- Email: [email protected]
Influences of ruthenium and crystallographic orientation on creep behavior of aluminized nickel-base single crystal superalloys
Materials Science & Engineering A 592 (2014) 143–152
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Influences of ruthenium and crystallographic orientation on creep behavior of aluminized nickel-base single crystal superalloys F.H. Latief a,n, K. Kakehi a, H. An-Chou Yeh b, H. Murakami c a
Department of Mechanical Engineering, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji-shi, Tokyo 192-0397, Japan Department of Materials Science and Engineering, National TsingHua University, 101, Section 2, Kuang-Fu Road, Hsinchu 30013, Taiwan c Hybrid Materials Center, National Institute for Materials Science (NIMS), 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 9 July 2013 Received in revised form 28 October 2013 Accepted 30 October 2013 Available online 13 November 2013
The influences of ruthenium and surface orientation on creep behavior of aluminized Ni-base single crystal superalloys were investigated by comparing two different types of NKH superalloys. The aluminized coated specimens were then subjected to creep rupture tests at a temperature of 900 1C and a stress of 392 MPa. The coating treatment resulted in a significant decrease in creep rupture lives for both superalloys. The diffusion zones between the coating and substrate led to changes in microstructure, which diminished the creep behavior of the aluminized superalloys. Because of the interdiffusion of Ru, Al and Ni, the solubility of some of the refractory elements, such as W, Re. Mo, Co and Cr decreased in the diffusion zone; the precipitation of topologically close-packed (TCP) phases was thus inevitable. In the present study, the addition of Ru increased the degree of Re and Cr supersaturation in the γ matrix. Consequently, the addition of Ru indirectly promoted the precipitation of TCP phases in aluminized Ni-base single crystal superalloys. Furthermore, the growth of TCP precipitates was greatly influenced by the specific surface orientations of the Ni-base single crystal superalloys. In conclusion, the {110} specimens showed shorter creep rupture life than the {100} specimens, this was due to the difference in the crystallographic geometry of {111}〈101〉 slip system and TCP precipitates between the two side-surface orientations of the specimens. & 2013 Elsevier B.V. All rights reserved.
Keywords: Ruthenium Ni-base superalloy Creep Crystallographic orientation Aluminide coating
1. Introduction Aluminide coatings have been applied to Ni-base single crystal superalloys used for turbine components in order to protect them from oxidation and corrosion during their service lives. Details of these coatings have been reported [1–5]. Problems have arisen with these coated superalloys, leading to the loss of the coating. These problems include the precipitation of topologically closepacked (TCP) phases, the formation of Secondary Reaction Zones (SRZs) and the ‘rumpling’ of the coating caused by thermal cycling. Such problems degrade the properties of coated Ni-based superalloys, and are potentially life-limiting to turbine blades. In particular, TCP phase precipitation and SRZ formation are regarded as serious problems in both coated and uncoated Ni-base superalloys. As is well known, topologically close-packed (TCP) phase is a collective designation for several intermetallic compounds rich in the elements W, Mo, Re and Cr precipitated in Ni-base single crystal superalloys [6]. TCP phases form during service at elevated temperatures in such superalloys with high concentrations of
n
Corresponding author. Tel./fax: þ 81 42 6772709. E-mail address: [email protected] (F.H. Latief).
0921-5093/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2013.10.092
these elements added to promote strength the loss of these elements from the matrix impairs the mechanical properties of the blade. Precipitation of the TCP phases [7–9] deteriorates the ductility and creep resistance of Ni-base superalloys. However, the damage mechanisms by aluminide coatings are not yet well understood. In the present study, we investigated the influence of crystallographic orientation on creep behavior of aluminized Ni-base single crystal superalloys, one Ru-free and the other Ru-containing. 2. Experimental materials and procedures Fully heat-treated (solution and aging) Ni-base single crystal superalloys NKH-304 and NKH-510 were used as substrate materials in this study. The details of chemical composition and heat treatment route for each superalloy are given in Table 1. NKH-304 is the base alloy, and NKH-510 is a 3 mass% Ru-addition alloy, where 3% Ru is substituted for Ni. The creep specimens with {100} and {110} side-surfaces (Fig. 1) were prepared by electric discharge machining (EDM), and had cross-section areas of 2.8 mm 2.8 mm and gauge length of 19.6 mm. The specimens were mechanically polished with emery paper prior to the aluminizing process.
144
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Table 1 Chemical composition and heat treatment of alloys. Alloy
Chemical composition (mass%)
Heat treatment
NKH-304
11Co, 6Cr, 6W, 5.4Al, 1.4Ti, 6.8Ta, 4.8Re, 0.12Hf and bal. Ni
Solution heat treatment: 1310 1C/10 h þ1320 1C/12 h þ 1325 1C/12 h þgas furnace cooling and followed by aging treatment: 1180 1C/4 hþ 870 1C/20 h þ air cooling
NKH-510
11Co, 6Cr, 6W, 5.4Al, 1.4Ti, 6.8Ta, 4.8Re, 0.12Hf, 3Ru and bal. Ni
Solution heat treatment: 1310 1C/10 hþ 1320 1C/12 h þ gas furnace cooling and followed by aging treatment: 1160 1C/4 h þ870 1C/20 h þ air cooling
the coating powders mixture and the outward diffusion of Ni from the substrate at the interface area, which was composed of β-NiAl due to the heat treatment. The thickness of the IDZ seemed to vary in both side-surface orientations as shown in Fig. 3. The IDZ thickness of the aluminized NKH-304 was about 3.0 μm on the {100} side-surface and 4.7 μm on the {110} side-surface; meanwhile, the IDZ thickness of aluminized NKH-510 was about 2.3 μm on the {100} side-surface and 3.9 μm on the {110} side-surface. In other words, differences in IDZ thickness ensued in all specimens. This means that the presence of the coating rendered the changes in the effective cross-section area that affects the creep behavior of Ni-base superalloys. 3.2. Creep behavior
Fig. 1. Arrangement of {111}〈101〉 slip systems for two kinds of creep specimens with (a) {100} and (b) {110} side-surface orientations.
The specimens were then embedded in an Al2O3 retort containing a mixture of 24.5 mass% Al, 24.5 mass% Cr, 49 mass% Al2O3 and 2 mass% NH4Cl powders, and heated at 1000 1C for 5 h in flowing argon for the aluminizing treatment called the pack cementation process. The coated superalloys were then subjected to creep test with a temperature of 900 1C and a stress of 392 MPa with a parallel load 〈001〉 direction. For comparison, the creep test was also conducted on uncoated specimen under same conditions in order to determine the effect of the aluminide coating on the creep behavior of Ni-base single crystal superalloys. The microstructural changes before and after the creep rupture tests were observed by scanning electron microscopy (SEM). The compositional profiles through the cross section of samples were analyzed by electron probe microanalysis (EPMA). The orientation of crystal grain was identified by a TSL-OIM electron back scattered diffraction (EBSD) analyzer.
