Study on Ni₃Al-Based Single Crystal Superalloy Joints Brazed by Vacuum Brazing with Zr-Containing Filler

Abstract

Melting point depressants (MPDs) are required to lower the melting point of filler for brazing. In this study, Zr was used as the MPD, and powder filler was prepared by adjusting the Zr and Mo content referring to Thermo-Calc calculations. The prepared filler was used to braze a high-Mo Ni₃Al-based single crystal superalloy, IC21, for 1200 °C/30 min. The effects of adjusting the Zr and Mo content on the microstructure and tensile properties of the joint were investigated. The increase in Zr content promotes the formation of Ni₇Zr₂ in the joint, leading to a decrease in the tensile strength of the joint. The increase in Mo content forms diffusion barriers between the BM and filler, resulting in an enhancement in the tensile strength of the joint. However, continued increases in Mo content leads to an increase in the P-topologically close packed phase, causing a decline in the tensile strength of the joint. When the Zr content was (11.8–12.2) wt.% and the Mo content was (7.3–7.7) wt.%, the tensile strength of the joint at 980 °C reached a maximum of 550 MPa. This study provides a potential direction for the design of brazing filler composition for high-Mo Ni₃Al-based superalloys.

1. Introduction

The Ni₃Al-based [1,2,3,4,5] single crystal superalloy is a newly developed high-temperature structural material utilized for turbine blades. Ni₃Al-based superalloys have a γ+γ′ two-phase structure [6] similar to Ni-based superalloys [7,8,9], but with the γ′ phase serving as the matrix in the Ni₃Al-based superalloys. Ni₃Al-based alloys consist of γ phase and a high-volume fraction of γ′ phase for precipitation strengthening, and contain Mo, Cr, Ta, and Re elements for solid solution strengthening [10]. The IC21 alloy investigated in the present study represents a new generation of Ni₃Al-based single crystal superalloys [11,12] characterized by elevated Mo concentrations. Replacing W and Re [13] with Mo in IC21 can reduce the cost and density [14] effectively.

Due to the complex configuration of aero-engine structures, welding emerges as a frequently utilized method for material joining [15,16]. Brazing represents a proficient technique for high-temperature alloy welding [17]. Vacuum brazing [18], a variant of brazing, involves heating both the filler material and the components to be joined to the welding temperature within a vacuum environment [19,20,21]. This process facilitates the melting of the filler material, which subsequently diffuses into the workpiece at the welding interface, thereby generating a dense and dependable joint. The research on Ni-based superalloy solder mainly focuses on using B and Si as melting point depressants (MPDs) to reduce the melting point of the filler [22,23,24,25,26,27,28,29,30], so that the filler can melt and fill the weld below the solid phase line of the base metal (BM) alloy. In this process, B tends to form a B compound phase in the weld area [31,32], while Si predominantly concentrates in the central region of the joint, forming a silicide phase [17]. This segregation adversely affects the homogeneity of the joint structure, thereby promoting crack initiation and propagation [33]. At the same time, B is easily diffused into the BM due to its small atomic radius and rapid diffusion rate [34]. A significant amount of B enters the BM to form B compounds, resulting in higher microhardness in the diffusion-affected zone compared to the weld area [35]. Moreover, the formed B compounds are prone to becoming initiation sites for cracks [36]. Therefore, it is recommended that alternative elements to B and Si be explored with the aim of reducing the melting point of the filler material.

Yajie Du et al. [37] used a Ti-Zr-Cu-Ni filler to prepare a Ti₂AlNb/TC17 joint and found that the addition of Zr could form a eutectic phase with Ni, thereby reducing the melting point of the filler. V.F. Khorunov et al. [38] used Ni-Cr-Zr solder to braze Ni-based alloys and found that compared to Ni-Cr-B-Si system brazing alloys, Ni-Cr-Zr solder exhibits higher adhesive activity, greater ductility, and lower hardness. Moreover, it has been observed that Zr exhibits excellent corrosion resistance and has the capability to form various solid solutions with non-metals and diverse metals, thereby serving as a secondary phase strengthening agent in the joint [39,40]. Therefore, it is feasible to incorporate Zr into Ni₃Al-based superalloy solder to reduce its melting point.

