Content uploaded by Hyokyung Sung
Author content
All content in this area was uploaded by Hyokyung Sung on Mar 27, 2022
Content may be subject to copyright.
Journal of Materials Science & Technology 117 (2022) 225–237
Contents lists available at ScienceDirect
Journal of Materials Science & Technology
journal homepage: www.elsevier.com/locate/jmst
A feasible route to produce 1.1 GPa ferritic-based low-Mn lightweight
steels with ductility of 47%
Kwang Kyu Ko
a , 1
, Hyo Ju Bae
a , 1
, Eun Hye Park
a
, Hyeon-Uk Jeong
b
, Hyoung Seok Park
c
,
Jae Seok Jeong
d
, Jung Gi Kim
a
, Hyokyung Sung
a
, Nokeun Parl
b , ∗, Jae Bok Seol
a , ∗
a
Department of Materials Engineering and Convergence Technology, Center for K-metal, Gyeongsang National University (GNU), Jinju 52828, South Korea
b
School of Materials Science and Engineering, Yeungnam University, Gyeongbuk 38541, South Korea
c
Materials Research Team R&D, Hyundai Mobis, Yong in 16891, South Korea
d
Materials Technology Development Team, Doosan Heavy Industries & Construction, Changwon 51711, South Korea
a r t i c l e i n f o
Article history:
Received 20 July 2021
Revised 10 November 2021
Accepted 25 November 2021
Available online 19 February 2022
Keywo rds:
Low-Mn lightweight steel
Carbon partitioning
Metastable austenite
Dislocation movement
a b s t r a c t
High- and medium-Mn (H/M-Mn) base lightweight steels are a class of ultrastrong structural materi-
als with high ductility compared to their low-Mn counterparts with low strength and poor ductility.
However, producing these H/M-Mn materials requires the advanced or high-tech manufacturing tech-
niques, which can unavoidably provoke labor and cost concerns. Herein, we have developed a facile
strategy that circumvents the strength–ductility trade-off in low-Mn ferritic lightweight steels, by em-
ploying low-temperature tempering-induced partitioning (LTP). This LTP treatment affords a typical Fe-
2.8Mn-5.7Al-0.3C (wt.%) steel with a heterogeneous size-distribution of metastable austenite embedded
in a ferrite matrix for partitioning more carbon into smaller austenite grains than into the larger austen-
ite ones. This size-dependent partitioning results in slip plane spacing modification and lattice strain,
which act through dislocation engineering. We ascribe the simultaneous improvement in strength and
total elongation to both the size-dependent dislocation movement in austenite grains and the controlled
deformation-induced martensitic transformation. The low-carbon-partitioned large austenite grains in-
crease the strength and ductility as a consequence of the combined martensitic transformation and
high dislocation density-induced hardening and by interface strengthening. Additionally, high-carbon-
partitioned small austenite grains enhance the strength and ductility by planar dislocation glide (in
the low strain regime) and by cross-slipping and delayed martensitic transformation (in the high strain
regime). The concept of size-dependent dislocation engineering may provide different pathways for de-
veloping a wide range of heterogeneous-structured low-Mn lightweight steels, suggesting that LTP may
be desirable for broad industrial applications at an economic cost.
©2022 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science &
Technology.
1. Introduction
Next-generation industrial technologies such as drones, ur-
ban air mobility system, and aerial transportation vehicles will
require highly sustainable structural metals that possess high-
strength/ductility and simultaneously have a reduced processing
cost [1–3] . These materials and their processing routes must meet
the critical demand for reducing anthropogenic CO
2
emissions be-
cause of circular economy [ 4 , 5 ].
Face-centered cubic (FCC, austenite)-based medium- or high-
Mn (M/H-Mn) lightweight steels including large amounts of Al ( >
∗Corresponding author at: Gyeongsang National University, Republic of Korea
E-mail addresses: nokeun_park@yu.ac.kr (N. Parl), jb.seol@gnu.ac.kr (J.B. Seol).
1 These authors contributed equally to this work.
5 wt.%) and Mn ( > 10 wt.%) are largely considered as potential
candidates for realizing these next-generation technologies [5–11] .
These M/H-Mn steels can achieve excellent tensile properties (ulti-
mate tensile strength levels of 80 0–150 0 MPa and total elongation
levels of 10%–80% [ 8 , 9 , 11 ]) by clever choices in alloy compositions
and processing routes, thus addressing the concerns related to the
well-known strength–ductility trade-off in these materials. This is
because the prominent twinning- or transformation-induced plas-
ticity (TWIP/TRIP) effects bestow them with mechanical proper-
ties superior to those of typical low-Mn body-centered cubic (BCC,
ferrite)-based lightweight series [12–14] . These outstanding prop-
erties drive the development of austenite-based lightweight steel
grade. Indeed, the M/H-Mn lightweight steels are remarkably bet-
ter than the tensile strength and elongation of ferrite-based low-
Mn steel series, which are 200–880 MPa (namely, below 1 GPa)
https://doi.org/10.1016/j.jmst.2021.11.052
1005-0302/© 2022 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
and 10%–38%, respectively. Despite such outstanding tensile prop-
erties, churning of the austenitic M/H-Mn alloy steels can belong
to the advanced and high-tech industrialization including welding
suitability and transporting (to carry the materials) facilitates, in
the existing general industrial production lines [ 11 , 15–17 ]. Since
these procedures are labor- or cost-intensive, using these to de-
velop the M/H-Mn steels unavoidably reduces the steel production
rates, and considerable efforts are being made to address these
limitations.
On the other hand, it is of great significance to tune the ex-
isting industrial routes capable of producing ultrastrong and duc-
tile ferrite-based lightweight steels (strength level > ∼1 GPa and
ductility > 40%) while maintaining low-Mn alloy compositions
and typically known lightweight steel compositions. Such low-Mn
lightweight steels, with bulk compositions of 2 wt.%–12 wt.% Mn,
3 wt.%–9 wt.% Al, and < 0.5 wt.% C, are strengthened by both the
coarse ferrite matrix and the metastable austenite grains [ 11 , 18 ]. In
the ferrite matrix, the solid solution strengthening by high amount
of Al increases the yield strength by ∼40 MPa/wt.% [7] . Meanwhile,
metastable austenite acts threefold: (i) high work-hardening capa-
bility stemming from the multiplication of dislocations inside lo-
cally strained austenitic grains [19] ; (ii) dynamic softening result-
ing from deformation-induced martensitic transformation via TRIP
effect [ 8 , 12 ]; and (iii) interface strengthening originating from hin-
dering slip transmission across phase boundaries during austenite
to martensitic transformations [ 12 , 14 , 20–22 ].
Unfortunately, alloying with high-Al compositions in the ferrite-
based low-Mn steels has two limitations for industrial applica-
tions. First, the thermal stability of austenite phase is drastically
reduced during thermomechanical processing at elevated temper-
atures, leading to a highly irregular (or heterogeneous) size distri-
bution of the metastable austenite grains during continuous casting
[ 23 , 24 ]. Second, the mechanical stability of the metastable austen-
ite grains can also deteriorate upon deformation [18] . Accordingly,
the TRIP effect may be manifested in an unpredictable manner due
to the microstructural inhomogeneity of the metastable austenite
grains embedded in the ferrite matrix, subsequently leading to the
decreased ultimate tensile strength ( < 1.0 GPa) [ 11 , 24 ]. Hence, the
development of a wide range of high-Al-containing ferrite-based
lightweight steel series with ultimate tensile strengths > 1.0 GPa
has remained elusive. Further, without improved choices in heat-
treatment strategies of typically known low-Mn lightweight steel
compositions, it will not be possible to alleviate the economic and
environmental concerns under the conventional industrial process-
ing conditions.
