Elastic Peak Electron Spectroscopy Research Articles

Corrected electron inelastic mean free paths (IMFPs) for selected wide band semiconductors

Elastic peak electron spectroscopy (EPES) has been widely used to determine the electron inelastic mean free paths (IMFPs) in solids. In this work, we investigated quantitatively the influence of surface excitations on electron IMFPs determined by EPES. We used IMFPs obtained from the early EPES measurements of the electron elastic backscattering probability from GaN and Cd0.88Mn0.12 Te wideband-gap semiconductors, and the Ni standard in the energy range 200–2000 eV. The total surface-excitation parameter (SEP) was evaluated using Chen and Werner approaches, and was applied for correcting the EPES IMFPs. These corrected values were then compared with those predicted by the TPP-2M formula. We found that implementation of the surface-excitation correction improved agreement between the resulting IMFPs for selected wide band semiconductors and the TPP-2M values at low-energy (E > 500 eV) electrons. The extent to which the IMFPs measured by EPES differ from the corresponding bulk values (on account of surface excitations) was found to depend on the semiconductor material with finite surface. Our results also clearly demonstrated the importance of accounting for surface excitations for accuracy of the IMFPs measured for GaN.

Quantification of surface excitation effects on the EPES‐determined IMFPs for GaN and SiC

AbstractSurface electron inelastic excitations, a consequence of electron surface interaction, effect substantially the measured peak intensities, or peak areas, in surface‐sensitive electron spectroscopies, and distort the quantitative information. To obtain reliable quantitative information from the electron spectra, the measured intensities and/or peak areas should be corrected for these effects.In the present work, we investigated, quantitatively, the influence of surface excitations on electron inelastic mean free paths (IMFPs) determined by elastic peak electron spectroscopy (EPES). We used IMFPs obtained from the previous EPES measurements of the electron elastic backscattering probability for GaN, SiC, and the Ni standard in the energy range 200–2000 eV. Calculated surface excitation parameters (SEPs) of both incident and emitted electrons were evaluated using the formulae of Chen, Tu et al., and Werner et al., and applied for correcting the EPES IMFPs. These corrected values were compared with those predicted by the TPP‐2M formula.We found that implementation of the surface excitation correction improved agreement between the resulting IMFPs for selected wide‐band semiconductors and the TPP‐2M values. The extent to which the IMFPs measured by EPES differ from the corresponding bulk values (on account of surface excitations) was found to depend on the semiconductor material with finite surface. The results also clearly demonstrated the importance of accounting for surface excitations for accuracy of the measured IMFPs. Copyright © 2008 John Wiley & Sons, Ltd.

Growth and crystalline structure of Co layers on Cu(0 0 1)

The crystalline structure of Co layers deposited on the Cu(0 0 1) surface was investigated with the use of the directional elastic peak electron spectroscopy (DEPES). For clean Cu(0 0 1) the experimental DEPES profiles obtained for different energies of the primary electron beam exhibit intensity maxima corresponding to the close packed rows of atoms. The Auger peak kinetics recorded during continuous Co deposition suggest the layer-by-layer growth mode. The DEPES profiles recorded for 10 monolayers (ML) of Co on Cu(0 0 1) reflect a short-range order in the adsorbate. Intensity maxima observed in the DEPES profiles for Co along [1 0 0], [0 1 0], and [1 1 0] azimuths of Cu(0 0 1) are characteristic of the face centered cubic (fcc) Co(0 0 1) layers. Low-intensity reflections and considerable background intensities were found in the low energy electron diffraction (LEED) patterns recorded from 10 ML of Co, which indicates a weak long-range order in the adsorbate. The adsorption of about 20 ML of Co results in considerable background contribution to DEPES. No reflections but a large background were observed with the use of LEED for this layer. The heating of the Co/Cu(0 0 1) system at T = 770 K leads to an increase of the short- and long-range order in the overlayer, observed in the DEPES profiles and LEED patterns, respectively. The theoretical DEPES profiles were obtained with the use of a multiple scattering approximation. A very good agreement between experimental and theoretical scans was found for the clean and covered copper substrate. The latter proves the epitaxial growth of Co layers on Cu(0 0 1).

