Engineering Materials and Their Applications
Influence of Isothermal Annealing on Mechanical Properties of Cu-Clad Al Wire
2020 Volume 61 Issue 6 Pages 1149-1157
Details
2020 Volume 61 Issue 6 Pages 1149-1157
The influence of isothermal annealing on mechanical properties of the Cu-clad Al (CA) wire was experimentally examined to better understand the annealing mechanisms that occur in the CA wire. In the experiment, the CA wire was prepared by drawing to decrease the diameter d from 10 mm to 2.9, 1.5 and 0.5 mm. The CA wires with these three different diameters were isothermally annealed at temperatures of 423–603 K (150–330°C) for various periods between 10.8 ks and 3456 ks (3 and 960 h). At room temperature, tensile tests were performed on the CA wire using an Instron type testing machine, and hardness tests were made on the Cu layer and the Al core of the CA wire utilizing a micro Vickers hardness testing machine. For the CA wire without annealing, the ultimate tensile strength su is 231, 215 and 198 MPa for d = 0.5, 1.5 and 2.9 mm, respectively, and the elongation eu is 0.4, 0.4 and 1.2% for d = 0.5, 1.5 and 2.9 mm, respectively. Due to isothermal annealing, recovery and recrystallization take place in both the Cu layer and the Al core of the CA wire, and an intermetallic layer consisting of various Cu–Al compounds forms at the Cu/Al interface. The intermetallic layer promotes formation of cavity at the Cu/Al interface during the tensile test. As the thickness l of the intermetallic layer reaches to the critical thickness lm, su attains to the minimum value, and eu takes the maximum value. Here, lm = 1–2 µm for d = 0.5–2.9 mm, respectively. For l > lm, however, su is insensitive to l, but eu decreases with increasing thickness l. Thus, the appropriate combination of the annealing temperature and time is essentially important to realize the optimal mechanical properties of the CA wire.
Fig. 9 The results in Fig. 8(b), 8(d) and 8(f) are represented as open circles, triangles and rhombuses, respectively. The corresponding results reported by Hug and Bellido7) are also shown as solid squares.
Due to high electrical conductivity and connectivity, Cu is widely used as a conductive material in the electronics industry. Although the electrical conductivity for Al is 62% of that for Cu, the mass density for Al is merely 30% of that for Cu.1) Thus, Al has an advantage of light conductive material. In the automobile industry, regulations on fuel efficiency in each country become strict year by year. Though the number of electric wires used in automobiles is increasing, the weight of the automobile is decreasing due to the conversion from Cu to Al.2–4) Recently, to derive the characteristics of various materials, different materials have often been utilized in combination. Thus, many combinations of various materials are particularly used to improve fuel economy through weight reduction. Consequently, cases of joining of Al wires and Cu for use in terminals are increasing. In addition to the method of connecting dissimilar materials, composite materials are built up with wire or plate.
One such composite material is the Cu-clad Al (CA) wire. The CA wire is an Al wire coated with Cu. It is widely used for electronic parts as a conductive material exhibiting both the lightness of Al and the good connectivity of Cu. For the CA wire, Cu and Al are solid-state bonded during the wire drawing process. At the state of only being drawn, almost no intermetallic compound (IMC) exists at the Cu/Al interface. If the CA wire is used in the temperature range of about 100–200°C (373–473 K) near the motor in automobile, however, the IMC layer forms at the Cu/Al interface and grows gradually. Since IMC is brittle, it has great influence on the mechanical properties of the CA wire. Thus, investigating the growth behavior of the IMC layer in such a temperature range is important to ensure the reliability of the product.
The combinations of Cu and Al are often used in solid-state bonding, and hence there are several reports on the relationship between the thickness of the IMC layer and the bonding strength. For instance, Wei and Daud5) experimentally observed such a relationship. In their experiment, Cu/(Al–Si) specimens consisting of pure Cu and a binary Al–1 mass% Si alloy were prepared by a thermosonic wire bonding technique, and then annealed at a temperature of 423 K (150°C) for various times up to 3.6 Ms (1000 h). An IMC layer with thickness of 0.07 µm forms at the Cu/(Al–Si) interface due to the thermosonic wire bonding, and then gradually grows during the annealing. According to their result,5) the thickness l of the IMC layer increases in proportion to the annealing time t, and finally reaches to 0.18 µm at t = 3.6 Ms. During the annealing, the shear force ps of the Cu/(Al–Si) interface increases with increasing thickness l, and takes the maximum value at l = 0.14 µm. However, at l > 0.14 µm, ps edgingly decreases with increasing thickness l.