This study focuses on a high-Mo Ni₃Al-based single-crystal superalloy, designated as IC21, where Zr is utilized as an MPD. The microstructure of the joint is adjusted by varying the concentrations of Zr and Mo. Tensile strength testing of the joints is performed at 980 °C. The fracture behavior of the joints is analyzed through observation of the joint fractures.

2. Experimental Procedures

2.1. Filler Preparation and Test

The filler composition is designed according to the IC21 composition. The phase diagrams, chemical potentials, and diffusion coefficients are calculated using the thermodynamic calculation software Thermo-Calc 2022a with the database TCNi 10. The design filler composition and alloy composition are shown in

Table 1. The nominal composition of fillers and the base metal IC21 (wt.%).

Sample	Re	Mo	Al	Ta	Cr	Zr	Ni
BNiZr-1		4.8–5.2	7.0–7.8	2.8–3.2	4.7–5.3	11.8–12.2	Bal.
BNiZr-2		4.8–5.2	7.0–7.8	2.8–3.2	4.7–5.3	13.8–14.2	Bal.
BNiZr-3		4.8–5.2	7.0–7.8	2.8–3.2	4.7–5.3	15.8–16.2	Bal.
BNiZr-4		7.3–7.7	7.0–7.8	2.8–3.2	4.7–5.3	11.8–12.2	Bal.
BNiZr-5		9.8–10.2	7.0–7.8	2.8–3.2	4.7–5.3	11.8–12.2	Bal.
IC21	1.2–1.8	9.0–10.0	7.0–8.0	2.5–3.5	1.2–1.8		Bal.

. During the filler design process, the expensive element Re was removed, while the Cr content was augmented to enhance the filler oxidation resistance. Additionally, the element Zr was introduced to lower the filler melting point. BNiZr-4 and BNiZr-5 adjusted the element Mo on the basis of BNiZr-1. The filler powder was prepared by argon atomization. The phase components of fillers were analyzed by X-ray diffraction (XRD, Bruker D8 DISCOVER, Karlsruhe, Germany). Differential scanning calorimeter (DSC, NETZSCH STA 449F5, Selb, Germany) tests were conducted on five fillers to determine their liquidus temperature, which was crucial for establishing the brazing temperature. The heating rate during the DSC tests was set at 10 K/min, and Ar gas was used to prevent oxidation during the experiment.

The heating curves of filler powders tested by DSC are shown in Figure 1. As can be seen from Figure 1, the main heat absorption peaks of fillers all end below 1160 °C. In order to ensure that the filler exhibits excellent fluidity during brazing, the brazing temperature is set at 1200 °C, approximately 40 °C above the end of the main heat absorption peak. Additionally, the brazing time is set to 30 min.

2.2. Joint Preparation and Test

The IC21 alloy was cut to 12 × 6.5 × 25 (mm) using wire cut electrical discharge machining (WEDM) before brazing, and the face of 12 × 6.5 (mm) was the brazing surface (Figure 2a). After cutting, the brazing surface was polished using 600-grit sandpaper and ultrasonically cleaned in acetone for 10 min to remove the oil on the workpiece. When assembling, nickel wire was utilized to maintain the weld gap at 50 μm, followed by the use of an energy storage welding machine to position the parts to be brazed. When preparing the brazing paste, the mass ratio of filler to Canadian resin to acetone was set at 10:1:20. The modulated filler paste was applied to the weld, and the sample was put into the vacuum furnace after drying. As shown in Figure 2b, the brazing temperature was 1200 °C, with a hold time of 30 min. The vacuum level of the vacuum furnace was maintained at less than 1 × 10⁻⁴ Pa. The samples were cooled in the furnace to below 200 °C after heating.

