Abstract
Phase stability of Ni-base single crystal superalloys was compared between Ir addition and Ru addition. Investigated alloys were TMS-238 and its derivative alloys. The thermodynamic equilibrium microstructures obtained by a strain aging showed that the volume fractions of γ, γ' and topologically close packed (TCP) phases in the alloys were virtually the same to each other without depending on Ir substitution for Ru. On the other hand, the time temperature transformation diagrams obtained by an ordinary aging showed a significant delay in TCP precipitation in the alloys with Ir substitution. The delay in the TCP precipitation might be attributed to the smaller interdiffusion coefficient of Ir-Ni, compared with one of Ru-Ni. The smaller interdiffusion coefficient may affect the kinetics of the TCP precipitation.
1. Introduction
Higher temperature capability is necessary for turbine blades and vanes of jet engines and land-base gas turbines in order to improve their thermal efficiencies. The turbine blades and the vanes are generally made of Ni-base superalloys. The Ni-base superalloys have the structure composed of γ phase (Ni solid solution) and γ' phase (L12 ordered phase based on Ni3Al). At the present time, the Ni-base single crystal (SC) superalloy having the highest creep strength in the world is TMS-2381) developed at National Institute for Materials Science (NIMS) in Japan. TMS-238 contains Ru so as to suppress precipitation of topologically close packed (TCP) phases2). The precipitation of TCP phases causes decline in creep strength of Ni-base SC superalloys. Because TCP phases are generally precipitated on {111} known as slip planes of an fcc structure, and allow dislocations to move more easily3).
Some of the authors found that not only Ru but also other platinum group metals (e.g. Ir) have chemical thermodynamic stabilization effect on γ phase and γ' phase, and suppress the TCP precipitation4). Based on this knowledge, Ni-base SC superalloys containing Ir substituting for equal at% Ru were designed and their creep strengths were investigated. The results show that the alloys containing Ir have much longer creep rupture life than those containing Ru under 900℃-392 MPa5,6). One of the reasons for it is that the Ni-base SC superalloys containing Ir have lower content of the TCP precipitates than that containing Ru6). However, it was not clear why Ir substitution for Ru suppresses the TCP precipitation.
Therefore, the purpose of this work is to clarify the reason for the suppression of TCP precipitation by substituting Ir for Ru in Ni-base SC superalloys.
2. Experimental Procedure
2.1 Sample preparation
Table 1 shows the nominal compositions of the investigated Ni-base SC superalloys. The alloys were TMS-2381), TMS-238Ir (designed by substituting Ir for equal at% Ru in TMS-238), TMS-238+ReRu6) and TMS-238+ReIr6) (designed by substituting Ir for equal at% Ru in TMS-238+ReRu). The four alloys were melted in a vacuum induction furnace for unidirectional solidification. The melt, held at about 1600℃, was poured into a lost wax mold kept at about 1500℃. The mold had a starter and a selector and was put on a copper chill plate cooled by water. After pouring, the mold was pulled down from a heating chamber to a cooling chamber at 200 mm/h for unidirectional solidification and cylindrical single crystal samples (length 130 mm, diameter 10 mm) were obtained. These samples were heat-treated under the following conditions.
Table 1
Nominal compositions of the investigated Ni-base single crystal superalloys (Ni bal., at%).
| |
Co |
Cr |
Mo |
W |
Al |
Ta |
Hf |
Re |
Ru |
Ir |
| TMS-238 |
6.99 |
5.60 |
0.73 |
1.38 |
13.85 |
2.66 |
0.04 |
2.18 |
3.13 |
- |
| TMS-238Ir |
6.99 |
5.60 |
0.73 |
1.38 |
13.85 |
2.66 |
0.04 |
2.18 |
- |
3.13 |
| TMS-238+ReRu |
7.10 |
5.70 |
0.74 |
1.40 |
14.08 |
2.70 |
0.04 |
2.56 |
4.46 |
- |
| TMS-238+ReIr |
7.10 |
5.70 |
0.74 |
1.40 |
14.08 |
2.70 |
0.04 |
2.56 |
- |
4.46 |
Solution treatment: 1345℃/20 h + Cooling
→First step aging: 1150℃/2 h + Cooling
→Second step aging: 870℃/20 h + Cooling
After the second step aging, the cylindrical samples of all the alloys were cut into disks (thickness 5 mm).
