1 Introduction

The joining of different materials is, due to the necessity of employing lightweight design concepts, increasingly gaining in importance. Due to the requirement of reducing weight and costs, novel joining methods and material combinations are, therefore, in the focus of industry and research [1]. Particular attention is paid to welded joints made of aluminum and copper. Among other things, this material combination is required for electrically conductive parts [2]. An important, future-orientated application lies in the field of electro mobility and in the production of high-capacity batteries [3]. Aluminum and also copper are characterized by good electrical conductivity; their specific density and the costs for the materials, however, differ significantly. The substitution of copper with aluminum results, on the one hand, in a weight reduction of the manufactured parts and, on the other hand, in cost savings. The positive-substance joining of the dissimilar materials makes special demands to the welding technique since the unwanted formation of intermetallic phases must be specifically controlled and monitored, Fig. 1. In the diagram, the red lines highlighted the different structures which may form during electron beam welding of aluminum to copper.

Fig. 1
figure 1

Phase diagram Al-Cu binary system [4]. The red lines clarify the different structures which form during electron beam welding of aluminum to copper. M means aluminum-mixed crystal, E means eutectic, and I represents the different intermetallic phases

Full size image

Tests made about the joining with laser beam [5–7] and with friction welding [8–10] show that it is not possible to completely avoid the formation of intermetallic phases. The detected phases are, at that, varying considerably with regard to number and dimension. Current research work about electron beam welding does not exist although this technology is highly suitable for the joining of dissimilar metal welds [11]. The objective of this work is to determine the characterization of the microstructure of Al-Cu joints on butt joint via electron beam welding and also its influence on the strength.

2 Materials and methods

In this study, high-purity aluminum and copper sheets of 3-mm thickness were used. The results from the optical emission spectrometry (OES) and the relevant mechanical properties of both materials are listed in Table 1. We are dealing with pure aluminum (99.5 %) and alloyed copper material with a purity of 99.8 %.

Table 1 OES analysis and mechanical properties of the base material
Full size table

In order to exclude a thermally induced change of the material properties, the specimens which had the size of 50 × 55 × 3 mm have been separated mechanically. For the guarantee of the zero gap, the specimens have been milled over their shorter side, Fig. 2. Immediately before the welding process takes place, the specimens have been cleaned chemically with acetone from adhering grease and impurities.

Fig. 2
figure 2

Schematic representation of EBW process and sheet dimension

Full size image

Samples were welded using a FOCUS MEBW micro electron beam welder featuring a maximum acceleration voltage of 60 kV and a beam current of 33 mA which complies with a maximum power of 2 kW. Since the position of the beam determines the composition of the molten metal significantly, the radius and the caustics must be identified. The dimensions and power density distribution of the electron beam were measured using DIABeam diagnostics system [12]. Figure 3 shows a representative measurement of the dimensions and power density of a 60-W e-beam as function of the beam focus current. To obtain better quality weldments, a near-circular (r xy = 0.5 mm)-shaped focus electron beam was used.

Fig. 3
figure 3

Beam caustic measurement of the used electron beam machine at 60-kV acceleration voltage and 1-mA beam current

Full size image

Electron beam welding of Al-Cu sheets was employed by varying the major process variables such as welding speed, e-beam energy applied per unit length of the weld seam, and beam offset with respect to the Al/Cu interface (shown as “x” in Fig. 2). Defect-free weldments were obtained within a restricted process window, as shown by the blue zone in Fig. 4a. The figure indicates that beam offset values greater than 0.5 mm result in incomplete melting of copper, whereas the offset values smaller than 0.3 mm may lead to formation of intermetallic phases and/or macrocracks. The region where the beam offset is too small is limited by the uncontrolled formation of intermetallic phases, Fig. 11. As from a distance of smaller than 0.3 mm, longitudinal and transversal hardness cracks, due to the formation of intermetallic phases, occur in the weld seam. The offset on the copper side, as in [13, 14], resulted in all cases in the failure of the weld seam during cooling. The asymptotic curve of the energy-per-unit length, plotted in Fig. 4, right, over the welding speed, is explained by the higher process efficiency degree during beam welding with vapor capillary and is in good compliance with [15, 16].

