Materials Science and Engineering A 527 (2010) 2725–2732

Contents lists available at ScienceDirect

Materials Science and Engineering A

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

Effect of microstructural features on the hot ductility of 2.25Cr–1Mo steel

X.-M. Chen a , S.-H. Song a,∗ , Z.-C. Sun a , S.-J. Liu a , L.-Q. Weng a , Z.-X. Yuan b
a
Shenzhen Key Laboratory of Advanced Materials, Department of Materials Science and Engineering, Shenzhen Graduate School,
Harbin Institute of Technology, Xili, Shenzhen 518055, China
b
School of Materials and Metallurgy, Wuhan University of Science and Technology, Wuhan 430081, China

a r t i c l e i n f o a b s t r a c t

Article history: Some factors contributing to the hot ductility losses of a 2.25Cr–1Mo steel were identified over the tem-
Received 21 October 2009 perature range 750–950 ◦ C, after the specimens were austenitized at 1000 ◦ C, furnace cooled to different
Received in revised form temperatures, and held there for sufficient periods of time, followed by tensile testing. There were two
20 December 2009
types of ferrite present in the microstructure, namely, pro-eutectoid ferrite and deformation-induced
Accepted 14 January 2010
ferrite. The pro-eutectoid ferrite was only formed below Ar3 (∼825 ◦ C), which was nucleated on the
inclusions and distributed uniformly. Nevertheless, the deformation-induced ferrite was formed in a
much wider temperature range. It was distributed mainly along austenite grain boundaries above Ar3 ,
Keywords:
Hot ductility
and around the pro-eutectoid ferrite below Ar3 . The deformation-induced ferrite had a primary effect
Steel on the hot ductility, which was mainly responsible for a hot ductility trough. There was a peak in the
Dynamic recrystallization quantity of deformation-induced ferrite between 800 and 900 ◦ C, which was just corresponding to the
Grain boundary segregation hot ductility trough. The morphology of ferrite was also essential. The net-like structure of ferrite formed
along austenite grain boundaries was the most deleterious to the hot ductility.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction micro-segregation, particularly of phosphorus. A study by Guo et

al. [5] demonstrated that a 2.25Cr–1Mo steel containing 0.06 wt.%P
Transverse cracking on the surface of continuously cast prod- has a lower hot ductility than the steel containing 0.01 wt.%P at
ucts is still a serious problem which causes ductility losses in the temperatures between 750 and 1000 ◦ C. However, as reviewed by
austenite–ferrite region for low alloy low carbon steels. Cracks Mintz [6], phosphorus is likely to improve the hot ductility and
can nucleate and propagate during the straightening operation reduce transverse cracking during straightening if the segregation
which is carried out normally in the range 700–950 ◦ C. Previous to grain boundary is controlled.
studies [1–6] have indicated that hot ductility deterioration in the Until the present time, there have been few studies that have
austenite–ferrite region is caused by the following factors: (1) for- distinguished the effects of deformation-induced ferrite and pro-
mation of a thin pro-eutectoid ferrite layer along austenite grain eutectoid ferrite on the hot ductility. It was the aim of the present
boundaries; (2) segregation of undesirable elements, such as S, Sb, work to investigate the effects of the morphology and magnitude
Sn, As, P, Cu and others, at ferrite/austenite interfaces or at austen- of these two different ferrites on the hot ductility of a 2.25Cr–1Mo
ite grain boundaries; and (3) intergranular precipitation of carbides, steel. Dynamic recrystallization and non-equilibrium segregation
carbonitrides, or nitrides. of phosphorus during high temperature deformation were also
Mintz et al. [1] suggested that the hot ductility trough is con- addressed.
trolled by the austenite–ferrite transformation. It was confirmed
[2] that the hot ductility trough in austenite–ferrite region depends
primarily on the thickness of pro-eutectoid ferrite formed along 2. Experimental procedures
austenite grain boundaries. Nachtrab and Chou [3] put forward that
grain boundary segregation of Sn, Cu and Sb can seriously reduce The experimental steel was prepared by vacuum induction
the hot ductility of C–Mn steel and deformation at high temper- melting with an ingot of 50 kg. The chemical composition of the
atures can enhance the grain boundary segregation. Harada et al. steel (wt.%) was 0.13C, 0.40Si, 0.53Mn, 0.028P, 0.0097S, 0.023Cu,
[4] pointed out that the origin of surface transverse cracks is the 2.45Cr and 0.97Mo. Obviously, the steel was doped with P. The
resulting ingot was hot rolled in the range 900–1000 ◦ C into a plate
20 mm in thickness. The tensile specimens were machined from
the plate with their axes perpendicular to the rolling direction. The
∗ Corresponding author. Tel.: +86 755 26033465; fax: +86 755 26033504. specimens had a gauge length of 50 mm with a gauge diameter of
E-mail address: shsonguk@yahoo.cn (S.-H. Song). 10 mm.

