Experimental investigation of phase equilibria in the Cu–Ni–Zr system

Abstract

Phase equilibria in the Cu–Ni–Zr ternary system have been measured through alloy sampling combined with diffusion couple approach. According to the phase relations identified with electron probe microanalysis and X-ray diffraction techniques, isothermal sections at both 1073 and 1293 K were constructed. It is evident that remarkable ternary solubility occurs in almost all binary intermetallic phases at both temperatures. The formerly reported ternary compounds T1 (Cu_20–40Ni_40–60Zr₂₀) and T2 (Cu_20–25Ni_60–65Zr₁₅) were not verified in this work. No other ternary compound was detected. In addition, continuous dissolution between Cu₁₀Zr₇ and Ni₁₀Zr₇ at 1073 K was observed.

Introduction

Owing to the unique mechanical and physical advantages of high elastic strain limits, high strength, and good fracture toughness, etc., bulk metallic glasses (BMGs) have attracted tremendous attention during the past decades. Among such BMGs, the Cu–Zr-based alloys preserve outstanding properties of ultrahigh yield strengths and corrosion resistance, applicable in structural and chemical fields. Due to high GFA in a wide composition range, the binary Cu–Zr alloys often behave as the base alloys for several other BMGs [1–5]. For instance, addition of Ni could further increase the glass-forming ability (GFA) of the Cu–Zr binary alloys [6]. For amorphous metals, their GFA is closely related to thermodynamics of all phases including the intermediate compounds. For instance, the melts with compositions at or close to these low-lying-liquidus surfaces would favor BMG formation. This indicates that, if phase diagrams of ternary alloy systems are known, potential alloy compositions can be identified for BMG formation. In order to design and synthesize new Cu–Ni–Zr BMG more effectively, extensive study of phase equilibria in the Cu–Ni–Zr ternary system is essential.

So far, phase diagrams of the boundary binary systems of the Cu–Ni–Zr system have been well studied. According to Mey [7], the Cu–Ni system is an isomorphous system where only two condensed stable phases, liquid and face-centered cubic (fcc) solid, exist. Phase diagram of the Ni–Zr binary system was first constructed by Nash [8] and then studied by many researchers [9–13]. Recently, Kosorukova [14] reconstructed phase diagram of the Ni–Zr binary system, where eight intermediate phases, i.e., Ni₅Zr, Ni₇Zr₂, Ni₂₁Zr₈, Ni₃Zr, NiZr, Ni₁₀Zr ₇, Ni₁₁Zr₉, and NiZr₂, were involved and all reaction temperatures were more accurately determined. Also, considerable work on phase equilibria in the Cu–Zr binary system has been carried out [15–23]. More recently, the relative stability of intermetallic phases in the Cu–Zr system was experimentally and theoretically investigated by Zhou [22], and the existence of the intermetallic compounds, i.e., Cu₅Zr, Cu₅₁Zr₁₄, Cu₈Zr₃, Cu₂Zr, Cu₁₀Zr₇, CuZr, Cu₅Zr₇, and CuZr₂, were convincingly confirmed.

Phase relations in the Cu–Ni–Zr ternary system have been measured by many groups. Available information about the stable solid phases in the ternary system is summarized in Table 1. Early in 1968, Fedorov [24] investigated phase equilibria of Cu-rich corner, emphasizing on the solubility of Zr in (Ni,Cu) solid solution, while Takeuchi [25] constructed the liquidus surface of Zr-rich corner. Independently, phase relationships in Zr-rich corner were investigated and a partial liquidus surface along with 4 partial isothermal sections at 1073, 1123, 1223, and 1573 K were reported by Vyal [26]. All the above-mentioned data were carefully reviewed and assembled by Ghosh [27]. Later in 2006, isothermal section of the Cu–Ni–Zr ternary system over the entire composition range at 1073 K was established by Liu [28], where 2 ternary phases T1(Cu_20–40Ni_40–60Zr₂₀) and T2(Cu_20–25Ni_60–65Zr₁₅) were reported, but the related phase relationships were not well measured. In addition, the phase boundaries of NiZr and CuZr in NiZr + CuZr phase field were only estimated in Ref. [28]. More recently, phase equilibria in the Cu–Ni–Zr system in the Zr-rich part at 1123 K were investigated once more [29], where phase relationships among CuZr, NiZr, Cu₁₀Zr₇, and Ni₁₀Zr₇ seemed to be in conflict with those in Ref. [28]. Obviously, phase diagrams of the Cu–Ni–Zr ternary system are far from complete. This paper is to measure phase equilibria in the Cu–Ni–Zr ternary system at 1073 and 1293 K.

