Improvement of stress–rupture property by Cr addition in Ni-based single crystal superalloys
J.Y. Chen,Q. Feng,L.M. Cao,Z.Q. Sun
DOI: https://doi.org/10.1016/j.msea.2011.01.060
2011-01-01
Abstract:The effects of Cr addition on the microstructure and stress–rupture property have been investigated in three experimental Ni-based single crystal superalloys containing various levels of Cr addition (0–5.7 wt.%). The Re partitioning ratio increased and the lattice misfit became more negative with increasing the Cr addition in both dendrite core and interdendritic region. The changes of elemental partitioning behaviors and the lattice misfit show good agreement with the change of γ′ morphology. Cr addition increased the stress–rupture life at 1100 °C/140 MPa significantly due to higher γ′ volume fraction, more negative lattice misfit and a well rafting structure as well as less width of γ channels. High Cr addition (5.7 wt.%) increased the degree of Re and Cr supersaturation in the γ phase and promoted the formation of topologically close-packed (TCP) phases significantly under thermal exposure and creep deformation at 1100 °C. Keywords Superalloys Chromium Elemental partitioning ratio Lattice misfit γ′ morphology Stress–rupture property 1 Introduction Ni-based single crystal superalloys have been used extensively for turbine blades in aircraft engines. Advanced Ni-based single crystal superalloys typically contain significant levels of refractory alloying additions such as Re, W, Ta and Mo in order to achieve a remarkable creep resistance at elevated temperature [1,2] . However, high levels of refractory alloying additions also result in an increased propensity for the formation of topologically close-packed (TCP) phases during high temperature thermal exposure [3,4] . TCP phases are detrimental to high-temperature performance of superalloys due to their brittle nature and depleting potent solid solution strengthening elements from the γ matrix. Chromium (Cr) is generally added to Ni-based superalloys because it is beneficial to hot corrosion and oxidation resistance [5] . However, high levels of Cr addition also result in detrimental TCP phases. In order to add more refractory alloying additions with keeping microstructural stability, modern commercial Ni-based single crystal [1] and directionally solidified [6] superalloys typically contain lower levels of Cr. In addition, Cr is an effective solid solution strengthener [5] . However, there were seldom systematic investigations on whether Cr addition has significant effect on high-temperature creep property of single crystal superalloys. Yoo et al. predicted the effect of Cr addition on the creep property of Ni-based single crystal superalloys using Bayesian neural network technique and indicated that the creep rupture life decreased with increasing levels of Cr addition [7] , but they did not consider influence factors such as γ′ volume fraction and phase stability. Furthermore, the mechanism by which Cr addition influences the creep property of single crystal superalloys is still not well understood. Early work regarding the effect of Cr addition on Ni-based single crystal superalloys suggested that Cr addition alter the partitioning behaviors of alloying elements (especially Re) between the γ and γ′ phases, and the Re partitioning ratio increased with increasing the Cr content [8–11] . The influence of Cr addition on the elemental partitioning ratio is of particular importance because of its close relationship with the γ/γ′ lattice misfit and γ′ morphology [12,13] . It is well known that the partitioning characteristics of alloying elements control both the sign and magnitude of lattice misfit, which determines the morphology of γ′ precipitates [12] . In addition, a non-zero lattice misfit is known to enhance the rafting process and increase the creep strength at high temperature and low stress because the rafting structure can be an effective barrier for confining dislocation activity within the γ rafts and leading to the steady state of the creep deformation, and the rafting direction is dependent on the direction of the applied stress and the sign of the lattice misfit [14] . Ni-based superalloys that raft perpendicularly to the applied tensile stress have a negative lattice misfit. And alloys with a positive lattice misfit raft parallelly to the applied tensile stress. If the lattice misfit is eliminated, the rafting should not occur. Recently, Carroll et al. reported that Cr addition (3.4–6.7 wt.