In modern engineering applications, monolithic materials often fail to meet the combined demands of electrical conductivity, thermal management, and lightweight structural efficiency. Bimetallic materials, which integrate two dissimilar metals into a single component, have emerged as a practical solution to this challenge. Among various bimetallic systems, Al/Cu bimetals are particularly attractive because they combine the excellent electrical and thermal conductivity of copper with the low density and favorable corrosion resistance of aluminum. By alloying these two metals, components can achieve superior performance in automotive electrical systems, power transmission, busbars, armoured cables, heat exchangers, and mechanical equipment, while simultaneously reducing mass and cost. For example, Al/Cu transition joints used in industrial power distribution benefit from both the conductivity of Cu and the cost-effectiveness of Al.
However, the fabrication of Al/Cu bimetals presents considerable difficulties. Traditional solid-solid bonding methods, such as roll bonding, extrusion cladding, diffusion bonding, and friction stir welding, often require complicated surface preparation, long processing times, or are limited in the geometry of the parts they can produce. In contrast, solid-liquid compound casting methods—which include sand mold casting, permanent mold casting, lost foam casting, and squeeze casting—offer more flexible and cost-effective routes for manufacturing Al/Cu bimetallic components. In such processes, molten aluminum is poured around a pre-positioned solid copper insert. The high temperature promotes interfacial diffusion and metallurgical reaction between the liquid Al and solid Cu, forming a metallurgically bonded interface. Because the liquid metal can conform to complex cavities, compound casting is particularly suitable for near-net-shape components with intricate geometries.
Among the various compound casting methods, lost foam casting (LFC) is especially promising. Lost foam castings are produced by evaporating a polystyrene foam pattern when molten metal is poured into the mold. This technique offers several distinct advantages for bimetallic fabrication. First, the copper insert can be stably embedded in the foam pattern, ensuring accurate positioning during pouring. Second, the decomposition of the foam pattern generates a reducing atmosphere that protects the surface of the copper insert from oxidation, thereby promoting a cleaner interfacial reaction. Third, lost foam casting requires no cores, reduces surface roughness, and is capable of producing complex components with high dimensional accuracy. These unique features make the combination of lost foam casting and compound casting, i.e., lost foam compound casting, an attractive route for the preparation of Al/Cu bimetallic materials.
One of the most critical processing parameters in solid-liquid compound casting is the melt-to-solid volume ratio. This ratio directly affects the thermal field during solidification, the extent of dissolution of the solid insert, and the duration of the interfacial reaction. Despite its importance, there have been relatively few studies that systematically examined the influence of melt-to-solid volume ratio on the microstructure and mechanical properties of Al/Cu bimetals fabricated through lost foam casting. In particular, the underlying interface formation mechanism remains poorly understood. In the present study, I addressed this gap by fabricating Al/Cu bimetallic materials via lost foam compound casting using four different melt-to-solid volume ratios (3:1, 5:1, 7:1, and 9:1). I then characterized the interfacial microstructures, phase constituents, hardness, and shear strength of the resulting Al/Cu bimetals, and I discussed the formation mechanism of the interface in relation to the processing conditions. The findings are expected to provide both theoretical and practical guidance for the production of high-quality Al/Cu bimetallic castings using lost foam technology.
Experimental Procedures
To systematically investigate the influence of melt-to-solid volume ratio on the microstructure and mechanical properties of Al/Cu bimetals, I designed a series of experiments in which T2 copper rods were embedded in foam patterns and cast with A356 aluminum alloy. The experimental geometry is shown schematically in the test setup: a T2 copper rod with a diameter of 10 mm and a length of 50 mm was pre-positioned inside the foam pattern as the solid insert. A356 aluminum alloy was melted in a resistance furnace, degassed and refined, and held at 800 °C before pouring. The entire lost foam casting process was conducted under a vacuum of −0.03 MPa to facilitate foam elimination and improve filling.
In total, I produced four groups of foam patterns with different melt-to-solid volume ratios: 3:1, 5:1, 7:1, and 9:1. The ratio was adjusted by varying the volume of the foam pattern cavity relative to the fixed volume of the copper rod. Because the volume of the copper insert was kept constant, the ratio directly controlled the amount of molten aluminum that would contact the solid copper during pouring. This in turn influenced the heat input, the melting rate of the Cu substrate, and the extent of interfacial reaction between the liquid Al and solid Cu.