Table 1 The stable solid phases in three binary systems

Experimental procedure

Alloy sampling along with diffusion couple approach were adopted to measure phase equilibria in the Cu–Ni–Zr ternary system. Copper (99.999 %), nickel (99.98 %), and zirconium (99.9 %), supplied by China New Metal Materials Technology Co., Ltd., were used as raw materials. The alloys were arc-melted on a water-cooled copper plate under argon atmosphere with titanium as getter material placed in the arc chamber. To ensure a good homogeneity of the samples, all samples were turned over before each melting and re-melted at least three times. Weight of each sample was limited to about 5 g and the weight losses did not exceed 1 %. Subsequently, the samples were sealed in evacuated quartz tubes and annealed in an electrical resistance tube furnace at 1073 K for 60 days and at 1293 K for 40 days, respectively. After annealing, the alloys were taken out and quenched in cold water.

To fabricate diffusion couples, Cu–Ni alloys of different composition (Cu₄₀Ni₆₀, Cu₈₀Ni₂₀, China New Metal Materials Technology Co., Ltd) and zirconium rod (99.9 %, China New Metal Materials Technology Co., Ltd) were machined into proper shapes (cuboid with 3 mm × 3 mm × 10 mm and cylindrical shells with rectangular openings), with wire-electrode cutting. The pieces were polished, cleaned, and then put into the cans made of commercial pure Cu. The assembled diffusion couple were then loaded into matching steel bush with a certain taper of the inner wall and then assembled by a hydraulic press machine. The so-obtained diffusion couples were subsequently sealed in an evacuated quartz tube, and annealed at 1073 K for 60 days in diffusion furnace. After annealing, the diffusion couples were taken out and quenched in cold water.

After standard metallographic preparation, the equilibrated alloys and annealed diffusion couples were examined with electron probe microanalysis (EPMA) (JXA-8800R, JEOL, Japan). The compositions reported in this study were the average values of five measurements. Standard deviations of the measured concentration is ±0.6 at.%. The total mass of elements Cu, Ni, and Zr in each phase is in the range of 97–103 %. Using a Cu Ka radiation on a Rigaku D-max/2550 VB + X-ray diffractometer at 40 kV and 300 mA, X-ray diffraction (XRD) was also performed for determining crystal structure of the existing phases in some selected annealed alloys.

Results and discussion

Phase relations at 1073 K

45 alloys were prepared to measure phase relationship at 1073 K. All the samples displayed a three-phase structure or two-phase structure and no more than three phases existed in the annealed alloys, implying that the equilibrium has been reached or nearly reached for the annealed Cu–Ni–Zr alloys. Back-scattered electron (BSE) images and XRD results of typical ternary Cu–Ni–Zr alloys are presented in Figs. 1 and 2. In Fig. 1a, which illustrates microstructure of the alloy Cu₁₀Ni₁₀Zr₈₀, three phases of different contrast exist. According to EPMA, the phases of α-Zr, CuZr₂, and NiZr₂ can be distinguished, consistent with XRD results shown in Fig. 2a. As shown in Fig. 1b, three-phased microstructure Ni₁₀Zr₇ + Ni₂₁Zr₈ + Ni₇Zr₂ occurs in the Cu₁₅Ni₅₇Zr₂₈ alloy, agreeing with XRD patterns (Fig. 2b). For the Cu₄₀Ni₄₃Zr₁₇ alloy, a three-phase equilibrium Ni₅Zr + Cu₅₁Zr₁₄ + (Cu,Ni) was identified as illustrated in Figs. 1c and 2c. BSE micrograph and XRD result of the alloy Cu₅Ni₆₀Zr₃₅ annealed at 1073 K were shown in Figs. 1d and 2d, respectively, both of which are featured with a three-phase equilibrium of Ni₂Zr + Ni₁₀Zr₇ + Ni₂₁Zr₈. It is worth mentioning that the Ni₂Zr phase was formerly regarded as an unstable phase in the binary Ni–Zr system [15, 19], but it can be stabilized by a third element [11]. According to EPMA, the content of Cu in Ni₂Zr can be up to 4.1 at.%, so it is reasonable that Ni₂Zr can be stabilized to 1073 K with Cu dissolving. Similar case also happens to the Cu₂Ti phase, which is unstable in Cu–Ti binary system but can be stabilized at 1023 K with Co substituting for Ti [30]. In addition, Fig. 1e presents microstructure of the Cu₂₄Ni₃₀Zr₄₇ alloy, which contains white NiZr, light gray NiZr₂, and gray Ni₁₀Zr₇, at the same time, the three-phase equilibrium of Cu₅Zr + Cu₅₁Zr₁₄ + (Cu,Ni) was also observed in the Cu₆₀Ni₂₅Zr₁₅ alloy (See Fig. 1f).

