- Email: [email protected]
Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy
Materials Science & Engineering A 639 (2015) 732–738
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy X.W. Li a, T. Liu a, L. Wang a,b,n, X.G. Liu a, L.H. Lou a, J. Zhang a,b a b
Superalloys Division, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
art ic l e i nf o
a b s t r a c t
Article history: Received 14 November 2014 Received in revised form 12 May 2015 Accepted 13 May 2015 Available online 27 May 2015
Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy has been investigated by the scanning electron microscope, X-ray computed tomography and electron probe microanalyzer. With the increase of the carbon content, MC carbides evolve from octahedral to well-developed dendrite, which promotes the formation of microporosity. Moreover, the volume fraction of porosity increases in the experimental alloys after solution heat treatment. As a result, the increase in the size of MC carbides and the porosity has a detrimental effect on the low temperature and high stress creep behavior of the alloys. The specimen crept at 850 °C and 586 MPa with the carbon content of 430 ppm shows the shortest rupture life due to the largest primary creep strain. However, the creep behavior of the alloy at 1120 °C and 140 MPa gets better as the carbon content increases from 88 to 430 ppm. TCP phase is observed near the fracture surfaces of the alloys, which explores as a potential cause for the creep rupture. However, the formation of TCP phase is effectively suppressed for decreasing segregation of the alloying elements, which results in the improvement of the creep life in the alloy with 430 ppm carbon at 1120 °C and 140 MPa. & 2015 Elsevier B.V. All rights reserved.
Keywords: Carbides Microporosity X-ray computed tomography Creep property Nickel-base superalloys
1. Introduction Single crystal (SX) superalloys have been widely used in advanced turbine engines for hot components. To further enhance their high temperature capability, the level of refractory elements such as tungsten and rhenium in the advanced SX superalloys has been gradually increased. However, there is a great tendency for the formation of grain defects (strays, low angle grain boundaries, freckles, etc.) in these alloys [1–3]. Therefore, trace elements such as carbon and boron are reintroduced into SX superalloys for grain boundary reinforcement [4,5] and reducing casting defects during solidification [6–8]. The formation of grain defects in the SX superalloys with the addition of carbon can be effectively suppressed [6,7]. However, conflicting results involving the role of carbon on the creep property of SX alloy have been reported. It has been found that the creep life in the carbon modified SX alloy at 850 °C/430 MPa is improved, but the creep life at 1050 °C/165 MPa decreases [9]. Contrary results were observed by Liu et.al [10]. Kong and Wasson also reported that carbon has a detrimental effect on creep life under different conditions [11,12]. It is necessary to clarify the n
Corresponding author. Tel.: þ 86 24 23971276; fax: þ 86 24 23971712. E-mail address: [email protected] (L. Wang).
http://dx.doi.org/10.1016/j.msea.2015.05.039 0921-5093/& 2015 Elsevier B.V. All rights reserved.
influence of carbon on the creep life in SX superalloys. Creep rupture is associated with a number of microstructural features, including MC carbides, porosity and TCP formation. It has been confirmed that morphology of MC carbides transforms from blocky to script-like and the volume fraction of γ/γ′ eutectic is reduced by carbon additions [13–15]. However, there are controversial opinions concerning the role of carbon during the formation of porosity. Liu and Chen [8,16] found that the formation of porosity is suppressed by the addition of carbon due to volume expansion of MC carbides. However, an increase volume fraction of porosity is observed by Al-Jarba and Culter [7,17] with carbon additions. Moreover, the H-pore formed during homogenization has been observed, and the volume fraction of H-pore exhibits the same level as that of S-pore formed during solidification [18–20]. The two kinds of pores have not been distinguished. In this study, the three dimensional information of S-pore and H-pore has been investigated by X-ray computed tomography (XCT). The topologically close packed (TCP) phase is prone to form during the rafting creep [21,22]. The brittle and needle-like phase depletes strengthening elements from the matrix and is susceptible to crack during creep deformation. Retardation of TCP formation in carbon modified alloy RR2072 has been observed [23], while Qin and Kong [24,25] reported that TCP phase initiates in the vicinity of MC carbides. Therefore, the aim of the present study is to illustrate the effect
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Table 1 The actual carbon contents of alloys in the present experiment.
C (ppm)
733
Table 2 Chemical composition of MC carbides (wt%).
SX-1
SX-2
SX-3
Phase
C
Ti
Cr
Co
Ni
Mo
Ta
88
220
430
MC
13
2
2
2
7
3
71
of carbon on the microstructure involving the porosity, the TCP phase and the creep properties in the SX alloys.
2. Experimental procedure A 3rd generation nickel-base single crystal (SX) superalloy DD33 with 4% Re (mass fraction, %) was used in the present experiment. The master alloys with three different amounts of carbon were directionally solidified (DS) into ϕ16 200 mm SX bars by the high rate solidification (HRS) process. The actual carbon contents of alloys were given in Table 1. Longitudinal orientation of all SX bars was within 10° deviating from [001] orientation. The samples for comparison of as-cast microstructures (γ/γ′ eutectic, porosity and MC carbides) were sectioned perpendicularly to the DS direction and taken from the same location on each SX bar in order to minimize the effect of solidification conditions. To investigate 3-D features of solidification pores (S-pore) and homogenization pores (H-pore) by XCT, the as-cast and solution heat treated samples were machined into cubes of 1 1 1 mm3. Partitioning coefficient of alloying elements in alloys with different carbon contents was measured by electron probe microanalyzer (EPMA). Point scans in dendrite cores and interdendritic regions were performed. The partitioning coefficient, ki, of each element was identified by dividing the weight percent of each element at the dendrite cores by that in the interdendritic regions. Ten points from each area were averaged to reduce variation. Following the same heat treatment (1335 °C/10 h, AC þ1150 °C/ 4 h, AC þ870 °C/24 h, AC), the first group of samples were exposed at 1100 °C for 1000 h to examine the formation of TCP phase. The second group of heat treated samples were machined into creep specimens with gauge diameter of 5 mm and gauge length of 25 mm along the [001] orientation. Creep tests were carried out at 870 °C/586 MPa and 1120 °C/140 MPa, using FC-20 high temperature creep-testing machines. The microstructure of crept specimens was examined using scanning electron microscope (SEM).
