the fracture surface, indicated by black arrows. The microstructures distribute along grain boundaries or at triple grain junctions, because the trace of grain boundaries can be seen on the microstructures, indicated by white arrows in Fig. 4B. The microstructures present a typical eutectic morphology, as shown in Fig. 4C, with a similar chemical composition to that shown in Fig. 3. It can be concluded the microstructures found on the fracture appearance were the same with those in thermal simulated specimens shown in Fig. 2E and F. Discussion The Origin of the Laves Phase in FB2 Steel In 9% Cr martensitic stainless steels, the Laves phase usually occurs after long-term, high-temperature exposure, and it was found the Laves phase nucleates at martensitic lath boundaries or around M23C6 carbides, coarsening at the expense of alloy atoms in the matrix or M23C6 (Refs. 7, 8). Some researchers found the formation of the Laves phase at 650ºC needed at least hundreds of hours (Ref. 7). As to virgin FB2 steel, it was impossible for the Laves phase to form during tempering due to the too short time. There is only one possibility left — that the Laves phase formed in prior processing, such as casting and forging. As mentioned in the introduction, a similar Laves phase was also found in virgin CB2 steel, a cast steel with similar chemical composition to FB2. In virgin CB2 steel, sparse, relatively large particles of the Laves phase (Fe, Cr2)Mo are present in interdendritic areas (Ref. 12), indicating that the Laves phase forms during the casting process. Considering that the FB2 specimens observed also went through the casting process, the Laves phase was most likely formed in this process. The results of our experiment showed the Laves phase remained stable at temperatures below 1200ºC, so the forging process following casting could not eliminate it. In forging, all the dendrites recrystallized, and large Laves particles remained within the austenitic grains. In addition, it was noted that the formation of the Laves phase should be attributed to dendritic segregation in the casting process. Fig. 5 — Equilibrium phase diagram of the eutectic constituent obtained by JMatPro®. Constitutional Liquation Resulting from the Eutectic Reaction between the Laves Phase and Austenite The occurrence of eutectic microstructures strongly suggests a liquation phenomenon in specimens experiencing peak temperatures above 1250ºC (including 1250ºC), implying a liquation crack tendency in the HAZ of FB2 steel during welding. The results of hot ductility tests further confirmed the existence of liquation crack in FB2 steel at 1350ºC. A prerequisite for liquation crack in the HAZ was the formation of discrete liquid regions within the solid metal experiencing enough high temperatures during welding. If this liquid exhibits a tendency to wet, and thereby form a continuous or semicontinuous thin film of liquid along grain boundaries, and if sufficient strains are present in this weld region, then intergranular separation will occur. The origin of such liquation in the HAZ was most often attributed to the “constitutional liquation” phenomenon, which was originally proposed by Savage and coworkers (Ref. 13). This phenomenon involves a eutectic reaction between a secondary constituent phase and the matrix, and mostly occurs in austenitic stainless steels such as Alloy 718, A-286, GH150, and so on. During the past decades, HAZ liquation in austenitic stainless steels containing Nb and Ti has been attributed to the constitutional liquation of Nbrich Nb(C, N) carbonitrides (Refs. 14–16), Ti-rich TiC carbides, and the Laves phase (Refs. 17, 18). In the present study, results of the thermal simulation with different peak temperatures suggest that constitutional liquation of the Laves phase in virgin FB2 steel accounts for the liquation phenomenon. The heating rate of the welding thermal simulation was nonequilibrium, which caused rapid decomposition of the Laves phase. This decomposition highly enriched the region adjacent to the dissolving Laves particles in solutes such as Cr and Mo. The equilibrium phase for this high-solute composition was a liquid that surrounded the dissolving Laves particles. The liquid region remained until the Laves particles dissolved completely. It was noted that the Ac3 of FB2 steel at a heating rate of 100ºC/s measured previously was about 960ºC, so the matrix had all transformed into austenite at peak temperatures of thermal simulation. Therefore, the constitutional liquation in the heating process of thermal simulation resulted from the eutectic reaction between the Laves phase and austenite. In specimens experiencing a peak temperature of 1250ºC, every eutectic structure was separated, and the size was similar to that of the Laves phase in virgin FB2 steel, as shown in Fig. 2A–D. This phenomenon can be explained by the poor flowing ability of liquid at 1250ºC. In the present work, the Laves phase remained stable at 1200ºC, and when the peak tempera- WELDING RESEARCH 260-s WELDING JOURNAL / JULY 2016, VOL. 95
Welding Journal | July 2016
To see the actual publication please follow the link above