Abstract
This thesis reports a study, of the atomical and electrical properties, as function of tilt angle for VLSI technology in silicon. The ion species investigated have been arsenic, antimony and boron. Particular attention has been given to the dose loss and electrical activation as a function of the tilt angle in relation to the annealing conditions and the nature of the samples (crystalline or pre-amorphised). All the wafers have been implanted on the new SWIFT single wafer implanter realised by Applied Materials. Thermawave (TW) measurements performed on all samples implanted with As, B, and Sb show high uniformity of the dose implanted (less than 0.5%) on pre-amoiphised (a-Si) and crystalline silicon (c-Si). Rutherford backscattering spectrometry (RBS) has been performed in order to characterise the absolute dosimetry of the SWIFT implanter. Results obtained by RBS with an error of 1.4% have found no significant error in the implanter dosimetry. RBS measurements have also been performed to measure the retained dose for all samples before and after annealing for all the tilt angles investigated. Secondary Ion Mass Spectroscopy (SIMS) has been performed on antimony and boron implanted samples to measure the atomic profile as a function of depth in order to quantify the retained dose, the projected range, the straggle, the peak concentration and the junction depth. Hall measurements have been conducted to study the sheet resistance (Rs), Hall mobility (?H) and sheet earner concentration (Ns) as a function of the tilt angle for all the annealing conditions. Results show that Rs and Ns are functions of the tilt angle whilst ?H is independent of the tilt angle. Pre-amorphised wafers with germanium have also shown a better electrical activation (5%) and a lower junction depth (20-50 nm) compared to the crystalline ones for all the implanted samples. Differential Hall Measurements (DHM) have been performed on some samples for each ion species for particular tilt angles in order to evaluate any diffusion of the dopant for different annealing temperatures and compared with SIMS profiles. Comparison between simulated and experimental data has shown a good agreement only for the as implanted profiles. Computer simulation software is not able to give accurate annealed profiles due to the complex mechanism that involves defects and diffusion during annealing.