Anda di halaman 1dari 6

Materials Characterization 101 (2015) 20–25

Contents lists available at ScienceDirect

Materials Characterization
journal homepage:

Solidification behavior of a Ni-based single crystal CMSX-4 superalloy

solidified by downward directional solidification process
F. Wang a,b, D. Ma b,⁎, J. Zhang a, S. Bogner b, A. Bührig-Polaczek b
State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Shaanxi, Xi'an 710071, PR China
Foundry Institute, RWTH University, Aachen 52072, Germany

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

Article history: The solidification behavior of a CMSX-4 superalloy single crystal solidified using the downward directional solid-
Received 29 October 2014 ification process has been investigated by using the directional solidification quenching technique. An analysis of
Received in revised form 19 December 2014 transverse sections within the mushy zone yielded the solidification path, the evolution of solid volume fraction
Accepted 5 January 2015
(fs) with the temperature decrease and the solidification sequence of eutectic pool. fs monotonically increased
Available online 7 January 2015
with decreasing temperature. Within the first 20% of the solidification interval, fs increased by about 60%. This in-
creasing ratio then significantly reduced. It became flat at the final portion up to the freezing of (γ + γ′) eutectic.
Superalloy The Lever rule and Scheil model cannot describe the actual evolution of fs observed in the experiment; however,
Directional solidification the Bower–Brody–Flemings model can, to some extent, depict this evolution. The solidification path was de-
Microstructure scribed as: primary crystallization of γ phase; termination of solidification with a eutectic reaction. The solidifi-
Phase transformation cation sequence of the eutectic pool commenced with a fine (γ + γ′) structure in the center of interdendritic
Quenching regions or adjacent to γ phases, then progressed spatially and developed a petal-like coarse (γ + γ′) structure.
Following this sequence, the γ′ particles precipitate from the γ phase between the impinging coarse petal-like
γ′ phases.
© 2015 Elsevier Inc. All rights reserved.

1. Introduction Thermo-Calc software [13,14]. The calculation is normally based on sim-

plified solidification models, which cannot precisely demonstrate the
Ni-based superalloys are widely used to fabricate single-crystal tur- actual solidification behavior. Several researchers [15,16] elicit the so-
bine blades of aero-engines as well as land-based power generators lidification behavior by using three-dimensional spatial (3-D) recon-
[1–4]. Owing to the increased demand for thermodynamic efficiency structions from a series of two-dimensional (2-D) planar transverse
of the engines, the operating temperatures of the engines are being con- sections of the directional solidified samples. Although 3-D images can
tinuously elevated. To meet these severe operating temperatures, the provide more information compared to 2-D observations, the solidifica-
high temperature mechanical strength of single-crystal blades must be tion sequence can only be deduced from these reconstructed results, but
enhanced. An excellent high temperature performance can be achieved is difficult to observe dynamically. In addition to this, no one unified
by adding refractory elements such as Re [5–10], which can increase conclusion has hitherto been reached about the solidification sequence
strength and creep resistance via solid solution strengthening within of the (γ + γ′) eutectic. Some investigations [17,18] indicate that the
the γ matrix and optimize the size of γ′ precipitates by acting as a re- onset of solidification of the eutectic is the crystallization of a coarse γ′
tarder [11,12]. However, under non-equilibrium solidification condi- phase via the peritectic reaction of L + γ → γ′, followed by a fine lamel-
tions, the sluggish diffusivity of these elements in the solid-state leads lar (γ + γ′) eutectic forming by epitaxial growth. Other studies [19–21]
to significant segregation and the formation of cored eutectic micro- suggest that a eutectic reaction of L → γ + γ′ is the original solidifica-
structures. Consequently, a subsequent heat treating process is essen- tion stage of the eutectic, which freezes the fine lamellar (γ + γ′) mor-
tial. To properly control the solidification and heat treating processes, phology. The reaction only occurs at the dendrite-melt interface. The
knowledge of the solidification behavior is both theoretically and indus- coarse γ′ particles then subsequently grow epitaxially from the fine
trially important. structure.
Over the past few decades, some studies have been carried out on The downward directional solidification process (DWDS) is a high
the solidification behavior of the conventional upward directional solid- thermal gradient method for growing single crystal superalloys. Previ-
ified superalloys. Some of these are based on a calculation using the ous studies [22,23] indicate that the segregation profile in the single
crystal castings is different from the conventional upward directional
⁎ Corresponding author. solidification process. The difference will lead to a distinct solidification
E-mail address: (D. Ma). behavior. However, no investigation has hitherto been conducted into
1044-5803/© 2015 Elsevier Inc. All rights reserved.
F. Wang et al. / Materials Characterization 101 (2015) 20–25 21