3. Results 3.1. As-cast microstructure Varying the carbon contents from 88 to 430 ppm, the alloys have similar dendrite arm spacing, but the morphology of MC
Table 3 Volume fractions of porosity in as-cast alloys measured by XCT and OM. C (ppm)
88
220
430
XCT OM
0.077 0.002 0.03 70.01
0.08 7 0.004 0.05 7 0.03
0.127 0.006 0.09 70.05
Fig. 2. Distribution of porosity diameter in as-cast alloys with different carbon contents.
carbides changes from blocky (Fig. 1a) to script-like (Fig. 1b). These results agree well with previous studies [13–15,26]. Further SEM observations on the deep-etched sample show that the script-like MC carbides actually adopt a dendritic structure (Fig. 1c). These dendritic carbides are rich in Ta and Ti, and have a small amount of Mo, Co and Cr (Table 2). The volume fraction of porosity measured by XCT and OM is given in Table 3. The volume fraction of S-pore measured by XCT is 0.07%, 0.08% and 0.12% in the alloys of SX-1, SX-2 and SX-3, respectively. The XCT results are higher, compared with the porosity metallographically measured by OM. The porosity measured by the two different methods are in agreement within the accuracy. Increasing volume fractions of porosity is observed in high carbon modified alloys. Distribution of porosity diameter is illustrated in Fig. 2. The maximum diameter of porosity increases from 30 to 46 μm, with the carbon content increasing from 88 to 430 ppm. Fig. 3 shows SEM and XCT images of the porosity in the alloy SX-1. SEM observation indicates that porosities are located in the interdendritic regions (Fig. 3a). XCT observation shows that there
Fig. 1. Morphology of MC carbides in the alloy with different carbon contents: (a) 88 ppm, (b) and (c) 430 ppm.
734
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Fig. 3. Morphology of porosity in the as-cast alloy SX-1: (a) SEM image, and (b) XCT results.
are two kinds of pores: the sphere pores and the irregular pores (Fig. 3b). Partitioning coefficient of alloying elements in as-cast alloys with different carbon contents is shown in Fig. 4. It can be seen that the addition of carbon decreases the segregation of Re and W, while it has a little influence on the segregation of elements such as Al and Ta. The results are identical to the findings of Tin et al. [27]. 3.2. Heat treated microstructure
Fig. 4. Partitioning coefficient of alloying elements in alloys with different carbon contents.
Fig. 5 presents that the volume fraction of porosity increases in the experimental alloys after solution heat treatment. The increment of porosity after solution heat treatment was 0.04%, 0.06% and 0.04%, which are named as volume fractions of H-pores. As shown in Fig. 6, the probability of small sphere-shaped pores increases after solution heat treatment. Morphology of porosity in the alloy SX-1 after solution heat treatment at 1335 °C for 10 h is illustrated in Fig. 7. Small sphere-shaped pores form and S-pores grow round. 3.3. Long-term exposed microstructure
Fig. 5. Volume fractions of porosity in as-cast and heat treated alloys with different carbon contents.
The formation of TCP phase during high temperature exposure at 1100 °C for 1000 h varies significantly in the alloys with different carbon contents. TCP precipitates are occasionally observed in the alloys with 88 and 220 ppm carbon as given in Fig. 8a and b. In the alloy with 430 ppm carbon, no TCP phase is observed (Fig. 8c). These TCP precipitates are rich in W, Re, Cr, and Co as listed in Table 4. Based on the chemical compositions and morphologies, these TCP precipitates are confirmed to be μ phase. Decomposition of MC carbides is not observed in the experimental alloys after long term exposure (Fig. 8c). 3.4. Creep behavior
Fig. 6. Distribution of porosity diameter in as-cast and heat treated alloy SX-1.
Fig. 9 gives the curves of time vs. creep strain for the experimental alloys at 850 °C and 586 MPa. Fig. 9a shows all curves consist of three stages. Significant primary creep strain develops in all alloys and the onset of primary creep occurs in a different manner as indicated in Fig. 9b. The specimens with 430 ppm carbon show the largest primary creep strain with the shortest rupture life of 276 h. The specimens containing 88 and 220 ppm carbon have shorter primary creep strain and longer creep life of 421 and 364 h, respectively. SEM images of the longitudinal section of the ruptured specimen reveal that cracks are normal to the applied stress direction, as shown in Fig. 10. In all the specimens, cracks preferentially initiate at MC carbides in the interdendritic regions. Cracks in the
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
735
Fig. 7. Morphology of porosity in the heat treated alloy SX-1: (a) SEM images and (b) XCT results.
