j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

Available online at www.sciencedirect.com

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

Original Article

Study of meta-dynamic recrystallization behavior

of GH5188 superalloy

Dahai Liu*, Jingdong Chen**, Haorui Chai, Yihang Jiang, Zhi Li,
Wenqing Qiu, Zhenghua Guo
School of Aeronautical Manufacturing Engineering, Nanchang Hangkong University, 330063, Nanchang, China

article info abstract

Article history: Cobalt-based superalloys have good service performance in environments where nickel-
Received 7 July 2021 based superalloys cannot work. A number of the studies focused on mechanical proper-
Accepted 29 August 2021 ties, microstructure evolution and hot deformation characteristics, but studies of meta-
Available online 3 September 2021 dynamic recrystallization (MDRX) behavior remain lacking. The MDRX behavior of
GH5188 superalloy was investigated using two-stage hot compression tests. The kinetic
Keywords: model and grain size model were built to better describe the MDRX process. The good
GH5188 superalloy agreement between the predicted values and experimental values was verified. The effects
Hot compression of holding time, temperature, strain rate and intervals time on MDRX microstructure
Meta-dynamic recrystallization evolution were analyzed and discussed. The results show that by descending order of
High-temperature deformation correlation was: intervals time, holding time, temperature and strain rate. The MDRX
fraction and grain size increase with the increase of holding time, temperature, strain rate
and intervals time. At low temperatures, MDRX grains grew up in the form of bowed grain
boundaries. At high temperatures, the growth of MDRX grains was often accompanied by
the generation of twinning bands. Twins were formed in the process of grain growth.
© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

and improves mechanical properties. Therefore, it is very

1. Introduction meaningful to study the MDRX behavior of GH5188 superalloy
to provide an experimental and theoretical basis for the
GH5188 superalloy is a cobalt-based superalloy that is optimization and control of grain size.
strengthened predominately by solution strengthening of M6C Some researchers have conducted studies on microstruc-
and M23C6 phase [1e4]. Compared with nickel-based superal- ture. Wang et al. [5] studied homogenization of superalloy
loys, GH5188 superalloy has stronger anti-creep, stronger anti- GH188 ingots, found that annealing at 1180
C for 20 h could
corrosion, better fatigue performance and higher-heat- reduce dendritic segregation. Gao et al. [6] supplemented the
strength. Under the increasingly harsh working environ- microstructure evolution of GH5188 cobalt-base superalloy,
ment, research on cobalt-based superalloys is imperative. The found that when the annealing temperature was higher than
MDRX makes great contributions to obtain uniform fine grain 1170
C, the recrystallized grain could grow in a normal way.

* Corresponding author.
** Corresponding author.
E-mail addresses: [email protected] (D. Liu), [email protected] (J. Chen).
https://doi.org/10.1016/j.jmrt.2021.08.134
2238-7854/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
1180 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

