Journal of Materials Science & Technology 34 (2018) 811–820

Contents lists available at ScienceDirect

Journal of Materials Science & Technology

journal homepage: www.jmst.org

Effect of cooling rate on microstructure, microsegregation and

mechanical properties of cast Ni-based superalloy K417G
Li Gong, Bo Chen ∗ , Long Zhang, Yingche Ma, Kui Liu
Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China

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

Article history: The effects of different solidification rates after pouring on the microstructures, microsegregation and
Received 25 April 2016 mechanical properties of cast superalloy K417G were investigated. Scheil-model was applied to calculate
Received in revised form 16 June 2016 the temperature range of solidification. The casting mould with different casting runners was designed
Accepted 27 February 2017
to obtain three different cooling rates. The microstructures were observed and the microsegregation
Available online 30 March 2017
was investigated. Also, high temperature tensile test was performed at 900 ◦ C and stress rupture test
was performed at 950 ◦ C with the stress of 235 MPa. The results showed that the secondary dendrite arm
Keywords:
spacing, microsegregation, the size and volume fraction of   phase and the size of /  eutectic increased
Ni-based cast superalloy
Cooling rate
with decreasing cooling rate, but the volume fraction of /  eutectic decreased. In the cooling rate range
Microstructure of 1.42 ◦ C s−1 –0.84 ◦ C s−1 , the cast micro-porosities and carbides varied little, while the volume fraction
Microsegregation and size of   phase and /  eutectic played a decisive role on mechanical properties. The specimen with
Mechanical property the slowest cooling rate of 0.84 ◦ C s−1 showed the best comprehensive mechanical properties.
© 2017 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science &
Technology.

1. Introduction elements and the growth of main phases. So the cooling rate has a
significant effect on the mechanical properties for casting superal-
Nickel-based superalloys are widely used in turbine compo- loy. Therefore, a complete description of the effect of cooling rate
nents of aircraft engines because of its excellent mechanical on K417G alloy is required.
properties under critical and severe service conditions [1–4]. Cast Most investigations focused on the effect of cooling rate during
superalloy K417G is used for the aeroengine turbine blade, which solidification process give a wide range of solidification cooling rate
is one of the most important parts of the engine working under [8,10–13]. However, the differences of cooling rate caused by dif-
the condition of high temperature up to 1100 ◦ C. The turbine blade ferent thickness in the turbine blade are small. So there is no critical
needs to sustain the centrifugal force generated by its own weight, conclusion which is closely linked to industrial production. Mean-
aerodynamic force generated by air and gas, thermal stress and while, most of these investigations are concerned on the effect of
vibration stress at the same time [5]. Therefore, K417G superalloy cooling rate on microsegregation and microstructure, but few are
requires excellent yield and tensile strength, high hot-corrosion concerned on mechanical properties.
and high-temperature oxidation resistance, enough ductility, ade- In this work, the casting mould with three different casting run-
quate endurance and creep strength, and good microstructure ners was designed to simulate different cooling rates in different
stability at high temperature [6,7]. parts of the turbine blade. All of the runners share a same runner
The common process of manufacturing K417G consists of vac- system, which excludes the influence of other factors. The effects of
uum induction melting (VIM) and then casting into turbine blades. cooling rate after pouring on the microstructures, microsegregation
The structure of turbine blade is complex and the thickness in dif- and mechanical properties of K417G superalloy were studied.
ferent parts is not the same [8]. During the solidification of melt,
the difference of thickness leads to different cooling rate [9]. Vari-
ous cooling rates result in the differences on the diffusion of alloy 2. Experimental

The chemical composition of K417G mother superalloy used in

the present work is listed in Table 1.
∗ Corresponding author. The solidification behavior of K417G alloy was calculated
E-mail address: [email protected] (B. Chen). according to the Scheil-model. Scheil calculations were performed

https://doi.org/10.1016/j.jmst.2017.03.023
1005-0302/© 2017 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.
812 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820

