Professional Documents
Culture Documents
Proceedings
https://www.nims.go.jp/group/sdg/icas2022/program/
Symp
No page session name title
osium
051 127-128 1 Martensite/bainite Hideyuki Ohtsuka First-principles calculation of the atomic interaction and tetragonality in an Fe-C system
052 129-131 1 Martensite/bainite Maria J. Santofimia Navarro Investigating the formation of bainite in the presence of martensite
053 132 2 Advanced steels 2 Mingxin Huang Negative strain rate sensitivity in steels
055 133 2 Advanced steels 2 Jae-Bok Seol Near atomic-scale comparison of passive film on a high-Cr high-Mn steel with those on typical stainless steels
056 134 2 Advanced steels 2 Shiro Torizuka MECHANICAL PROPERTUES of SINGLE VARIANT MARTENSITE from FERRITE + Mn-RICH CEMENTITE STRUCTURE in 0.1C-2Si-5Mn STEELS
057 135-136 2 Fracture Masaki Tanaka Delamination in wire-drawn pearlitic steels
058 137 2 Fracture Kazuki Shibanuma Fatigue life prediction of steels based on multiscale model synthesis
059 138-141 2 Fracture Yuuji Kimura Delayed Fracture Behavior of Ultrahigh-Strength Low-Alloy Steel Processed by Warm Tempforming
060 142 2 Fracture Yoshikazu Todaka Hydrogen Embrittlement Behavior of Nanostructured SUS316LN Austenitic Stainless Steel Produced by Heavy Cold-rolling and High-pressure Torsion Straining
061 143-144 2 Martensite Nobuo Nakada Heterogeneous cleavage fracture of lath martensite induced by Bain strain
062 145 2 Martensite Shimpei Yoshioka Formation of inhomogeneous microstructure in martensitic steel during cooling and its effect on yield strength
063 146-149 2 Martensite Osamu Idohara Estimation of true hardness of as-quenched martensite without auto-tempering
064 150-153 2 Martensite Hiroyuki Kawata Grain refinement strengthening on Fe?18Ni low carbon lath martensite
065 154-157 3 Fundamentals of Corrosion Masatoshi Sakairi Effects of metal cation on corrosion of structural metals in aqueous environments
066 158 3 Fundamentals of Corrosion Yoshinao Hoshi 3D Impedance Spectroscopy to Investigate Film Growth Rate of Chromate Film Formed on Steel
068 159-160 3 Fundamentals of Corrosion Mariko Kadowaki Roles of Interstitial Elements in Anodic Dissolution Resistance of Martensitic Steels
070 161-163 3 Corrosion resistant materials / evaluation methods Fumio Yuse Corrosion resistant steel for bridges and life cycle cost
071 164-167 3 Corrosion resistant materials / evaluation methods Yusuke Tsutsumi Fundamentals of corrosion engineering of metallic materials and development of ultrahigh corrosion resistant stainless steels by novel laser techniques
072 168-169 3 Corrosion resistant materials / evaluation methods Kotaro Doi Corrosion Behavior of Steel Rebar in Concrete Studied with Hyperbaric-oxygen Accelerated Corrosion Test
074 170-172 1 High alloyed steels Kyung-Tae Park Tensile Fracture of High Mn TWIP steel
075 173 1 High alloyed steels Hung-Wei (Homer) Yen Nano/microstructural engineering in lightweight high-entropy steel with multiple characteristics
076 174 1 High alloyed steels Ivan Gutierrez-Urrutia Study of microbanding in a Fe-30Mn-6.5Al-0.3C low-density steel deformed at cryogenic temperatures: Mechanisms and contribution to the mechanical behavior
077 175 1 High alloyed steels Mei Zhang Manufacturing superior property medium Mn TRIP steel free of Luders bands by warm deformation + IA process
078 176-178 2 Strength & Deformation 3 Heung Nam Han Mechanistic view on electric current induced kinetic enhancement and its various examples in steels
079 179-182 2 Strength & Deformation 3 Emmanuel De Moor Double Soaking of Medium Manganese Steels
080 183- 2 Strength & Deformation 3 Masato Wakeda Atomistic modeling and analysis of dislocation behaviors near grain boundary in iron
082 184-187 3 Fatigue Takashi Nakamura In situ observation of small fatigue cracks in high-strength metals using synchrotron radiation micro and nano computed tomography
083 188-190 3 Fatigue Hisao Matsunaga Various strength properties of low-alloy steels in high-pressure hydrogen gas and their hydrogen compatibility
084 191 3 Fatigue Motomichi Koyama Plastic deformation and dislocation motion at a fatigue crack tip of a bi-modal Ti-6Al-4V alloy during dwell loading
085 192-195 3 Fatigue Hideaki Nishikawa Understanding the effect of microstructure on fatigue properties of weld heat affected zone of low carbon steel based on microscopic fatigue crack initiation and growth behavior
086 196-199 Plenary Lecture 3 Setsuo Takaki (Japan) Yielding mechanism in polycrystalline ferritic steel
https://www.nims.go.jp/group/sdg/icas2022/program/
Role of Interstitials in Austenitic Stainless Steels: C vs N
Sung-Joon Kim1,*, Kyung-Shik Kim1 and Jee-Hyun Kang2
1
Graduate Institute of Ferrous and Energy Materials Technology, POSTECH, Pohang, 36763, Korea
2
School of Material Science and Engineering, Yeungnam University, Gyeongsan 35481, Korea
*email: sjkim1@postech.ac.kr
Austenitic stainless steels are known for their excellent corrosion resistance, toughness, and formability.
Furthermore, they are resistant to hydrogen embrittlement, which enables their application in hydrogen environment.
However, their strength is low; yield and tensile strengths are approximately 250 MPa and 600 MPa, respectively. To enhance
their strengths, interstitial alloying elements, such as carbon and nitrogen, are added.
The different behavior of C and N in austenitic stainless steels has been studied extensively. Gavriljuk et al. claimed
that different properties brought by the interstitials is attributed to their different influences on the electronic structure of the
steels. Because N and C change the distribution of density of states differently, free electron concentration is increased more
effectively by N than by C. Thus, nitrogen is known to promote metallic interatomic bonding while carbon induces covalent
bonding 1). Consequently, for neighboring carbon atoms, nearest interstitial sites are preferred, which results in short-range
clustering. On the other hand, nitrogen atoms tend to occupy second nearest interstices which would generate short-range
ordering 2).
Accordingly, various properties of austenitic stainless steels are altered in different ways. Precipitation of either
M23C6 or Cr2N was less likely to occur in N bearing steels than in C containing steels 3). Moreover, more lattice distortion was
caused by nitrogen, which resulted in higher strengthening effect than by the same concentration of carbon 4,5). Another
difference was the segregation tendency at grain boundaries. Carbon segregated to grain boundaries which would contribute
to intergranular corrosion.
Recently, different effects by C and N on austenite stability and staking fault energy (SFE) of austenitic stainless
steels have been intensively investigated. Masumura et al. reported that the effects by C and N on austenite stability depended
on the type of stability. When the steels are quenched to form athermal α'-martensite, nitrogen was more effective in
stabilizing austenite than carbon. On the other hand, a N containing steel induced less fraction of α'-martensite during
deformation 6). The trend of SFE alteration is controversial. It was reported that carbon increased 6-9), or did not affect 10-12) SFE,
while nitrogen increased 6,7), decreased 13-15), or non-monotonously affected 12-,16) SFE. This suggests that SFE variation may
depend on alloy systems.
Although different effects of C and N on austenite have been continuously studied, they were rarely investigated in
the same base alloy system. Inspired by this fact, the present study dealt with the austenitic stainless steels which only varies
with the type of interstitials while the concentrations of other alloying elements were equivalent. Deformation behavior at
room temperature and at high temperature as well as hydrogen embrittlement were studied by focusing on the different effects
of carbon and nitrogen.
Alloys were designed so that up to 0.3 wt% of either C or N is soluble. By using ThermoCalc with TCFE7
database, the suitable base alloy that satisfied the solubility condition was acquired to be Fe-15Cr-15Mn-4Ni-0.3Si (wt%).
The alloys with 0.1, 0.2, 0.3 wt% C were called as C1, C2, and C3, respectively, and their nitrogen counterpart was named as
N1, N2, and N3 (Table 1). All alloys were hot-rolled to 13 mm-thick sheets. For the room temperature deformation study, the
alloys were machined into tensile specimens, gauge dimension of which was 33×10×1.5 mm3. They were all annealed at
1100°C for 15 or 30 min so that the resulting grain size was ~50 μm. SFE values were obtained by thermodynamic
calculation and XRD (Table 1). Tensile tests were conducted with a constant crosshead speed of 1.5 mm min-1. For the
Hall-Petch relationship study, hot-rolled sheets were cold-rolled to either 1.4 or 2.3 mm, and solution-treated for 1-180 min at
1000 or 1200°C. Yield strength was obtained with tensile testing at a constant crosshead speed of 0.1 mm min-1. The gauge
dimension of tensile samples was 25×6×1 mm3. For the high temperature deformation study, hot compression testing was
performed. Cylindrical samples were prepared from hot-rolled sheets. They were annealed at 1200°C for either 1.5 or 5 min,
The interstitial effect on the tensile behavior at room temperature was most prominent in yield strength while it was
insignificant in strain hardening. The effect on yield strength was larger with N addition than C addition possibly due to higher
interstitial-dislocation binding energy for nitrogen 21). The difference was also noticeable in austenite stability. Although
ε-martensite was fully suppressed in the N containing steels on quenching, it appeared in the C containing steels and Base.
Moreover, the fractions of deformation-induced martensites were higher for the carbon bearing steels than for their
counterpart containing the equivalent concentrations of nitrogen. In other words, nitrogen was more effective for suppressing
both athermal and deformation-induced martensites, which was consistent with the variation in SFE (Table 1). Although the
martensites fractions were different, the morphologies were similar in both C and N containing steels. In all interstitial
containing steels, not only martensite but also mechanical twinning formed, and α'-martensite formed at the intersections of
ε-martensite. The different effects of the interstitials also appeared in dislocation structure. While the C bearing steels exhibited
wavy slip, N bearing steels showed planar slip. However, such difference insignificantly affected strain hardening.
To study Hall-Peth relationship, effective grain size was estimated by considering the number of twin boundaries
per grain. The effective grain size depended on the correction factor which implied the effectiveness of a twin boundary to act
as a dislocation barrier. When the steels were annealed at different temperatures and times to obtain various grain sizes, the
number of twin boundaries per grain was higher in N3 than in C3, which resulted in smaller effective grain size in N3. Both
C3 and N3 followed Hall-Petch relationship. For all values of the correction factor from 0 to 1, Hall-Petch slope was higher
for N3 than for C3. Hall-Petch slope is known to depend on Taylor factor, Burgers vector, shear modulus, Poisson ration, and
the stress τc required to activate dislocation source. Because all the factors except τc was similar, it is concluded that τc/G was
smaller for C3 than for N3. τc/G was suggested to be related to dislocation glide behavior. Because slip planarity tendency was
higher in N3, which implied the difficulty in cross slip, slip transmission would be suppressed which resulted in higher τc/G.
From high temperature compression, it was found that both C and N bearing steels followed the general trend; flow
stress was lowered by increasing temperature and decreasing strain rate. As the interstitial concentration increased, flow stress
increased accordingly; this trend appeared more strongly by N addition than by C addition. Moreover, activation energy for
deformation was higher for nitrogen containing steels. These suggested that deformation became more difficult by N addition
than by C addition. At the temperature ≥ 1000°C, strain hardening became negligible owing to dynamic softening. Especially,
dynamic recovery rate was revealed to be lower for N bearing steels, which was caused by suppression of cross slip, i.e.,
promotion of slip planarity by short-range ordering. The microstructure after the hot compression testing contained
recrystallized grains which nucleated along elongated grains. It was revealed that the size of the recrystallized grains increased
with increasing temperature and decreasing strain rate. Moreover, the size decreased with increasing interstitial concentration;
Acknowledgement
The authors are thankful for the financial support by National Research Foundation of Korea (NRF-2020R1A2B5B01001900).
References
1) V. Gavriljuk, B. Shanina and H. Berns: Acta Mater. 48 (2000) 3879.
2) V. Gavriljuk and V. Nadutov: Phys. Met. Metallogr. 55 (1983) 520.
3) K.-S. Kim, J.-H. Kang and S.-J. Kim: Mater. Sci. Eng. A 712 (2018) 114.
4) K. Irvine, T. Gladman and F. Pickering: JISI July (1969) 1017.
5) N. Saenarjhan, J.-H. Kang and S.-J. Kim: Mater. Sci. Eng. A 742 (2019) 608.
6) T. Masumura, N. Nakada, T. Tsuchiyama, S. Takaki, T. Koyano and K. Adachi: Acta Mater. 84 (2015) 330.
7) T.-H. Lee, E. Shin, C.-S. Oh, H.-Y. Ha and S.-J. Kim: Acta Mater. 58 (2010) 3173.
8) A. Abbasi, A. Dick, T. Hickel and J. Neugebauer: Acta Mater. 59 (2011) 3041.
In modern high strength steel, sophisticated control of transformation microstructure is required for better strength and
ductility/toughness balance. In addition to fraction, size and distribution of constituent phases, a degree of local partitioning of
carbon (C) and substitutional alloying elements (M) during phase transformations is an important factor in development of dual
phase and TRIP steels which are typical advanced high strength steels. Therefore, quantitative understanding of alloying effects
on phase transformation kinetics in Fe–C system becomes increasingly important. Recently, our group has been carrying out the
research on the effects of hetero-phase interface (i.e., interphase boundary) in various phase transformations and precipitation
mainly in low alloy steels1-4) . In this presentation, alloy partitioning and precipitation during austenite/ferrite interphase boundary
migration are discussed.
Interphase boundary nature has a direct impact on the transformation kinetics. The growth of grain boundary ferrite
in Fe-C alloys is contolled by volume diffusion of carbon in austenite by local equilibrium, but is slower than the expected from
calculation assuming the migration of incoherent boundaries The reason for this was considered to be effects of the boundary
structures mixed by partly coherent and incoherent ones. Boundary velocity (V) is proportional to driving force (ΔG) to move
the boundary forward as described below.
V = M・ΔG
where M is boundary mobility. If the interphase boundary can move freely, the entire chemical driving force is available for
boundary migration. However, there are various ‘frictions’ on the interphase boundary migration exerted from various
components of the microstructure as represented in Fig. 1(a). Atomic displacement at the boundary causes intrinsic friction. Solute
atoms segregated at the boundary provide another friction known as 'solute drag'. Dislocations or fine second-phase particles may
add further barriers for boundary migration. These kinds of friction or barrier result in the energy dissipation during boundary
migration and decrease in the net driving force ΔG. Fig. 1(b) shows the effect of such energy dissipations in boundary migration
on partitioning of carbon in the Fe-C binary system. The same situation can be imagined for the (Fe, M)-C pseudo-binary case
with para-equilibrium. Total energy dissipation caused from various sources is balanced by driving force (ΔGdvr) at migrating
Fig. 1: (a) Schematic illustration of microstructure components causing energy dissipation in boundary migration. (b) Deviation of
carbon partitioning from local equilibrium due to energy dissipation at a migrating boundary for γ→α forward transformation. (c)
Energy dissipation for BF and WF transformations with respect to para equilibrium 2, 3).
In the interphase precipitation, crystallographic relationship between ferrite (α), austenite (γ) and carbide plays a major
role to determine the carbide dispersion and corresponding strength. Recently, we have demonstrated that a larger driving force
for precipitation results in a finer carbide dispersion with incoherent α/γ boundary 4).. Fig. 2 shows the effect of boundary coherency
(here represented as deviation angle from the K-S OR, i.e., Δθ) on the VC particle dispersion and corresponding nano-hardness.
V is thought to have an attractive interaction to α/γ boundary and exert solute drag effect as in the case of Nb addition. More V
atoms segregate at non K-S boundary (Fig. 2(d)) and such a situation is advantageous for carbide precipitation thermodynamically
and kinetically. As the deviation angle becomes larger, which means that the α/γ boundary becomes less coherent, the number
density of VC particles increases drastically (Fig. 2(e)). Fig. 2(f) clearly shows that the increase of particle density leads to higher
hardness through the Orowan-type dispersion hardening.
It is important to make incoherent α/γ boundaries to obtain high strength through interphase precipitation. Although
this is rather challenging, one possible way is use of ferrite transformation from deformed austenite. Fine-grained austenite
structure inhibits formation of Widmanstätten ferrite so that recrystallization of austenite can be effective to obtain granular ferrite
structure. Ferrite transformation in unrecrystallized austenite can be promising since local orientation fluctuation in the deformed
austenite will accomplish loss of interfacial coherency during growth even though initial nucleus had near K-S OR with austenite.
Use of nucleation at particles is also a good way 5, 6). Nucleation at grain boundary precipitate or intragranular inclusion effectively
Acknowledgement
The authors gratefully acknowledge financial supports by the Innovative Structural Materials Project funded by the New Energy and
Industrial Technology Development Organization, CREST Basic Research Program “Creation of Innovative Functions of Intelligent
Materials on the Basis of Element Strategy” funded by Japan Science and Technology Agency, Grant-in-Aid for Scientific Research (B)
No. 23360316 (2011–2013) and (A) No. 17H01330 (2017-2019) funded by Japan Society for the Promotion of Science.
References
1) T. Furuhara:, Y. -J. Zhang, G. Miyamoto: IOP Conf Series: Mater Sci Engng, 580(2019), 012005.
2) G. Miyamoto, T. Furuhara: ISIJ International, 60(2020), pp. 2942–2953.
3) M. Sato, G. Miyamoto, T. Furuhara: Netsu Shori (J. Jpn. Soc. Heat Treat.), 61(2021), pp. 82-85.
4) Y. -J. Zhang, E. Chandiran, H. -K. Dong, N. Kamikawa, G. Miyamoto, T. Furuhara: JOM, 73(2021), pp. 3214-3227.
5) Y. -J. Zhang, G. Miyamoto, K. Shinbo, T. Furuhara: Acta Mater., 128(2017), pp. 166-175.
6) G. Miyamoto, R. Hor, B. Poorganji, T. Furuhara: Metall. Mater. Trans. A, 44(2013), pp. 3436-3443.
Fig. 2: (a) Ferrite orientation maps of the Fe–0.1C–1.5Mn–0.4V alloy transformed at 923 K. Circles represent near K-S boundaries
and triangles indicate non K-S boundaries. (b) Schematic illustration describing angular deviation from K-S OR. 1D concentration
profiles of V and C obtained by 3DAP analyses of (c) near K-S and (d) non K-S boundaries. Variations in (e) number density of VC
precipitate and (f) nanohardness of α grains with Δθ in Fe-0.1C-1.5Mn-0.4V 1,4).
Keywords: ultra-high-strength steel sheet, multiphase microstructure, ductility, stretch-flangeability, carbon mapping, spring back,
Bauschinger effect
1. Introduction
Due to the strong demand for minimization of global climate change, carbon dioxide (CO2) emission in automobile
industry has been required to be reduced drastically. Not only conventional countermeasures for energy consumption
reduction but also transition from gasoline to electric power train such as EV will be necessary for the challenging target of
CO2 emission reduction. However, even in EV, weight reduction of body weight is still a major issue to compensate the large
weight gain by pretty heavy battery. Thus, although weight reduction in automotive body has been the continuous
technological trend for these 30 years originating from corporate average fuel economy (CAFE)1) regulation, this trend will
continue in the coming century.
From the view point of material selection, non-ferrous light materials such as aluminum, magnesium, carbon fiber
reinforced plastic (CFRP) and so on might have been also considered. However, the major material for the weight reduction in
automobile would be steels taking into consideration of the stable and economical supply for the growing number of vehicles.
Without doubt, application of high-strength steel sheets has been an effective way in the past few decades and
would have been the same in the coming decades by thinning the automotive parts with maintaining or increasing crash
strength of vehicles. Especially, ultra-high-strength steels (UHSSs) with tensile strength over 980MPa have been widely used
for the structural automotive parts.
Due to the increasing demand of automobile weight reduction, maximization of strength of all the structural
automotive parts has been accelerated through the stronger collaboration than ever between steel makers and automobile
manufacturers.
Ferrite
0%
Martensite
5µm
0.4
0.2
0.1
0.0
0 10 20 30 40
Distance /μm
Fig. 4. Example of carbon quantitative mapping of a dual phase steel (a,b) and
line analysis (c) along the yellow line in b) using the developed FE-EPMA.
5. Summary
Demand for the application of UHSSs will become more and more aggressive in a carbon neutral society. At the
same time, not only advanced material design but also microstructural analyzing techniques and solution techniques in press
forming and welding will be required to be continuously developed. Furthermore, computational materials science and data
science will become essential for promoting the development of advanced UHSSs.
Acknowledgement
New FE-EPMA for quantitative carbon analysis was developed under a project of the Innovative Structural Materials
Association (ISMA) supported by the New Energy and Industrial Technology Development Organization (NEDO) in Japan.
References
1) https://www.nhtsa.gov/laws-regulations/corporate-average-fuel-economy.
2) https://www.mlit.go.jp/jidosha/content/001337988.pdf
3) https://www.jfe-steel.co.jp/en/release/2020/201223.html
4) T. Nakagaito, H. Matsuda, Y. Nagataki and K. Seto: ISIJ Int., 57 (2017), pp.380-387.
5) K. Hasegawa, K. Kawamura, T. Urabe and Y. Hosoya: ISIJ Int., 44 (2004), pp.603-609.
6) N. Tsuchida, T. Tanaka and Y. Toji: ISIJ Int., 60 (2020), pp.1349-1357.
7) D. K. Matlock and J. Speer: Proc. The 3rd ICASS, Gyeongju Korea (2006), pp.774-781.
8) T. Yamashita, Y. Tanaka, M. Nagoshi and K. Ishida: Sci. Rep., 6 (2016), Srep29825.
9) T. Yamashita, M. Enomoto, Y. Tanaka, H. Matsuda and M. Nagoshi: ISIJ Int., 58 (2018), pp.1079-1085.
10) S. Tobita, T. Shimmiya, Y. Yamasaki and J. Hiramoto: Mater. Trans., 62 (2021), pp.1750-1756.
1
ICAS 2022 The 7th International Conference on Advanced Steels -12-
0.75 wt. % Si were only partially stable, as a fraction of h carbides decomposed to q (Fe3C) (Figs.
1d-f). In the case of steel with 0.25 wt. % Si, only the existence of q precipitates was evident [3].
Fig. 1: TEM BF (a,d,e), DF (b,f) images and corresponding SADP of RA (c) of 0.75 wt. % steel.
Furthermore, the size of q precipitates was found to be relatively large compared to h
carbides, suggesting that the growth of cementite occurred during partitioning at 300 °C. The
carbon atom map in Fig. 2a shows some carbon agglomerated regions in the 0.75 wt. % Si steel
specimen. Owing to the respective stoichiometries of different carbides, a carbide with a carbon
content of about 25 at. % can be claimed as q and the other one with a carbon content of about 31
at. % can be claimed as h-carbide. Apart from carbides, segregation of carbon clusters in
martensite lath boundaries was clearly revealed by APT.
Fig. 2: (a,b) Iso-concentration surfaces & (c,d) corresponding proxigrams of 0.75 wt. % Si steel.
Additionally, an attempt has been made to detect the possible formation of a metastable,
hexagonal omega (w) phase within the boundaries of nano-twinned high carbon martensite. Fig.
3a represents a magnified image of twinned martensite structure in 1.5 wt. % Si steel. The selected
area diffraction pattern (SAED) on [011]BCC zone axis (inset in Fig. 3b) shows the twinned regions
having a (211) twin plane aligned in [111] twinning direction. In addition, there are two other spots
found to be present along with the (211) plane. The dark field image from one of the extra spots
showed the existence of nanosized thin plates in between the twinning planes (Fig. 3b). The
crystals in this twinned martensite most likely correspond to ω phase [7].
2
ICAS 2022 The 7th International Conference on Advanced Steels -13-
Fig. 3: (a) TEM bright, and (b) dark field images of DQ&P processed 1.5 wt. % Si steel.
3
ICAS 2022 The 7th International Conference on Advanced Steels -14-
strength and good ductility. The key factor in governing a good combination of strength and
ductility, however, lies in properly controlling the volume fraction and stability of RA that must
be finely divided in the matrix and hence, becomes a challenging objective in designing new
processing routes. Though the effects of different methods such as pre-existing austenite,
deformation and high temperature partitioning have been studied individually, a combination of
these process parameters on the processing and intercritical annealing of medium Mn steels has
still not been systematically investigated. A new combination of factors enhancing the stabilization
of austenite, including pre-existed austenite among quenched martensite, prior deformation, and
partitioning at high temperatures was employed to create a multi-component refined
microstructure in Fe-4Mn-0.31C-2Ni-0.5Al-0.2Mo steel [8]. The results demonstrated that the
combinatorial approach will be beneficial in accelerating austenite stabilization and microstructure
refinement. The final microstructure of the specimen after 0.4 strain deformation of 73%
martensite–27% austenite at 250 °C and subsequent partition-annealing at 650 °C for 20 min was
composed of ultrafine-grained equiaxed ferrite, 32% RA (avg. grain size of 250 nm) along with
some fresh martensite during final cooling and few carbide precipitates, as shown in Fig. 5.
(a) (b)
Fig. 5: EBSD IQ & Phase maps of the deformed & partition-annealed medium-Mn steel sample.
Acknowledgement
This funding of this research activity under the auspices of Genome of Steel (Profi3) project by
the Academy of Finland through grant # 311934 is gratefully acknowledged.
References
1) M.C. Somani, D.A. Porter, L.P. Karjalainen, D.K. Misra, Evaluation of DQ&P Processing Route for the
Development of Ultra-high Strength Tough Ductile Steels, Int. J. Metall. Eng. (2013)154–160.
2) Raabe, D., Sun, B., Kwiatkowski Da Silva, A. et al. Current Challenges and Opportunities in Microstructure-
Related Properties of Advanced High-Strength Steels. Metall Mater Trans A 51, 5517–5586 (2020).
3) I. Miettunen, S. Ghosh, M.C. Somani, S. Pallaspuro, J. Kömi, Competitive mechanisms occurring during Q&P of
three silicon variants of 0.4 wt.% carbon steels, J. Mater. Res. Technol 11 (2021) 1045-1060.
4) I. Miettunen, S. Ghosh, M.C. Somani, S. Pallaspuro, D. Porter, J. Kömi, Effect of Si content on the decomposition
of austenite in 0.4C steel during Q&P treatment. Mater. Sci. Forum 1016 (2020) 1361-1367.
5) S. Ghosh, I. Miettunen, M.C. Somani, J. Kömi, D. Porter, Nanolath martensite-austenite structures engineered
through DQ&P for developing tough, ultrahigh strength steels, Mater. Today Proceed. 46 (2021) 2131-2134,
6) S Ghosh, P Kaikkonen, V Javaheri, A Kaijalainen, I Miettunen, M Somani, J Kömi, S Pallaspuro, Design of tough,
ductile DQ&P AHSS with tailored Si content, J Mater. Res. Technol., 17, 2022, 1390-1407.
7) T. Liu, D. Zhang, Q. Liu, Y. Zheng, Y. Su, X. Zhao, J. Yin, M. Song, D. Ping. A new nanoscale metastable iron
phase in carbon steels, Sci. Rep. 5 (2015) 15331
8) S. Sadeghpour, M.C. Somani, J. Komi, L.P. Karjalainen, A new combinatorial processing route to achieve an
ultrafine-grained, multiphase microstructure in a medium Mn steel, J. Mater. Res. Tech. 15 (2021) 3426-3446.
4
ICAS 2022 The 7th International Conference on Advanced Steels -15-
Formation of coarse columnar austenite grains during solidification of peritectic steels
Munekazu Ohno1,*
1
Division of Material Science and Engineering, Hokkaido University, Sapporo, Hokkaido 060-8628, Japan
*email: mohno@eng.hokudai.ac.jp
Keywords: peritectic solidification, continuous casting, coarse columnar grains, phase-field simulations, directional solidification
Coarse columnar γ grain (CCG) structure is formed in the vicinity of continuously cast (CC) slab surface of peritectic
solidified carbon steels. Formation of the CCG structure is a serious problem, because it causes occurrence of surface cracking of
CC slabs1,2). Elucidation of formation mechanism of the CCG structure and also its kinetics is of great importance for production
of high-quality CC slabs.
It was found by means of a rapid unidirectional solidification experiment that
the formation of CCG structure is ascribed to a discontinuous grain growth3,4). The
microstructure during the solidification is schematically shown in Fig. 1. This was
drawn based on the microstructure observed in a sample quenched during rapid
directional solidification experiment3), the cooling condition of which is close to that in
continuous casting. The columnar dendrite of δ phase grows in the direction of
temperature gradient, which is followed by the formation of γ grain structure in δ
dendrite. Fine columnar γ grains (FCG) are formed ahead of CCGs. Importantly, the
change from FCG to CCG structures is discontinuous. In other words, there is a clear
distinction between FCG and CCG regions and, hence, the boundary between these
regions, called FCG/CCG region boundary (FCRB), can be identified. The temperature
at FCRB is close to the equilibrium temperature Tγ below which γ single phase is stable.
Namely, in the case of hyperperitectic steels, the liquid phase exists around FCG,
suppressing the coarsening of FCG. The liquid phase in FCG region also pin the motion
of FCRB in the direction of temperature gradient. As the cooling proceeds, FCRB
moves upward immediately after the liquid ahead of CCG vanishes, finally yielding Fig. 1 Schematic illustration of
the CCG structure. CCG formation process3).
Prevention of the formation of CCG structure in the CC slabs requires
understanding the kinetics of discontinuous grain growth process. It is especially important to quantitatively ascertain the critical
condition of the occurrence of the discontinuous grain growth. The purpose of this study is to elucidate the critical condition of
formation of CCG structure. To this end, the discontinuous grain growth process from FCG structure was systematically
investigated by means of phase-field simulations, mold casting experiment and directional solidification experiment.
The phase-field model is a diffuse interface model in which the entity of boundary is described by the spatial
distribution of phase-field variable. It enables description of moving boundary without explicit tracking of the position of boundary.
In this study, we utilized the multi-phase-field model proposed by Steinbach and Pezzola5) and further developed for the case of
grain growth by Kim et al.6). In this model, the grain structure is characterized by the spatial distribution of a set of phase field
variables {𝜙𝜙𝑖𝑖 (𝒓𝒓, 𝑡𝑡)}. 𝜙𝜙𝑖𝑖 represents a probability of finding a grain having a crystallographic orientation i (i = 1, 2, …, N). Here,
𝒓𝒓 is the spatial coordinates, t is the time and N is the total number of grain orientations considered in the simulation. The time
evolution of 𝜙𝜙𝑖𝑖 is described by the following equation
𝑁𝑁
𝜕𝜕𝜙𝜙𝑖𝑖 2𝐿𝐿 𝛿𝛿𝛿𝛿 𝛿𝛿𝛿𝛿
= − � 𝑠𝑠𝑖𝑖 𝑠𝑠𝑗𝑗 � − �
𝜕𝜕𝜕𝜕 𝑛𝑛 𝛿𝛿𝜙𝜙𝑖𝑖 𝛿𝛿𝜙𝜙𝑗𝑗
𝑗𝑗≠𝑖𝑖
where 𝐿𝐿 is the phase-field mobility, 𝑠𝑠𝑖𝑖 is a step function with 𝑠𝑠𝑖𝑖 = 1 for 𝜙𝜙𝑖𝑖 > 0 and 𝑠𝑠𝑖𝑖 = 0 otherwise and 𝑛𝑛 =
∑𝑁𝑁
𝑖𝑖 𝑠𝑠𝑖𝑖 (𝒓𝒓, 𝑡𝑡). The functional derivative on the right-hand side is given as,
Fig. 2. Grain structures obtained by phase-field simulations for 𝑇𝑇̇ = 1 K/s, (a) 𝐺𝐺 = 1.4 × 104 and (b) 6.0 × 103 K/m.
(c) The relation between the grain growth mode and the cooling condition obtained from the phase-field simulations7).
References
1) Y. Maehara, K. Yasumoto, Y. Sugitani, K. Gunji: ISIJ Int. 25 (1985) 1045.
2) L. Schmidt, Å. Josefsson: Scand J. Metall. 3 (1974) 193.
3) S. Tsuchiya, M. Ohno, K. Matsuura, K. Isobe: Acta Mater. 59 (2011) 3334.
4) S. Tsuchiya, M. Ohno, K. Matsuura: Acta Mater. 60 (2012) 2927.
5) I. Steinbach, F. Pezzola: Physca D 134 (1999) 385.
6) S.G. Kim, D.I. Kim, W.T. Kim, Y.B Park: Phys. Rev. E 74 (2006) 061605.
7) M. Ohno, S. Tsuchiya, K. Matsuura: Acta Mater. 59 (2011) 5700.
8) M. Ohno, M. Maruyama, K. Matsuura: Acta Mater. 61 (2013) 7334.
9) M. Ohno, S. Tsuchiya, M. Matsuura: ISIJ Inter. 55 (2015) 2374.
10) H. Yasuda, T. Suga, K. Ichida, T. Narumi and K. Morishita: IOP Conf. Ser. Mater. Sci. Eng. 861 (2020) 012051.
Keywords: 316L stainless steel, grain boundary engineering, tensile mechanical properties, strain rate; dynamic strain aging,
dynamic recrystallization
316L stainless steel has been widely used due to a combination of good corrosion resistance and mechanical properties.
However, with the stainless steel service conditions becoming increasingly complex, the demands for its performance are
increasing in many industrial fields. Especially for stainless steel as a structural material is concerned, the need to further
improve the mechanical properties should be considered. As one of the important microstructural features of polycrystalline
materials, grain boundary (GB) has an important influence on the performance of stainless steels. Grain boundary engineering
(GBE) can enhance the grain-boundary-related properties of metal materials which have face centered cubic (FCC) structure
with low stacking fault energy (SFE) by changing the twin-induced grain boundary character distribution (GBCD). In this
work, GBE was applied to a 316L stainless steel. Subsequently, the effects of the GBCD formed during GBE on the tensile
mechanical properties at room and elevated temperatures (300 ℃ ~ 900 ℃) at different strain rates (4×10-2 s-1 ~ 4×10-5 s-1)
were investigated. The results are as follows:
The proportion of low-ΣCSL (Σ≤29, coincidence site lattice) GBs can be increased to more than 70% in the GBE
specimen. Meanwhile, the large-size highly-twinned grain-cluster microstructure was formed through the treatment of GBE.
All of the grains inside of this kind of cluster have Σ3n (n=1, 2, 3…) mutual misorientations 1). During the recrystallization
process, the Σ3n GBs are developed associating with the recrystallization nuclei growth. After full recrystallization, the
specimens were in the grain growth stage. The proportion of low-ΣCSL GBs decreased with grain growth due to being swept
away of the existing low-ΣCSL GBs.
The GBE samples with enhanced proportion of Σ3n GBs displayed increased uniform elongation during tensile test at
room temperature, and the differences in the uniform elongations between the GBE and Non-GBE specimens
(UEGBE–UENon-GBE) increase with the decreasing of strain rates 2). The Σ3 GBs are effective for improving the extent of
uniform plastic deformation of the GBE specimens under the condition of low strain rates. During the tensile deformation at
room temperature, the GBE specimen showed lower hardness and more homogeneous local strain distribution. The abundant
Σ3 GBs in the interior of the grain-clusters have weaker interaction with dislocation. Dislocation slip can easily transfer most
Σ3 twin boundaries, preventing the piling up of dislocations. The large and deep dimple structure also indicates the improved
plasticity after GBE.
The results of high temperature tensile tests show that the high proportion of Σ3n GBs of GBE samples not only
increases the elongations but also affects the dynamic strain aging (DSA) behavior in the temperature range of 300 -500C °
and dynamic recrystallization (DRX) at 900 C ° . Delayed occurrence of serrated flow as a result of DSA was observed in the
GBE samples as compared to the Non-GBE samples. The GBE samples have smaller serrations on the stress-strain curves,
especially at lower strain rate. The GBE specimen has a more homogeneous strain distribution leading to better plasticity at
lower strain rate. When the test temperature increased to 700 C ° , the total elongations of GBE samples show a greater extent
of increase. Higher dynamic recrystallization (DRX) fractions were observed in the GBE samples as compared to the
Non-GBE samples during tensile tests at 900 C ° . The grains of DRX have an obvious Goss texture inside the deformed grains,
which have <001> and <111> fiber textures. The dynamically recrystallized grains have higher values of Schmid factor,
which can facilitate further deformation. The features of the high temperature fracture surface of GBE samples also suggested
improved plasticity of GBE specimens.
References
1) Z. Zhuo, S. Xia, Q. Bai and B. Zhou: J. Mate. Sci. 53(2018) 2844–2858.
2) S. Xia, B. Zhou and W. Chen: Metall. Mater. Trans. A, 40A (2009)3016-3030.
1. Introduction
Controlling recrystallization is important for controlling the mechanical properties of iron and steel. Moreover,
three-dimensional (3D) analysis of recrystallized ferrite grains is needed to precisely understand the recrystallization behavior.
It is well known that the presence of solute carbon retards recovery in iron and steel1). This means that the recovery rate in
low-carbon steel is lower than that in pure iron. However, it is unclear if the 3D morphology of recrystallized ferrite grains
varies when the presence of solute carbon retards the recovery progress. Thus, we performed a 3D analysis of recrystallized
ferrite grains in pure iron and low-carbon steel. In addition, we attempted to propose a mechanism for 3D microstructural
evolution during the recrystallization process.
1.5 1
2D 3D
1
0.8
0.5
n = 2.33
n = 0.37
ln(-ln(1-X))
Breadth/Length
0 0.6
-0.5
-1 n = 1.86 0.4
-1.5
0.2
3)
-2 Pure iron
Low-carbon steel
4) 1D ◯Pure iron3)
● Low-carbon steel4)
-2.5 0
0 1 2 3 4 5 6 0 0.2 0.4 0.6 0.8 1
ln(t) Width/Length
Fig. 1 JMAK plot of recrystallization kinetics. Fig. 2 Growth direction of recrystallized ferrite grains.
4. Conclusions
We investigated the 3D morphology of the recrystallized ferrite grains in pure iron and low-carbon steel, and the
following results were obtained.
(1) In the case of pure iron, the Avrami exponent was 1.86 and 0.37 in the early and later annealing stages,
respectively. Whereas, Avrami exponent in low-carbon steel was 2.33.
(2) The 2D and/or 3D growth of the recrystallized ferrite grains was confirmed in pure iron. In contrast, the 2D
growth of the recrystallized ferrite grains was mainly confirmed in low-carbon steel.
(3) The interaction between the recovery and recrystallization was attributed to the inconsistency between
theoretical and experimental results for pure iron.
Acknowledgement
This study was supported in part by JSPS KAKENHI Grant Number 21K14423.
References
1) N. Sugiura, N. Yoshinaga, K. Kawasaki, Y. Yamaguchi, J. Takahashi and T. Yamada: Tetsu-to-Hagané 94 (2008) 179-187.
2) J. E. Burke and D. Turnbull: Progress in Metal Physics 3 (1952) 220-292.
3) T. Matsui, T. Ogawa and Y. Adachi: Results in Materials 1 (2019) 100002.
4) K. Horiuchi, T. Ogawa, Z. Wang and Y. Adachi: Materials 14 (2021) 4154.
Grain-oriented silicon steel is the critical material for transmission and distribution transformer cores. It is well
known for long and complex manufacturing procedure with sharp secondary recrystallization Goss ({110}<001>) texture as
the final purpose1). Primary recrystallization microstructure and texture play a key role for secondary recrystallization texture.
In view of the significant inheritance of microstructure and texture2), the normalization annealing of hot bands is an effective
way to regulate primary recrystallization microstructure and texture, especially for low-temperature slab heating process.
During normalization, the occurrence of recrystallization not only improves the microstructure uniformity but also changes
the texture of hot bands. In contrast with the effect of normalization on macro-texture of primary recrystallization3-5), the
micro-texture and collaborative control of microstructure and texture have received less attention. In this study, the effect of
normalized annealing was analyzed to develop a more reasonable path for optimal control of primary recrystallization process
in grain-oriented silicon steel.
Industrial Fe-3.25 wt. % Si grain-oriented silicon steel slabs were heated at 1100 ℃ and hot rolled into 2.7 mm
bands. Then HR bands were normalized at 900 ℃ (L-NA) and 1100 ℃ (H-NA) for 5min. Cold rolled sheets with a reduction
of 87% were annealed at 830 ℃ for 3 min. Micro-texture analysis was carried out using HKL Technology Channel 5
software by EBSD on a JEOLJSM-7001F scanning electron microscope .
As shown in Figure 1, the L-NA and H-NA bands present coarse pancake grains and equiaxed grains respectively,
in contrast to fine elongated grains in HR bands. Goss texture in subsurface of L-NA bands is more enhanced than H-NA
bands, which can be attributed to the different recrystallization mechanism during normalization. Continuous recrystallization
during L-NA produces pancake grains and strong Goss texture. Discontinuous recrystallization during H-NA results in
equiaxed grains and weak Goss texture with few accurate Goss grains. The center layer in HR bands is dominated with strong
{001}<110> texture, which is sharpened and dispersed by L-NA and H-NA respectively due to the difference in
recrystallization mechanism.
