High silicon aluminum materials prepared by jet deposition method were selected as experimental materials, where aluminum silicon alloy containing 50% silicon content (mass fraction) is selected as the shell, the cover plate is made of silicon aluminum alloy with 27% silicon content. The main physical properties of high silicon aluminum are shown in Table 1. From the table, it can be seen that, with the improvement of silicon content, the coefficient of thermal expansion (CTE), density, and thermal conductivity, and yield strength decrease, while the elasticity module increase, as the silicon fillers are hard, light, and low conductive materials.

Figure 1 shows mechanical structure of the laser welding samples. The shell and the cover plate contact with each other using lap structure. A block was put on the cover plate to apply load to fix the plate against thermal deformation during welding process. A LPKF equipment was used to perform the laser welding process in an argon atmosphere protected glove box.

The morphology of joints was observed by a Leica optical microscope (OM). The properties of laser welding samples were analyzed by metallographic analysis. Firstly, the laser welding samples were cut by a wire cutting machine. Secondly, the samples were encapsulated with epoxy and treated by mechanical polishing with SiC abrasive sandpaper (50#, 200#, 800#, 1,000#, and 2,000#), and then the cross-sectional samples were polished with a .1 μm diamond suspension to eliminate surface scratches. The microstructure of prepared samples was observed by a TESCAN Vega three scanning electron microscope (SEM) equipment at 10 kV–20 kV. Vickers microhardness tester was used to test the microhardness of the welded joints. The loading load was 2.942 N and the loading time was 10 s. During the test, the joints were tested horizontally and longitudinally.

A LPKF P02500 laser welding equipment was used to carry out the study. The main two parameters are power and pulse width. The power is 1,300 W, 1,500 W, 1,800 W, 2,000 W, and 2,300 W respectively, while the pulse width is 3 ms, 4 ms, and 5 ms, respectively.

Figure 2 shows the surface morphologies of laser welding joints. It can be seen that the surface morphologies of all of the prepared joints are good, and the fish scale patterns on the joint surface are uniform and smooth, indicating that laser welding parameters selected in this study makes the high silicon aluminum have good fluidity during welding process. The forming of fish scale pattern is because of the weld solidification sequence during laser welding. The pulsed laser energy output is intermittent. When the laser beam moves forward, the pulsed laser energy will melt not only the base materials, but also the tail of the adjacent welding spot. Therefore, the welding seam is not smooth and has obvious pulse characteristics in front, and the welding spot texture is large and clear. The arrangement is dense and neat.

In Figure 2, a clear phenomenon is shown, the influence of power on joint width is minor than that of pulse width on joint width. For example, the joint width at parameter of 1,300 W and 3 ms is about 49.4 μm. When the power is 2,300 W and pulse width is 3 ms, the joint width is 57.4 μm. However, with parameters of 5 ms and 1,300 W, the joint width reaches at 84 μm. In summary, higher pulse width will form wider joint, while the power has slight influence on welding width. Although the higher power can increase the input energy, the simply increasing the power cannot obviously extend size of the welding spot, as the high thermal conductive base materials owns the ability of quick heat dissipation. By contrast, when the laser pulse width is enhanced, there will form larger welding spot, causing wider melting area.

Figure 3 shows the material microstructure characteristics of the Al-50wt.%Si and Al-27wt.%Si high silicon aluminum alloys. The light area in SEM image is the Si particle, part of which is skeleton-like and part of which is granular, while the dark area in SEM image is the Al phase. The average diameter of silicon particles is about 25 μm in the Al-50wt.%Si high silicon aluminum alloy, while most of the silicon particles are distributed between 10 μm and 40 μm. For the Al-27wt.%Si high silicon aluminum alloy, the size of silicon particles variates from 2 μm to 10 μm. The silicon particles are diffusely distributed in the Al matrix and two phases mix evenly, which is beneficial for improving the thermal conductivity of the material and reducing the linear CTE. Besides, β-Al₅FeSi phase is distributed in a small amount between Al phase and Si phase. According to the phase diagram of Al-Si binary alloy, aluminum alloys of Al-27wt.%Si and Al-50wt.%Si are hypereutectic alloys, and their solidification structure is mainly composed of primary silicon and Si-Al eutectic, in which Al-Si eutectic is composed of α-Al + Si. However, the solubility of aluminum in silicon is very small at room temperature. Therefore, the eutectic silicon in Si-Al eutectic can be regarded as pure silicon (Chen et al., 2014).

