International Journal of Fatigue 165 (2022) 107186

Contents lists available at ScienceDirect

International Journal of Fatigue

journal homepage: www.elsevier.com/locate/ijfatigue

Influence of pre-straining on the low-cycle fatigue performance of

Fe–0.1C–5Mn medium manganese steel
Mei Zhang a, b, *, Weijun Wang a, Bendao Zhang a, Qiongying Cen a, Jin Liu a
a
School of Materials Science and Engineering, Shanghai University, Shanghai 200444, China
b
State Key Laboratory of Advanced Special Steel, Shanghai University, Shanghai 200444, China

A R T I C L E I N F O A B S T R A C T

Keywords: The low-cycle fatigue (LCF) behavior of Fe–0.1C–5Mn medium manganese transformation-induced plasticity
Low-cycle fatigue steel in its as-received (AR: 0% pre-strained) and 10%, 15%, and 20% pre-strained conditions was analyzed.
Medium Mn steel Microstructural evolution during pre-straining and cyclic loading was characterized by X-ray diffraction, electron
Pre-straining
backscatter diffraction, and transmission electron microscopy. The LCF properties of the steel were observed to
Microstructure evolution
gradually improve as pre-straining increased. Moreover, during pre-straining, the blocky austenite grains were
successively split into thin film-like austenite by fresh lath martensite. It refined the microstructure, delayed
fatigue crack initiation, prevented crack growth, and resulted in notable improvements in the LCF performance.

1. Introduction forming processes before they are put into operation, and pre-straining
is usually introduced in the forming process [7]. The internal disloca­
Medium Mn (5–12% Mn) steels have attracted increasing attention tions sprout and proliferate during the pre-straining stage, thus hin­
from both steel manufacturing and automotive industries owing to their dering the slip motion and increasing the strength of steel [8]. Several
excellent comprehensive mechanical properties and relatively low studies have shown that the significant increase in tensile strength,
alloying costs. They are potential candidates for lightweight vehicles, which is related to pre-straining, does not translate into an increase in
energy conservation, and more reliable and economical industrial fatigue strength owing to the large loss in elongation; for example, high-
products [1–5]. strength low alloy (HSLA) steel (the total elongation (TE) of 0–5% pre-
Fatigue damage, due to the accumulation of cyclic plastic deforma­ strained HSLA steel decreases from 26% to 16%) [9,10], dual phase (DP)
tion, inevitably occurs in stress concentrations and weak locations of steel (the uniform elongation of 0–13% pre-strained DP steel decreases
structures during the normal use of vehicles [2]. The results of previous from 27% to 0.7%) [11], and complex phase (CP) steel (the TE of 0–8%
fatigue studies on medium manganese steel showed that the volume pre-strained CP steel decreases from 17.4 ± 0.6% to 9.1 ± 1.0%) [12]. It
fraction and stability of metastable austenite have various effects on the should be noted that the work-hardening phenomenon in medium
fatigue life under cyclic loading [3–6]. The absorbed energy [3,4] and manganese steel functions as mechanisms for dislocation strengthening
crack closure effect [5] in the martensitic transformation process have a and transformation-induced plasticity (TRIP) strengthening [13]. Sugi­
favorable impact on the fatigue life; however, brittle martensite, the moto [14] performed strain-controlled low-cycle fatigue (LCF) tests
phase transformation product of metastable austenite, can easily under 0–10% pre-strained conditions and reported that the pre-strain
become the source of crack initiation [6]. Moreover, multiple superpo­ was beneficial to the LCF life in TRIP steel. Well-maintained ductility
sitions complicate the effect of metastable austenite on the fatigue life and higher elastic-plastic strain ratio are observed when the pre-strain
[3–6]. increases, thus improving the fatigue life caused by phase trans­
Steels must undergo cold pressing, drawing, and other cold working formation [15]. For medium manganese steels containing metastable

Abbreviations: AR, As-received; CP, Complex phase; DP, Dual phase; EBSD, Electron backscatter diffraction; GND, Geometrically necessary dislocation; HSLA,
High-strength low alloy; KAM, Kernel average misorientation; LCF, Low-cycle fatigue; RA, Retained austenite; SAED, Selected-area-electron-diffraction; SEM,
Scanning electron microscopy; TE, Total elongation; TEM, Transmission electron microscopy; TRIP, Transformation-induced plasticity; XRD, X-ray diffraction; YS,
Yield strength.
* Corresponding author at: School of Materials Science and Engineering, Shanghai University, Shanghai 200444, China.
E-mail address: [email protected] (M. Zhang).

https://doi.org/10.1016/j.ijfatigue.2022.107186
Received 13 December 2021; Received in revised form 30 July 2022; Accepted 1 August 2022
Available online 4 August 2022
0142-1123/© 2022 Elsevier Ltd. All rights reserved.
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