Figure 2 Primary recrystallization microstructure from HR (a), L-NA(b) and H-NA(c) bands
The area fraction of primary recrystallization Goss texture is closely related with HR, L-NA and H-NA bands, as
shown in Figure 3. L-NA and H-NA increases and decreases Goss texture in subsurface compared with HR bands,
respectively. Preferred nucleation at grain boundaries during H-NA tends to strengthen fiber (<111>//ND) with higher stored
strain energy and weaken Goss texture in either small or large deviation angle ranges. Continuous recrystallization during
L-NA is evidently favorable for the nucleation and growth of Goss grains. After primary recrystallization, both L-NA and
H-NA can increase Goss texture. The coarse equiaxed grains after H-NA are responsible for the abundant shear bands in cold
rolled sheets and high nucleation rate of Goss texture during primary recrystallization. The pancake grains are not so efficient
in the formation of shear bands during cold rolling as equiaxed grains, and the increase of Goss texture after primary
recrystallization by L-NA is mainly due to the enhanced Goss texture in L-NA bands. Furthermore, the deviation angle
distribution of primary recrystallization Goss texture was statistically analyzed in Figure 5. Goss grains have a higher fraction
in small deviation angle range (0~10°) for L-NA, while the lowest fraction in small deviation angle range occurs without
normalization annealing. This suggests that shear bands formed within the normalized pancake grains can provide more
accurate Goss nuclei, resulting from the smaller deviation angle of Goss texture as well as the straight grain boundaries in
L-NAbands.
Figure 3 Fraction of Goss grains in HR, L-NAand H-NAbands and primarily recrystallized sheets
Therefore, Goss texture at subsurface and microstructure through sheet thickness after primary recrystallization
have a strong correlation with the normalized annealing of hot rolled bands in grain-oriented silicon steel. The regulation of
hot rolled bands needs to be carried out based on the recrystallization mechanism in normalized annealing so as to optimize
the microstructure gradient through sheet thickness and enhanced Goss texture in subsurface with small deviation angle after
primary recrystallization.
Acknowledgement
This work was performed under the support of the National Key Research and Development Program of China
(No. 2016YFB0300305) and the National Natural Science Foundation of China (No. 51931002).
.
References
1) Hayakawa Y. Science and Technology ofAdvanced Materials, 2017, 18(01): 480-497.
2) Matsuo M, Sakai T, SugaY. Metallurgical TransactionsA, 1986, 17A: 1313-1322.
3) Fukagawa T,Yashiki H. Transactions of the Iron & Steel Institute of Japan, 2007, 40(4): 402-408.
4) Chang S K. Materials Science and EngineeringA, 2007, 452: 93-98.
5) Shin S M, Birosca S, Chang S K, B. C. De Cooman. Journal of Microscopy, 2010, 230(03): 414-423.
Keywords: Inclusion morphology, Rolling contact fatigue, Artificial defect, Micro CT imaging
Introduction
Higher fuel efficiency and less CO2 emission have become increasingly required for road vehicles in recent consideration
of conservation of the global environment. Accordingly, it is important for steel products to be energy-saving, compact and
light-weight. To realize these needs, the steel bars and wire rods must be improved to have better functional properties and
higher strength. Steel bars and wire rods undergo various manufacturing processes such as forging and cutting on the user
side. Therefore, it is also necessary to improve workability and machinability of steel.
Control technique of nonmetallic inclusion morphology in steels is one of the effective solutions to realize high
performance and high strength 1), 2). In this presentation, we mainly introduce research examples on the extension life of
bearing with a focus on the inclusion morphology control3).
Bearings are the automotive parts subjected to rolling and sliding contact under high loaded conditions. It is well
known that rolling contact fatigue (RCF) life is improved by reducing the size of nonmetallic inclusions in the steels 4).
However, there is little research regarding the morphology effect on the RCF life 1), 5). This work focused on the influence of
the inclusion shape on the RCF life. Crack initiation and propagation in the steels having two different shapes of inclusions
and artificial defects during RCF were studied using cross-section observations and SR micro-CT imaging. We found that
RCF life was well correlated with inclusion length 6), 7).
RCF test specimens were taken from the center of the forged bar in the
20µm
longitudinal direction.
Speroidized type
Table 1 Chemical composition of steels (mass%) of MnS
Mark C Si Mn P S Cr Mo Ca
Steel B
20µm
Steel A 0.20 0.19 0.78 0.015 0.018 1.18 0.20 -
Steel B 0.20 0.19 0.78 0.013 0.017 1.18 0.19 0.0026
Fig.1 Optical micrographs of typical shapes
of MnS inclusions
RCF specimens were gas-carburized at 1203 K for 32.4 ks in a sealed quench furnace to a surface carbon content of
0.8 %, and then directly quenched in oil from 1093 K. After quenching, all specimens were tempered at 453 K for 7.2 ks.
We measured the hardness profile in the plane which was parallel to the tangential line of the rolling track center. Hard layer
thickness is defined as the depth from the surface where hardness reaches a value of 550 HV. The surface hardness of the
specimens was around 750 HV and the hard layer thickness was 1.5 mm for each steels.
The gas-carburizing and tempering treatment develops the residual stress at the surface of the specimen. We measured
the residual stress distribution in the subsurface of specimens by X-ray diffraction method. Residual stresses of both steel
specimens of each steel which contained two different shapes of MnS 1.00E+05 1.00E+06 1.00E+07 1.00E+08
Life (cycle)
were tested. We defined the RCF life as the number of cycles to
failure. Figure 2 shows the Weibull distribution of the RCF lives. Fig.2 Weibull distribution of RCF lives
The Weibull distribution shows clearly that the RCF life of Steel B
containing spheroidized-type MnS was longer compared to that of
Steel A containing stringer-type MnS. L50 life of Steel B was about
three times longer than that of Steel A.
In order to identify the occurrence of RCF crack initiated from (a) Steel A(N=3.4×10 cycles) 6
(b) Steel B(N=9.2×10 cycles)
6
direction in both steels. The width of the originated MnS was 30μm
(c) Steel A 【Section A】 (d) Steel B 【Section B】
for each steel. It was well known that the local stress field was
Ball Rolling Stringer Spheroidized
developed at the inclusion during RCF test 8). Tensile stress in Direction type MnS type MnS
rolling direction was occurred near the surface of specimen, and shear Perpendicular
crack Shear crack Perpendicular
stress in rolling and depth direction was also occurred in subsurface of crack
Ball Rolling
specimen. It seems that this perpendicular crack was initiated due to Shear crack
Direction 20μm
50μm
the tensile stress in rolling direction or shear stress in depth direction.
The shape of the originated MnS must affect the initiation and Fig.3 SEM images of occurrences of
propagation behavior of these “perpendicular cracks”, since the value RCF cracks
of local stress would be changed due to the shape of inclusion. To investigate the initiation and propagation behavior of
these perpendicular cracks, we conducted the sectional observation. The results are shown in Fig. 3-(c), (d). In the section
of Steel A, it was found that the perpendicular crack was propagated along the sidewall of the stringer-shaped MnS 150μm in
length and reached to the end of the stringer-shaped MnS. While in the section of Steel B, the propagation behavior of the
perpendicular crack was different from that of Steel A. It was found that the perpendicular crack was initiated from
spheroidized MnS and propagated to the depth direction. The length of the perpendicular crack initiated from spheroidized
MnS reached 40μm in depth, but its depth was shorter than that of the perpendicular crack initiated from stringer-shaped
MnS.
Present results provided evidence suggesting a strong possibility that the propagation behavior of the perpendicular crack
is well correlated with the length of the originated MnS because the value of tensile stress in rolling direction and shear stress
in depth direction depends on the length of MnS. In other words, this result may support that the perpendicular crack grew
acceleratory in mode I and mode III as the MnS length increased.
Next, attention was focused on the shear crack in the subsurface. It was found that several shear cracks appeared
parallel to the shear stress in rolling direction and were located at equal spacing in the section of Steel A. The shear crack
located at the depth of the maximum shear stress (=100μm) was the longest. However, in the section of Steel B, it should be
noted that shear cracks in the subsurface were initiated from the tip of the perpendicular crack, not from spheroidized MnS,
and the depth of the occurrence of the subsurface crack was 40μm from the surface, which was not reached at the maximum
Evaluation of the inclusion length effect on RCF crack behavior by using artificial defects
In order to clarify the RCF crack behavior, we evaluated RCF property of the material with artificial defects simulating
stringer type inclusions using synchrotron radiation micro-computed tomography (SR micro CT) imaging. We also carried
out RCF tests using a thrust-type machine under the constant load and well-controlled lubricant conditions. The specimens
were made of induction-hardened 0.55% carbon steel and the induction-hardened region had a Vicker’s hardness of 720 HV
and residual stress of -205 MPa. We introduced artificial defects of 15 m in diameter and between 30m and 200 m in
depth on the ball rolling track of the specimen by electro-discharge machining. The defect simulates a stringer type inclusion
(e.g. MnS) in terms of size and shape. The orientation of the defect, which is perpendicular to the rolling track, corresponds to
that of inclusions in thrust bearings.
Figure 4 shows the SEM images of artificial Ball rolling direction
Defect:φ15μm×L200μm
defect of 200 m in depth on rolling contact surface
after the RCF test. We found several small cracks
appeared due to the defect, almost perpendicular to the
rolling direction and the surface. We found that the Artificial
defect Vertical
20μm
From the specimens used for the SEM
Fig. 4 SEM photos of artificial defect on rolling contact
observation, we took a small sample which included
surface after the RCF test. (a) N=1x104 cycles. (b) N=1x105
the artificial defect and inspected the sample by SR
cycles. (c) N=1x106 cycles.
micro CT imaging at SPring-8. The inspection region
in the sample has a square cross section, 0.5 mm x 0.5 mm. The size is (a) N =1 x 104 cycles
determined from the penetrative power of X-rays in Fe material. We
conducted CT imaging at the BL19B2 beam line of SPring-8, the most PV V
P PV depth
powerful synchrotron radiation facility in Japan. For a 3-D PV direction
SH VP H
reconstruction, a set of 900 radiographs of a sample were recorded at
Artificial
rotations of over 180 °, where each rotation angle was 0.5 °. Slice defect
Ball rolling Ball rolling
images were reconstructed from the series of the projection images by a direction direction
Conclusion
In order to clarify the RCF crack behavior and to discuss the mechanism of RCF crack propagation, we carried out RCF
test with an artificial defect, SR micro CT imaging. We found the following results:
(1) Two different types of fatigue cracks - perpendicular crack and shear crack - were induced in RCF.
(2) Both perpendicular and shear cracks initiated before N=1x104 cycles. Perpendicular cracks propagated in the depth
direction and perpendicular to the rolling direction of the artificial defect from N=1x104 to 1x106 cycles. Shear cracks
propagated in a horizontal direction simultaneously.
(3) The perpendicular crack accelerates the propagation of shear cracks.
(4) Interaction between perpendicular and shear cracks would appear to dominate RCF crack propagation originating from
defects such as inclusions.
References
1) K. Hashimoto, K. Hiraoka, K. Kida and E. C. Santos: Materials Science and Technology, 28 (2012), pp.39-43.
2) N. Matsui: ISIJ Int. 46-11 (2016), pp.1720-1727.
3) Y. Neishi, T. Makino, H. Matsumoto, M. Higashida and H. Ambai: Metallurgical and Materials Transactions A 44-5(2013), pp.
2131-2140.
4) M. Nagao, K. Hiraoka and Y. Unigame: Sanyo Tech. Report 12-1 (2005), pp. 38-45.
5) K. Hashimoto, K. Hiraoka, and K. Kida: TMS2011 140th Annual Meeting & Exhibition (2011, San Diego, USA).
6) T. Makino, Y. Neishi, D. Shiozawa, Y. Fukuda, K. Kajiwara and Y. Nakai: Int. Journal of Fatigue 68 (2014), pp.168-177.
Abstract: Design of advanced alloys with simple compositions is a new strategy to improve the recyclability of the
alloys. In this study, a 2 wt% Nb bearing low carbon steel (2Nb steel) without addition of Ni, Cr and W was designed and
processed by hot forging, hot rolling and cold rolling followed by annealing. A fine grain size microstructure with NbC
precipitates formed. Microstructure and tensile properties were characterized at ambient and elevated temperatures. The good
strength was related to the pinning effect of large NbC precipitates present at the grain boundaries and small precipitates
formed in the grain interior that hinder the movement of grain boundaries and dislocations, respectively.
1. Introduction: In recent years, with the extensive development and application of a series of heat-resistant steels,
the design of heat-resistant alloys has gradually attracted the attention of researchers. Among some reported alloy designs,
such as Fe-20Ni-14Cr-2.5Al-(1-3) Nb (wt%) [1], Fe-18Cr-31Ni-4Nb (wt%) [2], Fe-20Cr-3Ni-3Si-2Nb-1Mn (wt%) [3],
Fe-21Cr-10Ni-2Nb-1Mn-(1-2) W (wt%) [4] add a large amount of alloying elements, including Ni, Cr, W, Nb, etc. However,
the more alloying elements are added, the more complex the material system is, and the more difficult it is to control the
microstructure and properties. For the sustainable development of global resources and environment, reducing alloy or using
simpler alloy to save material cost, increasing recyclability and improving sustainability without reducing the material
properties need considerations in material design [5].
Generally, niobium (Nb) is used as a microalloying element to enhance the mechanical properties of steel, but it is
added in large quantities in the composition design of heat-resistant steel. Precipitates such as Laves phase [2], G-phase
precipitates [3] and MC-carbides [6], are introduced in the grain interior or on the grain boundaries to improve the stability of
the microstructure at high temperatures. But Laves phase and G-phase generally lead to brittle failure of the material and are
detrimental to room temperature properties. For example, Padilha et al. [7] reportd that Fe2Nb phase is hard and fragile,
forming between austenite dendrites and result in loss of toughness and ductility in their investigated Fe-15Cr-15Ni-2Nb
(wt%) austenitic stainless steel. Thus, it is necessary to find better precipitates which can provide good properties at both room
temperature and high temperatures without sacrificing either. In the present study, a much plain material: Fe-0.2C-2Nb (wt%)
low carbon alloy steel was designed. The microstructure evolution of the alloy was systematically characterized by OM, SEM,
EBSD and TEM imaging techniques. The mechanical properties of the alloy at high temperature and room temperature were
tested to reveal the deformation mechanism and strengthening mechanism of the alloy.
2. Experimental: The designed chemical composition of the steel is 2 wt% Nb, 0.2 wt% C with balanced Fe,
shorted as 2Nb steel in the following part. After casting, the sample was hot forged, 80% hot rolled or 80% cold rolled. Heat
treatment was also performed at 200℃ to 400℃ to tune the microstructures and mechanical properties. Tensile tests were
performed at room temperature using dog-bone-shaped specimens, with gage dimension of 13 2.5 2 mm3, at an initial
strain rate of 6.4 × 10−4 s−1, and at high temperatures in the range of 600 ℃ - 1000 ℃ with cylindrical specimens (60 Ø 10
mm3, strain rate of 10−3 s−1). At least 3 specimens were used for each tensile test.
Microstructures of 2Nb steel were characterized by optical microscopy (OM), scanning electron microscopy
(SEM), electron backscatter diffraction (EBSD) and transmission electron microscopy (TEM). The specimen surface was
electro-polished in a solution of 5 vol.% HClO4 (perchloric acid) + 95 vol.% C2H5OH (ethanol) at -20 ℃ with a voltage of
20 V. SEM and EBSD measurements were carried out using JSM 7800 with Aztec program. For TEM observations, thin foils
were mechanically polished and then thinned by twin-jet in a solution of 5 vol.% HClO4 + 95 vol.% C2H5OH at -20 ℃ with
a voltage of 20 V. Conventional bright field (BF) images were taken by JEM 2100 operated at 200 kV.
Fig. 1 shows the phase diagram of Fe-Nb containing 0.2 wt% C using Thermo-Calc software, austenite transforms
into ferrite at around 920 ℃. NbC phase and Fe2Nb Laves phase may exist at low temperatures. Our microstructure and
XRD results found that the phases are mainly composed of ferrite matrix and NbC phase after casting and hot forging, Fe2Nb
phase was not found or with only a very small amount that can't be resolved by XRD. It can be seen from Fig.2 that much
equiaxed grains can be seen after hot forge (about 8.0 ± 1.0 μm) which is much finer than the as-cast sample.
Fig. 2 OP microstructures of 2Nb steel: (a) (b) as-cast; (c) (d) as-forged
From Fig. 2(d) and Fig. 3, it is easily seen that larger particles (~ 300 nm) are pinned at the grain boundaries and triple
junctions while the smaller precipitates are widely dispersed in the grain interior and at the dislocations. According to the
statistics, the average size of the small precipitates in the grain interiors is ~ 34 nm. The NbC precipitates underwent a small
amount of coarsening during the tensile test at 600 ℃. The detailed properties can be found in Table 1.
Fig. 4 TEM image of 80 % hot rolled sample RD-ND surface (a) and 200 ℃, 15 min heat treated sample (b)
Fig. 4(a) is the TEM image of 80 % hot rolled sample observed on RD-ND surface. Deformed ferrite matrix has a
large amount of dislocations. Large NbC particles can be found along the grain boundaries or in the grain interiors, as
indicated by yellow arrows. Fig. 4(b) is the 200 ℃, 15 min heat treated sample. Some Fe2Nb precipitates can be found after
annealing at low temperatures. The low magnification images can be found by SEM images as shown in Fig. 5, where Fig.
5(a) is the as-hot rolled sample, and others are 300 ℃ annealed samples. Fig. 6 is the TEM image of 80 % cold rolled sample.
More elongated ferrite matrix can be seen as comparing with hot rolled condition.
Fig. 6 TEM image of 80 % cold rolled sample (RD-ND surface): (a)TEM image and corresponding microstructural sketch lines (b)
Acknowledgement
The authors wish to thank the financial support of the National State Key Research and Development Program of China
(2016YFB0700403), the Natural Science Foundation of China (NSFC, Grants No. 51671039), and the Fundamental Research
Funds for the Central Universities (Project No. 2020CDJDPT001).
References
1) Y. Yamamoto, M. P. Brady, Z. P. Lu, P. J. Maziasz, C. T. Liu, B. A. Pint, K. L. More, H. M. Meyer, E. A. Payzant, Science, 316
(2007) 433-436.
2) S.W. Chen, C. Zhang, Z.X. Xia, H. Ishikawa, Z.G. Yang, Materials Science and Engineering: A, 616 (2014) 183-188.
3) M. Yang, J. Zhu, T. Yang, J. Luan, Z. Jiao, X. Fan, B. Kuhn, X. Xiong, C. Wang, C.T. Liu, X. Liu, Materials Science and
Engineering: A, 745 (2019) 390-399.
4) Yinhui Zhang, Mei Li, Larry A. Godlewski, Jacob W. Zindel, Qiang Feng, Materials Characterization, 139 (2018) 19-29.
5) Y.H. Gao, S.Z. Liu, X.B. Hu, Q.Q. Ren, Y. Li, Vinayak P. Dravid, C.X. Wang, Materials Science and Engineering: A, 759 (2019)
298-302.
6) Seungmun Jung, Yong Hee Jo, Changwoo Jeon, Won-Mi Choi, Byeong-Joo Lee, Yong-Jun Oh, Gi-Yong Kim, Seongsik Jang,
Sunghak. Lee, Materials Science and Engineering: A, 682 (2017) 147-155.
7) A. F. Padilha, I. F. Machado, R. L. Plaut, Journal of Materials Processing Tech, 170 (2005) 89-96.
Aim and scope: Solubility products (SP) of TiC and TiN in austenite are briefly reviewed.
The SPs of carbides and nitrides are commonly described by logarithms in a form of
log([%𝑇𝑖][%𝑋]) = 𝐴 − 𝐵/𝑇 , where X is C or N, A and B are constant parameters,
𝑇 represents temperature. The austenite grain growth is essentially related to the dissolution of
precipitates, which can be determined by SP. The present work aimed to quantify the
connection between the SP and austenite grain growth. The model was proposed by analyzing
the austenite grain growth with considering the pinning effect of precipitates. The austenite
grain size increased from 8.7 μm at 950 ℃ to 183.5 μm at 1300 ℃. With the statistical data,
SP equations of a Ti containing steel were thereafter obtained based on the statistical data and
compared with those in literature.
Results
Measured austenite grain size
It can be seen from Fig. 1a that the average austenite grain size increases from 8.7 μm at
950 ℃ to 183.5 μm at 1300 ℃. The coarsening rates of austenite grains during the heating
process were given the Fig. 1b. There was a decrease in coarsening rate when the temperature
increased from 1150 ℃ to 1200 ℃. The SP of TiN is so small below 1250 ℃ that its effect on
the austenite growth can be neglected [1]. Thus, it is reasonable to speculate that the
coarsening before 1150 ℃ mainly depends on the dissolution of TiC particles.
In the present work, a very slow heating rate of 1 ℃/s was used, thus one can assume
that the equilibrium of dispersing precipitates was kept at each step. For a selected heating
temperature, the maximum distance between two particles, 𝐿max , can be expressed as
follows:
𝑡 1/2
𝐿max = (∫𝑡 𝐷𝑇𝑖 𝑑𝑡) (5)
0
where 𝐷𝑇𝑖 is the diffusivity of Ti, given as 𝐷𝑇𝑖 (m2 · s−1 ) = 2.8 × 10−4 exp (−286000/
𝑅𝑇) [1]. According to the assumption of Nishizawa [3], the inter-particle distance can be
expressed as a function of particle radius r and volume fraction 𝑓𝑉 , that is:
Assuming 𝐿max = 𝐿 and combining Eqs. (4)-(6), the following equation can be obtained:
𝑡 1/2
1/3 8(∫𝑡 𝐷𝑇𝑖𝑑𝑡 )
𝑓𝑉 = 0
(7)
3𝑅
In addition, the total volume fraction of particles is essential related to the SPs of TiC
and TiN. Thus, the constant parameters A and B in the SP equations can be obtained by fitting
the measured austenite grain size R. The parameter t in Eq. (7) is heating time, which is
related to the heating rate and target austenitization temperature. The lower limit 𝑡0 of the
𝑡
integral formula ∫𝑡 𝐷𝑇𝑖 𝑑𝑡 can be termed as the beginning of dissolving precipitates in
0
Fig. 2 The calculated volume fractions of TiC (a) and TiN (b) as a fuction of heating
temperature.
Fig. 3 shows the comparison of the SP models in the present work with others from
literatures. It can be seen that the cited TiC and TiN SP models vary in a large range. The
present TiC model locates in a relatively dense region, which is neighbored to 2.75-7000/T
cited in most works. The TiN SP equation in the present work is close to 4.35-14890/T. It is
reported that alloying elements can affect the solubility of the second phase precipitates [3],
leading to variation of parameters A and B in the SP equation. Thus, the present SP model
could be used within similar composition. For the tested steel, the SP model was fitted by
austenite grains. This method is simplified compared to the way measuring precipitation
Fig. 3 Comparison of (a) TiC and (b) TiN SP model in the the present work with other works.
References
[1] M. Maalekian, R. Radis, M. Militzer, A. Moreau, W.J. Poole. In situ measurement and
modelling of austenite grain growth in a Ti/Nb microalloyed steel. Acta Materialia, 60 (2012)
1015-1026.
[2] K. Banerjee, M. Militzer, M. Perez, X. Wang. Nonisothermal Austenite Grain Growth
Kinetics in a Microalloyed X80 Linepipe Steel. Metallurgical And Materials Transactions A,
41 (2010) 3161-3172.
[3] T. Nishizawa, I. Ohnuma, and K. Ishida. Examination of the Zener Relationship between
Grain Size and Particle Dispersion. Materials Transactions, JIM, 38 (1997) 950-956.
Nobuhiro Tsuji1,2
1
Department of Materials Science and Engineering, Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto 606-8501, Japan
2
Elements Strategy Initiative for Structural Materials (ESISM), Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto 606-8501,
Japan
*email: tsuji.nobuhiro.5a@kyoto-u.ac.jp
Keywords: ultrafine grained steels; strength and ductility; recrystallization; deformation modes; strain hardening.
Fig.3 TEM images showing representative deformation microstructures in the ultrafine grained specimens at an early stage of
tensile deformation. (a) A grain with a size > 1 µm. (b) Grain size ~ 1 µm. (c) Grain size < 1 µm [10].
Their excellent mechanical properties are often given by a nucleation of new deformation modes, such as unusual slip
systems in HCP metals, nano deformation twins and deformation induced martensite all which were not expected to operate
in these metals and microstructures [5,7,8-10]. In case of the ultrafine grained Fe-31Mn-3Al-3Si austenitic steel, very thin
deformation twins with thickness of several nm ~ 30 nm nucleated from grain boundaries (Fig.3) [10], even though it has
been known that deformation twinning in FCC metals and alloys is suppressed by grain refinement. The activation of such
unexpected deformation modes leads to regeneration of strain-hardening ability. After these experimental results, a strategy to
overcome the strength-ductility trade-off is discussed on the basis of a new concept of plaston considering atomistic process in
nucleation of new deformation modes [11-13]. It is expected that the plaston concept could be a key to designing advanced
high-strength metallic materials managing both ultra-high strength and enough ductility/toughness.
Acknowledgement
This research work was financially supported by the Elements Strategy Initiative for Structural Materials (ESISM;
JPMXP0112101000), JST CREST (JPMJCR1994), and KAKENHI (Grant-in-Aid for Scientific Research from the Japan Society
for the Promotion of Science; 15H05767 and 20H00306), all from the Ministry of Education, Culture, Sports, Science and
Technology (MEXT), Japan. All the supports are gratefully appreciated.
References
[1] K.Saito: Materia Japan 53 (2014) 584-588.
[2] http://www.autosteel.com
[3] N.Tsuji, Y.Ito, Y.Saito, Y.Minamino: Scripta Materialia 47 (2002) 893-899.
[4] R.Saha, R.Ueji, N.Tsuji: Scripta Materialia 68 (2013) 813-816.
[5] Y.Z.Tian, L.J.Zhao, S.Chen, A.Shibata, Z.F.Zhang, N.Tsuji: Scientific Reports 5 (2015) 16707.
Keywords: steel, interphase precipitation, scanning transmission electron microscopy, tensile properties, strengthening mechanisms
Currently, there is a significant interest in both academia and industry related to the development of ferritic
microalloyed steels strengthened by nanometer-sized Interphase Precipitation (IP). The acceleration of these efforts started in
early 2000s with a report by JFE of new Ti-Mo microalloyed steels reaching 734 MPa yield strength (YS) 1). One of the major
contributors to the yield strength was deemed ~3 nm size IPs. Interphase precipitation can occur during the γ→α
transformation in steels alloyed with carbide-forming elements: Ti, Mo, Nb, V, Cr, etc.2,3) Experimental and theoretical data
suggest that IP could provide a significant strengthening effect (Table 1) presumably via operation of a dislocation looping
mechanism. At the same time, IP-strengthened steels display other attractive characteristics, such as cost-effectiveness,
formability and weldability.
Carbide-forming
Yield strength, MPa Reference
elements, wt.%
4)
0.06C-0.1Ti-0.5Cr 220-320
5)
0.06C-0.1Ti-0.2Mo 230-400
1)
0.04C-0.09Ti-0.2Mo 300
0.1C-0.2Ti 240
0.1C-0.2Ti-0.5Mo 260 6)
0.1C-0.2V 130
0.1C-0.2V-0.5Mo 240
7)
0.1C-0.04Nb-0.11Ti 112
8)
0.44C-0.3V 180
9)
0.023C-0.34Nb 210
0.08C-0.035Nb-0.085Ti- 10)
320
0.11Mo
In addition to steel composition, the thermo-mechanical processing (TMP) of these steels includes several
important stages (reheating temperature, deformation amount and temperature, coiling temperature and time), which could
directly affect the size, morphology and volume fraction of IP. With respect to the effect of deformation, while the majority of
studies on IP were conducted on fully annealed materials, a limited number of results indicated a decrease in size and
intersheet spacing accompanied by an increase in volume fraction of IP with an increase in the amount of strain in the
non-recrystallised region 11-13). Decreasing coiling temperature also generally leads to the IP size refinement and an increase in
its amount. However, these effects are not straightforward as modification of TMP parameters leading to the consumption of
Steel containing ~(0.08±0.003)C, 1.5Mn, 0.3Si, 0.2Ni, 0.03Al, 0.003S, 0.015P, (0.0131±0.0002)N, and 0.73 (Cr +
V + Nb) (wt. %) was used for this study. The TMP was carried out using a Gleeble 3500 simulator with various processing
parameters.14,15). Tensile testing was performed using 3 mm wide, 1 mm thick and 7 mm gauge length flat specimens on a
Kammrath and Weiss GmbH mini-tensile stage at the constant crosshead speed of 7 μms-1. The microstructures were
characterised using optical, scanning and transmission electron microscopies. Selected samples were also examined using a
high resolution scanning transmission electron microscope ARM 200F and local electrode atom probe.
Representative images of microstructure (Fig. 1a) and of interphase precipitation (Fig. 1b) confirmed
predominantly polygonal ferrite microstructure with a very small amount of pearlite and that the majority of ferrite grains
contained the IP of (V, Cr)C. However, the TMP parameters affected the characteristics of IP and consequently, the tensile
properties. For example, for the same amount of deformation, but varying the deformation finish temperature from 900 to
850°C, the ferrite grain size has reduced by ~1.5 µm, the particles became finer but their density lower. The strength values
were not affected by this variation in the deformation temperature (~870 MPa YS and ~1030 MPa tensile strength), but
elongation was slightly higher after higher finish temperature (23% vs 21%). The intersheet spacing was not affected and was
~ 22 nm for both schedules. On the other hand, an increase in the amount of deformation imposed in the recrystallization
region by 0.3 resulted in a small amount of ferrite grain refinement, but enormous increases in IP density (intersheet spacing
reduction to 12 nm after coiling at 600°C) and significant improvement in the mechanical properties (increase in YS from
717±8 to 862±62 MPa was accompanied by small reduction in elongation from 21 to 19%). A slightly longer coiling time of
30 min, compared to 15 min at higher coiling temperature of 650°C, also delivered a high volume fraction of IPs with 12 nm
intersheet spacing and even further increase in YS to 960 MPa.
a b
Figure 1 (a) Representative optical micrograph of polygonal ferrite with pearlite (dark). (b) Bright field image of
interphase precipitation. Adopted from Ref. 15)
To elucidate the mechanism of interaction between IP and dislocations, high resolution investigation of precipitates
before and after tensile tests were conducted (Fig. 2). A clear formation of steps at the particle/matrix interface was visible after
deformation (Fig. 2b), which led to the conclusion of dislocations cutting the ~1-3 nm thick particles, instead of looping
around. Assuming operation of coherency strengthening and modulus mismatch strengthening mechanisms has taken place in
this steel, an estimated ~395 MPa contribution to the yield strength has arisen from these nano-sized particles. However, it
should be noted that when applying the Pythagorean flow stress addition law, the actual strength contribution will be less.
Atom probe studies have shown the presence, in addition to IP rows, of clusters of V between the particles in the rows and
between the rows in ferrite matrix. TEM also confirmed the presence of random precipitates in the ferrite matrix. In addition
to these very fine precipitates and clusters, a relatively low density of carbides and carbo-nitrides > 20 nm size was also
detected. These were assumed to form in austenite during TMP. Thus, the contributions to YS in these steels arises from
Figure 2 STEM-HAADF images of VC (a) after TMP and (b) after tensile testing. Adopted from Ref. 16)
However, it needs to be considered that any change of one of the processing parameters may increase one of the
desirable microstructurural characteristics but have a detrimental effect on others. For example, increased deformation in the
non-recrystallisation region will lead to more nucleation sites for ferrite formation and its grain refinement, but it will also
trigger more abundant strain-induced precipitation. The latter will consume the microalloyed elements leaving a lesser amount
available for IP and random precipitation in ferrite, thus possibly reducing the volume fraction and density of IP, and their
associated strengthening effect. This may trigger a requirement to for additional modification of either the TMP and/or the
steel composition.
Acknowledgement
This work was funded by the Australian Research Council (ARC) Research Hub for Australian Steel Manufacturing under the
Industrial Transformation Research Hubs scheme (Project ID: IH130100017).
References
1) Y. Funakawa, T. Shiozaki, K. Tomita, T, Yamamoto and E. Maeda: ISIJ Int. 44 (2004) 1945-1951.
2) T.N. Baker: Ironmaking & Steelmaking 43, (2016) 264-307.
3) T.N. Baker: Mater. Sci. Techn. 25 (2009) 1083-1107.
4) C.-N Li, X.-L. Li, Y. Guo, R.D.K. Misra, J. Kang and G.-D. Wang: Mater. Sci. Eng. A 673 (2016) 213-221.
5) H.W. Yen, P.Y. Chen, C. Y. Huang, and J.R. Yang: Acta Mater. 59 (2011) 6264-6274.
6) P. Gong, X.G. Liu, A. Rijkenberg and W.M. Rainforth: Acta Mater. 161 (2018) 374-387.
7) S.-P. Tsai, T.-C. Su, J-R. Yang, C-Y. Chen, Y-T. Wang and C-Y. Huang: Mater. Des. 119 (2017) 319-325.
8) G. Miyamoto, R. Hori, B. Poorganji and T. Furuhara: ISIJ Int. 51 (2011) 1733-1739.
9) H.J. Kestenbach: Mater. Sci. Techn. 13 (1997) 731-739.
10) A.T. Davenport and R.W.K. Honeycombe: Proc. Royal Soc. A Math. Phys.Eng. Sci. 322 (1971) 191-205.
11) C.Y. Chen, C.C. Chen and J.R. Yang: Mater. Charact. 88 (2014) 69-79.
12) I. Bikmukhametov, H. Beladi, J. Wang, P.D. Hodgson and I.Timokhina: Acta Mater. 170 (2019) 75-86.
13) C-Y. Chen, S.-F., Chen, C.-C. Chen and J-R. Yang: Mater. Sci. Eng. A634 (2015) 123-133.
14) N. Singh, G. Casillas, D. Wexler, C. Killmore and E. Pereloma: Acta Mater. 201 (2020) 386-402.
15) A. Kostryzhev, C. Killmore and E. Pereloma: Metals 11 (2021) 107 (1-14).
16) E. Pereloma, D. Cortie, N. Singh, G. Casillas and F. Niessen: Mater. Res. Lett. 8, 341-347 (2020).
*email: shigesato.59d.genichi@jp.nipponsteel.com
1. Introduction
Steel plates are used for structural and construction applications, offshore structures, wind power plants, pressure vessels,
etc., for which safety and reliability are prime concern. Therefore, high toughness as well as high strength is required for the
material. In particular the properties of weld joints are important because the performance of welded structures is substantially
determined by them. Stress tends to concentrate at welded joints, and in addition, they are subject to thermal structural change
and residual stress due to welding. It is, therefore, not easy to obtain welded joints with high strength, toughness, fatigue prop-
erties, etc. The problem worsens when the heat input is increased to raise the efficiency of welding work. To solve these
problems, the development of steel materials and welding technology has to be done in a coordinated manner.
This paper focuses on the material side of development and outlines the investigations concerning the microstructure control
to obtain excellent toughness of the heat affected zones (HAZ) and weld metal. The toughness is governed by the properties of
the iron matrix, such as the grain size and the strength, and those of embrittlement phase acting as an initiation site of a crack. In
response to that, the grain refinement of ferrite grains and the reduction of embrittlement phase have been investigated to
improve the toughness of HAZ and weld metal. In the following section recent studies about the utilization and the mechanism
of intragranular ferrite transformation as a means of the grain refinement are shown. After that the carbon enrichment behavior
to austenite during bainite transformation will be presented, which is related to the formation mechanism of martensite-austenite
constituent (MA), which acts as an embrittlement phase.
Fig. 3 The microstructure (15° boundaries measured Fig. 4 Relationship between the number of Ti containing
by EBSD) of weld metals. inclusions and the effective grain size, deff.
Fig. 5 (a) Dilatation during isothermal holding at 550°C, SEM images after (b) 30 s, (c) 200 s, (d) 1000 s and (e) 10000 s.
surrounding austenite, which is the upper limit of the stored energy. As the transformation temperature
lowers, ΔGstrain is considered to approach to 400J/mol.
On the other hand, it is possible to interpret the results by the WBs theory. The WBs line in fig.4 was calculated using the
values proposed by Hillert as the barrier energy to the growth of acicular ferrite. The WBs line can be fitted to the current results
by lowering the values by about 100 J/mol (Fig.7). Thus, it is difficult to judge from the current results which model is
appropriate. The physical meaning of the barrier energy should be examined quantitatively.
4. Summary
Recent investigations concerning the mechanism of IGF generation and the carbon partitioning behavior during during
bainite transformation, which were connected with the microstructure control to obtain excellent toughness of the heat affected
zones (HAZ) and weld metal, were reviewed.
Ti oxide that have NaCl type crystal structure, TiO, can promote IGF nucleation due to the combined effects of good lattice
matching and the formation of Mn depleted zone.
The carbon concentration of untransformed austenite during the transformation stasis in Fe–0.1%C–0.5%Si–2.0%Mn
alloy were measured using EPMA. The results can be explained by either theory, T0’and WBs supposing the transformation
barrier energy are appropriately adjusted.
References
1) B. L. Bramfitt: Metall. Trans., 1 (1970), 1987.
2) A. R. Mills, G. Thewlis and J. A. Whiteman: Mater. Sci. Technol., 3 (1987), 1051.
3) H. Nako, Y. Okazaki, H. Hatano, K. Yamashita and H. Takauchi: Kobe Steel Eng. Rep., 67 (2018), 83.
4) T. Yamada, H. Terasaki and Y. Komizo: Tetsu-to-Hagané, 96 (2010), 608 (in Japanese).
5) G. Shigesato, M. Sugiyama, S. Aihara, R. Uemori and Y. Tomita: Tetsu-to-Hagané, 87 (2001), 93 (in Japanese).
6) K. Seo, Y. Kim, G. M. Evans, H. J. Kim and C. Lee: Weld. World, 59 (2015), 373.
7) N. Fujiyama and G. Shigesato: ISIJ International, 61 (2021), 1614.
8) J. Takamura and S. Mizoguchi: Proc. 6th Int. Iron and Steel Cong., 1, ISIJ, Tokyo, (1990), 591.
9) H. K. D. H. Bhadeshia and D. V. Edmonds: Acta Metall., 28 (1980), 1265.
10) H. K. D. H. Bhadeshia: Acta Metall., 29 (1980), 1117.
11) M. Hillert: The Growth of Ferrite, Bainite and Martensite, Swedish Institute for Metals Research, Stockholm, (1960).
12) M. Hillert, L. Höglund and J. Ågren: Metall. Mater. Trans. A, 35 (2004), 3693.
13) H. D. Wu, G. Miyamoto, Z. G. Yang, C. Zhang, H. Chen and T. Furuhara: Acta Mater., 133 (2017), 1.
14) S. Tanaka, H. Shirahata, G. Shigesato and M. Takahashi: Tetsu-to-Hagané 107 (2021), 835 (in Japanese)
Keywords: ultra-low carbon steel, low temperature, tensile property, aging, deformation twin
1. Introduction
Carbon is the most important alloying element for strengthening conventional steels and strongly affects their mechanical
properties, even in very small amounts such as tens of ppm. The several strengthening mechanism at room temperature by
aging of ultra-low carbon steel has been proposed such as precipitation strengthening by carbides [1], increase in the
Hall-Petch coefficient by segregation of carbon atoms to the grain boundary [2]. Furthermore, using a small ball rebound
hardness test, Koga et al. [3] demonstrated that a solid solution of ultra-low carbon (approximately 100 ppm) in the matrix
significantly enhanced the high-temperature hardness of steels with various carbon concentrations. The effect of ultra-low
carbon on the mechanical properties of steels above room temperature has been widely investigated; however, it is still unclear
whether ultra-low carbon influences the mechanical properties at low temperatures.
In this study, the microstructural changes and low-temperature tensile properties of ultra-low carbon steel with aging
treatment were examined and the relationship between the low-temperature tensile properties and the ultra-low carbon state
were discussed.
2. Experimental procedure
Ultra-low carbon steel with a carbon concentration of 75 mass ppm was used. The steel was cold rolled to a 90 % reduction
in thickness, annealed at 973 K for 1.8 ks, and then water-cooled. The annealed specimens exhibited equiaxial recrystallized
grains with an average grain size of approximately 50 μm. The annealed specimens were aged at 443 K for 0.06–600 ks and
subsequently water-cooled. The aged specimens were maintained at 77 K in liquid nitrogen to prevent further aging at room
temperature. Tensile test specimens were cut from plates where the tensile direction was parallel to the rolling direction (RD).
Tensile tests were conducted at an approximate initial strain rate of 5.2 × 10-4 s-1 at 293 K in air and at 77 K in liquid nitrogen.
The tensile tests were initiated after the specimens were maintained at the test temperature for 1.8 ks. The microstructure on
the surface of the normal direction (ND) specimen was observed using field-emission scanning electron microscopy
(FE-SEM). Electron back-scattered diffraction (EBSD) with FE-SEM was used to analyze the crystal orientation in the ferrite
matrix and twins. A twin-jet electro-polisher, using methanol-perchloric acid with a volume ratio of methanol, 1-butanol, and
perchloric acid of 12:7:1 at 233 K, was employed to (a) (b)
prepare the specimens for transmission electron
microscopy (TEM) observations.
50 μm 50 μm 50 μm
Figure 3 (a) Schematic illustration of a stress-strain curve with interruption conditions, and SEM images of an identical region
(b) before deformation, at the (c) elastic limit, (d) micro-yield stage, (e) macro-yield stage, and (f) work hardening stage in the
unaged specimen. [4]
4. Conclusion
Both the yield and tensile stresses at 77 K increased until the aging time of 6 ks, and then decreased until 600 ks. Under all
aging conditions, the nominal stress-nominal strain curves at 77 K exhibited low elastic limits, and the nominal stress plateau
regime appeared at a nominal strain of approximately 0.002, after which, work hardening occurred. In the unaged and 6 ks
aged specimens, several deformation twins occurred after the elastic limit and remarkably increased in the nominal stress
plateau regime, indicating macroscopic yielding. At the work hardening stage, the number of deformation twins increased
slightly. Therefore, macroscopic yielding occurred because of deformation twinning. In contrast, in the 600 ks aged specimen,
deformation twins negligibly increased at the macroscopic yield, indicating that macroscopic yielding occurred due to slip
deformation.