%) resulted in higher Re partitioning ratio, more cuboidal γ′ precipitates and the corresponding better creep resistance at 950 °C/290 MPa [11,15] . However, systematic investigations were limited regarding the influence of Cr addition on the γ′ morphology, γ/γ′ elemental partitioning ratio and lattice misfit as well as rafting behavior and creep resistance at high temperature and low stress. The experimental single-crystal superalloys containing different levels of Cr (0–5.7 wt.%) were investigated in the current study. The influence of Cr addition on microstructure and stress–rupture property at high temperature and low stress (1100 °C/140 MPa) has been investigated through changes in γ′ morphology, microstructural stability and elemental partitioning ratio as well as lattice misfit and rafting behavior in both dendrite core and interdendritic region. And the mechanisms by which Cr addition improve the stress–rupture property have been discussed. 2 Experimental procedures Three experimental Ni-based single crystal superalloys with different levels of Cr addition (0–5.7 wt.%) were directionally solidified (DS) as single crystal bars (14 mm in diameter and 150 mm in length). Their compositions were determined by an atomic absorption spectroscope and are listed in Table 1 . Alloy USTB-F1 is the baseline alloy without Cr addition. Alloys USTB-F2 and USTB-F3 contain intermediate level (2.8 wt.%) and high level (5.7 wt.%) of Cr, respectively. After solution treatment, thermal exposure tests were carried out at 1100 °C for different time. After the standard aging treatment (1140 °C/8 h + 870 °C/16 h), the specimens of the stress–rupture and creep tests were machined and the dimension of the gauge section was 5 mm in diameter and 25 mm in length. The stress–rupture testing was conducted at 1100 °C/140 MPa for three investigated alloys. Two specimens were tested for the same alloy. In order to investigate the effect of Cr addition on the tendency of the rafting behavior under the stress condition, the creep testing at 1100 °C was interrupted after the creep strain reached to 1% for three alloys. The maximum creep stress of 130 MPa was selected due to the limit of creep facilities. The stress–rupture and creep tests were conducted in the air. Metallographic specimens for microstructural investigation were prepared and etched in a solution of 1% HF, 33% CH 3 COOH, 33% HNO 3 and 33% H 2 O. Microstructural observations were conducted using a ZEISS SUPRA 55 field-emission scanning electron microscope (FE-SEM) equipped with an energy-dispersive spectroscope (EDS). A back-scattered electron (BSE) detector was used to differentiate between the dendrite core and the interdendritic region. In order to better describe the morphology of γ′ precipitates, the γ′ morphology was described as follows. γ′ precipitates were considered to be cuboidal and spherical if they were cubes and spheres, respectively. The intermediate shape was intervenient between cuboidal and spherical shape, and it has six planar γ/γ′ interfaces and eight rounded corners. The morphology between spherical and intermediate shape was defined as nearly spherical. And the morphology between intermediate and cuboidal was considered to be nearly cuboidal. The volume fraction of γ′ precipitates was determined using the standard point count method. Measurement of the γ′ size and γ channel width was performed on SEM images using the Image-Pro software, an image analyzer program (see the web site: http://www.mediacy.com/index.aspx?page=IPP for more information). At least five images were used for determining the size and volume fraction of γ′ precipitates as well as the width of γ channels in each alloy. The degree of the rafting behavior after 1% creep stain was characterized by the linear termination density, ρ l , which is represented by the following equation (1) ρ l = ρ a ⋅ λ where ρ a is the average number of γ′ lamella terminations per unit area and λ is the thickness of the γ′ lamella [16] . The smaller the linear termination