austenite, the effects of pre-straining on the microstructural evolution The microstructural evolution was characterized by XRD, SEM, EBSD
and fatigue behavior have not been clarified thus far. and TEM. The volume fraction of the retained austenite (RA) was
This study aims to introduce pre-strain to medium manganese steel examined using an X-ray diffractometer with a CuKα radiation source
during actual production, anticipating the occurrence of austenite to operated at 40 kV and 250 mA in the 2θ range of 40–100◦ , at a rate of 2◦ /
martensite phase transformation. In addition, this study focuses on the min. The volume fraction of RA was calculated based on the following
complex variations in the dislocation morphology during the subsequent equation: [16].
fatigue process, which is significant and non-negligible. The micro­ ( )
VA = 1/G Iα /Iγ (1)
structural evolution of Fe–0.1C–5Mn TRIP steel during pre-straining and
subsequent fatigue processes was investigated in detail using trans­
where VA denotes the volume fraction of RA; G is a parameter related to
mission electron microscopy (TEM), nanoindentation, X-ray diffraction
the crystal planes of austenite and ferrite; Iγ is the integrated intensity of
(XRD), and electron backscatter diffraction (EBSD), and the mechanism
(2 0 0)γ, (2 2 0)γ and (3 1 1)γ peaks; and Iα is the integrated intensity of
of its influence on the behavior of LCF was explored.
(2 0 0)α and (2 1 1)α peaks.
The paper is organized as follows. The experimental procedures are
The dislocation density measured by XRD can be calculated using Eq.
described in Section 2, and the experimental results about tensile
(2) [17]:
properties and low cycle fatigue properties are described in Section 3.
√̅̅̅
Section 4 involves the discussion with the support of TEM, scanning 2 3ε
electron microscopy (SEM), EBSD, XRD and nano-indentation charac­ ρx = (2)
Db
terization, including the influences of pre-straining on microstructural
evolution, cyclic strain on microstructural evolution, and pre-straining where D denotes the crystallite size, ε is the root-mean-square strain, and
on crack initiation and propagation rate, relationship between the b is the magnitude of the Burgers vector of dislocations.
microstructural evolution, gradually transformation of austenite-to- EBSD was performed using a field-emission SEM (Apollo 300)
martensite, as well as the mechanism of simultaneous improvement of equipped with the HKL Channel 5 software. The operating voltage and
strength and fatigue properties, and finally to end with conclusions in step size were 20 kV and 0.06 μm, respectively. The specimens for EBSD
Section 5. observations were prepared by mechanical grinding and ion polishing.
The method proposed by Kubin and Mortensen [18] was used to
2. Experiments calculate the geometrically necessary dislocation (GND) densities from
the kernel average misorientation (KAM) value using the following
In this study, the tensile mechanical properties and LCF behavior of a equation [19]:
Fe–0.1C–5Mn (in weight percent) medium manganese TRIP steel sheet αKAM
with a thickness of 1.25 mm were studied. To increase the stability of ρGND = (3)
b•R
austenite and control the cost of alloy, 5 wt% Mn was selected. The cold-
rolled steel sheet was slightly heated to a temperature above Ac3 and where α is a constant that depends on the nature of dislocation edge or
then water-quenched to room temperature to predominately produce a screw; the value of α was proposed to be 3 [20]. The KAM values were
martensitic structure. Subsequently, the steels were subjected to inter­ given by the EBSD software; moreover, R denotes the kernel size
critical annealing at 625 ◦ C for 12 h and then air-cooled to room tem­ considering the neighbor order considered in KAM mapping.
perature to obtain a dual-phase microstructure of ferrite and austenite. TEM observations were recorded using a JEM-2010 TEM under an
The test plan included pre-straining and subsequent fatigue testing. acceleration voltage of 200 kV with thin foils of diameter 3 mm, which
To investigate the influence of pre-straining on the static uniaxial tensile were ground to a thickness of 50 μm and electropolished with a solution
and fatigue properties, the specimens were pre-strained to different of 5 vol% perchloric acid in alcohol at − 30 ◦ C using a twin-jet machine.
engineering strains of 10%, 15% and 20% using uniaxial tension and The nanohardness of each phase was measured using a KLA Nano
then unloaded. Indenter system. A maximum load of 1.5 mN with a loading rate of 0.1
The section size of the pre-strained specimens was measured; then, a mN/s was applied to the specimens for 10 s. An array of indents in the 5
secondary static tensile test was conducted until fracture to explore the × 5 matrix was defined with an indent gap of 5 μm.
effect of pre-straining on the static mechanical properties according to
ASTM E8/E8M-16a. The partial specimens were subjected to a Vickers 3. Results
micro-hardness experiment using a load of 0.2 kgf (HV0.2).
The strain-controlled fatigue test pieces were cut from large pre- 3.1. Tensile properties
strained test specimens; the corresponding specimen geometry was
illustrated in Fig. 1. The fatigue test was performed using a MTS Land­ The engineering stress-strain curves of as-received (AR) and 10%,
mark fatigue-testing machine. The axial constant amplitude tension- 15%, and 20% pre-strained (PS10, PS15, PS20) Fe–0.1C–5Mn steels are
compression test with strain ratio R (εmin/εmax) = − 1 was adopted for illustrated in Fig. 2 and the corresponding static mechanical property
the strain-controlled fatigue test, according to ASTM E606/E606M- data are listed in Table 1. There is an associated increase in the yield
2012, in the strain amplitude range of 0.2–0.8% with a step size of strength (YS) and tensile strength (UTS) with an increase in pre-
0.1%. The gauge length of the extensometer and loading frequency were straining. The increments in YS and UTS were found to range between
10 mm and 1 Hz, respectively; moreover, the loading waveform was 186–313 and 36–78 MPa, respectively. The increase in YS was higher
triangular. than that in UTS, which increased the yield tensile strength ratio and

Fig. 1. Geometry schematic of uniaxial tensile (gray solid line) and fatigue test (dark solid line) specimens (dimensions in mm).

