296 Surface Science 251/252 (1991) 296-300 North-Hoiiand zyxwvuts Reflection electron energy loss spectra of silver: a quantitative analysis F. Yubero, J.M. Sam, E. Elizalde and L. Galhn zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK Deparramento Fisica Aplirada C-XII, Vnioersidad Autdnoma de Madrid, Can~tkhnco, E- 28049 M adrid, Spain Received 1 October 1990; accepted for publication 30 November 1990 Reflection electron energy kxs spectra (REELS) of ~o~y~~sta~~~~e silver have been measured at different primary beam energies ranging from 200 to 1500 eV and for different combinations of incidence and detection angles. The spectra have been Kramer%K.r&.ig analyzed and the corresponding ELF have been derived. It is shown that whereas variations in the geometry do not cause significant changes, the primary beam energy has to be > 600 eV in order to obtain an ELF in acceptable agreement with those determined optically 1, Introduction Electron energy loss spectroscopy in the reflec- tion mode (REELS) has been widely used as a surface and interface analytical technique [l-3] which provides also information on the electronic properties of the solid [1,4,53. In addition, REELS spectra carry valuable information on the interac- tion of electrons with the solid which can be retrieved if they are adequately processed [6]_ Thus quantitative analysis of REELS spectra has en- abled to obtain inelastic cross sections of low energy electrons [7,8] as well as optical properties of solids [9-111. However, important questions concerning the effect of the surface and non- dipole contributions to spectra measured under different experimental conditions as well as their consequences in the appli~abi~ty of the technique to obtain optical data have been raised [9-121. In the present paper REELS spectra of poly- crystalline Ag at different primary energies in the range 200 to 1500 eV and for different combina- tions of incidence and detection angles using a CHA have been quantitatively analyzed in order to observe possible distortions and influences due to changes in the primary beam energy and/ or the experimental geometry. Thus, energy loss func- tions (ELF) determined from the experimental spectra are compared with existing optical and TEELS (i.e. transmission electron energy loss spectroscopy) data in order to observe the limits of the technique to provide reliable optical data. 2. Experimental Experiments have been performed in a LHSlO ESCA spectrometer with a base pressure < lO^ rif Torr where a poly~~sta~line Ag sample w as mounted, In order to attain a clean surface (as determined by XPS) the sample was sputtered with 3.5 keV Ar’ ions. The REELS spectra were recorded in the non- derivative mode over an energy range of 50 eV including the zero loss peak using a hemispherical analyzer (CHA) operated in the constant pass energy mode (E,,,, = 20 eV) and a primary elec- tron beam ranging in energy from 200 to 1500 eV. The FWHM of the elastic peak was = 0.65-,0.7S eV for energies above 600 eV but increased up to = 0.85-l eV at 200 eV. The geometry of the experimental set up has been schematized in fig. 1. The scattering angle between the incident electron beam and the axis of the analyzer was Fixed ( a: = 120 * f. However by tilting the sample the incident and the collection angle were varied. Thus spectra corresponding to p = 0, 30 and 60 o and labelled as “ r”, “n” and ~39-6~Zg/91~~03.50 8 1991 - Ekevier Science Publishers B.V. ~or~-~~~~a~d~