Materials Science & Engineering A 743 (2019) 175–184

Contents lists available at ScienceDirect

Materials Science & Engineering A

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

Into the quenching and; partitioning of a 0.2C steel: An in-situ synchrotron T

study
⁎
Pierre Huyghea, Matteo Carusob, Jean-Louis Colletb, Sylvain Dépinoya, Stéphane Godeta,
a
4MAT, Materials Engineering, Characterization, Processing and Recycling, Université Libre de Bruxelles, 50 Avenue FD Roosevelt, CP194/03, B-1050 Brussels, Belgium
b
CRM Group, Avenue du Bois Saint Jean 21 P59, B-4000 Liège, Belgium

A R T I C LE I N FO A B S T R A C T

Keywords: Quenching and partitioning (Q&P) is an effective way for retaining austenite at room temperature. The actual
Quenching and partitioning mechanisms responsible for austenite stabilization are still under debate, since it is impossible to track carbon
Synchrotron diffusion with the conventional metallographic tools. The present work depicts several Q&P heat treatments
HEXRD performed on a 0.2 C commercial grade steel by in-situ High Energy X-Ray Diffraction (HEXRD). More speci-
Retained austenite
fically, the effect of three different initial quenching temperatures on the microstructural evolution occurring
Martensite
Bainite
during Q&P was scrutinized in details.
It was shown that about 50% of the initial carbon partitions effectively to austenite. A carbon enrichment up
to 0.8 wt% is sufficient to retain austenite at room temperature as grain size refinement contributes to further
stabilize austenite. The origin of carbon enrichment in retained austenite depends on the initial quench tem-
perature (QT). For low QT, corresponding to an initial martensite fraction larger than 0.75, austenite carbon
enrichment is ensured by carbon partitioning from supersaturated martensite. For higher QT, austenite carbon
enrichment results from both carbon partitioning from martensite and carbon rejected during the bainite
transformation. The former mechanism proceeds rapidly and already starts during the reheating stage to the
partitioning temperature. The latter is slower as the carbon enrichment is coupled to bainite formation.

1. Introduction years [2,12–21]. Hence, conventional characterization techniques such

as dilatometry, X-Ray diffraction (XRD), Scanning Electron Microscopy
The Quenching and Partitioning (Q&P) heat treatment has been (SEM) and Transmission Electron Microscopy (TEM) observations have
proposed in the early 2000s to produce cold-rolled sheets combining been widely applied to study the microstructural evolution during the Q
high-strength and formability for the automotive market. Q&P consists &P treatment.
of an interrupted quench between the martensite-start temperature (Ms) Our previous work [22], based on a combination of these post-
and the martensite-finish temperature (Mf) followed by a partitioning mortem conventional characterization techniques, showed that the
step in order to stabilize the untransformed austenite through carbon final Q&P microstructures typically consist in a mix of tempered mar-
enrichment. The transformation of retained austenite upon straining at tensite, bainite, fresh martensite and retained austenite. The relative
room temperature (TRIP effect) provides supplementary work-hard- amounts of these phases depend on the choice of QT, and has a big
ening and eventually improves the ductility [1–3]. The TRIP effect is impact on the mechanical properties and on the stretch-flange form-
dictated by the amount and the stability of the retained austenite, ability of Q&P steels [22,23]. More specifically, the thermal and me-
which depends in turn on its carbon content, size and surrounding chanical stability of retained austenite is of paramount importance.
phases [4–10]. In order to maximize carbon partitioning from mar- Retained austenite should be enough stable to avoid the formation of
tensite to austenite, the occurrence of competing phenomena as carbide fresh martensite during the final quench, as it drastically impacts the
precipitation and austenite decomposition have to be minimized. This is ductility and the local formability of the steel sheet. The mechanical
achieved by adding alloying elements delaying carbide precipitation stability should also be carefully tuned to ensure an effective TRIP-ef-
such as Si and Al and by optimizing the Q&P thermal cycle [11,12]. fect and actually improve the work-hardening rate. Understanding the
The Q&P heat treatment for the production of TRIP-assisted steel mechanisms responsible for the stabilization of the retained austenite is
proposed by Speer et al. has been extensively investigated in the past therefore of primary interest to design Q&P grades with superior

⁎
Corresponding author.
E-mail address: [email protected] (P. Huyghe).

https://doi.org/10.1016/j.msea.2018.11.065
Received 12 September 2018; Received in revised form 12 November 2018; Accepted 13 November 2018
Available online 16 November 2018
0921-5093/ © 2018 Elsevier B.V. All rights reserved.
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