Electron Backscattering in Sputter Depth Profiling Using AES, EPES, and REELS

The main applications of electron backscattering in sputter depth profiling using AES and related elec- tron spectroscopies such as elastic peak electron spectroscopy (EPES) and reflection electron energy loss spectroscopy (REELS) are briefly summarized. EPES depth profiling is particularly useful for improving the interfacial depth resolution in binary systems by selecting a suitable primary electron energy. With REELS depth profiling, the change of the IMFP during sputtering through an A/B interface can be quantita- tively determined. Recently, a method for implementing the backscattering effect in the MRI model has been introduced and applied to AES depth profiles of single layer and multilayer structures. Its capability to provide quantification of measured profiles including the backscattering effect is outlined.

Determination of Surface-Excitation Parameters for Elastic Peak Electron Spectroscopy (EPES) Using the Database of Goto

The inelastic mean free path (IMFP) can be determined from the elastic-backscattering probability Ie of electrons that can be measured by elastic peak electron spectroscopy (EPES). We calculated IMFPs from the absolute Ie measurements of Goto. Calculated values of the elastic backscattering-probability Iec can be obtained from the EPESWIN software of Jablonski. Iec denotes the number of electrons/incident electron, backscattered elastically and detected with the cylindrical mirror analyzer (CMA) of Goto. The Ie data deduced from the database of Goto are lower by 20-34% than the calculated Iec for Si, Ni, Cu, Ag and Au. This discrepancy is explained by surface-excitation losses characterized by the SEP parameter. We made corrections for the experimental Ie data for surface excitations using the material parameters of Chen (modified), Kwei, Ding, and Werner. The parameters of Chen were modified for achieving the minimum deviations between the calculated of Iec and the SEP corrected fsIe values. The material parameter of Nagatomi ach = 4.3 was confirmed for Ni by EPES. The SEP corrected IMFPs were deduced from the SEP corrected fsIe data. SEP correction of the IMFPs resulted in mean deviations from the calculated TPP-2M data of 3.9 % for Si, 6-7% for Ni, 8.2% for Ag and nearly 12 % for Cu and for Au.

Morphology, surface roughness, electron inelastic and quasielastic scattering in elastic peak electron spectroscopy of polymers

AbstractIn elastic peak electron spectroscopy (EPES), the nearest vicinity of elastic peak in the low kinetic energy region reflects electron inelastic and quasielastic processes. Incident electrons produce surface excitations, inducing surface plasmons, with the corresponding loss peaks separated by 1–20 eV energy from the elastic peak.In this work, X‐ray photoelectron spectroscopy (XPS) and helium pycnometry are applied for determining surface atomic composition and bulk density, whereas atomic force microscopy (AFM) is applied for determining surface morphology and roughness. The component due to electron recoil on hydrogen atoms can be observed in EPES spectra for selected primary electron energies. Simulations of EPES predict a larger contribution of the hydrogen component than observed experimentally, where hydrogen deficiency is observed. Elastic peak intensity is influenced more strongly by surface morphology (roughness and porosity) than by surface excitations and quasielastic scattering of electrons by hydrogen atoms. Copyright © 2007 John Wiley & Sons, Ltd.

