Diffusion bonding of AlCoCrFeNi2.1 eutectic high entropy alloy to TiAl alloy
School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, China
Corresponding authors: * E-mail address:donghg@dlut.edu.cn(H. Dong).
Received: 2019-08-27 Accepted: 2019-10-29 Online: 2020-05-15
High entropy alloys have special microstructure and remarkable properties. To explore their potential engineering application in high temperature structures, the microstructure evolution of bonding interface, the elemental diffusion behavior and mechanical property of the diffusion bonded joint between AlCoCrFeNi2.1 eutectic high entropy alloy (EHEA) and TiAl alloy were investigated. Four reaction layers (rodlike B2 phase, Al(Co, Ni)2Ti, τ3-Al3NiTi2 + TiAl, τ3-Al3NiTi2 + TiAl + Ti3Al) formed in the diffusion zone near FCC phase of EHEA, but three layers (Al(Co, Ni)2Ti, τ3-Al3NiTi2 + TiAl, τ3-Al3NiTi2 + TiAl + Ti3Al) formed near B2 phase. Al and Ni controlled the reaction diffusion of EHEA and TiAl alloy, coarsened the acicular precipitated B2 phase and turned TiAl phase into Al(Co, Ni)2Ti and τ3-Al3NiTi2 phases. All these reaction layers grew in a parabolic manner as a function of bonding temperature. Rodlike B2 phase has the lowest growth activation energy of 125.2 kJ/mol, and the growth activation energy of τ3-Al3NiTi2 + TiAl layer near B2 phase is much lower than that near FCC phase. The penetration phenomenon and convex structure formed in the diffusion zone, which resulted in interlocking effect and enhanced the strength of resultant joints. The highest shear strength of 449 MPa was achieved at 950 °C. And the brittle fracture generally initiated at the interface between Al(Co, Ni)2Ti and τ3-Al3NiTi2 + TiAl layers.
Keywords:
Cite this article
Li Peng, Wang Shuai, Xia Yueqing, Hao Xiaohu, Dong Honggang.
1. Introduction
As revolutionary new alloys, high entropy alloys (HEAs) have attracted extensive attention recently [1,2]. Compared with conventional alloys, HEAs are composed of no less than five principle elements and exhibit four core effects summarized as high-entropy effect, sluggish diffusion effect, severe lattice-distortion effect and cocktail effect [3]. Due to these special effects, HEAs have remarkable mechanical and physical properties at elevated or cryogenic temperature, excellent corrosion resistance, outstanding irradiation and thermoelectric properties [[4], [5], [6], [7], [8], [9]]. These prominent comprehensive properties make it possible for HEAs to be the ideal substitutions for traditional structural superalloys. Generally, HEAs have single-phased structures, such as body-centered-cubic (BCC), face-centered-cubic (FCC) and hexagonal-close-packing (HCP) structures [2,10]. Single-phased HEAs are usually difficult to exhibit reasonable strength and ductility simultaneously [11,12]. However, Lu et al. [13] designed a promising eutectic high entropy alloy (EHEA) AlCoCrFeNi2.1 which is composed of hard BCC and ductile FCC phases, thus realizing the combination of high tensile strength and high ductility. The superior mechanical properties and good castability also ensure that EHEA has a prospect in engineering application.
