Fig. 1 provides typical optical micrographs of the as-cast alloys. The disorder distribution of dendritic microstructure exhibits a typical equiaxed grain microstructure due to three-dimensional temperature field heating during the solidification. The dendrite core (DC), second dendrite arm (SDA) and interdendritic region (IR) are clearly observed. DC and SDA show higher lightness and IR displays lower lightness. By using an energy dispersive X-ray spectrometer (EDS) of SEM, it is confirmed that refractory elements such as W and Mo were enriched in DC. The magnitude of segregation of alloying element between DC and IR can be better characterized by average segregation coefficient (k_s). To ensure statistically significant mean values, an adequate number of measurements (at least ten) were taken within DC and IR. The average segregation coefficient (k_s) can be defined in Eq. (1). The k_s > 1 or k_s < 1 represents the element partition preferably to DC or IR, namely negative segregation or positive segregation. This method is straightforward and is generally accepted^{[34, 35, 36].}

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. Dendrite morphology of as-cast alloys: (a) 0Ru alloy; (b) 1Ru alloy; (c) 3Ru alloy; (d) 6Ru alloy.

k_s=C_DC/C_IR(1)

where k_s is the average segregation coefficient; C_DC is the chemical composition in the dendrite core area; and C_IR is the chemical composition in the interdendritic region.

Fig. 2 highlights the average segregation coefficient of alloying elements in the as-cast alloys. Fig. 2 shows that Co, Mo and W tend to segregate toward the dendrite core region during solidification in Ru-free alloy and display a negative segregation character. In other words, some phases rich in Co, Mo and W elements solidified first within the cores during the casting process. In contrast, Al, Ti and Ta elements present a positive segregation character in Ru-free alloy. These elements existed in greater quantity in the last liquid phase to solidify and result in large eutectic γ /γ ′ phases within the interdendritic regions, as shown in Fig. 3. The k_s values of Cr and Ni elements in 0Ru alloy are close to unity.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. Average segregation coefficients of alloying elements in the as-cast alloys.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. SEM images of the as-cast alloys: (a) 0Ru alloy; (b) 1Ru alloy; (c) 3Ru alloy; (d) 6Ru alloy.

With the addition of Ru, the segregation characteristic of Mo has changed from negative segregation element to positive segregation element and that of Cr becomes an obvious positive segregation characteristic. And the segregation characteristic of other alloying elements is unchanged. Ru represents a positive segregation characteristic, which is contrary to recent investigations^{[21, 37]}. But Li et al. had the idea that Ru was mainly enriched in interdendritic region during the solidification process of Ni₃Al single crystal alloy^[38]. With increasing Ru, k_s of W obviously increases and that of Al element slightly increases, which is not in agreement with the result of some reports that in the case of the Ru addition, only little or contrary influence on the segregation distribution coefficient of W or Al^{[3, 31, 39, 40]}. For an example, Liu et al. reported that the tendency for Al segregation to IR increases with the increase of Ru in bearing higher Re SC superalloys^[40]. The reason is possibly due to an addition of 0.5Ti wt% in our test alloys. In terms of alloy materials, it is worth noting that Ti probably plays an important role in controlling the partitioning behavior of Re and W^{[31, 41]}. Addition of Ti might be useful to reduce the nominal overall W concentration while maintaining a similar W content in the γ matrix. Furthermore, Kearsey et al. noted that Ru additions effectively reduced the severity of micro-segregation that occurs due to heavy addition of Re and high total refractory levels^[22]. To our experimental alloys, without Re, the lower total refractory elements (W + Ta + Co = 18.5 (wt%)) and high content Cr (12wt%) are of an inherent difference.

The degree of segregation of Co, Al and Ru decreased and that of Ti, Mo, Ta and W increased with the increase of Ru content. It demonstrates that Ru has a strong effect on the segregation behavior of other alloying elements in experimental alloys during solidification. But several researchers approved that Ru did not induce remarkable effect on the segregation behavior of other alloying elements in superalloys and it is important to note that these alloying investigations considered a relatively limited range of composition, with alloys typically containing lower Cr and higher Re^{[26, 29, 30, 31]}. Furthermore, the variation degree of segregation of Co, Cr and Ni is similar since the atomic radii of those elements are nearly equivalent^[42].