Fig. 8. Microstructure of alloys with different carbon contents after exposure at 1100 °C for 1000 h: (a) 88 ppm, (b) 220 ppm and (c) 430 ppm. Table 4 Chemical composition of TCP phase (wt%). Phase
Re
W
Mo
Cr
Co
Ni
μ
45.6
18.2
5.2
4.8
6.9
18.9
dendritic regions are not observed. Number of microcracks in the alloy with 430 ppm carbon is larger than that in the alloy with 88 ppm carbon. Fig. 11 shows the creep curves of three alloys tested at 1120 °C/ 140 MPa. The curves are very similar (Fig. 11a). Fig. 11b presents the plots of creep time vs. creep strain rate. The creep rates stay relatively constant for the first 60 h and increase more rapidly for alloys with 88 ppm carbon. The high carbon-containing specimens
show the smallest elongation of 14% as well as the longest rupture life of 91 h, while the specimens with 88 and 220 ppm carbon have the similar elongation of around 50% and creep life of 61 and 75 h, respectively. Fig. 12 shows SEM images of longitudinal sections, confirming the microcracks grow perpendicularly to the applied stress direction. Microcracks in both dendritic and interdendritic regions are observed in the experimental alloys with 88 and 220 ppm carbon, as indicated in Fig. 12a and b. Only creep cavities in the dendritic regions in the alloys with 430 ppm carbon are observed in Fig. 12c. Fig. 12d, e and f gives the microcracks or creep cavities originated from TCP phases in the dendritic areas, while Fig. 12h, i and j presents the microcracks in the interdendritic area originated from MC carbides.
Fig. 9. Creep curves of the alloys with different carbon contents at 850 °C/586 MPa.
736
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Fig. 10. SEM micrographs of longitudinal sections near the fracture surfaces of samples with different carbon contents crept at 850 °C/586 MPa: (a) 88 ppm, (b) 220 ppm and (c) 430 ppm.
Fig. 11. Creep curves of alloys with different carbon contents at 1120 °C/140 MPa.
4. Discussion 4.1. Microstructure According to Niyama criterion [28],
ΔP = − ηβ
V G
∫0
xc
fL K
⋅d x =
ηβΔT G/ V α
(
−2
)
(1)
where ΔP is pressure drop across the mushy zone, η is the viscosity of the liquid, fL is the volume fraction of liquid, K is the permeability of the liquid, G is temperature gradient and V is cooling velocity, ΔT is temperature difference across the mushy zone, α is a constant parameter. The total solidification shrinkage is defined as β ¼ (ρS ρL)/ρL, where ρL and ρS is the density of liquid and solid, respectively. The permeability of the liquid K proposed by Carlson and Madison [29,30] is defined as follows:
K = K0
f L3 (1 − fL )2
(2)
where initial permeability of the liquid K0 is proportional to equivalent diameter of the interdendritic regions. Integrating Eq. (1) substituted by expression (2), pressure drop across the mushy zone is written as follows:
ΔP0
ηβΔT K0
−2
(G/ V ) ηβΔT = (G/ V ) K
ΔP =
f
∫1 L
2
( ) 1 − fL fL
(
d fL = ΔP0 fL −
1 fL
− 2 ln fL
)
−2
0
permeability of the liquid, K0. Considering similar solidification parameters of three SX alloys, with the carbon content increasing from 88 to 430 ppm, the morphology of MC carbides evolves from blocky to dendritic-like (Fig. 1). The equivalent diameter of the interdendritic regions is decreased by these well-developed dendritic carbides. Therefore, the pressure drop increases with the reduction of the initial permeability of the liquid (K0) in the alloy with 430 ppm carbon (Eq. (3)). Although small pores may disappear due to the lattice expansion of MC carbides [8], the larger size of dendritic carbides is obviously an important reason for the high level of porosity in the alloy with 430 ppm carbon. It was found that the imbalance cross-diffusion of alloy elements between Al, Ta and Re, resulted in the Kirkendall pore, which may be the reason for the H-pore formation during solution heat treatment [19]. In the present study, although the carbon content is different, the primary chemical composition of SX alloys is identical. Thus, the level of imbalanced cross-diffusion of alloy elements during the same solution heat treatment is similar. TCP phases are prone to form during long term exposure or creep deformation because of the segregation of alloying elements, which is generally undesirable in SX superalloys [21–23]. The reduction of segregation of alloying elements with the increase of the carbon content gives the reason why the TCP phase formation in alloys with 430 ppm is suppressed (Fig. 4). On the other hand, creep deformation accelerates diffusion and thus promotes the formation of TCP during high temperature creep (Fig. 12f).
(3)
Porosity is prone to form when the pressure drop of molten alloy becomes larger or the remaining liquid is not enough to compensate the volume shrinkage during the last stage of solidification. The pressure drop of liquid is inversely proportional to the
4.2. Creep Most work has revealed that creep performance at low temperature and high stress is sensitive to the primary creep strain, and the primary creep behavior is controlled by shearing of γ′
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
737
Fig. 12. Microstructures of longitudinal section near the fracture surfaces in alloys with different carbon contents: (a), (d) and (h) 88 ppm; (b), (e) and (i) 220 ppm; (c), (f) and (j) 430 ppm; (d), (e) and (f) dendritic region; (h), (i) and (j) interdendritic region.
phase [31,32]. Therefore, the high carbon content alloy crept at 850 °C/586 MPa with the largest primary creep strains exhibits the shortest creep life, as shown in Fig. 9. It was reported that the formation of nano-size M23C6 carbides may prevent the dislocation movement and enhance the creep properties [10,34], but these nano-size M23C6 carbides were not observed in the experimental alloys. On the contrary, MC carbides act as stress concentration sites and induce premature microcracks during the low temperature creep deformation (Fig. 10c). Hence, the larger size of dendritic carbides in the interdendritic regions in the alloys with 430 ppm carbon may accelerate the local creep damage and have a detrimental effect on the creep behavior at 850 °C and 586 MPa [9]. Although microcracks induced by the dendritic carbides and porosity in the interdendritic regions were observed (Fig. 12), TCP phases near the fracture surfaces were observed in all the crept specimens at 1120 °C/140 MPa (Fig. 12), which would be expected to have the greatest impact on the high temperature creep behavior [33,34]. Usually, the precipitation of μ particles will lead to the depletion of strengthening elements from the matrix. The softened matrix was prone to deform, which in turn accelerates the formation of μ particles [35]. In addition, needle shaped μ phase was
reported to induce creep cavities and microcracks in the dendritic regions, which accelerates creep rupture [22,36]. Therefore, suppression of TCP phase during 1120 °C/140 MPa creep is a main reason for the longer creep life in the alloy with 430 ppm carbon. On the other hand, less microcracks in the interdendritic regions in the alloy with 220 and 430 ppm carbon are observed (Fig. 12a–c), which indicates less impact of dendritic carbides in the interdendritic regions on the creep life at 1120 °C/140 MPa in the alloy.
