The collision velocity has a remarkable effect on the strength and the interfacial morphology, including the thickness of the transition layer and the size of the interfacial wave [28–29]. Therefore, Fig. 6 presents the evolution of the collision velocity along the welding direction under different geometric features at the discharge voltage of 15 kV. Figure 6(a) presents the case without geometric features. The curve shows that the collision velocity remained constant at about 390 m/s within about 4 mm, increased within the following 1 mm, and then decreased sharply. After the geometric features were prefabricated, it can be seen from Figs. 6(b)-6(d) that the initial collision velocities under different geometric features were close to that without geometric features. On the whole, the collision velocity first increased and then decreased sharply with the progression of the collision point. Moreover, the collision velocity increased with the increase of α under all types of features. It is evident that before the maximum collision velocity was reached, the collision velocity under the convex wall feature from Fig. 6(d) increased the fastest, and the velocity under the concave wall feature shown in Fig. 6(c) increased the slowest. Thus, the collision velocity under the VWC was the greatest at the same value of α with a similar initial collision velocity. In addition, the position at which the maximum collision velocity occurred was closer to the initial collision end with the increase of the prefabricated angle, as shown in Fig. 6(b). According to Xu [30], the acceleration distance and time of the flyer tube decrease when the radial gap is too low, which makes it difficult to fully accelerate. Conversely, when the radial gap is too high, the circumferential stress increases significantly with the increase of deformation, which leads to the increase of the deformation resistance to slow down the velocity of the flyer tube. Thus, there exists an optimal radial gap to maximize the collision velocity. Based on Fig. 2, the geometric features enlarged the range of the radial gap to different degrees. Moreover, the radial gap increased the fastest under the convex wall feature, and increased the slowest under the concave wall feature. The larger the prefabrication angle, the faster the radial gap increased. Moreover, the faster the radial gap increased along the axial direction, the closer the position where the optimal radial gap appears is to the initial collision end. Thus, the collision velocity was found to be affected by the radial gap caused by the geometric features and their angle.
4.2 Evolution of the collision angle
The collision angle is a critical parameter for the initiation of the weld [31]. Therefore, the evolution of the collision angle was investigated in detail under different geometric features. Figure 7 presents the axial distribution of the collision angle under different geometric features. Overall, all of the curves can be divided into three parts based on the value of the collision angle, just as shown in the case without geometric features presented in Fig. 7(a). The collision angle broadly remained constant at no more than 3°. Figures 7(b)-7(d) show that with the increase of the prefabricated angle, the collision angle increased from about 6° to about 12° for the IWC (Fig. 7(b)) and VWC (Fig. 7(d)), and to below 12° for the CWC (Fig. 7(c)). Regarding the length of the constant zone at the same prefabricated angle, that of the CWC was about 3 mm, which was greater than those of the IWC (2–3 mm) and VWC (less than 1.5 mm). The collision angle then increased with the progression of the collision point in the rising zone, and increased faster with the increase of the prefabricated angle. The curves show that at the same prefabricated angle, the collision angle increased the fastest under the VWC, and increased the slowest under the CWC. The maximum collision angle was reached at about 6 mm away from the initial collision end, after which it decreased sharply in the sudden drop zone. Thus, compared with the case without geometric features, the CWC had a smaller effect on the size of the collision angle and the length of the constant zone than did the IWC and VWC. According to Psyk et al. [32], the evolution of the collision angle is related to the variation of the radial gap. Based on Fig. 2, the impact angle changed to different degrees, which resulted from the variation of the radial gap along the welding direction due to the changes of the types of geometric features and their geometric parameters.