Fig. 4
figure 4

Process window, the “brazing” area is characterized by an unmelted copper sheet, with lower distance of the beam macrocracks occuring in the weld. In the right picture, the energy-per-unit length is shown

Full size image

The examinations of the structures were carried out qualitatively by light and electron beam microscopy and quantitatively by energy dispersive X-Ray spectroscopy (EDS) and different electron backscatter diffraction (EBSD) analyses. The mechanical properties have been determined by hardness measurement in accordance with Vickers and also by tensile tests in accordance with DIN EN ISO 4136 [17]. The tensile tests were carried out by using five specimens in the same section. For the temperature measurements, a thermocouple type K (Ni/CrNi) with a measuring rate of 50 Hz in a distance of 2 mm to the edge of the joint at the top side of the copper sheet is used.

3 Results and discussion

3.1 Microstructure characterization

Figure 5 shows the top and bottom sides of the electron beam-welded Al-Cu joint made by setting a beam offset of 0.4 mm from the Al/Cu interface, towards the aluminum sheet. Complete root penetration and relatively uniform flaking were observed. Macro- and microcracks which normally occur with a smaller beam offset were not detected.

Fig. 5
figure 5

Weld upper side (a) and weld bottom side (b) of an electron beam-welded Al-Cu dissimilar material weld

Full size image

Figure 6a–d shows the backscattered electron image of the weld seam. The three recognizable regions are the two base materials aluminum and copper and also the weld seam which is limited on the aluminum side by the fusion line and on the copper side by a fine phase edge. Weld imperfections such as pores or cracks were not detected. Figure 6b is a magnified image showing Cu/weld seam interface. The weld seam on the left side of the interface features oversaturated aluminum-mixed crystal (M) composed of 3 at.% Cu. This aluminum-rich phase is streaked with copper-rich precipitations. The structure is explained by the fact that due to the decreasing border solubility at decreasing temperature, copper-rich phases are precipitated. Figure 6c shows formation of a layer featuring very fine lamellar structure (region E) adjusted to region M. EDS measurements indicate that region E is composed of 19.9–22.5 at.% Cu; this phase is identified as the eutectic (E) region from aluminum-mixed crystal and Al2Cu (θ-phase) which is also detected in [18] and [19]. The intermetallic phase I shown in Fig. 6d on the copper side has a fine phase edge. For the determination of the different phases, EBSD and EDS analyses are further carried out.

Fig. 6
figure 6

SEM analysis of an electron beam-welded Al/Cu joint. a A characterized overview. b–d The interface of the weld with pure copper at different magnifications

Full size image

Figure 6d shows formation of another intermediate layer “I” between region E and pure Cu. EBSD analysis indicates that this intermediate layer has a tetragonal crystal structure (Fig. 7) which corresponds to that of the θ-phase in the Al-Cu phase diagram. The identified phases are illustrated in different colors. The marking is made by means of the lattice structure which has occurred the most frequent in a region. EBSD mapping also identified aluminum-rich regions in yellow, pure copper regions in red, and Al2Cu-rich regions in green. Due to the homogeneous fcc-lattice structure of aluminum and copper, the differentiation is made by parallel EDS measurements. If a unique assignment to one of the mentioned phases is not possible, these regions are depicted in black. Regions where different lattice structures occur with the same frequency are also depicted in black. This applies to large areas of the eutectic region. The regions of the eutectic structure (E) where one of the lattice structures occurs more frequently are colored accordingly. It is observed that in the vicinity of the aluminum-mixed crystal, more θ-phase contents exist than in the eutectic structure. Total fraction counts all marks within the different phases. Partition fraction only counts the marks which can be clearly assigned to one structure. Sum-up of the total fraction shows that 26.9 % of the diffraction pattern cannot be assigned by the EBSD software.

