Analysis of reflection electron energy loss spectra (REELS) for determination of the dielectric function of solids: Fe, Co, Ni

A simple procedure is developed to simultaneously eliminate multiple scattering contributions from two reflection electron energy-loss spectra (REELS) measured at different energies or for different experimental geometrical configurations. The procedure provides the differential inverse inelastic mean free path (DIIMFP) and the differential surface excitation probability (DSEP). The only required input parameters are the differential cross section for elastic scattering and a reasonable estimate for the inelastic mean free path (IMFP). No prior information on surface excitations is required for the deconvolution. The retrieved DIIMFP and DSEP can be used to determine the dielectric function of a solid by fitting the DSEP and DIIMFP to theory. Eventually, the optical data can be used to calculate the (differential and total) inelastic mean free path and the surface excitation probability. The procedure is applied to Fe, Co and Ni and the retrieved optical data as well as the inelastic mean free paths and surface excitation parameters derived from it are compared to values reported earlier in the literature. In all cases, reasonable agreement is found between the present data and the earlier results, supporting the validity of the procedure.     

Simple algorithm for quantitative analysis of reflection electron energy loss spectra (REELS)

A detailed comparison of two different physical approaches for quantitative analysis of reflection electron energy loss spectra (REELS) is presented. The Tougaard–Chorkendorff (TC) algorithm [S. Tougaard, I. Chorkendorff, Phys. Rev. B35 (1987) 6570] is analyzed theoretically and applied to experimental spectra of four elemental solids (Si, Cu, Ag, and Au). A closed expression is derived for the quantity retrieved by the TC-algorithm, the so-called “effective” cross section, which was originally only given as a recursive procedure. Single scattering loss distributions are derived from the experimental spectra using the bivariate reversal method [W.S.M. Werner, Phys. Rev. B74 (2006) 075421]. The latter agree satisfactorily with density functional theory (DFT) calculations and other data found in the literature. Using these single scattering loss distributions, the TC “effective” cross section can be perfectly reproduced if the fact is taken into account that the effective cross section is not a single scattering loss distribution and is governed to a significant extent by elastic scattering. On the basis of the above results, a dramatically simplified deconvolution scheme for quantitative analysis of REELS is developed.     

Optical properties and electronic transitions of SnO2 thin films by reflection electron energy loss spectroscopy

Optical properties of SnO₂ thin films in the 4–60 eV energy range are determined by reflection electron energy loss spectroscopy. Bulk and surface electron loss functions, real and imaginary parts of the dielectric function, refraction index, extinction and absorption coefficients are obtained from the analysis of the electron energy loss spectra. Electronic transitions are identified through the interpretation of the optical data. The samples (250–500 nm thick) were produced by ion beam-induced chemical vapor deposition. It is found that the compacity of the SnO₂ thin films affects their optical properties and therefore the relative intensity of the observed electronic transitions. The advantages of this method to determine optical properties of thin films are discussed. Inelastic mean free paths (6.2, 17 and 41 Å for electrons traveling in SnO₂ with kinetic energies of 300, 800 and 2000 eV, respectively) are obtained from the corresponding inelastic electron scattering cross-sections.     

Determination of the inelastic mean free paths (IMFPs) in Ti by elastic peak electron spectroscopy (EPES): Effect of impurities and surface excitations

Surface excitations are important in surface sensitive electron spectroscopes, especially in elastic peak electron spectroscopy (EPES) since they may distort quantitative information. This phenomenon is more pronounced at low electron energy and glancing emission angles and should be appropriately corrected.In the present work we investigate quantitatively the role of contaminations, density and surface excitations on electron inelastic mean free paths (IMFPs) in Ti determined by elastic peak electron spectroscopy (EPES) using Cu standard. In the Monte Carlo algorithm the new NIST 3.1 database of electron elastic scattering cross sections was applied. It has been also shown that accounting for surface excitations, as well as for appropriate input parameters (surface composition, density, hydrogen) in the EPES method, is important for accuracy of evaluated IMFPs. Due to high reactivity of Ti, the IMFPs for contaminated Ti may be of interest. The authors indicate the magnitude of various corrections on the IMFPs derived by EPES.     