5. Conclusion Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy has been investigated. The following conclusions can be given: 1. With the increase of carbon content from 88 to 430 ppm, MC carbides evolve from blocky to script-like. The dendritic carbides promote the formation of S-pore due to the increase of the pressure drop. The volume fraction of porosity increases in the experimental alloys after solution heat treatment. The level of H-pores in SX alloys with various carbon contents is
738
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
identical. 2. The high carbon content alloy crept at 850 °C/586 MPa exhibits the shortest creep life with the largest primary creep strain. The dendritic carbides and pores are detrimental to the creep rupture behavior. 3. The TCP phase is expected to have significant impact on the creep behavior at 1120 °C/140 MPa. Formation of the TCP phase during high temperature exposure is suppressed by the addition of carbon, which results in the longer high temperature creep life in the alloy with 430 ppm carbon.
Acknowledgment This work was supported by the National Natural Science Foundation of China (51201164) and the National High Technology Research and Development Program of China (2012AA03A511).
References [1] T.M. Pollock, W.H. Murphy, E.H. Goldman, D.L. Uram, J.S. Tu, Superalloys 1992 (1992) 125–134. [2] P. Auburtin, T. Wang, S.L. Cockcroft, A. Mitchell, Metall. Mater. Trans. B 31 (2000) 801–811. [3] J.W. Aveson, P.A. Tennant, et al., Acta Mater. 61 (2013) 5162–5171. [4] Q.Z. Chen, C.N. Jones, D.M. Knowles, Mater. Sci. Eng. A 385 (2004) 402–418. [5] D.M. Shah, A. Cetel, Superalloys 2000 (2000) 295–304. [6] S. Tin, T.M. Pollock, W.T. King, Superalloys 2000 (2000) 201–210.
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
K.A. Al-Jarba, G.E. Fuchs, Mater. Sci. Eng. A 373 (2004) 255–267. Q.Z. Chen, Y.H. Kong, C.N. Jones, D.M. Knowles, Scr. Mater. 51 (2004) 155–160. Q.Z. Chen, N. Jones, D.M. Knowles, Acta Mater. 50 (2002) 1095–1112. L.R. Liu, T. Jin, et al., Mater. Sci. Eng. A 385 (2004) 105–112. Y.H. Kong, Q.Z. Chen, D.M. Knowles, J. Mater. Sci. 39 (2004) 6993–7001. A.J. Wasson, Ph.D. thesis, University of Florida, Gainesville, 2010. A.J. Wasson, G.E. Fuchs, Mater. Charact. 74 (2012) 11–16. K.A. Al-Jarba, G.E. Fuchs, JOM 56 (2004) 50–55. E.R. Cutler, A.J. Wasson, G.E. Fuchs, Scr. Mater. 58 (2008) 146–149. L.R. Liu, T. Jin, et al., Mater. Lett. 58 (2004) 2290–2294. E.R. Cutler, A.J. Wasson, G.E. Fuchs, J. Cryst. Growth 311 (2009) 3753–3760. D.L. Anton, A.F. Giamei, Mater. Sci. Eng. 76 (1985) 173–180. B.S. Bokstein, A.I. Epishin, et al., Scr. Mater. 57 (2007) 801–804. T.L.A.I. Epishina, Phys. Met. Metallogr. 115 (2014) 21–29. M.S.A. Karunaratne, C.M.F. Rae, R.C. Reed, Metall. Mater. Trans. A 32 (2001) 2409–2421. C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. Q.Z. Chen, D.M. Knowles, Metall. Mater. Trans. A 33 (2002) 1319–1330. Y.H. Kong, Q.Z. Chen, Mater. Sci. Eng. A 366 (2004) 135–143. X.Z. Qin, J.T. Guo, Mater. Sci. Eng. A 543 (2012) 121–128. X.W. Li, L. Wang, X.G. Liu, J.S. Dong, L.H. Lou, J. Mater. Sci. Technol. 30 (2014) 1296–1300. S. Tin, T.M. Pollock, Mater. Sci. Eng. A 348 (2003) 111–121. E. Niyama, T. Uchida, M. Morikawa, S. Saito, Cast Met. J. 6 (1981) 16–27. K.D. Carlson, C. Beckermann, Metall. Mater. Trans. A 40 (2009) 163–175. J. Madison, J.E. Spowart, et al., Metall. Mater. Trans. A 43 (2012) 369–380. V.A. Vorontsov, C. Shen, Y. Wang, D. Dye, C.M.F. Rae, Acta Mater. 58 (2010) 4110–4119. F. Diologent, P. Carbon, Mater. Sci. Eng. A 385 (2004) 245–257. R.C. Reed, D.C. Cox, C.M.F. Rae, Mater. Sci. Eng. A 448 (2007) 88–96. J.B. Le Graverend, J. Cormier, S. Kruch, F. Gallerneau, J. Mendez, Metall. Mater. Trans. A 43 (2012) 3988–3997. L. Wang, D. Wang., et al., Mater. Charact. 104 (2015) 81–85. S.G. Tian, M.G. Wang, T. Li, B.J. Qian, J. Xie, Mater. Sci. Eng. A 527 (2010) 5444–5451.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy X.W. Li a, T. Liu a, L. Wang a,b,n, X.G. Liu a, L.H. Lou a, J. Zhang a,b a b
Superalloys Division, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
art ic l e i nf o
a b s t r a c t
Article history: Received 14 November 2014 Received in revised form 12 May 2015 Accepted 13 May 2015 Available online 27 May 2015
Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy has been investigated by the scanning electron microscope, X-ray computed tomography and electron probe microanalyzer. With the increase of the carbon content, MC carbides evolve from octahedral to well-developed dendrite, which promotes the formation of microporosity. Moreover, the volume fraction of porosity increases in the experimental alloys after solution heat treatment. As a result, the increase in the size of MC carbides and the porosity has a detrimental effect on the low temperature and high stress creep behavior of the alloys. The specimen crept at 850 °C and 586 MPa with the carbon content of 430 ppm shows the shortest rupture life due to the largest primary creep strain. However, the creep behavior of the alloy at 1120 °C and 140 MPa gets better as the carbon content increases from 88 to 430 ppm. TCP phase is observed near the fracture surfaces of the alloys, which explores as a potential cause for the creep rupture. However, the formation of TCP phase is effectively suppressed for decreasing segregation of the alloying elements, which results in the improvement of the creep life in the alloy with 430 ppm carbon at 1120 °C and 140 MPa. & 2015 Elsevier B.V. All rights reserved.
