Effect of processing parameters on hot compressive deformation behavior of a new Ni–Cr–Co based P/M superalloy
K. Wu,G.Q. Liu,B.F. Hu,C.Y. Wang,Y.W. Zhang,Y. Tao,J.T. Liu
DOI: https://doi.org/10.1016/j.msea.2011.02.051
2011-01-01
Abstract:The hot compressive deformation behavior of a new Ni–Cr–Co based P/M superalloy was studied in the temperature range of 950–1150 °C and strain rate range of 0.0003–1 s −1 using hot compression tests. The effect of processing parameters on flow stress, activation energy, power dissipation η maps and instability maps were investigated. The results show that the flow stress decreases with increasing temperature and decreasing strain rate. The apparent activation energy of deformation decreases while the values of η in the peak zones increase with increasing strain, due to the occurrence of dynamic recrystallization. When the strain is 0.5, a domain with its peak η of 40% occurs at 1050 °C and 0.0003 s −1 , which corresponds to dynamic recrystallization with grain size of ASTM 10–11 and can be as the optimum processing parameters for good workability. Keywords P/M superalloy Flow stress Dynamic material model Dynamic recrystallization Processing parameters 1 Introduction The γ′ precipitate strengthened Ni–Cr–Co based superalloy combines excellent oxidation resistance with mechanical properties at high temperatures. The excellent working and mechanical properties make it a potential material for gas turbine applications as well as high temperature attaching parts. The powder metallurgy (P/M) route was found to be attractive for the manufacture of turbine components in Ni–Cr–Co based superalloys [1,2] . Based on the published chemical compositions of the typical third generation P/M superalloys [3–5] , a new Ni–Cr–Co based P/M superalloy was designed with the aid of the thermodynamic calculation results [6] and d-electron theory [7,8] . Compared with the first and second generations P/M superalloys, the new alloy was designed for thermal stability in long time service condition and controlled the precipitation of topologically close packed (TCP) phases. It produced the mass ratios of Al/Ti and Nb/Ta of be unity, in order to achieve the desired combination of properties [9] . It also added a certain amount of Hf, to enhance the comprehensive performance [10] . While hot isostatic pressing (HIP) and other techniques were developed, the need for a subsequent isothermal forging process was well recognized because it helped in mitigating the undesirable effects of prior particle boundaries (PPB) [11,12] . But this alloy is rather difficult to deform because of its poor workability. For this new alloy, controlling the forging process to attain a uniform microstructure is very important. Therefore, the effect of processing parameters on hot compressive deformation behavior is essential for the optimization of isothermal forging process. The trial and error methods were generally employed in selecting the hot deformation processing parameters while the recently developed processing map was found to be a very useful tool for optimizing workability and controlling microstructure in superalloys. Somani et al. [13,14] investigated the effect of processing conditions on microstructure evolution of P/M Nimonic AP-1 and IN-100 superalloys using processing map. Srinivasan et al. [15] studied the hot working characteristic of Nimonic 75, 80A and 90 using processing map in the ranges of temperature 850–1200 °C and strain rate 0.001–100 s −1 . Babu et al. [16] did the work on development and validation of processing maps for nickel based P/M superalloys IN 100, MA754 and AP-1. Cai et al. [17] developed the processing maps for GH708 in deformation temperature range 1050–1180 °C and strain rate range 0.01–10 s −1 , and pointed out that the strain had a slight effect on the processing maps. Aghaie-Khafri et al. [18] investigated the hot workability of Hastelloy X by calculating flow localization parameter and construction of workability map for shear band. Then it was compared with the developed processing map. Liu et al. [19] established the processing map of Haynes230 on the basis of dynamic material model. The optimum conditions for hot working were at deformation temperature of 1150 °C and strain rate of 0.001 s −1 . Processing map was developed on the basis of the dynamic material model (DMM), which aimed to correlate the constitutive behavior with microstructure evolution, flow instability and hot workability [13–19] . Briefly, the work-piece undergoing hot deformation was