Fig. 7
figure 7

EBSD color-coded map type. Phase identification. Pure copper dyed in red, Al2Cu in green, and pure aluminum in yellow

Full size image

A detail of the phase edge between Al2Cu and Cu made with EDS analysis is shown in Fig. 8. The maximum border solubility of Al in Cu is 19.7 at.% in accordance with [20]. Thus, the measured concentration is 15 at.% aluminum in the region of αCu. Adjacent, a concentration region of 63–67 at.% Al is measured. According to Fig. 1, the Al4Cu9 (γ1-phase) lies in this interval. The existence of the γ1-phase as a dominant phase beside Al2Cu has also been observed by [8, 21] and [22]. It is not possible to unambiguously clarify the third phase via EDS analysis. The high cooling gradient (compare Fig. 10) in electron beam welding may result in a solidification speed which is so high that meta-stable phases can also be established in the solid state. In [8] and [18], it reported about a phase Al2Cu3 which has been established via TEM and XRD and which, due to its chemical composition, comes into question as the third phase. According to [21, 23] and [24], Al-Cu is developing as the third phase after Al2Cu and Al4Cu9. Gueydan et al. [25] show the boundary layer structure: α-Al-Al2Cu-AlCu-Al4Cu9-α-Cu after thermal aging of copper-coated aluminum wires. While the eutectic structure and the θ-phase are highly uniform over the entire boundary layer, the phase edge is strongly varying between Al2Cu and Cu. The γ1- and the third phase can be detected via REM only in the upper part of the weld seam. The largest part of the boundary layer between Al2Cu and Cu looks similar as in Fig. 6d and consists thus mainly of a copper-mixed crystal.

Fig. 8
figure 8

SEM micrograph at the weld/Cu interface close to the upper surface. Three intermetallic structures can be observed

Full size image

The crystallographic orientation of the individual grains within the regions shown in Fig. 7 can be visualized in color by the inverse pole figure (IPF) orientation map, Fig. 9. The map allows to draw conclusions about the solidification direction and, thus, about the phase growth. Weld seam solidification occurs from the outside to the inside.

Fig. 9
figure 9

EBSD color-coded map type. Inverse pole figure (010) shows the growth direction during the solidification

Full size image

Due to the divergent heat conductivity values (Al99.5 at RT 224.7 (W/mK) [26] and Cu at RT 390 (W/mK)), the solidification is shifted in the direction of the copper. The θ-phase (I) and αAl show a directed growth into the eutectic phase. This must, as a consequence, have solidified as the last region of the weld seam at a temperature of 548 °C. It is not possible to unambiguously specify the growth of the γ1- and of the third phase in this paper. Since these phases are also detected in tests about friction stir welding where the process temperatures are clearly below the melting temperatures of both metals [27, 28], a phase growth during the cooling phase by solid-phase diffusion is also possible. [9] and [24] specify that proceeding from the assumption of a volume diffusion, the phase growth of the Al2Cu phase follows a parabolic growth law.

The temperature gradient which is depicted in Fig. 10 was measured using a thermocouple type K at a distance of 2 mm from the edge of the joint on the copper sheet. Due to the peak temperatures during electron beam welding of aluminum of approximately 1900 °C [29], the tactile measurement of the molten metal itself is not possible. In the diagram, moreover, the first dissipation of the temperature is plotted over the time (dT/dt). Curves which are typical of beam welding with high heating (>1290 °C/s) and cooling gradients (>400 °C/s) can be observed. The time-temperature integral has a significant influence on the formation of intermetallic phases [30, 31]. The specimens which have been examined in this work have intermetallic phase width values of 3–6 μm. An influence of the welding speed which may affect the energy-per-unit length and the cooling gradient has not been observed. The reason may be the small difference of the temperature-time integral of more/higher than 300 °C. The intermetallic phases start forming, at temperatures of 120 to 400 °C [23]. However, within this temperature range, the growth rates of these intermetallic phases were reported to be very slow [9].