2
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 2. Engineering stress-strain curves of AR, 10%, 15% and 20% pre-strained Fig. 3. Strain amplitude versus number of reversals to failure curves of AR and
(PS10, PS15, PS20) Fe–0.1C–5Mn steels. pre-strained Fe–0.1C–5Mn steels.

0.8%. This is related to the evolution of the microstructure during pre-

Table 1
straining and fatigue [6], which is discussed in detail in Section 4.
Static uniaxial tensile properties of Fe–0.1C–5Mn steel under various conditions.
The stress amplitude response, contributed by dislocation strength­
Specimens YS UTS TE YS/ Micro-hardness ening, as well as monotonic loading induced austenite to martensite
(MPa) (MPa) (%) UTS (HV0.2)
phase transformation strengthening, increased with increasing prestrain
AR: As-received 470 730 37.3 0.65 213 ± 9.0 [15]. Although pre-straining facilitates cyclic softening of Fe–0.1C–5Mn
PS10: 10% pre- 656 766 30.9 0.86 249 ± 8.5
steel, the blocked and accumulated dislocations due to pre-straining
strained
PS15: 15% pre- 710 797 26.2 0.89 265 ± 10.2 may get annihilated during subsequent cyclic deformation, but not
strained completely. In addition, the martensitic transformation of RA due to pre-
PS20: 20% pre- 783 808 16.3 0.97 272 ± 8.1 straining is irreversible, and the austenite-to-martensite phase trans­
strained
formation strengthening and dislocation strengthening caused by pre-
straining cannot disappear completely [24].
reduced the secondary strain hardening ability of Fe–0.1C–5Mn steel. In
terms of ductility, the TE decreased continuously with an increase in pre- 4. Discussion
straining. Meanwhile, the steel maintained good plastic deformation
capacity. The uniform elongation in the PS20 specimen during 4.1. Influence of pre-straining on microstructure evolution
restraining exceeded 10%. The decrease in TE was between 6.4 and
16.3%. Fig. 5 shows the TEM images of the microstructures of ferrite as well
as small and large austenite islands under the AR, PS10 and PS20 con­
ditions, along with the selected-area electron diffraction patterns of the γ
3.2. LCF properties and α’ phases.
The microstructure of Fe–0.1C–5Mn steel is composed of ferrite and
3.2.1. LCF life characteristics different types of RA (fine film-like and blocky RA); moreover, blocky
The LCF life variation in Fe–0.1C–5Mn steel was investigated via austenite is classified into two-level sizes, namely, large (over 2 μm in
strain-controlled fatigue tests. The fitting calculation of the LCF curve, length) and small (approximately 1 μm in length). Previous study indi­
followed by combining the Basquin [21] equation with the Manson[22]– cated that the Mn partitioning between austenite and ferrite is evidently
Coffin [23] can be expressed as follows: different; austenite contains much more Mn than ferrite [25]. According
Δεt /2 = (σ ’f /E)(2Nf )b + ε’f (2Nf )c (4) to EDS results (Fig. 6), it can be observed that the Mn concentration in
various regions of austenite is relatively different; lower Mn content
where Δεt /2 is the total strain amplitude, σ ’f is the fatigue strength co­ regions in the RA grain are preferentially transformed to martensite after
efficient, b is the fatigue strength exponent, ε’f is the fatigue ductility pre-straining owing to lower stability; thus, the fresh martensite trans­
coefficient, c is the fatigue ductility exponent, E is the elastic modulus, formed from austenite appears as bundles parallel to each other (Fig. 5b,
and 2 Nf is the number of reversals to failure (low-cycle fatigue life). c, g, and h). As pre-straining increased, the fresh martensite cut the
Fig. 3 plots strain amplitude versus number of reversals to failure blocky austenite into many film-like austenite.
curves under different pre-strain levels. It is evident from Fig. 3 that the Accordingly, the difference in the stability of blocky austenite islands
fatigue life increases with pre-straining. of different sizes was revealed; the smaller the grain size, the higher the
stability of austenite. The transformation of small RA islands requires
3.2.2. Cyclic deformation response larger mechanical free energy according to the Hall-Patch formula; in
The cyclic deformation responses of Fe–0.1C–5Mn steel during the addition, higher Mn content in small RA can effectively improve the
strain-controlled fatigue process under different conditions at a strain stability, indicating that the stability of small RA is significantly higher
amplitude of 0.2%, 0.4%, and 0.8% are shown in Fig. 4. The AR speci­ than that of large RA [26]. As shown in Fig. 5b and c, small blocky RA
mens underwent cyclic hardening at a strain amplitude of 0.2%, 0.4% can also transform to α’-martensite, but the amount of α’-martensite is
and 0.8%. After pre-straining, specimens PS10 and PS15 exhibited cyclic limited and the size of α’-martensite is smaller than that of large RA
softening at a strain amplitude of 0.2% and 0.4%, respectively; however, owing to its higher stability, as shown in Fig. 5g and h. It should be noted
they exhibited cyclic hardening at a strain amplitude of 0.8%. Specimen that dislocation lines and blocks were observed in the ferrite and
PS20 exhibited cyclic softening at a strain amplitude of 0.2%, 0.4% and austenite grains under pre-strained conditions.

3
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 4. Comparison of cyclic stress amplitude response of Fe–0.1C–5Mn steel at a strain amplitude of (a) 0.2%, (b) 0.4%, and (c) 0.8%.