According to the literature reviews on HEAs, it could be found that most of researchers put their attention on designing HEAs with new composition and optimizing the microstructure and properties of HEAs [[14], [15], [16], [17], [18], [19]], but seldom focus on the processing of HEAs. When it comes to material processing methods, welding is the indispensable one. Weldability is one of the most important criterions for evaluating the value of engineering application for new materials. The weldability, microstructure and mechanical properties of CoCrFeMnNi system HEAs joints are extensively explored. Wu et al. [20] studied the weldability of CoCrFeMnNi HEAs by electron beam welding. The joints without solidification cracking were acquired and their mechanical properties remained the same with BM even at cryogenic temperature. Then Wu et al. [21] further discussed the difference of microstructure and mechanical properties between electron beam welding and gas tungsten arc welding. The properties of gas tungsten arc welded joints were inferior to the results of electron beam welding. Kashaev et al. [22] achieved the laser beam welding of CoCrFeMnNi HEA and produced flawless joints. The remarkable increase of microhardness was found in fusion zone, which was caused by the precipitation of nanoscale B2 phase. Nam et al. [23,24] investigated the laser welding of CoCrFeMnNi HEA and effect of post-weld heat treatment (PWHT) on joints. The sound joints were obtained, but the hardness and tensile strength of them was deteriorated with the increase of PWHT temperature. Jo et al. [25] compared the results of friction stir welded (FSWed) CoCrFeMnNi HEAs joints with that of laser welded (LWed) joints and found the mechanical properties of FSWed joints were better than that of LWed joints. The most of the above researches about welding of HEAs principally focus on the microstructure and properties of similar joints. However, the dissimilar metal joints have been widely used in the key industrial areas. AlCoCrFeNi2.1 EHEA has great potential to be applied in high temperature structural component in aerospace industry, and TiAl alloy also has been widely employed in modern manufacture of aircraft engine [26]. To obtain the AlCoCrFeNi2.1/TiAl hybrid structure and enhance the engineering application of EHEA, this work investigates the feasibility of welding AlCoCrFeNi2.1 EHEA and TiAl alloy by vacuum diffusion bonding, and then analyzes the microstructure evolution and mechanical properties of resultant joints.
2. Materials and methods
The as-cast AlCoCrFeNi2.1 EHEA and commercially available TiAl alloy were selected as base metals, of which the chemical composition is displayed in Table 1. Fig. 1 presents the microstructures and XRD patterns of the base metals. As shown in Fig. 1(a), AlCoCrFeNi2.1 EHEA exhibits lamellar eutectic microstructure within which the white phase is FCC phase and the grey phase is B2 phase. Fig. 1(c) displays the microstructure of the dual-phase TiAl alloy that is composed of lath-shaped Ti3Al phase and dark grey TiAl phase. Then these metals were machined into blocks by wire cutting machine with the dimensions of 5 mm × 10 mm × 20 mm and 5 mm × 10 mm × 10 mm, respectively. The faying surfaces of blocks were polished with SiC papers up to grit 2000, and then ultrasonically cleaned in acetone to remove dust. The prepared samples were assembled in the structure as displayed in Fig. 2(a), and then put into a vacuum diffusion bonding furnace (ZTF2-10, Shaanxi Zituo Solid-State Additive Manufacturing Technology Company, China). Diffusion bonding experiments were conducted according to the process curve as displayed in Fig. 3. Generally, the diffusion bonding temperatures ranges from 0.6 Tm to 0.8 Tm (where Tm is the melting point) [27]. While for dissimilar metal couples, the bonding temperature mainly depends on the base metal with a lower melting point. Since the melting points of EHEA and TiAl alloy are 1361 °C and 1500 °C, respectively, the bonding temperature is determined by EHEA, which ranges from 700 °C to 1000 °C. To alleviate the influence of sluggish effect on the diffusion process, the bonding temperature should be higher. Therefore, 900, 950 and 1000 °C were selected as the bonding temperatures. The determination of bonding pressure relies on the TiAl alloy which exhibits inferior mechanical properties at elevated temperature. According to the research of Ustinov et al. [28] and previous experiments, the bonding pressure and time was determined as 15 MPa and 0.5 h to reduce the deformation of TiAl alloy and ensure the bonding efficiency. To further enhance the quality of bonded joints and decrease the deformation of the welding samples, the boding process was set as two steps in this study referring to preceding experience. First, the temperature was raised from 20 °C to 800 °C at the rate of 10 °C/min and preheated for 0.5 h under the high pressure of 30 MPa to ensure the close contact of faying surface. Then the assembled samples were heated to bonding temperature that was set from 900 °C to 1000 °C in a step of 50 °C to explore the temperature effect, and kept for 0.5 h under the axial pressure of 15 MPa. After bonding process, the resultant joints were cooled down in the furnace to room temperature.
Table 1 Chemical composition (at. %) of AlCoCrFeNi2.1 EHEA and TiAl alloy.
Materials | Al | Co | Cr | Fe | Ni | Ti | Nb |
---|---|---|---|---|---|---|---|
AlCoCrFeNi2.1 EHEA | 16.39 | 16.39 | 16.39 | 16.39 | 34.43 | - | - |
TiAl alloy | 42.0 | - | - | - | - | 50.9 | 7.1 |
Fig. 1.
Fig. 1.