mechanical properties. conventional techniques (SEM, EBSD and XRD). Several phases such as
However, metallurgical phenomena involving carbon as diffusion initial martensite (i.e. formed during the initial quench to QT), bainite,
and partitioning or carbides precipitation are unfortunately difficult to fresh martensite (i.e. formed during the final quench to RT) and re-
probe with conventional techniques as dilatometry or conventional tained austenite were identified and quantified. Their respective vo-
XRD. The kinetics of both martensite transformation and subsequent lume fractions are given in Table 2. The fractions of initial martensite
carbon partitioning being in the order of a few seconds, very fast ac- formed at a temperature between Ms and Mf were determined by using
quisition rates are required. Hence, previous studies have demonstrated the lever rule applied to the martensitic transformation profile on the
the great interest of using in-situ techniques to obtain time-resolved dilatometry curve. The dilatation observed during the subsequent par-
quantitative information during phase transformation. An elegant ap- titioning step was attributed to the formation of bainite, considering the
proach to investigate the partitioning stage in a 1C-3Mn-1.5Si grade temperature at which this transformation occurs (i.e. 400 °C) [16]. The
was proposed by De Knijf et al. [24]. The Mf temperature of this grade fraction of bainite formed during partitioning was shown to be pro-
was well below room temperature, allowing therefore an in-situ ana- portional to the amount of untransformed austenite present at the in-
lysis by TEM. The first relevant quantitative information on the mi- itial quench temperature. Some dilatation could be observed in some
crostructural evolution during a quench and partitioning in a low- cases during the final quench to room temperature. It corresponds to
carbon steel grade was reported by Epp et al. [25] and very recently by the formation of fresh untempered martensite from the insufficiently
Allain et al. [26,27]. These studies show that High-Energy X-Ray Dif- stable austenite. The volume fraction of fresh martensite was quantified
fraction (HEXRD) is the appropriate technique to investigate the me- using the average grain image quality in EBSD [28,29]. Finally, the
tallurgical phenomena occurring during a Q&P treatment, particularly volume fractions of retained austenite were measured by XRD at room
carbon partitioning. However, most of these works focus on high al- temperature. For more details regarding the developed phase quanti-
loyed grades, specifically designed to highlight the features of carbon fication methodology, please refer to our previous work [22].
partitioning from martensite to austenite. The in-situ experiments were carried out on the high resolution
The present work is aimed at investigating the microstructural ID11 beam line of the European Synchrotron Radiation Facility (ESRF
evolution and the mechanisms responsible for the austenite stabiliza- in Grenoble, France). The high energy monochromatic beam
tion during an industrial Q&P cycle in a 0.2 C commercial grade. Gaining (E = 65 keV, λ = 0.19 Å) permitted to work in transmission diffraction
information on the precise carbon distribution amongst the different mode. Hence, the synchrotron X-Ray beam probes all the grains
phases is of major interest to understand the metallurgical mechanisms through the thickness. High resolution of the patterns was obtained
involved in the Q&P process. More specifically, in-situ HEXRD coupled using a beam size of 0.2 × 0.2 mm2. A total volume of 0.032 mm3 was
with very high acquisition rates allows gaining further insight into the analyzed and is considered as statistically relevant knowing that the
evolution of austenite throughout the entire Q&P cycle: (i) its partial prior austenite grain size is about 15 µm. The high acquisition rate (i.e.
transformation to martensite during the initial quench, (ii) its behavior 5 Hz) allowed by the FReLoN CCD camera allowed a precise follow-up
from the initial quench temperature to the partitioning step, (iii) its of the microstructure throughout the Q&P treatment. The 2D-diffrac-
carbon enrichment and carbon homogenization during the partitioning tion rings were reduced to the 1D profile (intensity - 2θ) using the Fit2D
step as well as (iv) its stability during the final quench to room tem- program developed at ESRF [30]. The volume fraction of retained
perature. In this work, a particular attention is devoted to the effect of austenite and the lattice parameters were extracted from the integrated
three initial quench temperatures QT (corresponding to different frac- intensities and the scattering angles of three face-centered cubic peaks
tions of martensite formed during the initial quench) on partitioning of ({200}γ, {220}γ and {311}γ) and three body-centered cubic peaks
carbon from martensite to austenite. This is critically discussed and ({200}α, {211}α and {220}α) [5].
compared to quenching and austempering above Ms (QAT), which is a An Instron Electrothermal Mechanical Testing (ETMT) heating stage
more conventional processing route employed to stabilize retained was used to reproduce the various Q&P heat treatments. Samples were
austenite through bainitic transformation. heated up by Joule effect while the cooling step was achieved by heat
extraction through the device jaws. In order to apply a given Q&P heat
treatment as plotted in Fig. 1, the set-up was able to deliver specific
2. Experimental procedures
heating (10 °C/s) and cooling (50 °C/s) rates. The high soaking tem-
perature at 900 °C was performed under controlled atmosphere using
The material investigated is a 0.8 mm-thick cold-rolled sheet which
argon. The temperature was monitored using a Pt/Pt-13%Rh thermo-
composition (in wt%) and critical temperatures are given in Table 1.
couple spot-welded to the sample. Special attention was paid to analyze
The transformation temperatures reported in Table 1 were measured by
the material in the vicinity of the welded thermocouple.
dilatometry. Various heat treatments were considered in the present
After in-situ experiments, samples were prepared for metallographic
study as illustrated in Fig. 1:
investigations. The detailed metallographic preparation methods can be
found elsewhere [22].
1. A quench and austempering (QAT) treatment, in order to analyze
the bainitic transformation at 400 °C.
2. A set of Q&P heat treatments with different initial quench tem- 3. Results
peratures (i.e. QT = 280, 320 and 360 °C). These treatments were
performed with fixed partitioning parameters (i.e. PT = 400 °C and In-situ HEXRD coupled with very high acquisition rates allowed
Pt = 120 s) and a fixed heating rate of 10 °C/s. gaining further insight on the evolution of austenite throughout the
entire Q&P treatment. As an example, Fig. 2 illustrates the evolution of
Prior to HEXRD in-situ experiments, the different heat treatments in the diffractograms obtained at chronological key moments of a given Q
Fig. 1 were carried-out on a DIL805A Bahr quench dilatometer. The &P process (QT = 360 °C). From bottom to top, diffractograms corre-
final multiphase Q&P microstructures were characterized by means of sponding to the (1) soaking step at 900 °C, (2) the initial quench to QT

Table 1
Chemical composition (in wt%) and measured critical temperatures (°C) of the investigated steel.
C Si Mn Cr P S N Fe Ac3 Ac1 Ms

0.20 1.41 2.31 0.21 0.003 0.002 0.005 balance 853 ± 6 754 ± 5 370 ± 7

176
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

Fig. 1. Schematic heat treatment cycles used in the present study: Q&P = quenching & partitioning; QAT = quenching & austempering.

Table 2
Phase volume fractions in the Q&P microstructures (TM=tempered initial
martensite; RA=retained austenite B=bainite; FM=fresh martensite; CRA is
the carbon concentration in RA).
QT (°C) Time at 400 °C (s) TM B FM RA CRA (wt%)

Q&P 280 120 0.94 0 0 0.06 1.14

320 120 0.78 0.07 0.04 0.11 1.16
360 120 0.17 0.55 0.20 0.08 1.17

Fig. 3. Evolution of the austenite lattice parameter with temperature during the
initial quench. The thermal expansion parameter of austenite is measured from
the data recorded during the initial quench (between 900 °C and 400 °C).

estimation is done assuming no mechanical or chemical contribution to

the observed evolution of the austenite lattice parameter in the con-
sidered thermal window. The measured mean thermal expansion
coefficient is in good agreement with the values reported by Lu et al.
[31] and Allain et al. [26].
Fig. 4 presents the evolution of the volume fraction of the FCC-phase
Fig. 2. 1D diffractograms obtained throughout an entire Q&P heat treatment during the initial quench. The error made on the phase volume fraction
(from bottom to top). FCC = face-centered cubic; BCC = body-centered cubic. measurement is 1% in terms of absolute phase fraction [26]. As it can
be seen, a significant decrease of FCC-phase fraction occurs during the
= 360 °C, (3−5) the partitioning at 400 °C and (6) the final quench to initial quench as the martensitic transformation occurs when tem-
room temperature are presented. The characteristics (i.e. position, in- perature falls below Ms. The latter is found to be around 384 ± 5 °C
tensity, the full width at half maximums, shape) of the diffraction peaks
provide useful information regarding the microstructural evolution
taking place during the process. In the present work, two distinct phases
were considered as can be seen in Fig. 2: a face-centered cubic (FCC)-
phase corresponding to austenite and a body-centered cubic (BCC)-
phase. No recognizable peak splitting proving the martensite tetra-
gonality was observed. Hence, the BCC phase could therefore be ferrite,
bainite or martensite, depending on the temperature at which they were
formed.