In this study, we show that simple tuning of heat-treatment
processing routes can enhance the tensile strength and total elon-
gation (strength of ∼1.1 GPa and ductility of ∼48%) of the com-
monly known ferrite-based low-Mn lightweight steel compositions.
This outstanding strength–ductility balance is comparable to the
M/H-Mn steel series that can require more expensive process-
ing technologies. We propose a strategy in which the strength–
ductility trade-off is surmounted by low-temperature-tempering-
induced partitioning (LTP) while decreasing the bulk Mn com-
positions to below 3 wt.%. Particularly, reducing the bulk Mn
composition is counterintuitive, as previous attempts on avoiding
the strength–ductility dilemma focused on the increasing the Mn
compositions to above ∼8 wt.%. Our LTP treatment allows the C
atoms to partition into the metastable austenite with heteroge-
neous size distributions. Although this strategy can be similar to
the previously reported strategies for the fabrication of deformed–
partitioned medium-Mn steel [ 25 , 26 ] and quenched–partitioned
low-Mn TRIP steel with less heterogeneous microstructure [27] , the
C partitioning described in this study entails the mechanical sta-
bilization of size-dependent austenite grains, associated slip plane
spacing modification, size-dependent dislocation movement inside
metastable austenite or mechanically transformed martensite, and
controlled TRIP effect.
2. Material and experimental procedures
2.1. Fabrication and thermo-mechanical processing
Alloy ingots with nominal compositions of typical ferrite-based
low-Mn lightweight steels, i.e., Fe-2.8Mn-5.7Al-(0.1 and 0.3)C by
wt.%, were produced. The alloy was solution-treated at 1200 °C for
90 min, and then hot-rolled to reduce the thickness reduction
by ∼55% at temperatures of 90 0–110 0 °C. The hot-rolled samples
were coiled at ∼650 °C for 60 min, followed by air-cooling to room
temperature at a rate of 10 °C/s. The coiled samples were cold-
rolled to reduce the thickness reduction by ∼70%. A sketch of the
thermo-mechanical processing is provided in Fig. S1(a) (Supple-
mentary materials) [28] . The cold-rolled microstructure of the steel
plates shows a lamellar or band-like distribution of alternating δ-
ferrite (as a matrix phase) and κ-carbide (38.6% in volume frac-
tion) along the rolling direction (RD), Fig. S1(b). As reference sam-
ples, the cold-rolled sheet plates with 0.1 wt.% C were subjected
to intercritical annealing (ICA) for 90 s at 850 and 950 °C. The in-
tercritically annealed steel specimens were processed by isother-
mal holding for 50 s at 430 °C, followed by air cooling to room
temperature. The low-Mn steel samples subjected to this thermo-
mechanical processing were produced commercially by POSCO, a
steel-making company. The final microstructure of the reference
sample (0.1 wt.% C) with a dual-phase microstructure, where the
κ-carbide phase was dissolved completely, is shown in Fig. S1(c).
The ICA temperatures used in this study were chosen to dissolve
the κ-carbide phase completely (Fig. S2), based on the thermody-
namic calculations of the quaternary Fe-Mn-Al-C systems [29] .
Hereafter, the annealed samples will be denoted as x C- y , where
x and y represent the bulk C composition (wt.%) and ICA temper-
ature ( °C), respectively. The annealed Fe-2.8Mn-5.7Al-0.1C (wt.%)
samples were designated as 0.1C-850, and 0.1C-950. Meanwhile,
the cold-rolled sheet plates with 0.3 wt.% C (i.e., Fe-2.8Mn-5.7Al-
0.3C wt.%) were subjected to the ICA for 90 s at 850 °C; this sample
was labeled as 0.3C-850. We performed LTP on the 0.3C-850 spec-
imen that was processed by hot-rolling, cold-rolling, ICA at 850 °C,
and isothermal holding, for 665 s at 330 °C. This sample was la-
beled as 0.3C-850-LTP. The key reasons for selecting the LTP tem-
perature are as follows:
When tempered at > ∼340 °C for 600 s, the metastable austenite
was decomposed into α-ferrite and κ-carbide while scarifying the
austenite fraction, as revealed by SEM images (Fig. S3); When tem-
pered at 330 °C, there were no significant changes in the size and
volume fraction of metastable austenite. These results confirmed
no significant microstructural evolution during LTP route, which is
consistent with our novel design approach.
2.2. Room-temperature tensile tests
Samples with a gauge length of 12 mm and a width of 4 mm
were cut from the position of one-quarter through-thickness of the
sheets. The tensile samples were prepared parallel to the rolling
direction (RD). The tensile tests for all specimens were performed
under uniaxial tension in an Instron 8861 test machine with a 100-
kN load cell operating at a constant initial nominal strain rate of
10
−4 s
−1 at room temperature. All tensile tests were performed
along the RD of steels. The strain during the tensile tests was mea-
sured by an extensometer with a 10 mm gauge till fracture. The
data points reported were averaged from the values of the three
tests. The mechanical properties of all specimens are characterized
by the yield strength, ultimate tensile strength, and strain to frac-
226
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 1. Microstructure of the current samples before tensile tests: (a) Typical EBSD pole-figure maps of metastable austenite ( γ) grains with heterogeneous size distributions
in ferrite-based low-Mn steel samples with compositions of Fe-2.8Mn5.7Al-0.1C (wt.%; left) and its 0.3 wt.% C analogue (right) [28] , processed by hot-rolling, cold-rolling,
intercritical annealing (ICA) at 850 °C, and isothermal annealing. These samples were denoted as 0.1C-850 and 0.3C-850, respectively. Heterogeneously layered
metastable γ
grains are embedded in δ-ferrite matrix, where transverse direction is normal to the plane view of the EBSD maps. RD: rolling direction, ND: normal direction. (b) APT-
reconstructed maps of the C-partitioned small ( ∼0.5
μm in diameter, left) and large γgrains (3.0 μm in diameter, right) for low-temperature partitioning (LTP)-treated steel
(0.3C-850-LTP). (c) APT-measured values of C and Mn concentrations, obtained from small and large
γgrains in the undeformed steel samples of 0.3C-850-LTP, 0.3C-850,
and 0.1C-850. (d) Representative APT mass-spectrum obtained from the LTP steel, showing all peaks are fully revolved.
ture (total elongation). The yield strength was taken as 0.2% offset
stress. The data points are averages of three or four test results.
2.3. Microstructure characterization
The mean grain sizes of metastable austenite grains for all the
specimens in the annealed (or undeformed) and deformed states
were determined by electron backscattered diffraction (EBSD) us-
ing five or more data sets in each case. All the specimens for
measurements were prepared using electropolishing and chemical
etching (Lectropol-5, Struers
TM
) using a solution of acetic acid (90
vol.%) and perchloric acid (10 vol.%). EBSD measurements were per-
formed at an acceleration voltage of 10 kV, sample tilt angle of 70 °,
working distance of 10 mm, aperture of 50 μm, and a step size of
50 nm. The data were post-processed using the TSL-Optical Imag-
ing Microscopy software [ 30 , 31 ]. A confidence index value greater
than 0.15 was chosen.