2.2 Strain aging
In order to obtain thermodynamic equilibrium microstructures by recrystallization and coarsening, the disks of TMS-238 and TMS-238Ir were strain-aged by the following way. First, some punches were driven on the surfaces of the disks with a hummer and stamping punches so as to add strain. Then, the disks were heat-treated at 1100℃ for about 600 h and 1000 h, and water-quenched. After the water quenching, cross sections of the disks were observed using a scanning electron microscope (SEM, JEOL JSM-6060) and the volume fractions of γ, γ' and TCP were measured from 3 different back-scattered electron (BSE) images for each sample. The compositions of TCP phases were analyzed using an electron probe X-ray microanalyzer (EPMA, Shimadzu EPMA-1610). The number of the analysis points was 5 for each sample.
2.3 Ordinary aging
In order to obtain time temperature transformation (TTT) diagrams on TCP precipitates, an ordinary aging was performed. The disks of all the investigated alloys were heat-treated in an electric furnace at 900℃, 1000℃, 1100℃ and 1200℃ for 1 h, 4.5 h, 25 h, 72.5 h, 160–167.5 h, 302.5–310.5 h, 573–598 h, 952–1050 h, 2000 h and 2499–2541 h. After the ordinary aging, water quenching was performed. Then, the time temperature transformation (TTT) diagrams on TCP precipitates were obtained by checking the TCP precipitation in their dendrite core area using SEM.
3. Results and Discussion
Figure 1 shows BSE images of the investigated alloys after the second step aging without the strain and ordinary aging processes. All the images in Fig. 1 show cuboidal γ' precipitates (gray) and γ matrix (white). TCP precipitates were not observed in all of the samples.
Figure 2 shows BSE images of TMS-238 and TMS-238Ir after the strain aging at 1100℃. Comparing Fig. 1 and 2, it is clear that γ phase, γ' phase and TCP phase in Fig. 2 were recrystallized and coarsened. The size of the TCP precipitates is about 2～3 μm. The minimum beam size of EPMA in this work was 1 μm. The TCP precipitates size is thus larger than the minimum beam size. The volume fractions of γ, γ' and TCP in TMS-238 and TMS-238Ir after the strain aging at 1100℃ are shown in Fig. 3. The volume fractions of γ, γ' and TCP after the strain aging for 600 h were virtually equal to those after the strain aging for 1000 h. Therefore, this result suggests that the γ, γ' and TCP coexisted in the thermodynamic equilibrium at 1100℃ in the strain-aged samples. Meanwhile, the strain-aged samples of TMS-238Ir had virtually equal volume fractions of γ, γ' and TCP to ones of TMS-238 as shown in Fig. 3. Therefore, it can be explained by the equilibrium theory that Ir addition has virtually the same phase stabilization effect on γ phase and γ' phase as Ru addition.
Figure 4 shows the compositions of TCP precipitates in TMS-238 and TMS-238Ir after the strain aging at 1100℃. The TCP phase in TMS-238 was mainly composed of Re, Ni, Cr, Co and W. On the other hand, The TCP phase in TMS-238Ir contained Ir in addition to the elements composing the TCP phase in TMS-238. Also, Re content in TCP phase of TMS-238Ir was higher than that of TMS-238.
