Delta ferrite in modified 9Cr–1Mo steel weld metal

Abstract

The role of chemical composition on the presence of δ-ferrite in P91 weld metal in the as-welded condition was studied by systematically varying five alloying elements (Cr, Ni, Mn, Si, and Mo) in 28 gas–metal arc welds. The effect of thermal treatment on the δ-ferrite content was also investigated. Metallography analysis of the as-welded beads showed that 18 of the 28 welds contained 0.3% to 6.2% δ-ferrite in martensite. Twenty of the 28 welds fully complied with chemical composition specifications, including the limit of the sum of (Mn + Ni). Of the twenty welds that complied with the AWS A5.28/A5.28M:2020 (ER90S-B9) chemical composition specification, thirteen contained δ-ferrite. Based on these results, lower specification limits for Mo and Si are proposed to ensure a fully martensitic microstructure. Empirical formulae were applied to evaluate their accuracy in predicting the P91 as-welded microstructure: ten of the 28 welds (36%) were incorrectly predicted. Based on these results, the recommended Schaeffler Creq and Kaltenhauser ferrite factor maximum limits should be lowered to 11.5 and 6, respectively, to limit the presence of δ-ferrite in P91 as-welded microstructure. Phase-transformation temperatures were evaluated against the amount of δ-ferrite in beads in the as-welded condition. It was observed that a wider (Ae4 − Ae3) temperature range was beneficial in suppressing δ-ferrite in the final as-welded microstructure. Higher peak temperatures and slower cooling rate during welding also reduced the amount of δ-ferrite in P91 weld metal.