Materials Science & Engineering A 793 (2020) 139835

Contents lists available at ScienceDirect

Materials Science & Engineering A

journal homepage: http://www.elsevier.com/locate/msea

A novel 13Cr austenitic stainless steel with excellent mechanical properties

and high hydrogen embrittlement resistance via heterostructure and
TRIP effects
Jun He a, b, Lin Chen a, b, Zihui Guo a, b, Huihui Zhi a, b, Stoichko Antonov c, Yanjing Su a, b, *
a
Beijing Advanced Innovation Center for Materials Genome Engineering, University of Science and Technology Beijing, Beijing, 100083, PR China
b
Corrosion and Protection Center, University of Science and Technology Beijing, Beijing, 100083, PR China
c
Department of Microstructure Physics and Alloy Design, Max-Planck-Institut für Eisenforschung, Düsseldorf, 40237, Germany

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

Keywords: A 13Cr austenitic stainless steel with a good combination of high yield strength, ductility and hydrogen
Heterostructure embrittlement resistance was designed by combining hetero-deformation induced strengthening and martensitic
Martensitic transformation transformation induced plasticity effects. Room temperature metastable austenite is promoted by adding 8 wt%
Hydrogen embrittlement
Mn to a 13Cr–5Ni–2Mo supermartensitic stainless steel. A heterostructure consisting of remaining martensite
13Cr supermartensitic stainless steel
grains located between lath austenite grains and micrometer-sized austenite grains embedded inside ultrafine
grains, was fabricated by cold rolling (44% reduction) and annealing at 973 K through a shear reversion process.
The novel 13Cr austenitic stainless steel exhibits a yield strength of 923 MPa, tensile strength of 1085 MPa and
total elongation of 33.2%. The hydrogen embrittlement resistance of the novel steel was determined by incre­
mental step loading technique testing and was increased 40% compared with that of the 13Cr–5Ni–2Mo
supermartensitic stainless steel. This study provides an approach for the design and industrial fabrication of high
strength (125 ksi grade level), high ductility and hydrogen embrittlement resistant stainless steels that can be
used in harsh service environments containing hydrogen.

1. Introduction steels. For example, Zou et al. [10] presented a 13Cr–4Ni–1Mo

martensitic stainless steel with a good combination of tensile strength
Due to their high strength and relatively low cost, 13Cr martensitic (905 MPa), yield strength (832 MPa) and elongation (18%) achieved by
stainless steels are widely used for oil and gas pipelines [1,2], turbine introducing reversed austenite into the lath martensite matrix. Through
blades [3,4] and boiler components in power generation applications slow strain rate testing during which specimens were electrochemically
[5]. However, these steels are prone to premature failure fracture, as charged with hydrogen, it was further shown that the introduction of the
exhibited in many industrial applications, that is induced by stress additional reversed austenite in this steel can improve the HE resistance
corrosion cracking (SCC) combined with hydrogen induced cracking [11], and the corresponding fracture mode changed from intergranular
[6–8]. In general, it is widely recognized that the higher the strength of fracture to quasi-cleavage fracture. However, Solheim et al. [12]
the steel, the higher its susceptibility to hydrogen embrittlement (HE), claimed that the ductility of tensile test specimens was greatly reduced
and thus, although the martensite structure can be used to achieve ul­ by the presence of retained austenite when the microstructure was
trahigh strength levels, the actual yield strength of these material must charged with hydrogen prior to testing. Moreover, the HE susceptibility
be reduced during component design to compensate for its poor HE of 13Cr precipitation hardened martensitic stainless steel first decreases
resistance in industrial applications. For example, the typical yield and then increases with increasing aging time owing to the co-effect of
strength requirements of 13Cr martensitic stainless steel are only 550 precipitation and presence of reversed austenite [13]. Nevertheless, the
MPa (80 ksi) in harsh service environments containing hydrogen [9]. above studies have not been able to improve the ductility and HE
In the past few decades, many studies have focused on improving the resistance without decreasing the yield strength.
ductility and resistance to HE and SCC of 13Cr martensitic stainless Austenite has superior toughness and a higher HE resistance

* Corresponding author. Corrosion and Protection Center, University of Science and Technology Beijing, Beijing, 100083, PR China.
E-mail address: [email protected] (Y. Su).

https://doi.org/10.1016/j.msea.2020.139835
Received 24 April 2020; Received in revised form 23 June 2020; Accepted 25 June 2020
Available online 29 June 2020
0921-5093/© 2020 Elsevier B.V. All rights reserved.
J. He et al. Materials Science & Engineering A 793 (2020) 139835

