You are on page 1of 10

Materials Science & Engineering A 880 (2023) 145279

Contents lists available at ScienceDirect

Materials Science & Engineering A


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

The behaviour and deformation mechanisms for 316L stainless steel


deformed at cryogenic temperatures
Suning Li a, Philip J. Withers b, Saurabh Kabra c, Kun Yan a, *
a
Department of Materials, The University of Manchester, Manchester, M13 9PL, United Kingdom
b
Henry Royce Institute, Department of Materials, The University of Manchester, M13 9PL, United Kingdom
c
ISIS Facility, STFC-UKRI, Rutherford Appleton Laboratory, Harwell Campus, OX11 0QX, United Kingdom

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

Keywords: The microstructural evolution and deformation mechanisms of 316L stainless steel (SS) have been investigated at
Martensitic transformation 297, 173, 50 and 15 K by in situ neutron diffraction during tensile loading and correlative transmission electron
Cryogenic deformation microscopy (TEM). The yield strength and ultimate tensile strength of 316L stainless steel are significantly
Stacking fault energy
improved at cryogenic temperatures. In contrast to room temperature, deformation-induced martensitic trans­
Austenitic stainless steels
Neutron diffraction
formation was observed at all cryogenic temperatures. The γ-austenite (FCC) content decreases and α′-martensite
(BCC) increases with increasing strain, a fraction of this γ to α′ transformation is accompanied by the transient
appearance of ε-martensite (HCP) as an intermediate phase. The maximum volume fraction of ε-martensite in­
creases with decreasing deformation temperature and reaches 13% for deformation at 15 K. TEM results confirm
that γ → ε → α′ and γ → α′ martensitic transformations occur during cryogenic deformation, while mechanical
twins were observed only at 173 K. The evolution of lattice strain, phase volume fraction, stacking fault prob­
ability (SFP) and stacking fault energy (SFE) were quantified to investigate the correlation between deformation
mechanisms and mechanical behaviour of 316L SS as a function of deformation temperature.

1. Introduction (hexagonal close-packed (HCP) and α′-martensite (body-centred cubic


(BCC) structures) maintain a coherent relationship with the parent
Austenitic stainless steels (SS) are generally face-centred-cubic (FCC) γ-austenite (FCC structures). The close-packed structures of the FCC and
alloys at room temperature and can undergo martensitic transformation HCP phases mean that the introduction of stacking faults can easily
under plastic deformation, significantly enhancing both strength and occur by martensitic transformation since the fundamental difference
ductility, referred to as the transformation induced plasticity (TRIP) between the two phases is the stacking sequence of the close-packed
effect [1–3]. This effect is exploited to develop steels having an excep­ planes. During deformation, the passage of Shockley partial disloca­
tional combination of high ductility and strength. 316L SS is one of the tion with Burger vector 1/6<11 2 > on neighbouring FCC (111) planes
best-known austenitic steels and extensively used in chemical, nuclear, creates a stacking fault. After this, the activation of a second Shockley
transportation and aerospace industries [4]. While not considered a partial dislocation gliding on the second (111) plane will form HCP
TRIP steel per se, it can undergo martensitic transformation under phase. Therefore, the stacking fault energy (SFE) of the material plays a
plastic deformation at cryogenic temperatures, significantly changing crucial role in the martensitic transformation process and hence on the
the mechanical properties [5,6]. Das et al. [7] investigated the kinetics mechanical properties and deformation mechanism of FCC steels [8,9].
of the martensitic transformation in austenitic steels at 300 K and at When the SFE is below ~18 mJ/m2, the TRIP effect i.e., martensitic
various low temperatures by in situ neutron diffraction and found a transformation is dominant, affecting subsequent deformation [10–12].
distinct change in the transformation behaviour due to the suppression For the SFE between ~18 and ~45 mJ/m2, the twinning-induced plas­
of slip at low temperatures. ticity (TWIP) effect tends to occur, resulting in the formation of twins
During the martensitic transformation, the reorganization of the during deformation. When the SFE exceeds ~45 mJ/m2, dislocation
crystal lattice is achieved through the medium and short-range migra­ glide exclusively controls plastic deformation [13]. Galindo-Nava et al.
tion of the matrix atoms collectively. The newly formed ε-martensite

* Corresponding author.
E-mail address: kunyan.callaghan@manchester.ac.uk (K. Yan).

https://doi.org/10.1016/j.msea.2023.145279
Received 14 April 2023; Received in revised form 8 June 2023; Accepted 9 June 2023
Available online 10 June 2023
0921-5093/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