3.1. Initial quench to QT

The Q&P process starts with an initial quench from the austeniti-
zation stage at 900 °C to a given initial quench temperature QT. Note
that at the end of the soaking stage, the microstructure is fully auste-
nitic. The evolution of the austenite lattice parameter aγ with tem-
perature during the initial quench is plotted in Fig. 3. From these data, Fig. 4. Evolution of FCC volume fraction with temperature during the initial
an average thermal expansion coefficient of the FCC-phase is calculated quench to QT. The dashed line corresponds to the FCC volume fractions during a
and is found to be 2.349 × 10−5 K−1 between 900 °C and 400 °C. This full quench to room temperature.

177
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

340 °C, corresponding to about 30% of martensite formed, the tendency

is reversed and the lattice parameter decreases and falls below the pure
thermal contraction line at QT (i.e. 287 °C). At QT, the lattice parameter
is finally slightly lower compared to the value that is predicted when
considering only a thermal contribution (i.e. when neglecting other
contributions such as mechanical and chemical effects).

3.2. Re-heating from QT to 400 °C

Fig. 7 presents the evolution of the BCC and FCC lattice parameters
during the re-heating step for the three samples. During the reheating
stage, from QT to 400 °C, the evolution of the FCC and BCC lattice
parameters exhibit some clear differences. The BCC lattice parameter
exhibits a purely linear evolution with temperature while the evolution
of FCC lattice parameter deviates significantly from linearity above
360 °C, for all the QT investigated. The extent of the deviation from
Fig. 5. SEM micrographs of Q&P samples heat treated with the ETMT device linearity increases as the initial quench temperature decreases, i.e.
(in-situ HEXRD). Special attention is given to the area close to the surface of the when a larger fraction of martensite is present in the microstructure.
specimen. F = Ferrite. This can be further highlighted by comparing the difference between
the actual value of the austenite lattice parameter and the expected
with in-situ HEXRD experiments and closely corresponds to what was value at 400 °C if only thermal expansion had taken place. The devia-
measured using dilatometry with similar cooling conditions (Table 1). tion measured for QT = 280 °C is about 10 times higher than the one
The fraction of martensite formed depends on the stop quench tem- measured for QT = 360 °C (Table 3).
perature QT: BCC volume fractions of 0.84, 0.78 and 0.46 are measured The evolution of the FCC volume fractions during the reheating
after initial quenching to 287 °C, 310 °C and 358 °C, respectively. These stage is plotted in Fig. 8.
values are also in good agreement with the volume fractions found Some austenite is already consumed during the reheating stage. Its
using dilatometry (Table 2). As it can be seen in Fig. 4, the martensite extent depends on the initial quench temperature QT, i.e. on the mar-
transformation kinetics are quite fast. The fraction of martensite formed tensite fraction formed. For the lowest QT, a moderate decrease of 2.7%
is hence very sensitive to the stop quench temperature QT. A slight and 3.3% (for 280 °C and 320 °C, respectively) is measured. A more
deviation will lead to significant changes in fraction of martensite pronounced decrease (13.9%) is observed following quenching to
formed. 360 °C. This significant decrease in austenite fraction does not seem to
Moreover, as can be seen in Fig. 4, when cooling from 900 °C to influence the lattice parameter aγ (Fig. 7).
400 °C (i.e. above Ms) a small fraction (i.e. 6%) of BCC-phase is formed
at the expense of the austenite. This small BCC fraction can be attrib- 3.3. Partitioning at 400 °C
uted to ferrite that formed during the quench at the surface of the
specimen. Indeed, some ferrite grains can be observed in Fig. 5 in a Special attention has been paid to the evolution of the austenite
layer of approximately 20 µm under the surface of the sample. This is peaks during the partitioning step, in order to gain insight into the
due to the lack of hardenability of the outer layers following dec- phenomena taking place during partitioning. As can be seen in Fig. 9,
arburization inherited from the prior industrial processing of the steel the intensity of the austenite peaks decreases with partitioning time.
sheet [25,32]. The corresponding austenite volume fractions are illustrated in Fig. 10.
The evolution of the austenite lattice parameter (aγ) during the first As already noticed during the reheating stage, the austenite fraction
quench is presented in Fig. 6. aγ decreases linearly with temperature decreases during the partitioning stage but the extent of the austenite
above Ms, despite the small amount of ferrite formed. Its evolution consumption depends strongly on the initial quench temperature. The
slightly deviates from the pure thermal contraction when martensite is fraction of austenite consumed during the partitioning stage increases
formed: firstly, the austenite lattice parameter is slightly higher than when QT increases.
what is predicted if only thermal contraction is considered. Then, below The decrease in intensity of the austenite peaks is accompanied by a
shift to lower 2θ angles. The values of the lattice parameters de-
termined from the peaks positions are presented in Fig. 11. The FCC
lattice parameter significantly increases while the BCC lattice para-
meter only exhibits a small decrease at the early stages of partitioning.
The evolution of the BCC lattice parameter with partitioning time ex-
hibits a similar pattern for the three initial quenching temperatures. On
the contrary, the kinetics of austenite lattice parameter evolution is
different and depends on the initial quenching temperature QT. Hence,
the increase of aγ is almost completed within 60 s after initial quenching
to 280 °C while the increase of aγ after initial quenching to 360 °C is
significantly slower and does not seem to be completed within the
partitioning time investigated.

3.4. Final quench to room temperature

During the final quench to room temperature, about 2% of BCC

fraction is formed at the expense of austenite in the specimen that was
Fig. 6. Evolution of the austenite lattice parameter with temperature during the initially quenched to 360 °C. For lower initial quench temperatures, i.e.
initial quench to QT. The dashed line corresponds to the pure thermal con- 280 °C and 320 °C, the volume fraction of BCC phase formed is about
traction of austenite using the thermal expansion coefficient measured in Fig. 3. 0.5%, close to the experimental error.

178
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

Fig. 7. Evolution of (left) the BCC and (right) the FCC lattice parameters during the re-heating stage from QT to 400 °C. The dashed lines in the right graph represent
the evolution of austenite lattice parameter if only thermal expansion is considered.

The final microstructures obtained after in-situ Q&P processing on

the ETMT device are almost identical to those previously obtained with
dilatometry or with molten salt baths [14]. They consist of a matrix of
lath martensite containing retained austenite, which is highlighted in
green in the IQ maps presented in Fig. 12. Some retained austenite/
fresh martensite islands (appearing in dark in the IQ maps in Fig. 12)
can be recognized in the specimens quenched to 320 °C and 360 °C.