4.3 Axial distribution of the shear strength
Local shear tests were carried out to evaluate the axial distribution of the shear strength of the MPW joints. Shear tests were further conducted for the joints obtained at 15 kV under different types of geometric features, and the shear strength along the axial direction is presented in Fig. 8. The axial distribution of the shear strength without the prefabricated geometric features is shown in Fig. 8(a). The distribution of the shear strength was similar to the V-shape found in a previous study [27]. Samples 2 and 5 had a strength higher than 45 MPa, and Samples 1, 4, and 6 had a shear strength lower than 7 MPa. Figure 8b shows the shear strength under the IWC. Samples 2–5, 2–4, and 3 had a shear strength higher than 40 MPa when α was 5°, 7°, and 10°, respectively. Thus, the length of the zone with a strength higher than 40 MPa decreased with the increase of α. For the CWC, when α was 5°, as shown in Fig. 8(c), not only was the shear strength of Sample 2–5 higher than 45 MPa, but the shear strength of Sample 1 was closer to 40 MPa. Samples 3 and 2–4 had a shear strength higher than 40 MPa when α was 7° and 10°, respectively. The length of the zone with a shear strength higher than 40 MPa first decreased and then increase with the increase of α. Figure 8(d) shows that the strength under the VWC was generally below 40 MPa. To further evaluate the axial distribution of the shear strength, the maximum (max), mean (m), and standard deviation (std. D) of the shear strength of Samples 1–5 obtained under different types of geometric features, as well as their increment percentage relative to the cases with no geometric features, were calculated, and are listed in Table 3. Therein, the maximum, the mean, and the standard deviation are used to assess the strength of the joint and the uniformity of the strength along the axial direction of the joint. The maximum shear strength obtained without prefabricated features was still higher than that obtained with prefabricated features. Nevertheless, the strength of the joint obtained without features had a much lower mean (m = 24.78 MPa) and a rather higher standard deviation (std.D = 25.51 MPa) than those obtained with the IWC, CWC, and VWC, especially for IWC-z5 (m = 40.24 MPa, std.D = 9.16 MPa), CWC-a5 (m = 46.46 MPa, std.D = 4.96 MPa), and VWC-t10 (m = 27.63 MPa, std.D = 11.57 MPa). For the IWC, with the increase of α, the maximum and mean decreased, and the standard deviation increased. Thus, the strength and uniformity of the joint were the best under α = 5°, with the max of 52.41 MPa, m of 40.24 MPa, and std.D of 9.16 MPa. With regard to the CWC, with the increase of α, max first decrease and then slightly increased, m decreased gradually, and std.D first decreased and then significantly increased. This means that upon increasing α, the strength of the joint tended to be reduced, and the uniformity tended to be poor. Hence, the best case with the max of 52.32 MPa, m of 46.46 MPa, and std.D of 4.96 MPa occurred under α = 5°. Regarding the VWC, the max, m, and std.D values first decreased and then increased with the increase of α. The m and std.D values under α = 5° and 10° were similar, and the max value under α = 5° was higher than that under α = 10°. Hence, the best case with the max of 45.61 MPa, m of 27.63 MPa, and std.D of 11.57 MPa occurred under α = 10°. In terms of the maximum shear strength, the case with the IWC was close to that with the CWC and higher than that with the VWC. Regarding the mean and standard deviation, the case with the CWC was the best, especially for the case of CWC-a5, which had the greatest average of 46.46 MPa and a rather lower standard deviation of 4.96 MPa as compared to the other cases. In a previous study [27], only the strength of the first three samples exceeded 45 MPa, and the others had a low strength at the same overlap length. In the study by Lu et al. [33], only three discontinuous samples had a high strength, which was lower than 45 MPa, at the overlap length of 6 mm. This indicates that, in this work, the prefabrication of geometric features could improve the uniformity of the axial distribution of the shear strength of the joint, and could simultaneously maintain the high strength of the joint. The results were the best when the prefabrication feature was the concave wall feature and the initial angle was 5°.
Table 3. The data analysis of the shear strength under different geometric features.
Type of geometric feature
|
Items of analysis
|
max (MPa)
|
Increment percentage (%)
|
m (MPa)
|
Increment percentage (%)
|
std. D (MPa)
|
Increment percentage (%)
|
No features
|
63.50
|
--
|
24.78
|
--
|
25.51
|
--
|
IWC
|
z5
|
52.41
|
-17
|
40.24
|
62
|
9.16
|
-64
|
z7
|
48.04
|
-24
|
36.78
|
48
|
9.74
|
-62
|
z10
|
41.07
|
-35
|
19.17
|
-23
|
12.30
|
-52
|
CWC
|
a5
|
52.32
|
-18
|
46.46
|
87
|
4.96
|
-81
|
a7
|
41.34
|
-35
|
35.75
|
44
|
3.19
|
-88
|
a10
|
44.66
|
-30
|
34.33
|
39
|
11.90
|
-53
|
VWC
|
t5
|
43.03
|
-32
|
28.43
|
15
|
11.45
|
-55
|
t7
|
38.66
|
-39
|
26.67
|
8
|
10.30
|
-60
|
t10
|
45.61
|
-28
|
27.63
|
12
|
11.57
|
-55
|
4.4 Analysis of the evolution mechanism of the shear strength
It is well known that the evolution of the collision parameters along the welding direction is crucial for the interfacial morphology, and, further, the properties of the MPW joints are directly influenced by the interfacial morphology. Hence, the interfacial morphologies under different types of geometric features were observed. The full optical micrographs in the welding interface without geometric features are shown in Fig. 9(a). It can be seen that an unbonded zone of about 1 mm appeared here, which caused the low strength of Sample 1 presented in Fig. 8(a). The unbonded zone was generated by a too-high collision velocity (about 360 m/s) and a too-small collision angle (about 2°); these caused the rebound to occur here, thereby hindering the welding of the materials. Figures 9(b) and 9(c) show the metallurgical bonding zone with irregular waves and an uneven transition layer, which signify high strength. However, some micro-defects, such as micropores and cracks, existed here, which lowered the strength of Sample 2 presented in Fig. 8(a). Then, the metallurgical bonding zone disappeared, as shown in Fig. 9(d), which was caused by the rebound caused by the severe mismatch of the collision velocity and angle [33]. With the progression of the collision point, the collision velocity and angle increased at about 4 mm from the initial collision end, as shown in Figs. 6(a) and 7(a), and gradually matched each other; thus, the metallurgical bonding zone appeared again, as presented in Fig. 9(e). It can be observed that the metallurgical bonding zone exhibited a wavy interface with a thin transition layer without obvious micro-defects, which corresponded to the high shear strength of Sample 5 exhibited in Fig. 8(a).