Characteristics of base metals: microstructure (a) and XRD patterns (b) of AlCoCrFeNi2.1 EHEA; microstructure (c) and XRD patterns (d) of TiAl alloy.
Fig. 2.
Fig. 2.
Appearance of the bonded joint and schematic of the samples for shear test.
Fig. 3.
Fig. 3.
Process curve of diffusion bonding.
The samples for shear test were processed into the structure as shown in Fig. 2(b), and partial samples were ground and polished for metallographic analysis. The shear strength of joints was tested by a universal testing machine (DNS-100) with crosshead speed of 0.5 mm/min. The microstructure and elemental distribution of the interface were detected by electron probe micro-analyzer (EPMA, JXA-8530 F Plus). The fracture surface was examined by scanning electron microscope (SEM, ZEISS-SUPRA55) equipped with energy dispersive spectroscope (EDS). The thin foils of diffusion zone were prepared by focused ion beam (FIB, FEI-Helios G4 UX), then the transmission electron microscope (TEM, JEOL JEM-2100 F) was applied to observe the microstrcuture and selected area electron diffraction (SAED) patterns of different phases in diffusion zone.
3. Results and discussion
3.1. Microstructure of bonding interface
Fig. 4 shows the microstructure evolution of the joints bonded at different temperatures. There are no obvious welding defects existed in the bonding interface, suggesting that the reliable metallurgy bonding between EHEA and TiAl alloy was achieved under all the designed process parameters. The diffusion layer near EHEA is flat, while the layer close to TiAl alloy is rugged. Especially, when it comes to the diffusion zone between B2 phase and TiAl alloy, convex structure appears. The magnified pictures in Fig. 4 present more information of the base metals and diffusion zones. The acicular precipitates in FCC phase coarsened with the increase of temperature, while the precipitates in B2 extremely grew up at 950 °C and almost totally dissolved at 1000 °C. Wang et al. [29] investigated the phase transformation of AlxCoCrFeNi high entropy alloy at elevated temperature and also found that the σ phase rich in Fe and Cr would precipitate from the B2 phase at the temperature 600 °C but disappear at 1000 °C. In this work, the precipitates in B2 was actually σ phase. Therefore, the precipitates gradually grew up as a function of bonding temperature when the temperature was below 1000 °C, but dissolved in the B2 phase at 1000 °C. The thickness of whole diffusion zone and diffusion layers obviously increases with the augment of temperature. It is noteworthy that the diffusion zones near FCC phase and B2 phase in EHEA display different microstructures and thicknesses. To further explore their characteristics in diffusion bonding with TiAl alloy, these two diffusion zones are marked as Zone A and Zone B, respectively, and will be analyzed in following section.
Fig. 4.
Fig. 4.
Microstructure evolution of bonding interface at different temperatures of 900 °C (a), 950 °C (b) and 1000 °C (c).
3.2. Elemental distribution in Zone A and Zone B
Fig. 5 shows the distribution of alloying elements in Zone A at 950 °C. The segregation of Co, Cr and Fe elements is distinct in FCC phase of EHEA, while the acicular precipitates within FCC phase are rich in Al and Ni elements. The diffusion zone in Zone A can be typically divided into four layers based on the back scattered image (BEI). Layer I consists of two phases which are perpendicular to the bonding interface. The white phase is rich in Co, Cr and Fe elements, while the enrichment of Al and Ni is evident in grey rod phase. The elemental distribution of layer I is highly similar to FCC phase with acicular phase precipitated. The Co element largely diffused from layer I to layer II, causing the depletion of Co element in layer I. Layer II is almost one single phase on which white punctate phase disperses. Co element is abundant in this single phase which is also relatively rich in Al, Ni and Ti elements. Moreover, the white punctate phase consists of Co, Cr and Fe elements. Layer III is composed of blocky phase and grey lamellar phase. The content of Al is almost the same in these two phases, but the content of Ti in blocky phase is higher than that of grey lamellar phase. Ni element was also detected in grey lamellar phase. Layer IV is adjacent to TiAl alloy and two kinds of lath-shaped phases alternately grow in this layer. The grey lath-shaped phase composed of Al, Ni and Ti elements is similar to the grey lamellar phase in layer III. High content of Ti and Al elements gathered in the dark grey lath-shaped phase. Fig. 6 presents the elemental distribution of Zone B at 950 °C. The B2 phase of EHEA is rich in Al and Ni elements, while Co, Cr and Fe elements gathered at the blocky precipitates in B2. Compared with Zone A, Zone B is lack of layer I. The thickness of layer II decreased due to the bulge of layer III. The elemental distribution of layer II, III and IV in Zone B is almost the same with Zone A.