Pyczak et al. [7] studied the plastic deformation mechanism of

creep hardening cobalt-based superalloys and found that the
deformation was mainly caused by a/2<101> dislocation slip,
and the dislocations were mainly passed through the hin-
drance of g' phase. Favre et al. [8] investigated the evolution of
the microstructure of L-605 cobalt superalloy using mean-field
models combined with a reverse-engineering method, found
that this method provided some unknown material parame-
ters such as mobility and nucleation frequency, and helped to
quantify the occurrence of recrystallization. However, there is
no report on the research on MDRX process of GH5188
superalloy.
Although many researchers have studied the mechanical
properties and microstructure of cobalt-based superalloys, Fig. 1 e The initial microstructures of the GH5188
they have failed to establish a connection between the superalloy.
microstructure and mechanical properties. Therefore, some
researchers study the connection by establishing constitutive
equations. Lei et al. [9] established a new creep constitutive precipitate inside the grain and on the grain boundary. The
model, found that it shows better results than the norton grain size is about 16 mm.
model. Anil Kumar et al. [10] studied the hot workability and The cylindrical specimens of Ф10  15 mm were taken from a
microstructures evolution of cobalt-based superalloy forged billet. Two-stage hot compression tests were conducted
Coe20Cre15We10Ni using a thermo-mechanical simulator, on Gleeble-1500 thermal simulator produced by Dynamic Sys-
found that strain hardening was observed at lower strain rates tems Incorporated. Before the test, a thermocouple was welded
and flow softening at higher strain rates, and constitutive on the sample surface. To control the friction in the tests, the
equation had been established. Subsequently, Lyu et al. [11] ends of the specimens were lubricated with MoS2 powders. The
established the creep constitutive equation of GH188 alloy at two-stage hot compression process is shown in Fig. 2.
800
C, found that creep deformation and stress relaxation A heating rate of 20
C/s was heated for each specimen to
proceed at the early stage of forming process. He et al. [12] 1180
C and the specimens were held for a time (t1 ¼ 60, 120 and
established a constitutive model of GH5188 superalloy based 180 s). Then the temperature was cooled to the deformation
on backpropagation neural network, found that the flow temperature (T ¼ 1080, 1130 and 1180
C) and was held for 30 s.
stress of GH5188 superalloy was decreased greatly with the The initial-stage hot compression tests were carried out over
increase in temperature or decrease in the strain rate. Ouyang strain rates (_ε) of 1, 5 and 10 s1 and strain (ε1) of 0.4. In the
et al. [13] studied the hot deformation behavior of CoeNi- process of hot compression, to ensure the occurrence of
based superalloys, established the Arrhenius-type constitu- MDRX, the initial-stage hot compression amount was greater
tive equation and a processing map. In addition, studies have than the critical deformation amount of DRX (37%). Critical
found the optimum parameters (T ¼ 1050
C, ε_ ¼ 0.01 s1, deformation is the degree of deformation when the recrys-
ε ¼ 0.7). Zhang et al. [14] studied the influence of deformation tallized grains begin to be refined. At the same time, solution
parameters and Y2O3 content on the flow behavior and treatment is a heat treatment process to recrystallize the alloy.
microstructure of co-based ODS alloys. It was found that the In the basic knowledge of GH5188 superalloy, the maximum
flow softening could be accelerated by decreasing the strain deformation of solution treatment is 37%. Therefore, it is
rate. However, the MDRX dynamic model and grain size model determined that the critical deformation of GH5188 superalloy
have not been established.
A number of the above studies focused on mechanical
properties, microstructure evolution and hot deformation
characteristics, but studies of MDRX behavior remain lacking.
Thus, we studied the MDRX behavior of the GH5188 superalloy
using two-stage hot compression tests. The influence of
deformation parameters on MDRX was discussed. Besides, the
kinetic model and grain size model of MDRX for GH5188 su-
peralloy were established based on MDRX theory.

2. Materials and method

The material studied is GH5188 superalloy with a nominal

composition (wt.%) of 22.35Cr1.00Mn1.68Fe37.08Co
21.64Ni16.25W. The microstructure of the GH5188 superalloy
bar ingot was observed by SU510 scanning electron micro-
scope (SEM), as shown in Fig. 1 below. As can be seen from Fig. 2 e Schematic of two-stage hot compression tests of
Fig. 1, M6C and M23C6 carbides in the initial microstructures GH5188 superalloy.
 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9 1181

is 37%. After reaching the strain, unload and use different in-
tervals time (t2 ¼ 1, 2, 5 s) to ensure the progress of MDRX.
Subsequently, the two-stage hot compression tests were car-
ried out over a stain (ε2) of 0.1. The temperature and strain rate
were the same as the initial-stage hot compression tests. Upon
completion of the compression tests, all the specimens were
instantly water-quenched to maintain the microstructure
formed at the high testing temperature.
ZEISS MERLIN scanning electron microscopy (SEM) was
used to observe the microstructures. These compressed
specimens were sectioned longitudinally and cut the speci-
mens with a size of 7  7  2 mm using wire cutting. The
observation surfaces of the specimens were ground and pol-
ished mechanically. For electron backscattered diffraction
(EBSD) analysis, the surfaces to be observed were electro-
polished for 30 s in a mixture of 30% HNO3 þ 70% CH3OH, and
the sample was cooled using liquid nitrogen, to keep it at Fig. 3 e Relationship between ln½lnð1 XMDRX Þ and lnt
about 30
C, clean the surface with alcohol immediately after under different intervals time.
polishing. In an analysis of EBSD, the voltage of SEM was
30 kV, and the given steps size of 0.3 and 0.55 mm. The scanned
data were processed subsequently by Channel 5 software. The ln_ε and 1/T are shown as Fig. 4. The values of A8, n, m and
grain boundary, the values of grain size and recrystallization QMDRX were determined as 1.345  108, 1.503, 0.355 and
volume fraction were thus obtained. 168801.720, respectively.
Furthermore, the values of A8，n，m and QMDRX were
substituted into the Yada equation and the expression of t0.5.
3. MDRX model of GH5188 superalloy
The kinetic equation of MDRX for GH5188 superalloy could be
expressed as follows:
3.1. MDRX kinetic model