considered to be a dissipater of power supplied by a particular source. The true stress–true strain curve and the flow behavior of materials were strongly dependent on the strain rate and temperature. The general form of the flow behavior at constant strain and temperature was expressed as (1) σ = K ε ˙ m | ε , T where K and m were considered as constants. From Eq. (1) , m could be obtained as (2) m = ∂ ( ln σ ) ∂ ( ln ε ˙ ) The total power dissipated P could be calculated as (3) P = σ ε ˙ = G + J = ∫ 0 ε ˙ σ d ε ˙ + ∫ 0 σ ε ˙ d σ The first integral was defined as G content and represented the main power input dissipated in the form of a temperature rise. The second integral was defined as J co-content and was related to the power dissipated by metallurgical processes. The efficiency of power dissipation occurring through microstructural changes during deformation was derived by comparing the non-linear power dissipation occurring instantaneously in the work-piece with that of a linear dissipater for which the m value was unity, and was defined as (4) η = Δ J / Δ P ( Δ J / Δ P ) line = m / m + 1 1 / 2 = 2 m m + 1 The variation of η with deformation temperature and strain rate constituted the power dissipation map. This map represented the pattern in which the power was dissipated by metallurgical processes. A continuum criterion for the occurrence of flow instability was obtained by utilizing the principle of maximum rate of entropy production and given by (5) ξ ( ε ˙ ) = ∂ ln ( m / m + 1 ) ∂ ( ln ε ˙ ) + m > 0 where ξ ( ε ˙ ) was the instability parameter. Flow instabilities predicted to occur when ξ ( ε ˙ ) became negative. In this study, the hot compressive deformation behavior of a new Ni–Cr–Co based P/M superalloy was systematically investigated by hot compression tests, which were conducted in deformation temperature range of 950–1150 °C and strain rate range of 0.0003–1 s −1 . The effect of processing parameters on flow stress, activation energy, power dissipation η maps and instability maps was investigated. All of these provide a theoretical reference for the actual hot working process. 2 Experiments The hot isostatically pressed (HIPed) Ni–Cr–Co based P/M superalloy was selected as the experimental material with the following chemical composition: Cr 12.92, Co 20.83, Mo 2.64, W 3.85, Al 3.57, Ti 3.53, Nb 1.51, Ta 1.65, C 0.048, B 0.027, Zr 0.043, Hf 0.2, Ni-base. The details of the preparation procedures were as follows: the vacuum induction melting (VIM) and electroslag remelting (ESR) methods were used for alloy smelting; the plasma rotating electrode process (PREP) was used for powder preparation, and the vibration screening and electrostatic separation methods were combined to remove inclusions and produce powder with particle size of 50–150 μm. The powders were encapsulated in AISI 304 stainless steel capsules and sealed after vacuum degassing. The capsules were then HIPed in a single stage at 1180 °C for 4 h at a pressure of 120 MPa. The HIPed rods were then skinned to remove the stainless steel encapsulation. Hot compression tests were conducted using a Gleeble-1500 simulator at the temperatures of 950 °C, 1000 °C, 1050 °C, 1100 °C and 1150 °C and in the strain rate range of 0.0003–1 s −1 with 8 mm diameter, 15 mm high cylindrical specimens. All specimens were heated to the deformation temperatures at a heating rate of 10 °C/s and held for 5 min in order to gain the uniform microstructure before the hot compression test. Then they were deformed to a true strain of 0.7 and water cooled to freeze the microstructure. The selected deformation processing parameters for the new superalloy were similar with the forging parameters of ME3 [4] and LSHR [5] , in order to make the comparisons and assess the ability in achieving a better alloy for the intended application. The load-stroke data obtained in compression were processed to obtain true stress–true strain curves using standard equations. The deformed specimens were sectioned parallel to the compression axis and the cut surfaces were prepared for microstructure observation. The microstructure of as-HIPed and deformed samples was observed using an Olympus optical microscope and with the chemical etchant of CuCl 2 (10 g) + HCl (50 mL) + H 2 O (50 mL), the etching time of 30–60 s. Grain sizes were determined according to ASTM E112 linear intercept procedures using circular grid overlays. Besides, the microstructure of as-HIPed samples was observed using a XL30TMP scanning electron microscope (SEM), and the chemical composition analysis