The deformation behaviors of ferrite and austenite in medium Mn AL, AS, and ferrite grains demonstrated evident pop-ins during initial
steels were expected to be different. The deformation behavior of indi­ loading, i.e., the P-h curves deviated from the Hertzian solution, which
vidual phases cannot be studied directly by macroscopic tensile tests, corresponds to the onset of plasticity starting from dislocation nucle­
but it can be investigated by nanoindentation tests [27]. ation. Because the load required to trigger dislocation plasticity is
The Hertzian elastic contact solution was applied to investigate the significantly lower than that required for martensitic transformation,
elastic deformation in austenite and ferrite grains, as follows [28]: the second pop-in in austenite grains indicated martensite formation
[25]. The load value of the second pop-in in AS grains was significantly
4 √̅̅̅̅̅̅̅̅̅
P = Er Ri h 3 (5) higher than that in AL grains, indicating that AS grains have higher
3
mechanical stability, which is in agreement with TEM results. A higher
where P denotes the applied load, h is the indentation depth, Ri is the shear stress indicates a higher resistance to plastic deformation. More­
indenter tip radius, and Er is the reduced modulus given by. over, load, where the P-h curves deviate from the Hertzian solution, can
represent the resistance to the initiation of plasticity during uniaxial
1 1 − ν2i 1 − ν2s tensile deformation [28]. Therefore, dislocation plasticity is initiated by
= + (6)
Er Ei Es ferrite grains, followed by AL and AS grains (Fig. 7c). Therefore, the
initiation of plasticity (i.e., yielding) during the uniaxial tensile defor­
where ν denotes Poisson’s ratio; moreover, subscripts i and s represent
mation of the tested steel under the AR condition can be mainly
the indenter and specimen, respectively. Young’s modulus and Poisson’s
attributed to dislocation plasticity from ferrite grains. Additionally, the
ratio were 1141 GPa, 0.07 for the indenter [29]; 220 GPa, 0.3 for the
pop-ins during nanoindentation and yield drops during tensile tests were
ferrite [30]; and 187 GPa, 0.3 for the austenite [30], respectively.
manifestations of strain softening, and the yield-point phenomenon was
The maximum shear stress underneath the indenter, τmax, can also be
closely related to the initiation of plasticity [31]. The source of the sharp
calculated via Hertzian analysis [28]:
yield point in the AR specimen (Fig. 2) was dislocation nucleation in
(
0.31 6PEr2
)1/3 ferrite, while the subsequent Lüders strain of approximately 1% was the
τmax = (7) combined result of ferrite dislocation plasticity and austenite-to-
π R2i
martensite phase transformation [27].
By detailed microstructural characterization of indent positions Because the AL grains transformed into martensite after pre-
(Fig. 7a), the load-displacement (P-h) curves of indentations on large straining, the deformation behavior for further pre-straining is mainly
blocky austenite (AL) grains, small blocky austenite (AS) grains, and discussed in terms of AS and ferrite grains. After pre-straining, the
ferrite grains under the AR condition can be determined (Fig. 7b). austenite and ferrite deviated from the Hertzian solution at the begin­
Compared to ferrite (Fig. 7b), austenite exhibited higher nanohardness ning of deformation, showing only a slight change in slope and no sig­
owing to the higher concentration of C and Mn elements in it [25], and nificant pop-ins to mark the nucleation of new dislocations (Fig. 8); this
soft ferrite demonstrated a lower dislocation density [27]. indicates that plastic deformation occurred at the beginning of

4
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 5. TEM images showing the microstructures of ferrite and small austenite islands under (a) AR, (b) PS10, and (c) PS20 conditions; (d) selected-area-electron-
diffraction (SAED) pattern from the white dashed circle in (a); (e) SAED pattern from the white dashed circle in (b). Large RA islands under (f) AR, (g) PS10, and (h)
PS20 conditions. F: ferrite; γ: austenite; α’: α’-martensite.

Fig. 6. SEM micrograph and EDS analysis of medium manganese steel specimen. (a) SEM micrograph of the region selected for EDS analysis; (b) Concentration
profile of Mn along arrow marked in (a).

5
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 7. (a) SEM image showing indentation impression, (b) corresponding P-h curves of indentations, and (c) the maximum shear stress underneath indenter on large
austenite (AL) grains, small austenite (AS) grains, and ferrite (F) grains under the AR condition.

Fig. 8. P-h curves of indentations on (a) small austenite grains and (b) ferrite grains under the AR, PS10 and PS20 conditions.

indentation under pre-strained conditions. Sufficient mobile disloca­ transformation. Moreover, it is evident from Fig. 9b and c that when
tions were observed in the stressed volume after pre-straining; more­ specimens PS10 and PS20 were cyclically loaded at a low strain
over, the density of pre-existing mobile dislocations was high and the amplitude, the dislocations were rearranged from the large dislocation
radius of the spherical indenter tip was significantly larger than the block to the structure with partially sparse dislocation lines and small
dislocation spacing [31]. Plastic deformation can occur by the move­ dislocation block, resulting in a decrease in dislocation density. Speci­
ment of these pre-existing dislocations at considerably low loads mens PS10 and PS20 contained more film-like RA than the AR specimen,
(essentially immediately from the beginning of deformation) rather than and the segmented martensite formed by pre-straining played a pro­
the nucleation of new dislocations at high stresses [31]. tective role in the encapsulated austenite to “shield” the RA in the core
[32]. Protected austenite with enhanced strain-bearing capacity can
4.2. Influence of cyclic strain on microstructure evolution adjust the stress and strain partitioning behavior, thereby improving the
plasticity and strength properties. Hence, film-like RA cannot be trans­
Fig. 9 shows the TEM results for the AR, PS10 and PS20 specimens formed into martensite with volume expansion to increase the disloca­
after fatigue at an amplitude of 0.2% and 0.8%. It is evident from Fig. 9a tion density at low strain amplitudes [32]. Fig. 9d–f reveal that
and d that the AR specimens containing a large amount of blocky dislocation accumulation is still evident at high strain amplitudes in
austenite underwent martensite phase transformation, which increased comparison to low strain amplitudes, especially under AR and PS10
the dislocation density in ferrite grains near the ferrite-austenite inter­ conditions. At high strain amplitudes, with the accumulation of local­
face owing to volume expansion by austenite-to-martensite ized strain, film-like RA can also be transformed into martensite and