Acknowledgement
This work was performed under the support of the 28th ISIJ Research Promotion Grantt and the Grant-in-Aid for Scientific Research
(KAKENHI) Grant. (No. 20K1460.).
Reference
[1] N. Maruyama, M. Takahashi: Tetsu-to-Hagane, 93 (2007), 506.
[2] S. Araki, K. Fuji, D. Akama, T. Tsuchiyama, S. Takaki, T. Ohmura, J. Takahashi: ISIJ Int., 58 (2018), 1920.
[3] N. Koga, O. Umezawa, M. Yamamoto, T. Yamamoto, T. Yamashita, S. Morooka, T. Kawasaki, S. Harjo: Metall. Mater. Trans. A.,
52 (2021), 897.
[4] N. Koga, Y. Kanehira, P. T. H. Huyen, K. Hori, O. Umezawa: ISIJ Int., 61 (2021), 2308.
*email: park.myeongheom.8r@kyoto-u.ac.jp
Keywords: dual phase steels, mechanical properties, digital image correlation (DIC), strain distribution
Background
Low-carbon dual phase (DP) steels composed of soft ferrite and hard martensite are widely used in automotive industries
due to their good combination of strength and ductility. Over several decades, many studies of DP steels have been
achieving better mechanical properties by changing microstructural factors such as grain size, morphology, volume fraction
and strength (hardness) of each phase 1-4). It is believed that such a well-balanced strength-ductility of DP steels is attributed to
microscopically heterogeneous deformation accompanied with mechanical interaction between two phases having different
hardness. The difference in hardness (strength) between two phases may significantly affect local deformation behavior as
well as global mechanical response. For understanding underlying deformation nature of DP steels having different phase
hardness (strength), the local deformation behavior should be characterized. Digital image correlation (DIC) method 5, 6)
that can provide full-field local strain distributions may be one promising technique for quantitatively characterizing
deformation heterogeneity in DP steels 7, 8). In the present study, we prepared two kinds of DP structures having different
hardness of martensite through tempering heat treatment, and characterized global tensile properties and local deformation
behavior in terms of strain-partitioning evolution of each phase using microscopic DIC (μ-DIC) technique for understanding
local deformation nature of DP steels with the change of martensite hardness.
Experimental procedure
A low carbon steel with a chemical composition of Fe-2Mn-0.1C (wt. %) was used in the present study.
Homogenization at 1200℃ for 48 hrs + furnace cooling under a vacuum atmosphere and subsequent cold-rolling by 90%
reduction in thickness were carried out. Afterwards, the cold-rolled specimen was austenitized at 830℃ for 3 hrs and
subsequently furnace-cooled to obtain ferrite + pearlite (F+P) structure. DP structure was finally obtained by ferrite +
austenite two phases intercritical annealing at 750℃ for 1 hr and subsequent water-quenching (WQ) during which the
austenite transformed to martensite. This DP specimen obtained will be referred to as “as-quenched DP”. In parallel, the
as-quenched DP specimens were subjected to tempering heat treatment at 400 ℃ for 1 hr followed by WQ, and the obtained
specimen will be referred to as “tempered DP” hereafter. Microstructural observation was carried out by a field emission
type scanning electron microscope (FE-SEM) (JEOL, JSM-7800F) equipped with backscattered electron (BSE) detector at an
accelerated voltage of 15 kV. For SEM observation, the specimen was mechanically polished with 400-4000 grid-sized
emery papers and then electrolytically polished in a chemical solution of 90 vol% of CH3COOH and 10 vol. % of HClO4
with a voltage of 22 V for 30 s at room temperature. The ferrite grain size was measured by a linear interception method on
SEM microstructure images and the martensite volume fraction was estimated by the point counting method. Nano
hardness of each phase was measured by the use of a nano-indentation testing machine (TI 950 TriboIndenter, Hysitron) with
an indent load of 1000 μN and a dwell time of 0.5 s. Global mechanical properties of the DP specimens were characterized
by uniaxial tensile tests at room temperature with an initial strain rate of 8.3×10-4 s-1, Shimadzu AG-100kN Xplus tensile-test
machine. The tensile direction of the specimens was set to be parallel to the rolling direction of the sheets, and the tensile
specimens with a gage length of 10 mm and a cross section of 3 mm (width) × 1mm (thickness) were cut from the sheets.
Tensile strain of the tensile specimens was precisely measured by the digital image correlation (DIC) technique using a
dedicated DIC analysis software (Correlation Solutions, Vic-2D). For microscopic DIC (μ-DIC) analysis, surfaces of the
tensile specimen were polished prior to the tensile test, and identical area observations were carried out in SEM. For
Results
Figure 1 shows SEM microstructures of (a) as-quenched DP and (b) tempered DP specimens. F and M indicated on
the images denote ferrite and martensite, respectively. Two kinds of DP specimens showed very similar microstructural
features in martensite distribution (i.e., network-shaped morphology surrounding ferrite grains), martensite volume fraction (~
25%) and ferrite grain sizes (~15 μm) in the low magnification SEM images. In high magnification SEM images, on the
other hand, a great number of fine precipitates with bright contrast were observed in the interior of martensite in the tempered
DP specimen, which were considered to be carbides formed by tempering treatment. In ferrite grains, no significant changes
were found even at high magnification observations.
Figure 1 SEM microstructures of the (a) as-quenched DP specimen and (b) tempered DP specimen.
Nano hardness of each phase was investigated by nano-indentation testing, and obtained hardness results are shown in
Table 1. It was found that a significant decrease in martensite hardness by tempering heat treatment was recognized,
whereas ferrite showed nearly the same hardness value, indicating that softening effect by tempering took place mainly in the
martensite due to a decrease of dislocation density and solid-solution strengthening effect of carbon.
Table 1 Nano hardness of ferrite and martensite in the as-quenched DP and tempered DP specimens.
Figure 2 demonstrates engineering stress-strain curves of the (a) as-quenched DP and (b) tempered DP specimens.
From the results of Fig. 1 and Table 1, it is expected that the difference in stress-strain behavior between two kinds of DP
specimens is mostly attributed to martensite hardness (strength). Discontinuous yielding was observed in the tempered DP
specimen, in which Lüders (strain localization) band propagation was confirmed in DIC strain distribution map during tensile
deformation. The tempered DP specimen having softer martensite exhibited decreased strength but enhanced ductility
Figure 2 Engineering stress-strain curves of the (a) as-quenched DP specimen and (b) tempered DP specimen with an initial
strain rate of 8.3 × 10-4 s-1 at room temperature.
Figure 3 displays SEM microstructures and corresponding DIC strain distribution maps of the (a) as-quenched DP and
(b) tempered DP specimens tensile-deformed to ~ 2% plastic strain. Tensile direction is parallel to the horizontal direction of
images, and the colors superimposed on SEM images correspond to the value of normal strain (ε11) in the tensile-axis direction
according to the key strain color bar. The particles on SEM images are the colloidal silica (SiO2) acting as markers for
making the μ-DIC analysis precise. In the as-quenched DP specimen, some of ferrite grains experienced very strong strain
localization due to hardly deformed martensite phase. The tempered DP specimen, in contrast, exhibited relatively
homogeneous deformation in μ-DIC strain distribution maps as a result of more contribution of softened martensite to plastic
deformation. The significant difference in global ductility between two DP specimens can be understood by the local strain
distribution behavior, especially in terms of strain-partitioning between ferrite and martensite.
We confirmed the strain distribution difference between the as-quenched DP and tempered DP specimens using the μ-
-DIC technique in the current study. The evolution of internal stress and dislocation density in different constituent phases
obtained from in-situ XRD analysis will be presented in the presentation. Based on the experimentally obtained strain- and
stress-partitioning results, underlying deformation mechanism of the DP steels and an effective material design strategy that
can achieve simultaneous enhancement of strength and ductility will be also discussed.
Conclusions
In summary, we have investigated global mechanical properties of the DP structures having different martensite hardness
and characterized local deformation behavior in terms of strain-partitioning between ferrite and martensite by means of the
μ-DIC analysis. The main conclusions obtained in the current study are as follows:
(1) Two kinds of DP specimens having different martensite hardness were successfully fabricated.
(2) In the stress-strain behavior, the DP specimen obtained by tempering treatment (tempered DP) having softer martensite
exhibited decreased strength but enhanced ductility especially in post-uniform elongation, compared with the
as-quenched DP specimen with harder martensite.
(3) In the as-quenched DP specimen, some ferrite grains experienced strong strain localization due to hardly deformed
martensite phase. The tempered DP specimen exhibited relatively homogeneous deformation in μ-DIC strain
distribution maps, as a result of more contribution of martensite to plastic deformation. Homogeneous deformation in
the microstructure shown in the tempered DP specimen could be one reason for the enhancement of global ductility (total
elongation), inhibiting strain localization.
Acknowledgement
This research work was financially supported by the Elements Strategy Initiative for Structural Materials (ESISM;
JPMXP0112101000), JST CREST (JPMJCR1994), and KAKENHI (Grant-in-Aid for Scientific Research from the Japan Society
for the Promotion of Science; 15H05767, 19H02459 and 20K14608), all from the Ministry of Education, Culture, Sports, Science
and Technology (MEXT), Japan. All the supports are gratefully appreciated.
References
1) M. Calcagnotto, Y. Adachi, D. Ponge, D. Raabe: Acta Mater. 59 (2011) 658-670.
2) A. A. Sayed and Sh. Kheirandish: Mater. Sci. Eng. A. 532 (2012) 21-25.
3) M. Soliman and H. Palkowski: Mater. Sci. Eng. A. 777 (2020) 139044.
4) M. Calcagnotto, D. Ponge and D. Raabe: Mater. Sci. Eng. A. 527 (2010) 7832-7840.
5) J. Zhang, A. Sweedy, F. Gitzhofer and G. Baroud: Opt. Lasers. Eng. 100 (2018) 259-266.
6) S. Hamada, T. Fujisawa, M. Koyama, N. Koga, N. Nakada and T. Tsuchiyama: Mater. Charact. 98 (2014) 140-146.
7) C.C. Tasan, J.P.M. Hoefnagels, M. Diehl, D. Yan, F. Roters and D. Raabe: Int. J. Plast. 63 (2014) 198-210.
8) K. Park, M. Nishiyama, N. Nakada, T. Tsuchiyama and S. Takaki: Mater. Sci. Eng. A. 604 (2014) 135-141.
Noriyuki Tsuchida1,*
1
Graduate School of Engineering, University of Hyogo, 2167 Shosha, Himeji, Hyogo 671-2280, Japan
*email: tsuchida@eng.u-hyogo.ac.jp
1. Introduction
Steels are usually used as multi-phase or multi-microstructure steels which consist of different phases or microstructures.
Thus, it is very important to discuss their mechanical properties from the viewpoint of deformation behavior of each
constituent phase or microstructure 1,2). We have been studied the tensile deformation behavior of various multi-microstructure
steels from the phase strains and the lattice strains obtained by neutron diffraction experiments 3−5). From our previous studies,
it is found that the tensile deformation behavior of constituent phases and their difference, i.e., stress partitioning, play
important roles in the mechanical properties of various multi-phase steels. The stress partitioning also affects the
deformation-induced martensitic transformation (DIMT) behavior of austenite (γ) phase 6). In this study, I take notice of
transformation induced plasticity (TRIP) steels such as TRIP-aided multi-microstructure steels and metastable austenitic
stainless steels and try to discuss the improvement of mechanical properties due to the stress partitioning and TRIP effect.
neutron
Neutron
Slit
slit Load
dir
Collimator
collimator
-90- 90
degdeg +90
+90 degdeg
detector
detector detector
detector
sample
Sample
Fig. 1 Schematic illustration of in situ neutron Fig. 2 Phase stresses of γ, α and α' in 0.2C (steel A)
diffraction experimental setup. and 0.4C (steel B) TRIP steels as functions of applied
true strain.
0.1
(a)
Temperature change test
deformation-induced martensite
0.08
800
296 K
Volume fraction of
0.06
296 K
Nominal stress (MPa)
600 0.04
Enhancement of 0.02
400 uniform elongation
(+18%)
0
Temperature change 0.5
Transformation rate
(b)
200 0.4 Temperature
296 K change test
0.3
0.2
0 0.1
0 0.1 0.2 0.3 0.4 0.5
Nominal strain 0
0 0.1 0.2 0.3 0.4
True strain
2500
1200
60
Uniform elongation 2000
800 1500
40
?
0.2% proof stress
1000
400
20
500
Fig. 5 0.2% proof stress, tensile strength and uniform Fig. 6 Estimated phase stress of austenite (γ) and
elongation as functions of temperature in the deformation-induced martensite (α') phases vs. true strain
6Ni-0.2N-0.1C and SUS304 steels. in the 6Ni-0.2N-0.1C and SUS304 steels obtained by the
neutron diffraction experiments..
References
1) N. Tsuchida, S. Harjo, T. Ohnuki, Y. Tomota: Tetsu-to-Hagane 100 (2014) 1191.
2) R.G. Davies: Metall. Trans. A. 9 (1978) 451.
3) N. Tsuchida, T. Kawahata, E. Ishimaru, A. Takahashi, H. Suzuki and T. Shobu: ISIJ Int. 53 (2013) 1260.
4) S. Harjo, N. Tsuchida, J. Abe and W. Gong: Sci. Rep. 7 (2017) 15149.
5) N. Tsuchida and S. Harjo: Metals 11 (2021) 2053.
6) N. Tsuchida, T. Tanaka, Y. Toji: Tetsu-to-Hagane 105 (2019) 918.
7) S. Harjo, S. Kubota, W. Gong, T. Kawasaki, S. Gao: Acta Mater., 196 (2020) 584.
8) T. Yamashita, S. Morooka, S. Harjo, T. Kawasaki T, N. Koga, O. Umezawa: Scripta Mater. 117 (2020) 6.
9) N. Tsuchida, R. Ueji, T. Inoue: ISIJ Int. 61 (2021) 632.
10) N. Tsuchida, E. Ishimaru, M. Kawa: ISIJ Int. 61 (2021) 556.
Abstract:
The observation of deformation behavior, plays an important role in building basic knowledge and
providing feedback in material development. With the recent significant improvement in the properties
of advanced steel materials, the structure and deformation mechanism have become complicated.
Multi-phase structures have become commonplace, with microstructural changes and multiple
deformation modes occurring simultaneously during deformation. Such conditions have made
material characterization difficult.
Neutron diffraction is becoming one of the powerful probes to characterize various properties related
to the strength and microstructures of engineering materials, and its usefulness is also becoming well
known. Neutron has a high penetration depth and can be used to observe atomic arrangement using
diffraction. Careful analysis of the Bragg peaks in a neutron diffraction pattern can reveal important
structural details of a sample material such as internal stresses, phase conditions, dislocations, texture
etc. Such information is often crucial in engineering applications and the ability to carry out either ex
situ or in situ measurements makes neutron diffraction particularly useful in this respect. With the
application of neutron diffraction using the time-of-flight method, in situ neutron diffraction
measurement under various conditions became feasible, as well as to increase the accuracy by
measuring multi-diffraction peaks for analyses. Furthermore, due to the recent increase in the intensity
of the pulsed neutron source and the increase in the resolution of the neutron diffractometer, the
strength and microstructure of the material can be observed with finer changes, and more accurate
understanding of the materials mechanical or functional mechanisms can be achieved.
TAKUMI which was installed in J-PARC, is one of neutron diffractometers which are suited for studies
related to the observation of deformation behavior of steels. In the presentation, the current status of
TAKUMI, its capabilities, and some studies of deformation behavior of steels will be introduced.
Si Gao1), Yu Bai1) 2), Wenqi Mao1) 3), Wu Gong3) 5), Akinobu Shibata4) 5), Nobuhiro Tsuji1) 5)
Abstract:
Conventional austenitic stainless steels (ASSs) exhibit excellent corrosion resistivity and
deformability but low yield strength due to their FCC structure with coarse grain sizes. In recent years
it has been found that the ultra-fine grained (UFG) ASSs, usually fabricated by heavy rolling and
annealing processes, exhibit good combinations of strength and tensile ductility which make them
prospective materials for future industrial applications. Deformation induced martensitic
transformation (DIMT) is believed to play an important role in the excellent mechanical properties of
the UFG ASSs, however, the effect of the ultra-fine grain size on DIMT has not yet been clarified and
the corresponding excellent mechanical properties of UFG ASSs are lacking of interpretation. In the
present study, commercial 304 ASS specimens having different grain sizes ranging from 0.3 µm to 22
µm were fabricated by cold rolling and annealing process, and tensile tests with in-situ neutron
diffraction were carried in order to investigate the grain size dependence of the kinetics of DIMT and
the stress partitioning between austenite and martensite during tensile deformation. Tensile test results
showed that the yield strength increased from 250 MPa to 1000 MPa and the total elongation decreased
from 80% to 50% when the grain size of the specimen decreased from 22 µm to 0.3 µm. Neutron
diffraction revealed that stress partitioning behavior between austenite and martensite did not exhibit
notable difference among the specimens, while the kinetics of the DIMT dramatically increased by
decreasing the grain size, which was believed to be the main reason for the excellent tensile properties
of the UFG specimens. The kinetics of the DIMT will be discussed in terms of the mechanical stability
of austenite. Other interesting deformation behaviors associated with the DIMT, such as the huge
Lüders deformation in the UFG specimen, will also be discussed.
Division of Materials Science and Engineering, Hanyang University, Seoul 04763, Republic of Korea
1
2
Department of Advanced Materials Engineering, Korea Polytechnic University, Siheung 15073, Republic of Korea
3
Joining and Welding Research Institute, Osaka University, Osaka 567-0047, Japan
4
Department of Materials Science and Engineering, Northwestern University, Evanston, IL 60208, USA
5
Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic
of Korea
*email: jeonghohan@hanyang.ac.kr
Medium-Mn steels, which are composed of ~3-12 wt% Mn, <0.3 wt% C, and minor doping of steel alloying elements,
have been widely studied during last years, and they are suggested as next generation advanced high-strength steel and
cryogenic materials [1]. After hot rolling process, the steel normally reveals 100% α martensitic microstructure, however, it
shows two-phase microstructure with retained austenite (γR) and tempered martensite (αT) after intercritical annealing. Owing
to the large fraction of γR (< 40 vol%), the steel reveals an excellent combination of strength and ductility, and also shows the
high impact absorbed energy at ambient temperature. However, due to Mn segregation at pre-existed prior austenite grain
boundary (PAGB), the steel shows the severe intergranular crack when fractured at cryogenic condition (e.g., -196 °C),
leading to poor impact absorbed energy [2].
In the present study, to predict the possibility of application of medium-Mn steel for cryogenic materials, we investigated
the cryogenic impact toughness of medium-Mn steel after welding. Briefly, we propose a friction stir welding (FSW) is a
novel welding method of the medium-Mn steels for the cryogenic application [3]. The medium-Mn steel with two-phase
microstructure (retained austenite and martensite) was used as a base material, and the microstructures and impact toughness
after FSW and conventional tungsten inert gas (TIG) welding were systematically investigated and compared. The prime
scientific novelty of the present study is summarized below:
1. The FSW-processed steel revealed a greater impact absorbed energy than the TIG-welded steel owing to the presence
of retained austenite at weld zone.
2. The FSW-processed steel showed a higher toughness than base material, because the grain morphology tuning driven
by FSW could eliminate the weak microstructural feature of base material (i.e., prior austenite boundary) and lead to more
active transformation-induced plasticity.
Acknowledgement
This work was performed under the support of National Research Foundation of Korea (No. 2020R1F1A1070808) and Korea
Institute for Advancement of Technology (No. P0008425).
References
1) Y.-K. Lee and J. Han: Mater. Sci. Technol. 31 (2015) 843-856.
2) J. Han, A.K. da Silva, D. Ponge, D. Raabe, S.M. Lee, Y.-K. Lee, S.I. Lee, B. Hwang: Acta Mater. 122 (2017) 199-206.
3) M.S. Jeong, T.M. Park, D.-I. Kim, H. Fujii, H.J. Im, P.-P. Choi, S.-J. Lee, J. Han: J. Mater. Sci. Technol. 118 (2022) 243-254.
Abstract: Al-Si coated press hardening steel (PHS) is widely applied on carbody due to its high strength,
which tensile strength usually exceeds 1500MPa. In addition to strength, the fracture strain of PHS in
bending deformation characterized by the bending angle in tests of VDA238 standard is critical for the
crash performance of carbody. The Al-Si coating transforms into Fe-Al intermetallics during
austenitization heating in the hot stamping process, and therefore prevents severe oxidation of the sheet
surface. However, the Al-Si coating was found to seriously deteriorate the fracture strain. In this research,
the fracture process of a press-hardened Al-Si coated PHS during bending test was investigated. The Al-
enriched alloying layers involving the intermetallic layers and the ferrite layer are not able to inhibit the
crack propagation, but the cracks are then blunted by the hard martensite substrate. Finally, fracture
happens by the shear band initiation and propagation from the surface of martensite. The carbon
enrichment about several micrometers in martensite at the coating/substrate interface after press-
hardening was found for the first time to deteriorate the bending fracture strain of Al-Si coated PHS.
The coating/substrate interface migration during austenitization heating leads to the carbon diffusion
into the austenite substrate because carbon has large solubility in austenite compared to the negligible
solubility in Al-enriched alloying layers. A novel method of reducing the coating thickness has been
then invented to suppress the carbon enrichment by suppressing the coating/substrate interface migration.
This invention will change the commercial practice of PHS by improving 20% of bending fracture strain
and sacrificing neither of the other performances such as spot-weldabiltiy, paintability and corrosion
resistance after e-coating. The thin coating has another advantage that the austenitization soaking period
is reduced by 20%~30% due to the much less heat absorption of Al liquidation and the formation of Fe-
Al intermetallics.
*email: d.ponge@mpie.de
Keywords: medium manganese steel, TRIP, hydrogen embrittlement resistance, austenite/ferrite interface boundary
Material
A model medium Mn steel 0.2C-10.2Mn-2.8Al-1Si (in wt.%) was cast, hot and cold rolled and intercritically annealed (IA).
Slow strain rate tensile testing was performed for IA700 and IA800 with samples not charged with H (black curves in Fig. 2a)
and samples after pre-charging with H (red curves in Fig. 2a). The electrochemical pre-charging was performed in 3 wt.%
NaCl and 0.3 wt.% NH4SCN at current density of 5A/m2 for 24 h. IA800 shows a lower value of the yield stress but higher
strain hardening rates compared to IA700. The reason for the higher strain hardening rate is the lower chemical austenite
For H pre-charged samples, thermal desorption spectroscopy (TDS) revealed a lower total content of 6.2 wt. ppm H in IA800
than in IA700 (18.2 wt. ppm H) for identical charging conditions. This fits to the idea that in the case of IA800 ( matrix) the
fast ingress of H along will be blocked by grains (Fig. 2c). Important to stress here that, even with lower yield stress and
lower total H content in IA800, the charged material fails at a low strain (11%) than IA700 (24%) and the loss in ductility due
to HE in IA800 is very high. Unfortunately, this is due to the more active TRIP effect in IA800 and its lower austenite
mechanical stability. This is beneficial in the uncharged condition to postpone the necking. However, in the charged condition
this leads to early H cracking. The mechanism is shown and explained in Fig. 3. We report more details of the different
embrittling mechanisms in IA800 and IA700 in [1].
Fig. 3 (a) EBSD phase map with image quality
map shows partial transformation of austenite into
fresh martensite α'. This is due to a strain induced
martensitic transformation (or TRIP-effect).
A “fresh” α'/ α' grain boundary is highlighted by
two white arrows.
(b) Schematic of redistribution of H after partial
-> α' transformation: has a higher H solubility.
During the displacive -> α' transformation
all H atoms will be inherited from to α'.
The higher oversaturation and faster diffusion
in α' lead to a fast redistribution of H to
dislocations and boundaries.
(c, d) At larger local strain, cracks along H en-
riched α'/α' and /α' boundaries are observed,
very often initiated at triple junctions.
(e) Schematic of cracking mechanism in (c, d).
This HE mechanism in IA800 can be classified as a phase transformation-driven H-enhanced decohesion (HEDE). In contrast,
in IA700 we observed a less detrimental effect of hydrogen (a ductile fracture mode with increased void nucleation rate by H)
[1]. But this comes with the drawback to hinder the TRIP effect by the higher mechanical stability of the austenite (higher Mn
and C content) in IA700. Is there a way to still take advantage of the ductility enhancing TRIP effect without significantly
decreasing HE resistance? The answer is “yes” and a way is shown in the following (more details in [2] and [3]):
Fig. 4 One step IA leads to a homogeneous distribution of Mn Fig. 5 New multistep IA treatment creates inhomogeneous Mn
in when α' forms, H redistribution favors sharp cracks distribution. Mn-lean austenite allows for TRIP effect.
along boundaries and cleavage cracks inside α'. Possible cracks are blunted by the stable Mn-rich austenite.
Fig. 6 shows the modified treatment (Fig. 6 b,c,d) to create the HET microstructure and schematically the microstructure
changes along all processing steps (Fig. 6 e,f,g,h,i).
Fig. 7 (a) Phase map shows and after treatment shown in Fig. 6. Fig. 8 In comparison the EDX Mn map for the single
(b) corresponding EDX Mn map reveals heterogeneous Mn step IA annealing (cold rolled to same thickness as
distribution in (Mn-rich highlighted by red circles). HET and 750°C/5min).
In uncharged condition the stress-strain curves of HET and HOM samples are similar [3]. The transformed austenite fraction
versus local strain curves are also similar for HET and HOM samples [3], indicating a similar TRIP effect and mechanical
performance in H free condition.
In contrast, tensile tests of H charged HET and HOM samples reveal a significantly higher total elongation of HET samples in
comparison to HOM samples (Fig. 9) and samples of other advanced high strength steels of similar strength (Fig. 10).
Fig. 9 Results from slow strain rate tensile tests for HOM and Fig. 10 Comparison of total elongation versus H
HET samples; TDS shows similar total H content after H concentration (HET, HOM and other advanced high
charging (3 wt.% NaCl, 0.3 wt.% NH4SCN, 5A/m2, 24 h). strength steels with similar strength level).
Acknowledgement
Binhan Sun was supported by Alexander von Humboldt Foundation.
References
1) B. Sun, W. Krieger, M. Rohwerder, D. Ponge and D. Raabe: Acta Materialia 183 (2020) 313–328.
2) B. Sun, W. Lu, B. Gault, R. Ding, S. K. Makineni, D. Wan, C.-H. Wu, H. Chen, D. Ponge and D. Raabe: Nat. Mater. 20 (2021)
1629–1634. https://doi.org/10.1038/s41563-021-01050-y
3) Supplementary inf. to 2). https://static-content.springer.com/esm/art%3A10.1038%2Fs41563-021-01050-y/MediaObjects/41563_2021_1050_MOESM1_ESM.pdf
Keywords: Hydrogen embrittlement, Pearlitic steel, Lamella, Cold drawing, Axial crack
1. Introduction
The tensile strength of pearlitic steel is greatly improved by the refinement of lamellar spacing or cold drawing.
However, in general an increase in ultimate tensile strength (UTS) causes the reduction of the resistance to hydrogen
embrittlement (HE) 1-3). Nevertheless, pearlitic steel has a superior resistance to HE to tempered martensitic steel at the same
level of tensile strength due to its unique microstructure without prior austenite grain boundaries 4). Although many studies
have been performed on the HE of pearlitic steel, it is still needed to investigate the resistance to HE as functions of lamellar
spacing (λ) and prestrain. Therefore, in the present study the resistance to HE was evaluated using pearlitic steel specimens
with various lamellar spacings and cold drawing ratios (CDRs) by means of thermal desorption analysis (TDA), fractography
and silver decoration 5).
2. Experimental method
Pearlitic steel with a chemical composition of Fe-0.72C-0.20Si-0.70Mn-0.03Al (wt %) was used for this study. To
examine the effect of cold drawing on the resistance of HE, the wires of pearlitic steel were cold-drawn from 17 mm in
diameter to 14.2 mm, 12.6 mm and 10.8 mm after austenitizing at 1323 K and lead patenting at 923 K. The ASTM E-8M-03
subsized, round-shaped tensile specimens were machined using the cold-drawn wires. To investigate the effect of lamellar
spacing, undeformed pearlitic steel wires were machined to make tensile specimens. Tensile specimens were austenitized at
1273 K for 30 min, quenched to 823 K, 873 K and 923 K, and then held at a given temperature for 30 min, followed by air
cooling to room temperature. The specimens were electrochemically charged with H using a 3% NaOH aqueous solution
containing 0.3% NH4SCN. The H charging was conducted with a current density of 50 A/m2 at 300 K for 24 h. For thermal
desorption analysis (TDA), all H-charged specimens were heated at a constant rate of 100 K/h from 300 K to 693 K. slow
strain rate tensile tests (SSRTs) were performed at a crosshead speed of 0.1 mm/min. Fracture surfaces of tensile specimens
were observed using a FE-SEM after the SSRTs. The diffusion behavior of H was examined through the silver decoration
technique by immersed the samples into a 4.3 mM Ag[K(CN)2] aqueous solution for 1 h.
To analyze the reasons for the unusual results, we observed fracture surfaces and microstructures. In the H-charged,
undeformed specimen with a narrow λ, tearing topography surface (TTS) appeared inside the tensile specimen after SSRT
(Fig. 2a). The TTS is an initiation site for HE in the pearlitic steel 6), and its location is related to H diffusion 7). Accordingly, H
permeation occurred more actively in the specimen with a narrow λ during tensile deformation (Fig. 2b).
Fig. 2. (a) Macroscopic fracture surface of H-charged, undeformed specimen with a fine lamellar spacing after SSRT 8) and (b) SEM image
showing the microstructure of the specimen isothermally transformed at 823 K.
Meanwhile, the H-charged, cold-drawn specimens showed not only peak 1 at ~375 K, but also a second peak
(peak 2) at ~600 K (Fig. 3a) in their H desorption rate curves. When tensile strain increased, peak 1 decreased whereas peak 2
increased. The main H trap site of peak 2 is the strained FP/θ interface with interfacial dislocations. Unexpectedly, when the
CDR was above 30%, both tensile strength and the resistance of HE were simultaneously improved despite the higher
diffusible H content (Fig. 3b).
Fig. 3. Variations of (a) H content and (b) the losses of El and RA with cold drawing ratio in pearlitic steel.
To find the cause of the improvement of the resistance to HE, the fracture surface of the H-uncharged specimen
Fig. 1. (a) The surface of a micro-axial crack almost parallel to the direction of cold drawing in a H-uncharged, 60% cold-drawn specimen.
Schematic illustrations showing (b) pre-existing micro-cracks in as cold-drawn specimen, (c) additional formation of micro-cracks, (d) crack
coalescence, and (e) micro-axial cracking through a misaligned colony during tensile deformation 9).
4. Conclusions
1) As the λ decreased, the resistance to HE was improved despite the increase in tensile strength. This is because the formation
of TTS was delayed in the specimens with fine λ and short θ platelets.
2) Both the tensile strength and the diffusible H content of unnotched specimens was raised with increasing CDR.
Nevertheless, the resistance to HE was improved when the CDR was higher than 30% due to micro-axial cracking.
Acknowledgement
This work was performed under the support of the POSCO Technical Research Laboratories, Pohang, Republic of Korea (No.
2016Z008).
References
1) J. Toribio and A. Lancha: Mater. Struct. 26 (1993) 30-37.
2) J. Toribio and E. Ovejero: Corros. Sci. 49 (2007) 3539-3556.
3) J. Toribio: Eng. Fail. Anal. 94 (2018) 157-164.
4) J.S. Kim, Y.H. Lee, D.L. Lee, K.-T. Park, C.S. Lee: Mater. Sci. Eng. A 505 (2009) 105-110.
5) J. Yao and J. Cahoon: Metall. Trans. A 21 (1990) 603-608.
6) J. Toribio, A.M. Lancha, M. Elices: Metall. Trans. A 23 (1992) 1573-1584
National Materials Corrosion and Protection Data Center, Institute for Advanced Materials and Technology, University of Science
and Technology Beijing, Beijing 100083, China
*email: luohong@ustb.edu.cn
Traditional alloys use a metal element as the main element to make it a basic component, and then add different
alloying elements on this basis to obtain certain specific properties. Now it has been widely used by humans such as iron
alloys, aluminum alloys, Magnesium alloy, titanium alloy, etc. With the development of science and technology and the
improvement of human needs, especially the desire for space and deep-sea exploration, human beings have higher and higher
requirements for the physical and mechanical properties of materials. Traditional alloys and traditional preparation methods
have been difficult to meet this demand. Alloys face serious challenges. However, the design ideas of traditional alloy systems
are based on one or two components as basic components, and a small amount of other elements are added on this basis to
improve their properties. This design idea leads to the crystal structure, physical and mechanical properties of the alloy. All are
subject to the main element, which has an adverse effect on the overall performance of the alloy [1, 2]. It is based on this
thinking that multi-component alloys are developed. Multi-principal alloys are alloys that are formed by mixing equal or
relatively large proportions of five or more elements. These alloys are currently the focus of significant attention in materials
science and engineering because they have potentially desirable properties. Furthermore, research indicates that some alloys
have considerably better strength-to-weight ratios, with a higher degree of fracture resistance, tensile strength, and corrosion
and oxidation resistance than conventional alloys. Among them, the iron-based multi-principal alloys currently developed
account for a very high proportion of the entire multi-component alloy field. Like many alloys, multi-component alloys often
face corrosion, hydrogen embrittlement and cracking during environmental service, and their environmental sensitivities such
as corrosion and hydrogen embrittlement are closely related to the composition and structure of the passive film on the surface.
In the present work, we use traditional electrochemical methods, XPS and ICP-MS to study the environmental corrosion
behavior and mechanism of FeCoCrNiMn and carbon-doped FeCoCrNiMn multi-component alloys. The relationship
between the elements of Fe, Cr, Ni and the stable of the passive film are demonstrated. The mechanism of corrosion failure of
the multi-component alloy in acidic and neutral is well established. In addition, the influence of hydrogen element on the
surface microstructure and morphology of various iron-based multi-component alloys was in-situ observed by ECCI. The
hydrogen induced cracking of iron-based multi-component alloys were studied by SEM, TEM, ECCI and other analytical
testing and characterization methods. The influence of inclusions, precipitates and deformation structure on crack initiation
and crack propagation are invesitaged and the fracture behavior and mechanism of iron-based multi-component alloys in
hydrogen environment are also revealed [3-5].
Acknowledgement
This work was performed under the support of National Natural Science Foundation of China (No. 52071014).
References
1) Fu Y, Li J, Luo H, et al. Journal of Materials Science & Technology, 2021, 80: 217-233.
2) Luo H, Zhao B, Pan Z, et al. Materials Science and Engineering: A, 2021, 819: 141490.
3) Luo H, Lu W, Fang X, et al. Materials Today, 2018, 21(10): 1003-1009.
4) Luo H, Li Z, Mingers A M, et al. Corrosion Science, 2018, 134: 131-139.
5) Luo H, Zou S, Chen Y H, et al. Corrosion Science, 2020, 163: 108287.
*email: s-takagi@jfe-steel.co.jp
Keywords: delayed fracture, high strength steel, hydrogen content, local approach, high pressure hydrogen gas
The hydrogen society will come in the near future. Nowadays, the construction of hydrogen infrastructure and the
improvement of efficient energy consumption in various transportation and facilities are accelerated. In the situation, the
application of high strength steels must be expanded because the weight reduction of mobility and the downsizing of
equipment which are effective for the improvement of energy efficiency and the cost reduction of hydrogen infrastructure can
be realized with application of high strength steels. On the other hand, mechanical properties of high strength steels deteriorate
due to hydrogen atoms intruding into steels11-2). With the construction of a hydrogen society and the increase in the application
of high strength steels, the situation in which hydrogen atoms contact high strength steels increase. Therefore the issue of
hydrogen embrittelment in high strength steels becomes important more and more. For expansion of the application of high
strength steels, the risk of hydrogen embrittlement (hereinafter HE) have to be evaluated appropriately. In this topic, various
HE evaluation methods are shown and issue of HE evaluation are introduced.
There are many HE susceptibility evaluation index, for example, critical hydrogen content at fracture limit (Hc),
fracture stress, total elongation, reduction of area and fatigue crack growth rate. These properties are obtained by Constant
Load Test(CLT), Slow Strain Rate Technique(SSRT), Constant Strain Test (for example, 4-point bending test, U or V-bending
test) ,fatigue crack growth rate test.
By CLT with circumferential notch specimens under hydrogen pre-charging environment, various studies were
carried out on tempered martensitic steels with tensile strength of about 1500MPa for bolt3-6). The critical diffusible hydrogen
content (Hc) which is the maximum diffusible hydrogen content in the unfracture specimens can be determined. The value of
Hc is one of the index of hydrogen embrittlement resistance of steels and it becomes lower with increase of stress
concentration factor at notch root which is influenced notch root radios and specimen diameter and specimen diameter5-6).
Initial cracks are generated at the position where stress along tensile direction and hydrostatic stress are maximum7). It is
suggested HE critical condition can be determined by local hydrogen content and local stress around the fracture point.
Diffusible hydrogen atoms diffuse to high hydrostatic stress region and accumulate. In detail, stress distribution and hydrogen
distribution in the fracture process zone are important and if the material is more brittle, the process zone is limited close to the
maximum stress and the maximum hydrogen content position8-9). In CLT, when cracks are generated all at once, those cracks
absolutely propagate due to the severe stress and hydrogen content conditions at the crack tip7). Therefore, the crack initiation
condition is equal to the fracture critical condition.
SSRT has been using in various field10-12). SSRT is conducted under strain rate from about 10-4/s to 10-6/s because
HE phenomena is clearly shown under the strain rate lower than about 10-4/s12-13). Testing time of SSRT is about several hours.
It is shorter than that of CLT which is about several days to several month. So, by SSRT a lot of data can be obtained more
easily than by CLT. On the other hand, by SSRT unfracture data cannot be obtained14). HE is evaluated with decrease of
elongation, strength or reduction of area. Degree of HE increase with increasing hydrogen content in specimens or severe
hydrogen charging condition11).
Constant strain tests can be easily conducted with a small apparatus. High strength steel sheets for automobile are
used after press forming. Since microstructural damage introduced by press forming affects the HE failure limit, specimens
for the evaluation of HE sensitivity should better simulate the forming mode. A U-bending specimen has been used which
was tightened by bolt for applying stress in order to evaluate the HE susceptibility of roll-formed pipe applied to the impact
beam15-18). Nowadays, the application of ultra-high strength steel sheets is expanding and it is necessary to evaluate the HE
susceptibility in various forming modes. V-bending test19), 4-point bending test20), drawn cup method15, 21-23) and so on, are used
corresponding to the situation. Sometimes as-sheared specimen is also used for simulation of shear or punched edge after
References
1) A.R.Troiano:Tras.ASM, 52(1960), pp.54-80.
2) S.Matsuyama:Delayed Fracture (The Nikkan Kogyo Shimbun, LTD. Tokyo, (1989). (in Japanese)
3) S.Yamasaki and T.Takahashi:Tetsu-to-Hagané, 83(1997), pp.454-459. (in Japanese)
4) S. Yamasaki, M.Kubota and T.Tarui : Nippon Steel Technical Report, 80(July 1999), pp.50-55.
5) S. Takagi, T. Inoue, T. Hara, M. Hayakawa, K. Tsuzaki and T.Takahashi: Tetsu-to-Hagané, 86 (2000), pp.689-696. (in Japanese)
6) S. Takagi, T. Inoue, K. Tsuzaki and F. Minami: Proceedings of International Symposium on Today and Tomorrow in Science and
Technology of Welding and Joining (7WS), Kobe, Japan, (2001), pp.1159-1164.
7) S. Yamasaki and T. Takahashi: Tetsu-to-Hagané, 83(1997), 460-465. (in Japanese).
8) S. Takagi, T. Inoue, K. Tsuzaki and F. Minami: J. Jpn. Inst. Met., 65(2001), pp.1073-1081. (in Japanese)
1. Introduction
From economic and environmental viewpoints, the demand for high-strength steels is increasing. As the susceptibility to
hydrogen embrittlement of high-strength steels, particularly martensitic steels, is considerably high, hydrogen embrittlement is
one of the serial issues hindering wide application of high-strength martensitic steels. To develop advanced high-strength
steels with high resistance to hydrogen embrittlement, understanding the underlying mechanism of hydrogen embrittlement is
very important. However, the detailed fracture sequence of hydrogen embrittlement has not yet been clarified. The typical
modes of hydrogen-related fracture are classified into quasi-cleavage and intergranular. The hydrogen-related intergranular
fractures are often linked with significantly more severe brittle behaviors. Crack morphology is an important characteristic
which is closely related to the macroscopic mechanical properties. Although conventional 2D surface observation by using
SEM is appropriate for statistical analysis, it has inevitable limitations for analyzing crack morphology in detail. On the other
hand, 3D observation with different scales is a promising method for deeply understanding crack morphology. We employed
a multi-scale 3D analysis, i.e., X-ray computed tomography and FIB-SEM serial sectioning, to understand the morphology of
hydrogen-related intergranular crack in martensitic steel.
2. Experimental Procedure
The material used in the present study was an Fe-8Ni-0.1C alloy with a high nickel content to obtain high hardenability. This
is a model steel which exhibits intergranular fracture even in the uncharged state 1). Compact tension specimens with a
thickness of 12.5 mm (B) and a width of 25 mm (W) were cut from the heat-treated steel plates via spark wire cutting. The
pre-cracked CT specimens with side grooves were cathodically pre-charged with hydrogen in an aqueous solution of 3%
NaCl +3 g L-1 NH4SCN for three days at ambient temperature. Unloading compliance tests were performed to obtain the J –
a resistance curves based on ASTM E1820-17 standard. Unload / reload sequences were performed for a displacement
interval of 0.025 mm, and the rate of crack mouth opening displacement for the test was 2.5 × 10-5 mm s-1.