0921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2010.01.047
2726 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732

In this work, two groups of specimens were prepared. One group

was used for hot tensile tests, and the other was undeformed and
used merely for metallographic observation after undergoing the
same thermal history as the former for comparison. Hot tensile tests
were performed using a high temperature tensile testing machine.
Under an argon atmosphere, the specimens were austenitized at
1000 ◦ C for 15 min, then furnace cooled to different test temper-
atures between 750 and 950 ◦ C where the specimens were held
for adequate time (750 ◦ C for 60 min, 800 ◦ C for 30 min, 812 ◦ C
for 30 min, 818 ◦ C for 30 min, 825 ◦ C for 30 min, 832 ◦ C for 30 min,
840 ◦ C for 30 min, 850 ◦ C for 30 min, 900 ◦ C for 15 min, and 950 ◦ C
for 15 min) to enable phosphorus grain boundary segregation to
reach equilibrium, followed by tensile testing. According to McLean
[7], Song et al. [8] and Erhart et al. [9], the holding times at differ-
ent temperatures were long enough to acquire equilibrium grain
boundary segregation. The tensile tests were carried out with a
constant strain rate of 10−3 s−1 which is close to that used dur-
ing straightening operation [6], and the reduction in area (RA) was
employed to evaluate the hot ductility. After the specimens were
fractured, they were immediately quenched by a high speed argon
Fig. 1. Hot ductility as a function of temperature for the experimental steel.
flow with a cooling rate higher than 30 K/s [10] to prevent further
austenite–ferrite transformation during cooling. The undeformed
specimens were maintained for the same time as the deformed process of the testing. From that temperature, the magnitude of fer-
ones at corresponding temperatures, and then quenched. rite in the microstructure increases gradually with decreasing test
Microstructures near the fracture surfaces for the deformed temperature and the corresponding hot ductility drops. At 825 ◦ C,
and undeformed specimens were metallographically examined almost perfect ferrite networks along austenite grain boundaries
using optical microscopy. In addition, the fraction of ferrite in the are formed by ferrite films, which corresponds to the minimum
microstructure was measured. In the measurement, 15 fields with RA value (see Fig. 1). With further decreasing test temperature, the
an area of 2.7 × 105 ␮m2 each were randomly selected for each con- magnitude of ferrite increases continuously with the network mor-
dition and the mean value of data points obtained was taken as the phology of ferrite dying away gradually and the corresponding hot
measured result. Fractographic observations were performed using ductility starts to recover. For example, at 818 ◦ C there are few fer-
a Hitachi S-4700 field emission gun scanning electron microscope rite precipitates with a film-like morphology, and most of them are
(FEG-SEM). isolatedly distributed (see Fig. 2a), which would be beneficial to the
The austenization temperature used in the present work was ductility. Based on the above, it is claimed that the net-like struc-
1000 ◦ C, which is lower than the temperature usually used for ture of ferrite along austenite grain boundaries plays an important
the research of hot ductility. This is because the highest opera- role in the reduction of ductility. Fig. 3 represents the microstruc-
tion temperature of the hot tensile testing machine used is just tures before deformation at 825 and 818 ◦ C, respectively. Clearly,
above 1000 ◦ C. The austenization temperature is usually between there is almost no ferrite formed at 825 ◦ C while there is some
1320 and 1350 ◦ C for the investigation into hot ductility in order at 818 ◦ C. Hence, without deformation the highest temperature at
to explore transverse cracking present in continuous casting. Nev- which the ferrite can be formed is around 825 ◦ C. In addition, the
ertheless, a study by Yasumoto et al. [11] suggests that on raising morphology of ferrite is much different from that after deforma-
the solution treatment temperature, the width and depth of the tion. There are few ferrite precipitates formed preferentially along
ductility trough are greatly increased for low carbon steels, but grain boundaries.
the temperature corresponding to the minimum ductility does Three types of ferrite, namely, pro-eutectoid ferrite,
not change apparently. Transverse cracking on the surface of slab deformation-induced ferrite and Widmanstätten ferrite, may
is most likely produced around the minimum ductility tempera- precipitate during the experiment. According to Du et al. [12],
ture. Accordingly, the present work, which focuses on the effect of the morphology of Widmanstätten ferrite is mainly Widmanstät-
microstructural features on the hot ductility, is still of implication ten and allotriomorphs, which is not present in Figs. 2 and 3.
in exploring the susceptibility to transverse cracking in contin- Thus the ferrite existing in the present case can be divided into
uous casting although a lower austenization temperature was pro-eutectoid ferrite produced during isothermal holding before
employed. tensile testing and deformation-induced ferrite produced during
tensile testing. Previous work [12] demonstrated that not any
3. Results and discussion kind of ferrite can be formed in the microstructure if deformation
occurs at temperatures higher than Ae3 , and also at temperatures
The hot ductility values at different temperatures between 750 higher than Ar3 there can be no pro-eutectoid ferrite formed.
and 950 ◦ C are presented in Fig. 1. Since there were no data points Accordingly, it is seen from Figs. 2 and 3 that Ae3 is between 850
between 750 and 800 ◦ C and between 850 and 900 ◦ C, a dashed and 900 ◦ C and Ar3 is around 825 ◦ C. Using the Andrews’s empirical
line was used to link the data points from 750 to 800 ◦ C and from formula [13], we estimate Ae3 for the experimental 2.25Cr–1Mo
850 to 900 ◦ C. As seen, there is an evident ductility trough in the steel as approximately 885 ◦ C, which is in good agreement with the
range 750–900 ◦ C and the minimum ductility with an RA value of present observation. Therefore, the ductility shelf on the right side
63.3% occurs at 825 ◦ C. Fig. 2 shows typical microstructures after of the trough, as shown in Fig. 1, could start from this temperature.
hot tensile testing at different temperatures. Obviously, there is no As described by Mintz [6], the hot ductility trough may cover the
ferrite formed at 900 ◦ C while there is some at 850 ◦ C (see Fig. 2c and temperature range of about 30 ◦ C below Ar3 to Ae3 , which well
d), indicating that there is a temperature between 850 and 900 ◦ C supports the present results (see Fig. 1).
above which no austenite–ferrite transformation has taken place The magnitude of ferrite in the microstructure is shown in
both during isothermal holding prior to tensile testing and in the Fig. 4 as a function of temperature before and after deformation.
 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732 2727