3. Results 3.1. Microstructures after heat treatment and coating The microstructures of the γ' precipitates after aging heat treatment are shown in Fig. 2. No obvious differences in the morphology of γ′ precipitates were observed between the two alloys. The mean γ′ precipitate lengths were 345 nm and 310 nm in the alloys NKH-304 and NKH-510 respectively. The γ′ precipitate size of the Ru-containing alloy NKH-510 was a little smaller than that of the base alloy. SEM micrographs of as-aluminized Ni-base single crystal superalloys NKH are shown in Fig. 3. Three distinctive layers can be distinguished from the micrographs: (i) the coating as an outer protective layer, (ii) an interdiffusion zone (abbreviated as IDZ), and (iii) a substrate. IDZ was derived from the reaction between the inward diffusion of Al, produced by
Fig. 4 shows the variation in creep rupture lives performed at a temperature of 900 1C and a stress of 392 MPa. The two superalloys exhibited similar trends, while the bare specimens with the {100} side-surface orientation demonstrated higher creep rupture life than the {110} side-surface orientation. It was clear that the coating treatment resulted in a significant decrease in creep rupture life in both superalloys. Furthermore, the difference in creep rupture life between the two orientations indicated that anisotropy in creep had occurred. In general, the aluminized NKH510 had a higher creep rupture life than the aluminized NKH-304, in both the {100} and {110} side-surface specimens. The creep rupture lives of aluminized NKH-304 were 311 h for the {100} side-surface specimen and 253 h for the {110} side-surface specimen, while the measures for aluminized NKH-510 were 641 h for the {100} side-surface specimen and 539 h for the {110} sidesurface specimen. The differences in creep rupture lives for both side-surface orientations were 58 h (19%) for the aluminized NKH304 and 102 h (16%) for the aluminized NKH-510. 3.3. Microstructures after creep rupture test The SEM micrographs of aluminized specimens after creep rupture test are presented in Fig. 5. An additional zone formed beneath the IDZ in all specimens; this additional zone is called the substrate diffusion zone (SDZ). The SDZ must be differentiated from the IDZ in that it remains particularly γ′ and retains the orientation of the single crystal substrate. However, the secondary reaction zone (SRZ) could not be observed in the two superalloys, as the localized recrystallization had not yet occurred at this stage. The diffusion layer thickness was approximately 23.3 μm on the {100} side-surface specimen and 32.9 μm on the {110} side-surface specimen in the aluminized NKH-304 (Fig. 5a and b). Meanwhile the diffusion layer thickness was approximately 21.6 μm on the {100} side-surface specimen and 33.1 μm on the {110} side-surface specimen in the aluminized NKH-510 (Fig. 5c and d). In addition, the TCP phase was precipitated in the {100} and {110} side-surface specimens for both superalloys, but in differing amounts. The aluminized NKH-510 contained much more TCP phase than the aluminized NKH-304. As crystallographic orientation affected the growth of TCP phases, the formation of these phases was different in the two orientations. The TCP phase showed a
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
145
2 µm Fig. 2. SEM bright field images of (a) NKH-304 and (b) NKH-510.
Fig. 3. SEM micrographs of aluminized specimens: (a,b) NKH-304, and (c,d) NKH-510 with the side-surface orientations of (a,c) {100} and (b,d) {110}.
Fig. 4. Creep rupture curves of various superalloys: (a) NKH-304 and (b) NKH-510.
particular formation with a 451 angle to the substrate interface on the {100} side-surface specimen, but were vertically aligned on {110} the side-surface interface. EBSD crystallographic phase mapping of the aluminized NKH-304 and NKH-501 superalloys after creep rupture is shown in Figs. 6 and 7, respectively. The single-crystal substrates and orientation maps are shown from the reference direction (RD), which is normal to the substrate/coating
interface. EBSD mapping images are the corresponding orientation maps assigned as face-centered cubic (fcc) and body-centered cubic (bcc) structures. It was obvious from the figures that the surface aluminized layer consisted of bcc polycrystalline grains, while the diffusion layer consisted of bcc and fcc phases. EBSD can be utilized to elucidate the relationship between crystal orientation and TCP formation. The EBSD mapping images confirmed the
146
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 5. SEM micrographs of aluminized specimens after creep rupture test: (a,b) NKH-304, and (c,d) NKH-510 with the surface orientations of (a,c) {100} and (b,d) {110}.
Fig. 6. (a) Cross-sectional SEM image of aluminized NKH-304 and corresponding orientation maps assigned as (b) fcc and (c) bcc structures.
formation of TCP phases on both orientations but the SRZ was not observed in all specimens. The voids were apparent in the aluminized NKH-510; furthermore the specimen with the {110} side-surface (Fig. 5d) had many more voids rather than the specimen with the {100} side-surface (Fig. 5c). The schematic of TCP phase growth and voids in both orientations are illustrated in Fig. 8.
3.4. Fracture surfaces Fracture surfaces of aluminized specimens subjected to creep at 900 1C/392 MPa were observed on both the {100} and {110} sidesurface orientations. The aluminized NKH-510 specimens after creep rupture test were observed by SEM, as shown in Fig. 9. The fracture surface in both side-surface specimens showed a
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
147
Fig. 7. (a) Cross-sectional SEM image of aluminized NKH-510 and corresponding orientation maps assigned as (b) fcc and (c) bcc structures.
dimple pattern, and the dimple fracture dominated on the fracture surfaces. Furthermore, cleavage facets (slip planes) were found in both the {100} and {110} side-surface specimens. The difference in creep rupture life revealed that a structural instability occurred during the creep test. This structural instability was considered to be crack initiation at the surfaces. In the case of the coated specimens, creep fracture initiation would be virtually controlled by the crack nuclei in the coated layer.
4. Discussion 4.1. Effect of coating treatment In this study, we investigated the effects of ruthenium and crystallographic orientation on the creep behavior of aluminized Ni-base single crystal. Coating treatments are intended to protect the substrate surface from environmental attack when superalloys are exposed to high operating temperatures with aggressive oxidation and corrosive environment. However, the coating had been shown to degrade the mechanical properties of the coated superalloy. While the coating is not primarily formulated to improve the mechanical properties of the superalloy, the treatment (aluminizing) has been found to significantly impact the superalloy's mechanical properties due to the changes in microstructure. In addition, the coating treatment resulted in a decrease of the effective cross-section area by forming coating layers. This reduction in effective cross-section area due to the formation of coating diffusion layers might be a factor in the reduction of creep strength. However, this effect was tiny as can be seen in Table 2. It is therefore insufficient to explain the reduction of creep life. The change in microstructure in the coated specimens was due to the formation of the IDZ. The increased concentration of Al in the IDZ led to lower strength and eventually to premature failure of the specimen. For details, consider again the creep rupture curves in Fig. 4 where the coated Ni-base single crystal superalloy specimens have a lower creep rupture than the uncoated ones. In
Fig. 8. Schematic illustrations of TCP preferred precipitation orientation on {111} planes for: (a) {100} and (b) {110} side-surface orientations.
the coated specimen, the crack nucleation occurs in the coating layer and propagates according to available regions until final rupture. It is a certainty that this condition motivates the formation of microcracks in the coated layer. Hence, the main reason for the preferential formation of microcracks in the coated layer can be attributed to the significantly higher hardness of the diffusion zone in the coated layer relative to the substrate [10], as shown in
148
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 9. Fracture surfaces of NKH 510 aluminized specimens with: (a) {100} side-surface, and (b) {110} side-surface specimens subjected to creep rupture test at 900 1C and 392 MPa.
4.2. Effect of Ru content
Table 2 Decrease ratio in effective cross-section area (%). Alloy
Total diffusion layer thickness (IDZ þ SDZ) (μm)
Ratio of reduction in effective cross section area (%)
{100}
{110}
{100}
{110}
32.9 33.1
3.3 3.1
4.8 4.7
NKH-304 23.3 NKH-510 21.6
Fig. 10. Microhardness along the coating-substrate regions in aluminized superalloys: (a) as-aluminized and (b) after creep rupture test.
Fig. 10. It is also possible that the decrease of creep rupture might be due to a weakening of atomic bonds, because aluminum is a low melting metal, its effect could be considered similar to that of other low melting metals [11].
