Effects of Temperature and Secondary Orientations on the Deformation Behavior of Single-Crystal Superalloys

Abstract

The tensile behavior of single-crystal superalloys was investigated at room temperature (RT) and 850 °C, focusing on various secondary orientations. Transmission electron microscopy (TEM) and quasi in situ electron backscatter diffraction (EBSD) were employed to study the deformation mechanisms across length scales. Deformation at 850 °C enhanced the tensile ductility of the samples, evidenced by the more uniform coverage of dislocations across the γ and γ′ phases, and the fracture mode switched from pure cleavage at room temperature to mixed mode due to accelerated void growth. The influence of secondary orientations on mechanical properties is insignificant at room temperature. However, the ductility of the different secondary orientation samples shows significant variations at 850 °C, among which the one with [001] rotated 37° demonstrated superior ductility compared to others.

1. Introduction

Nickel-based single-crystal (Ni-based SC) superalloys have been widely employed in aircraft engines and gas turbines due to their structural stability, excellent mechanical properties, and corrosion resistance at high temperatures. The excellent elevated temperature performance of these alloys originates from the unique γ/γ′ two-phase coherent microstructure, where the ordered cubic phase γ′ (L1₂ structure) is coherent with the disordered γ matrix [1,2]. In terms of micromechanics, the strengthening phase γ′ and the coherent γ/γ′ interface to impede dislocation motions, and thus improving the overall mechanical properties of the alloy [3,4]. Tensile properties are regarded as a representative indicator of the comprehensive properties for SC superalloys, which can provide a reliable reference for fatigue life and creep properties. There have been many studies on the deformation mechanism of Ni-based SC superalloys under various experimental conditions, but the temperature-dependent deformation mechanism is not yet clearly understood. Especially at room and intermediate temperatures, different configurations of dislocations are closely related to the temperature, and these have significant impacts on the properties of the alloys [5,6,7,8].

Ni-based SC superalloys are mainly casted by directional solidification [9]. Due to the elimination of grain boundaries and the face-centered cubic (FCC) structure, the mechanical properties of the alloys show anisotropic characteristics, and it has been shown that the [001] orientation has the best overall performance [10,11]. During the past decades, a lot of studies have focused on the effects of orientations on mechanical properties, but most of them have been based on the influences of primary orientation [12,13]. However, studies in recent years have found that the secondary orientation also affects the mechanical properties [14,15,16,17,18,19,20,21]. Arakere et al. [14] have predicted that the appropriate secondary orientation can significantly increase the fatigue crack extension resistance of the component, without additional weight or cost. Sabnis et al. [15], using finite element analysis, argued that secondary orientation has a significant effect on stress distribution, and the slip mode as well as that the size and shape of the plastic zone are strongly dependent on the secondary orientation of the notch, which may have a significant impact on crack initiation. Zhou et al. [17] demonstrated that the tensile strength and fracture strain at room temperature are affected by the cooling holes and secondary orientations, with the (110) samples without holes exhibiting slightly higher tensile strength than the (100) samples. Suzuki et al. [19] tested fatigue crack propagation experiments on notched samples with secondary orientations (001) and (110) at different temperatures; the results showed the difference in crack propagation rates between (100) orientations was small at both 450 °C and 700 °C, while there was a significant difference in crack propagation rate for the (110) orientations, which consistently exceeded that of the (100) orientations at a variety of temperatures. Due to the anisotropy of the secondary orientation, Zhou et al. [20] demonstrated that samples with holes exhibited variations in tensile properties with secondary orientations, among which the (100) specimen demonstrated better tensile strength and plasticity compared to the (110) orientations. In summary, the optimal secondary orientation for different compositions of Ni-based SC superalloys may not be unique, and the activated slip systems vary for samples at different temperatures [19], thus requiring case-specific testing for different alloys.

Therefore, the main objective of this study is to test the diverged opinions on the effects of secondary orientations using a second-generation Ni-base superalloy. Additionally, we conducted the tests at elevated temperatures, hoping to uncover any related temperature effect. The influence of temperatures on the secondary orientation effect were explored by quasi in situ electron back scattering diffraction (EBSD), which was combined with TEM to investigate the dislocation mechanisms underlying secondary orientation effects at different temperatures.

2. Materials and Methods

2.1. Material Preparation

A second-generation Ni-based SC superalloy was used in the present work, and its composition is listed in

Table 1. Chemical composition of the alloy.