 0:587 
t
To quantify the MDRX fraction (XMDRX), it was necessary to XMDRX ¼ 1  exp  0:693 (5)
t0:5
establish a grain model. Senuma et al. [15] improved the
Avrami equation and proposed a new calculation equation  
168801:72
describing the grain size and fraction. This equation was t0:5 ¼ 1:345  108 ε1:503 ε_ 0:355 exp (6)
RT
called the Yada equation, as described as Eqs. (1) and (2).

 k  3.2. MDRX grain size model
t
XMDRX ¼ 1  exp  0:693 (1)
t0:5
Yada and Senuma [16] determined the MDRX grain
  size(dMDRX), as described as Eq. (7).
QMDRX
t0:5 ¼ A8 εn ε_ m exp (2)  
RT Q4
dMDRX ¼ A9 εn ε_ m dh0 exp  (7)
RT
where k is material-related constants, t is the intervals time;
t0.5 is the time to 50% MDRX fraction; A8, n and m are con- According to the literature [17], to a material, the grain
stants; QMDRX is the activation energy for MDRX; R is the gas size was only related to the holding time. Therefore, the
constant (R ¼ 8.314 J/(mol,K)); T is the temperature. MDRX grain size model directly used the holding time instead
Taking a natural logarithm on both sides for the Yada of the original grain size. The MDRX grain size was described
equation. The following Eq. (3) will be obtained. as Eq. (8).
 
ln½  lnðð1  XMDRX ÞÞ ¼ kðln t  ln t0:5 Þ þ ln 0:693 (3) Q4
dMDRX ¼ A9 ε_ p th exp  (8)
RT
Therefore, the Yada equation k could be obtained by the
regression method. The average slope between where A9, p and h are constants; d0 is the initial grain size; Q4 is
ln½½lnð1 =ð1 XMDRX ÞÞ and ln t is shown as Fig. 3. The average the corresponding activation energy.
value of k was 0.587. Take logarithms for both sides of Eq. (9):
To obtain the relationship between t0.5 and the deforma-
ln dMDRX ¼ ln A9 þ p ln_ε þ h ln t þ Q4 =RT (9)
tion parameter, taking the logarithm of both sides of the
expression of t0.5, the following Eq. (4) could be obtained. The average grain size under different deformation con-
ditions measured by the experiments was substituted into Eq.
ln t0:5 ¼ ln A8 þ n ln ε þ m ln_ε þ QMDRX =RT (4)
(9) to carry out linear regression analysis. The plot of dMDRX
Then, values of A8, n, m and QMDRX could be obtained by the versus ε_ p th expð  Q4 =RTÞ, ln_ε, ln t and 1/RT are shown as Fig. 5.
relationship between the ln t0:5 and εn ε_ m expðQMDRX =RTÞ, ln ε, ln_ε The value of A9, p, h and Q4 were determined as 1.150  104,
and 1/T, and the plot of ln t0:5 versus εn ε_ m expðQMDRX =RTÞ, ln ε, 0.442, 0.513 and 69423.500, respectively.
1182 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

Fig. 4 e Fitting of coefficients in the MDRX kinetic equation of GH5188 superalloy: (a) Relationship between ln t0:5 and
εn ε_ m expðQMDRX =RTÞ; (b) Relationship between ln t0:5 and lnε; (c) Relationship between ln t0:5 and ln_ε; (d) Relationship between
ln t0:5 and 1/T.

