Reactive ion beam sputtering of Ti: Influence of process parameters on angular and energy distribution of sputtered and backscattered particles

This work focuses on the angular and energy distribution of secondary particles during the ion beam sputtering of a Ti target in a reactive oxygen atmosphere. The influence of ion species, ion energy, and scattering geometry (ion incidence angle and polar emission angle) was investigated systematically. Ion energies of 0.5, 1.0, and 1.5 keV and an ion current of roughly 10mA were used to sputter deposit TiO2 films from a pure Ti target at an operating pressure of 5.0 x 10(-5) and 6.5 x 10(-5) mbar and an oxygen partial pressure of 1.0 x 10(-5) and 2.5 x 10(-5) mbar for sputtering with Ar and Xe ions, respectively. The angular distribution of the flux of sputtered Ti particles was determined by measuring the thickness of TiO2 films. The flux of sputtered Ti particles is described by the superposition of an isotropic and anisotropic part. The isotropic part increases when increasing ion energy, decreasing the incidence angle, or changing the sputtering gas from Xe to Ar. An energy-selective mass spectrometer was used to measure the mass and energy distribution of secondary ions. Several species of sputtered particles were detected, e.g., Ti, TiO, TiO2, Ti-2, O, and O-2 ions. The most prevalent species are Ti, TiO, and O ions. The energy distribution of the sputtered Ti and TiO ions shows a maximum at an energy between 10 and 30 eV followed by a high-energetic tail proportional to exp(-n.E). Decreasing the scattering angle leads to a decrease in the slope of the high-energetic tail, i.e., it extends to higher energies. The backscattered primary ions (Ar and Xe) and the sputtered atomic O ions show a maximum between 5 and 20 eV, which is followed by a sudden signal drop. For scattering angles less than 90 degrees, additional peaks at higher energy can appear in the energy distribution of Ar, Xe, and O ions. These peaks are related to direct scattering and sputtering events. The experimental results are compared with calculations based on a simple elastic two-particle-interaction theory and with simulations carried out using the Monte Carlo code SDTrimSP. (C) 2017 American Vacuum Society.