The joints brazed using the five different brazing fillers are designated as J1, J2, J3, J4, and J5, respectively. J represents joint, while the following number corresponds to the numerical identifier of the brazing filler used. After brazing, the samples were polished. The polished samples were then subjected to corrosion using an etching solution (HCl:CuSO₄:H₂O = 10 mL:2 g:20 mL). The microstructure of the joint was observed by scanning electron microscopy (SEM, ZEISS EVO 10, Oberkochen, Germany) and the composition was analyzed by energy dispersive X-ray spectroscopy (EDS, Oxford instruments X-max, Oxon, UK). Transmission electron microscopy (TEM, FEI Tecnai G2 F20, Hillsboro, OR, USA) was used to analyze the joint microstructure in bright field, and the characteristic phases were analyzed by selected-area electron diffraction (SAED) to determine the crystal structure. The mechanical property test samples were cut out using WEDM according to Figure 2a. Subsequently, the test samples were polished using 600-grit sandpaper, and the tensile tests were conducted at 980 °C, as 980 °C is a critical point for the mechanical properties of joints made from Ni-based or Ni₃Al-based superalloys in high-temperature applications. The tensile tests were conducted on an electronic universal testing machine (Wance ETM305D, Shenzhen, China) with a loading speed of 0.5 mm/min. Each test was performed at least twice in the same conditions to ensure the accuracy of the test data. The fracture morphology was observed by SEM to analyze the failure behavior of the joint.

3. Results

3.1. Characterizations of the Fillers

The microstructure of the brazing filler powders after embedding and grinding is shown in Figure 3, with the filler powder particles being spherical and having a diameter smaller than 120 μm. The XRD results of the fillers are presented in Figure 4. The results indicate that the prepared filler powder contains γ′ phase, P-topologically close-packed (P-TCP) phase, and Ni₇Zr₂ phase. However, no diffraction peaks of the γ′ phase are found in the XRD curve of the BNiZr-3 filler powder.

3.2. Microstructure of the Joint

The typical microstructure of the brazing joint J4 is shown in Figure 5. With reference to joints prepared using B-containing filler, the joint can be divided into four areas, including an isothermal solidified zone (ISZ), diffusion affected zone (DAZ), athermally solidified zone (ASZ), and BM [24]. The ASZ is located in the middle of the joint and constitutes the eutectic structure of the joint after solidification. The ISZ is located on both sides of the ASZ and mainly consists of a single-phase region. The DAZ is located at the junction between the joint and the BM, where the compositions and phase components are influenced by the elemental interdiffusion between the joint and the BM [41].

The compositions of spots in Figure 5 are shown in

Table 2. Composition of spots in Figure 5 (at. %).

Spot	Re	Mo	Al	Ta	Cr	Zr	Ni	Phase Component	Zone
A	0.00	1.41	2.08	0.19	1.16	20.09	75.08	Ni7Zr2	DAZ
B	0.80	12.06	7.30	0.18	10.85	0.29	68.53	γ+γ′ eutectic	DAZ
B1	0.23	4.64	14.04	0.21	4.64	2.48	73.95	γ+γ′ eutectic	DAZ
D	1.69	49.16	6.60	1.41	6.02	0.00	35.12	P-TCP	DAZ
E	5.59	32.84	3.61	0.27	10.68	0.38	46.61	NiMo	DAZ
C	0.00	2.62	18.28	1.12	2.35	1.35	74.27	γ′	ISZ
A1	0.00	1.17	2.39	0.20	1.23	20.52	74.49	Ni7Zr2	ASZ
B2	0.05	9.93	8.82	0.03	13.14	0.38	67.65	γ+γ′ eutectic	ASZ
D1	1.56	48.62	6.60	1.14	6.65	0.17	35.26	P-TCP	ASZ

, which also indicates the location of each spot within the joint. TEM analysis was performed on J4, and the bright-field images and SEAD results are depicted in Figure 6.

Table 3. Composition of spots in Figure 6 (at. %).

Spot	Re	Mo	Al	Ta	Cr	Zr	Ni	Phase Component
A	3.10	44.51	2.38	0.28	11.63	0.37	37.73	P-TCP
B	0.07	1.98	1.74	0.13	0.88	17.20	77.99	Ni7Zr2
C	0.10	4.13	16.13	1.63	0.97	0.00	77.04	γ′
D	2.47	51.29	7.08	1.17	5.5	0.51	31.97	NiMo

lists the chemical compositions of each spot in Figure 6 to ensure correspondence between the phases analyzed in TEM and SEM.

Table 4. Crystal structures and lattice parameters of some typical phases. [42,43,44,45,46].

