The lower panels show the depth distribution of elastically scattered electrons without energy loss (solid lines) and elastically scattered electrons with a single inelastic loss event (dashed lines) for incoming energies as shown. Results are given for a measurement geometry with the incoming beam along the surface normal (thick lines) and in a more surface sensitive geometry (rotated by 30°, as indicated in the inset in the 40 keV case). The mean free paths were calculated using the TPP formula for Ta2O5. 4 The upper panels show the integrated curves, i.e., the fraction of scattering events that occurs at a smaller depth.
Spectra of a thick, thermally grown, Ta2O5 film taken at 40 keV, 25 keV, and 5 keV. In the first two cases, the O recoil peak is well resolved but the measurement is not background free. The calculated amount of O elastic peak (based on Ta:O DCS ratio and Ta2O5 stoichiometry) was subtracted based on a fit of the main Ta and assuming a peak (blue dashed line). The resulting background (red line) is similar in shape to that observed at5 keV and, in this case, the O elastic peak has merged with the Ta elastic peak.
Energy loss spectra taken at 40 keV and 5 keV for Ta (top) and a 50 nm thick thermally grown Ta2O5 layer (middle). The lower panel shows the spectrum of a sputter-deposited and, subsequently, air-exposed Ta film with a native oxide layer.
Spectra (dots) taken at 25 keV for a sample implanted with 1 × 1017/cm2 (top panel) and 5 × 1017/cm2 (lower panel). Ta and O elastic peaks were simulated for different nominal stoichiometries). The stoichiometry that produced a smooth background after subtracting the O elastic peak was judged to be the correct one and is shown in the figure as a red (smooth) line. The spectrum after O peak subtraction is shown as a blue (noisy) line. The variation in the background curves for different choices of stoichiometry is shown as in inset for the 1 × 1017/cm2 sample.
Energy loss spectra taken at 40 keV for Ta, O implanted Ta samples (implantation dose as indicated) and Ta2O5. The incoming beam along the surface normal. All elastic peaks are normalized to equal area, but the plots are offset vertically for clarity. The dashed line on the left indicates the position of the oxygen elastic peak.
Energy loss spectra taken (dots) at 25 keV for O implanted Ta samples (implantation fluence as indicated) and thermally grown Ta2O5. The calculated O elastic peak contribution (thin line, based on the stoichiometries as shown in Table II ) and the background obtained after subtracting this calculated elastic peak from the measured spectrum (thick line) are shown. The lower right panel shows how the background develops as a function of implantation dose.
EELS spectrum, taken at 5 keV, of the sample implanted with 1017 O ions/cm2 before and after removal of the native oxide layer by Ar+ sputtering.
The Differential elastic scattering cross section (in atomic units) at 135° as calculated with the ELSEPA package. 14 Values are shown without absorption and as calculated within ELSEPA with the default parameters.
The oxygen concentration (expressed as the x value in Ta2O x ) obtained for the 25 keV measurement for the two different measurement geometries. The 1 × 1017 implant was measured as received and with the native oxide removed by sputtering.
Article metrics loading...
Full text loading...