Figure 4 shows the typical material microstructure characteristics of the joints. Usually, the joint can be divided into weld zone (WZ), fusion zone (FZ), and HAZ. Because laser welding has the characteristics of fast welding speed, high energy density, and small heat input, HAZ of laser welding joint is small, which is an important advantage of laser welding compared with other welding techniques. From Figure 4, it can be seen that the thickness of HAZ and FZ adjacent to the Al-50wt.%Si base material increases from 20 μm to 150 μm, and the highest value is near the surface of joint. When the HAZ and FZ near the Al-27wt.%Si base material, the thickness variation of HZA and FZ is from 20 μm to 110 μm. As shown in Figure 4, the microstructure of FZ is finer than that of both of Al-50wt.%Si and Al-27wt.%Si base materials, and the microstructure is mainly fine equiaxed crystal. Based on the analysis of the above results, it is believed that many complex metallurgical reactions occur in the molten pool at high temperature step in a very short time. Also, these changes will directly determine the microstructure, mechanical properties, and performance of the joints. Compared to other conventional welding processes, the pulsed laser beam energy is low and has less heat input to welding joints. Moreover, the high thermal conductivity of high aluminum alloys accelerates the cooling rate of the molten pool and leads to quick temperature drop. The residence time at high temperature and the grain growth time are shortened. Additionally, the homogeneous nucleation of silicon particles in melting pool makes that the microstructure of FZ is finer and more uniform than the base materials.

Under normal conditions, one major defect in aluminum alloy welding is the hydrogen porosity, which is characterized by sphericity, small size, and large number. The difference in the solubility of hydrogen in liquid aluminum and solid aluminum results in the formation of hydrogen pores. During laser welding of high silicon aluminum alloys, some hydrogen atoms escape due to the sharp heating and cooling. It is generally believed that the formation process of hydrogen stomata has three stages: hydrogen accumulates to form small hydrogen stomata; hydrogen pores grow to form bubbles; bubbles float up and escape. Due to the large thermal conductivity of Si-Al materials, the condensation is fast. Meanwhile, the viscosity of molten aluminum material is large and the bubble radius is small, so the floating speed of hydrogen bubbles is small in the solution pool. In summary, the low rising speed, the short rising time, and the relative long path make it difficult for bubbles to rise to the surface and escape, so some bubbles are frozen in the FZ.

Besides, a large number of cracks usually pass through the silicon particles and aluminum matrix. This is because the stress state of some silicon particles changed significantly during the welding cold-heat cycle, which changed sharply from compressive stress to tensile stress, leading to the fracture of silicon particles. In the part near the molten pool, the silicon was filled with molten aluminum after rupture, so the crack did not expand further. However, the material near the base material zone has high silicon content and is affected by welding heat. The aluminum matrix does not melt. Under the action of tensile stress during the cooling process, a large amount of brittle silicon breaks, expands, and passes through the aluminum matrix, resulting in macroscopic cracks in the HAZ. In the crack propagation, the propagation trajectory is along the boundary of silicon particles or penetrates the silicon particles. On the one hand, the silicon particles of Al-Si alloy have sharp edges and corners, which are easy to produce stress concentration, thus piercing the phase and forming cracks. On the other hand, the silicon particles are poor in plasticity and easy to fracture under the influence of stress.

According to the above analysis, the main defects in laser welding joints of high silicon aluminum alloy are pores and cracks. Thus, to forbid obvious hydrogen pores and cracks, the laser welding speed is setting at 3 mm/s∼5 mm/s. As shown in Figure 5, although the hydrogen pores and cracks are difficult to avoid, both defects do not form in prepared samples in this research, due to reasonable welding parameters. As the welding thermal cracks are the main factor of air tightness problem, it can speculate that the samples prepared in this study own high resistance to air loss. As shown in Figure 5, the joint depth is 41.1 μm at power of 1,300 W and pulse width of 3 ms. Just increasing the pulse width to 5 ms does slightly increase the joint depth, which is 49.6 μm. But just enhancing the input power to 2,300 w does obviously improve the joint depth to 125.2 μm. Therefore, the laser power is the key parameter to control the joint depth.