The values of A9, p, h and Q4 were substituted into Eq. (8). continuous variables and the relationship between the two
The grain size equation of MDRX for GH5188 superalloy could was linear in the scatter plot, there was a linear correlation
be expressed as follows: trend between the two. The coefficient of productemoment
correlation (such as Spearman correlation coefficient (r)) was
dMDRX ¼ 1:150  104 ε_ 0:442 t0:513 expð69423:5 = RTÞ (10) used to express the degree of correlation between the two. The
Spearman correlation coefficient was an index to quantita-
3.3. Verification of MDRX model
tively describe the degree of linear correlation between two
variables and expressed as Eq. (11).
To verify the accuracy of the MDRX model, the predicted values
P  
of MDRX fraction and grain size under different deformation i ðxi  xÞ yi  y
r ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi (11)
parameters were calculated. Then the predicted values were P 2 P 2
i ðxi  xÞ $ i yi  y
compared with the experimental values. Fig. 6 shows the
comparison between the predicted values of MDRX fraction where The i-th (1  i) values of the variable are represented by
and grain size of GH5188 superalloy and the experimental xi and yi , respectively. The average value of the sum of all
values. It can be seen from Fig. 6 that the predicted values of valued variables is represented by x and y, respectively.
MDRX fraction and grain size were in good agreement with the In statistical product and service solutions 24.0 (SPSS 24.0),
experimental values, and the average relative errors were the bivariate method in correlation analysis was used to
2.35% and 4.22%, respectively. Therefore, the MDRX model of realize the correlation analysis between the two. Then the
GH5188 superalloy can predict the MDRX behavior well. correlation results of different deformation parameters are
shown in Table 1. It can be seen from Table 1 that the defor-
mation parameters were all positively correlated with the
4. Results and discussion MDRX fraction. The MDRX fraction tended to increase when
the deformation parameters increased. By descending order
4.1. Correlation analysis of correlation was: intervals time, holding time, temperature
and strain rate.
According to the literature [18], bivariate correlation analysis
was used to describe the degree of connection between two 4.2. Flow feature analysis
different variables. It reflected the variation degree of the
other variable after controlling the value of one variable. Maric Fig. 7 shows the true stress-true strain curve of GH5188 su-
et al. [19] proposed that when the two variables were normal peralloy under different deformation conditions. It can be
 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9 1183

Fig. 5 e Fitting of coefficients in the MDRX grain size equation of GH5188 superalloy: (a) Relationship between dMDRX and
ε_ p th expð  Q4 =RTÞ; (b) Relationship between dMDRX and ln_ε; (c) Relationship between dMDRX and ln t; (d) Relationship between
dMDRX and 1/RT.

seen from Fig. 7(a) that with the increase of holding time, the was enhanced, so the true stress decreased with the lower
grains began to grow, and then stabilize to a maximum. The strain rate [21]. It can be seen from Fig. 7(d) that under the
grain size had a certain effect on the DRX and MDRX pro- same initial-stage compression parameters, the first segment
cesses. The true stress curve decreased with the increase of of the curve kept going; after some time, the true stress of the
the holding time. There is no significant difference between two-stage compression decreased significantly with the in-
the curve at 120s and 180s, indicating that the grains had fully crease of the intervals time, Which shows that the intervals
grown during the 60e120 s and a higher holding time had little time had a more obvious promotion effect on the MDRX.
effect on the grain size. It can be seen from Fig. 7(b) that as the
temperature increased, the true stress decreased substan- 4.3. Effect of holding time
tially. With the higher deformation temperature, the migra-
tion and annihilation of dislocations weaken the work Fig. 8 shows the influence of different holding times on the
hardening effects, and the dynamic softening behavior was microstructure evolution of GH5188 superalloy. The distinc-
enhanced, so the true stress decreased substantially [20]. It tion between different grain structures is determined by Grain
can be seen from Fig. 7(c) that with the increased strain rate, Orientation Spread (GOS). When GOS < 2
, the grain structures
the true stress increased. The low strain rate provided suffi- are recrystallized grains, Channel 5 software is shown in blue;
cient deformation time, and the dynamic softening behavior when 2
 GOS  10
, the grain structures are sub-grain grains,

Fig. 6 e Comparison between predicted and tested values of MDRX fraction (a) and grain size (b) for GH5188 superalloy.
1184 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

Table 1 e Correlation coefficient of different deformation parameters on MDRX fraction of GH5188 superalloy.
Parameter Holding time Temperature Strain rate Intervals time
Correlation coefficient 0.256 0.198 0.102 0.320