compared to martensite due to its face-centered cubic structure [14]. 2. Material and processing design
However, the strength of austenitic stainless steels is too low for appli­
cation in certain industries [15]. Therefore, novel ways of improving the A commercial grade 13Cr–5Ni–2Mo (termed S13Cr) super­
strength of these HE tolerant alloys to meet industrial requirements martensitic stainless steel obtained from TAI steel corporation was used
would be highly advantageous. The most utilized method for achieving as the base alloy; the composition is provided in Table 1.
high strength levels in these steels is through grain refinement via severe The martensitic transformation temperature (Ms) is a very useful
cold deformation and subsequent flash annealing, which cannot be parameter for estimating the presence and relative stability of austenite
widely used in industrial production. Recently, heterostructures, such as at room temperature and can be determined from the composition using
heterogeneous lamella structures [16], laminate structures [17], and the Hammond empirical equation [30], as shown below:
harmonic structures [18,19] have been reported to result in excellent
mechanical properties through a mechanism known as back stress Ms (K) ¼ 1578-41.7Cr-61.1Ni-33.3Mn-27.8Si-36.1Mo-1667(C þ N) (1)
strengthening [20]. Heterostructures convert the applied uniaxial stress To design a high-performance 13Cr austenitic stainless steel with a
to a multiaxial stress due to the evolution of incompatible deformation, heterostructure and TRIP effects, 8 wt% Mn was added to the
producing excellent intrinsic strength [21]. In particular, the yield 13Cr–5Ni–2Mo supermartensitic stainless steel. This resulted in a Ms of
strength of the heterostructure is much higher than that predicted by the 300 K, thereby allowing for a metastable austenite at room temperature.
rule of mixtures [22]. Zhu et al. [23] accurately described it as The designed alloy was named A13Cr and was manufactured by vacuum
hetero-deformation induced (HDI) strengthening. The creation of addi­ induction melting, chemical compositions also provided in Table 1.
tional interfaces is essential, such that the hard domains can effectively The cast ingot was subsequently subjected to hot forging at 1473 K
constrain the soft domains and increase the level of the HDI stress. Using and solution treatment at 1373 K for 1 h to obtain a nonporous and
the HDI strengthening approach, a combination of high strength and uniform austenitic microstructure. Then, the 18 mm thick slab was cold
ductility can be achieved by tailoring the microstructure without rolled to 10 mm (44% reduction) in order to obtain a martensitic matrix.
changing the chemical composition. Next, a suitable annealing temperature was determined, allowing for
Generally, high strength steels require sufficiently high strain hard­ certain amount of retained lath martensite to strengthen the austenite.
ening rates to achieve a high uniform elongation during plastic defor­ Compared to a diffusional reversion process, a shear reversion process
mation [24]. Although geometrically necessary dislocations induced by can produce retained lath martensite distributed between lath austenite
mechanical incompatibility can provide a certain amount of work [31], which creates additional phase interfaces and effectively increases
hardening, additional hardening mechanisms are also desirable to ach­ the strength by HDI strengthening.
ieve good property combinations. Transformation induced plasticity The reversion mechanism is mainly related to the difference between
(TRIP) is one of the most effective ways to improve the work hardening the Gibbs free energy of α and γ, which is a function of the chemical
rate and ductility, as observed in TRIP steels [25,26], bainite steel [27], composition and temperature as shown in Ref. [32]:

�
ΔGα→γ ¼ 10 2 ΔGαFe→γ 100 Creq Nieq 2
97:5Creq þ 2:02Creq 108:8Nieq þ 0:52Ni2eq (1)

� �
0:05Creq Nieq þ 10 3 T 73:3Creq 2
0:67Creq þ 50:2Nieq 0:84Ni2eq 1:51Creq Nieq (2)

Quenching-Partitioning steels [28] and stainless steels [29].

In this work, we propose a combination of HDI strengthening and the
TRIP effect to improve both the strength and ductility of an austenitic
stainless steel, that has an intrinsically good HE resistance. The micro­
Nieq ¼ xNi þ 0:6xMn þ 20xC þ 4xN 0:4xSi (3)
structural stability of the alloy was tailored by adding Mn to a
13Cr–5Ni–2Mo supermartensitic stainless steel, such that a shear
Creq ¼ xCr þ 4:5xMo (4)
reversion mechanism can be utilized to produce a heterostructure of
remaining martensite grains located between lath austenite grains and where ΔGαFe→γ is the free energy difference between the α and γ phases in
micrometer-sized austenite grains embedded in ultrafine grains during a
pure iron [33], T is the temperature in Kelvin, and the Nieq and Creq
medium reduction cold rolling and annealing treatment. As a result, a
represent the Ni and Cr equivalent concentrations as defined by Eq. (3)
novel 13Cr austenitic stainless steel that has the same yield strength as
and Eq. (4), respectively, where xj is the concentration of the element j.
the martensitic steel but is much more ductile and resistant to HE was
The calculated Gibbs free energy change for the α to γ transformation of
fabricated.
the designed A13Cr steel with respect to temperature is shown in Fig. 1

Table 1
Chemical compositions (wt.%) of the commercial 13Cr supermartensitic stainless steel the designed 13Cr austenitic stainless steel.
C Si Mn Cr Ni Mo Nb Ti V N

S13Cr 0.01 0.22 0.57 13.1 5.09 2.09 0.02 0.05

A13Cr 0.04 0.27 7.86 12.9 4.75 1.94 0.09 0.04 0.24 0.03

2
J. He et al. Materials Science & Engineering A 793 (2020) 139835

Fig. 1. (a) The relationship between the Gibbs free energy difference for the transformation from martensite to austenite and temperature for the designed steel. (b)
XRD results of the cold rolled sample and subsequently annealed samples at different temperatures for 1 h.

Fig. 2. A schematic illustration of the thermomechanical processing applied to the modified steel and the corresponding microstructure schematic at each processing
stage (A13Cr-CG: solution treatment; A13Cr-CR: cold rolling; A13Cr-HS: annealing treatment).