	C	Cr	Co	W	Al	Ta	Mo	Hf	B	Re	Ni
wt.%	0.05	7.05	7.5	5.4	6.24	6.58	1.51	0.16	0.0047	2.96	Bal

. Directionally solidified plate ingots were subjected to a solution treatment at 1300 °C for 2 h, followed by three-stage aging at 1120 °C, 1080 °C, and 900 °C for 4 h sequentially. All the cooling methods involved air cooling and temperature was controlled within a ±10 °C range. In order to determine the crystal orientation of the ingot, a small sample was cut from the top of the ingot and observed by EBSD. The results show that the solidification direction of the single crystal ingot was ~6° misaligned with the [001] orientation. Based on this orientation information, secondary orientation sampling was performed on the ingot. The outer surface of the plate ingot was used as the 0° reference plane; four different secondary orientations were, respectively, rotated around the [001] main axis by 5°, 37°, 47°, and 57° (Figure 1a). These orientations were chosen such that different slip systems are activated based on Schmid factors. Tensile specimens were prepared using electrical discharge machining (EDM), with the geometry shown in Figure 1b. Specimens were ground using SiC sandpapers ranging from 150# to 3000#, followed by polishing with 2.5 μm diamond paste and fine polishing with an 80 nm colloidal silica suspension until the sample surface was scratch-free. Subsequently, electrolytic polishing was performed at −20 °C using a 10% perchloric acid +90% alcohol solution for 50 s at −20 °C and 15 kV. The microstructures of the samples were shown in Figure 2a–c. As illustrated in Figure 2a, the scanning electron microscope (SEM) images show that a number of carbides are dispersed in the matrix in a needle or platelet-like morphology. The area fraction of carbides on the surfaces of samples in various secondary orientations was statistically analyzed, as depicted in Figure 2b. Variations in carbide proportions were observed among these orientations.

2.2. Mechanical Testing

The uniaxial tensile tests were performed using a MTS E45.105 electronic universal testing machine (MTS SANS, Shenzhen, China) at room temperature (RT) and a high temperature (HT, 850 °C) along the [001] orientation. Firstly, four specimens, each corresponding to one of the stated secondary orientations, were tested under a strain rate of 1.1$\times {10}^{-3}{\mathrm{s}}^{-1}$ until rupture to generate the stress–strain curve. Then, based on the stress–strain data, interrupted tensile tests were carried out under the same loading conditions, with the RT tests interrupted at 4%, 8%, and 12% of the accumulate strains, and the meso-scale GND density distribution was observed by EBSD at each of the strain levels. Tensile tests at 850 °C were interrupted at 4%, 8%, and 18% of the accumulate strains, and due to oxidation at the high temperature, the samples were ground and electro-polished before EBSD. The tensile rig compliance was corrected using procedures described in reference [22].

2.3. Microstructure Analysis

After the tensile test, the temperature dependence of the fracture surface morphology was observed using a Gmini-SEM460 scanning electron microscope (SEM) (Zeiss, Oberkochen, Germany). For TEM study, thin foils were cut parallel to the (001) orientation at 6 mm away from the fracture surface on the 37° orientation at RT and 850 °C. The foils were ground to a thickness of approximately 50 µm, and polished through a Leibow TJ100-SE-TMS (LEBO Science, Wuxi, China) standard temperature controlled electrolytic double-sprayer, then thinned in a Gatan 695 ion thinning apparatus, and the TEM experiments were performed in a FEI Talos F200X (FEI, Hillsboro, OR, USA).

2.4. Quasi In Situ EBSD

The samples were ground, chemical-mechanical polished, and electrolytic polished before the tensile testing. EBSD was performed using a FEI Versa5 scanning electron microscope (FEI, Hillsboro, OR, USA) and an EDAX velocity probe (EDAX, Mahwah, NJ, USA) under 20 kV, 13 nA, with a step size of 0.7 µm. For those tensile tests conducted at 850 °C, electro-polishing was conducted after the pre-determined strain before EBSD in order to remove the oxidation. All EBSD scans were maintained at the same region of interest during the interrupted testing. Data processing was performed using Matlab and GND density distributions were obtained using Mtex-5.7.0.