All these results suggest that segregation behaviors of alloying elements are very complex in multicomponent Ni-based superalloys over a broad range of compositions. As a result, the influence of Ru on segregation behavior of alloying elements in high Cr Re-free equiaxed Ni-based superalloys exhibits distinct characteristics.

Fig. 3 exhibits typical microstructure of as-cast alloys by SEM. The microstructures of four as-cast alloys are composed of γ , γ ′ , γ /γ ′ eutectic and carbide. The type of carbide is MC type by EDS analysis. The area percentages of carbides of the as-cast alloys are 0.057 ± 0.0041% (6Ru alloy), 0.028 ± 0.0021% (3Ru alloy) and 0.031 ± 0.0074% (1Ru alloy), respectively. The Ru-free alloy contains very little carbide due to serious carbon burning during casting process (as shown in Table 1) and such small quantities could not be easily detected. Addition of Ru has no effect on the phase composes of as-cast alloys. The area fractions of γ /γ ′ eutectic of as-cast alloys are 0.8 ± 0.014% (0Ru), 1.13 ± 0.018% (1Ru), 1.52 ± 0.012 (3Ru) and 2.12 ± 0.016% (6Ru), respectively. The so-called eutectic fraction displays the remaining liquid at the last stage of solidification^[31]. Without Ru addition, the amount of remaining liquid at the end of solidification is marginally lower. With the increase of Ru, the area fraction of γ /γ ′ eutectic increases, which is different from the result of Shi et al.^[21] but similar to recent studies^{[31, 40]}. Although with addition of Ru, Al concentration in interdendritic region is slightly less, that of Ti and Ta strongly tends to enrich in this area (as shown in Fig. 2). As a consequence, an increase in the amount of eutectic occurs.

Fig. 4 is the γ ′ morphology of as-cast alloys. The primary γ ′ phase of as-cast samples shows butterfly and the size of γ ′ in IR is larger than that in DC. The influence of Ru on the size and the volume fraction of primary γ ′ of as-cast alloys is illustrated in Fig. 5. In light of the image analysis results, the γ ′ size and its volume fraction in DC and IR decrease, the γ ′ shape is more regular and the distribution of γ ′ particles in DC is more uniform with increasing Ru (as shown in Fig. 4), which is different from the reported results in the literature^{[43] and [44]} and consistent with the study of Liu et al. ^{[45, 46]}. This effect may not result from modified alloying element fractions since both elements were added at the expense of Ni and all other alloying elements were kept as constant (as shown in Table 1 ). Thus, the decreasing volume fraction and size of γ ′ can be an effect of the Ru-addition itself. One possible suggestion is the reduction in the γ ′ volume fraction by addition of Ru, accompanied by decreased γ supersaturation due to a higher γ volume fraction^[25].

Fig. 4

	Figure Option ViewDownloadNew Window
	Fig. 4 γ ′ morphology of as-cast alloys in dendrite core (DC) and interdendritic region (IR): (a) DC in 0Ru alloy; (b) IR in 0Ru alloy; (c) DC in 1 Ru alloy; (d) IR in 1 Ru alloy; (e) DC in 3 Ru alloy; (f) IR in 3 Ru alloy; (g) DC in 6 Ru alloy; (h) IR in 6 Ru alloy.

Fig. 5

	Figure Option ViewDownloadNew Window
	Fig. 5 Size and volume fraction of γ ′ phase in as-cast alloys with various Ru content.

Fig. 6 displays morphologies of heat-treated alloys. After heat treatment, the primary γ ′ and γ /γ ′ eutectic dissolved completely. It is very difficult to distinguish DC and IR, which presents the uneven distribution of all elements between DC and IR after heat treatment. The phase of four heated alloys is composed of γ , γ ′ and carbide. The type of carbide has no change. Fig. 7 illustrates alloying element distribution plotted vs distance between the adjacent grains of the heat-treated alloys. Cr, Co and W elements display relatively large variation and Ta shows intermediate variation as well as Al, Mo, Ti and Ru nearly keep constant. It reveals that most of alloying elements present very uniform distribution after heat treatment. It is interesting to note that the content of Cr and Co is higher than their average content and that of W and Ta is lower than their average value. It can be attributed to a discrepancy of standard samples of EPMA.