[11] established deformation mechanism maps for the formation of ε, α′


and twinning over different strain levels and explored SFE effects in
microstructure by modifying deformation temperature and chemical
composition in austenitic steels.
The SFE can be affected by factors such as chemical composition,
temperature and grain size [14–16]. In general, the SFE decreases with
the falling temperature [17–19]. Numerous researchers have found a
positive correlation between the SFE of austenitic steels and tempera­
ture using first-principles or thermodynamic calculations and have
inferred that the SFE will be low or even negative at extremely cryogenic
temperature [17,20,21]. Therefore, low SFE plays a prominent role in
determining the competing deformation mechanisms of the 316L SS at
cryogenic temperature, which is significant for the mechanical proper­
ties and strengthening effect involving martensitic transformation [22].
Although there have been some studies looking at the effect of SFE on
the deformation of 316L SS as a function of temperature by in situ
neutron diffraction, these have been at room and elevated temperature
deformation. For example, Molnár et al. [16] found that as the defor­
mation temperature decreases (500-20 ◦ C), deformation twins gradually
replace dislocation slip as the main contributor to the plastic deforma­
tion of 316L SS due to the decrease in SFE. On the other hand, there have
been many post-mortem microstructural characterisation studies of the
deformation of 316L SS at cryogenic temperatures. Nalepka et al. [3]
inferred that 316L SS undergoes deformation-induced martensitic
transformation during deformation at 77 and 4 K by electron backscatter
diffraction and X-ray diffraction. Consequently, while the
deformation-induced martensitic transformation for low stacking fault
metals and alloys is well known, quantitative in situ studies relating the Fig. 1. (a) EBSD phase map and (b) inverse pole figure (IPF) map of the as-
sequence and extent of the martensitic transformation in 316L SS to the received sample for longitudinal direction (LD) and transverse direction (TD)
cross-section.
stacking fault energy during deformation at cryogenic temperatures are
lacking.
In this study, we have conducted in situ time-of-flight neutron illustrates the experimental setup. All the step-loading tests were per­
diffraction during tensile loading of 316L SS at 297, 173, 50 and 15 K, in formed at a strain rate of 10− 4 s− 1. The in situ stress rig has a load
conjunction with post-mortem microstructure analysis by transmission capability of ±100 kN. The cold chamber was built in-house and com­
electron microscopy (TEM). We have monitored and quantified elastic bined with the stress rig to give a controlled cryogenic temperature
and plastic deformation parameters, the evolution of lattice strain and environment [23]. The neutron scattering gauge volume was 1 × 1 × 1
phase volume fraction as a function of deformation. Stacking fault mm3, which was defined by the 1 × 1 mm2 incident slit and the 1 mm
probability and energy at these four temperatures have also been receiving radial collimators [24] mounted in front of the detector. The
calculated. This is correlated with the deformed microstructures. They loading axis was parallel to the longitudinal direction and oriented 45◦
provide evidence for the transition of the deformation mechanism at relative to the incident neutron beam (as shown in Fig. 2). The long
different cryogenic temperatures. In the light of these experiments, the duration of the tests (12 h per sample) precluded the testing of many
relationship between SFE, temperature and deformation mechanisms of samples. In each case 2 samples were tested, one before mounting the rig
316L SS deformed at 173, 50 and 15 K is discussed. on the neutron diffractometer and one while neutron diffraction data
was acquired. Nevertheless, the difference between the two tests in each
2. Experimental details case was small. The scattering angles of the two detector banks (longi­
tudinal and transverse detectors) relative to the incident neutron beam
A commercial type 316L stainless steel (SS) was studied and the were ±90◦ . Diffraction spectra were collected over 20 min intervals at
chemical composition (in weight percent) is listed in Table 1. It was each loading step to obtain good counting statistics for the diffraction
manufactured by hot extrusion and annealing at 1040 ◦ C followed by air patterns, until the sample failed. During loading, load-control was used
cooling to room temperature. The as-received material is completely in the elastic region and subsequently switched to strain-control after
austenitic (Fig. 1a) and the grain structure is shown in Fig. 1b. The grain yielding.
size is approximately 18 μm (excluding twin boundaries) and a few Neutron diffraction spectra were analysed using the Rietveld method
annealing twins can be observed. LD and TD represent longitudinal and peak fitting within TOPAS [25] to obtain the lattice plane spacing
(loading) direction and transverse direction, respectively. dhkl , corresponding to each (hkl) lattice plane. Lattice strain (εexp
hkl ) is the
Dog bone-shaped cylindrical samples were machined having a gauge change in interplanar distance due to the applied stress, which was
length of 42 mm and diameter of 8 mm and with M12 screw-threaded calculated by the following equation:
ends. In situ time-of-flight neutron diffraction measurements were con­ 0
dhkl − dhkl
ducted during tensile deformation at 297, 173, 50 and 15 K on the εexp
hkl = 0
(1)
dhkl
ENGIN-X diffractometer at the ISIS spallation neutron source. Fig. 2
where d0hkl refers to the strain-free interplanar spacing.
A field emission gun scanning electron microscope (TESCAN Mira3)
Table 1
equipped with an electron backscatter diffraction (EBSD) system was
Chemical composition of the 316L SS (wt.%).
used for EBSD mapping. The beam parameters for the EBSD measure­
Sample C Cr Ni Mo Mn Si Fe ments were set to an accelerating voltage of 15 kV, current of 11 nA and
316L SS 0.03 18 10 3 1 – Bal. a step size of 400 nm. Transmission electron microscopy (TEM) tests

2
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

3. Results

3.1. Mechanical properties

The engineering stress-strain curves for 316L SS deformed at 15, 50,


173 and 297 K during the in situ neutron measurement are shown in
Fig. 3a. The yield strength (YS), ultimate tensile strength (UTS) and
elongations at each temperature are summarised in Table 2. It is clear
that the YS and UTS increased significantly with decreasing tempera­
ture. The YS and UTS were enhanced to 1010 and 1260 MPa at 15 K
while the largest elongation of 42% was recorded at 173 K. The strain
hardening rate (SHR, dσtrue /dεtrue ) is plotted in Fig. 3b. The SHR drop
rapidly at the elastic-plastic stage with similar trends at all tempera­
tures. Beyond yielding, the SHR remains stable at 297 K, while at
cryogenic temperature the SHR rises to a peak before falling again with
increasing strain. At 15 and 50 K, stress decreases slightly upon yielding
and then remain almost constant until 0.1 strain whereupon strain-
hardening rate increases sharply, compared to that at 173 and 297 K.
Besides, secondary strain hardening increases with decreasing temper­
ature and occurs at lower strains. At 15 K, the peak value of SHR after
yielding was 6400 MPa at a strain of 0.15.