of carbides was performed using a Phoenix energy dispersive spectroscopy (EDS). For SEM and EDS studies, the preparation method of sample started with electropolishing with 20% sulfuric-methanol and followed by electrolytic etching in a CrO 3 –H 3 PO 4 –sulfuric acid solution, the etching time of 15 s. 3 Results and discussion 3.1 Initial Microstructure The typical microstructure of as-HIPed Ni–Cr–Co based P/M superalloy is shown in Fig. 1 . It consists primarily of recrystallized grains with a uniform dispersion of a few apparently undeformed powder particles that have retained as-cast dendritic structure. The HIP treatment at 1180 °C reduces carbide precipitation at PPBs but gives a coarse grained structure. The grain size is about ASTM 7–8. SEM examination of the specimens reveals cuboidal γ′ precipitates ( Fig. 2 (a) ) and carbide particles, mainly at the grain boundaries ( Fig. 2 (b)). The microstructure of Fig. 2 (b) reveals a prior particle boundary (PPB) network by the large primary γ′ precipitates and white carbides. The volume fraction of γ′ phase is estimated to be 55%. The EDS recorded at the carbides at the grain boundaries show that the carbides are Cr, Mo and W based, while the carbides in the matrix are Ti, Ta and Nb based. 3.2 Effect of processing parameters on flow stress The true stress–true strain curves of the Ni–Cr–Co based P/M superalloy can be easily transformed from the load-displacement data recorded by Gleeble-1500 simulator during hot compression test. Fig. 3 shows the typical true stress–true strain curves obtained from hot compression tests of the alloy. These flow stress curves exhibit a flow softening feature, which is more significant at lower temperatures and higher strain rates. The basic changing trends of flow stress are as follows: the stress increases rapidly with increasing strain at first, then decreases when peak stress is reached and almost be stable under certain variables of strain finally. It means that a typical flow stress curve is composed of three stages: stage I (work hardening stage), stage II (softening stage) and stage III (steady stage). This is a combination effect of work hardening, which is primarily caused by second order pyramidal slip system dislocation motion, and thermally activated softening. In the stage I, the initial rapid rise in stress is associated with an increase in dislocation density and the formation of poorly developed subgrain boundaries, as a result of work hardening and dynamic recovery. As the Ni–Cr–Co based P/M superalloy has low stacking fault energy, the dynamic recovery proceeds slowly. The high dislocation density stimulates the occurrence of dynamic recrystallization once a critical strain (usually less than 0.1) is exceeded, then the flow stress decreases which means the coming of stage II. In this stage, the dislocations are annihilated in large numbers through the migration of a high angle boundary and the stress drops steeply. Finally, stage III (steady stage): the stress becomes steady when a new balance between softening and hardening is obtained [20,21] . From the experimental results (shown in Fig. 3 ), it can be found that the flow stress represents as the characteristic of dynamic crystallization with increasing strain at the deformation temperature lower than 1100 °C and strain rates higher than 0.0003 s −1 . The variation of flow stress with deformation temperatures and strain rates can be seen more clearly in Fig. 4 , where the flow stress at a true strain of 0.4 is plotted as a function of the logarithm of the strain rate and the deformation temperature for all tested conditions in Fig. 4 (a) and (b), respectively. It can be found that the effect of the deformation temperature and strain rate on the flow stress is significant. For the Ni–Cr–Co based P/M superalloy, the lower the deformation temperature, the higher the flow stress; while the higher strain rate, the higher the flow stress. This is because low strain rates and high temperatures provide longer time for energy accumulation and higher mobility at boundaries for the nucleation and growth of dynamically recrystallized grains and dislocation annihilation, and thus reduce the flow stress level. So the effect of deformation temperature and strain rate on flow stress can be explained by the terms of dynamic recrystallization and dislocation mechanism. 