Keywords: Carbides Microporosity X-ray computed tomography Creep property Nickel-base superalloys
1. Introduction Single crystal (SX) superalloys have been widely used in advanced turbine engines for hot components. To further enhance their high temperature capability, the level of refractory elements such as tungsten and rhenium in the advanced SX superalloys has been gradually increased. However, there is a great tendency for the formation of grain defects (strays, low angle grain boundaries, freckles, etc.) in these alloys [1–3]. Therefore, trace elements such as carbon and boron are reintroduced into SX superalloys for grain boundary reinforcement [4,5] and reducing casting defects during solidification [6–8]. The formation of grain defects in the SX superalloys with the addition of carbon can be effectively suppressed [6,7]. However, conflicting results involving the role of carbon on the creep property of SX alloy have been reported. It has been found that the creep life in the carbon modified SX alloy at 850 °C/430 MPa is improved, but the creep life at 1050 °C/165 MPa decreases [9]. Contrary results were observed by Liu et.al [10]. Kong and Wasson also reported that carbon has a detrimental effect on creep life under different conditions [11,12]. It is necessary to clarify the n
Corresponding author. Tel.: þ 86 24 23971276; fax: þ 86 24 23971712. E-mail address: [email protected] (L. Wang).
http://dx.doi.org/10.1016/j.msea.2015.05.039 0921-5093/& 2015 Elsevier B.V. All rights reserved.
influence of carbon on the creep life in SX superalloys. Creep rupture is associated with a number of microstructural features, including MC carbides, porosity and TCP formation. It has been confirmed that morphology of MC carbides transforms from blocky to script-like and the volume fraction of γ/γ′ eutectic is reduced by carbon additions [13–15]. However, there are controversial opinions concerning the role of carbon during the formation of porosity. Liu and Chen [8,16] found that the formation of porosity is suppressed by the addition of carbon due to volume expansion of MC carbides. However, an increase volume fraction of porosity is observed by Al-Jarba and Culter [7,17] with carbon additions. Moreover, the H-pore formed during homogenization has been observed, and the volume fraction of H-pore exhibits the same level as that of S-pore formed during solidification [18–20]. The two kinds of pores have not been distinguished. In this study, the three dimensional information of S-pore and H-pore has been investigated by X-ray computed tomography (XCT). The topologically close packed (TCP) phase is prone to form during the rafting creep [21,22]. The brittle and needle-like phase depletes strengthening elements from the matrix and is susceptible to crack during creep deformation. Retardation of TCP formation in carbon modified alloy RR2072 has been observed [23], while Qin and Kong [24,25] reported that TCP phase initiates in the vicinity of MC carbides. Therefore, the aim of the present study is to illustrate the effect
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Table 1 The actual carbon contents of alloys in the present experiment.
C (ppm)
733
Table 2 Chemical composition of MC carbides (wt%).
SX-1
SX-2
SX-3
Phase
C
Ti
Cr
Co
Ni
Mo
Ta
88
220
430
MC
13
2
2
2
7
3
71
of carbon on the microstructure involving the porosity, the TCP phase and the creep properties in the SX alloys.
2. Experimental procedure A 3rd generation nickel-base single crystal (SX) superalloy DD33 with 4% Re (mass fraction, %) was used in the present experiment. The master alloys with three different amounts of carbon were directionally solidified (DS) into ϕ16 200 mm SX bars by the high rate solidification (HRS) process. The actual carbon contents of alloys were given in Table 1. Longitudinal orientation of all SX bars was within 10° deviating from [001] orientation. The samples for comparison of as-cast microstructures (γ/γ′ eutectic, porosity and MC carbides) were sectioned perpendicularly to the DS direction and taken from the same location on each SX bar in order to minimize the effect of solidification conditions. To investigate 3-D features of solidification pores (S-pore) and homogenization pores (H-pore) by XCT, the as-cast and solution heat treated samples were machined into cubes of 1 1 1 mm3. Partitioning coefficient of alloying elements in alloys with different carbon contents was measured by electron probe microanalyzer (EPMA). Point scans in dendrite cores and interdendritic regions were performed. The partitioning coefficient, ki, of each element was identified by dividing the weight percent of each element at the dendrite cores by that in the interdendritic regions. Ten points from each area were averaged to reduce variation. Following the same heat treatment (1335 °C/10 h, AC þ1150 °C/ 4 h, AC þ870 °C/24 h, AC), the first group of samples were exposed at 1100 °C for 1000 h to examine the formation of TCP phase. The second group of heat treated samples were machined into creep specimens with gauge diameter of 5 mm and gauge length of 25 mm along the [001] orientation. Creep tests were carried out at 870 °C/586 MPa and 1120 °C/140 MPa, using FC-20 high temperature creep-testing machines. The microstructure of crept specimens was examined using scanning electron microscope (SEM).
