Search

The energy loss function (ELF) is a key parameter for the calculations of energy losses undergone by electrons in matter. It is often the only input in the calculations performed within the models based on the semi-classical dielectric response theory. Its perfect knowledge is thus of primordial importance. To evaluate the ELF, it is usual to consider as a model an expansion in Drude-Lindhard (DL) type oscillators with fixed values of the strength, width, energy and dispersion for the various oscillators. However, for materials that are characterized by a single sharp oscillator as aluminium or indium, it has been shown [Phys. Rev. B 46 (1992) 2486] that the damping parameter that corresponds in the DL model to the width of the oscillator increases for decreasing incident electron energy. To emphasize this effect, we perform in this work systematic calculation of the ELF for an indium target and for incident electron energies between 200 and 2000-eV. The ELF is determined by comparing REELS (reflection electron energy loss spectroscopy) experimental inelastic electron scattering cross sections with cross sections calculated within the semi-classical dielectric response model, which is implemented in the QUEELS-e(k,ω)-REELS software (Quantitative analysis of Electron Energy Losses at Surfaces) [Surf. Interface Anal. 36 (2004) 824]. We also perform measurements and calculations for varying incident and exit angles of the electron, namely for angles to the surface normal between 15o and 75o, to check the validity of our results for all geometries. Our results show that the damping parameter of indium in the DL model for electron energy of 200-eV is three times larger than for energy of 2000-eV. Copyright © 2014 John Wiley & Sons, Ltd.

Numerous data on the inelastic mean free path (IMFP) of electrons have been published, mainly for elemental solids and some inorganic and organic compounds. However, there is limited information available for IMFP values of binary alloy systems. Although IMFP values for different energies can be calculated for alloys from predictive formulae, IMFPs can be measured experimentally by elastic peak electron spectroscopy (EPES). In the present work, the IMFPs for three Au-Cu alloys-Au (25 at.%)-Cu (75 at.%), Au (50 at.%)-Cu (50 at.%), Au (75 at.%)-Cu (25 at.%)-were determined for the first time. The EPES experiments have been performed in three laboratories using spectrometers with different geometries and energy resolutions. The experiments were performed with the use of an Au standard (relative measurements) and without a standard material (absolute measurements). Two Monte Carlo algorithms were used to calculate the IMFPs. The IMFP energy dependence for the alloy samples has been obtained in the primary electron energy range 200-2000 eV. The measured IMFPs were compared with IMFPs calculated from two predictive formulae, i.e. the TPP-2M equation of Tanuma et al. and the G-1 equation of Gries. The scatter between the measured IMFPs and those calculated from the predictive formulae has been analysed statistically. The root-mean-square deviation from IMFP values calculated from the TPP-2M equation was 1.81-5.97 A, depending on the alloy surface composition and the particular EPES measurement, and it was smallest for the Au 25 Cu 75 alloy. The mean percentage deviation from the TPP-2M IMFPs was ∼27%.

Intensity ratios of electrons backscattered elastically from thick, amorphous Si, polycrystalline Ag and Au, as well as various (vacuum evaporated, electrodeposited layers and metallic sheets) polycrystalline Ni samples, were measured using high energy resolution and two different experimental geometries (0° and 50° primary and 50° and 0° scattered beam angles relative to the surface normal, respectively) in the range of 1-5 keV primary electron beam energy. The elastic yield ratios obtained experimentally are compared to the results computed by Monte Carlo simulations using different algorithms and different sets of parameters. In the case of measured elastic yield ratios Si/Ni, Ag/Ni and Au/Ni, systematic differences were observed between the respective values obtained for two experimental geometries, whereas each Monte Carlo calculation predicted similar ratios for the different geometries. For the Si/Ag, Si/Au and Ag/Au elastic yield ratios, however, good agreement is observed between experiment and calculation, resulting in similar ratios for each geometry. As a conclusion to our study, we suggest that when using a polycrystalline Ni standard reference sample for measuring elastic yield ratios the homogeneity of the sample should be ensured. Our results confirm the applicability of elastic peak electron spectroscopy in the case of these solids for deriving IMFP values, i.e. no significant role of surface effects on the elastic yield ratios can be identified in the primary beam energy range studied.

Elastic peak electron spectroscopy (EPES) is based on measurements of elastic electron reflection coefficients. The angular distribution in EPES depends on different parameters such as atomic number of the substrate, the energy of the electron beam and the geometry of the analysis system. In this study, we compare the measured angular distribution of electrons elastically reflected from Pd and In surfaces with Monte-Carlo simulations. We have used a theoretical approach based on knowledge of electron trajectories and interaction effects inside the samples. For the experiments, we have constructed a special rotary energy analyser with a small aperture. The angular distributions of elastically backscattered electrons at 500 eV resulting from Monte-Carlo simulations of electron transport for Pd and In (atomic numbers 46 and 49) compare very well with the corresponding experimental measurements.

The usefulness of elastic peak electron spectroscopy (EPES) has been demonstrated by application to a semiconductor study of indium phosphide or InAsP. In this paper, we present some results obtained with EPES. The inelastic mean free paths of indium phosphide have been calculated using EPES and the Monte-Carlo simulation or electron trajectories in solids. The second part of this work deals with the determination of the concentration of elements in heterostructures such as Au/InSb/InP(100) and InAs x P 1-x , and the results are compared to those obtained using AES.

Alumina was evaporated on InP(100) and InSb/InP(100) substrates. The InSb buffer layer permits thermal and structural protection of the indium phosphide substrate. This layer was obtained by condensation of antimony atoms on the indium-rich InP surface after ionic etching. The evaporation of antimony leads to Sb-In bonds. Modellization of the Auger signal variations during alumina condensation was compared to the experimental results. Auger measurements indicate the improvement of the elaborated structure. This fact has been confirmed by the electrical C(V) characterizations.

Efficient visible luminescence from porous Si (PSi) layers is associated with the quantum size and the presence of hydrogen as a passivant on the inner surface. The increase of the amount of H was performed by wet chemical etching in HF, producing the increase of photoluminescence intensity. The detection of H is not possible by AES or XPS and is difficult even with EELS. Elastic peak electron spectroscopy (EPES) showed a large decrease of the elastic reflection coefficient reappearing in the elastic peak, and measured with a retarding field analyser (RFA). The effective elastic backscattering coefficients of a monolayer of H, Si, O were calculated for the RFA angular window and resulted in reasonable agreement with literature data available

The mechanisms for the formation of the systems 〈Au/InP(100)〉 and 〈Au/InSb/InP(100)〉 are connected with their electrical properties. By using quantitative AES, a precise knowledge of the atomic mechanisms allows us to perform the desired component such as a Schottky diode. The best results are obtained for the 〈Au/heated InSb/InP〉 system with a barrier height of 0.7 eV.