In this study, we were interested in investigating the effect of the addition of Ru on the creep behavior of aluminized Ni-based single crystal superalloys. The addition of 3 mass% Ru into superalloy NKH-510 improved the creep behavior of this superalloy when compared to the superalloys without Ru NKH-304. Ru is a solid solution strengthening element that has been added to the next generation of superalloys. The addition of Ru at a higher level than the solubility limit imposed the solvus of the γ/γ′ two-phase region. Early studies on Ru-containing nickel-base single crystal superalloys suggested that Ru additions alter the partitioning of elements between the γ and the γ′ phases [12].The substrate hardness values of Ru-containing alloy (NKH-510) were higher than those of the base alloy (NKH-304) as shown in Fig. 10. Ru addition improved strength in both the γ and γ′ phases [13]. There was also a significant increase in flow stresses of the single-phase γ alloy with Ru, and a modest increase in the flow stress of the single-phase γ′ alloy was also measured [13]. These results provide supportive evidence for the solid solution strengthening effect of Ru. Since the element Ru also has an hcp structure, its addition would be expected to decrease the stacking fault energy of the matrix [14]. The a/2〈110〉 dislocation motion in the matrix channel leaves interfacial dislocations around cuboidal precipitates and forms homogeneous γ–γ' interfacial dislocation networks [15,16]; the addition of Ru extends the width between the partial dislocations in the matrix [17], and would inhibit the cross slip of dislocation at this temperature. Moreover, the addition of Ru into Ni-base single crystal superalloys could enhance microstructural stability, augment resistance to the formation of TCP phases [18], improve the high temperature mechanical properties or environmental resistance [19], and increase the liquidus temperature [20]. During exposure to elevated temperatures, the formation of TCP phases provokes serious deterioration in creep strength. The TCP phases were slightly precipitated in aluminized NKH-304, but considerably precipitated in aluminized NKH-510, as shown in Fig. 5c and d. TCP phases are composed mainly of the elements Ni, Cr, Mo, Co, W, and Re, and unfortunately this list contains the elements that are most effective at conferring resistance to creep. Therefore, the presence of TCP during service must deplete the matrix of these elements and reduce their solid strengthening effect. We also consider the Ru addition to have contributed to the formation of the SDZ in the aluminized NKH-510. For details, see the schematic illustration of the SDZ formation and the interaction with TCP growth in aluminized NKH-510 presented in Fig. 8. The reaction between outward diffusion of Ru and inward diffusion of Al forms β-RuAl nearby IDZ, with respect to SDZ formation. Furthermore, voids were observed in aluminized NKH-510 (Fig. 5c and d). The formation of such voids is generally referred to as the Kirkendall
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
149
Fig. 11. EPMA profiles of ruptured aluminized NKH-510 superalloys: (a) {100} side-surface and (b) {110} side-surface specimen.
effect, in which voids are formed due to the different diffusion rates of two or more metal atoms [21]. When coated specimens are exposed to high temperatures, the volume changes at the interface of the coating and substrate by the transformation of β-NiAl into γ ′-Ni3Al lead Al to diffuse outward from the substrate diffusion zone (SDZ) as it oxidizes, forming the voids in the SDZ. The results of the EPMA analysis (Figs. 11 and 12) showed that Re and Cr contents were enriched in the two regions where the TCP phases were present; the first region was in IDZ, where Ru and Al were enriched, and the other was the Ru depletion region within the SDZ. When NKH-510 (Fig. 11) is compared to NKH-304 (Fig. 12), it is clear that
the underlying mechanism that explains why for the Ru-bearing NKH-510 has more TCP precipitates beneath the coating has to do with the increase in supersaturation of Re and Cr in the IDZ due to the presence of RuAl. The absence of Ru in the SDZ would allow the TCP precipitates to propagate more easily from the IDZ into the SDZ, resulting in a greater depth of TCP-phase region beneath the coating of NKH-510 (Fig. 13). As a consequence, the addition of Ru indirectly promoted the precipitation of TCP phases in the aluminized NKH-510 specimens. It has previously been reported that the addition of Ru suppressed the TCP phase precipitation [18]. In contrast, in the
150
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 12. EPMA profiles of ruptured aluminized NKH-304 superalloy with {100} side-surface specimen.
Fig. 13. Schematic illustrations of the SDZ formation process: (a) outward diffusion of Ru from the substrate to the IDZ, and (b) SDZ formation with the precipitation of TCP phase in the γ/γ′ substrate.
present study, the addition of Ru seemed to promote the formation of TCP phases beneath the coating. For the bulk microstructures, the effect of Ru addition on the elemental partitioning ratios appears to be minimal; however Ru does decrease the propensity for the precipitation of TCP phases [22]. From a thermodynamic perspective, the degree of supersaturation in the γ phase, which is engendered particularly by elements such as Re, W, and Cr, can control TCP phase precipitation. From the point of view of kinetics, it has been suggested that the addition of Ru can also decrease the coarsening rate of the TCP phases [23]; since Ru and Re are both hcp in crystal structure, a strong Ru–Re bond can possibly slow down the diffusion of Re. 4.3. Effect of crystallographic orientation The crystallographic orientation affected the mechanical properties of the Ni-base single crystal superalloys. Ni-base single-crystal superalloys could be characterized by crystallographic anisotropy
since their fatigue properties [24] and creep behavior [25] are evidently crystallographic-orientation dependent. The orientation dependence of creep rupture lives of Ni-base single crystal superalloys has been reported by MacKay and Mayer [26] and Caron et al. [27].They reported long lives for orientation near [001] and short lives for samples oriented close to [011]. An orientation relationship between TCP precipitates and matrix superalloys has also been reported [28]. Consequently, the aluminized Ni-based superalloys exhibit orientation dependence in terms of microstructure. In fact, in this study, the TCP phase precipitated with a distinct layout as shown by EBSD mapping orientation (Figs. 6 and 7). TCP phases formed coherently on {111} planes in platinum-modified aluminized Ni-based single-crystal substrates [29]. The TCP preferred precipitation orientation affected the extension of zone under the IDZ, termed the SDZ, which occurred during creep test in aluminized NKH-510. The SDZ formed instead of SRZ, since recrystallization was not observed in this study. A schematic of TCP phase growth in both orientations is illustrated in Fig. 8. On
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
8.0
Table 3 Resolved shear stress (τ) on each of {111}〈101〉 slip systems for the two specimens types.
{100}
Groupa T
Principal slip system
τ
1. (111)[101]
p1ffiffiðs22 s33 Þ 6
2. (111)[110] 3. (111)[101] 4. (111)[101] W
5. (111)[011]
p1ffiffiðs22 s11 Þ 6
6. (111)[011] 7. (111)[011] 8. (111)[011] {110}
T
1. (111)[101]
p1ffiffiðs22 s33 Þ 6
2. (111)[011]
{100}
7.5 Effective cross-section area, mm2
Side-surface orientation
7.0 6.5
{110}
6.0 5.5 5.0 4.5 4.0 0
3. (111)[101]
5. (111)[011]
50
100
150
200
Crack length on {111} plane, µm
4. (111)[011] W
151
p1ffiffiðs22 s11 Þ 6
6. (111)[101] 7. (111)[101] 8. (111)[011] a
If one of the T group slip systems operate preferentially, this results in a contraction of the thickness of the specimen. If the W group slip systems operate, this results in a contraction of the width.
