No abstract

QUASES Analyze and Generate were used to model the extrinsic loss structures for XPS spectra of oxide films grown on iron in such a way that their thickness and structure could be determined. The Generate program used in conjunction with spectra of model iron oxides allowed for both magnetite (Fe 3 O 4 ) and maghaemite (g-Fe 2 O 3 ) structures to be identified in all films studied. These structures were identified as overlying layers in the oxide films and were usually intermixed at their interface. The absence of other iron oxide structures within the film could be tested based on their goodness of fit to the experimental spectrum. Comparison of the thickness values obtained using Generate with those found using nuclear reaction analysis suggested that the Generate results were higher by 20%. This difference likely resulted from the use of a calculated inelastic mean free path value for Fe 2p electrons in the Generate calculation rather than using the real attenuation length. For oxide films whose thickness approached 10 nm, the QUASES results for photoelectron spectra obtained with a Zr achromatic x-ray source were compared with those from the standard Al monochromatic source. In this particular case, the oxide thicknesses obtained using Generate and Analyze were found to be more consistent when the Zr source was used. Copyright INTRODUCTIONIn the 1980s, Tougaard and co-workers 1 were able to demonstrate that the contributions of layered and/or laterally distributed surface phases could be calculated from the extrinsic background of an XPS spectrum. The Tougaard algorithms have been found to provide in-depth information to an approximate maximum depth of 5-10 (IMFP, inelastic mean free path) 2 using loss structures extending as much as 200 eV in kinetic energy below the main photoelectron peak. Such algorithms were incorporated into a software package called QUASES. 3Two different programs are included within the QUASES software package: the first program, 'Analyze', allows the user to model portions of the XPS spectral background using as variables the energy loss cross-section and the IMFP of the electron of interest. Although other crosssections are available for use, the universal cross-section developed by Tougaard 4 is the most convenient for most materials. Apart from these variables, depth profile models such as ''buried layer'', ''islands; passive substrate'', ''islands; active substrate'', ''exponential'', and ''several buried layers'' Ł Correspondence to: A. P. Grosvenor, Surface Science Western, Room G-1, Western Science Centre, University of Western Ontario, London, Ontario N6A 5B7, Canada. E-mail: agrosven@uwo.ca can be used with their resulting calculated extrinsic loss structure being compared with the experimental curve.Also available with the QUASES software is the 'Generate' program, which allows for a model spectrum to be calculated using varying contributions of reference spectra along with the alterable parameters discussed above.5 To date, most of the depth profiles analysed using...

Gadolinium oxide films were deposited on Si͑100͒ substrates from a rod-fed electron beam evaporator using a pressed-powder Gd 2 O 3 target. Films 25 nm thick were shown to be stoichiometric Gd 2 O 3 by Rutherford backscattering and had a dielectric constant at 100 kHz of 16.0 Ϯ 0.3. Transmission electron microscopy and X-ray reflectivity measurements showed that films 7-13 nm thick annealed in oxygen consisted of three distinct layers, an interfacial silicon dioxide layer next to the substrate, a second amorphous oxide layer containing silicon, gadolinium, and oxygen above this, and a polycrystalline Gd 2 O 3 layer on top. Annealing in oxygen reduced the leakage currents, increased the thickness of the silicon dioxide layer, and increased the grain size of the top Gd 2 O 3 layer. The characteristics of the leakage currents through the gadolinium oxide were consistent with a Frenkel-Poole conduction mechanism with a silicon-Gd 2 O 3 band offset of 1.8 V. Interfaces with excellent electrical properties, characteristic of good SiO 2 , were obtained after annealing in oxygen. Annealing of the films in vacuum prior to oxygen annealing reduced the thickness of the interfacial silicon dioxide.

No abstract

Stopping powers have been measured at common ion velocities v = 0.18 cm ns−1 (v/v0 = 0.82) and v = 0.22 cm ns−1 (v/v0 = 1.01) for all projectiles 6 ≤ Z1 ≤ 20 in five solid materials: carbon, aluminum, nickel, silver, and gold. The Z1-oscillation was observed and several new features for the systematics of the oscillation emerge: (i) the amplitude of the oscillation varies with target material and is weak in nickel, (ii) the amplitude changes with ion velocity but this dependence varies with material, (iii) the phase of the oscillation varies by up to 2 units in Z1 according to the material, and (iv) the phase is independent of the velocity over the range studied.We conclude that at low velocity the effect of projectile (Z1) electronic structure on the stopping power cannot be separated from that of the material (Z2).An apparent variation of the stopping power with target thickness was observed and is interpreted in terms of multiple scattering through a recent theoretical treatment.