Phases	Crystal Structure	Space Group	$\mathbf{Lattice} \mathbf{Parementer} \mathbf{(}\mathit{Å}$)	α, β, γ
P-TCP	Orthorhombic	Pnma	a = 16.98; b = 4.75; c = 9.07	90°, 90°, 90°
Ni7Zr2	Monoclinic	C2/m	a = 4.70; b = 8.24; c = 12.19	90°, 96°, 90°
γ′	Cubic	Pm$\bar{3}$m	a = b = c = 3.52	90°, 90°, 90°
γ	Cubic	Fm$\bar{3}$m	a = b = c = 3.48	90°, 90°, 90°
NiMo	Orthorhombic	P212121	a = b = 9.11, c = 8.85	90°, 90°, 90°

presents the crystal structures and lattice parameters of some phases reported in the literature. Combining SEM and TEM test results, it can be determined that the joint is composed of five phases: γ, γ′, P-TCP, NiMo, and Ni₇Zr₂.

Table 5. Crystal structures and lattice parameters of some typical phases.

Phases	Interplanar Distance Calculated		Interplanar Distance Measured
P-TCP	$d_{\left(1 0 0\right)}^{P-TCP}$$=16.98Å$	$d_{\left(\mathrm{0,1},1\right)}^{P-TCP}$$=4.20Å$	${d_{\left(1 0 0\right)}^{P-TCP}}_{meas}$$=17.81Å$	${d_{\left(0 1 1\right)}^{P-TCP}}_{meas}$$=4.41Å$
Ni7Zr2	$d_{\left(1 1-3\right)}^{{Ni}_7{Zr}_2}$$=3.00Å$	$d_{\left(0 2 2\right)}^{{Ni}_7{Zr}_2}$$=3.40Å$	${d_{\left(1 1-3\right)}^{{Ni}_7{Zr}_2}}_{meas}$$=3.12Å$	${d_{\left(0 2 2\right)}^{{Ni}_7{Zr}_2}}_{meas}$$=3.89Å$
γ′	$d_{\left(2 0 0\right)}^{\gamma ΄}$$=d_{\left(0 2 0\right)}^{\gamma ΄}=$$1.76Å$		${d_{\left(2 0 0\right)}^{\gamma ΄}}_{meas}$$=1.90Å$	${d_{\left(0 2 0\right)}^{\gamma ΄}}_{meas}$$=1.87Å$
NiMo	$d_{\left(2 0 0\right)}^{NiMo}$$=d_{\left(0 2 0\right)}^{NiMo}=$$4.56Å$		${d_{\left(2 0 0\right)}^{NiMo}}_{meas}$$=4.11Å$	${d_{\left(0 2 0\right)}^{NiMo}}_{meas}$$=4.12Å$

shows the calculated and measured interplanar distances of the main phases in the joint. According to the literature, P-TCP belongs to the orthorhombic crystal system with the Pnma space group. The interplanar distances calculated from its lattice parameters are $d_{\left(1 0 0\right)}^{P-TCP}$ = 16.98$Å$ for the (1 0 0) plane, and $d_{\left(\mathrm{0,1},1\right)}^{P-TCP}$ = 4.20$Å$ for the (0 1 1) plane. However, the measured distances are ${d_{\left(1 0 0\right)}^{P-TCP}}_{meas}$ = 17.81$Å$ and ${d_{\left(0 1 1\right)}^{P-TCP}}_{meas}$ = 4.41$Å$. Combined with composition analysis, this discrepancy is possibly due to lattice expansion caused by elements such as Ta. Ni₇Zr₂ and γ′ may have undergone lattice expansion caused by elements such as Mo, and the change in the interplanar distance of NiMo may be caused by elements such as Cr.

