020s.pdf

Welding Journal | January 2014

Fig. 7 — SEM micrograph of particles extracted from the weld metal following dissolution in acid. A B 36). However, this ideal Al/O ratio is based on the fraction of acicular ferrite being maximized, as long as titanium is present (Ref. 37). Since the Al/O ratio was extremely high and there was negligible titanium content, no acicular ferrite formed in the weld metal. Precipitate particles with dimensions >1 μm could not be observed in the inclusions extracted from the weld metal by dissolution, or on the fracture surfaces, suggesting that the presence of submicron sized (Mg, Al)O particles may have suppressed the coarsening of oxide inclusions. The oxygen content measured in the weld metal (120 ppm) is within the range observed for gas metal arc welds; however, the particularly low nitrogen content (64 ppm) is attributed to the use of CO2 shielding gas along with the high aluminum content (0.557 wt-%). For any given level of oxygen content, a transition from a small number of large inclusions to a large number of fine inclusions will result in lower room temperature fracture energy values since the number of initiation points for fibrous fracture will increase (Ref. 38), particularly if decohesion has already occurred at the particle interface upon cooling of the weld metal. In the case of cleavage fracture, fracture stress increases dramatically with decreasing inclusion size, particularly those <0.5 μm in diameter (Ref. 39), as in the case of the weld metal examined here. Further analysis could not directly correlate the size and spacing of the inclusions to any other microstructural features. It is interesting to note a shelled inclusion structure similar to the one observed in this work (but without Mg) was observed in a prior study of flux cored consumables containing Ti and Zr by Narayan et al. (Ref. 40). They showed that the formation of a core/shell structure prevents coarsening and agglomeration of inclusions by “capping” the aluminum oxide particles and suppressing their growth. In that study, a much higher fraction of nitrogen (0.018 to 0.020 wt-%) was present in the weld metal, promoting a shell of (Zr,Ti)N to cap the inclusions. In the present work, it appears that magnesium may have a similar effect in suppressing the coarsening or agglomeration of the oxides in the liquid weld metal, as suggested by the thermodynamic calculations in Fig. 4. The amount of retained austenite measured (2.9%) is comparable to 9.4% measured using the same technique previously in steel welds containing 9 wt-% Ni (Ref. 41), where no peaks corresponding with martensite could be detected (Ref. 42). However, this does not necessarily indicate that martensite was absent since the low carbon content of the weld metal JANUARY 2014, VOL. 93 20-s WELDING RESEARCH Fig. 8 — Charpy impact energy values for material extracted from the middle region of the fusion zone. Fig. 9 — Fracture surfaces from central region of the Charpy sample for tests. A — –18°C; B — –62°C.


Welding Journal | January 2014
To see the actual publication please follow the link above