Fig. 6

	Figure Option ViewDownloadNew Window
	Fig. 6 SEM morphology of four heat-treated alloys: (a) 0Ru alloy; (b) 1Ru alloy; (c) 3Ru alloy; (d) 6Ru alloy. The heat treatment process is follow: 1270 ° C/2 h + 1290 ° C/2 h, air cooling (AC) + 1120 ° C /2 h, AC + 900 ° C/2 h, AC.

Fig. 7

	Figure Option ViewDownloadNew Window
	Fig. 7 Measured alloying element distribution plotted vs distance between the adjacent grains according to constant interval distance to select tested point with by EPMA: (a) 0Ru alloy; (b) 1Ru alloy; (c) 3Ru alloy; (d) 6Ru alloy.

Fig. 8 compares the γ ′ morphology of the heat-treated alloys. The morphology of γ ′ has been noticed to be more cuboidal in Ru bearing alloy as compared to Ru-free alloy and the degree of cubic of γ ′ with the evolution of Ru addition. When the content of Ru reaches 6 wt%, the cuboidal degree of γ ′ becomes the optimum. It can be observed that smaller γ ′ exists between larger γ ′ phases and small quantity of uneven planes can be also found due to spalling of partial larger γ ′ . Thus, using electrode chemical etched samples to evaluate the volume fraction of γ ′ easily leads to bigger discrepancy. In this study, we analyzed the volume fraction of γ ′ using chemical etched samples due to the accuracy of the method.

Fig. 8

	Figure Option ViewDownloadNew Window
	Fig. 8 γ ′ morphology of heat-treated alloys: (a) 0Ru alloy; (b) 1Ru alloy; (c) 3Ru alloy; (d) 6Ru alloy. The heat treatment process is as follows: 1270 ° C/2 h + 1290 ° C/2 h, air cooling (AC) + 1120 ° C/2 h, AC + 900 ° C/2 h, AC.

Fig. 9 reveals the variation of the size and the volume fraction of γ ′ in the heat-treated alloys with the content of Ru. With the addition of Ru, the volume fraction and size of γ ′ in heat-treated alloys decrease and the case is the same as that in the as-cast alloys. The morphology of γ ′ particles depends on both the interfacial energy and elastic strain energy, which are related to the interfacial area and volume of precipitates, respectively^[47]. It is well accepted that the composition of γ and γ ′ phases controls the sign and magnitude of the lattice misfit, which influences the morphology of γ ′ precipitates^[48]. The difference in the γ /γ ′ partitioning coefficients could give a reasonable explanation for why there is a decreasing γ ′ volume fraction with increasing Ru content. Thus, it is very important to understand the γ /γ ′ partitioning coefficient of all alloying elements and the γ /γ ′ lattice misfit δ of alloys. Normally, the γ /γ ′ partitioning coefficient of alloying elements can be detected by three dimensional atomic probing (3DAP) or using EPMA to measure the chemical composition of γ and γ ′ after growing treatment (the size of γ /γ ′ is at least 1 µ m, which can minimize the matrix effect ^[49]). As to δ , it can be calculated with different model equations or detected by TEM or XRD that concludes synchrotron radiation ^[50]. It has been reported that Ru partitioned preferentially to the matrix (γ ) resulting in more negative lattice misfit (δ ) between γ ′ and γ ^{[10, 51, 52, 53]}, an effect that can be attributed to the low diffusion coefficient of Ru and the increased γ /γ ′ lattice misfit^{[54, 55, 56, 57]}. It can be seen from Fig. 8 that with increasing Ru content, the morphology of γ ′ particles evolved from spherical in 0Ru alloy and 1Ru alloy to intermediately shaped in 3Ru alloy, to cuboidal in the 6Ru alloy, which reveals that the addition of Ru prompts the lattice misfit becoming more negative. In the Ru-free alloy, γ ′ precipitates as sphere, so the lattice elastic energy is deduced to be negligible, which gives a small value of the lattice misfit between γ ′ and γ . With the increase of Ru content, the lattice parameter of γ would increase due to larger atom radius of Ru incorporated into γ phase^{[2, 58]}; as a consequence, the lattice misfit would become bigger, resulting in the increase of lattice elastic energy.