Monte Carlo backscattering yield (BY) calculations applying continuous slowing down approximation (CSDA) and experimental data

The backscattering yield (BY) is an important factor in elastic peak electron spectroscopy. Monte Carlo (MC) algorithm describing the electron elastic and inelastic backscattering from surfaces is presented applying the continuous slowing down approximation, i.e., an assumption that an electron loses energy continuously along the trajectory length. Energy losses calculations were performed using recently published stopping power (SP) functions, calculated from the experimental optical data by Tanuma et al., and the experimental SP functions published elsewhere. In the MC algorithm, calculating the values of BY, the input parameters are the elastic scattering cross sections and the SP function. The MC calculations were performed at selected energies ranging from 200 to 30 000 eV. Ten elemental solids were considered: Al, Si, Cr, Ni, Cu, Ge, Pd, Ag, Pt, and Au. All MC simulations were performed for normal incidence of the beam on a sample and the backscattered electron emission into negative hemisphere. Results of BY calculations were compared to available experimental data taken from Joy database of electron–solid interactions. In most cases, the results deviate less than 10% from the available experimental data.

Angle-resolved elastic peak electron spectroscopy: Role of surface excitations

We analyze the possibility of determining the surface excitation parameter (SEP) from the dependence of the elastic backscattering signal intensity on the emission angle. It has been found that the shape of this dependence is reasonably well described by the theoretical model implemented in a typical Monte Carlo simulation strategy. As shown recently, the mean percentage deviation between the experimental angular dependence and the theoretical dependence is equal to 8.82% at 200 eV, 6.28% at 500 eV and 4.69% at 1000 eV. In the theoretical model used, the surface energy losses were ignored. Close inspection of the deviations between theory and experiments indicates systematic trends that can be ascribed to the surface energy losses. We found here that taking into the account the surface energy losses further improves the agreement between theory and experiment. The total mean percentage deviation, equal to 6.65%, decreases to 5.59% if the mathematical form of the Chen formula for SEP is used, or to 5.16% if the Oswald expression is used. The material dependent coefficients in the expression of SEP derived from the emission angle dependence of the elastic peak intensity differ from these coefficients resulting from other methods. We conclude that the determination of SEP from shape of the angular dependence requires the experimental data of high quality, and the reliable theoretical model describing elastic electron backscattering.

Evaluation of the inelastic mean free path (IMFP) of electrons in polyaniline and polyacetylene samples obtained from elastic peak electron spectroscopy (EPES)

Abstract The inelastic mean free path (IMFP) of electrons was determined experimentally for selected polyaniline and polyacetylene samples with Ag and Ni references using elastic peak electron spectroscopy (EPES). The surface composition was determined by XPS and density by helium pycnometry. The high resolution hemispherical ESA-31 and ADES-400 spectrometers were used for measurements in the energy range E = 0.5–3.0 keV and E =0.4 − 1.6 keV, respectively. The integrated elastic peak intensity ratios for sample and reference were calculated using the Monte Carlo (MC) algorithm based on the electron elastic scattering cross-sections database NIST SRD64 version 3.1 and applying TPP-2M IMFPs for polymers. Surface excitation parameters (SEP) and material parameters (ach) for polymers were determined, using the model of Chen, from comparison of measured and MC calculated elastic peak intensity ratios. These corrections proved to be efficient in decreasing the percentage deviations between the obtained IMFPs and the TPP-2M formula IMFPs. The elastic peak of hydrogen was observed in the EPES spectra of polymers. The experimental contribution of the hydrogen to the total elastic peak was 0.58%, while this value obtained from the MC simulations was 1.98%.

Multiple scattering events of primary electrons in directional elastic peak electron spectroscopy