Fig. 5.
Fig. 5.
Distribution of alloying elements in Zone A at 950 °C.
Fig. 6.
Fig. 6.
Distribution of alloying elements in Zone B at 950 °C.
3.3. Diffusion behavior of diffusion layers
Fig. 7 shows the effect of bonding temperature on the growth of diffusion layers in Zone A and Zone B. The thickness of all diffusion layers in these two zones increased with the augment of temperature. Their average values of thickness are marked in the pictures measured by using the region function of Auto CAD software. As for Zone A, the growth of layer I actually is the coarsening of acicular B2 phase precipitated in FCC phase. From scanning line near FCC in Fig. 8, the content of Al dramatically decreases from TiAl alloy to EHEA and small peaks of Al element arise in layer I. The large amount of Al element diffusing into FCC phase accelerated the growth of acicular B2 phase, the growth direction of which was opposite to diffusion direction of Al element. The chemical composition and possible phases at locations A and B shown in Table 2 are consistent with the preceding explanation. In layer II, the content of Al, Ni, Co and Ti elements rises up with platforms forming as shown in scanning line near FCC. In consideration of the composition at location C and the XRD results of fracture surface on EHEA side as shown in Fig. 9, it could be preliminarily deduced that Al(Co, Ni)2Ti phase formed in layer II. In the blue box marked in Fig. 7, the white punctate phase dispersed on Al(Co, Ni)2Ti phase tended to gather at edge of layer III and gradually disappeared with the temperature rising up. Its elemental distribution is similar with FCC phase, suggesting that it probably be FCC phase and fell off from EHEA matrix during diffusion bonding process. Li et al. [30] also mentioned that element diffusion could induce penetration phenomenon. Along the interface between layer I and layer II, penetration is evident and extrudes FCC phase out of matrix, thus forming the white punctate phase dispersed on layer II. With the penetration proceeding, Al(Co, Ni)2Ti phase cut through FCC phase and got in touch with coarsened rodlike B2 phase, simultaneously forming a more effective diffusion tunnel for Al elements and interlocking layers I and II. For layers III and IV near the TiAl alloy, the Co, Fe and Ni elements coming from EHEA affect the types of phase, among which Ni element plays a dominant role in reaction with Ti and Al elements. According to Ref. [31], the phase at location D could be brittle τ3-Al3NiTi2 phase with the TiAl phase at location E remained in layer III. TiAl and Ti3Al phases at location F in layer IV were just matrix of TiAl alloy, and τ3-Al3NiTi2 phase also grew from layer III into layer IV.
Fig. 7.
Fig. 7.
Growth of diffusion layers in Zone A and Zone B at different temperatures: (a) Zone A at 900 °C; (a1) Zone A at 950 °C; (a2) Zone A at 1000 °C; (b) Zone B at 900 °C; (b1) Zone B at 950 °C; (b2) Zone B at 1000 °C.
Fig. 8.
Fig. 8.
Results of line scanning analysis in Zone A and Zone B at 950 °C.
Table 2
Chemical composition (at. %) analysis at the locations marked in
Locations | Al | Co | Cr | Fe | Ni | Ti | Nb | Possible phases |
---|---|---|---|---|---|---|---|---|
A | 8.9 | 14.2 | 34.0 | 21.5 | 18.2 | 2.8 | 0.4 | FCC |
B | 22.6 | 12.7 | 16.9 | 14.2 | 29.7 | 3.7 | 0.3 | B2 |
C | 22.9 | 14.8 | 1.7 | 7.8 | 26.8 | 24.1 | 1.8 | Al(Co, Ni)2Ti |
D | 36.3 | 3.6 | 2.0 | 6.8 | 14 .4 | 31.6 | 5.3 | τ3-Al3NiTi2 |
E | 33.2 | 7.1 | 0.6 | 8.1 | 9.0 | 39.5 | 2.4 | TiAl |
F | 34.6 | 1.3 | 0.6 | 2.3 | 5.2 | 50.3 | 5.6 | TiAl + Ti3Al |
Fig. 9.