this field. Consequently, in this study we combined the differential scan- sample. This zone possesses a typical dendrite array. The evolution of
ning calorimetry technique with the quenched directional solidified mi- fs as a function of the temperature reduction (Tonset − T) is shown in
crostructure to investigate the evolution of the microstructure. By Fig. 1(f); where Tonset is the onset temperature of solidification, and T
sequential transversal sectioning the mushy zone, the variation of is the temperature of isothermal solidification which corresponds to dif-
solid volume fractions was determined as a function of the temperature ferent transverse sections of the solidification interval. Both tempera-
decrease. Simultaneously, we also ascertained the detailed solidification ture values can be determined from the DSC results (shown in
path. Fig. 2(b)) and the thermal gradient. Fig. 1(f) illustrates a smooth mono-
tonic increase in fs with decreasing temperature within the mushy zone.
2. Material and methods At the beginning of the solidification, γ phase grows so quickly that fs
sharply increases. Within the first 20% of the solidification interval, fs in-
The material used in the present investigation was the CMSX-4 su- creases by about 60%. In this freezing range, γ dendrite can grow rela-
peralloy (Ni–5.6Al–9.0Co–6.5Cr–0.6Mo–3.0Re–6.5Ta–6.0W, wt %). The tively freely without great impediment from the adjacent dendrites.
experiments were performed in an in-house modified DWDS furnace, The increasing ratio of fs then significantly reduces because of impinge-
schematically shown in reference [22]. During the experiment, the ment on the dendrites. Therefore, the curve of fs vs (Tonset − T) becomes
CMSX-4 superalloy was overheated to 1773 K in a crucible and covered flat at its final portion until the (γ + γ′) eutectic solidifies.
with hollow Al2O3 particles (1–3 mm in diameter) acting as a dynamic The evolution of fs during the solidification process can be described
baffle. A ceramic tube (1 mm thick, 200 mm high, and 9 mm inside di- by the Scheil model, the Lever rule and the Bower–Brody–Flemings
ameter), having a single crystal seed of CMSX-4 superalloy with [100] (BBF) model. These theoretical models were actually developed from
crystal orientation in the longitudinal direction was inserted into the the binary system. Ma et al. [24] suggested that these models can be
melt. The other end of the tube was wrapped and sealed using a nickel analogously applied in the multicomponent system such as a superal-
foil in order to prevent the penetration of the dynamic baffle. When the loy. The Scheil equation can be expressed as Eq. (1):
foil melted, the alloy melt flowed into the mold and made contact with
the seed. Each seed was instrumented with a thermocouple of type B C s ¼ kC 0 ð1−f s Þ ð1Þ
(Pt–30%Rh/Pt–6%Rh) located along its center line. The thermocouple
was inserted into an alumina tube which was sealed at one end. This where, Cs is the solute concentration in the solid phase; k is the equiva-
sealed end was located 65 mm from the chill end. Thermocouple read- lent partition coefficient for the entire system which was determined by
ings were recorded each second during the entire process. To obtain the Ma et al. [25] for the multicomponent system; C0 is the initial solute
thermal gradient ahead of the L/S interface (GL), the cooling rate (Ṫ) was concentration. This model is based on the hypothesis that the solute
first calculated from the cooling curve's slope at the liquidus tempera- can completely diffuse into the liquid phase and there is no diffusion
ture. The thermal gradient (GL) was then calculated by dividing the into the solid phase.
cooling rate by the withdrawal rate (V): GL = Ṫ/V. When a part of the The Lever rule can be expressed as Eq. (2) and is applicable to an ab-
seed had melted, the tube was elevated at a withdrawal rate of solute equilibrium system; that is, in this system the solute can
0.4 mm min− 1 (without Ar cooling) and quenched in a water bath. completely diffuse into the liquid and the solid phases.
The thermal gradient at the liquid/solid (L/S) interface was estimated
C s ¼ kC 0 =½1−f s ð1−kÞ ð2Þ
as 13.1 K mm−1. After quenching, the bars were sectioned longitudinal-
ly (parallel to the growth direction) and transversally (perpendicular to
the growth direction). The characterization of microstructural evolution The BBF-model is suitable for a directional solidification system in
and the variation in the solid volume fraction (fs) was obtained by se- which the solute diffusion into the liquid and solute accumulation
quentially studying transversally milled sections from the tip of the den- ahead of the L/S interface are considered. The simplified equation is
drite array to the bottom of the mushy zone using metallurgical analysis shown as:
software (Image Pro Plus). The material removed in each milling cycle h i
from 0.05 to 0.2 mm depended on the detail requirement. The etching C s ¼ C 0 V c =V þ kð1−V c =V Þð1− f s Þ ð3Þ
solution used in the metallographic examination was 1.5 g CuSO4:
40 mL HCl: 20 mL C2H5OH. The morphological evolution of the where, V is the dendrite growth velocity; Vc is the critical velocity of pla-
(γ + γ′) eutectic phase was examined on a Zeiss Ultra-55 scanning elec- nar–cellular morphological transition.
tron microscope (SEM). Energy dispersive X-ray analysis (EDX) was According to the functional relationship between the temperature
carried out to examine the element distribution within the and concentration in the monophase solidification and the
interdendritic structure. undercooling:
A Netzsch 409 differential scanning calorimeter (DSC) was used to
ðT onset −T Þ=ΔT 0 ¼ ðC s =k−C 0 Þ=ðC 0 =k−C 0 Þ or ð4Þ
predict the phase transformation temperatures. Approximately
100 mg of CMSX-4 superalloy was employed in the analysis. During
T onset −T ¼ ½kΔT 0 =ð1−kÞðC s =kC 0 −1Þ: ð5Þ
the experiment, a dynamic high purity Ar atmosphere was applied to
minimize the potential oxidizing effects. The samples were subjected
The forms of these theoretical models can be modified as follows:
to the following heating/cooling cycles: Room temperature → heating
(20 K min−1) to 1000 °C → isothermal holding at 1000 °C for T onset −T ¼ ½kΔT 0 =ð1−kÞf1=½1− f s ð1−kÞ−1g ðLever ruleÞ ð6Þ
10 min → heating (5 K min− 1) to 1400 °C → isothermal holding at
h i
1400 °C for 10 min → cooling (5 K min−1) to 1000 °C → isothermal k−1
T onset −T ¼ ½kΔT 0 =ð1−kÞ ð1− f s Þ −1 ðScheilÞ ð7Þ
holding at 1000 °C for 10 min → cooling (20 K min− 1) to room
T onset −T ¼ ½kΔT
nh0 =ð1−kÞ i o
3. Results and discussion  V c =kV þ ð1−V c =V Þð1−f s Þ −1 ðBB FÞ ð8Þ