3. Results 3.1. As-cast microstructure Varying the carbon contents from 88 to 430 ppm, the alloys have similar dendrite arm spacing, but the morphology of MC
Table 3 Volume fractions of porosity in as-cast alloys measured by XCT and OM. C (ppm)
88
220
430
XCT OM
0.077 0.002 0.03 70.01
0.08 7 0.004 0.05 7 0.03
0.127 0.006 0.09 70.05
Fig. 2. Distribution of porosity diameter in as-cast alloys with different carbon contents.
carbides changes from blocky (Fig. 1a) to script-like (Fig. 1b). These results agree well with previous studies [13–15,26]. Further SEM observations on the deep-etched sample show that the script-like MC carbides actually adopt a dendritic structure (Fig. 1c). These dendritic carbides are rich in Ta and Ti, and have a small amount of Mo, Co and Cr (Table 2). The volume fraction of porosity measured by XCT and OM is given in Table 3. The volume fraction of S-pore measured by XCT is 0.07%, 0.08% and 0.12% in the alloys of SX-1, SX-2 and SX-3, respectively. The XCT results are higher, compared with the porosity metallographically measured by OM. The porosity measured by the two different methods are in agreement within the accuracy. Increasing volume fractions of porosity is observed in high carbon modified alloys. Distribution of porosity diameter is illustrated in Fig. 2. The maximum diameter of porosity increases from 30 to 46 μm, with the carbon content increasing from 88 to 430 ppm. Fig. 3 shows SEM and XCT images of the porosity in the alloy SX-1. SEM observation indicates that porosities are located in the interdendritic regions (Fig. 3a). XCT observation shows that there
Fig. 1. Morphology of MC carbides in the alloy with different carbon contents: (a) 88 ppm, (b) and (c) 430 ppm.
734
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Fig. 3. Morphology of porosity in the as-cast alloy SX-1: (a) SEM image, and (b) XCT results.
are two kinds of pores: the sphere pores and the irregular pores (Fig. 3b). Partitioning coefficient of alloying elements in as-cast alloys with different carbon contents is shown in Fig. 4. It can be seen that the addition of carbon decreases the segregation of Re and W, while it has a little influence on the segregation of elements such as Al and Ta. The results are identical to the findings of Tin et al. [27]. 3.2. Heat treated microstructure
Fig. 4. Partitioning coefficient of alloying elements in alloys with different carbon contents.
Fig. 5 presents that the volume fraction of porosity increases in the experimental alloys after solution heat treatment. The increment of porosity after solution heat treatment was 0.04%, 0.06% and 0.04%, which are named as volume fractions of H-pores. As shown in Fig. 6, the probability of small sphere-shaped pores increases after solution heat treatment. Morphology of porosity in the alloy SX-1 after solution heat treatment at 1335 °C for 10 h is illustrated in Fig. 7. Small sphere-shaped pores form and S-pores grow round. 3.3. Long-term exposed microstructure
Fig. 5. Volume fractions of porosity in as-cast and heat treated alloys with different carbon contents.
The formation of TCP phase during high temperature exposure at 1100 °C for 1000 h varies significantly in the alloys with different carbon contents. TCP precipitates are occasionally observed in the alloys with 88 and 220 ppm carbon as given in Fig. 8a and b. In the alloy with 430 ppm carbon, no TCP phase is observed (Fig. 8c). These TCP precipitates are rich in W, Re, Cr, and Co as listed in Table 4. Based on the chemical compositions and morphologies, these TCP precipitates are confirmed to be μ phase. Decomposition of MC carbides is not observed in the experimental alloys after long term exposure (Fig. 8c). 3.4. Creep behavior
Fig. 6. Distribution of porosity diameter in as-cast and heat treated alloy SX-1.
Fig. 9 gives the curves of time vs. creep strain for the experimental alloys at 850 °C and 586 MPa. Fig. 9a shows all curves consist of three stages. Significant primary creep strain develops in all alloys and the onset of primary creep occurs in a different manner as indicated in Fig. 9b. The specimens with 430 ppm carbon show the largest primary creep strain with the shortest rupture life of 276 h. The specimens containing 88 and 220 ppm carbon have shorter primary creep strain and longer creep life of 421 and 364 h, respectively. SEM images of the longitudinal section of the ruptured specimen reveal that cracks are normal to the applied stress direction, as shown in Fig. 10. In all the specimens, cracks preferentially initiate at MC carbides in the interdendritic regions. Cracks in the
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
735
Fig. 7. Morphology of porosity in the heat treated alloy SX-1: (a) SEM images and (b) XCT results.
Fig. 8. Microstructure of alloys with different carbon contents after exposure at 1100 °C for 1000 h: (a) 88 ppm, (b) 220 ppm and (c) 430 ppm. Table 4 Chemical composition of TCP phase (wt%). Phase
Re
W
Mo
Cr
Co
Ni
μ
45.6
18.2
5.2
4.8
6.9
18.9
dendritic regions are not observed. Number of microcracks in the alloy with 430 ppm carbon is larger than that in the alloy with 88 ppm carbon. Fig. 11 shows the creep curves of three alloys tested at 1120 °C/ 140 MPa. The curves are very similar (Fig. 11a). Fig. 11b presents the plots of creep time vs. creep strain rate. The creep rates stay relatively constant for the first 60 h and increase more rapidly for alloys with 88 ppm carbon. The high carbon-containing specimens
show the smallest elongation of 14% as well as the longest rupture life of 91 h, while the specimens with 88 and 220 ppm carbon have the similar elongation of around 50% and creep life of 61 and 75 h, respectively. Fig. 12 shows SEM images of longitudinal sections, confirming the microcracks grow perpendicularly to the applied stress direction. Microcracks in both dendritic and interdendritic regions are observed in the experimental alloys with 88 and 220 ppm carbon, as indicated in Fig. 12a and b. Only creep cavities in the dendritic regions in the alloys with 430 ppm carbon are observed in Fig. 12c. Fig. 12d, e and f gives the microcracks or creep cavities originated from TCP phases in the dendritic areas, while Fig. 12h, i and j presents the microcracks in the interdendritic area originated from MC carbides.
