1. INTRODUCTION

The joining of dissimilar metals is widely applied in industrial construction and manufacturing, where the characteristic features of the different metals are optimized for specific applications to result in value addition and cost efficiency [¹]. In particular, Ni/Al dissimilar metal joints can be applied in electronic packaging and electrified automobile since nickel has excellent thermal stability and corrosion resistance while aluminum has low density and excellent thermal dissipation behavior [²-⁵]. However, brittle IMCs easily form at the interface between nickel and aluminum during the joining process of Ni/Al dissimilar metals [⁶]. Similarly, defects tend to readily form at the joint during the joining process of dissimilar metals [⁷]. These defects within brittle intermetallic compounds (IMCs) at the interface can serve as stress concentration zones, accelerating crack propagation and leading to early fracture under mechanical stress [⁸]. Since the defects at IMCs greatly reduce the strength and toughness of the joint, the formation of IMCs should be controlled [⁹]. Achieving defect-free joints is particularly critical in industrial applications where mechanical reliability and cost efficiency are essential, as defective joints can lead to premature failures, increased maintenance costs, and reduced overall performance of components in demanding environments.

In the past decade, additive manufacturing (AM), an emerging technique that allows the fabrication of complex, solid, and functional parts, has gradually gained attention as an alternative to conventional manufacturing techniques [¹⁰,¹¹]. Among the various types of AM technologies, including laser powder bed fusion, wire arc additive manufacturing, electron beam melting, binder jetting, and L-DED, powder-based laser direct energy deposition (L-DED) has been commonly used in industrial settings [¹²]. Three-dimensional components can be built with the L-DED process by directly injecting metallic powder materials on a melt pool produced by a laser. The rapid solidification provided by this process allows the formation of distinct microstructures and phases that cannot be attained through equilibrium cooling [¹³]. Additionally, the highly focused energy input results in a durable metallurgical bond between the deposited material and the substrate while the mechanical properties of the base metal are nearly unaffected owing to the reduced heat-affected zone, especially when the chemical composition of the deposited material is similar to that of the substrate [¹⁴,¹⁵]. In contrast to conventional welding processes such as laser welding, the L-DED process can be applied not only for the joining and repair of components but also for the fabrication of metallic components. The L-DED process is particularly well-suited for fabricating complex geometries such as lattice structures, compared to conventional welding processes. One of the notable advantages of AM is its capability to join dissimilar metals through precise control of process parameters, are thereby overcome the limitations of conventional welding processes, such as the formation of brittle IMCs, interfacial cracking, and poor bonding strength. This approach can significantly enhance the performance of dissimilar metal joints and broaden their applicability within various fields. For instance, in the automotive industry, the combination of high-strength steels and aluminum alloys can simultaneously achieve crash safety and light weight. In biomedical applications, orthopedic implants often require site-specific properties depending on their functional demands. By applying dissimilar metals, desired localized properties can be achieved without using high-cost materials and both functional performance and economic efficiency can be enhanced [¹⁶].

Aluminum and nickel alloys are two of most commonly used metals in the L-DED process. Various aspects of the behavior and performance of aluminum and nickel alloys fabricated through the L-DED process, have been extensively studied. For instance, Zhang et al. [¹⁷] deposited AlSi10Mg using L-DED and found that the increased scanning speed improved tensile properties while laser rescanning significantly reduced the anisotropy of the material. Fu et al. [¹⁸] fabricated an Al-7075 alloy by L-DED and reported that the heat treatment process improved the strength and elongation of the as-printed Al-7075 alloy. Hu et al. [¹⁹] deposited Inconel 625 alloy by L-DED and observed strong anisotropy of the microstructure and the mechanical properties of Inconel 625. The main cause of the anisotropy was the difference in grain boundary strengthening effect and the distribution of Laves phases.

Most prior studies on joining Ni/Al dissimilar metals have focused on conventional welding processes. For instance, Chen et al. [²⁰] performed laser welding of Al-5052 on pure Ni and found that the tensile strength of the weld initially increased and then decreased, with a peak strength of 136.2 MPa. In the weld, IMCs such as Al₃Ni, Al₃Ni₂, and AlNi were observed. Bataev et al. [²¹] fabricated multilayer Ni-Al composites with explosion welding and found that unique metastable phases such as Al₉Ni₂ formed at the interfaces between the layers. However, less attention has been paid to L-DED for Ni/Al dissimilar metals. In particular, there is a significant lack of research on interfacial reactions and defect characteristics of Ni/Al dissimilar joints under varying L-DED process parameters.