Fig. 9

	Figure Option ViewDownloadNew Window
	Fig. 9 Size and volume fraction of γ ′ phase in various Ru content heat-treated alloys.

For the alloy with smaller magnitude of lattice misfit, interfacial energy predominates and spherical γ ′ particles are shown due to their minimum surface area with the same volume. Inversely, for the alloy with larger magnitude of lattice misfit, elastic strain energy induced by lattice misfit predominates and cuboidal γ ′ particles are exhibited. The cuboidal degree of γ ′ increases with the increase of lattice misfit. Furthermore, in the characteristics of cuboidal γ ′ precipitation, the degree of parallelism of γ /γ ′ interfaces is improved with the increase of magnitude of lattice misfit (Fig. 8). Consequently, the distribution of γ ′ is the most uniform in 6Ru alloy.

In accordance with Heckl et al.^[44], the γ ′ -solubility of the typical γ ′ forming elements (Al, Ti, Ta and Ni) is lowered by the addition of Ru and the γ solubility is increased, which illustrates that the tendency for supersaturation and precipitation of γ ′ in the matrix is lowered. Thus, with the addition of Ru, decreasing the available amount of Ni for the γ ′ -formation in alloy leads to a lower volume fraction of γ ′ in alloys.

The effect of Ru addition on thermo-physical properties and high-temperature transformations was investigated by DSC instrument. The DSC heating curves of four alloys are displayed in Fig. 10. The initial γ ′ solvus temperature of alloy (T_{γ ′ solvus}) is defined as the first endothermic peak. The γ /γ ′ eutectic transformation temperature of alloy (T_{γ /γ ′ eutectic}) is identified as the second endothermic reaction. The liquidus temperature of alloy (T_Liquidus) is named as the peak of the maximum temperature difference. The solidus temperature of alloy (T_Solidus) is described by extrapolation of the onset temperature of the liquidus. The range of melting temperature (Δ T) has a difference between T_Liquidus and T_Solidus. Above defined temperatures are marked in Fig. 10.

Fig. 10

	Figure Option ViewDownloadNew Window
	Fig. 10 DSC heating curves of as-cast alloys: (a) 0Ru; (b) 1Ru; (c) 3Ru; (d) 6Ru. Tγ ′ — The γ ′ solvus temperature, Tγ /γ ′ — The γ /γ ′ eutectic transformation temperature, TL— The liquidus temperature, TS— The solidus temperature, Δ T— The range of melting temperature.

It can be seen from the heating curves of the alloys that Ru addition decreases the initial γ ′ solvus temperature and solidus of alloys, but has a negligible effect on the liquidus temperature of alloy. This result is different from the results of some other reports^{[3, 32, 59]}. Feng et al. and Chen et al. found that addition of Ru enhanced the solidus and liquidus temperatures of Ni-based SC superalloys^{[3, 59]}, and Cui et al. found that Ru only increased the liquidus temperature in Ni-based disc alloys^[32]. On the contrary, Shi et al. observed that addition Ru can reduce the solidus and liquidus temperatures^[21]. Furthermore, Zheng et al. investigated the solidification behavior of Ru-containing Ni-based Re-free superalloy with lower Cr high W and found that addition of Ru decreased the precipitation temperature of γ /γ ′ eutectic and secondary γ ′ in alloy and had no effect on the temperature of solidus and liquidus^[60], and Ai et al. maintained that the addition of Ru tended to lower the solidus and liquidus temperatures and to increase freezing range, but the influence of Ru was negligible^[39]. The difference between γ ′ solvus and the solidus is considered to be the solution treatment “ window” ^[55]. As seen in Fig. 10, these alloys typically have at least a 90 ° C solution treatment window and the window increases with the increase of Ru. Microscopy analysis confirmed that all alloys were completely homogenized by the chosen heat treatments.

The Δ T of the alloy with Ru addition is wider than that of Ru-free alloy, which is in agreement with other experimental data ^[31]. In particular, from Fig. 10(d), an obvious γ /γ ′ eutectic endothermic peak can be seen and there are no findings in other heating curves (as shown in Fig. 10(a-c)). It indicates that the quantity of γ /γ ′ eutectic in 6Ru alloy is much more than that of other alloys. Compared with our experimental alloys, most studies have investigated the Re-containing SC superalloys^{[3, 39, 59]}. Different results can be attributed to the multicomponent of superalloys. It reveals that different superalloy systems show different solidification characteristic.