The application of multiple scattering (MS) calculations is presented in order to describe the scattering process of the primary electron beam in crystalline samples and obtain theoretical directional elastic peak electron spectroscopy (DEPES) profiles. Succeeding scattering orders were considered in MS calculations. The contribution of particular sample layers to the calculated intensities is discussed considering forward focusing, attenuation, and defocusing of primary electrons in the crystalline sample. The calculations were performed for a single $[110]$ chain of Al, Cu, Ag, and Au atoms. The results show the importance of multiple scattering events in DEPES profiles. A reduction of the calculated intensities for the $[110]$ direction was observed when the multiple scattering of primaries was considered. Convergence of the recursion method of MS evaluations for the increasing scattering order was ensured by means of $1∕\sqrt{N}$ normalization factor, where $N$ is the number of cluster atoms taken into account. Above the fourth scattering order, no significant changes of the calculated intensities were observed for all the considered kinds of atoms. Theoretical DEPES profiles obtained for the $\mathrm{Au}(111)$ crystal show a similar reduction of intensities for the $[110]$ atomic row as was observed for a single $\mathrm{Au}[110]$ chain. The consideration of MS events in calculations improve the correspondence between experimental and theoretical DEPES scans for $\mathrm{Cu}(111)$ at ${E}_{P}=1.2\phantom{\rule{0.3em}{0ex}}\mathrm{keV}$.

Experimental confirmation of the EPES sampling depth paradox for overlayer/substrate systems

Elastic-peak electron spectroscopy (EPES) has been one of the main tools for obtaining the inelastic mean free path of electrons in solids. Recently it has become clear that, if this type of experiment is done using an energetic electron beam (20–40 keV) and large scattering angles, then the recoil energies of the elastic scattering event for different elements can be resolved. This recoil energy is mass dependent and this fact makes it possible to separate the elastic-peak contributions due to electrons scattered from light and heavy elements. Here we use this energy separation to determine experimentally the sampling depth for an overlayer/substrate system. The sampling depth for a (high- Z) Au overlayer on a (low- Z) C substrate is found to be about two orders of magnitude smaller than for a C overlayer on a Au substrate, whereas the inelastic mean free path of electrons in both materials differ much less. This effect is shown to be a consequence the strong Z dependence of the elastic scattering cross section. The dependence of the spectra on the electron kinetic energy and sample rotation is also dramatically different for both sample geometries.

Determination of the surface excitation correction in elastic peak electron spectroscopy for selected conducting polymers

Inelastic mean free paths (IMFPs) determined by elastic peak electron spectroscopy (EPES) have been frequently evaluated neglecting surface excitations that affect the elastic peak intensity for a sample and a reference material. The surface excitation correction is defined by the surface excitation parameter, P s, denoted by SEP. SEPs for eight selected conducting polymers (polythiophenes, polyaniline and polyethylene) undoped and doped with Pd were determined by EPES using Ag, Ni and Si reference materials for electron energies between 0.2 and 2.0 keV. The mean percentage deviations between IMFPs uncorrected for surface excitations and those calculated with the predictive formulae of Gries and Tanuma et al. were 4.32 and 27.32%, respectively. Relevant deviations for IMFPs corrected for surface excitations were 2.97 and 22.90%, respectively.

Growth and crystalline structure of Au layers on Ni(1 1 1)

We present the results of investigations concerning the growth and crystalline structure of Au layers deposited on the Ni(1 1 1) surface at RT obtained with the use of Auger electron spectroscopy (AES) and directional elastic peak electron spectroscopy (DEPES). For clean Ni(1 1 1) experimental DEPES profiles exhibit intensity maxima corresponding to the close packed rows of atoms. Theoretical DEPES scans were obtained with the use of single scattering cluster (SSC) calculations by means of phase shifts, inelastic mean free paths and scattering cross-sections. A good correspondence between experimental and theoretical scans was obtained for the [ 2 1 ¯ 1 ¯ ] and [ 1 1 ¯ 0 ] azimuths of the Ni(1 1 1) surface. Auger peak kinetics recorded during continuous Au deposition suggests the formation of a flat Au layer till 0.54 ML and then the simultaneous growth of a few layers. Very low intensity reflections and considerable background intensities were found in the LEED patterns, which indicate a weak long-range order in the adsorbate. DEPES profiles obtained for 8 ML of Au on Ni(1 1 1) reflect a short-range order in the adsorbate. Intensity maxima observed in DEPES profiles are characteristic of two mutually rotated Au(1 1 1) domains. Domain populations were found by an R-factor analysis giving n ¯ 1 = 49 ± 2 % for rotated and n ¯ 2 = 51 ± 2 % for unrotated domains.