Fig. 2. Microstructures of the specimens after hot tensile testing at different temperatures: (a) 818 ◦ C, (b) 825 ◦ C, (c) 850 ◦ C, and (d) 900 ◦ C.

The ferrite observed before deformation represents pro-eutectoid pro-eutectoid ferrite. As seen in Fig. 3, the pro-eutectoid ferrite
ferrite and the ferrite observed after deformation involves both is not formed apparently along grain boundaries and therefore in
pro-eutectoid ferrite and deformation-induced ferrite. The differ- the present case the main nucleation sites should be carbides or
ence in quantity between the latter and the former stands for the inclusions. After austenitizing for 15 min at 1000 ◦ C, chromium and
magnitude of deformation-induced ferrite, which is represented molybdenum carbides can fully dissolve in the austenite [15,16].
in Fig. 5 as a function of test temperature. As seen, the mag- However, manganese sulfide (MnS), i.e., inclusions, may exist in
nitude of deformation-induced ferrite increases with decreasing the steel. An equation to describe the solubility product of MnS is
test temperature from Ae3 and attains a maximum value around expressed as [17]
820 ◦ C and then decreases with further decreasing test temper-
ature until 800 ◦ C where there is nearly no deformation-induced −10, 590
log(Ksp ) = log[(Pct Mn) × (Pct S)] = + 4.092 (1)
ferrite present in the microstructure. Comparing Fig. 5 with Fig. 1, T
one can claim that the deformation-induced ferrite plays a primary where Ksp is the solubility product at a given temperature T
role in the appearance of the hot ductility trough. Of course, as (in Kelvin); Pct denotes the weight percent. With this equation,
mentioned above, the morphology of ferrite is also important. The log(Ksp ) at 1000 ◦ C (1273 K) can be calculated as −4.23, while for
net-like structure of ferrite along austenite grain boundaries is the the experimental 2.25Cr–1Mo steel the value of log[(Pct Mn) × (Pct
most detrimental to the hot ductility. S)] is −1.289, which is greater than −4.23. This implies that there are
It is well known [14] that grain boundaries and second phase still some MnS inclusions which would not dissolve in the austen-
particles or inclusions may act as potential nucleation sites for ite after austenitization at 1000 ◦ C. After hot rolling, the inclusion