Fig. 9.
XRD analysis of facture surface on EHEA side.
As for Zone B, the rugged diffusion layer and convex structure was caused by the Kirkendall effect. Fig. 8 shows that the concentration gradient of Al in the diffusion zone near FCC is much higher than that near B2, indicating that the diffusion coefficient of Al in the diffusion zone near FCC is greater than that near B2. The unequal diffusion coefficient results in the different growth rate of Al(Co, Ni)2Ti phase. The rapid growth of Al(Co, Ni)2Ti phase in the diffusion zone near FCC accelerates the movement of the interface between layer II and layer III, while the movement of this interface in the diffusion zone near B2 is slower, thus causing the rugged and convex structure in the diffusion zone between B2 phase and TiAl alloy. In addition, the higher concentration gradient of Ni in the diffusion zone near B2 benefited the growth of layer III, which also contributed to the forming of convex structure. With the increase of temperature, the difference in the thickness of layer II in Zone A and Zone B gradually decreased and the convex structure became unobvious as shown in Fig. 7. When temperature raised to 1000 °C, the alloying elements largely diffused from Zone A to Zone B along the layer II parallel to the interface, thus accelerating the growth of Al(Co, Ni)2Ti phase in Zone B and decreasing the difference of thickness of layer II in Zone A and Zone B. Furthermore, the rapid growth of Al(Co, Ni)2Ti phase in Zone B resulted in the disappearance of convex structure. In fact, the interface between layer I (rodlike B2 phase) and Al(Co, Ni)2Ti phase is the initial faying surface of EHEA, while the interface between Al(Co, Ni)2Ti phase and τ3-Al3NiTi2 phase is the initial faying surface of TiAl alloy. The sluggish diffusion effect of EHEA restrains the interdiffusion of alloying elements to a certain extent. For Al-Co-Cr-Fe-Ni system, the sequence of the slowest to fastest component is: Ni<Co<Fe<Cr, but the energy of activation of Ni is the lowest [32]. The diffusion coefficient of Al cannot be correctly estimated due to the absence of the concentration gradient. However, Al has lower activation energy value of self-diffusion (142.0 kJ/mol) and this decreases the average level of activation energy of diffusion in the lattice of the present system [33]. In this research, contrary to the preceding results, the diffusion of Al and Ni is more obvious than that of Co, Cr and Fe in EHEA. The reason of this phenomenon is that reaction diffusion accelerates the diffusion of Al and Ni and generates Al(Co,Ni)2Ti phase in the interface. Furthermore, the high concentration gradient of Al and Ni between EHEA and TiAl alloy is also beneficial to their diffusion. Thus, reaction diffusion between EHEA and TiAl alloy is mainly controlled by the diffusion of Al and Ni elements. A lot of Al atoms diffused from TiAl alloy to EHEA, resulting in the coarsening of acicular B2 phase precipitated in FCC matrix. Meanwhile, Ni element diffused from EHEA to TiAl alloy. With the decrement in concentration of Ni element, Al(Co, Ni)2Ti phase formed in the layer II, which is possibly τ4-AlNi2Ti phase with Co element dissolved in. τ3-Al3NiTi2 phase generated in layer III with less Ni element dissolved in TiAl phase. Layer IV possibly used to be the matrix of TiAl alloy, but turned into τ3-Al3NiTi2, Ti3Al and TiAl phases due to the diffusion of Ni element.
The TEM test was conducted to further analyze the diffusion behavior of atoms and the formation of phases. The position of TEM sample was marked in red box as shown in Fig. 7(a1). Fig. 10 exhibits the distribution of elements in diffusion zone detected by the EDS in TEM. The segregation of Co and Ni occurred in layer II and the punctate phase dispersed on layer II is rich in Cr, Fe, Nb. Fig. 11 shows the TEM images and SAED patterns of diffusion zone. From the dark field image, it can be seen that the interface between layer II and layer III is obvious while the interface between layer III and layer IV is ambiguous due to the grains in layer III growing into layer IV. There are lots of fine grains near the interface between layer II and layer III, indicating that the nucleation of Al(Co, Ni)2Ti phase occurred here. The SAED patterns of positions 2, 3 and 4 presented in Fig. 11(c) are corresponding to the single-phase FCC, τ3-Al3NiTi2 and Ti3Al, respectively. However, the superlattice spots emerges in the SAED pattern of position 1. According to the Ref. [34], the SAED pattern along [110] zone axis indexed is in accord with the structure of L21-Ni2TiAl, suggesting that the phase at position 1 could be AlNi2Ti phase. Meanwhile, the crystal structure of AlCo2Ti is same as that of AlNi2Ti phase, which means it is possible for Co atoms to take the place of Ni in AlNi2Ti to form the hypothetical phase Al(Co, Ni)2Ti. In consideration of the SAED pattern and the composition of location C in layer II, it can be firmly demonstrated that the Al(Co, Ni)2Ti phase formed in layer II during diffusion bonding process.