the cross section, the TCP phases grow in a ‘x’ shape on the {100} side-surface specimen and in a ‘þ ’ on the {110} side-surface, due to the arrangement of {111} planes (Fig. 1) on which the TCP precipitates. The TCP preferred precipitation orientation affected the extension of SDZ (Table 2), which occurred during plastic deformation. The SDZ thickness will then affect the creep rupture life of aluminized single crystal through the reduction in effective cross-section area. However, the differences of reduction in effective cross-section area between the two orientations were very small (Table 2). As a result, side-surface orientation dependence of creep rupture life could not be explained by the thickness difference of the diffusion layers between two orientations. In aluminized specimens, the crack nucleation preferentially occurs in the coated layer rather than in the parent substrate for the conventionally and directionally solidified nickel-base alloys [10]. The main reason for the preferential formation of microcracks in the coated layer is the significantly higher hardness of the diffusion zone in the coated layer relative to the substrate (Fig. 10). These higher hardness values in the coated layers suggest an explanation for the phenomenon of anisotropic creep properties. As shown in Fig. 9, the fracture surfaces exhibited two different fracture modes. In the mode I region, small micro-voids formed and coalesced in the interior of the substrate. In contrast, in the mode-II region, cracks propagated along {111} crystallographic planes, which in the normal direction makes an angle of 54.71 with the tensile stress axis. Thus, the operation of the {111}〈101〉 slip system would result in a slip-band de-cohesion fracture along {111} the slip planes near the surface. The ratio or percentage of the {111} plane (compared to the overall area) was estimated (Fig. 9). After measurement, the percentage of crystallographic {111} facets associated with the activated slip system was about 11% for the specimen with {100} side-surface and 27% for the specimen with {110} side-surface. This confirmed that the specimens with the {110} side-surface had a larger percentage of crystallographic {111} facets than the specimens with the {100} side-surface. The arrangement of slip systems on {111} affects the creep strength [30]. Each specimen has eight equivalent slip systems (Fig. 1). The resolved shear stresses on the slip planes under a multi-axial stress state are shown in Table 3. Under a multi-axial
Fig. 14. Relationship between a crack length on a {111} plane and the effective cross-section area, assuming that a crack propagates on a {111} plane along the maximum shear-stress direction.
stress state, the resolved shear stress τ is decreased by the lateral stresses: s11 in the width direction or s33 in the thickness direction. Consequently, a smaller τ will be applied on the slip plane. In this study, a free surface of the specimen will be under the plane stress condition, i.e., s11 and s33 are approximately equal to zero; therefore, a large shear stress τ will be applied on the slip plane near the specimen surface. This large shear stress will promote crack initiation and growth. Fig. 14 shows the relationship between crack length nucleated on the {111} planes and the effective cross section area, assuming that a crack propagates on a {111} plane along the maximum shear-stress direction. Crack propagation on the {111} planes causes a larger decrease in the effective cross section area of the {110} side-surface specimen than of the {100} side-surface specimen. As shown in Fig. 4, the side surface effect was more pronounced in NKH-510 containing Ru than in NKH-304.The higher-hardness layer produced by the Ru solid solution strengthening (Fig. 10) might have amplified the side-surface orientation effect through the ease of crack initiation and propagation. In conclusion, the {110} specimens exhibited lower creep rupture life than the {100} specimens. This was due to the difference in the crystallographic geometry of the {111} slip planes on which microcracks form in the IDZ hard layer and along the interface of the TCP. However, regardless of the very slight amount of TCP precipitation, NKH-304 showed the secondary orientation effect; therefore, hard layers in the diffusion zones can be deemed a common degradation mechanism in aluminized superalloys. TCP precipitation would be an additional degradation factor. Ru would amplify the sidesurface-orientation effect by hardening the diffusion zones even more.
5. Conclusions The creep properties of Ru-containing and Ru-free Ni-base single crystal superalloys coated by the pack cementation process were investigated. Based on the experimental studies, the conclusions can be summarized as follows: 1. The creep strength was decreased by the aluminized coating treatment. Surface micro-cracks caused by the higherhardness layer in IDZ and TCP phase formation in the SDZ decreased creep rupture life in aluminized specimens. 2. The {110} specimen exhibited shorter creep rupture life than the {100} specimen, this was due to the difference between the two kinds of specimens in the crystallographic geometry of the {111} slip planes on which micro-cracks formed.
152
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
3. The addition of Ru improved creep strength in the aluminized NKH-510 superalloy, as Ru is a solid solution strengthening element. On the other hand, the addition of Ru seemed to indirectly provoke the formation of TCP phase beneath the coating. The influence of the secondary orientation on creep life was more pronounced in NKH-510 than in NKH-304 because of TCP precipitation, which developed profoundly, and the harder diffusion layers in the former.
References [1] G.R. Krishna, D.K. Das, V. Singh, S.V. Joshi, Mater. Sci. Eng. A 251 (1998) 40–47. [2] J. Angenete, K. Stiller, Mater. Sci. Eng. A 316 (2001) 182–194. [3] Y. Zhang, J.A. Haynes, W.Y. Lee, I.G. Wright, B.A. Pint, K.M. Cooley, P.K. Liaw, Metall. Mater. Trans. A 30 (1999) 2679–2687. [4] B.M. Warnes, Surf. Coat. Technol. 146–147 (2001) 7–12. [5] R.C. Reed, The Superalloys: Fundamentals and Applications, Cambridge University Press, London, UK, 2006. [6] A.K. Sinha, Progress in Materials Science,, Pergamon, New York (1972) 81. [7] C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. [8] T.M. Pollock, Mater. Sci. Eng. B 32 (1995) 255–266. [9] M.S.A. Karunaratne, C.M.F. Rae, R.C. Reed, Metall. Mater. Trans. A 32 (2001) 2409–2421.
[10] S.B. Kang, Y.G. Kim, Mater. Sci. Eng. A 83 (1986) 75–86. [11] V.I. Nikitin, T.N. Grigor’eva, Mater. Sci. 10 (1975) 5–9. [12] K.S. O’Hara, W.S. Walston, E.W. Ross, R. Darolia, US Patent no. 5,482,789, General Electric Company, 1996. [13] A.C. Yeh, S. Tin, Scr. Mater. 52 (2005) 519–524. [14] R.A. Hobbs, L. Zhang, C.M.F. Rae, S. Tin, Mater. Sci. Eng. A 489 (2008) 65–76. [15] T. Link, M. Feller-Kniepmeier, Metall. Trans. A 23A (1992) 99–105. [16] T.M. Pollock, A.S. Argon, Acta Metall. Mater. 40 (1992) 1–30. [17] N. Tsuno, K. Kakehi, C.M.F. Rae, R. Hashizume, Metall. Mater. Trans. A 40 (2009) 269–272. [18] A.C. Yeh, S. Tin, Metall. Mater. Trans. A 37 (2006) 2621–2631. [19] L.J. Carroll, Q. Feng, T.M. Pollock, ll. Mater. Trans. A 39 (2008) 1290–1307. [20] R.A. Hobbs, L. Zhang, C.M.F. Rae, S. Tin, Metall. Mater. Trans. A 39 (2008) 1014–1025. [21] A.D. Smigelskas, E.O. Kirkendall, Trans. AIME 171 (1947) 130–142. [22] R.C. Reed, A.C. Yeh, S. Tin, S.S. Babu, M.K. Miller, Scr. Mater. 51 (2004) 327–331. [23] A.C. Yeh, S. Tin, Metall. Mater. Trans. A 37 (2006) 2621–2631. [24] M.B. Henderson, J.W. Martin, Acta Mater. 44 (1996) 111–126. [25] V. Sass, V. Glatzel, M. Feller-Kniepmeier, Acta Mater. 44 (1996) 1967–1977. [26] R.A. MacKay, R.D. Mayer, Metall. Mater Trans. A 13 (1982) 1747–1754. [27] P. Caron, Y. Ohta, Y.G. Nakagawa, T. Khan, Superalloys 1988, in: S. Reichmann, et al., (Eds.), The Minerals, Metals and Materials Society, Warrendale, 1988(p. 215). [28] C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. [29] C.M.F. Rae, M.S. Hook, R.C. Reed, Mater. Sci. Eng. A 396 (2005) 231–239. [30] K. Kakehi, Metall. Mater. Trans. A 30 (1999) 1249–1259.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Influences of ruthenium and crystallographic orientation on creep behavior of aluminized nickel-base single crystal superalloys F.H. Latief a,n, K. Kakehi a, H. An-Chou Yeh b, H. Murakami c a
Department of Mechanical Engineering, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji-shi, Tokyo 192-0397, Japan Department of Materials Science and Engineering, National TsingHua University, 101, Section 2, Kuang-Fu Road, Hsinchu 30013, Taiwan c Hybrid Materials Center, National Institute for Materials Science (NIMS), 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 9 July 2013 Received in revised form 28 October 2013 Accepted 30 October 2013 Available online 13 November 2013
The influences of ruthenium and surface orientation on creep behavior of aluminized Ni-base single crystal superalloys were investigated by comparing two different types of NKH superalloys. The aluminized coated specimens were then subjected to creep rupture tests at a temperature of 900 1C and a stress of 392 MPa. The coating treatment resulted in a significant decrease in creep rupture lives for both superalloys. The diffusion zones between the coating and substrate led to changes in microstructure, which diminished the creep behavior of the aluminized superalloys. Because of the interdiffusion of Ru, Al and Ni, the solubility of some of the refractory elements, such as W, Re. Mo, Co and Cr decreased in the diffusion zone; the precipitation of topologically close-packed (TCP) phases was thus inevitable. In the present study, the addition of Ru increased the degree of Re and Cr supersaturation in the γ matrix. Consequently, the addition of Ru indirectly promoted the precipitation of TCP phases in aluminized Ni-base single crystal superalloys. Furthermore, the growth of TCP precipitates was greatly influenced by the specific surface orientations of the Ni-base single crystal superalloys. In conclusion, the {110} specimens showed shorter creep rupture life than the {100} specimens, this was due to the difference in the crystallographic geometry of {111}〈101〉 slip system and TCP precipitates between the two side-surface orientations of the specimens. & 2013 Elsevier B.V. All rights reserved.