6
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 9. TEM images showing the microstructures of fractured specimens under the (a) AR, (b) PS10 and (c) PS20 conditions fatigued with a strain amplitude of 0.2%;
and (d) AR, (e) PS10, and (f) PS20 conditions fatigued with a strain amplitude of 0.8%.

Fig. 10. XRD showing the volume fraction of RA and dislocation density in Fe–0.1C–5Mn steel at a strain amplitude of (a) 0.2%, (b) 0.4%, and (c) 0.8%.

7
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

contribute to the TRIP effect, which can relieve the stress concentration 7.7% and 6.3%, respectively, which is almost consistent with the XRD
and postpone plastic localization to extend the LCF life [33]. results. Fig. 11d–f illustrate the KAM maps of specimens under these
Interrupted fatigue tests were performed with a significant change in conditions. Combined with Fig. 11a–c, it can be observed that the KAM
the cyclic deformation response of Fe–0.1C–5Mn steel. XRD was used to values under the AR condition were significantly small. For the
determine the volume fraction of RA [16] and dislocation density of deformed specimens, the maximum KAM values were found in
ferrite [17]; the results are shown in Fig. 10, including the specimens martensite islands, while regions of high KAM values were observed at
before fatigue, after fatigue failure, and interrupted fatigue for cycles the ferrite-martensite and austenite-martensite phase boundaries. This is
104, 5 × 102, and 102 with a strain amplitude of 0.2%, 0.4% and 0.8%, because large martensite islands produce large absolute volume
respectively. It was found that specimens AR, PS10, PS15 and PS20 expansion during austenite-to-martensite transformation, which affects
underwent substantial austenite-to-martensite phase transformation at the subsequent deformation of their adjacent ferrite and austenite
the early stage of the cycle; however, the phase transformation variables phases [32]. The GND density values shown in Fig. 11g–i varied from
changed at later stages; this is because RA underwent selective approximately 1.9 × 1013 m− 2 in ferrite to 3.3 × 1014 m− 2 in martensite
martensitic transformation according to its stability in the early cyclic particles under the three conditions. The dislocation densities calculated
loading stage, and RA with lower stability transformed into martensite using EBSD were clearly lower than those obtained using XRD. This is
[34]. Therefore, after the early fatigue process, the remaining film-like because KAM imaging involves the advantage of providing spatial dis­
RA was more stable, which weakened the further transformation of tribution in dislocations [19], and statistically stored dislocations were
RA during later cyclic deformation and reduced the strengthening effect. not included in the EBSD results [35].
The dislocation density of ferrite in the AR specimen increased Fig. 12a and b show the size distribution and average grain diameter
throughout the cycle, while that of ferrite in specimens PS15 and PS20 of austenite for AR and interrupted fatigue at cycle 102 with a total strain
was originally very high and decreased at low strain amplitudes. amplitude of 0.8% under the PS10 and PS20 conditions. As shown in
Therefore, specimens PS15 and PS20 exhibited cyclic softening at low Fig. 12, the austenite grain size for the three conditions was uneven, and
strain amplitudes. the pre-strained and fatigued specimens exhibited a higher proportion of
Fig. 11a–c illustrate the EBSD phase maps + band contrast maps for fine grains, mainly concentrated in the range of 200–600 nm, account­
the AR condition without deformation, as well as the PS10 and PS20 ing for nearly 80% of grains. The average austenite grain size decreased
conditions fatigued at a total strain amplitude of 0.8% interrupted at 102 sequentially as the deformation increased, further proving that more
cycles. The green, light red and dark red lines represent austenite, ferrite unstable AL grains are gradually transformed after larger deformation,
and martensite, respectively. These EBSD maps show that the content of resulting in grain refinement strengthening, which improves the LCF life
austenite in Fe–0.1C–5Mn steels under the three conditions was 26.9%, [36].

Fig. 11. (a, b, c) Phase maps + band contrast maps; (d, e, f) KAM maps and (g,h,i) GND densities, calculated from KAM values under AR cases, and the interrupted
fatigue at cycle 102 with a total strain amplitude of 0.8% in specimens PS10 and PS20.

8
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

Fig. 12. (a) Size distribution and (b) average grain diameter of austenite grains for the AR case, and interrupted fatigue at cycle 102 with a total strain amplitude of
0.8% for specimens PS10 and PS20.

Fig. 13. Schematic of the effect of pre-straining and subsequent cyclic loading on the evolution of martensitic transformation and dislocation arrangements in
Fe–0.1C–5Mn steel.