Fig. 10
figure 10

Temperature gradient, measured on the copper sheet with a distance of 2 mm to the edge of the joint

Full size image

The influence of the beam offset, however, has significant influence on the seam formation and on the width of the intermetallic phases, Fig. 11. Based on the backscattered image contrast, it can be deduced that as the beam offset decreases, the Cu content of the weld seams becomes greater (brighter gray tones). Al concentrations in the α-mixed crystal are constant at approximately 3 %. The eutectic structure in the weld seam is, however, increasing considerably. In the specimen with the distance of 0.3 mm, agglomeration of the θ-phase occurs. The phase γ1 and the third phase develop, at that, exclusively in the boundary layer to the copper.

Fig. 11
figure 11

SEM micrographs of electron beam-welded Al-Cu seams with different distances towards the edge of the joint

Full size image

3.2 Mechanical properties and fracture behavior

The investigation about the influence of the beam offset on the mechanical properties is carried out via hardness measurement and quasi-static tensile tests. Figure 12a shows the hardness measurements taken across the joints electron beam welded by using different beam offset values of 0.4 and 0.3 mm. The measurements have been carried out centrically across the weld seam. The hardness profile across the weld seam made by using 0.4-mm offset exhibited a linear-like increase indicating an increase in Cu concentration and increased fraction of eutectic structure (α-Al + θ-phase) towards the weld seam/Cu interface. On the other hand, the hardness profile of the weld seam made by using 0.3-mm beam offset can be attributed to even dispersion of the eutectic structure (region E) throughout the weld seam. Figure 12b depicts a hardness measurement with HV0.01 along the dashed line in Fig. 6c. The hardness of the eutectic structure (E) was 276 HV which is consistent with the measurements reported in [28] and [32]. The hardness of the θ-phase region was 551 HV. Minor deviations from the measurements reported in [30] and [33] may be explained by the tongue-shaped morphology of the phase and, caused by that, the lack of the supportive effect.

Fig. 12
figure 12

Hardness measurements HV0.1 across the complete weld (a) and HV0.01 of the interface weld/pure copper (b)

Full size image

The evaluation of the tensile tests is depicted in Fig. 13. The stress-strain diagram shows exemplary two graphs, a welded aluminum specimen and an Al-Cu joint. It is shown that the formation of Al-Cu-mixed crystal leads to higher tensile strength. The maximum tensile strength of 104 MPa has been achieved with a beam offset of 0.4 mm. Almost all tensile specimens failed in brittle fracture tests with an elongation after fracture of maximally 3 %. In the tests with 0.4 mm, some of these specimens show areas of ductile failure. The minor scattering within the individual specimens is explained by the high beam quality and the exact positioning of the beam with the used equipment technique. Larger distances to the edge of joints resulted in the decrease of strength values. As from 0.5-mm distance, the face of the copper sheet is no longer completely molten and a joint is developing which is composed, on the aluminum side of a welded seam and, on the copper side, of a brazed joint. Since these joints did not fit to the requirements of a weld, no further analysis was performed. The influence of the microstructure in the weld seam on the tensile strength is discussed by means of the fracture surfaces.