density, the more complete the rafting structure. Again, at least five SEM images were used for determining the linear termination density in each alloy. The elemental partitioning ratio of γ and γ′ phases, k i , is defined as (2) k i = C i γ C i γ ′ where i represents each of the alloying elements, C i γ and C i γ ′ correspond to the atomic concentration of element i in γ and γ′ phases, respectively. The partitioning ratio k i greater or less than unity, indicates that the constituent element i preferentially partitioning to the γ or γ′ phase, respectively. Concentration profiles of γ and γ′ phases were measured by a JEOL JXA-8800R electron microprobe analyzer (EPMA) with four wavelength-dispersive spectrometers (WDS) under an acceleration voltage of 20 kV, a beam current of 20 nA and a spot size of 1 μm. Concentrations were obtained using a ZAF correction program, with pure element standards. In order to minimize the matrix effect, some γ and γ′ phases were apparently coarsened to at least 3 μm in size after thermal exposure at 1100 °C for 800 h, as shown in Fig. 1 . All the EPMA analyses were carried out in polished but unetched samples. WDS spectra were collected from γ and γ′ phases in at least five areas for each experimental alloy. Values of C i γ and C i γ ′ were averaged from above WDS analyses. Subsequently, lattice parameters of the γ and γ’ phases were calculated by the formula presented by Ref. [17] according to their compositions. And then, the γ/γ′ lattice misfit ( δ ) is calculated as follows: (3) δ = 2 ( a γ ′ − a γ ) a γ ′ + a γ where a γ and a γ′ are the lattice parameters of the γ and γ′ phases, respectively. 3 Results 3.1 Morphology of γ′ precipitates Fig. 2 shows typical microstructures in the dendrite core and interdendritic region of three investigated alloys after aging at 1100 °C for 8 h. In the dendrite core, the morphology of γ′ precipitates in the Cr-free alloy (USTB-F1) was nearly spherical ( Fig. 2 a1). The γ′ morphology was intermediate in alloys USTB-F2 and USTB-F3 ( Fig. 2 b1 and c1), while it was slightly more cuboidal in alloy USTB-F3 than that of alloy USTB-F2. The present results indicate that the morphology of γ′ precipitates in the dendrite core changed to be more cuboidal with increasing levels of Cr addition. Meanwhile, the volume fraction of γ′ precipitates increased with increasing the Cr content. It was 45.3% in alloy USTB-F1 and increased to 55.1% and 57.4% in alloys USTB-F2 and USTB-F3, respectively. However, the size of γ′ precipitates decreased with increasing levels of Cr addition. The γ′ size in alloys USTB-F1 and USTB-F2 was 0.39 μm and 0.37 μm, respectively. It was significantly larger than that in alloy USTB-F3 (0.27 μm, Fig. 2 c1). In addition, it is worthy of noting that the width of γ channel became narrower with increasing levels of Cr addition, and they were 101 nm, 77 nm and 42 nm in alloys USTB-F1, USTB-F2 and USTB-F3, respectively. The γ′ morphology in the interdendritic region of alloys USTB-F1 and USTB-F2 was nearly cuboidal ( Fig. 2 a2) and nearly spherical ( Fig. 2 b2), respectively, and it was different from that in the dendrite core. However, the γ′ morphology in the interdendritic region of alloy USTB-F3 was intermediate ( Fig. 2 c2), and it was similar to that in the dendrite core. In addition, the γ′ volume fractions of Cr-containing alloys USTB-F2 and USTB-F3 were in the same level, and they were higher than that of Cr-free alloy USTB-F1. Meanwhile, the size of γ′ precipitates decreased with increasing the Cr content, while the width of γ channel was in the same level in the interdendritic region of three investigated alloys. The morphology, volume fraction and size of γ′ precipitates as well as the width of γ channels of the investigated alloys are given in Table 2 . Both the size and the volume fraction of γ′ precipitates in the interdendritic region were larger than those in the dendrite core. 3.2 Long-term microstructural stability The typical microstructures in the dendrite core of the investigated alloys after thermal exposure at 1100 °C for different time are shown in Fig. 3 . The nearly spherical γ′ precipitates in alloy USTB-F1 did not change in morphology and coarsened to 1.50 μm after thermal exposure for 2000 h ( Fig. 3 a), while