To complete this study, metallurgical analyses were needed in view of various opinions that came to us. We had to de- termine if differences other than sheet thickness explained the greater suscepti- bility of DP20 toward weld interfacial frac- ture. Since steel chemical composition first appeared to be important (Equation 1), variations in hardness (i.e., strength), microstructure, and local composition were examined to attempt determining all causes, primary or secondary, of spot weld fracture in DP600 steels.
Variations in Microhardness
For both DP18 and DP20, Fig. 9 de- picts microhardness profiles for welds produced with both regular and unusually fast welding schedules (12.5 kA, 20 cycles vs. 40 kA, 2 cycles). The average hardness values of the weld fusion zones are also shown by the horizontal straight lines. Av- erage hardness was higher in the rapidly- made high-current weld, as seen by com- paring the upper and lower graphs. Also, for all welds, microhardness fluctuated slightly and always peaked in the HAZ, marked as HAZ. Near the fusion zone, where microstructure was the hardest, hardness exceeded 400 kg/mm2 and was about twice greater than in the base mate- rial. Like many non-AHSS, which have all consistently produced buttons after chisel testing, no measurable softening was de- tected near the base metal (Refs. 6, 27). A softening in the HAZ would have helped plastic strain to localize and fracture to follow outside the hard fusion zone (Refs. 27–29). Therefore, this feature alone would have explained why welds in DP600 fracture more interfacially than tradi- tional automotive steels.
Regardless of the welding parameters, the HAZ and the fusion zone were harder in DP18, a result that agrees with the CE values of Table 1. While microhardness in both fusion zone and HAZ were greater in DP18, the hardening with respect to the initial microstructure (or percent increase in hardness) was greater in the welds of DP20. This hardening was 2.17 in DP20 vs. 2.05 in DP18. If comparing strength of fu- sion zone or HAZ with that of base metal is important, as suggested by mathemati- cal modeling (Refs. 28, 29), then examin- ing hardening was not helpful here in ex- plaining the greater susceptibility of DP20 to weld interfacial fracture, because a greater hardening generally correlates with less interfacial fractures (Ref. 27). If, instead, the CE of Equation 1 is used to differentiate steels by types of spot weld fracture, then CE cannot explain the su- perior performance of DP18 during chisel
testing. Of all factors, only sheet thickness has to this point explained the dissimilar behavior and different weldability of DP18 and DP20.
Variations in Microstructure
Examination by optical microscopy of weld cross sections neither distinguished DP18 from DP20, but did reveal mi- crostructural features that potentially af- fected weld fracture in DP18 and DP20. Fig. 10 depicts four optical micrographs of a characteristic spot weld in DP20. The low-magnification micrograph of Figure 10A shows that microstructure was more heterogeneous in the fusion zone than in the HAZ. Figure 10B, for the colder sec- tion of the HAZ, demonstrates that its mi- croconstituents were considerably finer than those of either the base metal or the fusion zone (Figs. 10C and 10D). This is explained by the fact that austenitizing was incomplete in the colder section of the HAZ, and even when austenite grains formed, grain growth was restricted by the thermal cycles. In this section of the HAZ where austenite grains are fine, the result- ing high density of grain boundaries con- stitute obstacles against the formation of large martensite laths. As opposed to the coarser martensite of Fig. 10E, observed in the fusion zone, the fine martensite of the HAZ was not well resolved by SEM. As shown in Fig. 9, hardness correspond- ing to the fine microstructures of the HAZ was also greatest.
Figure 10C shows a region of the fusion zone near the HAZ. Large aggregates of a white blocky phase (presumably ferrite) can be observed in between finer micro- constituents. Due to their unique mor- phology, these fine microconstituents were also identified by SEM as martensite. Figure 10D depicts a narrow view of the fusion zone center. Tiny cracks (approxi- mately 50 µm long) were regularly seen, especially in the DP20 steel welds. The ir- regular topography seen in these cracks and revealed by Fig. 10E suggests shrink- age cracks, or solidification cracks; the later being more likely, as indicated by EDS measurements of sulfur, calcium, potassium, and silicon (i.e., many ele- ments that are found in slags for deoxidiz- ing and desulfurizing steel and that are normally removed to minimize solidifica- tion cracking tendency) (Ref. 2).
Although all our observations indicate that the role played by microscopic cracks on weld fracture is unlikely significant compared to that of the sheet thickness, this section has brought another explana- tion for the greater susceptibility of DP20 for weld interfacial fracture. These micro- scopic cracks, most distinctively encoun- tered in the DP20 steel welds, were given
further attention as chemical composition nearby internal defects was further investigated.
Variations in Chemical Composition
For DP20, Fig. 11 shows six X-ray com- positional maps captured at a weld center, and three horizontal line scans across the maps for the major alloying elements. In addition to voids and cracks along the weld centerline, Fig. 11 shows that the most abundant element, manganese, was most distinctively microsegregated. Its re- distribution within the microstructure re- vealed a fine columnar substructure that grew from the fusion line to the weld cen- terline. Close examination of the map and line scan of manganese indicates periodic depletion and accumulation every 10 to 20 mm. This microstructural feature alone can be well explained by the classic Scheil’s model (Ref. 15), where the man- ganese-lean regions would be first to so- lidify and the manganese-rich regions last. Manganese profiles could thus be used to track the solidification substructure.
The line scan of Fig. 11 shows that the peaks for manganese, silicon, and chromium were superimposed. Unlike manganese, chromium distribution was quite uniform, as explained by the iron- chromium partitioning coefficient (nearly one) (Ref. 30). Although zinc was also fairly well dispersed throughout the fusion zone, Fig. 11 raises the possibility of zinc buildup near voids; an observation that is well justified by zinc and iron’s limited mu- tual solubility, and zinc’s property to re- main liquid long after iron has solidified. In Fig. 11, note that the regions of high X rays for carbon, silicon, and oxygen not only correlated with one another, but were also aligned with the solidification sub- structure (revealed by the microsegre- gated manganese). Another important feature is that all three elements were clearly identified in the larger voids, thereby suggesting that they were proba- bly in the form of carbides and silica par- ticles left by the sample preparation. By tracking these last three elements, micro- scopic voids could be also revealed. In Fig. 10, recall that sulfur, calcium, and potas- sium were detected in microscopic voids and cracks; an observation that first sug- gested that these elements could be linked to the formation of internal defects. In this section, we observe that only the contri- bution of zinc to the formation of internal defects could not be ruled out yet.
Spot Weld Interfacial Fracture
The numerous weld fracture surfaces
180 -s NOVEMBER 2005