3.2. In situ neutron diffraction

3.2.1. Neutron spectra


Fig. 4 shows the neutron diffraction patterns collected for the lon­
gitudinal direction during tensile straining at 297, 173, 50 and 15 K. No
new phases are observed upon cooling to cryogenic temperature con­
firming that the cooling process did not induce martensitic trans­
formation for 316L SS. Unsurprisingly, the γ-phase lattice parameter
falls with decreasing temperature as summarised in Table 3. For
straining at 297 K, no new phases appeared during the deformation
(Fig. 4d). By contrast for straining at cryogenic temperatures new peaks
corresponding to the ε phase such as (1010)ε peak start to appear upon
plastic straining. At larger strains still, the α′ peaks start to appear and
Fig. 2. (a) Schematic illustration of the tensile specimen dimensions (mm) and
the ε phase and the γ phase diffraction peaks weaken. The (110)α′ peak
(b) in situ neutron diffraction measurement set-up at ENGIN-X.
nearly overlaps (111)γ peak. At the same time the diffraction peak po­
sitions of γ-phase constantly shift toward larger d-spacing.
were conducted on a FEI Tecnai TF30 operated at an acceleration
voltage of 300 kV to examine the microstructure of the samples after the
in situ neutron diffraction experiments. The TEM samples were extracted
from regions of the failed specimens near and far from the facture sur­
faces along the deformation direction. They were mechanically ground Table 2
down to ~80 μm thickness and then disks of 3-mm diameter were Yielding strength (YS), ultimate tensile strength (UTS), and elongation of the
punched out. The foils were prepared by twin-jet electropolishing with a 316L SS at all temperatures.
TenuPol-5 using an electrolyte of 5% (volume fraction) perchloric acid Temperature (K) YS (MPa) UTS (MPa) Elongation (%)
(HClO4) and 95% methanol at − 35 ◦ C and the setting voltage of 35 V.
15 1010 1260 27
The strain values of the measured TEM samples in this study were
50 990 1205 28
estimated by finite element method (FEM) simulation in ABAQUS. The 173 860 1080 42
parameters used and simulated images of the longitudinal section at 15, 297 660 750 34
50 and 173 K are described in the Appendix.

Fig. 3. Mechanical performance of the 316L SS at 297, 173, 50 and 15 K: (a) engineering stress-strain curves and (b) strain hardening rate (SHR) curves.

3
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

Fig. 4. Diffraction patterns collected using the longitudinal detector for different applied strain levels displaced vertically for clarity at (a) 15, (b) 50, (c) 173 and (d)
297 K.

(BCC) phases as a function of strain during straining at 15, 50, 173 and
Table 3
297 K is illustrated in Fig. 6a. As observed in connection with Fig. 4d, at
The lattice parameter (a0) and elastic modulus corresponding to the various
297 K no martensitic transformation occurs during straining. For
grain orientations as a function of straining temperature.
deformation at 173 K, the martensitic transformation begins after
Temperature a0 E111 E200 E220 E311 E222 generalised yielding with an almost linear decrease in austenite with
(K) (nm) (GPa) (GPa) (GPa) (GPa) (GPa)
plastic strain. It is evident that a small amount (<6%) of ε-phase forms
15 0.3566 272 160 265 205 260 before declining between strains of 0.15–0.30, while the level of α′ rises
50 0.3570 275 155 260 205 248
approximately linearly with plastic strain. At 50 K, and even more so at
173 0.3591 265 145 275 197 256
297 0.3597 229 133 235 182 218 15 K, martensite starts to form during ‘elastic’ deformation. During the
deformation-induced martensitic transformation, the load re-
partitioning will occur between different phases, and the strength of α′
3.2.2. Lattice strain evolution is greater than that of γ, which contributes to the improvement of the
The evolution of lattice strains as a function of the engineering stress strength of 316L SS. With decreasing temperature, the degree of
for different crystallographic planes along longitudinal and transverse martensitic transformation increases such that at failure there are 100%,
directions during tensile deformation at 15, 50, 173 and 297 K are 40%, 26%, 18% of γ-phase and the balance of martensitic phase when
shown in Fig. 5. Because of the appearance of new diffraction peaks that deformed at 297, 173, 50 and 15 K, respectively. The phase interfaces
overlap the original peaks (e.g., (111)γ and (110)α′) during plastic act as obstacles to block dislocation motion and the increased α′ enhance
deformation at cryogenic temperature, giving ambiguity in lattice strain the strain hardening at cryogenic temperatures [29].
determination, the lattice strains are only shown for the elastic and At cryogenic temperatures, more ε is formed with the maximum level
elasto-plastic part of the loading curve. In all cases the trends are similar occurring at progressively lower strain. The sample deformed at 15 K has
with the curves for each lattice plane family being broadly linear but up to 13% ε-phase, more than twice that of 173 K. In this context it has
having different slopes, which is due to the cubic elastic anisotropy [26]. been reported that suppressing the formation of ε-martensite during the
For the longitudinal strains, the elastic moduli corresponding to each of deformation-induced martensitic transformation can enhance the
the different grain families (Ehkl) are shown in Table 3. The elastic cryogenic-temperature tensile ductility [30]. In addition, it can be seen
moduli broadly increase with decreasing temperature, in which the that the transformation rate (dVγ/dεEng) at 173 K is more stable than that
(200) grains display the lowest elastic modulus and highest lattice strain at 15 and 50 K (Fig. 6b). The γ-austenite fraction at 173 K decreases
followed by (311), (222), (220) and (111) in accordance with theory approximately linearly and the transformation rate remains stable. It is
from previous work [27]. Before the yield point, the non-linear increase reported that the stability of the rate of martensitic transformation
of the lattice strain was also observed. For all temperatures observed in contributes to the maximum elongation [31]. Therefore, our results
this study, the lattice strain curves begin to deviate prior to generalised suggest that the greater ductility at 173 K may be attributed to the low
yielding (indicated by the dashed line), which is due to the transfer of level of ε-martensite and the stable martensitic transformation rate.
tensile loading from the harder grains (e.g., 111) to the softer grains (e.
g., 200) [28].
3.3. Microstructure characterization
3.2.3. Evolution of phase volume fractions
The variation in the volume fractions of the γ (FCC), ε (HCP) and α′ The microstructures of the specimens after in situ tensile loading
were analysed by TEM. TEM samples were extracted at different

4
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

Fig. 5. Lattice strain evolution relative to the unloaded state along the longitudinal (LD) and transverse (TD) directions with engineering stress for the major crystal
planes at (a) 15, (b) 50, (c) 173 and (d) 297 K.

