Characterization of twin boundaries in an Fe–17.5Mn–0.56C twinning induced plasticity steel

Highlights

• Performed tensile tests to assess mechanical performance of TWIP alloy

• Analyzed tensile specimens using EBSD, TEM, and ECCI

• EBSD showed that most twinning occurred at or near the < 111 >//TA orientation.

• EBSD, TEM and ECCI were used to measure average twin density.

• Compared spatial resolution of EBSD, ECCI and TEM for the instrumentation used

Abstract

A twinning-induced plasticity steel of composition Fe–17.5 wt.% Mn–0.56 wt.% C–1.39 wt.% Al–0.24 wt.% Si was analyzed for the purpose of characterizing the relationship between tensile strain and deformation twinning. Tensile samples achieved a maximum of 0.46 true strain at failure, and a maximum ultimate tensile strength of 1599 MPa. Electron backscatter diffraction (EBSD) analysis showed that the grain orientation rotated heavily to < 111 > parallel to the tensile axis above 0.3 true strain. Sigma 3 misorientations, as identified by EBSD orientation measurements, and using the image quality maps were used to quantify the number of twins present in the scanned areas of the samples. The image quality method yielded a distinct positive correlation between the twin area density and deformation, but the orientation measurements were unreliable in quantifying twin density in these structures. Quantitative analysis of the twin fraction is limited from orientation information because of the poor spatial resolution of EBSD in relation to the twin thickness. The EBSD orientation maps created for a thin foil sample showed some improvement in the resolution of the twins, but not enough to be significant. Measurements of the twins in the transmission electron microscopy micrographs yielded an average thickness of 23 nm, which is near the resolution capabilities of EBSD on this material for the instrumentation used. Electron channeling contrast imaging performed on one bulk tensile specimen of 0.34 true strain, using a method of controlled diffraction, yielded several images of twinning, dislocation structures and strain fields. A twin thickness of 66 nm was measured by the same method used for the transmission electron microscopy measurement. It is apparent that the results obtain by electron channeling contrast imaging were better than those by EBSD but did not capture all information on the twin boundaries such as was observed by transmission electron microscopy.

Introduction

Twinning-induced plasticity (TWIP) steels came into existence in the late 1990s in the automotive industry's search for materials with advantageous combinations of mechanical properties. Modern automobiles are designed with safety factors such as crumple zones that absorb kinetic energy upon impact, which are meant to protect the occupants during a collision. For this reason the combination of ductility, strength, and energy absorption capacity under high deformation rates is highly desirable. Fully-austenitic TWIP steels strain harden during deformation while still maintaining more toughness and ductility than many other alloys such as dual phase (DP) steel, transformation-induced plasticity (TRIP) steels and other advanced high strength alloys [1], [2], [3], [4], [5], [6]. High yield and tensile strengths can be utilized for weight reduction by reducing material thickness, while ductility can be employed to accommodate increased complexity of component design. The strain hardening rates achievable during plastic deformation of TWIP alloys show a very high capacity for energy absorption during high-rate deformation [1], [2], [4], [7], [8], [9], [10], [11], [12]. Since their fruition, much effort has been made to characterize the properties and performance of TWIP steels to better understand their capabilities and how they may be tailored to meet the needs of different industries and applications.

One major influence on properties and performance of TWIP alloys is the stabilization of deformation twins at room temperature and during mechanical deformation, which is a function of the stacking fault energy (SFE). Annealing twins are also stabilized and are inherently present in the microstructure. The alloy composition, specifically the amounts of manganese (Mn), aluminum (Al), silicon (Si) and carbon (C), as well as deformation temperature, significantly affect the SFE and can be used to tailor which mechanisms, i.e. phase transformation, twinning, or dislocation slip, dominate deformation [4], [12]. The manganese content typically ranges from 15 to 35 weight percent (wt.%) and acts as the main austenite stabilizer. It has been reported that deformation by twinning is favored in the range of 16–40 mJ/m². Below that the TRIP effect dominates, whereby austenite (γ) changes to hexagonal close-packed (hcp) martensite (ε) and then to body-centered cubic (bcc) martensite (α′) [13]. When the deformation temperature is increased, the SFE is also increased to a level where dislocation slip dominates [12].

