Gast ::
| Hauptseite > Publikationsdatenbank > Spectrum imaging of Helium pores in amophous Silicon-coatings |
| Hauptseite > Publikationsdatenbank > Spectrum imaging of Helium pores in amophous Silicon-coatings |
| Conference Presentation (Other) | FZJ-2014-04425 |
; ; ; ;
2014
Abstract: In order to probe the helium distribution in porous amorphous coatings of silicon grown by magnetron, we present an extraction method of the Helium signal obtained from STEM-EELS spectrum images [1]. The goal of the work is to get a rough estimation of the Helium pressure inside the pores and correlate this to the deposition parameters. For this we modified the procedure described by Walsh [2] and David et al. [3] and integrated this in MATLAB. With our present architecture it is possible to read in images in dm3 format recorded on a with DigiScan by Gatan and undergo several data treatment. For our purpose we selected centering the zero loss peak and integrating it, deconvolution, fitting of the plasmon intensity with one narrow peak at ≈ 23 eV attributed to the Silicon bulk plasmon and a wider one at ≈ 24 to 25 eV to adapt contributions from surface oxide layer and carbon contamination, and fitting of the residual intensity arsing from the He-K edge at ≈ 22 eV with a gaussian. Part of the procedure is visualized in Figure 2, which shows two spectra from the same spectrum image one at the matrix position (Figure 2 (a)) and the other at the pore center (Figure 2 (b)). The spectra were are already deconvoluted and the fit to the Silicon plasmon is plotted red and the fit for the SiO2 and the carbon contamination is plotted green. For both positions the fit is satisfactory and for the pore position and also the residual signal around 22 eV is well described by the gaussian fit.The procedure allows to plot maps of all fitting parameters and also to extract EFTEM images. Figure 1 shows (a) thickness map (b) the gaussian integral and (c) the gaussian peak position for a selected spectrum image of a single pore. The Helium density can be derived in two ways, from the ratio of the Helium K-edge intensity/ to the ZLP-intensity and by the energy shift of the edge position. Both methods suffer from large errors around 30 % so the cross check is an advantage. We will complete our results by additional tomography experiments and correlate them to compositional depth profiles measured with Rutherford backscattering.
|
The record appears in these collections: |