Fig. 10.
Fig. 10.
Map of elemental distribution in diffusion zone analyzed by TEM equipped with EDS.
Fig. 11.
Fig. 11.
TEM images of diffusion zone at 950 °C: (a) the bright field image (BFI), (b) the dark field image (DFI) and (c) the SAED patterns of marked positions.
To further reveal the effect of temperature on the diffusion behavior of reaction layers, the fitted growth curve and growth activation energy were investigated. For diffusion bonding, the relationship between thickness of diffusion layers and reaction time can be described as following equations [35]:
where w is the thickness of diffusion layers, k is the growth rate, k0 is the constant of growth rate, t is the reaction time, Q is the growth activation energy and R is the ideal gas constant of 8.314 J/(mol∙K). In this work, the reaction time is constant of 0.5 h and the bonding temperature is variable. By combining Eqs. (1) with (2), a new equation about the relationship between thickness of diffusion layers and temperature can be deduced as follows:
Fig. 12 shows the fitting results about the growth behavior of reaction layers. Almost all the reaction layers in Zone A and Zone B grew in a parabolic manner as a function of temperature. Reaction layer of rodlike B2 phase in Zone A has the lowest growth activation energy of 125.2 kJ/mol, indicating that it grew first when the bonding temperature rising up. And Al(Co, Ni)2Ti phase in Zone A is easier to grow than it in Zone B. In Zone B, the growth activation energy of τ3-Al3NiTi2 phase is lower than that of Al(Co, Ni)2Ti phase which means τ3-Al3NiTi2 phase is preferential to grow up, resulting in the convex structure of diffusion layers in Zone B. Moreover, the high content of Ni element in B2 phase is beneficial for the growth of τ3-Al3NiTi2 phase in Zone B. Thus, the growth activation energy of τ3-Al3NiTi2 phase in Zone B is much lower than that in Zone A. The growth activation energy of layer IV (τ3-Al3NiTi2 + Ti3Al + TiAl) in two zones are both high.
Fig. 12.
Fig. 12.
Fitted growth curve and growth activation energy of different layers in Zone A and Zone B: (a) growth curve in Zone A; (b) growth curve in Zone B;(c) growth activation energy in Zone A; (d) growth activation energy in Zone B.
3.4. Mechanical properties and fracture morphology
Fig. 13 displays the evolution of shear strength as a function of bonding temperature. The shear strength of joints bonded at 950 °C is the highest among all the joints. The reasons for the highest shear strength achieved at 950 °C are summed up in two aspects. First, the penetration and convex structure generated in bonding interface actually forms interlocking effect contributing to the enhancement of shear strength. The joints bonded at 900 °C have weak interlocking effect due to insufficient element diffusion, which also corresponds to the low shear strength of joints. Another reason is the thickness of diffusion zone. The thickness of reaction layers almost dominates the mechanical properties of dissimilar joints. Moreover, there exists a critical thickness of IMCs. When the thickness of IMCs exceeds the critical value, the mechanical properties of joints will be deteriorated [36,37]. At 1000 °C, the overgrowth of reaction layers resulted in that the thickness of reaction layers exceeded the critical value and the shear strength of the boned joints was impaired. Apart from this, the disappearance of convex structure at 1000 °C also weakened the interlocking effect and deteriorated the shear strength of bonded joints. While the joints bonded at 950 °C experienced appropriate elemental diffusion, which formed relatively strong interlocking effect and reasonable thickness of diffusion layers.
Fig. 13.
Fig. 13.
Average shear strength of joints as a function of bonding temperature.