Fig. 3. Microstructures of the specimens before deformation at (a) 825 ◦ C, and (b) 818 ◦ C.
2728 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732

Fig. 6. Dependence of ultimate strength on temperature.

Fig. 4. The magnitude of ferrite in the microstructure as a function of temperature
before and after deformation.
formed in advance during the holding period before tensile test-
particles are distributed uniformly. This is why the pro-eutectoid ing (see Fig. 3). This is because, in the dual phase region, strain
ferrites formed before tensile testing are isolatedly distributed concentration takes place mainly in the pro-eutectoid ferrite due
throughout the microstructure. to its lower strength and lower work hardening rate [20]. The above
It is demonstrated [18] that free energy change at ␥/␣ phase phenomena may be seen clearly in Fig. 2, which shows that below
transformation in the deformation process consists of free energy Ar3 (∼825 ◦ C) the ferrite in the microstructure is not distributed
change in volume, elastic strain energy change, surface free energy preferentially in the austenite grain boundary regions, but it is at
change and deformation energy. The first three parts are claimed Ar3 or above.
as ‘chemical’ energy which is the driving force for the formation The ultimate strengths at different temperatures are shown in
of pro-eutectoid ferrites, while in the deformed specimens the Fig. 6. From 800 to 750 ◦ C and 950 to 825 ◦ C, the ultimate strength
deformation could increase the driving force of transformation. increases with decreasing temperatures, while it decreases abnor-
At temperatures higher than Ar3 , the chemical energy is not high mally with decreasing temperature from 800 to 825 ◦ C. Normally,
enough to form the pro-eutectoid ferrite [12], so the deformation lower temperature causes higher strength if there is no apparent
energy plays a crucial role in deformation-induced ferrite trans- phase change. From 800 to 750 ◦ C, the dominant phase is ferrite
formation on the grain boundaries [19]. This can be accounted and the volume fraction of ferrite and austenite is not changed a
for by considering that the grain boundary regions are subject lot, and hence the strength of the steel is mainly determined by the
to more restraint than the grain interiors, causing stress concen- strength of ferrite. From 950 to 825 ◦ C, there is also no evident phase
tration, and thus more deformation may occur along the grain transformation (see Fig. 4), and the strength of the steel is mainly
boundaries, enhancing the nucleation of ferrite there. However, at determined by the strength of austenite. Nevertheless, from 825 to
temperatures lower than Ar3 , the pro-eutectoid ferrite had been 800 ◦ C there is intense ferrite transformation (see Fig. 4), and thus
formed during isothermal holding before deformation and the at this stage the transformation of ferrite has a dominant effect on
deformation energy would encourage transformation of austenite the ultimate strength. The abnormal strength decrease from 825 to
to deformation-induced ferrite. In this case, the nucleation sites 800 ◦ C is attributed to the lower strength of ferrite than austenite.
for deformation-induced ferrite should be pro-eutectoid ferrites Based on the above, it is plausible to predict the microstructural
features in the steel at different test temperatures.
In addition, as suggested by Suzuki et al. [20], the work hard-
ening rate of ferrite is lower than that of austenite. So the
deformation-induced ferrite can be formed preferentially on the
pro-eutectoid ferrite in the ferrite/austenite range due to its lower
strength and lower work hardening rate. During tensile testing,
heavy stain concentration may occur in the pro-eutectoid ferrite
phase regions, and thus deformation-induced ferrite nucleates and
grows in these high-deformed regions, resulting in isolatedly dis-
tributed ferrites (see Fig. 2a).
Previous studies [21,22] indicated that the process of dynamic
recovery and recrystallization can affect the hot ductility in the
ferrite/austenite and low austenite region. Fig. 7 shows the mor-
phology of ferrites after deformation at different temperatures. Fine
and equiaxed ferrite grains shown for the specimens deformed at
800 and 825 ◦ C indicate that significant dynamic recrystallization
of ferrite has occurred at temperatures higher than 800 ◦ C, but no
dynamic recrystallization can be seen at 750 ◦ C. For the microstruc-
ture with a large amount of ferrite, e.g., at 800 ◦ C, the dynamic
Fig. 5. The magnitude of deformation-induced ferrite as a function of test temper-
recrystallization of ferrite has a beneficial effect on the hot ductil-
ature. ity due to the release of the deformation energy. However, if the
 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732 2729