In contrast, Xie et al.6) prepared Cu-clad Al (CA) sheets by a vacuum roll bonding technique, and precedently annealed them at temperatures of 753–803 K (480–530°C) for various periods of 5–60 s. After the precedent annealing, annealing was further conducted at temperatures of 753 and 773 K (480 and 500°C) for times of 3.6–36 ks (1–10 h). Their result indicates that an IMC layer composed of CuAl2, CuAl and Cu9Al4 is produced at the Cu/Al interface of the CA sheet owing to the precedent and further annealing. Unlike the result reported by Wei and Daud,5) however, the thickness l of the IMC layer is proportional to the square root of the annealing time t. According to the tensile test along the direction parallel to the Cu/Al interface, the ultimate tensile strength su of the CA sheet almost linearly decreases with increasing thickness l up to l = 8 µm within t = 0–60 s. However, the further annealing up to t = 36 ks (10 h) remarkably decreases the value of su.
The mechanical strength of the Cu/Al bond was experimentally examined also by Hug and Bellido7) using CA wires with diameter of 0.3 mm. In their experiment, the CA wires were annealed at a temperature of 573 K (300°C) for various times of 0.6–25.2 ks (10 min to 7 h). During the annealing, an IMC layer consisting of CuAl2, CuAl, Cu4Al3 and Cu9Al4 forms at the Cu/Al interface of the CA wire. Although the dependence of the thickness l of the IMC layer on the annealing time t is not reported by Hug and Bellido,7) they insist that l is proportional to the square root of t. The ultimate tensile strength su of the CA wire decreases almost half at t = 0.2 ks and remains mostly constant at t > 0.2 ks. In contrast, the elongation eu of the CA wire becomes almost five times greater at t = 1.8 ks and rather insensitive to t at t = 1.8–7.2 ks. However, at t > 7.2 ks, eu gradually decreases with increasing annealing time t. Nevertheless, the relationship between the thickness l and the values of su and eu was not discussed in their study.
In the case of the CA wire, the volume fraction of the IMC layer varies depending on the wire diameter, even though the thickness of the IMC layer is equivalent among the wires with different diameters. Therefore, the bonding strength may depend not only on the layer thickness but also on the wire diameter. However, no reliable information on the dependencies of the bonding strength on the wire diameter and the layer thickness is available. To obtain such information, the mechanical properties of the CA wire were experimentally examined in the present study. For the experiment, CA wires with diameters of 0.5, 1.5 and 2.9 mm were isothermally annealed at temperatures of 423–603 K (150–330°C) for various times of 10.8–3456 ks (3–960 h). Tensile tests of the CA wires were conducted using an Instron type testing machine. The relationship among the mechanical properties, the layer thickness and the wire diameter was discussed on the basis of the experimental result. This relationship provides information on the annealing conditions yielding the optimal mechanical properties of the CA wire.
A pure Al rod equivalent to A 1050 was used as the core material and a Cu tape equivalent to C 1020 was longitudinally added; the butt portion was continuously welded, and metal bonding between Cu and Al was achieved by subsequent wire drawing. The thickness of the Cu-clad layer was uniform not only in the circumferential direction but also in the longitudinal direction, and there was no discontinuous section in the welded portion of the Cu tape. The volume fractions of Cu and Al were roughly 0.15 and 0.85, respectively. Using this method, CA wire base material with diameter of 10 mm was produced, and thereafter, wire drawing with reductions in area of more than 90% was performed. After drawing, the CA wires with diameters of d = 0.5, 1.5 and 2.9 mm were obtained, in which Cu and Al were bonded strongly. Here, d is the diameter of the CA wire measured in mm. The CA wires were isothermally annealed in the temperature range of T = 423–603 K (150–330°C) for various periods of t = 10.8–3456 ks (3–960 h). Here, T is the annealing temperature measured in K, and t is the annealing time measured in s.
