The elemental signals do not necessarily localize at atomic-column positions because the spatial resolution of an EELS signal is constrained by the delocalization of inelastic scattering and electron channeling process. These complexities make it difficult to perform quantitative analysis with atomic resolution. When we estimate the exact value of elemental ratio with atomic scale, full quantum mechanical simulations combined with experimental result are necessary. On the other hand, if there is a criterion about accuracy of experimental result about elemental ratio for an atomic column without simulation, it would be very useful. 

In this study, atomic-resolution quantification of the elemental ratio of Fe to Mn at octahedral and tetrahedral sites in brownmillerite Ca₂Fe_1.07Mn_0.93O₅ (Fig. 1) is demonstrated using STEM-EELS. It is known that Fe and Mn ions are nearly all ordered but not fully ordered, i.e., a small number of Fe and Mn ions reside in octahedral and tetrahedral sites, respectively. It was found that a considerable oversampling of the spectral imaging data yields a spatially resolved area that very nearly reflects atomic resolution (~1.2 Å in radius) for Fe and Mn L_2,3-edge (Fig. 2). And the average relative compositions of Fe to Mn within the region were 17.7 to 82.3 ± 13.1 in octahedral sites and 80.7 to 19.3 ± 9.9 in tetrahedral sites. The actual atomic ratio was estimated by calculating the mixing of signals from nearest-neighbor columns using simple simulation based on multislice technique. It was concluded that the ratio of Fe to Mn was 14 to 86 at octahedral sites. It agrees well with the previous neutron diffraction experiment (14.4 to 85.6) which can correctly decide such information for bulk sample [1]. On the other hand, the experimental value and the estimation value from tetrahedral site have relatively large error compare with the result of neutron diffraction experiment (92.2 to 7.8). This means that an experimental oversampling SI data of Fe and Mn L_2,3-edge from octahedral site in perovskite-like structure is probably interpreted with an uncertainty of approximately 10% without simulation.

[1] Hosaka, Y.; Ichikawa, N.; Saito, T.; Haruta, M.; Kimoto, K.; Kurata, H.; Shimakawa, Y. Bull. Chem. Soc. Jpn. 2015, 88, 657-661

Acknowledgements

This work was supported by JSPS KAKENHI Grant Numbers 26706015, 19GS0207 and 22740227. It is also supported by a grant for the Joint Project of Chemical Synthesis Core Research Institutions from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by the Japan Science and Technology Agency, CREST.

Figures:

Fig1. Crystal structure of Ca2FeMnO5 along [101] projection

Fig.2 (a) HAADF and (b) Elemental <010> line scans for the Fe L2,3-edge and Mn L2,3-edge. (c) The profile of relative composition of Fe and Mn.

« Back to The 16th European Microscopy Congress 2016

EMC Abstracts - https://emc-proceedings.com/abstract/determination-of-elemental-ratio-in-an-atomic-column-by-stem-eels/