longitudinal locations, each representing different strain levels, as faults are distributed parallel to the ε-martensite. The lath ε-martensite
inferred by FEM analysis (see appendix). Bright field (BF) images and in this case exhibits an orientation relationship close to the Shoji-
selected area diffraction patterns (SADP) at 15, 50 and 173K are shown Nishiyama (S–N) relationship of (111)γ//(0001)ε and [110]γ// [1120]ε
in Fig. 7. to the γ-austenite matrix [36]. Furthermore, unlike the microstructure at
At 0.025 strain for the sample deformed at 15 K (Fig. 7a), disloca­ 15 K, twins can be seen distributed parallel to each other between the
tions are densely trapped inside the γ-austenite. The nano-sized blocky ε-martensite separating the grain into small regions (see inset). After
regions comprise α′-martensite as confirmed by selected area diffraction. 0.37 engineering strain (Fig. 7f), there is no evidence of ε-martensite
This indicates that γ → α′ martensitic transformation occurs during the phase. Instead α′-martensite having a lamella structure transformed
deformation at 15 K. In Fig. 7a, the γ-austenite and newly formed from γ-austenite during deformation and deformation twins are
α′-martensite follow the established Pitsch orientation relationship, observed. Meanwhile, a high-density of dislocations are observed, which
namely, (100)γ//(110)α′ with the zone axis [110]γ//[111]α′. Similar are trapped within, and intersect the twins. The orientation relationship
crystallographic orientation relationships between γ-austenite and between γ-austenite and α′-martensite shown here is the
α′-martensite have also been observed in other stainless steels and alloys Nishiyama-Wassermann (N–W) relationship of (111)γ//(110)α′ and
subjected to various plastic strains [3,22,32–34]. At 0.20 engineering [110]γ//[100]α′ [37].
strain, parallel ε-laths can be observed clearly from the images, sand­ Fig. 8 shows TEM images of stacking faults at low strain for the three
wiched between γ-austenite and α′-martensite, as indicated by red ar­ samples deformed at cryogenic temperatures. At 15 K, stacking faults
rows in Fig. 7b. This observation is closely related to the transformation can be observed embedded between ε-martensite laths near the dis­
mechanism following the sequence γ → ε → α′. This confirms that some located areas. This corroborates the fact that deformation mechanism of
of the α′ phase is formed by the transformation from already-formed martensitic transformation at cryogenic temperature, involving the
ε-phase. It is notable that γ-austenite is only present in a small number creation of ε-martensite embryos and new sites of α′. The α′-martensite
of dislocated blocky areas with α′-martensite predominating at this nucleation sites are largely confined to the intersections of shear bands
higher strain level. consisting of stacking faults bundles and ε-martensite. At 50 K (Fig. 8b)
For the sample deformed at 50 K (Fig. 7c), a high-density of dislo­ stacking faults similar to those at 15 K are observed. For low strain
cations that are trapped inside the γ-austenite can be found in sample deformation at 173 K (Fig. 8c), numerous overlapping stacking faults are
after strained to 0.025, similar to the microstructure shown in 15 K observed along with parallel ε-martensite. Interestingly, it is likely that
sample. Thin parallel strips of ε-martensite are distributed in the the overlapping stacking faults represent the broad faces of strongly
γ-austenite, and these appear to grow towards the insides of the grains. inclined ε-martensite laths. That is, many of these stacking faults may be
At 0.21 engineering strain (Fig. 7d), only a small amount of γ-austenite embryonic or nascent ε-martensite laths. In addition, the average
remains as the martensitic transformation progresses. The γ-austenite stacking fault widths (i.e., the distance between the Shockley partial
and α′-martensite shown here follows the Kurdjumov–Sachs (K–S), i.e. dislocations) are ~14.6 ± 2.7 nm at 15 K, which are ~10.8 ± 2.1 nm at
(111)γ//(110)α′ and [110]γ//[111]α′ orientation relationship [35]. At 50 K and ~4.5 ± 1.6 nm 173 K.
neither 15 nor 50 K, do we find clear evidence of deformation twins by
TEM.
At 0.025 strains for the 173 K sample (Fig. 7e) there is unambiguous
evidence for the coexistence of ε-martensite, twins, dislocations and a
multitude of stacking faults. The parallel ε-martensite plates (red ar­
rows) obstruct partial dislocation glide. A small number of stacking

5
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

temperatures can be obtained from Eq. (2). The estimates of SFP at 15,
50, 173 and 297 K are plotted in Fig. 9a. It is clear that at all tempera­
tures the SFP increases with straining during the plastic deformation
although the number of stacking faults formed at 297 K is insignificant.
The uncertainty in the extremely small strain results in fluctuation of
SFP in elastic stage. The value of SFP reached 0.041, 0.027, 0.019 and
0.0052 for deformation at 15, 50, 173 and 297 K respectively. In this
respect, the cooling-to-cryogenic temperatures can strongly affect the
evolution of dislocations during deformation hence the capacity for the
formation of stacking faults. As the temperature falls, the partial dislo­
cations increasingly separate such that the extended dislocation be­
comes harder to move, leading to the formation of stacking faults [40].
The stacking fault energy (SFE), γ SFE , controls the ease of dissociation
into partial dislocations and the tendency to form of stacking faults. It
plays an important role in determining the stability of austenite because
it controls the formation of nucleation sites for martensite, which can be
estimated by Reed and Schramm’s relationship [41]:
( )− 0.37 2 ( )
6.6a0 2C44 〈ε50 〉111 C44 + C11 − C12
γ SFE = √̅̅̅ (3)
3π C11 − C12 SFP 3

