Effect of Grain-boundary Phases on the Tensile and Compressive Creep Behavior of a Ni–Co Based Superalloy with High W Content

Abstract

The effects of intergranular W-rich compounds (W₆C/W₁₂C and µ phases) on the creep properties of a γ’-precipitation-strengthened, wrought Ni-Co-based superalloy were examined. Two microstructures with different intergranular precipitates (but the same volume fraction and size of intragranular γ’ precipitates) were prepared by adjusting aging conditions, leading to two distinct grain-boundary precipitation microstructures, with grain-boundaries exhibiting: (i) W-rich compounds (W₆C/W₁₂C and µ), or (ii) W-free compounds (Cr₂₃C₆ and γ’ phases, as in conventional Ni-based alloys). In compressive creep tests, lower minimum creep rate and longer 1% creep life were observed for alloys with the second grain-boundary microstructure (with Cr₂₃C₆ and γ’) than for the first grain-boundary microstructure (with W₆C/W₁₂C and µ). On the other hand, in tensile creep tests, longer creep rupture life was observed for alloys with the first grain-boundary microstructure than for the second, especially at higher stresses. Based on the grain boundary precipitation strengthening mechanism, it is suggested that the difference in compressive minimum creep rate between both microstructures can be explained by the difference in grain-boundary coverage. In tensile creep tests, the creep rate due to cavity and crack growth appears to be the dominant factor in the overall creep rate at higher stresses, consistent with Stress Assisted Grain-Boundary Oxidation (SAGBO). It appears that the grain-boundary precipitation of W₆C/W₁₂C and µ phases is more effective in suppressing SAGBO than grain-boundary precipitation of Cr₂₃C₆ and γ’.

1. Introduction

Wrought, heat-resistant Ni-base superalloys are widely used in combustors for aircraft and power generation plants, engine valves for automobiles, and other applications. Particularly, medium-strength class wrought heat-resistant Ni-based superalloys such as Alloy 263 and Alloy 520 are generally reinforced with coherent intragranular γ’-Ni₃Al, while the grain boundaries are reinforced with carbides (such as Cr₂₃C₆) and γ’ precipitates.¹⁾ Recently, an increase in creep resistance due to increased coverage of grain boundaries by second-phase particles (grain-boundary coverage) has been proposed as a mechanism of Grain-Boundary Precipitation Strengthening (GBPS).^{2,3,4,5,6,7,8)} Takeyama reported that an austenitic heat-resistant steel with intragranular intermetallic γ’’-Ni₃Nb and intergranular intermetallic Laves-Fe₂Nb with high grain-boundary coverage exhibits comparable creep rupture strength with Ni-base superalloys.⁸⁾ Zhang et al. confirmed that precipitation of grain-boundary carbides is effective for increasing the creep resistance of austenitic heat-resistant steels.^9,10,11) In addition, GBPS mechanism is studied in some γ’ precipitation strengthening Ni-based heat-resistant alloys. Ito et al.¹²⁾ conducted tensile creep tests using Ni-base alloys with different carbon contents and reported that the creep rate in the accelerated creep region was decreased and the rupture life was increased when the grain-boundary coverage by Cr₂₃C₆ was high. Elbatahgy et al.¹³⁾ also reported that grain-boundary precipitation strengthening is also effective due to γ’ precipitates at grain-boundaries in Alloy 80A.

In this study, we used a newly developed γ’ precipitation strengthened wrought Ni–Co based heat-resistant alloy, COWALOY.¹⁴⁾ This alloy has better hot forgeability and creep rupture properties than Alloy 520 and it is expected to be applied to automotive exhaust valves and heat-resistant springs. Unlike conventional wrought Ni-based heat-resistant alloys, COWALOY contains a high tungsten concentration (16 wt.% W). Therefore, it is reported that not only Cr₂₃C₆ and γ’ but also W-rich phases precipitate, such as carbides (W₆C, W₁₂C) and the μ-Co₇W₆ intermetallic compound.¹⁵⁾ Here, W₁₂C is a metastable phase which does not appear in the phase diagram. The effect of these W-rich phases on creep properties is not yet clarified. According to the grain-boundary precipitation strengthening mechanism, it is suggested that the creep rate is determined only by the grain-boundary coverage, not by the type of precipitating phase. Hence the grain-boundary precipitation of these W-rich compounds is expected to reduce the creep rate in this alloy. However, variations in grain-boundary precipitates, such as crystal structure and chemical composition, may cause differences in physical and chemical properties, such as precipitation morphology, interface strength between the matrix and precipitated phase, and toughness of the precipitated phase itself, which may affect creep rupture life and ductility. Therefore, the effectiveness of the precipitation of W-rich phases at grain boundaries should be examined in detail.