4. Discussion

In-situ HEXRD allowed continuous monitoring of the microstructure

evolution during Q&P. The changes in austenite fraction and its lattice
parameter measured during each step of the Q&P thermal cycle are
summarized in Table 3. In the next paragraphs, the metallurgical phe-
nomena occurring during Q&P will be firstly discussed considering the
evolution of the phase fractions and their respective lattice parameters. Fig. 8. Evolution of the FCC volume fraction during the re-heating stage from
The effect of the initial quench temperature QT will be depicted. Finally, QT to 400 °C.
the extent of carbon partitioning is evaluated by means of simple
carbon mass balance and the origin of austenite stabilization in Q&P By subtracting the thermal contribution to the actual austenite lat-
steels is discussed. tice parameters, the actual deviation daγ is resulting either from a
While the austenite fraction can be easily retrieved from the HEXRD change in carbon content or the generation of type II internal stresses.
patterns, the interpretation of the evolution of austenite lattice para- The non-thermal related lattice parameter variations will be discussed
meters during the Q&P cycle is more complicated as various phe- through the four different steps of the Q&P process leading eventually
nomena could contribute. Indeed, in addition to the thermal contribu- to the stabilization of the retained austenite: (1) the evolution in daγ
tion, any change in its stress state and/or in its chemical composition observed during the initial martensitic transformation, (2) the reheating
may affect the austenite lattice parameter. Fig. 13 presents the evolu- stage (from QT to 400 °C), (3) the partitioning at 400 °C and (4) the final
tion of the austenite lattice parameter after subtracting the contribution quench to room temperature.
from thermal expansion throughout the entire Q&P heat treatment
[25,26,33]. The choice of the reference austenite thermal expansion is 4.1. Initial quench
of paramount importance for the further determination of the austenite
carbon content and stress state. Allain et al. [33] reported that a con- During the initial quench, a given fraction of martensite is formed
stant thermal expansion coefficient (CTE) measured from Fig. 2(i.e. below the initial Ms. The martensite fraction formed depends only on
= 2.3493 10−5 K−1) over the entire temperature range in-
1 daγ
aγ dT the undercooling below Ms as it was observed in Fig. 4. The martensite
vestigated leads to an overestimation of the austenite thermal con- volume fractions measured by HERXD are generally in good agreement
traction, especially at low temperatures. Therefore, the thermal auste- with those measured previously by dilatometry. However, it has to be
nite expansion is here modeled using a temperature-dependent CTE and noticed that a slight deviation of quench temperature QT will lead to a
taking into account the constraining effect of martensite, following the significant shift of martensite volume fraction due to the fast transfor-
same approach proposed by Allain et al.. mation kinetics. This could explain the difference found for the QT

Table 3
Evolutions of austenite volume fraction and austenite lattice parameter during the re-heating stage and the partitioning step at 400 °C. The error made on the phase
volume fraction measurement is ± 1% in terms of absolute phase fraction. The uncertainty on the lattice parameter is ± 0.0002 Å [26].
QT (°C) Time at 400 °C (s) Quenching Re-heating Partitioning

ΔFCC vol. fraction (%) ΔFCC vol. fraction (%) daγ (Å) ΔFCC vol. fraction (%) Δaγ (Å)

Q&P 280 120 84.5 2.5 0.0083 3.2 0.0111

320 120 78.3 3.2 0.0053 6.7 0.0155
360 120 45.9 13.9 0.0006 26.0 0.0157

179
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

Fig. 9. 1D diffractograms of the FCC peaks at selected times of the partitioning step at 400 °C (after initial quench at QT = 320 °C).

0.2 C here investigated, austenite is likely to accommodate a larger part

of the transformation induced strain.

4.2. Reheating step

Some decomposition of austenite is already observed during re-

heating from QT to the partitioning temperature (400 °C). The amounts
of austenite consumed are summarized in Table 3. The exact nature of
the BCC product formed during reheating is debatable as it occurs
above the initial QT to the partitioning temperature of 400 °C. Previous
studies have shown that the formation of prior athermal martensite
could have a strong accelerating effect on the subsequent formation of
bainite [37–39]. As can be seen in Table 3, the volume fraction of BCC-
phase is small for low quenching temperatures (i.e. 280 °C and 320 °C).
Fig. 10. Evolution of the FCC volume fraction during the partitioning step at Therefore, the question of whether it is martensite, bainite or a mixture
400 °C. of both will not have a great impact on the final desired product.
However, when initial quenching at 360 °C, that fraction increases up to
= 360 °C specimen between the two techniques. almost 14%. Given the high fractions of BCC-phase formed and the high
The martensitic transformation induces some variations in the transformation temperatures involved (i.e. from 360 °C to 400 °C), it is
austenite lattice parameters, as already reported in literature [33–35]. difficult to explain this by the formation and/or growth of athermal
In the very first stage, an expansion of the austenite lattice is observed martensite or by the migration of austenite/martensite interfaces. For
while below 340 °C a contraction is recorded. This finding is interpreted these reasons, the phase formed during re-heating is assumed to be
in terms of micro-stress type II generating during the martensitic bainite.
transformation, as recently reported by Villa et al. [36]. At the start of During the re-heating step a significant deviation from non-linear
the transformation, martensite is the minority phase and undergoes evolution of the FCC lattice parameters with temperature was observed.
compressive stresses, which are balanced by a very small tensile stress It can reasonably be assumed that carbon partitioning, taking place at
in austenite. As martensite transformation progresses, the state of stress temperature above 360 °C, is responsible for the additional austenite
is reversed. Martensite experiences tensile stress while compression lattice expansion recorded. By applying the relationship developed by
stresses develop into austenite. The extent of the compression in aus- Toji et al. [40], it is possible to extract the corresponding carbon en-
tenite is however lower than that reported by Allain et al. [33] in a richment from the measured lattice parameter. This was highlighted by
0.3 C steel. In fact, due to the higher transformation temperature in the considering the difference, at 400 °C, between the actual austenite lat-
tice parameter that is measured and the austenite lattice parameter that

Fig. 11. Evolution of (left) the BCC and (right) the FCC lattice parameters during the partitioning step at 400 °C.