Fig. 9. Creep curves of the alloys with different carbon contents at 850 °C/586 MPa.
736
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
Fig. 10. SEM micrographs of longitudinal sections near the fracture surfaces of samples with different carbon contents crept at 850 °C/586 MPa: (a) 88 ppm, (b) 220 ppm and (c) 430 ppm.
Fig. 11. Creep curves of alloys with different carbon contents at 1120 °C/140 MPa.
4. Discussion 4.1. Microstructure According to Niyama criterion [28],
ΔP = − ηβ
V G
∫0
xc
fL K
⋅d x =
ηβΔT G/ V α
(
−2
)
(1)
where ΔP is pressure drop across the mushy zone, η is the viscosity of the liquid, fL is the volume fraction of liquid, K is the permeability of the liquid, G is temperature gradient and V is cooling velocity, ΔT is temperature difference across the mushy zone, α is a constant parameter. The total solidification shrinkage is defined as β ¼ (ρS ρL)/ρL, where ρL and ρS is the density of liquid and solid, respectively. The permeability of the liquid K proposed by Carlson and Madison [29,30] is defined as follows:
K = K0
f L3 (1 − fL )2
(2)
where initial permeability of the liquid K0 is proportional to equivalent diameter of the interdendritic regions. Integrating Eq. (1) substituted by expression (2), pressure drop across the mushy zone is written as follows:
ΔP0
ηβΔT K0
−2
(G/ V ) ηβΔT = (G/ V ) K
ΔP =
f
∫1 L
2
( ) 1 − fL fL
(
d fL = ΔP0 fL −
1 fL
− 2 ln fL
)
−2
0
permeability of the liquid, K0. Considering similar solidification parameters of three SX alloys, with the carbon content increasing from 88 to 430 ppm, the morphology of MC carbides evolves from blocky to dendritic-like (Fig. 1). The equivalent diameter of the interdendritic regions is decreased by these well-developed dendritic carbides. Therefore, the pressure drop increases with the reduction of the initial permeability of the liquid (K0) in the alloy with 430 ppm carbon (Eq. (3)). Although small pores may disappear due to the lattice expansion of MC carbides [8], the larger size of dendritic carbides is obviously an important reason for the high level of porosity in the alloy with 430 ppm carbon. It was found that the imbalance cross-diffusion of alloy elements between Al, Ta and Re, resulted in the Kirkendall pore, which may be the reason for the H-pore formation during solution heat treatment [19]. In the present study, although the carbon content is different, the primary chemical composition of SX alloys is identical. Thus, the level of imbalanced cross-diffusion of alloy elements during the same solution heat treatment is similar. TCP phases are prone to form during long term exposure or creep deformation because of the segregation of alloying elements, which is generally undesirable in SX superalloys [21–23]. The reduction of segregation of alloying elements with the increase of the carbon content gives the reason why the TCP phase formation in alloys with 430 ppm is suppressed (Fig. 4). On the other hand, creep deformation accelerates diffusion and thus promotes the formation of TCP during high temperature creep (Fig. 12f).
(3)
Porosity is prone to form when the pressure drop of molten alloy becomes larger or the remaining liquid is not enough to compensate the volume shrinkage during the last stage of solidification. The pressure drop of liquid is inversely proportional to the
4.2. Creep Most work has revealed that creep performance at low temperature and high stress is sensitive to the primary creep strain, and the primary creep behavior is controlled by shearing of γ′
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
737
Fig. 12. Microstructures of longitudinal section near the fracture surfaces in alloys with different carbon contents: (a), (d) and (h) 88 ppm; (b), (e) and (i) 220 ppm; (c), (f) and (j) 430 ppm; (d), (e) and (f) dendritic region; (h), (i) and (j) interdendritic region.
phase [31,32]. Therefore, the high carbon content alloy crept at 850 °C/586 MPa with the largest primary creep strains exhibits the shortest creep life, as shown in Fig. 9. It was reported that the formation of nano-size M23C6 carbides may prevent the dislocation movement and enhance the creep properties [10,34], but these nano-size M23C6 carbides were not observed in the experimental alloys. On the contrary, MC carbides act as stress concentration sites and induce premature microcracks during the low temperature creep deformation (Fig. 10c). Hence, the larger size of dendritic carbides in the interdendritic regions in the alloys with 430 ppm carbon may accelerate the local creep damage and have a detrimental effect on the creep behavior at 850 °C and 586 MPa [9]. Although microcracks induced by the dendritic carbides and porosity in the interdendritic regions were observed (Fig. 12), TCP phases near the fracture surfaces were observed in all the crept specimens at 1120 °C/140 MPa (Fig. 12), which would be expected to have the greatest impact on the high temperature creep behavior [33,34]. Usually, the precipitation of μ particles will lead to the depletion of strengthening elements from the matrix. The softened matrix was prone to deform, which in turn accelerates the formation of μ particles [35]. In addition, needle shaped μ phase was
reported to induce creep cavities and microcracks in the dendritic regions, which accelerates creep rupture [22,36]. Therefore, suppression of TCP phase during 1120 °C/140 MPa creep is a main reason for the longer creep life in the alloy with 430 ppm carbon. On the other hand, less microcracks in the interdendritic regions in the alloy with 220 and 430 ppm carbon are observed (Fig. 12a–c), which indicates less impact of dendritic carbides in the interdendritic regions on the creep life at 1120 °C/140 MPa in the alloy.