Conventional 2D observations on the fracture surface and crack propagation were performed using SEM and SEM / EBSD.
The macroscopic 3D crack propagation morphologies were analyzed by computed tomography using X-ray microscopy. A
specimen with the dimensions of ~ 2 mm (W) × 16 mm (L) × 1 mm (T) was cut from the mid-thickness region of the tested
compact tension specimen by spark wire cutting. For microscopic analysis, a pillar with the dimensions of ~ 25 m × 25 m
× 25 m containing cracks was fabricated from the tested compact tension specimen using an FIB. Then, the 3D crack
propagation behavior was analyzed by FIB-SEM serial sectioning.
The macroscopic crack morphology in the uncharged specimen reconstructed by X-ray computed tomography is presented
in Figure 2. The fracture propagated discontinuously with a
certain area of un-cracked ligaments as indicated by yellow
arrows 3). We have confirmed that the crack propagation tended
to be more continuous with increasing hydrogen content.
Accordingly, we can propose that the macroscopic continuous
crack propagation resulted in the decrease of crack-growth
resistance with increasing hydrogen content.
One typical example of the morphology of hydrogen-related
intergranular crack (HD = 0.42 wt ppm) obtained by FIB-SEM
serial sectioning is presented in Figure 3 (blue: intergranular
crack component, red: transgranular (quasi-cleavage) crack
component). On a microscopic scale, the crack frequently Figure 2 3D crack propagation morphology
deviated from prior austenite grain boundary. The prior austenite reconstructed by X-ray computed tomography in the
grain boundary is the position of the grain boundary in the uncharged specimen.
austenite state before transformation. Because several martensite
variants form by transformation, it exhibits an inhomogeneous
structure. The crystallographic analysis revealed that the
low-angle grain boundary plane segment acted as an obstacle to
crack propagation. Based on the results, we concluded that local
crack-arrestability of low-angle grain boundary plane segment
could be the possible reason for the stable crack propagation even
when the hydrogen content was large.
Acknowledgement
This study was financially supported by JST PRESTO (Grant
Number JPMJPR2096), JSPS KAKENHI (Grant Numbers
JP19H02459 and JP20K21083). In addition, this study is based on
results obtained from a project commissioned by the New Energy
and Industrial Technology Development Organization (NEDO) Figure 3 3D morphology of the hydrogen-related
[JPNP14014]. intergranular crack (HD = 0.42 mass ppm)
reconstructed by FIB-SEM serial sectioning.
The ultrahigh-strength steel sheets are expected to apply for the mobile frame parts due to the wight reduction of
frames, the improvement of collision safety and the protection of fuel cell batteries against collision damages. The tensile
strength of 1500 MPa, the total elongation of more than 20% and the excellent cold formability are required for the
ultrahigh-strength steel sheets. To achieve these mechanical properties and formability, it is expected that the ultrahigh-strength
transformation-induced plasticity (TRIP)-aided steels associated with TRIP of retained austenite are applied for the
automotive frame parts. On the other hand, the occurrence of hydrogen embrittlement which is caused by the absorption of
small amount of hydrogen becomes serious problem in the ultrahigh-strength steel sheets with tensile strength of 1500 MPa.
In this presentation, the hydrogen embrittlement properties of the ultrahigh-strength TRIP-aided steels which are
systematically investigated will be reported.
References
1) T. Hojo, S. Song, K. Sugimoto, A. Nagasaka, S. Ikeda, H. Akamizu and M. Mayuzuki: Tetsu-to-Haganè 90 (2004) pp.177-182.
2) T. Hojo, K. Sugimoto, Y. Mukai, H. Akamizu and S. Ikeda: Tetsu-to-Haganè 92 (2006) pp.83-89.
3) T. Hojo, R. Kikuchi, H. Waki, F. Nishimura, Y. Ukai and E.Akiyama: ISIJ Int. 58 (2018) pp. 751-759.
4) T. Hojo, B. Kumai, M. Koyama, E. Akiyama, H. Waki, H. Saitoh, A. Shiro, R. Yasuda, S. Takahisa, A. Nagasaka: Int. J. Fracture,
224 (2020), pp.253-260
Satoru Kobayashi
Tokyo Institute of Technology
Abstract:
Laves phase, such as Fe2Mo and Fe2W, is often reported to precipitate during creep deformation
within tempered martensitic matrix in conventional high Cr ferritic steels. Different Laves phase
morphologies have been observed in the steels: fine particles along the lath/block boundaries and
coarse globular ones. The effects of Laves phase formation on the creep resistance of ferritic steels are
under debate and have not yet been fully understood.
We have recently found the formation of periodically arrayed rows of very fine Fe2Hf Laves phase
particles in the 9 % Cr ferritic matrix through interphase precipitation along the eutectoid type reaction
of delta-ferrite - gamma-austenite + Fe2Hf and a subsequent gamma-alpha phase transformation.
In the present paper, we will show our recent results on the microstructure formation along the new
route, alloying effect, creep propeties, and alloy design perspective.
Abstract:
The relationship between changes in the magnetic properties and the creep strength with the addition
of 3 or 6 mass% Co was investigated for ferritic steel containing 15 mass% Cr. Co addition up to 6
mass% hardly contributed to solid solution strengthening or precipitation strengthening. However, in
the range of 650 to 750°C, the steel with the larger amount of Co exhibited higher creep strength,
which is explained by a reduction in the diffusion rate associated with a change in magnetic properties
by Co addition. An increase of the magnetization of the steel with increasing Co content in the range
from room temperature to ~800°C was confirmed. Comparing the difference in magnetization and the
ratio of the creep strain rate for steels with different amounts of Co, a clear correlation was found. That
is, at the temperature at which the difference in magnetization reached a maximum, the peak of the
creep strain rate ratio was also observed. This result is explained as follows. In a low temperature
region where the magnetization is large or in a high temperature region above the Curie point of both
steels, the steels exhibit no significant difference in the creep strength. However, at a temperature
where one steel loses its ferromagnetism but the other steel retains it, a significant difference in the
creep strength is observed.
1
Graduate School of Engineering, Tohoku University, Sendai, Miyagi 980-8579, Japan
2
Graduate School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
3
Advanced Technology Research Laboratories, Nippon Steel Corporation, Amagasaki, Hyogo, 660-0891, Japan
*email: sekido@material.tohoku.ac.jp
Introduction
The development of the creep strength-enhanced ferritic (CSEF) steels has led to increased steam temperature and
pressure, and thereby enhanced power generation efficiency in coal-fired power plants 1). Recently, there has been a
growing demand for the development of new high-Cr heat-resistant ferritic steels that can be used at temperatures up to
650°C. Some promising candidates include 9Cr-3W-3Co-V-Nb-B-N steels (MARBN steels) 2), 9Cr-3W-3Co-V-Nb-Ta
-Nd-B steels (SAVE12AD)3), and 0.035C-9Cr-2.4W-1.8Co-V-Nb steels 4).
It is reported that trace B additions (~ 100 wt. ppm) to 9-12%Cr ferritic steels are beneficial for improving the
long-term creep strength 2,5-7). Although the detailed mechanism of this creep strength enhancement by B addition
remains unclear yet, it is believed that boron atoms segregated on austenite grain boundaries (PAGBs) during
austenitizing are partitioned into M23C6 carbides (M23(C,B)6) during tempering and creep deformation, which then
suppresses the coarsening of M23(C,B)6 carbides 7-10). Since the B concentration of M23(C,B)6 carbides precipitated near
PAGBs is higher than that formed in the grain interiors of a 9Cr-3W-B steel 7), the B content in M23C6 can be affected by
the presence of B segregation when they nucleates on the PAGBs. The purpose of this study is to investigate the
evolution of B segregation during tempering by using nano-SIMS.
Experimental procedures
A steel with the composition of Fe-0.081C-0.27Si-0.51Mn-8.96Cr-2.80Co-2.82W-0.20V-0.051Nb-0.013B-0.007N
(wt.%) 11) was used in this study. It should be noted that the present steel falls within the range of composition, where
boron nitride (BN) will not form 12). The steel was austenitized at 1100°C for 4 h, followed by water-quenching (AQ
specimen). The AQ specimens were tempered at 600 and 790°C from 10 min to 8 h, followed by water-quenching. All
the specimens were sealed in quartz tubes under vacuum (< 10-1 Pa) during the austenitizing and tempering treatments.
Microstructures of the steel were characterized by optical microscopy (OM), scanning electron microscopy (SEM),
and transmission electron microscopy (TEM). The acceleration voltage of TEM was 200 kV, and the TEM specimens
were prepared by the focused ion beam (FIB) lift-out method. The spatial distribution of B was visualized by
nano-SIMS (NanoSIMS 50L 13)). The 11B16O2- and 11B12C- ions were used for visualizing the B distribution. Since the
ionization efficiency of 11B16O2- is affected by both B and O, the distribution of O atoms on the surface must be uniform
for precise measurement of B distribution. Therefore, the specimen was pre-sputtered for removing contamination and
surface oxide layers, and left until micro-leak gaseous O2 molecules were adsorbed on the surface 14). An open-source
software, “Open MIMS,” was used for processing the nano-SIMS images 15). Further detailed information on the
nano-SIMS measurement was described elsewhere 16).
the same areas. The specimens show a typical lath martensite microstructure that consists of laths, blocks, and packets
within the prior-austenite grains (PAGs). High intensity of the 11B16O2- ions was depicted on the PAGBs of the (c) AQ
and (d) tempered specimens, suggesting that B segregation has developed on PAGBs during austenitizing and it retains
after the tempering at 790°C for 10 min. The enrichment of B was not observed on the block and packet boundaries in
the AQ specimen because they were newly formed boundaries by the martensite transformation. No precipitates are
observed in the AQ specimen, while some are observed in the tempered specimen. The precipitates formed during
tempering show the high intensity of the 11B16O2- ions as shown in Fig. 1(d). Along with TEM analysis, these
precipitates are found to be B-containing M23C6 carbides; i.e., M23(C,B)6.
Although the present tempering temperature of 790°C was within a typical temperature range of tempering for 9%Cr
steels, M23(C,B)6 precipitation has initiated within a short tempering duration of 10 min. To facilitate the investigation of
the boron segregation evolution and its relation to the carbide precipitation, the tempering temperature was lowered to
600°C in this study.
Figure 2 summarizes EBSD-IPF maps, and the corresponding 11B16O2- and 11B12C maps for the specimen tempered at
600°C for (a,d,g) 10 min, (b,e,h) 2 h, and (c,f,i) 8 h. High intensities of 11B16O2- ions are depicted on PAGBs for all the
specimens shown in Fig. 2, suggesting that the B segregation retains after tempering at 600°C for, at least, 8 h. It should
be noted that the B segregation can be distinguished from M23(C,B)6 by comparing the 11B16O2- and 11B12C- ion maps, i.e.,
B segregation results in high 11B16O2- intensity and low 11B12C- intensity, but M23(C,B)6 yields high intensity for both
11 16
B O2- and 11B12C- 16). Although some M23(C,B)6 carbides have formed at the grain interiors in the specimens tempered
at 600°C for 10 min and 2 h, few were confirmed on the PAGBs. Besides, the block boundaries near the PAGBs exhibit
high 11B16O2- signals, suggesting that B atoms segregated at PAGBs have oozed through the block boundaries during
tempering.
The present study indicated that, when tempered at 600°C, the amount of M23(C,B)6 carbides formed on the PAGBs
was fewer than those on block and packet boundaries. In addition, the M23(C,B)6 carbides formed on block and packet
boundaries were coarser than those formed on PAGBs. This tendency was supported by OM and SEM observations as
shown in Fig. 3. Here, the specimen tempered at 790°C for 10 min show a typical tempered martensite microstructure
(Fig. 3(a)), however, the specimen tempered at 600°C for 8 h shows a precipitation-free zone (PFZ) like microstructure,
where the PAGBs were enclosed by bands with a brighter contrast, as shown in Fig.3(b). The brighter contrast on
PAGBs is originated from the difference in the number of carbide precipitates. That is, the volume fraction of
precipitates near PAGBs is much lower than that at the grain interior, as depicted by the SEM micrograph shown in Fig.
3(c), so the etching rate difference by the Villela reagent between the area near PAGBs and the grain interior causes
some contrast in OM micrographs.
It has been demonstrated that the size of M23C6 precipitates in tempered Gr. 91 steels (without B additions) becomes
larger in the order of the ones formed at PAGBs, packet boundaries, block boundaries, and lath boundaries 17). This
indicates that nucleation and/or growth of M23C6 precipitates are kinetically faster on the PAGBs of the Gr. 91 steels.
However, this tendency appears to be the opposite in the present steel; i.e., M23C6 precipitation was not observed on the
PAGBs when tempered at 600°C for 8 h. This suggests that the B segregation suppresses precipitation and growth of
Fig. 3: OM micrographs of the specimens tempered at (a) 790°C for 10 min and (b) 600°C for 8 h, and (c)
a magnified SEM image of (b).
M23C6 carbides in the vicinity of PAGBs. This consideration is supported by the previous studies, where B addition
results in the reduced coarsening rate of M23C6, but this is only evident in the vicinity of PAGBs 2,10).
The mechanisms for the suppression of carbide precipitation and growth by B segregation are not clear, but it is
speculated that B segregation results in the reduction of other elemental segregation or diffusivity on grain boundaries. It
has been demonstrated that the segregation of solute atoms with weaker segregation energy is reduced by the presence
of the other solute atoms with stronger segregation energy because the segregation sites are preferentially occupied by
the latter (site competition) 18-20). Since the segregation energy of B at austenite grain boundaries is higher than that of C
Conclusions
The evolution of B segregation during tempering was investigated using nano-SIMS in a B-doped 9Cr steel. In the
specimen quenched from the austenitizing temperature, B segregation was confirmed on the prior-austenite grain
boundaries (PAGBs) but not on block and packet boundaries. Carbide precipitation was not confirmed at the early stage
of tempering at 600°C. Although precipitation of M23C6 was confirmed to occur at the grain interior of the tempered
specimens, the precipitation and/or growth of M23C6 carbide were suppressed near PAGBs. It is suggested that the B
segregation on PAGBs is effective to suppress the carbide precipitation and growth during tempering at 600°C.
Acknowledgments
This work was supported by ISIJ Research Promotion Grant, Iketani Science and Technology Foundation Research
Grant, JSPS KAKENHI (JP18K18930), and the University of Tokyo Advanced Characterization Nanotechnology
Platform, as a program of “Nanotechnology Platform” of the MEXT Japan (JPMXP09A19UT0286).
References
1) F. Masuyama: ISIJ International 41 (2001) 612-625.
2) F. Abe: STAM 9 (2008) 013002.
3) A. Iseda, M. Yoshizawa, H. Okada, T. Hamaguchi, H. Hirata, K. Joutoku, T. Ono, K. Tanaka, The Thermal and
Nuclear Power Generation Convention Collected Works 12 (2016) 49-55.
4) K. Asakura, T. Koseki, T. Sato, M. Arai, T. Horiuchi, K. Tamura, T. Fujita: ISIJ International 52 (2012) 902-909.
5) N. Takahashi, T. Fujita, T. Yamada, Tetsu-to-Hagané 61 (1975) 2263-2273.
6) T. Azuma, K. Miki, Y. Tanaka, T. Ishiguro, Tetsu-to-Hagané 88 (2002) 678-685.
7) T. Horiuchi, M. Igarashi, F. Abe: ISIJ International 42 (2002) S67-S71.
8) R. Sahara, T. Matsunaga, H. Hongo, M. Tabuchi: Metall. Mater. Trans. A 47 (2016) 2487-2497.
9) K. Maruyama, K. Sawada, J. Koike: ISIJ International 41 (2001) 641-653.
10) H. Semba, F. Abe: Energy Materials 1 (2006) 238-244.
11) M. Yonemura, Y. Hasegawa, M. Igarashi, M. Tabuchi, F. Abe: Proc. 7th Intnl. Conf. Advances in Materials
Technology for Fossil Power Plants (EPRI 2013), ASM International, 2014, pp.1329-1340.
12) K. Sakuraya, H. Okada, F. Abe: Energy 1 (2006) 159.
13) M. Takeuchi, Plant Morphol. 26 (2014) 19-23.
14) L. Sangely, B. Boyer, E. de Chambost, N. Valle, J.-N. Audinot, T. Ireland, M. Wiedenbeck, J. Aleon, H. Jungnickel,
J.-P. Barnes, P. Bienvenu, U. Breuer, "Sector Field Mass Spectrometry for Elemental and Isotopic Analysis", The
Royal Society of Chemistry, Cambridge, 2015, pp.439-499.
15) C. Poczatek, Z. Kaufman, C. Lechene, "Open MIMS Image-J Plugin Guide", National Resource for Imaging Mass
Spectrometry (2012).
16) T. Osanai, N. Sekido, M. Yonemura, K. Maruyama, M. Takeuchi, K. Yoshimi: Materials Characterization 177
(2021) 111192.
17) T. Okano, T. Ito, M. Mitsuhara, M. Nishida: Proc. 1st Intnl. Conf. on Advanced High-Temperature Materials
Technology for Sustainable and Reliable Power Engineering (123HiMAT-2015), 2015, pp.130-133.
18) S. Suzuki, M. Obata, K. Abiko, H. Kimura: Scripta Metallurgica 17 (1983) 1325-1328.
19) C. M. Liu, T. Nagoya, K. Abiko, H. Kimura: Metallurgical Transactions A 23 (1992) 263-269.
20) M. Paju, H. Viefhaus, H. J. Grabke: Steel Research 59 (1988) 336-343.
21) L. Karlsson, H. Nordén: Acta Metallurgica 36 (1988) 13-24.
Keywords: creep strength, ferrite matrix, ferritic heat-resistant steels, Laves phase, M23C6 carbide, steam oxidation resistance
The steel was melted in a vacuum induction furnace. An ingot weighing 10 kg was hot-forged into square bars with
20-mm long sides and solution-treated for 0.5 h at 1473 K, followed by water quenching. The creep test specimens were
round-bar test pieces with a gauge diameter of 6 mm and a gauge length of 30 mm. The creep tests were conducted at 923,
973, and 1023 K at stresses ranging from 30 to 240 MPa for up to about 65 000 h in air. The microstructure of the
creep-ruptured specimens was observed with a cold cathode field-emission type scanning electron microscope.
The specimens used for the steam oxidation tests were cut from the water-quenched bars into coupons of about 10
mm 20 mm 2 mm by electrical discharge machining, and then finished with 600 grit abrasive paper. The specimens
were washed with methanol in an ultrasonic bath and then placed in ceramic crucibles, which were carefully set in the furnace
chambers for a steam oxidation test. The air in chambers were replaced with argon gas with a dew point of about 210 K and
including 0.3 ppm oxygen. Then, the chambers were heated to the test temperatures in about 3 or 4 h. When the chamber
temperature exceeded 473 K, steam was introduced to replace the argon gas. The flow rate of the steam was controlled at
about 240 ml/h by controlling the water supplied to the steam generator. The oxygen concentration of the water in the tanks
was maintained at below 30 ppb by argon bubbling, and its electrical conductivity was controlled at below 10.0 μS/m by
passing it through ion exchange resins. The steam oxidation duration was defined as the time period between reaching the test
temperatures and shutting down the power to the furnace chambers. The weight gains of the specimens were measured with a
microbalance for various oxidation durations up to 3 000 h.
Results
Figure 1 shows creep stress versus time to rupture curves for 15Cr steels (solid curves) and conventional 9Cr ferrit-
ic heat-resistant steel Grade T92 11) (dashed curves) at 923 K (circles), 973 K (triangles), and 1023 K (squares). The Grade T92
steel possesses the highest creep strength of practical ferritic heat-resistant steels. The creep strength of the 15Cr steels is about
twice that of T92 steel at every temperature. The creep rupture lives of the steels are approximately one hundred times longer
than those of T92 steel at 923 and 973 K, and from ten to one hundred times longer than those of T92 steel at 1023 K. By ex-
trapolating the curves in Fig. 1, the creep rupture strength of the 15Cr steels after 100 000 h can thus be estimated to be 130
MPa at 923 K and 45 MPa at 973 K. This strength level is double that of conventional ferritic heat-resistant steels, and com-
parable to that of conventional austenitic steel 12).
Figure 2 (a) shows the scanning electron micrograph in the vicinity of a triple junction of grain boundaries for the
15Cr steel creep-ruptured at 973 K-80 MPa after 23 592 h. The blocks with white and light gray contrast as shown by the
white arrows in the Fig. 2 (a) were identified as Laves phase and M23C6 carbide, respectively. Most precipitates formed on
grain boundaries, and few precipitates were observed within grains. Then, it is supposed that a precipitation strengthening
effect on grain boundaries contributed to the high creep strength of the 15Cr steel under this condition as shown in Fig. 1. Fig-
ure 2 (b) is an enlarged figure of the grain boundary shown in Fig. 2 (a). Some small voids were observed in the interfaces
between ferrite matrix and M23C6 carbide as shown by the white arrows in Fig. 2 (b), and few voids were formed around the
Laves phase. The details of these voids’ formation mechanism are unclear. The large incoherency of the lattice structure be-
tween the ferrite phase and M23C6 carbide is supposed to be the reason for the voids’ formation. These voids could be an origin
of creep rupture, and we believe that it is necessary to reduce the coarsened M23C6 carbide. We are measuring the creep
strength of 15Cr steels with a lower carbon content than 0.05 mass%.
Fig. 2 The scanning electron micrographs of the 15Cr steel creep-ruptured at 973 K-80 MPa after 23 592 h. (a)
shows the microstructure in the vicinity of a triple junction of grain boundaries, and (b) is the enlarged figure of
the grain boundary in (a).
Figure 3 shows the mass gain of the 15Cr steel (solid curves) and conventional 9Cr ferritic heat-resistant steel
Grade T91 (dashed curves) at 923 K (circles), 973 K (triangles), and 1023 K (squares) for different steam oxidation durations.
At 923 K, the mass gain of the 15Cr steel after 3 000 h is about 60% that of the T91 steel. At temperatures higher than 973 K,
the weight of the T91 steels continues to increase like power math functions. By contrast, the mass gains of the 15Cr steels are
relatively small at 973 and 1023 K. The differences between the mass gains of 15Cr steels and T91 steels are very significant
at elevated temperatures. In addition, higher oxidation resistance than that of conventional ferritic steels was obtained by a
ferrite matrix with a higher chromium content than that of the conventional one.
Conclusions
We have proposed precipitation strengthened 15Cr ferritic steels that have a solution-treated ferrite matrix and that
are strengthened by the precipitation of intermetallic compounds of the Fe2W Laves phase to replace conventional 9Cr ferritic
heat-resistant steels with a tempered martensite phase. The creep strength of the 15Cr steels is twice that of conventional steels.
Also, the steam oxidation resistance of the 15Cr steels is higher than that of conventional steels. We then successfully im-
Fig. 3 Mass gain of 15Cr steels (solid curves) and conventional ferritic heat-resistant steel Grade T91 (dashed
curves) at 923 K (circles), 973 K (triangles), and 1023 K (squares) for various steam oxidation durations.
Acknowledgement
A part of this study was supported by Grants-in-Aid for Research (KAKENHI) of the Japan Society for the Promotion of Science
(Grant Number 17K06581). We would like to thank Dr. M. Auchi at National Institute for Materials Science (Present address: Kyu-
shu University) for his measuring the steam oxidation resistance.
References
1) M. Fukuda: Proceedings of the 9th Liege Conference: Materials for Advanced Power Engineering, (2010) pp. 5-19.
2) H. Tschaffon: Proceedings of the 9th Liege Conference: Materials for Advanced Power Engineering, (2010) pp. 20-28.
3) Y. F. Yin and R. G. Faulkner: Physical and elastic behavior of creep-resistant steel: From Creep-resistant steels, ed. by F. Abe,
T.-U. Kern and R. Viswanathan, (Woodhead Publishing Limited, 2008) pp. 217-240.
4) F. Abe: Strengthening mechanisms in steel for creep and creep rupture: From Creep-resistant steels, ed. by F. Abe, T.-U. Kern and
R. Viswanathan, (Woodhead Publishing Limited, 2008) pp. 279-304.
5) K. Kimura, K. Seki, Y. Toda and F. Abe: ISIJ Int., 41 (2001) S121-S125.
6) Y. Toda, K. Seki, K. Kimura and F. Abe: ISIJ Int., 43 (2003) 112-118.
7) Y. Toda, H. Tohyama, H. Kushima, K. Kimura and F. Abe: JSME Int. J., A48 (2005) 35-40.
8) K. Kimura, Y. Toda, H. Kushima and K. Sawada: Int. J. Pressure Vessels and Piping, 87 (2010) 282-288.
9) M. Shibuya, Y. Toda, K. Sawada, H. Kushima, K. Kimura: Mater. Sci. Eng., A652 (2016) 1-6.
10) Japan Society of Mechanical Engineers: Rule on Thermal Power Generation Facilities (2000).
11) National Institute for Materials Science: Data sheets on the elevated-temperature properties of 9Cr-0.5Mo-1.8W-V-Nb steel tube
for power boilers (ASME SA-213/SA-213M Grade T92) and 9Cr-0.5Mo-1.8W-V-Nb steel pipe for high temperature service
(ASME SA-335/SA-335M Grade P92) (2002).
12) National Research Institute of Metals (present address: National Institute for Materials Science): Data sheets on the elevat-
ed-temperature properties of 18Cr-12Ni-Mo stainless steel tubes for boilers and heat exchangers (SUS 316H TB) (2000).
ASME Grade 91 steel has been used as structural materials at ultra-super-critical (USC) plants and has
considerably contributed to the improved efficiency of plants. However, the decrease in the creep strength of the steel with
increasing time (long-term creep strength) and the decrease in strength at welded joints have become an issue for maintenance
of USC plants. To solve the problem, the long-term creep strength has been evaluated by committees in Japan, and creep life
curves as well as allowable tensile stress and the weld strength reduction factor were changed on the bases of the evaluation
results. However, the number of data obtained in long-term creep tests was not sufficient and it was pointed out that the
long-term creep strength should be continuously examined. Therefore, we have conducted studies on microstructure and
creep strength of ASME Grade 91 steel in a region exceeding 100,000 hours, utilizing materials used at USC plants for a long
time because in practice it is difficult to conduct creep tests exceeding 100,000 hours.
Six straight and three elbow pipes that had been used as hot reheat pipes (HRP) were used as the materials. For
elbow pipes, the creep strength on the intrados side, where the stress is higher than that on the extrados side, was basically
examined. The hoop stresses applied to the straight and elbow pipes were calculated using the equation for the mean diameter
with the dimensions of the pipes and the steam pressure. Table 1 shows a summary of the temperature, hoop stress, and
operation time for each material. Table 2 shows chemical compositions of the materials used, where a product form of those is
plate (KA-SCMV28).
500 500
1)
Assessment in 2015 Plant Assessment in 2015 1) Plant
Average property ○ A Average property ○ A
300 - - - 99% lower limit 300
□ B - - - 99% lower limit □ B
+ C1 + C1
Stress (MPa)
Stress (MPa)
◇ G1 ◇ G1
⊿ L ⊿ L
△ N △ N
100 ▽ Q 100 ▽ Q
S S
⊿
⊿
80 80
60 60
40 Rupture location
40 Rupture location
Black : Weld metal Black : Weld metal
Red : HAZ Red : HAZ
20 1 20 1
10 102 103 104 105 106 10 102 103 104 105 106
Time to rupture (h) Time to rupture (h)
(a) 600oC (b) 650oC
Figure. 1 Creep test results on welded joint of long-term used materials.
Then, remaining creep lives of the materials for operating condition were calculated on a fitting curve using
Larson-Miller parameter. Total creep life of the material, which means a creep life at initial state, was presumed to be a
summation of the service time at the plants and the remaining creep life. The estimation was conducted for welded joints used
at five plants for long-term, and all results lay within 98% confidential band (between 99% lower limit and 99% upper limit)
by the creep life evaluation curve 1) as shown in Fig.2, while a significant heat-to-heat variation of creep strength was found
even in the region exceeding 100,000 hours. Creep tests on base metals related to each welded joint were also conducted, and
the estimation results of the total creep life of the base metals were compared to those of the welded joints. Figure 3 shows a
relationship between the creep property of welded joint and the base metal. The abscissa is the value obtained by dividing the
estimated initial creep life of the base metal by the value obtained from the curve used to estimate the creep life of the base
metal (plate) in the 2015 assessment 2) (normalized creep life of the base metal). The ordinate is the value obtained by dividing
the estimated initial creep life of welded joint under the operating conditions of an actual plant (temperature, hoop stress) by
the value obtained from the curve in the 2015 assessment 1) (normalized creep life of welded joint). Although the number of
data is limited, there is a positive correlation between the normalized creep life of the base metal and that of the welded joint
under the conditions examined in this study. Namely, the initial creep property of the welded joint depends on that of the base
metal; in other words, as the initial creep strength of the base metal increases, the initial creep strength of the welded joint also
increases. The above findings indicate the possibility of estimating the creep property of welded joint if the creep property of
the base metal of pipes can be evaluated by a certain method.
Stress (MPa)
100
80
60
40 Present
Data (10) estimation Temperature
△ 550℃
○ □ 600℃
▽ 650℃
20 1 2 3
10 10 104 105 10 106
Time to rupture (h)
Figure. 2 Estimation results in comparison with creep data and master curves on welded joint of Grade 91 steel 1).
10.0
Normalized creep life property
of welded joint
1.0
0.1
0.1 1.0 10.0
Normalized creep life property
of base metal
Figure. 3 Relationship between creep property of welded joint and that of base metal.
Microstructural analyses were conducted using spherical aberration-corrected TEM. Because Grade 91 steel has a
multilevel structure consisting of prior austenite grain boundaries, packet/block boundaries and laths, the precipitations densities
significantly vary with the observation site. Figure 4 shows images of high-strength and low-strength base metals. As Cr-type
precipitations (mostly M23C6) and Mo-type precipitations (mostly the Laves phase) are irregularly distributed, the condition of
the precipitations appears to be completely different depending on the observation site. To reduce an influence of differences in
the observation site on the analysis results, an observation area is specified for each precipitation, ≤380 μm2 for the Cr-type and
the Mo-type and ≤65 μm2 for the V-type. Then, we have obtained results that there is no correlation between Cr-type and the
normalized creep property in base metal, and no correlation between Mo-type and the normalized creep property in base metal,
but a positive correlation is observed between the densities of V-type precipitations and the normalized creep property in base
metal, which suggests that creep strength of the material is improved due to pinning of dislocations by V-type precipitations.
Figure 5 shows the relationship between the normalized creep property of Grade 91 steel base metal and the densities of V-type
precipitations obtained from the analyses of the observation area specified above. The results indicate the possibility of
estimating the creep property of the material by determining the densities of V-type precipitations.
1.5
Normalized creep property of base metal
0.5
0
0.0 5.0 10.0 15.0
Density of precipitations (mm-2)
Fig. 5 Relationship between normalized creep property and number density of V-type precipitations in Grade 91 steel base
metal.
Here are summaries of the present study. The total creep lives of Grade 91 steel welded joint, which was presumed
to be the sum of the service time at the plants and the remaining creep life, lay within the 98% confidential band of the creep life
evaluation curve of the material adopted in Japan, while a significant heat-to-heat variation of creep strength was found even in
the region exceeding 100,000 hours. The comparison of the total creep life property of the welded joint with that of base metal
suggested that the heat-to-heat variation of the welded joint depends on the creep life property of the corresponding base metal.
From microstructural analysis on the material, it was found that there was a strong correlation between the density of V-type
precipitation and the creep life property of the base metal.
References
1) Yaguchi, M., Nakamura, K. and Nakahashi, S., “Re-evaluation of Long-term Creep Strength of Welded Joint of ASME Grade 91 Type
Steel”, Proc. PVP2016, Vancouver, Canada, PVP2016-63316, (2016).
2) Kimura, K. and Yaguchi, M., “Re-evaluation of Long-term Creep Strength of Base Metal of ASME Grade 91 Type Steel”, Proc.
PVP2016, Vancouver, Canada, PVP2016-63355, (2016).
Abstract:
The recovery behavior of lath martensite microstructure during creep in 9Cr ferritic heat-resistant
steels is controlled by the M23C6 carbides on grain boundaries. So, the growth behavior of M23C6
carbides should be essentially taken into account to completely understand the creep deformation
mechanism of the steels. In this presentation, we focused on the effect of grain boundary character,
stress, and pre-tempering on the growth behaviors of M23C6 carbides in the lath martensite
microstructure. There are four types of boundaries in the lath martensite, those are the prior-austenite
grain boundary (PAGB), the packet boundary, the block boundary, and the lath boundary. We clarified
that the growth rates of M23C6 carbides during both aging and creep strongly depended on the
boundary character. The growth rate of carbides on the PAGB was most rapid, and that on the lath
boundary was slowest. The stress applying obviously accelerated the growth rate of carbides on the
PAGB and the packet, though that on the block and lath boundaries were apparently insensitive. It
means that the growth of carbides was controlled by not only atomic diffusion but also the
microstructural factors. In the presentation, we will show the detail analysis of crystal orientation
relationship (COR) between M23C6 carbides and matrix and discuss the effect of deviation of COR
by the creep on the growth rate of carbides..
Fig. 1 Creep rate vs time curves of 18Cr–9Ni–3Cu–Nb–N Fig. 2 Relationship between stress and time to rupture of
steel at 873 K.1) 18Cr–9Ni–3Cu–Nb–N steel at 873–1073 K.2)
Fig. 4 Change in the average KAM value Fig. 5 Coverage ratio of the precipitate on grain boundary
as a function of time.1) as a function of time.1)
4. Conclusions
This study focused on the effect of microstructure change on the creep behavior. Identification of the
microstructural factors that cause double local minima in creep rate suggested that GBPS by M23C6 phase and precipitation
strengthening by Cu phase are effective to reduce the creep rate. However, 3D observation of the long-term aged specimen
clarified that the contributions of GBPS by M23C6 and precipitation strengthening by Cu phase are locally reduced around the
σ phase, result in the premature failure in the longer-term side.
Acknowledgement
The authors thank Ms. Yuka Hara (NIMS) and Ms. Akiko Nakamura (NIMS) for their support with the microstructural
observations. This work was performed under the support of JSPS KAKENHI (No. JP20K22474).
References
1) T. Hatakeyama, K. Sawada, K. Sekido, T. Hara and K. Kimura: Mater. Sci. Eng. A, 831 (2021) 142225.
2) NIMS Creep Data Sheet, Metallographic Atlas of Long-Term Crept Materials, No. M-11 (2016).
3) KA-SUS304J1 HTB, The material specification in the Technical Regulation of Equipment for Thermal Power Plant, 2007.
4) T. Hara, K. Tsuchiya, K. Tsuzaki, X. Man, T. Asahata, and A. Umeto: J. Alloys Compd., 577S (2013) S717–S721.
5) I. Arganda-Carreras, V. Kaynig, C. Rueden, K. W. Eliceiri, J. Schindelin, A. Cardona, H. S. Seung: Bioinformatics, 33 (2017)
2424–2426.
6) J. Schindelin, I. Arganda-Carreras, E. Frise et al.: Nature methods, 9 (2012) 676–682.
7) T. Hatakeyama, K. Sawada, T. Hara K. Sekido, and K. Kimura: Scr. Mater., 200 (2021) 113904.
Keywords: Super 304H, Mechanical property, Microstructure, Aged material, Boiler tube
Introduction
Super 304H steel (18Cr-9Ni-3Cu-Nb-N, ASME Code Case 2328) has been widely used as superheater (SH) and
reheater (RH) boiler tubes in ultra-supercritical (USC) power plants due to its excellent creep strength, oxidation/hot corrosion
resistance and weldability 1). This steel was developed to have superior creep strength to that of SUS 304H steel by the
addition of copper, niobium and nitrogen. The superior creep strength of Super 304H steel is obtained by the formation of
finely distributed coherent copper phase, Nb (C, N), M23C6 and NbCrN particles, which results in higher allowable stress than
other austenitic stainless steels used in boiler tubes, such as SUS 347HTB and SUS321HTB. The steel has been applied to
actual plants for more than 20 years, and several steam leakage accidents have occurred in the superheaters with the operating
time more than 100,000 hours in Japan. Recent studies have indicated that times to creep rupture in the longer term are shorter
than those predicted by straight extrapolation of the short-term data points to the lower stresses or to the lower temperatures,
and the long-term creep strength of Super 304H may be lower than originally assumed 2, 3). This may result from the
difference in local deformation at the interface between matrix and σ phase that precipitates in creep exposure, which can lead
to crack initiation easily reported by Kubushiro 4) and Kimura 5) et al. It is important to understand the long-term
microstructural evolution and stability, and one way to investigate the microstructural evolution is to study service exposed
materials used in actual plants.
In this study, microstructures and mechanical properties of several long-term used Super 304H steel boiler tubes
were presented. The aged superheater tubes that have been in service for about 140,000 hours at the approximate metal
temperature ranged from 550°C to 640°C, have been investigated. The microstructures were observed by optical, scanning
and transmission electron microscopies including chemical analysis techniques. In addition, the hardness, impact value and
tensile properties of aged materials were measured, and these properties were discussed in relation with the microstructures.
10
55 15 10 30 10 15
7.5
5~7.5
σ phase σ phase
σ phase
M 23C6
Z phase
σ phase
σ phase
10µm 2µm 500nm
Flame
Note: Average value is
shown for D2, because side
flame side is unknown.
Rear
side
N D2 X N D2 X
Figure 4 Area fraction of σ phase at the middle Figure 5 Mean and the maximum diameters of Cu
part in thickness direction of aged materials phase of aged materials
Figure 4 compares area fraction of σ phase at the middle part in thickness direction of aged materials. The total area
and the area at grain boundary are shown to distinguish the amount of σ phase at the grain boundary and ingrain. The area
fraction of σ phase increased in the order of N, D2 and flame side of X, and the value on the rear side was smaller than that on
the flame side. This is due to the difference in metal temperature of each material. That is, the N plant has the lowest operating
temperature, the X plant has the highest operating temperature as shown in Table 1, and the metal temperature on the rear side
is lower than that on the flame side.
Figure 5 shows mean and the maximum diameters of Cu phase of aged materials. Note that the value of X plant is
measured on the flame side. Both mean and the maximum diameters increased in the order of N, D2 and X as well as the area
fraction of σ phase. However, the maximum value does not necessarily depend only on temperature since Cu phase
preferentially precipitates around σ phase as shown in Fig. 2(b).
Figure 6 presents hardness distributions in thickness direction of new and aged tubes. The hardness on the outer
surface side of new tube were higher than the rest part, possibly due to cold calibration during the tube manufacturing process.
The hardness on the outer surface side of aged tubes were comparable to those of the new tube, indicating that the effect of the
cold calibration remains. In addition, the hardness on the inner surface side of D2 and X were also higher, which may be due
to work hardening from the shot peening process. On the other hand, the hardness at the middle part in thickness direction of
all the aged tubes were higher than those of the new tube, approximately 200 HV, and no difference was confirmed depending
on the metal temperature of aged tube.
Figure 7 shows average values of impact energy of new and aged materials. The impact energy decreased
significantly due to long-term exposure, to about 60% of new material for the N plant and about 20% for the D2 and X plants.
The impact energy on the flame side was smaller than that on the rear side. This means that the impact energy decreases with
increasing the metal temperature of aged material. The change in impact energy corresponds to the microstructural evolution
of σ phase shown in Fig. 4. That is, material with larger amount of σ phase in Fig. 4 had the smaller impact energy in Fig. 7.
New N D2 X
Figure 6 Hardness distributions of all materials Figure 7 Impact energy of all materials
Requirement value
at R.T
New N D2 X
Figure 8 Stress-strain curves of all materials at 600°C Figure 9 0.2% proof stresses of all materials
Figure 8 shows stress-strain curves of new and aged materials at 600°C as a representative case. Ultimate tensile
strength of the N plant was larger than that of new material while those of the D2 and X plants were smaller. Rupture
elongations of all the aged materials, on the other hand, were smaller than that of new material. Similar to the impact energy,
the rupture elongations of the D2 and X plants were smaller than that of the N plant, which is also due to the difference in the
amount of σ phase.
Figure 9 compares 0.2% proof stress (σ0.2) of new and aged materials at R.T, 600°C and 650°C. The 0.2% proof
stresses of the aged materials at each temperature were larger than those of new material. This is resulted from precipitation
strengthening by fine M23C6, Cu phase, Nb (C, N) and Z phase precipitated during long-term exposure.
Conclusions
Microstructures and mechanical properties of several long-term used Super 304H steel tubes were investigated.
Several typical phases were observed in aged materials such as M23C6, Cu phase, Nb (C, N), Z phase and σ phase. σ phases
were precipitated not only at grain boundaries but also within grains in the aged materials with high operating temperature.
The hardness at the middle part of tube thickness of aged materials were larger than those of new material as well
as the 0.2% proof stress. The elongation and impact energy, on the other hand, were significantly lower than those of the new
material, and the reduction increased with increasing the metal temperature of the aged material. The change in impact energy
and elongation corresponds to the microstructural evolution of σ phase.
References
1) H. Semba, Y. Sawaragi and K. Ogawa, Materia Japan, 48-2 (2007) pp. 99-101.
2) NIMS Creep Data Sheet, No. 56A, (2018).
3) K. Sawada, K. Sekido, M. Murata, K. Kamihira and K. Kimura, Materials Characterization, 141 (2018) pp. 279-285.
4) K. Kubushiro, Y. Shioda and K. Nomura, Journal of the Society of Materials Science, Japan, 66-2 (2017) pp. 179-184.
5) T. Kimura, Y. Shioda and K. Kubushiro, Journal of the Society of Materials Science, Japan, 68-2 (2019) pp. 142-147.
6) JIS G 0567: 2012, (2012).