The chemical compositions of the alloys used in this study are listed in Table 1. The T2 copper is essentially pure copper with trace amounts of Fe and Pb, while the A356 alloy is a typical Al-Si hypoeutectic alloy containing 6.81% Si and minor additions of Ti, Fe, and Mg.
| Alloy | Si | Ti | Fe | Mg | Pb | Cu | Al |
|---|---|---|---|---|---|---|---|
| A356 | 6.81 | 0.017 | 0.205 | 0.439 | − | − | Bal. |
| T2 Cu | − | − | 0.002 | − | 0.005 | Bal. | − |
Before assembly, the copper rods were subjected to a rigorous surface treatment procedure. Each rod was first ground with 2000-grit sandpaper to remove the outer oxide layer and then ultrasonically cleaned in acetone for 15 minutes. Following this degreasing step, the rods were acid-pickled at room temperature in a mixed solution consisting of 30% H₂SO₄, 40% HNO₃, and 4 g/L HCl for approximately 40 s. This acid treatment served to further remove residual oxide films and to generate a clean, slightly roughened surface. The roughened surface is beneficial for enhancing the mechanical interlocking and metallurgical reaction between the molten aluminum and the copper substrate. The treated copper rods were then combined with the foam patterns to yield the desired melt-to-solid volume ratios, after which the assembled patterns were coated with refractory paint, dried, and embedded in a sand box, ready for pouring.
After the lost foam casting experiments were completed, the Al/Cu bimetallic castings were sectioned transversely at the midpoint of the copper rod. Metallographic samples were prepared using standard grinding and polishing techniques. The microstructures were observed using a Zeiss Gemini 300 field-emission scanning electron microscope (SEM) equipped with energy-dispersive X-ray spectroscopy (EDS) for compositional analysis. The phase constituents of the fracture surfaces were identified using X-ray diffraction (XRD). Shear tests were performed on a Zwick Z1000 universal testing machine to evaluate the mechanical integrity of the interface. The shear strength τ was calculated using the following equation:
$$ \tau = \frac{F}{S} $$
where F is the maximum load recorded during the test and S is the nominal contact area between the aluminum matrix and the copper substrate. The fracture surfaces were subsequently examined by optical microscopy and SEM to identify the failure location.
Results and Discussion
Macroscopic Interface Morphology
Figure 1 shows the macroscopic appearance of the Al/Cu bimetallic samples produced using different melt-to-solid volume ratios. When the ratio was 3:1, the surface of the copper rod did not exhibit noticeable melting or deformation, and no obvious transition layer was observed at the boundary between aluminum and copper. This indicates that the amount of molten aluminum was insufficient to generate a high-temperature condition at the interface capable of promoting diffusion and metallurgical reaction. Under these conditions, the Al/Cu joint was primarily a mechanical interlock rather than a metallurgical bond. Therefore, further detailed microstructural analysis was not performed on the 3:1 samples, and my subsequent investigations focused on the samples with volume ratios of 5:1, 7:1, and 9:1, where a certain degree of metallurgical bonding was observed.
In contrast to the 3:1 ratio, when the melt-to-solid volume ratio was increased to 5:1 or higher, the surface of the copper rod showed clear signs of localized melting, and a distinct transition layer appeared along the interface between the aluminum matrix and the copper substrate. However, this transition layer was not fully continuous at the 5:1 ratio; some regions still exhibited purely mechanical contact. This suggests that at 5:1, the thermal input at the interface was sufficient to initiate local metallurgical reactions, but not strong enough to produce a uniform interfacial bond over the entire contact area. As the melt-to-solid volume ratio increased to 7:1, the metallurgical reaction layer became continuous across the entire interface, indicating that the heat input was now sufficient to trigger complete interfacial melting and diffusion. When the ratio was further increased to 9:1, the transition layer continued to grow in thickness, but more gas porosity was visible in the interfacial region. The presence of porosity at the higher ratio is likely related to the longer solidification time, which allowed more time for gas evolution and coalescence at the interface, as well as the greater dissolution of the copper substrate, which may have trapped gases released during the eutectic reaction.