180
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

Fig. 12. EBSD micrographs of Q&P microstructures showing combined Image Quality (fresh martensite appearing in dark) and austenite phase (in green) maps: QT
= (a) 280 °C, (b) 320 °C and (c) 360 °C.

is expected if only thermal expansion is occurring. The increase in the The austenite lattice parameter significantly increases during the
austenite lattice parameter is thus converted into an increase in the partitioning step as was seen in Fig. 11. This is attributed to the carbon
austenite carbon concentration. Carbon enrichments of 0.25, 0.16 and enrichment of austenite resulting from both carbon partitioning from
0.02 occurred during the reheating step from 280, 320 and 360 °C, martensite and carbon partitioning during the bainitic transformation.
respectively. These carbon contents are in wt% and the uncertainties As shown in the previous section, the increase of the austenite lattice
are around 0.015% [26]. As can be seen, the largest and most rapid parameter is thus converted into an increase in the austenite carbon
carbon enrichment occurred during reheating from the lowest initial concentration. Carbon enrichments of 0.34, 0.47 and 0.48 occurred
quench temperature (i.e. QT = 280 °C). Conversely, during reheating during the partitioning step for the QT = 280, 320 and 360 °C speci-
from QT = 360 °C, the increase in daγ is the smallest, although it is mens, respectively.
associated to the largest austenite consumption (and bainite formation). Similarly to what was observed during reheating, the kinetics of the
Two reasons may explain the fastest enrichment kinetics following austenite lattice parameter evolution during partitioning depends on
quench to low QT: (1) the larger amount of martensite available, and the initial quenching temperatures QT (Fig. 11). Hence, the increase of
consequently the larger amount of carbon potentially available for aγ is almost immediate after initial quenching at 280 °C while the in-
partitioning, (2) the larger density of austenite/martensite interfaces, crease of aγ is significantly slower after initial quenching at 360 °C, just
through which the carbon may diffuse (carbon diffusion in austenite below Ms. From a general point of view, when the initial quench tem-
being the limiting step of the partitioning process [41]). perature QT is low, such as 280 °C and 320 °C, the principal contribution
to carbon enrichment is provided by carbon partitioning from mar-
4.3. Partitioning step tensite. On the contrary, when initial quenching is achieved at 360 °C,
the austenite decomposition into bainite during partitioning also con-
During the partitioning step, the constrained carbon equilibrium tributes to the carbon enrichment of austenite. In this case, the carbon
(CCE) model developed by Speer et al. supposes that carbon partitions enrichment of retained austenite is therefore a combination of carbon
from the supersaturated martensite to the residual untransformed aus- partitioning from martensite and from the bainitic transformation.
tenite [11,14,42]. The partitioning of carbon is assumed to take place Fig. 14 highlights this statement as the evolution of the austenite lattice
without any other competing reactions (i.e. no carbide precipitation, no parameter in the QP360C specimen is intermediate between the
bainite transformation, no C trapped at lattice defects and interfaces are QP320C and the quenching and austempering QAT specimens (pure
assumed to remain immobile). However, it was observed in Fig. 10 that carbon enrichment due to bainite transformation). Hence, the stabili-
a significant fraction of austenite is consumed during the partitioning zation mechanism of metastable retained austenite differs between the
step, suggesting the occurrence of bainite transformation [16,19]. The QAT and the Q&P process. While the carbon enrichment of austenite is
fraction of bainite formed during the partitioning step depends on the achieved through the formation of bainitic ferrite and subsequent C
stop quench temperature (i.e. on the martensite fraction present), and diffusion during austempering, the carbon diffuses directly from the
thus on the amount of untransformed austenite present at the beginning supersaturated martensite into the surrounding austenite during the
of the partitioning step. partitioning step. Unlike austempering, carbon partitioning and

Fig. 13. Evolution of day throughout the entire Q&P heat treatment. The small figure in the QT = 280 °C case corresponds to a zoom of day during the initial quench
from 400 °C to 280 °C.

181
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

largest carbon enrichment took place during reheating for low QT while
it occurred during partitioning for high QT. Hence, the largest and most
rapid carbon enrichment occurred during reheating from the lowest
initial quench temperature (i.e. QT = 280 °C). The fast carbon en-
richment results from a rapid carbon depletion of martensite. Con-
versely, during reheating for the QT = 360 °C cycle, the increase in daγ
is the smallest, although it is associated to the largest austenite con-
sumption (and bainite formation). During partitioning, the opposite
observation was made: both the carbon enrichment and austenite
consumption were the largest for high QT (i.e. QT = 360 °C).
By summing the carbon enrichments measured during the reheating
step and the partitioning step to the initial austenite carbon con-
centration (i.e. 0.2 wt%), the final carbon concentration of retained
austenite at the end of the partitioning can be calculated. The final
Fig. 14. Comparison of the evolution of the FCC lattice parameters during carbon concentrations in austenite, summarized in Table 4, are equal to
partitioning and austempering (QAT) at 400 °C. 0.79 wt%, 0.83 wt% and 0.69 wt% when the initial quenching is
achieved at 280, 320 and 360 °C, respectively.
microstructure development are decoupled in the Q&P process. Indeed, Assuming that only the carbon has diffused during the Q&P process,
in quenching and partitioning, the carbon originates from martensite the corresponding martensite-start temperature Ms of the residual
which was formed during the initial quench, while in quenching and austenite is calculated using an empirical relation [43,44]. Surprisingly,
austempering, the carbon originates from bainite which still has to form all the temperatures were found to be higher than room temperature:
at 400 °C. 124 °C, 105 °C and 165 °C. The carbon enrichment up to 0.8% alone
does not justify the retention of austenite at room temperature, other
4.4. Final quench effects contribute to the stabilization of austenite.
The sub-micron size explains the higher stability of the retained
During the final quench to room temperature, a further increase in austenite comparing to “bulk” austenite: recent work [45] carried-out
daγ is observed. The latter is in this case very unlikely to be due to an in order to include the grain size effects in the computation of the Ms
additional carbon enrichment of austenite but is rather the consequence temperature, demonstrates that actual Ms value of a 0.83% C containing
of mechanical stresses rising into austenite. Indeed, the low tempera- austenite drops to room temperature by decreasing the grain size to
tures involved (i.e. less than 200 °C) as well as no significant changes in 0.35 µm, i.e. a value consistent with that measured by EBSD.
FCC volume fractions do not give any reason to support any chemical Finally, a mass carbon balance can be carried out to estimate the
changes in austenite during the final quench. Type II internal stresses part of carbon that has not been partitioned to the residual austenite.
are possibly induced during the cooling step by the difference in Hence, the total carbon content remaining in the BCC-phases (i.e.
thermal expansion coefficients between austenite and martensite [27]. martensite and bainite) as precipitated carbides or C trapped on dis-
Hence, austenite wants to contract more than what the surrounding locations at room temperature is estimated by subtracting the retained
matrix of martensite/bainite allows. This constrained contraction austenite carbon content from the nominal carbon content. As can be
during the final quench generates internal hydrostatic tensile stresses in seen in Table 4, the amount of carbon trapped in the BCC-phases is not
austenite. The formation of some fresh martensite during the final negligible as almost 50% of nominal carbon content present in the
quench may explain the higher increase in daγ observed in the QT specimen does not participate to the stabilization of austenite. The
= 360 °C specimen when comparing to the two other specimens. In- amount of carbon trapped in remaining BCC-phases estimated is in
deed, the formation of high carbon martensite can generate additional good agreement with the large numbers of carbides found in martensite
stresses into the retained austenite. in our previous investigation [22] and with what has been reported in
previous works [19,27,46].
4.5. Total carbon enrichment and mass balance