Transmission electron miscopy (TEM) samples were prepared
by mechanical polishing of thin foils down to a thickness of < ∼90
μm, followed by twin-jet technique in perchloric electrolytes (10%
perchloric acid in ethanol) at - 10 °C, and finally thinned by Ar
+
beam in a Precision Ion Polishing (GATAN) system [30] . For the
undeformed samples, the dislocation configuration inside the small
and large austenite grains were imaged by bright-field or dark-field
conventional TEM, wherein a parallel incident electron beam with
an acceleration voltage of 200 kV was used under the given low-
indexed diffraction conditions. In addition, for the bend-contour-
free dislocation imaging [ 32 , 33 ], annular bright-field (ABF) and an-
nular dark-field (ADF) techniques of the scanning (S)TEM mode
were employed on the deformed austenite grains. Because the con-
vergent electron beams in the STEM mode illuminate the thin TEM
sample, the dislocation contrast is enhanced and the elastic strain
contrast in the image background can be reduced [34–36] . Viewing
the dislocation configuration is easier under a simple operation of
the STEM conditions than the conventional TEM conditions [37] .
The STEM-ABF and -ADF images were obtained at a camera
length of 20 or 40 cm, and the corresponding collection angle
ranged from 35 to 93 mrad or from 17.5 to 46.7 mrad, respectively
[37] . The specimens were tilted using a double-tilt sample holder
for both the TEM and the STEM modes to satisfy either the sys-
tematic excitation condition or the two-beam excitation condition
that is appropriate for observing the dislocations [38–41] . There
are invisible or blurred dislocations that satisfy ( g
•b = 0), where
g and b denote a diffraction vector and Burgers vector of an ob-
served grain, respectively. In this study, we determined the dislo-
cation density of FCC-austenite and BCC-structures (including fer-
rite matrix together with deformation-induced martensite) in the
undeformed and deformed steels via synchrotron XRD patterns-
based peak-broadening approach, because TEM and STEM methods
for measuring dislocation density are limited to metals including
dislocation density < 10
14 cm
−2 [37] .
2.4. Dislocation density determination
Synchrotron XRD experiments were conducted on the electro-
polished samples in reflection mode at the BL5A beamline of the
Pohang Light Source II (PLS-II) in POSTECH, South Korea, so as to
determine the dislocation density of the annealed and deformed
steel samples. The energy of the monochromatic X-ray beam was
20 keV, which corresponds to a wavelength of 0.61985 ˚
A. The
beam had a spot diameter of 0.5 mm. Because the detailed dis-
location density measurements via synchrotron XRD are well de-
scribed in the literature [25] , we avoid the repetition; instead, we
addressed different features (Supplementary materials).
2.5. Carbon concentration measurement
Samples for atom probe tomography (APT) were prepared by
electropolishing, followed by treatment with a focused-ion beam
(FIB, FEI Helios Nano-Lab
TM
). The APT analyses were conducted us-
ing a local electrode atom probe (LEAP 40 0 0X HR, CAMECA
TM
) in
227
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 2. Room-temperature tensile properties of the current ferritic base duplex lightweight steels. Te ns ile properties for the LT P steel (0.3C-850-LTP) and non-LTP counterpart
(0.3C-850) with the same alloy composition are provided by red and blue curves [28] . Curves of oth er non-LTP steels with 0.1 wt.% C subject to the ICA at 850 °C (0.1C-850;
cyan) and 950 °C (0.1C-950; black) are also included. Bottom inset: Steel alloy compositions and heat-treatment processing strategies applied in this study. Loading direction
was parallel to the RD.
the voltage-pulsing mode. The experimental parameters were set
to maintain a 1.0% detection rate, 20% pulse fraction, and 125-kHz
pulse repetition. All measurements were performed at −223 °C and
< 10
–7 Pa. At least two successful measurements were performed
and evaluated for small and large austenite grains at each speci-
men type, two of which contained more than 5-million-collected
ions. The commercial IVAS® software (ver. 3.8.4) by Cameca was
employed following the previous protocol [42] to visualize the to-
mographic reconstruction of alloying elements. Statistical errors for
the measured atom counts were calculated as σ= ( C
i
×( 1 –
C
i
) /N )
–1/2
, where C
i
corresponds to the measured atomic concen-
tration fraction of the individual element i and N denotes the total
number of ions collected in the bin.
3. Results and discussion
3.1. Microstructure of undeformed steel samples
All the steel samples (0.3C-850-LTP, 0.3C-850, 0.1C-850, and
0.1C-950) exhibited dual-phase lamellar microstructures of layered
metastable austenite grains embedded in a coarse ferrite matrix.
The metastable austenite grains of the 0.3C-850 and 0.1C-850 sam-
ples without LTP treatment are heterogeneous, with a size distri-
bution ranging from 0.25 to 4.2 μm, as shown in Fig. 1 (a). For
both samples, 57% of the metastable austenite grains had diam-
eters of approximately 0.6 μm or less, while 39% had diameters
of 2.4–3.2 μm. According to this distribution, the austenite grains
are divided by small austenite grains (mean diameter of ∼0.5 μm)
and large grains (mean diameter of ∼3.1 μm). As expected, this
heterogeneity resulted from the layering of austenite grains along
the cold-rolled direction in the as-rolled state (see Fig. S1 and Refs
[ 11 , 18 , 24 , 28 ].). The volume fraction of metastable austenite grains
was estimated to be ∼23% (see Figs. S2 and S3). In line with our
proposed design concept, the LTP allowed more carbon to accu-
mulate in the small metastable austenite grains than in the large
grains because of the higher surface-to-volume ratio and shorter
diffusion path of the small-sized grains. The APT results support
this hypothesis, demonstrating that more C atoms partition into
the small austenite grains than into the large ones during the LTP
treatment ( Fig. 1 (b) and (c)). The Mn concentrations did not show
the trend of grain size dependency for a given heat treatment. As
shown in the mass spectrum ( Fig. 1 (d)), C peaks detected at 6,
6.5, 12, and 13 amu were assigned as a single count (C
2 + and C
+
),
while at 18, 18.5, 36 and 37 amu as triple counts (C
3
2 + and C
3
+
).
In the mass spectrum acquired from austenite phase, the peaks
at 24.0 amu were decomposed into C
2
+ and C
4
2 + because of the
C
4
2 + peak detection at 24.5 amu [43–45] . The detected peaks at
27 amu from the APT results of the present steel samples were
deconvoluted using the abundance ratios of Fe isotopes because
strong peak overlaps exist between
54
Fe
2 +
and Al
+
near 26.98 amu
in the APT mass spectrum [25] . To accurately determine the con-
centration of Al and Fe in the mass spectrum from the constituent
phases, a pure Fe metal was additionally analyzed by APT under
the same analysis parameters. The abundance ratios of Fe
2 + are
5.84% at 26.97 amu and 91.75% at 27.99 amu, as shown in Supple-
mentary materials Fig. S4(a). With these data, we determined the
contributions of Al
+
and Fe
2 +
at 27 amu in the APT spectrum; Fe
2 +
: Al
+
= 28 : 72 at 27 amu (Supplementary materials Fig. S4(b)). For
example, if 4609 ions were collected at 27 amu in the APT spec-
trum, then 12 91 and 3010 of those ions were Fe
2 +
and Al
+
, respec-
tively.
