Ryuichi Shimizu

Backscattering correction for quantitative Auger analysis: I. Monte Carlo calculations of backscattering factors for standard materials

Abstract The electron backscattering effect which is important for the quantitative interpretation of “matrix effects” in AES is investigated by applying the Monte Carlo calculation technique. The present calculation model is based on the use of a precise elastic scattering cross-section obtained by the partial wave expansion method, as well as on the combined use of Gryzinskis excitation function and Bethes stopping power for inelastic scattering. Systematic calculations of the backscattering factors were performed for over 25 materials including pure elements, compounds, and alloys, which have been widely used as standard materials for practical Auger analysis. The results should enable the accuracy of quantitative analysis by AES to be improved.

Monte Carlo modelling of electron-solid interactions

First, Monte Carlo modelling of electron-solid interactions is briefly reviewed regarding its historical development, followed by quantification in microbeam analysis with various types of signals generated by electron penetration in solids. Second, typical models stimulated by these improvements are explained by demonstrating some of the Monte Carlo calculations which are, the authors believe, still of practical interest and use. Then, a typical calculation procedure for Monte Carlo simulation is described in detail from the standpoint of a physicist who does his or her own programming. Third, up-to-date Monte Carlo calculations as applied to modern surface analysis, high-resolution scanning electron microscopy, Auger electron microscopy and X-ray photoelectron spectroscopy are also presented.

Quantitative Analysis by Auger Electron Spectroscopy

A short review is presented of the theoretical background of a physical model for the quantification of Auger electron spectroscopy (AES) for surface analysis. The recent studies on the data-base for the inelastic mean free paths (IMFP) by Seah and Dench and systematic calculations of the backscattering factors (R) by Shimizu and Ichimura have now enabled standard quantitative corrections comparable to those widely used in electron probe microanalysis, to be accomplished. For quantitative corrections of wider practical use, the present paper proposes the use of functional representations of the backscattering factors for different angles of incidence (ψ) for primary energies ranging from 3 to 10 keV as follows: R=1+(2.34-2.10Z0.14)×U-0.35+(2.58Z0.14-2.98) for ψ=0°, R=1+(0.462-0.777Z0.20)×U-0.32+(1.15Z0.20-1.05) for ψ=30°, R=1+(1.21-1.39Z0.13)U-0.33+(1.94Z0.131.88) for ψ=45°, where U is the ratio of the primary energy to the binding energy, and Z is the atomic number of a sample.

Monte Carlo Calculations on Electron Scattering in a Solid Target

The Monte Carlo technique using a simple single scattering model, different from Reimers, is applied to the field of the Electron Probe Microanalyzer (EPMA) and the Scanning Electron Microscope (SEM). The calculations show fairly good agreement with the experiments on back-scattered electrons, energy dissipation, characteristic X-ray production and secondary electron emission. Also the electron trajectories are depicted on the chart by the plotter connected with the digital computer.

Computer simulation of atomic mixing during ion bombardment

The atomic mixing in the target under ion bombardment is assumed to result from cascades of atomic collision events. Computer simulations have been applied to collision cascades to estimate the depth resolution of surface analysis with an ion probe. The Monte Carlo method based on a single scattering model has been used mainly in the calculation under the assumptions of random collision process, no diffusion and no target saturation processes. High-energy collisions are characterized by a Lenz-Jensen or a Thomas-Fermi potential, while a Born-Mayer potential is used in the low energy region. The simulations have been performed for the bombardment of Ar ions withE0=5 keV and 10 keV at angles of incidence θ=0° and 60° on Si targets. The depth resolutions [the definition of which is explained by (15) in the text] are about 140Å for the Lenz-Jensen cross section and about 80Å for the Thomas-Fermi one for θ=0° atE0=5 keV, and decrease by 20–40% at θ=60° and increase by 70–90% forE0=10 keV.

A Monte Carlo approach to the direct simulation of electron penetration in solids

A new Monte Carlo approach to the direct simulation of scattering process of a penetrating electron in matter has been attempted. The theoretical stopping powers for elementary excitation processes (conduction-electron, plasmon and L-shell electron excitations) were used instead of Bethes stopping power equation to describe the energy loss process in the electron penetration. The result obtained for aluminium describes the energy as well as the angular distribution of the transmitted electrons satisfactorily. This approach has also been applied to copper as a preliminary example. The theory describes qualitatively the energy and angular distributions of the transmitted electrons fairly well, but there is some discrepancy between theory and experiment in the energy distribution.

Dependence of coercivity on the anisotropy field in the Nd2Fe14B-type sintered magnets

The temperature dependence of the coercivity (HcI)of Nd15(Fe1−xCox)77B8 and (Nd1−xDyx)15Fe77B8 sintered magnets and that of the saturation magnetization (Ms) and the anisotropy field (HA) of Nd2(Fe1−xCox)14B and (Nd1−xDyx)2Fe14B single crystals have been observed in the temperature range between 295 and 800 K. The dependence of the coercivity on the major magnetic properties of the matrix phase in the Nd‐Fe‐B based magnets are investigated using the μ0HA vs μ0HcI+Ms plot. It is demonstrated that this method of analysis is useful in studying the coercivity mechanism of the Nd‐Fe‐B based sintered magnets.

Inelastic collisions of kV electrons in solids

Abstract We have calculated electron stopping powers and electron inelastic mean free paths in solids for energies between 1 eV and 10 keV above the Fermi energy using the wave-vector- and frequency-dependent energy loss function recently proposed by Penn. In addition, the energy loss distribution and angular distribution necessary for specifying an inelastic scattering event are consistently obtained. By this approach, the energy loss function determined by the experimentally measured optical dielectric data at long wave-length limit was extrapolated to finite wave-length as a sum of infinite terms of Drude-Lindhard type dielectric functions under a single-pole approximation. It is shown that this method is useful for evaluating inelastic collisions of fast and slow electrons in solids, particularly, in transition metals and noble metals for which the dielectric function of the free electron gas is difficult to be directly extended. Results for the solid media Au, Cu, Ag, Ni and Si are presented.

Ion Beam Assisted Deposition of Metal Organic Films Using Focused Ion Beams

50 keV Ar+ or focused Au+ were irradiated in a trimethyl aluminum atmosphere to investigate characteristics of ion beam assisted deposition. About 80 nm thick films were deposited at a dose of 1×1016/cm2. The film composition and its depth dependence was measured by Auger electron spectroscopy. It was found that the film contains oxygen, carbon and aluminum, and the atomic ratio varies across the film depth. The atomic ratio was 0.8(O):1.1 (C):1.4(Al) at the surface and was 0.8(O):3.3(C):1(Al) inside. A direct maskless pattern deposition was also done using 50 keV focused ion beams.

Monte Carlo study of secondary electron emission

Based on our previous Monte Carlo simulation model of electron interactions with solids, including cascade secondary electron production, in which an optical dielectric function was used to describe electron energy loss and the associated secondary electron excitation, we have systematically investigated secondary electron generation and emission for 19 metals. The calculated secondary yield curve for primary beam energy ranging from 100 eV to 2 keV was found to correspond with the experimental universal curve. The dependence of the secondary yield on the work function was studied numerically, leading to a remarkable scattered deviation from Baroody’s relationship. This deviation shows that the secondary yield relates to different aspects of behavior by electrons in a metal, such as the cascade production process, the stopping power and specific energy loss mechanism for a sample, and the dependence on the electron density of states. The results provide an explanation for the scattered data on the experim...