The microstructural evolution of Fe–0.1C–5Mn steel following pre- blocky RA exhibited less transformation, whereas fine film-like RA
straining and subsequent cyclic loading is schematically summarized hardly transformed into martensite owing to its extremely high stability
in Fig. 13. Under the AR condition, the microstructure was composed of [36]. Under continuous cyclic loading, austenite-to-martensite phase
ferrite (white color) and RA (light gray color) with a small number of transformation led to volume expansion and ferrite squeezing, thus
dislocations, and RA comprised fine film-like, as well as large and small resulting in numerous dislocations in ferrite near the ferrite-martensite
blocky types. During fatigue, the strain amplitude was applied from low phase boundaries [32].
to high, and the unstable large blocky RA was partially transformed into With pre-straining from low to high levels, fresh parallel martensite
martensite gradually, where dislocations previously accumulated. Small beams and fine film-like austenite were gradually formed in the blocky

9
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

austenite, with increasing dislocations in ferrite and RA. When pre- Moreover, for fatigue crack propagation, the crack propagation zone
strained specimens are cyclically loaded at low strain amplitudes, increased with pre-straining (Fig. 14a-c), which implies that the fatigue
some entangled dislocations are rearranged or dissimilar dislocations process is considerably slower during crack propagation and ensures a
are annihilated to disappear, resulting in a lower dislocation density longer fatigue life for Fe–0.1C–5Mn steel. Local stress concentration at
[37], whereas film-like RA is rarely transformed into martensite. When the crack tip in the vicinity of RA can lead to austenite-to-martensite
pre-strained specimens are cyclically loaded at high strain amplitudes, transformation [39]. Therefore, film-like RA can be further trans­
more blocky RA is transformed to martensite owing to uneven austenite formed into martensite, which can make the crack tip blunt; it is also
stability and accumulation of localized strain [33]. Under low-level pre- beneficial in preventing crack growth [15,40]. In addition, thin film-like
strained conditions, the number of dislocations was enormous, with the austenite after plastic deformation can absorb energy during fatigue and
formation of several dislocation blocks at high strain amplitudes. contribute to decelerating fatigue crack propagation when it reaches a
boundary [36]. It is evident from Fig. 5 that the formation of fine film-
4.3. Influence of pre-straining on crack initiation and propagation rate like RA reduces the lamellar spacing, and the low crack growth rate of
pre-strained specimens can be attributed to thin lamellar spacing, which
Fig. 14 shows the fatigue fracture morphology under the AR, PS10, limits the movement of dislocations caused by crack tips [41].
and PS20 conditions at a strain amplitude of 0.4%. It is evident that after
pre-straining, the fatigue source is shinier owing to higher number of 5. Conclusions
cycles in this region and the source is constantly rubbed and squeezed,
indicating that the resistance to crack initiation is stronger. Without pre- The effects of pre-straining on the static uniaxial tensile and LCF
straining, the microstructure is under a low stability condition, which behaviors of Fe–0.1C–5Mn medium Mn TRIP steel were systematically
can easily produce significant deformation under cyclic tensile and investigated. In addition, microstructural evolution during the pre-
compressive loads, resulting in cracking in a local weak area. Fatigue straining and fatigue processes was observed. Based on the experi­
crack initiation can also be attributed to elastic mismatch associated mental results and discussion, the following conclusions were drawn:
with a significant difference in hardness between the matrix phase and
fresh martensite that transformed from large RA, observed at austenite- (1) Pre-straining (10% to 20%) increased the strength of
related boundaries [38]. After pre-straining, the microstructure was Fe–0.1C–5Mn steel, and the increments in YS and UTS were be­
deformed and strengthened, exhibiting uniform strength. Moreover, tween 186–313 and 36–78 MPa, respectively. Meanwhile, the
almost no weak areas and cracks were observed. Owing to the high steel maintained good plastic deformation capacity, with a
hardness of thin film-like austenite, local stress concentration cannot be decrease of 6.4–16.3% in the TE. During re-straining, uniform
easily generated because of the stress-strain distribution between soft/ elongation in the PS20 specimen remained greater than 10%.
hard phases, which can effectively hinder the initiation of crack sources, (2) Compared to small RA, it was easier to transform large blocky
thus improving the LCF life [38]. The film-like austenite caused by pre- austenite into segment martensite with pre-straining, leaving
straining can also lead to a roughness-induced crack closure effect and partially untransformed stable film-like austenite in the interior
prevent crack initiation due to elastic mismatch between the newly and subsequently refining the microstructure.
formed and previous martensite [38]; thus, cracks start propagating
earlier in the AR specimen in comparison to pre-strained specimens.

Fig. 14. Macroscopic fracture morphology under the conditions of (a) AR, (b) PS10 and (c) PS20 at a strain amplitude of 0.4%.