3.3 Effect of strain on activation energy From the analysis of flow curves of the Ni–Cr–Co based P/M superalloy ( Fig. 3 ), it is found that the main factors affecting flow stress are deformation temperature, strain rate and strain, and the effect of the former two are more significant ( Fig. 4 ). As both the hot deformation and creep are belong to thermally activated process, hot deformation can be regarded as the extension of creep under the condition with larger strain rate and higher stress. Therefore, the flow stress mainly depends on deformation temperature and strain rate, and there exists dynamic balance between strain hardening and dynamic softening. In this study, the research [22,23] on the constitutive equation between strain rates, temperatures with flow stress are summarized as below. The hyperbolic sine model: (6) ε ˙ = A [ sinh ( α σ ) ] n exp − Q R T The power function model: (7) ε ˙ = A 1 σ n 1 ( low stress , α σ < 0.8 ) The exponential function model (8) ε ˙ = A 2 exp ( β σ ) ( high stress , α σ > 1.2 ) In which, A , A 1 , A 2 , n , n 1 , α and β are the material constants which are independent of temperature and α = β / n 1 , R is the universal gas constant (8.31 J mol −1 K −1 ), T is the absolute temperature (K), Q is the activation energy of hot deformation (kJ/mol) which reflects the difficult degree of hot deformation. Many studies [24–26] show that Eq. (6) can be used to describe the conventional hot deformation process of materials, and also can be simplified to Eq. (7) and Eq. (8) under different stress conditions. As we know, dynamic changes of microstructure and flow stress with the variation of strain are obvious during hot deformation. However, it is commonly accepted that the effect of strain on the flow stress will not be considered in Eq. (6) . Therefore, this study assumes that A , n , α and Q are the material constants which depend on strain. Taking the logarithm of both sides of Eqs. (7) and (8) gives (9) ln σ = ln ε • n 1 − ln A 1 n 1 (10) σ = ln ε • β − ln A 2 β The corresponding flow stress of strain ranging from 0.05 to 0.7 is selected from the flow curves of Fig. 3 . The values of β and n 1 can be obtained by Eqs. (9) and (10) from the slope of the lines in the σ − ln ε • and ln σ − ln ε • plots, respectively. Then, further work on the relations of α with ɛ can be done. Fig. 5 shows both the experimental data and regression results at the strain of 0.3. Obviously, σ with ln ε • and ln σ with ln ε • have good linear relationship, and the correlation coefficients are 0.981–0.993. As the value of Q is independent of temperature, taking the logarithm of both sides of Eq. (6) gives (11) ln ε • = ln A − Q R T + n ln [ sinh ( α σ ) ] For the given strain rate conditions, differentiating Eq. (11) gives (12) Q = R ∂ ln ε • ∂ ln sinh ( α σ ) T ∂ ln sinh ( α σ ) ∂ ( 1 T ) ε • By substituting the values of flow stress and strain rate for all the tested temperatures into Eq. (11) , the relations of ln sinh ( α σ ) with ln ε • and ln sinh ( α σ ) with 1 T can be obtained, as shown in Fig. 6 . From this kind of figure, the values of n and A can be easily obtained, and Q can be calculated by Eq. (11) under different strain. The values of Q with the variation of strain are shown in Fig. 7 . It is found that their values decrease with increasing strain. The main reasons for the change of Q can be considered as: In the initial stage, only a small number of soft-oriented slip systems and grain take part in the deformation, so the value of Q is relatively high. While with the increase of deformation strain, grain rotation occurs and more slip systems are activated, leading to the decrease of Q . Herein, the relation of Q with strain is fitted and gives as Eqs. (13) . (13) Q = 811.3 − 86.25 ε − 1401 ε 2 + 2502 ε 3 − 1634 ε 4 The values of Q can be expressed as 4th order polynomial function, as shown in Fig. 7 . It is easy to find that the regression curves fit the experimental data very well and the values of mean square deviation are almost equal to zero. 3.4 Effect of strain on power dissipation maps The flow stress data obtained from the flow curves of Fig. 3 were friction-revised by Gubkin formula. Then they were corrected for adiabatic temperature rise using Aresenault method, in order to ensure the accuracy of the experimental data. The log of flow stress vs. log of strain rate data were fitted using a cubic spline function and the strain rate sensitivity m was calculated as a function of strain rate according to Eq. (2) . This was repeated at different temperatures. The calculated efficiency of power dissipation η was plotted as isoefficiency contour maps by the software of Surfer 8.0. As shown in Fig. 8 , the values of η increase with increasing