3. Results

3.1. Comparison of Tensile Properties

The true stress–strain curves are demonstrated in Figure 3. Obviously, the alloy exhibits completely different tensile behaviors at different temperatures, where the flow stress and work hardening at room temperature and 850 °C exhibit different features after yielding. The RT specimens all demonstrate a similar yield strength of ~867 MPa, an initial slow work hardening spanning ~1% of the plastic strain, followed by a work hardening rate of ~3910 MPa until fracture. The HT specimen demonstrated similar yield strength at ~865 MPa compared to the RT specimens, but the 37° sample demonstrated a higher loss of yield strength by 2.5% compared to the average of the other three orientations. In general, for the HT samples, the work hardening rate increased from ~267 MPa to ~731 MPa after yielding, and then gradually decreased after reaching 9% strain. Compared to the RT tests, the HT tests show strain softening before fracture, possibly because of necking due to the growth of voids. Significant improvement of ductility is seen from the samples tested at the HT, with an average fracture strain of 9.6% at RT compared to an average of 17% at the HT. A feature to notice is that among the secondary orientation samples tested at the HT, the fracture strain shows significantly higher dispersion (2.32 standard deviations) compared to those tested at RT (0.32 standard deviation). It is also interesting to note that the 37° sample has higher fracture strain at both the HT and RT.

3.2. Fracture Patterns at Different Temperatures

The fracture surfaces at different temperatures are displayed in Figure 4. On the fracture surface at RT (Figure 4a,b), there are some precipitate particles resting inside the micro-voids, indicating that those precipitate particles contributed to the micro-void nucleation and growth (indicated by the yellow arrows in the inset of Figure 4a). A significant number of casting voids were also observable in Figure 4a (indicated by the green arrows). In certain regions depicted in Figure 4a, the fracture surface exhibits a smooth texture devoid of river-like patterns or dimples; furthermore, micro-cracks and instances of crack bifurcation are discernible; these features indicate characteristics of a brittle fracture. In addition, some cracks were initiated at the location of large-sized precipitates (indicated by the white arrows in Figure 4b), which extended to the matrix during subsequent tensile loading. Therefore, the main type of fracture at room temperature is a brittle fracture, but the micro-dimples induced by carbides provide certain plasticity to the alloy.

As the temperature increased to 850 °C (Figure 4c,d), the fracture surface became less flat (Figure 4c) and not as faceted as in that from the RT test, and the local magnified view (yellow dashed box in Figure 4c, which is shown in Figure 4d) indicated some tearing ledges and cleavage steps together with a dimpled zone showing evidence of relatively large void growth (indicated by the red arrows). It can be inferred that the fracture surface is composed of multiple (111) planes (i.e., the dislocation pile-up plane) separated by tearing ledges due to the intersection of dislocation pile-ups from adjacent slip planes. Dimples are a characteristic of ductile fracture and are formed due to the casting voids and precipitates. The enlargements of the dimples were facilitated by the tensile loading at the higher temperature. The contrast of void growth at the HT to that at RT is consistent with the larger sample elongation observed at the HT (Figure 3b). The mechanism behind the formation of these dimples is subjected to further study but can be related to the more active dislocation slip at the elevated temperature which provides the prismatic dislocation loops and the passage for vacancy diffusion, both of which facilitate void growth.

3.3. Deformed Microstructure After Tensile Fracture

The bright-field TEM images of the deformed microstructure after tensile fracture at room temperature are shown in Figure 5. At first sight, high density dislocations are tangled in the narrow γ matrix, and some dislocation pairs sheared into the γ′ precipitates along the <110> directions (indicated by the red triangles in Figure 5c). Occasionally, some γ′ demonstrated elevated dislocation punch-in and entanglement, as highlighted by the white dashed boxes in Figure 5a,b. More frequently, stacking faults are found in the γ′ phase (marked white triangle in Figure 5c). The chemical compositions in the dislocation-enriched γ′ were analyzed by scanning transmission electron microscopy (STEM) and the corresponding energy dispersive x-ray spectroscopy (EDS) elemental mapping (at.%) is shown in Figure 6. Interestingly, the heterogeneously deformed regions where dislocations shear into the γ′ precipitates (in the white dashed boxes) exhibit significant enrichments in Co, Cr, and Re, which are the γ stabilizing elements. In contrast, Al and Ta, which are the γ′ stabilizers, are depleted in those regions.

