Influence of thickness and interface on the low-temperature enhancement of the spin Seebeck effect in YIG films Guo, E.-J., Cramer, J., Kehlberger, A., Ferguson, C. A., MacLaren, D., Jakob, G., and Kläui, M. (2016) Influence of thickness and interface on the low-temperature enhancement of the spin Seebeck effect in YIG films. Accepted for publication in Physical Review X, 2016 ------------------------ Only the transmission electron microscopy data used to create figure 5 is available in this repository; other data may be available by contacting the lead author: Prof. Dr. M. Kläui Institute of Physics Johannes Gutenberg-University Mainz 55128 Mainz Germany Tel. +49 6131 39 23633 Fax. +49 6131 39 24076 Email. Klaeui@uni-mainz.de ------------------------ Raw data TEM files are in Gatan's Digital Micrograph file format (details at the time of writing are advertised at: http://www.gatan.com/products/tem-analysis/gatan-microscopy-suite-software). Data was acquired using a JEOL ARMcFEG instrument operated at 200kV, using a Gatan Quantum electron spectrometer for EELS measurements (see manuscript for details). Spectrum imaging is performed by first collecting a dark field 'survey image' then acquiring data within a sub-region indicated within the survey. For the datasets used here, two EELS spectra (low-loss and core loss) an EDS spectrum and dark field images from two detectors ('Gatan DF' and 'Gatan HAADF') are recorded; the file names generally indicate the nature of each subset of the data. See references 26 & 27 for descriptions of the spectrum imaging technique. The files included in the repository are as follows. * Fig5a_CTEM.dm3 & Fig5b_CTEM.dm3 Conventional TEM images collected from the two samples under bright field conditions and without an objective aperture. * Fig5bd_EDS Spectrum Image.dm3 * Fig5bd_EELS Spectrum Image (high-loss).dm3 * Fig5bd_EELS Spectrum Image (low-loss).dm3 * Fig5bd_Gatan BF_DF.dm3 * Fig5bd_Gatan HAADF.dm3 * Fig5bd_SI Survey Image.dm3 Spectrum Image files used to derive the data presented in the inset to figure 5b and the line profiles in figure 5d. Note that the area of data collection is similar to - but not exactly the same as - the area indicated in the conventional TEM image. Image contrast profiles from the TEM and STEM images were used to align the two data sets. The metadata regarding instrument settings contained within the files may not be accurate. A 40 micron objective aperture, spot size 7 and 2cm camera length are typically used to collect spectrum image data (these values are particular to the ARM instrument.) * Fig5bd_Fe L Signal.dm3 * Fig5bd_O K Signal.dm3 * Fig5bd_Pt L_EDS Signal.dm3 * Fig5bd_Pt M Signal.dm3 * Fig5bd_Y L Signal.dm3 Elemental signals calculated from the spectrum image data sets - all are EELS signals except for that labelled EDS, which was used to plot the Pt profile in this case. Standard EELS processing was applied, including zero loss alignment, background substraction and deconvolution. EDS signals were derived from an energy window applied to the data set without subtraction of the generally low background. * Fig5bd_Profiles.dm3 * Fig5bd_RGB_Fe_null_PtEDS.dm3 * Fib5bInset_RGB_Fe_null_PtEDS.jpg Processed data using the above elemental maps. The line profiles are generated by binning the signals along the direction parallel to the interface. All line profiles were normalised to their maximum signal level for ease of comparison. The RGB file is generated by overlaying false-coloured maps of the distributions of Pt (EDS signal, blue) and Y (EELS signal, red). * Fig5ac_EDS Spectrum Image.dm3 * Fig5ac_EELS Spectrum Image (high-loss).dm3 * Fig5ac_EELS Spectrum Image (low-loss).dm3 * Fig5ac_Gatan BF_DF.dm3 * Fig5ac_Gatan HAADF.dm3 * Fig5ac_SI Survey Image.dm3 Spectrum Image files used to derive the data presented in the inset to figure 5a and the line profiles in figure 5c. Note that the area of data collection is similar to - but not exactly the same as - the area indicated in the conventional TEM image. Image contrast profiles from the TEM and STEM images were used to align the two data sets. The metadata regarding instrument settings contained within the files may not be accurate. A 40 micron objective aperture, spot size 7 and 2cm camera length are typically used to collect spectrum image data (these values are particular to the ARM instrument.) * Fig5ac_Cu L Signal.dm3 * Fig5ac_Fe L Signal.dm3 * Fig5ac_O K Signal.dm3 * Fig5ac_Pt L Signal.dm3 * Fig5ac_Pt M Signal.dm3 * Fig5ac_Y L Signal.dm3 Elemental signals calculated from the spectrum image data sets - all are EELS signals except for that labelled Pt L, which is an EDS signal. The EELS signal was used in this case to to plot the Pt profile and distribution. Standard EELS processing was applied, including zero loss alignment, background substraction and deconvolution. EDS signals were derived from an energy window applied to the data set without subtraction of the generally low background. * Fig5ac_RGB_Fe_Cu_Pt.dm3 * Fig5ac_RGB_Fe_Cu_Pt.jpg * Fig5ac_Profiles.dm3 Processed data using the above elemental maps. The line profiles are generated by binning the signals along the direction parallel to the interface. All line profiles were normalised to their maximum signal level for ease of comparison. The RGB file is generated by overlaying false-coloured maps of the distributions of Pt (EELS signal, blue), Cu (EELS signal, green) and Y (EELS signal, red).