The microstructure of the joints brazed with five fillers are shown in Figure 7. According to the cross-sectional microstructure images in Figure 7a, numerous granular NiMo phases can be observed in the DAZ of J1 (Zr content of filler: 11.8–12.2 wt.%). γ+γ′ eutectic, P-TCP, and Ni₇Zr₂ are observed in the ASZ. The P-TCP phases are marked by red arrows and dashed lines, while Ni₇Zr₂ phases are marked by yellow arrows and dashed lines. Increasing the Zr content to 13.8–14.2 wt.% in the BNiZr-2 filler results in an increase in the Ni₇Zr₂ phase within J2, while reducing the number of granular NiMo phases in the DAZ. Moreover, when the Zr content in the BNiZr-3 filler increases to 15.8–16.2 wt.%, the size of the Ni₇Zr₂ phase in J3 significantly increases compared to J2.

Increasing Mo content in BNiZr-4 filler to 7.3–7.7 wt.% results in the disappearance of numerous granular NiMo phases in the DAZ, and an increase in P-TCP phases in the ASZ and DAZ of J4. With an increase in Mo content in the BNiZr-5 filler to 9.8–10.2 wt.%, the presence of P-TCP in J5 continues to increase compared to that in J4.

3.3. Mechanical Properties and Failure Behavior of Joint

The tensile tests of five kinds of samples brazed with different brazing fillers are conducted at 980 °C, and the test results are shown in Figure 8. The tensile strength of J1 is 481 MPa at 980 °C. With the increase in Zr in the filler, the tensile strength of J2 decreases to 461 MPa. When Zr in the filler continues to increase, the tensile strength of J3 decreases to 307 MPa. In comparison to J1, an increase in Mo content results in elevated tensile strength for J4, reaching 550 MPa at 980 °C. However, as the Mo content further increases, the tensile strength of J5 declines to 474 MPa. This trend indicates an initial rise followed by a subsequent decrease in joint tensile strength with increasing Mo content at 980 °C.

The tensile fracture morphologies of joints at 980 °C were observed by SEM (Figure 9). The fracture site of J1 (Figure 9a) is located in the periphery of the γ′ phase. As the Zr content increases, J2 fails at the interfaces of ISZ and DAZ (Figure 9b). With further increases in Zr content, besides failure at the interfaces of ISZ and DAZ, pores are also found in the ASZ of J3 (Figure 9c). Upon increasing the Mo content, it is observed that both J4 (Figure 9d) and J5 (Figure 9e) fail at the interfaces of ISZ and DAZ, with the fracture surface of J5 being smoother.

4. Discussion

4.1. The Effects of Mo and Zr Content on the Fillers

Before designing the composition of the brazing filler, phase diagrams with the addition of Zr on the basis of the BM IC21 composition were calculated using the commercial software Thermo-Calc 2022a (Figure 10a). It was found that the addition of Zr effectively lowered the liquidus line of the system. When the Zr content increases to 16%, the liquidus line of the system decreases to 1145 °C. Furthermore, the mole fraction of different phases in the system at 1200 °C was calculated in Figure 10b, and the results show that the mole fraction of the liquid phase is maximized when the Zr content is about 15%. These calculation results indicate that the liquidus line can be effectively reduced by adding Zr to the BM composition, allowing the designed brazing filler to exist in a liquid form of over 95% at 1200 °C.

The phase diagrams and phase mole fraction obtained from Thermo-Calc by adjusting the Zr content based on the composition of the BNiZr-1 filler are shown in Figure 10c,d. When the mass fraction of Zr is ~0.4%, the Ni₇Zr₂ phase appears in the system. With an increase in Zr content, the transformation temperature of the Ni₇Zr₂ phase stabilizes around 1145 °C until the Zr content reaches 16%. Throughout this process, the liquidus line of the system remains less than 1160 °C, consistent with the results of DSC testing. According to the mole fraction of the existing phase in Figure 10d, it can be observed that at 1200 °C, when the mass fraction of Zr is between 13% and 18%, the system exists entirely in the liquid phase. When the mass fraction of Zr exceeds 19%, the Ni₇Zr₂ phase appears in the system. Furthermore, as indicated by Figure 10c, when the Zr content ranges from 4 to 16 wt.%, the system predominantly precipitates the γ′ phase during the cooling process. However, when the Zr content reaches 16–20 wt.%, the γ′ phase no longer precipitates preferentially. This may explain the absence of γ′ diffraction peaks in the XRD curve of the BNiZr-3 filler as discussed in Section 3.1.