Fig. 7. Microstructures showing the ferrite morphology in the specimens after deformation at different temperatures: (a) 750 ◦ C, (b) 800 ◦ C, and (c) 825 ◦ C.

amount of ferrite is small, e.g., at 825 ◦ C or above, the dynamic equilibrium segregation during the testing [3]. Nachtrab and Chou
recrystallization of ferrite would not show an apparent influence [3] found deformation-induced grain boundary segregation of cop-
on the hot ductility (see Fig. 1). Fig. 8 shows the stress–strain curves per, tin and antimony at 900 ◦ C in C–Mn steels. In their experiments,
at different deformation temperatures. When tested at 900 and the samples were solution treated for 60 min at 1200 ◦ C, followed
950 ◦ C, the wavy part of the curves (marked) are caused by dynamic by furnace cooling to 900 ◦ C, and held there for 30 min and then
recrystallization of austenite [19], leading to an excellent RA value quenched directly or after 10% tensile deformation by an argon
(see Fig. 1), while the smooth curves with no wavy part when tested stream. Auger electron spectroscopy measurements demonstrated
at 825 and 850 ◦ C imply that there is no dynamic recrystallization that there is apparent grain boundary segregation of copper, tin
occurring. As a consequence, at 850 ◦ C or below, the hot ductility and antimony for the deformed samples, but there is not for the
has not been affected by dynamic recrystallization of austenite. undeformed samples.
SEM fractographs of the specimens tested at different temper- In 1957, McLean first proposed a model for equilibrium segrega-
atures are given in Fig. 9. As seen, they all exhibit ductile fracture, tion [7]. The maximum equilibrium grain boundary concentration
but the difference in the fracture appearance are evident. With of a solute at a given temperature (T), Ceq , is given by
an excellent hot ductility at 750 or 900 ◦ C, the dimples are quite
Cg exp(Egb /kT )
deep, while for the minimum hot ductility at 825 ◦ C they almost die Ceq = (2)
away and are very shallow. Cracks propagating along the austenite 1 + Cg exp(Egb /kT )
grain boundary can be seen in the microstructure for the specimen where Cg is the solute concentration in the matrix; Egb is the free
deformed at 825 ◦ C (see Fig. 10). As described above, the deforma- energy of segregation of solute and k is the Boltzmann constant.
tion is concentrated on the ferrite in the austenite/ferrite range. As seen from Eq. (2), the equilibrium segregation decreases with
Because of the thin ferrite layer with a net-like structure formed increasing temperature, implying that in the hot ductility trough
along austenite grain boundaries at 825 ◦ C, the cracks may propa- range the level of equilibrium segregation is low.
gate easily in the ferrite layer, leading to shallow dimples on the Recently, based on the vacancy–solute complex effect produced
fracture surfaces. during plastic deformation, Song et al. [27] developed a theoretical
In addition to the thin ferrite layer, segregation of undesir- model for deformation-induced non-equilibrium grain boundary
able elements at austenite grain boundaries can weaken the grain segregation. The maximum segregation concentration produced
boundaries [6,23–26]. According to the chemical composition of the during high temperature plastic deformation at a given strain rate
experimental steel, phosphorus is the main segregating species. (ε̇) and temperature (T), Cbm , is given by
In the present case, two dominant processes of grain bound-
  
ary segregation are equilibrium segregation during isothermal Qsv 2 2 ˝o A1/n Q
holding before tensile testing [7] and deformation-induced non- Cbm = Ceq + Cg 1+ exp − def
Qf Dv B nkT
   