Fig. 14 exhibits the fracture path and fracture morphology of the joints after tensile test. The cracks mostly initiated at the interface of layer II/III (Al(Co, Ni)2Ti /τ3-Al3NiTi2 + TiAl), and the detailed fracture paths of joints bonded at different temperatures are displayed in Fig. 14(a-c). For the joints bonded at 900 °C, the cracks initiated at interfaces of layer I/II and II/III due to the weak bonding between Al(Co, Ni)2Ti phase and matrix of alloys, and then secondary cracks extended from Al(Co, Ni)2Ti phase into the B2 phase. The cracks generated at the interface of layer II/III for the joints bonded at 950 °C, while some secondary cracks extended through layer I caused by the fact that coarsening of acicular B2 phase decreased the ductility of FCC phase. The crack propagated into B2 phase was hindered by the precipitated blocky FCC phase. When the bonding temperature increased to 1000 °C, the crack was uniform and generated at layer II/III interface. However, the secondary cracks just extended through layer I and B2 phase due to their inferior mechanical properties. The fracture morphology on TiAl alloy side of joints bonded at 950 °C is shown in Fig. 14(d-f). The macrostructure of fracture surface is almost flat and the magnified pictures indicate a typical characteristic of brittle rupture. The phases at locations G, H and I as shown in Table 3 also confirm the fracture path of joints bonded at 950 °C.
Fig. 14.
Fig. 14.
Fracture path of joints bonded at different temperatures of 900 °C (a), 950 °C (b), 1000 °C (c) and fracture morphologyon TiAl alloy side of resultant joints at 950 °C (d-f).
Table 3 Chemical composition (at. %) of marked locations on fracture surface of TiAl alloy side.
Locations | Al | Co | Cr | Fe | Ni | Ti | Nb | Possible phases |
---|---|---|---|---|---|---|---|---|
G | 38.8 | 3.3 | 1.2 | 4.9 | 14.0 | 31.6 | 6.2 | τ3-Al3NiTi2 |
H | 42.3 | 0.3 | - | 0.4 | 0.1 | 49.9 | 7 | TiAl |
I | 8.1 | 13.4 | 39.7 | 23.8 | 9.2 | 5.1 | 0.8 | FCC |
4. Conclusions
This work investigated the microstructure evolution, diffusion behavior and mechanical properties of the diffusion bonded joints of AlCoCrFeNi2.1 eutectic high entropy alloy (EHEA) and TiAl alloy, and the following conclusions are drawn:
(1) Sound joining of AlCoCrFeNi2.1 EHEA to TiAl alloy was achieved by diffusion bonding. With the temperature rising up, the shear strength of the resultant joints increases first and then decreases, and the highest shear strength of 449 MPa was obtained at 950 °C. The brittle fracture generally initiated at the interface of Al(Co, Ni)2Ti and τ3-Al3NiTi2 + TiAl layers.
(2) Diffusion zone near FCC phase of EHEA was composed of four reaction layers (rodlike B2 phase, Al(Co, Ni)2Ti, τ3-Al3NiTi2 + TiAl, τ3-Al3NiTi2 + TiAl + Ti3Al), but only three layers (Al(Co, Ni)2Ti, τ3-Al3NiTi2 + TiAl, τ3-Al3NiTi2 + TiAl + Ti3Al) near B2 phase. The penetration phenomenon and the convex structure formed in diffusion zone, which resulted in interlocking effect and enhanced the shear strength of the resultant joints.
(3) The thickness of diffusion layers increases in a parabolic manner with bonding temperature rising up. Rodlike B2 phase has the lowest growth activation energy of 125.2 kJ/mol. The growth activation energy of Al(Co, Ni)2Ti layer near FCC matrix of EHEA is lower than that near B2 phase, whereas the growth activation energy of τ3-Al3NiTi2 + TiAl layer near B2 phase is lower than that near FCC matrix.
Acknowledgments
This work was financially supported by the National Natural Science Foundation of China (Nos. 51605075 and 51674060), the China Postdoctoral Science Foundation (No. 2018T110217), the Scientific Research Foundation for Doctor, Liaoning Province of China (No. 20170520375), the Fundamental Research Funds for the Central Universities (Nos. DUT18RC(4)032 and DUT18LAB01), and technically supported by the Collaborative Innovation Center of Major Machine Manufacturing in Liaoning Province.
Reference
/
〈 | 〉 |