(a). When the value is negative, γ is more stable than α. It has been mechanical properties, monotonic tension tests were performed at a
pointed out that the shear mechanism occurs when the Gibbs free energy strain rate of 10 4/s using an MTS Landmark machine. The HE sus­
difference change for the α to γ transformation is greater than 500 J/mol ceptibility was gauged by the hydrogen induced threshold stress ob­
[31]. Therefore, the transition temperature from the diffusional reverse tained through the incremental step loading technique (ISLT) testing
mechanism to shear reverse mechanism in the A13Cr steel was deter­ during which the specimen is electrochemical hydrogen charged at 2
mined to be 910 K. To maintain a certain amount of martensite upon mA/cm2 in 0.2 mol/L NaOH in accordance with the ASTM F1624
annealing, the temperature should not be much higher. To confirm this, standard. A load-unload-reload (LUR) test was carried out to deter­
the cold rolled sample was annealed at various temperatures for 1 h and mine the HDI stress [34]. The sample was preloaded to a given strain
X-ray diffraction (XRD) was used to determine the microstructure con­ level at a strain rate of 10 4/s and immediately unloaded by load control
stituents. Fig. 1(b) illustrates that all martensite is fully transformed at a speed of 40 N/s to 4 MPa, and then reloaded to the next studied
back to austenite during annealing at 1053 K—notice the absence of the strain point with the same strain rate.
(110)α at this temperature in the inset of Fig. 1(b). Therefore, an Multi-scale microstructural characterization was conducted using a
annealing temperature of 973 K was selected. Fig. 2 illustrates the Zeiss Auriga 39–87 Gemini field emission scanning electron microscope
thermomechanical processing after the hot forging of the designed steel. (SEM) equipped with an electron backscatter diffraction (EBSD) detec­
For convenience, the samples of the A13Cr steel after each treatment tor, and with a JEM G2100 transmission electron microscope (TEM)
step are named as A13Cr-CG (coarse grain), A13Cr-CR (cold rolling) and capable of energy dispersive x-ray spectroscope (EDX). The EBSD mea­
A13Cr-HS (heterostructure) base on their microstructure, as schema­ surements were performed at 20 kV accelerating voltage and a step size
tized in Fig. 2. of 0.04–0.5 μm, depending on the fineness of the microstructure. The
XRD measurements were carried out at a scan rate of 2� /min ranging
3. Mechanical properties and microstructural characterization from 40� to 100� . Rietveld analysis within the TOPAS software package
was used to quantify the austenite content of each sample. The samples
Dog-bone-type samples with dimensions of 25 � 6 � 2 mm3 were for EBSD and XRD tests were prepared using electropolishing in 12.5%
prepared according to the ASTM E8M standard. To evaluate the HClO4 þ 37.5% CH3COOH þ50% CH3CH2OH solution at 25 V, 278 K for

3
J. He et al. Materials Science & Engineering A 793 (2020) 139835

Fig. 3. EBSD obtained inverse pole figure, phase and band contrast maps for the (a) A13Cr-CG sample (0.5 μm step size) and (b) A13Cr-HS sample (0.2 μm step size),
with (c) a higher magnification scan (0.04 μm step size) in the area indicated in (b).

30 s. regions. The coarse-grained regions represent austenite that was

retained after cold rolling. Since the stored strain energy and crystal
4. Results defects in the deformed austenite are less than those in the deformed
martensite during cold rolling [35], only recovery and dislocation
4.1. Microstructure annihilation occurs in those regions during intercritical annealing,
resulting in the coarse austenitic grains. A higher magnification of a
The coarse austenite grain microstructure in the A13Cr-CG sample is fine-grained region in Fig. 3(b) is provided in Fig. 3(c). It shows that a
shown in Fig. 3(a), where the average size of the austenite grains is 35 certain amount of martensite was retained and embedded in the bands of
μm. The A13Cr-CG sample does not fully comprise of austenite, but the reversed austenite. A direct comparison of the grain size distribu­
contains a certain amount of martensite, formed during quenching, in­ tions (Fig. 4) shows that the prior coarse austenite grains were refined to
side the austenitic grains. the nanoscale by recrystallization. To gain further insight into the sub­
After 44% cold rolling, most of the austenite is transformed into structure of these recrystallized fine-grained regions, TEM analysis was
martensite, and only 16.03% retained austenite is detected during employed.
analysis of the A13Cr-CR sample XRD spectrum in Fig. 1(b). Fig. 5(a) shows a TEM micrograph of a typical the fine-grained re­
The microstructure of the heterogeneous structured sample, A13Cr- gion in the A13Cr-HS sample, consisting of distinct equiaxed and lath
HS, is shown in Fig. 3(b), and it consist of fine- and coarse-grained grain areas. The selected area diffraction (SAD) patterns in Fig. 5(c)

