Materials Science & Engineering A 729 (2018) 257–267

Contents lists available at ScienceDirect

Materials Science & Engineering A

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

Effect of microstructures to tensile and impact properties of stir zone on T

9%Cr reduced activation ferritic/martensitic steel friction stir welds
⁎
Chao Zhang, Lei Cui , Dongpo Wang, Yongchang Liu, Huijun Li
Tianjin Key Laboratory of Advanced Joining Technology, School of Materials Science and Engineering, Tianjin University, Tianjin 300354, PR China

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

Keywords: The present study reports the microstructural evolution and mechanical properties of the stir zone (SZ) in friction
Reduced activation ferritic/martensitic steel stir welding (FSW) reduced activation ferritic/martensitic (RAFM) steel. The effects of the grain size, retained
Friction stir welding austenite (RA), and precipitates on the SZ tensile properties and impact toughness are discussed. The SZ of the
Stir zone weld exhibits a good balance of strength and ductility at 600 °C temperatures and a slightly decrease of impact
Microstructures
toughness comparing with the base metal (BM). The RA plays a critical role in effecting the ductility and impact
Tensile properties
Impact toughness
toughness of SZ. The influence of RA to the ductility of SZ becomes more effective as the carbon concentration of
RA is increased to 1% or above where the thermal and mechanical stabilities are improved. The highly dispersed
and refined MX carbonitrides and M3C carbides during FSW are good for the ductility and impact toughness of
SZ as well. It indicates the SZ of FSW RAFM steel weld has the potential to have good mechanical properties.

1. Introduction microstructural evolution and mechanical properties of similar steel

Friction stir welding (FSW) process was invented by the welding
institute (TWI) in 1991. The process principle of FSW includes frictional
heating and severe plastic flow with a low welding peak temperature.
During FSW, the function of stirring tool probe is producing the severe
plastic deformation (SPD) for the materials in stir zone (SZ) under high
temperatures and high strain rates. Therefore, FSW could not deterio-
rate seriously the microstructure and mechanical properties of the raw
material [1–3]. Such process could result in the occurrence of dynamic
recrystallization, and further grain refinement and dislocation multi-
plication. Thus, FSW process has been used for modifying the micro-
structure for several alloys in some specific engineering fields to im-
prove the mechanical properties of the materials [4–6].
Recently, martensitic heat-resistant steels with high chromium
content (9–12 wt%) were developed to achieve high creep properties at
elevated temperatures and regarded as candidate structure materials in
fossil fuel and power industry applications [7–11]. However, owing to
the effect of welding thermal cycle, conventional fusion welding pro-
cesses always lead to recrystallization, precipitate coarsening and dis-
solution in joint regions which can rapidly lose strength in high tem-
perature and stress service conditions [12,13]. Reduced activation
ferritic/martensitic (RAFM) steel with tempered martensite micro-
structure exhibits excellent high temperature mechanical properties
and then will potentially be used in power plants at elevated tem-
peratures [14–18]. At present, many studies have reported
under different condition, such as the effect of additional alloy element
and heat treatment [19–21]. Furthermore, it is also essential to in-
vestigate the high-temperature tensile properties and impact toughness
for FSW RAFM steel joints, which will provide important data to im-
prove the reliability of RAFM steel for future applications.
However, limited works have been done on the high-temperature
tensile properties and impact toughness of SZ for FSW RAFM steel
joints. Therefore, in this paper, we use a kind of 9% Cr RAFM steel as
the BM and carrying out FSW experiments with optimized welding
parameters. Based on these, the characteristics of microstructures in-
cluding the retained austenite (RA), grain size, and precipitates are
examined. The effects of the above microstructural respects on high-
temperature tensile properties and impact toughness of SZ are dis-
cussed.
2. Experimental procedures
In this study, the base metal (BM) is selected as a 9%Cr RAFM steel
with the chemicals as shown in Table 1. The as received BM is a 5 mm
thick plate processed by hot rolling, normalizing and then tempering. In
the alloying system, the C, N and Mn are used for stabilizing the aus-
tenite, the Cr, V, W, Ta are the solution and precipitation alloy elements
[22]. The addition of Cr can improve oxidation and corrosion re-
sistance, high temperature strength and generate precipitate phase of
M23C6. The V and Ta may react with C and N to precipitate nanometer

⁎
Corresponding author.
E-mail address: leicui@tju.edu.cn (L. Cui).

https://doi.org/10.1016/j.msea.2018.05.043
Received 28 January 2018; Received in revised form 9 April 2018; Accepted 12 May 2018
Available online 14 May 2018
0921-5093/ © 2018 Elsevier B.V. All rights reserved.
C. Zhang et al. Materials Science & Engineering A 729 (2018) 257–267

