The microstructural characterization of diffusion bonding, the appeared intermetallic compounds (IMCs), and mechanical characterization, including micro-hardness and shear strength, of Al/Cu bimetals at the Al/Cu interface, have been studied. As a result, the Al matrix composite reinforced by Cu lattice structure is produced using optimal parameters.
— Microstructural Characterization
Figure 4 shows cross-sections images of the Al/Cu bimetals produced with a casting temperature of T = 680°C, and soaking time of t = 30 s, as well as T = 750°C and t = 30 s. Figure 4 (a-c) show the cross-section of Al-D3Cu, Al-D2.5Cu, and Al-D2Cu bimetals at T = 680°C, t = 30 s, respectively. Additionally, Fig. 4 (d-f) show the cross-section of these bimetals at T = 750°C, t = 30 s.
As shown in Fig. 4 (d-f), the Cu wire cross-section was not observed in the specimens with casting condition of T = 750°C, t = 30 s, in comparison with samples at T = 680°C, t = 30 s, according to Fig. 4 (a - c). Temperature is an important factor in the diffusion process. Atoms diffuse across the interface into the reverse side if the temperature is high enough, according to the Arrhenius equation, Eq. (1), as follows 26:
$$D={D}_{0}{exp}^{(-\frac{Q}{RT})}$$
1
D,\({ D}_{0}\), Q, T, and R are diffusion coefficient, pre-exponential factor, activation energy, temperature of diffusion, and gas constant.
|
Figure 4. SEM images of Al/Cu bimetals at ((a), (b), (c)) T = 680°C, t = 30 s and ((d), (e), (f)) T = 750°C, t = 30 s. |
In this study, the temperature of mold and molten Al were set above the Al melting point (660°C) at two temperatures of 680°C and 750°C to investigate the temperature effect on the Al/Cu interface. As revealed in Fig. 4, the Al/Cu interface became indistinct with increasing temperature, which is related to increased atom diffusion. Through the diffusion of Al and Cu atoms across the Al/Cu interface, intermetallic compounds of Al and Cu were formed in the diffusion layer. The chemical composition, name, and structure of some phases suggested by Murray 15,16 and Liu 27 are mentioned in Table 1. As reported by other researchers, different AlxCuy intermetallic compounds have different formation activation energy (Ef), which Ef of some IMCs mentioned in Table 2. Regarding Table 2, CuAl2 has the lowest Ef among Al-Cu intermetallic compounds. Thus, it appears earlier than the others. Besides, the second lowest activation energy belongs to Cu9Al4. On the other hand, Cu4Al3 has the highest Ef. Hence, it will appear rarely compared to other IMCs 28.
Table 1
Chemical composition, name and structure of some intermetallic in Al-Cu system suggested by Murray 15,16 and Liu 27
Stoichiometry | Symbol | Name | Symmetry | Composition (at. % Cu) |
Cu | \({\alpha }_{Cu}\) | Alpha | FCC | > 80.31 |
Cu9Al4 | \({\gamma }_{1}\) | Gamma 1 | Cubic | 59.8–70 |
Cu4Al3 | \({\zeta }_{2}\) | Zeta 2 | Body-Centered Orthorhombic | 55.2–56.3 |
CuAl | \({\eta }_{2}\) | Eta 2 | Monoclinic | 49.8–52.3 |
CuAl2 | \(\theta\) | Theta | Body-centered Tetragonal | 31.9–33 |
Al | \({\alpha }_{Al}\) | Alpha | FCC | 0-2.48 |
Table 2
Formation activation energies of (Ef) different AlxCuy intermetallic compounds reported in other researches
Activation energy of formation (Ef, eV) | Reference |
CuAl | CuAl2 | Cu9Al4 | Cu4Al3 |
- | 0.78 | 0.83 | - | 28,36 |
- | 1.01 | 1.23 | - | 28,37 |
- | - | - | 1.70 | 28,38 |
0.93 | 0.74 | 0.88 | - | 28,39 |
Pure Cu wire is only observed in the cross-section of the Al-D3Cu specimen at T = 680°C, t = 30 s (Fig. 4 (a)) with a decrease in the diameter of the Cu wire by about 15.6% (from 3 to 2.53 mm) due to solid solution and diffusion bonding. Hence, the substructure and properties of this specimen are investigated in the following. To study Al and Cu diffusion at the Al/Cu interface of this specimen, a line scan near the Al/Cu interface is shown in Fig. 5 (a-b). According to it, the concentration of Cu elements gradually decreases across the Al/Cu interface from the Cu side to the Al side, in contrast to the Al elements. Besides, there is no Al element at the center of the Cu wire, which shows diffusion has occurred just across the Al/Cu interface, and diffusion bonding with a specific thickness is observed. This phenomenon reveals that adequate interdiffusion between Al and Cu elements happens at the diffusion layer and promotes the formation of the interfacial reaction layer 29.