10
M. Zhang et al. International Journal of Fatigue 165 (2022) 107186

(3) Pre-straining delays crack initiation and reduces the crack growth [16] Cheng L, Böttger A, Keijser TD, Mittemeijer EJ. Lattice parameters of iron-carbon
and iron-nitrogen martensites and austenites. Scr Metall Mater 1990;24:509–14.
rate owing to improved strength and the formation of film-like
https://doi.org/10.1016/0956-716X(90)90192-J.
austenite. [17] Williamson GK, Smallman III RE. Dislocation densities in some annealed and cold-
(4) The remaining film-like austenite can still contribute to the TRIP worked metals from measurements on the X-ray debye-scherrer spectrum. Philos
effect during further straining and fatigue processes to avoid Mag 1956;1:34–46. https://doi.org/10.1080/14786435608238074.
[18] Britton TB, Wilkinson AJ. Stress fields and geometrically necessary dislocation
stress concentration, leading to improvements in the LCF life. density distributions near the head of a blocked slip band. Acta Mater 2012;60:
5773–82. https://doi.org/10.1016/j.actamat.2012.07.004.
[19] Zribi Z, Ktari HH, Herbst F, Optasanu V, Njah N. EBSD, XRD and SRS
characterization of a casting Al-7wt%Si alloy processed by equal channel angular
Declaration of Competing Interest extrusion: Dislocation density evaluation. Mater Charact 2019;153:190–8. https://
doi.org/10.1016/j.matchar.2019.04.044.
The authors declare that they have no known competing financial [20] Konijnenberg JP, Zaefferer S, Raabe D. Assessment of geometrically necessary
dislocation levels derived by 3D EBSD. Acta Mater 2015;99:402–14. https://doi.
interests or personal relationships that could have appeared to influence org/10.1016/j.actamat.2015.06.051.
the work reported in this paper. [21] Basquin OH. The exponential law of endurance tests, Acoustic Emission: Current
Practice and Future Directions. Proc ASTM 1910;10(II):625-630.
[22] Manson SS. Behavior of materials under conditions of thermal stress. USA: National
Data availability Advisory Committee for Aeronautics 1954. https://doi.org/10.1016/S1386-9477
(99)00404-X.
The data that has been used is confidential. [23] Coffin L. A study of the effects of cyclic thermal stresses on a ductile meta. Trans
ASME 1954; 76(6): 931-949. https://doi.org/10.1115/1.4015020.
[24] Ly AL, Findley KO. The effects of pre-straining conditions on fatigue behavior of a
Acknowledgments multiphase TRIP steel. Int J Fatigue 2016;87:225–34. https://doi.org/10.1016/j.
ijfatigue.2016.02.004.
[25] Sun XY, Zhang M, Wang Y, Jiang Y, Song Y, Ge NF. Effect of deep cryogenic
This study was supported by the National Key R&D Program of China pretreatment on microstructure and mechanical properties of warm-deformed 7
(2017YFB0304402). Mn steel after intercritical annealing. Mater Sci Eng A 2019;764:
138202.1–138202.11. https://doi.org/10.1016/j.msea.2019.138202.
[26] Lee S, Lee SJ, Cooman B. Austenite stability of ultrafine-grained transformation-
References induced plasticity steel with Mn partitioning. Scripta Mater 2011;65:225–8.
https://doi.org/10.1016/j.scriptamat.2011.04.010.
[1] Lee JY, Kim M, Lee YK. Design of high strength medium-Mn steel using machine [27] He BB, Liang ZY, Huang MX. Nanoindentation investigation on the initiation of
learning. Mater Sci Eng C 2022;843:143138. https://doi.org/10.1016/j. yield point phenomenon in a medium Mn steel. Scr Mater 2018;150:134–8.
msea.2022.143148. [28] Barnoush A, Yang Bo, Vehoff H. Effect of hydrogen and grain boundaries on
[2] Liu CT, Song Y, Li FS, Wang SQ, Mao RC. Development method and experimental dislocation nucleation and multiplication examined with a NI-AFM. In: Haug R,
verification of fatigue load spectrum of anti-yaw damper for rail vehicles. Eng Fail editor. Advances in Solid State PhysicsAdvances in Solid State Physics. Berlin,
Anal 2022;138:342–53. https://doi.org/10.1016/j.engfailanal.2022.106298. Heidelberg: Springer Berlin Heidelberg; 2008. p. 253–69.
[3] Huo YC, Gao LH. Strain-induced martensitic transformation in fatigue crack tip [29] Shim S, Bei H, George EP, Pharr GM. A different type of indentation size effect.
zone for a high strength steel. Mater Charact 2005;55:12–8. https://doi.org/ Scripta Mater 2008;59:1095–8. https://doi.org/10.1016/j.
10.1016/j.matchar.2005.02.004. scriptamat.2008.07.026.
[4] Hu ZG, Zhu P, Meng J. Fatigue properties of transformation-induced plasticity and [30] Lani F, Furnémont Q, Rompaey TV, Delannay F, Jacques PJ, Pardoen T. Multiscale
dual-phase steels for auto-body lightweight: Experiment, modeling and mechanics of TRIP-assisted multiphase steels: II. Micromechanical modelling. Acta
application. Mater Des 2010;31:2884–90. https://doi.org/10.1016/j. Mater 2007;55:3695–705. https://doi.org/10.1016/j.actamat.2007.02.015.
matdes.2009.12.034. [31] Ahn TH, Oh CS, Lee K, George EP, Han HN. Relationship between yield point