Table 1 homogenize the melt and then poured into the preheated mould.
Chemical compositions of K417G superalloy (wt%).
The pouring temperature was 1426 ◦ C. Molten alloys came into the
Ni C Cr Co Mo Al Ti V Zr B mould shell from the pouring gate P, flowed through the runners
Bal. 0.166 9.37 9.65 3.08 5.21 4.56 0.71 0.086 0.023 and the sprues, and then simultaneously filled into the specimen
A, B and C from the bottom of the sprues. After kept for 5 min,
the mould was removed from the vacuum induction furnace. In
the next day, the rods were removed from the sandbox. In theory,
there are three different cooling rates corresponding to specimen
A–C. The travel distance of the melt from the pouring gate to the
specimen A–C is the same. However, as shown in Fig. 1(b), spec-
imen A is the nearest to the furnace wall, C is the farthest and B
is in the middle. Since the room temperature was about 25 ◦ C and
the furnace wall was cold, it was easy for specimen A to transport
heat outside. Thus, specimen A should have the fastest cooling rate,
while C has the slowest. In order to trace the temperature during
casting, the B-type thermocouples were placed in the midsection of
specimen A–C (as shown in Fig. 1(a)). The temperature-time curves
were recorded by KSL/A-16EM1V0 temperature recorder during
the casting process.
The rod specimens were machined into gauge of
5 mm × 25 mm for high temperature tensile test and stress
rupture test. High temperature tensile test was performed at
900 ◦ C on WDW-T100 test machine with the initial displacement
speed of 0.15 mm min−1 , and then increasing to 3 mm min−1 after
yielding. Rupture test was performed at 950 ◦ C with the stress of
235 MPa on RDL-50 test machine.
Metallographic examinations were made on the longitudinal
section both at the screw position and the gauge position of the
high temperature tensile specimens. The samples were observed
by Olympus optical microscopy (OM) and FEI XL-30FEG scanning
electron microscopy (SEM). To quantitatively determine the vol-
ume fraction of the   phase, the /  eutectic and cast porosity, 20
images captured at different areas of each specimen were needed
by Sisc IAS image analysis system, and the average was taken. 100
particles were selected randomly in the SEM images and measured
to confirm the size of   phase. The element concentrations and
distribution on the un-etched samples were investigated by JEOL
JXA8530F electron probe microanalysis (EPMA). Three points at the
dendrite core and the interdendritic region were selected randomly
in each sample and the average was taken. The samples were pre-
pared by jet-beam electro-polishing in a solution of 10% perchloric
acid and 90% ethanol at about −20 ◦ C, and then observed on FEI G2
20 transmission electron microscopy (TEM). The fracture surfaces
at 900 ◦ C tensile tests with various cooling rates were observed by
scanning electron microscopy (SEM) after ultrasonic cleaning.

3. Results and discussion

3.1. Scheil-model simulation and cooling rates of solidification

The solidification process of K417G superalloy calculated by the

Scheil model is presented in Fig. 2. It was predicted by this model
Fig. 1. Shell structure of casting mould designed for different cooling rates after that the formation temperature of the  matrix (1346 ◦ C) corre-
pouring: (a) stereogram; (b) profile of casting mould. sponded to the liquidus temperature, and the solidus temperature
was 1150 ◦ C as the solid fraction was 1.0. From these results, major
on TTNI8 database. The program proceeded to step down every 1 ◦ C phases precipitated from the liquid between 1346 ◦ C and 1175 ◦ C,
until the solidification was completed. At each step, the program so this temperature range was selected to calculate the cooling
would determine the phase according to the lowest Gibb’s energy. rates.
The Scheil model calculated under the non-equilibrium condition Fig. 3 shows the temperature-time curves recorded by the
based on the correction of the segregation coefficient. thermocouple during the casting experiment. Before pouring, tem-
Remelting and casting experiments were performed in the vac- peratures of the cast mould at specimens A, B and C had fallen from
uum induction furnace. The dewaxing mould buried in the Al2 O3 900 ◦ C to 754 ◦ C, 836 ◦ C and 869 ◦ C, respectively. After pouring, the
sands was preheated to 900 ◦ C for 4 h. As shown in Fig. 1(a), temperature rose quickly to the highest in a few seconds and then
12 rod specimens (10 mm × 100 mm) uniformly distributed in decreased at different cooling rates. With extension of time, heat
the mould. The alloys were superheated to 1450 ◦ C for 10 min to would spread all over the furnace, so cooling rates of all the three
 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820 813

Table 2
Element segregation analysis of different cooling rates (v) between dendrite core
and interdendritic regions.

