May_EDFA_Digital

edfas.org 29 ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 23 NO . 2 The valence state of Mn in the Mn-rich phase is also investigated. The Mn- L 2,3 edge of the Mn-rich phase is background subtracted and subsequently corrected for plural scattering [11] to highlight the electron loss near edge structure (ELNES). The processed Mn- L 2,3 edge is given in Fig. 10. The Mn- L 2,3 edge acquired from bulk LSM phase is superimposed for comparison. It is seen that the peak energy of the Mn- L 2,3 edge is shifted slightly to lower energy for the Mn-rich phase when compared with the LSM phase. In addition, a larger L 3 / L 2 ratio is recognized for the Mn-rich phase. For a quantitative comparison and valence state determination, themeasured edge position and calculated L 3 / L 2 ratio are summarized in Table 1 for both the LSM and Mn-rich phases. The valence of Mn is determined to be approximately +3 for LSM with a slight portion of +4, and a mixture of +2 and +3 for the Mn-rich phase. [20,21] Accordingly, the Mn-rich phase is likely to be a Mn 3 O 4 phase. DISCUSSION Two substantiallydifferent LSM-YSZ interfacial reaction mechanisms were observed for the two tested cells. The difference in LSM-YSZ reactions is suggested to correlate with the different behavior of Mn transportation, including the Mn solubility in the YSZ phase and the Mn excess in the LSM phase. Both properties vary with oxygen partial pressure and temperature and therefore play important roles in affecting Mn transportation, LSM stability, and the resulting LSM-YSZ reactions. To achieve a better understanding of the Mn ion diffusion mechanism in YSZ and LSM at these testing conditions, the Mn solubility in 8YSZ and the Mn excess in LSM were calculated with respect to a wide range of oxygen partial pressures. The plots are presented in Fig. 11. The Mn solubility in 8YSZ is presented in cation percentage and the Mn excess in LSM corresponds to x in (La 0.75 Sr 0.25 ) 1 Mn 1+ x O 3±δ . As shown in Fig. 11, the Mn solubility in YSZ is calculated to be ~5.5 cat.% at 1050°C (sintering), ~2.1 cat.% at 750°C (air- tested), and ~1.7 cat.% at 750°C (oxygen-tested). The Mn excess in LSM is calculated as x ~1.5% atomic fraction at 1050°C in air (sintering), ~-0.05% atomic fraction at 750°C (air-tested), and ~1.0% atomic fraction at 750°C (oxygen- tested). It is noted that the P O 2 for the testing condition is the value in the vicinity of cathode/electrolyte interface, which is calculatedby taking the cathode polarization into account. [5] By incorporating the Mn solubility in the YSZ and LSM phases as a function of temperature and oxygen partial pressure, the experimentally observed LSM-YSZ interactions and their impact on cathode degradation can thus be interpreted as the following. Table 1 Mn- L 2,3 edge datameasured from quantitative EELS in bulk LSMand Mn-rich phase. Bulk LSM phase Mn-oxide phase L 3 (eV) 646.8 645.6 L 2 (eV) 657.6 656.8 ΔE ( L 2 - L 3 ) (eV) 10.8 11.2 L 3 / L 2 2.33 3.24 Valence state +3 (slightly +4) +3, +2 Fig. 9 The STEM-EELS spectra acquired across the same YSZ (red)/Mn-rich grain (dark green)/LSM (blue) phases in Fig. 8. I 1 and I 2 are designated to the YSZ/Mn-rich grain and Mn-rich grain/LSM interface, respectively, (oxygen flow cell). Fig. 10 Comparison of ELNES fine structure of Mn- L 2,3 edge corresponding to LSM and Mn-rich phase, (oxygen flow cell).