Figure 10e,f show the phase diagram adjustment of Mo based on the BNiZr-1 filler and the calculated phase mole fraction of the system at 1200 °C. The phase diagram indicates that despite an increase in Mo content, the liquidus line of the system remains stable at around 1140 °C, and the Ni₇Zr₂ phase boundary stabilizes at around 1130 °C, which is consistent with the results of DSC test. According to the calculated phase fractions, when the mass fraction of Mo in the system is ~(7–8) wt.%, the system exists entirely in the liquid phase, indicating that within this range, the filler can fully melt to fill the weld gap.

4.2. The Formation Process of the Joint

According to the analysis in Section 4.1, at the brazing temperature, the BNiZr-4 filler is completely in the liquid state. During the brazing process, there is interdiffusion of elements between the BM and the filler. As the BM, IC21 contains ~77.0 wt.% Ni, which is higher than the Ni content in the BNiZr-4 filler (~65.0 wt.%), resulting in concentration gradients between the BM and the filler. For a one-dimensional problem, Fick’s law reduces to [47]:

$$J=-D{\displaystyle \frac{dC}{dx}}$$

(1)

where $J$ is diffusion flux, $D$ presents effective diffusivity, and $dC$/$dx$ denotes concentration gradient. When there are concentration ($C)$ differences in the solid, atoms will diffuse from regions of high concentration to regions of low concentration. Before brazing, the Ni concentration in the BM ($C_{Ni-BM}$) is higher than the Ni concentration in the filler ($C_{Ni-filler})$. Therefore, during brazing, Ni from the BM diffuses towards the center of the joint.

Since the Al content in both the BM and the filler are the same, it is necessary to consider the influence of chemical potential on diffusion. The driving force for diffusion $F$ always causes diffusion to proceed in the direction of decreasing chemical potential until $∆u_i=0$. Chemical potentials of Al in the BM (${\mu }_{Al-BM}$) were calculated using Thermo-Calc. Chemical potentials of Ni and Mo at 1200 °C were simultaneously calculated to further determine the direction of chemical driving force (Figure 11). Additionally, the influence of Mo variation in the filler on the change in element chemical potentials (${\mu }_i$) at 1200 °C was calculated for subsequent discussion. The chemical potentials of elements in the BNiZr-1, BNiZr-4, and BNiZr-5 fillers have been indicated in Figure 11. To emphasize the magnitude of element’s chemical potentials in both the BM and filler, dashed lines were used to represent the chemical potentials of elements in the BM. The data marked with dashed lines are independent of the x-axis.

Based on the calculations from Figure 11, it can be inferred that at 1200 °C, Al diffuses from the filler to the BM, while Ni diffuses from the BM to the filler, forming single-phase γ′ regions known as the ISZ on both sides of the ASZ. The Ni that has not fully diffused from the BM into the joint, along with elements like Re and Mo in the BM used for solid solution strengthening of γ, form the granular NiMo phase in the DAZ. The substrate of the DAZ is the γ′ phase formed in the BM after Ni diffusion into the joint. During the cooling process after brazing, the γ′ phase precipitates and grows to form the ISZ. Spherical γ′ phases are prominently observed in Figure 12. Mo tends to partition into the γ phase, while Zr, with a significant atomic size difference from Ni, tends to segregate at the γ phase boundaries [7]. Zr reacts with Mo-containing γ phase along the γ′ grain boundaries to form significant quantities of the P-TCP phase and the Ni₇Zr₂ phase within the γ+γ′ eutectic. On both sides of the ISZ, enriched regions of P-TCP, Ni₇Zr₂, and γ+γ′ eutectic constitute the DAZ and ASZ.

4.3. The Influence of Elemental Adjustments on the Joint

The phase constitutes of each area in the joint were determined in Section 3.1. Each area of the joint has been marked out in Figure 7. Figure 13 analyzes the percentage of the width of each area to the width of the joints brazed with different fillers according to Figure 7.