Sample v (◦ C s−1 ) V Co Al Cr Mo Ti

A 1.42 Cd (%) 0.78 10.75 5.28 9.39 2.46 3.89

Ci (%) 0.80 10.03 5.36 9.54 2.52 4.46
k 0.98 1.07 0.98 0.98 0.98 0.87

B 1.06 Cd (%) 0.83 10.90 5.31 9.60 2.51 3.45

Ci (%) 0.81 10.19 5.30 9.96 2.63 4.36
k 1.02 1.07 1.00 0.96 0.95 0.79

C 0.84 Cd (%) 0.80 11.09 5.26 9.87 2.34 3.41

Ci (%) 0.78 10.49 5.16 10.47 2.60 4.51
k 1.02 1.06 1.02 0.94 0.90 0.76

Table 2, elements such as Ti, Mo and Cr tend to segregate towards

Fig. 2. Simulation results of K417G superalloy calculated by Scheil model. the interdendritic regions, so these elements are more likely to
gather in the liquid phase during solidification. The segregation of
Ti is the severest, while Mo and Cr just slightly segregate to the
interdendritic region. The k value of Co is about 1.07 which means
that during solidification Co would segregate to the dendrite core
slightly. The other elements such as V and Al can be considered as
neutral elements, because they have no obvious segregation ten-
dency. It is well known that the solidification rate has a significant
effect on the segregation level and lower solidification rate would
result in more severe segregation. As shown in Table 2, the segre-
gations of Ti, Mo and Cr become more serious with lower cooling
rate.

3.3. Microstructure

After etching with the solution of 20 g CuSO4 + 100 ml HCl + 5 ml

H2 SO4 + 100 ml H2 O, the microstructure at the gauge position of the
high temperature tensile specimens with different cooling rates
were analyzed by OM. Fig. 4 shows that K417G superalloy has a
typical dendritic morphology with  matrix,   reinforced phase,
Fig. 3. Temperature vs. time curves during casting experiment for K417G superalloy.
/  eutectic, carbides and cast porosity.
The dendrite arm spacing, as a critical microstructure param-
specimens became slower and slower. As shown in Fig. 3, cooling eter, has an important effect on the mechanical properties [18].
rates of all the three specimens tended to be the same after 160 s. The secondary dendrite arm spacing () decreased with increased
However, the slopes of the three temperature-time curves were cooling rate, as shown in Fig. 5. The specimen A with 1.42 ◦ C s−1
obviously different between 1310 ◦ C to 1175 ◦ C. That is to say, the cooling rate has the finest microstructure and the average value of
cooling rates of specimens A, B and C during solidification were  is 44.3 ␮m, while the specimen C with 0.84 ◦ C s−1 cooling rate has
different. Just as expected, specimen A which was the nearest one the coarsest microstructure and the average value of  is 53.8 ␮m,
to the furnace wall had the fastest cooling rate, while the farthest as listed in Table 3.
specimen C had the slowest cooling rate. The temperatures of spec- It is well known that the   phase is the main strengthening
imen A decreased from 1310 ◦ C to 1175 ◦ C in 95 s, B from 1328 ◦ C to phase of Ni-based superalloys and it coherently precipitates from
1175 ◦ C in 144 s, and C from 1349 ◦ C to 1175 ◦ C in 208 s. Therefore, the supersaturated  matrix during the solidification process. As
the cooling rates ranging from the highest temperature to 1175 ◦ C shown in Fig. 6, the typical morphology of   particles showed good
of specimens A, B and C were 1.42, 1.06 and 0.84 ◦ C s−1 , respectively. cubic structure both in the dendritic and interdendritic regions with
As the temperature of the cast mould at the specimen A was lower various cooling rates. With different cooling rates, the volume frac-
than specimen C before pouring, the highest temperature at spec- tion of the   phase in the dendrite cores changed from 63.3% to
imen A after pouring captured by the temperature recorder was 64.7%, and the average sizes were 521 nm, 569 nm and 612 nm,
lower than that of specimen C. respectively, as listed in Table 3. The volume fraction of the   phase
in the interdendrite changed from 64.0% to 65.5%, and the aver-
3.2. Microsegregation age sizes were 550 nm, 601 nm and 640 nm. With increased cooling
rate, the volume fraction and average size of the   phase decreased
The chemical compositions of the dendritic and interdendritic gradually. As is known, the degrees of supercooling and supersat-
regions were determined by EPMA to investigate the microseg- uration play a significant role in the nucleation of   phase. Higher
regation with different cooling rates. The elemental segregation supercooling degree promotes the nucleation, but results in insuf-
coefficients (k) are listed in Table 2, which are defined as the ratio ficient time for   phase to grow [19]. For this work, the variation
of the average concentration of the elements in the dendrite core range of cooling rate is not very large, so its effect on the promo-
(Cd ) to that in the interdendritic region (Ci ), i.e., k = Cd /Ci [14,15]. tion of nucleation is not obvious. On the other hand, slower cooling
The value of k less than 1 indicates that the elements segregate to rate leads to coarser particles, which results in increased volume
the interdendritic region, while k greater than 1 indicates that the fraction. Because of microsegregation, Al and Ti were enriched in
elements segregate to the dendrite core [16,17]. From the results in the interdendrite. Therefore, the average size and volume fraction
814 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820

