Microstructure and properties of ultrasonic-assisted gallium-based alloys for room-temperature bonding of Cu/Cu

ABSTRACT Room-temperature welding with gallium is regarded as a new strategy because of the low melting point of gallium and its affinity for forming intermetallic compounds with other metals. Benefitting from liquid gallium-based solder and an ultrasonic drive, the establishment of Cu/Cu diffusion-bond connections can be quickly completed. The underlying characteristics for the weld quality of room-temperature bonding were preliminarily clarified from the perspectives of welding process parameters and the solder composition. To further determine these characteristics, the research directions including features of welding product formed by ultrasonic-driven room-temperature bonding, the interface microstructure, the element distribution and the shear strength of the weld joint were then determined.


Introduction
Ultrasonic welding is considered to be the fastest method of connecting materials, owing to its characteristics of simple operation, low-temperature, small deformation of the materials and environmental friendliness [1][2][3][4].As for suitable industrial applications, ultrasonic welding can not only realise the connection of the same and dissimilar metals but also achieve welding between composite, ceramic materials, as well as difficult-to-connect metals [5].At present, when it comes to ultrasonic welding that requires solder, it is necessary to heat the solder before welding, and there is no evidence for room-temperature solder.Room-temperature welding can avoid various problems associated with high temperatures, such as component warpage, degradation of temperature-sensitive devices, etc. Elaborate exploration and rational development of room-temperature liquid solder are essential to realise room-temperature welding.
Gallium's liquid fluidity, its metallic properties and its low melting point (302.95K), have been utilised to provide a new direction in welding.The melting points of GaIn and GaInSn are 288.55K and 286.35K, respectively [6].These alloys can easily form intermetallic compounds with most metals (Cu, Au, Mg, Al, Ni and Co) [7][8][9][10].In this work, in order to obtain roomtemperature solidification and enhance bonding strength, ultrasonic treatment is introduced as an assistant in the welding process.Ultrasonics can not only remove the oxide film but can also provide a driving force for solder diffusion and accelerate the reaction [11][12][13].At present, various researches on ultrasonic welding concentrate on refining the grain size and improving the weld strength [14][15][16].However, further advances in ultrasonic welding are greatly hampered by its high suitability for thin base metal plates (≈1 mm).In this work, we chose thick plates (4 mm) to testing the expansion of ultrasonic welding industrial applicability.Furthermore, we added an appropriate amount of copper powder to the liquid solder, which avoided excessive corrosion of the base metal and played a role in particle reinforcement [17].This also, shortened the reaction time and improved the interface structure.

Materials and methods
Pure copper plates were used as base metals (up: 15 × 10 × 4 mm 3 , down: 15 × 20 × 4 mm 3 ).The welding surface was polished and carefully cleaned using 400 grit sandpaper.The solder was a GaIn eutectic alloy (Ga 75 wt %-In 25 wt %) doped with 2-3 μm pure atomised copper powder.The surfaces to be welded were evenly scribbled with solder (0.1 mm thickness) and set in a mould.The lower base metal was completely fixed while the other could slide in one direction.Then the mould was placed under and ultrasonic generator and a 200 kg pressure applied in a direction perpendicular to the base metal.The temperature during welding was room-temperature (298.15K) maintained by a temperature-control device.The ultrasonic amplitude and duration were 40 μm and 0.3-1.2s, respectively.The ultrasonic frequency and power were unchanged throughout the experiments.The homemade welding system consisted of four parts, including an ultrasonic intelligent numerical control driving power supply, an ultrasonic time controller, a pressure sensor and ultrasonic excitation equipment(2 kW power, 20 kHz frequency), as schematically shown in Figure 1a.Based on previous experimental data, measurements were taken after four days following completion of the welding to allow for aging.
The fracture morphologies, microstructures and composition distributions of the joints were observed with a scanning electron microscope (SEM, VEGA3 XMU, TESCAN, Czech Republic) equipped with an energy-dispersive spectrometer (EDS).The fracture phases were examined by X-ray diffraction (XRD, D/Max 2500 PC, Rigaku, Japan).The shear strength values of the joints were tested using a manual pressure shearing equipment in air at room temperature with a constant speed of 0.1 mm/min.The average of three measurements was taken.

