Influence of Replacing Molybdenum with Tungsten on the Creep Fracture Property of Waspaloy Nickel-Based Alloy

Abstract

Alloys meeting the requirements of “700 °C and above” advanced ultra-super-critical technology, with higher thermal efficiency, have been developed in recent years. Here, a new wrought Ni-based superalloy with excellent high-temperature creep strength based on Waspaloy has been developed and is proposed as a candidate material for application in 700 °C class advanced ultra-super-critical steam turbine blades. In this new alloy, the Molybdenum (Mo) in Waspaloy is partially replaced by Tungsten (W). Creep tests have shown that this new Ni-based alloy has a 70 MPa higher creep-rupture strength than that of Waspaloy at 700 °C by extrapolating the experimental data. Detailed creep-rupture mechanisms have been analyzed by means of scanning electron microscopy, transmission electron microscopy, and chemical phase analysis with a view to devising potential approaches for performance improvements. The results showed that the partial replacement of Mo by W had negligible effect on the composition of carbides precipitated in the alloy. Instead, the amount of the γ′ phase was significantly increased, and mismatch between the γ and γ′ phases was reduced. In this way, the stability of the γ′ phase was increased, its coarsening rate was reduced, and its critical shear stress was increased. As a result, the high-temperature creep-fracture strength of the new alloy was increased.

1. Introduction

In recent years, many countries, including China, have begun to focus on the research and development of advanced ultra-super-critical technology with a steam temperature of 700 °C, whereby the thermal efficiency of power stations is expected to break through the 50% threshold. The extremely harsh service conditions of this grade of power station have led to the development of specifically designed superalloys as candidate materials [1,2,3].

As the core components of turbine blades, the main candidate materials in various countries are Waspaloy, Nimonic105, USC141, M252, and so on [4,5]. Waspaloy (Ni-20Cr-13Co-4Mo-1.4Al-3Ti-B-Zr), a precipitation-hardening nickel-based superalloy, was developed by Pratt and Whitney in the United States in the 1950s [6] for utilization in aviation engines and power machines. Such alloys are now essential materials for manufacturing turbine disks and blades, showing good tensile and endurance properties below 760 °C and good oxidation resistance above 870 °C. An advantage of Waspaloy is that it has good strength and toughness matching [7].

Waspaloy is mainly used in aero-engine turbine blade materials. M. Razumovskii et al. [8]. found that increasing the mass fraction of W would greatly improve the melting temperature of the alloy, and the creep-fracture strength would be more advantageous. Alloying element W can improve the heat resistance of the alloy, but also has the properties of being cheap, very cost-effective [9]. The Mo content increases the strength and durability of Waspaloy by solid-solution strengthening. A new type of Ni-based alloy has been designed by our research group by partially replacing Mo in Waspaloy with W. The previous study showed that the type of precipitated phases in Waspaloy alloy would not change, and the alloy maintained good structure stability during long-term aging when adding 2% W (mass fraction) to the alloy. Meanwhile, W could refine the size and reduce the growth rate of γ′ phase, and improved the tensile strength of the alloy [10,11,12,13].

The influence of W on the microstructural stability and properties of such alloys has yet to be fully characterized. Therefore, in this study, we have monitored the effects of long-term aging on carbide separation, precipitation-strengthening phase formation, and the properties of the new alloy. Furthermore, the strengthening mechanism of W substitution of Mo in Waspaloy alloy is discussed.

2. Experimental Material and Procedures

The chemical composition of the test alloy in

Table 1. Mass percentage compositions of alloys (wt%).

Alloy	C	Cr	Co	Al	Ti	W	Mo	Zr	B	Ni
2Mo2W	0.04	19.49	13.56	1.44	2.95	1.97	2.19	0.028	0.01	Bal.
Waspaloy [14,15]	0.056	19.64	13.68	1.40	3.18	-	4.5	0.044	0.0063	Bal.

was determined by thermodynamic calculation and combining with the alloys with different W content in the previous study [9,11]. The new alloy (designated as 2Mo2W) was obtained by replacing Mo in Waspaloy with an equal mass of W. The test alloy was melted by a “double-vacuum” process (vacuum induction smelting + vacuum arc remelting), poured into 50 kg ingots, and subsequently forged into a rod of diameter 18 mm. The forging temperature was 1150 °C, the final forging temperature was not lower than 950 °C, and the cooling method is heap cooling. The heat-treatment process involved the following sequential steps: 1080 °C for 4 h (oil cooling), 845 °C for 24 h (air cooling), and 760 °C for 16 h (air cooling). Thereafter, alloy specimens were exposed to 700 °C for 1000–10,000 h or 750 °C for 1000–5000 h in a resistance furnace, then air-cooled to room temperature.