4
J. He et al. Materials Science & Engineering A 793 (2020) 139835

indicates that the equiaxed grain area contains only fine austenitic total elongation of the A13Cr-HS sample is approximately twice that of
grains with an average grain size of approximately 500 nm. The lath the S13Cr sample, and the uniform elongation is much higher, 28.6% vs.
grain area has a relatively higher dislocation density than the equiaxed 9.2%, respectively. Moreover, the heterostructure effectively increase
grain area. A higher magnification micrograph from this area is shown in the yield strength (three times as much) compared to that of the A13Cr-
Fig. 5(d). The SAD patterns reveals that it consists of mostly austenite CG sample.
with a small amount of martensite. Two types of martensite grains, α1 The corresponding work hardening rate curves are shown in Fig. 6
and α2, are observed. The α1 martensite lath is smaller than the SAD (b). The work hardening curve of S13Cr keeps declining with logarith­
aperture, hence its diffraction pattern also includes the adjacent γ1 mic due to low work hardening capacity of martensite. At the same time,
austenite. The SAD of α1 and adjacent γ1 is indexed as (110)α//(111)γ, the work hardening curves of A13Cr-CG and A13Cr-HS samples conform
[111]α//[110]γ, indicating that the reversed lath austenite maintains a to the typical three stage TRIP steel deformation mechanism [38]. Take
typical K–S orientation relationship with the remaining lath martensite A13Cr-HS sample as an example. The first rapid drop is due to the low
[36]. There are multiple diffraction spots that lies on a perfect circle in work hardening capability of dislocation hardening. The austenite
the SAD pattern of the α2 martensite area, arising from different {110} content of A13Cr-HS sample at different strains was obtained by inter­
crystal planes in the martensitic orientation family, which indicates that rupting the tensile tests and the values are shown in Fig. 6(d). Once the
the α2 martensite is a martensite bundle composed of multiple martensitic transformation starts, the work hardening rate begins to rise.
martensitic laths. For easier visualization of the lath structured area in The third stage, where the work hardening rate curves start to decrease
Fig. 5(d), a simplified phase distribution diagram based on the SAD re­ again, occur at 15% engineering strain. EBSD data obtained from this
sults is shown in Fig. 5(e). It should be noted that no precipitation was strain level, presented in Fig. 7, shows that coarse austenitic grains no
observed for this material. During the annealing process, the shear longer exist. At this point, some dislocations have already accumulated
reverse formed lath austenite can be subdivided into subgrains or within the small austenitic grains. Since the martensitic transformation
dislocation cells by a recovery process. The subgrains eventually change involve the movement of glissile dislocations at the austenite/martensite
to recrystallized equiaxed grains by lattice rotation or cell boundary interface, the dislocations that have formed in the small austenitic grains
merging [37]. This explains the existence of equiaxed austenite grains may hinder further motion of glissile dislocations and thus inhibit the
that form during the shear reversion mechanism. completion of the martensitic transformation [39,40]. Consequently, the
Fig. 5(f) shows the compositional profile of Ni and Mn along the entangled dislocations present in the fine grains will lead to a decrease in
lime-green line in Fig. 5(d). The content of these austenite stabilizing the austenite transformation rate.
elements is lower in the remaining lath martensite than in the sur­ The ISLT measurement curves for the S13Cr and A13Cr-HS samples
rounding reversed austenite. The shear reversion mechanism is a dif­ tested in a 0.2 M NaOH solution at 2 mA/cm2 are shown in Fig. 8. The
fusionless phase transformation. Thus, we suggest that due to the hydrogen induced threshold stress is determined as the maximum
common inhomogeneous composition of the steel, the martensite con­ loading stress without fracture in the 4th grade test. The threshold stress
taining higher content of austenite stabilizing elements (Ni and Mn) has for HE for the A13Cr-HS sample is 860 MPa, which is approximately 250
a higher chemical driving force for phase transformation and is the first MPa higher than that of the S13Cr sample (613 MPa). This difference is
region to be transformed into lath austenite. The shear reversion, being a ascribed to the austenitization of the microstructure of A13Cr-HS
diffusionless transformation, is rather quick, but during the annealing sample.
process, subsequent diffusion of austenite stabilizing elements to the The fracture surfaces of the 4th grade test sample are shown in Fig. 8
austenite phase occurs and depletes these solutes from the martensite. (b) and (d) for S13Cr and A13Cr-HS, respectively. For the S13Cr sample,
This makes it more difficult for the remaining martensite to reverse into intergranular (Figs. 8(b-1)) and quasi-cleavage (Figs. 8(b-2)) are the
austenite and it consequently remains untransformed. dominant fracture modes, along with some dimples at the edge of the
fracture surface, Figs. 8(b-3). As shown in the inset of Figs. 8(b-1), slip
4.2. Mechanical properties bands can be seen on the crystal surfaces. This is to say, the S13Cr
sample exhibits mainly a brittle fracture. Meanwhile, most of the frac­
The engineering stress-strain curves of the S13Cr, A13Cr-HS, and ture surface of the A13Cr-HS sample exhibits dimples, Figs. 8(d-3),
A13Cr-CG samples are shown in Fig. 6(a), where the respective XRD suggestive of some ductility, with brittle fracture features at the edge. It
spectra of these samples before and after the tensile testing are shown in can be seen from Figs. 8(d-1) and (d-2) that this brittle zone is in the
Fig. 6(c). The mechanical properties and determined austenite content form of quasi-cleavage fracture. The edges, showing uneven tearing at
are listed in Table 2. While exhibiting comparable strength levels, the the quasi cleavage fracture is caused by the strong dislocation movement

Fig. 4. (a) Grain size distribution of the coarse austenitic grains in the A13Cr-CG sample. (b) Grain size distribution of the fine austenitic grains in the A13Cr-
HS sample.

5
J. He et al. Materials Science & Engineering A 793 (2020) 139835

Fig. 5. (a) TEM micrograph from the fine-grained region of the A13Cr-HS sample showing that it consists of equiaxed grains and lath grains; (b) The equiaxed region
is composed of austenitic grains as shown in the (c) corresponding SAD pattern; (d) The lath region is comprised of martensite and austenite phase as determined by
the (g–i) SAD patterns obtained from the regions marked with respective color arrows; (e) is a simplified phase distribution diagram of (d), and (e) is the
compositional profile along the lime-green solid line in (d). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version
of this article.)

at the intersection of two cleavage planes [41]. The maximum depth of but also a low stacking fault energy (SFE) was obtained.
the hydrogen influenced zone is approximately 260 μm. In general, Saeed-Akbari et al. [42]. directly illustrated a polynomial trend for
fracture of the A13Cr-HS sample occurred at the elastic-plastic transition the SFE with respect to increases in the Mn content, where it decreased
stage. The large amount of dimples indicates that the A13Cr-HS sample at first and then increased again. The inversion point was around a Mn
still has good toughness, while the S13Cr sample is completely brittle. content of 12%–15%. It is therefore clear that the addition of 8 wt% Mn
to the 13Cr–5Ni–2Mo supermartensitic stainless steel decreased the SFE
5. Discussion of the obtained austenite. The plastic deformation of low SFE austenite is
characterized by the dissociation of perfect dislocations to Shockley
5.1. Effect of Mn on the austenite stability partial dislocations and the formation of wide stacking faults [43]. In
addition, strain induced martensite nucleates at the interaction of the
Based on the presented experimental results, the designed 13Cr shear bands during plastic deformation [44]. Shear bands is a collective
austenitic stainless steel exhibits poor microstructural stability. After term for the planar defects that form as a result of the overlapped
solution treatment, the A13Cr-CG sample is not fully austenitic phase, stacking faults [14]. Since there is a critical thickness for a martensite
but contains 8.0% of newly formed martensite. Furthermore, a large nucleus, which is 13.5 [45] and 27 [46] closed-packed atom planes,
portion of the austenite in the A13Cr-CG sample is transformed into significant overlapping of the stacking faults must occur before
martensite upon tensile deformation, resulting in only 14.3% of retained martensite nucleation can take place. A decreased SFE due to the addi­
austenite. The Ms temperature is a very useful parameter for estimating tion of Mn can promote the formation of shear bands with a sufficient
the stability of the austenite with respect to the martensitic phase thickness. The growth of martensite is described as the repeated nucle­
transformation at room temperature. The microstructural stability is ation of new nuclei and their coalescence [47]. Therefore, the addition
governed by the difference in e Gibbs free energies of the two phases. of 8% Mn can promote strain induced martensite. The XRD results in
Reducing Ms below 300 K allows for austenite to be present in the Table 2 show that the austenite content of the A13Cr-CG sample is only
microstructure at the room temperature, with its stability increasing 14.3% after the fracture. The maximum engineering strain of the
with decreasing of Ms. According to Hammond’s empirical equation, all A13Cr-CG sample in tensile deformation is 0.546, which is equivalent to
common alloying elements can reduce the Ms temperature, however, Mn a logarithmic strain of 0.436. Therefore, we can achieve a martensitic
was selected to promote austenite metastability in this study. By adding matrix with only 44% cold rolling. This procedure does not require a
Mn, not only was a small difference in the Gibbs free energy achieved, heavy reduction, and can be easily accomplished in industrial

