clusions observed in Figs. 5 and 6, along with iron-chloride residue, which may have reprecipitated during particle extraction. EDX analysis revealed mainly the presence of oxygen, carbon, iron, and chlorine, with a small fraction of zirconium, magnesium, and aluminum in the extracted residue. Based on the SEM and XRD observations, it would be expected that the cuboidal particles correspond with ZrC, with an average size of 221 ± 45 nm (n = 10). Charpy Toughness Measurements and Fractography The impact testing results are shown in Fig. 8. The upper shelf extends to –40°C, and the ductile to brittle transition temperature, if defined as the temperature at which toughness is intermediate between the upper and lower shelves, is at –60°C or below. The upper shelf value is approximately 130 J, and the lower shelf was never reached, despite tests being conducted down to –73°C. The fracture surface of Charpy specimens tested at –18°C exhibited mainly a fibrous fracture surface, while those tested at –62°C exhibited a combination of fibrous failure and quasicleavage fracture, as shown in Fig. 9. Spherical particles could be observed in bottoms of many of the dimples observed in the fibrous fracture surfaces, in addition to a few randomly distributed cuboidal particles, as shown in Fig. 10. The quasicleavage fracture surface shown in Fig. 11 had facets with dimensions comparable to the ferrite packed diameters observed by SEM in Fig. 2. Instrumented Charpy Testing of Top and Bottom Region of Weld The top (near the cap) and bottom (near the root) of the welds tested showed significantly different microstructures. The bottom of the weld shows a much higher amount of reheated material, as shown in Fig. 1A, which results in much different balances between FS(A) and FS(NA) microconstituents (higher FS(NA) in reheated material). The differences in the fracture strength of these two microstructures were investigated using instrumented impact tests. In these tests, the evolution of force during the breaking of the sample is recorded. The resulting curves are illustrated in Fig. 12. These curves provide much richer detail than a report of only total impact energy values. In particular, the dynamic fracture toughness or J-integral value (J1d) may also be calculated from the data based on the methodology proposed by Moitra et al. (Ref. 33). In this approach, the standard Charpy sample has a notch and no precrack is present, which requires A B C D one to estimate when the actual fracture has initiated based on the force-displacement data collected during impact. For this type of specimen, the dynamic fracture toughness is given by: J1d = η (ES)i Bbo (2) where η is a constant, (ES)i is the energy absorbed up to the crack initiation point, the sample thickness B is 10 mm, and bo is the remaining ligament length of 8 mm. It has been shown that in the case of Charpy impact specimens, η =1.45 (Refs. 33, 34), and that the crack initiation point for ferritic steel specimens can be taken as the point corresponding with (PMAX + PGY)/2, where PMAX is the maximum load during impact, and PGY is the general yield load (Ref. 35). In some cases, resonance in the impact tester produced large oscillations in the force output, so the force output data was averaged to remove these oscillations and allow PMAX and PGY to be readily determined. The area directly under the force-displacement curve up to the point (PMAX + PGY)/2 was then quantified to directly measure (ES)i. The measured values for the impact performance of the top and bottom regions of the weld are summarized in Table 3. The J1d and the total impact energy are similar for both regions of the weld (slightly higher for the bottom region, with the difference more marked at lower temperatures). The fracture surfaces are also comparable, with slightly finer features (average size of dimples) in the bottom. The similarity in fracture toughness and fracture surface between the top and the bottom of the weld, despite having such different balances of FS(A) and FS(NA) is consistent with a fracture mechanism dominated by inclusions and carbides, which are stable during reheating and are expected to have a similar distribution in the top and bottom of the weld. In a mechanism dominated by carbides and inclusions, smaller inclusions result in higher toughness values, and the small size of the inclusions and carbides observed here (all below 0.5 μm) are an important factor in the high-impact values observed. Discussion In prior investigations, Koseki and Thewlis have shown that toughness and strength degrades when the weld metal Al/O ratio exceeds 1.0 (Ref. 4), since these will promote a spinel structure that does not favor acicular ferrite nucleation (Ref. WELDING JOURNAL 19-s WELDING RESEARCH Fig. 6 — AES analysis of element distributions of the Mg- and Al-rich oxide in weld metal, with A — Fe; B — O; C — Mg; D — Al maps shown.
Welding Journal | January 2014
To see the actual publication please follow the link above