Table 1 ∂ρuj ∂ρuj ∂P ∂ ⎛ ∂μuj ⎞ ∂uj

Chemical composition of the experimental steel.
+ ui =− + ⎜ ⎟ − ρU1
∂t ∂x i ∂x j ∂x i ⎝ ∂x i ⎠ ∂x1 (1)
C Cr Mn V W Ta Si Zr N S P Fe
where ρ, P, μ, and U1 are the density, the pressure, the non-Newtonian
0.1 9 0.5 0.2 1.5 0.15 0.05 0.005 0.007 ＜0.002 ＜0.002 Bal viscosity and the welding velocity, respectively. Shi et al. [24] and
Hasan et al. [25] found that the viscosity is closely related to the
temperature and the strain rate for the non-Newtonian fluid. The
viscosity can be calculated using the following equation:
σ
μ=
MX carbo-nitrides, which pins the mobile dislocations, stabilizes mar- 3ε ̇ (2)
tensitic microstructures and enhances creep rupture strength and creep
where σ and ε ̇ are flow stress and is the effective strain rate, respec-
life.
tively. The Zener–Hollomon field variable contains of the effect of the
The FSW experiments were carried out on a FSW-3LM-020 welding
temperature and strain rate on the flow stress in the following equation
machine. The welding tool is made of W-25%Re alloy and contains a
to define σ:
shoulder of 18 mm diameter and a 4.7 mm length tapered probe with
inverted spin threads. During the FSW experiment, the rotational Q
Z = ε ė xp ( )
speeds were selected of 200 rpm, 300 rpm and 400 rpm. The tool travel RT (3)
speed, tool tilt angle, and axial load were universe of 60 mm/min, 2.5°,
1 Z
and 10 kN, respectively. After welding, the SZs in welds with different σ= sinh−1 ⎡ ( )1/ n⎤
α ⎣ A ⎦ (4)
parameters were cut for microstructural observations, tensile and im-
pact tests. For convenience of discussion, the SZ samples were named where A, α and n are material constants and Z is the Zener–Hollomon
for SZ-200 (200 rpm), SZ-300 (300 rpm) and SZ-400 (400 rpm), re- parameter. A, α and n are obtained from open literature [26], as shown
spectively. in Table 3. Q, T, R and ε ̇ were the Activation energy, temperature, gas
The SZ samples were mechanically ground and polished, and then constant and strain rate, respectively. The thermal energy conservation
etched with a solution of 5 g FeCl3, 20 mL HCl and 100 mL water for equation can be written as:
90 s. A JEOL-7800 field emission scanning electron microscope (SEM)
∂ (ρCp T ) ∂ ∂T ∂ ⎛ ∂T ⎞
was used for briefly observations of grain structure and distribution of + ui (ρCp T ) = −ρCp U1 + ⎜k ⎟ + Si + Sb
precipitates for all SZ specimens. Synchrotron X-ray diffraction (XRD) ∂t ∂x i ∂x1 ∂x i ⎝ ∂x i ⎠ (5)
was used for detecting the volume fraction of RA in SZ specimens, with where Cp, k, Si, and Sb is the specific heat, the thermal conductivity, the
a scanning step size of 0.01° per second and the 2θ range from 40° to source term and the heat generation rate. The thermal conductivity and
120°. Transmission electron microscope (TEM) experiments were pre- specific heat capacity of the RAFM steel were provided by Lei Yu-cheng
formed using Tecnai G2F30 microscope that equipped with an energy [27] and shown in the Table 2.
dispersive spectrum (EDS) to explore the characteristics of grain
structures, RA, precipitates and local chemical compositions in the SZ 3.2. Heat source mode
samples. The tensile tests were then carried out at 600 °C with constant
strain rate of 0.05 s−1. The Charpy V-notch impact tests were carried The material was assumed to be in partial sticking with the tool,
out at − 20 °C. Details of cutting locations and dimensions of the tensile which rotated at the experimental rotation speed. The heat generated at
and impact specimens of SZ are shown in Fig. 1. the interface between the vertical and horizontal surfaces of the tool
and the base metal, Si, may be defined as [28]:

3. Mathematical model Ar
Si = [(1−δ ) ητ + δμf PN ](ωr − U1sinθ)
V (6)
3.1. Governing equations where Ar is any small area on the tool–base metal interface. r is the
radial distance of the center of the area from the axis. V is the control-
The conservation of the momentum equation in index form is ex- volume enclosing the area. θ is the angle with the negative x-axis in the
pressed as follows [23]: counter-clockwise direction. η is the mechanical efficiency which is
assumed to be 0.5 guided by the range of the previous work in FSW

Fig. 1. Dimensions and positions of tensile and impact toughness testing samples in the weld.

258
C. Zhang et al. Materials Science & Engineering A 729 (2018) 257–267

Table 2
Thermo-physical properties of the RAFM steel [27].
Temperature/K 298 373 473 573 673 773 873

−1 −1
Specific heat capacity/J kg K 430 478 520 561 625 689 776
Thermal conductivity/W m−1 K−1 26.3 26.2 26.2 25.8 25.6 25.0 24.6

Table 3
Properties and parameters used in the simulation.
Properties and Parameters Value

Density ρ/kg m−3 7800

Stefan-Boltzmann constant σb/J K−4 5.67 × 10−8
m−2 s−1
Emissivity ε 0.35
Ambient temperature Ta/K 300
Constant A/s−1 3.1014 × 107
Constant α/Pa−1 1.07 + 1.70 × 10−4T − 2.81 × 10−7T Fig. 3. Variations of the computed peak temperature in SZ-200, SZ-300 and SZ-
Constant n 0.2 + 3.966 × 10−4T 400.
Gas constant R/J mol−1 K−1 8.314
Activation energy Q/J mol−1 471,205
3.3. Boundary conditions

process [28,29]. δ is the fractional slip between the tool and the base In this model, the heat flux was added to FLUENT by the user-de-
metal interface. μf represents the coefficient of friction. ω is the angular fined-functions (UDF). The walls of the filled hole were set as the
velocity of tool and PN is the normal pressure on the surface. The ve- coupling walls and the other walls were all assumed to be the no-slip
locity (ωr −U1sinθ ) represents the local velocity of a point on the tool boundary condition. The simulation was started when the tool reached
with the origin fixed at the tool axis. a valued of 4.7 mm. Before this, a temperature field had formed. This
The heat generated at the interface between the vertical and hor- temperature field obtained from a steady calculation was considered as
izontal surfaces of the tool and the base metal, Si, may be defined as an initialization for later calculations.
heat flux, q1, and can be expressed as: Heat exchange occurred both at the top and at the side surfaces of
the base metal outlet. The thermal boundary conditions including the
q1 = [(1−δ ) ητ + δμf PN ](ωr − U1sinθ) (in the range of r < Rp) (7) heat convection and radiation were expressed by the following equa-
tions:
where RP is the pin radius. At the shoulder–base metal interface, a
∂T
significant amount of heat will be transported to the tool material. −k |top = σε (T 4 − Ta4 ) + ht (T − Ta)
∂z (9)
Thus, the rate of heat generation due to friction and plastic work at the
shoulder–base metal interface, q2, may be written as:
∂T
JB −k |side = σε (T 4 − Ta4 ) + hs (T − Ta)
q2 = (in the range of Rp < r < Rs ) ∂z (10)
JB + JT (8)
where σ is the Stefan–Boltzmann constant, ε is the external emissivity,
where J represents the transported extent of heat generated at the as shown in Table 3, and hs and ht are the heat transfer coefficient of
shoulder–base metal interface or the effusivity of the material. The 30 W m−2 K−1. At the bottom surface, the heat transfer coefficient hb of
subscripts B and T denote the base metal and the tool, respectively. 50 W m−2 K−1 was used:

Fig. 2. The computed temperature profiles on the top surface of the workpiece under different conditions: (a) SZ-200, (b) SZ-300 and (c) SZ-400.