Fig. 4. Microstructure at gauge position of high temperature tensile specimens with different cooling rates: (a, b) specimen A, 1.42 ◦ C s−1 ; (c, d) specimen B, 1.06 ◦ C s−1 , (e, f)
specimen C, 0.84 ◦ C s−1 .

Fig. 5. OM images showing secondary dendrite arm spacing () for specimen A (a), specimen B (b) and specimen C (c) with cooling rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and
0.84 ◦ C s−1 , respectively.

Table 3
Microstructure features at gauge position of specimens under different cooling rates.

Sample v (◦ C s−1 )  (␮m) Size of   phase (nm) Volume fraction of   phase (%) Size of /  Volume fraction Volume fraction
eutectic (␮m) of /  eutectic of cast porosity
(%) (%)

dendrite interdendrite dendrite interdendrite

A 1.42 44.3 ± 0.3 521 ± 16 550 ± 11 63.3 ± 0.4 64.0 ± 0.8 31.7 ± 2.2 6.0 ± 1.3 0.24 ± 0.1
B 1.06 48.7 ± 0.3 569 ± 11 601 ± 15 63.9 ± 0.8 64.7 ± 0.6 32.2 ± 2.3 5.6 ± 1.3 0.25 ± 0.1
C 0.84 54.3 ± 0.8 612 ± 19 640 ± 9 64.7 ± 0.5 65.5 ± 0.6 41.3 ± 4.9 4.7 ± 1.2 0.25 ± 0.1
 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820 815

Fig. 6. Morphologies of   phases in dendrite (a, c, e) and interdendrite (b, d, f) for specimen A (a, b), specimen B (c, d) and specimen C (e, f) with cooling rates of 1.42 ◦ C s−1 ,
1.06 ◦ C s−1 and 0.84 ◦ C s−1 , respectively.

of the   phase in the interdendrite are slightly larger than those in [19]. At higher cooling rate, the shorter diffusion time may weaken
the dendrite. the effect of back-diffusion, and Al and Ti have insufficient time to
To better analyze the /  eutectic, the samples were electrolyt- dissolve into  matrix and carbides. The enrichment of Al and Ti in
ically etched with 10% phosphoric acid. As shown in Fig. 7, most the liquid would lead to more /  eutectic. Therefore, the volume
of the /  eutectic distributed at dendrite boundaries, and the fraction of the /  eutectic is relatively larger with higher cooling
micro-porosities and carbides could be generally observed nearby, rate. On the other hand, the lower cooling rate leads to longer diffu-
which indicated that the /  eutectic were formed during the last sion distance, so Al and Ti tend to gather together. Therefore, larger
stage of solidification. 10 images were measured at each sample /  eutectic will precipitate there.
and the results were listed in Table 3. The average size of the /  K417G is subjected to forming cast porosity because of its wide
eutectic varied from 31.7 ␮m to 41.3 ␮m and the average volume solidification range. As shown in Fig. 4, the micro-porosities were
fraction varied from 6.0% to 4.7%. It indicates that the volume frac- mostly presented in the interdendritic regions or grain boundaries.
tion of the /  eutectic tends to be smaller and the size is larger From quantitative calculation listed in Table 3, the volume frac-
with relatively lower cooling rate. The formation of the /  eutec- tions of cast porosity with different cooling rates were 0.24%, 0.25%
tic is strongly related to the microsegregation. The degree of the and 0.25%, respectively. It can be seen that the volume fraction of
microsegregation is greatly influenced by the back-diffusion in the the micro-porosity had a slight increase at lower cooling rate. The
solidified metal, and there are two factors that play a significant extent of the micro-porosity is strongly related to the solidification
part in this diffusion: the diffusion distance and the diffusion time parameters, such as grain size, precipitation and development of
816 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820

Fig. 7. Morphologies and distribution of /  eutectic specimen A (a, b), specimen B (c, d) and specimen C (e, f) at low (a, c, e) and high (b, d, f) magnification with cooling
rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and 0.84 ◦ C s−1 , respectively.

