Experimental investigation of a Portevin-Le Chatelier band in Ni‒Co-based superalloys in relation to γʹ precipitates at 500 ℃
The macroscopically localized deformation behaviors of Ni-Co-based superalloys with different γ′ precipitate content were investigated at 500 °C and 1 × 10-4 s-1 via an in situ method namely, digital image correlation (DIC). The DIC results showed that the serrated flow of the stress-strain curves was accompanied by localized deformation of the specimens. The fracture morphology was characterized mainly by transgranular fracture with numerous dimples in the low γ′ content alloy, and intergranular fracture with large fracture section in the high γ′ content alloy. The Portevin-Le Chatelier (PLC) effect occurred in the investigated Ni-Co-based superalloys. Furthermore, the localized deformation of the high γ′ content alloy was more severe than that of the low γ′ content alloy, and the band width was slightly larger. Moreover, for the first-time ever, a special propagation feature, namely ±60° zigzag bands characterized by head-to-tail connections, was observed in the high γ′ content alloy.
Cite this article
Yanke Liu, Yulong Cai, Chenggang Tian, Guoliang Zhang, Guoming Han, Shihua Fu, Chuanyong Cui, Qingchuan Zhang.
Nickel-cobalt (Ni-Co)-based superalloys with excellent mechanical properties at elevated temperatures have been developed for use as turbine-disk components of aircraft engines and industrial gas turbine engine applications [, , , , , , ]. γ′ precipitates, an important type of dispersed particles in the γ matrix, play a critical role in improving the mechanical properties of Ni-Co-based superalloys. However, the turbine disks are large components that experience a broad spectrum of temperatures, and hence, exhibit an undesirable plastic instability during their service life [, , ].
The aforementioned phenomenon, generally referred to as the Portevin-Le Chatelier (PLC) effect, is manifested as a jerky or serrated flow in the stress-strain curves and localized deformation on the specimen surfaces [, , , , , ]. Many experimental observations have revealed that the PLC effect in Al-based alloys and Ni-based superalloys occurs over a certain range of strain rates, testing temperatures, and even grain sizes. Usually, the microscopic origin of the PLC effect is attributed to the dynamic interactions between the mobile dislocations and the diffusing solutes, i.e., the dynamic strain aging (DSA) [, , , ]. The foundational theory of DSA stipulates that the mobile dislocations are blocked and released repeatedly by the solutes, resulting in repeated serrations in the stress-strain curves. Based on the origin of the first serration, researchers classified the PLC effect into two categories: the normal and the inverse PLC or DSA effect [, , ].
γ′ (Ni3Al) precipitates are of coherent and ordered L12 crystal structure embedded in the disordered solid-solution face centered cubic γ matrix. Due to the special crystal structure of γ′ precipitates, which act as effective barriers to dislocation motion, these precipitates have significant influence on the micro-scale deformation mechanism of Ni-based superalloys. For example, the main deformation modes occurring in creep tests include γ′ precipitates bypassing dislocations via Orowan loops, cooperative climbing, and dislocation shearing of γ′ precipitates under certain temperatures and loading stresses [, , ]. Cui et al.  and Xu et al.  found that high densities of dislocations and stacking faults (SFs) are generated with the normal and inverse PLC effect at intermediate or high temperatures, respectively, in TMW-2 alloys. Fu et al.  investigated the size effect of γ′ particles on the deformation mechanism at intermediate temperatures. The results revealed that, as the precipitate size increases, the micro-deformation mode changes from anti-phase boundary shearing to SF shearing and then to Orowan bypassing.
γ′ precipitates play a significant role in affecting the microstructures and further influencing the mechanical properties. Nevertheless, there are few systematic studies in relation to the γ′ content. The influence of γ′ content on the strengthening features and serration characteristics has been investigated in our previous study . However, variations in the macro-scale localized deformation with the γ′ content at intermediate temperatures have not been illustrated yet.
The digital image correlation (DIC) method, with the advantages of slightest requirements in measurement environments, simple experimental setup, full-field measurement and high accuracy, has been widely used in non-contact deformation measurements . Since Lyons et al.  have proved the validity of DIC in measuring the high-temperature deformation via experiments, and this full-field non-contact measurement method has been extensively used to characterize the mechanical behavior of materials in extreme environments. Considering the black-body radiation that occurs at high temperatures, Grant et al.  and Pan et al.  introduced a band-pass filter and blue illumination in order to obtain accurate DIC measurements. Pataky et al. , Swaminathan et al. , and Leplay et al.  used DIC to investigate the creep and tensile properties of Haynes alloy, Hastelloy X alloy, and a ceramic. In this work, we employed the DIC method to explore the influence of γ′ precipitates on macro-scale localized deformation (the spatial and temporal characteristics of localized deformation were emphasized).