6
J. He et al. Materials Science & Engineering A 793 (2020) 139835

Fig. 6. (a) Engineering stress vs. strain curves of the A13Cr-HS, A13Cr-CG, and S13 samples; (b) True stress vs. logarithmic strain curves and working hardening rate
curves of the A13Cr-HS, A13Cr-CG, and S13 samples; (c) XRD spectra of the A13Cr-HS, A13Cr-CG, and S13 samples before and after the tensile testing; and (d)
Austenite content vs. logarithmic strain curve of the A13Cr-HS sample.

production. Moreover, the austenite remains metastable in A13Cr-HS

Table 2
sample since there is no modification of the alloy composition. The
The mechanical properties and austenite content of S13Cr supermartensitic
strain induced martensitic transformation can provide sufficient work
stainless steel and the designed A13Cr-HS, A13Cr-CG austenitic stainless steel.
hardening rate, so that the material can maintain a good uniform
Sample YS (MPa) UTS (MPa) Elongation % Austenite content % elongation under high strength.
Total Uniform Original Fracture

S13Cr 912 988 16.3 9.2 0 0

5.2. Heterostructure induced strengthening
A13Cr-HS 923 1085 33.2 28.6 87.7 16.3
A13Cr-CG 278 904 54.6 51.2 92.0 14.3
The yield strength of the designed A13Cr-HS sample is 923 MPa,
YS: Yield Strength.
which can be considered quite high for an austenitic stainless steel. At
UTS: Ultimate Tensile Strength.
the beginning of the deformation, both soft austenite and hard
martensite domains are elastically deformed, similar to conventional

Fig. 7. EBSD acquired inverse pole figure, phase and band contrast maps (0.1 μm step size) of the A13Cr-HS sample upon 15% engineering strain.

7
J. He et al. Materials Science & Engineering A 793 (2020) 139835

Fig. 8. ISLT results for the (a) S13Cr sample and (c) A13Cr-HS sample, with (b) and (d), the fracture surfaces of the S13Cr sample and A13Cr-HS sample from the 4th
ISLT, respectively.

homogeneous materials. As the stress continues to increase, the soft The yield strength of the CG sample σCG can be calculated using the
domains are first to generate plastic deformation through dislocations rule of mixtures:
slip. However, these domains cannot freely deform plastically, as the
σ CG ¼ V1 σLG þ ð1 V1 Þσα (5)
interfacial coherency needs to be preserved. Hence the adjacent domains
need to be deformed together with the soft domains, resulting in a where σα is the yield strength of the martensite, and σ LG is the yield
plastic strain gradient arises that needs to be accommodated by strength of the large austenite grains. The austenite content, V1 , in the
geometrically necessary dislocations. This inevitably leads to an in­ A13Cr-CG sample was determined to be 92%. When calculating the yield
crease in the measured global yield strength. strength of the A13Cr-HS sample, the HDI stress provides an additional
intrinsic strength σHDI , and σHS is calculated as follow:
σ HS ¼ V2 σα þ V3 σLG þ ð1 V2 V3 Þσ SG þ σHDI (6)
The martensite content, V2 , is 12.3%, and the large austenite grain
content, V3 , is the retained austenite content of the CG sample after cold
rolling, which is 16.3%. The yield strength of small austenite grains, σSG ,
can be calculated using the Hall-Petch relationship:

σ SG ¼ σLG þ kdSG1=2 (7)

1=2
kdLG
As already mentioned, the small austenite grain size, dSG , is 0.5 μm,
and the large austenite grain size, dLG , is 35 μm. The constant k is 274
MPa μm1/2 according to Rajasekhara et al. [48]. Combining Eqs. (5)–(7),
the heterostructure induced extra stress σHDI can be estimated as:
σ HDI ¼ 413 4:7%σα (8)
Considering σ α can vary with respect to many factors, we take the
maximum true stress of A13Cr-HS sample (1380 MPa) as the lower limit
of σ α , and 1500 MPa [48] as the upper limit of σ α . Hence, the estimated
Fig. 9. The unloading and reloading test hysteresis loops measured at varying σ HDI range is 343–348 MPa, which represents a very high contribution to
tensile strains and the corresponding HDI stress calculated using different the strength.
slope reductions. The HDI stress can be measured by the LUR test because it has the