where 〈ε250 〉111 is the mean square strain calculated from the integral
breadth of pseudo-voigt diffraction peak fit [42], a0 is the lattice
parameter, C11, C12 and C44 are the single crystal elastic constants,
which varies only marginally among 15, 50, 173 and 297 K according to
Ref. [43]. The C11, C12 and C44 were found to be 204.6, 137.3 and 126.2
GPa for 316L SS [44].
The evolution of the inferred SFEs is plotted as a function of engi­
neering strain at different temperatures obtained by Eq. (3) in Fig. 9b.
All four SFE curves display similar trends in that they fall sharply at the
beginning before stabilising to a constant value with increased straining.
At low strains, it has been shown previously that inferring SFE from the
diffraction peaks is unreliable because they are not sufficient to signif­
icantly affect the diffraction peaks and small strain at elastic deforma­
tion were used [45]. The attainment of stable values with further
straining suggest that the stacking faults become sufficient to provide a
reliable estimate of the SFE. This suggests that a good estimate of the SFE
at 15, 50, 173 and 297 K are 4.9, 7.8, 17.7 and 28.2 mJ/m2, respectively.
As the SFE falls with decreasing temperature, it leads to a larger and
larger separation of the partial dislocations thereby hindering disloca­
tion motion and cross-slip (numerous dislocations can be preserved).
The present SFE calculation values are reasonable when compared
with the values reported in literatures [17,46,47]. For example, Olson
Fig. 6. (a) Evolution of the γ-FCC (squares), ε-HCP (hexagons) and α′-BCC et al. [46] obtained the SFE variation with temperature finding it to drop
(diamonds) phases with engineering strain at 15 K (black), 50 K (red), 173 K from 30 to 9 mJ/m2 with temperature from 300 to 100 K for a similar
(blue) and 297 K (green) and (b) rate of the martensitic transformation as a FeNiCr alloys. Walter et al. [48] measured the SFE at room temperature
function of engineering strain. (For interpretation of the references to colour in
to be about ~25.9 mJ/m2 for 316 SS. Vitos et al. [49] calculated that the
this figure legend, the reader is referred to the Web version of this article.)
SFE of austenitic stainless steel rises from 3 to 29 mJ/m2 from 50 to 300
K, using an axial interaction model. Since lower SFE materials show
4. Discussion wider stacking faults and are more difficult to cross-slip [50], with
decreasing SFE, stacking faults tend to overlap on every second (111)
4.1. Stacking fault energy planes, reducing the overall energy of these stacking fault bundles to
form ε-martensite. This can explain why more ε-martensite is formed
Stacking faults are known to affect the shifts in diffraction peak po­ from γ-austenite with decreasing temperature (Fig. 6). Furthermore,
sition for different orders of the same reflection (e.g. 111 and 222) according to Refs [11,17,22,51], as the SFE decreases, twinning and/or
differently [38]. The observed diffraction peak shifts have two main phase transformation become the preferred deformation mechanism.
contributions: that due to the elastic stress field (εstrain hkl ) and that due to The deformation mechanism corresponding to the SFE Fig. 9b calculated
the SFs (εSF hkl ), where in this study is consistent with the observed deformation microstructure
√̅̅̅ ∑ in TEM. The samples deformed at 15 and 50 K are inhibited from
3 ±(h + k + l)
εexp
hkl = εstrain
hkl − ε SF
hkl = εstrain
hkl − ( ) SFP (2) forming deformation twins and more prone to phase transformation due
4π (u + b) h2 + k2 + l2
to the extremely low SFE, while sample deformed at 173 K can generate
both deformation twins and phase transformation (Fig. 7e). Conse­
and SFP is stacking fault probability and the b and u are the numbers of
quently, the deformation mechanisms change as a result of the
broadened and non-broadened components due to stacking faults [39].
decreasing SFE. These results clearly demonstrate that different trans­
Considering the (111) and (222) planes, the evolution of the SFP as a
formation mechanisms operate at different cryogenic temperatures and
function of engineering strain under tensile deformation at different
the operating deformation mechanisms depend strongly on the SFE for

6
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

Fig. 7. TEM micrographs of 316L SS deformed at 15K


to (a) 0.025 and (b) 0.20 engineering strain, at 50 K
to (c) 0.025 and (d) 0.21 engineering strain and at
173 K to (e) 0.025 and (f) 0.37 engineering strain,
with the corresponding selected area diffraction pat­
terns (SADP). The inset in (e) shows twins. The lo­
cations at which the SADP were acquired are
indicated in the corresponding micrographs by red
circles while the red arrows indicate ε-martensite.
(For interpretation of the references to colour in this
figure legend, the reader is referred to the Web
version of this article.)

316L SS. pronounced secondary strain hardening rate, reaching 6400, 5000, and
2800 MPa at 15, 50, and 173 K, respectively (Fig. 3b). The α′-martensite
exhibits superior tensile strength and work hardening when contrasted
4.2. Deformation mechanisms with the initial γ-austenite, highlighting the significance of phase
transformation in augmenting the mechanical strength of austenitic
The mechanical behaviours and deformation mechanisms of 316L SS steels.
are closely related to the characteristic of the microstructure, i.e. dis­ Finally, with respect to the above-mentioned improvement in
locations density, formation of deformation-induced martensite and strength, the ductility of these cryogenic temperature deformed samples
twins. Firstly, upon decreasing the deformation temperature to 173 K has not been sacrificed under the normal ‘trade-off’ between strength
and 15/50 K, the yield stresses have increased by ~30% and 50% and ductility. In fact, the total strain for the sample deformed at 173 K
respectively compared with the yield stress at 297 K. This can be reached 0.42, which is the highest of all the samples being ~24% higher
attributed to the extra energy required for activating dislocations at than sample deformed at 297 K. The reason behind this outstanding
cryogenic temperatures. As the deformation temperature drops, the ductility is the formation of deformation twins in the γ-austenite.
energy required for dislocation to overcome the Peierls lattice potential Comparing to deformation at 15 and 50 K, twins are thermodynamically
friction and the pinning effect of atoms increases, which resulting in favoured such that for sample deformed at 173 K, its SFE lies in the
higher stress to initiate plastic deformation [52]. region where both the martensitic transformation and deformation
Secondly, the ultimate strength also improved significantly for twins can be activated. In contrast to the strengthening effect provided
samples deformed at cryogenic temperatures due to the strengthening by the newly formed α′-martensite, the deformation twins contribute to
effect of newly formed martensite via TRIP effect. Under cryogenic both improved ductility and strength via the TWIP effect such that the
temperatures, the metastable γ-austenite in the 316L SS undergoes a increased strength does not see a decrease in ductility [53]. Besides, a
martensitic transformation with strain following either of the two rather stable phase transformation rate, as well as the lower level of
pathways (γ → ε → α′ or γ → α′). The volume fraction of deformation- ε-martensite (~4.5% volume fraction) are believed to associate with the
induced α′-martensite reached 78%, 71%, and 58% (Fig. 6a) at the maximum elongation found in 173 K deformed sample. Similar phe­
failure of 15, 50, and 173 K deformed sample, respectively, with cor­ nomenon of complexed microstructure enhancing ductility has been
responding ultimate tensile strength reaches 1260, 1205 and 1010 MPa reported in various TRIP steels [32,33].
(Fig. 3a). In contrast, sample deformed at 297 K that did not undergo
martensitic phase transformation had an ultimate tensile strength of
780 MPa. Furthermore, we can see that the cryogenic-deformation
samples with more α′-martensite at similar strain exhibited a more

7
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

Fig. 8. Higher magnification bright field images recorded on the [110] zone axis showing stacking faults (SF) of (a) 15, (b) 50 and (c) 173 K each deformed to an
engineering strain of 0.025.