Figure 2 presents the quantitative evolution of the average thickness of the transition zone and the melting rate of the copper substrate as a function of melt-to-solid volume ratio. Both parameters increased monotonically with increasing ratio. The thickness of the interface layer increased because a larger melt volume delivers more heat to the Cu substrate, promoting greater dissolution of copper into the liquid aluminum and enhancing interdiffusion. The enhanced mass transport and longer solidification time at higher ratios allow the interface layer to grow thicker. These observations are consistent with the general expectation that the melt-to-solid volume ratio is a powerful lever for controlling the extent of metallurgical reaction at the solid-liquid interface.
| Melt-to-solid volume ratio | Average thickness of transition zone (μm) | Melting rate of Cu substrate (%) |
|---|---|---|
| 5:1 | ≈ 420 | ≈ 8 |
| 7:1 | ≈ 900 | ≈ 15 |
| 9:1 | ≈ 1500 | ≈ 24 |
Interfacial Microstructure
The microstructural features of the Al/Cu interface region are critical in determining the mechanical properties of the bimetallic joint. SEM observations of the samples produced at the 5:1, 7:1, and 9:1 volume ratios revealed that the interface consists of two distinct zones: an intermetallic compound (IMCs) layer adjacent to the copper substrate and a eutectic reaction layer adjacent to the aluminum matrix. The IMCs layer is further subdivided into three sublayers with different contrasts under backscattered electron imaging, as identified by EDS analysis. Table 3 presents the EDS point analysis results at different positions within the interfacial region. Regardless of the melt-to-solid volume ratio, the three sublayers were consistently identified as Al4Cu9 (Layer I), AlCu (Layer II), and Al2Cu (Layer III), proceeding from the Cu substrate toward the aluminum side. The chemical compositions of these intermetallic phases did not change with the melt-to-solid volume ratio, indicating that the phase identities are determined by local thermodynamic equilibrium rather than by the processing parameters. In addition to these IMCs, the eutectic reaction layer (Layer IV) was found to be composed of α(Al), Al2Cu, and Si phases, as confirmed by EDS.
| Region | Al (mol.%) | Cu (mol.%) | Si (mol.%) | Phase |
|---|---|---|---|---|
| 1 | 37.74 | 62.26 | − | Al4Cu9 |
| 2 | 49.34 | 49.63 | 1.03 | AlCu |
| 3 | 65.04 | 32.03 | 2.93 | Al2Cu |
| 4 | 37.12 | 62.88 | − | Al4Cu9 |
| 5 | 51.13 | 48.87 | − | AlCu |
| 6 | 66.11 | 33.89 | − | Al2Cu |
| 7 | 38.00 | 62.00 | − | Al4Cu9 |
| 8 | 49.55 | 50.45 | − | AlCu |
| 9 | 66.68 | 32.51 | 0.81 | Al2Cu |
The morphology of the IMCs layer varied notably with increasing melt-to-solid volume ratio. At a ratio of 5:1, Layer III (Al2Cu) exhibited an irregular, faceted morphology with fine dendrite-like protrusions extending into the adjacent layer. This faceted growth has been associated with a high degree of constitutional undercooling at the solidification front, which promotes the attachment of atoms at faceted crystal surfaces. When the ratio was increased to 7:1, the Al2Cu layer transformed into a more cellular or columnar growth pattern. The transition from faceted dendrites to cellular morphology can be attributed to a more complete distribution of copper in the liquid due to enhanced convection and diffusion. A more homogeneous copper distribution reduces the degree of constitutional undercooling at the Cu substrate surface, thereby suppressing the formation of dendrite branches. At the 9:1 ratio, the Al2Cu layer showed further thickening due to the extended solidification time, which allowed more time for diffusion-controlled growth of the IMCs layer.
The eutectic reaction layer, Layer IV, also underwent microstructural changes as the melt-to-solid volume ratio was varied. At 5:1, the eutectic structure consisted of relatively fine α(Al)-Al2Cu eutectic colonies. When the ratio was increased from 5:1 to 7:1, the eutectic structure became noticeably coarser, with larger interlamellar spacing. The coarsening is principally a consequence of the slower cooling rate associated with the larger melt volume; a larger mass of molten aluminum releases a greater amount of latent heat, thereby prolonging the solidification time. Slower cooling provides more time for the eutectic lamellae to coarsen by diffusion. However, upon further increasing the ratio to 9:1, the eutectic spacing appeared to decrease again. This unexpected refinement can be explained by the enhanced dissolution of the copper substrate. At the 9:1 ratio, a significantly higher amount of copper dissolves into the liquid aluminum, enriching the eutectic reaction zone in copper. When the copper concentration approaches the eutectic composition, the undercooling required for eutectic solidification is reduced, and the eutectic front advances more rapidly, leading to a finer lamellar structure. Thus, the refinement at 9:1 is a result of the competing effects of cooling rate and liquid composition, with the latter becoming dominant at high copper dissolution levels.