259
C. Zhang et al. Materials Science & Engineering A 729 (2018) 257–267

Fig. 4. The SEM observations of microstructure in BM and SZ: (a)-(d) BM, SZ-200, SZ-300 and SZ-400, respectively.

Fig. 5. The TEM observations of microstructure in BM and SZ: (a)-(d) are for BM, SZ-200, SZ-300 and SZ-400, respectively.

260
C. Zhang et al.
Fig. 6. The martensite lath width and prior austenite size observed in BM, SZ-
200, SZ-300 and SZ-400.
Fig. 7. The XRD patterns for samples BM, SZ-200, SZ-300 and SZ-400.
∂T
k |bottom = hb (T − Ta)
∂z
3.4. Model description
Thermomechanical simulation of the FSW process for the RAFM
steel sheet was performed using the computational fluid dynamics
(CFD) code, Fluent based on the governing equations and the boundary
conditions as mentioned above. Conservation principles of momentum
and energy were calculated considering the continuity equation for the
incompressible single-phase flow.
The whole process is composed of a rotating tool and moving base
metal. If it is assumed that the pin diameter and the pin length used in
the numerical model are equal to the pin diameter of tool and the
plunging depth, respectively, the shoulder diameter could be de-
termined from both the cross-sectional macrograph of the welded
sample [30] and the plunging depth. Thus, the shoulder diameter and
the pin length of the simplified cylindrical tool used in the numerical
studies are 18 mm and 4.7 mm, respectively. The pin exhibits conical
shape. The welding conditions are the same as the experimental con-
ditions.
Materials Science & Engineering A 729 (2018) 257–267
4. Results
4.1. Analysis of simulation results of temperature field
The computed temperature profiles on the top surface of the
workpiece under different conditions are shown in Fig. 2. The tem-
perature profiles on the top surface of the workpiece are compressed in
the front of the tool and are expanded behind the tool. The computed
results demonstrate that heat is transported rapidly to the cold region
ahead of the tool while heat is transported at a slower rate to the weld
already welded behind the tool. Moreover, the temperature profiles
exhibit asymmetry, which is that the temperature in advancing side
(AS) is always higher than that in retreat side (RS). This asymmetry
becomes more prominent at high welding rotational speeds. It indicates
that the welding rotational speed is strongly related to the temperature
distributions.
From Fig. 2, we can obtain the relationship of peak temperature in
the centerline and welding rotational speed, as shown in Fig. 3. The
peak temperature of friction stir welding SZ under 200 rpm, 300 rpm
and 400 rpm is 1108 K, 1152 K and 1199 K, respectively. The calcula-
tion results is agree with the ones from relevant literatures [31,32]. It
can be found that the peak temperature is increasing with the welding
rotational speed in the range of 200–400 rpm, due to the increase of
heat input. Hoon-Hwe Cho et al. [33] analyzes temperature profiles of
FSW to obtain the peak temperature and the effect of welding traveling
speed on the computed temperature profiles using this similar mathe-
matical model.
4.2. Evolution of grain structures
Fig. 4 show the SEM morphologies of grain structures for BM and
the stir zones obtained with different welding parameters. Comparing
with BM, the prior austenite grain in SZ appears to be significantly
refined. The most obvious refining of prior austenite grain is found in
sample SZ-200, which is welded with 200 rpm rotational speed. The
average size of prior austenite grain in SZ-200 is below 5 µm. As shown
in Fig. 4b~d and Fig. 6, the prior austenite grain in SZ is coarsening in
range of 4.1–11 µm with the increase of the rotational speed.
The higher magnifications under TEM for the corresponding loca-
tions in Fig. 4 are shown Fig. 5. It is found that, in SZ of the joint, the
initial tempered marteniste of BM transformed to lath martensite after
(11)
FSW. At the same time, a lot of M23C6 carbides in BM initially dis-
tributed along the prior austenite grain boundaries and interfaces of
martensitic laths could no longer be observed in SZ samples. In sample
SZ-200, the microstructure is composed of lath martensite and a little
ferrite. When the rotational speed increases up to 300 rpm, the micro-
structure of SZ is more likely to be of fully lath martentise. The increase
of rotational speed may also increase the width of martensite in SZ
samples. The estimated date from TEM observations indicate the
average width of martensite increases in the range of 210–380 nm, as
the rotational speed is increased from 200 rpm to 400 rpm.
The grain refinement of SZ for such RAFM steel is associated with
the dynamic recrystallization and very limiting grain growth caused by
SPD and rapid cooling rate during FSW process. There might be two
pathways in performing the recrystallization of the material in SZ
during FSW. One was proposed as geometric dynamic recrystallization
generated by the fragmentation of initial grains. The other was con-
tinuous dynamic recrystallization performed by the transformation of
low angle boundaries into high angle boundaries [34]. Since the cooling
rate in SZ during FSW for steels may be higher than 100 K/s, the chance
to austenite grain growth might be highly reduced. It can promote the
refinement for both prior austenite grains and martensite laths. In ad-
dition, owing to the fact that nucleation and growth of martensite laths
could not break through the prior austenite grain boundaries in cooling
process, refining the prior austenite grain may also contribute to the
refinement of martensitic laths. The above two aspects are believed as
261
C. Zhang et al. Materials Science & Engineering A 729 (2018) 257–267