Fig. 11 represents DSC cooling curves of four alloys. During cooling, the temperature of initial solidification (T_i) is defined as the end temperature of overall liquid. The temperature of final solidification (T_f) is represented as the onset temperature of overall solid. The temperature of forming γ /γ ′ eutectic (T_{p γ /γ ′ eutectic}) is determined as an endothermic reaction near final solidification region and the starting point of the eutectic solidification was set as the first appearance of the eutectic γ ′ phase. The eutectic temperature and eutectic fraction were taken from the solidification curve at this point. In previous works^{[31, 61, 62]}, the eutectic fraction was evaluated by calculating the eutectic solidification enthalpy in proportional to the overall released enthalpy from DSC cooling curves of alloys. The temperature of precipitating γ ′ from solid (T_{p γ ′} ) is named as the temperature of last endothermic peak. The range freezing temperature (Δ T^′ ) is a difference between T_i and T_f. All defined temperatures are illustrated in Fig. 11. It can be seen from Fig. 11 that Ru has a negligible effect on T_i but has pronounced influence on T_f, T_{p γ ′} , T_{p γ /γ ′ eutectic} and Δ T^′ . With the addition of Ru, T_f and T_{p γ ′} are lowered and Δ T^′ is widened. For Ru-free alloy, there is no finding of any γ /γ ′ eutectic endothermic reaction due to tiny γ /γ ′ eutectic phase quantity in the alloy. With the increase of Ru, the value of T_{p γ /γ ′ eutectic} changes initially from 1324 ° C (1Ru alloy) to 1321 ° C (3Ru alloy) and then increases to 1326 ° C (6Ru alloy). It is interesting to note that the reaction of precipitating carbide can be seen in the DSC cooling curve of 1Ru alloy, and the temperature of precipitating carbide (T_Carbide) is approximately 1338 ° C, which is not presented in other alloys. It is possible that the DSC sample of 1Ru contains relative high carbon content. According to previous works^{[31, 63]}, the concentration C_DC or C_IC of the first solid formation can be assumed to be controlled by the liquid diffusion coefficient. An increase in the freezing range with the addition of Ru provokes more time for liquid diffusion, which in turn leads to an increase in W microsegregation into the dendrite core due to its solubility limit. As a result, the value k_s of W element increases with the addition of Ru. The calculated eutectic fractions are 0% (0Ru), 9.54% (1Ru), 10.36% (3Ru) and 12.54% (6Ru), respectively. The agreement between the calculated and experimentally derived values of eutectic volume fraction is rather poor. It is believed that there are other low-melting-point phases in equiaxed superalloys besides γ /γ ′ eutectic, for example boride. But, with the addition of Ru, the change of eutectic fraction shows similar trend. Table 2 concludes the results of DSC.

Fig. 11

	Figure Option ViewDownloadNew Window
	Fig. 11 DSC cooling curves of as-cast alloys: (a) 0Ru; (b) 1Ru; (c) 3Ru; (d) 6Ru. Ti— The temperature of initial solidification, Tf— The temperature of final solidification, Tp γ /γ ′ eutectic— The temperature of forming γ /γ ′ eutectic, Tp γ ′ — The temperature of precipitating γ ′ from solid, Δ T′ — The range freezing temperature, TCarbide— The temperature of precipitating carbide.

Table 2.

Table 2.

Table 2. Results of DSC heating and cooling curves of as-cast alloys (° C) Alloys Tγ ′ Tγ /γ ′ TL TS Δ T Ti Tf Tp γ /γ ′ eutectic Tp γ ′ Δ T′ TCarbide 0Ru 1243 N/A 1377 1336 41 1376 1300 N/A 1142 76 N/A 1Ru 1209 N/A 1378 1333 45 1376 1297 1324 1139 79 1338 3Ru 1203 N/A 1328 1378 50 1376 1295 1321 1132 81 N/A 6Ru 1157 1289 1326 1376 50 1377 1290 1326 1125 87 N/A Note: N/A: no detection.