Formation Sequence of Intermetallic Phases
To understand the formation mechanism of the IMCs layer in the Al/Cu bimetals produced by lost foam casting, I employed the effective heat of formation (EHF) model to predict the sequence of phase formation. The EHF model, originally proposed by Pretorius et al., determines which phase is likely to form first based on the effective heat of formation, which is defined for each phase as:
$$ \Delta H’ = \Delta H^0 \cdot \frac{c_e}{c_0} $$
where ΔH⁰ is the standard heat of formation of the phase, cₑ is the effective concentration of the “limiting element” at the interface, typically taken as the concentration at the lowest eutectic point in the binary system, and c₀ is the concentration of the limiting element in the phase of interest. The phase with the highest effective heat of formation is expected to be the first phase to nucleate. Using this model, I calculated the effective heat of formation for Al2Cu, AlCu, and Al4Cu9. The relevant thermodynamic data are summarized in Table 4.
| Phase | Compound composition | ΔH⁰ (kJ·mol⁻¹) | ΔH′ (kJ·mol⁻¹) |
|---|---|---|---|
| Al2Cu | Al₀.₆₆Cu₀.₃₄ | −13.05 | −6.72 |
| AlCu | Al₀.₄₉Cu₀.₅₁ | −19.92 | −6.64 |
| Al4Cu9 | Al₀.₃₄Cu₀.₆₆ | −21.69 | −5.59 |
The calculation results indicate that Al2Cu has the highest effective heat of formation among the three IMCs, suggesting that Al2Cu is the first phase to nucleate at the Cu substrate surface. This prediction is consistent with the SEM observations: the Al2Cu layer is found adjacent to the copper substrate, and primary Si particles are sometimes observed to penetrate through the Al2Cu, AlCu, and Al4Cu9 layers, indicating that Al2Cu formed early on the substrate before the other phases. After the initial formation of Al2Cu, the continued interdiffusion of Al and Cu atoms between the Al2Cu layer and the copper substrate leads to the gradual transformation of Al2Cu into AlCu, followed by Al4Cu9. This sequential transformation was accompanied by the development of Kirkendall voids at the interface between the different IMCs sublayers. The Kirkendall effect arises from the unequal diffusion rates of Al and Cu atoms, which creates a net vacancy flux and leads to void formation in the region where atoms are depleted. The presence of Kirkendall voids was observed at the boundaries between the Al2Cu, AlCu, and Al4Cu9 layers, and these voids can inhibit the further growth of the AlCu layer, explaining the relatively small thickness of Layer II compared with Layers I and III. This solid-state transformation mechanism is consistent with previous findings in the Al-Cu diffusion couple literature.
Interface Formation Process
Based on the microstructural observations and thermodynamic analyses, I propose the following mechanism for the formation of the Al/Cu interface in lost foam castings. The formation process can be divided into several distinct stages:
First, when the molten aluminum comes into contact with the solid copper substrate, a steep concentration gradient is established at the interface. At the high pouring temperature of 800 °C, both Al and Cu atoms begin to diffuse rapidly across the interface. The mutual diffusion lowers the local melting point of the copper surface, creating a region with a composition that is molten or semi-molten at the casting temperature. Consequently, the surface layer of the copper substrate begins to melt, forming a copper-enriched liquid zone adjacent to the solid copper.
Second, further diffusion of copper atoms from the enriched zone into the bulk liquid aluminum leads to the formation of a diffusion layer with a gradually decreasing copper concentration gradient. Upon cooling, the solidification sequence begins within this diffusion layer. Due to the high melting point of silicon and its low solubility in the copper-rich phases, primary silicon particles nucleate early in the solidification process, appearing as small polygonal particles dispersed in the intermetallic layer. Meanwhile, in the copper-depleted regions farther from the interface, α(Al) dendrites nucleate and grow in the liquid, forming the basis of the eutectic reaction layer.