Fig. 8. TEM observations of retained austenite (RA) in BM and SZ samples in bright-field and dark-field modes: (a) the morphology of RA in BM, (b) the corre-
sponding diffraction pattern of RA, (c)-(d) for the RA in sample SZ-200, (e)-(f) for the RA in sample SZ-300, and (g)-(h) for the RA in SZ-400, respectively.

the main reasons for grain refining in SZ when FSW for RAFM steels.
The literatures proposed that, when using different welding para-
meters, the peak temperature during FSW for steels may be in a very
wide range from lower than the AC1 point to higher than the AC3 point
[35–37]. For ferritic steels, an extremely low welding peak temperature
below the AC1 point, can be achieved by using a polycrystalline cubic
boron nitride tool combined with very low rotational speed. In such
situation, the steels could be welded without any phase transformations
by means of the absence of transformations from α to γ in heating stage
and from γ to α in cooling stage. It refers to that when the peak tem-
perature is below AC1, the microstructural evolution in SZ merely
involves the recrystallization of ferrite, but non allotropy transforma-
tions of iron [37]. The previous works and simulation results proved
that when FSW for RAFM steel with a W-Re tool and a relatively low
rotational speed (200 rpm in this study), the peak temperature in SZ is
more likely to be in a temperature range of AC1~AC3. When applying
higher rotational speed (300 rpm or above), the peak temperature in SZ
may exceed the AC3 point easily [38]. This is believed as the main
reason for the difference on grain structures of SZ in different welding
parameters.