Channel 5 software is represented in yellow; when 10
< GOS, temperatures. Chen et al. [23] proposed that the higher the
the grain structures are deformed grains, Channel 5 software temperature, the greater the driving force of DRX. The more
is represented in red. According to the Channel 5 software, the recrystallized grain nuclei and the mobility of grain boundary
average grain size at t1 ¼ 60, 120, 180 s were 5.5462, 8.6699 and increased with the increased temperature, which was
9.5812 mm, respectively. With the holding time increase, the conducive to the process of MDRX. It can be seen from Fig. 9(a)
recrystallization fraction was 92%, 22.4% and 24.6%, respec- that incomplete DRX occurred in the microstructure of
tively. It can be seen from Fig. 8(aec) that with the holding GH5188 superalloy at a low temperature. The recrystallized
time increase, the growth of MDRX was fast and slow. As the grains were distributed around the large grains in a "necklace"
holding time increases, the stored energy of the grain in- shape, and there were many small angle grain boundaries in
creases, so the MDRX grain size increases. Yuan et al. [22] the original grains and sub-structured grains. With the in-
proposed that twinning was more likely to occur in smaller crease of deformation temperature, the rate of recrystalliza-
grains so that smaller grains could be obtained, and it was tion nucleation increased considerably. The recrystallized
more conducive to MDRX. However, The original grains were grains not only appear at the grain boundaries of the original
segmented and rotated mainly through the cross of twin grains but also nucleate at the twin boundaries, which can be
boundaries to form recrystallized nuclei and grow up during found in some sub-structured grains, as shown in Fig. 9(b).
the intervals time to reduce the dislocation strength. There- The movement of high angle grain boundaries was not
fore, after a certain holding time, the grain size growth rate noticeable at 1080
C, and more small angle grain boundaries
slows and the recrystallization fraction changes little. were observed in the sub-structured grains, which indicated
that the softening mechanism of GH5188 superalloy at low
4.4. Effect of temperature temperature was mainly recovery. The MDRX mechanism at
1130
C was dominated by the migration of the grain bound-
Fig. 9 illustrates the influence of different deformation tem- ary. During the DRX process, the nucleated grains grew up at
peratures on the microstructure evolution of GH5188 super- this temperature, and there was a twin structure in some
alloy. It can be seen from Fig. 9(aec) that the recrystallization recrystallized grains. When the temperature reached 1180
C,
fraction and grain size are increasing with higher the grain boundaries were substantially flatter, indicating that

Fig. 7 e The true stress-true strain curves of GH5188 superalloy at various deformation conditions: (a) holding time; (b)
deformation temperature; (c) strain rate; (d) intervals time.
 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9 1185

Fig. 8 e The microstructure of GH5188 superalloy obtained at T ¼ 1080
C, ε1 ¼ 0.4, ε_ ¼ 1 s¡1, t2 ¼ 1 and ε2 ¼ 0.1 with t1 ¼ 60
(a), 120 (b) and 180 s (c). The blue represents recrystallized grains, the yellow represents sub-structured grains and the red
represents deformed grains.

the recrystallization mechanism had been changed. The the movement of sub-structured grain boundaries. The
higher temperature provides energy for the formation of recrystallization nucleation was conducted by the way of sub-
twins, which separates the original grains, and swallows the structured grains growth, and the growth continued to occur
adjacent deformed matrix and sub-structured grains through in the process of MDRX.

Fig. 9 e The microstructure of GH5188 superalloy obtained at t1 ¼ 60 s, t2 ¼ 5 s, ε1 ¼ 0.4, ε_ ¼ 5 s¡1 and ε2 ¼ 0.1 with T ¼ 1080
(a), 1130 (b), 1180
C (c).
1186 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