As previously reported by Frommeyer [6] who formulated the
density of austenite grain ρaustenitic
(g/cm
3
) = 8.15 –0.101Al –
0.41C – 0.0085Mn (by wt.%), the LTP reduces the density of small
and large austenite grains by ∼2.1% and ∼0.9%, respectively, as
compared to those in non-LTP steel sample. The LTP is very effec-
tive for the density reduction for the small grains than large ones,
i.e. the effectiveness for small grains is approximately two times
higher than that for the large ones.
228
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
This grain-size dependent partitioning strategy effectively re-
duces the diffraction angle of the (220)
FCC
plane peak, as revealed
by the synchrotron XRD in Fig. S5(a). The calculated interplanar
(220)
FCC
d -spacings of non-LTP (0.3C-850) sample and its LTP-
treated version (0.3C-850-LTP) were 0.12859 and 0.12880 nm, re-
spectively. Based on our APT and synchrotron XRD results, the ad-
dition of 1 at.% C to metastable austenite increases the interval
of (220)
FCC
d -spacing by ∼1.8 ×10
−4 nm. Further, by combin-
ing a general empirical equation [46] written as lattice parame-
ter a = 3.5780 + 0.033C + 0.0 0 095Mn + 0.0056Al + 0.0031Mo –
0.0 0 02Ni (by wt.%) and our APT results, the difference in the d -
spacing interval of the (111 )
FCC
slip planes in the small and large
grains was calculated to be 0.004573 nm and 0.019624 nm for
non-LTP (0.3C-850) and LTP-treated (0.3C-850-LTP) samples, re-
spectively. Thus, the difference of 1 at% C between the small and
large grains in the LTP sample increased the ( 111 )
FCC
d -spacings
by 0.01905 nm (see Fig. S5(b)). Consequently, the interplanar d -
spacing interval of the {111 }
FCC
planes—key dislocation slip planes
in most FCC structures—is effectively extended in the small austen-
ite grains through our LTP. This suggests that extending the dislo-
cation slip planes offers a potential to increase the mean glide dis-
tance of mobile dislocations, which can lead to increased plastic
strain (or severe lattice strain) owing to the general relationship
between plastic strain ( ε) and mobile dislocation density [ 25 , 47 ].
Furthermore, the enlarged dislocation slip plane interval provides
a potential for the easy passage of mobile dislocations in the small
austenite grains, which softens the present steel [48] . Hence, the
combined APT and synchrotron XRD results verify the basis of
our size-dependent C-partitioned steel design concept, i.e., more
C atoms partition into the small metastable austenite grains (mean
diameter of ∼0.5 μm) than into the large ones (mean diameter
of ∼3.1 μm) by LTP, resulting in the size-dependent severe lattice
strain effect and slip plane spacing modification.
3.2. Tensil e properties
Room-temperature engineering stress–strain curves of the cur-
rent low-Mn steels (0.3C-850-LTP, 0.3C-850, 0.1C-850, and 0.1C-
950) are presented in Fig. 2 . We begin with reducing the ICA
temperature from 950 to 850 °C of our low-Mn steel (i.e., Fe-
2.8Mn-5.7Al-0.1C intercritically annealed under 950 °C- and 850 °C-
ICA conditions) to improve the tensile properties (black vs. cyan
curves). Further, we increased the nominal C content from 0.1
to 0.3 wt.% (i.e., Fe-2.8Mn-5.7Al-0.3C intercritically annealed at
850 °C-ICA conditions) to enhance the tensile properties (blue vs.
cyan curves). The tensile properties of the LTP-treated 0.3C-850
sample (0.3C-850-LTP; red curve) improved notably, with ∼31% in-
crease in yield strength (from 610 to 798 MPa), ∼24% increase in
ultimate tensile strength (from 900 MPa to 1.12 GPa), and ∼11 %
increase in total elongation (from 42.5% to 47% in absolute values),
compared to those of the non-LTP specimen (0.3C-850; blue curve)
[ 28 ].
Consequently, our LTP steel (0.3C-850-LTP) exhibits an excel-
lent strength-ductility balance, highlighting the potential of the
proposed LTP treatment for resolving the strength–ductility trade-
off in typical lightweight steels ( Fig. 3 ). Our low-Mn LTP steel
can outperform the room-temperature strength–ductility combi-
nation (ultimate tensile strength ×uniform elongation) of pre-
vious lightweight steels [ 18 , 49 , 50 ] (e.g., nanobainitic steel [51] ),
medium-Mn lightweight alloys [ 22 , 24 ] (e.g. deformed and parti-
tioned steels [25] ), and other dual-phase low-Mn alloys [52] (e.g.
quenched and partitioned steels [53–55] ). The current LTP steel
without expensive raw elements such as Ni and Co shows better
strength–ductility combination than the nanoprecipitate-hardened
maraging medium-Mn (12 at.% Mn) TRIP steel [2] and ultrahard
Ni(Fe, Al)-maraging TRIP steel [3] . Moreover, the strength–ductility
balance of LTP steel is comparable to that of FCC-based materials,
such as high-Mn steels [ 19 , 53 ] (e.g. high-Mn TWIP steel [56] , high-
Mn stainless steels [ 53 , 57 ], nano-twinned steel [58] , high-Co (10
at.%) TRIP-assisted high-entropy alloy (HEA) [59] and recently de-
veloped high-Mn HEAs [ 60 , 61 ].
3.3. Dislocation density of LTP- treated low-Mn lightweight steel
The synchrotron XRD profiles of the 0.3C-850-LTP sample were
recorded as a function of the applied strain from the end of the
gauge region at different strain levels, i.e., ε= 0% (undeformed
state), 13.5% (deformed state), 25.2% (deformed state), and 47.1%
(fractured state). Based on the resultant synchrotron XRD profiles
in Fig. 4 (a), we calculated dislocation density during deformation
using the modified Warren–Averbach (MWA) method (see Supple-
mentary materials or Ref [62] . for details). BCC phases are consti-
tuted as a combination of ferrite matrix and deformation-induced
martensite. In principle, the ferrite and martensite phases can be
differentiated using separated XRD peaks at a high angle region.
However, in this study, integrated peaks appeared without separa-
tion due to a lack of tetragonality. Therefore, the dislocation den-
sity of the BCC phase was not differentiated individually. To de-
termine the distortion Fourier coefficients of the FCC-austenite and
BCC (ferrite matrix, together with deformation-induced martensite)
phases, each point was obtained by the MWA method from the
synchrotron XRD profiles, and the overall data showed excellent fit
to the Wilkens-model-based Berkum function ( Fig. 4 (b) and (c)).
This indicates that our dislocation calculations via synchrotron XRD
are reliable. In the undeformed LTP sample comprising austenite
and ferrite phases, the metastable austenite mostly contains edge
dislocations with a density of 3.13 ×10
15 m
−2
, whereas the ferrite
matrix includes screw dislocations with a density of 4.48 ×10
14
m
−2 ( Fig. 4 (d)). As the tensile strain extended beyond ε= 25.2%,
the density of screw dislocations inside austenite grains increased
substantially (blue curve). In contrast, our dislocation calculations
confirmed that the deformation-induced α
-martensite predomi-
nantly comprised screw dislocations, even after sample fracture.
This is because the lattice invariant shear can primarily induce
screw-component dislocations, implying that the prevailing screw
dislocations inside the BCC phases, shown in Fig. 4 (d), might be
generated predominantly from the austenite-to-martensite trans-
formation rather than from the deformed ferrite matrix. This hy-
pothesis becomes clear from the TEM image of the deformed fer-
rite matrix ( Fig. 6 ), which clarifies much higher density of dislo-
cations inside deformation-induced α
-martensite than inside the
deformed ferrite matrix.
