an ~1 μm diameter. It was determined that the eutectic lamellae in the backfilled regions is much less than 1 μm, so EDS measurements taken in this region are likely to more closely represent the average Nb level. Near the edges, an average representation is less likely, as lamellae spacing appears to increase. As such, differences from predicted values were encountered in some samples, especially those with a thinner backfilled region to collect data from. Depletion of the other major elements, Ni, Cr, and Fe coincides with the Nb enrichment. Similar trends were seen with the other compositions tested. As with the 8Nb0Mo sample, the backfilled crack for the 6Nb4Mo pin also exhibited a large Nb enrichment, but to a lesser extent with spots 1, 2, and 6 averaging at 21.2 wt-% Nb. Molybdenum enrichment of an average of 7.11 wt-% was measured for the same spots. ThermoCalc™ predicted a niobium level around 20 wt-% and a molybdenum level around 13 wt-% for this alloy. While the niobium prediction is consistent with the measurement, the molybdenum prediction is overestimated. This may indicate several things: 1) the calculated partition coefficient is too low, 2) the Scheil simulation predicted a phase with a higher stoichiometric ratio of molybdenum than actually occurs, and 3) the EDS analysis may be inaccurate. Sulfur Kalpha lines overlap molybdenum L lines, but since sulfur levels in the Alloy 690 base composition were less than 0.002 wt-%, such a sulfur K-alpha contribution should be minimal. Coinciding depletion of Ni, Cr, and Fe is again observed in the backfilled region of this alloy. Discussion The results presented here indicate for Ni-based alloys that form eutectic constituents at the end of solidification, the backfilling effect makes predictions of cracking susceptibility difficult to determine simply as a function of the solidification temperature range. Previous data from DuPont et al. (Ref. 54) have shown a correlation between the solidification temperature range (STR) and cracking susceptibility in Ni-based alloys measured using the Varestraint test, as shown in Fig. 11. While this data may be relevant, it should be pointed out that at the strains used in generating this data, crack healing is probably not possible. When the cracking susceptibility curve from Fig. 2 is plotted with the corresponding predicted STR for each composition, it is apparent that STR is not the dominating factor affecting susceptibility at high-alloy additions, as shown in Fig. 12. Note that cracking susceptibility and solidification temperature range correlate well with small additions of Nb, with increased solidification temperature range resulting in increased susceptibility. However, as alloying addition increases, a relatively small decrease in STR is predicted compared to the dramatic decrease in susceptibility and the correlation is much weaker. This is because of the following: 1) the liquid composition has reached the terminal eutectic point, so further increase in STR is not possible, and 2) the amount of eutectic increases allowing for crack healing to occur. As a result, higher strains (longer pins) are required to promote solidification cracking. It should be noted that the STR values in Fig. 12 are predicted, not measured. Ongoing work will determine the actual STR values using the single sensor differential thermal analysis (SS DTA) technique (Ref. 55). The specific effect of Mo on cracking susceptibility remains unclear. It has little influence on either expanding the solidification temperature range or increasing the fraction eutectic. It does segregate to the liquid and is present in higher than nominal concentration in the eutectic liquid. As shown in Fig. 4, it seems to have no influence in the 2 and 4 wt-% Nb alloys, but its effect is dramatic in reducing cracking susceptibility in the 6Nb4Mo alloy. One possibility is that the presence of Mo somehow affects the wetting characteristics of the eutectic liquid and helps promote crack healing. It should be noted that filler metal 625, which has good solidification cracking resistance despite a wide solidification temperature range, contains on the order of 4 wt-% Nb and 10 wt-% Mo. Further work is underway to better understand the effect of Mo in high-Cr, Nb-bearing Ni-based alloys. Conclusions 1) The fraction eutectic in Ni-30Cr weld metals increases dramatically with increasing niobium content based on both the measured fraction eutectic and values calculated with ThermoCalc™. 2) Additions of Nb up to 4 wt-% result in an increase in cracking susceptibility associated with an increase in solidification temperature range and the presence of continuous liquid films along the solidification grain boundaries. 3) Niobium additions above 4 wt-% result in a decrease in cracking susceptibility due to an increase in the fraction eutectic, which promotes crack healing. 4) The peak in cracking susceptibility corresponds to a fraction eutectic of ~7.5 vol-%. At levels above ~15 vol-%, crack healing by eutectic backfilling is significant. 5) Additions of Mo to alloys containing 6 wt-% Nb resulted in a decrease in cracking susceptibility. This effect was not observed at lower Nb levels. 6) There was a good correlation between cracking susceptibility and solidification temperature range at lower Nb levels. At higher Nb levels (> 4 wt-%), the correlation is not good because of crack healing effects. 7) The cast pin tear test appears to be an effective method to evaluate the influence of backfilling on cracking susceptibility in materials where sufficient liquid of eutectic composition forms at the end of solidification. The lead author (RW) would like to thank the Graduate School at The Ohio State University for providing a fellowship during the first year of graduate studies and to the Department of Energy (DOE) Nuclear Engineering University Program (NEUP) for providing continuing support through an NEUP Fellowship. Our thanks to Paul Mason from Thermo- Calc™ for his prompt and helpful responses regarding the computational work performed in this study. Fellow graduate students Adam Hope, David Tung, and Daniel Tung are recognized for their instruction and guidance WELDING RESEARCH 236-s WELDING JOURNAL / JULY 2016, VOL. 95 Acknowledgments
Welding Journal | July 2016
To see the actual publication please follow the link above