2. Materials and experimental procedure
Two nickel‒cobalt-based superalloys designed with various γ′ volume fractions (5% and 30 %) were considered in this work. The chemical components and the heat treatment of these alloys have been presented elsewhere . For DIC observations, 2-mm-thick sheet dumbbell specimens (gauge length: 25 mm, width: 5 mm) were subjected to tensile tests at a temperature and strain rate of 500 °C and 1 × 10-4 s-1, respectively. The force data were recorded at a sampling rate of 10 Hz. All the tests were started after holding the furnace at the test temperature for ~10 min. Samples for optical microscopy (OM) observation were etched in a solution of modiﬁed Kalling reagent (100 mL HCl, 100 mL methanol, and 50 g CuCl2). Fresh fracture surfaces were obtained by ultrasonic cleaning the fracture surfaces for 5 min. For transmission electron microscopy (TEM) observations, 50-μm-thick foils were prepared by means of a standard twin jet polishing technique using a solution of 10 % perchloric acid and 90 % ethanol (operating current: ~16 mA, temperature: -20 °C).
A DIC software developed by ourselves, referred to as PMLAB, was employed . In this system, images of the deformed specimens are captured at an image sampling rate of 2 fps using a synchronous acquisition trigger in this work. Prior to the tensile tests, sheet specimens were sprayed in flat white lacquer oversprayed with random black spots. Note that the DIC record ended at the strain of ~0.45 for 5% γ′ alloy. The photograph of the high-temperature DIC measurement system is shown in Fig. 1. This system consists of: a tensile machine (RGM-4050), a heating furnace, a temperature controlling box, a camera with a standard 50 mm lens, a light source, a trigger collector, and a high-performance computer for acquiring images and performing post-processing procedures. The array dimensions of each image = 2048 pixel × 2048 pixel; calculation grid size = 3 pixel; patch size = 29 pixel × 29 pixel; strain calculation window size = 15 point × 15 point. A value of ~34 pixel/mm was obtained for the correspondence between the actual dimension and the acquired image. The right-handed coordinate system was defined as follows: the transverse direction and the inverse tensile direction are X-axis and Y-axis, respectively.
Experimental setup of tensile test and DIC observation systems, which consist of a tensile machine (RGM-4050), a heating furnace, a temperature controlling box, a camera with a standard 50 mm lens, a light source, a trigger collector, and a high-performance computer for acquiring images and performing post-processing procedures.
3. Results and discussion
3.1. Microstructures observed via OM and TEM
Fig. 2 shows optical micrographs of the microstructures comprising the 5% (Fig. 2(a)) and 30 % (Fig. 2(b)) γ′ alloys (see insets for the corresponding grain size distributions). The insets indicate that grain size mainly concentrates into a range of 100-200 μm, and the average grain size of these two alloys is about 200 μm.
Optical micrographs showing the grain sizes of the microstructures comprising the different γ′ content alloys: (a) 5% and (b) 30 %. The micrographs were captured by Vitt: DM1000-M.
Fig. 3 shows the γ′ precipitate morphologies of the 5% (Fig. 3(a)) and 30 % (Fig. 3(b)) γ′ alloys, the chemical components of the matrix (Fig. 3(c)), and the γ′ precipitate size and fraction (Fig. 3(d)). Spherical γ′ precipitates with mean radius of ~15 nm and 35 nm were uniformly dispersed in the matrices of the 5% and 30 % γ′ alloys, respectively. Moreover, similar chemical components were present in the matrices of the two alloys (Fig. 3(c)), and the measured radius and area fractions of γ′ precipitates increased with the nominal volume fraction (Fig. 3(d)). These results demonstrated that the alloys accord with the designed value, which are consistent with our designed purpose.
(a) and (b) TEM images revealing the γ′ precipitate morphologies of the 5% and 30 % γ′ alloys; (c) chemical components of the matrix; (d) mean radius and area fraction of the γ′ precipitates.