The purpose of this study is to clarify the effect of intergranular W compounds (W₆C/W₁₂C and μ-Co₇W₆) on the creep properties of COWALOY. The creep properties were compared for two distinct microstructures: one in which W compounds are intentionally precipitated at grain boundaries and the other in which Cr₂₃C₆ and γ’ precipitate at grain boundaries, as in conventional Ni-based alloys. In addition to tensile creep properties which are expected to cause differences in creep rupture life, compressive creep properties, which do not cause rupture, are also examined in this study.

2. Experimental Procedures

The chemical composition of COWALOY used in this study is given in Table 1. At first, a square forged bar with a thickness of 150 mm was prepared. After homogenization heat treatment at 1280°C/168 h, round bars with a diameter of 20 mm were made by hot forging. Then, after further solution treatment at 1200°C/30 min and removing oxide scale, round bars with a diameter of 13 mm were made by cold swaging. Figure 1 shows the Time-Temperature-Precipitation (TTP) diagram of the investigated alloy prepared by the preliminary experiments (details are described in Supporting Information). According to this diagram, only grain-boundary W-compounds (W₆C/W₁₂C and/or μ) precipitate at temperatures above ~1000°C and only grain-boundary γ’ and/or Cr₂₃C₆ precipitate at 800°C up to ~200 h by aging heat treatment of solution treated samples. Therefore, the cold-swaged round bars were heat-treated at 1200°C for 5 min and water cooled to adjust the grain size to ~100 μm, and then two types of microstructures were obtained by aging heat treatment. One bar was at first aged at 1020°C/8 h and water cooled and then aged at 800°C/24 h and water cooled to form W₆C/W₁₂C and μ phases at grain boundaries and γ’ precipitates in the matrix (hereinafter referred to as “2-step aging condition”). The other bar - intended to show a microstructure with γ’ and Cr₂₃C₆ precipitated at grain boundaries and γ’ formed in the matrix - was only aged at 800°C/24 h and water cooled (hereafter referred to as “1-step aging condition”).

Table 1. Chemical composition of the investigated alloy (wt.%).

Ni	Co	Cr	W	Al	C	B	Zr
47.3	Bal.	14.7	16.1	3.6	0.012	0.007	0.002

Fig. 1. Experimental TTP diagram of the investigated alloy. Only grain-boundary precipitates are assessed. The two dashed lines indicate the C-curves of W compounds (W₆C/W₁₂C/μ) and γ’/Cr₂₃C₆, respectively.

Creep tests were performed on specimens aged to either of these two aging conditions below 800°C to avoid microstructural changes during the tests. Compressive creep tests were performed on mirror-polished specimens (6 mm in diameter and 14 mm in length) at 800°C in air, for 8 stresses between 320 and 635 MPa. To avoid stress drop during the test, the load was increased every ~0.5% strain increment and interrupted at a maximum strain of 5%. Tensile creep tests were conducted in air at two temperatures (700°C/520–630 MPa and 800°C/294–450 MPa), using round bar specimens in accordance with JIS Z 2271, with a gauge length of 30 mm and a diameter of 6 mm. In addition, tensile creep tests were also conducted under inert gas (Ar, with <0.1% residual O₂) at 800°C/409 MPa to investigate the effect of oxidation on tensile creep behavior. Tensile creep tests were conducted under constant load conditions until rupture.

Microstructural observations of the pre- and post-creep specimens were conducted using a FEI Quanta 650 Scanning Electron Microscope (SEM), equipped with Electron Backscatter Diffraction (EBSD). Secondary Electron Images (SEIs) and Backscattered Electron Images (BEIs) were captured at a 10 mm working distance under 5, 15 or 20 kV accelerating voltage. EBSD scans were performed at 25 kV and a step size of 1 mm; acquired EBSD data were analyzed using OIM Analysis software. The SEM specimens were embedded in epoxy, mechanically polished, and then vibratory polished to a mirror finish for 2 h, using a colloidal silica solution with a particle size of 0.06 mm. Also, some tensile-crept specimens were embedded in epoxy after being pre-plated with Ni to observe the near-surface area. For each grain boundary studied, the sum of the length of grain-boundary precipitates relative to the grain-boundary length was measured to obtain the grain-boundary coverage, with an average value determined over 30 grain boundaries. Elemental analysis by a JEOL JXA-8500F Field Emission Electron Probe Micro Analyzer (FE-EPMA) was used for phase identification. Since it was impossible to distinguish between W₆C and W₁₂C, the W carbides will be referred to as W₆C/W₁₂C. The microstructures of pre-creep specimens were also observed in etched conditions using Kalling’s no. 2 reagent to image intragranular precipitates. Hardness tests were performed using a micro-Vickers hardness tester with a load of 1 kg and a loading time of 10 s (averaged over 5 measurements within multiple grains).

