al–1.3C
1 3 5
20
21 20 23
194 72 70
700 221 210
6.2 1012 2.3 1013 2.7 1013
Fe–28Mn–2.7Al–1.3C
1 3 5
20
33 18 19
50 57 22
350 130 60
1.1 1013 2.5 1013 8.0 1013
As a rule, under low stress and strain, there is a non-uniform distribution of microhardness values across the diameter of a sample, with significantly lower values in the center of the disks for metals and alloys with slow recovery and low SFEs. The relationship is inverse for HPT materials with a rapid recovery rate and high stacking-fault energy [2]. Furthermore, this difference decreases as the pressure and number of revolutions under HPT increase [2]. The results available to date suggest that complete homogeneity across the HPT fcc pure Ni and Al samples required torsional straining through more than 5 revolutions [2,36]. For austenitic/ferritic duplex stainless steel (C 0.017, Si 0.3, Mn 0.5, P 0.015, S 0.001, Ni 7.0, Cr 25, Mo 3.3, W 2.0, and N 0.28, wt%) processed by HPT (4–8 GPa, 1–20 rev.), Cao [37] found that after 5 revolutions of HPT, the hardness in the disk center is appreciably lower (6 GPa) than at the periphery (8 GPa) and that after 6 revolutions, the average hardness values at locations E0.5 mm away from the disk center become saturated. For austenitic TWIP-steel Fe– 24Mn–3%Al–2%Si–1%Ni–0.06%C, Matoso [12] reported that under HPT (6 GPa, RT) there are variations in the hardness distribution in the through-thickness direction of the disks up to 10 revolutions. According to their data, pronounced twinning was found in the TWIP-steel in the early stages of deformation, and twin boundaries cover almost the entire surface after processing for 1/4 revolution (with a minimal twin boundary spacing of E400 nm). Thereafter, the twin saturation was reached at a strain of 1 revolution, and the fraction of twined volumes decreased with strain [12]. In contrast to previous data, the Hadfield steel exhibits a highly homogeneous hardness distribution in the given HPT strains from 1 to 5 revolutions (Fig. 6a). This is a result of pronounced mechanical twinning and reduced twin thickness at a given strain. Differences in hardness homogeneity for the Hadfield steel and TWIP-steel [12] arise from the different twinning morphologies. The structure of the Hadfield steel contains a high concentration of carbon atoms, which results in a reduced twin thickness (5–15 nm) compared to TWIP-steel ( E400 nm [12]). For the Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C steels, it is reasonable to conclude that the slight inhomogeneity in hardness values around the disks (Fig. 6b, c) is primarily due to the increase in twin boundary spacing. It is reasonable to assume that the increase in strain and stress under torsion (N 45, P4 6 GPa) could destroy the twinning grid in the investigated steels while increasing the number of HPT revolutions. However, because the destruction of twin boundaries is accompanied with the formation of a nanocrystalline structure with a grain size of 10–15 nm [5], it is difficult to expect further inhomogeneity in the mechanical properties because, despite the differences in the mechanisms of fragment formation, the structural parameters (size of fragments) remain constant.
174
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
homogeneous and an equiaxed one, the majority of grain boundaries have high-angles of misorientation (Σ3n), and the average distance between twin boundaries is less than 1 μm. All three steels showed pronounced twinning across the disks, and the density of twin boundaries, twin thickness and mean distance between twins increased when the Fe–13Mn–1.3C steel was alloyed with aluminum and manganese. The microstructure formed under HPT in the Fe–13Mn–1.3C steel with a low stacking-fault energy is very homogeneous independent of the strain that provides a constant microhardness values through the disk diameter. A lower hardness value in the disk centers for the two Fe–13Mn–2.7Al–1.3C and Fe–28Mn– 2.7Al–1.3C steels with higher stacking-fault energies was revealed, which is due to changes in twin morphology. An additional increase in the dislocation density was found in the Fe–13Mn–1.3C steel compared with the two other steels, which was a result of torsion in the temperature range of dynamic strain aging and small values of the distance between twin boundaries. Small fractions of ε and α0 phases were found in single crystals of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al–1.3C high-carbon austenitic steels as a result of severe plastic deformation by HPT (release of 0.2 wt% of carbon from the solid solution) and some initial local inhomogeneity of the single crystals.