Tensile tests were performed on the CA wire at room temperature using an Instron type testing machine. For the tensile test, the gauge length was 250 mm, and the cross-head speed was 10 mm/min. After obtaining the nominal stress s versus the nominal strain e (SS) curve, the maximum value of s was used as the ultimate tensile strength su. The elongation eu was calculated from the measured length of the broken pieces divided by the gauge length. The tensile test was performed using three test specimens with the equivalent annealing temperature and time. The average value and the standard error for su and eu were calculated from the results of the three test specimens. Hardness tests were also made on the Cu layer and the Al core of the CA wire at room temperature using a micro Vickers hardness testing machine with load of 10 gf.
Cross-sections of the CA wire were mechanically polished using # 180–1000 emery papers and diamond paste with size of 1 µm. The mechanically polished cross-section was chemically etched with a solution containing ammonia and hydrogen peroxide as the main components. The microstructure of the chemically etched cross-section was observed by optical microscopy (OM) and scanning electron microscopy (SEM). The fracture surface of the CA wire was observed after the tensile test.
Typical SS curves for three different CA wires with diameter of d = 0.5 mm are shown in Fig. 1. Here, Wire A is the CA wire without annealing, Wire B is that annealed at T = 513 K (240°C) for t = 86.4 ks (24 h), and Wire C is that annealed at T = 543 K (270°C) for t = 691.2 ks (192 h). The ultimate tensile strength su and the elongation eu for Wires A, B and C were evaluated from the SS curves in Fig. 1. The evaluation is listed in Table 1. In this table, the Vickers hardness number HV is also indicated for the Cu layer and the Al core of Wires A, B and C. As can be seen, su = 231 MPa and eu = 0.4% for Wire A, su = 129 MPa and eu = 19% for Wire B, and su = 130 MPa and eu = 1.6% for Wire C. Furthermore, HV = 108 and 50 for the Cu and Al, respectively, of Wire A, HV = 55 and 40 for the Cu and Al, respectively, of Wire B, and HV = 58 and 39 for the Cu and Al, respectively, of Wire C. Thus, the Cu and Al of Wires B and C were sufficiently softened by annealing at T = 513–543 K (240–270°C) for t = 86.4–691.2 ks (24–192 h). Such softening causes the values of su = 129–130 MPa for Wires B and C smaller than that of su = 231 MPa for Wire A. In contrast, the value of eu = 1.6% for Wire C is much smaller than that of eu = 19% for Wire B, and it is merely slightly greater than that of eu = 0.4% for Wire A. This means that Wire C is rather brittle. Therefore, at first glance, the brittleness of Wire C seems to contradict softening of the Cu and Al. Consequently, there should exist a certain reason for the embrittlement of Wire C.
The nominal stress s versus the nominal strain e at room temperature shown as solid, dotted and dashed curves for Wires A, B and C, respectively.
The values of su and eu for the CA wires with various diameters annealed at T = 423–603 K (150–330°C) for t = 10.8–3456 ks (3–960 h) are plotted against the annealing time t as different open symbols in Fig. 2. Figure 2(a) and 2(b) shows the results of su and eu, respectively, for d = 0.5 mm; Fig. 2(c) and 2(d) indicates those of su and eu, respectively, for d = 1.5 mm; and Fig. 2(e) and 2(f) represents those of su and eu, respectively, for d = 2.9 mm. In Fig. 2, the vertical axis shows su or eu, and the horizontal axis indicates the logarithm of t.
The ultimate tensile strength su or the elongation eu versus the annealing time t for the CA wire: (a) su and (b) eu for d = 0.5 mm at T = 423–573 K (150–300°C); (c) su and (d) eu for d = 1.5 mm at T = 423–603 K (150–330°C); and (e) su and (f) eu for d = 2.9 mm at T = 423–603 K (150–330°C).