3. Results

3.1. Initial Microstructures after Aging

Figure 2 shows the pre-creep microstructures of the 2-step and 1-step aging conditions: bright-contrast, coarse W₆C/W₁₂C particles of ~500 nm thickness and flake-like μ-Co₇W₆ particles of <100 nm diameter are found at the grain boundaries of the 2-step aging condition, and between the W₆C/W₁₂C particles small amount of gray-contrast Cr₂₃C₆ particles are also precipitated (Fig. 2(a)). The μ particles appear to precipitate independently of W₆C/W₁₂C, without contacting each other. On the other hand, for the 1-step aging condition, 100 nm thick, globular Cr₂₃C₆ particles with gray contrast and dark-contrast film-like γ’ particles are observed at the grain boundaries (Fig. 2(b)). Although nodule-like γ’ particles^16,17) are also observed at the grain boundaries (Fig. 2(c)), the amount of precipitation is small. The average grain-boundary coverage in the 2-step and 1-step aging conditions is 67 and 82%, respectively, and the precipitates covering the grain boundaries are mostly W₆C/W₁₂C in the 2-step aged specimen and mostly Cr₂₃C₆ in the 1-step aged specimen. Within grains, both aged specimens exhibit γ’ particles with a diameter of ~50 nm and the average hardness inside grains is ~415 HV (Figs. 2(d), 2(e)). In addition, no precipitation-free zone (PFZ) of intragranular γ’ particles is observed in the vicinity of grain boundaries. From these observation, it is apparent that the intragranular microstructures of both aging conditions are nearly identical, but the precipitated phases and their precipitation ratios at grain boundaries are different. Here, each grain boundary phase was identified by comparing the results of elemental analysis by FE-EPMA (Fig. 3) with the precipitation morphology. Namely, W is enriched together with C in bright-contrast coarse particles, suggesting that these particles are W₆C/W₁₂C, while only W is enriched in bright-contrast flake-like particles, indicating that these are μ phase in 2-step aging condition (Fig. 3(a)). On the other hand, Cr is concentrated with C in gray-contrast particles (Cr₂₃C₆) and Al is enriched in dark-contrast particles (γ’) in 1-step aging condition (Fig. 3(b)). In the subsequent microstructural analysis, the phases were identified based on the precipitation morphology only.

Fig. 2. Pre-creep microstructures of (a, d) 2-step aging condition and (b, c, e)1-step aging condition: (a, b) BEIs of grain boundaries (after mechanical polishing), (c) SEI focusing on grain boundary γ’ nodule (after etching), (d, e) SEIs within grains (after etching).

Fig. 3. Elemental mappings of pre-creep microstructures by FE-EPMA: (a) 2-step aging condition and (b) 1-step aging condition. (Online version in color.)

3.2. Compressive Creep Behavior in Air

Figure 4(a) shows, for both aging conditions, the relationship between stress and minimum creep rate in compressive creep tests performed in laboratory air at 800°C. The apparent stress exponents of both aging conditions are 7–8 below 550 MPa. These values are slightly higher than the value for dislocation creep (power-law creep) of pure metals (4–7),¹⁸⁾ but it is reported that the existence of grain-boundary precipitates increases stress exponents.^9,10,11) Although a power-law breakdown¹⁸⁾ is observed above 550 MPa, where the apparent stress exponent is 12–16, we focus here only on the dislocation creep stress range. The minimum creep rate of the 2-step aging condition is approximately twice that of the 1-step aging condition. Figure 4(b) shows the relationship between stress and 1% creep life: the creep life of the 2-step aging condition is shorter than that of the 1-step aging condition, and the difference increases at lower stresses, reaching a factor ~1.5 at 409 MPa. Figures 4(c), 4(d) shows creep curves (strain vs. time) and creep-rate vs time curves for both aging conditions at 409 MPa as representative examples. A tertiary stage is observed at later times, and the difference in creep rate between both aging conditions is particularly large in this stage. Since the initial intragranular microstructure is almost the same between the two aging conditions (Figs. 2(d), 2(e)), it can be concluded that the difference in grain-boundary precipitates affects the compressive creep properties.

Fig. 4. (a) Minimum creep rate vs. applied stress plots, (b) applied stress vs. 1% creep time plots, (c) 409 MPa creep curves (strain vs. time), and (d) 409 MPa creep rate vs. time curves of the compression creep tests at 800°C for 2-step and 1-step aging conditions.

