MLLS fitting on plasmon pic for mapping hydrides in a Zr alloy with a complex αZr+βZr acicular microstructure obtained by water quenching
Résumé
Zr alloys are used to manufacture fuels claddings in nuclear plants. These tubes have to insure safety and prevent any fuel leak even in severe conditions such as Loss Of Coolant Accident. In such conditions the fuel cladding tubes are submitted to high temperature steam oxidation before core reflooding. This results in an oxidation, α-β phase transformations and partitioning of oxygen/hydrogen in a very complex microstructure (Fig. 1). At 1200°C, the hydrogen has low affinity for α-O and a high solubility in the β phase and most of the hydrogen moves towards the β-phase and is expected to remain in the prior β-phase after quenching. Macroscopic analysis (Elastic Recoil Detection Analysis-ERDA) evidenced hydrogen enrichment of the β-phase and depletion of the α-O. Because of the lack of spatial resolution of this technic, studying sub-micrometric hydrides needles in such complex mixture of α-O, β-Zr phases, highly strained needs to perform TEM (Fig.2). Furthermore, very thin specimens are needed to limit features overlapping in the lamella thickness. The multiphase sample cannot be thinned using electropolishing and has to be prepared using ions. In another hand, the high reactivity of zirconium with hydrogen implies accurate preparation conditions. Any additional hydrides formation during thinning is achievable with a very high vacuum in the thinning devices and with cold traps. This is achieved by a two steps specimen preparation combining Focused Ion bean and low kV Ar+ ions cleaning. A 100-150 nm thick lamella is first extracted using FEI Helios Nanolab Dual Beam at 30kV. We performed post FIB cleaning and thinning in PIPS II from GATAN with Ar ions at 500V during 60 min to eliminate the 20-30 nm amorphous and Ga+ implanted layer on both sides of the samples and to reach a thickness of about 40 nm (Fig.2). Diffraction studies are very complicated and time consuming even on such thin specimens. Based on the difference of plasmon energy of α-O, β-Zr in comparison with hydrides (ζ : 17,3-17,5 eV, γ : 18,3 eV, δ : 19,2 eV and ε : 19,6 eV reported in [1, 2 & 3]), hydrides are identified as the δ-ZrH2 and mapped using on MLLS fitting on plasmon pics. Since α-O and β-Zr phase have the same plasmon energy, the ambiguity between these two allotropic Zr compounds is solved using Fe and Cr mapping by EDXS with Super X detectors in a XFEG TECNAI OSIRIS. The high brightness of the XFEG gun and the high angle collection of the four EDXS detectors allows us to distinguish α-Zr from β-Zr based on their Fe and Cr solubility difference (few percent in β-Zr, and few hundred ppm in α-Zr [4, 5]). Correlation between all these analytical technics and microdiffraction patterns give us the possibility to confirm ERDA analysis concluding that sub-micrometric needles of hydrides, identified as δ-ZrH2, where mainly located at the interface between residual β-Zr and α-Zr (Fig. 3).