Utilizing L-DED for joining Ni/Al dissimilar metals offers several advantages over other joining methods. Geometrically complex structures, such as lattice structures, that are challenging to produce using conventional manufacturing methods can be fabricated. Such complex structures can contribute to component optimization and can be used to manufacture parts with superior engineering performance. Additionally, L-DED provides greater potential for the repair and maintenance of dissimilar metal components. This approach can extend the lifespan of high-value metal components, such as Ni/Al dissimilar metal parts, thereby maximizing their cost-effectiveness and resource efficiency [²²]. To exploit the advantages of manufacturing Ni/Al dissimilar metal components through L-DED, it is crucial to investigate the interfacial and defect characteristics of the joints formed under various L-DED process parameters.

The aim of this study is to fabricate defect-less Ni/Al dissimilar metals through the L-DED process. AlSi10Mg (wt.%) was deposited an on Inconel 625 substrate with different L-DED processing parameters to investigate the interfacial reactions between AlSi10Mg and Inconel 625. The microstructure of the interface was observed to analyze the relationship between the process parameters and interfacial reaction behavior. The microhardness and interfacial tensile strength were measured to analyze the mechanical properties of the interface and the chemical composition of the interface was analyzed with an energy dispersive spectrometer (EDS).

2. MATERIALS AND METHODS

3. RESULTS AND DISCUSSIONS

Cross-sectional OM images of the samples near the interfaces between the Inconel 625 substrate and deposited AlSi10Mg are shown in Fig 3. A phase that was presumed to be the IMC was observed at the interfaces in all samples. The average thickness of the IMC phase for each sample is listed in Table 4.

As seen in Figs 3(a) – (d), delamination cracks were observed at the edge of the interface. These cracks are ascribed to the high residual stress caused by thermal shrinkage of the deposited aluminum and the brittle IMCs interlayer [²²,²³]. Moreover, occasional pores with a spherical shape were observed in the deposited AlSi10Mg. In a previous study, near fully dense aluminum alloys could be produced with the L-DED process parameters used in this study [¹⁵]. Considering this, the pores observed in the aluminum part are determined to be corrosion pits that occurred during the mechanical grinding process. Since the Inconel 625 substrate with relatively high electric potential and low oxygen affinity was electrically connected to the deposited AlSi10Mg, pitting corrosion of aluminum could be accelerated by the galvanic corrosion process. In Fig 3(a), some pores with irregular shapes are also observed in the interface. These pores are likely process-induced defects caused by hot cracking. This type of porosity defect was significantly reduced when the VED was less than 250 W/mm³, as can be seen in Figs 3(b) - (e).

It can be clearly seen that as the laser power is increased and the scanning speed decreased, the interface phase tends to become thicker since the higher VED results in higher heat input. On the other hand, as the VED decreases, the length of the delamination cracks tends to become shorter. At the lowest VED, no delamination cracks were observed (Fig 3(e)). One possible reason for this is process-inherited residual stress, which occurs due to the shrinkage of deposited material during the process followed by localized melting and solidification. The high thermal gradient inherent to the L-DED process, combined with the mismatch in the coefficient of thermal expansion between aluminum and nickel, can result in the development of residual stress along the interface between Ni and Al. This residual stress can concentrate at the brittle interface between aluminum and nickel, promoting delamination cracking in the regions with internal defects[²³].

Fig 4(a) shows the thickness of the IMC phase increased almost linearly as the VED increased. This is presumably because a higher heat energy input promotes the formation of a thicker IMC phase during the L-DED process. Fig 4(b) shows the area ratio representing the proportion of defects relative to the total interface area. These results indicate that the fraction of internal defects at the interface tends to increase with increasing VED.