Fig. 1

BSE images of typical Cu–Ni–Zr ternary alloys annealed at 1073 K for 60 days. a The Cu₁₀Ni₁₀Zr₈₀ alloy, b the Cu₁₅Ni₅₇Zr₂₈ alloy, c the Cu₄₀Ni₄₃Zr₁₇ alloy, d the Cu₅Ni₆₀Zr₃₅ alloy, e the Cu₂₄Ni₃₀Zr₄₇ alloy, f the Cu₆₀Ni₂₅Zr₁₅ alloy

Fig. 2

X-ray diffraction patterns of typical Cu–Ni–Zr ternary alloys annealed at 1073 K for 60 days. a The Cu₁₀Ni₁₀Zr₈₀ alloy, b the Cu₁₅Ni₅₇Zr₂₈ alloy, c the Cu₄₀Ni₄₃Zr₁₇ alloy, d the Cu₅Ni₆₀Zr₃₅ alloy

In order to confirm phase relations in the Cu–Ni–Zr system at 1073 K, diffusion couples composed of pure Zr and Cu–Ni alloys of two different compositions were further analyzed. Both of the two couples were respectively annealed at 1073 K for 60 days.

The diffusion couple technique for constructing isothermal cross sections of ternary system is based on the assumption of local equilibria at the interface between two phases in the diffusion zone. Details of this technique were given by Jin [31] and discussed by Kodentsov [32]. From BSE images of the diffusion couples shown in Fig. 3, it is clear that extensive interdiffusion across the initial interface between Zr and Cu–Ni alloy developed, resulting in the formation of a series of intermetallic phase layers. In the Cu₄₀Ni₆₀–Zr diffusion couple, 6 intermetallic phase layers were formed, i.e., Ni₅Zr/Cu₅₁Zr₁₄/(Cu,Ni)₁₀Zr₇/CuZr/CuZr₂/NiZr₂ (see Fig. 3a), which were identified through the composition measured by EPMA, while in the Cu₈₀Ni₂₀–Zr couple illustrated in Fig. 3b, only five diffusion layers, i.e., Cu₅Zr/Cu₅₁Zr₁₄/(Cu,Ni)₁₀Zr₇/CuZr/CuZr₂ can be observed without NiZr₂. In the diffusion couple, each interface between two phases represents a two-phase equilibrium in the isothermal section. The tie-lines between the phases in equilibrium in the isothermal section were obtained by extrapolating the composition profile to the phase interface in the diffusion couple. Accordingly, several two-phase equilibria were additionally obtained, which supplement data for constructing the isothermal section at 1073 K.

Fig. 3

Backscatter electron SEM image of the Cu–Ni–Zr diffusion multiples annealed at 1073 K for 60 days. a Cu₄₀Ni₆₀/Zr, b Cu₈₀Ni₂₀/Zr

Constituent phases and their compositions in the equilibrated Cu–Ni–Zr ternary alloys and in diffusion couples at 1073 K are summarized in Tables 2 and 3, respectively. Accordingly, the isothermal section at 1073 K is constructed in Fig. 4. It is evident that, seven three-phase regions were completely determined, including α-Zr + NiZr₂ + CuZr₂, NiZr₂ + CuZr₂ + CuZr, NiZr₂ + NiZr + (Cu,Ni)₁₀Zr₇, Ni₁₀Zr₇ + Ni₂Zr + Ni₂₁Zr₈, (Cu,Ni)₁₀Zr₇ + Ni₂₁Zr₈ + Ni₇Zr₂, Cu₅₁Zr₁₄ + Ni₅Zr + (Cu,Ni), and Cu₅₁Zr₁₄ + Cu₅Zr + (Cu,Ni). Additionally other three-phase regions i.e., NiZr₂ + CuZr + (Cu,Ni)₁₀Zr₇, (Cu,Ni)₁₀Zr₇ + Ni₇Zr2 + Cu₅₁Zr₁₄, and Ni₇Zr₂ + Cu₅₁Zr₁₄ + Ni₅Zr could be further deduced from the determined tie-triangle. Although Cu₈Zr₃ and Ni₃Zr were not detected in this work, the phase relations involving them, i.e., Ni₂₁Zr₈ + Ni₃Zr + Ni₇Zr₃ and (Cu,Ni)₁₀Zr₇ + Cu₈Zr₃ + Cu₅₁Zr₁₄, are still deduced from the obtained phase areas, because phases Cu₈Zr₃ and Ni₃Zr should be stable at 1073 K.