b2 (1−1/n) Cj A1/n Q
× ε̇ + exp − def ε̇(1−2/n)
Qf 4b nkT
 Q 
f
× exp (3)
kT

where Qsv is the vacancy–solute binding energy; Qf is the vacancy

formation energy;  is a structural parameter describing the dis-
tribution of dislocations, which is equal to 1 in the case of a
homogeneous distribution;  is the shear modulus; ˝o is the
atomic volume; A and n are empirical constants; Dv is the diffu-
sion coefficient of vacancies; B is a constant associated with the
vibrational frequency of the atoms around a vacancy; Qdef is the
apparent activation energy for deformation;  is a constant associ-
ated with the mechanical production of jogs; b is the Burgers vector;
 is a parameter describing the neutralization effect produced by
the presence of vacancy emitting and vacancy absorbing jogs; and
Cj is the concentration of thermal jogs which is described by

Ej
Cj = exp − (4)
Fig. 8. Stress–strain curves at different test temperatures. kT
2730 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732

Fig. 10. Microstructure showing cracks propagating along the austenite grain
boundary for the specimen deformed at 825 ◦ C (marked by black arrows).

where Cb (t) is the boundary concentration at time t; Dc is the dif-

fusion coefficient of complexes in the matrix; d is the boundary
thickness, and ˛ is the enrichment ratio, given by ˛ = Cbm /Cg . In Eq.
(5), erfc(x) = 1 − erf(x), where erf(x) is the error function of x.
In the model, the concept of the critical time tc is introduced.
At this critical time, the forward diffusion flux of complexes from
the matrix to the boundary is balanced by the backward diffusion
flux of solute atoms, i.e., the segregation level reaches a stable value
during deformation. tc is expressed as
ˇ2 ln(Dc /Ds )
tc = (6)
4ı(Dc − Ds )
where ˇ is the grain size; Ds is the diffusion coefficient of solute
atoms in the matrix; ı is a dimensionless material constant, termed
the critical time constant.
In light of the model, the boundary concentration of solute
increases with increasing deformation until the critical time. After
that it will remain unchanged, equal to Cb (tc ). This theoretical
model is now used to predict phosphorus grain boundary segrega-
tion during tensile testing. Data used in the calculations are listed
in Table 1. As shown in Fig. 8, the magnitude of homogeneous
deformation is about 20% which is employed in the calculations.
Fig. 11 represents predicted phosphorus grain boundary seg-
regation as a function of temperature during tensile testing at a

Fig. 9. SEM fractographs for the specimens tested at different temperatures: (a)
750 ◦ C, (b) 825 ◦ C, and (c) 900 ◦ C.

where Ej is the jog formation energy. It is assumed in the model that

the maximum equilibrium segregation has been attained before
deformation, which applies to the present situation.
To gain this maximum concentration, a kinetic process is
required. The kinetic equation is described by

 
Cb (t) − Ceq
 4D t  2 Dc t
c
= 1 − exp erfc (5)
Cbm − Ceq ˛2 d2 ˛d Fig. 11. Predicted phosphorus grain boundary segregation during hot tensile testing
at a strain rate of 1 × 10−3 s−1 .
 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732 2731

Table 1
Data used in the segregation calculations.

/J m−3 8.1 × 1010 (1 − 0.91(T − 300)/1810) [28] Qdef /eV 4 [28]
˝o /m3 1.21 × 10−29 [28] b/m 2.58 × 10−10 [28]
 0.5 [28] к 1 [28]
␹ 0.1 [28] Dv /m2 s−1 5 × 10−5 exp(−1.24/kT) [29]
Qf /eV 1.4 [29] Dc /m2 s−1 1.7 × 10−5 exp(−1.63/kT) [29]
Qsv /eV 0.41 [30] Ds /m2 s−1 2.83 × 10−3 exp(−3.03/kT) [29]
Egb /eV 0.33–1.76 × 10−4 T [9] ˇ/m 70
d/nm 1 Ej /eV 0.75 [28]
ı 30 [31] A 1 [23]
B 4 [32] n 4.3 [28]