Fig. 9. (a) Stacking fault probability curve and (b) stacking fault energy curve with strain at 15, 50, 173 and 297 K for 316L SS.

5. Conclusions MPa at 15 K, respectively. The maximum elongation (42%) was


achieved at 173 K. Significant secondary hardening is observed after
In situ neutron diffraction measurements during tensile testing have yield at cryogenic temperature.
been conducted to investigate the martensitic transformations and their 2. No martensitic transformation was observed for deformation at 297
effect on the tensile loading behaviour of 316L SS across a range of K. Martensitic transformation occurs during the cryogenic defor­
temperatures from 293 K to 15 K. These changes and effects on micro­ mation with the ε-martensite acting as a transient phase. As tem­
structures are correlated with TEM observations. Based on the analysis perature drops, the degree of martensitic transformation increases as
of the experimental results, the following conclusions can be drawn from does the peak level of ε.
present work. 3. From TEM analysis, both the γ → ε → α′ and γ → α′ martensitic
transformation can be observed at 173, 50 and 15 K. Many additional
1. An increase in the tensile strength was observed as the deformation twins are formed only at 173 K. The K–S, N–W and Pitsch orientation
temperature decreases. Yield strength and ultimate tensile strength relationships are demonstrated between γ-austenite and
rose from 660 and 750 MPa at room temperature to 1010 and 1260

8
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

α′-martensite, while the relationship between γ-austenite and Conceptualization, Methodology, Formal analysis, Writing – review &
ε-martensite is close to the S–N orientation relationship. editing.
4. The stacking fault probability of 316L SS increases with decreasing
temperature while the stacking fault energy is estimated to be 28.2, Declaration of competing interest
17.7, 7.8 and 4.9 mJ/m2 at 297, 173, 50 and 15 K, respectively.
5. Taken together, SFE decreases as the temperature falls, leading to the The authors declare that they have no known competing financial
variation in deformation mechanisms, which in turn results in the interests or personal relationships that could have appeared to influence
improved mechanical behaviour at cryogenic temperatures. At 15 the work reported in this paper.
and 50 K, martensitic transformation dominates the deformation
mechanism, corresponding to extremely low SFE, while at 173 K, the Data availability
deformation mechanism combines both martensitic transformation
and twins. The deformation-induced α′-martensite greatly contrib­ Data will be made available on request.
utes to the enhanced strength, while the higher ductility at 173 K is
probably related to the formation of twins, low level of ε-martensite Acknowledgement
and stable rate of martensitic transformation.
This work was also supported by the Henry Royce Institute for
CRediT authorship contribution statement Advanced Materials, funded through EPSRC grants EP/R00661X/1, EP/
S019367/1, EP/P025021/1, and EP/P025498/1. The neutron beamtime
Suning Li: Methodology, Investigation, Data curation, Software, at ENGIN-X of ISIS neutron and muon source (the Rutherford Appleton
Formal analysis, Visualization, Writing – original draft, Writing – review Laboratory, UK) under experiment RB1810732 is acknowledged. S. Li
& editing. Philip J. Withers: Writing – review & editing. Saurabh would like to express his gratitude for the support of China Scholarship
Kabra: Investigation, Writing – review & editing. Kun Yan: Supervision, Council (CSC).

Appendix

The strain distribution inside the tensile specimen along the loading direction were estimated by finite element method (FEM) simulation in
ABAQUS based on the results of tensile test. This was used to determine the strain values corresponding to the analysed TEM sample areas, as shown in
Fig. A.1. The dimensions of the virtual specimen were the same as the experimental ones. The number of elements was 105732 for tensile specimens. A
Poisson’s ratio of 0.3, Young’s modulus of 210000 and density of 7.87 g/cm3 were used as material properties for elasticity, while an isotropic strain
hardening model was used for plasticity.

Fig. A1. Simulated images of the longitudinal section of the (a) 15, (b) 50 and (c) 173 K deformed samples using Abaqus: strain values along the loading direction
corresponding to the measured TEM samples in this study are marked on the right.