Keywords: Ruthenium Ni-base superalloy Creep Crystallographic orientation Aluminide coating
1. Introduction Aluminide coatings have been applied to Ni-base single crystal superalloys used for turbine components in order to protect them from oxidation and corrosion during their service lives. Details of these coatings have been reported [1–5]. Problems have arisen with these coated superalloys, leading to the loss of the coating. These problems include the precipitation of topologically closepacked (TCP) phases, the formation of Secondary Reaction Zones (SRZs) and the ‘rumpling’ of the coating caused by thermal cycling. Such problems degrade the properties of coated Ni-based superalloys, and are potentially life-limiting to turbine blades. In particular, TCP phase precipitation and SRZ formation are regarded as serious problems in both coated and uncoated Ni-base superalloys. As is well known, topologically close-packed (TCP) phase is a collective designation for several intermetallic compounds rich in the elements W, Mo, Re and Cr precipitated in Ni-base single crystal superalloys [6]. TCP phases form during service at elevated temperatures in such superalloys with high concentrations of
n
Corresponding author. Tel./fax: þ 81 42 6772709. E-mail address: [email protected] (F.H. Latief).
0921-5093/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2013.10.092
these elements added to promote strength the loss of these elements from the matrix impairs the mechanical properties of the blade. Precipitation of the TCP phases [7–9] deteriorates the ductility and creep resistance of Ni-base superalloys. However, the damage mechanisms by aluminide coatings are not yet well understood. In the present study, we investigated the influence of crystallographic orientation on creep behavior of aluminized Ni-base single crystal superalloys, one Ru-free and the other Ru-containing. 2. Experimental materials and procedures Fully heat-treated (solution and aging) Ni-base single crystal superalloys NKH-304 and NKH-510 were used as substrate materials in this study. The details of chemical composition and heat treatment route for each superalloy are given in Table 1. NKH-304 is the base alloy, and NKH-510 is a 3 mass% Ru-addition alloy, where 3% Ru is substituted for Ni. The creep specimens with {100} and {110} side-surfaces (Fig. 1) were prepared by electric discharge machining (EDM), and had cross-section areas of 2.8 mm 2.8 mm and gauge length of 19.6 mm. The specimens were mechanically polished with emery paper prior to the aluminizing process.
144
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Table 1 Chemical composition and heat treatment of alloys. Alloy
Chemical composition (mass%)
Heat treatment
NKH-304
11Co, 6Cr, 6W, 5.4Al, 1.4Ti, 6.8Ta, 4.8Re, 0.12Hf and bal. Ni
Solution heat treatment: 1310 1C/10 h þ1320 1C/12 h þ 1325 1C/12 h þgas furnace cooling and followed by aging treatment: 1180 1C/4 hþ 870 1C/20 h þ air cooling
NKH-510
11Co, 6Cr, 6W, 5.4Al, 1.4Ti, 6.8Ta, 4.8Re, 0.12Hf, 3Ru and bal. Ni
Solution heat treatment: 1310 1C/10 hþ 1320 1C/12 h þ gas furnace cooling and followed by aging treatment: 1160 1C/4 h þ870 1C/20 h þ air cooling
the coating powders mixture and the outward diffusion of Ni from the substrate at the interface area, which was composed of β-NiAl due to the heat treatment. The thickness of the IDZ seemed to vary in both side-surface orientations as shown in Fig. 3. The IDZ thickness of the aluminized NKH-304 was about 3.0 μm on the {100} side-surface and 4.7 μm on the {110} side-surface; meanwhile, the IDZ thickness of aluminized NKH-510 was about 2.3 μm on the {100} side-surface and 3.9 μm on the {110} side-surface. In other words, differences in IDZ thickness ensued in all specimens. This means that the presence of the coating rendered the changes in the effective cross-section area that affects the creep behavior of Ni-base superalloys. 3.2. Creep behavior
Fig. 1. Arrangement of {111}〈101〉 slip systems for two kinds of creep specimens with (a) {100} and (b) {110} side-surface orientations.
The specimens were then embedded in an Al2O3 retort containing a mixture of 24.5 mass% Al, 24.5 mass% Cr, 49 mass% Al2O3 and 2 mass% NH4Cl powders, and heated at 1000 1C for 5 h in flowing argon for the aluminizing treatment called the pack cementation process. The coated superalloys were then subjected to creep test with a temperature of 900 1C and a stress of 392 MPa with a parallel load 〈001〉 direction. For comparison, the creep test was also conducted on uncoated specimen under same conditions in order to determine the effect of the aluminide coating on the creep behavior of Ni-base single crystal superalloys. The microstructural changes before and after the creep rupture tests were observed by scanning electron microscopy (SEM). The compositional profiles through the cross section of samples were analyzed by electron probe microanalysis (EPMA). The orientation of crystal grain was identified by a TSL-OIM electron back scattered diffraction (EBSD) analyzer.