In the fractured state ( ε= 47.1%), the total dislocation density
was calculated to be ∼5.42 ×10
16 m
−2
, with 4.17 ×10
16 m
−2 in
the metastable austenite and 1.25 ×10
16 m
−2 in the BCC phases
including martensite plus ferrite matrix. More specific, the dislo-
cation density in the deformed austenite grains calculated in this
study is approximately ten times higher than the previously re-
ported dislocation density (4.8 ×10
15
m
−2
) in an FCC-single phase
Fe-18Mn-0.55C (wt.%) TWIP steel even at similar plastic strain level
[63] . Such a high dislocation density in the deformation struc-
tures of the current ferritic steel system (comprising ferritic ma-
trix, metastable austenite, and mechanically transformed marten-
site) might be due to a combination of the strain partitioning effect
and displacive shear transformation in the absence of mechanical
twinning.
3.4. Transformation kinetics of metastable austenite
Based on the synchrotron XRD results ( Fig. 4 (a)), the volume
fraction of austenite (V
γ) was measured as a function of applied
229
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 3. Tens il e strength-ductility map including the values of alloy strength–ductility balance. Besides our LT P-tr eate d steel sample with a composition of Fe-2.8Mn5.7Al0.3C
(wt%), the tensile properties of several lightweight steels with bulk Mn compositions of < 5 wt% (low-Mn grade; gold clouds), 5–12 wt% (medium-Mn grade; cyan clouds),
and > 12 wt% (high-Mn grade; silver clouds) are included. Our LT P steel, which can be realized with existing industrial production lines, exhibits a superior strength–ductility
balance (see gray dotted lines) to that of bcc-based low-Mn lightweight alloys, even those of deformed and partitioned medium-Mn (10 wt%) steel [25] , low-C medium-Mn
TRIP
steel [24] , typical quenching and partitioning steel [51] , ultrafine-grained steel [48] , Ni(Fe,Al)-maraging steel [3] , nanoprecipitate-hardened medium-Mn (12 wt%) steel
[2] , while a comparable strength–ductility balance to that of high-Mn base nanotwinned steel [54] , high-Mn TWIP steels [53] , and high-Co bearing high-entropy alloys
(HEAs) [55] .
strain using the below Eq. (1) [64] .
V
γ=
I
220
γ
1 . 42I
200
α+I
220
γ
+
I
220
γ
0 . 71I
211
α+I
220
γ
+
I
311
γ
1 . 62I
200
α+I
311
γ
+
I
311
γ
0 . 81I
211
α+I
311
γ
4
(1)
Where I
αand I
γare the integrated intensity of ferrite and
austenite, respectively. The calculated values were 28.6%, 22.8%,
9.3%, and 4.0% for the ε= 0%, 13.5%, 25.2%, and 47.1% LTP sam-
ples as shown in Fig. 5 . These results are comparable with the
austenite volume fraction in the annealed specimen revealed by
conventional XRD ( V
γ= 23.1%, Fig. S3) [28] . Even after sample
fracture ( ε= 47.1%), 4% of heterogeneous austenite grains did not
undergo martensitic transformation. Therefore, in order to confirm
transformation-resistance depending on the grain size, the volume
fractions of small and large austenite grain during deformation un-
til ε= 13.5% were examined using the EBSD phase map (Sup-
plementary Information Fig. S6(a)) [28] . The smaller metastable
austenite grains exhibited greater resistance to martensitic trans-
formation upon deformation at the same plastic stress/strain level
applied (Supplementary Information Fig. S6(b)).
3.5. Dislocation features inside deformed austenite grains
Next, we examined the dislocation movements in the deformed
0.3C-850-LTP sample. The transformation-resistant austenite grains
with sizes of ∼0.5 μm (small grains) and ∼3.0 μm (large grains)
in the ε= 13.5% (deformed) and ε= 47.1% (fractured) LTP steel
samples were considered. For ε= 13.5% ( Fig. 7 (a)), the planar
configuration of dislocations is distinctly decorated in the small
transformation-resistant austenite grain. Further, some planar dis-
locations penetrate and glide across the phase boundaries in the
small-grained austenite grain, which implies dislocation slip trans-
mission across the boundaries. These planar dislocations can be
regarded as edge-components, as revealed by synchrotron XRD
profiles ( Fig. 4 (d)). Hence, we attribute the delay in deformation-
induced martensitic transformation (or TRIP effect) in the strained
small austenite with ε= 13. 5% to the lack of planar edge disloca-
tion impingement at the interfaces.
The bright-field TEM image from the small transformation-
resistant austenite grain at ε= 47.1% shows the stress-assisted
cross-slip of dislocations in the initial state amongst succes-
sive cross-slipping steps ( Fig. 7 (b)). In the TEM image, one
constriction—the joining of the Shockley partials—was observed;
this confirms the cross-slipping in the small austenite grains at
a high strain during the tensile test. The cross-slipping of mobile
screw dislocations has been typically observed in the FCC lattice
with a high stacking fault energy [65] . Constrictions only arise if
glissile screw dislocations move from one slip plane to another at
undissociated jogs or at nodes of attractive junctions in the pri-
mary plane [66] . Additionally, we found that some dislocations
were free to move, probably in other {111 } slip planes of the small
transformation-resistant grains. The corresponding selected area
electron diffraction (SAED) pattern showed strong streaks around
typical disordered FCC reflections, which result from the presence
of faulted ribbons. Thus, mobile dislocations readily glide in the
small transformation-resistant austenite grains owing to the ex-
tended interval of the { 111 }
FCC
slip planes by LTP, accordingly im-
proving ductility.
In contrast to the small transformation-resistant austenite grain,
the large transformation-resistant austenite in the LTP steel sam-
ple exhibits three distinct deformation structures under ε= 13. 5
% ( Fig. 7 (c)): (i) a high number of strongly curved edge dislo-
cations are decorated at the low strain regime; (ii) deformation-
induced martensite or premature TRIP effect operates on the
early deformation; (iii) a high density of edge dislocations is im-
pinged and accumulated at the phase boundary between the early
transformed martensite and the austenite grains. At a high-strain
regime ( ε= 47.1%), STEM-ABF and -ADF images ( Fig. 7 (d)) show
that the large transformation-resistant austenite mainly possesses
wavy configurated dislocations. We note that the dislocations ob-
served from the STEM analysis of the 47.1%-deformed austenite,
comprise both edge and screw components ( Fig. 4 (d)).
230
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 4. (a) Synchrotron XRD profiles of the LTP steel at different plastic strain levels, i.e., ε= 0 (undeformed), ∼13.5 %, 25.2%, and 47.1% (fractured). Distortion Fourier
coefficients of (b) FCC-austenite and (c) BCC-based structures, including ferrite matrix together with mechanically transformed martensite, as a function of Fourier length
under applied strains, determined from synchrotron XRD of tensile tested LTP steel. Each point was obtained by the
MWA method from synchrotron XRD profiles, and the
corresponding curves were fitted by the Wilkens model-based Berkum function. (d) Increased density of dislocations at different strains in BCC and FCC-structures. Details
of dislocation density determination are given in the Supplementary materials.