3.1. Evolution of solid volume fraction (fs) where, ΔT0 is the solidification interval.
Using Eqs. (6), (7) and (8), the evolution of fs was calculated and
Fig. 1(e) and (a)–(d) characterizes the morphologies of the mushy comparatively plotted with the experimental data in Fig. 2(a). The pa-
zone on the longitudinal and transverse sections of the quenched rameters used in the calculation were ΔT0 = 67 K, Vc =
22 F. Wang et al. / Materials Characterization 101 (2015) 20–25

(a) (b) (c) (d)



Fig. 1. Longitudinal section and transverse sections at different positions in the mushy zone, and the evolution of the solid volume fraction fs vs temperature-reduction (Tonset − T).

0.035 mm min−1 and k = 0.28; which are given in the previous study model cannot describe the actual evolution of fs in the experiment,
[25]. The plots illustrate that the Scheil model and Lever rule deviate which is also reported in other investigations [26–30]. By considering
considerable from the experimental data. The BBF model can character- the effect of the diffusion of alloying elements into the liquid and the
ize the evolution of fs in the initial and final portions of the solidification solute accumulation in front of L/S interface, the BBF model is closer to
curve obtained from the experiment. However, this model cannot de- the actual solidification conditions. This model can, to some extent, pre-
scribe the increasing tendency of fs between about 50% and 80% of the dict the evolution of fs. However, apart from diffusion, the convection is
experimental data. As mentioned above, the theoretical models used stronger in the DWDS process than that in the conventional solidifica-
to characterize fs are normally based on specific hypotheses. Subject to tion process [31]. This has an effect on the distribution of solutes and
actual solidification conditions, perfect system-equilibrium is difficult further influences the evolution of fs. Convection is believed to be the
to achieve. In addition to this, complete diffusion into the liquid appears factor leading to the deviation between the BBF model and the experi-
to be acutely difficult. For this reason, the Lever rule and the Scheil mental data.