3. Results 3.1. Microstructures after heat treatment and coating The microstructures of the γ' precipitates after aging heat treatment are shown in Fig. 2. No obvious differences in the morphology of γ′ precipitates were observed between the two alloys. The mean γ′ precipitate lengths were 345 nm and 310 nm in the alloys NKH-304 and NKH-510 respectively. The γ′ precipitate size of the Ru-containing alloy NKH-510 was a little smaller than that of the base alloy. SEM micrographs of as-aluminized Ni-base single crystal superalloys NKH are shown in Fig. 3. Three distinctive layers can be distinguished from the micrographs: (i) the coating as an outer protective layer, (ii) an interdiffusion zone (abbreviated as IDZ), and (iii) a substrate. IDZ was derived from the reaction between the inward diffusion of Al, produced by
Fig. 4 shows the variation in creep rupture lives performed at a temperature of 900 1C and a stress of 392 MPa. The two superalloys exhibited similar trends, while the bare specimens with the {100} side-surface orientation demonstrated higher creep rupture life than the {110} side-surface orientation. It was clear that the coating treatment resulted in a significant decrease in creep rupture life in both superalloys. Furthermore, the difference in creep rupture life between the two orientations indicated that anisotropy in creep had occurred. In general, the aluminized NKH510 had a higher creep rupture life than the aluminized NKH-304, in both the {100} and {110} side-surface specimens. The creep rupture lives of aluminized NKH-304 were 311 h for the {100} side-surface specimen and 253 h for the {110} side-surface specimen, while the measures for aluminized NKH-510 were 641 h for the {100} side-surface specimen and 539 h for the {110} sidesurface specimen. The differences in creep rupture lives for both side-surface orientations were 58 h (19%) for the aluminized NKH304 and 102 h (16%) for the aluminized NKH-510. 3.3. Microstructures after creep rupture test The SEM micrographs of aluminized specimens after creep rupture test are presented in Fig. 5. An additional zone formed beneath the IDZ in all specimens; this additional zone is called the substrate diffusion zone (SDZ). The SDZ must be differentiated from the IDZ in that it remains particularly γ′ and retains the orientation of the single crystal substrate. However, the secondary reaction zone (SRZ) could not be observed in the two superalloys, as the localized recrystallization had not yet occurred at this stage. The diffusion layer thickness was approximately 23.3 μm on the {100} side-surface specimen and 32.9 μm on the {110} side-surface specimen in the aluminized NKH-304 (Fig. 5a and b). Meanwhile the diffusion layer thickness was approximately 21.6 μm on the {100} side-surface specimen and 33.1 μm on the {110} side-surface specimen in the aluminized NKH-510 (Fig. 5c and d). In addition, the TCP phase was precipitated in the {100} and {110} side-surface specimens for both superalloys, but in differing amounts. The aluminized NKH-510 contained much more TCP phase than the aluminized NKH-304. As crystallographic orientation affected the growth of TCP phases, the formation of these phases was different in the two orientations. The TCP phase showed a
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
145
2 µm Fig. 2. SEM bright field images of (a) NKH-304 and (b) NKH-510.
Fig. 3. SEM micrographs of aluminized specimens: (a,b) NKH-304, and (c,d) NKH-510 with the side-surface orientations of (a,c) {100} and (b,d) {110}.
Fig. 4. Creep rupture curves of various superalloys: (a) NKH-304 and (b) NKH-510.
particular formation with a 451 angle to the substrate interface on the {100} side-surface specimen, but were vertically aligned on {110} the side-surface interface. EBSD crystallographic phase mapping of the aluminized NKH-304 and NKH-501 superalloys after creep rupture is shown in Figs. 6 and 7, respectively. The single-crystal substrates and orientation maps are shown from the reference direction (RD), which is normal to the substrate/coating
interface. EBSD mapping images are the corresponding orientation maps assigned as face-centered cubic (fcc) and body-centered cubic (bcc) structures. It was obvious from the figures that the surface aluminized layer consisted of bcc polycrystalline grains, while the diffusion layer consisted of bcc and fcc phases. EBSD can be utilized to elucidate the relationship between crystal orientation and TCP formation. The EBSD mapping images confirmed the
146
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 5. SEM micrographs of aluminized specimens after creep rupture test: (a,b) NKH-304, and (c,d) NKH-510 with the surface orientations of (a,c) {100} and (b,d) {110}.
Fig. 6. (a) Cross-sectional SEM image of aluminized NKH-304 and corresponding orientation maps assigned as (b) fcc and (c) bcc structures.
formation of TCP phases on both orientations but the SRZ was not observed in all specimens. The voids were apparent in the aluminized NKH-510; furthermore the specimen with the {110} side-surface (Fig. 5d) had many more voids rather than the specimen with the {100} side-surface (Fig. 5c). The schematic of TCP phase growth and voids in both orientations are illustrated in Fig. 8.
3.4. Fracture surfaces Fracture surfaces of aluminized specimens subjected to creep at 900 1C/392 MPa were observed on both the {100} and {110} sidesurface orientations. The aluminized NKH-510 specimens after creep rupture test were observed by SEM, as shown in Fig. 9. The fracture surface in both side-surface specimens showed a
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
147
Fig. 7. (a) Cross-sectional SEM image of aluminized NKH-510 and corresponding orientation maps assigned as (b) fcc and (c) bcc structures.
dimple pattern, and the dimple fracture dominated on the fracture surfaces. Furthermore, cleavage facets (slip planes) were found in both the {100} and {110} side-surface specimens. The difference in creep rupture life revealed that a structural instability occurred during the creep test. This structural instability was considered to be crack initiation at the surfaces. In the case of the coated specimens, creep fracture initiation would be virtually controlled by the crack nuclei in the coated layer.
4. Discussion 4.1. Effect of coating treatment In this study, we investigated the effects of ruthenium and crystallographic orientation on the creep behavior of aluminized Ni-base single crystal. Coating treatments are intended to protect the substrate surface from environmental attack when superalloys are exposed to high operating temperatures with aggressive oxidation and corrosive environment. However, the coating had been shown to degrade the mechanical properties of the coated superalloy. While the coating is not primarily formulated to improve the mechanical properties of the superalloy, the treatment (aluminizing) has been found to significantly impact the superalloy's mechanical properties due to the changes in microstructure. In addition, the coating treatment resulted in a decrease of the effective cross-section area by forming coating layers. This reduction in effective cross-section area due to the formation of coating diffusion layers might be a factor in the reduction of creep strength. However, this effect was tiny as can be seen in Table 2. It is therefore insufficient to explain the reduction of creep life. The change in microstructure in the coated specimens was due to the formation of the IDZ. The increased concentration of Al in the IDZ led to lower strength and eventually to premature failure of the specimen. For details, consider again the creep rupture curves in Fig. 4 where the coated Ni-base single crystal superalloy specimens have a lower creep rupture than the uncoated ones. In
Fig. 8. Schematic illustrations of TCP preferred precipitation orientation on {111} planes for: (a) {100} and (b) {110} side-surface orientations.
the coated specimen, the crack nucleation occurs in the coating layer and propagates according to available regions until final rupture. It is a certainty that this condition motivates the formation of microcracks in the coated layer. Hence, the main reason for the preferential formation of microcracks in the coated layer can be attributed to the significantly higher hardness of the diffusion zone in the coated layer relative to the substrate [10], as shown in
148
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 9. Fracture surfaces of NKH 510 aluminized specimens with: (a) {100} side-surface, and (b) {110} side-surface specimens subjected to creep rupture test at 900 1C and 392 MPa.
4.2. Effect of Ru content
Table 2 Decrease ratio in effective cross-section area (%). Alloy
Total diffusion layer thickness (IDZ þ SDZ) (μm)
Ratio of reduction in effective cross section area (%)
{100}
{110}
{100}
{110}
32.9 33.1
3.3 3.1
4.8 4.7
NKH-304 23.3 NKH-510 21.6
Fig. 10. Microhardness along the coating-substrate regions in aluminized superalloys: (a) as-aluminized and (b) after creep rupture test.
Fig. 10. It is also possible that the decrease of creep rupture might be due to a weakening of atomic bonds, because aluminum is a low melting metal, its effect could be considered similar to that of other low melting metals [11].