The above samples were cleaned with a mixture of CuCl₂ (10 g) + HCl (100 mL) + CH₃CH₂OH (100 mL) or electrolytically etched with HNO₃ (40%) + H₂O (60%) solution. This allowed the microstructures of carbides and the γ′ phase to be observed by scanning electron microscopy (SEM, Hitachi S4300, Tokyo, Japan). Qualitative and quantitative analyses of carbides and γ′ phase particles were carried out by X-ray diffraction (XRD, APD-10, Milan, Italy; employing Co-K_α radiation and operated at 30 kV and 25 mA) and chemical phase analysis methods. Equilibrium precipitations in the alloys were investigated by thermodynamic calculations using the Ni-9 database.

Two samples were used for each observation, three samples were used for each performance test, performance is averaged over tests.

The creep-fracture strength was tested on the CMT-105 (100 KN) creep test machine, and the samples were Φ5 mm standard high temperature tensile sample (GB6395-86). The test temperature was 700 °C, and the stresses were 550 MPa, 500 MPa, 475 MPa, 450 MPa, and 425 MPa.

Figure 1 shows the metallographic and scanning microstructure of 2Mo2W and Waspaloy alloys before long-term exposure. The difference of microstructure between the two alloys is not obvious. There are some M₂₃C₆ carbides distributed on the grain boundary. The large precipitations, precipitated in the grain, are unmelted primary MC carbides.

3. Results and Discussion

3.1. Characteristics of the Precipitation after Long-Term Exposure

(1)

Evolution of carbides

Thermal-CALC thermodynamic software was applied to calculate the equilibrium precipitate composition of the new alloy containing 2% W and 2% Mo, as shown in Figure 2. The precipitate comprised small amounts of the M₂₃C₆ phase, MC phase, and µ-phase and a large amount of the γ′ phase. Figure 2b shows an enlarged image of the pertinent region of Figure 2a, in which it can be seen more clearly that the M₂₃C₆ phase precipitated at 920 °C and then remained almost unchanged with increasing temperature. The µ-phase content increased in a parabolic manner as the sample was cooled, while the γ′ phase content increased rapidly.

Thermodynamic calculations can only explain the theoretical equilibrium state, and the actual precipitation process will differ. Figure 3 shows the XRD patterns of the precipitated phase in the alloy before and after long-term aging. From inspection of Figure 3, it can be seen that the carbides in the alloy were M₂₃C₆ and TiC both before and after aging. This indicates that long-term aging at 700 °C or 750 °C has no significant effect on the type of carbides precipitated.

No precipitation of σ, τ, or other harmful phases was observed, indicating that their contents were insignificantly small and any effect on the strength of the alloy could be ignored.

The variations in the contents of M₂₃C₆ and MC in the experimental alloy with aging time at 700 °C are shown in Figure 4. The M₂₃C₆ content increased steadily with aging time from 0 to 3000 h, reached a maximum at around 5000 h, and thereafter remained essentially unchanged. The MC content slowly increased with aging time from 0 to 3000 h, and thereafter showed a slight decrease.

TiC has a high melting point and is usually stable in the range 700–750 °C [16], but during long service of Ni-based alloys, the MC phase will degenerate according to MC + γ→M₂₃C₆ + γ′ or MC + γ→M₆C + γ′. In some high-W alloys, the MC + γ→M₂₃C₆ + γ′ degradation reaction will become MC + γ→ M₂₃C₆ + η [17,18]. The η and M₆C phases are not stable and usually have adverse effects on the creep-fracture strengths of alloys.

The MC phase in the long-term aged alloy and its associated precipitation were analyzed by energy-dispersive spectrometry (EDS), and the results are shown in Figure 5 and

Table 2. Mass percentage compositions of precipitations (wt%).

Precipitation	C K	AlK	WL	TiK	CrK	CoK	NiK	MoL
η	0.04	1.6	1.86	7.17	12.62	11.99	64.72	-
M23C6	2.97	1.49	1.66	3.17	19.03	12.15	56.26	3.27
MC	1.12	0.49	12.35	50.59	8.90	4.65	16.25	6.64

.