High performance low alloying structural steel has many applications in humanity activities. Increasing the strength
of steel is a trend for reduce structural weight, save cost and declining environment load. However, with the increasing the
strength, toughness always degenerating, especially for the heavy gauge plate. For the safety of the high-rise building, long
span bridge, large container ship and long-distance pipeline etc., excellent lower temperature toughness is one of the key
challenges of the metallurgist.
Low carbon bainitic and/or martensitic phases are the dominant microstructure to achieve high strength and high
toughness. Conventionally, prior austenite grain, packet, block and variant boundaries are the crystallographic microstructure
of bainite and/or martensite. Definitely, chemical elements (C and Ni), recrystallization (prior austenite grain size), and cooling
rate dramatically influence the phase transformation of bainite and/or martensite. Although, the prior austenite grain size and
the cooling rate can be controlled properly for the thin plate, it is obviously very difficult to control the homogeneity
microstructure of throughout cross section in a heavy gauge plate.
The quantitative characterization of bainite and/or martensite microstructure is based on parents and daughters (K-S
or N-W) relationship. By the computational materials engineering approach, homemade software and model have been
developed to digitalization the variants and boundaries in bainite and/or martensite 1,2).
To simulating the heavy gauge plate, Jominy end quenching test has been carried out to reveal the effects of
composition, prior-austenite grain size and cooling rate on the variant selection and boundary characteristics. In this
presentation, the crystallographic characteristic of boundaries of low carbon bainitic and/or martensitic microstructure will be
investigated. There will be three parts in the presentation:
(1) Effect of carbon and Ni content on the variant selections 3,4)
(2) Effect of prior austenite grain size refinement on the variant selection 5)
(3) Effect of kinetic and thermodynamic on the variant selection 6)
References
1)Wu B B, Wang Z Q, Wang X L, et al. Toughening of martensite matrix in high strength low alloy steel: Regulation of variant pairs.
Materials Science and Engineering: A, 2019, 759: 430-436.
2)Li X C, Zhao J X, Cong J H, et al. Machine learning guided automatic recognition of crystal boundaries in bainitic/martensitic
alloy and relationship between boundary types and ductile-to-brittle transition behavior[J]. Journal of Materials Science &
Technology, 2021, 84: 49-58.
3)Yu Y S, Wang Z Q, Wu B B, Zhao J X, et al. Tailoring variant pairing to enhance impact toughness in high-strength low-alloy
steels via trace carbon addition, Acta Metallurgica Sinica (English Letters), 2021, 34: 755-764.
4)Huang S, Yu Y S, Wang Z Q, et al. Crystallographic insights into the role of nickel on hardenability of wear-resistant steels,
Materials Letters, 2022, 306: 130961.
5)Wu B B, Wang X L, Wang Z Q, et al. New insights from crystallography into the effect of refining prior austenite grain size on
transformation phenomenon and consequent mechanical properties of ultra-high strength low alloy steel. Materials Science and
Engineering: A, 2019, 745: 126-136.
6)Wu B B, Wang Z Q, Yu Y S, et al. Thermodynamic basis of twin-related variant pair in high strength low alloy steel. Scripta
Materialia, 2019, 170: 43-47.
Keywords: Scanning transmission electron microscopy, differential phase contrast, electromagnetic field, steels, magnetic materials
Acknowledgement
The author thanks all the collaborators of this research, especially Y. Kohno (JEOL Ltd.), T. Seki, T. Matsumoto, Y. Ikuhara (U.
Tokyo), S.D. Findlay (Monash U.) for their contribution to the works shown in this presentation. The author acknowledges support
from the JST SENTAN Grant Number JPMJSN14A and the JSPS KAKENHI Grant numbers 20H05659, 19H05788, 17H06094.
References
1) S. Morishita et al., Microscopy, 67, 46 (2018).
2) N. Shibata et al., Nature Comm. 10, 2380 (2019).
3) N. Shibata et al., Nature Phys., 8, 611-615 (2012).
4) N. Shibata et al., Nature Comm. 8, 15631 (2017).
5) Y. Kohno et al., Nature 602, 234-239 (2022).
1. INTRODUCTION
High strength low alloy (HSLA) steels have a good combination of strength and formability, so they have many applications
in the building, energy transportation, and automotive industries. In HSLA steels, microalloying elements such as Ti, V and
Nb are used to control microstructure and mechanical properties. In particular, Ti and Nb are used to control the austenite
grain size or to induce precipitation strengthening. The improvement effects of Ti and Nb carbides in ferrite on strength have
been extensively studied 1-3). According to one recent study, TiC precipitates appeared after the Ti-clustering process, and it
was confirmed that the strengthening effect of Ti clustering was more effective than that of TiC type precipitates 4). Although
many theoretical studies have been conducted on the formation of MC (M=Ti, Nb) type carbide and on interfacial properties 5),
studies on clustering in the early stages of the precipitation process have rarely been reported. In this study, first principles
calculations were used to analyze the generation of Ti and Nb clusters in the ferrite matrix. In addition, the process of
changing into MC type precipitates from cluster was analyzed, and this was experimentally confirmed by observing Nb
microalloyed steel using high resolution transmission electron microscopy.
2. EXPERIMENTAL
First-principles modeling was employed to estimate the formation energy and the lattice misfit at the ferrite/cluster interface.
The spin-polarized Vienna ab-initio simulation package (VASP) was used in projector augmented wave (PAW) formalism.
To calculate the lattice parameter and formation energy of the bcc M (M=Ti, Nb) with carbon, the atomic structure of a 2x2x2
supercell of conventional bcc unit cell was used. It was assumed that the carbon atom was located at the octahedral interstitial
site of the bcc lattice, and the arrangement of additional carbon was assumed to be at the site with the greatest distance from
the nearest atom among 16 octahedral interstitial sites. The internal atomic positions and lattice parameters were fully relaxed
using the total energy and force minimization scheme with the gradient quasi-Newton method. We assumed that convergence
was achieved when the forces acting on the atoms did not exceed 0.01 eV/ Å.
Lattice parameter / A
4.0
Nb c-direction
Nb a-direction
3.6 Ti c-direction
Ti a-direction
3.2
a
2.8
0.0 0.5 1.0
Site fraction of C
References
1) J. H. Jang, C.-H. Lee, Y.-U. Heo, D.-W. Suh: Acta. Mater. 60 (2012) 208-217
2) C. Y. Chen, C. C. Chen, J. R. Yang: Mater. Char. 88 (2014) 69-79
3) S. Jukherjee, I. Timokhina, C. Zhu, S. P. Ringer, P. D. Hodgson: J. Alloys Compd, 690 (2017) 621-632
4) J. Wang, M. Weyland, I. Bikmukhametove, M. K. Miller, P. D. Hodgson, I. Timokhina: Scr. Mater. 160 (2019) 53-57
5) S. Lu, J. Agren, L. Vitos: Acta Mater. 156 (2018) 20-30
Keywords: martensitic transformation, neutron diffraction, phase stress, lattice parameter, Fe-18Ni alloy
Austenite to martensite transformation in steels is a volume expansion phase transformation. Under the prediction given by the
Eshelby’s inclusion theory, a hydrostatic compressive stress will be generated in martensite during the martensitic transformation,
because the martensite grain cannot expand to its ideal volume under the constraint of surrounding austenite. As a results, a tensile
hydrostatic stress will be generated in the untransformed austenite due to the stress balance condition. The phase stresses generated
by martensitic transformation have been studied experimentally for a long time, however, the puzzling results have still been
reported. In 1980, Tanaka and Shimizu claimed the generation of hydrostatic compressive stress in austenite with MT during sub-
zero cooling estimated from the change in austenite lattice parameter measured by in-situ X-ray diffraction below room
temperature (RT) for Fe-Mn-C alloys, in which the austenite lattice parameter was smaller than that extrapolated from thermal
expansion coefficient 1). On the other hand, Hayakawa and Oka observed a decreasing of the austenite lattice parameter for a
rapidly cooled sample while little change in a slowly cooled sample and then concluded that the macroscopic residual stress near
the specimen surface generated by temperature gradient during quenching after solution treatment was relaxed with MT 2). That
is, macroscopic residual stress should be considered to evaluate phase stresses, correctly. To avoid the surface effect, synchrotron
X-ray diffraction with high penetration ability has been employed for studying this issue in recent years. Several studies have
reported that the austenite phase stress change from tensile hydrostatic to compressive with increasing of martensite volume
fraction 3,4). To clarify this puzzling topic, we carried out neutron diffraction experiments having higher penetration ability than
above methods to investigate the lattice parameter changes in real time during martensitic transformation. A high-purity Fe-18Ni
binary alloy was selected to avoid the chemical contribution from the interstitial element. Several measuring processes were
carefully designed and performed by in-situ neutron diffraction.
Chemical composition of the alloy used in this study was Fe-17.76Ni-0.037Al-0.0016C-0.0012N (wt%). In-situ neutron
diffraction experiments were conducted at the beam-line 19 ‘TAKUMI’ in MLF, J-PARC with a thermomechanical processing
simulator installed on it as shown in Fig. 1(a). Cylindrical specimens of the alloy with height of 11 mm and diameter of 6.6 mm
were used. As shown schematically
in Fig. 1(b), an optical dilatometry
was used to simultaneously measure
the change in diameter of the
specimen with in-situ neutron
diffraction to compare volume
change of a whole sample with
crystalline diffraction data.
Diffraction profiles were analyzed
via Rietveld refinement to obtain the
volume fractions and lattice
Fig. 1 Schematic illustration showing (a) the thermomechanical processing simulator
parameters of austenite and
installed on TAKUMI (b) the diffraction geometry of in-situ neutron diffraction
martensite. Single peak fitting was
experiment.
also carried out to obtain the
Fig. 2 (a) Temperature history and change in volume fraction of martensite. (b) Changes in neutron diffraction profile during
continuous cooling.
Fig. 4 (a) Temperature history and changes in volume fraction of martensite. Changes in lattice parameters of (b) austenite and
(c) martensite, and (d) total stress.
Using the stress and defects free lattice parameters (𝜀) as the reference, the lattice strains of austenite and martensite were
calculated. Then, the lattice stress is given by,
!
𝜎= "#$% 𝜀
where, E and υ refer to the Young’s modulus and the Poisson’s ratio, respectively. Therefore, the total stress 𝜎& can be calculated
by the volume fraction weighted average method as follows,
)
𝜎& = 𝑓 '( ∙ 𝜎 '( + (1 − 𝑓 '( ) ∙ 𝜎
)
where, 𝜎 and 𝜎 '( refer to the lattice stresses of austenite and martensite, respectively. The 𝜎& as a function of time is shown
in Fig. 4(d). Despite the total stress is expected to be zero, 𝜎& deviated from zero with the progress of martensitic transformation
Acknowledgement
This work was primarily supported by the Elements Strategy Initiative for Structural Materials (ESISM) in Kyoto University. The
authors would like to thank Prof. Goro Miyamoto of Tohoku University, Japan for providing the alloy. The neutron diffraction
experiments at the Materials and Life Science Experimental Facility of the J-PARC were performed under a general program (Project
No. 2018L0400).
References
1) Y. Tanaka, K. Shimizu: Trans. JIM 21 (1980) 42-51.
2) M. Hayakawa,M. Oka: Acta metall., 31(1983), 955-959.
3) M. Villa, F. Niessen, M.A.J. Somers: Metall. Mater. Trans. A 49 (2018) 28-40.
4) S.Y.P. Allain, S. Gaudez, G. Geandier, J.C. Hell, M. Gouné, F.Danoix, S. Michel, A. Samy, A. Poulon-Quintin: Mater. Sci. Eng. A
710(2018)245-250.
At the end of the hot rolling process, microalloyed steel will form nanoscale precipitates of compounds such as
NbC and TiN. These precipitates have the potential to improve properties by grain refinement and precipitation hardening.
Precipitates with smaller size and higher number density typically yield higher strength. The rate of cooling after hot rolling
will control the precipitation and coarsening of these particles. One of the challenges is to obtain a statistically meaningful
distribution of their size, shape and composition as up to now this has only been possible through tedious, manual analysis.
In this study two electron microscopy workflows are presented. In the first workflow, a FIB-SEM dual beam is used
to prepare a site specific TEM lamella with a highly automated process. Thin foil sections of roughly 25 square microns can be
prepared in 2-3 hours. In the second workflow, a TEM is used to acquire tile-by-tile images and chemical information in a fully
automated process, even for overnight, unattended analysis. Lamellae prepared by focused ion beam have advantages
compared to the carbon replica technique. For example, particle location is maintained relative to a grain boundary, and
particles are not lost or destroyed during sample preparation.
Instead of characterizing just a few precipitates per day manually, the automated TEM can process a statistically
relevant number of particles (thousands) in an overnight run. With these workflows, the time to develop a new steel product can
be reduced and R&D manpower can be freed up for new responsibilities.
Figure 1: S/TEM image of FIB-prepared steel lamella (left) and STEM-in-SEM image of nanoparticles in the steel matrix.
1
Applied Physics Research Laboratory, Kobe Steel, Ltd., 1-5-5, Takatsukadai, Nishi-ku, Kobe, 651-
2271, Japan
2
Research Network and Facility Services Division, National Institute for Materials Science, 1-2-1,
Sengen, Tsukuba, 305-0047, Japan
*Email: murata.yuya@kobelco.com
Keywords: steel, grain boundary segregation, STEM-EDS, ξ-factor method
Results
The composition maps of the solute atoms of the Fe-Mn-Cr-P sample treated by isothermal
heating at 773K are shown in Fig. 1. The composition of the solute atoms increases at a thin area
with a width of about 5 nm, which seems to be a grain boundary. The absorption-correction term
was calculated from the sample thickness at each measurement point. In this sample, the range of
the absorption-correction term for P distributed in 1.25-1.4.
The grain boundary segregations Γ (atom/nm2) with its errors, and the minimum detectable Γ are
shown in Table 1. The values were calculated from the composition, the composition error, and
the minimum detectable composition, in the grain boundary area and the grain area.3,6,7 In order
to improve the accuracy, the 14000 data points were binned in the grain boundary area and the
grain area with a width of 5 nm each, equivalent to an acquisition time of about 550 seconds. In
a reasonable acquisition time, we could obtain the error and the minimum detectable Γ lower than
the ones in the previous reports, owing to the large solid-angle of the EDS detector used in this
study. Table 2 showed that the Γ of P and Mn decrease with increasing temperature and Cr addition.
The grain boundary segregation energy −𝛥𝑔𝑏 calculated from the Γ of P and Mn according to
equation (1) is shown in Table 2.8
𝛤⁄(𝛤 𝑚𝑎𝑥 − 𝛤) = 𝑥 𝑒𝑥𝑝(− 𝛥𝑔𝑏 ⁄𝑅𝑇) (1)
Conclusion
We achieved grain boundary segregation measurement of trace elements in steels with the error
and the minimum detectable value lower than the ones in the previous reports, owing to the large
solid-angle of the STEM-EDS detector and ζ-factor method. We confirmed that the P grain
segregation measured by the improved method falls in the range in the previous reports based on
AES, in terms of its temperature dependence and the influence to the embrittlement property.
In future, the improved segregation measurement should be applied to grain boundaries with more
defined characters, in alloys with constant plastic flow resistance, in order to further reduce the
uncertainty in the investigations of the segregation and the embrittlement properties.
Figure 1
Composition maps of the CR sample treated by isothermal heating. The area is 100 nm × 100 nm.
Mn P Cr
Acknowledgement
The authors thank Prof. M. Watanabe of Lehigh University for providing the SRM2063a standard
sample and the analysis software for ξ-factor method. This study was carried out as a part of
Materials Open Platform steel project at National Institute for Materials Science in Japan.
References
1
P. Lejček, M. Šob, and V. Paidar, Progress in Mater. Sci. 87, 83 (2017).
2
L. Zhang, B. Radiguet, P. Todeschini, C. Domain, Y. Shen, and P. Pareige, J. Nucl. Mater. 523, 434
(2019).
3 M. Watanabe and D. B. Williams, Z. Metallkd. 94, 3 (2003).
4 C.-Y. Hsu, J. Stodolna, P. Todeschini, F. Delabrouille, B. Radiguet, and F. Christien, Micron 153,
103175 (2022).
5 P. Lejček, S. Hofmann, and A. Krajnikov, Mater. Sci. and Eng. A234, 283 (1997).
6 U. Alber, H. Müllejans, M. Rühle, Ultramicroscopy 69, 105 (1997).
7 V. J. Keast, D. B. Williams, J. Microscopy 199, 45 (2000).
8
J. R. Rice and J.-S. Wang, Mater. Sci. and Eng. A107, 23 (1989).
*email: crh@monash.edu
Keywords: austenite stability, medium-Mn steels, deformation-induced martensitic transformation, retained austenite,
strain partitioning, digital image correlation
Steel researchers continually strive to develop improved high-strength high-elongation structural components with applications
for the industrial world, especially in the automotive industry. This drives the development of sheet steel from conventional High
Strength steels (HSS) to the 3rd generation Advanced High Strength Steels (AHSS). Throughout this development, a key design
consideration is to take advantage of the transformation-induced plasticity (TRIP) effect arising from the deformation-induced
martensitic transformation from metastable austenite. The dynamically evolving microstructure during deformation can both
increase the base strength of the material and also significantly improve the strain hardening rate. These two contributions are
important requirements for designing high-strength high-elongation alloys.
Even though researchers have long realized the importance of understanding the TRIP effect, many important
questions governing the the mechanical stability of austenite remain unresolved. This is because of the complexities involved in
deconvoluting the different factors affecting the stability. It is known that the mechanical stability of austenite depends on the
chemical composition (1-5), grain size (6–9), and morphology(10–12) of the grains and also the constraints from the surrounding phases
(13–16)
. However, deconvoluting these different effects is not straightforward, and few studies dedicated to this deconvolution
exist. In addition, there is a lack of systematic characterization combining both macroscopic and local measurements of the
transformation. The variation between local and average measurements of strain-induced martensite can be significant due to
stress and strain partitioning between phases and an understanding must be achieved at both scales of the microstructure.
Recently, microscopic-digital image correlation (micro-DIC) combined with scanning electron microscopy (SEM)
has been proven as a useful technique for studying the local deformation behaviors with sub-micron resolution(17–22). In our study,
we couple micro-DIC with quasi in-situ electron backscattered diffraction (EBSD) testing. The patterning we chose for micro-
DIC is SiO2 nanoparticles (22). This speckle pattern does not interfere with the EBSD data acquisition, and the speckle density is
sufficiently high to provide good spatial resolution in the local strain measurements.
In this talk, we discuss the grain size effect on austenite mechanical stability in an intercritically annealed Fe-0.1C-5Mn
(wt. %) steel prepared under conditions where the composition of the austenite islands is the same and spatially uniform withing
the austenite. At each macroscopic deformation level, we obtain both the EBSD map (Figure 1(a), revealing the microstructure)
and DIC strain map for multiple regions. We use the open-source python routines (23) to conduct the data analysis. An example
coupling the DIC and the EBSD datasets is shown in Figure 1(b).
(a) (b)
Tensile direction
)
TD
ND RD
Figure 1 (a) EBSD map showing the phase contrast of the 120 hour-intercritically annealed 0.1C5Mn steel after 1st strain level. (b) The EBSD
map in (a) overlapped onto the DIC strain map for the same area.
001 101
(c) (d)
Figure 2 (a) The grain size versus the local effective shear grain average strain (after which the austenite grain transforms to martensite) for all
the austenite grains examined in the studied 0.1C5Mn steel. The grains shown in (a) can be categorized into three groups based on their grain
orientations, i.e. (b) grains of which the <111> direction is (nearly) parallel to the rolling direction (RD), (c) grains of which the <101> direction
is (nearly) parallel to the rolling direction (RD) and (d) grains of which the <001> direction is (nearly) parallel to the rolling direction (RD)
In this talk, we discuss this approach and the insights that can be obtained in the effect of different variables on the mechanical stability
of austenite in advanced high strength steels.
1. Introduction
Sub-micron cellular structure is a unique microstructure in additive manufactured (AMed) alloys. Abundant efforts
have been made to reveal the physical origin of cellular microstructures and their influence on mechanical response of AMed
materials. However, these researches are mainly based on austenitic stainless steels1,2,3), aluminium alloys4) and Co-Cr-Mo alloys5),
which are almost free of solid-state phase transformations. The effect of cellular microstructure on solid-state transformation and
their potential role in microstructure architecting are rarely explored.
In this work, the 18Ni300 maraging steel is fabricated by selective laser melting (SLM) and post-fabrication heat
treatment. The formation mechanism of cellular microstructures and the role of cellular microstructure in austenite reversion is
demonstrated by microstructure characterization and thermo-dynamic modeling.
Table 1: The chemical composition ( in wt.%) of 18Ni300 powder used in this research
Fe Co Ni Ti Mo
Bal. 9.2 18.7 1.0 4.8
3. Results
Sub-micrometer cellular microstructures are found in as-SLMed samples. The cell boundaries are enriched in Ti and
Mo. Unlike materials with almost no solid-state phase transformations, no dislocation networks along the cell boundaries are
found in as-SLMed samples. The elimination of dislocation networks is supposed to be due to cyclic martensite ↔ austenite
transformations.
The cellular microstructure shows its crucial effect on promoting the austenite reversion during aging. Only small
amount (less than 3 vol.%) of austenite is detected in the as-SLMed samples (SLM samples). When aged at 480 ℃ (SLM-Age
samples), most cell boundaries transform into reverted austenite, and the austenite fraction exceeds 10 vol.% within 1 hour. The
austenite on the cell boundaries formed a distinctive 3D network. In contrast, conventional 18Ni300 steel and SLMed samples
aged after the solution treatment (Sol-Aged samples) show negligible austenite reversion.
Thermodynamic calculations and phase-field simulations clarify the mechanism of the fast austenite reversion in
SLM-Age samples. The high Ti and Mo content can increase the driving force for austenite reversion effectively. Further phase-
field simulations show that when austenite grows along the cell boundaries, the austenite-martensite interface migrates quickly in
a massive manner. When the austenite-martensite interface moves towards cell interiors whose composition is almost the same
as the bulk composition, the interface will migrate sluggishly with significant element partitioning across it. The latter case can
also describe the austenite reversion at 480 ℃ in Sol-Aged samples and conventional 18Ni300 maraging steels.
4. Conclusions
Sub-micro cellular microstructure presents in SLMed Maraging steels. The cell boundaries are found to be rich in Ti
and Mo, but free of dislocation networks. These cells can significantly promote the austenite reversion when the SLMed samples
are aged at 480 ℃. During aging, the reverse austenite prefers to grow along the cell boundaries for the Ti and Mo enrichment in
these regions increases the driving force of austenite reversion effectively. When austenite-martensite interface moves along cell
boundaries (with high Ti, Mo), the interface moves in a massive manner with only slight element redistribution. When the
austenite-martensite interface moves towards regions with the bulk composition, the interface migration is sluggish accompanied
with notable element partitioning.
References
1) Y.M. Wang, T. Voisin, J.T. McKeown, J. Ye, N.P. Calta, Z. Li, Z. Zeng, Y. Zhang, W. Chen, T.T. Roehling, R.T. Ott, M.K. Santala,
P.J. Depond, M.J. Matthews, A.V. Hamza and T. Zhu, Nat Mater 17(1) (2018) 63-71.
2) K.M. Bertsch, G. Meric de Bellefon, B. Kuehl and D.J. Thoma, Acta Materialia 199 (2020) 19-33.
3) T. Voisin, J.-B. Forien, A. Perron, S. Aubry, N. Bertin, A. Samanta, A. Baker and Y.M. Wang, Acta Materialia 203 (2021).
4) J. Wu, X.Q. Wang, W. Wang, M.M. Attallah, M.H. Loretto and Acta Materialia 117 (2016) 311-320.
5) Z.W. Chen, M.A.L. Phan and K. Darvish, Journal of Materials Science 52(12) (2017) 7415-7427.
Multi-step intercritical heat treatment has been introduced in low carbon low manganese steels in some studies to obtain
a considerable amount of stable retained austenite and enhance mechanical properties 1,2). Retained austenite is usually
introduced in microstructure 2) when the sample is subjected to the first intercritical annealing which was also called
pre-annealing 3). It was reported that the amount of cementite precipitated during heating was reduced when retained austenite
was a constituent of the initial microstructure 4,5). However, the underlying reason continues to be unclear to the best of our
knowledge. The objective of the present study is to elucidate the precipitation and dissolution of cementite during annealing
when the initial microstructure with or without retained austenite was introduced as a constituent by pre-annealing. The
microstructure evolution during the heat treatment and the change of the stability of retained austenite will also be
investigated.
The chemical composition of experimental steel is 0.2C -2.18Mn -1.01Ni -0.91Cu -1.11Al -0.58Si -0.032Nb -0.255Mo.
The steel was melted in a vacuum induction furnace and forged into ingots with a thickness of 80 mm. The ingot was
homogenized at 1200˚C for 2 h in a muffle furnace, and then hot rolled to 12 mm plate followed by quenching to room
temperature. The Ac1 and Ac3 temperatures of experimental steel were determined to be 680˚C and 874˚C, respectively, by
using thermal dilatometer (DIL805A). In the present study, two types of heat treatment process were carried out. In type I,
samples were intercritically annealed at 690˚C for 30 min, 60 min and 120 min, respectively, and then air cooled to room
temperature, which were referred henceforth as IA30, IA60, and IA120, respectively. In type II, the sample was first
pre-annealed at 740˚C for 30 min and quenched to room temperature, referred as sample P. Then, sample P was annealed at
690˚C for 30 min and cooled to room temperature in air, labeled as P-IA30. Microstructure was observed by scanning
electron microscope (ZEISS ULTRA-55, SEM) at an accelerated voltage of 15 kV and transmission electron microscope
(FEI F20, TEM) equipped with an energy dispersive X-ray spectroscopy at an acceleration voltage of 200 kV. Specimens for
SEM observations were mechanically polished and etched by 3% nital solution. Specimens for TEM observations were
mechanically ground to 50 μm and punched into small disks of 3 mm diameter. Then, small disks were electropolished using
a solution of 5% perchloric acid and 95% ethanol (in vol.) at a voltage of 22 V and a temperature of -25˚C. The size and
distribution of retained austenite were studied by electron back scatter diffraction (EBSD) at an acceleration voltage of 20 kV
and a step size of 50 nm. Specimens for EBSD observation were mechanically polished and electro-polished using a mixed
solution of 85% ethanol, 10% perchloric acid and 5% glycerine (in vol.) at 20 V for 15 s to remove the mechanically damaged
surface layer. HKL Channel 5 software was used to process the EBSD data.
The microstructure of hot rolled sample was martensite, while the microstructure of sample P was composed of ferrite,
martensite and retained austenite. The statistical analysis indicated that the width of reversed austenite was 300 ± 97 nm. The
reversed austenite formed at 740˚C was not stable and partly transformed into martensite. The XRD results showed that the
volume fraction of retained austenite in sample P was about 8%, while retained austenite could be hardly detected in hot rolled
sample. Microstructure of samples subjected to different heat treatments characterized by SEM is presented in Fig. 1.
According to previous analysis, the microstructure of sample IA30 was composed of reversed austenite and ferrite, as shown
in Fig. 1a. Both granular and lath-like reversed austenite were observed. There were a number of small particles of size
ranging from ~50-100 nm dispersed in the matrix, which are considered as cementite based on previous studies 5,6). With the
increase of annealing time, the size of reversed phase increased and the fraction of cementite decreased. After annealing for
120 min, cementite was hardly observed in the microstructure. The microstructure of sample P-IA30 was mainly comprised
of reversed austenite and ferrite. Majority of reversed austenite was lath-like or small rod-like and very few cementite particles
(1)
where λ (=0.15418 nm) is the wave length of X-ray, θ is the diffraction angle of (hkl) peak. The carbon concentration in
austenite can be calculated by equation (2):
(2)
where the atomic symbols represent the concentration of carbon and alloying elements in austenite in weight percent. Mn and
Ni concentration in austenite were measured by TEM-EDS, while the Si, Al and Cu content in austenite were given by
Thermo-Calc simulation results. No Mo was detected in austenite based on TEM-EDS results. The carbon concentration in
austenite calculated based on above equations were 0.57 wt.% for sample IA30 and 0.68 wt.% for sample P-IA30,
respectively. The carbon concentration in ferrite was calculated using equations (3) and (4):
(3)
(4)
where xMn and xC are weight percent of Mn and carbon in ferrite, respectively. The calculated Mn concentration in ferrite was
1.31 wt.% in sample IA30 and 1.34 wt.% in sample P-IA30, respectively. The carbon concentration in ferrite was calculated
to be 0.011 wt.% for sample IA30 and 0.014 wt.% for sample P-IA30, respectively. The total mass fraction of carbon enriched
in austenite were 0.110 wt.% for sample IA30 and 0.156 wt.% for sample P-IA30, respectively, which indicated that
pre-annealing was beneficial for carbon enrichment in austenite.
In order to verify whether cementite precipitated in this study during the heating period, a sample was heated to 690˚C
and then quenched to room temperature without isothermal holding. The microstructure of this sample is presented in Fig. 2.
A number of cementite particles were observed along the lath boundaries and within the martensite lath. It is reasonable to
believe that cementite precipitated during heating the sample to 690˚C in this study, which did not dissolve completely even
after annealing for 30 minutes at this temperature. And dilatometric curve showed that cementite formed at ~450˚C during
heating at a heating rate of 10 ˚C/s. TEM-EDS results showed that Mn concentration in cementite was ~10.1 ± 3.2 wt.%,
which was lower than that of sample IA30. Given that the cementite can rapidly grow during heating, which is driven by rapid
partitioning of carbon from martensite into cementite, then Mn would further partition from martensite into cementite during
the early stage of annealing. With the increase of annealing time, cementite dissolves and is accompanied by austenite
formation due to thermodynamic constraint. The austenite stability can be assessed by martensite transformation temperature
(MS). MS of samples IA30, IA60, IA120 and P-IA30 was calculated to be 28˚C, 15˚C, 50˚C and -14˚C, respectively. It is
reasonable to conclude that the retained austenite in sample P-IA30 has a higher stability, which may result in a sustained
Acknowledgement
This research was funded by the Key Research and Development Program of Shandong Province, China (No. 2019JZZY020238).
References
1) Z.J. Xie, S.F. Yuan, W.H. Zhou, J.R. Yang, H. Guo, C.J. Shang, Materials & Design 59 (2014) 193-198.
2) W. Zhou, H. Guo, Z. Xie, X. Wang, C. Shang, Materials Science and Engineering: A 587 (2013) 365-371.
3) D.T. Han, Y.B. Xu, F. Peng, Y. Zou, R.D.K. Misra, Journal of Materials Science 55(10) (2019) 4437-4452.
4) N. Yan, H. Di, R.D.K. Misra, H. Huang, Y. Li, Materials Science and Engineering: A 753 (2019) 11-21.
5) X.L. Wang, Z.Q. Wang, A.R. Huang, J.L. Wang, X.C. Li, S.V. Subramanian, C.J. Shang, Z.J. Xie, Materials Characterization 169
(2020).
6) X. Zhang, G. Miyamoto, Y. Toji, S. Nambu, T. Koseki, T. Furuhara, Acta Materialia 144 (2018) 601-612.
Fig. 1. SEM micrographs of (a) IA30, (b) IA60, (c) IA120 and (d) P-IA30 samples.
Heterogeneity attracts a deep interests as an alternative concept for microstructure control in structural materials1). It is because,
the heterogeneous microstructure, if precisely controlled, can present extraordinary properties which are difficult to be
expected in conventional monolithic microstructure. The advantages in structural properties led by heterogeneity is believed to
be driven by synergetic effect of mixed domains with dissimilar characteristics. In that sense, combining more than one
heterogeneous characteristics can open a high chance for the development of steels having mutually exclusive properties. In
the present study, we attempted to disclose the influence of chemical heterogeneity in dissimilar microstructure scale in
martensitic steel. It revealed that the combined heterogeneity in chemistry and size can contribute to the enhanced
combination of strength and ductility in martensitic steel, compared to that in conventional one. Detailed analysis on the effect
of chemical heterogeneity in micro and nano-scale demonstrates the importance of sophisticated control of heterogeneity in
various microstructure scales for the development of innovative structural materials.
Chemical composition of investigated alloy was 0.18 wt.% C based low carbon steel. First intercritical annealing at a
temperature, where ferrite, austenite and cementite coexist, created a multi-scale chemical hegerogeneity of Mn distribution
thanks to the dissimilar Mn solubility in each phase. Then second annealing was conducted for austenitization, which
generated martensitic microstructure with micro-scale and nano-scale heterogeneities, resided in the Mn-enriched region
originating from the initial austenite and cementite. Microstructure characterization was conducted by using a SEM with
EBSD and TEM observations and austenite fraction was evaluated by using magnetic saturation method. Tensile test was
performed to examine the influence of heterogeneity-driven characteristic microstructure on the mechanical properties of
martensite.
Fig. 1 illustrates a schematic diagram of the microstructure evolution during the two-step heat treatment for the generation
multi-scale heterogeneity of Mn distribution, compared with the conventional heat treatment to generate martensitic
microstructure. Intercritical annealing successfully introduced multi-scale heterogeneous Mn distribution (Fig. 2), which ends
up with characteristic features in the final microstructure.
Micro-scale heterogeneity before the second annealing resulted in soft and hard region in the martensitic microstructure led by
difference solid solution hardening effect of Mn, which produced strain partitioning and back stress, promoting the strength of
steel. On the other hand, the nano-scale heterogeneity generated the tiny austenite particles after the second annealing, of
which extraordinary stability provided a persistent work hardening by TRIP effect, contributing improvement of ductility. It
made it possible to achieve the simultaneous improvement of tensile strength and ductility in martensitic steel, which have
been regarded as mutually exclusive properties (Fig. 3).
Acknowledgement
This work was performed under the support of POSCO
References
1) X. Wu, M. Yang, F. Yuan, G. Wu, Y. Wei, X. Huang and Y. Zhu: PNAS 112 (2015) 14501-14505.
2) J. H. Kim, G. Gu, M.H. Kwon, M. Koo, E. Y. Kim, J. K. Kim, J. S. Lee and D. W. Suh: Acta Mater. 201 (2022) 117506.
Keywords: first-principles calculation, atomic interaction, tetragonality, martensite, Fe-C interaction, C-C interaction
1. Introduction
The atomic interaction affects the microstructures and mechanical properties of materials. One of the most important factors that
has an influence on the atomic interaction is the chemical bonding between atoms. Whether bonding or antibonding is formed
between atoms determines the attractive or repulsive interaction between the atoms. Tetragonality is closely related with the
atomic interaction and is one of the most important research topics of Fe-C martensite, which is a vital microstructure of high-
strength steels. Many researches have been conducted for the tetragonality of Fe-C martensite[1]. In spite of these researches, we
still have some unsolved issues, such as, how alloying elements affect the tetragonality, how carbon configuration affects the
tetragonality and total energy, and what is the atomic interactions in Fe-C systems. In this study, the total energy for various Fe-C
systems is evaluated by the first-principles calculation and the tetragonality value is determined. The dependence of energy on
the configuration of carbon atoms is investigated and the effects of Al, Ni and Mn addition on the tetragonality of Fe-C martensite
are calculated. The degree of bonding and antibonding between Fe-C and C-C interactions is evaluated, and based on these
calculation results, the origin of atomic interactions in Fe-C system is discussed.
2. Calculation
The plane-wave base Vienna Ab-initio Simulation Package (VASP) with the pseudo potential method was used for the
calculation[2,3]. The density functional theory was applied using the projector-augmented wave (PAW) method based on the
generalized density gradient approximation (GGA-PBE) with spin polarization. Three kinds of supercells, a 2 × 2 × 2 supercell
with 15 Fe atoms and 1 substitutional atom (Al, Ni or Mn) and 1 interstitial carbon (k-point mesh is 8 × 8 × 8), a 3 × 3 × 3 supercell
with 54 Fe atoms and 1 or 2 interstitial carbon atoms (k-point mesh is 6 × 6 × 6) and 4× 4 × 4 (k-point mesh is 4× 4 × 4 ) supercell
with 128Fe atoms and 1 interstitial carbon atom were used for the calculation. These supercells are abbreviated as Fe15M1C1 (M=
Al, Ni or Mn), Fe54C1, Fe54C2, Fe128C1, respectively. The cut-off energy is 400 eV. Both the shape and size of the supercell are
varied during calculation. The degree of bonding and antibonding was quantitatively evaluated by COHP (Crystal Orbital
Hamilton Population) analysis using the software “LOBSTER”[4,5].
4. Summary
The total energy and tetragonality of Fe-C systems have been evaluated using the first-principles calculation and the effects of
carbon configuration on the total energy and tetragonality have been discussed based on the atomic interaction. The tetragonality
of the Fe-C system obtained from the first-principles calculation agrees well with the experimental results. When the Fe-C-Fe
pairs are parallel, the calculated tetragonality value becomes large and agrees with the experimental value. Al or Ni increases and
Mn decreases the tetragonality. The bonding is formed between Fe and C atoms, but the antibonding is formed between C atoms.
This antibonding is the one of the reasons why the C-C interaction is repulsive.
REFERENCES
[1] H.Ohtsuka and K.Tsuzaki: ISIJ International, 61(2021), 2677-2686.
[2] G.Kresse and J.Furthmuller: Phys. Rev., B54(1996), 11169-11186.
[3] G.Kress and D.Joubert: Phys. Rev., 59(1999), 1758-1775.
[4] R.Dronskowski and P.E.Bl Ö chl: J. Phys. Chem., 97(1993), 8617.
[5] S. Maintz, V. L. Deringer, A. L. Tchougreeff, R. Dronskowski: J. Comput. Chem., 34(2013), 2557.
[6] H.Ohtsuka, V. A. Dinh, T. Ohno, K. Tsuzaki, K. Tsuchiya, R. Sahara, H. Kitazawa and T. Nakamura et al.:
ISIJ International, 55(2015), 2483-2491.
Multiphase microstructures combining a strong phase such as martensite and bainite, with the more ductile retained
austenite, are key for the simultaneous increase of strength and ductility in advanced high strength steels. Steels combining bainite
and retained austenite can be produced through austenitisation followed by an isothermal treatment at a temperature in the bainitic
range. Alloying with silicon, with a low solubility in cementite, prevents that the carbon diffusion into the austenite during bainite
formation leads to significant carbide precipitation. Instead, austenite becomes rich in carbon and (meta)stable at room
temperature. Steels with higher hardenability can lead to combinations of martensite and retained austenite through the quenching
and partitioning process, in which an austenitic microstructure is partially transformed into martensite, while further annealing
leads to the partitioning of carbon from martensite to austenite. Also in this case, the carbon enriched austenite remains
(meta)stable at room temperature.
Investigations on the quenching and partitioning process showed that, in some alloys, the process of carbon partitioning
can be affected by the formation of bainite. Both processes (carbon partitioning from martensite and bainite formation) lead to
carbon enrichment of the austenite, although only the second leads to a significant change on the volume fractions of phases in
the microstructure. Further investigations showed that bainite formation seemed promoted and even accelerated by the presence
of martensite, when compared with bainite kinetics at the same conditions but in absence of martensite [1-3]. Moreover, bainite
can even form isothermally below the martensite start temperature [4]. Literature showed that these were not new observations,
and that the acceleration of bainite kinetics in the presence of martensite was considered a “well stablished phenomenon” already
in the year 1959 [5].
The present work presents ten years of investigations on the formation of bainite in the presence of martensite at the
University of Delft. The main aim of this research was the application of this phenomenon for the development of new advanced
high strength steels through shortened processing routes presenting controlled combinations of martensite, bainite and retained
austenite. This implied investigations at the level of microstructural characterisation, microstructure control and modelling, and
analysis of mechanical properties.
Investigated steels were composed in its majority by 0.2 C-3.5Mn-1.5Si (in wt.%), with different additions of Mo and
Al, among other elements. Heat treatments were applied with a Bähr 805A dilatometer. Thermal treatments included full
austenitisation followed by the formation of bainite in presence of different fractions of martensite. The analysis of the dilatometry
curves allowed the identification of an increase in the nucleation rate of bainite in presence of martensite at the very initial stages
of the transformation [6]. The identification of martensite and bainite after the application of this type of heat treatments is difficult
[7-8], due to the partial tempering of martensite during the isothermal treatment, which makes it resemble as bainite in the final
microstructure. However, some specific features belonging to bainite or martensite were identified [9]. Figure 1 displays a
microstructure isothermally formed below the martensite start temperature in a low carbon steel.
Modelling the formation of bainite in presence of martensite requires good understanding of bainite formation.
However, the nature of the bainitic transformation is still a topic of much discussion in the scientific community. Nevertheless,
the displacive approach has been proved to efficiently predict the kinetics of bainite formation in low carbon steels at relatively
low temperatures. This approach assumes that bainite formation is controlled by the successive displacive nucleation of individual
bainitic subunits, while carbon partitioning into the surrounding austenite occurs soon afterwards. Based on this approach, Ravi
et al. [10] developed a kinetic model for the kinetics of bainite formation able to distinguish nucleation at austenite grain
boundaries and at bainite/austenite interfaces. In Ravi et al. [11], the mentioned model was applied to elucidate the influence of
martensite presence on bainite kinetics by also introducing the role of martensite/austenite interfaces as possible nucleation sites.
Figure 1. Microstructure of a low carbon steel alloyed with 1.5 wt.% Si after bainite formation below the martensite start
temperature: (a) scanning electron micro micrograph. (b-c) electron backscatter diffraction analysis of one prior austenite
grain from the location shown in (a). Specifically: (b) inverse pole figure map (only BCC structure), (c) phase distribution
map (BCC: green; FCC: red), and (d) kernel average misorientation map.