deformation temperature and decreasing strain rate, except for the lower right corner of the contour maps at high temperatures and low strain rates. It may have a relation with grain growth, dissolution and coarsen of γ′ phases. In addition, within the deformation conditions of this study, it is also found that the peak zones in the upper right corner of η contour maps change from a semi-circle at the strain of 0.1 into a closed circle at the strain of 0.7. The peak zones move to the region of lower temperatures and higher strain rates. The values of η gradually increase with increasing strain, and good hot workability can be obtained. The Ni–Cr–Co based P/M superalloy is a γ–γ′ alloy with an estimated volume fraction of γ′ of 55%. The γ′ dissolution of the Ni–Cr–Co based P/M superalloy starts occurring from 1155 °C and completes at 1165 °C [6] . As the carbides were not significantly present in the microstructure, the alloy may be essentially considered as a two phase alloy within the experimental conditions of this study. The characteristics of a two phase alloy under hot working conditions were analyzed using processing maps [27] . A typical power dissipation map generated in the temperature range of 950–1150 °C and strain rate range of 0.0003–1 s −1 at a strain range from 0.1 to 0.7 is shown in Fig. 8 . In the Fig. 8 (c), the map Exhibits 4 different domains according to the characteristic values of η : domain I occuring at 1050–1150 °C and 0.1–1 s −1 (upper right), with the η value of 38–40%; domain II occuring at 1000–1060 °C and 0.0003–0.001 s −1 (middle bottom), with the η value of 36–40%; domain III occuring at 1080–1150 °C and 0.0003–0.01 s −1 (lower right), with the η value of less than 30%; domain IV occuring at 950–1050 °C and 0.1–1 s −1 (upper left), with the η value of less than 34%. The power dissipation maps are continuum maps and are interpreted in terms of the microstructural processes. For domain I, the sample at 1050 °C and 0.1 s −1 shows obvious dynamic recrystallization characteristics ( Fig. 9 (a) ), with the grain size of ASTM 8.5–9.0. For domain II, according to the atomistic processing map of Raj [28] , the observed domain may correspond either to superplasticity or DRX. The efficiency of power dissipation for superplasticity is generally very high (>60%). This is not observed in the map of Fig. 8 (c) since the peak efficiency is not higher than 40%. In view of this observation, the domain observed does not represent superplastic deformation. Hence, the domain in the processing map can be interpreted in terms of DRX. On the basis of the hot deformation microstructural observations ( Fig. 9 (b)), the recrystallization region can be identified in the power dissipation map at a strain of 0.5 ( Fig. 8 (c)). The average grain size of the hot deformed sample ( Fig. 9 (b)) is ASTM 10–11. The grain refinement effect of the Ni–Cr–Co based P/M superalloy is slightly better than ME3 [4] and LSHR [5] under the same deformation conditions. For domain III, as shown in Figs. 8 (c), there are low value of η under different strain. Further, they decrease with increasing strain. This is because the significant grain growth appears at high temperature and low strain rate condition. The grain size is ASTM 6.0–6.5 ( Fig. 9 (c)). The plasticity deteriorates and the hot workability becomes worse. In addition, intercrystalline cracking is also detected in the sample under 1150 °C and 3 × 10 −4 s −1 , as shown in positions A and B of Fig. 9 (d). For the Ni–Cr–Co based P/M superalloy, there is some residual dendrite which leads to the chemical segregation along the grain boundaries. It results in intergranular cracking. Meanwhile, wedge cracking occurs at grain boundary triple junctions, as shown in positions C and D of Fig. 9 (d). It is similar with the result of P/M IN-100 [14] . Fig. 10 shows the peak η of the Ni–Cr–Co based P/M superalloy changes with strain. When the strain is not higher than 0.5, the positions of peak η do not change and at deformation temperature of 1050 °C and strain rates of 0.0003 s −1 , as shown in Fig. 10 (a). Besides, it can be seen from Fig. 3 that the flow stress increases with increasing strain and reaches the peak value almost at the strain of 0.1. In this stage, dislocation motion is considered as the dominant deformation mechanism. But with the further increasing of strain, dynamic recrystallization occurs and leads to both flow stress and activation energy ( Fig. 7 ) decreasing fast. So the peak η of the Ni–Cr–Co based P/M superalloy occurs at the strain of 0.2, as shown in Fig. 10 (b). 