Interface effect for EPES sampling depth for overlayer/substrate systems

We found that the interface effects for the sampling depth of elastic peak electron spectroscopy (EPES) for the overlayer/substrate system, first noticed by Zommer and Jablonski for the Rh/Al and Au/Ni systems, differ profoundly in their magnitude and dependence on the overlayer thickness when the overlayer and substrate materials are swapped. Monte Carlo calculations performed for the Al/Rh and Ni/Au systems show that their sampling depths, in contrary to the Rh/Al and Au/Ni, can be substantially greater than those of the elements constituting the respective system. The magnitude of the interface effect increases with the energy of the primary electron beam. It is interesting to mention that the sampling depth can be equivocal for a system with overlayer.

Angle-resolved elastic-peak electron spectroscopy: Solid-state effects

It has been frequently reported that characteristics of electrons elastically backscattered from solid surfaces (e.g., the angular distribution) are well described by Monte Carlo simulations of electron trajectories in solids. The theoretical model implemented in these simulations requires knowledge of accurate differential elastic-scattering cross sections (DCSs). In computational practice, the DCSs for isolated neutral atoms constituting the solid are used to simplify the calculations. In reality, the interaction potential between an electron and an atom inside a solid is different from the interaction between an electron and an isolated atom. In the present work, we study changes of the DCSs due to agglomeration of atoms. The interaction between an atom and an electron in the solid is approximated by the muffin-tin potential. It has been found that the DCSs are considerably influenced by the agglomeration of atoms for small scattering angles. The difference for silicon reaches 500% for silicon at 200 eV. On the other hand, electron elastic-backscattering probabilities calculated using DCSs from two potentials were only slightly affected. Calculations and measurements of elastically backscattered intensity were compared for 10 elemental solids, a number of emission angles from 35° to 74°, and three energies (200 eV, 500 eV, and 1000 eV). The experimental angular distributions compare very well with the calculated distributions; the mean percentage deviation between them was about 10% at 200 eV, and decreased to about 5% at 1000 eV. Agreement between theory and experiment was not improved when DCSs determined from muffin-tin potentials were used in the calculations. This result justifies the use of DCSs for isolated atoms in theoretical description of elastic-electron backscattering from surfaces.

Short-range atomic order in the surface layer of sputter-deposited Al0.6W0.4 thin film investigated using directional elastic peak electron spectroscopy

A mechanism of creating maxima in directional elastic peak electron spectroscopy (DEPES) polar profiles for textured polycrystalline samples and for amorphous samples with a short-range atomic order is proposed and discussed. DEPES was used in investigating the atomic order in Al and W polycrystalline samples. The maxima found in DEPES polar profiles correspond with the texture expected for these samples on the basis of literature data. A method of processing DEPES data of amorphous metallic alloy with a short-range atomic order is proposed and successfully used for an Al0.6W0.4 alloy. The results obtained indicate that nanocrystals with an average linear size in the range of 1nm and with a particular orientation with respect to the sample surface are present in the surface layer of the sample investigated. A concentration of nanocrystals with such orientation equal to 1.9×1019/cm3 was found when using of the contrast obtained for the maximum in the DEPES polar profile, corresponding to an incidence angle of zero. The contrast of DEPES profiles for the Al0.6W0.4 sample increases remarkably during annealing at 373K. An increase in the amount of preferentially-oriented nanocrystals is suggested as the probable reason for this increase.

