be related to impurity content. At all three travel speeds, at the Creq/Nieq for the transition between primary ferrite and primary austenite solidification, a variation in solidification mode with impurity content was observed. For a constant Creq/Nieq of 1.55 at 21 mm/s and 1.75 at 42 and 85 mm/s, primary ferrite solidification was observed at low impurity contents and dual or primary austenite solidification modes are observed at high impurity content. The change in solidification mode could be due to impurity content affecting the solidification mode, or some other factors not captured in the Cr and Ni equivalencies such as elements not included in the regression or interaction effects between elements. Possible elements not captured in the Espy equivalents include Ti, a ferrite stabilizer, and Co, an austenite stabilizer. Considering the compositions of Alloys 1 and 59, which exhibit the variability in solidification mode at constant Creq/Nieq for 42 and 85 mm/s travel speed welds, there is only minor variation in the Ti and Co contents between alloys. If anything, the slightly higher Co content of Alloy 1 should slightly increase the tendency for primary austenite solidification relative to Alloy 59. Literature (Refs. 24, 25) has shown that P and S can lower Cr content of the matrix by forming Crrich sulfides and phosphides during solidification, effectively reducing the Cr equivalent. Brooks et al. (Ref. 6) observed a similar shift in solidification mode as impurity level increased at constant Creq/Nieq for welding of freemachining austenitic stainless steels and incorporated a similar slope to the vertical portion of the cracking boundary in the weldability diagram developed in that work. The influence of P and S levels of solidification behavior are likely the cause of the shift in solidification mode at constant Creq/Nieq, and the slight slope of the vertical cracking boundary line is appropriate. The outlier (compared to the trend observed for 21-6-9 solidification mode) in the 21 mm/s travel speed diagram of Alloy 35 with some primary austenite solidification at Creq/Nieq of 1.70 agrees with work from Robino et al. (Ref. 26) where it was observed that Gall-Tough®, an alloy similar to Nitronic 60, shifted to primary austenite solidification with pulsed laser welding at a Creq/Nieq of 1.79. The reason for the outlier in the 85 mm/s diagram with primary austenite solidification in Alloy 30 at Creq/Nieqof 1.92 is unknown. The uncertainty values of ±0.019 Creq/Nieq and ±0.0015 wt-% impurity content discussed above represent one standard deviation Fig. 7 — Total crack length as a function of travel speed for all alloys showing cracking. of the chemical composition analysis from the three analyses on the same laboratory equipment. The variation in chemical composition analysis between laboratories must also be considered when applying the results of the weldability diagrams. Recent work from Kotecki and Zhang (Ref. 27) highlights the results on chemical composition analysis variability from round robin interlaboratory testing of chemical composition for several stainless steel alloys. To consider the possible interlaboratory variability and bias, the Creq/Nieq uncertainty was estimated using published standard deviations for chemical analysis from both round robin interlaboratory chemistry testing (Refs. 27, 28) and ASTM E1086 (Ref. 29), the relevant standard. Calculating the uncertainty of Cr and Ni equivalents, the possible range of Creq/Nieq is approximately ±0.13 Creq/Nieq. Using the interlaboratory standard deviation for P and S and using OES analysis from ASTM E1086 gives an uncertainty of ±0.0018 wt-% P + 0.2S, slightly higher than the impurity content uncertainty calculated from the single laboratory value. The interlaboratory impurity content uncertainty should be reduced slightly when using inert gas fusion technique to measure S as was done in this work. The interlaboratory variability of impurity content is not likely to make a large difference when applying the weldability diagram to predict solidification crack susceptibility, but the large interlaboratory variation in Creq/Nieq WELDING RESEARCH could impair the applicability of the diagrams. Such a large variation in measured chemical composition would not be expected typically, but the large magnitude in possible variation of Creq/Nieq must be taken into account when applying the weldability diagram. Calibration of the OES instrument to a 21-6-9 type standard when measuring the chemical compositions should reduce the uncertainty closer to the levels calculated for the single instrument analysis of ±0.019 Creq/Nieq and ±0.0015 wt-% impurity content. As reported in Part 1, average solidification rates of 6 mm/s at 21 mm/s travel speed, 13 mm/s at 42 mm/s travel speed, and 25 mm/s at 85 mm/s travel speed were observed at 50% penetration depth, which approach the solidification rates reported for pulsed laser welding. Knowing that the solidification rates at 85 mm/s are close to pulsed laser welding that was used to develop a weldability diagram for 300 series stainless steels (Ref. 5) allows for comparison between that diagram and the 85 mm/s diagram. With pulsed laser welding, Lienert and Lippold reported a minimum Creq/Nieq of 1.69 (H&S) for primary ferrite solidification, which is similar to the 1.75 Espy Creq/Nieq observed for 21-6- 9 at 25 mm/s solidification rate. Comparing diagrams developed with different equivalents is subjective, and the lower solidification rates of this work compared to pulsed laser welding must also be considered. For pulsed laser welding of 21-6-9 or higher travel speeds with continuous-wave laser NOVEMBER 2016 / WELDING JOURNAL 415-s
Welding Journal | November 2016
To see the actual publication please follow the link above