The gettering efficiency of copper and platinum by cavities formed in silicon after high dose helium implantation and thermal processing has been investigated. The formation of helium bubbles and their evolution into cavities has been investigated by transmission electron microscopy; the measured values of void density, diameter and the width of the void layer can be interpreted by assuming a simple coalescence model. Metal impurities intentionally introduced in silicon by ion implantation are efficiently gettered inside these cavities, probably due to the large amount of unsatured bonds at the void internal surface. Processing at temperatures higher than 1000oC causes a release of the trapped metal atoms which can be gettered again by repeating the process. The method is demonstrated on real devices such as large area diodes (a particle detector) and bipolar transistors. The capability to localize in depth and across the wafer surface on the gettering sites allows the development of a new gettering engineering.

The photoluminescence of erbium-doped semi-insulating polycrystalline and amorphous silicon containing 30 at. % oxygen is studied. The films were deposited on single-crystal Si substrates by chemical vapor deposition, implanted with 500 keV Er to fluences ranging from 0.05 to 6×1015 ions/cm2, and annealed at 300–1000 °C. Upon optical pumping near 500 nm, the samples show room-temperature luminescence around 1.54 μm due to intra-4f transitions in Er3+, excited by photogenerated carriers. The strongest luminescence is obtained after 400 °C annealing. Two classes of Er3+ can be distinguished, characterized by luminescence lifetimes of 170 and 800 μs. The classes are attributed to Er3+ in Si-rich and in O-rich environments. Photoluminescence excitation spectroscopy on a sample with 1×1015 Er/cm2 shows that ∼2% of the implanted Er is optically active. No quenching of the Er luminescence efficiency is observed between 77 K and room temperature in this Si-based semiconductor. The internal quantum efficiency for the excitation of Er3+ via photogenerated carriers is 10−3 at room temperature. A model is presented which explains the luminescence data in terms of trapping of electrical carriers at localized Er-related defects, and subsequent energy transfer to Er3+ ions, which can then decay by emission of 1.5 μm photons.

The transition to single crystal of ion-implanted amorphous Si and Ge layers is described in terms of a liquid-phase epitaxy occurring during pulsing-laser irradiation. A standard heat equations including laser light absorption was solved numerically to give the time evolution of temperature and melting as a function of the pulse energy density and its duration. The structure dependence of the absorption coefficient and the temperature dependence of the thermal conductivity were accounted for in the calculations. In this model the transition to single crystal occurs above a well-defined threshold energy density at which the liquid layer wets the underlying single-crystal substrate. Experiments were performed in ion-implanted amorphous layers of thicknesses ranging between 500 and 9000 Å. The energy densities of the Q-switched ruby laser ranged between 0.2 and 3.5 J/cm2; time durations of 20 and 50 ns were used. The experimental data are in good agreement with the calculated values for the amorphous thickness–energy−density threshold. The model deals mainly with plausibility arguments and does not account for processes occuring in the near-threshold region or below the melting temperature.

No abstract

Experiments are reported for Te and Ag implantation in silicon, as examples of slow and fast diffusers, after furnace or laser annealing. Slow diffusers are substitutionally located at concentrations in great excess of the maximum solid solubility after both processes. Fast diffusers inhibit the solid-phase epitaxial regrowth or are rejected at the sample surface after laser irradiation. Although the epitaxial growth occurs with velocities which differ up to ten orders of magnitude after furnace or laser annealing, the supersaturation is interpreted as due to the same basic mechanism: solute trapping at the moving interface when the residence time is larger than the one monolayer regrowth time. This process is controlled by the diffusion coefficient in the two adjacent phases.