Fig. 8. Morphologies and distribution of the carbides for specimen A (a), specimen B (b) and specimen C (c) with cooling rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and 0.84 ◦ C s−1 ,
respectively.

secondary phase, element segregation, temperature gradient and These three different cooling rates have no significant effect on
solidification velocity[20,21]. In present study, lower cooling rate the morphology and distribution of carbides in K417G superalloy.
leads to more serious segregation, which results in the precipitation As shown in Fig. 8, the carbides showed blocky and rod-like shape
and development of the secondary phases. They serve as bridging and they mainly distributed in dendrite boundaries or near the / 
obstacles in the dendritic regions and hinder the smooth flow of eutectic region. From the energy spectrum analysis (Fig. 9), the
feed metal [22,23], and thus more micro-porosities would gener- rod-like and blocky carbides are rich in Mo and Ti. As illustrated
ate. However, the variation range of cooling rate in this work is not in Fig. 10, the blocky carbide is of fcc structure with the lattice con-
very large, so the volume fraction of cast porosity varies slightly. stant a of 0.429 nm, which falls between 0.4273 (MoC) and 0.4327
 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820 817

Fig. 10. TEM image and SAED pattern of MC carbide in specimen C with cooling rate
of 0.84 ◦ C s−1 .

Fig. 9. SEM images of carbides (a) and EDS analysis of different areas (b, c).

(TiC), so the probability is that they are MC ((Ti, Mo)C) carbides.

As is known, MC carbide is a kind of common carbides in Ni-based
superalloy [24]. During the process of solidification, the segrega-
tion of Ti and Mo in the interdendrite resulted in a large amount
of MC carbides precipitation. The average volume fraction of MC
carbides is about 2.1%. The fine MC carbides contribute to strength-
ening grain boundaries at elevated temperature, but degenerated
MC carbides accelerate the microcracks [25] and facilitate the fail-
ure of specimens on the creep test [26], so degenerated MC carbides
are detrimental to the mechanical properties. The MC carbides in
the present work remained stable and did not degenerate during
high temperature tensile test, as shown in Figs. 8 and 9.
In addition, some tiny rod-like carbides precipitating in the Fig. 11. SEM image showing morphology of carbides in specimen C with cooling
grain boundaries are rich in Cr, as shown in Fig. 11. These car- rate of 0.84 ◦ C s−1 (a) and corresponding EDS analysis (b).
bides also have fcc structure but the lattice constant a is 1.073 nm
(the lattice constant of M23 C6 carbide is 1.066 nm), as illustrated
818 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820

Table 4
Microstructure features at screw position of specimens under different cooling rates.

Sample v (◦ C s−1 )  (␮m) Size of   phase (nm) Volume fraction of   phase (%) Size of /  Volume fraction Volume fraction
eutectic (␮m) of /  eutectic of cast porosity
(%) (%)

dendrite interdendrite dendrite interdendrite

A 1.42 47.1 ± 0.5 565 ± 12 589 ± 17 64.3 ± 0.4 64.9 ± 0.6 33.2 ± 2.4 4.9 ± 1.2 0.21 ± 0.1
B 1.06 50.9 ± 0.7 594 ± 15 622 ± 10 64.9 ± 0.7 65.7 ± 0.7 34.7 ± 2.8 4.8 ± 0.9 0.24 ± 0.2
C 0.84 54.7 ± 0.3 627 ± 10 653 ± 13 65.9 ± 0.6 66.2 ± 0.9 44.4 ± 3.7 4.7 ± 0.5 0.24 ± 0.1

Table 5
High temperature tensile properties and stress rupture properties of K417G superalloy with various cooling rates.

Sample v (◦ C s−1 ) Ultimate tensile strength (MPa) Elongation (%) Reduction of area (%) Rupture life (h)

A 1.42 761.0 7.4 8.8 67.5

B 1.06 735.5 8.2 8.9 68.0
C 0.84 714.0 11.3 15.8 70.5

geneous compositions of casting alloy, but they have the same

tendency.