For d = 0.5 mm, su = 231 MPa at t = 0 s as shown in Table 1. However, su monotonically decreases with increasing annealing time t and then reaches to the minimum value sum at t = 3456 ks (960 h) for T = 423–483 K (150–210°C) as indicated in Fig. 2(a). As the annealing temperature T increases, the dependence of su on t becomes more remarkable. Thus, sum = 173, 155 and 134 MPa for T = 423, 453 and 483 K (150, 180 and 210°C), respectively. In contrast, for T = 513 K (240°C), su attains to sum = 129 MPa at t = 86.4 ks (24 h) and then becomes insensitive to t. For T = 543–573 K (270–300°C), sum = 126 MPa is realized at the shortest annealing time of t = 10.8 ks (3 h). In contrast, eu = 0.4% at t = 0 s (0 h) as indicated in Table 1. For T = 423 K (150°C), eu monotonically increases with increasing annealing time t and then achieves the maximum value eum of 0.9% at t = 3456 ks (960 h) as shown in Fig. 2(b). However, eu amounts to eum = 4.3 and 13.6% for T = 453 and 483 K (180 and 210°C), respectively, at t = 691.2 ks (192 h), and then decreases to 3.5 and 3.6% for T = 453 and 483 K (180 and 210°C), respectively, at t = 3456 ks (960 h). For T = 513 K (240°C), eu reaches to eum = 18.5% at t = 86.4 ks (24 h), and then decreases to 13.5% at t = 691.2 ks (192 h). On the other hand, eu attains to eum = 21.5 and 14.9% for T = 543 and 573 K (270 and 300°C), respectively, at t = 10.8 ks (3 h), and then decreases to 1.6 and 2.5% for T = 543 K (270°C) at t = 691.2 ks (192 h) and for T = 573 K (300°C) at t = 86.4 ks (24 h), respectively.
For d = 1.5 mm, su = 215 MPa and eu = 0.4% at t = 0 s. As shown in Fig. 2(c), su monotonically decreases with increasing annealing time t and then reaches to sum = 160, 139 and 123 MPa for T = 423, 453 and 483 K (150, 180 and 210°C), respectively, at t = 3456 ks (960 h) and sum = 116 MPa for T = 513 K (240°C) at t = 691.2 ks (192 h). In contrast, su attains to sum = 118, 114 and 112 MPa for T = 543, 573 and 603 K (270, 300 and 330°C), respectively, at t = 10.8 ks (3 h), and then becomes insensitive to t. As indicated in Fig. 2(d), eu monotonically increases with increasing annealing time t and then reaches to eum = 5.2% at t = 3456 ks (960 h) for T = 423 K (150°C). In contrast, eum = 8.4% at t = 86.4 ks (24 h) for T = 453 K (180°C), eum = 15.3% at t = 691.2 ks (192 h) for T = 483 K (210°C), eum = 16.7% at t = 86.4 ks (24 h) for T = 513 K (240°C), eum = 22.9% at t = 86.4 ks (24 h) for T = 543 K (270°C), eum = 22.0% at t = 10.8 ks (3 h) for T = 573 K (300°C), and eum = 21.5% at t = 10.8 ks (3 h) for T = 603 K (330°C).
For d = 2.9 mm, su = 198 MPa and eu = 1.2% at t = 0 s. Thus, the value of su at t = 0 s decreases with increasing diameter d. The larger the diameter d of the CA wire is, the smaller the strain hardening due to wire drawing becomes. As indicated in Fig. 2(e), su monotonically decreases with increasing annealing time t and then attains to sum = 153, 133 and 109 MPa for T = 423, 453 and 483 K (150, 180 and 210°C), respectively, at t = 3456 ks (960 h). In contrast, su reaches to sum = 99, 99, 98 and 97 MPa at t = 691.2, 86.4, 10.8 and 10.8 ks (192, 24, 3 and 3 h) for T = 513, 543, 573 and 603 K (240, 270, 300 and 330°C), respectively. As shown in Fig. 2(f), eu gradually increases with increasing annealing time t and then amounts to eum = 7.6, 17.6 and 26.9% for T = 423, 453 and 483 K (150, 180 and 210°C), respectively, at t = 3456 ks (960 h). In contrast, eu reaches to eum = 22.8, 27.3, 22.0 and 20.1% at t = 691.2, 10.8, 86.4 and 10.8 ks (192, 3, 24 and 3 h) for T = 513, 543, 573 and 603 K (240, 270, 300 and 330°C), respectively. As the diameter d increases at each annealing temperature, sum decreases, but eum increases. Thus, under the equivalent annealing conditions, the CA wire is more ductile for larger values of d than for smaller values of d.