alloy USTB-F2 with intermediate γ′ precipitates coarsened more severely, and the γ′ size was 1.95 μm after 2000 h ( Fig. 3 b). In addition, no secondary phase precipitated in the dendrite core of alloys USTB-F1 and USTB-F2 even after 2000 h. However, TCP phases precipitated in the dendrite core of alloy USTB-F3 only after 50 h ( Fig. 3 c). The EDS analyses indicated that TCP phases were enriched in Re, W and Cr. 3.3 Elemental partitioning ratio and lattice misfit The compositions and the associated elemental partitioning ratios of γ/γ′ phases in the dendrite core and interdendritic region of three investigated alloys after thermal exposure at 1100 °C for 800 h are listed in Table 3 . The partitioning behaviors of Al, W and Co did not change obviously under the influence of Cr addition. Although Ta and Cr partitioned preferentially to γ′ phase and γ phase with increasing the Cr content, respectively, the partitioning behavior of Re changed most significantly due to Cr addition in the dendrite core of these investigated alloys. The Re concentration in the γ phase of the Cr-free alloy was 3.0 at.%. More Re partitioned to the γ phase with increasing levels of Cr addition, and the Re content in the γ phase of alloy USTB-F3 was 7.1 at.%. Meanwhile, the Re content of γ′ precipitates changed slightly in despite of different levels of Cr addition. Accordingly, the Re partitioning ratio increased with increasing levels of Cr addition, significantly. It was 3.75 in the Cr-free alloy and it reached to 14.20 in the alloy containing high level of Cr (5.7 wt.%). In the interdendritic region, the partitioning behaviors of Al, W, Co and Cr changed very slightly from alloy USTB-F1 to alloy USTB-F3. However, Re partitioned preferentially to the γ phase with increasing levels of Cr addition. The partitioning behavior of Ta changed with the Cr addition, but with less pronounced characteristics than Re. The lattice misfits ( δ ) were estimated by the formula presented by Kablov and Petrushin [17] , according to the compositions of γ and γ′ phases, they are also listed in Table 3 . In the dendrite core, the lattice misfit of the baseline alloy with nearly spherical γ′ precipitates was 0.04%. And it became more negative with increasing the Cr content. The lattice misfit of alloy USTB-F3 was −0.49%. In the interdendritic region, the lattice misfit of alloy USTB-F1 was 0.23% and it also became more negative with increasing levels of Cr addition. The lattice misfit of alloy USTB-F2 was close to zero and it was −0.18% in alloy USTB-F3. 3.4 Stress–rupture properties The stress–rupture properties of three investigated alloys at 1100 °C/140 MPa are listed in Table 4 . The average stress–rupture life of alloys USTB-F1, USTB-F2 and USTB-F3 was 29.5 h, 73.0 h and 137.6 h, respectively. The current results suggest that the average stress–rupture life increased with increasing levels of Cr addition. It was about 3.7 times in alloy USTB-F3 (high level of Cr addition) longer than that of alloy USTB-F1 (Cr-free). In addition, all the investigated alloys possessed good high-temperature ductility. 3.5 Post-creep microstructures Time to 1% creep strain under 1100 °C/130 MPa for three investigated single crystal superalloys were 9 h, 39 h and 80 h, respectively. Subsequently, post-creep microstructures were examined on (0 1 0), (1 0 0) and (0 0 1) sections. Fig. 4 shows the three-dimensional post-creep microstructures of three investigated alloys after 1% creep strain. In the dendrite core ( Fig. 4 a1–c1), the Cr-free alloy did not raft on the surfaces of (1 0 0) and (0 1 0) after the creep deformation ( Fig. 4 a1). However, alloys USTB-F2 and USTB-F3 rafted perpendicularly to the stress axis in the longitudinal-section ( Fig. 4 b1 and c1). In the interdendritic region ( Fig. 4 a2–c2), alloy USTB-F1 formed a nearly rafted structure, and it was parallel to the stress axis ( Fig. 4 a2). Alloy USTB-F2 did not raft ( Fig. 4 b2). However, alloy USTB-F3 rafted perpendicularly to the stress axis ( Fig. 4 c2). The linear termination densities of the investigated alloys are listed in Table 5 . In the dendrite core, the Cr-free alloy USTB-F1 did not raft and the linear termination density could not be determined. The linear termination density in alloy USTB-F2 containing 2.8 wt.