RBS analysis of thin amorphous YBaCuO films: comparison with direct determination of oxygen contents by NRA☆

Abstract YBaCuO superconductor thin films are usually grown by high temperature deposition on a substrate containing oxygen (MgO, SrTiO 3 , …). In order to measure the oxygen content by fitting the RBS spectra corresponding to the cations with the RUMP simulation program, the stopping power of the film elements must be known with a good precision. We therefore tested the accuracy of the stopping power data in RUMP (version 3.502) by analysing the results from well defined YBaCuO samples of various thicknesses prepared by room temperature cathodic sputtering on silicon substrate. A similar experiment was performed on a Ta 2 O 5 /Si reference target. We used 2.2 MeV 4 He + , 2.4 MeV p, or 0.86 MeV d RBS and 16 O(d,p) 17 O ∗ NRA to measure the relative and absolute content of cations and oxygen of the thin films. In the particular case of substrates containing no oxygen, the relative oxygen contents of YBaCuO/Si layers were measured with high precision using as a reference a thin layer of tantalum oxide of a well known stoichiometry (Ta 2 O 5 ) deposited on silicon. The absolute quantities of Y, Ba, Cu and O atoms were determined using a RBS reference target (Bi implanted in Si) and a reference target of oxygen (Ta 2 O 5 /Ta). It was found that the relative and the absolute measurements give very similar results. The 2.2 MeV 4 He + spectra from the YBaCuO/Si samples and from the Ta 2 O 5 /Si reference were analysed using the simulation program RUMP with normalizing factors based on the above quoted Bi standard. It was systematically observed that the RUMP calculations for 4 He result in a relative oxygen content about 10% lower than that deduced from direct measurements. A good fit can be obtained by decreasing the oxygen stopping power in the data base by 12%.