Selective Excitation of a Symmetric Interference Plasmon Mode in Two Close Planar SiO₂/Si Interfaces Observed by Electron Energy-Loss Spectroscopy

Abstract

We have studied the effects of the SiO2/Si interface parallel to an electron beam on transmission electron energy-loss spectra of a SiO2 area for poly-Si/SiO2/Si samples. The dependence of the energy-loss spectra on the distance from the interface to the probe position and on the distance between two interfaces was investigated. Spectra obtained from the center of the thick (150 nm) SiO2 layer had no peak in the energy region of 4–10 eV. However, a peak at about 7 eV was observed in the spectra obtained at a position of 7.5 nm from the SiO2/Si interface for the same specimen. This peak was assigned to a SiO2/Si interface plasmon excitation. For the thin (15 nm) oxide poly-Si/SiO2/Si sample, on the other hand, the peak appeared at 8.7 eV. Furthermore, this peak shifts to higher energies as the oxide thickness decreases. This peak was assigned to an excitation of a symmetric interference plasmon mode in two close planar SiO2/Si interfaces.

Competition between bulk and interface plasmonic modes in valence electron energy-loss spectroscopy of ultrathinSiO2gate stacks

“…9 Because they are less easily interpretable, low-loss spectra have not been used as extensively to probe gate stacks. Previous studies of Si/ SiO 2 interfaces have identified a strong spectral feature around 8 eV ͑below the SiO 2 band gap en-ergy͒ originating from an interface plasmon, [10][11][12][13][14] and have also reported changes at higher energy loss. 15,16 Delocalized signals, originating from the excitation of modes located relatively far from the electron trajectory, contribute to valence EELS.…”

Fermi level pinning characterisation on ammonium fluoride-treated surfaces of silicon by energy-filtered doping contrast in the scanning electron microscope

“…Furthermore, the removal of the multiple kinks at relatively high kinetic energies after NH 4 F-treatment (see Fig. 2(b) ) is striking, which a posteriori indicates that, far from being correctly imputed to tertiary SE3 signals originating from the pole piece or objects other than the sample as claimed in refs 29 and 30 , the former are most likely associated with plasmon contributions from: oxides which form rapidly on the surface 31 ; or the Si-SiO 2 interface and/or oxide surface 32 ; or beam-induced contamination on the sample. The disappearance of these effects may correspond to the dissolution of the native oxide layer or any other organic contamination, followed by passivation of the surface.…”