Correlating Laboratory-Based X-Ray and Electron Microscopy to Characterize Defects and Structural Evolution in Solid Oxide Fuel Cells

Abstract

Nano- and micro-scale structural characterization, in 3D, are critical to understanding the structure-property relationships in energy materials. In solid oxide fuel cells, long-term structural degradation due to such effects as Ni coarsening or oxidation mean that additionally these structures are not static. Such issues have plagued widespread SOFC adoption, and are inherently difficult to study due to the length and time scales of the heterogeneous material systems. Recent approaches to the problem have largely employed synchrotron X-ray tomography and FIB-SEM tomography as microstructure characterization techniques. To date, however, there has been very limited application of laboratory X-ray microscopy which has the potential to offer the combined benefits of round-the-clock access and nondestructive 3D nanoscale imaging. Recently, simultaneous improvements in X-ray optics and detection have greatly increased the application space of lab nanoscale XRM, which now serves as a viable solution for SOFC characterization. In this work, multiple lab-based imaging techniques are combined in a correlative manner to probe SOFC microstructure. Specifically, the nondestructive nature of laboratory XRM is combined with the high resolution of FIB-SEM microscopy and EDS to characterize SOFC components through a 4D evolution experiment and spanning the relevant length scales from microns to nanometers. As an initial test, an as-fabricated un-reduced SOFC sample (NiO/YSZ anode) was examined using nanoscale XRM operating at 5.4 keV and 50 nm spatial resolution. Whereas synchrotron imaging approaches have frequently exploited the tunable nature of the incident beam to perform absorption edge imaging and enhance material contrast, the laboratory-based instrument used in this work operates at a single, fixed beam energy. Nonetheless, it was found that the 5.4 keV X-rays produced sufficient contrast to discern solid phases of the multiple cell components contained in the sample: NiO/YSZ anode, YSZ electrolyte, and LSM cathode. Furthermore, additional data collection in phase contrast mode by employing a Zernike phase ring was used to highlight the existence of small features including nanoporosity (defects) in the electrolyte layer. A second SOFC anode sample was sourced, this time in the reduced Ni/YSZ state with small particle sizes characteristic of the active layer of the electrode. A 3D scan was first performed with the XRM, and an advanced segmentation algorithm based on both grayscale and texture-based contrast was used to segment the Ni, YSZ, and pore phases. The sample was then mounted in a FIB-SEM instrument which was used to shave a small amount of material from the sample to create a flat, polished surface for EDS mapping. An EDS map was generated and compared to the corresponding 2D plane of the XRM dataset to validate the XRM capability to accurately characterize the structure. The sample was then degraded by an oxidation process by holding the sample at 700C in ambient air for 1 hour. A second XRM scan with the same imaging conditions was performed after oxidation. The before and after datasets were spatially registered and compared to evaluate the effects of Ni oxidation. Substantial swelling of the Ni upon oxidation, at the expense of the void space, with minimal modification of the YSZ were observed as expected. Furthermore, it is realized there may be additional modification of the microstructure at a length scale smaller than can be viewed with the XRM (less than 50 nm). To probe these finer structures, the sample was subsequently brought back to the FIB-SEM, which was used to perform 2D EDS as well as 3D FIB tomography. Results from the FIB tomography were then correlated back to the 4D before-and-after data set obtained by XRM.