The Influence of Liquid Lead-Bismuth Environment on Mechanical Properties and Corrosion Resistance of Austenitic Stainless Steel Welds

In the context of the development of MYRRHA, the new Generation IV research reactor that will be built at the Belgian Nuclear Research Centre (SCK CEN), structural materials need to be selected. Since the reactor will be cooled by a heavy liquid metal, more specifically Lead-Bismuth-Eutectic (LBE), the selected materials need to be resistant to possible degradation effects in this environment. The two degradation effects of main interest are Liquid Metal Corrosion (LMC) and Liquid Metal Embrittlement (LME). For this purpose, austenitic stainless steels seem promising candidates. In this work, the effect of the liquid LBE environment was assessed on the welds of such materials, as a structure is only as strong as its weakest link and welds have a tendency to act as such. Two 316L-type annealed austenitic stainless steel plates with a 32 mm and 75 mm thickness, welded by Gas Tungsten Arc Welding (GTAW) and Submerged Arc Welding (SAW), respectively, have been investigated and were compared to the so-called DEMETRA heat as reference 316L material. The chemical composition, microstructure, mechanical properties in air and in LBE and the corrosion properties in LBE of all materials and their welds were characterized and compared. Both welded plates had a higher Cr-, Mo- and Ni-content than the DEMETRA reference heat. In the case of the 75 mm thick plate, this resulted in a homogeneous solution annealed plate with the appropriate microstructural features of a 316L-type material. For the 32 mm thick plate, the increased Cr- and Mo-content, together with chemical segregation in the centre of the plate, led to the formation of intermetallic σ-phase. Because of the similar appearance of δ-ferrite and σ-phase, the σ-phase can remain undetected when not analysed in depth. However, this additional phase can significantly influence the mechanical behaviour of the material, as was experienced in this work. In-depth Electron Backscatter Diffraction (EBSD) analysis and mechanical testing in the through-thickness orientation have proven helpful in solving this issue of identifying the phase, where classical etchants and testing of standard oriented tensile samples in the rolling and transverse orientation
fall short. Therefore, adding those procedures to the standard qualification procedures for these specific steel grades is strongly recommended.
Based on Slow Strain Rate Tensile (SSRT) tests and fracture toughness C(T) tests in the liquid metal environment, no effect of the LBE on the mechanical behaviour within the investigated conditions was observed on any of the base and weld materials. During the reference SSRT tests in air, the deformation behaviour of the base and weld materials at different temperatures was investigated by 3D Digital Image Correlation (3D-DIC). This technique has shown to be an effective tool to characterize the local deformation behaviour of complex materials such as welds. It was proven to be more sensitive (and for some tests even crucial) for observing localized effects directly on the specimen surface. An example of such a localized effect is the Portevin-Le Châtelier (PLC) effect, where bands of localized strain move through the material. The DIC technique allows to observe more ieasily the presence of this effect, compared to if only the size and shape of the serrated stress response on the stress-strain curves due to Dynamic Strain Ageing (DSA) are considered. In the tested temperature and strain rate regimes, type A PLC bands, identified by DIC, move continuously through the homogeneous base material, as reported in literature. However, in case of heterogeneous materials such as welds, these type A PLC bands
move in a discontinuous way through the material, due to the heterogeneous deformation during the initial stage of the mechanical tests at high temperature. Additionally, the weld material itself seems more susceptible to DSA and the PLC effect, due to its heterogeneous character and consequently induced strain rate gradients across the material. The investigations into the corrosion behaviour of the materials were preliminary; the differences in behaviour between 316L base and weld metal were sought. It should be underlined that the test conditions with respect to temperature and flow pattern of the liquid metal did not necessarily correspond to real operating conditions for the welds in an LBEcooled reactor such as MYRRHA and only reflect the current experimental possibilities in the laboratory. In stagnant LBE with a low dissolved oxygen content, dissolution is the active corrosion mechanism. Under these conditions, the weld metal was proven more prone to corrosion than the base metal. After an incubation period, the LBE progressed faster in the weld material than in the base material, at the expense of the 3D interconnected dendritic ferrite structure. In flowing LBE with a controlled dissolved oxygen content, the weld metal seemed to be more resistant to corrosion than the base metal. Under these conditions, the microstructure of the weld played a less crucial role than when dissolution was the active corrosion mechanism. Local flow variations of the LBE in the regions near the material surface were put forward as a possible reason for the significant erosion degradation observed, since the most degraded samples were located in regions where turbulent flow can be expected. For stagnant conditions, no degradation was observed both for the base and weld metals, for at least 2000 h of exposure at an oxygen content of 3 · 10-8 wt.%. The turbulent flow should therefore be avoided, which means equivalently avoiding sharp turns, edges or uneven finishing of welds by overlay welding or excess weld reinforcement.
This work provides an initial but dedicated view at the behaviour of weld metal in LBE environment; however, further investigations are necessary. The full joint includes a range of zones with differences in local microstructures (mushy zone, unmixed zone, heat affected zone). In this study, mainly the fusion zone was investigated, however the other zones might behave differently under these conditions. In addition, in a real life component, the weld will not be exposed to LBE in a manner that was tested here: i.e. the central part of the fusion zone with a polished surface. Although this work showed that the welds did not necessarily behave worse than the base materials in the liquid metal environment, still care should be taken when designing an LBE-cooled reactor. Welds might still be the weakest link and proper weld design is needed with any kind of structure.