Shear strength analysis
Based on earlier extensive experiments, it was found that the ultrasonic action time and the solder composition (referred to by the copper powder addition) were the two most important factors in affecting the weld quality.Their main influences on shear strength and microstructure were determined in this work.When studying the effects of the ultrasound, the copper powder concentration was fixed (5%), as was the ultrasound time (0.6 s).
The shearing test (as shown in Figure 1b) revealed that the joint strength was low without the assistance of copper powder or ultrasound.An appropriate addition of copper powder plays a role in fine-grain strengthening and avoidance of excessive dissolution of the base metal, as shown in Figure 4. Cu and Ga can react at room temperature [18], and so they still added a certain bonding strength without ultrasound assistance.No ultrasonic action cannot ensure better reaction of the solder.With increase in the content of Cu powder and the ultrasonic time, the shear strength curves presented an inverted Vshape.A shear strength of 15 MPa was the maximum value with 5% Cu powder addition and 0.6 s ultrasound.

Phase analysis
As shown in the XRD patterns (Figure 1c), with no copper powder addition, CuGa 2 diffraction peaks appeared on the fracture surface, indicating Ga had successfully diffused into the copper matrix and produced a metallurgical reaction with the base metal to form a unique room-temperature product, namely CuGa 2 .For the XRD patterns from all fracture phases at room temperature (Figure 1c and d), all peaks could be identified as corresponding to In and Cu, except for the peaks from CuGa 2 .Precipitation of In occurs because of a metallurgical reaction between Ga and Cu according to the reaction Cu + (Ga + 25%In) → CuGa 2 + In [19].Furthermore, there were no phase transitions on changing the copper powder addition or the ultrasonic time, as confirmed in the literature [18].Owing to the relatively thin interface layer, the X-rays penetrated this and reached the substrate, resulting in a relatively strong diffraction peak associated with Cu.

Fracture analysis
Figure 2a-d shows the shear fracture morphologies of the welded joint with different parameters.The content of the elements at different positions are shown in Table 1. Figure 2a and b shows that a honeycomb network structure composed of elemental In, in which CuGa 2 particles were embedded, acts as a barrier to prevent interfacial shear sliding.During the reaction, In is continuously precipitated from the eutectic, covering the surface of CuGa 2 to form an interesting sandwich structure In/CuGa 2 /Cu.This structure can be approximated as a combination of soft/hard/soft phases (with hardness in the order In<Cu<CuGa 2 ) [19,20], which prevents interface sliding during shearing.The existence of tear ridges created onside the copper matrix can be observed in Figure 2c.Moreover, combining Figure 2d and its embedded diagram, an obvious plastic flow phenomenon inside the joint is evident.The CuGa 2 particles coordinate with each other and form a river-like shape as shown in Figure 2d during the shear process.This unsmooth surface also confirmed the shearing process was severely resistant.All these results demonstrate the fracture mode in this sample occurred partially in the substrate layer and partially in the interface layer.The fracture morphology with 20% copper powder addition and 0.6 s ultrasound is shown in Figure 2e.It can be observed that the precipitates In and the product CuGa 2 exist in the smooth fracture without obvious In net, indicating a poor plastic deformation ability under these conditions.The dotted frame area shows that all the CuGa 2 peeled off.Similarly, no network structure was found and the interface was smooth (Figure 2f).Therefore, 20% copper powder addition and 1.2 s ultrasound weaken the intermetallic combination, resulting in brittle fracture at the interface and a low shear strength (as Figure 1b).

Microstructure under different sonication times
Figure 3a-d displays the cross-sections under 0.3-1.2s ultrasound, respectively.Figure 3c is the EDS result of Figure 3b.Table 2 shows the relative content of  elements at different positions.With the extension of the ultrasound time, the pores in the weld gradually become inconspicuous.Meanwhile, the shedding between the weld and the base metal gradually becomes serious and the layered structure in the weld obvious.When the ultrasonic time was 0.6 s (Figure 3b), the weld was flat without obvious dense pores.According to the EDS analysis, In evenly dispersed around CuGa 2 particles and large copper particles (formed by the action of agglomeration) formed a reinforced structure.In this structure, soft and hard tissues were intertwined (as A, B, C).Except for obvious dissolution pits, typical wave peaks and troughs were also seen after  the longer ultrasonic action (Figure 3c and d).Moreover, under the long-term mechanical action of 1.2 s, wavy delamination in the weld was also observed, which destroyed the originally complete weld structure.Such ultrasonic effects have also been confirmed in other studies [21,22].