|
Figure 5. (a, c-d) SEM microstructural characterization, and (b) line scan of Al-D3Cu specimen at T = 680°C, t = 30 s. |
Figure 5 (a, c-d) exhibits FE-SEM images of Al-D3Cu bimetal at T = 680°C, t = 30 s with varying magnifications. The diffusion layer, with an average thickness of 1061.08 µm, formed at the Al/Cu interface (Fig. 5 (a, c)). This layer is composed of two main layers: layer I on the Cu side of the Al/Cu interface with a thickness of about 596.98 µm (\({\text{L}}_{1}\)+ \({\text{L}}_{2}\)) and layer II on the Al side of the Al/Cu interface with a thickness of about 464.10 µm (\({\text{L}}_{3}\)). Both are mainly composed of two sublayers, including light and dark gray phases. The amount of dark gray phase on the Al side of the interface layer is more significant than that on the Cu side, unlike the light gray phase. Besides, in contrast to layer II, the light gray phase is the matrix phase in layer I. A magnified image of layer I (red dashed square in Fig. 5 (c)) is shown in Fig. 5 (d). This layer is composed of lamellar eutectic microstructures. Considering previous research, this lamellar eutectic structure was often witnessed at the interface of Al/Cu couples 1,10,30. To investigate the chemical composition of these layers, EDS analysis has been done at the Al/Cu interface layer, and the results of points A, B, C, D, and E (Fig. 5 (c-d)) are mentioned in Table 3. Comparing weight percentages of Al and Cu elements of these points with other research 15,16,31 in Table 1, it was comprehended that points A, B, and E (light gray phase) are related to CuAl2 (\({\theta }\)), intermetallic phase of Al-Cu, since the atomic ratio of Al and Cu was 2:1. The maximum solubility limit of Cu atoms in Al is around 2.5 at%. Regarding the low solubility of Cu atoms in Al, saturation of Cu will occur in Al solid solution 31. Hence, points C and D (dark gray phase) are related to a combination of α-Al and CuAl2 (\({\theta })\) because their EDS results exhibited that Cu atomic percent is much lower than 33.3% 27.
Table 3
EDS analysis of points in Fig. 5
Point | Al (at %) | Cu (at %) | Possible Phase |
A | 56.46 | 43.55 | \({\theta }\)-CuAl2 |
B | 62.92 | 37.07 | \({\theta }\)-CuAl2 |
C | 81.43 | 18.57 | α-Al + \({\theta }\)-CuAl2 |
D | 80.07 | 19.93 | α-Al + \({\theta }\)-CuAl2 |
E | 67.45 | 32.55 | \({\theta }\)-CuAl2 |
To disclose the diffusion layer growth at the Al/Cu interface and the effect of soaking time in the diffusion of atoms, the microstructure of the Al-D3Cu specimen at T = 680°C & t = 15 s condition has been shown in Fig. 6. According to it, the average diffusion layer thickness of this specimen is about 618.18 µm. It is less than the diffusion layer thickness in the Al-D3Cu specimen at T = 680°C & t = 30 s. Results show that the increase in soaking time of the casting process caused an increase in the diffusion depth of Cu and Al atoms. According to earlier research, the diffusion depth and time have a parabolic relationship, as shown in Eq. (2) 26,32:
$$d=k*{t}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$$
2
Where t is diffusion duration, k is the diffusion constant of solute atoms or temperature-dependent growth constant of the interface layer thickness, and d is the diffusion depth of solute atoms or the diffusion layer thickness. According to this equation, by doubling the time, the diffusion depth will be 1.4 times. However, according to experimental results, at a soaking time of 30s, the diffusion layer thickness was about 1.7 times that of 15 s.
At the Al/Cu interface of the Al-D3Cu specimen at T = 680°C & t = 15 s, some voids are seen with a maximum width of about 0.30 mm, which are related to the manufacturing process condition. The EDS analysis results of points in Fig. 6 are listed in Table 4. Comparing the EDS results with Table 1, it is understood that points H, J, and F (on the Cu side of the interface layer) are related to Cu-based solid solution considering the maximum solubility limit of Al in Cu is around 19 at% 27. Moreover, point E is related to the Cu9Al4 intermetallic phase. The magnified image of the interface layer is displayed in Fig. 6 (c-d). Figure 6 (d) is a magnified image of the yellow oval area in Fig. 6 (c). Points D and k are relevant to \({\theta }\)-CuAl2 phase, and the interface layer has a lamellar eutectic structure consisting of \({\theta }\)-CuAl2 and α-Al + \({\theta }\)-CuAl2 phases. Points G and I on the Al side of the interface layer are related to α-Al + \({\theta }\)-CuAl2 and CuAl phases. As shown in Fig. 6 (a), the copper wire diameter decreased by about 8% (from 3 to 2.782 mm) due to solid solution and diffusion bonding, lower than the desired decrease of 10 ~ 20%.
