(V)EELS analysis for mapping properties in semiconductor materials.

Recent technological advances in transmission electron microscopy (TEM) are allowing us to obtain more and more information about semiconductor materials analyzing the electron energy loss spectra (EELS) in the low-loss regime. This revolution is mainly driven by the generalized implantation of  aberration corrector and monochromator in the instruments that have (respectively) enhanced the spatial and energetic resolutions of the instruments below the (once again, respectively) nanometer and electronvolt barriers. These are very relevant properties for typical semiconductor structures like ternary compond-based (AlGaN, InAlN … ) structures or silicon-based structures for optoelectronics or photonics applications.

Ideally, low-loss EELS of a semiconductor material at high resolution will give information on the valence band properties of the materials – hence the “V” in (V)EELS – allowing us to measure band gap energies, intraband transitions, plasmon excitation peaks and more. Once the (V)EELS data is acquired, our aim is to analyze it to bring out the great amount information enclosed in it. For this purpose, we work on developing useful software, testing it in real-case scenarios. Using various specialized techniques we have been able to uncover a wide range of chemical and structural properties of the semiconductor materials, e.g. simultaneously measuring band gap energies and composition of a ternary compound semiconductor (InxAl1-xN), obtaining the complex dielectric function in the examined energetic range through Kramers-Kronig analysis, and getting clues to uncover the polytypic nature of a using structure-specific intraband transitions [1]. All these from a single experiment, and a relatively small amount of data acquired, in the form of sub-nm resolved spectrum lines.

Given the sometimes complex nature of the examined structures there are times when a small amount of data will not suffice to uncover all the demanded information. This was the case when we examined some anomalous segregations that were typically 2D defects in AlxGa1-xN structures [2]. Sub-nm resolved spectrum images were acquired, covering the regions of interest. These represent a vast amount of data, that could be analyzed in a smart way thanks to the computer tools developed. Finally, the properties that were calculated from the spectrum images allowed to look the examined defects in new angles. Please examine Figs. 1 and 2 for two exemplary one and two (respectively) dimensional analyses.

(a) STEM-HAADF image of the full structure, from the GaN buffer layer at right hand side to the top of the DBR. Lower panel (c) shows a detail of two successive periods (susbtrate to the upper side). Top graph, (b), shows the aluminum ratio profiles (circles) calculated through Vegard Law analysis of the plasmon excitation energy position along with the HAADF intensity profile (blue). Below, (d) shows experimental and HAADF simulated intensity profiles for 3 periods.




Universitat de Barcelona