8
J. He et al. Materials Science & Engineering A 793 (2020) 139835

same physical origin as the Bauschinger effect [49]. However, the real concentration of intergranular and cleavage cracking is much higher
hysteresis loop is not as well as defined loop. Thus we adapt the than that of martensite microstructure. Thus, the hydrogen influenced
approach from literature [34] and choose 5%, 10% and 15% slope re­ brittle zone of the designed A13Cr-HS sample exhibits a quasi-cleavage
ductions from the effective Young’s modulus. The results are shown in structure.
Fig. 9. The HDI stress contributes a significant amount of strength to the
flow stress at different strains. Specifically, at 3% strain, the measured 6. Conclusions
HDI stress is 454–485 MPa, while the flow stress is 955 MPa.
The measured HDI stress is slightly higher than that calculated using A novel 13Cr austenitic stainless steel, possessing excellent me­
the rule of mixtures. This is because the measurement was made at 3% chanical properties and high hydrogen embrittlement resistance, was
strain, which inevitably leads to dislocation multiplication, as evident by designed and fabricated with the aim of replacing traditional 13Cr
the 32 MPa higher flow stress than the yield strength. In addition, the supermartensitic stainless steel. The conclusions can be summarized as
fine-grained austenite also exists in two types, namely, equiaxed and follows:
laths grains. However, the HDI calculations assumes just one type, which
introduces an additional source of error. Nevertheless, despite the dif­ 1. The heterostructure of remaining martensite grain dispersed be­
ference, both the measured and calculated HDI stress attest to the large tween lath austenite grains and micrometer sized austenite grains
flow stress contribution, accounting for 37%–50% of the yield strength. embedded inside ultrafine grains was design by cold rolling at me­
dium reduction and annealing at a low temperature.
5.3. Hydrogen embrittlement mechanism 2. The metastable austenite renders the cold rolling process simple and
provides good uniform elongation. This is attributed to the addition
The HE resistance of the designed steel is greatly improved, as of Mn, which results in both a low Gibbs free energy difference from
evident by the ISLT test results in Fig. 8. By austenitizing the material, martensite to austenite transformation and low SFE of austenite.
the hydrogen influenced brittle zone is limited to the surface layer, and 3. The yield strength of A13Cr-HS sample was measured as 923 MPa.
is only ~260 μm deep. More importantly, even under a higher stress The primary source for such superior strength compared to the
state and long hydrogen attack time, the hydrogen induced brittle zone coarse grained material was the fine-grained strengthening and HDI
forms quasi cleavage features instead of intergranular cracks. strengthening. The HDI stress plays an important role in improving
It was noted that intergranular cracking are the dominant fracture yield strength, accounting for 37%–50%.
modes in the central part of the S13Cr sample. This is because hydrogen 4. The austenitized A13Cr-HS sample limits hydrogen uptake to a depth
effectively decreases the cohesive energy along crystallographic planes of ~260 μm from the specimen surface and the fracture mode is
or material interfaces when the hydrogen content exceeds the critical quasi-cleavage cracking due to the low hydrogen diffusivity, while
concentration. At the same time, quasi-cleavage features and slip bands S13Cr sample mainly exhibits intergranular cracking determined by
are also observed (Figs. 8(b-1) and (b-2)), suggesting that hydrogen hydrogen diffusion and re-distribution driven by stress.
promotes dislocation nucleation, emission and movement. Hence, to
account for the hydrogen embrittlement fracture behavior, a synergistic Declaration of competing interest
action and the interplay of two mechanisms, hydrogen enhanced deco­
hesion (HEDE) [50,51] and hydrogen enhanced localized plasticity The authors declare that they have no known competing financial
(HELP) [52,53] need to be considered. interests or personal relationships that could have appeared to influence
Multiple studies have shown that martensite is an extremely strong the work reported in this paper.
structure with high sensitivity to HE [13,54]. The hydrogen induced
threshold stress of the S13Cr sample, as measured by the ISLT test is 613 CRediT authorship contribution statement
MPa—indicating a low HE resistance. The central region of the fracture
surface is characterized by intergranular cracking, and at a certain dis­ Jun He: Methodology, Formal analysis, Investigation, Writing -
tance from the center, the surface transitions into quasi-cleavage original draft. Lin Chen: Writing - review & editing, Resources. Zihui
cracking—indicating a brittle nature. The entire test time of the fourth Guo: Investigation. Huihui Zhi: Investigation. Stoichko Antonov:
stage of the S13Cr sample was 76 h, and when considering the hydrogen Writing - review & editing. Yanjing Su: Conceptualization, Supervision,
diffusion coefficient in martensite, approximately 10 8 cm2/s [1], it Funding acquisition.
allows for ample time for hydrogen diffusion throughout the sample.
That is, hydrogen has enough time to diffuse to where the stress is Acknowledgements
concentrated and accumulate up to reach the critical hydrogen con­
centration. The maximum stress occurs in the center of the rectangular This work was financially supported by the National Natural Science
cross-section sample. The existing stress field promotes hydrogen Foundation of China (Grant No. 51571028), and 111 Project (No.
diffusion, and the hydrogen atoms can lower the interaction energies B170003). S.A. acknowledges financial support from the Alexander von
and forces between dislocations by elastic shielding [55] and enhance Humboldt Foundation.
dislocation mobility. Consequently, hydrogen increases the dislocation
density close to the prior austenite grain and lath boundaries which in Appendix A. Supplementary data
turn attracts more hydrogen, until the critical hydrogen concentration is
achieved and the hydrogen induced crack is generated. Supplementary data to this article can be found online at https://doi.
Although the designed A13Cr-HS sample contains 8.0% martensite, org/10.1016/j.msea.2020.139835.
the martensite is dispersed throughout the austenite and is not inter­
connected. Therefore, it cannot act as a fast hydrogen diffusion channel. References
Instead, the martensite/austenite interface may be a hydrogen trap site
that hinders the hydrogen transportation. Based on such a structural [1] M. Kimura, Y. Miyata, T. Toyooka, Y. Kitahaba, Effect of retained austenite on
corrosion performance for modified 13% Cr steel pipe, Corrosion 57 (2001)
design and low hydrogen diffusion rate of austenite, the hydrogen 433–439.
influenced brittle zone is limited to only 260 μm from the surface layer, [2] C. Gesnouin, A. Hazarabedian, P. Bruzzoni, J. Ovejero-Garcı ́a, P. Bilmes,
even though the test time and stress state of the A13Cr-HS sample are C. Llorente, Effect of post-weld heat treatment on the microstructure and hydrogen
permeation of 13CrNiMo steels, Corrosion Sci. 46 (2004) 1633–1647.
longer/larger than those of the S13Cr sample. As the austenite structure
has a strong ability to dissolve hydrogen, the critical hydrogen