Fig. 4 shows fractographs of the 5% (Fig. 4(b, c)) and 30 % (Fig. 4(e, f)) γ′ alloys, and sketches of the macroscopic fracture (see Fig. 4(a) and (d), respectively). The fracture morphology of the 5% γ′ alloy is characterized mainly by plastic dimples associated with transgranular fracture, which is mixed with a smoother section. The fracture morphology of the 30 % γ′ alloys appears to be the plastic dimples of the transgranular fracture and the large fracture section of intergranular fracture simultaneously. The occurrence of plastic dimples confirmed that both alloys exhibit a certain level of plasticity. However, the intergranular fracture of the high γ′ precipitate alloy indicated that the plasticity of this alloy is poor, consistent with the tensile test results. Yuan et al.  found that high Co content induced the change of deformation mechanisms from dislocation pairs cutting to SF shearing and/or deformation twinning. In our previous study , the SF density of the 30 % γ′ alloy was much higher than that of the 5% γ′ alloy at a temperature and strain rate of 500 °C and 3 × 10-4 s-1, respectively, and the micro-deformation mechanism switched from dislocation slip to SF deformation with decreasing strain rate. It can be speculated that the SF density of the 30 % γ′ alloy was also much higher than that of the 5% γ′ alloy at lower strain rate of 1 × 10-4 s-1. According to Tian et al. , the extension of microtwins can be prevented by grain boundary at intermediate temperatures, thereby promoting strain localization, and consequently, intergranular fracture. The intergranular fracture of the 30 % γ′ alloy may be attributed to high density SFs that can be blocked by grain boundaries, thereby leading to stress concentrations. Therefore, cracks were generated along the grain boundaries. Moreover, a cut section (Fig. 4(a)) and a normal section (Fig. 4(d)) were easily generated during the testing process of the 5% γ′ alloy and the 30 % γ′ alloy, respectively.
Sketches of the macroscopic fracture observed for the (a) 5% and (d) 30 % γ′ alloys and the fractographs of the (b, c) 5% and (e, f) 30 % γ′ alloys.
3.2. Stress-strain curves and the corresponding serration amplitudes
Fig. 5 shows the stress-strain curves of the 5% and 30 % γ′ alloys at a temperature and strain rate of 500 °C and 1 × 10-4 s-1, respectively. The yield strength and the ultimate strength of the 30 % γ′ alloy are considerably greater than those of the 5% γ′ alloy due to the second-phase strengthening effect. However, the elongation decreased with increasing γ′ content, and the serrated flow of both alloys began in the work-hardening stage. The average serration amplitudes of the 5% and the 30 % γ′ alloys are ~8.5 MPa and 12.5 MPa, respectively. The insets of Fig. 5 reveal that the serration amplitude distribution of both alloys are the peak-shaped distributions, and the value of distribution histogram shifts to the larger magnitude with the increase of γ′ content.
Engineering stress-strain curves of the 5% and 30 % γ′ alloys. The insets show the distributions of the serration amplitude, which indicate the peak-shaped distribution.
3.3. Localized deformation bands associated with γ′ precipitates
We selected serrations at strains of ~0.1 and 0.2 for comparison of the localized deformation generated in the two alloys. The plots in Fig. 6(a)-(c) show the morphologies of the serrations generated at strains of ~0.1, ~0.2, and ~0.4, respectively. The amplitude of the 30 % γ′ alloy is larger than that of the 5% γ′ alloy during the tensile test. Considering the effect of precipitates in impeding dislocation motion, the effective aging time for solute diffusion to the temporarily arrested dislocations increases, which leads to a significant interaction during the DSA process and consequently large magnitude serrations. In addition, both alloys are Co-rich, so the SF energy should be relatively low, thereby favoring the occurrence of SFs. According to Xu et al. , numerous interactions between SFs and Suzuki segregation can also retard dislocation motion and subsequently introduce DSA effects. This may lead to enhancement of the serration amplitude in high γ′ content alloys. Therefore, as expected, the localized deformation of the high γ′ content alloy was more severe than that of the low γ′ content alloy.
Selected serrations for DIC observations at strains of: (a) ~0.1, (b) ~0.2, (c) ~0.4.
Fig. 7 shows the obtained DIC strain maps for different loading stages of the 5% and 30 % γ′ alloys: Fig. 7(a), (c) and (e) correspond to the 5% γ′ content alloy at 0.1, 0.2, and 0.4, respectively; Fig. 7(b) and (d) correspond to the 30 % γ′ content alloy at 0.1 and 0.2, respectively; Fig. 7(f) shows the corresponding strain distributions. The localized deformation was concentrated into an inclined strip (as shown in Fig. 7) and the band inclination was similar to that occurring in Al-based alloys. Unlike the constant band width of the localized bands occurring in the Al-based alloys, the band width of the considered superalloys varied (albeit slightly) through the specimen width (see Fig. 7(a-e)). A comparison of the PLC bands observed in the low and high γ′ content alloys (Fig. 7(a-d)) revealed that the band width of the latter is slightly greater than that of the former. Fig. 7(f) shows the strain distributions of center lines corresponding to Fig. 7(a-e). These distributions demonstrate that, for a given loading stage, the maximum strain within the bands and the band width of the 30 % γ′ alloy are both greater than those of the 5% γ′ alloy. Here, we also give the statistic of the PLC band width (i.e., the width at half height) and band inclination, which have been well-defined in our previous study . Fig. 8 shows the variation in the band width and the band inclination with the loading procedure and γ′ content. The average band width of the 30 % γ′ alloy, ~73.6 ± 2 pixel (2.16 ± 0.06 mm), is modestly larger than that (~60.5 ± 4.5 pixel, 1.78 ± 0.13 mm) of the 5% γ′ alloy. The band inclination of these two alloys with respect to the tensile direction is ~54.3° ± 3°.