262
C. Zhang et al.
Fig. 9. The volume fraction of RA and their carbon content for BM, SZ-200, SZ-
300 and SZ-400 samples.
4.3. Characteristics of retained austenite
Fig. 7 show the XRD patterns for BM and SZ samples. It involves
diffraction peaks of α-Fe from the crystal faces of (110), (200), (211),
(220) and (310) and of γ-Fe from the crystal faces of (111) and (200) in
the scanning range of 2θ = 40–120°. This proves that, besides mar-
tensite and ferrite, there might be some of the retained austenite (RA) in
BM and SZ samples. The austenite diffraction peaks appear mainly at
when 2θ in ranged in 40–50°. And the γ-Fe diffraction peaks are de-
tected as more significant in SZ samples than those in BM, showing that
the fraction of RA is increased in SZ after FSW process. The details
under TEM observations of RA in BM and SZ are shown in Fig. 8. The
selected area electron diffraction (SAED) results of reveal that BM in-
cludes a little RA characterizing by “film-like” shaped between the
tempered martensite. In SZ samples, in Fig. 8c~h, an interlayered mi-
crostructure consists of martensite lath and RA could be observed. Most
of the RA in SZ are “needle-like” or “film-like” shaped and in a width
ranged in 30–100 nm.
Fig. 9 show the variation trend of RA volume fraction and its carbon
content in SZ with varying rotational speed. It is found that the volume
fraction of RA can be much higher in SZ than that in BM. The highest
volume fraction of RA is obtained for ~6.49% in SZ-200 sample, where
a relatively low rotational speed of 200 rpm is used. As the rotational
speed is increased to 300–400 rpm, the volume fraction of RA in SZ
exhibits a trend of decrease to ~4.12% and ~3.61% in SZ-300 and SZ-
400, respectively. It is also observed that the volume fraction of RA in
BM is the lowest among all the examined samples. Fig. 9 also reveal the
variation carbon content of RA with the variation of welding condi-
tions. To all SZ samples, the carbon content in RA is found as increased
with the increase of rotational speed, showing an opposite variation
trend of carbon content to the variation of RA volume fraction with the
variation of rotational speed.
Similar results were reported by Miura and Fujii [39,40], revealing
that the severe SPD of SZ during FSW could stabilize the RA in steels.
The main reasons were believed as the fine-grained prior austenite
causes the increase of high angle grain boundaries as barricades for
martensite growth. According to the Koistenen and Marburger [41], the
fraction of transformed martensite during cooling or quenching is
highly dependent on the temperature to be quenched. The general rule
is the lower the quenching temperature is, the higher fraction of
transformed matensite and certainly the fewer RA. Obviously, when all
the SZs were proposed to be cooled to room temperature, there should
be very limited chances for the variation of RA fraction in SZ by
changing the FSW parameters. Therefore, the difference of RA fraction
263
Materials Science & Engineering A 729 (2018) 257–267
in SZ might be caused by the SPD process, the peak temperature and
cooling rate in different FSW parameters. As the hot plastic deformation
and peak temperature are enhanced with the increase of rotational
speed, the driving force of martensitic transformation can be improved
accordingly. This may result in the reduction of RA fraction and in-
crease of transformed martensite. When FSW with higher rotational
speed, the holding time at high temperature of SZ could also be ex-
tended. In this case, the time for carbon diffusion from martensite to
austenite might be more adequate [42,43]. Therefore, to SZ samples,
the carbon content of RA might be increased with the increase of ro-
tational speed.
4.4. Precipitates evolution
Fig. 10a&b show the main precipitates observed in BM. According to
the SAED analysis results, these spherical particles distributing at grain
boundaries and within grains are M23C6 carbides and MX carbonitrides,
respectively. As shown in Fig. 10c~f, the M23C6 precipitates could
merely be found in sample SZ-200 (Fig. 10d). However, the MX pre-
cipitates could be observed in all the samples. As shown in Fig. 10c, the
MX particles in SZ are much finer than that in BM. As shown in
Fig. 10d~f, a lot of needle-like precipitates could be observed in all the
SZ samples. The EDS pattern as shown in Fig. 10g suggests that these
needle-like precipitates are of M3C carbides ((Fe, Cr)3C), a kind of
metastable precipitation in similar steels. The M3C carbides are more
likely to be formed during air cooling process and the initial stage of
tempering for high Cr steels. They might transform to M23C6 carbides
with a transformation sequence of (Fe, Cr)3C → Cr7C3 → Cr23C6 in
tempering process.
As proposed in previous sections, the peak temperature in SZ during
FSW might be near or above the AC3 point of the experimental steel
[44,45]. In most cases of this study, the peak temperature of SZ may
reach or exceed 850 °C, which is the complete dissolution temperature
of M23C6 carbides for RAFM steels. Therefore, austenite recrystalliza-
tion and M23C6 carbides dissolution may occur simultaneously in SZ
during FSW process. Owing that the dissolution temperature of MX is
much higher than the peak temperature of SZ, the variation for the
characteristics of MX phase is found as rare with varying FSW para-
meters. However, the variation of rotational speed significantly influ-
ences the size of M3C particles in SZ samples. A general trend observed
in this study is that the M3C carbides within marteniste would become
larger when the rotational speed is increased.
4.5. Tensile properties at 600 °C
Fig. 11a&b show the stress-strain curves and the comparison results
for the tensile properties of BM and SZ samples tested at 600 °C. The
yield strength (YS) and ultimate tensile strength (UTS) of BM are in
average of 327 ± 12 MPa and 401 ± 21 MPa, and the elongation after
fracture is in average of 25.5 ± 0.3%. To all welding parameters, the
SZ samples show much higher YS and UTS, but lower elongation than
that of BM. The sample SZ-200 has the lowest YS, UTS, and elongation
among the three series of SZ samples. The increase of rotational speed
results in the increase for both strength and ductility in SZ of the weld.
The most outstanding tensile results from this study are found in sample
SZ-300, which involves 433 MPa YS, 643 MPa UTS, and 15.5% elon-
gation.
4.6. Impact toughness
Fig. 12 illustrate the − 20 °C Charpy V-notch impact testing results
for BM and SZ samples in 3.3 mm thick impact specimens. It is found
that the impact absorbed energy of BM (34.54 J) is the highest among
all the specimens. The impact toughness of SZ is lower than that in BM.
The highest SZ impact toughness value of 32.75 J is obtained when
using 200 rpm rotational speed. With the increase of rotational speed,
C. Zhang et al.
Fig. 10. The TEM observations of precipitates: (a)&(b) M23C6 and MX in BM, (c) MX in SZ-300, (d)-(f) M3C in SZ-200, SZ-300 and SZ-400, respectively, and (g) EDS
result of M3C.
the impact toughness might be decreased. When FSW with 400 rpm
rotational speed, the impact toughness of SZ reaches the minimum of
26.95 J. The above results indicate that the FSW process can reduce the
SZ impact toughness to 90–95% of the BM when using a rotational
speed lower than 300 rpm. A further increase of rotational speed to
400 rpm or higher may lead to an obvious reduction of impact tough-
ness of SZ.
5. Discussion
Table 4 summarizes the main results for tensile and impact prop-
erties and the key microstructural aspects of the SZ samples obtained in
this study. Clearly, the obtained SZ properties show obvious correla-
tions with the microstructural aspects, including the amount of RA,
prior austenite size and martensite size. It is found that, the decrease of
RA fraction is SZ results in a simultaneous increase of high temperature
strength and ductility, but a decrease of impact toughness. Moreover,
the increase of the sizes for martensitic laths and prior austenite grains
264
Materials Science & Engineering A 729 (2018) 257–267
may also lead to an increase of high temperature strength whereas a
decrease of impact absorbed energy.
The improved high temperature strength of SZ is associated with the
martensitic transformation during FSW. As is revealed in Fig. 5, instead
of tempered martensite in BM, the microstructure in SZ might transform
to lath martensite substantially after FSW (a volume fraction more than
93%). Because the lath martensite is normally harder than tempered
martensite, the SZ could have much higher strength than BM. The in-
crease trend of high temperature strength of SZ samples is also close
related to the fraction of transformed martensite caused by the increase
of rotational speed. As discussed above, increasing the rotational speed
of FSW may increase the welding peak temperature and cooling rate in
SZ. As a result, the difference value between the initial cooling tem-
perature and the Ms point is enlarged (T0-Ms), which could enhance the
driving force of martensitic transformation. Therefore, when the vo-
lume fraction of transformed martensite and internal stress caused by
martensitic transformation and growth are increased, the strength of SZ
is consequentially enhanced. In addition, as the fraction of transformed
C. Zhang et al. Materials Science & Engineering A 729 (2018) 257–267

Fig. 11. (a) Stress-strain curves and (b) correlation data of tensile properties correlation data test at 600 °C for samples BM, SZ-200, SZ-300 and SZ-400.