3.2 MicrostructureTypical cross-sectional microstructure images are shown in Fig. 3. Figure 3(a) and 3(b) indicates the OM and SEM images, respectively, of Wire A; Fig. 3(c) and 3(d) represents the OM and SEM images, respectively, of Wire B; and Fig. 3(e) and 3(f) shows the OM and SEM images, respectively, of Wire C. In the OM images, the bottom side is the Al core of the CA wire, and the outermost curved layer is the Cu layer of the CA wire. On the other hand, in the SEM images, the bottom side is the Al, and the top side is the Cu. As can be seen in Fig. 3(a) and 3(b), the Cu shows polycrystalline microstructure consisting of very fine grains. Such fine-grained polycrystalline microstructure is attributed to heavy plastic deformation due to wire drawing. The heavy plastic deformation causes strain hardening and thus the large value of su and the small value of eu for Wire A shown in Table 1. The grain size of the Cu becomes greater for Wire B in Fig. 3(c) and 3(d) than for Wire A in Fig. 3(a) and 3(b) and for Wire C in Fig. 3(e) and 3(f) than for Wire B in Fig. 3(c) and 3(d). This means that grain growth as well as recovery occurs in the Cu during isothermal annealing. Although each polycrystalline grain of the Al cannot be distinguished in Fig. 3, grain growth and recovery should take place also in the Al. Consequently, as listed in Table 1, the values of HV for the Cu and Al are smaller for Wires B and C than for Wire A.
Cross-sectional OM and SEM images: (a) and (b) Wire A; (c) and (d) Wire B; and (e) and (f) Wire C.
In contrast, an intermetallic layer composed of various Cu–Al compounds is recognized at the original Cu/Al interface for Wires B and C in Fig. 3(d) and 3(f), respectively. In a previous study,8) the growth behavior of the intermetallic layer was experimentally observed in the temperature range of T = 423–543 K (150–270°C). According to the observation,8) the intermetallic layer consists of the α2 (Cu3Al), γ1 (Cu9Al4), δ (Cu3Al2), η2 (CuAl) and θ (CuAl2) phases. Furthermore, the total thickness l of the intermetallic layer is proportional to a power function of the annealing time t as follows.8)
| \begin{equation} l = k \left( \frac{t}{t_{0}} \right)^{n} \end{equation} | (1) |
| \begin{equation} k = k_{0}\exp \left( -\frac{Q_{k}}{\textit{RT}} \right) \end{equation} | (2) |
In addition to the intermetallic layer, a fine-grained region alloyed with Al forms in the Cu matrix neighboring the intermetallic layer for Wire C in Fig. 3(e) and 3(f). This fine-grained region is the Cu-rich solid-solution phase produced by diffusion induced recrystallization (DIR). Here, DIR is the phenomenon that new fine grains with discontinuously different solute concentrations form behind moving grain boundaries owing to recrystallization combined with diffusion of solute atoms along the moving and stationary boundaries surrounding the fine grains.9) In various binary alloy systems, DIR occurs at temperatures where volume diffusion is frozen out but boundary diffusion occurs practically. The occurrence of DIR in the Cu(Al) system was reported also by Gueydan et al.,10) Vandenberg et al.11) and den Broeder et al.12) Here, the notation A(B) shows that a solute B diffuses into either a pure metal A or a binary A–B alloy of the A-rich single-phase according to convention. In a previous study,13) the kinetics of DIR in the Cu(Al) system was experimentally observed at T = 483–543 K (210–270°C), and the observation was theoretically analyzed using a diffusion model. The theoretical analysis satisfactorily reproduces the observation.13)
As previously mentioned, the intermetallic layer is observed at the Cu/Al interface for Wires B and C in Fig. 3(d) and 3(f), respectively. The thickness l of the intermetallic layer is about 1 µm for Wire B in Fig. 3(d) and about 6 µm for Wire C in Fig. 3(f). Thus, l is almost six times greater for Wire C than for Wire B. According to the result in Table 1, eu is mostly one order of magnitude smaller for Wire C than for Wire B, though su is almost equivalent to each other between Wires B and C. Since most of the Cu–Al compounds are brittle, the smaller value of eu may be attributed to the larger value of l.