4.5. Effect of strain rate indicating that the recrystallized grain had not fully grown at
this time.
Zhang et al. [24] proposed that at a high strain rate, most of the According to the Channel 5 software, the S3 twin bound-
dislocations generated have not yet slipped and piled up. The aries at ε_ ¼ 1, 5 and 10 s1 were 12.3%, 26.3%, and 40%,
dislocation density was high so that the grains had a high respectively. It can be inferred that the twins in the MDRX
orientation gradient, which provided conditions for the structure were by grain growth. In general, the formation of
nucleation of recrystallization, while the grains that had not twin boundaries during recrystallization mainly includes
grown in time would continue to grow in the gap time. Fig. 10 interface rotation mechanism, twin defect mechanism and
illustrates the influence of different strain rates on the accidental growth mechanism. The formation of twin
microstructure evolution of the GH5188 superalloy. It can be boundaries during recrystallization mainly includes the
seen from Fig. 10(aec) that with a higher strain rate, the mechanism of interface rotation, twin defect and accidental
recrystallization fraction and average grain size are both growth. The growth of twins was related to the crystal struc-
increasing. Zhang et al. [25] proposed that the strain rate was ture and the ratio of the interface energy between twin and
higher, the nucleation number was more, while the newly normal grain boundaries. For the GH5188 superalloy with a
generated grains got less time for growing up. Then, due to face-centered cubic crystal structure, the twin nucleation
more multiplication of dislocation density, the stored energy occurred mainly through the change of orientation of adjacent
became more. As shown in Fig. 10(a), there were large-sized grains when the MDRX fraction was small. When the MDRX
original grains in the structure, and more small angle grain fraction was large, twins were caused by accidental
boundaries were observed in the grains. Twin bands existed in misalignment that occurred during atomic diffusion.
the smaller recrystallized grains and were abundant. In
Fig. 10(b), the deformed grains almost completely dis- 4.6. Effect of intervals time
appeared, replaced by sub-structured grains with smaller
orientation differences within the grains. With the increased Fig. 11 illustrates the influence of different intervals time on
strain rate, the strain energy was higher, and the nucleation the microstructure evolution of GH5188 superalloy. It can be
rate of dynamic recrystallization was faster. The recrystalli- seen from Fig. 11(aec) that the MDRX fraction increased with
zation core was formed by the sub-crystalline nucleation the intervals time increased. The intervals time had an
mechanism, the deformed grains gradually become sub- obvious promotion effect on the occurrence of MDRX.
structured grains. Therefore, the deformed grains almost Although the intervals time was very short, the softening
completely disappear and were replaced by sub-structured degree of the material was high. Quan et al. [26] proposed that
grains. The recrystallized grains distributed at grain bound- grain growth played a major role in this process. The rate of
aries and twin boundaries grew up and expanded into the grain boundary migration determines the softening rate of the
grain interior. In Fig. 10(c), the recrystallized fraction was the material, and DRX usually required nucleation first, so the
largest, but there was still a large size of sub-structured grains, MDRX fraction increased with the intervals time increased.

Fig. 10 e The microstructure of GH5188 superalloy obtained at t1 ¼ 60 s, T ¼ 1080
C, ε1 ¼ 0.4, t2 ¼ 5 s and ε2 ¼ 0.1 with ε_ ¼ 1
(a), 5 (b) and 10 s¡1 (c).
 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9 1187

Fig. 11 e The microstructure of GH5188 superalloy obtained at t1 ¼ 60 s, T ¼ 1080
C, ε1 ¼ 0.4, ε_ ¼ 5 s¡1 and ε2 ¼ 0.1 with
t2 ¼ 1 (a), 2 (b) and 5 s (c).

According to the Channel 5 software, the S3 twin boundaries it was found that the MDRX fraction increases with the
at t2 ¼ 1, 2 and 5 s were 23.7%, 19.2% and 12.7%, respectively. increase of deformation parameters. By descending order
With the increase of the intervals time, the proportion of S3 of correlation was: intervals time, holding time, tempera-
twin boundaries gradually decreased. It can be inferred that ture and strain rate.
the disappearance of annealing twins in the GH5188 superal- 2. With the higher holding time, the growth of MDRX is fast
loy was related to the grain growth process of the MDRX and slow, the MDRX grain size is larger.
process. This phenomenon could be explained from two per- 3. With the increase of deformation temperature, the rate of
spectives. On the one hand, the migration of recrystallized recrystallization nucleation increased considerably. The
grains nucleating near the twin boundaries through the recrystallized grains not only appear at the grain bound-
interface leads to the annihilation of the annealing twin aries of the original grains but also nucleate at the twin
boundary, which reduced the total interfacial energy of the boundaries.
system. So the fraction of the twin boundary decreased with 4. With the higher strain rate, the average grain size is
the increase in the recrystallized grain size, such as the increasing. Twins were formed in the process of grain
migration of recrystallized grains [27]. On the other hand, the growth.
formation of annealing twins requires time, and the grain 5. With the extension of the intervals time, the recrystallized
boundaries moved fast under high-temperature conditions, so nuclei grow more fully and the grain size is more consis-
there was not enough time, such as the formation of anneal- tent. A decrease in the fraction of the twin boundary with
ing twins [28]. the increase of the intervals time.