Figure 5 shows the TTT diagrams on TCP precipitates in dendrite core area of TMS-238 and TMS-238Ir (a), and TMS-238+ReRu and TMS-238+ReIr (b). White plots represent that TCP precipitates were observed in neither Ru addition alloy nor Ir addition one. Gray plots represent that TCP precipitates were observed only in Ru addition alloy. Black plots represent that TCP precipitates were observed in both alloys. The TTT diagrams indicate that precipitation of TCP phase in TMS-238Ir and TMS-238+ReIr was significantly delayed as compared to that in TMS-238 and TMS-238+ReRu in the range from 900℃ to 1100℃, respectively. It is thus clarified that Ir substitution for Ru in Ni-base SC superalloys causes delay in start of TCP precipitation. Taking it into account that the equilibrium volume fractions of γ, γ' and TCP in the Ir addition alloy are the same as the Ru addition alloy, the delay in the TCP precipitation should be explained by decline in mobility of alloying elements due to Ir substitution for Ru. The interdiffusion coefficient of Ir-Ni is about one fourth and one sixth as large as that of Ru-Ni at 1100℃ and at 900℃, respectively7). The smaller interdiffusion coefficient may affect mobility of other alloying elements in the Ir addition alloys, and the kinetic delay of the TCP precipitation. However, in order to clarify this hypothesis, further studies are necessary.
Figure 6 shows BSE images showing TCP precipitates in dendrite core area of TMS-238+ReRu and TMS-238+ReIr after the ordinary aging at 900℃ (a) and 1100℃ (b). TCP precipitates of which some part indicated by arrows in Fig. 6 were observed in TMS-238+ReRu but not in TMS-238+ReIr under 166 h-900℃ and under 4.5 h-1100℃. TMS-238+ReRu had many fine TCP precipitates at 900℃ and 1100℃. On the other hand, TMS-238+ReIr had a fewer coarser ones. The reason for the difference in the size and the distribution of TCP precipitates between Ru addition and Ir addition is also an interesting subject for future investigation.
4. Conclusions
The phase stability of Ni-base single crystal superalloys was compared between Ir addition and Ru addition. The following conclusions were obtained.
(1) The thermodynamic equilibrium microstructures obtained by the strain aging showed that the volume fractions of γ, γ' and TCP in the Ni-base SC superalloys containing Ir were virtually the same as those in the alloy containing Ru.
(2) The TTT diagrams obtained by the ordinary aging shows that Ir substitution for Ru causes the significant delay in the TCP precipitation in the alloys with Ir substitution.
(3) The delay in the TCP precipitation might be attributed to the smaller interdiffusion coefficient of Ir-Ni than that of Ru-Ni. The smaller coefficient may affect the kinetics of the TCP precipitation.
Acknowledgments
The first author: Yuhi Mori is a Research Fellow of Japan Society for the Promotion of Science (JSPS). This work was supported by JSPS Research Fellowship for Young Scientists under grant number 15J01523. All the experiments were carried out at NIMS in Japan. We acknowledge Dr. M. Osawa of NIMS for his advices on this work.
REFERENCES
- 1) K. Kawagishi, A.C. Yeh, T. Yokokawa, T. Kobayashi, Y. Koizumi and H. Harada: Superalloys 2012 (2012) 189–195. 10.7449/2012/Superalloys_2012_189_195
- 2) K.S. O'hara, W.S. Walston, E.W. Ross and R. Darolia: Nickel Base Superalloy and Article, U.S. Patent 5482789A (1996).
- 3) C.M.F. Rae, N. Matan and R.C. Reed: Mater. Sci. Eng. A 300 (2001) 125–134. 10.1016/S0921-5093(00)01788-3
- 4) T. Yokokawa, M. Osawa, K. Nishida, T. Kobayashi, Y. Koizumi and H. Harada: Scr. Mater. 49 (2003) 1041–1046. 10.1016/S1359-6462(03)00437-8
- 5) T. Kobayashi, H. Harada, M. Osawa and A. Sato: J. Jpn. Inst. Metals 69 (2005) 1099–1103. 10.2320/jinstmet.69.1099
- 6) Y. Takebe, T. Yokokawa, T. Kobayashi, K. Kawagishi, H. Harada and C. Masuda: J. Jpn. Inst. Met. Mater. 79 (2015) 227–231. 10.2320/jinstmet.JBW201402
- 7) M.S.A. Karunaratne and R.C. Reed: Acta Mater. 51 (2003) 2905–2919. 10.1016/S1359-6454(03)00105-8