Longitudinal cross-sectional OM images near the fracture surface after the tensile test were shown in Fig. 4. Figure 4(a) and 4(b) indicates the OM images of Wire A, Fig. 4(c) and 4(d) represents those of Wire B, and Fig. 4(e) and 4(f) shows those of Wire C. Figure 4(b), 4(d) and 4(f) corresponds to the enlargements of Fig. 4(a), 4(c) and 4(e), respectively. As indicated in Fig. 3(a) and 3(b), the intermetallic layer is not recognized at the Cu/Al interface for Wire A. Consequently, for Wire A, sound bonding is realized at the Cu/Al interface, and thus the delamination of the Cu/Al interface hardly occurs even near the fracture surface. In contrast, as represented in Fig. 3(c) and 3(d), the intermetallic layer with thickness of about 1 µm is observed at the Cu/Al interface for Wire B. The intermetallic layer promotes formation of cavity at the Cu/Al interface during the tensile test as shown in Fig. 4(c) and 4(d). The cavity formation at the Cu/Al interface promoted by the intermetallic layer occurs also for Wire C in Fig. 4(e) and 4(f). For Wire C, however, a rather straight crack penetrates the cavities in the intermetallic layer and thus results in the delamination of the Cu/Al interface.
Longitudinal cross-sectional OM images: (a) and (b) Wire A; (c) and (d) Wire B; and (e) and (f) Wire C.
The formation of cavity and/or crack in the intermetallic layer during the tensile test is schematically shown in Fig. 5. Figure 5(a) and 5(b) corresponds to the thin and thick intermetallic layers, respectively. According to the observation in Fig. 4, only small cavities form in the thin intermetallic layer during the tensile test as indicated in Fig. 5(a). In such a case, the cavities are isolated one another and hence hardly influence the ductility of the Cu layer and the Al core in the CA wire. For the thick intermetallic layer, however, the cavities will be connected to one another and then produce cracks. During the tensile test, the crack production proceeds steadily, and hence the number of cracks increases monotonically. As a result, the thick intermetallic layer acts as a crack source. Such a crack source will considerably deteriorate the mechanical properties of the Al core, even though the Al core is sufficiently ductile. Consequently, as shown in Table 1, eu is almost one order of magnitude smaller for Wire C than for Wire B, though su is close to each other between Wires B and C. Thus, the smaller value of eu is surely attributed to the larger value of l.
Schematic for formation of cavity and/or crack in the intermetallic layer during tensile test: (a) thin intermetallic layer, and (b) thick intermetallic layer.
As previously mentioned, for different values of the annealing temperature T, the thickness l of the intermetallic layer is calculated as a function of the annealing time t from eqs. (1) and (2) using the parameters of n = 0.5, k0 = 9.92 × 10−4 m and Qk = 53.3 kJ/mol.8) The calculation is shown as various straight lines in Fig. 6. On the basis of this calculation, the values of su and eu in Fig. 2 are plotted against the thickness l as the corresponding open symbols in Fig. 7. Figure 7(a)–7(f) shows the results correlated with Fig. 2(a)–2(f), respectively. The horizontal axis of t in Fig. 2 is replaced with that of l in Fig. 7.
The ultimate tensile strength su or the elongation eu versus the thickness l of the intermetallic layer for the CA wire: (a) su and (b) eu for d = 0.5 mm at T = 423–573 K (150–300°C); (c) su and (d) eu for d = 1.5 mm at T = 423–603 K (150–330°C); and (e) su and (f) eu for d = 2.9 mm at T = 423–603 K (150–330°C).