Comparison with experiments showed that the formation of bainite at martensite/austenite interfaces facilitates further
bainite nucleation by the creation of new bainite/austenite interfaces, being this influence dependent on the fraction of pre-existing
martensite. Further, the overall rate of bainite formation was found to quickly decrease as the available martensite/austenite
interfaces are consumed and carbon enrichment of austenite increases due to carbon partitioning from martensite to austenite.
The influence of the pre-existing martensite on the mechanical properties of steels developed by this approach was
found to depend on its phase fraction and tempering degree [12]. In general, the prior formation of martensite leads to an increase
in strength due to the refinement of the resulting microstructure and the presence of carbon in solid solution. However, relatively
long isothermal treatments lead to an increase on the bainite fraction but also to the tempering of the pre-existing martensite. In
this case, the microstructures display a lower strain hardening, possibly due to softening of the pre-existing martensite due to
carbon segregation (reduction of tensile strength) and the presence of carbide precipitation (increase of yield strength). Further
investigations have demonstrated that steels in which bainite has been formed in presence of a small fraction of martensite (less
than 0.05) can lead to similar strength and ductility combinations than in absence of pre-existing martensite, while the addition of
pre-existing martensite has the advantage of significantly reducing the whole processing time [13].
The presented investigations helped to clarify the effect of the presence of martensite on the kinetics of bainite
formation, and contributed to the industrial applicability of this phenomenon for the development of new advanced high strength
steels. However, some fundamental questions remain unanswered. One of them is what is the specific mechanisms leading to the
observed acceleration of bainite kinetics. Although its relation with the introduced martensite/austenite interfaces in the system is
clear, the question whether bainite nucleation is induced by the interface itself or by the dislocations next to it remains.
Acknowledgement
This work has been performed in the framework of several projects. Firstly, under the support of Netherlands Organization for Scientific
Research (NWO) and the Dutch Foundation for Applied Sciences (STW) through the VIDI-Grant 12376, and under the support of the
European Research Council under the European Union’s Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement n.
[306292]. Currently, this topic is being further investigated under the support of the National PPS-HTSM Program and of the Research
Fund for Coal and Steel for funding this research under the contract RFCS-02-2019 (Project No. 899251).
1) W. Gong, . Tomota, S. Harjo, Y. H. Sua and K. Aizawa, Acta Materialia 85 (2015) 243
2) H. Kawata, K. Hayashi, N. Sugiura, N. Yoshinaga and M. Takahashi, Materials Science Forum 638-642 (2010) 3307
3) Y. Toji, H. Matsuda and D. Raabe, Acta Materialia 116 (2016) 250
4) S. M. C. van Bohemen, M. J. Santofimia and J. Sietsma, Scripta Materialia 58 (2008) 488
5) S. Radcliffe, E. Rollason, J. Iron Steel Inst. 191 (1959) 56
6) A. Navarro-Lopez, J. Sietsma and M. J. Santofimia, Metallurgical and Materials Transactions A 47 (2016) 1028
7) D. Kim, J. G. Speer and B. C. De Cooman, Metallurgical and Materials Transactions A 42 (2011)1575
8) S. Samanta, P. Biswas, S. Giri, S. B. Singh and S. Kundu, Acta Materialia 105 (2016) 390
9) A. Navarro-Lopez, J. Hidalgo-Garcia, J. Sietsma and M. J. Santofimia, Materials Characterization 128 (2017) 248
10) A. M. Ravi, J. Sietsma and M. J. Santofimia, Acta Materialia 105 (2016) 155
11) A. M. Ravi, A. Navarro-Lopez, J. Sietsma and M. J. Santofimia, Acta Materialia 188 (2020) 1
12) A. Navarro-Lopez, J. Hidalgo-Garcia, J. Sietsma and M. J. Santofimia, Materials Science & Engineering A 735 (2018) 343
13) A. Navarro-Lopez, J. Hidalgo-Garcia, J. Sietsma and M. J. Santofimia, Materials and Design 188 (2020) 108484
Abstract:
Understanding the deformation behavior of metals over a wide range of strain rates is of critical
importance in a variety of engineering applications, such as vehicle collision, projectile penetration,
explosion shock, and so on. Decades ago, studies on the strain rate dependent deformation behavior
of various metals, including bcc metals (e.g. Pure Iron), fcc metals (e.g. Copper, Nickel), and hcp
metals (e.g. Titanium), were reported, and the results show that all of these metals exhibit positive
strain rate sensitivity. Here, we investigated the strain rate dependent mechanical behavior of several
steels, including bcc steels (e.g. IF steel and PHS steel), fcc steels (e.g. TWIP steel and 316L steel),
and multiphase steels (e.g. Q&P steel and medium Mn steel). We found that the strain rate dependent
deformation behavior of most of these steels differs from that of the pure metals previously mentioned,
which exhibits abnormally negative strain rate sensitivity. The underlying mechanism has been
extensively investigated. This discovery will have a significant impact on the use of steels as structural
materials in components that are likely to be subjected to high strain rate impact, such as automobile
anti-collision beams.
Abstract:
The structural and chemical features of passive films formed on a high Cr-added high-Mn (18 wt%)
twinning-induced plasticity (TWIP) steel and 304L, 316L austenitic stainless steels, i.e., 304L, 316L,
were revealed using Cs-corrected scanning transmission electron microscopy (STEM) and atom probe
tomography (APT). Based on these results, we investigated the cause of lower passive ability of the
high Cr-added TWIP steel than the typical stainless steels. Moreover, based on the chemical difference
of passive films between the 316L and the 304L, we suggested the primary cause of the improved
pitting resistance of 316L than 304L.
To increase the corrosion resistance of typical high-Mn TWIP steel series, high amount of Cr (17wt%)
was added to typical 18 (wt%) TWIP steel, i.e., HCr-HMnS. When comparing the passive abilities of
the HCr-HMnS, 304L and 316L by potentiodynamic polarization curves, the thickness and chemical
compositions of the material passivation films were performed via Cr-corrected STEM and APT.
Although the passive films of all the samples with similar Cr concentrations had the same thickness,
unprecedented hexagonal wurtzite MnO inside the passive film of HCr-HMnS specimen was
susceptible to corrosion cracking; this was not observed in the passive films of typical stainless steels.
This MnO caused crack formation during potentiodynamic polarization test, suggesting that reducing
the harmful MnO by adding Mo and Ni facilitates the development of high-Mn base stainless steels.
Furthermore, higher MoO2 composition of the passive films on 316L than 304L might would result
in primarily the improved pitting resistance. We propose that Mo and Ni, which can reduce the MnO-
density, may be beneficial for developing high-Mn base stainless steels..
1. INTRODUCTION
Microstructure of martensite contains packets, blocks, and laths. The structure consisting of several packets and many blocks
with various orientations (variants) are usually transformed from one of the prior austenite (γ) grains. Martensite with single-
packet and multi-block structure is formed when the prior austenite grain becomes fine. Further refinement in the austenite grain
size results in the formation of single-block (single-variant) martensite. In this study, we examined the structure and mechanical
properties of the martensite formed from various grain size prior γ obtained from ultrafine ferrite + Mn rich cementite steels with
the chemical composition of 0.1-0.2%C-2%Si-5-7%Mn. Particularly, the purpose of this study is to form a single-variant
martensite with a reduction in the prior γ grain size and to investigate the mechanical properties of the single-variant martensite1).
2. EXPERIMENTAL
First, a 20 kg ingot of steels were prepared by vacuum-melting with the chemical composition of 0.1-
0.2%C-2%Si-5-7%Mn . Subsequently, the ingots sample were subjected to hot forging at 1200 °C.
Then, multi-pass warm caliber-rolling was performed at 550℃. The resulting structure was an
ultrafine ferrite + MN rich cementite structure. Second, these bars with ferrite + Mn rich cementite
structure were heated and soaked at 700, 750, 800, 900, and 1200 °C for 1 h and air
Fig. 2 Single variant
cooled. Variant analysis was carried out to the microstructures obseved by EBSD to martensite in 0.1C-2Si-5%Mn
clear the prior austenite gtain size. Transmission X-ray diffraction (XRD) experiments steels heat treated at 750 ℃1).
on 0.5-mm-thick specimens were carried out using the synchrotron radiation
facility at SPring-8. A quantitative analysis of the retained was carried out
using the results.
*email: tanaka.masaki.760@m.kyushu-u.ac.jp
Pearlite exhibits a lamellar structure comprising alternating layers of ferrite and cementite. In the wire manufacturing
process, the widths of the cementite lamellae are gradually reduced, resulting in heteronanostructures on the submicron scale. The
increase in strength obtained during the wire-manufacturing process is explained by the decrease in the lamellar spacing, which
occurs as the wire is drawn[2]. Wire-drawn pearlitic steels have high strength and resistance to delayed fractures[3]; however,
they experience delamination—a type of vertical cracking—during torsion tests. A delamination mechanism has not been
definitively identified, although controlling factors have been suggested. Therefore, the mechanisms underlying the delamination
are discussed in this paper, with a focus on the plastic-deformation capability of ferrite in the pearlite structure.
Factory-produced piano wires were employed as a testing material (SWRS82B). After hot rolling, the diameter of the wires
was reduced to 9 mm. The wires were then kept in a tubular furnace at 1223 K for 240 s and then subjected to lead patenting at 833 K
for 60 s to obtain a full pearlite microstructure. The patented material was drawn until the diameter reached 3 mm. Delamination is
promoted by an increase in the wire-drawing strain; however, as the amount of wire-drawing strain in this study was small, the sample
did not exhibit delamination during drawing. Therefore, the drawn sample was aged at 423 K for 3.6 ks to induce delamination during
torsion. Torsion testing was conducted using self-made equipment, in which both ends of the rod material were held by chucks placed
160 mm apart. Then, one side was firmly affixed in a chuck, and the other side was rotated at a fixed speed. The torsion test was conducted
at room temperature. The torque applied to the test specimen was recorded using a torque meter mounted to the rotating part, and the
number of rotations was calculated using the rotation speed. To control the scattering of the crack initiation points, a 0.2-mm-diameter
hole was formed at the centres of the torsion-tested specimens (80 mm from the grip points) using electrical-discharge machining.
Fig. 1 shows that the torque from the as-drawn specimen increased slightly owing to work hardening after yielding
and that the specimen fractured at a strain of 0.34. The elastic limit was increased for the aged specimen compared with the as-
drawn specimen, which was attributed to the ageing. The aged specimen exhibited almost no work hardening after yielding, and
it exhibited a sudden decrease in torque caused by the delamination. The aged specimen fractured at a shear deformation of 0.49;
it was not destroyed immediately after the onset of delamination.
In order to investigate the mechanism underlying the delamination, the microstructures immediately beneath the
normal fracture surface and beneath the delamination crack were
observed. First, a torsion test was performed, which was terminated
immediately after the drop of the torque, and the microstructure
directly beneath the delamination crack was observed using BSE
images. The observed surface was a C-section, as shown in Fig. 2(a).
The delamination crack propagated along both the longitudinal and
radial directions of the specimen. Fig. 2(a) shows a C-section for which
the drawing direction was the depth direction. The cementite lamellae
exhibited complex shapes. For the cementite lamellae at the wake of
the delamination crack, the contrasts vanished. Transmission electron
microscopy revealed that the cementite lamellae that were beneath the
delamination crack[4] before the rotation were no longer there. This
suggests that the decomposition of cementite is closely related to the
propagation of delamination cracks. The microstructure beneath the Fig.1 Torque-shear strain curves from as-
drawn and aged specimens, where the
normal fracture surface was also examined, to determine whether the drawing strain was 2.2 [1].
decomposition of cementite in this manner occurred only in the area
References
[1] M. Tanaka, Y. Manabe, T. Morikawa, K. Higashida, ISIJ
International 60 (2020) 2596-2603.
[2] A.R. Marder, B.L. Bramfitt, Metall. Trans. A 7 (1976) 365-
372.
[3] D. Hirakami, T. Manabe, K. Ushioda, K. Noguchi, K. Takai,
Y. Hata, S. Hata, H. Nakashima, ISIJ International 56
(2016) 893-898. Fig. 3 State of stress at the specimen surface
[4] M. Tanaka, H. Saito, M. Yasumaru, K. Higashida, Scripta under the torsion as in the figure. (a) Shear
Mater. 112 (2016) 32-36.
stress under the torsion on the surface. (b)
Dominant shear stress causing normal
[5] M. Umemoto, Y. Todaka, T. Takahashi, P. Li, R. Tokumiya, K. fracture. (b) Dominant shear stress causing
Tsuchiya, Mater. Sci. Eng. A 375-377 (2004) 894-898. delamination crack propagation [1].
Kazuki Shibanuma
The University of Tokyo
Abstract:
This study presents a multiscale modelling strategy for accurately predicting the high-cycle fatigue
strengths of steels based on only microstructural information, tensile properties, and loading
conditions, without any adjustable material constants. In the proposed strategy, total fatigue life is
estimated from crack growth life alone. The entire model comprises three sub-models, for: (i) a
macroscopic finite element analysis, (ii) microstructure, and (iii) crack growth. The model was strictly
validated against the results of experiments performed on three different steels under various loading
conditions. Although the experimental fatigue life results exhibited wide variation, the predicted and
experimental data were accurately matched over the entire range. The results demonstrate that the
fatigue life of steels under high-cycle fatigue can be accurately predicted from crack growth life alone.
Furthermore, the proposed strategy is capable of effectively explaining the dependence of fatigue
strength on microstructure and loading conditions based on the fracture mechanics.
Keywords: steel, thermomechanical treatment, ultrafine grained microstructure, strength, delayed fracture
Warm tempforming is a thermomechanical treatment that plastically deforms tempered martensitic structure at an
elevated temperature. We have developed the warm tempforming process with multi-pass caliber rolling as a method to create
ultrafine elongated grain (UFEG) structure with a strong <110>// rolling direction (RD) fiber texture in medium-carbon
low-alloy steels. It has been demonstrated that warm tempformed (TF) steel bars with UFEG structures have been
demonstrated to exhibit excellent mechanical performance.1) In this presentation, we report on the delayed fracture behavior of
0.4%C-2%Si-1%Cr-1%Mo steel (mass %) with UFEG structure processed by warm tempforming. The delayed fracture
property was evaluated by comparing the relationship between the critical hydrogen concentration for
hydrogen embrittlement and the concentration of hydrogen intrusion from an atmospheric corrosive
environment. 2) Delayed fracture behavior of the ultra-high-strength bolts produced from the TF steels in long-term outdoor
exposure tests is also be presented. 1, 3)
The quenched bars for 0.4%C-2%Si-1%Cr-1%Mo steel were tempered at 500 °C for 1 h and subsequently warm
tempformed by multi-pass caliber rolling with a rolling reduction of 762)~78%3), followed by air cooling (TF sample). The
details of the warm tempforming process were described elsewhere.1-3) The TF samples were annealed 570 °C for 1 h, where
Mo-rich precipitates significantly precipitated, 2, 4) and then water cooled (TFA sample).
The microstructures were observed by FE-SEM and TEM. FE-SEM/EBSP analysis was carried out at a step size
of 50 nm. Tensile tests were performed at a crosshead speed of 0.85 mm/min for JIS Z 2201-14A specimens with a diameter
of 6 mm and gauge length of 30 mm. Representative nominal stress-strain curves and tensile properties are presented in Fig. 1
and Table 1, respectively. 2) Data for the sample quenched and tempered at 500 °C for 1 h (QT sample) are also shown for
reference. While the TF sample exhibited discontinuous yielding, the TFA and QT samples exhibited continuous yielding.
Total elongation (EL) and reduction in area (RA) of the TF and TFA samples were higher than those of the QT sample
although the tensile strength, σB, of these samples were almost the same at 1.8 GPa.
Fig. 2 Shape and dimension of notched bar specimen with a stress concentration factor, Kt, of 4.9
Fig. 3 shows the microstructures of the QT and TF samples. 2) The QT samples exhibit tempered
martensitic structure characterized by equiaxed prior-austenite grains consisting of packets that are further
subdivided by fine blocks. On the other hand, the TF sample exhibits a UFEG structure with a strong
<110>//RD fiber texture. It is difficult to distinguish the boundaries of the prior-austenite grains and packets.
Film-like intergranular carbides as observed in the QT sample are not observed. The carbide particles in the TF
sample are spheroidized and most of the intergranular carbides are aligned along the RD. These carbide
particles are mainly cementite. A similar UFEG structure was observed in the TFA sample. The transverse grain
size in the UFEG structure and cementite particle distribution were measured to be almost the same between
the TF and TFA samples. However, the precipitation state of the Mo-rich precipitates in the TF sample was suggested to
be different from that in the TFA sample because the hydrogen absorption capacity in the TFA sample was larger than that in
the TF sample.
Fig. 3 Inverse pole figure maps for the RD and TEM images of of the QT and TF samples.
Inverse pole figure for the RD of the TF sample is also shown.
at the notch. In all samples the σNB decreases with increasing the HD and hydrogen embrittlement is observed.
The σNB–HD relationship can be used to estimate the critical hydrogen content, HC, below which fracture does
not occur at the applied stress corresponding to σNB. Furthermore, the maximum concentration of diffusible
hydrogen that accumulated locally in the stress concentrated region near the notch root, HC* can be estimated
by substituting the HC value into the following equation:3-4, 5)
HC* = HC exp (–Δσh ΔVH/RT) (1)
where Δσh is the difference between the maximum hydrostatic stress in the vicinity of the notch root,
σm*, and the hydrostatic stress in the region far from the notch root, σp, ΔVH is the partial molar volume of
hydrogen in bcc–Fe (= 2×10−6 m3/mol), R is the gas constant (= 8.31 J/K/mol), and T is the testing temperature
(= 300 K). Fig. 5 shows the relationship between the peak value of the maximum principal stress, σmax*, in the
vicinity of the notch root and Hc*, calculated from the data points in Fig. 4.
Fig. 4 Relationship between the diffusible hydrogen Fig. 5 Relationship between the peak value of
content, HD, and notch tensile strength, σNB locally accumulated diffusible hydrogen
concentration, HC*, and the peak value of the
maximum principal stress, σmax*
On the other hand, the maximum concentration of diffusible hydrogen that accumulates locally in the
stress concentration region near the notch root from an atmospheric corrosive environment, HE*, can be
estimated by the following equation: 3-4, 5)
HE* = HE exp (–σm*ΔVH/RT) (2)
where HE is the maximum content of diffusible hydrogen intrusion from an atmospheric corrosive
environment. The respective HE values were estimated by CCT to be 0.3 mass ppm for the QT sample, 0.4 for
TF, 0.6 for TFA sample.
Fig. 6 shows the relationship between the applied stress, σa and the HC*/HE*. Based on this index, if
the value of HC*/HE* is 1 or less, it can be determined that a delayed fracture could occur. The HC*/HE* values
of the QT samples are less than 1 at σa of greater than 1.1 GPa, and thus the QT sample is prone to a delayed
fracture. By contrast, the TF and TFA samples are expected to be free from delayed fracture up to σa of 1.9 and
2.0 GPa, respectively. Therefore, it can be concluded that the delayed fracture resistance of the TF and TF
samples are significantly higher than that of the QT sample under the atmospheric corrosive environment,
considering the hydrogen uptake.
Ultra-high-strength bolts (JIS-M12-hexagon head bolts) with a σB of 1.7 and 1.8 GPa were produced
by warm forging the QT and TF steels, respectively.1, 3) Outdoor exposure tests of the bolted fasteners were
Fig. 6 Changes in HC*/HE* as a function of the Fig. 7 Changes in cumulative ratio of fracture
applied stress, σa. with exposure time
The warm tempforming using multi-pass caliber rolling at 500 ℃ resulted in the formation of an ultrafine
elongated grain (UFEG) structure with a strong <110>//RD fiber texture and a dispersion of fine spheroidized carbide
particles for 0.4%C-2%Si-1%Cr-1%Mo steel. When the warm tempformed (TF) sample was annealed at 570 ℃ for 1 h
(TFA sample), the hydrogen absorption capacity was enhanced significantly due to nanoscale Mo-rich precipitates in the
UFEG matrix. The results of the slow-strain-rate test and cyclic corrosion test showed that TF and TFA samples with UFEG
structures had a high delayed fracture resistance at an ultra-high tensile strength of 1.8 GPa in an atmospheric corrosion
environment. Excellent delayed fracture resistance was further demonstrated by the long-term exposure test for the
ultra-high-strength bolts processed by warm forging the TF steels.
Acknowledgement
The study was partly supported by grants from the JSPS KAKENHI Grant Number 19H02468.
References
1) Y. Kimura and T. Inoue: ISIJ Int. 60 (2020) pp.1108-1126.
2) Y. Kimura, T. Inoue and E. Akiyama: Mater. Sci. Eng. A 703 (2017) pp.503-512.
3) Y. Kimura, T. Inoue, T. Nakata and M. Fukuda: Trans. Jpn. Soc. Mech. Eng. 84 (2018) 17-00493 (1-15).
4) Y. Kimura, T. Moronaga and T. Inoue: ISIJ Int. 62 (2022) pp.377-388.
5) Guide book for evaluation of delayed fracture property of high strength bolts, JSSC Tech. Rep. No. 91, (2010).
Keywords: austenitic stainless steel, mechanical property, hydrogen embrittlement, nanostructure, plastic deformation
Austenitic stainless steels (ASSs) are one of the most important materials and widely utilized as basic a structural
material. Much attention has been drawn in the past decades to strengthen the ASSs, and various approaches have been
developed, such as grain refinement. To achieve grain refinement down to nanometer-size, heavy cold-rolling and
high-pressure torsion (HPT) straining were performed in SUS316LN ASS (316LN). The heavy cold-rolling of 316LN to
92 % reduction produced characteristic heterogeneous nanostructure composed of twin domains, shear bands and low-angle
lamellae. 1) The 316LN after the HPT straining for 20 turns has homogeneous nanostructure with grain sizes of several tens
of nanometers. 2) The heterogeneous and the homogeneous nanostructured 316LN specimens showed the tensile strengths
of 1.8 and 2.0 GPa with the large fracture elongation of about 15 %. The effect of these nanostructures on the hydrogen
embrittlement behavior was investigated. 2) Thermal desorption spectroscopy showed that the electrochemical hydrogen-
charged 316LN specimens with these nanostructures had almost the same concentration of diffusible hydrogen. While the
tensile properties of the homogenous nanostructure were severely affected by diffusible hydrogen, it had less impact on the
tensile properties of the heterogeneous nanostructure, as shown in Fig. 1. Hydrogen embrittlement is suppressed in the
heterogeneous nanostructure because crack propagation seems to be hindered by twin domains.
3000 3000
(a) (b)
Uncharged Uncharged
Nominal stress, σ / MPa
2000 2000
1500 1500
1000 1000
500 500
0 0
0.00 0.05 0.10 0.15 0.20 0.00 0.05 0.10 0.15 0.20
Acknowledgement
This study was conducted under the financial supported by the Japan Science and Technology Agency (JST) under Collaborative
Research Based on Industrial Demand “Heterogeneous Structure Control: Towards Innovative Development of Metallic Structural
Materials” (Grant No. JPMJSK1413).
References
1) H. Miura, M. Kobayashi, Y. Todaka, C. Watanabe, Y. Aoyagi, N. Sugiura and N. Yoshinaga: Scripta Materialia 133 (2017) 33.
2) F.J. Mweta, N. Adachi, Y. Todaka, H. Sato, H. Miura, M. Kobayashi, C. Watanabe and Y. Aoyagi: MRS Advances, 6 (2021), 682.
1 Graduate student, Tokyo Institute of Technology, 4259, Nagatsuta-cho, Midori-ku, Yokohama 226-8503, Japan
2 Department of Materials Science and Engineering, Tokyo Institute of Technology, 4259, Nagatsuta-cho, Midori-ku, Yokohama 226-
8503, Japan
*email: nakada.n.aa@m.titech.ac.jp
Keywords: martensitic transformation, Bain lattice correspondence, transformation-induced internal stress, Charpy impact testing
1 Introduction
In iron and steel, austenite (A) with face-centered cubic lattice (fcc) transforms to martensite (M) with body-centered cubic lattice (bcc)
or body-centered tetragonal lattice (bct), and the Bain lattice correspondence is established between them. Some of the authors[1] reported
that a small part of Bain strain remains even after lattice invariant deformation, and microscopic internal stresses are then distributed
among Bain groups in Fe–16mass%Ni alloy having lath martensite. It is thought that the stresses affect the mechanical properties of
martensitic steels. In this study, the {001} cleavage fracture under mode I condition was evaluated and its anisotropy was discussed in
terms of microscopic transformation-induced internal stress.
2 Experimental Section
An Fe–0.1mass%C–5mass%Mn alloy was used in this study. The test pieces were austenitized at 1373 K, followed by water quenching
to fully obtain as-quenched lath martensite. Charpy impact test was carried out at ambient temperature immediately after holding the test
piece at 77 K in liquid nitrogen. The crystal orientation was analyzed by means of electron backscattered diffraction (EBSD) technique.
At this time, martensite variants with <001>M roughly parallel to [100]A, [010]A, and [001]A of prior austenite were classified as Bain-1,
-2, and -3, respectively. Similarly to the previous study[1], in order to evaluate microscopic internal stress developed in lath martensite, a
cylindrical micropillar with 10 µm in diameter was fabricated within a Bain group of lath martensite using a field emission-type SEM
with a focused ion beam system (FIB/FE-SEM), and two-dimensional plane strain released by the fabrication was measured accurately
using digital image correlation method (DIC, GOM Collection Pro.).
3 Results
3.1 Transformation-induced Internal Stress in Lath Martensite
The FIB-DIC analysis revealed that the anisotropic elastic strain remained in a Bain group, and the average principal strains were
evaluated at e11 = e22 = –0.27% and e33 = 0.15%, where their directions are parallel to [001]M, [010]M and [100]M, respectively. According
to these values, it was calculated that the transformation-induced internal stresses were s11 = s22 = –744 MPa and s33 = –72 MPa, which
leads to a large stress deviation, s33–s11, ca.700 MPa. Considering that the same phenomenon was confirmed in the Fe–16mass%Ni
alloy[1], it should be thought that the transformation-induced internal stress attributed to the Bain strain is one of the characteristics of the
lath martensite regardless of the presence of solute carbon.
Acknowledgment
This work was supported by JST, PRESTO, Grant
Number JP20345123, Japan, and the Society of Hot Dip
Galvanized High Strength Bolt Engineering.
References
1) [1] D. Fukui, N. Nakada, S. Onaka: Acta Mater.,
196(2020), pp. 660-668.
Keywords: steel, martensite, cooling rate, yield strength, grain refinement, carbon partitioning
1. Introduction
Applications of ultra high strength martensitic steel sheets for automotive structural parts are expanding due to their
outstanding yield strength which strongly affects crash performance. The yield strength of martensitic steel sheets should be
influenced by many microstructural factors. Although cooling rate during martensite transformation is one of the basic control
in producing martensitic steel, but there are few researches have reported about its effect on microstructure or strength of
martensitic steel sheets. Therefore, the objective of this study is to clarify the effect of cooling rate on formation of
inhomogeneous microstructure leading to yield strength of martensitic steel sheets.
2. Experimental procedure
The tensile properties and microstructures were investigated using martensitic steels with chemical compositions of
Fe-C-0.2%Si-2.0%Mn-0.02%Ti-0.002%B (wt%) produced in different cooling rate during martensitic transformation.
Water-quenched (WQ) steels were produced by quenching into water, and air-cooled (AC) steels were produced by air
cooling after annealing in γ region. Low temperature tempering was performed in both steels. Microstructures were
investigated by means of SEM, EBSD analysis and high accuracy FE-EPMA (C-analyzer)1).
Introduction
The strength of as-quenched martensite increases with an increase of carbon content but untransformed
austenite retains in steels with carbon more than 0.6% because Mf temperature becomes lower than room
temperature1). On the other hand, in steels with high Ms temperature, auto-tempering occurs after
martensitic transformation and this leads to the decrease of hardness in as-quenched martensite2). Since
auto-tempering is attributed to the diffusion of carbon atoms during cooling, the degree of auto-tempering
can be evaluated by the integrated diffusion area; S (=∑Dt), where D and t are the diffusion coefficient of
carbon in bcc-Fe and holding time respectively. It is found that the degree of auto-tempering is so large in
steels with Ms temperature higher than 350℃3). In commercial martensitic steels, Ms temperature can be
estimated by the following equation as a function of chemical composition4).
The above equation suggests that auto-tempering is significant in steels with carbon less than 0.45%. In
this study, steels with various carbon content (S25C, S35C, S45C, S55C) were used to estimate the
degree of auto-tempering in as-quenched martensite and true hardness of as-quenched martensite without
auto-tempering was proposed for martensitic steels with carbon less than 0.6%.The amount of retained
austenite is so little that the effect of retained austenite may be negrected with respect to the hardness of
specimens.
Experimental procedure
Material used in this study was commercial structural steels (S25C, S35C, S45C, S55C) having
different carbon content from 0.27% to 0.55%. Chemical composition and transformation temperature of
steels are listed in Table 1. Transformation temperature was measured by dilatometry during cooling
(60℃/s) using an induction heat treating simulation apparatus. Specimens with the dimension of 16mm
diameter and 3mm thickness were prepared for Vickers hardness testing. The specimens were solution-
treated at 890 ℃ for 30min in Nitrogen atmosphere and then quenched into 5%-NaCl cold water.
Subsequently they were tempered for 60 min at the temperature in the range of 150℃ through 600℃.
Vickers hardness testing was performed on the cutting section (3×16 mm) of specimens under the load
2.9N. Measurement was performed 5 times and the average value was adopted as the hardness of each
specimen.
D0, Q, R are the frequency factor, activation energy in atomic diffusion, gas constant (8.3145 J/mol/K),
respectively. In the tempering reaction at A-D stages, it is sure that carbon diffusion relates the tempering
reaction. Regarding the diffusion of carbon in bcc-Fe, D0 =5.7×10-6 [m2/s] and Q = 91000 J/mol were
obtained from reliable data6). Since the value of D is given as a function of tempering temperature T (Eq.
2), the diffusion area S is calculated putting t = 3600s in the case of 1h tempering. The data of Fig. 1 are
re-plotted in Fig. 2 as a function of diffusion area S in the temperature range below 400℃. Regarding
S45C, a smooth curve is obtained against the S-value so that it was selected as a standard tempering curve,
that is expressed by the following equation.
HV = A×f(S) (4)
Hs = HV / HV* (6)
Fig. 5 Relations between diffusion area and the
It is interesting that all data follow an standardized hardness in martensite with 1h
identical curve in each steel except for tempering.
specimens with auto-tempering. Grange’s
In the case of quenching, S-value is given as an integrated diffusion area; ∑Dt based on a cooling
curve, thus the hardness of martensite with auto-tempering can be estimated applying Eq. 6 and 7.
Conclusion
It was confirmed that marked softening has occured in steels with carbon less than 0.4% even if water-
quenching has been done to obtain fresh martensite. True hardness of as-quenched martensite without
auto-tempering HV* is expressed by the follwing equation as a function of carbon content (%C).
HV* [GPa] = 2.0+7.8×(%C)1/2
The standardized hardness Hs (= HV / HV*) can be determined identically independent of carbon content
in steels under the same tempering condition. Regarding the martensitic steels with auto-tempering,
Vickers hardness can be estimated by the equation.
Hs = 1.0-0.0052×{18+(log S)}2
References
1) M.Maki:Materia Japan 54(2015) 557-563.
2) B.Hutchinson, J.Hagström, O.Karlsson, D.Lindell, M.Tornberg, F.Lindberg, and M.Thuvander:
Acta Materialia 59(2011) 5845-5858
3) S.Takaki, D.Akama and T.Tsuchiyama: Journal of the Japan Society for Heat Treatment, 56(2016)
340-344
4) N.Murai and T.Tsumura:Tetsu-to-Haganѐ 84(1998) 446-451
5) S.Nagakura:Denshikenbikyo 21(1986) 113-120
6) T.Kunitake: Materia Japan 3(1964) 466-476
7) R.A.Grange, C.R.Hribal and L.F.Porter:Metallurgical Transactions A, 8A(1977)1775-1785
8) Y.Imai and S.Sawada:Materia Japan 8(1969) 820-832.
9) H.Suto: Zairyou. 21(1972) 159-167.
10) M.Ueno, K.Nakamura, K.Ito and H.Minematsu: Tetsu-to-Haganѐ 68(1982) 155-161
11) G.R.Speich and P.R.Swann:J.Iron Steel Inst., 203(1965), 480
12) K.Sakuma, J.Kumagai and T.Nishizawa: Tetsu-to-Haganѐ 60(1974) 2153-2164
1. Introduction
Lath martensite in steel is a beneficial structure to design high strength steels, e.g. dual phase steels and hot-stamping steels.
Although it is important to understand strengthening mechanisms of lath martensite for improving these high strength steels,
these mechanisms are unclear, because of its complex and hierarchical microstructure1). The increasing of carbon content in
lath martensite significantly increases the strength2), and carbon would increase the strength via several routes, as shown in Fig.
1. Typically, the solution hardening effect of carbon is regarded as the most effective route3,4). Similarly, grain refinement
induced by carbon increasing would be important factor in lath martensite strengthening. Morito et al. reported that the
effective grain of lath martensite5) is a block comprising laths that exhibit similar crystallographic orientations and the same
habit plane, and that a block size6) decreases with carbon increasing. The strengthening of lath martensite with carbon content
increasing depends on solution hardening and grain refinement strengthening; however, their proportions are unknown. In this
study7,8), we focused on the hardening effects in ultra-low- and low-carbon lath martensite structure.
2. Experimental procedure
Table 1 shows chemical compositions and martensite transformation start temperature of four Fe–18Ni. These alloys contain
several carbon contents between 4 and 570 mass-ppm. Alloy A contains 0.017 mass% titanium to trap 4 mass-ppm carbon
and 8 mass-ppm nitrogen as huge Ti(CN) particles The ingots melted by a vacuum induction furnace were hot-rolled over
900 °C. The grinded hot-rolled sheets with 3.0 mm thickness were austenitized at 800, 900, 1000 °C for 1.2 ks and quenched
by water. We applied an additional heat treatment, holding at 500 °C for 10.8 ks to alloy A for the growth of Ti(CN) particles.
Martensite transformation in these alloys occurred from 300 °C to room temperature; thus these alloys quenched did not
contain retained austenite evaluated via X-ray diffraction.
We observed high angle boundaries, whose misorientations were larger than 10 degree via a field-emission electron
microscopy (JEOL-6500F) with electron backscattering diffraction analysis system (OIM Data Collection ver. 7), as effective
grain boundaries corresponding to block boundaries in lath martensite structure5,6).
We evaluated tensile properties via tensile test with JIS 13B tensile test specimens (JIS Z 2201). The strain gage length was
50 mm, and the nominal strain rate was 3.3 × 10-3 s-1.
3. Results
Fig. 2 shows high angle boundary maps of lath martensite. The decreasing of austenitizing temperature decreased the grain
diameter (deff), corresponding to the average spacing of high angle boundaries, because deff depends on the austenite grain
diameter. At each austenitization temperature, the increasing of carbon content decreased deff, because packets became divided
by multi blocks in high carbon lath martensite6). In this study, we got lath martensite structures whose deff were from 6 to 30
µm.
Fig. 3 shows the relationship between deff and tensile properties, namely, yield point (YP), 0.2% proof stress (σ0.2%), 0.6%
proof stress (σ0.6%), and maximum tensile strength (TS) in each alloy. YP and TS correspond to the elastic limit and the onset
of plastic instability, respectively. In ultra-low-carbon alloy, Alloy A, YP, σ0.2%, σ0.6%, and TS did not depend on deff. These
stresses were completely constant in three samples austenitized at several temperatures. On the other hand, in Alloy D
containing 570 mass-ppm carbon, these stresses increased with deff refinement. This strengthening behavior seems to be
Hall-Petch relationship, σ = k·deff–1/2 + A5). Although this relationship similarly appeared in Alloys B and C, the strengthening
with grain refinement is clearer with the increasing of carbon content.
4. Summary
In this study, we evaluated the tensile properties of as-quenched martensite on ultra-low- and low-carbon Fe–18Ni alloys.
1. Tensile properties (YP, σ0.2%, σ0.6%, and TS) were constant in ultra-low-carbon alloy specimens containing several effective
grain size.
2. YP, σ0.2%, σ0.6%, and TS increased with grain refinement in low-carbon alloys. This strengthening behavior corresponded to
Hall-Petch relationship, and became clear with the increasing of carbon content.
3. Hall-Petch coefficient was almost zero in ultra-low-carbon alloy, and increased with the increasing of carbon content. This
behavior is similar to Hall-Petch coefficient for YP in polygonal ferritic steel.
4. The strengthening with carbon in low-carbon Fe–18Ni lath martensite mainly depends on the grain refinement
strengthening. The increasing of carbon content refined the effective grain, and increased the effectivity of grain refinement
strengthening.
References
1) E.C. Bain, H.W. Paxton : Alloying Elements in Steel, 2nd Ed., ASM, 1961, 33–36.
2) S. Morito, H. Saito, T. Ogawa, T. Furuhara, and T. Maki : ISIJ Int., 45, 2005, 91–94.
3) J.M. Chilton, and P.M. Kelly : Acta Metall, 16, 1968, 637–656
4) G. Krauss : Mater. Sci. Eng. A, 273–275, 1999, 40–57
5) S. Morito, H. Yoshida, T. Maki, and X. Huang : Mater. Sci. Eng. A, 438–440, 2006, 237–240
6) S. Morito, H. Tanaka, R. Konishi, T. Furuhara, and T. Maki : Acta Mater., 51, 2003, 1789–1799
7) H. Kawata, Y. Honda, and K. Takeda : Minerals, Metals and Materials Series, TMS2019 148th Annual Meeting &
Exhibition Supplemental Proceedings, 2019, 529–535
8) H. Kawata, Y. Honda, K. Nakano, K. Takeda, K. Hikida : ISIJ Int., 62, 2022, in press.
9) K. Takeda, N. Nakada, T. Tsuchiyama, and S. Takaki : ISIJ Int., 48, 2008, 1122–1125.
10) S. Takaki, D. Akama, N. Nakada, and T. Tsuchiyama : Mater. Trans. JIM, 55, 2014, 28–34
Fig. 1 Schematic flow of the strengthening mechanism by carbon in lath martensite structure.
Ms : martensite transformation start temperature, σγ : strength of prior austenite during martensite transformation,
ΔV : volume expansion during martensite transformation
Fig. 2 High angle boundary maps of alloys A and D samples austenitized at 800 and 1000 °C.
black line : high angle boundary (θ : ≥ 10 degree), gray line : low angle boundary (θ : 5~10 degree)
Fig. 3 The relationship between inverse square root of the average effective grain diameter and tensile properties in lath
martensite structure. YP : yield point (elastic limit), σ0.2% : 0.2% proof stress, σ0.6% : 0.6% proof stress,
TS : maximum tensile strength (onset of plastic instability)
Fig. 6 Modified schematic flow of the strengthening mechanism by carbon in lath martensite structure.
Ms : martensite transformation start temperature, σγ : strength of prior austenite during martensite transformation,
ΔV : volume expansion during martensite transformation
Keywords: steel, aluminum alloy, metal cation, corrosion, surface analysis, electrochemistry
Structural metals such as carbon steels, stainless steels, and aluminum alloys are suffer from corrosion during usage.
The corrosion resistance of the structural metals in aqueous and atmospheric environments depends on the stability of passive
film. However, the passive films on the metals are destroyed by the chloride ions in chloride containing environments. It is
recognized that chloride ions are very much aggressive and destroy the oxide films1, 2), therefor, many researches have been carried
out focused on effects and role of chloride ions on corrosion behavior of structural metals3-5). The corrosion rate of structural
metals is changing in despite of similar concentrations of chloride ions in fresh water6-8). This may be due to difference of contained
metal cations. Some researchers stated that the corrosion inhibition of zinc and carbon steels is attained by the kind of metal
cations in aqueous medium at room temperature9-12). Tsuru13) also reported that Zn2+ dissolved from corrosion products on zinc
coated steel prevents further corrosion of steel substrate. The same observations14) have noticed while mild steel is exposed in the
aqueous solution with Zn2+. However it is not fully elucidated the metal cation effects on corrosion of structural metals. In this
presentation, recent authors group results related to effects of metal cations on corrosion behavior of carbon steel, stainless steel
and aluminum alloys will be explain.
Aluminum alloys
Aluminum alloys such as A3003 and A6061 are widely used because of the high strength/weight ratios, and they show
good corrosion resistance in most environments. As previously mentioned, the good corrosion resistance is explained by
protective oxide films, passive films, covering the aluminum alloy surface15-17). Generally, the corrosion rate of aluminum alloys
also depends on the concentration of chloride ions and it is considered that only little corrosion of aluminum alloys occurs in very
dilute chloride solutions such as tap waters. There are only few studies focused on corrosion of aluminum and aluminum alloys
in fresh water. However, corrosion of aluminum alloys sometimes does occur in fresh water and this causes leakage of water
when the alloy is used in containers and water supply tubes. As well know, fresh water contains very low concentrations of anions
and metal cations. It is considered that corrosion is affected by not only chloride ions but also metal cations, which are in the fresh
water. Otani et al. reported influences of metal cations in model fresh waters on the corrosion mechanisms of A3003 by metal
cation hardness theory and XPS surface analysis18, 19). Neutral model fresh waters used in the paper was 0.5 kmol m−3 H3BO3−
0.05 kmol m−3 Na2B4O7 with each one of six different salts; 1 mol m−3 KCl, 1 mol m−3 NaCl, 0.5 mol m−3 CaCl2, 0.5 mol m−3
MgCl2, 0.5 mol m−3 ZnCl2, and 0.5 mol m−3 NiCl2. Here, the metal cation hardness is based on the hard and soft acid and base
(HSAB) concept20). According to the Lewis rule concerning acid and bases, cations and anions act as acids and bases, respectively.