Fig. 2. (a) Comparison of theoretical, calculated fs vs temperature reduction with the experiment results and (b) DSC thermal analysis of the experimental CMSX-4 superalloy.
F. Wang et al. / Materials Characterization 101 (2015) 20–25 23

(a) (b)



Fig. 3. Secondary electron images focusing on the periphery of the secondary dendrite lobes and interdendritic regions. (a)–(c) Different morphologies of (γ + γ′) eutectic; (d) the details
of the region between impinging coarse petal-like (γ + γ′) phases; (e) the morphology of (γ + γ′) core formed in the center of the interdendritic area.

3.2. Solidification sequence to the γ dendrites, while Ta and Al segregate into the remaining liquid of
the interdendritic regions. This leads to a separately identifiable
The extent of microsegregation of alloying elements has a direct interdendritic constituent containing both the γ and γ′ phases. In gener-
bearing on the resultant solidification sequence [18]. Fig. 2(b) shows al, these morphologies are aggregated together and are characterized as
the typical differential scanning calorimetry (DSC) curves of the exam- a eutectic. Several morphologies in the interdendritic regions are shown
ined CMSX-4 superalloy under the stabilized heating and cooling pro- in Fig. 3. Similar to the conventional morphology of the upward direc-
gram. Phase transition temperatures were obtained from the cooling tional solidified eutectic pool, there are three distinct morphologies:
curve. In terms of the exotherm, solidification commences at Tonset = (i) fine (γ + γ′) morphology; (ii) coarse petal-like γ′ morphology;
1651.5 K (1378.5 °C), which is defined as the solidification onset tem- (iii) γ′ precipitates in γ matrix between the impinging coarse petal-
perature. A second sharp transition was observed at T = 1592.2 K like γ′ phases. Apart from this, another much finer (γ + γ′) morphology
(1319.2 °C), which is related to the onset of freezing of the non- observed ahead of the coarse γ′ morphology (illustrated as A in
equilibrium phases. Solidification terminates at the solidus tempera- Fig. 3(a) and (c)) is believed to form during the quenching process
ture, Ts = 1583.7 K (1310.7 °C). The undercooling is approximately and not in the solidification process.
9 K for the onset of solidification.
The solidification path begins with the crystallization of the γ phase,
which constitutes the dendritic structures. This transition can be Table 1
expressed as: Experimentally measured composition within the different interdendritic structure.

Structure Elements, wt %
L→γ þ L1 1592:2 KbT ≤1651:5 K :
Al Cr Co Mo Ta W Re Ni

As the solidification proceeds, the γ phase gradually grows and the Fine γ′ 6.4 3.8 7.9 0.4 11 4.1 0.9 Bal.
Coarse γ′ 7.0 2.9 7.4 0.4 12.8 3.9 0.2 Bal.
solid volume fraction significantly increases. Cr, Co, Re and W segregate
24 F. Wang et al. / Materials Characterization 101 (2015) 20–25

Fig. 4. The solidification model based on the experimental results.