In this study, we were interested in investigating the effect of the addition of Ru on the creep behavior of aluminized Ni-based single crystal superalloys. The addition of 3 mass% Ru into superalloy NKH-510 improved the creep behavior of this superalloy when compared to the superalloys without Ru NKH-304. Ru is a solid solution strengthening element that has been added to the next generation of superalloys. The addition of Ru at a higher level than the solubility limit imposed the solvus of the γ/γ′ two-phase region. Early studies on Ru-containing nickel-base single crystal superalloys suggested that Ru additions alter the partitioning of elements between the γ and the γ′ phases [12].The substrate hardness values of Ru-containing alloy (NKH-510) were higher than those of the base alloy (NKH-304) as shown in Fig. 10. Ru addition improved strength in both the γ and γ′ phases [13]. There was also a significant increase in flow stresses of the single-phase γ alloy with Ru, and a modest increase in the flow stress of the single-phase γ′ alloy was also measured [13]. These results provide supportive evidence for the solid solution strengthening effect of Ru. Since the element Ru also has an hcp structure, its addition would be expected to decrease the stacking fault energy of the matrix [14]. The a/2〈110〉 dislocation motion in the matrix channel leaves interfacial dislocations around cuboidal precipitates and forms homogeneous γ–γ' interfacial dislocation networks [15,16]; the addition of Ru extends the width between the partial dislocations in the matrix [17], and would inhibit the cross slip of dislocation at this temperature. Moreover, the addition of Ru into Ni-base single crystal superalloys could enhance microstructural stability, augment resistance to the formation of TCP phases [18], improve the high temperature mechanical properties or environmental resistance [19], and increase the liquidus temperature [20]. During exposure to elevated temperatures, the formation of TCP phases provokes serious deterioration in creep strength. The TCP phases were slightly precipitated in aluminized NKH-304, but considerably precipitated in aluminized NKH-510, as shown in Fig. 5c and d. TCP phases are composed mainly of the elements Ni, Cr, Mo, Co, W, and Re, and unfortunately this list contains the elements that are most effective at conferring resistance to creep. Therefore, the presence of TCP during service must deplete the matrix of these elements and reduce their solid strengthening effect. We also consider the Ru addition to have contributed to the formation of the SDZ in the aluminized NKH-510. For details, see the schematic illustration of the SDZ formation and the interaction with TCP growth in aluminized NKH-510 presented in Fig. 8. The reaction between outward diffusion of Ru and inward diffusion of Al forms β-RuAl nearby IDZ, with respect to SDZ formation. Furthermore, voids were observed in aluminized NKH-510 (Fig. 5c and d). The formation of such voids is generally referred to as the Kirkendall
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
149
Fig. 11. EPMA profiles of ruptured aluminized NKH-510 superalloys: (a) {100} side-surface and (b) {110} side-surface specimen.
effect, in which voids are formed due to the different diffusion rates of two or more metal atoms [21]. When coated specimens are exposed to high temperatures, the volume changes at the interface of the coating and substrate by the transformation of β-NiAl into γ ′-Ni3Al lead Al to diffuse outward from the substrate diffusion zone (SDZ) as it oxidizes, forming the voids in the SDZ. The results of the EPMA analysis (Figs. 11 and 12) showed that Re and Cr contents were enriched in the two regions where the TCP phases were present; the first region was in IDZ, where Ru and Al were enriched, and the other was the Ru depletion region within the SDZ. When NKH-510 (Fig. 11) is compared to NKH-304 (Fig. 12), it is clear that
the underlying mechanism that explains why for the Ru-bearing NKH-510 has more TCP precipitates beneath the coating has to do with the increase in supersaturation of Re and Cr in the IDZ due to the presence of RuAl. The absence of Ru in the SDZ would allow the TCP precipitates to propagate more easily from the IDZ into the SDZ, resulting in a greater depth of TCP-phase region beneath the coating of NKH-510 (Fig. 13). As a consequence, the addition of Ru indirectly promoted the precipitation of TCP phases in the aluminized NKH-510 specimens. It has previously been reported that the addition of Ru suppressed the TCP phase precipitation [18]. In contrast, in the
150
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
Fig. 12. EPMA profiles of ruptured aluminized NKH-304 superalloy with {100} side-surface specimen.
Fig. 13. Schematic illustrations of the SDZ formation process: (a) outward diffusion of Ru from the substrate to the IDZ, and (b) SDZ formation with the precipitation of TCP phase in the γ/γ′ substrate.
present study, the addition of Ru seemed to promote the formation of TCP phases beneath the coating. For the bulk microstructures, the effect of Ru addition on the elemental partitioning ratios appears to be minimal; however Ru does decrease the propensity for the precipitation of TCP phases [22]. From a thermodynamic perspective, the degree of supersaturation in the γ phase, which is engendered particularly by elements such as Re, W, and Cr, can control TCP phase precipitation. From the point of view of kinetics, it has been suggested that the addition of Ru can also decrease the coarsening rate of the TCP phases [23]; since Ru and Re are both hcp in crystal structure, a strong Ru–Re bond can possibly slow down the diffusion of Re. 4.3. Effect of crystallographic orientation The crystallographic orientation affected the mechanical properties of the Ni-base single crystal superalloys. Ni-base single-crystal superalloys could be characterized by crystallographic anisotropy
since their fatigue properties [24] and creep behavior [25] are evidently crystallographic-orientation dependent. The orientation dependence of creep rupture lives of Ni-base single crystal superalloys has been reported by MacKay and Mayer [26] and Caron et al. [27].They reported long lives for orientation near [001] and short lives for samples oriented close to [011]. An orientation relationship between TCP precipitates and matrix superalloys has also been reported [28]. Consequently, the aluminized Ni-based superalloys exhibit orientation dependence in terms of microstructure. In fact, in this study, the TCP phase precipitated with a distinct layout as shown by EBSD mapping orientation (Figs. 6 and 7). TCP phases formed coherently on {111} planes in platinum-modified aluminized Ni-based single-crystal substrates [29]. The TCP preferred precipitation orientation affected the extension of zone under the IDZ, termed the SDZ, which occurred during creep test in aluminized NKH-510. The SDZ formed instead of SRZ, since recrystallization was not observed in this study. A schematic of TCP phase growth in both orientations is illustrated in Fig. 8. On
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
8.0
Table 3 Resolved shear stress (τ) on each of {111}〈101〉 slip systems for the two specimens types.
{100}
Groupa T
Principal slip system
τ
1. (111)[101]
p1ffiffiðs22 s33 Þ 6
2. (111)[110] 3. (111)[101] 4. (111)[101] W
5. (111)[011]
p1ffiffiðs22 s11 Þ 6
6. (111)[011] 7. (111)[011] 8. (111)[011] {110}
T
1. (111)[101]
p1ffiffiðs22 s33 Þ 6
2. (111)[011]
{100}
7.5 Effective cross-section area, mm2
Side-surface orientation
7.0 6.5
{110}
6.0 5.5 5.0 4.5 4.0 0
3. (111)[101]
5. (111)[011]
50
100
150
200
Crack length on {111} plane, µm
4. (111)[011] W
151
p1ffiffiðs22 s11 Þ 6
6. (111)[101] 7. (111)[101] 8. (111)[011] a
If one of the T group slip systems operate preferentially, this results in a contraction of the thickness of the specimen. If the W group slip systems operate, this results in a contraction of the width.