3.4. Mechanical property

The ultimate tensile strength, elongation and reduction of area

in tensile tests at 900 ◦ C and the typical rupture life in rupture tests
at 950 ◦ C with the stress of 235 MPa are listed in Table 5. Specimen C
with the slowest cooling rate of 0.84 ◦ C s−1 has the best comprehen-
sive mechanical properties, because of the minimal volume fraction
of /  eutectic, the maximum volume fraction and the largest size
of   phase.
The three fracture surfaces at 900 ◦ C tensile tests with various
cooling rates are shown in Fig. 13. It can be seen that the speci-
mens with different cooling rates all had a ductile fracture manner.
Fig. 13(b), (d) and (f) exhibited that there were cast porosities on the
fracture surfaces. Also, these three images showed that the crack
expanded in the interdendrite in all specimens. It can be found that
the cooling rate has no obvious effect on the fracture manner. As
shown in Fig. 13(a), (c) and (e), more shallow dimples can be iden-
tified on the fracture surface of specimen C. The crystal lattice of
specimen C generated larger deformation under high tensile test,
so its plastic property (elongation and reduction of area) was higher
than that of the other specimens.
Abbadi et al. [29] discovered that the cracks grow faster in the
coarse microstructure alloy than in the fine microstructure alloy in
the Ni-based superalloys at elevated temperature. However, spec-
imen C with the coarsest dendrite arm spacing showed the best
comprehensive mechanical properties. The variation range of grain
Fig. 12. TEM image and SAED pattern of M23 C6 carbide in specimen C with cooling size in this work was small, so it had little influence on the mechan-
rate of 0.84 ◦ C s−1 . ical properties.
The high temperature strength of Ni-based superalloy is pri-
marily determined by the size, volume fraction and morphology of
in Fig. 12. Therefore, these tiny rod-like precipitations should be the   precipitates [30–32]. The   particles act as barriers against
M23 C6 carbides. At low cooling rate in the sandbox, the segrega- movement of dislocation [33]. With slow cooling rate, the   phase
tion of Cr and C at the grain boundaries would create favorable becomes coarser, and it seems easier for the dislocation to move
conditions for the nucleation and growth of M23 C6 carbides. The around the   phase rather than shear it. Therefore, the deformation
M23 C6 carbides precipitated from the  matrix in the reaction resistance of the  matrix is low and the crack is difficult to form and
of 23 M + 6C→M23 C6 [27]. Continuous M23 C6 chains in the grain expand. So the elongation and reduction of area are high with slow
boundaries would induce initiation and propagation of the cracks cooling rate. On the other hand, with rapid cooling rate, the   phase
under service, which is detrimental to mechanical properties [28]. becomes fine and the dislocation is easier to shear it. Therefore, the
M23 C6 carbides in this alloy showed tiny rod-like shape and did not high deformation resistance of the  matrix results in high tensile
harm the mechanical properties. strength and low ductility property at high temperature tensile test.
Microstructures at the screw position of the high temperature The longitudinal profiles of three fractures at 900 ◦ C tensile tests
tensile specimens (both ends of the specimens) were also ana- with various cooling rates are shown in Fig. 14. It could be seen that
lyzed in the same way as the samples from the gauge position. cracks in all specimens expanded along the grain boundary and
As illustrated in Table 4, samples in the screw position have some the interdendrite, and the direction of tensile stress axis and the
differences from those in the gauge position because of inhomo- direction of crack propagation are about 45◦ . Under the action of
 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820 819

Fig. 13. SEM images of fracture surfaces after tensile test at 900 ◦ C for specimen A (a, b), specimen B (c, d) and specimen C (e, f) at low (a, c, e) and high (b, d, f) magnification
with cooling rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and 0.84 ◦ C s−1 , respectively.

Fig. 14. OM images of fracture surfaces after tensile tests at 900 ◦ C for specimen A (a), specimen B (b) and specimen C (c) with cooling rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and
0.84 ◦ C s−1 , respectively.

shear stress, the slip and deformation in the grain boundary and the because of stress concentration, which leads to the formulation and
interdendrite lead to microcracks. The expansion of microcracks propagation of microcracks, so the increased volume fraction of
along the grain boundary and the interdendrite leads to the fracture /  eutectic in the grain boundary and the interdendritic region
of materials. Therefore, with the cooling rate increasing, the volume results in low ductility property of K417G alloy.
fraction of /  eutectic increases a lot, which would hinder the The favorable stress rupture life mainly depends on the
movement of dislocations. Thus, a larger deformation will generate increased volume fraction of   phase [34]. Specimen C with the
820 L. Gong et al. / Journal of Materials Science & Technology 34 (2018) 811–820