5. Conclusion Effect of carbon content on the microstructure and creep properties of a 3rd generation single crystal nickel-base superalloy has been investigated. The following conclusions can be given: 1. With the increase of carbon content from 88 to 430 ppm, MC carbides evolve from blocky to script-like. The dendritic carbides promote the formation of S-pore due to the increase of the pressure drop. The volume fraction of porosity increases in the experimental alloys after solution heat treatment. The level of H-pores in SX alloys with various carbon contents is
738
X.W. Li et al. / Materials Science & Engineering A 639 (2015) 732–738
identical. 2. The high carbon content alloy crept at 850 °C/586 MPa exhibits the shortest creep life with the largest primary creep strain. The dendritic carbides and pores are detrimental to the creep rupture behavior. 3. The TCP phase is expected to have significant impact on the creep behavior at 1120 °C/140 MPa. Formation of the TCP phase during high temperature exposure is suppressed by the addition of carbon, which results in the longer high temperature creep life in the alloy with 430 ppm carbon.
Acknowledgment This work was supported by the National Natural Science Foundation of China (51201164) and the National High Technology Research and Development Program of China (2012AA03A511).
References [1] T.M. Pollock, W.H. Murphy, E.H. Goldman, D.L. Uram, J.S. Tu, Superalloys 1992 (1992) 125–134. [2] P. Auburtin, T. Wang, S.L. Cockcroft, A. Mitchell, Metall. Mater. Trans. B 31 (2000) 801–811. [3] J.W. Aveson, P.A. Tennant, et al., Acta Mater. 61 (2013) 5162–5171. [4] Q.Z. Chen, C.N. Jones, D.M. Knowles, Mater. Sci. Eng. A 385 (2004) 402–418. [5] D.M. Shah, A. Cetel, Superalloys 2000 (2000) 295–304. [6] S. Tin, T.M. Pollock, W.T. King, Superalloys 2000 (2000) 201–210.
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
K.A. Al-Jarba, G.E. Fuchs, Mater. Sci. Eng. A 373 (2004) 255–267. Q.Z. Chen, Y.H. Kong, C.N. Jones, D.M. Knowles, Scr. Mater. 51 (2004) 155–160. Q.Z. Chen, N. Jones, D.M. Knowles, Acta Mater. 50 (2002) 1095–1112. L.R. Liu, T. Jin, et al., Mater. Sci. Eng. A 385 (2004) 105–112. Y.H. Kong, Q.Z. Chen, D.M. Knowles, J. Mater. Sci. 39 (2004) 6993–7001. A.J. Wasson, Ph.D. thesis, University of Florida, Gainesville, 2010. A.J. Wasson, G.E. Fuchs, Mater. Charact. 74 (2012) 11–16. K.A. Al-Jarba, G.E. Fuchs, JOM 56 (2004) 50–55. E.R. Cutler, A.J. Wasson, G.E. Fuchs, Scr. Mater. 58 (2008) 146–149. L.R. Liu, T. Jin, et al., Mater. Lett. 58 (2004) 2290–2294. E.R. Cutler, A.J. Wasson, G.E. Fuchs, J. Cryst. Growth 311 (2009) 3753–3760. D.L. Anton, A.F. Giamei, Mater. Sci. Eng. 76 (1985) 173–180. B.S. Bokstein, A.I. Epishin, et al., Scr. Mater. 57 (2007) 801–804. T.L.A.I. Epishina, Phys. Met. Metallogr. 115 (2014) 21–29. M.S.A. Karunaratne, C.M.F. Rae, R.C. Reed, Metall. Mater. Trans. A 32 (2001) 2409–2421. C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. Q.Z. Chen, D.M. Knowles, Metall. Mater. Trans. A 33 (2002) 1319–1330. Y.H. Kong, Q.Z. Chen, Mater. Sci. Eng. A 366 (2004) 135–143. X.Z. Qin, J.T. Guo, Mater. Sci. Eng. A 543 (2012) 121–128. X.W. Li, L. Wang, X.G. Liu, J.S. Dong, L.H. Lou, J. Mater. Sci. Technol. 30 (2014) 1296–1300. S. Tin, T.M. Pollock, Mater. Sci. Eng. A 348 (2003) 111–121. E. Niyama, T. Uchida, M. Morikawa, S. Saito, Cast Met. J. 6 (1981) 16–27. K.D. Carlson, C. Beckermann, Metall. Mater. Trans. A 40 (2009) 163–175. J. Madison, J.E. Spowart, et al., Metall. Mater. Trans. A 43 (2012) 369–380. V.A. Vorontsov, C. Shen, Y. Wang, D. Dye, C.M.F. Rae, Acta Mater. 58 (2010) 4110–4119. F. Diologent, P. Carbon, Mater. Sci. Eng. A 385 (2004) 245–257. R.C. Reed, D.C. Cox, C.M.F. Rae, Mater. Sci. Eng. A 448 (2007) 88–96. J.B. Le Graverend, J. Cormier, S. Kruch, F. Gallerneau, J. Mendez, Metall. Mater. Trans. A 43 (2012) 3988–3997. L. Wang, D. Wang., et al., Mater. Charact. 104 (2015) 81–85. S.G. Tian, M.G. Wang, T. Li, B.J. Qian, J. Xie, Mater. Sci. Eng. A 527 (2010) 5444–5451.