3.5 Effect of strain on instability maps The instability maps for the Ni–Cr–Co based P/M superalloy are shown in Fig. 11 . These maps reveal the limiting temperature and strain rate conditions where fracture and instabilities occur during the deformation of materials. It is easy to find that the instability zone shrinks to the region of lower temperatures and higher strain rates with increasing strain. As hot compressive deformation is a thermally activated process, accompany with grain rotation, more slip systems are activated when the strain increases. It results in the decrease of activation energy ( Fig. 7 ), making it easier for the progress of hot deformation. Besides, it is easy to find that the instability zone indicated by Fig. 11 is also considered as domain IV in Fig. 8 (c). The 45° shear cracking ( Fig. 12 (a) ) is observed in the hot-deformed sample in the instability zone. This is caused by the maximum shear stress along the 45° direction from the compression axis during the hot compression test. The lower the deformation temperatures and the higher the strain rates, the greater the tendency for shear cracking it is. Also there are longitudinal cracking ( Fig. 12 (b)) and mixed-type cracking ( Fig. 12 (c)). The direction of longitudinal cracking is parallel to the compression axis. It mainly results from the tensile stress in the hoop direction caused by the compressive deformation process. When the superalloy deforms at lower temperatures and higher strain rates, the mixed-type cracking appears by the combining effect of the shear stress along the 45° direction from the compression axis and the hoop tensile stress. The phenomenon of large bulk caving can be seen, as shown in Fig. 12 (c). After the full observation of the characteristics of power dissipation η maps and instability maps under different strain, it is found that there are two peak domains locating at the same positions as domain I and II when the strain ranges from 0.1 to 0.5. The peak domain with lower strain rates will disappear within the experimental conditions, if the strain exceeds 0.5. The power dissipation η map and instability map for the Ni–Cr–Co based P/M superalloy at the strain rate of 3 × 10 −4 s −1 are plotted, as shown in Fig. 13 . The alloy has relatively high power dissipation efficiency ( Fig. 13 (a)) when the temperature is about 1050 °C. But the instability region appears when the strain is higher than 0.5 ( Fig. 13 (b)). In order to validate the predictability of hot processing maps, two samples at 1050 °C and 3 × 10 −4 s −1 with the strain of 0.5 and 0.6, respectively are subjected to hot compression test. The former has no cracks while the latter has obvious longitudinal cracks, as shown in Fig. 14 . This is in agreement with the hot processing maps under the corresponding strain. Therefore, it is recommended that the first-step forging strain of the Ni–Cr–Co based P/M superalloy should be not higher than 0.5. Generally, with the full consideration in the effect of strain rate, temperature and strain on flow stress, activation energy, power dissipation η maps and instability maps, the optimum condition for good workability is at 1050 °C and 0.0003 s −1 with the strain of 0.5. The microstructure and flow instability of the Ni–Cr–Co based P/M superalloy are consistent with hot processing maps based on DMM within the experimental conditions. The results can provide a theoretical reference for the actual hot working process optimization of the Ni–Cr–Co based P/M superalloy. 4 Conclusions The hot compressive deformation behavior of a new Ni–Cr–Co based P/M superalloy by means of the hot compression test have been investigated in the temperature range of 950–1150 °C and strain rate range of 0.0003–1 s −1 by Gleeble-1500 thermal simulator. The following conclusions have been drawn from the results of this investigation. 1. The true stress–true strain flow curves of the Ni–Cr–Co based P/M superalloy shows the typical characteristics of dynamic recrystallization and dynamic recovery. The flow stress decreases with increasing deformation temperature and decreasing strain rate. The flow stress represents as the characteristic of dynamic recrystallization with increasing strain at the deformation temperature lower than 1100 °C and strain rates higher than 0.0003 s −1 . 