THE DIELECTRIC-CONSTANT OF PD-NI ALLOYS

In the last few years a lot of work has been devoted to the deduction of the dielectric constant of materials exploiting the electron energy loss spectroscopy in the reflection mode (REELS). The transition metals palladium, nickel and their alloys represent a particular class of solids in which bulk plasmons and interband transitions are strongly superimposed to their surface counterparts, generating very complex energy loss spectra. In order to deduce information on the dielectric constant E of such materials, an adequate spectral deconvolution is required in order to obtain their pure bulk contribution which is proportional to the optical loss function Im (-1/epsilon). The main assumption made in deducing the dielectric constant by REELS measurements is that the momentum transfer q of incoming electron is negligible; this means to work in the optical limit [epsilon(omega, q) --> epsilon(omega, 0)]. This condition may be accomplished by using suitable primary energies of about some keV. In this frame, a procedure is presented in this work which exploits the simultaneous analysis of two REEL spectra taken at different primary energies (0.7 and 2.0keV) and/or incidence angles (40 and 80 degrees). The sensitivity of surface effects to these two last parameters makes our deconvolution technique a good tool for the removal of their corresponding contributions to the spectra. The analysis of three compositions of PdxNi1-x system (x = 0.2, 0.4 and 0.8) has shown no particular trend in the energetic positions of the observed structure as a function of the composition, whilst an increase of the low energy loss function intensity is found for increasing Pd content. The dielectric constants for the three compositions are also deduced by a conventional Kramers-Kronig analysis and compared with those of pure Pd and Ni as reported in literature.