cooling rate of 0.84 ◦ C s−1 has relatively larger volume fraction of   Acknowledgements
phase, so it results in a longer rupture life. However, this improve-
ment is not obvious because there are small differences in the The authors thank Jiangsu Tunan Alloy Co., Ltd for providing
volume fraction of   phase with different cooling rates. alloy and their assistance on the remelting and cast procedures.
The /  eutectic is a phase with lower strength and is prone to Prof. Shunnan Zhang is especially acknowledged for his advice on
form cracks [20,35]. Moreover, /  eutectic contains an amount the solidification experiments.
of   formation elements, such as Al and Ti, which results in a
decreased volume fraction of   phase [20,35]. Therefore, /  References
eutectic is detrimental to mechanical properties. As shown in
Fig. 14, an amount of /  eutectics could be observed on the sur- [1] S.M. Seo, H.W. Jeong, Y.K. Ahn, D.W. Yun, J.H. Lee, Y.S. Yoo, Mater. Charact. 89
(2014) 43–55.
face fracture in all specimens, so the /  eutectic has a negative [2] X.R. Guan, E.Z. Liu, Z. Zheng, Y.S. Yu, J. Tong, Y.C. Zhai, J. Mater. Sci. Technol. 27
influence on the mechanical properties for this alloy. Therefore, (2011) 113–117.
specimen C with minimal volume fraction of /  eutectic has bet- [3] D.W. Maclachlan, D.M. Knowles, Mater. Sci. Eng. A 302 (2001) 275–285.
[4] T.M. Pollock, S. Tin, J. Propuls. Power 22 (2006) 361–374.
ter comprehensive properties. [5] C. Yuan, in: J.T. Guo (Ed.), Materials Science and Engineering for Superalloys
Ш, Science Press, Beijing, 2010, pp. 491–531 (in Chinese).
4. Conclusions [6] P. Zhang, Q. Zhu, C. Hu, C.J. Wang, G. Chen, H.Y. Qin, Mater. Des. 69 (2015)
12–21.
[7] F.J. Xu, Y.H. Lv, Y.X. Liu, F.Y. Shu, P. He, B.S. Xu, J. Mater. Sci. Technol. 29 (2013)
After analyzing K417G specimens at different location of cast 480–488.
mould, there were three cooling rates of 1.42 ◦ C s−1 , 1.06 ◦ C s−1 and [8] W.G. Jiang, J.S. Dong, L. Wang, L.H. Lou, J. Mater. Sci. Technol. 27 (2011)
831–840.
0.84 ◦ C s−1 after pouring, and the microstructure, microsegregation [9] J. Campbell, Castings, second ed., Reed Educational and Professional
and mechanical properties varied evidently with different cooling Publishing, London, 2003, pp. 117–177.
rates. The results obtained as follows: [10] L. Zheng, J. Aeronaut. Mater. 26 (2006) 7–11 (in Chinese).
[11] Y.J. Zhang, Y.J. Huang, L. Yang, J.G. Li, J. Alloys Compd. 570 (2013) 70–75.
[12] S. Behrouzghaemi, R.J. Mitchell, Mater Sci. Eng. A 498 (2008) 266–271.
(1) Mo and Cr tend to segregate towards the interdendritic regions, [13] N. EI-Bagoury, A. Nofal, Mater. Sci. Eng. A 527 (2010) 7793–7800.
and the segregation becomes more serious with lower cool- [14] E.C. Caldwell, F.J. Fela, G.E. Fuchs, in: K.A. Green, T.M. Pollock, H. Harada, T.E.
ing rate. Co slightly segregates towards the dendrite cores. The Howson, R.C. Reed, J.J. Schirra, S. Walston (Eds.), Superalloy 2004, TMS, 2004,
pp. 811–818, Champion.
other elements such as V and Al can be considered as neutral [15] R.M. Kearsey, J.C. Beddoes, K.M. Jaansalu, W.T. Thompson, P. Au, in: K.A.
elements. With lower cooling rate, microsegregation becomes Green, T.M. Pollock, H. Harada, T.E. Howson, R.C. Reed, J.J. Schirra (Eds.),
more serious. Improved homogeneity in the microstructure is Superalloy 2004, TMS, 2004, pp. 801–810, Champion.