The bright-field TEM images of the deformed microstructure after tensile fracture at 850 °C are shown in Figure 7. The dislocation configurations exhibit significant differences compared to those at RT. At 850 °C, the dislocations are more homogeneously distributed across both the γ and γ′ phases (Figure 7a). Notably and in contrast to room temperature deformation, stacking faults are absent, which may be attributed to an increase in stacking fault energy at higher temperatures [23,24], although the anti-phase boundary (APB) dislocation pairs (indicated by the white triangles in Figure 7b) are present, similar to the RT case. Furthermore, the misfit between γ and γ′ increases at high temperatures, prompting a formation of irregular dislocation networks at the γ/γ′ interfaces due to the release of misfit strain, as indicated by the arrow in Figure 7b. Additionally, some dislocations enter the γ′ phase through cross-slip under thermal activation, as demonstrated by dislocations CD and EF in Figure 7c. During deformation, certain dislocations extended into the γ′ precipitate parallel to the gliding direction in the γ phase, exemplified by dislocation AB in Figure 7c. In the micrograph shown in Figure 7d, some [110] curved dislocation pairs glide on (111) planes (indicated by the black triangles), while another portion of curved dislocation pairs initiated cross-slip (indicated by the red triangles). Importantly, unlike the RT case, there is no significant dislocation pile-up in both γ and γ′ at the HT. Dislocations sheared into the γ′ precipitates by a mixture of slip transmission and cross slip.

The composition of the region where dislocations sheared into γ′ was also analyzed by STEM and the corresponding EDS elemental mapping (at.%) is shown in Figure 8. It seems that the element distribution is more homogeneous in the sample tested at the HT. The Co, Cr, and Re preferential segregation into some of the γ′ phases, as observed in the RT test, is absent at the HT.

3.4. Quasi In Situ EBSD with Different Orientations

Figure 9 shows the meso-scale geometrically necessary dislocation (GND) density distributions during room temperature tensile tests for different secondary orientations, investigated at strains of 3%, 8%, and 12%. Following the concepts proposed by Hughes [25] and Gao [26], the additional dislocation storage required for lattice rotation arising from non-uniform plastic deformation in the presence of strain gradients is termed geometrically necessary dislocation. The local GND density changes indicated the degree of strain localization in the alloy. As depicted in Figure 9, although the total GND density increased with increasing tensile strain, the variation in GND density for each orientation was uneven.

As illustrated in Figure 9(a1–a4), the GND density ($\rho GND$) started low before loading. Around yielding at ~3% strain (Figure 9(b1–b4)), the 47° and 57° orientations accumulated higher dislocation densities compared to the other two orientations. When the strain reached 8% (Figure 9(c1–c4)), the dislocation accumulation on the 37° sample started to catch up. The features of slip bands were not clearly observed in the GND density maps because dislocations in the slip bands are theoretically not “geometrically necessary”, meaning they do not induce lattice rotations apart from a relative lattice shift. A few slip trace features can be observed on the maps and are indicated by the red arrows in Figure 9(b1,b4). Those features are better named “shear bands” rather than slip bands, as the interactions between multiple dislocation types within those bands gives rise to local lattice curvature, hence the observed GND hot spots. A few cases of GND density concentration were observed around the precipitates, as indicated by the white arrows in Figure 9(a2,b2). The overall strain accumulation for the 5° orientation sample was not as significant as those of the other orientations as indicated by the lack of dislocation cell build-up evidenced in Figure 9(d1) compared to Figure 9(d2–d4). Moreover, the rectangular-shaped dislocation cells possibly indicate that the cell formation at RT is primarily influenced by the deformation within the γ channel.

Figure 10 illustrated the GND density distribution during tensile testing at 850 °C for the same sets of secondary orientation samples, obtained at 3%, 8%, and 18% strains. Irregularly high dislocations are seen on the undeformed samples, which are likely due to the residual scratches from mechanical polishing. These scratches are less likely to affect other GND maps at higher strain levels as they are subsequently removed by electro polishing (to remove oxides from HT testing). The total GND density also increased with strain, but, in sharp contrast to the RT distributions in Figure 9, the GNDs from HT testing are more uniformly distributed. Yielding occurs at approximately 3% strain; a notable feature was the presence of precipitates in the 5° sample which accelerated the GND build-up at 3% strain (Figure 9(b1)) compared to the precipitate-free cases. A dendrite boundary (indicated by the white arrow) was captured in the HT cases, which serves as a strain concentrator as indicated by the elevated GND densities along the boundary. The accumulation of GND near precipitates could be attributed to the pile up of dislocations at the interface together with strain gradient plasticity due to the differing elastic properties between the precipitate and the matrix. An abrupt increase in GNDs was observed when the strain went beyond 8%. The GNDs at 18% strain are more uniformly distributed in the matrix with a smaller dislocation cell size (compared to the RT case) and local GND hot spots at the dendrite boundary and precipitates.