strain rate of 1 × 10−3 s−1 . Obviously, there is a phosphorus con- rus [5], which could deepen the ductility trough. In addition, it is
centration peak in the range 800–850 ◦ C. It is seen clearly from demonstrated by Yasumoto et al. [11] that the solution treatment at
Figs. 1 and 11 that the hot ductility trough is corresponding to the a lower temperature can make more sulphur atoms associated with
predicted phosphorus segregation peak. Mn atoms, forming randomly dispersed MnS particles and improv-
Phosphorus segregation at grain boundaries reduces the inter- ing the hot ductility. This means that the lower solution treatment
granular cohesive strength of the steel, and the reduction in temperature (1000 ◦ C) used in the present work could shallow the
intergranular fracture energy is in proportion to the segregation hot ductility trough and enlarge the difference between Guo et al.’s
level [23]. So a high segregation level of phosphorus can result in a and our results. The temperature corresponding to the minimum
low grain boundary cohesive strength, accelerating crack propaga- hot ductility is dependent on Ar3 which is affected by cooling rates
tion along the grain boundary, and then reducing the hot ductility [6,35]. Normally, a lower cooling rate leads to a higher Ar3 , which
of the steel. causes a narrower hot ductility trough and shifts the minimum duc-
The ductility trough present in the austenite–ferrite region has tility temperature to a higher temperature. In the present work,
received much attention [5,33]. However, the width and minimum the slow furnace cooling could be responsible for the narrower hot
point of the trough is quite different depending on chemical com- ductility trough and the higher minimum ductility temperature as
position and heat treatment condition. An investigation by Guo compared with those obtained by Guo et al. [5].
et al. [5] showed that there is a wider ductility trough between
700 and 1000 ◦ C for a steel with a similar chemical composition 4. Conclusions
to ours but with a higher phosphorus content (0.06 wt.%P), where
the minimum ductility is located at about 750 ◦ C. A comparison The ferrite concerned in the present work is divided into
of their results with the present ones is shown in Fig. 12. In their pro-eutectoid ferrite and deformation-induced ferrite. The pro-
experiments, an extra procedure of solution treatment at 1320 ◦ C eutectoid ferrite is formed below Ar3 (∼825 ◦ C) and distributed
for 3 min was involved, followed by cooling down to different test uniformly throughout the microstructure. The deformation-
temperatures at a rate of 10 K s−1 , and the specimens were held for induced thin ferrite layer is formed along austenite grain
10 s before tensile testing. According to Eq. (1), log Ksp at 1320 ◦ C is boundaries above Ar3 , while it is nucleated on the pro-eutectoid
−2.56, while the value of log[(Pct Mn) × (Pct S)] is −2.72 for their ferrite below Ar3 , resulting in an isolated distribution. The quan-
steel. Therefore, besides carbides, all MnS inclusions are dissolved tity of deformation-induced ferrite has a primary effect on the hot
in the austenite. It is suggested that precipitation of MnS at all tem- ductility. In general, the hot ductility decreases with increasing
peratures proceeds with an incubation time of 5–30 min [34], which amount of deformation-induced ferrite, leading to the appear-
is much longer than the cooling and holding time used in their ance of a hot ductility trough. There is a peak in the quantity of
experiment. Thus no inclusions left before tensile testing, which deformation-induced ferrite between 800 and 900 ◦ C, which is just
enables both the pro-eutectoid and deformation-induced ferrites to corresponding to the hot ductility trough. The morphology of ferrite
nucleate mainly along austenite grain boundaries, widening the hot is also important. The net-like structure of ferrite along austen-
ductility trough. Also the cooling rate of 10 K s−1 caused apparent ite grain boundaries is the most detrimental to the hot ductility.
thermal non-equilibrium grain boundary segregation of phospho- Dynamic recrystallization of ferrite takes place during deformation
above 800 ◦ C, yet it has no considerable effect on the hot ductility
trough. There is no dynamic recrystallization of austenite occurring
below 850 ◦ C, and hence it has no effect on the hot ductility trough.
In addition, predictions of phosphorus grain boundary segregation
demonstrate that there is a phosphorus concentration peak in the
range 800–850 ◦ C, which is also corresponding to the hot ductility
trough.

Acknowledgement

This work was supported by the National Natural Science Foun-

dation of China under grant no. 50671033.

References

[1] B. Mintz, R. Abushosha, M. Shaker, Mater. Sci. Technol. 9 (1993) 907.