plastically and transform to martensite, therefore maintaining a high

ductility without any loss of strength. To enhancing the thermal and
mechanical stability of RA, partitioning the carbon between martensite
and RA is essential. The increase of the carbon concentration can im-
prove the strength, thermal stability and mechanical stability of RA
[43,48,49]. At such a high temperature above MS, the chance for RA to
martensite transformation can be rare. However, since the thermal
stability and strength of RA are improved by the increase of carbon
concentration, both strength and ductility could be enhanced simulta-
neously.
The variation for the strength and ductility to SZ samples observed
in the present study indicates the important role of carbon content in
RA to effect the high temperature tensile properties. As shown in Fig. 9
and Table 4, in SZ of the FSW joint, the carbon concentration in RA
varies from 0.41% to 1.06% with the increase of rotational speed. To
sample SZ-200, where the carbon content of RA is nearly of 0.41%, the
RA is more likely to transform into martensite during high temperature
Fig. 12. The impact absorbed energies of BM, SZ-200, SZ-300 and SZ-400 at tensile deformation owing to the low thermal and mechanical stabi-
− 20 °C. lities. Since the carbon concentration of RA is increased to 1% or above,
in samples SZ-300 and SZ-400, the increased stability of RA starts to
Table 4 playing an effective role to enhance both strength and ductility and
The main mechanical properties and microstructural characteristics for SZ opposes the effect of the reduction of RA fraction.
samples. Another reason for enhancing the high temperature strength of SZ is
by the precipitates. As is shown in Fig. 10, although the grain boundary
Sample Number SZ-200 SZ-300 SZ-400
strengthening particles M23C6 in SZ have been dissolved during FSW
The RA volume fraction (%) 6.49 4.12 3.61 process, the highly dispersed nano-sized MX carbonitrides and the
Carbon concentration in RA (%) 0.41 0.98 1.06 newly and vastly precipitated M3C carbides could strengthen the SZ
Martensite width (nm) 210 310 380
materials at high temperatures. Combined with the strengthening effect
Prior austenite diameter (μm) 4.1 5.6 10.5
Yield strength (MPa) 386 433 439 of grain refinement and lath martensite, a high strength of SZ material
Ultimate tensile strength (MPa) 544 643 636 is easily to be obtained. Meanwhile, benefit by plasticity induced by γ→
The total elongation (%) 10 15.5 13.5 M transformation during deformation, a good balance of strength and
Impact absorbed energy (J) 32.75 31.1 26.95 ductility can be achieved. These fine precipitates are also helpful to
improve the impact toughness of SZ samples.
On the basic of the above results and discussions, we found that the
martensite and the grain size in SZ are increased, the impact toughness
tensile and impact properties of SZ for the FSW RAFM steel joint
of SZ could be decreased accordingly.
without any post welding heat treatment (PWHT) may reach a similar
One interesting phenomenon observed in tensile test is the variation
to the weld metal in fusion-welded joints after PWHT [50–53]. Fur-
of ductility of the SZ samples. As illustrated in Table 4, the elongation of
thermore, the microstructures and mechanical properties could be
SZ is enhanced with the decrease of RA fraction, a rule betraying the
controlled easily by optimizing the FSW parameters. It indicates FSW
normal relationship between the ductility and RA fraction in steels that
technology is promising to be used in welding RAFM and similar steels
the increase of RA fraction increases the ductility. This is mainly be-
for the applications in fusion component fabrications. One thing to be
cause of the strength and stability of RA. For steels containing mar-
noted is that creep resistance of the FSW joint for this kind of steel is
tensite and RA, such as Q&P and TRIP steels, the RA plays an important
unknown. This may be a very important respect to be examined in the
role for the ductility when plastic deformation is performed at room
future works.
temperature [46,47]. During tensile, the RA in the steel might deform