Around the large TiC particles, there are two precipitated phases of two components. Compositional analysis showed one to be the M₂₃C₆ phase and the other to be the η phase, rich in Ti [19]. It has been shown previously that, during degradation of the MC phase with high Ti content, the η phase is precipitated. Thus, it can be surmised that the MC phase was degraded here.

(2)

Evolution and coarsening kinetics of the γ′ phase

Figure 6 shows high-resolution Transmission Electron Microscopy (TEM) images of the interface structure between the γ and γ′ phases in the new alloy. From the diffraction pattern in Figure 6b, it can be seen that the γ′ particles and the matrix in Figure 6a are both [110] zone axis. Moreover, it can be seen from the enlargement in Figure 6d of the area within the rectangle in Figure 6c that the interface between the γ and γ′ phases is a perfect co-lattice, with parallel stripes and no defects such as distortions or dislocations.

Ostwald ripening coarsening follows the diffusion-controlled particle growth rule: r³∝t, where r is the particle size and t is time. In Ni-based superalloys, the coarsening rate can be written as r³ − r₀³ = kt, where r is the average size of γ′ after long-term exposure, r₀ is the average γ′ particle size before exposure, and k is the linear coefficient of a plot of r³−r₀³ versus exposure time [20,21].

Table 3. γ′ phase size of test alloys before and after treatment.

Test Alloy (r/nm)	Before Aging	1000 h	2000 h	5000 h	10,000 h
2Mo2W	30.75	32.15	-	28.90	34.50
Waspaloy	31.40	-	31.80	33.50	41.35

shows the size of the γ′ phase of test alloys after long-time aging. Figure 7 compares the coarsening rates of the γ′ phase in 2Mo2W alloy and Waspaloy over 10,000 h at 700 °C. The coarsening rate of the γ′ phase in Waspaloy is k = 5.76, more than twice that of the 2Mo2W alloy (k = 2.47).

3.2. Discussion on the Mechanism of Creep-Fracture Reinforcement

Figure 8 compares the stress–time enduring logarithmic plots for Waspaloy and the 2Mo2W alloy. Waspaloy has a durable strength of about 340 MPa at 10,000 h, around 70 MPa lower than that of the new alloy.

The strengthening modes of the Ni-based alloy include solid-solution strengthening, γ′-phase precipitation strengthening, and grain boundary precipitation strengthening. To probe the reasons for the creep-fracture strength improvement after replacing Mo with W, the strengths, precipitated phases, and carbide compositions of the two alloys were compared in detail. From Figure 9a,b, one can see the tensile strength, yield strength and hardness of the 2Mo2W alloy before and after aging were higher than those of Waspaloy.

The γ′ phase contents of the two alloys prior to aging and after 5000 h and 10,000 h are compared in Figure 9c. The γ′ phase content in the 2Mo2W alloy was significantly higher than that in the Waspaloy. After 10,000 h of aging, the γ′ phase content of the 2Mo2W alloy was about 5% higher than that of Waspaloy. Figure 9d shows that the M₂₃C₆ carbide contents of the two alloys were not significantly different.

Mainly due to the large atomic radius of the added W, additional lattice distortion is generated, which increases the resistance to dislocation motion and increases the solid-solution strengthening effect. Meanwhile, more of the γ′ phase leads to an increased precipitation-strengthening effect, imparting the new 2Mo2W alloy with greater strength.

The model of critical shear stress was used to estimate the precipitation-strengthening effect of the γ′ phase in the Ni-based alloy, which was mainly estimated by the cutting mechanism or the critical shear stress induced by the bypass mechanism when the γ′ phase acts by dislocation.

In the cutting mechanism, for small particles, the weakly coupled dislocation (WCD) model is applicable, whereas for large particles, the strongly coupled dislocation (SCD) model is applicable [22,23].