[2] C. Ouchi, K. Matsumoto, Trans. ISIJ 22 (1982) 181.
[3] W.T. Nachtrab, Y.T. Chou, J. Mater. Sci. 19 (1984) 2136.
[4] S. Harada, S. Tanaka, H. Misumi, S. Mizoguchi, H. Horiguch, ISIJ Int. 30 (1990)
Fig. 12. Comparison of Guo et al.’s hot ductility results [5] with the present ones. 310.
2732 X.-M. Chen et al. / Materials Science and Engineering A 527 (2010) 2725–2732

[5] A.-M. Guo, Y.-H. Wang, D.-D. Shen, Z.-X. Yuan, S.-H. Song, Mater. Sci. Technol. [21] B. Mintz, R. Abushosha, J.J. Jonas, ISIJ Int. 32 (1992) 241.
19 (2003) 1553. [22] J. Calvo, J.M. Cabrera, A. Rezaeian, S. Yue, ISIJ Int. 47 (2007) 1518.
[6] B. Mintz, ISIJ Int. 39 (1999) 833. [23] M.P. Seah, Acta Metall. 28 (1980) 955.
[7] D. McLean, Grain Boundaries in Metals, Oxford University Press, London, 1957. [24] W.T. Nachtrab, Y.T. Chou, Metall. Trans. A 17A (1986) 1995.
[8] S.-H. Song, H. Zhuang, J. Wu, L.-Q. Weng, Z.-X. Yuan, T.-H. Xi, Mater. Sci. Eng. A [25] C. Nagasaki, J. Kihara, ISIJ Int. 37 (1997) 523.
486 (2008) 433. [26] S.-H. Song, Z.-X. Yuan, J. Jia, A.-M. Guo, D.-D. Shen, Metall. Mater. Trans. A 34A
[9] H. Erhart, M. Paju, Scr. Metall. 17 (1983) 171. (2003) 1611.
[10] M. Gojic, L. Kosec, ISIJ Int. 37 (1997) 412. [27] S.-H. Song, Q. Zhang, L.-Q. Weng, Mater. Sci. Eng. A 473 (2008) 226.
[11] K. Yasumoto, Y. Maehara, S. Ura, Y. Ohmori, Mater. Sci. Technol. 1 (1985) 111. [28] M. Militzer, W.P. Sun, J.J. Jonas, Acta Metall. Mater. 42 (1994) 133.
[12] L.-X. Du, C.-B. Zhang, H. Ding, X.-H. Liu, G.-D. Wang, ISIJ Int. 42 (2002) 1119. [29] S.-H. Song, L.-Q. Weng, Mater. Sci. Technol. 21 (2005) 305.
[13] K.W. Andrews, J. Iron Steel Inst. Jpn. 203 (1965) 721. [30] R.G. Faulkner, S.-H. Song, P.E. Flewitt, Mater. Sci. Technol. 12 (1996) 904.
[14] S.C. Hong, S.H. Lim, H.S. Hong, K.J. Lee, D.H. Shin, K.S. Lee, Mater. Sci. Eng. A 355 [31] T.-D. Xu, B.-Y. Cheng, Prog. Mater. Sci. 49 (2004) 109.
(2003) 241. [32] D.A. Porter, K.E. Easterling, Phase Transformations in Metals and Alloys, 2nd
[15] H.M. Lee, S.M. Allen, M. Grujicic, Metall. Trans. A 22A (1991) 2869. ed., Chapman & Hall, London, 1992.
[16] D.V. Shtansky, G. Inden, Acta. Mater. 45 (1997) 2879. [33] B. Mintz, S. Yue, J.J. Jonas, Int. Mater. Rev. 36 (1991) 187.
[17] H.A. Wriedt, H. Hu, Metall. Trans. A 7A (7) (1976) 11. [34] D.A. Mirzaev, N.I. Vorob’ev, I.L. Yakovleva, O.K. Tokovoi, D.V. Shaburov, E.A.
[18] Y.Q. Weng, ISIJ Int. 43 (2003) 1675. Fominykh, O.V. Samoilova, Phys. Met. Metall. 103 (2007) 499.
[19] J. Lewis, J.J. Jonas, B. Mintz, ISIJ Int. 38 (1998) 300. [35] S.C. Hong, S.H. Lim, K.J. Lee, D.H. Shin, K.S. Lee, Mater. Sci. Eng. A 355 (2003)
[20] H.G. Suzuki, S. Nishimura, J. Mamura, Y. Nakamura, Trans. ISIJ 24 (1984) 169. 241.