% Cr was 3.13 × 10 2 mm −1 and it was 1.51 × 10 2 mm −1 in alloy USTB-F3 (5.7 wt.% Cr). These experimental results suggest that the linear termination density decreased and the degree of microstructural rafting behavior increased with increasing levels of Cr addition. In the interdendritic region, the linear termination densities of alloys USTB-F1 and USTB-F3 were 6.18 × 10 2 mm −1 and 1.83 × 10 2 mm −1 , respectively. Alloy USTB-F2 did not raft and the linear termination density could not be determined. It should be noted that there were no TCP precipitates in the creep specimens of alloys USTB-F1 and USTB-F2 ( Fig. 4 a1, a2, b1 and b2), while TCP phases were observed in the dendrite core of creep specimens of alloy USTB-F3 ( Fig. 4 c1) with the absence of TCP precipitates in the interdendritic region ( Fig. 4 c2). 4 Discussion The γ/γ′ lattice misfit ( δ ) is an important characteristic of Ni-based single crystal superalloys, and it influences the morphology of γ’ precipitates [12,13,18] and mechanical properties [19,20] as well as microstructural stability [18] . It is well accepted that the elemental partitioning ratio (γ–γ′ composition), the lattice misfit and the precipitate morphology are interrelated. The influence of the sign and magnitude of lattice misfit on the precipitate morphology has been investigated in details in Ni–Al–Mo ternary alloys [12] . Spherical precipitates were observed in alloys with a near zero misfit, whereas when the lattice misfit deviated from zero in either the positive or negative magnitude, γ′ precipitates became cuboidal. Recently, the simulated results using phase-field simulation showed good agreement with the above mentioned experimental results [13] . 4.1 Elemental partitioning ratio, lattice misfit and γ′ morphology Chromium addition was found to affect the elemental partitioning ratios (especially Re), significantly [8–11] . Re partitioned much more strongly to the γ phase in a Ni–Al–Cr–Re quaternary alloy, compared to a Ni–Al–Re ternary alloy [8] . Huang and Chang predicted by thermodynamic modeling that Re partitioning ratio in the Ni–Al–Cr–Re quaternary system is close to 1 and it increased with increasing the Cr content [9] . Furthermore, Dreshfield and Thomas reported that Re partitioning ratio increased from 10.5 to 19.0 with increasing the Cr content from 1.6 wt.% to 4.1 wt.% in multicomponent Ni-based single crystal superalloys containing 4.5 wt.% Ru and 2.1 wt.% Mo [10] . Recently, Carroll et al. found that Re partitioning ratio increased from 5–8 to 12–17 when Cr addition increased from 3.4 wt.% to 6.7 wt.% in UM-F series multicomponent Ni-based single crystal superalloys containing 5.7 wt.% Ru and 3.0 wt.% W [11] . The results that Re partitioning ratio increased with increasing levels of Cr addition are of particular importance in terms of the lattice misfit. According to Eq. (3) , the strong Re partitioning to the γ matrix caused by Cr addition results in more negative lattice misfit since Re has a large atomic radius. In the current study, the morphology of γ′ precipitates changed from nearly spherical in the Cr-free alloy ( Fig. 2 a1) to be more cuboidal ( Fig. 2 b1 and c1) with increasing levels of Cr addition in the dendrite core of the investigated alloys. Meanwhile, Cr addition resulted in more Re partition to the γ matrix ( Table 3 ). The change of the above two factors (γ′ morphology and Re partitioning ratio) shows good agreement with the change of the lattice misfit ( Table 3 ). Similar features have been reported in UM-F series alloys [21] . Despite the Ta partitioning ratio changed due to the Cr addition, it was neglectable compared to the influence of Re in the current study. In addition, the Cr partitioning ratio varied among the investigated alloys ( Table 3 ). However, Cr is not expected to have strong influence on lattice parameters of either γ phase or γ′ phase, and therefore the lattice misfit and γ′ morphology since Cr has an atomic radius very close to that of Ni. In the interdendritic region, the variation in partitioning behaviors for Re and Ta ( Table 3 ) and γ′ morphology ( Fig. 2 a2–c2) also show good agreement with the change of lattice misfit. In addition, the γ′ morphology in the interdendritic region of alloy USTB-F3 was similar to that of dendrite core ( Fig. 2 c1 and c2). However, the morphology of γ′ precipitates in the dendrite core and interdendritic region of alloys USTB-F1 and USTB-F2 was different. The variation in γ′ morphology suggests that the lattice misfit varies between the dendrite core and the interdendritic region due to the residual segregation of refractory alloying elements during solidification, which cannot be eliminated by solution heat treatment. 