Table 2. Results of DSC heating and cooling curves of as-cast alloys (° C)

The densities of as-cast and heat-treated alloys were determined using the ‘ Archimedes-method’ in deionized water at 23.5 ° C. Fig. 12 shows the densities of four alloys under different conditions. It can be seen from Fig. 12 that the density of alloys increases with the increase of Ru. One reason is that the density of Ru element (12.45 g/cm³) is larger than that of Ni (8.88 g/cm³).

Fig. 12

	Figure Option ViewDownloadNew Window
	Fig. 12 Densities of alloys vs Ru content.

As we know, the quantity of defects in alloys has a closer relationship with the density of alloy. Fig. 13 displays the area fraction of porosity of as-cast and heat-treated alloys. It exhibits that the area fraction of porosity decreases with the addition of Ru, which suggests that Ru can reduce the formation of porosity in alloys. This result is in agreement with several investigations^{[15, 22, 64]}. Thus, the other reason of improving density by Ru addition is that Ru addition is effective in suppressing the tendency of grain defect formation. The size of porosities of the as-cast and heat-treated alloys is presented in Fig. 14. The size of porosities in the as-cast alloys is not more than 2 µ m and decreased with the increase of Ru. After heat treatment, the size of porosities is larger than that in the as-cast alloys and the change trend with increasing Ru is the same as that in the as-cast alloys. It reveals that Ru addition suppresses the formation of porosity and the reason is that Ru strongly affects the segregation behavior of refractory alloying elements, which possibly lowers the density inversion between the solute and bulk liquid in mushy zone^[63] and thus reduces the driving force for grain defect formation. Although the value k_s of W in Fig. 2 increases with the addition of Ru, Liu et al. found that the partition coefficients from point matrix scanning technique are quite different from the directional solidification followed quenching (DSQ) technique due to solid backdiffusion^[63]. They found that the partition coefficient of W slightly decreased from 1.23 (Ru-free alloy) to 1.22 (Ru-containing alloy) using DSQ technique and clearly increased from 1.26 (Ru-free alloy) to 1.37 (Ru-containing alloy) using point matrix scanning technique^[63]. It is a future work to use DSQ technique to measure test alloys.

Fig. 13

	Figure Option ViewDownloadNew Window
	Fig. 13 Area fraction of porosity vs Ru content.

Fig. 14

	Figure Option ViewDownloadNew Window
	Fig. 14 Size of porosity in as-cast and heat-treated alloys vs Ru content.

Furthermore, it is interestingly noted that the area fraction and size of porosity in the as-cast alloys is lower than that of heat-treated alloys. The reason could be dissolving of γ /γ ′ eutectic after solution treatment, which leads to form much micro pore. As a result, the density of the as-cast alloy is higher than that of heat-treated alloys as shown in Fig. 12.

This is confirmed by the hardness of four alloys at different conditions in Fig. 15.With increasing Ru, the hardness of alloys increases. Several literatures presented similar result that the hardness of γ and γ ′ phases increased with the addition of Ru^[45]. They thought that the Ru additions appear to provide a certain amount of solid solution strengthening to both the γ and γ ′ phases^{[48, 49, 54]}. Since the element Ru also has a hexagonal close-packed structure (HCP) structure, its addition would be expected to decrease the stacking fault energy of the matrix^[2]. The a/2[110] dislocation motion in the matrix channel leaves interfacial dislocations around cuboidal precipitates and forms homogeneous γ /γ ′ interfacial dislocation networks ^{[65, 66]}; the addition of Ru extends the width between the partial dislocations in the matrix^[1] and would inhibit the cross slip of dislocation at this temperature. Addition of Ru reduced the volume fraction of γ ′ as shown in Fig. 5 and Fig. 9 and a low γ ′ volume fraction also widens the width of the γ channels between the γ ′ precipitates. Intrinsic reason could be the fact that Ru addition promotes larger lattice distortion, and then the hardness of alloy increases^[67]. As mentioned above, Ru addition improves the hardness of experimental alloys. For the same alloy, the hardness of the as-cast alloy is bigger than that of heat-treated alloy, so, the degree supersaturation of the as-cast alloy is larger than that of heat-treated alloy.

Fig. 15

	Figure Option ViewDownloadNew Window
	Fig. 15 Vickers hardness vs Ru content.