2. The values of Q decrease with increasing strain. In the initial stage, only a small number of soft-oriented slip systems and grain take part in the deformation, so the value of Q is relatively high. While with the increase of deformation strain, grain rotation occurs and more slip systems are activated, leading to the decrease of Q . The values of Q can be expressed as 4th order polynomial function. 3. The values of η increase with increasing deformation temperature and decreasing strain rate. Within the deformation conditions of this study, it is also found that the peak zones in the upper right corner of η contour maps change from a semi-circle at the strain of 0.1 into a closed circle at the strain of 0.7. The peak zones move to the region of lower temperatures and higher strain rates. The values of η also gradually increase with increasing strain and the peak η occurs at the strain of 0.2. 4. The power dissipation map at a strain of 0.5 Exhibits 4 different domains according to the characteristic values of η : domain I occuring at 1050–1150 °C and 0.1–1 s −1 (upper right), with the η value of 38–40%, the domain shows obvious dynamic recrystallization characteristic; domain II occuring at 1000–1060 °C and 0.0003–0.001 s −1 (middle bottom), with the η value of 36–40%, the domain can be interpreted in terms of DRX and grain refinement; domain III occuring at 1080–1150 °C and 0.0003–0.01 s −1 (lower right), with the η value of less than 30%, the domain reveals the phenomena of grain growth, intergranular and wedge cracking; domain IV occuring at 950–1050 °C and 0.1–1s −1 (upper left), with the η value of less than 34%. The domain belongs to the flow instability zone which shrinks with the increasing strain, and reveals the phenomena of shear cracking, longitudinal cracking and mixed-type cracking. Acknowledgements This work was financially supported by National Pre-research Funds for Weapons and Equipment and High Technology Research and Development Program of China (No. 2007AA03A223). References [1] W.C. Robert Nature 276 1978 209 210 [2] C.H. Symonds J.M. Eggar G.J. Lewis R.J. Siddall Powder. Metall. Int. 15 1983 30 35 [3] T.P. Gabb J. Gayda NASA-TM-211218 National Aeronautics and Space Administration 2001 Glenn Research Center Celeveland, OH, USA [4] T.P. Gabb K. O’Connor NASA-TM-210901 National Aeronautics and Space Administration 2001 Glenn Research Center Celeveland, OH, USA [5] T.P. Gabb J. Gayda J. Falsey NASA-TM-213649 National Aeronautics and Space Administration 2005 Glenn Research Center Celeveland, OH, USA [6] K. Wu G.Q. Liu B.F. Hu H. Wu Y.W. Zhang Y. Tao J.T. Liu J. Univ. Sci, Technol. Beijing 31 2009 722-710 (In Chinese) [7] N. Yukama M. Morinaga H. Ezaki Proceedings of the International Conference on High Temperature Materials for Power Applications 1986 Dordrecht Holland pp935 [8] N. Yukama M. Morinaga Y. Murata Proceedings of the Sixth International Symposium on Superalloy TMS, Pennsylvania 1988 pp. 225 [9] T.P. Gabb J. Gayda J. Telesman P.T. Kantzos W.A. Konkel NASA-TM-2120 86 National Aeronautics Space Administration 2003 Glenn Research Center Celeveland, OH, USA [10] J. Radavich T. Carneiro D. Furrer J. Lemsky A. Banik proceeding of the eleventh international symposium on advanced superalloys-production and application The effect of Hafnium, Niobium, and heat treatment on advanced powder metallurgy superalloys, edited by the Chinese Society for Metals (Shanghai) 2007 [11] T.L. Prakash Y.N. Chari E.S. Bhagiradha Rao R. Thamburaj Metall. Trans. A 14 1983 733 742 [12] R. Thamburaj W. Wallace Y.N. Chari T.L. Prakash Powder. Metall. 27 1984 169 180 [13] M.C. Somani N.C. Birla Y.V.R.K. Prasad V. Singh J. Mater. Process. Technol. 52 1995 225 237 [14] M.C. Somani K. Muraleedharan Y.V.R.K. Prasad V. Singh Mater. Sci. Eng. A 245 1998 88 99 [15] N. Srinivasan Y.V.R.K. Prasad J. Mater. Process. Technol. 51 1995 171 192 [16] N.S. Babu S.B. Tiwari B.N. Rao Powder. Metall. 49 2006 160 166 [17] D.Y. Cai L.Y. Xiong W.C. Liu G.D. Sun M. Yao Mater. Charact. 58 2007 941 946 [18] M. Aghaie-Khafri N. Golarzingh Mater. Sci. Eng. A 486 2008 641 647 [19] Y. Liu R. Hu J.S. Li H.C. Kou H.W. Li H. Chang H.Z. Fu Mater. Sci. Eng. A 209 2009 4020 4026 [20] T. Sakai J.J. Jonas Acta. Mater. 32 1984 189 209 [21] H.J. Mcqueen Mater. Sci. Eng. A 387–389 2004 203 208 [22] F. Garofalo Trans. AIME 227 1963 3351 3356 [23] C.M. Sellars W.M. Tegart Acta. Metall 14 1966 1136 1138 [24] Z.J. Gronostajski J. Mater. Process. Technol. 78 1998 84 89 [25] C.A.C. Imbert H.J. Mcqueen Mater. Sci. Eng. A 313 2001 88 103 [26] Y.C. Lin M.S. Chen J. Zhong J. Mater. Process. Technol. 205 2008 308 315 [27] J. Luo M.Q. Li H. Li W.X. Yu Mater. Sci. Eng. A 505 2009 88 95 [28] R. Raj Metall. Trans. A 12 1981 1089 1097