Microstructural characterization of welded unions of cast refractory steels

This work presents a microstructural study of welded unions of cast heat resistant (refractory) steels, type HC (29Cr-2Ni-0,4C) and HD (28Cr-5Ni-0,4C) using quantitative metallography and X-ray diffraction. Two series of alloys were prepared in an induction furnace and cast in sand molds with phenolic resin, in the classic 'Y' shape, with thicknesses of 12, 25 and 50 mm, following ASTM A395. These samples were chemically and micro structurally characterized, in order to verify their chemical composition and the presence and distribution of phases in the material. Test pieces were then cut from the test section of the 'Y' samples to produce the welded unions, which were done with a single and double bevel (X) butt, using a manual arc and one with electrodes AWS E 309 and AWS E 410. These unions were also submitted to thermal treatments of 780oC, in order to study the forced precipitation of the sigma phase, simulating service conditions in a production environment. The presence and distribution of phases in the fusion zone (welding cord) and in the thermally affected zones (TAZ) next to both sides of the cord were studied. They displayed a relatively low dilution compared to the composition of the electrodes. The presence of austenite in a ferritic matrix, with relatively small amounts of Cr23C6 and Cr7C3 carbides was shown together with an incipient presence of sigma phase. The annealing of the test pieces at temperatures close to those for the use of these materials, resulted in a new sigma precipitation, in different relative amounts, depending on the time of exposure to these temperatures. Sigma phase precipitates in low relative amounts simply from welding, but this event deserves more attention when the material is submitted to post-welding thermal treatments, or more intensely, to prolonged exposures to higher working temperatures. Quantitative metallographic techniques with contrast and coloring of phases can be used to quantify the micro-constituents; the process is optimized even more when the phases are clearly identified by X-ray diffraction. However, the identification of sigma phase with X-ray diffraction is accurate as long as the possible texturization of the main phases, ferrite and austenite, does not interfere by reducing the characteristics signs of sigma. In this case a correction of the effects of texture associated with the solid solution phases, is essential for obtaining the successful comparison of the diffraction model with the experimental data, ensuring a highly reliable quantification of the component phases (CW)