Fig. 3(e) shows the formed core of (γ + γ′) eutectic, which was ob- forming elements, which leads to the thicken of the γ′ particles in the
served prior to the coarse γ′ morphology. This indicates that after the final solidification. Therefore, the contents of Al, Ta in the coarse γ′ par-
crystallization of the γ phase, the (γ + γ′) core, which is the fine ticle are higher than that in the fine γ′ particle, while the amount of Cr,
(γ + γ′) morphology, was crystallized by means of a eutectic reaction. Co, W and Re are less, as shown in Table 1.
The formation sites were not only adjacent to the γ phase (correspond- Furthermore, from thermodynamic considerations, it follows that
ing to Fig. 3(c)), but also in the center of the interdendritic regions (cor- the γ′ particles (shown in Fig. 3(d)), in the γ matrix between the im-
responding to Fig. 3(a)). At this time, the volume fraction of the pinging coarse petal-like γ′ phases, are supersaturated precipitates oc-
remaining liquid is approximately 8%. This reaction can be presented as: curring after solidification since, if the fine (γ + γ′) structure is the
 product of a eutectic reaction, the subsequent solidification cannot be
L1 → γ þ γ þ L2 1583:7 KbT ≤1592:2 K : followed by another two-phase eutectic reaction [17]. Moreover, it is
found that Fig. 3(b) has a strikingly different eutectic morphology
These findings provide evidence that the solidification sequence of from that in Fig. 3(a) and (c). Heckl et al. [16] suggested that the 2-D
eutectic pools does not start with a coarse γ′ phase but instead, the morphology of a eutectic pool depends on both the orientation of the
growth direction should start with the fine (γ + γ′) structure. The dif- examined plane and the distance from the center of the 3-D eutectic.
ferent eutectic solidification sites then spread three dimensionally in The sectioned position in the 3-D eutectic for Fig. 3(b) should be in
different directions until they grow together or meet the boundary of the petal-like γ′ regions.
a neighboring γ dendrite. Another aspect that confirms the coarse γ′
phase as the last constituent to form is the spatial location of cavity 4. Conclusions
formed during quenching process observed in the interdendritic re-
gions, as shown in Fig. 3(a) and (c). The cavity is an excellent indication In this paper, the solidification behavior of DWDS single crystal of
of the location of the last liquid to solidify. These images reveal that cav- CMSX-4 superalloy was investigated. fs monotonically increased with
ity appears at locations between coarse γ′ particles and adjacent γ den- decreasing temperature. Within the first 20% of the solidification inter-
drite lobe periphery. There is a marked absence of such cavity in the val, fs increased by about 60%. This increasing ratio of fs then significantly
vicinity of the fine (γ + γ′) morphology, which suggests that the fine reduced due to the impingement of the dendrites. The curve of fs vs tem-
(γ + γ′) morphology is formed earlier. In addition to this, this solidifica- perature reduction became flat at the final portion up to the freezing of
tion sequence can also be verified by the liquidus temperatures of the (γ + γ′) eutectic. The Lever rule and Scheil model cannot describe the
fine γ′ morphology and the coarse γ′ morphology which were calculat- actual evolution of fs observed in the experiment; however, the BBF
ed in terms of the measured composition (listed in Table 1) by using the model can, to some extent, depict this evolution. The solidification
Thermo-Calc software. The calculated results suggest that the liquidus path was determined as: primary crystallization of γ phase with the
temperature of the fine γ′ lies around 9 °C above the liquidus tempera- transition L → γ + L1 (1592.2 K b T ≤ 1651.5 K); termination of solidi-
ture of the coarse γ′ eutectic structure. Strictly speaking, the composi- fication with a eutectic reaction L1 → (γ + γ′) + L2 (1583.7 K b T ≤
tions are averaged for different structures and the calculated liquidus 1592.2 K). The solidification sequence of the eutectic pool commenced
temperatures are thus not an accurate value for the specific beginning with a fine (γ + γ′) structure in the center of interdendritic regions or
of the eutectic solidification. Nevertheless, they are a good estimation adjacent to γ dendrites, then progressed spatially and developed a
to verity the solidification direction from the fine γ′ into coarse γ′. petal-like coarse (γ + γ′) structure. Following this sequence, the γ′ par-
Enclosed between the solidified γ dendrite, the eutectic nucleation ticles precipitate from the γ phase between the impinging coarse petal-
site will be controlled by compositional fluctuations reaching the eutec- like γ′ phases.
tic valley first [16]. Nucleation of the remaining liquid can occur on the
surface of γ dendrites or directly from the liquid, which strongly de- Acknowledgments
pends on the local undercooling of the remaining liquid.
Based on the experimental results, the solidification sequence can be This research was supported by German Research Foundation
precisely demonstrated in the following model; illustrated in Fig. 4. On through Grant No. MA 2505/3-1. One of the authors (FW) would like
commencing, the remaining liquid is enriched with sufficient γ′- to acknowledge the China Scholarship Council for supporting his stays
forming elements (Al, Ta) and has a large undercooling. Solidification in Germany. The authors would like to acknowledge the aid of Maria
proceeds faster and the lamellar structure is finer. The γ′ phase in the Schaarschmidt, Elke Schaberger-Zimmermann and Elke Breuer.
eutectic structure has not enough time to solutionize much more Al,
Ta elements. Meanwhile, the γ-phase forming elements Cr, Co, W and
Re cannot be liberated to the utmost. As this reaction proceeds, due to
the formation of fine (γ + γ′) morphology, the latent heat is gradually [1] G. Xie, L. Wang, J. Zhang, L.H. Lou, Orientational dependence of recrystallization in
released, which causes a reduction in the nucleating undercooling and an Ni-base single-crystal superalloy, Scr. Mater. 66 (2012) 378–381.
[2] Z.Y. Liu, H. Qi, Effect of processing parameters on crystal growth and microstructure
retards the eutectic reaction. The γ′ particles have sufficient time to ab- formation in laser power deposition of single-crystal superalloy, J. Mater. Process.
sorb much more Al, Ta elements, and maximally release the γ-phase Technol. 216 (2015) 19–27.
F. Wang et al. / Materials Characterization 101 (2015) 20–25 25