The final retained austenite volume fractions stabilized at room 4.6. Stress state of austenite at room temperature
temperature measured by in-situ HEXRD are slightly higher than those
measured by conventional XRD (Table 2). Indeed, values of retained The final stress state of austenite is the results of different me-
austenite at room temperature of 9.9%, 13.2% and 12.2% (instead of chanisms inducing internal stresses along the Q&P cycle. Austenite
6%, 11% and 8%) are measured following Q&P cycles with stop quench undergoes compressive hydrostatic stresses following the initial
temperature of 280, 320 and 360 °C, respectively. This difference is quench, which are partially accommodated by deformation. Then,
possibly due to the fact that transmission HEXRD allows bulk mea- during the reheating step hydrostatic stress (“eigenstrain”) can rise
surement over the entire specimen thickness, while conventional XRD is because of the different CTE of austenite and ferrite [33], which are also
a surface measurement being more sensitive to the sample preparation. possibly accommodated because of the relatively high temperature. For
The carbon enrichments measured during the reheating and the the same reason, no stresses are generated during the partitioning step,
partitioning steps are summarized in Table 4. As it can be seen, the the entire deviation daγ could be attributed to carbon enrichment.

Table 4
Final volume fractions (%) and carbon concentrations (wt%) of the residual austenite at room temperature, after quenching and partitioning with various QT. The
corresponding Ms (°C) of the retained austenite is also summarized as well as the carbon trapped in the BCC-phase.
QT (°C) Time at 400 °C (s) Reheat. ΔC (wt%) Partit. ΔC (wt%) Final RA vol. (%) Final CRA (wt%) Corresp. Ms (°C) C trapped in BCC (wt%)

Q&P 280 120 0.25 0.34 9.9 0.79 124 0.12

320 120 0.16 0.47 13.2 0.83 105 0.09
360 120 0.02 0.48 12.2 0.70 165 0.11
QAT 400 1000 – 0.55 13.3 0.75 141 0.10

182
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

During the final quench, the “eigenstain” leads to significant tensile partitioning is provided from carbon partitioning from martensite while
stresses in austenite, the thermal strain being difficult to accommodate it is a combination of carbon partitioning from martensite and carbon
at lower temperatures. The final stress state at room temperature is the rejected from bainite transformation for high QT (just below Ms).
sum of each single contribution. However, its exact computation is a
tricky issue because it is very sensitive to the assumptions made for the Acknowledgements
stress-free CTE of austenite, as recently demonstrated by Allain et al.
[32]. It can be argued that a final small compressive state of austenite as We acknowledge the European Synchrotron Radiation Facility for
well as the increase in dislocation density in austenite may further assist provision of synchrotron radiation facilities and we would like to thank
the stabilization of austenite at room temperature. P. Sedmak and T. Buslaps for assistance in using beamline ID11.