Acids and bases are classified into hard and soft according to the HSAB concept. The hardness of metal cations can be expressed
as follows :
" &
#X! + (ΣI# )$/& ) ,
X= 10
0
Where X M is the electronegativity of the metal atom, and In (eV) is the ionization potential from the neutral metal atom to the
given oxidized state, n. They withdraw the following results. 1) Corrosion pits of about 10 μm in diameter were observed in
solutions with soft metal cations such as Na+ and K+. The corrosion behavior of A3003 in the model tap waters changed with
metal cation hardness, X. 2) The corrosion potential increases and corrosion current decreases with increasing metal cation
hardness. 3) Hard metal cations were observed by XPS in the oxide films formed on A3003 after immersion in the model tap
waters containing these cations.
From several researches related to corrosion of A300318, 19, 21), it was established that Zn2+ is superior to inhibit the fresh
water corrosion of A3003 among the metal cations used in the studies. However, the minimum concentration of Zn2+ to inhibit
the corrosion of A3003 in model fresh water has not been determined. Establishing the minimum concentration of metal cations
Carbon steels
Fresh water corrosion of mild steel has been investigated by many researchers because the corrosion has caused serious
problems in industries. Important factors for the fresh water corrosion of mild steel are the corrosive nature of dissolved oxygen
and the concentrations of anions such as Cl- and SO42-. On the other hand, there are reports that Zn2+ and Al3+ have corrosion
inhibiting ability24, 25). For this reason, the effect of metal cations on the corrosion of mild steel should be considered in fresh water.
However, there has been no study in which the quantitative effects of all metal cations on corrosion of mild steel in model fresh
water were investigated. The effects of metal cations on corrosion of mild steel in model fresh water were investigated by EIS,
immersion tests, and surface observation and analysis with SEM and XPS26). Following results were drown. The corrosion rate
of mild steel decreased with increase in the hardness of cations, X, in model fresh water except for solutions containing Mg2+ and
Zr2+. Therefor, The novel indicator, Y = X × ∆V, was introduced to explain the different degrees of protection of metal cations.
Where X is metal cation hardness and ∆V is molar volume difference between passive film and hydroxides of metal cations. XPS
analysis showed that hard metal cations are incorporated in the passive films of the mild steel, and hydroxides of metal cations
were formed on the mild steel after immersion for a short time. Hydroxides of metal cations that have large ∆V protect the passive
films of mild steel from penetration by chloride ions.
The effects of metal cations on the corrosion behavior of carbon steel SM490Y in 10 molm-3 Cl- model fresh water
were also investigated by immersion tests, electrochemical tests, and surface analysis27). In this experiment, three different metal
cation containing solutions were used as 10 molm-3 NaCl, 5 molm-3 MgCl2 and 5 molm-3 ZnCl2. Immersion tests and
electrochemical tests results showed that Zn2+ had a significant inhibition effect on corrosion of SM490Y. The results of EDS and
XPS showed that Zn2+ was incorporated in the oxide films by making a strong bond and formed zinc hydroxide or zinc oxide
which can cover the surface of carbon steel. It is suggested that Zn2+ forms a protective film which can efficiently prevent break
down of surface film by Cl- and inhibit the corrosion.
Stainless steel
SUS304 is widely used in applications requiring high corrosion resistance. However, corrosion of stainless steel is a
serious problem in constructions such as liquid supply pipes and storage tanks, ships, and bridges. Corrosion of stainless steel
depends on the various environmental factors. One of the important factor is presence of anions (Cl- and SO42-) in the aqueous
solution. Cl- plays a major role in corrosion of stainless steel because the passive films are easily destroyed by this anion. The
effects of metal cations on corrosion of SUS304 in 0.5 kmolm-3 Cl- solution remain unknown and the corrosion inhibition
mechanism of metal cations with the passive films are not fully elucidated. The effects of metal cations on corrosion of SUS304
in 0.5M Cl- aqueous solution were investigated by immersion tests, electrochemical tests, SEM, atomic force microscope (AFM)
and XPS. In this study, three 0.5 kmolm-3 Cl- solutions with different metal cations were used as test solutions; 0.1 kmolm-3 MgCl2,
0.1 kmolm-3 ZnCl2, and 0.1 kmolm-3 AlCl3, and their Cl- concentration was adjusted to 0.5 kmolm-3 by NaCl. Therefor, 0.5 kmolm-
3
NaCl was used as a reference solution. The corrosion behavior of the stainless steel in the 0.5 kmolm-3 Cl- aqueous solution was
changed with X, and corrosion rates were closely correlated with X. Different surface morphologies were observed by SEM after
immersed in the solutions which were depended on the metal cations. Zn2+ and Al3+ contained solutions showed higher impedance
and higher charge transfer resistance as compared to other solutions. From the XPS results, Zn2+ and Al3+ were existed on the steel
as hydroxides and formed a layer by incorporating in the passive films through de-hydroxylation process. The layer of metal
cation would have good protecting ability against Cl- attack as well as corrosion reactions on the steel surface.
The corrosion behaviors of SUS 310 in model seawater (0.1 kmolm-3 MgCl2, 0.1 kmolm-3 ZnCl2, and 0.1 kmolm-3
AlCl3, and their Cl- concentration was adjusted to 0.5 kmolm-3 by NaCl) with different metal cations also investigated by
gravimetric tests and surface analysis31). Different corrosion rates were observed in different metal cationic solutions, and Zn2+
containing solution showed the lowest corrosion rate among used solutions. SEM observations showed the pits on the surface
that were different from each other. White pits were observed on the specimen immersed in the Zn2+ containing solution. XPS
results showed the presence of Zn2+ and Al3+ on the specimen immersed in the corresponding solution. Metal cations that have
large X, showed the tendency to form a fence layer on the specimen surface. The layer of metal cation would have good protecting
ability against Cl- attack as well as corrosion reactions on the steel surface.
Acknowledgement
Surface analysis of this work was conducted at Laboratory of XPS analysis, Hokkaido university, supported by “Nanotechnology
Platform” Program of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
References
1) R. T. Foley, Corrosion, 26 (1970) 58-70 .
2) D. D. Macdonald, J. Electrochem. Soc., 139 (1992) 3434-3449 .
3) M. Abdallah, E. A. Helal, and A .S. Fouda, Corros. Sci., 48 (2006) 1639-1654.
4) T. Balusamy and T. Nishimura, Electrochim. Acta, 199 (2016) 305-313.
5) S. Zhang, L. Hou, H. Du, H. Wei, B. Liu, and Y. Wei, Corros. Sci., 167 (2020) 108531.
6) Japan Aluminium Association edited, Aluminum Handbook, 7th ed. (2007) 73.
7) R. Lindström, J. E. Svensson, and L. G. Johansson, J. Electrochem. Soc., 149 (2002) B57-B64.
8) T. Prosek, D. Thierry, C. Taxén, and J. Maixner, Corros. Sci., 49 (2007) 2676-2693.
9) G. K. Gomma, Materials Chem. Phys., 55 (1998) 131-138.
10) J. Telegdi, M. M. Shaglouf, A. Shaban, F. H. Ka'rma'n, I. Betro'ti, M. Mohai, and E. Ka'lma'n, Electrochim Acta, 46 (2001) 3791-
3799.
11) T. Prosek, D. Thierry, and C. Taxe'n, and J. Maixner, Corros Sci, 49 (2007) 2676-2693.
Department of Engineering, Graduate School of Engineering, Nagoya Institute of Technology, Nagoya, Aichi 466-8555, Japan
*email: hoshi.yoshinao@nitech.ac.jp
Keywords: steel, electrochemical impedance spectroscopy, 3D impedance, chromate film, film growth rate
An improvement of corrosion resistance of steel is required for the long-term use in the structural materials. The
chemical conversion coatings are one of the most effective methods for surface modification of steel substrates. In these
methods, the chromate conversion treatment is useful method to improve the corrosion resistance of steel. It is well known
that the dissolution from the base steel due to the scratched or defective parts on the chromate film can be suppressed by the
self-healing effect of chromate film without further treatment. As the development of a new chemical conversion coatings,
which shows self-healing effect, has been required for further improvement of the corrosion resistance of steel in recent years,
it is important to clarify the mechanisms related to self-healing effect of chromate film. Thus, the aim of the present study was
to investigate the time variation of film resistance of chromate film related to the self-healing effect on the steel surface by a
three-dimensional (3D) impedance spectroscopy 1,2) .
An electrochemical impedance spectroscopy (EIS) is useful method to investigate dissolution mechanisms of steel
because the time constant observed in the impedance spectrum can be discriminated. The parameters involving film resistance
can be estimated by the curve-fitting of the measured impedance spectrum using an equivalent circuit. In this method, a time
stability is required in the electrode condition, indicating no variation of electrode reaction rate during the impedance
measurement. As the electrode reactions related to the self-healing effect of chromate film on the steel do not satisfy the time
stability, the distortion of impedance spectrum is observed in the low frequency range on the Nyquist plane. In this case, it is
difficult to estimate the film resistance related to the self-healing effect of the chromate film formed on the steel surface by the
curve-fitting of the measured impedance spectrum using the equivalent circuit.
We applied the 3D impedance spectroscopy 3-5) for investigation of the self-healing effect of the chromate film
formed on steel surface. In the impedance measurement, the impedance spectra were measured successively, and the
impedance was plotted on the 3D diagram whose axes are real, imaginary and time. The plots of impedance are connected by
the spline under tension function at each frequency. Because the instantaneous impedance at an arbitrary time can be obtained
from 3D diagram, the parameters involving film resistance can be estimated by the curve-fitting of the instantaneous
impedance at an arbitrary time using an equivalent circuit. We employed the 3D impedance spectroscopy for the investigation
of the time variation of the film resistance related to the self-healing effect of chromate film formed on the steel surface 1,2). We
prepared the scratched electrodes of chromated steel and chromated galvanized steel for the 3D impedance measurement in
order to examine the self-hearing effect of the chromate film. As we could obtain the instantaneous impedance at arbitrary
time from 3D impedance plot, the film resistance of each scratched electrode could be estimated by the curve fitting of the
instantaneous impedance using the equivalent circuit. I will present the 3D impedance of different types of scratched
electrodes of chromated steel and chromated galvanized steel, and I will discuss the film growth rate of chromate film owing
to self-hearing effect based on the time variation of film resistance estimated from instantaneous impedance at arbitrary time.
References
1) Y. Hoshi, Y. Endo, I. Shitanda, M. Itagaki, Y. Suzuki and S. Yamaguchi: J. Surf. Finish. Soc. Jpn., 67, (2016)
494-498.
2) Y. Hoshi: J. Surf. Finish. Soc. Jpn., 72, (2021) 462-466.
3) Z. B. Stoynov and B. S. Savova – Stoynov: J. Electroanal. Chem. 183 (1985) 133-144.
4) M. Itagaki, N. Kobari, S. Yotsuda, K. Watanabe, S. Kinoshita, M. Ue: J. Power Sources 135 (2004) 55-261.
5) M. Itagaki, K. Honda, Y. Hoshi, I. Shitanda: J. Electroanal. Chem. 737 (2015) 78-84.
Keywords: carbon steel, corrosion resistance, martensite, polarization measurements, first-principles calculations
3. Results
In the potentiodynamic polarization curves of the low and high carbon martensitic specimens in 1mM NaCl
containing boric-borate buffer at pH 8.0, the region of active dissolution appeared on both specimens at about -0.5 V. The
active dissolution current of high carbon martensite was lower than that of low carbon martensite, indicating that active
dissolution was suppressed by interstitial carbon.
Next, to clarify the reason for the superior dissolution resistance obtained by the interstitial carbon, first-principles
calculations were conducted, and the effect of interstitial carbon on the electronic density of states (DOS) of iron was analyzed.
It has been widely known that the valence electrons at Fermi level are readily transferred among ions and metals during
electrochemical reactions. Therefore, the DOS at and near the Fermi level is important for determining the electrochemical
reactivity 6, 7 ). In other words, it is understood that a high DOS at and near the Fermi level increases the electrochemical
reactivity. Since the active dissolution of steels is the oxidation reaction of iron, the DOS at and near Fermi level is one of the
important parameters for determining the resistance against active dissolution. The calculation results clearly showed that the
DOS of the iron near and at the Fermi level decreased with the increases in the interstitial carbon content. That is, it is highly
likely that the electrochemical reactions such as the active dissolution are considered to be suppressed by interstitial carbon.
4. Conclusions
1) According to the polarization measurements, the interstitial carbon suppressed the dissolution reaction of
martensite.
2) First-principles calculations showed that the electronic density of states (DOS) of iron near the Fermi energy
decreased with increasing interstitial carbon.
3) The change in the electronic structure of iron appeared to be related with the superior dissolution resistance
provided by interstitial carbon.
Acknowledgement
This work was performed under the support of a Grant-in-Aid for JSPS Research Fellow (grant No. JP18J20518). The
authors acknowledge the Center for Computational Materials Science of the Institute for Materials Research, Tohoku University for
the support of the supercomputing facilities.
References
1) F. J. Martin, P. M. Natishan, E. J. Lemieux, T. M. Newbauer, R. J. Rayne, R. A. Bayles, H. Kahn, G. M. Michal, F. Ernst,
and A. H. Heuer: Metall. Mater. Trans. A 40A (2009) 1805.
2) A. Chiba, M. Koyama, E. Akiyama, and T. Nishimura: J. Electrochem. Soc. 165 (2018) C19.
3) M. Kadowaki, I. Muto, Y. Sugawara, T. Doi, K. Kawano, and N. Hara: J. Electrochem. Soc. 164 (2017) C962.
4) G. Kresse and J. Furthmüller: Phys. Rev. B, 54 (1996) 11169.
5) G. Kresse and D. Joubert: Phys. Rev. B, 59 (1999) 1758.
6) P. C. Searson, P. V. Nagarkar, and R. M. Latanision: Int. J. Hydrogen Energy 14 (1989) 131.
7) L. M. C. Pinto, P. Quaino, E. Santos, and W. Schmickler: ChemPhysChem 15 (2014) 132.
*email: yuse.fumio@kobelco.com
Keywords: bridge, corrosion, life cycle cost, coating, long-life steel sheet for coating
1. Introduction
In recent years, in the field of bridges, the recognition of maintenance, painting, and anticorrosion has been
increasing, and in addition to the reduction of initial construction cost, the reduction of maintenance costs and life cycle cost
(LCC) is required 1,2). In steel bridges, coating is carried out from the viewpoint of corrosion protection and appearance, but it
is necessary to repaint the coating because the coating film deteriorates over time and the corrosion protection performance
declines. Even in heavy anticorrosion coating systems, which have a longer coating life than general coating systems, the
period until the anticorrosion performance deteriorates is shorter than the service period of the bridge. Therefore, repainting
must be performed during the service period to maintain the anticorrosion performance. Therefore, prolonging the repainting
cycle is effective for LCC reduction.
In the examination of LCC, it is necessary to clarify the criteria for the degradation level of coating film and the
recoating time for the required performance. It is said that the repainting time will be decided by conducting a step-by-step
evaluation of each item of rust, peeling, discoloration and staining by inspection 3). Various factors are involved in the
deterioration of the coating film of actual steel bridges, and the deterioration progresses. It is said that local corrosion as well as
uniform corrosion occurring on the surface of the steel plate should be taken into consideration. Therefore, in many cases,
standards are set by each entity that conducts maintenance management. In addition, it is often difficult to continuously
observe the deterioration of coating film and subsequent corrosion evaluation of steel over a long period of time.
In this study, as a consideration of LCC, we report the results of corrosion resistance evaluation of long-term
exposure test for 15 years and examination of repainting life using ordinary steel and developed steel 4) (long-life steel sheet
for coating, which satisfies all the relevant JIS standards for conventional welded structural steel (JIS G 3106; SM) and has a
new function of inhibiting corrosion under the coating film added to the steel itself).
2. Experimental procedure
Coating is a corrosion prevention method that protects steel by blocking corrosion-promoting substances such as
oxygen, water and chloride ions from the coating film formed on the steel surface. As time passes, the coating film
deteriorates and its corrosion protection performance declines. In other words, the degree of deterioration, which is a corrosion
index, increases with the number of years passed. When the index reaches a certain level, repainting should be considered (Fig.
1). Rusting, peeling, and blistering are used as indicators of deterioration in the corrosion protection performance of the
coating film, and the area ratio of rusting, peeling, and blistering is used as an index.
In general, the corrosion of steel bridges tends to start from the scratches and corners of the coating, and it has been
reported in some cases that the initial defects such as scratches and partial lack of coating thickness are the starting point for
the corrosion of heavily corrosion-protected coating systems, rather than the deterioration from the coating surface. In this
study, blister width was used as a corrosion index of coated steel.
Small specimens of ordinary steel and developed steel were placed under the girder of a bridge in Kobe City,
Hyogo Prefecture, and exposure tests were conducted (Fig. 2). The small specimens (150 x 70 x 6 mm) were sealed on the
back and sides with tape and coated with the old C-4 coating system (75 μm inorganic zinc, 120 μm epoxy resin, and 55 μm
fluorocarbon resin for a total of 250 μm). After curing, artificial coating defects were imparted with a box cutter knife to
simulate the scratched areas where rust tends to spread. As a comparison, a bare (unpainted) specimen was also placed on the
same exposure stand.
4. Conclusion
A 15-year exposure test was conducted on an actual bridge. The results show that the developed steel, a steel plate
for bridges with long-life coating, exhibits excellent resistance against coating corrosion compared with ordinary steel. The
results, including the results of rust analysis, indicate that the mechanism of inhibiting under-film corrosion is working as
expected. The repainting life of the developed steel is considered to be about 1.3 times longer than that of the ordinary steel at
present, so it is expected to enable re-painting cycles to be lengthened and reduce the life-cycle cost of steel bridges.
References
1) Japanese Society of Steel Construction: JSSC Technical Report, No.55(2002) pp.24-41.
2) T. Tamakoshi : Report of General Studies on Land and Infrastructure Management. Ministry of Land, Infrastructure,Transport and
Tourism. No.294,(2006) p.1-3.
3) Japanese Society of Steel Construction : Heavy Duty Coating (Gihodo-books Japan, 2012)pp.153-163.
4) F.Yuse, M Matsushita and M Izumi: KOBELCO TECHNOLOGY REVIEW NO. 34 (MAR. 2016) pp.6-11.
5) T. Ishikawa: Zairyo-to-Kankyo. Vol.52, No.3 (2003), p.140-154.
6) F.Yuse: Rust Rrevention & Control Japan, Vol.63, No.6 (2019) pp.22-29.
Fig. 2 Example of installation location Fig. 3 Relationship between exposure years and Fig. 4 Relationship between exposure
and exposure test conditions. blister width of ordinary steel and developed steel. years and thickness loss.
Research Center for Structural Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan
email: TSUTSUMI.Yusuke@nims.go.jp
1. Introduction
In the field of materials engineering, physical properties such as mechanical properties are considered as the most
important properties, however, chemical properties such as corrosion resistance are also critical properties that cannot be
ignored. In particular, Japan is island country where metallic materials are frequently used in coastal areas. Therefore, the
problems of failures and accidents caused by corrosion reaction of structural materials in our countries are more serious than
in other countries. The author and the researchers in our group have been studying the corrosion evaluation and the
development of techniques to improve the corrosion resistance of biomaterials and structural materials. In this presentation,
the efforts of these researches will be introduced.
Fig. 1 Schematic illustration of the electrochemical surface treatment performed in this study1)
Fig. 3 Surfaces of 316L stainless steels prepared from commercially available sheet and LPBF process observed via laser
microscopy (upper) and SEM (lower) 2)
5. Conclusions
The present study demonstrated two-individual techniques for improvement of metallic materials: the
electrochemical treatments comprising anodic and cathodic polarization cycles for removement of inclusions, and LPBF
process for fabricating inclusion-free metallic structures. Both techniques are highly effective in inhibiting localized corrosion
in chloride ion environments. In other words, inclusions on the surface and inner matrix of the materials are key factor in the
occurrence of localized corrosion. If such inclusions can be selectively removed, the materials can reclaim their original and
intrinsic corrosion resistance. Authors believe that these techniques will be utilized in the establishment of a sustainable
societies in future.
References
1) Manaka T, Tsutsumi Y, Ashida M, Chen P, Katayama H, and Hanawa T: Mater. Trans.. 62 (2021) 788-796.
2) Tsutsumi Y, Ishimoto T, Oishi T, Manaka T, Chen P, Ashida M, Doi K, Katayama H, Hanawa T, and Nakano T: Additive
Manufacturing 45 (2021) 102066.
Keywords: steel rebar, concrete, corrosion, mill scale, hyperbaric-oxygen accelerated corrosion test
1. Introduction
Concrete structures containing steel rebar are widely used for large-scale constructions such as buildings, bridges and
highways. Concrete structures often lose their long-term reliability by corrosion of steel rebar. Nowadays, many concrete
structures require repair or reconstruction because of their aging. Generally, in concrete, alkalinity of the environment is
helpful for reinforcing steels to form the passive films and to exhibit superior corrosion resistance. Therefore, it is difficult to
demonstrate corrosion of reinforcing steel in concrete in the laboratory in a short time. A new method for acceleration of
corrosion of reinforcing steel is necessary.
Recently, some researchers tried to accelerate the corrosion of steel rebar in concrete by enhancement of metal dissolution.
On the other hand, we developed a novel accelerated corrosion test method (Hyperbaric-oxygen Accelerated Corrosion Test,
HOACT)1). In the HOACT, the oxygen reduction reaction which is the rate-limiting process of steel rebar corrosion is
enhanced by preparing the pressurized oxygen gas, and the corrosion of steel rebar in concrete is accelerated.
In this presentation, the effect of oxygen gas pressure and Cl- concentration in mortar on the corrosion of iron in mortar
was examined. In addition, the role of mill scale on the corrosion of steel rebar in mortar and the mechanism of steel rust
formation in mortar was also examined.
2. Effect of oxygen gas pressure and Cl- concentration in mortar on the corrosion of iron in mortar2)
A pure iron (99.5 %) coupon embedded in mortar with cover thickness of 5 mm was used for HOACT. The mortar was
produced by mixing of ordinary Portland cement paste and fine aggregates with a weight ratio of water-cement to fine
aggregates of 0.6:1:3. In the mortar, NaCl solution was used as mixing water to induce breakdown of the passive film on the
sample. Cl- concentration of mixing NaCl solution was 0, 0.6, and 2.0 M. In the HOACT, the sample was exposed to a humid
pressurized O2 gas or ambient air at over RH 95% in the high-pressure container. The O2 gas pressure was 0.1, 0.4, 0.6, 1.1,
and 2.1 MPa. After 14-day HOACT, the sample was extracted from mortar, and surface of the sample was observed using the
optical microscope. The cross section of the samples was characterized using scanning electron microscope (SEM) by
detecting back scattered electron.
The sample was hardly corroded in ambient air without Cl- because the passive film protected the sample surface. The
sample with 0.6 M Cl- was corroded in 0.1 and 0.4 MPa pressurized O2 gas. In addition, the corrosion on the sample with 2.0
M Cl- was obviously accelerated in 0.1, 0.4 0.6, and 1.1 MPa pressurized O2 gas. The passive film breakdown and metal
dissolution were enhanced by Cl- and the oxygen reduction was enhanced by supplied pressurized O2 gas. However, corrosion
of the sample was little observed in case of containing 0.6 M Cl- over 0.4 MPa O2, and 2.0 M Cl- in 2.1 MPa O2 contrary to
expectation. The thickness of rusts of the sample containing Cl- was increased with an increase in oxygen pressure until it
reached to a peak value, then the thickness of rusts adversely decreased over the peak value of oxygen pressure. The peak
value of oxygen pressure increases with an increase in the amount of Cl-. With 2.0 M Cl-, the thickest rust was formed in 0.6
MPa O2 gas. Up to 0.6 MPa, the thickness of rusts increased in proportion to the oxygen pressure. This result indicates that the
HOACT has an adequate oxygen pressure to efficiently enhance the corrosion of sample in the case of relying on Cl- addition
to mortar to breakdown of passive film.
Acknowledgement
This work was performed under the support of Cross-ministerial Strategic Innovation Promotion Program (SIP),
“Infrastructure maintenance, renovation and management” (Funding agency: JST) and Leading Initiative for Excellent Young
Researchers (Funding agency: JSPS).
References
1) K. Doi, S. Hiromoto and E. Akiyama: Materials Transactions, 59(6) (2018) 927-934.
2) K. Doi, S. Hiromoto, H. Katayama and E. Akiyama; Journal of the Electrochemical Society, 165(9), (2018) C582-C589.
3) K. Doi, S. Hiromoto, T. Shinohara, K. Tsuchiya, H. Katayama and E. Akiyama; Corrosion Science, 177 (2020) 108995.
Dept. of Mater. Sci. Engr., Hanbat National University, Daejeon, 34158, Republic of Korea
*email: ktpark@hanbat.ac.kr
Keywords: TWIP steel, Notch tensile tests, Plastic localization, Dimple fracture, Stress triaxiality
The two most appreciable fracture characteristics of Fe-Mn-C(-Al) twinning induced plasticity (TWIP) steel under
uniaxial tension are the near-absence of diffused necking and the fracture surface entirely consisting of fine dimples. These
characteristics are quite at odd in light that fracture of usual ductile austenitic steels such as austenitic stainless steel exhibits
considerable necking and relatively large dimples. TWIP steel and austenitic stainless steel possess relatively high stain
hardenability leading to remarkable combinations of elongation and tensile strength. The dominant mechanisms of strain
hardening and deformation of both steels are either similar or different depending on the stacking fault energy, i.e. transformation
induced plasticity (TRIP) or TWIP or planar glide. TWIP steel usually has the higher C content than austenitic stainless steel.
Accordingly, the unique fracture behavior of TWIP steel is thought to be closely related to mechanical twinning and its higher C
content. In the presentation, the fracture characteristics of TWIP steel under uniaxial tension reported in literature are reviewed
and discussed.
Macroscopic fracture
Macroscopic features of fracture of the smooth specimen of an Al-bearing TWIP steel (Fe-15Mn-0.65C-2Al) plate
under room temperature uniaxial tension (the nominal strain rate of 10-3 s-1) are shown in Fig. 1. The steel was annealed after cold
rolling and the grain size was ~25 m. A consecutive series of digital image correlation images showing effective strain
distribution at the moment of fracture (Fig. 1a) reveals (a) uniform deformation without PLC bands up to fracture, (b) near-
absence of diffused necking, (c) the highest effective strain at the center of the gage width at the fracture area, and (d) fracture
normal to the loading direction. Besides, the thickness-wise slant fracture is evident (Fig. 1b) and shallow surface cracks mostly
normal to the loading direction are observed near the fracture front (Fig. 1c). These features reveal a brittle nature of TWIP steel
macroscopically in spite of high strain hardenability and large uniform elongation.
Fig. 1 (a) Effective strain distribution at the moment of fracture, (b) SEM micrographs showing the thickness-wise slant fracture, (c) and the
front surface cracks of the Fe-15Mn-0.65C-2Al TWIP steel plate.
The brittle fracture features of TWIP steel (in particular without Al) are often explained by dynamic strain aging (DSA)
associated with C - dislocation interactions, i.e. Portevin - Le Chatelier (PLC) effect 1). DSA results in plastic localization in the
form of PLC band during deformation. In the stress-strain response, the serrated flow and negative strain rate sensitivity reflect
repetitive formation and propagation of PLC bands. The plastic localization is favored by plastic instability (Considére criterion),
strain rate softening (negative strain rate sensitivity) and thermal softening (due to adiabatic heating) 2,3), i.e. Eq. 1.
𝛿𝜀 𝛿𝜀 𝛿𝑇 0 (1)
Thermal softening is not significant compared to other contributions in the usual quasi-static tensile tests. The formation and
Microscopic fracture
TWIP steel exhibits a typical dimple fracture microscopically. Dimples are usually fine (typically less than ~ 2 m) and
shallow except some large dimples associated with second phase particles (Fig. 2a) 7). It is indicative that fracture is primarily
attributed to void nucleation rather than growth. For the smooth specimen, entire fracture surface is covered with such fine dimples
almost without the shear lips at the edge of fracture surface. The intersections of mechanical twins and slip bands with grain
boundaries are known to be the primary void nucleation sites 8). As strain increases, the density of mechanical twin increases,
creating more potential void nucleation sites. Meanwhile the void growth rate increases with increasing the stress triaxiality 9).
However, void growth during tensile deformation of the smooth specimen of TWIP steel is suppressed because the stress
triaxiality remains almost constant (~0.33 for the smooth specimen) due to lack of necking 7,10). Accordingly voids remain fine
even after prolonged uniform elongation and the final fracture occurs by bridging or interlinkage of these fine voids. Void
interlinkage in ductile materials is often explained in terms of the void sheet mechanism 11). The void sheet is a planar feature
consisting of fine voids, usually oriented 45º with the tensile axis. The void sheet is formed by shear localization between primary
large voids and void nucleation inside shear localized region. It was reported that ~70% of fracture surface of the smooth specimen
of TWIP steel consists of void sheets 12).
Fig. 2 (a) A low magnification SEM micrograph showing the fracture surface of the smooth specimen of the Fe-15Mn-0.65C-2Al TWIP
steel plate consisting of fine dimples and few isolated large dimples, (b) SEM micrographs taken at the center and edge of the fracture
surfaces of the smooth and notched specimens of the same steel.
The void growth rate increases with increasing the stress triaxiality (𝜂 𝜎 /𝜎, 𝜎 : mean stress, 𝜎: effective stress),
resulting in a drastic decrease of the fracture strain for dimple fracture. In the case of the smooth specimen of TWIP steel,
remains as ~0.33 up to fracture because of no necking, as aforementioned. The effects of on dimple fracture of TWIP steel can
References
1) L. Chen, H. S. Kim, S. K. Kim, and B. C. De Cooman: ISIJ Int., 47 (2007) 1804-1812.
2) A. K. Chakrabarti and J. W. Spretnak: Metall. Trans. A, 6A (1975) 737-747.
3) G. Scavino, F. D’aiuto, P. Matteis, P. Russo Spena, and D. Firrao: Metall. Trans. A, 41A (2010) 1493-1501.
4) H. Y. Yu, S. M. lee, J. H. Nam, S. J. Lee, D. Fabrégue, M, H. Park, N. Tsuji, and Y. K. Lee: Acta Mater., 131 (2017) 435-444.
5) D. Rittel and I. Roman: Metall. Trans. A, 19A (1988) 2269-2277.
6) M. Abbasi, S. Kheirandish, Y. Kharrazi, and J. Hejazi: Mater. Sci. Eng. A., 513-514 (2009) 72-76.
7) D. Fabrégue, C. Landron, O. Bouaziz, and E. Maire: Mater. Sci. Eng. A., 579 (2013) 92-98.
8) M. Madivala, A. Schwedt, U. Prahl, and W. Bleck: Metals, 9 (2019) 367 (1-24)
9) F. A. McClintock: J. Appl. Mech., 35 (1968) 363-371.
10) D. Fabrégue, C. Landron, O. Bouaziz, and E. Maire: Steel Res. Int., 86 (2015) 1197-1203.
11) T. B. Cox and J. R. Low. Jr.: Metall. Trans., 5 (1974) 1457-1470.
12) Z. C. Luo and M. X. Huang: Steel Res. Int., 89 (2018) 1700433 (1-7).
13) S. Park, J. Jung, W. Cho, B. S. Jeong, H. Na, S. I. Kim, M. G. Lee, and H. N. Han: Int. J. Plast., 136 (2021) 102900 (1-24).
High-entropy steel, also named TRIPLEX or Fe-Al-Mn-C steel, has been defined to meet multiple engineering
requirements such as ultrahigh strength, excellent ductility and lower density. In the past 10 years, many studies focused on
alloy design with respect to density reduction, phase stability and mechanical behaviors in this grade of steel. Actually, there a
large space of nano/microstructural engineering to optimize the overall performance of high-entropy steels.
In this work, Fe-Al-Si-Mn-C-(Ni) alloys will be the newly-developed high-entropy lightweight steels meeting
multiple mechanical behaviors through various nano/microstructural engineering. When it is nearly austenitic steel, high
strength and supra ductility of over 70 % can be achieved under a slip-to-twinning plasticity1). When it is duplex steel with
austenite and D03 ordered phase, the steel own an ultrahigh ultimate strength of over 1300 MPa and excellent ductility of
about 45 % through an enhanced effect of geometrically necessary dislocations. When κ carbides are induced into the steels,
precipitation hardening brings about yield strength of 1500 MPa, ultimate tensile strength of over 1700 MPa and total
elongation of 15 % for this steel. Importantly, all these novel properties are achieved in steels with density lower than 7.0
g.cm-3, about 10% lighter than conventional steels. Therefore, this is a new high-entropy lightweight steel with multiple
characteristics in mechanical behaviors.
We will shortly discuss the behavior of hydrogen-induced delayed cracking (HIDC) in this new steel.
Hydrogen-induced plasticity seems observed when it is nearly austenitic steel2). Besides, excellent resistance to HIDC is also
obtained in 1300 MPa-grade duplex steel. In summary, high-entropy or compositionally complex concept, indeed, is able
drives new novelties for steels especially under effective nano/microstructural engineering.
Acknowledgement
This work was performed under the support of Grant Nos. 110-3116-F-002-002 and 109-2224-E-002-002.
References
1) Z.-H. Lai, Yi-Hsuan Sun, Yi-Ting Lin, Rui-Fan Tu, H.-W. Yen: Acta Materialia. 210 (2021) 116814
2) Z.-H. Lai, Yi-Ting Lin, Yi-Hsuan Sun, Rui-Fan Tu, H.-W. Yen: Scripta Materialia. 213 (2022) 114629
Abstract:
Austenitic (face-centered cubic) Fe-Mn-Al-C low-density steels have been considered as structural
materials for cryogenic applications due to the attractive combination of mechanical properties such
as mechanical strength (yield strength of 0.5 – 1.0 GPa; ultimate tensile strength of 1.2 – 1.5 GPa),
tensile elongation (30 - 70%) and Charpy impact energy (40 - 140 J), which are comparable to that of
conventional stainless steels. Several studies have shown that the room-temperature deformation
behavior of austenitic Fe-Mn-Al-C low-density steels is characterized by the formation of microbands
(MBs) ascribed to deformation localization phenomena. These works have indicated that MB
formation results in a further plastic accommodation, and hence, enhances the ductility of these steels.
Due to the current interest in Fe-Mn-Al-C low-density steels as structural materials for cryogenic
applications, the investigation of MB formation at cryogenic temperatures can provide further insight
into the deformation behavior of these steels. The present study aims at studying the MB formation
mechanisms in austenitic Fe-Mn-Al-C low-density steels deformed at a cryogenic temperature. The
MB structure in an austenitic Fe-30Mn-6.5Al-0.3C (wt.%) low-density steel tensile deformed at -
196°C was investigated by combined electron channeling contrast imaging (ECCI) and electron
backscatter diffraction (EBSD). The interplay between the deformation structure and MB formation
was evaluated. Several microstructural aspects of microbanding such as the MB structure, MB
morphology, and MB alignment were quantitatively analyzed. From this analysis, several MB
formation mechanisms are proposed. Finally, the contribution of MBs to the mechanical behavior is
analyzed.
Abstract:
A novel process combination of thermal deformation and intercritical annealing (IA) is proposed to
optimize the mechanical properties and avoid Lüders bands of 7Mn steel. The test results show that
the prior austenite grain (PAG) size, carbides precipitation, dislocation density and stored deformation
energy before IA have great effects on the nucleation and growth of austenite during IA, and ultimately
influence the grain size, morphology, volume fraction and stability of retained austenite (RA) after IA.
The relatively high strain rate and low temperature deformation condition sample shows a mixture of
lamellar and equiaxed morphology structure after IA, and exhibits the best comprehensive mechanical
properties, with a product of ultimate tensile strength and total elongation (UTS×TE) value of
(53GPa∙%), owing to the 2-stage discontinuous transformation-induced plasticity (TRIP) effects.
Because the lamellar austenite is more stable than the equiaxed, block-like austenite dominates the
1st-stage plastic deformation, while the film-like dominates the 2nd-stage. It is further confirmed by
the interrupted tensile tests and nanoindentation investigation. This novel process is simple with low
industrial cost, which can provide a new guidance for development of lower cost fabricating
technology of medium Mn TRIP steel.
Department of Materials Science and Engineering & Research Institute of Advanced Materials, Seoul National University,
Seoul 08827, Republic of Korea
*email: hnhan@snu.ac.kr
Electroplastic phenomenon has been demonstrated by that the elongation increases remarkably during deformation
under electric current without significant temperature rise due to Joule heating. Since the 1960s, the electroplasticity has been
actively investigated until recently 1,2); however, full explanation of electroplasticity mechanism has been lacking. In this
presentation, the origin of electroplasticity was elucidated based on numerical and experimental approaches 3). Ab-initio
calculations show that a charge imbalance near defects weakens drastically atomic bonding. The electroplastic behavior could
be well reproduced with microstructure-based FE simulations. In addition, the weakening of atomic bonding was confirmed
by measuring elastic modulus under electric current, which is inherently related to bond strength.
0.6
0.4
0.2
0.0
0 60 120 180 240
2
Electric current density (A/mm )
Fig. 2. Measured and calculated recrystallization fractions (left) and schematic diagram showing changes in thermal and
athermal effects as a function of electric current density (right) 4,5)
As one of the challenging applications of electroplasticity, in addition, we utilized sub-second electric pulsing,
which can enhance the kinetics of microstructural change to infinitely reset the damaged microstructure as a non-autonomous
self-healing method. The principle of microstructure resetting was explained based on three categories of resetting cores:
phase transformation, dislocation recovery, and recrystallization. Microstructure resetting assisted infinite reuse was
successfully realized using 301L and 316L stainless steels, which are applicable materials of the resetting core. This is a new
concept combining extreme simplicity, rapidness, and infinite repetition, which cannot be achieved by conventional methods.
In addition, various application examples such as electric current assisted trimming, joining and sintering will be introduced in
this presentation.
Fig. 3. Infinite reuse concept using microstructure resetting based on electric current-induced kinetic enhancement 6)
References
1) C. Rudolf, R. Goswami, W. Kang, J. Thomas: Acta Materialia 209 (2021) 116776.
2) S. Zhao, R. Zhang, Y. Chong, X. Li, A. Abu-Odeh, E. Rothchild, D. C. Chrzan, M. Asta, J. W. Morris Jr., A. M. Minor: Nature
Materials 20 (2021) 468-472.
3) M. –J. Kim, S. Yoon, S. Park, H. –J. Jeong, J. –W. Park, K. Kim, J. Jo, T. Heo, S. –T. Hong, S. H. Cho, Y. –K. Kwon, I. –S. Choi,
M. Kim, H. N. Han: Applied Materials Today 21 (2020) 100874.
4) J. –W Park, H. –J. Jeong, S.-W. Jin, M. –J. Kim, K. Lee, J. J. Kim, S. –T. Hong, H. N. Han: Materials Characterization 133 (2017)
70-76.
5) K. Jeong, S.-W. Jin, S.-G. Kang, J.-W. Park, H.-J. Jeong, S.-T. Hong, S. H. Cho, M.-J. Kim, H. N. Han: Acta Materialia (2022)
under revision
6) H. –J. Jeong, M. –J. Kim, S. –J. Choi, J. –W. Park, H. Choi, V. T. Luu, S. –T. Hong, H. N. Han: Applied Materials Today 20
(2020) 100755.
Keywords: medium manganese steel, retained austenite, solute partitioning, martensite, yield point elongation
Medium manganese steels have received substantial attention from academia and the steel producing and using
communities. These steels are viable candidates for vehicle light weighting due to attractive properties combinations which
could enable downgauging of automobile components. Manganese alloying levels are usually between 5-15 wt pct, much
lower than fully austenitic steels such as stainless and high manganese twinning induced plasticity (TWIP) steels. Thermal
processing typically involves intercritical annealing and resulting microstructures are ultrafine grained ferrite/austenite
mixtures with potentially some martensite present.1) If a cold rolled material is used as starting material, the ferrite/austenite
grains tend to be equiaxed following heat treating. High elongation levels can be obtained as a resultant of substantial strain
hardening due to transformation induced plasticity (TRIP) when substantial austenite fractions are present with favorable
mechanical stability.
More recently, alternate thermal processing routes have been considered for medium manganese steels including
quenching and partitioning (Q&P), hot stamping, and double soaking (DS) amongst others.2,3,4,5,6) The DS heat treatment
will be considered in the present contribution and Figure 1 schematically shows the approach.
Figure 1 Schematic of double soaking heat treating. The primary soak is conducted in the intercritical range and the
secondary soak is conducted at a higher temperature. The two steps can be conducted discontinuously with intermediate
cooling to room temperature or continuously. Anticipated microstructural changes are indicated with ferrite, p and s
primary and secondary austenite respectively and ’ martensite.4)
The primary soak consists of intercritical annealing to introduce austenite in the microstructure and to partition
solute manganese and carbon from ferrite into austenite. The secondary soak is conducted at a temperature in excess of the
first soaking temperature and aims at replacing the solute-lean ferrite present at the end of the primary soak with martensite
following final quenching to room temperature. This is achieved by transformation of the ferrite into austenite during the
(short duration) high temperature secondary soaking. The two steps can be separated by cooling to room temperature
(discontinuous) or can be done continuously.
DS has been applied discontinuously to a 0.14C-7.14Mn-0.23Si steel with an initial as-received microstructure of
ultrafine-grained ferrite and retained austenite (46 vol pct) produced by cold rolling followed by batch annealing in the
Figure 2 Retained austenite evolution as a function of secondary soaking hold time at the indicated temperatures.
As-received austenite volume percent in the cold rolled batch annealed condition is indicated by the dashed line. 4)
The character of microstructural evolution is further reflected in the transmission Kikuchi diffraction (TKD) phase
maps presented in Figure 3 which are overlaid with image quality maps.7) Green indicates ferrite/martensite whereas red
indicates austenite in the phase map. The as-received cold rolled and batch annealed (AR-BA) condition shown in Figure 3a
is a typical intercritically annealed medium manganese steel microstructure with equiaxed ultrafine ferrite and austenite grains.
