suboxide eels and t-ebsd

Download suboxide EELS and t-EBSD

Post on 11-Aug-2015




2 download

Embed Size (px)


  1. 1. Micron 69 (2015) 3542 Contents lists available at ScienceDirect Micron journal homepage: Identifying suboxide grains at the metaloxide interface of a corroded Zr1.0%Nb alloy using (S)TEM, transmission-EBSD and EELS Jing Hua, , Alistair Garnerb , Na Nic , Ali Gholiniab , Rebecca J. Nichollsa , Sergio Lozano-Pereza , Philipp Frankelb , Michael Preussb , Chris R.M. Grovenora a Department of Materials, University of Oxford, Parks Road, Oxford, UK b Materials Performance Centre, School of Materials, University of Manchester, Manchester, UK c Department of Materials, Imperial College London, Royal School of Mines, London, UK a r t i c l e i n f o Article history: Received 3 September 2014 Received in revised form 20 October 2014 Accepted 20 October 2014 Available online 29 October 2014 Keywords: Zirconium alloys (S)TEM Transmission-EBSD Transmission Kikuchi Diffraction EELS Suboxide Oxidation mechanism a b s t r a c t Here we report a methodology combining TEM, STEM, Transmission-EBSD and EELS to analyse the struc- tural and chemical properties of the metaloxide interface of corroded Zr alloys in unprecedented detail. TEM, STEM and diffraction results revealed the complexity of the distribution of suboxide grains at the metaloxide interface. EELS provided accurate quantitative analysis of the oxygen concentration across the interface, identifying the existence of local regions of stoichiometric ZrO and Zr3O2 with varying thickness. Transmission-EBSD conrmed that the suboxide grains can be indexed with the hexagonal ZrO structure predicted with ab initio by Nicholls et al. (2014). The t-EBSD analysis has also allowed for the mapping of a relatively large region of the metaloxide interface, revealing the location and size distribution of the suboxide grains. 2014 Elsevier Ltd. All rights reserved. 1. Introduction Zirconium alloys are used extensively as cladding materials in modern light water reactors. The purpose of the cladding mate- rial is to separate the uranium dioxide (UO2) fuel rods and the coolant water in order to prevent the escape of ssion products whilst maintaining heat transfer to the coolant. The widespread use of these alloys is mainly due to their low thermal neutron cap- ture cross section and good corrosion performance in the aggressive reactor environment (Pickman, 1994; Cox, 1961; Cox et al., 1998). With increasing demand for high burn-up in modern nuclear reactors, environmental degradation of these alloys is now the life- limiting factor for fuel assemblies (Pickman, 1994). The general waterside corrosion kinetics of zirconium alloys has two stages: the rst is an initial pre-transition period of slow parabolic or cubic corrosion which can be tted using the power law equation, W = Ktn (W is the weight gain in mg/dm2, K depends on both alloying elements and reactor temperature, and t is the oxidation time in days). At a critical thickness, the previously Corresponding author. Tel.: +44 07583260056. E-mail address: (J. Hu). protective oxide begins to break down and there is an abrupt increase in corrosion rate, often termed the transition. The pro- tective oxide then builds up again and post-transition corrosion is composed of several cycles that mimic pre-transition corrosion until a period of rapid linear growth eventually develops (Hillner, 1977; Garzarolli et al., 2012). Post-transition corrosion generally follows an expression of form W = Kt + C (where C is the initial weight gain). The Zr1.0%Nb alloy used in this investigation has been shown to exhibit a delayed transition when compared to other Nb-containing alloys (Wei et al., 2012). During corrosion, an adherent oxide lm is formed by the dis- sociation of water molecules at the outer surface of the oxide. Oxygen ions then diffuse inwards, forming an oxide that grows into the metal. The metaloxide interface is thus of particular inter- est as it is the location of the oxidation reaction and is expected to play an important role in controlling the corrosion behaviour. Yilmazbayhan et al. (2006) used Transmission Electron Microscopy (TEM) to show there are 100150 nm wide rectangular blocky suboxide grains near the metaloxide interface on a Zr2.5%Nb alloy. Abolhassani et al. (2010) also identied a substochiomet- ric phase with 4060 at.% oxygen at the interface of irradiated Zr2.5%Nb using Energy Dispersive X-ray Spectroscopy (EDS). Ni and Hudson (Ni et al., 2012) used Electron Energy Loss Spectroscopy 0968-4328/ 2014 Elsevier Ltd. All rights reserved.
