3.1. Weld appearance
With the addition of Ni foil of different thicknesses, the weld profiles of the welded joints are shown in
Figure 5. The weld shape was X-shape without Ni foil (No.1) while the weld shape changed into Y-shape with Ni foil (No.2 and 3). After measuring the widths of the welded joints, the top width (B
1) had little change. The cross-section width (B
2) decreased slightly, and the bottom width (B
3) showed a decreasing trend, as shown in
Figure 5.
The shape and size of the weld are mainly affected by the value of the heat input. When Ni foil was not added, the heat input was large enough, and the depth of the key hole was large. A strong vortex occurred on the upper and lower surfaces of the weld to obtain a completely penetrating weld. Both the top width (B1) and the bottom bead width (B3) were wide, so the weld shape is an X-shape. After adding Ni foil at the same heat input, Ni foil would absorb some energy which led to a decrease of heat at the bottom of the weld, therefore, the weld is Y-shaped (No.2 and No.3).
3.2. Microstructure
After laser welding of dissimilar steels, the welded joint can be divided into fusion zone (FZ), heat affected zone (HAZ) and base material(BM).
Figure 6 gives the optical images in FZ with 0, 30 and 100μm Ni foil. Without Ni foil, FZ and FB (fusion boundary) comprised a lath-like structure and strip-like white phase (
Figure 6 bc). After adding a little of Ni foil (30μm), the amount and size of white phase obviously decreased. At the same time, a new grey phase (FM) appeared. However, no white phases were observed with more thickness of Ni foil(100μm).
To thoroughly analyze the microstructure of the FZ, the SEM and EDS analytical results are given in
Figure 7 and
Table 2. The color of “white phase” in
Figure 6 changed to black color. This was because these phases contained light-weight elements (Al). From the EDS results, P
1,P
2, P
3 and P
4 all contained some Al (7.65%, 6.73%, 3.76%, 3.11%). The same as reference [
20], Al element belonged to the ferrite-forming element promoted the formation of δ-ferrite(white band-like). During the laser welding, Al-Si coating was melted into FZ, the gathering of Al in FZ, especially in FB, δ-ferrite phases were formed. However, Zn did not exist in FZ. This was because the boiling point of Zn was very low(903℃). Nearly all of them were vaporized during the welding process. Therefore, little or no Zn existed in FZ after welding.
Except for the “black phases”, other phases expressed lath-like. Based on the composition of base material and the quick cooling speed of laser welding, these phases were martensite. But carefully comparing the lath-like phases, they were obviously different. From
Figure 6, lath-like martensite was gray without Ni foil. But part of martensite turned greyish-white with 30, especially with 100μm Ni foil, the proportion of greyish-white martensite was very high. From the SEM images of 100μm Ni foil, the size of greyish-white lath (PM) became smaller than gray lath(FM). The EDS results further confirmed that the two martensites were different because the greyish-white lath martensite contained more Ni than grey martensite. Ni element can expand the austenitic phase region which can increase the temperature of A4 and form more γ phases at high temperatures. Since Ni can be infinitely soluble in the γ phase, and the martensite transition temperature (M
S) is reduced, thereby increasing the martensite phase transition hysteresis width (ΔT=A
S-M
S). During cooling, the martensite at high temperatures was called the previous martensite (PM) transition which contained low Ni content. But in high Ni content areas, the Ms point decreased. The martensite transition occurred at low temperatures. This martensite was called Fresh Martensite (FM) transition. The similar phenomenon was observed in reference [
21].
The evolution of the microstructure in FZ and FB can be explained as that: During laser welding, both the base material and coatings were melted. Little or No Zn went into the FZ due to its low boiling point. However, Al-Si coating went into the melted welding pool. Al is possibly gathered in FZ, especially at the FB due to the weld pool flow of laser welding (
Figure 8ab). Without Ni foil adding the joint, the gathering extent of Al was serious. Al was a ferrite-forming element which promoted the forming of δ-ferrite. Therefore, a large amount of δ-ferrite was formed in FZ, especially big bank-like δ-ferrite were formed due to the hysteresis effect at FB(
Figure 9a).
After adding Ni foil between the two base materials, Ni was melted into the welding pool(
Figure 8b). Since Ni element can expand the austenitic phase region which can inhibit the nucleation and growth of δ-ferrite(
Figure 9b). Studies have shown that the same phenomenon is obtained by electron probe microscopy (EPMA), and it can be seen that Ni can well inhibit the macroscopic segregation of δ-Fe caused by the partial polymerization of Al elements [
22]. When the amount of Ni was low(30μmNi foil), it was not enough to completely inhibit the macroscopic segregation phenomenon of δ-ferrite caused by Al elements, so a small amount of dot-like δ-Fe was also observed in FZ and FB, as shown in
Figure 6ef and
Figure 7cd. However, when the amount of Ni exceeded a value (such as 100umNi foil), the δ-Fe phase would disappear.as shown in
Figure 6hi and
Figure 7ef.