Third, as the temperature continues to drop, Al2Cu nucleates heterogeneously on the surface of the copper substrate. The nucleation of Al2Cu is favored because it has the highest effective heat of formation, as discussed earlier. With further cooling, the Al2Cu layer grows by consuming both copper from the substrate and aluminum from the liquid. When the temperature falls below the eutectic temperature, the remaining liquid in the diffusion layer undergoes a eutectic reaction to form the α(Al)-Al2Cu eutectic structure. Silicon, which is present in the A356 alloy, can either precipitate as primary Si or be incorporated into the eutectic structure, depending on the local composition.
Finally, after the eutectic reaction is complete and the temperature drops further, solid-state diffusion between the Al2Cu layer and the copper substrate continues. This solid-state reaction leads to the sequential formation of AlCu and Al4Cu9 phases. The Al2Cu layer adjacent to the copper substrate is gradually consumed, being first transformed into AlCu and then into Al4Cu9 as the copper concentration increases. The outward diffusion of copper atoms and the inward diffusion of aluminum atoms are not perfectly balanced, resulting in the formation of Kirkendall voids at the interfaces between the newly formed phases. This solid-phase transformation is a relatively slow process and is more pronounced at higher melt-to-solid volume ratios because of the longer cooling and solidification times.
Mechanical Properties
Figure 3 shows the Vickers hardness profiles measured across the Al/Cu interface for samples produced with melt-to-solid volume ratios of 5:1, 7:1, and 9:1. The hardness of the interfacial reaction layer lies in the range of 140–190 HV, which is considerably higher than that of both the aluminum matrix and the copper substrate. This high hardness is attributed to the presence of hard and brittle intermetallic compounds, including Al2Cu, AlCu, and Al4Cu9. These phases possess high intrinsic hardness due to their strong intermetallic bonding and complex crystal structures. The hardness of the eutectic reaction layer is also relatively high compared with the aluminum matrix, primarily because of the fine mixture of α(Al) and Al2Cu eutectic phases, as well as the presence of silicon particles. I did not observe a systematic change in the average hardness of the interface layer among the different melt-to-solid volume ratios, suggesting that the intrinsic hardness of the constituent phases is not strongly affected by the processing parameters.
Figure 4 presents the average shear strength of the Al/Cu bimetallic samples as a function of melt-to-solid volume ratio. At a ratio of 5:1, the shear strength was the lowest, approximately 28 MPa. This low value is consistent with the observation that only partial metallurgical bonding was achieved at this ratio. The remaining regions of the interface were mechanically bonded, which cannot efficiently transfer load between the aluminum and copper substrates, resulting in premature failure. When the melt-to-solid volume ratio was increased to 7:1, the shear strength improved dramatically to approximately 80 MPa. This improvement is the direct result of achieving a fully continuous metallurgical bonding layer without significant defects. At this processing condition, the interface layer is well-developed and provides a strong metallurgical bond between the aluminum and copper matrices.
However, upon further increasing the ratio to 9:1, the shear strength declined to roughly 52 MPa. This reduction in mechanical performance can be attributed to two main factors. First, the thickness of the brittle IMCs layer increases at higher volume ratios. Thick intermetallic layers are more prone to cracking because they accumulate more internal stresses and are more likely to contain microcracks, which then propagate under external loading, leading to a reduction in the overall shear strength. Second, the presence of increased porosity at the interface at the 9:1 ratio, particularly near the copper substrate, provides additional sites for crack initiation and reduces the effective load-bearing area of the interface. These pores are likely formed as a result of the prolonged solidification time, which allowed more gas to accumulate and become trapped at the interface. The combination of a thicker brittle intermetallic layer and increased porosity leads to earlier failure during shear testing.