Acknowledgments The authors wish to thank Professor Y. Chumlyakov for providing the single crystals and for fruitful discussions. This research was partially supported by the Russian Ministry of Education and Science (Contract no. 8749, 01.10.2012) and Russian President Scholarship (SP-4384.2013.1). References
Fig. 8. Variations of strain-rate sensitivity of flow stress (a–c) with temperature in steels under investigation. Data for polycrystalline steels (red symbols) are taken from [23,25]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
5. Conclusions Disks of Fe–13Mn–1.3C, Fe–13Mn–2.7Al–1.3C and Fe–28Mn–2.7Al– 1.3C high-carbon austenitic steels with different SFEs (30, 45 and 60 mJ/m2, respectively) were processed by HPT for 1–5 revolutions under an applied pressure of 5–6 GPa. The evolution of the deformation structure and hardening as a function of strain was investigated. High-pressure torsion did not produce a true ultrafine-grained microstructure with high-angle boundaries of a common type. It was found that HPT deformation results in the formation of an ultrafinegrained microstructure that includes a high fraction of twin boundaries and a high dislocation density. This microstructure is fully
[1] R.Z. Valiev, Y. Estrin, Z. Horita, T.G. Langdon, M.J. Zehetbauer, J. Mater. 58 (4) (2006) 33–39. [2] A.P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53 (2008) 893–979. [3] R.Z. Valiev, Yu V. Ivanisenko, E.F. Rauch, B. Baudelet, Acta Mater. 44 (12) (1996) 4705–4712. [4] Yu Ivanisenko, R.Z. Valiev, H.-J. Fecht, Mater. Sci. Eng. A 390 (2005) 159–165. [5] V.A. Teplov, L.G. Korshunov, V.A. Shabashov, R.I. Kuznetsov, V.P. Pilyugin, D.I. Tupitsa, Phys. Met. Metall. 66 (1988) 563–571 (in Russian). [6] E.G. Astafurova, M.S. Tukeeva, G.G. Zakharova, E.V. Melnikov, H.J. Maier, Mater. Charact. 62 (2011) 588–592. [7] O.V. Rybal’chenko, S.V. Dobatkin, L.M. Kaputkina, G.I. Raab, N.A. Krasilnikov, Mater. Sci. Eng. A 387–389 (2004) 244–248. [8] A. Etienne, B. Radiguet, C. Genevois, J.-M. Le Breton, R. Valiev, P Pareoge, Mater Sci. Eng. A 527 (2010) 5805–5810. [9] H. Wang, I. Shuro, M. Umemoto, H.-H. Kuo, Y. Todaka, Mater. Sci. Eng. A 556 (2012) 906–910. [10] Yu. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.-J. Fecht, Acta Mater. 51 (2003) 5555–5570. [11] E.G. Astafurova, S.V. Dobatkin, E.V. Naydenkin, S.V. Shagalina, G.G. Zakharova, Yu.F. Ivanov, Nanotechnol. Russ. 4 (1–2) (2009) 109–120. [12] M. Matoso, R.B. Figueiredo, M. Kawasaki, D.B. Santos, T.G. Langdon, Scr. Mater. 67 (2012) 649–652. [13] N.R. Tao, K. Lu, Scr. Mater. 60 (2009) 1039–1043. [14] Y.T. Zhu, X.Z. Liao, X.L Wu, Prog. Mater. Sci. 57 (2012) 1–62. [15] Z.W. Wang, Y.B. Wang, X.Z. Liao, Y.H. Zhao, E.J. Lavernia, Y.T. Zhu, Z. Horita, T.G. Langdon, Scr. Mater. 60 (2009) 52–55. [16] J.W. Christian, S. Mahajan, Prog. Mater. Sci. 39 (1995) 1–157. [17] S.M. Schlegel, S. Hopkins, M. Frary, Scr. Mater. 61 (2009) 88–91. [18] C.X. Huang, G. Yang, Y.L. Gao, S.D. Wu, Z.F. Zhang, Mater. Sci. Eng. A 485 (2008) 643–650. [19] I. Karaman, G.G. Yapici, Y.I. Chumlyakov, I.V. Kireeva, Mater. Sci. Eng. A 410–411 (2005) 243–247. [20] E.G. Astafurova, I.V. Kireeva, Y.I. Chumlyakov, H.J. Maier, H. Sehitoglu, Int. J. Mater. Res. 98 (2007) 144–149. [21] I. Karaman, H. Sehitoglu, K. Gall, Y.I. Chumlyakov, H.J. Maier, Acta Mater. 48 (2000) 1345–1359. [22] K.S. Raghavan, A.S. Sastri, M.J. Marcinkowski, Trans. Metall. Soc. AIME 245 (1969) 1569–1575. [23] Y.N. Dastur, W.C. Leslie, Metall. Trans. A 12 (1981) 749–759. [24] P.H. Adler, G.B. Olson, W.S. Owen, Metall. Trans. A 17 (1986) 1725–1737.
E.G. Astafurova et al. / Materials Science & Engineering A 604 (2014) 166–175
[25] B.K. Zuidema, D.K. Subramanyam, W.C. Leslie, Metall. Trans. A 18 (1987) 1629–1639. [26] A. Saeed-Akbari, J. Imlau, U. Prahl, W. Bleck, Metall. Mater. Trans. A 40 (2009) 3076–3090. [27] D.B. Williams, C.B. Carter, Transmission Electron Microscopy, Springer, USA (1996) 2009. [28] G.K. Williamson, R.E. Smallman, Philos. Mag. 1 (1956) 34–46. [29] K. Mumtaz, S. Sakahashi, J. Echigova, Y. Kamada, L.F. Zhang, H. Kikuchi, K. Ara, M. Sato, J. Mater. Sci. 39 (2004) 85–97. [30] J.D. Yoo, K.T. Park, Mater. Sci. Eng. A 496 (2008) 417–424.
175
[31] L. Balogh, T. Ungár, Y. Zhao, Y.T. Zhu, Z. Horita, C. Xu, T.G. Langdon, Acta Mater. 56 (2008) 809–820. [32] F.A. Mohamed, S.S. Dheda, Mater. Sci. Eng. A 558 (2012) 59–63. [33] W.S. Owen, M. Grujicic, Acta Mater. 47 (1999) 111–126. [34] D. Canadinc, C. Efstathiou, H. Sehitoglu, Scr. Mater. 59 (2008) 1103–1106. [35] K. Otsuka, C.M. Wayman (Eds.), Cambridge University Press, 1998. [36] M. Kawasaki, R.B. Figueiredo, T.G. Langdon, Acta Mater. 59 (2011) 308–316. [37] Y. Cao, Y.B. Wang, R.B. Figueiredo, L. Chang, X.Z. Liao, M. Kawasaki, W.L. Zheng, S.P. Ringer, T.G. Langdon, Y.T. Zhu, Acta Mater. 59 (2011) 3903–3914.