4.7. Consequences of the present results on the understanding of the References

mechanical properties
[1] M.J. Santofimia, L. Zhao, R. Petrov, C. Kwakernaak, W.G. Sloof, J. Sietsma,
The effect of the Q&P process on the mechanical properties was Microstructural development during the quenching and partitioning process in a
newly designed low-carbon steel, Acta Mater. 59 (2011) 6059–6068, https://doi.
extensively studied in our previous works, in particular the tensile org/10.1016/j.actamat.2011.06.014.
properties [22] and the stretch-flange formability [23]. The results [2] E. de Moor, S. Lacroix, A.J. Clarke, J. Penning, J.G. Speer, Effect of retained aus-
presented in the present work give new insight on the effect of the Q&P tenite stabilized via quench and partitioning on the strain hardening of martensitic
steels, Metall. Mater. Trans. A. 39 (2008) 2586–2595, https://doi.org/10.1007/
parameters on the mechanical properties. For a more complete dis- s11661-008-9609-z.
cussion, the reader is referred to these references. [3] V. Zackay, E.R. Parker, D. Fahr, R. Busch, The enhancement of ductility in high-
One of the highlighted result was that for QT = 280 °C and 320 °C strength steels, Trans. ASM 60 (1967) 252–259.
[4] D. De Knijf, C. Föjer, L.A.I. Kestens, R. Petrov, Factors influencing the austenite
(corresponding to fractions of initial martensite larger than 0.75), the stability during tensile testing of Quenching and Partitioning steel determined via
ultimate tensile strength decreases and the uniform elongation in- in-situ Electron Backscatter Diffraction, Mater. Sci. Eng. A. 638 (2015) 219–227,
creases when increasing the partitioning time from 10 s to 120 s. For a https://doi.org/10.1016/j.msea.2015.04.075.
[5] N.H. Van Dijk, A.M. Butt, L. Zhao, J. Sietsma, S.E. Offerman, J.P. Wright, S. Van Der
longer partitioning time up to 1000 s, the stress-strain curves were very
Zwaag, Thermal stability of retained austenite in TRIP steels studied by synchrotron
similar. These trends are in very good agreement with the results X-ray diffraction during cooling, Acta Mater. 53 (2005) 5439–5447, https://doi.
showed in Fig. 11: the carbon enrichment of austenite, mainly ensured org/10.1016/j.actamat.2005.08.017.
by partitioning of carbon from martensite, is completed within 100 s. [6] X.C. Xiong, B. Chen, M.X. Huang, J.F. Wang, L. Wang, The effect of morphology on
the stability of retained austenite in a quenched and partitioned steel, Scr. Mater. 68
Longer partitioning times hardly affect the amount and the carbon (2013) 321–324, https://doi.org/10.1016/j.scriptamat.2012.11.003.
concentration of retained austenite and therefore have only a slight [7] P.J. Jacques, F. Delannay, J. Ladrière, On the influence of interactions between
effect on the final mechanical properties. phases on the mechanical stability of retained austenite in transformation-induced
plasticity multiphase steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 32
For lower fractions of initial martensite (i.e. QT = 360 °C), the yield (2001) 2759–2768, https://doi.org/10.1007/s11661-001-1027-4.
strength continuously decreases and the total elongation increases with [8] D. De Knijf, R. Petrov, C. Föjer, L.A.I. Kestens, Effect of fresh martensite on the
increasing partitioning times. This is a direct consequence of the slower stability of retained austenite in quenching and partitioning steel, Mater. Sci. Eng.
A. 615 (2014) 107–115, https://doi.org/10.1016/j.msea.2014.07.054.
kinetics of austenite carbon enrichment when carbon diffusion is cou- [9] H.S. Zhao, W. Li, X. Zhu, X.H. Lu, L. Wang, S. Zhou, X.J. Jin, Analysis of the re-
pled with the bainitic transformation. For short partitioning times, lationship between retained austenite locations and the deformation behavior of
austenite is insufficiently stabilized and transforms into fresh marten- quenching and partitioning treated steels, Mater. Sci. Eng. A. 649 (2016) 18–26,
https://doi.org/10.1016/j.msea.2015.09.088.
site during the final quench, leading to poor mechanical properties, in [10] R. Blondé, E. Jimenez-Melero, L. Zhao, J.P. Wright, E. Brück, S. Van Der Zwaag,
general, and to low formability levels, in particular. Partitioning times N.H. Van Dijk, High-energy X-ray diffraction study on the temperature-dependent
significantly longer than 120 s (in the order of 1000 s) are needed to mechanical stability of retained austenite in low-alloyed TRIP steels, Acta Mater. 60
(2012) 565–577, https://doi.org/10.1016/j.actamat.2011.10.019.
complete the bainitic transformation and stabilize austenite at room
[11] J. Speer, D.K. Matlock, B.C. De Cooman, J.G. Schroth, Carbon partitioning into
temperature. austenite after martensite transformation, Acta Mater. 51 (2003) 2611–2622,
https://doi.org/10.1016/S1359-6454(03)00059-4.
5. Conclusions [12] M.J. Santofimia, L. Zhao, J. Sietsma, Overview of mechanisms involved during the
quenching and partitioning process in steels, in: Metall. Mater. Trans. A, Springer
US: 2011: pp. 3620–3626. doi: 10.1007/s11661-011-0706-z.
High Energy X-RAY diffraction is a powerful technique to scrutinize [13] A.S. Nishikawa, M.J. Santofimia, J. Sietsma, H. Goldenstein, Influence of bainite
the phenomena taking place during Quenching and Partitioning. As was reaction on the kinetics of carbon redistribution during the Quenching and
Partitioning process, Acta Mater. 142 (2018) 142–151, https://doi.org/10.1016/j.
shown in the present work, the phase transformations and the parti- actamat.2017.09.048.
tioning of carbon occurring during the reheating and the partitioning [14] J.G. Speer, D.V. Edmonds, F.C. Rizzo, D.K. Matlock, Partitioning of carbon from
steps are making the in-situ approach essential. More conventional supersaturated plates of ferrite, with application to steel processing and funda-
mentals of the bainite transformation, Curr. Opin. Solid State Mater. Sci. 8 (2004)
techniques are shown to be a good tool for a first phase quantification of 219–237, https://doi.org/10.1016/j.cossms.2004.09.003.
the complex Q&P microstructures. However, the HEXRD allows a “live” [15] M.J. Santofimia, L. Zhao, R. Petrov, J. Sietsma, Characterization of the micro-
analysis of the microstructural evolution in such way that no reliable structure obtained by the quenching and partitioning process in a low-carbon steel,
Mater. Charact. 59 (2008) 1758–1764, https://doi.org/10.1016/j.matchar.2008.
information is lost in comparison to these conventional techniques.
04.004.
The present work demonstrates that a final carbon concentration in [16] M.J. Santofimia, T. Nguyen-Minh, L. Zhao, R. Petrov, I. Sabirov, J. Sietsma, New
austenite of around 0.7–0.8% is sufficient to retain austenite at room low carbon Q&P steels containing film-like intercritical ferrite, Mater. Sci. Eng. A.
527 (2010) 6429–6439, https://doi.org/10.1016/j.msea.2010.06.083.
temperature, as the reduction in grain size contributes to stabilize
[17] H.Y. Li, X.W. Lu, X.C. Wu, Y.A. Min, X.J. Jin, Bainitic transformation during the
further the retained austenite. two-step quenching and partitioning process in a medium carbon steel containing
Two mechanisms responsible for carbon enrichment of austenite are silicon, Mater. Sci. Eng. A. 527 (2010) 6255–6259, https://doi.org/10.1016/j.msea.
identified in the studied commercial 0.2%C commercial grade: bainite 2010.06.045.
[18] E. Paravicini Bagliani, M.J. Santofimia, L. Zhao, J. Sietsma, E. Anelli,
transformation and martensite carbon depletion. The kinetics of carbon Microstructure, tensile and toughness properties after quenching and partitioning
enrichment following martensite carbon depletion is found to be fast treatments of a medium-carbon steel, Mater. Sci. Eng. A. 559 (2013) 486–495,
and can already start during the reheating stage. The carbon enrich- https://doi.org/10.1016/j.msea.2012.08.130.
[19] F. HajyAkbary, J. Sietsma, G. Miyamoto, T. Furuhara, M.J. Santofimia, Interaction
ment following bainite transformation is slower as bainite formation of carbon partitioning, carbide precipitation and bainite formation during the Q&P
and austenite carbon enrichment are coupled. process in a low C steel, Acta Mater. 104 (2016) 72–83, https://doi.org/10.1016/j.
The specific contribution of each mechanism depends on the initial actamat.2015.11.032.
[20] K.O. Findley, J. Hidalgo, R.M. Huizenga, M.J. Santofimia, Controlling the work
QT: for low QT's, the major contribution to carbon enrichment during