The microhardness of the deposited AlSi10Mg, the Inconel 625 substrate and the interface phase were measured. Sample 125-600 was used for the measurement. The average microhardness of each part and the micro-indentation marks are shown in Fig 5. The average microhardness of the deposited AlSi10Mg and the Inconel 625 substrate was lower compared to the microhardness of the interface phase. In previous studies, the microhardness of L-DED processed AlSi10Mg and Inconel 625 was found to be approximately 100 and 260 HV, respectively [¹⁵,²¹]. The microhardness of the Ni-Al IMCs fabricated by laser cladding were found to be around 500 – 750 HV [²⁴]. The microhardness values of the deposited AlSi10Mg, the Inconel 625 substrate and the interface in this study are comparable with the microhardness values reported in the previous studies for the Inconel 625, AlSi10Mg, and Ni-Al IMCs. Therefore, the phase observed between the deposited AlSi10Mg and the Inconel 625 substrate is inferred to be Ni-Al IMCs. Significant differences in the microhardness were not detected among the different types of Ni-Al IMCs.

Chemical compositions of the Ni-Al IMC phase were analyzed by EDS line scanning for sample 125-600. From the chemical composition analysis results shown in Fig 7, it was observed that the IMC phase consisted of two different phases. The IMC phase close to the deposited AlSi10Mg is thought to be NiAl since the average atomic percentages of the Al and Ni were about 49.6 and 50.4 %, respectively. The IMC phase close to the Inconel 625 substrate is inferred to be Al₃Ni₅ since the average atomic percentages of the Al and Ni were close to 35.7 and 64.3 %, respectively. These phase identifications are further supported by a Ni-Al phase diagram analysis in addition to atomic composition measurements, as shown in Fig 6. A linear compositional gradient was observed in the Al₃Ni₅ phase, with lower nickel content closer to the NiAl phase and higher content closer to the Inconel 625 substrate, while a nearly uniform chemical composition was observed in the NiAl phase. The same behavior is observed for all the other samples used in this study.

A previous study reported that the diffusion rate of nickel in Ni-Al IMCs is extremely slow, being less than 2 μm per hour at 1,300°C [25]. Considering this, the solid-state growth of NiAl phase due to Ni diffusion directly from the Inconel 625 substrate is not likely to occur. It is more likely that most of the IMC(s) formed when the first aluminum L-DED layer was deposited, due to partial remelting of the top part of the Inconel 625 substrate. It is presumed that when the top part of the Inconel 625 substrate is melted by the laser and the aluminum powders that are injected into the molten pool, the NiAl phase is immediately produced. The latent heat produced during solidification of the NiAl phase, can locally increase the temperature at the NiAl/Inconel 625 interface. This can promote diffusion of Ni from the Inconel 625 to the NiAl phase at a temperature near the melting point of Inconel 625, resulting in the formation of the Al₃Ni₅ phase. Linearly graded distributions of Al and Ni compositions in the Al₃Ni₅ phase strongly support the hypothesis that the NiAl phase is formed first and then the Al₃Ni₅ phase is grown due to solid-state diffusion. The nearly linear correlation of IMC thickness with the VED (Fig 4) also supports the above hypothesis since the increased heat input results in a larger amount of molten Inconel 625 that can react with the aluminum during the L-DED process.

XRD patterns of the interface between Inconel 625 and AlSi10Mg are shown in Fig 8. The surface of the Al10Si1Mg deposit that was in contract with the Inconel 625 substrate was examined after separating it from the substrate. XRD peaks corresponding to the NiAl and Al₃Ni₅ phases, previously identified by the EDS analysis, were again detected, confirming that the IMC phase consists of NiAl and Al₃Ni₅. Additionally, XRD peaks corresponding to Al and Al₂O₃ phases were observed. It is presumed that the detected Al phase is from the AlSi10Mg deposit, while the Al₂O₃ phase is ascribed to the oxidation of aluminum on the exposed surface.

The interfacial strengths of the thin-wall tensile specimens, fabricated with L-DED process parameters of 175-600, 175-720, 175-840, 150-840, and 125-840 are shown in Fig 9. The 175-600 specimen exhibited very low tensile interfacial strength of less than 12 MPa in comparison to the other tensile specimens produced with a lower VED. The longer delamination cracks and relatively severe process-induced defects in the IMC phase in comparison to the other tensile specimens may cause earlier fracture in this case. The specimens processed with VEDs lower than 250.0 W/mm³ (i.e. with the parameters of 150-600, 125-600, 125-720 and 125-840) showed interfacial tensile strength ranging from 26.5 to 34.3 MPa. The specimen processed with lower VED tend to have slightly higher tensile interfacial strength. This is likely due to the higher VED resulting in an increased number of defects and thicker IMC phases at the interface between AlSi10Mg and Inconel 625.