4.2 Stress–rupture properties Traditionally, Cr is added to improve hot corrosion and oxidation resistance of Ni-based single crystal superalloys [5] . Although Cr is an effective solid solution strengthener, there is no report regarding Cr improving the stress–rupture property significantly, to the current knowledge of the authors. High temperature creep and stress–rupture property of Ni-based single crystal superalloys is affected by several factors, including volume fraction of γ′ precipitates, γ/γ′ lattice misfit, rafting behavior under stress, γ channel width, degree of solid solution strengthening and microstructural stability. The present results indicate that all these factors were influenced by Cr addition. The relationships among these associated changes and the stress–rupture property of the investigated alloys are discussed as follows. 4.2.1 Effect of γ′ volume fraction As precipitation strengthening alloys, γ′ precipitates coherently embedded in the γ matrix are the primary strengthening phase in Ni-based superalloys, and their creep properties are mainly dependent on the volume fraction of γ′ precipitates. Murakumo and co-authors reported that the creep life was longest in alloy with about 65–70% volume fraction of γ′ precipitates in both 900 °C/392 MPa and 1100 °C/137 MPa [22] . In this study, the γ′ volume fraction in alloys USTB-F1, USTB-F2 and USTB-F3 was in between 45.3–55.3%, 55.1–63.5% and 57.4–62.1%, respectively, since the volume fraction of dendritic regions was distinctively larger than that of interdendritic regions. Therefore the γ′ volume fraction of alloy USTB-F2 containing intermediate level of Cr was about 10% higher than that of the Cr-free alloy USTB-F1 ( Table 2 ). This factor mainly contributed to the longer stress–rupture life of alloy USTB-F2 than that of alloy USTB-F1. However, alloys USTB-F2 and USTB-F3 have the same level of the γ′ volume fraction. Thus, additional factors must be responsible for the longer stress–rupture life of alloy USTB-F3. 4.2.2 Effect of lattice misfit, morphology and rafting behavior The Cr-free alloy did not raft with nearly spherical γ′ precipitates and close to zero lattice misfit in the dendrite core ( Fig. 4 a1). However, alloys USTB-F2 and USTB-F3 with intermediate γ′ precipitates as well as negative lattice misfit rafted perpendicularly to the applied tensile stress, as shown in Fig. 4 b1 and c1, respectively. It is worthy of noting that the linear termination density of alloy USTB-F2 was larger than that of alloy USTB-F3 ( Table 5 ). Then, the rafting structure became more completely with the increase of Cr content and the magnitude of lattice misfit. In the interdendritic region, alloy USTB-F1 with nearly cuboidal γ′ precipitates ( Fig. 2 a2) and positive lattice misfit ( Table 3 ) rafted parallelly to the stress axis ( Fig. 4 a2). Alloy USTB-F2 with nearly spherical γ′ precipitates ( Fig. 2 b2) and close to zero lattice misfit did not raft ( Fig. 4 b2). However, alloy USTB-F3 with intermediate γ′ precipitates ( Fig. 2 c2) and negative lattice misfit rafted perpendicularly to the stress axis ( Fig. 4 c2). Therefore, the rafting behaviors in both dendrite core and interdendritic region are consistent with the γ′ morphology and lattice misfit in the current study. Harada and Murakami reported that the temperature capability of single crystal superalloys was improved by changing the lattice misfit towards more negative [23] . This is because the creep deformation is controlled by dislocation activity in the γ channels, which is superimposed on the misfit stress. The net stress in the γ channels of alloys with negative lattice misfit is lower than that of alloys with positive lattice misfit under the loading along the <0 0 1> direction [1] . In addition, precipitate shearing and the creation