5. Conclusions Data availability

The microstructures evolutions of the GH5188 superalloy The data supporting the findings of this study are available
during the MDRX were investigated. The kinetic model and from the corresponding authors upon request.
grain size model of MDRX for GH5188 superalloy were estab-
lished. The good agreement between the predicted values and
experimental values was verified. The effects of holding time, Credit authorship contribution statement
temperature, strain rate and intervals time on the MDRX were
studied. Conceptualization: Dahai Liu, Jingdong Chen. Methodology:
Dahai Liu, Jingdong Chen, Haorui Chai. Validation: Haorui
1. Through the correlation analysis of the influence of Chai, Zhi Li, Yihang Jiang; Data curation: Dahai Liu, Jingdong
different deformation parameters on the softening effect, Chen. Writing - original draft: Dahai Liu, Jingdong Chen.
1188 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9

Writing - review and editing: Jingdong Chen, Haorui Chai, [11] Lyu FG, Shang JQ, Sun B, Teng JF. Numerical simulation for
Wenqing Qiu, Zhenghua Guo. All authors have read and agree thermal creep forming of GH188 alloy. J Plast Eng
to the published version of the manuscript. 2017;24:187e91. https://doi.org/10.3969/j.issn.1007-
2012.2017.02.031.
[12] He L, Zhang RY, Zhao GY, Cheng YB, Jin CB, Guo ZH.
Constitutive model of GH5188 superalloy based on BP neural
Funding network. Spec Cast Nonferrous Alloy 2021;41:223e6. https://
doi.org/10.15980/j.tzzz.2021.02.020.
This work was supported by the National Natural Science [13] Ouyang LX, Luo R, Gui YW, Cao Y, Chen LL, Cui YJ, et al. Hot
Foundation of China, China (Award Number: 52065045), the deformation characteristics and dynamic recrystallization
Key Research and Development Program of Jiangxi Province, mechanisms of a CoeNi-based superalloy. Mater Sci Eng A
2020;788:139638. https://doi.org/10.1016/j.msea.2020.139638.
China (Award Number: 20202BBEL53011) and Major Scientific
[14] Dong YP, Wang CY, Song XJ, Chen YH, Zhang MC, Wang SY.
and Technological Research and Development Program of Study on plastic deformation behavior of GH5188 superalloy.
Jiangxi Province, China (Award Number: 20194ABC28001). Forg Stamp Technol 2013;38:116e21. https://doi.org/10.3969/
j.issn.1000-3940.2013.06.025.
[15] Senuma T, Yada H, Matsumura Y, Futamura T. Structure of
Declaration of Competing Interest austenite of carbon steels in high speed hot working
processes. Tetsu-To-Hagane/Journal Iron Steel Inst Japan
1984;70:2112e9. https://doi.org/10.2355/
The authors declare that they have no known competing
tetsutohagane1955.70.15_2112.
financial interests or personal relationships that could have [16] Yada H, Senuma T. Computer modeling for predicting
appeared to influence the work reported in this paper. microstructure and mechanical properties of steels. Bull
Japan Inst Met 1990;29:430e9. https://doi.org/10.2320/
materia1962.29.430.
references [17] Hodgson PD, Gibbs RK. A mathematical model to predict the
mechanical properties of hot rolled C-Mn and microalloyed
steels. ISIJ Int 1992;32:1329e38. https://doi.org/10.2355/
[1] Rao KBS, Castelli MG, Ellis JR. On the low cycle fatigue isijinternational.32.1329.
deformation of haynes 188 superalloy in the dynamic strain [18] Heinen A, Valdesogo A. Spearman rank correlation of the
aging regime. Scr Metall Mater 1995;33:1005e12. https:// bivariate student t and scale mixtures of normal
doi.org/10.1016/0956-716X(95)00320-U. distributions. J Multivar Anal 2020;179:104650. https://
[2] Chen LJ, Liaw PK, He YH, Benson ML, Blust JW, Browning PF, doi.org/10.1016/j.jmva.2020.104650.
et al. Tensile hold low-cycle fatigue behavior of cobalt-based [19] Maric M, de Haan E, Hogendoorn SM, Wolters LH,
HAYNES® 188 superalloy. Scr Mater 2001;44:859e65. https:// Huizenga HM. Evaluating statistical and clinical significance
doi.org/10.1016/S1359-6462(00)00702-8. of intervention effects in single-case experimental designs:
