Fig. 1. Determination of intrinsic stacking-fault forming probability via electron diffraction shift method used in this study: (a) A sketch of typical selected area electron diffraction (SAED) pattern including face-centered cubic (FCC, gold spheres) and hexagonal close-packed (HCP, green spheres) spots, designated as (hkl)FCC planes on h -k = 3n + 1 (blue) and h -k = 3n -1 (red) along [011]FCC zone axis. (b) A sketch of the φ and ψ angles in the SAED patterns, showing that the (111)FCC diffraction reflection deviates from its original ψ = 70.52°, and thereby, it gradually comes close to the (101)HCP ε-martensite diffraction spot (ψ = 62.06°) in the case of h -k = 3n -1 [25,26]. (c) Relationship between the measured ψ angle and the stacking fault formation probability, as developed by Huang and Wang et al. [25,26]. The measured ψ angle → 54.74° indicates stacking fault formation probability → 0 in the case of h -k = 3n + 1 (blue curve), meaning that no stacking faults (or thermally-induced ε-martensite) appear in FCC structure.

Fig. 2. Correlative analysis of the location of carbon in Fe-17Mn-1.56C (at%) steel subjected to annealing at 1150 °C followed by air cooling: (a) Bright-field transmission electron microscopy (TEM) image of an atom probe tomography (APT) tip taken of a region containing a high-angle grain boundary (GB, yellow dotted line). (b) APT-reconstructed carbon map and the rotated view of the same as the TEM-observed tip shown in (a) The GB in the APT map was the same as the TEM-observed GB. (c) APT-reconstructed carbon maps including GBs in the 17Mn-1.1C, 17Mn-0.4C, and 17Mn-0.06C (at%) steel samples. Note that the 17Mn-0.06C (at%) steel was designated as an undoped 17Mn specimen in this study because the carbon content of 0.06 at% signifies an impurity. (d) One-dimensional concentration profiles of carbon across the APT-observed GBs for all the steel samples.

Fig. 3. (a) Variation in the GB carbon content (red) and soluble carbon content (black) within grains determined via APT and as a function of nominal carbon-doping content. (b) Calculated concentration ratio of GB carbon to soluble carbon shown in (a). (c) Relationship between the GB coverage factor (βc, red) of carbon and the nominal carbon-doping content in carbon-doped 17Mn steels. Also included are βc (black) values from the previously reported in ferritic steels [53]. (d) Plot of the segregation free energy of carbon, ΔGcGB = ?R·T·ln (βc), where R is the gas constant, T is the temperature, and βc is the GB coverage factor of carbon shown in (c) against the inverse of temperature for carbon-doped 17Mn steels.

Fig. 4. HRTEM or HAADF-STEM images (top panel), Fast Fourier transform patterns (FFT; right corner), and the corresponding FFT images (bottom panel), obtained from (a) undoped 17Mn, (b) 17Mn-0.4C, (c) 17Mn-1.1C, (d) 17Mn-1.56C steel samples. All the samples are in the annealed state.

Fig. 5. (a) Variation in FCC interplanar d-spacings of {220} planes in the undoped Fe-17Mn steel system and its carbon-doped versions, determined from STEM techniques, as a function of nominally doped soluble C content. (b) Dependence of the FCC lattice parameter on chemical composition.

Fig. 6. Stacking fault formation probabilities in undoped and carbon-doped 17Mn steels in the annealed state, determined from ψ angle measurements via electron diffraction: (a) typical SAED patterns along the [110]FCC zone axis, obtained from co-existing regions of austenite and thermally induced ε-martensite phases in the steels. All the electron patterns included FCC (yellow spheres and arrows) and HCP (green arrows) spots, showing the ψ angle between the diffraction spot of the FCC structure and transmitted spot. (b) Relationship between the ψ angle and the nominal carbon-doping content. (c) Variation in stacking-fault formation probability as a function of nominal carbon-doping content, determined from the relationship between the measured ψ angle and the stacking-fault forming probabilities in (b) and Fig. 1(c), respectively.

Fig. 7. (a) Representative low-magnified weak-beam dark-field TEM image and corresponding SAED pattern (right corner) showing plentiful intrinsic stacking faults, acquired from 17Mn-1.56C sample. (b) A sketch of intrinsic stacking faults for measuring the width of identified stacking faults. (c) Variation in stacking-fault width as a function of bulk carbon-doping content. (d) Variation in stacking-fault formation probability as a function of bulk carbon-doping content, determined from two different measurements, i.e., the stacking-fault width measurement and the ψ angle measurement in Fig. 6(c).

Fig. 8. Room-temperature SFE values of undoped and carbon-doped 17Mn steels, based on measurements of the stacking fault forming probabilities. The non-linear SFE variation as a function of nominal carbon-doping content is seen.

Fig. 9. (a) SFE map of Fe-Mn binary systems, determined from theoretical calculations or modeling (dotted lines) [20,26,63] and from TEM dislocation node method (continuous lines) [58-60]. The SFE value (9.9 ± 1.3 mJ/m2) of undoped 17Mn steel, estimated in this study, agrees with values estimated in two earlier theoretical studies via the supercell method [63] and thermodynamic considerations [30]. (b) SFE map of Fe-Mn-C ternary systems (by at%), determined from earlier theoretical calculations or modeling (dotted lines) [20,21,62] and experiments (continuous lines) [22,27].