The image quality is high in most grains which suggests a low dislocation density. The DS conditions shown in Figures 3b,
c, and d reflect microstructural evolution for secondary soaking at 750 °C, 775 °C, and 800 °C respectively at a fixed time of
30 s. In the DS-750 °C condition a number of ferrite grains with high image quality have been replaced by low image
quality ferrite constituents with traces of austenite. These areas are interpreted to have transformed during final quenching to
room temperature from the additional (secondary) austenite that likely formed during the secondary soak from solute lean
ferrite. A number of high image quality ferrite areas remain as well as equiaxed austenite grains. As the secondary soak
temperature increases, the ferritic areas with low image quality become predominant indicating increased martensite
formation while equiaxed grains of austenite remain in the microstructure.
Engineering stress/strain curves are shown in Figure 4a for a fixed secondary soaking time of 30 s at the indicated
temperatures. It is clear that the tensile behavior is greatly dependent on the employed secondary soaking temperature. At
the lower temperatures (< 725 oC) pronounced yield point elongations are observed followed by substantial strain hardening
and high tensile elongation similar to the tensile behavior of “conventional” intercritically annealed medium manganese steels.
With an increase in secondary soaking temperature, the stress-strain curve shapes systematically change to reduce or eliminate
the yield point elongation and transition the curve shapes to continuous yielding, higher yield strengths, and high strain
hardening rates leading to overall high ultimate tensile strengths and elongation levels (e.g. for T > 800 oC a UTS of
approximately 1700 MPa with 15 pct total elongation is obtained). The properties for the double-soaked steels exceed levels
observed for traditional intercritical annealing and hot-stamped low alloy steels. It is further noted that post-uniform
elongation is absent or limited for the higher temperature annealed samples which may also indicate that, although not
investigated here, local formability/hole expansion may be limited. The low post-uniform elongation may relate to the fresh
martensite introduced during final quenching and tempering/austempering following DS were thus investigated to reduce
a) AR-BA
a) b)
Figure 4 Engineering stress-strain curves developed using a 0.14C–7.17Mn–0.21Si cold rolled and batch annealed steel and
a) double soaking for 30 s using the indicated secondary soaking temperatures and b) double soaking at 800 °C/30 s followed
by austempering at 450 °C for 300 s (DS-A) or, tempering at 450 °C for 300 s and the as-received, batch annealed (BA) steel
as well as a DS sample. Secondary soaking was conducted at 800 °C for 30 s in all conditions.4,2,5)
Acknowledgement
The authors gratefully acknowledge the support of the sponsors of the Advanced Steel Processing and Products Research
Center, an industry-university cooperative research center at the Colorado School of Mines.
References
1) R.L. Miller: Metall. Trans. 3 (1972) (905-912).
2) J.G. Speer, R. Rana, D.K. Matlock, A. Glover, G.A. Thomas, E. De Moor: Metals 9 (2019) 771 (1-9).
3) E. De Moor, J.G. Speer, D.K. Matlock: Proc. of the ICAS 2016 & HMnS 2016, Jeju, Korea, 5–10 September 2016
(182–185).
4) A. Glover, J.G. Speer, E. De Moor: Proc. of the Int. Symp. on New Developments in Advanced High-Strength Sheet Steels,
Keystone, CO, USA, 30 May–2 June 2017, AIST, Warrendale, PA, USA (2017) (189–197).
5) A. Glover, J.G. Speer, E. De Moor: Front. in Mater. 7 (2021) 622131 (1-11).
6) A. Glover, P.J. Gibbs, C. Liu, D.W. Brown, B. Clausen, J.G. Speer, E. De Moor: Metals 9 (2019) 761 (1-12).
7) A. Glover, PhD thesis, Colorado School of Mines, (2020).
The interactions between dislocations and other lattice defects have large effects on the mechanical properties of
iron. For instance, the solute atom1,2), vacancy, precipitate, and other dislocations affect dislocation behaviors and macroscopic
yield strength. The grain boundary (GB) is one of the important lattice defects interacting with dislocations and inducing grain
boundary strengthening. The GB-dislocation interaction also affects the fracture of polycrystalline metals. Therefore, further
understating of GB-dislocation interaction should provide a necessary basis for the strengthening design of steel.
In this presentation, I introduce recent approaches of atomistic modeling and analysis of dislocation behaviors near
grain boundary in iron. The GB-dislocation interaction contains various stages, such as elastic interaction, absorption, pile-up,
transmission across GB, and nucleation from GB. In addition, the GB-dislocation interactions depend on GB structure,
dislocation type, temperature, alloy system (solute element), and stress state. In addition, the GB structure has various degrees
of freedom even in a single element system. For these reasons, the GB-dislocation interaction should have various physical
aspects and depend on various factors. We here focus on the dislocation transmission behaviors across GB in body-centered
cubic (bcc) iron and reveal the mechanism and dominating factors of the transmission behaviors. In experimental
nanoindentation, there are distinct differences in transmission behaviors depending on the GB types: 9-221 exhibits a higher
barrier effect for dislocation transmission across GB than 9-114 3). This is interesting because the two GBs exhibits a small
difference in GB energy. Using molecular static simulations, we investigated the origin of the GB-dislocation interaction in
terms of dislocation types, slip direction, and stress fields. In atomistic nanoindentation simulations of bicrystal models with
GB, the dislocations nucleate beneath the indenter; and the nucleated dislocations have slip directions, which agree well with
those predicted by the elastic stress fields before the dislocation nucleation. In addition, it is interesting that dislocations
approaching GB have different components between the two GB types. In the case of 9-114 GB, dislocations approaching
the GB mainly have an edge component. We confirmed that the edge dislocation makes large atomic displacement at GB,
which may enhance dislocation transmission across GB. This result suggests the important role of the dislocation component
on the transmission behaviors. The dislocation transmission behaviors across GB in bcc iron depend on misorientation angle
(GB structures) as well as the dislocation type. We investigated dislocation transmission behaviors at high angle GB and
small-angle GB in bcc iron using molecular dynamics simulations, which can consider temperature effects. In the case of
small-angle GB, the GB contains an array of dislocations or dislocation networks, and the GB-dislocation interaction has an
aspect of dislocation-dislocation interaction. In the atomistic simulations, the transmission behaviors change depending on the
dislocation types (i.e., edge or screw) and applied stress. In addition, dislocation transmission behaviors are different between
small-angle GB and large-angle GB. This atomistic knowledge should be fundamental for the strengthening design of iron
alloys based on the GB-dislocation interactions.
Acknowledgment
Some of the calculations for this study were performed using the Numerical Materials Simulator at the National Institute for
Materials Science. A part of this work was supported by JST, PRESTO Grant Number JPMJPR219A, Japan.
References
1) M. Wakeda, H. Kimizuka, S. Ogata, Journal of the Japan Institute of Metals and Materials 77 (2013) 409-414.
2) M. Wakeda, T. Tsuru, M. Kohyama, T. Ozaki, H. Sawada, M. Itakura, S. Ogata, Acta Mater. 131 (2017) 445-456.
3) M. Wakeda, Y.L. Chang, S. Ii, T. Ohmura, Int. J. Plast. 145 (2021) 103047.
*email: nakamut@eng.hokudai.ac.jp
Keywords: Gigacycle fatigue, Internal crack, Nondestructive inspection, Synchrotron radiation, Tomography, Steel, Titanium alloy
1. Introduction
Generally, the S–N curve of steel materials exhibits a horizontal line—i.e., a fatigue limit—at about 106–107 cycles.
However, it has been recognized that fatigue failure occurs in regions exceeding 107, even at stresses below the fatigue limit. This
phenomenon is called very high cycle fatigue (VHCF) or gigacycle fatigue, and it has been actively studied both in Japan and
abroad1 since the phenomenon was identified in the 1980s.2 VHCF in high-strength metals, such as high-strength steels and
titanium alloys, exhibits a peculiar phenomenon: fractures originate in the interior of the material. For example, nonmetallic
inclusions of several µm to several tens of µm in high-strength steels and crystal grains of several tens of µm in titanium alloys
are recognized as crack-initiation sites.3 However, it is difficult to detect such small cracks using conventional nondestructive
inspections, such as industrial X-ray computed tomography (CT) or ultrasonic CT. Due to the limitations of the observation
technique, the internal-fracture behaviors and their underlying mechanisms are not yet fully understood.
Third-generation synchrotron radiation facilities equipped with insertion devices, such as SPring-8, ESRF, and APS,
have recently attracted a great deal of attention. Thanks to its high spatial resolution and high-contrast imaging capability, X-ray
CT using high-brilliance synchrotron radiation, so-called SR-CT, has been applied to the field of material fractures.4 The author’s
group has been investigating the nondestructive detection of internal cracks in VHCF using SPring-8.5-7 In a series of studies, the
author’s group has used mainly projection CT (micro-CT) with a spatial resolution of approximately 1 µm. In recent years, the
group has applied phase-contrast imaging CT (nano-CT)8, 9 to observe small fatigue cracks and the surrounding microstructure
with a spatial resolution of approximately 200 nm or higher. The combination of micro-CT and nano-CT has the potential to
clarify the initiation and propagation processes of internal fatigue cracks. The present study introduces the recent SR-CT technique
optimized to detect small fatigue cracks based on an in situ fatigue-testing device available at SPring-8. Small fatigue cracks’
initiation and growth behaviors are discussed with reference to several images of cracks in titanium alloy and high-strength steel.
2. Material selection
The ultimate objective of the author’s group is to develop a universal method for evaluating internal fatigue fractures
in the VHCF regime.10 However, fatigue tests in the VHCF region and SR-CT experiments require considerable time, labor, cost,
etc., and it is impossible to obtain data for every kind of material. Therefore, we examined only three metals—(α+β) titanium
alloy, β titanium alloy, and precipitation-hardened martensite stainless steel—as representative materials in which internal
fractures are frequently observed. More specifically, we used (α+β) type Ti-6Al-4V, β type Ti-22V-4V, and SUS630 as
representatives of dual-phase titanium alloys, single-phase titanium alloys, and vitally important high-strength steel, respectively.
3. Experimental procedure
The imaging experiment was performed at BL20XU of SPring-8. This beamline provided the projection CT and the
phase-contrast imaging CT.8,9 In both cases, the synchrotron radiation from the undulator was monochromatized into
monochromatic X-rays and irradiated to the specimen. The projection CT utilized the absorption of X-rays as they passed through
the specimen, and the physical limit of the spatial resolution was 1 µm or higher, with a millimeter order of field of view. The
imaging CT used a Fresnel zone plate, and the phase shift of the X-ray traveling through the sample was converted to contrast
using a Zernike phase plate. Although the field of view was limited (approximately a few tens of micrometers) compared to
projection CT, the spatial resolution was extremely high (200 nm or higher).
An in situ piezo fatigue-testing device was developed to perform fatigue tests on the beamline.11 The capacities of the
tensile and compressive loads were 600 N and 3800 N, respectively. The entire system was set on a rotation stage of the beamline
Field of view
(Imaging type-CT) Existed at
1.10×107
A
(α+β) α
Fig. 1. Internal fatigue crack detected by micro-CT. (a) N = 1.10 ×107 cycles. (b) N = 1.12 Fig. 2. Magnified view of the dashed
×107cycles. The loading direction is the vertical direction in the figures. square in Fig. 4 (b) via nano-CT.
Fig. 1 shows examples of the internal crack images at a virtual cross-section parallel to the loading direction. During
CT imaging, the tensile load equivalent to the maximum stress in the fatigue test (σmax = 650MPa) was applied. X-ray energy,
projection number, exposure time, voxel size, and field of view were 30.0 keV, 1800, 0.025 s, 0.5 µm, and 1 × 1 mm2, respectively.
Fig. 1 (a) was observed at 1.10 ×107 cycles and Fig. 4 (b) at 1.12 ×107 cycles. In both figures, the fatigue crack and the surrounding
microstructure were identified. The dark and bright parts corresponded to the α-grain and (α+β) grain, respectively. Crack growth,
which originated from the right-side crack tip in Fig. 1 (b), was observed from 1.10 ×107 to 1.12 ×107 cycles. This crack extension
likely occurred in the α-grain. Fig. 2 shows a magnified view of the white-dashed square in Fig. 1 (b) observed using nano-CT.
X-ray energy, projection number, exposure time, voxel size, and field of view were 30.0 keV, 1800, 0.5 s, 0.0387 µm, and about
63 × 63 µm2, respectively. The crack shape and surrounding microstructure were much more clearly visible than in the micro-CT
image. Very straight crack growth was confirmed to occur in the α-grain, as the dark part of Fig. 1 (b) suggested. The newly
generated crack reached and stopped at the adjacent grain boundary. Considering the linear and inclined crack extension, this part
corresponded to the faceted feature12 that was particularly evident on the fracture surface of the internal crack.
Including this crack, a total of 63 internal cracks were detected over the course of 1.18 × 107 cycles. Almost all the
initiation sites corresponded to the α-grain, and their features were very similar to those of the extended portion of the crack tip
shown in Fig. 2. The phenomenon of many internal cracks originating in one specimen has been reported in a previous study7;
however, the number of detected cracks was significantly increased in the present study, from 287 to 63. This increase resulted
from the improvement in image quality, including the effect of using a thinner specimen (ϕ 0.7).
Fig. 3. Fatigue crack originating from an inclusion observed using micro-CT. (a) Top view of the crack (plane perpendicular to
loading axis). (b) Magnified view of the crack in (a). (c) Side view of the same crack (plane parallel to the loading axis, which
corresponds to the vertical direction in the figure). (d) Magnified view of the inclusion and crack in (b).
When fatigue testing on the same specimen was repeated, the specimen fractured only after an additional cyclic load
of 0.19 × 106 cycles. After the fracture, the specimen was observed again using micro-CT, and the fractured surface was observed
using a scanning electron microscope. On the fracture surface, another 10.8-μm internal inclusion was observed. This inclusion
was slightly larger than the 9.9-μm surface inclusion (Fig. 3 (d)) that was responsible for the fracture. The specimen exhibited
other internal inclusions that were larger than the surface inclusion shown in Fig. 3; however, these inclusions did not exhibit any
crack initiations. Although further investigations are required, it is likely easier for cracks to originate from surface inclusions than
from internal inclusions. Importantly, this observation also indicated that the crack initiation and its propagation up to the very
short distance of 5.3 μm consumed more than 90% (2.0 × 106 cycles) of the total fatigue life (2.19 × 106 cycles). The crack
5. Summary
This paper presented a recently developed SR-CT technique: the use of micro-CT and nano-CT to nondestructively
observe small fatigue cracks in high-strength metals. Example images of cracks in (α+β) Ti-6Al-4V and SUS 630 were introduced
to discuss the initiation and growth behaviors related to VHCF phenomena. The main findings are summarized below.
1. A crack extension in the α grain from an internal crack tip was detected in (α+β) Ti-6Al-4V. Its straight and inclined growth
to the loading direction matches the α-facet characteristics usually observed on the internal-fracture surface. This result
suggests that the so-called multiple facets on fracture surfaces in the VHCF regime are not formed by the coalescence of
plural individual microcracks but rather induced by the crack’s propagation from a single origin.
2. A crack initiation with a length of approximately 5 µm was successfully detected at a surface inclusion with a size of about
10 µm. Although larger internal inclusions were found in the control volume of the specimen, no clear sign of crack initiation
was observed around them. Moreover, the initiation of this very small crack consumed more than 90% of the fatigue life;
therefore, taking extreme care not to overlook cracks and continuous improvement of image quality will be indispensable
to observing internal cracks in this material.
Acknowledgments
The author expresses sincere gratitude to Daido Steel Co., Ltd., for providing (α+β) Ti-6Al-4V and to Dr. Y. Furuya at NIMS for
providing SUS630 with beneficial suggestions. The author is grateful to Dr. F. Yoshinaka, Dr. H. Oguma at NIMS, and Dr. N.
Fujimura at Hokkaido University for fruitful discussion. The synchrotron radiation experiments were performed at the BL20XU,
SPring-8 (Proposal No. 2017B1421, 2018B1289, 2019B2042, 2020A0172, 2021A0172, 2021B0172) with the approval of the
Japan Synchrotron Radiation Research Institute (JASRI). The author expresses sincere appreciation to Dr. A. Takeuchi, Dr. M.
Uesugi, and Dr. K. Uesugi at JASRI for providing their ideas and suggestions for utilizing SR-CT. This study was funded by the
Japan Society for the Promotion of Science, Japan (grant numbers: Scientific Research [A, 18H03748, and 21H04529]). The
author is grateful to several graduate students—Mr. Y. Tomoda, Mr. S. Modi, Mr. H. Morishita, Mr. T. Yamazaki, and Mr. G.
Xue—for their support in fatigue tests and computed tomography.
References
1) Sakai T, Nakagawa A, Oguma N, Nakamura Y, Ueno A, Kikuchi S, Sakaida A. Int J Fatigue. 2016;93:339–351.
2) Naito T, Ueda H, Kikuchi M. Metall Trans. 1984;15A:1431–1436.
3) Murakami Y. Metal Fatigue: Effects of Small Defects and Nonmetallic Inclusions, Second Edition, Academic Press, Elsevier
Ltd. (2019).
4) Wu S, Xiao T, Withers PJ. Eng Fract Mech. 2017;182:127–156.
5) Nakamura T, Yoshinaka F, Nakayama S, Oguma H, Shiozawa D, Nakai Y, Uesugi K. Mech Eng Lett. 2016;2:16-00233-16–
00233.
6) Yoshinaka F, Nakamura T, Nakayama S, Shiozawa D, Nakai Y, Uesugi K. Int J Fatigue. 2016;93:397–405.
7) Yoshinaka F, Nakamura T, Takeuchi A, Uesugi M, Uesugi K. Fatigue Fract Eng Mater Struct. 2019;42:2093–2105.
8) Takeuchi A, Suzuki Y. Microscopy 2020;69:259–279.
9) Takeuchi A, Uesugi K, Uesugi M, Toda H, Hirayama K, Shimizu K, Matsuo K, Nakamura T. Rev Sci Instrum.
2021;92:023701.5
10) SPring-8 Users Office, Brief Description of Long-term Proposals Approved for 2020A, SPring-8/SACLA Information,
2020;25-2:154–159.
11) Xue G, Nakamura T, Fujimura N, Takahashi K, Oguma H, Takeuchi A, Uesugi M, Uesugi K. Eng Fract Mech.
2022;263:108308.
12) Liu X, Sun C, Hong Y. Int J Fatigue. 2016;92:434–41.
13) National Institute for Materials Science, Japan. NIMS Fatigue Data Sheet No. 118. 2014.
14) Modi S, Fujimura N, Takahashi K, Nakamura T, Takeuchi A, Uesugi M, Uesugi K. Final Brief Paper of the Eighth
International Conference on Very High Cycle Fatigue (VHCF8). 2022:2B12.
Keywords: low-alloy steels, hydrogen gas, ductility, fatigue life, fatigue crack growth, fracture toughness
Quenched and tempered low-alloy steels having martensitic microstructure are promising materials, primarily in
the context of pressure vessels destined for the storage of hydrogen gas in hydrogen refueling stations 1). According to the
database established by the National Aeronautics and Space Administration (NASA) in the USA 2), low-alloy steels (e.g., AISI
4140) are classified as either “severely-” or “extremely-embrittled” materials with the occlusion of hydrogen. Thus, when
low-alloy steels are destined for use in hydrogen gas environment, the hydrogen-induced degradation of their mechanical
performance should be thoroughly taken into account in the strength-design.
For components used in high-pressure gas system, two different design methodologies have been developed: (i)
Design by Rule and (ii) Design by Analysis 3, 4). The former provides an infinite life design with a high safety factor, while a
low safety factor can be used as a premise for a finite life design in the latter. Even though the Design by Analysis involves
complex procedures such as leak-before-break (LBB), fatigue life and fatigue crack-growth (FCG) analyses 5, 6), it offers the
advantage of weight-saving and resultant better rationality. Therefore, a proactive application of Design by Analysis is more
desirable in any attempt to achieve superior economic efficiency of high-pressure hydrogen systems.
The strength properties of low-alloy steels in high-pressure hydrogen gas have been addressed by several
researchers over the past decade. Matsunaga et al. 7) carried out slow strain-rate tensile (SSRT) tests in 115-MPa-hydrogen gas
at ambient temperature using low-alloy steel, JIS-SCM435, with a tensile strength (TS) of ≈ 820 MPa. They reported that the
TS of the steel was not degraded in hydrogen gas, although its ductility loss was significant. Moreover, they conducted
tension-compression fatigue tests for the same material and demonstrated that, even in 115-MPa-hydrogen gas, the fatigue life
was not degraded in a relatively high-cycle regime (i.e., N > 105), while a remarkable degradation was noted in a
comparatively low-cycle regime (i.e., N = 104 ~ 105) 7). Subsequently, Ogawa et al. 8) also subjected the same material to
tension-compression fatigue tests in 115-MPa-hydrogen gas, using specimens with tiny semi-circular slits. The fatigue limit
was determined by the non-propagation condition of micro-cracks that initiated at notch-tips, for which the threshold stress
intensity factor range, ΔKth, was invariable, regardless of the presence or absence of hydrogen. Yamabe et al. 9) performed FCG
tests in 90-MPa-hydrogen gas, as well as cyclic pressurization tests between 0.6 and 45 MPa, using hydrogen vessels of
JIS-SCM435 with TSs of 830 ~ 950 MPa. According to their findings, the fatigue lives of the vessels could successfully be
predicted according to FCG-rate data, while the LBB condition was satisfied when the crack-fronts reached the outer surfaces
of the vessels. Furthermore, Ogawa et al. 10) investigated FCG behavior in JIS-SCM435 after single overloading in air and in
90-MPa-hydrogen gas. They discovered that substantial FCG retardation occurred even in hydrogen gas, the extent of which
was similar to that encountered in air. Considering the above-mentioned series of experimental facts, low-alloy steels with TSs
of less than or in the range of 900 MPa seem to be suitable for use in components exposed to high-pressure hydrogen gas. In
fact, such knowledge has partially been incorporated in standards 3) and technical guidelines 1) for the selection of materials
and the strength-design of hydrogen utilization systems.
The hydrogen sensitivity of low-alloy steels varies depending on their strength levels and microstructural
morphologies 2, 11, 12). In the aforementioned design methodologies, TS value is used in the definition of safety factor 3, 4). For
this reason, the hydrogen-induced TS degradation is not preferred in view of the hydrogen compatibility. It has been revealed
by a number of experiments that TS of material is degraded in high-pressure hydrogen gas when the TS exceeds 1000 MPa 2),
as shown in Fig. 1, in which low-alloy steels are no exception. Also, such a variation in susceptibility is particularly prominent
for FCG properties as elaborated by Matsuoka et al., who carried out FCG tests of low-alloy steels JIS-SCM435 and
JIS-SNCM439 with TSs of 820 ~ 1130 MPa at various test frequencies ranging from 0.001 to 1 Hz 11). Their key discoveries
Fig. 1 Relative tensile strength (RTS) as a function of TS in hydrogen gas at pressures of 69−115 MPa at ambient temperature.
10000
Relative fatigue crack growth rate, (da/dN H )/(da/dN Air)
JIS-SM490B
(K = 30 MPa m1/2 )
100 JIS-S25C
(K = 25 MPa m1/2 )
JIS-S55C
(K = 25 MPa m1/2 )
10
Insusceptible to
test frequency Time-dependent
1
400 500 600 700 800 900 1000 1100 1200
Tensile strength, B (MPa)
Fig. 2 Relationship between relative fatigue crack-growth rate and tensile strength in low-alloy and carbon steels 2).
Acknowledgement
This work was partially supported by the New Energy and Industrial Technology Development Organization (NEDO),
Hydrogen Utilization Technology (FY2013∼FY2017).
References
1) JPEC-TD 0003, Technical document for safety use of pressure storage vessels made of low alloy steels used in hydrogen
refueling stations. Japan Petroleum Energy Center (2020).
2) NASA Safety Standard for Hydrogen and Hydrogen Systems: Guidelines for Hydrogen System Design, Materials
Selection, Operations, Storage and Transportation, NSS 1740.16-Rev. National Aeronautics and Space Administration
(2005).
3) KHK technical standard, KHKS0220. High Pressure Gas Safety Institute of Japan (2020).
4) J. Yamabe, H. Matsunaga, Y. Furuya, S. Hamada, H. Itoga, M. Yoshikawa, et al. Int. J. Hydrogen Energy 40 (2015)
pp.719-28.
5) Yamabe J, Itoga H, Awane T, Matsuo T, Matsunaga H, Matsuoka S. J. Press. Vessel Technol. Trans. ASME 138 (2016)
011401.
6) K. Ma, J. Zheng, Z. Hua, C. Gu, R. Zhang, Y. Liu. Int. J. Hydrogen Energy 45 (2020) 20132-41.
7) H. Matsunaga, M. Yoshikawa, R. Kondo, J. Yamabe, S. Matsuoka. Int. J. Hydrogen Energy 40 (2015) pp.5739-48.
8) Y. Ogawa, H, Matsunaga, J. Yamabe, M. Yoshikawa, S. Matsuoka. Int. J. Hydrogen Energy 43 (2018) pp.20133-42.
9) J. Yamabe, O. Takakuwa, H. Matsunaga, H. Itoga, S, Matsuoka. Int. J. Hydrogen Energy 42 (2017) pp.13289-99.
10) Y. Ogawa, K. Iwata, S. Okazaki, M. Nakamura, K. Matsubara, O. Takakuwa, et al. Mater. Lett. 308 (2022) 131115.
11) S. Matsuoka, H. Matsunaga, J. Yamabe, S. Hamada, T. Iijima. Trans. JSME 83 (2017) 17-00264.
Keywords: titanium alloy, metal fatigue, fatigue crack growth, electron channeling contrast imaging, in situ observation
Ti-6Al-4V alloy shows characteristic fatigue crack propagation associated with the anisotropic plasticity evolution
at the crack tip1). In addition, the dwell effect, which is practically important to use Ti alloys, further enhances the complexity
of the fatigue crack propagation behavior. In this study, micro-mechanical and microstructural analyses were performed to
clarify the dwell fatigue behavior of a bi-modal Ti-6Al-4V alloy. Specifically, how dwell fatigue loading accelerates the crack
propagation was investigated. Dwell fatigue tests that contain a dwell process were conducted at room temperature for
understanding the effect of the dwell fatigue on two type of crack propagation modes: (1) crack propagation controlled by
crack tip dislocation emission (Mode I) and (2) crack propagation along a slip plane, associated with dislocation/damage
accumulation (Mode II). The fatigue tests were carried out under a load control including a 10-min displacement holding at
the peak force for only one cycle at several stress intensity factor ranges. To understand the relationship between crack tip
deformation and dislocation microstructure evolution during Mode I crack propagation, we applied digital image correlation
based strain mapping during the 10-min holding, secondary electron imaging to observe the fatigue striation, and electron
channeling contrast imaging of beneath the fracture surface. Furthermore, focusing on Mode II crack propagation, the
deformation behavior during the displacement holding was accurately observed through SE imaging under an in situ loading
test. As a factor controlling the dwell-induced acceleration of crack propagation, we visually clarified that the enhanced
dislocation emission at the crack tip by the thermally activated process is important. This creates the time dependency of the
plastic strain evolution. More specifically, the factors controlling the dwell-induced crack propagation acceleration were
concluded as follows.
(1) During Mode I fatigue crack propagation, spacing of the striations on the fracture surface was locally extended by the
displacement holding, indicating a significant acceleration of crack propagation. Moreover, the acceleration of crack
propagation remained during several cycles even after displacement holding, regardless of stress intensity factor range.
(2) The crack tip strain, which evolved along the loading axis, increased with time during the displacement holding. The
increase in striation spacing was proportional to the increase in strain at the crack tip. The dislocation structure near the
fracture surface observed through the electron channeling contrast imaging indicated that the number of dislocations
emitted from the crack tip was increased by displacement holding. The results conclude that the dwell process accelerates
crack propagation by assisting crack opening. The observed extra crack tip opening during the displacement holding
caused a delay of crack surface contact during the unloading process, resulting in the decrease in the plasticity-induced
crack closure effect.
(3) The in-situ SE imaging indicated that the displacement holding assisted shear deformation at the crack tip and caused a
delay of crack surface contact during the unloading process. This effect resulted in the increase in the effective stress
intensity factor range and the subsequent plasticity-related damage accumulation for Mode II crack propagation.
Acknowledgement
This work was performed under the support of JSPS KAKENHI (JP16H06365; 20H02457).
References
1) A. Maenosono, M. Koyama, Y. Tanaka, S. Ri, Q. Wang and H. Noguchi: Int. J. Fatigue 128 (2019) 105200.
Keywords: low carbon steel, weld, heat affected zone, microstructure, fatigue, small fatigue crack
Fatigue crack of the welding initiate and grow at weld toe where has various microstructure according to heat history
of welding, called heat affected zone (HAZ). Hence, it is important to understand the effect of HAZ microstructure on fatigue
property for long term reliability of the welding. Especially, it is necessary to understand the small fatigue crack initiation and
growth behavior which is considered to be sensitive to microstructure.
Recently, we have been studied the effect of HAZ microstructure on fatigue properties by using the specimen with
HAZ microstructure which thermally simulated by heat treatment. In these study, we have investigated the effect of HAZ
microstructure on cyclic yielding behavior1), low-to-high cycle fatigue property2), giga-cycle fatigue property3) and microscopic
fatigue initiation and growth behavior4,5) by using newly developed our automatic microscope system6). From these research,
negative influence on fatigue life of coarse grain of HAZ were clarified. This paper briefly review these recent research from the
view point of microstructural small fatigue crack behavior.
In this study, more than 15 types of simulated HAZ microstructures were prepared by heat treatment which controlling
maximum temperature and cooling rate. Figure 1 shows example of microstructures images. Higher cooling rate generate
martensitic microstructure and lower cooling rate generate bainitic, pearlitic and ferritic microstructures. Coarse grain HAZ
(CGHAZ) microstructures were generated by higher maximum temperature and fine grain HAZ (FGHAZ) microstructures were
generated by lower maximum temperature. Average prior austenite grain size of prepared microstructures were 30 to 600 µm,
Vickers hardness were 150 to 400.
Macroscopic yielding property of these microstructure where comparable to other usual steels. Figure 2 shows
relationship between Vickers hardness, tensile strength and cyclic yielding stress of HAZ microstructures. Tensile strength was
proportional to Vickers hardness and cyclic yielding strength were proportional to tensile strength similar to ordinal steels
regardless of microstructural features. On the other hand, prior austenite grain size effects were revealed in low, high and giga
cycle fatigue properties as shown in Figure 3. As shown in Figure 3 a), fatigue life of CGHAZ microstructure group shows lower
fatigue life than FGHAZ group. In addition, internal origin giga cycle fracture occurred in CGHAZ microstructure in relatively
lower strength region. Hence, fatigue limit of CGHAZ were lower than FGHAZ microstructure as shown in Figure 3 b). In
addition, relationship between fatigue life and internal structure, such as martensite and bainite, was not clear.
Such macroscopic fatigue life difference between CG and FGHAZ microstructure were able to be understood by
microscopic fatigue crack initiation and growth behavior. Figure 4 shows examples of fatigue crack initiation behaviors.
According to crack and fracture surface orientation, fatigue cracks seems initiated along slip plane by shear stress regardless of
microstructure. In other words, fundamental fatigue crack initiation and growth mechanism of FG and CGHAZ were same. On
the other hand, early stage fatigue crack growth rate of each microstructures were different. Figure 5 shows fatigue crack growth
behavior of FG and CGHAZ microstructures. Regardless of microstructure, fatigue cracks initiated at very early stage of total
fatigue life as shown in Figure 5 a). However, early stage fatigue crack growth rate of CGHAZ were faster than that of FGHAZ
as shown in Figure 5 b). In this figure, average prior grain size were described by dotted line. Acceleration of early stage fatigue
crack growth of CGHAZ were mainly appeared in the region where the fatigue crack was smaller than prior austenite grain size.
In this region, fatigue crack were propagated along shear stress direction as shown in Figure 4 b).
On the other hand, tensile mode fatigue crack growth rate of HAZ microstructure were seems to follow the fracture
mechanics parameter6,7). Figure 6 shows an example of fatigue crack growth evaluation results of FGHAZ microstructure. Our
microscope observation system is applicable to direct measurement of fatigue crack opening point by using Digital Image
Correlation (DIC) technique6) as shown in Figure 6 a). Fatigue crack growth rates were well organized by using effective J integral
range as shown in Figure 6 b).
1500 1500
Tempered martensite
Carbon and Low alloy steel : Carbon steel, N
: Carbon steel, QT σyc =0.9 σB
Normalized Steel Tmax CR2
Carbon steel : Low alloy steel, QT
Cyclic yielding stress,σyc (MPa)
0 0
0 100 200 300 400 0 500 1000 1500
Vickers hardness, HV Tensile strength, σB(MPa)
Figure 2 Relationship between Vickers hardness, tensile strength and cyclic yielding stress of HAZ microstructures2).
B10-1.TXT H900
B10-30.TXT H1075
B14-10.TXT SCM440, 180T
a) 0.05
b) 1400 C10-30.TXT
D10-30.TXT
S40C, 180T
Steel Tmax CR2 D14-1.TXT A7075-T5
Solid mark: Internal origin
Heat Treated Low Carbon Steel UniaxialD14-3.TXT
loading, R=-1 SCM440,
Open mark: Surface550T
origin
B 1273 K 1 K/s
Fatigue limit, σW (MPa)
(0.15C-1.5Mn, 0.07C-1.5Mn)
B 1273 K 30 K/s 1200 D210-10.TXT S40C, 550T
RT, R=-1
Strain amplitude εa (-)
Figure 3 a) low to high cycle fatigue life2) and b) giga cycle fatigue strength3) of HAZ microstructures. Open mark of a) and solid mark
of b) indicate the results of CGHAZ microstructure and the others indicate that of FGHAZ microstructures.
Figure 4 Examples of microscopic fatigue crack initiation behavior and corresponds fracture surface morphologies for a) FGHAZ and
b) CGHAZ microstructure4). Arrows indicate initially detected crack tip position.
Figure 5 Small fatigue crack growth behaviors of FG and CGHAZ microstructures of a) fatigue crack growth curve and b) normalized
fatigue crack growth rate4). One of the typical CGHAZ data was highlighted in the figures.
a) b) c)
Figure 6 Fatigue crack growth evaluation of FGHAZ microstructure7). a) direct measurement of crack opening stress by using DIC
technique and small fatigue crack growth rate represented with b) stress intensity factor and c) effective J integral range calculated from
directly measured crack opening stress.
References
1) H. Nishikawa and Y. Furuya: Mater. Trans., 60 (2019).
2) H. Nishikawa, Y. Furuya, S. Igi, S. Goto, F. Briffod, T. Shiraiwa, M. Enoki and T. Kasuya: Fatigue Fract. Eng. Mater. Struct., 43
(2020) , 1239.
3) H. Nishikawa and Y. Furuya: ISIJ Int., 59 (2019) , 1926.
4) H. Nishikawa, Y. Furuya, T. Kasuya and M. Enoki: Mater. Sci. Eng. A, 832 (2022) , 142363.
5) H. Nishikawa and Y. Furuya: Tetsu-To-Hagane/Journal Iron Steel Inst. Japan, 105 (2019) , 1179.
6) H. Nishikawa and Y. Furuya: ISIJ Int., 60 (2020) , 2090.
7) H. Nishikawa and Y. Furuya: Tetsu-To-Hagane/Journal Iron Steel Inst. Japan, 105 (2019).
Keywords: ferritic steel, yielding mechanism, Hall-Petch relation, Bailey-Hirsch relation, grain size, dislocation density
1. Introduction
In my presentation, three topics will be introduced on the yielding behavior of polycrystalline ferritic steel. First one
is the effect of purity of ferritic steels. Discontinuous yielding of ferritic steels has been believed due to Cottrell locking of
dislocations by interstitial elements such as C and N, because yield stress decreases with purifying1). However, authors et al.
have proved that Hall-Petch (HP) coefficient decreases with reducing the total amount of C and N below 50 ppm (0.005%)
and this results in the decrease of yield stress even at the same ferrite grain size2). The yielding based on grain refinement
strengthening is characterized by Step-type yielding that accompanies yield elongation.
Second topic relates the effect of temper rolling (about 1% of cold rolling). Temper rolling was developed to
suppress the stretcher strain in pressed thin-plate steels. By temper rolling, the step-type yielding has disappeared and yield
stress drops greatly3). The role of temper rolling has been explained in connection with the Cottrell locking until now, but it is
true that such a yield drop occurs in high purity poly crystalline iron4). This result suggests that the behavior of yield drop by
temper rolling is not due to the Cottrell locking but due to the change of yielding mechanism.
Third topic is the effect of grain size on the yielding behavior in cold worked ferritic steels. In annealed steels,
yielding occurs by dislocation emission from grain boundary (grain boundary dislocation yielding) but, in cold worked ferritic
steels, it occurs by the movement of dislocations in matrix (inner grain dislocation yielding)5). With decreasing grain size, the
dislocation density increases even at the same percentage of cold rolling and this leads to the increase of yield stress based on
the mechanism of inner grain dislocation yielding6,7).
3. Effect of cold rolling on the yielding behavior of poly crystalline ferritic steels
Yielding of annealed ferric steels is characterized by the step-type yielding, as shown in Fig. 1. However, the step
yielding disappears by temper rolling and yield stress drops greatly. Fig. 2 shows the effect of prior cold rolling on the yield
stress4). It is interesting that yield drop appears in not only 0.04%C steel but also de-carburized steel (high purity iron). This
result indicates that the behavior of yield drop does not relate the existence of carbon. The difference of yield stress (0.07 GPa)
in annealed specimens can be explained by the change of HP coefficient at the grain size 40μm. In the deformation range
after the yield drop, yield stress is higher in 0.04%C steel than de-carburized steel. Since the data of dislocation density ρ
are reported in the original paper4), data of yield stress are re-plotted in Fig. 3 as a function of ρ. The data after yield drop
4. Effect of grain size on the yield stress of cold worked ferritic steels
Grain size affects the yield stress of cold worked specimens as well as annealed specimens. Fig. 4 shows the effect
of grain size on yield stress and dislocation density in cold rolled specimens6,7). At a same percentage of deformation, yield
stress becomes higher with decreasing grain size although it also depends on the amount of deformation. This result makes us
imagine that some additional relationship exists between grain refinement strengthening and dislocation strengthening.
However, it should be noted that dislocation density is markedly changed depending on the grain size: At a same percentage
of deformation, dislocation density is higher in specimens with smaller grain size. Especially in the case of 1.3μm specimen,
dislocation density has been heightened to 1×1015/m2 by 10% cold rolling, that is same level as that in martensite10).
Regarding the effect of grain size on dislocation density, two mechanisms may be considered6,7). One is an increase of misfit
dislocations around grain boundary with increasing the total grain boundary area. The other one is a decrease in moving
The contribution of dislocation strengthening should be evaluated by the increment of strength Δσ (=σy -σ0).
Fig. 6 summarizes the values of Δσ for the specimens used in Fig. 3 and Fig. 511). The result in Fig. 6 proves that the yield
stress of cold worked specimens is governed by only dislocation density, although that is changeable depending on the
chemical composition and grain size of steels even at a same percentage of deformation.
Summary
Fig. 8 shows the yielding mechanisms in poly crystalline metals11). In annealed metals, yielding is governed by the
mechanism of grain boundary dislocation yielding and yield stress follows Hall-Petch relation as a function of grain size d.
Hall-Petch coefficient ky is affected by the elements segregated at grain boundary. However, in metals containing dislocations
within matrix, yielding is governed by the mechanism of inner grain dislocation yielding and yield stress follows
Bailey-Hirsch relation as a function of dislocation density ρ. The character and distribution of dislocations should affect the
Bailey-Hirsch coefficient kd and the term of kd√ρ is affected by 1) deformation condition (temperature, strain rate,
deformation mode, etc.) 2) chemical composition and 3) microstructure (grain size, texture, precipitates, etc.). When annealed
steels have been cold worked, yielding mechanism changes from grain boundary dislocation yielding to inner grain
dislocation yielding. In the case the amount of deformation is small, yield drop appears due to the change of yielding
mechanism.
References
1) J.R.Low Jr. and M.Gensamer: Trans. AIME, 158 (1944), 207-249.
2) K.Takeda, N.Nakada, T.Tsuchiyama and S.Takaki: ISIJ Int., 48 (2008), pp.1122-1125.
3) A.J.K.Honeyman: Sheet Metal Industries, 34 (1957), 51-65.
4) A.J.Goldman: Trans. ASM, 57 (1964), 900-908.
5) S.Takaki: J. Japan Inst. Met. Mater., 83 (2019), pp.107-118.
6) Y.Tanaka, S.Takaki, T.Tsuchiyama and R.Uemori: TETSU-To-HAGANE, 104 (2018), 284-291.
7) Y.Tanaka, S.Takaki, T.Tsuchiyama and R.Uemori: ISIJ Int., 58 (2018),1927-1933.
8) W.B.Morrison: Trans. ASM, 59 (1966), 824-846.
9) M.Eto, S.Fukushima, T.Sasaki, Y.Haraguchi, K.Miyata, M.Wakita, T.Tomida, N.Imai, M.Yoshida and Y.Okada: ISIJ Int.,
48 (2008), 1142-1147.
10) D.Akama, T.Tsuchiyama and S.Takaki: ISIJ Int., 56 (2016), 1675-1680.
11) S.Takaki: Ferrum, 26 (2021), 423-426.
12) T.Ungár and A.Borbély: Appl. Phys. Letter, 69 (1996), 3173-3175.
13) S.Takaki: TETSU-To-HAGANE, 107 (2021), 780-783.