References transformation in 301 stainless steel, Mater. Sci. Technol. 34 (17) (2018)
2114–2125.
[8] D.T. Pierce, J.A. Jiménez, J. Bentley, D. Raabe, J.E. Wittig, The influence of
[1] M. Calmunger, G. Chai, R. Eriksson, S. Johansson, J.J. Moverare, Characterization
stacking fault energy on the microstructural and strain-hardening evolution of
of austenitic stainless steels deformed at elevated temperature, Metall. Mater.
Fe–Mn–Al–Si steels during tensile deformation, Acta Mater. 100 (2015) 178–190.
Trans. 48 (10) (2017) 4525–4538.
[9] J.T. Benzing, W.A. Poling, D.T. Pierce, J. Bentley, K.O. Findley, D. Raabe, J.
[2] D.J. Wang, Y.H. Wang, L.G. Liu, W.J. Liu, X.Q. Zhao, Q.F. Wang, Phase
E. Wittig, Effects of strain rate on mechanical properties and deformation behavior
Transformation Behavior and Microstructure Characterization of 9Ni Cryogenic
of an austenitic Fe-25Mn-3Al-3Si TWIP-TRIP steel, Mater. Sci. Eng., A 711 (2018)
Steel, Advanced Materials Research, Trans Tech Publ, 2012, pp. 1183–1187.
78–92.
[3] K. Nalepka, B. Skoczeń, M. Ciepielowska, R. Schmidt, J. Tabin, E. Schmidt,
[10] R. Schramm, R. Reed, Stacking fault energies of seven commercial austenitic
W. Zwolińska-Faryj, R. Chulist, Phase transformation in 316L austenitic steel
stainless steels, Metall. Trans. A 6 (7) (1975) 1345–1351.
induced by fracture at cryogenic temperatures: experiment and modelling,
[11] E. Galindo-Nava, P. Rivera-Díaz-del-Castillo, Understanding martensite and twin
Materials 14 (1) (2020) 127.
formation in austenitic steels: a model describing TRIP and TWIP effects, Acta
[4] F. Yan, G. Liu, N. Tao, K. Lu, Strength and ductility of 316L austenitic stainless steel
Mater. 128 (2017) 120–134.
strengthened by nano-scale twin bundles, Acta Mater. 60 (3) (2012) 1059–1071.
[12] J. Lu, L. Hultman, E. Holmström, K.H. Antonsson, M. Grehk, W. Li, L. Vitos,
[5] S. Hecker, M. Stout, K. Staudhammer, J. Smith, Effects of strain state and strain rate
A. Golpayegani, Stacking fault energies in austenitic stainless steels, Acta Mater.
on deformation-induced transformation in 304 stainless steel: Part I. Magnetic
111 (2016) 39–46.
measurements and mechanical behavior, Metall. Trans. A 13 (4) (1982) 619–626.
[13] O. Grässel, L. Krüger, G. Frommeyer, L. Meyer, High strength Fe–Mn–(Al, Si) TRIP/
[6] N. Hashimoto, T.S. Byun, Deformation-induced martensite formation and
TWIP steels development—properties—application, Int. J. Plast. 16 (10–11) (2000)
dislocation channeling in neutron-irradiated 316 stainless steel, J. Nucl. Mater. 367
1391–1409.
(2007) 960–965.
[7] Y.B. Das, A.N. Forsey, J. Kelleher, S. Kabra, M.E. Fitzpatrick, T.H. Simm, S. Gungor,
R.J. Moat, The influence of temperature on deformation-induced martensitic