Metal interface formation studied by high-energy reflection energy loss spectroscopy and electron Rutherford backscattering

We demonstrate that high-energy, high-resolution reflection electron energy loss spectroscopy can provide unique insights into interface formation, especially for the case where an extended interface is formed. By changing the geometry and/or electron energy the electronic structure can be probed over a range of thicknesses (from 10s of Å to more than 1000 Å). At the same time one resolves the elastically scattered electrons into different components, corresponding to scattering of atoms with different mass (so-called ‘electron Rutherford backscattering’). Thus these high-energy REELS/elastic scattering experiments obtain information on both the electronic structure and the atomic composition of the overlayer formed.     

Splitting the plasmon peak in high-energy reflection electron energy loss experiments

Elastic scattering of energetic electrons over large angles (in this study 40 keV and 120°$120°$) implies momentum and hence energy transfer from an electron to a nucleus. Due to the large mass of the nucleus (relative to the mass of an electron) this energy transfer is small, but it has recently been shown that it can be resolved in a modern spectrometer. Hence the elastic peak of an overlayer/substrate system splits into different components corresponding to atoms with different mass. Here we extend this type of experiment to the plasmon part of a reflection energy loss spectroscopy (REELS) spectrum. It is shown that, for suitable systems, the plasmon peak of an overlayer/substrate system is split by the same amount as the elastic peak. This is a consequence of the fact that detection of an electron in REELS always requires a large-angle elastic scattering event. Moreover, we show that the relative intensity of the plasmon components contains information on the depth distribution of the scatterers.     

Frenkel exciton model of electron energy loss spectroscopy in -PTCDA

Recent experimental findings concerning electron energy loss spectroscopy in -Perylene-tetracarboxylic-dianhydride are analysed in terms of a Frenkel exciton model. Taking into account the energy dispersion of excitations with finite momentum transfer, the k-dependence of the dielectric tensor and the corresponding electron energy loss functions can be calculated. The exciton dispersion with a minimum at k≠0 yields a red shift of the lineshape of loss functions at large k, as observed experimentally.     

Background subtraction in electron spectroscopy by use of reflection electron energy loss spectra

The use of reflected electron energy loss spectra (REELS) in deconvoluting the inelastic background signal from XPS and AES spectra from homogeneous samples is studied. It is demonstrated that under certain assumptions, the cross section for inelastic electron scattering can be extracted from a REELS spectrum. This cross section is applied to deconvolute an experimental XPS spectrum of aluminium. The method, its limitations and its relation to other methods are discussed.     

A comparative analysis of electron spectroscopy and first-principles studies on Cu(Pd) adsorption on MgO

Ultrathin MgO films were grown on a W(1 1 0) substrate while metastable impact electron (MIES) and photoelectron (UPS) spectra were measured in situ; apart from the valence band emission, no additional spectral features were detected. The oxide surface was exposed to metal atoms (Cu, Pd) at RT. A comparison with the DOS extracted from first-principles DFT calculations shows that the metal-induced intensity developing above the top of the O 2p valence band in the UP spectra under Cu(Pd) exposure is caused by Cu 3d (Pd 4d) emission. The emission seen in the MIES spectra is attributed to the ionization of Cu 3d and 4s states of adsorbed neutral Cu atoms in an Auger process, Auger neutralization, involving two electrons from the surface, at least one of them from the metal adsorbate. The shape of the MIES spectra suggests metallic island growth even at the lowest studied exposures, which is supported by the first-principles calculations.     

Very low energy electron reflection at Cu(001) surfaces

Measurements of the coefficient of elastic reflection of very low energy electrons at Cu(001), Cu(001) (2 × 4)45°O and Cu(001)c(2 × 2)N surfaces are reported. The measurements refer to normally-incident electrons with kinetic energies E in the range 0.5–22 eV. The elastic reflection coefficient R_el was determined from separate observations of the total reflection coefficient and of the energy distribution of reflected electrons. For Cu(001) R_el is 0.55 at E = 0.5 eV and drops monotonically to 0.03 with increasing E, the maximum slope being at E = 3 eV. Theoretical calculations of R_el are reported. The reflection amplitude of the substrate crystal was parameterized using existing results of accurate band structure calculations, and the surface scattering matrix was evaluated for assumed surface scattering potentials. It is shown that to fit the observed R_el it is necessary to take account of both the image potential and the extension of the imaginary part of the crystal scattering potential into vacuum. From the fit, the range of the imaginary potential is 1.0 Å. For Cu(001) (2 × 4)45°O and Cu(001)c(2 × 2)N the values of R_el at E = 0.5 eV were 0.35 and 0.15, respectively. The effect of adsorption in reducing R_el is especially marked for E < 2 eV. Adsorption of either O or N results in an additional peak in R_el near E = 12 eV.     