EDS maps of Al-D3Cu specimens at three different conditions are shown in Fig. 7, with primary conditions of T = 750°C & t = 30 s, T = 680°C & t = 30 s, and T = 680°C & t = 15 s to discuss the distribution of Al and Cu atoms across the Al/Cu interface in the Al/Cu bimetals. At the temperature of 680°C (Fig. 7 (b-c), the atoms of Cu and Al diffused across the initial interface. Furthermore, the diffusion depth of atoms and, as a result, the thickness of the interface layer grew by increasing time from 15 s to 30 s. According to former studies, the diffusion depth of Cu atoms is more sensitive to time than Al atoms. Hence, the diffusion layer growth is controlled by Cu atom diffusion in liquid Al 26.
Table 4
EDS analysis of points in Fig. 6
Point | Al (at %) | Cu (at %) | Possible Phase |
A | 100 | 0 | Al |
B | 64.58 | 35.42 | \({\theta }\)-CuAl2 |
C | 0.96 | 99.04 | Cu |
D | 56.34 | 43.66 | \({\theta }\)-CuAl2 |
E | 35.03 | 64.97 | Cu9Al4 |
F | 10.76 | 89.24 | Cu-based solid solution |
G | 90.66 | 4.83 | α-Al + \({\theta }\)-CuAl2 |
H | 1.46 | 96.93 | Cu-based solid solution |
I | 46.13 | 53.87 | CuAl |
J | 18.67 | 81.33 | Cu-based solid solution |
K | 68.1 | 31.9 | \({\theta }\)-CuAl2 |
— Mechanical Characterization
As mentioned, the Al-D3Cu bimetal at T = 680°C, t = 30 s had the optimum reduction in copper wire diameter (about 15.6%). Hence, the mechanical properties of Al-D3Cu specimens at T = 750°C, t = 30 s, T = 680°C, t = 30 s, and T = 680°C, t = 15 s were characterized at the Al/Cu interface to understand the effect of temperature and soaking time on their mechanical properties. The micro-hardness (HV) of Al-D3Cu specimens has been assessed across the Al/Cu interface with distances of 20, 50, 100, 200, 400, and 500 µm on both sides of Al and Cu, as shown in Fig. 8 (a). At T = 750°C, t = 30 s, the Al-D3Cu specimen had the highest micro-hardness at the Cu side of the Al/Cu interface and the lowest micro-hardness at the Al side of the Al/Cu interface among all three Al/D3Cu specimens. As shown in Fig. 4 (d), the Al/Cu interface is not distinctive at T = 750°C, t = 30 s and the copper core was completely dissolved during soaking time due to diffusion of atoms. In the specimens at T = 680°C, t = 30 s, and T = 680°C, t = 15 s, there is a sharp increase in the micro-hardness profile at the Al/Cu interface in comparison with the micro-hardness of the Cu side (around 45 HV) and the Al side (around 22.3 HV) of the interface, which is related to micro-hardness of Cu and Al. This is due to the formed intermetallic phases at the Al/Cu interface and observed lamellar eutectic structure, including α-Al + θ-CuAl2 and θ-CuAl2 phases (Fig. 5 and Fig. 6). The maximum micro-hardness in the Al/Cu interface of the Al-D3Cu specimen at T = 680°C, t = 15 s (328 HV) is higher than the specimen at T = 680°C, t = 30s (164 HV). This is due to the observed CuAl, Cu9Al4, and CuAl2 intermetallic compounds at the Al/Cu interface of the Al-D3Cu specimen at T = 680°C, t = 15 s as shown in Fig. 6 and Table 4, while only CuAl2 phase was observed at the Al/Cu interface of the specimen at T = 680°C, t = 30 s (according to Fig. 5 and Table 3). These intermetallic are harder than the Aluminum base and Cu wire according to previous studies, and among them, CuAl, Cu9Al4, and CuAl2 have the highest micro-hardness, respectively 33.
The diffusion layer at the Al/Cu interface and Cu core plays a considerable role in the mechanical properties of bimetals. Hence, a punch test is used to assess the bonding quality of the Al/Cu interface in the Al/Cu bimetals. The shear strength of the interface (\({\tau }_{int}\)) was calculated using the following Eq. (3) 10:
$${\tau }_{int}= \frac{{F}_{max}}{2\pi rl}$$
3
Where l is the specimen thickness (5 mm), r is the insert radius, and \({F}_{max}\) is the maximum load.