9
S. Li et al. Materials Science & Engineering A 880 (2023) 145279

[14] G.M. De Bellefon, J. Van Duysen, K. Sridharan, Composition-dependence of [34] Y. Jo, W. Choi, D. Kim, A. Zargaran, S. Sohn, H. Kim, B. Lee, N. Kim, S. Lee, FCC to
stacking fault energy in austenitic stainless steels through linear regression with BCC transformation-induced plasticity based on thermodynamic phase stability in
random intercepts, J. Nucl. Mater. 492 (2017) 227–230. novel V10Cr10Fe45CoxNi35− x medium-entropy alloys, Sci. Rep. 9 (1) (2019)
[15] D.T. Pierce, J.A. Jiménez, J. Bentley, D. Raabe, C. Oskay, J. Wittig, The influence of 1–14.
manganese content on the stacking fault and austenite/ε-martensite interfacial [35] G. Kurdjumow, G. Sachs, Über den mechanismus der stahlhärtung, Z. Phys. 64 (5)
energies in Fe–Mn–(Al–Si) steels investigated by experiment and theory, Acta (1930) 325–343.
Mater. 68 (2014) 238–253. [36] Z. Nishiyama, M.E. Fine, M. Meshii, C.M. Wayman, in: Morris E. Fine, M. Meshii, C.
[16] D. Molnár, X. Sun, S. Lu, W. Li, G. Engberg, L. Vitos, Effect of temperature on the M. Wayman (Eds.), Martensitic Transformation/Zenji Nishiyama, Academic Press,
stacking fault energy and deformation behaviour in 316L austenitic stainless steel, New York, 1978.
Mater. Sci. Eng., A 759 (2019) 490–497. [37] Z. Nishiyama, X-ray investigation of the mechanism of the transformation from
[17] S. Curtze, V.-T. Kuokkala, Dependence of tensile deformation behavior of TWIP face centered cubic lattice to body centered cubic, Sci. Rep. Tohoku Univ. 23
steels on stacking fault energy, temperature and strain rate, Acta Mater. 58 (15) (1934) 637.
(2010) 5129–5141. [38] L. Tang, L. Wang, M. Wang, H. Liu, S. Kabra, Y. Chiu, B. Cai, Synergistic
[18] M. Linderov, C. Segel, A. Weidner, H. Biermann, A. Vinogradov, Deformation deformation pathways in a TWIP steel at cryogenic temperatures: in situ neutron
mechanisms in austenitic TRIP/TWIP steels at room and elevated temperature diffraction, Acta Mater. 200 (2020) 943–958.
investigated by acoustic emission and scanning electron microscopy, Mater. Sci. [39] B.E. Warren, X-Ray Diffraction, Courier Corporation, 1990.
Eng., A 597 (2014) 183–193. [40] D. Hull, D.J. Bacon, Chapter 5 - dislocations in face-centered cubic metals, in:
[19] A. Saeed-Akbari, J. Imlau, U. Prahl, W. Bleck, Derivation and variation in D. Hull, D.J. Bacon (Eds.), Introduction to Dislocations, fifth ed., Butterworth-
composition-dependent stacking fault energy maps based on subregular solution Heinemann, Oxford, 2011, pp. 85–107.
model in high-manganese steels, Metall. Mater. Trans. 40 (13) (2009) 3076–3090. [41] R. Reed, R. Schramm, Relationship between stacking-fault energy and x-ray
[20] S. Allain, J.-P. Chateau, O. Bouaziz, A physical model of the twinning-induced measurements of stacking-fault probability and microstrain, J. Appl. Phys. 45 (11)
plasticity effect in a high manganese austenitic steel, Mater. Sci. Eng., A 387 (2004) (1974) 4705–4711.
143–147. [42] S. Harjo, Y. Tomota, P. Lukáš, D. Neov, M. Vrana, P. Mikula, M. Ono, In situ
[21] M. Olsson, Thermodynamic Modeling of the Stacking Fault Energy in Austenitic neutron diffraction study of α–γ Fe–Cr–Ni alloys under tensile deformation, Acta
Stainless Steels, 2014. Mater. 49 (13) (2001) 2471–2479.
[22] Y.F. Shen, X.X. Li, X. Sun, Y.D. Wang, L. Zuo, Twinning and martensite in a 304 [43] H. Ledbetter, S. Kim, Handbook of Elastic Properties of Solids, Liquids, and Gases,
austenitic stainless steel, Mater. Sci. Eng., A 552 (2012) 514–522. vol. 2, Academic Science, New York, 2001.
[23] O. Kirichek, J. Timms, J. Kelleher, R. Down, C. Offer, S. Kabra, S. Zhang, Sample [44] B. Clausen, T. Lorentzen, T. Leffers, Self-consistent modelling of the plastic
environment for neutron scattering measurements of internal stresses in deformation of fcc polycrystals and its implications for diffraction measurements of
engineering materials in the temperature range of 6 K to 300 K, Rev. Sci. Instrum. internal stresses, Acta Mater. 46 (9) (1998) 3087–3098.
88 (2) (2017), 025103. [45] L. Tang, K. Yan, B. Cai, Y. Wang, B. Liu, S. Kabra, M.M. Attallah, Y. Liu,
[24] P.J. Withers, M.W. Johnson, J.S. Wright, Neutron strain scanning using a radially Deformation mechanisms of FeCoCrNiMo0. 2 high entropy alloy at 77 and 15 K,
collimated diffracted beam, Phys. B Condens. Matter 292 (3) (2000) 273–285. Scripta Mater. 178 (2020) 166–170.
[25] K. Yan, K.-D. Liss, I.B. Timokhina, E.V. Pereloma, In situ synchrotron X-ray [46] G.B. Olson, M. Cohen, A general mechanism of martensitic nucleation: Part I.
diffraction studies of the effect of microstructure on tensile behavior and retained General concepts and the FCC→ HCP transformation, Metall. Trans. A 7 (12)
austenite stability of thermo-mechanically processed transformation induced (1976) 1897–1904.
plasticity steel, Mater. Sci. Eng., A 662 (2016) 185–197. [47] G.G. Ribamar, T.C. Andrade, H.C. De Miranda, H.F.G. de Abreu, Thermodynamic
[26] M.T. Hutchings, P.J. Withers, T.M. Holden, T. Lorentzen, Introduction to the stacking fault energy, chemical composition, and microstructure relationship in
Characterization of Residual Stress by Neutron Diffraction, CRC press, 2005. high-manganese steels, Metall. Mater. Trans. 51 (9) (2020) 4812–4825.
[27] B. Clausen, T. Lorentzen, M.A. Bourke, M.R. Daymond, Lattice strain evolution [48] M. Walter, L. Mujica Roncery, S. Weber, L. Leich, W. Theisen, XRD measurement of
during uniaxial tensile loading of stainless steel, Mater. Sci. Eng., A 259 (1) (1999) stacking fault energy of Cr–Ni austenitic steels: influence of temperature and
17–24. alloying elements, J. Mater. Sci. 55 (27) (2020) 13424–13437.
[28] Y. Wang, B. Liu, K. Yan, M. Wang, S. Kabra, Y.-L. Chiu, D. Dye, P.D. Lee, Y. Liu, [49] L. Vitos, J.-O. Nilsson, B. Johansson, Alloying effects on the stacking fault energy in
B. Cai, Probing deformation mechanisms of a FeCoCrNi high-entropy alloy at 293 austenitic stainless steels from first-principles theory, Acta Mater. 54 (14) (2006)
and 77 K using in situ neutron diffraction, Acta Mater. 154 (2018) 79–89. 3821–3826.
[29] N. Koga, T. Nameki, O. Umezawa, V. Tschan, K.-P. Weiss, Tensile properties and [50] P. Thornton, T. Mitchell, P. Hirsch, The dependence of cross-slip on stacking-fault
deformation behavior of ferrite and austenite duplex stainless steel at cryogenic energy in face-centred cubic metals and alloys, Phil. Mag. 7 (80) (1962)
temperatures, Mater. Sci. Eng., A 801 (2021), 140442. 1349–1369.
[30] M. Koyama, T. Sawaguchi, K. Tsuzaki, TWIP effect and plastic instability condition [51] D. Molnar, X. Sun, S. Lu, W. Li, G. Engberg, L. Vitos, Effect of temperature on the
in an Fe–Mn–C austenitic steel, ISIJ Int. 53 (2) (2013) 323–329. stacking fault energy and deformation behaviour in 316L austenitic stainless steel,
[31] N. Tsuchida, Y. Morimoto, T. Tonan, Y. Shibata, K. Fukaura, R. Ueji, Stress-induced Mater. Sci. Eng., A 759 (2019) 490–497.
martensitic transformation behaviors at various temperatures and their TRIP [52] I.-C. Jung, B.C. De Cooman, Temperature dependence of the flow stress of
effects in SUS304 metastable austenitic stainless steel, ISIJ Int. 51 (1) (2011) Fe–18Mn–0.6 C–xAl twinning-induced plasticity steel, Acta Mater. 61 (18) (2013)
124–129. 6724–6735.
[32] K. Verbeken, L. Barbé, D. Raabe, Evaluation of the crystallographic orientation [53] Y. Wei, Y. Li, L. Zhu, Y. Liu, X. Lei, G. Wang, Y. Wu, Z. Mi, J. Liu, H. Wang, H. Gao,
relationships between FCC and BCC phases in TRIP steels, ISIJ Int. 49 (10) (2009) Evading the strength–ductility trade-off dilemma in steel through gradient
1601–1609. hierarchical nanotwins, Nat. Commun. 5 (1) (2014) 3580.
[33] D. Borisova, V. Klemm, S. Martin, S. Wolf, D. Rafaja, Microstructure defects
contributing to the energy absorption in Cr M n N i TRIP steels, Adv. Eng. Mater. 15
(7) (2013) 571–582.

10

You might also like