265
C. Zhang et al.
6. Conclusions
In this study, the microstructural evolution and mechanical prop-
erties of SZ in FSW of RAFM steel were investigated. The main results
are as follows:
(1) The SZ of the RAFM steel FSW joint exhibits a good balance of
strength and ductility at high temperatures. In optimized para-
meters, to sample SZ-300, the YS and UTS are nearly 50 MPa and
100 MPa higher than the BM. The elongation of SZ-300 is 15.5%,
nearly 10% lower than BM. Since the fraction of transformed
martensite and grain size are increased with the increase of the
rotational speed, the strength of SZ is increased accordingly.
(2) The impact toughness of SZ is slightly lower than that of BM, and
may show a decrease trend with the increasing of rotational speed.
The main reasons for such variation of impact toughness are be-
lieved as the decrease of RA fraction and the grain coarsening as the
peak temperature is increased with the increase of rotational speed.
(3) The RA plays a critical role in effecting the ductility of SZ. The
influence of RA to the ductility of SZ becomes more effective as the
carbon concentration of RA is increased to 1% or above where the
thermal and mechanical stabilities are improved. Therefore, al-
though the fraction of RA is decreased and the grain size is in-
creased with increasing the rotational speed, the ductility of SZ is
improved.
(4) The highly dispersed strengthening phases including MX carboni-
trides and M3C carbides are also good for the ductility and impact
toughness of SZ. The FSW process makes the MX particles finer and
more evenly distributing within the grains. However, the needle-
like M3C particles might be larger when the rotational speed is in-
creased. The coarsening of M3C phase is believed as another reason
for the increase of strength and decrease of impact toughness of SZ.
Acknowledgements
The authors were supported by the National Natural Science
Foundation of China (Granted Nos. U1660201 and 517011741) and the
National Magnetic Confinement Fusion Energy Research Project
(Granted No. 2015GB119001) for grant and financial support.
References
[1] R.S. Mishra, M.W. Mahoney, Friction stir processing: a new grain refinement
technique to achieve high strain rate superplasticity in commercial alloys, Mater.
Sci. Forum 357–359 (2001) 507–514.
[2] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng. R 50
(2005) 1–78.
[3] S.M. Aktarer, D.M. Sekban, O. Saray, T. Kucukomeroglu, Z.Y. Ma, G. Purcek, Effect
of two-pass friction stir processing on the microstructure and mechanical properties
of as-cast binary Al–12Si alloy, Mater. Sci. Eng. A 636 (2015) 311–319.
[4] D.H. Choi, B.W. Ahn, Y.M. Yeon, S.B. Jung, Microstructures and mechanical
properties of friction stir welded 2.25 Cr–1Mo steel, Mater. Trans. 53 (2012)
1022–1025.
[5] M. Pourali, A. Abdollah-Zadeh, T. Saeid, F. Kargar, Influence of welding parameters
on intermetallic compounds formation in dissimilar steel/aluminum friction stir
welds, J. Alloy. Compd. 715 (2017) 1–8.
[6] B. Cai, Z.Q. Zheng, D.Q. He, S.C. Li, H.P. Li, Friction stir weld of 2060 Al-Cu-Li
alloy: microstructure and mechanical properties, J. Alloy. Compd. 649 (2015)
19–27.
[7] F. Liu, D.H.R. Fors, A. Golpayegani, H. Andren, G. Wahnstrom, Effect of boron on
carbide coarsening at 873 K (600 °C) in 9 to 12 pct chromium steels, Metall. Mater.
Trans. A 43 (2012) 4053–4062.
[8] S.S. Wang, D.L. Peng, L. Chang, X.D. Hui, Enhanced mechanical properties induced
by refined heat treatment for 9Cr–0.5 Mo–1.8 W martensitic heat resistant steel,
Mater. Des. 50 (2013) 174–180.
[9] F. Abe, M. Tabuchi, S. Tsukamoto, Mechanisms for boron effect on microstructure
and creep strength of ferritic power plant steels, Energy Mater. 4 (2012) 166–174.
[10] L. Milovic, T. Vuherer, I. Blacic, M. Vrhovac, M. Stankovic, Microstructures and
mechanical properties of creep resistant steel for application at elevated tempera-
tures, Mater. Des. 46 (2013) 660–667.
[11] F.S. Yin, W.S. Jung, Nanosized MX precipitates in ultra-low-carbon ferritic/mar-
tensitic heat-resistant steels, Metall. Mater. Trans. A 40 (2009) 302–309.
[12] S. Kumar, R. Awasthi, C.S. Viswanadham, K. Bhanumurthy, G.K. Dey, Thermo-
266
Materials Science & Engineering A 729 (2018) 257–267
metallurgical and thermo-mechanical computations for laser welded joint in
9Cr–1Mo(V, Nb) ferritic/martensitic steel, Mater. Des. 59 (2014) 211–220.
[13] B. Arivazhagan, G. Srinivasan, S.K. Albert, A.K. Bhaduri, A study on influence of
heat input variation on microstructure of reduced activation ferritic martensitic
steel weld metal produced by GTAW process, Fus. Eng. Des. 86 (2011) 192–197.
[14] J. Henry, S.A. Maloy, 9-Irradiation-resistant ferritic and martensitic steels as core
materials for generation iv nuclear reactors A2 - yvon, Pascal, in: Pascal Yvon (Ed.),
Structural Materials for Generation IV Nuclear Reactors, Woodhead Publishing,
2017, pp. 329–355.
[15] R.L. Klueh, M.A. Sokolov, Mechanical properties of irradiated 9-Cr-2WVTa steel
with and without nickel, J. Nucl. Mater. 367 (2007) 102–106.
[16] Z. Zhong, T. Hinoki, Y. Park, et al., Interfacial reaction and diffusion control be-
tween SiC and F82H steel, Fus. Eng. Des. 85 (2010) 992–997.
[17] H. Yoon, Y. Kong, S. Kim, A. Kohyama, Mechanical properties of friction welds of
RAFs (JLF-1) to SUS304 steels as measured by the acoustic emission technique, Fus.
Eng. Des. 81 (2006) 945–950.
[18] Y.F. Li, Q.Y. Huang, Y.C. Wu, Study on impact and tensile properties of CLAM steel,
Nucl. Phys. Rev. 23 (2006) 151–154.
[19] J. Vanaja, K. Laha, M. Nandagopal, S. Sam, M.D. Mathew, T. Jayakumar,
E.R. Kumar, Effect of tungsten on tensile properties and flow behaviour of RAFM
steel, J. Nucl. Mater. 433 (2013) 412–418.
[20] M. Kondo, M. Ishii, Y. Hishinuma, T. Tanaka, T. Nozawa, T. Muroga, Metallurgical
study on corrosion of RAFM steel JLF-1 in Pb-Li alloys with various Li concentra-
tions, Fus. Eng. Des. 125 (2017) 316–325.
[21] J. Chen, C. Liu, Y. Liu, B. Yan, H. Li, Effects of tantalum content on the micro-
structure and mechanical properties of low-carbon RAFM steel, J. Nucl. Mater. 479
(2016) 295–301.
[22] J. Speer, D. Matlock, B. De Cooman, J. Schroth, Carbon partitioning into austenite