Fig. 5. The volume fraction of austenite at different plastic strain levels measured by X-ray diffraction (4 peaks method)
231
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 6. Bright-field TEM image of the dislocations in the deformed ferrite matrix at ε= 47 %. The yellow arrows indicate the Burgers vectors of the observed grain.
Fig. 7. Conventional TEM bright-field images of the deformed, untransformed small austenite ( γ) grains in an LTP steel at different strains: (a) ε= ∼13. 5%, which shows
both the easy passage of planar dislocations across boundaries and the planar configuration of dislocations (the enlarged view; inset); (b)
ε= ∼47% (sample fractured
state), which reveals the constriction in the initial cross-slipping state. The corresponding SAED pattern (right), captured from the constriction region, includes the streaks
of diffraction scattering due to the faulted ribbons. (c) Conventional TEM image and corresponding SAED pattern from the deformed, untransformed large austenite grains
in an LTP steel at ε= ∼13. 5%, showing presence of a common K–S relationship between the mechanically transformed α’-martensite and its parent austenite. (d) Scanning
(S)TEM annular bright field (ABF; left) and annual dark field (ADF; right) images of the large austenite grains in a fractured LTP steel upon
ε= ∼47%, which displays the
wavy configuration of dislocations.
232
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 8. (a) Typical SAED pattern along a low-index [ 1
¯
1 1 ]
BCC zone, obtained from the deformation-induced
α’-martensite from small austenite grains in the tensile-tested
LTP steel (
ε= ∼47%) to reach three series of two-beam conditions, i.e., three diffraction vectors, g
1 = (101); red arrow, g
2 = (110); green arrow, and g
3 =
( 01
¯
1 ) ; blue
arrow, respectively. (b) Corresponding TEM dark-field images, where most dislocations in the transformed
α’-martensite are regarded as screw-components. Three Burgers
vectors are indicated by the yellow arrows. (c) Typical SAED pattern along a low-index [ 1
¯
1 1 ]
BCC
zone (left), obtained from deformation-induced
α’-martensite from large
austenite grains in the tension-fractured LTP steel (
ε= ∼47%). (d) Corresponding TEM bright-field image and additional images (bottom panel), revealing the formation of
fine dislocation cells (red arrows) and dense dislocation wall (green arrows)
3.6. Dislocations inside deformation-induced α
-martensite
Fig. 8 (a) is the representative SAED pattern along the [ 1
¯
1 1 ]
BCC
zone axis, taken from deformation-induced α
-martensite in the
47.1%-strained LTP sample. Bragg conditions were realized by tilt-
ing the low-index zone [ 1
¯
1 1 ]
BCC to reach the three series of two-
beam conditions, i.e., three reflectors having diffraction vectors,
g
1
= (101), g
2
= (110), and g
3
= ( 01
¯
1 ) , respectively. The corre-
sponding diffraction patterns and resultant dark-field images are
provided in the insets of Fig. 8 (a) and (b), and the bottom panel,
respectively. Under each two-beam condition, some dislocations
were distinguished according to the visible criterion ( g
•b = 0).
Following the well-known TEM approach [ 34 , 35 , 37 , 67 ], we de-
termined Burger vectors in the α
-martensite. The TEM analysis
indicated that the majority of dislocations inside the martensite
from the small austenite were screw-components, which is in good
agreement with the synchrotron XRD results ( Fig. 4 ). Further, the
synchrotron XRD and TEM results indicate that the dislocation
densities in α
-martensite are considerably lower than that inside
the untransformed small austenite grains at a given strain level
( ε= 47.1%).
The transformed α
-martensite from the large austenite in the
47.1%-strained 0.3C-850-LTP sample displayed the substantial den-
sity of mobile and immobile screw dislocations than those in
α
-martensite from the small austenite. This is evident from the
bright-field TEM images along the low-indexed [001]
BCC
zone axis
( Fig. 8 (c) and (d)). The conventional TEM images revealed the for-
mation of fine dislocation cells and dense dislocation wall in the
martensite from the large austenite. The size-dependent difference
in the density of screw dislocations can be attributed to the com-
bination of plastic deformation and displacive shear transformation
[25] . The early transformed α
-martensite from the large austen-
ite would suffer from successive external plastic strains than the
later transformed α
-martensite from the small austenite, based
on a general relationship between dislocation density and ap-
plied plastic strain [47] . Hence, the dislocation forest strengthen-
ing that originates as a result of the cell-forming dislocations-
induced Taylor hardening law could be higher in the martensite
derived from the large austenite than that derived from the small
one.
3.7. Interstitial carbon distributions in the tensile-tested LTP steel
For the transformation-resistant small austenite grains in the
47.1%-deformed 0.3C-850-LTP sample, an irregular distribution of
interstitial C atoms (red spheres) is seen in an APT-reconstructed
map ( Fig. 9 (a)). The iso-concentration envelopes of 4.5 at.% Mn
are chosen to outline the austenite/ferrite interface, while the C-
dislocation interaction is selected by the iso-concentration en-
velopes of 8.0 at.% C. A different view and corresponding iso-
concentration wireframe revealed the presence of wolf-shaped
interaction between the dislocations and the C atoms in the
transformation-resistant small austenite grains. This resembles the
constriction that appears in the initial cross-slipping state, as
shown in the TEM image ( Fig. 7 (b)). The planar-featured three-
dimensional (3D) decoration of C atoms at dislocations was ob-
served near the phase boundary. The concentration profile across
the phase boundary ( Fig. 9 (b)) suggests that the C concentra-
tion peaks inside the small austenite grains and at the inter-
face zone were at ∼8.5 at.% and ∼10 at%, respectively. This re-
sult supports the C-compositional spike near the interface for the
transformation-resistant small austenite.
The APT results of the transformation-resistant large austenite,
as indicated by the iso-concentration envelope in Fig. 10 (a), display
the interplay between C atoms and dislocations. An enlarged differ-
ent view and the corresponding iso-concentration wireframe sug-
gest that the 3D decoration of C atoms in the APT map resembles
the dislocation cell formation, as revealed by TEM in Fig. 8 (b). As
indicated in Fig. 10 (b), the C concentration peaks inside the large
austenite grains and at the interface zone were ∼7.1 at.% and ∼7.5
at.%, respectively. This reveals the absence of C-compositional spike
at the interface for the transformation-resistant large austenite.
233
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 9. APT results of small austenite ( γ) grain zone in the fractured LTP steel under ε= ∼47 %: (a) APT-reconstructed C (red sphere) map with iso-concentration envelopes
of 4.5 at% Mn (cyan) and 8.0 at% C (gold). The iso-concentration wire-frame from the outlined red box shows the wolf-shaped interaction between dislocations and C atoms
(right inset); (b) 1D concentration profiles of Mn and C across the
BCC/FCC interface in (a), revealing the strong concentration spike of C at the interface region. By the same
iso-concentration analysis, the distinct C-dislocation interaction did not visible in the APT results of the LTP-processed, undeformed austenite grains.
Furthermore, it should be noted that using iso-concentration en-
velopes of 8.0 at. % C, the distinct C-dislocation interaction did not
visible in the APT results from the undeformed LTP steel ( Fig. 1 (b)).
Hence, the APT result that shows the profound C decoration at dis-
locations in the transformation-resistant small and large austenite
grains suggested that the high density of C-decorated dislocations
predominantly originated from the applied plastic strains. How-
ever, the C segregation at the austenite/ferrite interface in the un-
deformed samples will be analyzed in the next study.