  2. 2. 36 J. Hu et al. / Micron 69 (2015) 3542 (EELS) and Atom Probe Tomography (APT) to study oxidised ZIRLOTM and Zircaloy-4 alloys, and found a suboxide layer with composition of ZrO at the metaloxide interface in pre-transition samples which disappears after transition. Nishino et al. (1996) used Auger Electron Spectroscopy (AES) and X-ray Photoelectron Spectroscopy (XPS) to study the initial oxidation of Zircaloy-2 at room temperature, and identied Zr2O, ZrO and Zr2O3 suboxide phases on the oxidised surfaces. The equilibrium ZrO binary phase diagram contains no stable phases with these stoichiometries, but Density Functional Theory (DFT) modelling performed by Puchala and Van der Ven (2013) and Nicholls et al. (2014) predicts a stable ZrO suboxide phase with a hexagonal structure as well as other stoi- chiometric phases not in the ZrO binary phase diagram, including Zr2O. It has previously been demonstrated that the micro-chemistry and crystallography at this metaloxide interface is extremely com- plex and localised (Ni et al., 2012). The techniques used to examine the metaloxide interface are usually focused on a very small region and so it is difcult to correlate the local appearance of suboxide phases to the overall corrosion performance of different alloys. In this paper we present a methodology combining TEM, STEM, t-EBSD and EELS analysis to characterise the chemistry and crystal- lography of suboxide phases at the metaloxide interface. T-EBSD also allows for characterisation of a relatively large suboxide region, giving a more general view of what is happening at this particular stage of the oxidation process. 2. Materials and methods The sample used for this investigation was prepared from a WestinghouseTM developmental alloy with composition of Zr0.9Nb0.01Sn0.08Fe (wt%) in the recrystallized condition. The sample was oxidised in an autoclave at EDF Energy under simu- lated Pressurised Water Reactor (PWR) water conditions at 360 C for 360 days, and shows no sign of accelerated oxidation rate and so is considered to be in the pre-transition stage of the corrosion cycle (Wei et al., 2012). A TEM sample was prepared by in situ lift out method on an FEI FIB 200 using milling currents of 7000100 pA at 30 kV, and further thinned down to a thickness below 100 nm in a Zeiss Nvision 50 dual beam FIB using a beam current of 150 pA at 30 kV and then 250 pA at 5 kV to create a homogeneous and electron-transparent sample for TEM, STEM and t-EBSD analysis. The sample was further thinned down to 50 nm using the low- voltage conditions for nal EELS analysis on selected regions of interest. TEM and STEM analysis were performed on a JEOL 2100 LaB6 microscope operated at 200 kV. EELS analysis was performed on an FEI Titan microscope operated at 300 kV and equipped with a Gatan image lter. The convergence half-angle was 10 mrad and collection half-angle was 12 mrad with an energy dispersion of 0.5 eV per channel and a step size of 10 nm. T-EBSD (Trimby, 2012) was performed on an FEI Magellan FEG-SEM XHR 400L at 30 kV with a probe current of 1.6 nA. The step size for t-EBSD was 15 nm, with an acquisition speed of 33.5 Hz, so that the time to acquire a map time was 70 mins. In order to achieve optimal spatial res- olution, the EBSD measurements were performed in transmission geometry. The sample was tilted at 20 away from the EBSD detec- tor with a working distance of 2 mm. The Kikuchi patterns were indexed using the AZtec software suite developed by Oxford Instru- ments. The EBSD measurements in transmission geometry allowed an optimal spatial resolution of 10 nm, necessary for the study of the extremely ne microstructure at the metaloxide interface (Garner et al., 2014). 3. Experimental results 3.1. (Scanning) transmission electron microscopy Fig. 1(a) shows a bright eld (BF) image of the TEM sample. Below the Pt layer, which is used to protect the sample during FIB milling, a ZrO2 layer of thickness 2.8 m can clearly be seen. The oxide layer consists of equiaxed grains on the outer surface and columnar grains towards the metaloxide interface. There are some lateral cracks visible throughout the oxide, the largest of which are located near the metaloxide interface. The high angle annu- lar dark-eld (HAADF) STEM image of the same area in Fig. 1(b) shows metallic second phase particles embedded in the oxide with voids above them. Below the metaloxide interface, large zirconium metal grains are visible in both images, with embed- ded metallic second phase particles. These second phase particles are -Nb and Zr(Nb,Fe)2 particles. A similar general microstruc- ture has been seen in oxidised Zr alloys by many previous authors (Yilmazbayhan et al., 2006; Ni et al., 2012). The insets in Fig. 1(a) and (b) are higher magnication TEM bright eld and STEM HAADF images of the region chosen for EELS analysis. In all these images, there is a slight contrast change at the metaloxide interface but it is difcult to distinguish if there are phases other than ZrO2 and metallic Zr in this region. Convergent beam electron diffraction (CBED) patterns were acquired from a region close to the metaloxide interface just between the two cracks shown in the higher magnication TEM bright eld image in Fig. 2(a). A suboxide grain which is strongly diffracting is highlighted by the dotted lines. A STEM dark