4. Discussion

4.1. Deformation Mechanism at Different Temperatures

Based on the experimental results, Figure 5 depicts the complete structural defects observed at room temperature. Since the higher intrinsic strength of γ′, plastic deformation primarily occurs in the γ channels and results in local dislocation multiplications. The confined dislocation multiplication together with the higher RT shear strength of the γ′ phase resulted in pronounced work hardening as shown in Figure 3a. Stacking faults are usually difficult to observe in the γ channels attributed to the high density of dislocations [27]. With the applied stress further increased, some dislocations are able to shear into the γ′ phase in the form of partial dislocation pairs.

It is well known that the formation of APB and a stacking fault (SF) is related to their formation energies, which in turn are highly dependent on the deformation temperature [23,24,28,29]. APBs and SFs are identified as the main features in the γ′ phase from the fractured specimens deformed at RT. On the one hand, due to the low stacking fault energy at RT [23,24,30], the super-dislocations [10$\overline{1}$] shear into γ′ precipitates and then decompose into a couple of super-Shockley partial dislocations on the (111) plane, corresponding to the Burgers vectors of 1/3[11$\overline{2}$] and 1/3[2$\overline{1}\overline{1}$], with SFs distributed between the two super-partials dislocations. This reaction in the γ′ can be expressed as follows:

$$\mathrm{a} [10\overline{1}] \to (\mathrm{a}/3)[11\overline{2}]+(\mathrm{SISF})+(\mathrm{a}/3)[2\overline{1}\overline{1}]$$

(1)

where a is the lattice constant.

On the other hand, considering the highly heterogeneous plastic deformation, dislocations with the same type of Burgers vector tend to pile up at the γ/γ′ interface. When the local shear stress in front of the pile up becomes high, super-dislocations shear into the γ′ phase and may decompose into APB pairs, such as the red triangles indicated in Figure 5c. As the APB pairs cut into the γ′ phase, the reaction can be expressed as follows:

$$\mathrm{a} [\overline{1}10] \to (\mathrm{a}/2)[\overline{1}10]+(\mathrm{APB})+(\mathrm{a}/2)[\overline{1}10]$$

(2)

Although different APBs may possess the same scalar Burgers vector, b = [110], the activation of different slip systems during tension causes APBs to slip and decompose on various slip planes. Additionally, the a/2[110] dislocations of different slip systems shear into the γ′ phase and interact with each other to form a dislocation network within the γ′ phase, as indicated by the white dotted box in Figure 5a,b. These dislocation networks hinder the movement of subsequently punched-in dislocations, contributing to a certain level of work hardening.

Compared to room temperature deformation, deformations at 850 °C exhibit two notable characteristics. Firstly, there is a higher dislocation density within the γ′ precipitates. Notice that the APB energy decreases with temperature [23,28]; hence, as the external stress increases, more super-dislocations break through the γ/γ′ interface and decompose into APB couple pairs, which explains the drop in work hardening in the later stages of the tensile testing. The initial working hardening of the HT test samples may be similar to that of those tested at RT, i.e., due to dislocation multiplication in the γ phase and pile up at the γ/γ′ interface, but due to the easier transmission into the γ′ phase, the overall work hardening rate is much lower. The dislocations participated uniformly in both the γ and γ′ phases as can be seen from Figure 7, indicating a more homogeneous deformation at 850 °C. This dislocation patterning characteristic may explain the improved HT ductility. The other notable feature is the presence of curved dislocations as mentioned in Section 3.3. This phenomenon has been observed in other studies [8,30], and is potentially related to the dislocation self-arrangement into a lower energy state. Referring to the linear dislocation energies formula [31]: $E_s=\left(1-\nu \right)E_e$, where ν is the Poisson’s ratio, $E_s$ represents the elastic energy of screw dislocations and $E_e$ denotes that of the edge dislocations. It follows that the creation of curved dislocations from an initially straight edge dislocation is energetically favorable. By comparing the dislocation configurations between Figure 5 (RT, more straight dislocations) and Figure 7 (HT, more curved dislocations) it may be the case that thermal activation plays an important role in this process. The formation of a higher proportion of curved dislocations can also be seen by contrasting the mesoscale dislocation configuration between Figure 9 and Figure 10. When considering the cross-slip behavior of certain mixed dislocation pairs such as those indicated by the red triangles in Figure 7d, they initiate cross slip from the octahedral (111) planes to the cubic (100) planes due to the higher APB energy on the former [32,33]. This process occurs under thermal activation. However, the (100) planes are not easy glide planes for FCC nickel-based superalloys, making further slip of APB pairs on the (100) planes difficult and thus hindering subsequent dislocation motion, thereby forming Kear–Wilsdorf (K-W) locks [34]. This effect may have countered the strength decrease due to the easier dislocation activation at 850 °C and therefore maintained a similar yield strength level as that of the RT tests.