A comparison of Figs. 9 and 10 clarifies that there are significant
differences in the transformation-resistant small and large austen-
ite grains in the 47.1%-deformed state:
234
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
Fig. 10. (a) APT map of C atoms, taken from the large austenite zone, showing a high density of C decoration at dislocations inside the large austenite grain zone for the
fractured LTP steel under
ε= ∼47%. The (b) 1D concentration profiles of C and Mn across the austenite/BCC interface in (a). The iso-concentration wire-frame from the
outlined red box shows the cell formation by the C-decorated dislocations.
(i) Size-dependent dislocation pinning, i.e., the C distribution fre-
quency at the dislocations is more distinct in the large austenite
grains than in the small ones.
(ii) The peak concentration of C inside the large grains is higher
than that in the small ones.
(iii) A clear concentration spike of C at the interface is visible for
the large grains, and this was mainly driven by dislocation im-
pingement, leading to interface shielding effect.
These underlying differences stem from the grain-size depen-
dent dislocation activities, i.e., high dislocation density-enabled C
235
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
distribution and plastic strain accommodation at the interface in-
duced by the stress concentration.
4. Conclusions
The LTP-based processing route overcomes the strength–
ductility trade-off in typical ferritic-based low-Mn lightweight
steels with a heterogeneous metastable microstructure. The
LTP treatment facilitates higher C-partitioning into the small
metastable austenite (with a mean diameter of ∼0.5 μm) grains
than into the large grains (with mean diameter of ∼3.1 μm). This
size-dependent C partitioning allows the small austenite grains
to increase the slip plane spacing and enhance the severe lat-
tice strain effect compared to the large grains, leading to the
size-dependent dislocation movement in the austenite grains upon
tensile deformation. We attribute the simultaneous increase in
strength and total elongation of the typical low-Mn lightweight
steel upon LTP to both the size-dependent dislocation movement
in the austenite grains and the controlled deformation-induced
martensitic transformation. The increased strength and ductility,
arising from the large metastable austenite grains, are achieved
through the TRIP effect on the early deformation, wavy dislocation
configuration in the deformed austenite phase in the high-strain
regime, and high density of C-decorated dislocations in the α’-
martensite at a high strain (namely, dislocation forest strengthen-
ing originating as a result of the cell-forming dislocation-induced
Tayl or hardening law).
The additional ductility and strength, which result from the
small austenite grains, are attributed to the gliding of dislocations
along slip planes (upon the early deformation), interface shielding
effect of C-dislocations, delayed TRIP effect, and the cross-slip mo-
tions of mobile dislocations (at the high strain level). Particularly,
the size-dependent dislocation movement postpones the TRIP op-
eration in the small austenite grains at the early stages of defor-
mation. This size-dependent dislocation activity in austenite grains
stemming from LTP ensures a low raw-materials cost compared to
those of the high-Ni maraging steels. It also ensures a low thermo-
mechanical processing cost compared to that of the high-Mn TWIP
steels, while maintaining an analogous strength–ductility balance
for lightweighting technologies.
Supplementary materials
Supplementary material associated with this article can be
found, in the online version.
Declaration of competing interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared to
influence the work reported in this paper. Patent application (Ko-
rean Patent application number 10-2020-0172118) has been filed
based on the results of this study.
Data availability statement
All data included in this study are available upon request con-
tact with the corresponding authors.
Acknowledgements
. This work was financially supported by the National Research
Foundation of Korea (NRF) grant funded by the Korea government
(MSIT) ( No. 2021R1A2C4002622 ). N.G. Park acknowledges finan-
cial supports from the Industrial Strategic Technology Development
Program (20 0 09993), funded by the Ministry of Trade, Industry &
Engergy (MOTIE, Korea).
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi: 10.1016/j.jmst.2021.11.052 .
References
[1] D. Raabe , C.C. Tasan , E.A. Olivetti , Nature 575 (2019) 64–74 .
[2] D. Raabe , D. Ponge , O. Dmitrieva , B. Sander , Scr. Mater. 60 (2009) 1141–1144 .
[3] S. Jiang , H. Wang , Y. Wu , X. Liu , H. Chen , M. Yao , B. Gault , D. Ponge , D. Raabe ,
A. Hirata , M. Chen , Y. Wang , Z. Lu , Nature 544 (2017) 460–464 .
[4] J.G. Canadell , C.L. Quéré, M.R. Raupach , C.B. Field , E.T. Buitenhuis , P. Ciais ,
T.J. Conway , N.P.
Gillett , R.A. Houghton , G. Marland , Proc. Natl. Acad. Sci. U.
S. A. 104 (2007) 18866–18870 .
[5] A. Devaraj , V.V. Joshi , A. Srivastava , S. Manandhar , V. Moxson , V.A. Duz ,
C. Lavender , Nat. Commun. 7 (2016) 1117 6 .
[6] G. Frommeyer , E.J. Drewes , B. Engl , Re v. Met. Paris 97 (20 0 0) 1245–1253 .
[7] G. Frommeyer , U. Brüx , Steel Res. Int. 77 (2006) 627–633 .
[8] P. Wen , B. Hu , J. Han , H. Luo , J. Mater. Sci. Technol. 97 (2022) 54–68 .
[9] B. Hu , H. Luo , F. Yang , H. Dong , J. Mater. Sci. Technol. 33 (2017) 1457–1464 .
[10] D. Raabe , H. Springer , I. Gutierrez-Urrutia , F. Roters , M. Bausch , J.-B. Seol ,
M. Koyama , P.- P. Choi , K. Tsuzaki , JOM 66 (2014) 1845–1856 .
[11] S. Chen , R. Rana , A. Haldar , R.K. Ray , Prog. Mater. Sci. 89 (2017) 345–391 .
[12] B.C. De Cooman , Curr. Opin. Solid State Mater. Sci. 8 (2004) 285–303 .
[13] F. Lu , P. Yang , L. Meng , F. Cui
, H. Ding , J. Mater. Sci. Technol. 27 (2011) 257–265 .
[14] O. Bouaziz , S. Allain , C.P. Scott , P. Cugy , D. Barbier , Curr. Opin. Solid State Mater.
Sci. 15 (2011) 141–168 .
[15] L. Mujica , S. Weber , H. Pinto , C. Thomy , F. Vollertsen , Mater. Sci. Eng. A 527
(2010) 2071–2078 .
[16] L. Mujica , S. Weber , C. Thomy , F. Vollertsen , Sci. Technol. Weld. Joining 14
(2009) 517–522 .
[17] M. Dahmen , S. Lindner , D. Monfort , D. Petring , Phys. Proc. 83 (2016) 344–351
.
[18] S.S. Sohn , B.-J. Lee , S. Lee , N.J. Kim , J.-H. Kwak , Acta Mater 61 (2013)
5050–5066 .
[19] I. Gutierrez-Urrutia , D. Raabe , Scr. Mater. 68 (2013) 343–347 .
[20] M.J. Yao , E. We lsc h , D. Ponge , S.M.H. Haghighat , S. Sandlöbes , P. Choi , M. Her-
big , I. Bleskov , T. Hickel , M. Lipinska-Chwalek , P. Shanthraj , C. Scheu , S. Zaef-
ferer , B. Gault , D. Raabe , Acta Mater 14 0 (2017) 258–273 .
[21] T. Furuhara , Y. Zhang , G. Miyamoto , IOP
Conf. Ser.: Mater. Sci. Eng. 580 (2019)
012005 .