Another significant influence on performance is the interaction between dislocations and twin boundaries, as well as interactions between multiple twin systems [4]. When mechanical twins are formed, they impede dislocation motion by reducing the mean free path, increasing the strain hardening rate [2], [4], [5], [6], [7], [9], [12], [14], [15], [16]. Gutierrez-Urrutia et al. observed that deformation twins pass through dislocation cell structures as they form, creating more refined structures in the lattice [5]. As the twins thicken during the deformation process they also decreases the spacing between the individual twins. This refinement of the lattice structure is frequently referred to as the dynamic Hall–Petch effect [6], [9], [12], [13], [14], [15]. Zhu et al. suggested that dislocations of specific orientations can either grow or shrink the thickness of a twin boundary depending on the sample loading parameters and energy barriers, which are a function of the SFE, and may produce zigzag-like steps in the twin as it undergoes cross-slip [16]. Twin–dislocation interactions are very prominent in the lower strain levels, evidenced by a high strain hardening rate and serrations frequently present in the plastic region of the uniaxial stress–strain curve [9]. With higher levels of plastic strain, however, when multiple non-coplanar twin systems have formed in single parent grains, the twin systems begin to interact with one another, preventing further twinning and decreasing the strain hardening rate [4], [6], [9], [12]. Bundles of nano-scale twins and multiple twin systems have been observed in other studies and were expected to be observed here using various methods of electron microscopy [4], [5], [6], [7], [12], [15], [17].

Electron backscatter diffraction (EBSD) is a useful method for obtaining microstructural and orientation data, and has a spatial resolution of approximately 20–30 nm when using field emission sources [5], [18], [19]. Several research groups have successfully used EBSD to image twin bundles in different TWIP alloys [2], [9], [11], [14], [15]. The spatial resolution is limited when a large beam diameter is needed to deliver the necessary probe current, and a highly-deformed lattice structure, as is the case here, decreases the clarity of the diffraction patterns [18]. For this reason, transmission electron microscopy (TEM) is frequently used to image individual nano-scale twins within the bundles [2], [3], [7], [9], [14], [17], [20], [21], [22], [23], [24]. Diffraction spot patterns from TEM can also give useful information on lattice structures and defects. However, thin foil samples are necessary in order for the beam to transmit through the sample, which requires labor-intensive sample preparation.

Electron channeling contrast imaging (ECCI), also a SEM-based characterization technique, has been shown to achieve spatial resolutions much smaller than EBSD when using low accelerating voltage and small working distances (WD) in a low-tilt configuration, but with larger sampling areas and less sample preparation than TEM [5], [13], [19], [25], [27]. The backscattered electron yield, η, also called the electron channeling coefficient, is dependent on crystal orientation and diffraction. The further the primary electron beam can penetrate the specimen, the lower the η because the electrons are less likely to escape from the sample surface. However, if the crystal is oriented in the Bragg condition, penetration will be shallower. The BSEs will lose less energy and the diffraction contrast will be high [26]. Any change in orientation and atomic packing density within the grain, such as twins or bending of the lattice due to strain fields and dislocation structures will cause modulation in BSE yield, thus making it possible to image the defects [26], [27], [28], [29], [30].

For this study, by employing a controlled diffraction technique, explained in further detail in later sections, specific grains are oriented into the Bragg condition to image the deformation twins. Contrast is optimized by using an accelerating voltage between 10 keV and 15 keV, a high current, and a large aperture diameter. Surface deformation and oxidation due to mechanical preparation must also be removed so as not to obscure the microstructure [28]. The BSE yield and thus contrast is also better with a small WD, between 3 and 7 mm, and the sample surface nearly orthogonal to the incident beam [5], [13], [19], [25], [26], [27]. This study uses EBSD, ECCI, and TEM to quantitatively and qualitatively analyze deformation-induced twin boundaries.