Fig. 13
figure 13

Stress-strain curve and values of electron beam-welded Al-Cu joints with a welding speed of 15 mm/s over the distance to the weld seam

Full size image

The copper-side fracture surface of a specimen with a distance of 0.4 mm is shown in Fig. 14a, and it can basically be divided into three different regions. Those are clearly differing by their different copper concentrations. The region A1 is increased in Fig. 14b and is characterized by fine lamellar eutectic structure (E) with included mixed crystals. The fracture path in this region leads through a region where both phases exist. In the considered specimen, the distance to the copper is approximately 25 μm. The identification of the phase regions A2 and A3 is carried out in addition to the morphology via EDS-point measurements, Table 2. In Fig. 14c, the middle region is identified as eutectic due to its morphology and copper content of 23 at.% The medium-gray region shows the typical topography of a brittle fracture. The measured copper content is on the copper and also on the aluminum side at 62–63 at.%. This means that the fracture path at this point is going through the intermetallic phase Al2Cu as has also been observed by [21]. The region A3, as shown in Fig. 14c, exhibited the highest Cu concentration and thus can be identified as Al4Cu9. A fine-grained structure is observed which exists also on the aluminum side, Fig. 14d. The EDS measurements show, however, that at this point, we are dealing with Al2Cu. The crack path may thus be in the boundary layer between the γ1- and the θ-phases. According to [23], a phase width Al4Cu9 smaller than 100 nm allows to produce good metallurgical joints between Al and Cu. The phase width values which were detected in this phase are similar.

Fig. 14
figure 14

Fracture surfaces of a specimen with a distance of 0.4 mm. a Overview (Cu). b Magnification region A1 (Cu). c Magnification region A2 and A3 (Cu). d region Al

Full size image
Table 2 EDX measurement of the fracture surfaces
Full size table

The weld seams 0.4 and 0.5 mm consist mainly of the aluminum-mixed crystal which was dispersed with eutectic precipitations. The exception is the approx. 5 % of the copper-side weld seam which consists of the specified structure (E) + (I), Fig. 6c. The proportion of the eutectic precipitations is in the specimens with a distance of 0.4 mm higher than in those with a distance of 0.5 mm. A possible explanation for the higher strength values is the ductility of the eutectic structure which can reduce internal stresses by plastic deformation. Highly deformed and recrystallized regions in the microstructure show typically high values of the local misorientation [34]. These misorientations can be represented with EBSD KAM maps (Kernel Average Misorientation) by the comparison of crystal orientation of two neighboring measuring points, Fig. 15. Green/red represents regions with residual stresses while the blue regions have low stresses. The eutectic structure shows increased but, however, homogeneous stress distribution. The low stress within the intermetallic phase Al2Cu is remarkable. The highest stresses are in the region of the copper where the grain boundaries of the intermetallic meet one another.

Fig. 15
figure 15

EBSD color-coded map type. Kernel Average Misorientation shows the internal stresses

Full size image

The fracture of this specimen may thus have occurred under tensile load, as in Fig. 16. Crack initiation has occurred at highly deformed grain boundaries within the intermetallic phase (1). The propagation is, in the beginning, along the boundary line between the γ1- and the θ-phases and subsequently through the θ-phase (2) into the eutectic phase region [12] (3). This means that the binding between Al2Cu and Al4Cu9 is worse than that between Al4Cu9 and the Cu-mixed crystal.

Fig. 16
figure 16

Possible fracture mechanism of an electron beam-welded Al-Cu dissimilar joint

Full size image

4 Conclusion

Sound welds of the material combination aluminum-copper were produced by micro electron beam welding. It has been shown that the controlling of the melting ratio of metals is an important factor for defect-free welding of these dissimilar metals. The ratio can be controlled by the beam position. A deflection of 0.4 mm on the aluminum sheet achieves a tensile strength of 104 MPa which is higher compared to a similar aluminum weld.

Due to locally restricted energy of the high-power electron beam, the heat input can be controlled which directs to short interaction times. Thus, an uncontrolled formation of brittle intermetallic can be avoided.

The study of the fracture surface permits the presentation of a possible fracture mechanism of Al-Cu dissimilar joints. After that, the crack starts in highly deformed regions inside the intermetallic layer.

Besides demonstration of the influence of the microstructure to the mechanical properties, the results of the present study can be used for further investigations, like the corrosion, crash, and electronic behavior of this combination.