Figure 5 shows Kernel Average Misorientation (KAM) maps for both aging conditions after creep was interrupted at 1% (early in the tertiary stage), 2% (mid- tertiary) and 5% (late- tertiary) strain in the 800°C/409 MPa compression creep tests. At the beginning of the tertiary stage, regions of high KAM values suggesting local deformation are observed near grain boundaries (Figs. 5(a), 5(d)); as creep progresses, the regions of high KAM values expand within the grains (Figs. 5(b), 5(c), 5(e)). Next, Fig. 6 shows the microstructures near grain boundaries to illustrate the correspondence between the KAM maps and the microstructures. In both aging conditions, ledges are formed at grain boundaries and, at the tips of the ledges, strain contrast (suggesting localized plastic strain accumulation) is observed. Neither fracture of intergranular precipitates nor cavities at the interface are observed in any of the interrupted specimens. In the 1-step aging condition, bright-contrast W₆C/W₁₂C or μ particles (as indicated by arrows in Figs. 6(d), 6(e)), which may have been newly precipitated during creep, are observed at the grain boundaries, but in small amounts. Figure 7 shows the intra-grain microstructure of the 2-step aging condition interrupted at 5% strain, where the increase in KAM value in the grains is significant. Many linear contrasts suggest slip bands in the grains (Fig. 7(a)), indicating the occurrence of strain localization^19,20) reported for some Ni-based alloys. Furthermore, the dissolution of γ’ particles was observed inside some of slip bands (Fig. 7(b)). Thus, tertiary compressive creep is clearly associated with microstructural changes near grain boundaries and within grains. Finally, no indication of grain-boundary oxidation was observed in both aging conditions, as illustrated in Fig. 8 which shows vertical sections near parallel section surface of the compressive-crept samples tested at 800°C/409 MPa/195 h.

Fig. 5. Kernel Average Misorientation (KAM) maps of the compressive crept specimens tested at 800°C/409 MPa: (a–c) 2-step aging condition, and (d, e) 1-step aging condition. (Online version in color.)

Fig. 6. BEIs of the compressive crept specimens tested at 800°C/409 MPa: (a–c) 2-step aging condition, and (d, e) 1-step aging condition.

Fig. 7. (a, b) BEIs of the compressive crept specimens for the 2-step aging condition tested at 800°C/409 MPa and interrupted at 245 h. (b) high-magnification image of the black frame in (a).

Fig. 8. BEIs of the vertical sections near surface of the compressive crept specimens tested at 800°C/409 MPa and interrupted at 195 h: (a) 2-step aging condition and (b) 1-step aging condition.

3.3. Tensile Creep Behavior in Air

Figure 9(a) shows the relationship between stress and minimum creep rate of tensile creep tests performed in laboratory air, for both aging conditions. The tensile minimum creep rate for the 2-step aging condition is slightly greater than (or equal to) the 1-step aging condition. The apparent stress exponent is ~9 at both 700 and 800°C, which is slightly greater than the apparent stress exponent in compressive creep at 800°C, and the minimum creep rate itself is also greater in tension. The difference in minimum creep rate between compressive/tensile creep test increases with increasing stress and there is ~3 fold difference at 800°C/409 MPa. To describe tensile fracture, Fig. 9(b) shows the stress/Larson Miller parameter relationship and Fig. 9(c) shows the creep rupture elongation/stress relationship. Here, the test temperature and creep rupture life were used to calculate the Larson Miller parameters. The creep rupture life of the 2-step aging condition tends to be the same as, or longer than, that of the 1-step aging condition, and the higher the stress, the larger the difference in creep rupture life. In other words, the aging conditions providing a better creep performance is opposite for the tensile and compressive creep conditions. For the 700°C/520 MPa test of the 1-step aging condition, early rupture occurred, which is different from the trend of rupture life under the other test conditions. Furthermore, at 800°C, the creep rupture elongation for the 2-step aging condition is greater than that for the 1-step aging condition and is about three times greater in the 409 MPa test. Figure 10 shows creep curves and creep rate/time curves for both aging conditions at 800°C/409 MPa as representative examples. For both aging conditions, onset of tertiary creep starts earlier in tension than for compressive creep, for the same temperature and stress conditions.

Fig. 9. (a) Minimum creep rate vs. applied stress plots, (b) applied stress vs. Larson Miller parameter plots, and (c) creep rupture elongation vs. applied stress plots of the tensile creep tests under air atmosphere for 2-step and 1-step aging conditions.

Fig. 10. (a) Creep curves (strain vs. time) and (b) creep rate/time curves of the tensile creep tests at 800°C/409 MPa under air atmosphere for 2-step and 1-step aging conditions. The dashed lines indicate the creep curves and creep rate/time curves of the compressive creep tests from Fig. 4.

Figure 11 shows the microstructures near the fracture surface in parallel cross-sections of the 800°C/409 MPa tensile crept specimens for both aging conditions. In both cases, creep rupture occurred mainly at grain boundaries, and crack propagation starting from the gauge surface was observed in some areas (Figs. 11(a), 11(b)). Unlike the compressive creep test, grain-boundary oxidation is observed on the parallel surfaces of both aging conditions (Figs. 11(c), 11(d)), which may have been the origin of the fracture. However, any difference in the oxidation resistance between the two aging conditions cannot be determined from these results. In addition, Fig. 12 shows the microstructures focusing on the difference in precipitated phases near the fracture surface. Under the 2-step aging condition, cavities are observed near μ particles at grain boundaries but the fracture tends to occur within grains near grain boundaries rather than at grain boundaries (Fig. 12(a)). At grain boundaries with W₆C/W₁₂C particles, cracks tend to propagate within W₆C/W₁₂C particles (Fig. 12(b)). The fracture under the 2-step aging condition occurs mainly at grain boundaries with W₆C/W₁₂C particles, and the fracture near grain boundaries with μ particles is minor. On the other hand, under the 1-step aging condition, cracks tend to propagate at the interfaces between the matrix and grain-boundary Cr₂₃C₆ or γ’ particles (Fig. 12(c)). In other words, the difference in precipitated phases affects the crack propagation path.