9
J. He et al. Materials Science & Engineering A 793 (2020) 139835

[3] L. Tsay, Y. Chang, S. Torng, H. Wu, Improved impact toughness of 13Cr martensitic [29] A. Hamada, J. K€ omi, Effect of microstructure on mechanical properties of a novel
stainless steel hardened by laser, J. Mater. Eng. Perform. 11 (2002) 422–427. high-Mn TWIP stainless steel bearing vanadium, Mater. Sci. Eng. 718 (2018)
[4] S. Prifiharni, H. Perdana, T.B. Romijarso, B. Adjiantoro, A. Juniarsih, E. Mabruri, 301–304.
The hardness and microstructure of the modified 13Cr steam turbine blade steel in [30] C. Hammond, Mise au point d’aciers inoxydables du type maraging contenant du
tempered conditions, Int. J. Eng. Technol. 8 (2016) 2672–2675. cobalt, Cobalt 25 (1964) 195–202.
[5] L.B. Niu, K. Nakada, Effect of chloride and sulfate ions in simulated boiler water on [31] K. Tomimura, S. Takaki, Y. Tokunaga, Reversion mechanism from deformation
pitting corrosion behavior of 13Cr steel, Corrosion Sci. 96 (2015) 171–177. induced martensite to austenite in metastable Austenitic stainless steels, ISIJ Int. 31
[6] J.W. Kim, Y.S. Kim, C.Y. Park, Failure analysis of cracking at volute tongues of (1991) 1431–1437.
feedwater pump casings, Eng. Fail. Anal. 9 (2002) 17–30. [32] M.C. Somani, P. Juntunen, L.P. Karjalainen, R.D.K. Misra, A. Kyr€ ol€
ainen, Enhanced
[7] D. Nie, X. Chen, Z. Fan, Q. Wu, Failure analysis of a slot-welded impeller of recycle mechanical properties through reversion in metastable Austenitic stainless steels,
hydrogen centrifugal compressor, Eng. Fail. Anal. 42 (2014) 1–9. Metall. Mater. Trans. 40 (2009) 729–744.
[8] C.O. Alonso, M. Lucio-Garcia, I. Hermoso-Diaz, J. Chacon-Nava, A. MArtinez- [33] A.T. Dinsdale, SGTE data for pure elements, Calphad 15 (1991) 317–425.
Villafa~ne, J. Gonzalez-Rodriguez, Detection of sulfide stress cracking in a [34] M. Yang, Y. Pan, F. Yuan, Y. Zhu, X. Wu, Back stress strengthening and strain
supermartensitic stainless steel by using electrochemical noise, Int. J. Electrochem. hardening in gradient structure, Mater. Res. Lett. 4 (2016) 145–151.
Sci. 9 (2014) 6717–6733. [35] I. Shakhova, V. Dudko, A. Belyakov, K. Tsuzaki, R. Kaibyshev, Effect of large strain
[9] P. Rhodes, Environment-assisted cracking of corrosion-resistant alloys in oil and cold rolling and subsequent annealing on microstructure and mechanical
gas production environments: a review, Corrosion 57 (2001) 923–966. properties of an austenitic stainless steel, Mater. Sci. Eng. 545 (2012) 176–186.
[10] D.N. Zou, H. Ying, W. Zhang, X.D. Fang, Influence of tempering process on [36] H. Liu, J. Liu, C. Luo, Z. Liu, Microstructure, crystallography of phase
mechanical properties of 00Cr13Ni4Mo supermartensitic stainless steel, J. Iron transformations and multiple precipitations in PH 15-7Mo stainless steel, J. Alloys
Steel Res. Int. 17 (2010) 50–54. Compd. 672 (2016) 386–392.
[11] Y.H. Fan, B. Zhang, H.L. Yi, G.S. Hao, Y.Y. Sun, J.Q. Wang, E.H. Han, W. Ke, The [37] S.J. Lee, Y.M. Park, Y.K. Lee, Reverse transformation mechanism of martensite to
role of reversed austenite in hydrogen embrittlement fracture of S41500 austenite in a metastable austenitic alloy, Mater. Sci. Eng. 515 (2009) 32–37.
martensitic stainless steel, Acta Mater. 139 (2017) 188–195. [38] J. Shi, X. Sun, M. Wang, W. Hui, H. Dong, W. Cao, Enhanced work-hardening
[12] K.G. Solheim, J.K. Solberg, J. Walmsley, F. Rosenqvist, T.H. Bjørnå, The role of behavior and mechanical properties in ultrafine-grained steels with large-
retained austenite in hydrogen embrittlement of supermartensitic stainless steel, fractioned metastable austenite, Scripta Mater. 63 (2010) 815–818.
Eng. Fail. Anal. 34 (2013) 140–149. [39] J. Breedis, W. Robertson, Martensitic transformation and plastic deformation in
[13] X. Li, J. Zhang, J. Chen, S. Shen, G. Yang, T. Wang, X. Song, Effect of aging iron alloy single crystals, Acta Metall. 11 (1963) 547–559.
treatment on hydrogen embrittlement of PH 13-8 Mo martensite stainless steel, [40] R. Lagneborgj, The martensite transformation in 18% Cr-8% Ni steels, Acta Metall.
Mater. Sci. Eng. 651 (2016) 474–485. 12 (1964) 823–843.
[14] K.H. Lo, C.H. Shek, J.K.L. Lai, Recent developments in stainless steels, Mater. Sci. [41] M.L. Martin, J.A. Fenske, G.S. Liu, P. Sofronis, I.M. Robertson, On the formation
Eng. R 65 (2009) 39–104. and nature of quasi-cleavage fracture surfaces in hydrogen embrittled steels, Acta
[15] G. Huang, D. Matlock, G. Krauss, Martensite formation, strain rate sensitivity, and Mater. 59 (2011) 1601–1606.
deformation behavior of type 304 stainless steel sheet, Metall. Trans. A 20 (1989) [42] A. Saeed-Akbari, J. Imlau, U. Prahl, W. Bleck, Derivation and variation in