| Melt-to-solid volume ratio | Shear strength (MPa) | Fracture location |
|---|---|---|
| 5:1 | ≈ 28 | IMCs layer |
| 7:1 | ≈ 81 | IMCs layer |
| 9:1 | ≈ 52 | IMCs layer |
Examination of the fracture surfaces after shear testing provided further information about the failure mode. Optical micrographs of the copper-side fracture surfaces revealed residual pieces of the IMCs layer adhering to the copper substrate, with clear evidence of crack propagation through the intermetallic phases. This observation confirms that the shear failure of the Al/Cu bimetallic joints occurred within the IMCs layer rather than at the interface between the IMCs and the eutectic reaction layer. The XRD analysis of the fracture surfaces for the 7:1 sample further corroborated this finding. On the aluminum side, the fracture surface was found to contain α(Al), Al2Cu, Si, and Al4Cu9, whereas on the copper side, the surface comprised α(Al), Al2Cu, Si, AlCu, and Al4Cu9. The presence of Al4Cu9 on both sides of the fractured sample indicates that the crack propagated through the Al4Cu9 sublayer or through a location that allowed both sides to retain some of the same intermetallic phases. In other words, the IMCs layer is the weakest link in the interfacial structure under shear loading, and the crack path is primarily controlled by the brittleness and microstructural defects of the intermetallic phases.
The fact that fracture consistently occurred in the IMCs layer, even when the interface thickness and porosity varied, suggests that the inherent brittleness of the intermetallic phases is the dominant factor determining the shear strength of the joint. Since the IMCs layers are unavoidable in Al/Cu composite castings, the optimization of the shear strength should focus on controlling the thickness and morphology of the IMCs layer, as well as on minimizing porosity within the reaction zone. The present study demonstrates that a melt-to-solid volume ratio of 7:1 provides the optimal balance between obtaining a continuous metallurgical bond and avoiding excessive growth of the brittle intermetallic layer. This processing condition yields an Al/Cu bimetal with enhanced mechanical performance, which is significant for the practical application of lost foam castings in the production of Al/Cu bimetallic components.
Conclusions
In this study, I fabricated Al/Cu bimetallic materials using lost foam compound casting technology and systematically investigated the influence of the melt-to-solid volume ratio on the interface morphology, phase constitution, and mechanical properties. The key findings are as follows:
(1) I found that no effective metallurgical bonding was achieved when the melt-to-solid volume ratio was 3:1. When the ratio was increased to 5:1, metallurgical bonding only occurred in localized regions of the interface. Starting from 7:1, the metallurgical bonding layer became continuous across the entire interface. The average thickness of the interface reaction layer increased with increasing melt-to-solid volume ratio, reaching the maximum value at 9:1.
(2) The microstructure of the Al/Cu interface in lost foam castings was found to consist of an IMCs layer and a eutectic reaction layer. The IMCs layer comprised three sublayers: Al4Cu9, AlCu, and Al2Cu, in sequence from the copper substrate toward the aluminum matrix. The eutectic reaction layer was composed of α(Al), Al2Cu, and Si phases. The morphology of the Al2Cu layer changed from faceted dendrites at the 5:1 ratio to a cellular structure at the 7:1 ratio, and the eutectic structure first coarsened and then refined as the melt-to-solid volume ratio increased from 5:1 to 9:1.
(3) Using the effective heat of formation model, I determined that the first phase to nucleate at the copper substrate was Al2Cu. During the subsequent solid-state transformation, Al2Cu was progressively converted into AlCu and then Al4Cu9 due to the interdiffusion of Al and Cu atoms, with concurrent formation of Kirkendall voids at the phase boundaries. This solid-state transformation was identified as the dominant mechanism for the formation of the AlCu and Al4Cu9 layers.
(4) The Vickers hardness of the interface reaction layer was measured to be within 140–190 HV, which is significantly higher than the hardness of the aluminum matrix and the copper substrate, due to the presence of the hard intermetallic phases. The shear strength of the Al/Cu bimetals exhibited an initial increase followed by a decrease as the melt-to-solid volume ratio increased. I achieved a maximum average shear strength of approximately 81 MPa at the ratio of 7:1. All the shear-tested samples fractured within the IMCs layer, and the fracture path was mainly controlled by the brittleness of the intermetallic compounds and the presence of interfacial defects.
(5) The findings of this work highlight the importance of melt-to-solid volume ratio as a processing parameter in lost foam compound casting for fabricating Al/Cu bimetals. A moderate ratio of 7:1 offered the best balance between achieving a fully continuous metallurgical bond and limiting the thickness of the brittle IMCs layer, thereby ensuring optimal mechanical performance. These results provide useful practical guidance for the design and fabrication of high-quality Al/Cu bimetallic components by lost foam castings.