183
P. Huyghe et al. Materials Science & Engineering A 743 (2019) 175–184

hardening of martensite to increase the strength/ductility balance in quenched and S. Aoued, A. Poulon-Quintin, Internal stresses and carbon enrichment in austenite of
partitioned steels, Mater. Des. 117 (2017) 248–256, https://doi.org/10.1016/j. Quenching and Partitioning steels from high energy X-ray diffraction experiments,
matdes.2016.12.065. Mater. Sci. Eng. A. 710 (2018) 245–250, https://doi.org/10.1016/j.msea.2017.10.
[21] D.T. Pierce, D.R. Coughlin, D.L. Williamson, K.D. Clarke, A.J. Clarke, J.G. Speer, 105.
E. De Moor, Characterization of transition carbides in quench and partitioned steel [34] N. Nakada, Y. Ishibashi, T. Tsuchiyama, S. Takaki, Self-stabilization of un-
microstructures by Mössbauer spectroscopy and complementary techniques, Acta transformed austenite by hydrostatic pressure via martensitic transformation, Acta
Mater. 90 (2014), https://doi.org/10.1016/j.actamat.2015.01.024. Mater. 110 (2016) 95–102, https://doi.org/10.1016/j.actamat.2016.03.048.
[22] P. Huyghe, L. Malet, M. Caruso, C. Georges, S. Godet, On the relationship between [35] M. Villa, K. Pantleon, M.A.J. Somers, In situ investigation of the martensitic
the multiphase microstructure and the mechanical properties of a 0.2C quenched & transformation in Fe-12 wt%Ni-0.6 wt%C steel at subzero temperatures, J. Alloy.
partitioned steel, Mater. Sci. Eng. A. 701 (2017) 254–263, https://doi.org/10.1016/ Compd. 577 (2013), https://doi.org/10.1016/j.jallcom.2011.12.162.
j.msea.2017.06.058. [36] M. Villa, F. Niessen, M.A.J. Somers, In situ investigation of the evolution of lattice
[23] P. Huyghe, S. Dépinoy, M. Caruso, D. Mercier, C. Georges, L. Malet, S. Godet, On the strain and stresses in austenite and martensite during quenching and tempering of
Effect of Q&P processing on the stretch-flange-formability of 0.2C ultra-high steel, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 49 (2018) 28–40, https://doi.
strength steel sheets, ISIJ Int. 58 (2018) 1341–1350, https://doi.org/10.2355/ org/10.1007/s11661-017-4387-0.
isijinternational.ISIJINT-2018-121. [37] H. Kawata, K. Hayashi, N. Sugiura, N. Yoshinaga, M. Takahashi, Effect of martensite
[24] D. De Knijf, M.J. Santofimia, H. Shi, V. Bliznuk, C. Föjer, R. Petrov, W. Xu, In situ in initial structure on bainite transformation, Mater. Sci. Forum 638–642 (2010)
austenite-martensite interface mobility study during annealing, Acta Mater. 90 3307–3312, https://doi.org/10.4028/www.scientific.net/MSF.638-642.3307.
(2015) 161–168, https://doi.org/10.1016/j.actamat.2015.02.040. [38] A. Navarro-López, J. Sietsma, M.J. Santofimia, Effect of prior athermal martensite
[25] J. Epp, T. Hirsch, C. Curfs, In situ X-ray diffraction analysis of carbon partitioning on the isothermal transformation kinetics below M s in a low-C High-Si steel,
during quenching of low carbon steel, Metall. Mater. Trans. A. 43 (2012) Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 47 (2016) 1028–1039, https://doi.
2210–2217, https://doi.org/10.1007/s11661-012-1087-7. org/10.1007/s11661-015-3285-6.
[26] S.Y.P. Allain, G. Geandier, J.C. Hell, M. Soler, F. Danoix, M. Gouné, In-situ in- [39] Y. Toji, H. Matsuda, D. Raabe, Effect of Si on the acceleration of bainite transfor-
vestigation of quenching and partitioning by High Energy X-Ray Diffraction ex- mation by pre-existing martensite, Acta Mater. 116 (2016) 250–262, https://doi.
periments, Scr. Mater. 131 (2017) 15–18, https://doi.org/10.1016/j.scriptamat. org/10.1016/j.actamat.2016.06.044.
2016.12.026. [40] Y. Toji, H. Matsuda, M. Herbig, P.P. Choi, D. Raabe, Atomic-scale analysis of carbon
[27] S. Allain, G. Geandier, J.-C. Hell, M. Soler, F. Danoix, M. Gouné, Effects of Q&P partitioning between martensite and austenite by atom probe tomography and
processing conditions on austenite carbon enrichment studied by In situ high-en- correlative transmission electron microscopy, Acta Mater. 65 (2014) 215–228,
ergy X-ray diffraction experiments, Met. (Basel) 7 (2017) 232, https://doi.org/10. https://doi.org/10.1016/j.actamat.2013.10.064.
3390/met7070232. [41] M.J. Santofimia, L. Zhao, J. Sietsma, Model for the interaction between interface
[28] J. Wu, P.J. Wray, C.I. Garcia, M. Hua, A.J. Deardo, Image quality analysis: a new migration and carbon diffusion during annealing of martensite-austenite micro-
method of characterizing microstructures, ISIJ Int. 45 (2005) 254–262, https://doi. structures in steels, Scr. Mater. 59 (2008) 159–162, https://doi.org/10.1016/j.
org/10.2355/isijinternational.45.254. scriptamat.2008.02.045.
[29] M.J. Santofimia, R.H. Petrov, L. Zhao, J. Sietsma, Microstructural analysis of [42] D.V. Edmonds, K. He, F.C. Rizzo, B.C. De Cooman, D.K. Matlock, J.G. Speer,
martensite constituents in quenching and partitioning steels, Mater. Charact. 92 Quenching and partitioning martensite-A novel steel heat treatment, Mater. Sci.
(2014) 91–95, https://doi.org/10.1016/j.matchar.2014.03.003. Eng. A. 438–440 (2006) 25–34, https://doi.org/10.1016/j.msea.2006.02.133.
[30] ESRF Fit2D, (n.d.). 〈http://www.esrf.eu/computing/scientific/FIT2D/〉. [43] K.W. Andrews, Empirical formulae for the calculation of some transformation
[31] X. Lu, M. Selleby, B. Sundman, Theoretical modeling of molar volume and thermal temperatures, J. Iron Steel Inst. (1965) 721–727.
expansion, Acta Mater. 53 (2005) 2259–2272, https://doi.org/10.1016/j.actamat. [44] S.M.C. van Bohemen, Bainite and martensite start temperature calculated with
2005.01.049. exponential carbon dependence, Mater. Sci. Technol. 28 (2012) 487–495, https://
[32] J. Epp, H. Surm, O. Kessler, T. Hirsch, In situ X-ray phase analysis and computer doi.org/10.1179/1743284711Y.0000000097.
simulation of carbide dissolution of ball bearing steel at different austenitizing [45] Toolmart RFCS Project - Task 2, 2018.
temperatures, Acta Mater. 55 (2007) 5959–5967, https://doi.org/10.1016/j. [46] J.G. Thomas, Grant Aaron;Danoix, Frederique;Speer, Carbon atom re-distribution
actamat.2007.07.022. during quencing and partitioning, ISIJ Int. 54 (2014) 2900–2906.
[33] S.Y.P. Allain, S. Gaudez, G. Geandier, J.C. Hell, M. Gouné, F. Danoix, M. Soler,

184