Metallographic samples under different types of geometric features were collected, and the prefabricated angles of 5° and 7° were selected for all types of geometric features. The interfacial micrographs of the joint obtained under inclined wall features of α = 5° and α = 7° are presented in Figs. 10 and 11, respectively. It can be seen from Figs. 10(a) and 11(a) that only one metallurgical bonding zone existed at the interface. Figures 10(b)-10(e) and Figs. 11(b)-11(e) reveal a similar irregular wavy interface with an uneven transition layer, which led to the similar shear strength between z5 and z7. Nevertheless, compared to Figs. 10(b), 10(c), and 10(e), in Figs. 11(b), 11(c), and 11(e), the wavelength increased significantly, which caused the interface to be flat under z7. This was caused by the higher collision angle and velocity under z7, which raised the wavelength [34]. Figure 10(d) presents an irregular wavy interface with an uneven transition layer, and the maximum thickness reached about 35 µm. No thick transition layer was found at the interface under z7. Thus, the area with an apparent transition layer decreased, and the interface tended to be straight with the increase of α. Based on a previous study [27], a too-thin transition layer weakens the strength of the joint. Moreover, a flat interface usually has a lower strength than a wavy interface [33], which caused the lower local strength of Samples 3 and 5 under z7 as compared to that under z5.
Figure 12a shows the full interfacial morphology of the joint obtained under the concave wall feature of a5. Figures 12(b) and 12(c) indicate that the thickness of the discontinuous transition layer increased gradually with the progressive collision point, which was caused by the increased collision velocity and the broadly constant collision angle shown in Figs. 6(c) and 7(c). Subsequently, a broadly uniform transition layer with a thickness of about 25–40 µm, a length of about 1 mm, and no apparent micro-defects were generated, as shown in Fig. 12(d). This is the reason why the collision velocity (about 450 m/s) and angle (about 4–6°) matched each other well and exhibited little fluctuation within this zone. Beyond this zone, the collision parameters increased gradually, and the thickness of the transition layer decreased, as shown in Fig. 12(e). Figure 13(a) displays the full interfacial morphology of the joint obtained under the concave wall feature of a7. Figure 13(b) shows a similar morphology as that under a5, leading to a similar shear strength in this zone. With the progressive collision point, no apparent transition layer was formed, as shown in Fig. 13(c). Then, as shown in Fig. 13(d), an uneven transition layer with a maximum thickness of about 35 µm and micro-defects (pores and cracks) was generated, which was related to the violent shear deformation due to the higher collision velocity (about 520 m/s) and angle (about 9°) than those under a5 in this zone. With the sharp increase of the collision velocity and angle, a morphology with a disordered interfacial wave and no apparent transition layer was formed, as shown in Fig. 13e. Hence, it is indicated that with the increase of α, the area in the interface with an apparent transition layer decreased, and the micro-defects in the transition layer increased, which possibly lowered the average strength of the joint.
Compared with the case without geometric features, the unbonded zone in the middle of the joint was removed after the geometric features were prefabricated, which caused the area of the joint with high strength to converge. This is the reason why the geometric features changed the matching range of the collision parameters, and the prefabrication angle α played a key role in the initial position where the welding was achieved. The thickness of the transition layer under the CWC was greater than those under the IWC and VWC, and decreased with the increase of α. The presence of a transition layer usually signifies the achievement of metallurgical bonding and a high bonding strength [33]. According to the study by Fan [35], the formation of the transition layer is closely related to the generation and development of the jet. When the collision point velocity is too high, the jet will be captured and then participate in the formation of the transition layer. The results of Wang et al. [36] revealed that the jet is easily captured for the flat target, and the jet can escape by using an appropriate collision angle. In this work, geometric features were prefabricated on the target tube to provide a great impact angle and space for the jet to escape. The greater impact angle and space under the IWC and VWC than those under the CWC resulted in smaller thicknesses of their transition layer or even the absence of a transition layer, which possibly lowered the bonding strength. In addition, the collision parameters had the smallest fluctuation under the CWC, which caused the collision parameters to match well within a greater range than those under the IWC and VWC. When the prefabrication angle of α was 5° under the CWC, the collision parameters exhibited the smallest fluctuation and the greatest matching range. Thus, good metallurgical bonding was realized to different degrees within the greatest scope under a5 although the interfacial microstructure was uneven along the axial direction. When good metallurgical bonding is achieved, shear fracture generally occurs at the side of the weak parent material, meaning that the shear strength is the strength of the weak parent material [19, 37]. Hence, the shear strength under a5 was higher and had better uniformity along the axial direction than that under the other features.