[3] Wang YQ, Suneson M. Oxidation behavior of Hf-modified an SPSS method to analyze univariate data. Behav Ther
aluminide coatings on Haynes-188 at 1050
C. Surf Coatings 2015;46:230e41. https://doi.org/10.1016/j.beth.2014.09.005.
Technol 2013;215:7e15. https://doi.org/10.1016/ [20] He DG, Lin YC, Chen MS, Li L. Kinetics equations and
j.surfcoat.2012.07.091. microstructural evolution during metadynamic
[4] Lee WS, Kao HC. High temperature deformation behaviour of recrystallization in a nickel-based superalloy with d phase. J
Haynes 188 alloy subjected to high strain rate loading. Mater Alloys Compd 2017;690:971e8. https://doi.org/10.1016/
Sci Eng A 2014;594:292e301. https://doi.org/10.1016/ J.JALLCOM.2016.08.096.
j.msea.2013.11.076. [21] He DG, Lin YC, Wang LH. Microstructural variations and
[5] Wang ZY, Zhang WQ, Shen X. Microstucture and kinetic behaviors during metadynamic recrystallization in a
homogenization of GH188 superalloy. J Shenyang Ligong nickel base superalloy with pre-precipitated d phase. Mater
Univ 2011;30:12e6. Des 2019;165:107584. https://doi.org/10.1016/
[6] Gao YW, Dong JX, Yao ZH, Zhang MC, Niu YJ. Microstructure J.MATDES.2019.107584.
characteristics and microstructure evolution during hot and [22] Yuan Y, Gu Y, Cui CY, Osada T, Zhong ZH, Tetsui T, et al.
cold working process of GH5188 superalloy. Rare Met Mater Influence of Co content on stacking fault energy in NieCo
Eng 2017;46:2922e8. base disk superalloys. J Mater Res 2011;26:2833e7. https://
[7] Pyczak F, Bauer A, Go € ken M, Neumeier S, Lorenz U, doi.org/10.1557/jmr.2011.346.
Oehring M, et al. Plastic deformation mechanisms in a crept [23] Chen L, Long HJ, Wang JF, Mao TQ, Li F, Zhang YJ.
L12 hardened Co-base superalloy. Mater Sci Eng A Metadynamic recrystallization behavior of a novel austenite
2013;571:13e8. https://doi.org/10.1016/j.msea.2013.02.007. heat resistant stainless steel. J Cent South Univ Technol
[8] Favre J, Fabregue D, Yamanaka K, Chiba A. Modeling 2016;47:3995e4000. https://doi.org/10.11817/j.issn.1672-
dynamic recrystallization of L-605 cobalt superalloy. Mater 7207.2016.12.005.
Sci Eng A 2016;653:84e92. https://doi.org/10.1016/ [24] Zhang HB, Zhang KF, Jiang SS, Zhou HP, Zhao CH, Yang XL.
J.MSEA.2015.12.003. Dynamic recrystallization behavior of a g0 -hardened nickel-
[9] Lei H, Hu XT, Song YD. Study and application of creep based superalloy during hot deformation. J Alloys Compd
constitutive model of GH188 superalloy. Mech Sci Technol 2015;623:374e85. https://doi.org/10.1016/
Aerosp Eng 2011;30:1623e8. https://doi.org/10.13433/ j.jallcom.2014.11.056.
j.cnki.1003-8728.2011.10.025. [25] Zhang F, Liu D, Yang Y, Liu C, Zhang Z, Wang H, et al.
[10] Anil Kumar V, Gupta RK, Narayana Murty SVS, Prasad AD. Investigation on the meta-dynamic recrystallization
Hot workability and microstructure control in behavior of Inconel 718 superalloy in the presence of d phase
Co20Cr15W10Ni cobalt-based superalloy. J Alloys Compd through a modified cellular automaton model. J Alloys
2016;676:527e41. https://doi.org/10.1016/ Compd 2020;817:152773. https://doi.org/10.1016/
J.JALLCOM.2016.03.186. j.jallcom.2019.152773.
 j o u r n a l o f m a t e r i a l s r e s e a r c h a n d t e c h n o l o g y 2 0 2 1 ; 1 5 : 1 1 7 9 e1 1 8 9 1189

[26] Quan GZ, Shi Y, Zhao L, Mao YP, Zhou J. Dynamic Mater Sci Eng A 1995;203:408e14. https://doi.org/10.1016/
recrystallization behavior of 42CrMo steel induced by 0921-5093(95)09821-6.
thermal deformation. Trans Mater Heat Treat 2009;30:71e5. [28] Zhang HB, Zhang KF, Zhou HP, Lu Z, Zhao CH, Yang XL. Effect
https://doi.org/10.13289/j.issn.1009-6264.2012.10.029. of strain rate on microstructure evolution of a nickel-based
[27] Horton D, Thomson CB, Randle V. Aspects of twinning and superalloy during hot deformation. Mater Des 2015;80:51e62.
grain growth in high purity and commercially pure nickel. https://doi.org/10.1016/j.matdes.2015.05.004.