after martensite transformation, Acta Mater. 51 (2003) 2611–2622.
[23] W. Zhang, G.G. Roy, J.W. Elmer, T. DebRoy, Modeling of temperature field and
solidified surface profile during gas–metal arc fillet welding, J. Appl. Phys. 93
(2003) 3022.
[24] L. Shi, C.S. Wu, X.C. Liu, Modeling the effects of ultrasonic vibration on friction stir
welding, J. Mater. Process. Technol. 222 (2015) 91–102.
[25] A.F. Hasan, C.J. Bennett, P.H. Shipway, S. Cater, J. Martin, A numerical metho-
dology for predicting tool wear in Friction Stir Welding, J. Mater. Process. Technol.
241 (2017) 129–140.
[26] P.F. Kozlowski, B.G. Thomas, J.A. Azzi, W. Hao, Simple constitutive equations for
steel at high temperature, Metall. Trans. A 23 (1992) 903.
[27] Lei Yu-cheng, Zhu Qiang, Zhang Ling, Gu Kang-jia, Ju Xin, Simulation on the
welding of CLAM steel, Fusion. Eng. Des. 85 (2010) 1503–1507.
[28] R. Nandan, G.G. Roy, T.J. Lienert, T. Debroy, Three- dimensional heat and material
flow during friction stir welding of mild steel, Acta Mater. 55 (2007) 883–895.
[29] X.K. Zhu, Y.J. Chao, Numerical simulation of transient temperature and residual
stresses in friction stir welding of 304L stainless steel, J. Mater. Process. Technol.
1456 (2004) 263–272.
[30] H.H. Cho, H.N. Han, S.T. Hong, J.H. Park, Y.J. Kwon, S.H. Kim, et al.,
Microstructural analysis of friction stir welded ferritic stainless steel, Mater. Sci.
Eng. A 528 (2011) 2889–2894.
[31] M. Ghosh, K. Kumar, R.S. Mishra, Process optimization for friction-stir-welded
martensitic steel, Metall. Mater. Trans. A: Phys. Metall. Mater. Sci. 43 (2012)
1966–1975.
[32] Yasuyuki Miyano, Hidetoshi Fujii, Yufeng Sun, Yasuyuki Katada, Shuji Kuroda,
Osamu Kamiya, Mechanical properties of friction stir butt welds of high nitrogen-
containing austenitic stainless steel, Mater. Sci. Eng. A 528 (2011) 2917–2921.
[33] Hoon-Hwe Cho, Sung-Tae Hong, Jae-Hun Roh, Hyun-Sik Choi, Suk Hoon Kang,
Russell J. Steel, Heung Nam Han, Three-dimensional numerical and experimental
investigation on friction stir welding processes of ferritic stainless steel, Acta Mater.
61 (2013) 2649–2661.
[34] W. Han, A. Kimura, N. Tsuda, H. Serizawa, D. Chen, H. Je, et al., Effects of me-
chanical force on grain structures of friction stir welded oxide dispersion
strengthened ferritic steel, J. Nucl. Mater. 455 (2014) 46–50.
[35] L. Cui, H. Fujii, N. Tsuji, K. Nogi, Friction stir welding of a high carbon steel, Scr.
Mater. 56 (2007) 637–640.
[36] H. Fujii, L. Cui, K. Nogi, Welding of high carbon steel without transformation, Key
Eng. Mater. 345–346 (2007) 1411–1415.
[37] L. Cui, H. Fujii, N. Tsuji, K. Nakata, K. Nogi, R. Ikeda, M. Matsushita,
Transformation in stir zone of friction stir welded carbon steels with different
carbon contents, ISIJ Int. 47 (2007) 299–306.
[38] C. Zhang, L. Cui, Y. Liu, C. Liu, H. Li H, Microstructures and mechanical properties
of friction stir welds on 9% Cr reduced activation ferritic/martensitic steel, J. Mater.
Sci. Technol. (2017), http://dx.doi.org/10.1016/j.jmst.2017.11.049.
[39] H. Fujii, R. Ueji, Y. Morisada, H. Tanigawa, High strength and ductility of friction-
stir-welded steel joints due to mechanically stabilized metastable austenite, Scr.
Mater. 70 (2014) 39–42.
[40] T. Miura, R. Ueji, H. Fujii, Enhanced tensile properties of Fe–Ni–C steel resulting
from stabilization of austenite by friction stir welding, J. Mater. Process. Technol.
216 (2015) 216–222.
[41] X. Zhou, C. Liu, L. Yu, Y. Liu, H. Li, Phase transformation behavior and micro-
structural control of high-Cr martensitic/ferritic heat-resistant steels for power and
nuclear plants: a review, J. Mater. Sci. Technol. 31 (2015) 235–242.
[42] H.F. Lan, L.X. Du, Q. Li, C.L. Qiu, J.P. Li, R.D.K. Misra, Improvement of strength-
toughness combination in austempered low carbon bainitic steel: the key role of
refining prior austenite grain size, J. Alloy. Compd. 710 (2017) 702–710.
[43] Y.J. Li, J. Kang, W.N. Zhang, D. Liu, X.H. Wang, G. Yuan, et al., A novel phase
transition behavior during dynamic partitioning and analysis of retained austenite
C. Zhang et al.
in quenched and partitioned steels, Mater. Sci. Eng. A 710 (2018) 181–191.
[44] T. Miura, R. Ueji, H. Fujii, H. Komine, J. Yanagimoto, Stabilization of austenite in
low carbon Cr–Mo steel by high speed deformation during friction stir welding,
Mater. Des. 90 (2016) 915–921.
[45] S. Noh, M. Ando, H. Tanigawa, H. Fujii, A. Kimura, Friction stir welding of F82H
steel for fusion applications, J. Nucl. Mater. 478 (2016) 1–6.
[46] X. Li, L. Chen, Y. Zhao, X. Yuan, R.D. Kumar, Misra, Influence of original austenite
grain size on tensile properties of a high-manganese transformation-induced plas-
ticity (TRIP) steel, Mater. Sci. Eng. A (2018), http://dx.doi.org/10.1016/j.msea.
2017.12.107.
[47] A. Arlazarov, M. Ollat, J.P. Masse, M. Bouzat, Influence of partitioning on me-
chanical behavior of Q&P steels, Mater. Sci. Eng. A 661 (2016) 79–86.
[48] W. Cui, D. San-Martín, P.E.J. Rivera-Díaz-del-Castillo, Stability of retained austenite
in martensitic high carbon steels. Part I: thermal stability, Mater. Sci. Eng. A 711
(2018) 683–695.
267
Materials Science & Engineering A 729 (2018) 257–267
[49] W. Cui, M. Gintalas, P.E.J. Rivera-Diaz-del-Castillo, Stability of retained austenite in
martensitic high carbon steels. Part II: mechanical stability, Mater. Sci. Eng. A 711
(2018) 696–703.
[50] J.S. Yoon, S.-K. Kim, E.H. Lee, H.G. Jin, D.W. Lee, S. Cho, Mechanical properties of
ARAA steel after electron beam welding, Fusion. Eng. Des. 124 (2017) 774–778.
[51] J.J. Vora, V.J. Badheka, Experimental investigation on microstructure and me-
chanical properties of activated TIG welded reduced activation ferritic/martensitic
steel joints, J. Manuf. Process. 25 (2017) 85–93.
[52] C.R. Das, S.K. Albert, S. Sam, P. Mastanaiah, G.M.S.K. Chaitanya, A.K. Bhaduri,
et al., Mechanical properties of 9Cr–1W reduced activation ferritic martensitic steel
weldment prepared by electron beam welding process, Fusion. Eng. Des. 89 (2014)
2672–2678.
[53] V. Thomas Paul, S. Saroja, S.K. Albert, T. Jayakumar, K.E. Rajendra, Microstructural
characterization of weld joints of 9Cr reduced activation ferritic martensitic steel
fabricated by different joining methods, Mater. Charact. 96 (2014) 213–224.