[3] H.U. Hong, J.G. Yoon, B.G. Choi, I.S. Kim, C.Y. Jo, On the mechanism of secondary re- [17] N. D'Souza, H.B. Dong, Solidification path in third-generation Ni-based superalloys,
action zone formation in a coated nickel-based single-crystal superalloy containing with an emphasis on last stage solidification, Scr. Mater. 56 (2007) 41–44.
ruthenium, Scr. Mater. 69 (2013) 33–36. [18] N. D'Souza, H.B. Dong, An analysis of solidification path in the Ni-base superalloy,
[4] M. Gell, D.N. Duhl, A.F. Giamei, The development of single crystal superalloy turbine CMSX-10 K, in: R.C. Reed, K.A. Green, P. Caron, T.P. Gabb, M.G. Fahrmann, E.S.
blades, in: J.K. Tien, S.T. Wlodek, H. Morrow, M. Gell, G. Maurer (Eds.), Superalloys Huron, S.A. Woodard (Eds.), Superalloys 2008, TMS, PA, 2008, pp. 261–269.
1980, ASM, Metals Park, OH, 1980, pp. 205–214. [19] X.Y. Zhu, S.N. Zhang, L.Y. Xu, J. Bi, Z.Q. Hu, C.X. Shi, Superalloys with low segregation,
[5] D. Blavette, P. Caron, T. Khan, An atom probe investigation of the role of rhenium ad- in: S. Reichman, D.N. Duhl, G. Maurer, S. Antolovich, C. Lund (Eds.), Superalloys
ditions in improving creep resistance of Ni-base superalloys, Scr. Metall. 20 (1986) 1988, TMS, PA, 1988, pp. 703–712.
1395–1400. [20] N. Warnken, D. Ma, M. Mathes, I. Steinbach, Investigation of eutectic island forma-
[6] L. Jiarong, T. Dingzhong, L. Riling, L. Shizhong, W. Zhongtang, Effects of Rhenium on tion in SX superalloys, Mater. Sci. Eng. A 413–414 (2005) 277–286.
creep rupture life of a single crystal superalloys, J. Mater. Sci. Technol. 15 (1999) [21] S.M. Seo, J.H. Lee, Y.S. Yoo, C.Y. Jo, H. Miyahara, K. Ogi, Solute redistribution during
53–57. planar and dendritic growth of directionally solidified Ni-base superalloy CMSX-
[7] A.C. Yeh, S. Tin, Effect of Ru and Re additions on the high temperature flow stresses 10 K, in: R.C. Reed, K.A. Green, P. Caron, T.P. Gabb, M.G. Fahrmann, E.S. Huron, S.A.
of Ni-base single crystal superalloys, Scr. Mater. 52 (2005) 519–524. Woodard (Eds.), Superalloys 2008, TMS, PA, 2008, pp. 261–269.
[8] A. Heckl, S. Neumeier, M. Göken, R.F. Singer, The effect of Re and Ru on γ/γ′ micro- [22] F. Wang, D.X. Ma, J. Zhang, S. Bogner, A. Bührig-Polaczek, A high thermal gradient
structure, γ-solid solution strengthening and creep strength in nickel-base superal- directional solidification method for growing superalloy single crystals, J. Mater.
loys, Mater. Sci. Eng. A 528 (2011) 3435–3444. Process. Technol. 214 (2014) 3112–3121.
[9] A.F. Giamei, D.L. Anton, Rhenium additions to a Ni-base superalloy: effects on micro- [23] F. Wang, D.X. Ma, J. Zhang, A. Bührig-Polaczek, Investigation of segregation and den-
structure, Metall. Trans. 16 (1985) 1997–2005. sity profiles in the mushy zone of CMSX-4 superalloys solidified during downward
[10] S. Neumeier, F. Pyczak, M. Göken, The influence of ruthenium and rhenium on the and upward directional solidification processes, J. Alloys Compd. 620 (2015) 24–30.