Microstructure under different copper powder content
Figure 4a-d displays cross-sections of welds with different additions of copper powder.The upper and lower base metals were dissolved by Ga, forming corrosion pits, as shown in Figure 4a (0% Cu powder).The CuGa 2 particles were relatively independent, accompanied by poor compactness.With an increase in the copper powder content, densification of the weld was promoted.It can be seen in Figure 3b that the substances in the weld are closely combined with each other under the addition of 5% copper powder.However, the sections under high concentrations of copper powder showed gradually severe peeling and an agglomeration structure of copper powder (Figure 4b-d), especially clear in Figure 4d.This was inseparable from the fact that excessive addition of copper powder led to the reduction of the liquid-phase content so that it became difficult for Ga to further diffuse into the matrix.Figure 4 shows the interface results corresponding to the complete reaction of different coated solders in their natural state.It can be seen that the addition of copper powder not only significantly reduces the reaction time, but also makes the products finer, more uniform, and closely arranged.Therefore, by tuning the ultrasonic time and the copper powder concentration, the microstructure and the shear strength of the weld can be adjusted and enhanced.Under the ultrasonic vibration, the base metal, the solder and the inside of the solder rub against each other.This process removes the oxide film on the surface, scours the surface of the base metal and mixes the solder significantly.Meanwhile, ultrasound provides the driving force for diffusion as well as a special chemical environment for metallurgical reactions.With ultrasonic vibrations, the interface particles contact each other, react and grow.Too-short or too-long times directly influence the input of ultrasonic energy into the workpiece, thereby affecting the mechanical action and cohesion, finally reducing the shear strength and damaging the microstructure of the joint.A short ultrasonic time leads to a large area of residual oxide on the metal contact surface.This leads to bumps in the weldment instead of more effective large-area welder spots.With fewer particles in contact, more voids are inevitable.In the long-term, large-amplitude ultrasonic vibration will cause the material to 'harden', internal dislocations of the material will accumulate and stack, and finally cause the material to become plastically deformed.This is also the root cause of the poor shear performance of the weldment.The complete weld structure will also be damaged by particles in the weld during long-time mechanical action.Adding an appropriate amount of copper powder not only prevents the base metal from being overly eroded, but also acts as a particle reinforcement phase to form a soft-hard structure with In and CuGa 2 to improve joint performance during the ultrasonic process.However, when the copper powder content exceeds a critical value, a reduction of the liquid-phase content inevitably leads to excessive aggregation of particles.At the same time, excessive intermetallic compounds in the weld seriously inhibit further diffusion of the solder.Therefore, only with an appropriate ultrasonic vibration time and a small amount of reinforcing particles, can more reliable joints be obtained with room-temperature welding.

Conclusions
In conclusion, a method is proposed to rapidly achieve room-temperature welding.Using ultrasonic machinery and coagulation as the driving force for a Cu-Ga interface chemical reaction, combined with its unique room-temperature liquid properties and affinity with Cu, Ga can quickly diffuse to the matrix and achieve fast metallurgical bonding with the base metals.Under the condition of 0.6 s ultrasonic action and 5% copper powder addition, the weld particles connect to the maximum extent and form a soft/hard/soft inlaid interface structure to achieve a room-temperature shear strength of 15 MPa.

Figure 1 .
Figure 1.(a) Schematic diagram of welding system; (b) Variation curves of shear strength with the addition of copper powder and ultrasonic time (c, d) XRD results of fractures.

Figure 3 .
Figure 3.The micro-morphology of the interfaces under different ultrasonic time and 5% addition of copper powder: (a) 0.3 s; (b) 0.6 s; (d) 0.9 s; (e) 1.2 s; (c) EDS mapping of b.

Table 1 .
Elements content at different fracture positions.

Table 2 .
Elements content at different interface positions.