1239–1246. composition-dependent stacking fault energy maps based on subregular solution
[16] X. Wu, M. Yang, F. Yuan, G. Wu, Y. Wei, X. Huang, Y. Zhu, Heterogeneous lamella model in high-manganese steels, Metall. Mater. Trans. 40 (2009) 3076–3090.
structure unites ultrafine-grain strength with coarse-grain ductility, Proc. Natl. [43] J. Talonen, H. H€ anninen, Formation of shear bands and strain-induced martensite
Acad. Sci. U. S. A 112 (2015) 14501–14505. during plastic deformation of metastable austenitic stainless steels, Acta Mater. 55
[17] X. Ma, C. Huang, J. Moering, M. Ruppert, H.W. H€ oppel, M. G€oken, J. Narayan, (2007) 6108–6118.
Y. Zhu, Mechanical properties of copper/bronze laminates: role of interfaces, Acta [44] G. Olson, M. Cohen, Kinetics of strain-induced martensitic nucleation, Metall.
Mater. 116 (2016) 43–52. Mater. Trans. 6 (1975) 791–795.
[18] C. Sawangrat, S. Kato, D. Orlov, K. Ameyama, Harmonic-structured copper: [45] G. Olson, M. Cohen, A general mechanism of martensitic nucleation: Part II.
performance and proof of fabrication concept based on severe plastic deformation FCC→BCC and other martensitic transformations, Metall. Trans A 7 (1976)
of powders, J. Mater. Sci. 49 (2014) 6579–6585. 1905–1914.
[19] S.K. Vajpai, M. Ota, T. Watanabe, R. Maeda, T. Sekiguchi, T. Kusaka, K. Ameyama, [46] K. Staudhammer, L. Murr, S. Hecker, Nucleation and evolution of strain-induced
The development of high performance Ti-6Al-4V alloy via a unique microstructural martensitic (bcc) embryos and substructure in stainless steel: a transmission
design with bimodal grain size distribution, Metall. Mater. Trans. 46 (2015) electron microscope study, Acta Metall. 31 (1983) 267–274.
903–914. [47] L. Murr, K. Staudhammer, S. Hecker, Effects of strain state and strain rate on
[20] X. Wu, Y. Zhu, Heterogeneous materials: a new class of materials with deformation-induced transformation in 304 stainless steel: Part II. Microstructural
unprecedented mechanical properties, Mater. Res. Lett. 5 (2017) 527–532. study, Metall. Trans. A 13 (1982) 627–635.
[21] X. Wu, P. Jiang, L. Chen, F. Yuan, Y.T. Zhu, Extraordinary strain hardening by [48] S. Rajasekhara, P.J. Ferreira, L.P. Karjalainen, A. Kyr€ol€
ainen, Hall–Petch behavior
gradient structure, Proc. Natl. Acad. Sci. U. S. A 111 (2014) 7197–7201. in ultra-fine-grained AISI 301LN stainless steel, Metall. Mater. Trans. 38 (2007)
[22] X.L. Wu, P. Jiang, L. Chen, J.F. Zhang, F.P. Yuan, Y.T. Zhu, Synergetic 1202–1210.
strengthening by gradient structure, Mater. Res. Lett. 2 (2014) 185–191. [49] F. Mompiou, D. Caillard, M. Legros, H. Mughrabi, In situ TEM observations of
[23] Y. Zhu, X. Wu, Perspective on hetero-deformation induced (HDI) hardening and reverse dislocation motion upon unloading in tensile-deformed UFG aluminium,
back stress, Mater. Res. Lett. 7 (2019) 393–398. Acta Mater. 60 (2012) 3402–3414.
[24] G. Niu, Q. Tang, H.S. Zurob, H. Wu, L. Xu, N. Gong, Strong and ductile steel via [50] R. Oriani, P. Josephic, Equilibrium and kinetic studies of the hydrogen-assisted
high dislocation density and heterogeneous nano/ultrafine grains, Mater. Sci. Eng. cracking of steel, Acta Metall. 25 (1977) 979–988.
759 (2019) 1–10. [51] W. Gerberich, R. Oriani, M.J. Lji, X. Chen, T. Foecke, The necessity of both
[25] K. Tao, H. Choo, H. Li, B. Clausen, J.E. Jin, Y.K. Lee, Transformation-induced plasticity and brittleness in the fracture thresholds of iron, Philos. Mag. A 63
plasticity in an ultrafine-grained steel: an in situ neutron diffraction study, Appl. (1991) 363–376.
Phys. Lett. 90 (2007) 101911. [52] C. Beachem, A new model for hydrogen-assisted cracking (hydrogen
[26] Y.S. Jung, Y.K. Lee, Effect of pre-deformation on the tensile properties of a “embrittlement”), Metall. Mater. Trans. B 3 (1972) 441–455.
metastable austenitic steel, Scripta Mater. 59 (2008) 47–50. [53] H.K. Birnbaum, P. Sofronis, Hydrogen-enhanced localized plasticity—a mechanism
[27] H. Bhadeshia, Nanostructured bainite, Proc. R. Soc. A 466 (2009) 3–18. for hydrogen-related fracture, Mater. Sci. Eng. 176 (1994) 191–202.
[28] E.J. Seo, L. Cho, Y. Estrin, B.C. De Cooman, Microstructure-mechanical properties [54] S. Lynch, A fractographic study of gaseous hydrogen embrittlement and liquid-
relationships for quenching and partitioning (Q&P) processed steel, Acta Mater. metal embrittlement in a tempered-martensitic steel, Perspectives in Hydrogen in
113 (2016) 124–139. Metals (1986) 739–750.
[55] P. Sofronis, H.K. Birnbaum, Mechanics of the hydrogen-dislocation-impurity
interactions-I. Increasing shear modulus, J. Mech. Phys. Solid. 43 (1995) 49–90.

10