4.3.1. ASZ

According to the analysis in Section 3.1, the ASZ is mainly composed of Ni₇Zr₂, TCP, and γ+γ′ eutectic. Compared to J1, J2, and J3, with the increase in Zr content, the amount of Ni-Zr phase in the joints progressively increases. According to the Ni-Zr binary phase diagram (Figure 14a), when the atomic proportion of Zr in the Ni-Zr system is (16–27.5) at. % at 1200 °C, the Ni-Zr phase all exists in the form of Ni₇Zr₂. EDS analysis was performed on Ni-Zr phase of five joints, and the changes in compositions of Ni-Zr phases are shown in Figure 14b. It can be seen that the changes in Ni and Mo in filler did not affect the compositions of Ni-Zr phases obtained by brazing at 1200 °C for 30 min. This result demonstrates that the Ni-Zr phases present in the five brazed joints exist primarily in the form of Ni₇Zr₂.

For J1, J2, and J3, the increase in Zr content in the joint leads to an increase in Ni₇Zr₂ in the ASZ, as shown in Figure 7. Consequently, the proportion of ASZ width increases with the increase in Zr content. Similarly, for J1, J4, and J5, an increase in Mo content also results in a higher presence of the P-TCP phase in the joint. Therefore, with the increase in Mo content, the proportion of ASZ width increases as well.

4.3.2. DAZ and ISZ

According to the discussion in Section 4.2, the ISZ is primarily composed of a single-phase γ′ region formed by Ni diffusing from the BM to the filler and Al diffusing from the filler to the BM. The DAZ is the region where the reaction occurs between the BM and the filler, as well as where the phases precipitate between the ISZ and the BM. In the BNiZr-1 filler, the Ni content is ~67.5 wt.%. Ni is added as a balance component during the formulation. So, as the Zr content increases, the Ni content decreases. In the BNiZr-3 filler, the Ni content has been reduced to ~63.5 wt.%. The Ni concentration gradient between the BM and the filler increases. More Ni diffuses from the BM into the weld seam, causing the Ni in the BM to more readily transform into γ′. However, with the increase in Zr content in the filler, the Zr concentration gradient between the filler and the BM also increases. Consequently, more Zr forms Ni₇Zr₂ at the DAZ. The presence of Ni₇Zr₂ and the supersaturated Zr in the DAZ impedes the diffusion of Ni from the BM into the joint. This may explain why the proportion of DAZ width decreases with the increase in Zr content.

With an increase in Mo content, the proportion of DAZ width decreases significantly. In J4 and J5, there are no large amounts of granular NiMo phase observed in the DAZ. This contrasts with the noticeable changes in the DAZ region observed in J1 and is the primary reason for the reduced DAZ width in J4 and J5. In this context, the effect of Mo addition on the chemical potentials of elements during the brazing process should be considered.

According to the results from Thermo-Calc calculations (Figure 11), it is observed that with an increase in Mo content, the chemical potential of Ni decreases, while that of Al increases in the filler. Therefore, as the Mo content increases, the driving force for diffusion of Ni from the BM to the filler, as well as Al from the filler to the BM, also increases. For J1, Section 4.2 has already explained that the NiMo phase in the DAZ is formed by the interaction of Mo and other elements with Ni, originally from the γ phase in the BM.

Figure 15 illustrates the variation in diffusion coefficients of Al and Ni in the filler liquid phase at 1200 °C with adjustments in Mo content (based on BNiZr-1 filler). It can be observed that with the increase in Mo content, both the diffusion coefficients of Al ($D_{Ni Ni}^{Al}$) and Ni ($D_{Al Al}^{Ni}$) in the filler liquid phase at 1200 °C increase. The increase in Mo content leads to a decrease in the chemical potential of Ni in the filler, and an increase in the diffusion coefficient of Ni in the filler liquid phase, which causes the BM to be more inclined to transform into γ′. Additionally, the increase in the chemical potential of Al in the filler and the diffusion coefficient of Al in the liquid phase also contribute to the preference of the BM towards γ′ transformation, leading to the release of refractory elements such as Mo and the precipitation of TCP phases [49].