of an anti-phase boundary (APB) in γ′ ordered precipitates are not necessary for continued deformation if the matrix channels parallel to the stress axis have not been eliminated [24] . Alloy USTB-F1 has slightly positive and positive lattice misfit in the dendrite core and interdendritic region, respectively, and there were a large amount of γ channels parallel to the stress axis after 1% creep strain ( Fig. 4 a1 and a2). Therefore, the stress–rupture life of alloy USTB-F1 was poor. The lattice misfit became more negative with increasing levels of Cr addition in both dendrite core and interdendritic region ( Table 3 ), and resulted in more perfect rafting structure perpendicular to the applied tensile stress during the primary creep ( Fig. 4 ), and accordingly, the average stress–rupture life became longer. Similar features have been founded in UM-F series alloys [15,21,24] . 4.2.3 Effect of γ channel width The total resolved shear resistance τ tot on the active slip planes is defined as: (4) τ tot = τ OR + τ ss + τ dis where τ OR is the Orowan resistance, τ ss is the solid solution resistance and τ dis is the dislocation resistance. For the case where precipitates are stable in shape and resist shearing, the Orowan resistance to bowing of dislocations through the narrow matrix channels must be overcome. It can be calculated as follows [19] : (5) τ OR = 2 3 μ b l where μ is the shear modulus, b is the Burgers vector and l is the width of γ channel. In this study, the γ channel in the dendrite core of alloy USTB-F3 ( Fig. 2 c1) was narrower than that of alloy USTB-F2 ( Fig. 2 b1) and it was in the same level in the interdendritic region ( Fig. 2 b2 and c2). Assuming μ = 48.2 GPa and b = 0.254 nm [2,19,24] , the uniaxial applied stress required for bowing dislocations through the matrix channel is 318 MPa and 583 MPa in the dendrite core of alloys USTB-F2 and USTB-F3, respectively, taking into account of the Schmid factor of [0 0 1] direction (0.408). 4.2.4 Effect of solid solution strengthening Another factor to influence the stress–rupture life is the solid solution strengthening, it inhibits dislocation gliding through the matrix. Re is known to be the most effective solid solution strengthening element in Ni-based single crystal superalloys due to its tendency to form clusters [25] and its low diffusivity [26] . It is also well accepted that Cr is an effective solid solution strengthening element [5] . In the current study, Re partitioned more severely to the γ phase and the corresponding Re concentration of the γ matrix in the dendrite core and interdendritic region increased for 4.1 at.% (10.6 wt.%) and 2.9 at.% (7.9 wt.%), respectively, from the Cr-free alloy to the alloy containing 5.7 wt.% Cr ( Table 3 ). In addition, more Cr partitioned to the γ matrix with increasing levels of Cr addition. For these reasons, a stronger contribution of Re and Cr to solid solution strengthening would be expected in alloy USTB-F3 than that of alloys USTB-F1 and USTB-F2. In summary, the combination of more completely rafting structure, higher Orowan resistance and solid solution strengthening by Re and Cr contributed to the longer stress–rupture life of alloy USTB-F3 than that of alloy USTB-F2. 4.2.5 Effect of microstructural stability It is known that Cr is an effective former of TCP phases, and high levels of Cr addition promoted the formation of TCP phases, significantly [5] . In the current study, no TCP precipitates existed after thermal exposure at 1100 °C for 2000 h and creep testing to 1% strain at 1100 °C/130 MPa in Cr-free alloy USTB-F1 and alloy USTB-F2 containing intermediate level of Cr addition, while TCP phases were observed in the dendrite core of high Cr-containing alloy USTB-F3 ( Fig. 3c and 4c1 ). These results confirmed that high level of Cr addition (5.7 wt.%) promoted the precipitation of TCP phases under thermal exposure [5] and creep deformation significantly at 1100 °C, since it influenced much more Re and Cr partitioning to the γ phase and increased the extent of Re and Cr supersaturation in the γ phase [27] . The stress–rupture life of alloy USTB-F3 was the longest among the investigated alloys despite of the precipitation of TCP phases. 