the cross section, the TCP phases grow in a ‘x’ shape on the {100} side-surface specimen and in a ‘þ ’ on the {110} side-surface, due to the arrangement of {111} planes (Fig. 1) on which the TCP precipitates. The TCP preferred precipitation orientation affected the extension of SDZ (Table 2), which occurred during plastic deformation. The SDZ thickness will then affect the creep rupture life of aluminized single crystal through the reduction in effective cross-section area. However, the differences of reduction in effective cross-section area between the two orientations were very small (Table 2). As a result, side-surface orientation dependence of creep rupture life could not be explained by the thickness difference of the diffusion layers between two orientations. In aluminized specimens, the crack nucleation preferentially occurs in the coated layer rather than in the parent substrate for the conventionally and directionally solidified nickel-base alloys [10]. The main reason for the preferential formation of microcracks in the coated layer is the significantly higher hardness of the diffusion zone in the coated layer relative to the substrate (Fig. 10). These higher hardness values in the coated layers suggest an explanation for the phenomenon of anisotropic creep properties. As shown in Fig. 9, the fracture surfaces exhibited two different fracture modes. In the mode I region, small micro-voids formed and coalesced in the interior of the substrate. In contrast, in the mode-II region, cracks propagated along {111} crystallographic planes, which in the normal direction makes an angle of 54.71 with the tensile stress axis. Thus, the operation of the {111}〈101〉 slip system would result in a slip-band de-cohesion fracture along {111} the slip planes near the surface. The ratio or percentage of the {111} plane (compared to the overall area) was estimated (Fig. 9). After measurement, the percentage of crystallographic {111} facets associated with the activated slip system was about 11% for the specimen with {100} side-surface and 27% for the specimen with {110} side-surface. This confirmed that the specimens with the {110} side-surface had a larger percentage of crystallographic {111} facets than the specimens with the {100} side-surface. The arrangement of slip systems on {111} affects the creep strength [30]. Each specimen has eight equivalent slip systems (Fig. 1). The resolved shear stresses on the slip planes under a multi-axial stress state are shown in Table 3. Under a multi-axial
Fig. 14. Relationship between a crack length on a {111} plane and the effective cross-section area, assuming that a crack propagates on a {111} plane along the maximum shear-stress direction.
stress state, the resolved shear stress τ is decreased by the lateral stresses: s11 in the width direction or s33 in the thickness direction. Consequently, a smaller τ will be applied on the slip plane. In this study, a free surface of the specimen will be under the plane stress condition, i.e., s11 and s33 are approximately equal to zero; therefore, a large shear stress τ will be applied on the slip plane near the specimen surface. This large shear stress will promote crack initiation and growth. Fig. 14 shows the relationship between crack length nucleated on the {111} planes and the effective cross section area, assuming that a crack propagates on a {111} plane along the maximum shear-stress direction. Crack propagation on the {111} planes causes a larger decrease in the effective cross section area of the {110} side-surface specimen than of the {100} side-surface specimen. As shown in Fig. 4, the side surface effect was more pronounced in NKH-510 containing Ru than in NKH-304.The higher-hardness layer produced by the Ru solid solution strengthening (Fig. 10) might have amplified the side-surface orientation effect through the ease of crack initiation and propagation. In conclusion, the {110} specimens exhibited lower creep rupture life than the {100} specimens. This was due to the difference in the crystallographic geometry of the {111} slip planes on which microcracks form in the IDZ hard layer and along the interface of the TCP. However, regardless of the very slight amount of TCP precipitation, NKH-304 showed the secondary orientation effect; therefore, hard layers in the diffusion zones can be deemed a common degradation mechanism in aluminized superalloys. TCP precipitation would be an additional degradation factor. Ru would amplify the sidesurface-orientation effect by hardening the diffusion zones even more.
5. Conclusions The creep properties of Ru-containing and Ru-free Ni-base single crystal superalloys coated by the pack cementation process were investigated. Based on the experimental studies, the conclusions can be summarized as follows: 1. The creep strength was decreased by the aluminized coating treatment. Surface micro-cracks caused by the higherhardness layer in IDZ and TCP phase formation in the SDZ decreased creep rupture life in aluminized specimens. 2. The {110} specimen exhibited shorter creep rupture life than the {100} specimen, this was due to the difference between the two kinds of specimens in the crystallographic geometry of the {111} slip planes on which micro-cracks formed.
152
F.H. Latief et al. / Materials Science & Engineering A 592 (2014) 143–152
3. The addition of Ru improved creep strength in the aluminized NKH-510 superalloy, as Ru is a solid solution strengthening element. On the other hand, the addition of Ru seemed to indirectly provoke the formation of TCP phase beneath the coating. The influence of the secondary orientation on creep life was more pronounced in NKH-510 than in NKH-304 because of TCP precipitation, which developed profoundly, and the harder diffusion layers in the former.
References [1] G.R. Krishna, D.K. Das, V. Singh, S.V. Joshi, Mater. Sci. Eng. A 251 (1998) 40–47. [2] J. Angenete, K. Stiller, Mater. Sci. Eng. A 316 (2001) 182–194. [3] Y. Zhang, J.A. Haynes, W.Y. Lee, I.G. Wright, B.A. Pint, K.M. Cooley, P.K. Liaw, Metall. Mater. Trans. A 30 (1999) 2679–2687. [4] B.M. Warnes, Surf. Coat. Technol. 146–147 (2001) 7–12. [5] R.C. Reed, The Superalloys: Fundamentals and Applications, Cambridge University Press, London, UK, 2006. [6] A.K. Sinha, Progress in Materials Science,, Pergamon, New York (1972) 81. [7] C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. [8] T.M. Pollock, Mater. Sci. Eng. B 32 (1995) 255–266. [9] M.S.A. Karunaratne, C.M.F. Rae, R.C. Reed, Metall. Mater. Trans. A 32 (2001) 2409–2421.
[10] S.B. Kang, Y.G. Kim, Mater. Sci. Eng. A 83 (1986) 75–86. [11] V.I. Nikitin, T.N. Grigor’eva, Mater. Sci. 10 (1975) 5–9. [12] K.S. O’Hara, W.S. Walston, E.W. Ross, R. Darolia, US Patent no. 5,482,789, General Electric Company, 1996. [13] A.C. Yeh, S. Tin, Scr. Mater. 52 (2005) 519–524. [14] R.A. Hobbs, L. Zhang, C.M.F. Rae, S. Tin, Mater. Sci. Eng. A 489 (2008) 65–76. [15] T. Link, M. Feller-Kniepmeier, Metall. Trans. A 23A (1992) 99–105. [16] T.M. Pollock, A.S. Argon, Acta Metall. Mater. 40 (1992) 1–30. [17] N. Tsuno, K. Kakehi, C.M.F. Rae, R. Hashizume, Metall. Mater. Trans. A 40 (2009) 269–272. [18] A.C. Yeh, S. Tin, Metall. Mater. Trans. A 37 (2006) 2621–2631. [19] L.J. Carroll, Q. Feng, T.M. Pollock, ll. Mater. Trans. A 39 (2008) 1290–1307. [20] R.A. Hobbs, L. Zhang, C.M.F. Rae, S. Tin, Metall. Mater. Trans. A 39 (2008) 1014–1025. [21] A.D. Smigelskas, E.O. Kirkendall, Trans. AIME 171 (1947) 130–142. [22] R.C. Reed, A.C. Yeh, S. Tin, S.S. Babu, M.K. Miller, Scr. Mater. 51 (2004) 327–331. [23] A.C. Yeh, S. Tin, Metall. Mater. Trans. A 37 (2006) 2621–2631. [24] M.B. Henderson, J.W. Martin, Acta Mater. 44 (1996) 111–126. [25] V. Sass, V. Glatzel, M. Feller-Kniepmeier, Acta Mater. 44 (1996) 1967–1977. [26] R.A. MacKay, R.D. Mayer, Metall. Mater Trans. A 13 (1982) 1747–1754. [27] P. Caron, Y. Ohta, Y.G. Nakagawa, T. Khan, Superalloys 1988, in: S. Reichmann, et al., (Eds.), The Minerals, Metals and Materials Society, Warrendale, 1988(p. 215). [28] C.M.F. Rae, R.C. Reed, Acta Mater. 49 (2001) 4113–4125. [29] C.M.F. Rae, M.S. Hook, R.C. Reed, Mater. Sci. Eng. A 396 (2005) 231–239. [30] K. Kakehi, Metall. Mater. Trans. A 30 (1999) 1249–1259.