As mentioned above, γ′, in which there is γ stabilizer (Co, Cr, and Re) enrichment, is more heavily deformed at RT. Two possibilities can be considered here. Firstly, the diffusion rate is higher along the dislocation line compared to the bulk material. This results in element segregation in the heavily deformed region, commonly referred to as pipe diffusion [35,36]. The second possibility is that elemental segregation results from the casting process, where the γ stabilizer-enriched γ′ precipitates have lower stacking fault energy and therefore make dislocation generation and dislocation transmission easier from the neighboring γ phase.

4.2. Effects of Temperature and Secondary Orientations on Deformation Behavior

The true stress–strain curves during tensile testing at RT and 850 °C show minimal variation in yield strength, a phenomenon commonly observed in other superalloys as well [30,37]. Abnormal yield strengths of superalloys are observed at 600–800 °C, due to thermally activated dislocations that cross-slip from {111} planes to {001} planes, forming K-W locks. These locks tend to be loosened at higher temperatures, returning the yield strength to the room temperature level; therefore, our test at 850 °C did not capture the abnormal yield phenomenon.

To quantify the extent of work hardening at the two temperatures, the work hardening rate θ was calculated and plotted as a function of true plastic strain in Figure 11a,b, i.e., θ = $\mathrm{d}\mathsf{\sigma }/\mathrm{d}\mathsf{ϵ}$, and the work hardening response can be divided into three stages. In Stage 1, dislocations are impeded by the γ′ precipitate, leading to an increased hardening rate at RT. In contrast, as dislocations are relatively easy to shear into γ′ at 850 °C, a drop in the work hardening rate is observed in Stage 1. Stage 2 involves initial softening followed by hardening at RT, attributed to the localized strain concentration [38], as depicted in the distribution of GNDs in Figure 9. Hardening at high temperatures is due to the formation of K-W locks. When it comes to Stage 3, the higher applied load causes the unlocking of the K-W locks, resulting in softening at 850 °C. The capability of dislocations shearing into the γ′ phase at 850 °C facilitates uniform plastic deformation, as shown in Figure 10, and ultimately leads to a gradual decrease in the work hardening rate until fracture.

The effect of the secondary orientation on mechanical behavior remain unclear throughout the literature. According to some research in the literature, it seems that the effect of secondary orientation takes effect only through local structural irregularities. For example, Guo et al. [21] discovered that in cylindrical hole specimens, the secondary orientation influences the location of creep crack initiation at hole edges. Zhou et al. [20] confirmed that the presence of a circular hole induces a multi-axil stress condition that is favorable for anisotropic plastic deformation around the hole. Furthermore, secondary orientation influences the activation of slip systems adjacent to the hole, with a greater number of activated slip systems leading to more uniform plastic deformation and higher fracture strains. Zhou et al. [17], by conducting tensile testing on a specimen with circular holes, concluded that the three-dimensional tensile stress parallel to the stress axis around the holes promotes stress concentration. Under these conditions, the <100> slip systems experience higher resolved shear stresses, leading to greater activation of slip systems, enhanced work hardening, and higher strength.