Table 2 Constituent phases and their compositions in the Cu–Ni–Zr alloys annealed at 1073 K

Table 3 Phase equilibria obtained from the diffusion couple after annealed at 1073 K

Fig. 4

Experimentally determined isothermal section of the Cu–Ni–Zr system at 1073 K. (Dashed lines are estimated phase equilibria.)

Compared with the result in Liu [28], the present phase relations show several differences. Firstly, the reported ternary phases T1 (Cu_20-40Ni_40-60Zr₂₀) and T2 (Cu_20-25Ni_60-65Zr₁₅) were not observed in this work. It should be pointed out that the samples used in Ref. [28] were annealed at 1073 K only for two weeks (14 days), much shorter than the 60 days in this work. So, it is thought that T1 and T2 should be non-equilibrium phases at 1073 K.

Secondly, according to the equilibrium compositions of the phases measured with EPMA, there was not a deviation at the end of Ni₁₀Zr₇, and the three-phase region NiZr + Ni₁₀Zr₇ + (Cu,Ni)₁₀Zr₇ deduced by Liu [28] should not exist at 1073 K. Due to that the phases Cu₁₀Zr₇ and Ni₁₀Zr₇ are isostructural, and it is reasonable that a continuous solid solution (Cu,Ni)₁₀Zr₇ can be formed. The relation between Cu₁₀Zr₇ and Ni₁₀Zr₇ in this work agrees with the result of Ref. [29]. What’s more, the three-phase regions NiZr₂ + NiZr + (Cu,Ni)₁₀Zr₇, Ni₁₀Zr₇ + Ni₂Zr + Ni₂₁Zr₈, Cu₅₁Zr₁₄ + Ni₅Zr + (Cu,Ni), and Cu₅₁Zr₁₄ + Cu₅Zr + (Cu,Ni) were first detected, making the isothermal section more comprehensive and credible.

By the way, at 1073 K, the maximum solubility of Cu in NiZr₂, NiZr, Ni₂₁Zr₈, and Ni₇Zr₂ were found to be up to 11.3, 12.5, 13.1, and 20.4 at.%, respectively. Similarly, the Ni solubility in CuZr₂, CuZr, Cu₅₁Zr₁₄, and Cu₅Zr can be up to 7.6 , 22.5, 56.6, 30.4 at.%, respectively. All these data were in agreement with those in Ref. [27, 28]. The only difference happened in the phase of Ni₅Zr, where the solubility of Cu was about 33.2 at.%, much higher than 20 at.% reported by Liu [28]. In addition, all the single-phase regions extended along the isoconcentrate of Zr, implying that Cu and Ni atoms could substitute for each other in the compounds.