Surface and bulk plasmon coupling observed in reflection electron energy loss spectra

Reflection electron energy loss spectra have been measured for a semiconductor and some metals (Si, Cu, Ag and Au). A novel procedure is presented to rigorously decompose the spectra into contributions corresponding to surface and volume excitations. The resulting distributions of energy losses in an individual surface loss are in good agreement with theory. In particular, the begrenzungs effect occurring at the boundary of a solid state plasma, i.e. the reduction of the intensity of bulk modes due to the coupling with surface modes, can be clearly observed in the retrieved energy loss distribution.     

Local thickness measurement through scattering contrast and electron energy-loss spectroscopy

Scattering contrast measurements were performed on thin films of amorphous carbon and polycrystalline Au, as well as single-crystal MgO nanocubes. Based on the exponential absorption law, mass-thickness can be obtained within 10% accuracy by measuring the incident and transmitted intensities in the same image. For mass-thickness measurement of a thin amorphous specimen, a small collection semiangle improves the measurement sensitivity, whereas for the measurement of polycrystalline or single-crystal specimens, a large collection semiangle should be used to reduce diffraction-contrast effects. EELS thickness measurements on MgO nanocubes suggest that the Kramers-Kronig sum-rule method (with correction for plural and surface scattering) gives 10% accuracy at medium collection semiangles but overestimates the thickness at small collection semiangles, due to underestimation of the surface-mode scattering. The log-ratio method, with a formula for inelastic mean free path proposed by Malis et al. (1988), provides 10% accuracy at small collection semiangle, while that proposed by Iakoubovskii et al. (2008a) is preferable for medium and large collection semiangles. As a result of this work, we provide recommendations of preferred methods and conditions for local-thickness measurement in the TEM.     

Comparison of cross sections in high resolution electron energy loss spectroscopy and infrared reflection spectroscopy

Electron energy losses and absorption of infrared radiation are both caused by the dipole moment of the surface vibration. A comparison of absolute intensities between both techniques should therefore be possible. In this paper the appropriate formulas are derived. For the adsorption system CO on Pt(111) which has been investigated by both techniques a perfect agreement is found. For a number of adsorbate systems the effective ionic charge is calculated from previously published electron energy loss data.     

Angular and energy dependences of the surface excitation parameter for semiconducting III-V compounds

Sum-rule-constrained extended Drude dielectric functions were used to study surface excitations generated by energetic electrons moving across surfaces of semiconducting III-V compounds. Parameters in the dielectric functions were determined from fits to experimental optical data and electron energy-loss spectra. Electron inelastic mean free paths (IMFPs) in GaN, GaP, GaAs, GaSb, InAs and InSb were calculated for electron energies between 200 and 2000 eV, and the results were found to follow the simple formula, i.e., λ = kE^p, where λ is the IMFP and E is the electron energy. Surface excitation parameters (SEPs), which describe the total probability of surface excitations by electrons crossing the surface and travelling in vacuum, were also calculated for different electron energies and crossing angles. The SEP was found to follow the simple formula, i.e., , where P_s is the SEP and α is the crossing angle relative to the surface normal.     

Auger electron spectroscopy and electron energy loss spectroscopy studies on carbonization of Si(100) and (111) surfaces with ethylene

The reactions of Si(100) and Si(111) surfaces at 700 °C (973 K) with ethylene (C₂H₄) at a pressure of 1.3×10⁻⁴ Pa for various periods of time were studied by using Auger electron spectroscopy (AES) and electron energy loss spectroscopy (ELS). For a C₂H₄ exposure level, the amount of C on the (111) surface was larger than that on the (100) surface. The formation of β-SiC grain was deduced by comparing the C_KLL spectra from the sample subjected to various C₂H₄ exposure levels, and from β-SiC crystal.     

Verification of CuInSe2 density of states features by reflection electron energy loss spectroscopy

Reflection electron energy loss measurements of CuInSe₂ single crystals are reported and when possible, the losses are compared with former results by Kazmerski and Jamjoum et al. Their identification of the transitions is confirmed and refined by spectrum simulations with total and partial densities of states from band structure calculations of Jaffe, Bendt, Zunger, and Bullett.New conduction-band maxima at energies not covered by the available calculations, are proposed and their consistency with the experimental loss energies is shown by another simulation with Lorentzian density peaks.     