Figure 8 (b) shows force-displacement curves of punch test for pure Al and Al-D3Cu specimens at three process condition, i.e., T = 750°C, t = 30 s, T = 680°C, t = 30 s, and T = 680°C, t = 15 s. Firstly, a nearly linear loading happens, and after reaching the maximum load, the load drops. The abrupt descent of the load at T = 680°C, t = 30 s, and T = 680°C, t = 15 s conditions reveals that the debonding of the Cu core and the Al matrix occurs 12. Since the Cu core is completely dissolved in the Al matrix at T = 750°C, t = 30 s due to high temperature and extended soaking time, the load's abrupt descent is not observed in this condition.
The shear strength-strain curves of pure Al and Al-D3Cu bimetals at the Al/Cu interface at different casting conditions, including T = 750°C, t = 30 s, T = 680°C, t = 30 s, and T = 680°C, t = 15 s are shown in Fig. 8 (c). The shear strength increases with strain increment up to reaching maximum strength, then decreases without showing noticeable plastic deformation at T = 680°C, t = 30 s, and T = 680°C, t = 15 s conditions. However, apparent plastic deformation occurs at T = 750°C, t = 30 s, which can be related to the complete solid solution of Cu core in the Al matrix and disappearing the Al/Cu interface in the specimen. According to Fig. 8 (d), the shear strength of Al-D3Cu bimetals at mentioned different casting conditions are higher than pure Al (about 40.36 MPa). The specimen at T = 680°C, t = 30 s condition had the highest maximum shear strength (about 116.81 MPa) among the three Al-D3Cu specimens, which is related to the appeared lamellar eutectic microstructure at the Al/Cu interface including intermetallic compounds. However, the shear strength of the Al-D3Cu specimen at T = 680°C, t = 15 s (about 80.23 MPa) is lower than the specimen at T = 680°C, t = 30 s, which can be related to the lower thickness of the Al/Cu interface layer.
The area under the shear strength/strain curve of Al-D3Cu bimetals, which indicates toughness, is obtained from Fig. 8 (c). It can be calculated from the area under the stress-strain curve through the following Eq. 4 34:
$$W={\int }_{0}^{\epsilon }\sigma \left(\epsilon \right).d\epsilon$$
4
Where W,\(\epsilon\), and \(\sigma \left(\epsilon \right)\) are strain energy, strain, and stress, separately. The highest toughness is related to the specimen at T = 750°C, t = 30s (78.41\(\text{m}\text{J}.{\text{m}\text{m}}^{-3}\)), and the lowest toughness is related to the specimen at T = 680°C, t = 15 s (32.92 \(\text{m}\text{J}.{\text{m}\text{m}}^{-3}\)). Also, the toughness of the specimen at T = 680°C, t = 30 s is 35.64 \(\text{m}\text{J}.{\text{m}\text{m}}^{-3}\). The difference in mechanical properties of the Al-D3Cu bimetals at different casting conditions can be related to the variation in their microstructure. Figure 4 shows that Cu wire completely dissolved in liquid Al at T = 750°C, t = 30 s due to the diffusion of atoms at high temperature and extended soaking time, leading to solid solution formation and disappearance of the interface between Al matrix and Cu wire, as well as the lamellar eutectic structure formation. However, solid solution and lamellar eutectic structure have been formed at the interface of Cu wire and Al matrix through diffusion bonding in the Al-D3Cu bimetals at T = 680°C, t = 30 s and T = 680°C, t = 15 s conditions. This leads to an increase in the micro-hardness of bimetals at the Al/Cu interface through the formation of hard and brittle IMCs, causing a decrease in their toughness 33,35.
By comparing the microstructure and mechanical properties of all Al/Cu bimetals under different casting conditions, it was found that the Al-D3Cu bimetal had the optimum reduction in Cu wire thickness at T = 680°C and soaking time of t = 30 s, as well as a diffusion bonding with noticeable thickness at the Al/Cu interface. This interface layer improved the mechanical properties of the Al/Cu bimetallic composite. Since it is vital to partially dissolve the ligament of the lattice structure in the composite to have appropriate diffusion bonding between the ligament of the lattice structure and the Al matrix, this casting condition was selected to produce the Al matrix composite reinforced by the Cu lattice structure.
Moreover, images of the utilized Cu lattice structure, Al matrix composite reinforced by Cu lattice structure after casting, produced Al/Cu composite after machining, and a cross-section of this composite are shown in Fig. 9 (a-d), respectively. The microstructural and mechanical characterization of the produced Al/Cu composite will be presented in our future publication.