The representative microstructure near the fracture surface of the tensile specimen is shown in Fig 10. The tensile specimen processed with the process parameters of sample 150-600 was used for the observation. From the fractographic analysis, the fracture surface morphologies of the five specimens fabricated under different process parameters were found to be similar, regardless of the process parameters. It clearly shows that the fracture occurs along the IMC phase. The same behavior was observed in all the other thin-wall tensile specimens. The fractography of the tensile sample (sample 150-600) shown in Fig 11 indicates cleavage fracture without any noticeable dimple formation, suggesting that the delamination crack propagated rapidly through the IMC layer. These results indicate that the tensile fracture was caused by the propagation of the delamination crack along the brittle IMC phases. Considering this, the smaller delamination crack near the edge in the specimen with lower VED may be attributed primarily to the higher tensile interfacial strength.

In previous studies[¹⁵,²¹], the tensile strength of L-DED processed AlSi10Mg and Inconel 625 were found to range from 990 to 1,100 MPa for Inconel 625 and around 100 - 300 MPa for AlSi10Mg. These properties are significantly higher compared to the tensile interfacial strength observed in this study, and indicate that relatively low tensile strength of the Ni/Al interface that contains an IMC interlayer. Nevertheless, despite the formation of the IMCs, the interface obtained in this study still exhibits sound tensile interfacial strength with a maximum value of 34.3 MPa for the case with a laser power of 125 W and a scanning speed of 840 mm/min. This shows that the Ni/Al dissimilar metal joint fabricated with the L-DED process can provide reliable mechanical interfaces. The results also indicate that when AlSi10Mg alloy is deposited on Inconel 625 using L-DED, an adequately low VED can produce a dissimilar joint with a defect-less interface. This approach is expected to be beneficial for the Ni-Al multi-material L-DED process and for producing Ni and Al dissimilar joint structures.

4. CONCLUSIONS

In this study, AlSi10Mg was L-DED processed on an Inconel 625 substrate with five different process parameters. The microstructure, chemical composition and mechanical properties of the interface were investigated.

With the deposition of AlSi10Mg, an IMC layer formed between the deposit and the substrate, without critical delamination. Since a higher VED results in increased heat input, the interface phase tended to become thicker and the delamination cracks became longer. In the sample with the lowest VED, no delamination cracks were observed. When the VED was lower than 250W/mm³, process-induced defects were significantly reduced.

In the IMC layer, two IMC phases with different chemical compositions were observed. The IMC phase close to the deposited AlSi10Mg was inferred to be NiAl while the composition of the IMC phase close to the Inconel 625 substrate was close to that of Al₃Ni₅. Considering that the diffusion rate of nickel in Ni-Al IMCs is extremely slow, it was thought that the NiAl initially formed due to the reaction between molten Inconel 625 and the AlSi10Mg powder followed by subsequent formation of the Al₃Ni₅ phase due to solid-state diffusion.

Thin-wall structures consisting of Inconel 625 and AlSi10Mg were successfully fabricated by L-DED. The thin-wall structures were subjected to tensile tests in the tensile direction perpendicular to the interfaces between the two dissimilar metals. It was found that the lower the VED was, the higher the tensile interfacial strength of the joint was. This was likely due to the higher VED resulting in higher heat input, which causes a thicker IMC layer and increases the number of process-induced defects. Despite the formation of brittle IMCs at the interface, the dissimilar joint obtained in this study exhibited reasonably high tensile interfacial strength with a maximum value of 34.3 MPa, obtained under process parameters of a laser power of 125 W and a scanning speed of 840 mm/min.

L-DED deposition of AlSi10Mg alloy on Inconel 625 with a sufficiently low VED could produce a defect-less dissimilar joint. This technique is expected to be useful for fabricating multi-material Ni-Al structures using the L-DED process.