The adsorption of CO on Cu surfaces studied by electron energy loss spectroscopy

Electron energy loss spectroscopy (ELS) in the energy range of electronic transitions (primary energy 30 < E₀ < 50 eV, resolution ΔE ≈ 0.3 eV) has been used to study the adsorption of CO on polycrystalline surfaces and on the low index faces (100), (110), (111) of Cu at 80 K. Also LEED patterns were investigated and thermal desorption was analyzed by means of the temperature dependence of three losses near 9, 12 and 14 eV characteristic for adsorbed CO. The 12 and 14 eV losses occur on all Cu surfaces in the whole coverage range; they are interpreted in terms of intramolecular transitions of the CO. The 9 eV loss is sensitive to the crystallographic type of Cu surface and to the coverage with CO. The interpretation in terms of d(Cu) → 2π¹(CO) charge transfer transitions allows conclusions concerning the adsorption site geometry. The ELS results are consistent with information obtained from LEED. On the (100) surface CO adsorption enhances the intensity of a bulk electronic transition near 4 eV at E₀ < 50 eV. This effect is interpreted within the framework of dielectric theory for surface scattering on the basis of the Cu electron energy band scheme.     

Adsorption and decomposition of HNCO on Cu(111) surface studied by auger electron,electron energy loss and thermal desorption spectroscopy

No detectable adsorbed species were observed after exposure of HNCO to a clean Cu(111) surface at 300 K. The presence of adsorbed oxygen, however, exerted a dramatic influence on the adsorptive properties of this surface and caused the dissociative adsorption of HNCO with concomitant release of water. The adsorption of HNCO at 300 K produced two new strong losses at 10.4 and 13.5 eV in electron energy loss spectra, which were not observed during the adsorption of either CO or atomic N. These loses can be attributed to surface NCO on Cu(111). The surface isocyanate was stable up to 400 K. The decomposition in the adsorbed phase began with the evolution of CO₂. The desorption of nitrogen started at 700 K. Above 800 K, the formation of C₂N₂ was observed. The characteristics of the CO₂ formation and the ratios of the products sensitively depended on the amount of preadsorbed oxygen. No HNCO was desorbed as such, and neither NCO nor (NCO)₂ were detected during the desorption. From the comparison of adsorption and desorption behaviours of HNCO, N, CO and CO₂ on copper surfaces it was concluded that NCO exists as such on a Cu(111) surface at 300 K. The interaction of HNCO with oxygen covered Cu(111) surface and the reactions of surface NCO with adsorbed oxygen are discussed in detail.     

Hydrogen chemisorption on Ni(110) by high-resolution electron energy loss spectroscopy

High-resolution electron energy loss spectra of hydrogen-covered Ni(110) surfaces both at 100 and 300 K are presented. The adsorbed sites of hydrogen atoms are discussed.High-resolution electron energy loss spectra of hydrogen covered Ni(110) surfaces have been studied. Tentative models for the adsorbed sites of hydrogen atoms are as follows: (1) For the (2 × 1)-H surface, hydrogen is adsorbed in the three-coordinated sites of the rudimentary (111) face of the unreconstructed Ni(110) substrate. (2) For the low-temperature (1 × 2)-H surface, hydrogen is adsorbed in the three-coordinated sites and, probably, in the two-fold hollow sites of the distorted Ni(110) substrate. (3) For the room-temperature (1 × 2)-H surface, hydrogen is disorderedly adsorbed in the three-coordinated, two-fold hollow and short-bridge sites and, possibly, in the octahedral sites of the distorted Ni(110) substrate. Some of the unresolved problems in the above assignments are summarized: (1) Strictly, the three-coordinated sites above are somewhat different from those discussed in the molecular-beam diffraction study [5]. (2) For the low-temperature (1 × 2)-H surface, the loss associated with hydrogen in the two-fold hollow sites is apparently not observed. (3) Intensity changes of the three losses for the room-temperature (1 × 2)-H surface with increasing hydrogen pressure (Fig. 2) are not well understood.     

Spin-polarized electron energy loss spectroscopy of high energy, large wave vector spin waves in ultrathin fcc Co films on Cu(001)

The realm of high energy, large wave vector spin waves in ultrathin films and at surfaces is unexplored because a suitable method was not available up to now. We present experimental data for an 8 ML thick Co film deposited on Cu(001) which show that spin-polarized electron energy loss spectroscopy can be used to measure spin-wave dispersion curves of ultrathin ferromagnetic films up to the surface Brillouin zone boundary.