5 Conclusions (1) The morphology of γ′ precipitates was nearly spherical in the dendrite core of Cr-free alloy after aging at 1100 °C for 8 h. With increasing the Cr content, Re partitioned preferentially to the γ phase, which in turn resulted in more negative lattice misfit and more cuboidal γ′ precipitates. In the mean time, the γ′ volume fraction increased while the precipitate size and the width of γ channel decreased with increasing levels of Cr addition. In the interdendritic region, the changes of Re partitioning ratio and lattice misfit due to Cr addition showed the similar trend to the dendrite core. The changes of elemental partitioning behavior and lattice misfit showed good agreement with the change of γ′ morphology. (2) The average stress–rupture life of the Cr-free baseline alloy was only 29.5 h at 1100 °C/140 MPa. Cr addition increased the stress–rupture life significantly since it increased the γ′ volume fraction, decreased the γ channel width and resulted in more negative lattice misfit in both dendrite core and interdendritic region with a well rafting structure perpendicular to the applied tensile stress. And the average stress–rupture life of high Cr-containing (5.7 wt.%) alloy was 3.7 times longer than that of Cr-free baseline alloy. (3) High level of Cr addition (5.7 wt.%) and the corresponding high Re and Cr supersaturation in the γ phase promoted the formation of TCP phases under thermal exposure and creep deformation significantly at 1100 °C. Acknowledgements The authors are grateful to S. Tin and Y.R. Zheng for useful discussions. The financial support provided by National Natural Science Foundation of China (Grant No.: 50671015 ), New Century Excellent Talents in University, Chinese Ministry of Education (Grant No.: NCET-06-0079 ), National Basic Research Program of China (No.: 2010CB631201 ) and High Technology Research and Development Program of China (No.: 2007AA03A225 ) is also gratefully acknowledged. References [1] R.C. Reed The Superalloys: Fundamentals and Applications 2006 Cambridge University Press Cambridge, UK [2] T.M. Pollock S. Tin J. Propul. Power. 22 2006 361 374 [3] R. Darolia D.F. Lahrman R.D. Field S. Reichman D.N. Duhl G. Maurer S. Antolovich C. Lund Superalloys, 1988 1988 TMS Champion, PA 255 264 [4] C.M.F. Rae R.C. Reed Acta Mater. 49 2001 4113 4125 [5] C.T. Sims N.S. Stoloff W.C. Hagel Superalloys II 1987 Wiley-Interscience New York [6] L. Zheng S.S. Li C.B. Xiao D.Z. Tang C.Q. Gu Key Eng. Mater. 353–358 2007 507 510 [7] Y.S. Yoo C.Y. Jo C.N. Jones Mater. Sci. Eng. A336 2002 22 29 [8] J. Smith, PhD dissertation, University of Illinois at Urbana-Champaign, 1987. [9] W. Huang Y.A. Chang Mater. Sci. Eng. A259 1999 110 119 [10] R.L. Dreshfield K.J. Thomas NASA report TM-2005-213288, Cleveland, OH 2005 [11] L.J. Carroll Q. Feng J.F. Mansfield T.M. Pollock Mater. Sci. Eng. A457 2007 292 299 [12] M. Fährmann P. Fratzl O. Paris E. Fährmann W.C. Johnson Acta Metall. Mater. 43 1995 1007 1022 [13] T. Wang G. Sheng Z.K. Liu L.Q. Chen Acta Mater. 56 2008 5544 5551 [14] M. Fährmann W. Hermann E. Fährmann A. Boegli T.M. Pollock H.G. Sockel Mater. Sci. Eng. A260 1999 212 221 [15] L.J. Carroll Q. Feng T.M. Pollock Metall. Mater. Trans. 39A 2008 1290 1307 [16] M. Nathal Metall. Mater. Trans. 18A 1987 1961 1970 [17] E.N. Kablov N.V. Petrushin R.C. Reed K.A. Green P. Caron T.P. Gabb M.G. Fahrmann E.S. Huron S.A. Woodard Superalloys 2008 2008 TMS Champion, PA 901 908 [18] J.Y. Chen B. Zhao Q. Feng L.M. Cao Z.Q. Sun Acta Metall. Sin. 46 2010 897 906 (in Chinese) [19] T.M. Pollock A.S. Argon Acta Metall. Mater. 40 1992 1 30 [20] J.X. Zhang T. Murakumo H. Harada Y. Koizumi Scripta Mater. 48 2003 287 293 [21] L.J. Rowland Q. Feng T.M. Pollock K.A. Green T.M. Pollock H. Harada T.E. Howson R.C. Reed J.J. Schirra S. Walston Superalloys 2004 2004 TMS Champion, PA 697 706 [22] T. Murakumo T. Kobayashi Y. Koizumi H. Harada Acta Mater. 52 2004 3737 3744 [23] H. Harada H. Murakami T. Saito Computational Materials Design 1999 Springer-Verlag Heidelberg, Berlin 53 55 [24] L.J. Carroll Q. Feng J.F. Mansfield T.M. Pollock Metall. Mater. Trans. 37A 2006 2927 2938 [25] D. Blavette P. Caron T. Khan Scripta Metall. 20 1986 1395 1400 [26] A. Giamei D. Anton Metall. Mater. Trans. 16A 1985 1997 2005 [27] J.Y. Chen Q. Feng Z.Q. Sun Scripta Mater. 63 2010 795 798