[22] M. Liu , C. Song , Z. Cui , J. Mater. Sci. Technol. 78 (2021) 247–259 .
[23] M.X. Yang , F.P. Yuan , Q.G. Xie , Y.D. Wang , E. Ma , X.L. Wu , Acta Mater 109 (2016)
213–222 .
[24] S.S. Sohn , K. Choi , J.-H. Kwak , N.J. Kim , S. Lee , Acta Mater 78 (2014) 181–189 .
[25] B.B. He , B. Hu , H.W. Ye n , G.J. Cheng , Z.K. Wang , H.W. Luo , M.X. Huang , Science
357 (2017) 1029–1032 .
[26] L. Liu , B. He , M. Huang , Adv. Eng. Mater. 20 (2018) 170 108 3 .
[27] Y. Wang , Z. Guo , N. Chen , Y. Rong , J. Mater. Sci. Technol. 29 (2013) 451–457 .
[28] H.J. Bae , K.K. Ko , H.S. Park , J.S. Jeong , J.-K. Kim , H.-K. Sung , J.B. Seol, Korean J.
Met. Mater. 59 (2021) 683–691 .
[29] K.-G. Chin , H.-J. Lee , J.-H. Kwak , J.-Y. Kang , B.-J. Lee , J. Alloy. Compd. 505 (2010)
217–223 .
[30] J.B. Seol , J.G. Kim , S.H. Na , C.G.
Park , H.S. Kim , Acta Mater 131 (2017) 187–196 .
[31] O. Matsumura , Y. Sakuma , H. Take chi , Scr. Metall. Mater. 21 (1987) 1301–1306 .
[32] Y. Zhu , C. Ophus , M.B. To loczk o , D.J. Edwards , Ultramicroscopy 193 (2018)
12–23 .
[33] P.J . Phillips , M.C. Brandes , M.J. Mills , M. De Graef , Ultramicroscopy 111 (2011)
1483–1487 .
[34] M. Tana ka , K. Higashida , K. Kaneko , S. Hata , M. Mitsuhara , Scr. Mater. 59
(2008) 901–904 .
[35] Y. Miyajima , M. Mitsuhara , S. Hata , H. Nakashima , N. Tsuji
, Mater. Sci. Eng. A
528 (2010) 776–779 .
[36] P.J . Phillips , M.J. Mills , M.D. Graef , Philos. Mag. 91 (2011) 2081–2101 .
[37] Y. Meng , X. Ju , X. Yang , Mater. Charact. 17 5 (2021) 1110 65 .
[38] J.E. Bailey , P.B . Hirsch , Philos. Mag. 5 (1960) 4 85–4 97 .
[39] P.B . Hirsch , A. Howie , R.B. Nicholson , D.W. Pashley , M.J. Whelan , Electron Mi-
croscopy Thin Cryst. (1965) .
[40] D.J.H. Cockayne , I.L.F. Ray , M.J. Whelan , Philos. Mag. 20 (1969) 1265–1270 .
[41] R. Zhang , J.Ding S.Zhao
, Y. Chong , T. Jia , C. Ophus , M. Asta , R.O. Ritchie ,
A.M. Minor , Nature 581 (2020) 283–287 .
[42] B. Gault , D. Haley , F. de Geuser , M.P. Moody , E.A. Marquis , D.J. Larson ,
B.P. Geiser , Ultramicroscopy 111 (2011) 448–457 .
[43] J. Takahashi , K. Kawakami , Y. Kobayashi , Ultramicroscopy 111 (2011)
1233–1238 .
[44] M.J. Yao , P. Dey , J.-B. Seol , P. Choi , M. Herbig , R.K.W. Marceau , T. Hickel ,
J. Neugebauer , D. Raabe , Acta Mater 106 (2016) 229–238 .
[45] G. Miyamoto , A. Goto , N. Takaya ma , T. Furuhara , Scr. Mater. 154 (2018)
168–171 .
[46] J. Yoo , W.-M . Choi , B.-J. Lee , G.-Y. Kim , H. Kim , W.- D. Choi , Y.J. Oh , S. Lee , Met.
Mater. Int. 26 (2019) 1506–1514 .
[47] J.P. Hirth , J. Lothe , Theory of Dislocations, 2nd ed, Wiley, New York , 1982 .
[48] J.R. Greer , W.C. Oliver , W. D. Nix , Acta Mater 53 (2005) 1821–1830 .
[49] H. Kim , D.-W. Suh , N.J. Kim , Sci. Technol. Adv. Mater. 14 (2013) 014205 .
[50] S.Y. Han , S.Y. Shin , S. Lee , N.J. Kim , J.H. Kwak , K.G. Chin , Metall. Mater. Trans. A
42 (2011) 138–146 .
236
K.K. Ko, H.J. Bae, E.H. Park et al. Journal of Materials Science & Technology 117 (2022) 225–237
[51] C. García-Mateo , F.G. Caballero , Mater. Trans. 46 (2005) 1839–1846 .
[52] R. Rana , C. Liu , R.K. Ray , Acta Mater 75 (2014) 227–245 .
[53] J. Moon , H.-Y. Ha , K.-W. Kim , S.-J. Park , T.- H. Lee , S.-D. Kim , J.H. Jang , H.-H. Jo ,
H.-U. Hong , B.H. Lee , Y.-J . Lee , C. Lee , D.-W. Suh , H.N. Han , D. Raabe , C.-H. Lee ,
Sci. Rep. 10 (2020) 12140 .
[54] Y. Toji , G. Miyamoto , D. Raabe , Acta Mater 86 (2015) 137–147 .
[55] E.D. Moor , J.G. Speer , D.K. Matlock , J.-H. Kwak , S.-B. Lee , ISIJ Intl 51 (2011)
137–144 .
[56] J.H. Hwang , T.T.T. Trang , O. Lee , G. Park , A. Zargaran , N.J. Kim , Acta Mater 191
(2020) 1–12 .
[57] J. Moon , S.-J. Park , J.H. Jang , T.-H. Lee , C.-H. Lee , H.-U. Hong , H.N. Han , J. Lee ,
B.H. Lee , C. Lee , Acta Mater 147 (2018) 226–235 .
[58] H.T. Wa ng , N.R. Tao , K. Lu , Acta Mater 60 (2012) 4027–4040 .
[59] Z. Li , K.G. Pradeep , Y. Deng , D. Raabe , C.C. Tasan , Nature 534 (2016) 227–230 .
[60] Z. Li , C.C. Tasan , H. Springer , B. Gault , D. Raabe , Sci. Rep. 7 (2017) 40704 .
[61] J.B. Seol , J.W. Bae , Z. Li , J.Chan Han , J.G. Kim , D. Raabe , H.S. Kim , Acta Mater
151 (2018) 366–376 .
[62] T. Shintani , Y. Murata , Acta Mater 59 (2011) 4314–4322 .
[63] S.K. Oh , M.E. Kilic , J.B. Seol , J.S. Hong , A. Soon , Y.K . Lee , Acta Mater 188 (2020)
366–375 .
[64] R.L. Miller , Trans. ASM 57 (1964) 893–899 .
[65] J. Bonneville , B. Escaig , Acta Metall. Mater. 27 (1979) 1477–1486 .
[66] B. Escaig , J. Phys. (Paris) 29 (1968) 225–239 .
[67] J.K. Kim , M.-H. Kwon , B.C. De Cooman , Acta Mater 141 (2017) 4 4 4–455 .
237