[16] Z.X. Shi, J.X. Dong, M.C. Zhang, L. Zheng, J. Alloys Compd. 571 (2013) 168–177.
observed with the highest cooling rate of 1.42 ◦ C s−1 . [17] X. Zhao, L. Liu, Z. Yu, Q. Zhang, J. Zhang, H. Fu, J. Mater. Sci. 45 (2010)
(2) In the cooling rate range of 1.42 ◦ C s−1 –0.84 ◦ C s−1 , the   phase 6101–6107.
shows good cubic structure in all specimens. With the cooling [18] B.N. Du, J.X. Yang, C.Y. Cui, X.F. Sun, Mater. Sci. Eng. A 623 (2015) 59–67.
[19] F. Wang, D. Ma, J. Zhang, L. Liu, J. Hong, S. Bogner, A. Bührig-Polaczek, J. Cryst.
rate decreasing, the secondary dendrite arm spacing, the size Growth 389 (2014) 47–54.
and volume fraction of the   phase and the size of the /  [20] J. Zhang, J.G. Li, T. Jin, X.F. Sun, Z.Q. Hu, J. Mater. Sci. Technol. 26 (2010)
eutectic increase, but the volume fraction of the /  eutectic 889–894.
[21] J. Lecomte-Beckers, Metall. Mater. Trans. A 19 (1988) 2341–2348.
decreases. [22] Q.Z. Chen, Y.H. Kong, C.N. Jones, D.M. Knowles, Scr. Mater. 51 (2004) 155–160.
(3) All the specimens with different cooling rates contain a large [23] M.J. Donachie, S.J. Donachie, Superalloys: A Technical Guide, second ed., ASM
number of blocky and rod-like carbides at the dendrite bound- International, Materials Park, 2002, pp. 224–257.
aries or near the /  eutectic region. The carbides are probable [24] J. Wang, L.Z. Zhou, L.Y. Sheng, J.T. Guo, Mater. Des. 39 (2012) 55–62.
[25] X.Z. Qin, J.T. Guo, C. Yuan, C.L. Chen, J.S. Hou, H.Q. Ye, Mater. Sci. Eng. A 485
(Ti, Mo)C, which are fcc crystal structure with the lattice con- (2008) 74–79.
stant a of 0.423 nm. A few tiny M23 C6 carbides can be observed [26] L.R. Liu, T. Jin, N.R. Zhao, X.F. Sun, H.R. Guan, Z.Q. Hu, Mater. Sci. Eng. A 361
in the interdendrite region, which are of fcc crystal structure (2003) 191–197.
[27] G.H. Bo, R. Hu, J.S. Li, H. Zhong, H.C. Kou, Z.H. Fu, Rare Met. Mater. Eng. 40
with the lattice constant a of 0.429 nm. (2011) 1737–1741 (in Chinese).
(4) In the cooling rate range of 1.42 ◦ C s−1 –0.84 ◦ C s−1 , the volume [28] L.Z. He, Q. Zheng, X.F. Sun, G.C. Hou, H.R. Guan, Z.Q. Hu, J. Mater. Sci. 32 (1997)
fraction and distribution of cast micro-porosities and carbides 4917–4923.
[29] M. Abbadi, P. Hähner, S. Belouettar, M. Zenasni, Mater. Des. 32 (2011)
vary little. Therefore, the volume fraction and size of   phase 2710–2716.
and /  eutectic play a decisive role on mechanical proper- [30] D.M. Collins, R.K. Heenan, H.J. Stone, Metall. Mater. Trans. A 42 (2011) 49–59.
ties. With the slowest cooling rate of 0.84 ◦ C s−1 , the specimen [31] Y.L. Xu, Q.M. Jin, X.S. Xiao, X.L. Cao, G.Q. Jia, Y.M. Zhu, H.J. Yin, Mater. Sci. Eng.
A 528 (2011) 4600–4607.
C shows good mechanical properties of ductility and stress rup- [32] F. Pyczak, B. Devrient, F.C. Neuner, H. Mughrabi, Acta Mater. 53 (2005)
ture life due to the minimal volume fraction of the /  eutectic, 3879–3891.
and the maximum fraction and the largest size of the   phase. [33] S.A. Sajjadi, H.R. Elahifar, H. Farhangi, J. Alloys Compd. 455 (2008) 215–220.
[34] F.X. Yang, E.Z. Liu, Z. Zhi, J. Tong, L.K. Ning, Mater. Des. 61 (2014) 41–49.
(5) Considering the microstructure and mechanical properties, the [35] W.S. Walston, I.M. Bernstein, A.W. Thompson, Metall. Mater. Trans. A 22
optimized cooling rate is 0.84 ◦ C s−1 . (1991) 1443–1451.