On the other hand, some experiments have also demonstrated that the secondary orientation has minimal impact on the mechanical properties of single-crystal superalloys. Guo et al. [39] found that during RT tensile testing of specimens with circular holes, the secondary orientation had no influence on the stress–strain curve. The only difference observed was in the initiation of slip bands at the edges of the holes. The lack of influence on mechanical properties may be attributed to the rapid initiation and propagation of cracks in the large circular holes, which could mask any secondary orientation effects. Zhou et al. [17], through the in situ tensile testing of smooth specimens, discovered that the secondary orientation has a minor impact on yield stress and ultimate tensile strength. However, it notably influences the activation of slip systems, as well as the initiation and propagation of edge cracks in the samples. Sabnis et al. [15], through three-dimensional finite element simulations and experimental data, found that the tensile curves of edge U-notch specimens with varying secondary orientations exhibit striking similarity.

The lack of consistency in data in the literature on the effect of secondary orientations may be attributable to the specific testing conditions and chemical compositions of the specimen used, but it is clear that the most prominent effect of secondary orientations is that they lead to differences in the activations of slip systems. However, despite the distinct activation patterns of slip systems across different secondary orientations, this variability does not always translate into differences in mechanical properties. Secondary orientation effects may be amplified through methods such as reducing hole size, pre-notching, conducting tests at elevated temperatures, extending test durations, and increasing specimen thickness.

The results in the current study indicated minimal variation in strength among different secondary orientations, but the fracture plasticity exhibited greater dispersion across different secondary orientations at 850 °C compared to RT. Specifically, the 37° orientation demonstrated superior plasticity at both temperatures. Additionally, all four orientations exhibited similar work hardening behavior at the same temperatures, but all the samples showed a continuous decrease in the work hardening rate θ (obtained by taking the derivative of stress to strain; the derivative of θ with respect to strain therefore indicates changes in the work hardening rate) during Stage 3 prior to fracture. The inset in Figure 11a,b depicts the derivative of θ with respect to strain in Stage 3. Notably, the 37° samples exhibited the lowest decrease in θ at both temperatures. This slower reduction in strength suggests reduced susceptibility to fracture under high-stress conditions, resulting in higher fracture strain. Moreover, to accurately depict the change in GND in the matrix during tensile deformation, the average growth rate of GND density in the matrix, excluding carbides, was calculated using data from Figure 9 and Figure 10, and presented in Figure 11c,d. It is evident that the 37° orientation generally shows a higher growth rate in GND density compared to the other orientations, which is consistent with the slower drop of the working hardening rate. The number of activated slip systems and their sustained multiplication capability play a crucial role in the continuous accumulation of GND density and correlate closely with plasticity [38,40,41]. Among the batch of secondary orientation samples tested, the 37° orientation meet those conditions and therefore demonstrates higher fracture strain. Notice from Figure 2b that the 37° sample also has the highest carbide volume fraction, which may be counterintuitive to the higher ductility, but the local uniformity of slip activation may override the reduced ductility due to the higher carbide area fraction in this specific orientation, which is consistent with the results indicated by Zhou et al. [20].

5. Conclusions

The influences of temperature and secondary orientations on the plastic deformation of Ni-based SC superalloys were investigated using quasi in situ tensile tests and multiscale characterizations. The main conclusions are summarized as follows:

• Deformation at 850 °C leads to higher tensile ductility and higher variance in tensile ductility among secondary orientations. Apart from this variation in ductility at higher temperatures, we found a weak influence of secondary orientation on mechanical properties.
• The fracture mode at RT is brittle fracture, but the micro-dimples formed at 850 °C provide certain plasticity to the alloy, leading to a mixture of cleavage and ductile fractures mode.
• At RT, dislocations accumulate mainly at the γ phase with limited dislocation transmission into the γ′ phase and the γ′ phase deform by stacking faults formation. At 850 °C, stacking faults are absent in the γ′ phase, but it is relatively easy for dislocation transmission to occur through paired partial dislocation mechanisms, leading to a more uniform dislocation distribution, resulting in a lower work hardening rate but improved fracture strain.
• Local entanglements of high density dislocations were found in some of the γ′ phase. This local high deformation zone is correlated with the local segregation of γ stabilizing elements.
• The secondary orientation sample with the [100] direction rotated 37° demonstrated superior tensile ductility at 850 °C, which was supported by the high and sustained accumulation of GNDs even though this sample contained the highest area fraction of precipitates.