In both models, the critical shear stress is determined as that required for the edge dislocation pair to cut through the γ′ phase along the <110> direction in the plane (111). In the WCD model, the critical shear stress can be described as:

$$\Delta \tau =\frac{1}{2}A{(\frac{\Gamma }{b})}^{\frac{3}{2}}{(\frac{bdf}{T})}^{\frac{1}{2}}-\frac{1}{2}(\frac{\Gamma }{b})f$$

(1)

where A is the shape factor of the particle, 0.72 is the value of the spherical particles, Γ is the antiphase domain bounding energy of the γ′ phase in the (111) plane, b is the Bergs vector of the edge dislocation in the matrix, and f and d are the volume fraction and diameter of the γ′ phase, respectively. T is the linear tension of the dislocation, which can be calculated as follows:

$$T=0.5G{b}^2$$

(2)

where G is the shear modulus, taken as 68 GPa. In the WCD model, the critical shear stress can be described as:

$$\Delta \tau =\frac{1.72}{2}(\frac{T{f}^{\frac{1}{2}}\omega }{bd}){(\frac{1.28d\Gamma }{\omega T}-1)}^{\frac{1}{2}}$$

(3)

where ω is a constant unit.

In the bypass mechanism (Orowan), the critical shear stress can be described as:

$$\Delta \tau =\frac{Gb}{L}$$

(4)

where L is the spacing of the nearest neighbor atom, which can be calculated by:

$$L=\frac{2\left(1-f\right)d}{3f}$$

(5)

In the calculation, Γ was set at 0.28 J/m² and b was set at 0.254 nm. The volume fraction f was taken as 18% for Waspaloy and 25% for 2Mo2W. The calculated data were substituted into the above three models to calculate the corresponding critical shear stresses, as shown in Figure 10.

Given the diameter d of the γ′ phase particles, the critical shear stresses corresponding to the three models can be calculated. According to the principle of energy minimization, the dislocation preferentially moves in the direction of least resistance. Therefore, for a given value of d, the actual critical shear stress is the minimum in the three models.

As can be seen from Figure 10, both alloys are dominated by the WCD model when the particles are at their initial size. Figure 10a shows that when 2Mo2W particles grow to about 35 nm, the dislocation bypass model plays a dominant role, and the critical shear stress at this stage is 0.355 MPa. Figure 10b shows that Waspaloy is similar to 2Mo2W, with only two mechanisms of strong coupling, the dislocation model and the Orowan dislocation bypass model, and the critical shear stress is 0.208 MPa.

Compared with the above results, for the same particle radius of the γ′ phase, the critical shear stress of 2Mo2W is larger and the strengthening effect is better. This explains why the increase of the γ′ phase content is the main reason for the high creep-fracture strength of 2Mo2W.

The stability of the γ′ phase can also affect the creep-fracture performance of Ni-based alloys [24], and the γ′ phase coarsening rate and the mismatch degree of the γ′ and γ phases (matrix) of the 2Mo2W alloy and Waspaloy were compared.

Small lattice mismatch is beneficial to the creep-fracture properties of the alloy, which is expressed by δ, defined as follows:

$$\delta =2\frac{{\alpha }_{{\gamma }^{\prime }}-{\alpha }_{\gamma }}{{\alpha }_{{\gamma }^{\prime }}+{\alpha }_{\gamma }}$$

(6)

Figure 11 shows the results of dynamics software calculations of the mismatch degrees of the γ and γ′ phases in alloys with different W and Mo contents. It can be seen that an alloy with 4% W and no Mo showed the greatest degree of mismatch, followed by an alloy with 4.5% Mo and no W. The least mismatch was seen for the new alloy with 2% W and 2% Mo. This corroborates that the compound addition of Mo and W is conducive to reducing the mismatch of the γ and γ′ phases, stabilizing the γ′ phase, consistent with the finding that the coarsening rate of the γ′ phase of 2Mo2W alloy is lower than that of Waspaloy.

4. Conclusions

A new alloy, designated as 2Mo2W, in which Mo is partially replaced by W, is proposed. During long-term heating of this alloy at 700 °C, the main precipitates that separate have been identified as M₂₃C₆ and MC. The creep-fracture strength of 2Mo2W is higher than that of Waspaloy after 10,000 h, which can be mainly attributed to the compound addition of W and Mo, more of the precipitation-strengthening γ′ phase, and reduced mismatch of the γ and γ′ phases. This renders the γ′ phase more stable and lowers the coarsening rate during long-term service at 700 °C. Since the atomic radius of W is larger than that of Mo, its addition reduces the diffusion rates of other alloying elements, increases the stability of the microstructure, and thus improves the creep-fracture strength of the alloy.