Fig. 11. BEIs of the vertical sections of the tensile crept specimens for (a, c) 2-step aging condition and (b, d) 1-step aging condition at 800°C/409 MPa under air atmosphere: (a, b) near fracture surface and (c, d) near gauge surface.

Fig. 12. BEIs near the fracture surface of the tensile crept specimens tested at 800°C/409 MPa under air atmosphere: (a, b) 2-step aging condition and (c) 1-step aging condition.

3.4. Tensile Creep Behavior under Ar

As described in the previous section, it appears that grain-boundary oxidation at the specimen gauge surface was the starting point of fracture in tensile creep tests in air, but the difference in grain-boundary oxidation resistance under both aging conditions is unclear. Therefore, we attempted to suppress specimen oxidation during tensile creep tests by using Ar as the creep atmosphere. Figure 13 shows the creep rupture life and creep rupture elongation of 800°C/409 MPa tensile creep under Ar, compared to the results of tensile creep tests under air atmosphere. It is clear that oxidation during the test affects tensile creep properties, as both rupture time and creep rupture elongation are higher under Ar as compared to air, for both aging conditions. The increase in rupture time is particularly significant for the 1-step aging condition, for which the rupture times is equivalent to the 2-step aging condition under Ar. However, even under Ar, the creep rupture elongation of the 2-step aging condition is approximately twice that of the 1-step aging condition.

Fig. 13. Creep rupture life and creep rupture elongation of 800°C/409 MPa tensile creep tests under Air and Ar gas atmosphere.

Figure 14 shows the appearance of the specimens after the tensile creep tests. Compared to the 2-step aging condition, the 1-step aging condition shows numerous cracks on the gauge surface (Fig. 14(d)) and an oxidized intergranular fracture surface on the outside of the fracture surface (Fig. 14(f)). This oxidized fracture surface is presumed to be caused by a small amount of oxygen contained in the Ar gas. In any case, it is suggested that the 1-step aging condition is more prone to grain-boundary oxidation during tensile creep tests than the 2-step aging condition.

Fig. 14. Photographs of 800°C/409 MPa tensile creep specimens under Ar atmosphere for (a–c) 2-step aging condition and (d–f) 1-step aging condition: (a, d) gauge surface, (b, e) fracture surface, and (c, f) high-magnification of the black frame in (b, e). (Online version in color.)

4. Discussion

4.1. Effect of Grain-boundary Phases on Compressive Creep Behavior

From the experimental results, it was found that the compressive and tensile creep properties differ when the microstructure within grains is almost the same but only the type of precipitated phase at grain boundaries changes. However, since the average grain-boundary coverage under both aging conditions is different in this study, it is necessary to separately examine whether these differences in creep properties are due to the difference in grain-boundary coverage or to the difference in the type of precipitated phases. Therefore, we first consider whether the compressive creep properties in this study can be explained only by the grain-boundary coverage. The equation for grain-boundary precipitation strengthening is expressed as:⁵⁾

  

$$\dot{\epsilon }/{\dot{\epsilon }}_o=(1-\rho )$$(1)

where $\dot{\epsilon }$ is creep rate, ρ is grain-boundary coverage, ${\dot{\epsilon }}_o$ is creep rate for ρ=0. Since the grain-boundary coverage can change with the progress of creep deformation, the grain-boundary coverage of the pre-creep microstructures (Fig. 2) is used as ρ and the minimum creep rate as $\dot{\epsilon }$, and it is assumed that the grain-boundary coverage at the minimum creep rate is the same with as for the pre-creep condition. By using Eq. (1), the relationship between the minimum compressive creep rate of the 1-step and 2-step aging conditions (${\dot{\epsilon }}_{1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$ and ${\dot{\epsilon }}_{2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$, respectively) and the grain-boundary coverage (ρ_{1 step AG} and ρ_{2 step AG}, respectively) is given as

  

$${\dot{\epsilon }}_{1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}/{\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}=(1-{\rho }_{1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG})$$(2)

  

$${\dot{\epsilon }}_{2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}/{\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}=(1-{\rho }_{2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG})$$(3)

In this study, it was difficult to measure the creep rates for ρ=0 (${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$ and ${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$). However, since the intragranular microstructure (Figs. 2(d), 2(e)), hardness within grains, and average grain size of the two aging conditions are almost the same, it is reasonable to suppose that ${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$ and ${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$, which should be mainly controlled by the strength of grain interior, are nearly identical:

  

$${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}\text{≒}{\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$$(4)

Then, deleting ${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$ and ${\dot{\epsilon }}_{o\mathrm{\hspace{0.17em} }2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}$ using Eqs. (2), (3), (4) and substituting the values ρ_{1 step AG}=0.82, ρ_{2 step AG}=0.67 yields:

  

$$\frac{{\dot{\epsilon }}_{1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}}{{\dot{\epsilon }}_{2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}}=\frac{1-{\rho }_{1\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}}{1-{\rho }_{2\mathrm{\hspace{0.17em} }step\mathrm{\hspace{0.17em} }AG}}\text{≒}0.55$$(5)

Therefore, if the difference of the compressive minimum creep rate between both aging conditions can be solely explained by the grain-boundary coverages, the compressive minimum creep rate for the 1-step aging condition can be predicted from the compressive minimum creep rate for the 2-step aging condition. Figure 15 shows the calculated compressive minimum creep rates in the stress range of 320–550 MPa for the 1-step aging condition using Eq. (5), together with the experimental values. The good agreement between the experimental and calculated values suggests that the difference in compressive minimum creep rate between both aging conditions is due to the difference in grain-boundary coverage rather than to the difference in the nature, size and shape of the grain-boundary precipitation phases. In other words, even if W₆C/W₁₂C or μ phases rather than γ’ or Cr₂₃C₆ phases precipitate at grain boundaries, the compressive minimum creep rate may be reduced if the grain-boundary coverage is large.

Fig. 15. (a) Compressive minimum creep rate/applied stress plots, and (b) relationship between the calculated and experimental values of the creep rates for 1-step aging condition at 800°C.

4.2. Effect of Grain-boundary Phases on Tensile Creep Behavior

In tensile creep tests under air, the creep rupture life is longer at higher stresses in the 2-step aging condition than in the 1-step aging condition, even though the grain-boundary coverage is smaller in the 2-step aging condition (Fig. 9(b)). As these results cannot be explained by the difference of the grain-boundary coverage of both aging conditions, the difference in precipitated grain-boundary phases is dominant. Therefore, we first discuss the factors affecting tensile creep rate. As shown in Figs. 11 and 12, under both aging conditions, tensile creep causes cavity and crack growth at, and near, grain boundaries. Hence, the tensile creep rate ${\dot{\epsilon }}_{ten}$ is assumed to be the sum of the creep rate due to bulk deformation and microstructural changes ${\dot{\epsilon }}_{bulk}$ and the creep rate due to cavity and crack growth ${\dot{\epsilon }}_{cv}$:^21,22)

  

$${\dot{\epsilon }}_{ten}={\dot{\epsilon }}_{bulk}+{\dot{\epsilon }}_{cv}$$(6)

In addition, ${\dot{\epsilon }}_{bulk}$ can be approximated to the compressive creep rate ${\dot{\epsilon }}_{comp}$ as no cavities or cracks were observed in the compressive creep tests:

  

$${\dot{\epsilon }}_{bulk}\text{≒}{\dot{\epsilon }}_{comp}$$(7)

The minimum creep rate in tensile creep is greater than that in compressive creep at higher stresses and is about three times greater at 409 MPa (Fig. 9(a)). In other words, according to Eqs. (6) and (7), the contribution of ${\dot{\epsilon }}_{cv}$ to the overall tensile creep rate ${\dot{\epsilon }}_{ten}$ is expected to be larger than that of ${\dot{\epsilon }}_{bulk}$ at higher stresses. A schematic representation of this mechanism is shown in Fig. 16. Now, in the tensile creep tests, the creep rupture time and rupture elongation increased more under Ar than under air (Fig. 13). Thus, it is clear that oxidation at the specimen surface is one of the factors that affects ${\dot{\epsilon }}_{cv}$. Furthermore, the 1-step aging condition exhibited clear grain-boundary oxidation even under Ar atmosphere with very low oxygen partial pressure (Fig. 14). This suggests that, for tensile creep tests, specimens in the 1-step aging condition with Cr₂₃C₆ and γ’ precipitate at grain boundaries, have higher grain-boundary oxidation and increased ${\dot{\epsilon }}_{cv}$ as compared to specimens in the 2-step aging condition, with W₆C/W₁₂C and μ precipitates at grain boundaries. Since grain-boundary oxidation is not clearly observed in the compressive creep tests (Fig. 8), the grain-boundary oxidation observed in the tensile creep tests (Fig. 11) may be due to Stress Assisted Grain-Boundary Oxidation (SAGBO)²³⁾ or Stress Assisted Intergranular Oxidation Cracking (SAIOC).²⁴⁾ The degradation of creep properties by SAGBO is reported in some Ni-base superalloys,^25,26) and the oxidation rate is reported to be dependent on the applied tensile stress.^27,28) The increase of tensile creep rupture life and rupture elongation under Ar atmosphere compared to under Air (Fig. 13) is the result of the suppression of SAGBO. From the above, it appears that the grain-boundary precipitation of W₆C/W₁₂C and μ phases is more effective in suppressing SAGBO than grain-boundary precipitation of Cr₂₃C₆ and γ’, and this is one of the reasons why the tensile creep rupture life is longer at higher stresses under the 2-step aging condition, which had smaller grain-boundary coverage, than under the 1-step aging condition, which had larger grain-boundary coverage. In addition, considering that the crack propagation path differs depending on the grain-boundary precipitation phase (Fig. 12), ${\dot{\epsilon }}_{cv}$ may also be affected by differences in size, morphology, interfacial strength, and toughness of precipitates. In this study, since multiple grain-boundary phases precipitated under both 2-step and 1-step aging conditions, it was difficult to examine these factors individually, so a detailed study of these factors will be the subject of future work.