Influence of Recoil Effect and Surface Excitations on the Inelastic Mean Free Paths of Electrons in Polymers

In this work, the influence of recoil effect and surface excitations on the inelastic mean free paths for polythiophenes is investigated. The inelastic mean free paths of electrons in polythiophenes are measured with the elastic peak electron spectroscopy method using the Ag standard and the electron elastic scattering cross-sections from the database NIST 3.1 in the electron kinetic energy range 200–5000 eV. The Monte Carlo model is applied for evaluating the electron backscattering intensities from the polymers and the Ag standard, as well as for evaluating electrons quasi-elastically backscattered from atoms of different atomic numbers (the recoil effect). The surface excitation corrections are accounted for using the formalism of Chen, with the material parameters for polythiophenes evaluated from the elastic peak electron spectroscopy method. Deviations due to recoil effect and surface excitations to the inelastic mean free paths are compared and discussed. Correction to the inelastic mean free paths due to recoil effect is considerable but is smaller, however, than the correction due to surface excitations. Accounting for recoil effect and surface excitations leads to improvement of the inelastic mean free paths, as compared to the inelastic mean free paths resulting from the predictive formulae of Gries.

Elastic electron backscattering from flat and rough Si surfaces

In earlier studies, it has been found that the surface roughness influences only weakly the inelastic mean free path values resulting from elastic peak electron spectroscopy. To address this issue, in the present work, the elastic electron backscattering intensity has been systematically studied in numerous experimental configurations. Measurements were made for the flat Si sample, and for the Si sample with well-defined roughness in the form of long trapezoidal channels. Such structure is a good representation for one-dimensional models used in theoretical description of the surface roughness effects in XPS. It has been found that there is a pronounced influence of the surface roughness on the electron backscattered intensity at the primary beam incidence different from normal. These results were observed at energies of interest for surface sensitive electron spectroscopies, i.e. 200, 500 and 1000 eV. Additionally, the elastic backscattering probability for a flat surface was found to agree well with the theoretical predictions. Correction for the surface energy losses further improved the agreement between theory and experiment.

Investigation of CoPd alloys by XPS and EPES using the pattern recognition method

The electrons elastically backscattered from a solid surface are visualised in the elastic peak electron spectroscopy (EPES) spectra, which intensity recorded at certain geometry and energy of primary electrons, is a measure of electron backscattering probability. The theoretical description of electron interaction with a solid surface reflected in the EPES spectrum and its vicinity is of a complex character. In the present work the EPES method was applied for quantification of the in-depth composition in sputter-cleaned CoPd alloys and annealed Co50Pd50 at different measurement geometries and electron energies. The present approach applies the pattern recognition method called the fuzzy k-nearest neighbour (f kNN) rule to the line shape analysis of the EPES spectra recorded at Co, Pd and CoPd alloys. The proposed method was successful for distinguishing qualitatively the effect of the surface roughness, the texture and the grain size, and for identifying quantitatively the surface composition. Discrepancies between the results are discussed, within the limitations and sensitivity of the applied method, and the accuracy of the electron elastic cross-sections.

Angular‐resolved elastic peak electron spectroscopy: experiment and Monte Carlo calculations

AbstractElastic peak electron spectroscopy (EPES) is widely applied for determining the inelastic mean free path (IMFP) of electrons, which is of crucial importance for quantitative electron spectroscopy. For this reason, it is highly desirable to verify the reliability of the EPES method for a variety of elements and different experimental geometries. In the present work, elastically backscattered electrons from Au, Ag, Cu, Fe and Si were measured at different experimental geometries (fixed incidence angle along the surface normal and various emission angles) for selected electron energies: 200, 500 and 1000 eV. The experimental angular distributions of elastically backscattered electrons were compared to the Monte Carlo (MC) calculations by the conventional and the reverse simulations.Agreement between the results of conventional, reverse MC calculations and the experimental values of electron backscattering intensity was obtained. Application of the procedure for the surface‐excitation corrections resulted in better agreement between the simulated and the experimental data. Larger values of deviation were observed for grazing emission angles, indicating inaccuracy of the correcting procedure and the values of the material parameter, independent of the electron energy and the geometry of measurement. Copyright © 2006 John Wiley & Sons, Ltd.