As shown in Figure 6A, the microhardness test was carried out along the horizontal direction of joint, the points were successively taken through the cover plate, the HAZ near cover plate, the FZ near cover plate, the WZ, the FZ near the shell, the HAZ near shell, and the shell. When the points were taken, multiple samples were tested and the average value was calculated. After comparing the testing results in Figure 6B, it is found that the microhardness of two parts is different due to the different Si contents of the cover plate and the shell, which are Al-27wt.%Si and Al-50wt.%Si, respectively. The microhardness of the cover plate is 51 HV, while the microhardness of the shell is 91 HV. Additionally, it can be seen from the microhardness distribution line diagram, the microhardness gradually increases from the FZ to the WZ, and the microhardness closer to the center of WZ is higher. Among them, the microhardness of the FZ is about 130 HV, and the microhard-ness of the WZ is about 167 HV. The results show that laser welding could significantly improve the joint microhardness. Because the microstructure of the WZ in the center of the joint is fine, the microhardness is higher. Besides, the Si particles are distributed evenly in joints, the microhardness measurement results in the weld retain a slight fluctuation.

In Figure 7A, the microhardness test was carried out along the center line of the joint longitudinally, and points were successively taken from the upper surface of the joint to the bottom of the joint. It can be seen from Figure 7B, from the top surface to the bottom of the joint, the microhardness value is relatively stable, but the overall fluctuation is slightly. The test position near the joint center has the highest hardness value, and the microhardness value is about 165 HV, followed by the part near the top surface, whose microhardness value is about 156 HV. The lowest microhardness value is the joint bottom with microhardness of about 139 HV. In the center zone, the microstructure is the primary silicon and aluminum silicon eutectic tiny grains, so the hardness is generally on the high level. The position at the bottom of the joint is close to the joint gap between the shell and the cover plate, where there are many defects, such as pores. Although the defects are efficiently reduced, the existence of pores will affect the mechanical properties of the sur-rounding zone to some degree, causing microhardness decrease.

The tensile tests were carried out to the Al-27wt.%Si and Al-50wt.%Si base materials and welding samples. The experiments were conducted using a Instron mechanical test machine. The experimental speed was .05 mm/min and the standard distance was 50 mm. To minimize the experimental errors, five samples were selected to each group, the tensile strength and extension rate were the average value of five samples. The Al-27wt.%Si and Al-50wt.%Si base materials are the control samples and welding samples are the experimental samples. The tensile strengths of the Al-27wt.%Si and Al-50wt.%Si samples are 189.3 MPa and 218.8 MPa, respectively, and the extension rate of Al-27wt.%Si and Al-50wt.%Si samples are 6.2% and 11.4%, respectively. As the grain size of the Al-27wt.%Si base material is smaller than that of the Al-50wt.%Si base material, the mechanical strength and elongation of the former material is greater than that of the latter material, which has been proved in the above research content. By comparison, the tensile properties of laser welding samples and Al-50wt.%Si base materials are almost the same. The welding samples have tensile strength of 187.5 MPa and extension rate of 6.1%.

Figure 8 shows the representative fracture morphologies of base materials and welding samples. For the Al-27wt.%Si base materials, which have less silicon content, relatively fine dimples exist in the fracture location, indicating that the fracture mechanism is ductile fracture. For the Al-50wt.%Si base materials, the crack is splitting through the silicon particles, resulting in brittle fracture. In summary, the less silicon content and smaller silicon particle size could enhance the mechanical strength and elongation property of silicon-aluminum alloy. By observing the welding samples, the fracture locates at the Al-50wt.%Si base material near the joints, and the fracture morphology is similar to the morphology of the Al-50wt.%Si samples. In this study, the spot of the pulsed laser beam is small, and the residual stress is small, effectively reducing the deformation and restraint during laser welding. Additionally, few of pore forms in the joints by pulsed laser welding, and the grain size of welding sample is smaller than both base materials. Thus, the joints of the welding samples are stronger than the base materials. By contrast, the Al-50wt.%Si base material is the weakest part in the welding samples, and the uneven surface leads to stress concentration. Thus, the fracture initiates at the Al-50wt.%Si base material near the welding joint.