Based on the analysis above, theoretically, an increase in Mo content should result in more γ phases in the BM transforming into γ′ phases, leaving behind a significant amount of NiMo phases in the original position of the γ phase in the BM, thus increasing the proportion of DAZ width in J4. However, in reality, when Mo is increased, the proportion of DAZ width in J4 significantly decreases. This is because a large amount of Mo in the filler is excluded from the ISZ due to its limited solubility, with one side located in the ASZ and the other side in the DAZ. The accumulation of Mo in the DAZ forms diffusion barriers [50], hindering the mutual diffusion between the BM and filler and causing a significant precipitation of P-TCP phases in the DAZ. The Ni blocked by Mo from diffusing into the filler maintains the integrity of the γ+γ′ eutectic in the BM, while the Al blocked from diffusing into the BM leads to the formation of more ISZ in the joint. Therefore, compared to J1, the proportion of ISZ width in J4 significantly increases. The increase in Mo content also increases the amount of P-TCP in the DAZ. Thus, relative to J4, the proportion of DAZ width in J5 increases.

4.4. Joint Failure Behavior

In Section 3.3, the tensile strength at 980 °C and fracture morphology of the joints were provided. According to the fracture morphology in Figure 9, it can be observed that the failure occurred exclusively at the edges of the γ′ phase. This occurs because in the joint, the P-TCP and Ni₇Zr₂ phases, which act as sources of cracks, primarily accumulate at the γ′ phase grain boundaries, and the cracks propagate along these γ′ phase boundaries. In this study, the P-TCP phase is mainly composed of Ni, Mo, and Cr. The precipitation of the TCP phase consumes excessive amounts of solid solution strengthening elements [51]. Microcracks readily generate near the TCP phases [52] and the hard and brittle nature of the TCP phase can easily lead to alloy cracking [53,54]. Simultaneously, the NiMo phase also acts as a source of cracks and is dispersed throughout the DAZ.

As the Zr content increases, the presence of the Ni₇Zr₂ phase in the joint also increases, leading to a higher number of crack initiation sites and a consequent decrease in the tensile strength of the joint. The presence of pores in the ASZ of the J3 fracture morphology further confirms the increase in crack sources within the joint. When the Mo content is increased, compared to J1, a significant transformation occurs in the joint where the dispersed NiMo phase transitions from being widely distributed to forming a continuous distribution of P-TCP phase, leading to a significant decrease in the proportion of DAZ width in J4. This suggests a reduction in crack sources in J4, thereby significantly enhancing its tensile strength compared to J1. However, as the Mo content continues to increase, the increased presence of the P-TCP phase in the joint leads to more crack sources, thus reducing the tensile strength of J5 compared to J4.

5. Conclusions

The brazing joint of the Ni₃Al-based single crystal superalloy IC21 was prepared using a Zr-containing filler at a brazing temperature of 1200 °C for 30 min. The effect of adjusting the content of Zr and Mo in the filler on the microstructural evolution and 980 °C tensile properties of the joints was investigated. The following conclusions were drawn:

(1)

Thermo-Calc calculations and experimental results suggest that Zr can be utilized as an MPD in the preparation of the filler for brazing Ni₃Al-based single crystal superalloys. This results in the formation of robust joints characterized by a combination of γ′, γ, P-TCP, Ni₇Zr₂, and NiMo phases. The joint can be divided into four regions: BM, ASZ, ISZ, and DAZ.

(2)

When adjusting the Zr and Mo contents within appropriate ranges, the filler can completely transform into a liquid state and enter the joint at 1200 °C. Increasing the Zr content in the joint leads to a higher concentration of the Ni₇Zr₂ phase, which results in an increase in the proportion of the ASZ width. Increasing the Mo content theoretically should increase the diffusion of Ni from the BM to the filler and Al from the filler to the BM. However, the diffusion barrier effect of Mo impedes this phenomenon, leading to a transformation of NiMo phases in the DAZ to P-TCP phases. Nonetheless, a continued increase in Mo content also leads to a higher concentration of the P-TCP phase in the joint.

(3)

Ni₇Zr₂, NiMo, and P-TCP phases are the failure locations of joints. When the Zr content is increased, the rise in the Ni₇Zr₂ phase within the joint leads to a reduction in the tensile strength. As the Mo content increases, there is a reduction in the NiMo phase, resulting in an increased tensile strength of J4 compared to J1. However, the continuous increase in Mo leads to an increase in the P-TCP phase in the joint, causing a decrease in the tensile strength of J5 compared to J4.