DIC strain maps of 5% and 30 % γ′ alloys at different stages of loading: (a), (c), and (e) correspond to the 5% γ′ content alloy at 0.1, 0.2, and 0.4, respectively; (b) and (d) correspond to the 30 % γ′ content alloy at 0.1 and 0.2, respectively; (f) shows the corresponding strain distribution. The three red points (P1, P2, P3) in (a) and (b) are used in the analysis of the local strain rate evolutions shown in
Variation in the PLC band width and the band inclination with loading procedure and γ′ content. A value of ~34 pixel/mm was obtained for the correspondence between the actual dimension and the acquired image.
3.4. Temporal evolution of local strain rate
As stated previously, similar to the case of Al-based alloys, localized bands were also observed in Ni-Co-based superalloys at 500 °C. We assessed the temporal characteristics of the PLC band. Fig. 9 shows the evolution of the local strain rate corresponding to three selected points (as shown in Fig. 7(a, b)) of different γ′ content alloys. These points are located at the upper, center, and lower regions of the specimen. The sharp peak indicates the appearance of the localized band at the corresponding time, and hence, the shift of these peaks can reflect the propagation features of the bands. As shown in Fig. 9, the local strain rate increased slightly in the early stage of plastic deformation, and remained constant in the intermediate and later stages of plastic deformation. The corresponding sharp increments observed for the 30 % γ′ alloy are higher than those of the 5% γ′ alloy, suggesting that the localized deformation in the former is more severe than that in the latter.
Evolutions of the local strain rate corresponding to three selected points (as shown in
3.5. Propagation features of PLC bands
Fig. 10 shows the propagation of PLC bands in different γ′ content alloys subjected to a strain of ~0.15 in our previous study . As shown in the figure, the bands propagate continuously in the low γ′ content alloy and intermittently in the high γ′ content alloy. The band inclination of the two alloys is ~60°, consistent with the previously mentioned observations. In the present study, for the first-time ever, a new interest phenomenon, i.e., special propagation of PLC bands in high γ′ content alloy, was observed in the late stage of a tensile test. As Fig. 11(a) indicates, zigzag bands characterized by head-to-tail connections with ±60° band inclination occur in the specimen. The strain evolution of the center lines (Fig. 11(b)) reveals the special propagation feature, which occurs only in the high γ′ content alloy. Furthermore, the PLC effect of high γ′ content alloy is towards B/C type development, indicating that the spatial coupling of the localized deformation band has decreased. Moreover, the areas swept by the localized deformation band are strengthened, and the strengthening effect increases with the content of γ′ precipitates. The zigzag band propagates forward with high probability. In addition, as Fig. 4(e) shows, the 30 % alloy underwent mainly intergranular fracture. This suggests that the special propagation of the PLC bands is correlated with the grain interaction deformation. Besides, intergranular fracture has certain directivity, which may result in the special propagation of PLC bands. However, further investigation of the mechanism governing zigzag band propagation is required.
DIC strain maps of different γ′ content alloys at a strain of ~0.15 in our previous study : (a) 5% γ′ alloy, (b) 30 % γ′ alloy.
Special propagation of PLC bands in the 30 % γ′ alloy: the (a) strain maps obtained for the tensile direction and (b) strain evolution corresponding to the center lines.
In summary, this study focuses on the macroscale deformation behaviors of localized bands in Ni-Co-based superalloys with different γ′ content at 500 °C. The spatial and temporal features of the localized deformation were characterized via surface strain maps obtained from DIC. The major conclusions of this study can be summarized as follows:
(1)The fracture morphology was characterized mainly by transgranular fracture with numerous dimples in the low γ′ content alloy, and intergranular fracture with large fracture section in the high γ′ content alloy.
(2)Spatial and temporal characteristics revealed that the localized deformation and the band width in the high γ′ content alloy are more severe and moderately larger, respectively, than those of the low γ′ content alloy.
(3)A special propagation feature, namely zigzag bands consisting of head-to-tail connections and characterized by a ±60° band inclination, occurred in the high γ′ content alloy.
This work was supported by the National Natural Science Foundation of China (Grant Nos.11335010, 51271174, 11802080 and 11627803), and the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB22040502).
In summary, this study focuses on the macroscale deformation behaviors of localized bands in Ni-Co-based superalloys with different γ′ content at 500 °C. The spatial and temporal features of the localized deformation were characterized via surface strain maps obtained from DIC. The major conclusions of this study can be summarized as follows: ...