3.2.1. Tensile Shear Strength
After determining the weld lobe, the effects of both preheating and slow cooling stages (according to Fig. 2) on the stability of the nugget diameters were assessed. For this purpose, ten joints were welded for each mode. The results of statistical processing of the nugget diameters are shown in Fig. 5. For all modes, the average values differed within 5%, which indicated an insignificant contribution of the preheating and slow cooling stages to the total heat generation. However, the implementation of the preheating stage made it possible to reduce the dispersion of the nugget diameters by 60%. In turn, mode 3, characterized by both preheating and slow cooling stages, enabled to decrease it by 160% in comparison with the rectangular-shaped current pulse (mode 1).
Then, the tensile shear tests were carried out using the specimens welded in all three modes according to Fig. 2. The obtained results are presented in Fig. 6. Since the SSRSW joints were simultaneously affected by both shear and tear forces, as well as torque [50], the displacement vs tensile shear strength dependences were not converted into typical stress–strain diagrams. These data are also summarized in Fig. 7, which presents the peak load levels and the fracture energy values, calculated as the area under the displacement vs tensile shear strength dependences (Fig. 6) till peak load. In addition, Table 5 and Fig. 8 show the nugget diameters and the peak load levels at the fracture stage for all the welded specimens. According to these data, the average fracture energy increased by 30% for mode 2 (with the preheating stage), and by 48% for mode 3, which included both preheating and slow cooling stages, compared with those for mode 1 (the rectangular-shaped pulse current). In the tensile shear tests, the peak loads were 2.87 and 3.05 kN for modes 2 and 3, respectively, in contrast to 2.67 kN for mode 1. Also, the dispersion of the peak load values decreased by a factor of 2 for both modes 2 and 3, which included the preheating stage, comparing with mode 1.
Table 5
The average both nugget diameters and peak load levels at the failure stage for the specimens welded using the modes according to Fig. 2.
|
Mode
|
Correlation coefficient
|
1
|
2
|
3
|
Peak load, kN
|
2.67
|
2.87
|
3.05
|
0.99
|
Nugget diameter, mm
|
3.07
|
3.15
|
3.21
|
After the tensile shear tests, the failed SSRSW joints were examined in two ways: by evaluating their cross sections via OM investigations (Fig. 9) and by SEM fractography of the fracture surfaces (Fig. 10). According to Fig. 9, a, the nugget had been pulled out through both weld and base metals for mode 1, while fracture had occurred via both heat-affected zone and base metal for mode 2 and through the base metal only for mode 3.
All pulled out nuggets were characterized by rather uniform macroreliefs with a pronounced rim in the center (Fig. 10, a). At the microlevel, their fracture surfaces were of a mixed nature, both brittle and ductile in different degrees and volumes, for all three modes (Fig. 10, b–d). For mode 1 (Fig. 10, b), most of the microrelief was of a fibrous-banded type due to the brittle intergranular fracture. The reason of such a failure had been the structural inhomogeneity of the deformed material and/or an increased concentration of impurities, which contributed to the formation of brittle inclusions, characteristic of titanium and its alloys [1, 2], at grain boundaries. They had broken the continuity of the boundaries, causing the intergranular fracture. In some cases, the boundary embrittlement could be associated with the impurity segregation without the formation of second phase particles (SPPs) [52].
For mode 2, the microrelief consisted of two regions occupying approximately equal areas (Fig. 10, c). The first one had a similar brittle fibrous-striated type, while the second part was in a ductile pit fracture form (i.e., exposed surfaces of microvoids), formed during the plastic flow of the metal. Non-metallic inclusions or SPPs, as well as microdiscontinuities at the boundaries of grains, subgrains and shear planes could serve as a source of such microvoids [52].
In the case of mode 3 (Fig. 10, d), the microrelief was characterized by two similar areas described above, but the ductile one occupied about 80% of the fracture surface. Depths of the observed pits (heights of the bridges between them) were much greater in comparison with those for mode 2, which indicated a high degree of the material’s ability to plastic deformation. The energy intensity was much higher for such a ductile fracture, because, during the ductile crack propagation, the metal had been plasticly strained not only near its tip, but also over a significant specimen volume. As a result, the work required to propagate the main crack had been much greater here than that in the brittle fracture case, when plastic deformation had been localized in a narrow layer near the main crack tip [52]. According to the authors, this phenomenon correlated well with the results of the tensile shear tests shown in Figs. 6 and 7.
3.2.2. The Microstructure, Microhardness, Distribution the Alloying Elements and Phase Composition of the Nuggets
For a deeper understanding of the influence of the preheated and slow cooling stages on the formation of nuggets, their microstructure (Fig. 11), microhardness (Fig. 12), distributions of the alloying elements (Fig. 13) and phase compositions (Fig. 14 and Table 6) were examined. The distributions of the alloying elements were measured similarly to the microhardness tests strictly through the nugget diagonal axis with a step of 200 µm (Fig. 12).
In all investigated cases, the microstructure of the welded joints had three main clearly distinguishable zones: a nugget, a heat-affected zone, and the base metal. The nuggets consisted of a large acicular martensitic α′ microstructure (Fig. 11). This fact was confirmed by X-ray diffraction patterns (Fig. 14). In the nuggets, grain sizes decreased towards their periphery, i.e. in the heat removing direction upon the SSRSW process. For mode 1, a crack was observed at the nugget boundary between the welded plates (Fig. 11, a). The reason for its appearance was the high cooling rate upon solidification with insufficient electrode compression force, but such cracks were not found for mode 2 with similar cooling conditions (Fig. 11, b). After a decrease in the cooling rate due to the prolonged down slope, no significant visually detectable changes in the nugget microstructure were observed (Fig. 11, c). This fact was confirmed by the comparable microhardness levels (Fig. 12), but the lattice microdistortions, reflecting residual stresses of the second kind inside grains, decrease slightly for mode 3 (Table 6). The distributions of the alloying elements were rather uneven in all cases and no obvious dependences on the heat input algorithm were revealed either (Fig. 13).
Table 6
The calculated values of the lattice parameters, unit cell volumes, sizes of coherent scattering regions (CSR) and microdistortions, reflecting residual stresses of the second kind inside grains.
Phase
|
Mode
|
Lattice parameters a, Å
|
Lattice parameters c, Å
|
Unit cell volume V, Å3
|
CSR dimensions, nm
|
Microdistortions Δd/d
|
α-Ti
|
1
|
2.9407 ± 0.0001
|
4.6765 ± 0.0060
|
35.0230 ± 2.0
|
35 ± 20
|
3.8·10–3
|
2
|
2.9400 ± 0.0001
|
4.6758 ± 0.0030
|
35.0000 ± 2.0
|
34 ± 25
|
3.8·10–3
|
3
|
2.9410 ± 0.0020
|
4.6689 ± 0.0070
|
34.9730 ± 2.0
|
40 ± 32
|
3.6·10–3
|