local properties of the γ- and γ′-phase in nickel-base superalloys and their conse- [24] D.X. Ma, P.R. Sahm, Single-crystal solidification of a Ni-base superalloy SRR 99 Part
quences for alloy behavior, in: R.C. Reed, K.A. Green, P. Caron, T.P. Gabb, M.G. 3: evolution of microstructure during directional solidification, Int. J. Mater. Res.
Fahrmann, E.S. Huron, S.A. Woodard (Eds.), Superalloys 2008, ASM, Metals Park, 87 (1996) 640–644.
OH, 2008, pp. 109–119. [25] D.X. Ma, U. Grafe, Dendrite growth and microsegregation during directional solidifi-
[11] W.Z. Wang, T. Jin, J.L. Liu, X.F. Sun, H.R. Guan, Z.Q. Hu, Role of Re and Co on micro- cation: an analytical model and experimental studies on the superalloy CMSX-4, Int.
structures and γ′ coarsening in single crystal superalloys, Mater. Sci. Eng. A 479 J. Cast Met. Res. 13 (2000) 85–92.
(2008) 148–156. [26] D. Ma, P.R. Sahm, Microstructural evolution during directional solidification of su-
[12] A. Mottura, N. Warnken, M.K. Miller, M.W. Finnis, R.C. Reed, Atom probe tomogra- peralloy CMSX-6, Mater. Sci. Forum 215–216 (1996) 223–228.
phy analysis of the distribution of rhenium in nickel alloys, Acta Mater. 58 (2010) [27] D. Ma, U. Grafe, Microsegregation in directionally solidified dendritic-cellular struc-
931–942. ture of superalloy CMSX-4, Mater. Sci. Eng. A 270 (1999) 339–342.
[13] S.M. Seo, J.H. Lee, Y.S. Yoo, C.Y. Jo, H. Miyahara, K. Ogi, A comparative study of the γ/ [28] Z.X. Shi, J.X. Dong, M.C. Zhang, L. Zheng, Solidification characteristics and segrega-
γ′ eutectic evolution during the solidification of Ni-base superalloys, Metall. Mater. tion behavior of Ni-based superalloy K418 for auto turbocharger turbine, J. Alloys
Trans. A 42 (2011) 3150–3159. Compd. 571 (2013) 168–177.
[14] S.H. Kang, Y. Deguchi, K. Yamamoto, K. Ogi, M. Shirai, Solidification process and be- [29] N. D'Souza, M. Lekstrom, H.J. Dai, B.A. Shollock, H.B. Dong, Quantitative characteriza-
havior of alloying elements in Ni-based superalloy Inconel 718, Mater. Trans. 45 tion of last stage solidification in nickel base superalloy using enthalpy based meth-
(2004) 2728–2733. od, Mater. Sci. Technol. 23 (2007) 1085–1092.
[15] H.T. Pang, H.B. Dong, R. Beanland, H.J. Stone, C.M.F. Rae, P.A. Midgley, G. Brewster, N. [30] C. Walter, B. Hallstedt, N. Warnken, Simulation of the solidification of CMSX-4,
D'Souza, Microstructure and solidification sequence of the interdendritic region in a Mater. Sci. Eng. A 397 (2005) 385–390.
third generation single-crystal nickel-base superalloy, Metall. Mater. Trans. A 40 [31] F. Wang, D.X. Ma, J. Zhang, L. Liu, J.P. Hong, S. Bogner, A. Bührig-Polaczek, Micro-
(2009) 1660–1669. structural evolution of aluminium-copper alloys during the downward directional
[16] A. Heckl, R. Rettig, S. Cenanovic, M. Göken, R.F. Singer, Investigation of the final solidification process, Int. J. Mater. Res. 105 (2014) 168–176.
stages of solidification and eutectic phase formation in Re and Ru containing
nickel-base superalloys, J. Cryst. Growth 312 (2007) 2137–2144.