Except for the forming of δ-ferrite, the remained phase featured a lath-like shape microstructure in FZ. This was due to the fact that the cooling speed of laser welding was much faster than the crucial rate of martensitic transition. Initially, generated austenitic would undergo a shear-type phase transition into martensite. As a result, martensite has a lath-like shape in FZ. However, when the joint was added of Ni, Ni could expand the γ phase indefinitely, and formed two different shapes of martensite (PM, FM).The cross-sectional morphology and hardness distribution of HAZ on the side of galvanized steel and 22MnB5 were analyzed with 100μm nickel foil, as shown in
Figure 10. The transformation pathways and the resultant microstructure heat-affected zone are given in
Table 3. Due to the different base materials and thermal cycles in different positions, the structure of HAZ on the two sides of base materials was different. On the 22MnB5 side, HAZ was separated into a complete quenching zone (UCHAZ), incomplete quenching zone (ICHAZ) and tempering zone (SCHAZ). Among them, according to the grain size, complete quenching zone can be divided into coarse grain area (CGHAZ) and fine grain area (FGHAZ). Because the peak heating temperature was higher than the austenitic complete transition temperature A
C3 (CGHAZ and FGHAZ), the martensite first transformed austenitic and subsequently transformed martensite due to the rapid cooling rate(M→γ→M). But the size of grain in CGHAZ was very large due to the high temperature of abnormal grain growth (above 1100℃). So the hardness in the UCHAZ was very high up to 500-550HV
0.1(
Figure 10e). When the temperature was between A
C3 and A
C1(Ac
1<T<Ac
3), the incomplete austenitic transformation occurred. The material was heated to the double region. The martensite firstly transformed the two phases: austenitic and α-Fe, then the austenitic transformed martensite ( M→γ+α-Fe→M+α-Fe). If the peak temperature of the tempering zone was lower than Ac
1 (T
temper<T<Ac
1), the transformation of tempered martensite occurred(M→TM). Its microhardness was very low( 332 HV
0.1), as shown in
Figure 10e.
On the side of galvanized steel, it is difficult to distinguish different zones due to the quick heating and cooling of laser welding process. The microstructure of BM mainly contained ferrite and very little pearlite. Only part of F+P was transformed austenitic during heating, then recrystallized to the small size of F+P (α-Fe+P→γ+α-Fe+P→F+P). Therefore, only one HAZ-GA was obtained(Incomplete recrystallization zone). As illustrated in
Figure 10de, it comprised ferrite and pearlite and displayed very little hardness (150-170 HV0.1).
3.3. Mechanical properties
Table 4 and
Figure 11 show the tensile-shear strength and fracture locations of welded joints with variable Ni foil. The tensile strength of the welded joint was deficient without Ni foil, and the fracture location was FB. However, the tensile strength increased with the increasing thickness of Ni foil. The fracture location was from FB changed to interface. The maximum strength can be up to 679MPa under the thickness of 100 μm nickel foil.
Figure 12 and
Table 5 give the images and its composition of fracture. In the absence of Ni foil, river patterns (E
1, 8.07 wt.% Al )and big cleavage planes (E
2, 3.05 wt.% Al )were observed. Meanwhile, some dimples also seemed on the fracture surface (E
3, little or no of Al). As a result, the fracture mechanism was a mixture of brittle cleavage, quasi-cleavage, and ductile fracture. It was discovered that the fracture most likely began at the FB near interface with many -ferrite phases and then traveled higher along the FB. That adequately accounted for why the mechanical property of the welded joints without Ni was so low. As added 30μm Ni foil, the cleavage planes were small. Most of the fracture surfaces were dimples. The composition of small cleavage planes also contained Al(E
4, 2.04 wt.% Al ). The fracture pattern with 30μm Ni foil mainly contained quasi-cleavage and ductile fracture. Furthering increasing the thickness of Ni(100μm), the fracture surface mainly contained dimples. Compared with the small dimples (E
5, 2.59 wt.% Ni ), the big dimples consisted of more Ni(E
6, 8.07 wt.% Ni ). The fracture pattern with high Ni was a ductile fracture. All of the above indicated that with no or low amount of Ni, the accumulation of Al at the FB or in the FZ would be harmful to the joint which resulted in low strength due to the low strength and ductility of δ-ferrite(an Al-rich phase). After adding Ni, the amount of δ-ferrite decreased or even disappeared. The strength of the welded joint increased accordingly.