For d = 0.5 mm, su reaches to sum ≈ 130 MPa at l ≈ 1 µm independent of T as shown in Fig. 7(a). Although eum takes different values of 13.6–21.5% for T = 483–573 K (210–300°C), eu also attains to eum at l ≈ 1 µm independent of T as indicated in Fig. 7(b). Hereafter, the thickness l corresponding to su = sum and eu = eum is denoted by lm. As can be seen in Fig. 7, lm is rather insensitive to T. This means that the mechanical properties of the CA wire are predominantly determined by the thickness l of the intermetallic layer. As the diameter d increases from 0.5 mm to 2.9 mm, lm slightly increases from 1 µm to 2 µm. Thus, the layer-thickness dependencies of the mechanical properties become lightly less sensitive for larger values of d.
As mentioned earlier, the thickness of the Cu layer in the CA wire is uniform not only in the circumferential direction but also in the longitudinal direction. If this is the case also for the thickness l of the intermetallic layer, the area fraction of the intermetallic layer on the cross section is equal to the volume fraction of that in the CA wire. According to the observation in a previous study,8) the layer growth occurs almost equivalently towards both the Cu and Al sides. In such a case, the volume fraction f of the intermetallic layer in the CA wire is evaluated as follows.
| \begin{equation} f = \cfrac{\pi \left( r_{\text{Al}} + \cfrac{l}{2} \right)^{2} {}- \pi \left( r_{\text{Al}} - \cfrac{l}{2} \right)^{2}}{\pi \left( \cfrac{d}{2} \right)^{2}} = \cfrac{8r_{\text{Al}}l}{d^{2}} \end{equation} | (3) |
The ultimate tensile strength su or the elongation eu versus the volume fraction f of the intermetallic layer in the CA wire: (a) su and (b) eu for d = 0.5 mm at T = 423–573 K (150–300°C); (c) su and (d) eu for d = 1.5 mm at T = 423–603 K (150–330°C); and (e) su and (f) eu for d = 2.9 mm at T = 423–603 K (150–330°C).
As previously mentioned, Hug and Bellido7) experimentally observed the mechanical properties of the CA wire with d = 0.3 mm under the annealing conditions of T = 573 K (300°C) and t = 0.6–25.2 ks (10 min to 7 h). Under such annealing conditions, f was calculated from eqs. (1)–(3). Their values of eu7) are plotted against f as solid squares in Fig. 9. The results in Fig. 8(b), 8(d) and 8(f) are also represented as open circles, triangles and rhombuses, respectively, in Fig. 9. As can be seen, the result for d = 0.3 mm rather well coincides with that for d = 0.5 mm. Thus, we may expect that Fig. 9 is applicable to the CA wire in the range of d = 0.3–3 mm. From the experimental results in Figs. 2, 7 and 8, we can estimate the annealing conditions for the CA wire with the optimal mechanical properties.
To find the adequate annealing conditions for the optimal mechanical properties of the CA wire, the CA wires with diameters of d = 0.5, 1.5 and 2.9 mm were isothermally annealed in the temperature range of T = 423–603 K (150–330°C) for various times of t = 10.8–3456 ks (3–960 h). At ambient temperature, an Instron type testing machine was used to perform the tensile test, and a micro Vickers hardness testing machine was utilized to conduct the hardness test. Before annealing, the ultimate tensile strength su of the CA wire is 231, 215 and 198 MPa for d = 0.5, 1.5 and 2.9 mm, respectively, and the elongation eu of that is 0.4, 0.4 and 1.2% for d = 0.5, 1.5 and 2.9 mm, respectively. Isothermal annealing causes recovery and recrystallization of the Cu layer and the Al core in the CA wire. Furthermore, owing to isothermal annealing, the intermetallic layer consisting of various Cu–Al compounds forms at the Cu/Al interface. During the tensile test, formation of cavity at the Cu/Al interface is promoted by the intermetallic layer. If the thickness l of the intermetallic layer increases to the thickness lm, su reaches to the minimum value, and eu attains to the maximum value. Here, lm = 1–2 µm for d = 0.5–2.9 mm, respectively. When l exceeds lm, however, su becomes insensitive to l, but eu decreases with increasing thickness l. The optimal annealing conditions can be determined on the basis of such information.