[5] Padmanabhan KA, Sankaran S. Fatigue behavior of a multiphase medium carbon V- phenomena and the nanoindentation pop-in behavior of steel. J Mater Res 2012;
bearing microalloyed steel processed through two thermomechanical routes. 27:39–44. https://doi.org/10.1557/jmr.2011.208.
J Mater Process Technol 2008;207:293–300. https://doi.org/10.1016/j. [32] Wang M, Huang MX. Abnormal TRIP effect on the work hardening behavior of a
jmatprotec.2008.06.052. quenching and partitioning steel at high strain rate. Acta Mater 2020. https://doi.
[6] Tsuzaki K, Nakanishi E, Maki T, Tamura I. Low-cycle fatigue behavior in org/10.1016/j.actamat.2020.02.035. https://doi.org/188:551-559.
metastable austenitic steel accompanying deformation-induced martensitic [33] Glage A, Weidner A, Biermann H. Effect of austenite stability on the low cycle
transformation. Trans Iron Steel Inst Jpn 1983;23:834–41. https://doi.org/ fatigue behavior and microstructure of high alloyed metastable austenitic cast
10.2355/isijinternational1966.23.834. TRIPsteels. Procedia Eng 2010;2(1):2085–94.
[7] Wang B, Zhang P, Duan QQ, Zhang ZJ, Yang HJ, Pang JC, et al. High-cycle fatigue [34] Grigorescu AC, Hilgendorff PM, Zimmermann M, Fritzen CP, Christ HJ. Cyclic
properties and damage mechanisms of pre-strained Fe-30Mn-0.9C twinning- deformation behavior of austenitic Cr–Ni-steels in the VHCF regime: Part I –
induced plasticity steel. Mater Sci Eng A 2017;679:258–71. https://doi.org/ Experimental study. Int J Fatigue 2016;93:250–260. https://doi.org/ 10.1016/j.
10.1016/j.msea.2016.10.043. ijfatigue.2016.05.005.
[8] Kamaya M. Characterization of microstructural damage due to low-cycle fatigue by [35] Ji DP, Zhang M, Zhu DL, Luo SW, Li L. Influence of microstructure and pre-
EBSD observation. Mater Charact 2009;60:1454–62. https://doi.org/10.1016/j. straining on the bake hardening response for ferrite-martensite dual-phase steels of
matchar.2009.07.003. different grades. Mater Sci Eng A 2017;708:129–41. https://doi.org/10.1016/j.
[9] Fredriksson K, Melander A, Hedman M. Influence of prestraining and ageing on msea.2017.09.127.
fatigue properties of high-strength sheet steels. Int J Fatigue 1988;10:139–51. [36] Zhao P, Zhang B, Cheng C, Misra RDK, Gao G, Bai B, et al. The significance of
https://doi.org/10.1016/0142-1123(88)90056-4. ultrafine film-like retained austenite in governing very high cycle fatigue behavior
[10] Sherman AM. Fatigue properties of high strength-low alloy steels. Metall Mater in an ultrahigh-strength MN–SI–Cr–C steel. Mater Sci Eng A 2015;645:116–21.
Trans A 1975;6:1035–40. https://doi.org/10.1007/BF02661357. https://doi.org/10.1016/j.msea.2015.07.095.
[11] Le Q, Hong TK, Kridli G, Khosrovaneh AK, Yan B. Effect of prestrain paths on [37] Lambers HG, Ruesing CJ, Niendorf T, Geissler D, Freudenberger J, Maier HJ. On
mechanical behavior of dual phase sheet steel. Int J Fatigue 2009;31:607–15. the low-cycle fatigue response of pre-strained austenitic Fe61Mn24Ni6.5Cr8.5 alloy
https://doi.org/10.1016/j.ijfatigue.2008.03.028. showing TWIP effect. Int J Fatigue 2012;40:51–60. https://doi.org/10.1016/j.
[12] Sun H, Wei K, Yang X, Xiao Z, Wu Y. Effects of pre-strain and annealing on the ijfatigue.2012.01.002.
fatigue properties of complex phase steel CP800. Int J Fatigue 2019;131:105364. [38] Zhang Z, Koyama M, Wang MM, Tsuzaki K, Tasan C, Noguchi H. Effects of lamella
https://doi.org/10.1016/j.ijfatigue.2019.105364. size and connectivity on fatigue crack resistance of TRIP-maraging steel. Int J
[13] Aydin H, Essadiqi E, Jung IH, Yue S. Development of 3rd generation AHSS with Fatigue 2017;100:176–86. https://doi.org/10.1016/j.ijfatigue.2017.03.040.
medium Mn content alloying compositions. Mater Sci Eng A 2013;564:501–8. [39] Yang J, Wang TS, Zhang B, Zhang FC. High-cycle bending fatigue behaviour of
https://doi.org/10.1016/j.msea.2012.11.113. nanostructured bainitic steel. Scripta Mater 2012;66:363–6. https://doi.org/
[14] Song SW, Lee JH, Lee T, Chong SL. Effect of the amount and temperature of 10.1016/j.scriptamat.2011.11.033.
prestrain on tensile and low-cycle fatigue properties of Fe-17Mn-0.5C TRIP/TWIP [40] Huo CY, Gao HL. Strain-induced martensitic transformation in fatigue crack tip
steel. Mater Sci Eng A 2017;696:493–502. https://doi.org/10.1016/j. zone for a high strength steel. Mater Charact 2005;55:12–18. https://doi.org/
msea.2017.04.099. 10.1016/j.matchar.2005.02.004.
[15] Robertson L, Hilditch T, Hodgson P. The effect of prestrain and bake hardening on [41] Abareshi M, Emadoddin E. Effect of retained austenite characteristics on fatigue
the low-cycle fatigue properties of TRIP steel. Int J Fatigue 2008;30:587–94. behavior and tensile properties of transformation induced plasticity steel. Mater
https://doi.org/10.1016/j.ijfatigue.2007.06.002. Des 2012;32:5099–105. https://doi.org/10.1016/j.matdes.2011.06.018.

11