Fig. 16. Schematic illustration of creep curve under tensile creep test. (Online version in color.)

5. Conclusions

To clarify the effect of intergranular W-rich compounds (W₆C/W₁₂C and μ phases) on the creep properties of a γ’-precipitation-strengthened, wrought Ni-Co-based superalloy (COWALOY), two microstructures with different intergranular precipitates were prepared by manipulating aging conditions: a microstructure in which W-rich compounds form at grain-boundaries, and a microstructure in which intergranular Cr₂₃C₆ and γ’ phases precipitate, as in conventional Ni-based alloys. The difference in compressive and tensile creep properties between the two microstructures is compared, leading to the following main conclusions:

(1) The average grain-boundary coverages are 67 and 82% for alloys with (i) grain-boundary W₆C/W₁₂C and μ and (ii) grain-boundary Cr₂₃C₆ and γ’, respectively.

(2) In compressive creep tests, lower minimum creep rate and higher 1% creep life were observed for the microstructure with grain-boundary Cr₂₃C₆ and γ’ than for the microstructure with grain-boundary W₆C/W₁₂C and μ.

(3) In tensile creep test under air, longer creep rupture life was observed for the microstructure with grain-boundary W₆C/W₁₂C and μ than for the microstructure with grain-boundary Cr₂₃C₆ and γ’, especially at higher stresses.

(4) In tensile creep test under Ar, longer creep rupture life and higher rupture elongation were observed for both microstructures compared to under Air atmosphere. However, the crept specimen with grain-boundary Cr₂₃C₆ and γ’ phases exhibited oxidized grain-boundary fracture surface at 800°C.

(5) Supposing that the intragranular creep strength of both microstructures is nearly identical, it is suggested that the difference in compressive minimum creep rate between the two microstructures can be explained by the difference in grain-boundary coverage rather than the difference in grain-boundary precipitation phase.

(6) In tensile creep test, the creep rate due to cavity and crack growth appears to be the dominant factor in the overall creep rate at higher stresses, consistent with Stress Assisted Grain-Boundary Oxidation (SAGBO). It appears that the grain-boundary precipitation of W₆C/W₁₂C and μ phases is more effective in suppressing SAGBO than grain-boundary precipitation of Cr₂₃C₆ and γ’.

Supporting Information

The experimental data used for the construction of the TTP diagram (Fig. 1). This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2024-298.

Acknowledgments

This work made use of the MatCI Facility which receives support from the MRSEC Program (NSF DMR-1720139) of the Materials Research Center at Northwestern University; and the EPIC facility of Northwestern University’s NUANCE Center, which has received support from the SHyNE Resource (NSF ECCS-2025633), the IIN, and Northwestern’s MRSEC program (NSF DMR-2308691).