Phase relations at 1293 K

Only alloy sampling was adopted here to study the isothermal section of the Cu–Ni–Zr ternary system at 1293 K. According to the Ni–Zr [14] and Cu–Zr [22] binary systems, liquid can be formed in three composition ranges at this temperature i.e., 72–77 at.%Zr for Ni–Zr alloys, 5–13 at.%Zr and 33–72 at.%Zr for Cu–Zr alloys. So, during annealing at 1293 K, several samples were wasted due to melting and reaction with silicon. Figure 5 illustrates BSE images of some typical reliable samples annealed at 1293 K. Phase identification was based on the measured equilibrium composition by EPMA assisted with XRD results. In the Cu₂₅Ni₆₅Zr₁₀ alloy, two-phased microstructure Ni₅Zr + (Cu,Ni) was observed (Fig. 5a), in agreement with the XRD patterns shown in Fig. 6a. Figure 5b shows the BSE micrograph of the alloy Cu₂₃Ni₃₀Zr₄₇, which featured with a three-phase equilibrium of NiZr₂ + Ni₁₀Zr₇ + NiZr, agreeing with the XRD (Fig. 6b). While a three-phase equilibrium Ni₅Zr + Cu₅₁Zr₁₄ + (Cu,Ni) was observed in Fig. 5c for the Cu₃₀Ni₅₀Zr₂₀ alloy, phase equilibrium of Ni₂Zr + Ni₁₀Zr₇ + Ni₂₁Zr₈ occurs in the Cu₅Ni₆₀Zr₃₅ alloy (see Fig. 5d). As shown in Fig. 5e, a two-phase microstructure Ni₁₁Zr₉ + Ni₁₀Zr₇ exists in the Cu₄Ni₅₁Zr₄₅ alloys. The micrograph of the alloy Cu₅₀Ni₂₀Zr₃₀ in the Fig. 5f, where the white phase is Cu₅₁Zr₁₄, the gray phase is Cu₅Zr, and the rest black area is hole, which is deduced to be liquid phase at this temperature. Because the liquid phase solidified to be brittle phase after quenching from the high temperature and got peeled off during grinding, resulting in the holes in the sample.

Fig. 5

BSE images of typical Cu–Ni–Zr ternary alloys annealed at 1093 K for 40 days. a The Cu₂₅Ni₆₅Zr₁₀ alloy, b the Cu₂₃Ni₃₀Zr₄₇ alloy, c the Cu₄₀Ni₄₃Zr₁₇ alloy, d the Cu₅Ni₆₀Zr₃₅ alloy, e the Cu₄Ni₅₁Zr₄₅ alloy, f the Cu₅₀Ni₂₀Zr₃₀ alloy

Fig. 6

X-ray diffraction patterns of typical Cu–Ni–Zr ternary alloys annealed at 1093 K for 40 days. a The Cu₂₅Ni₆₀Zr₁₀ alloy, b the Cu₂₃Ni₃₀Zr₄₇ alloy

Constituent phases and their compositions in alloys at 1293 K are summarized in Table 4, with which the isothermal section of the Cu–Ni–Zr system at 1293 K was established as presented in Fig. 7. There were five three-phase equilibria detected, i.e., NiZr₂ + NiZr + Ni₁₀Zr₇, Ni₁₀Zr₇ + Ni₂Zr + Ni₂₁Zr₈, Ni₁₀Zr₇ + Ni₂₁Zr₈ + Ni₇Zr₂, Cu₅₁Zr₁₄ + Ni₅Zr + (Cu,Ni), and Cu₅₁Zr₁₄ + Cu₅Zr + (Cu,Ni). By the way, a liquid region at 1293 K was estimated according to the Ni–Zr [14] and Cu–Zr [22] binary system. Hence, other three-phase regions including L + NiZr₂ + Ni₁₀Zr₇, Ni₁₀Zr₇ + L + Cu₅₁Zr₁₄, NiZr + Ni₁₁Zr₉ + Ni₁₀Zr, Ni₁₀Zr₇ + Ni₇Zr₂ + Cu₅₁Zr₁₄, Ni₇Zr₂ + Cu₅₁Zr₁₄ + Ni₅Zr, and Cu₅Zr + L + (Cu,Ni), could be further deduced, which are all shown in Fig. 7.

Table 4 Constituent phases and their compositions in the Cu–Ni–Zr alloys at 1293 K

Fig. 7

Experimentally determined isothermal section of the Cu–Ni–Zr system at 1293 K. (Dashed lines are estimated phase equilibria.)

No ternary phase is formed in the two sections, and ternary solubility of the third element in all the binary intermetallic phases vary little with temperature. For instance, the maximum solubility of Cu in Ni₅Zr can be up to 39.5 at.% at 1073 K, while only 33.2 at.% at 1073 K. The most important difference between these two sections at 1073 and 1293 K is whether liquid forms or not.

Conclusions

The phase equilibria of the Cu–Ni–Zr ternary system at 1073 and 1293 K were experimentally determined in the whole composition range. No ternary phase was detected. All binary intermetallic phases show remarkable ternary solubility. A Cu₁₀Zr₇ and Ni₁₀Zr₇ can form a continuous solution at 1073 K. At 1293 K, liquid with a large composition range may form. The newly determined phase relationship provides new data necessary for thermodynamic assessment of the Cu–Ni–Zr system and also helps in design of new Cu–Ni–Zr-based BMGs.