References

• 1)   C.T.  Sims,  N.S.  Stoloff and  W.C.  Hagel: Superalloys II, John Wiley & Sons, New York, NY, (1987), 124. ISBN978-0471011477
• 2)   R.  Tanaka,  M.  Kikuchi,  T.  Matsuo,  S.  Takeda,  H.  Nishikawa,  T.  Ichihara and  M.  Kajihara: Superalloys 1980, TMS, Warrendale, PA, (1980), 481. ISBN978-0871701022
• 3)   M.  Takeyama,  K.  Kawasaki,  T.  Matsuo and  R.  Tanaka: Tetsu-to-Hagané, 72 (1986), 1605 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.72.10_1605
• 4)   T.  Matsuo,  M.  Ueki,  M.  Takeyama and  R.  Tanaka: J. Mater. Sci., 22 (1987), 1901. https://doi.org/10.1007/BF01132914
• 5)   I.  Tarigan,  K.  Kurata,  N.  Takata,  T.  Matsuo and  M.  Takeyama: MRS Online Proceedings Library, 1295 (2011), mrsf10-1295-n06-03. https://doi.org/10.1557/opl.2011.558
• 6)   M.  Takeyama: Materia Japan, 56 (2017), 145 (in Japanese). https://doi.org/10.2320/materia.56.145
• 7)   M.  Takeyama: Materia Japan, 60 (2021), 281 (in Japanese). https://doi.org/10.2320/materia.60.281
• 8)   M.  Takeyama: Proceedings of the AM-EPRI 2024, ASM, Bonita Springs, FL, (2024), 1033. https://doi.org/10.31399/asm.cp.am-epri-2024p1033
• 9)   J.  Zhang,  W.  Chen,  Z.  Cao and  R.  Tanaka: Tetsu-to-Hagané, 75 (1989), 545 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.75.3_545
• 10)   J.S.  Zhang,  P.E.  Li,  W.X.  Chen and  J.Z.  Jin: Scripta Metallurgica, 23 (1989), 547. https://doi.org/10.1016/0036-9748(89)90449-3
• 11)   J.S.  Zhang,  P.E.  Li and  J.Z.  Jin: Acta Metall. Mater., 39 (1991), 3063. https://doi.org/10.1016/0956-7151(91)90039-4
• 12)   T.  Ito,  S.  Yamasaki,  M.  Mitsuhara,  M.  Nishida and  M.  Yonemura: Mater. Trans., 58 (2017), 52. https://doi.org/10.2320/matertrans.M2016291
• 13)   A.M.  Elbatahgy,  T.  Matsuo and  M.  Kikuchi: Tetsu-to-Hagané, 76 (1990), 767 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.76.5_767
• 14)   Y.  Sumi,  K.  Yokota,  M.  Osaki,  H.  Takabayashi,  G.  Toyoda,  T.  Akizuki,  T.  Omori and  K.  Ishida: Trans. JSSE, 63 (2018), 1 (in Japanese). https://doi.org/10.5346/trbane.2018.1
• 15)   M.  Osaki,  H.  Takabayashi,  T.  Omori and  K.  Ishida: Denki-Seiko (Electr. Furn. Steel), 86 (2015), 17 (in Japanese).
• 16)   H.L.  Danflou,  M.  Macia,  T.H.  Sanders and  T.  Khan: Superalloys 1996, TMS, Warrendale, PA, (1996), 119. ISBN978-0873393522
• 17)   R.J.  Mitchell,  H.Y.  Li and  Z.W.  Huang: J. Mater. Process. Tech., 209 (2009), 1011. https://doi.org/10.1016/j.jmatprotec.2008.03.008
• 18)   M.E.  Kassner: Fundamentals of Creep in Metals and Alloys (3rd ed.), Butterworth-Heinemann, Oxford, (2015), 8. https://doi.org/10.1016/B978-0-08-099427-7.00002-5
• 19)   J.C.  Stinville,  N.  Vanderesse,  F.  Bridier,  P.  Bocher and  T.M.  Pollock: Acta Mater., 98 (2015), 29. https://doi.org/10.1016/j.actamat.2015.07.016
• 20)   Z.  Zhang,  Z.  Yang,  S.  Lu,  A.  Harte,  R.  Morana and  M.  Preuss: Nat. Commun., 11 (2020), 4890. https://doi.org/10.1038/s41467-020-18641-z
• 21)   A.H.  Chokshi: Mat. Sci. Tech., 3 (1987), No. 8, 656. https://doi.org/10.1179/mst.1987.3.8.656
• 22)   J.U.  Rakhmonov,  S.  Bahl,  A.  Syam and  D.C.  Dunand: Acta Mater., 228 (2022), 117788. https://doi.org/10.1016/j.actamat.2022.117788
• 23)   R.H.  Bricknell and  D.A.  Woodford: Metall. Trans. A, 12 (1981), 1673. https://doi.org/10.1007/BF02643573
• 24)   M.M.  Morra: PFC/PR-95-7, MIT, (1995).
• 25)   W.  Carpenter,  B.S.-J.  Kang and  K.M.  Chang: Proceedings of the International Symposium on Superalloy 718, 625, 706 and Derivatives, TMS, Pittsburgh, PA, (1997), 679. ISBN978-0873393768
• 26)   J.  Milanese,  E.  Andrieu,  J.B.  Osio,  J.  Alexis and  D.  Bardel: Proceedings of the 9th International Symposium on Superalloy 718 & Derivatives: Energy, Aerospace, and Industrial Applications, TMS, Warrendale, PA, (2018), 805. ISBN978-3030077822
• 27)   H.E.  Evans,  H.Y.  Li and  P.  Bowen: Scr. Mater., 69 (2013), 179. https://doi.org/10.1016/j.scriptamat.2013.03.026
• 28)   K.  Kusabiraki,  K.  Sakuradani and  S.  Saji: Tetsu-to-Hagané, 84 (1998), 291 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.84.4_291