Characterization of macroscopic defects in silicon after processing for CMOS and bipolar circuits

Abstract

We have been engaged in a systematic and careful study of the early stages of oxygen precipitation at temperatures up to 850°C which involved transmission electron microscopy correlated with diffraction, optical and IR spectroscopy. We have simultaneously made a detailed study of macroscopic defects induced in Czochralski silicon wafers after repeated heat cycling during the fabrication of circuits. Even though there have been a number of such studies there remain many questions about the effects which are observed, such as why do stacking faults sometimes occur while on other occasions prismatic loop punching predominates? Variable precipitate morphologies are also observed such as {100} plates or octahedra without any general agreement on the reasons for such differences. The question of how precipitates nucleated at lower temperatures grow and develop after repeated high temperature cycles has also not often been addressed. Further questions exist over the heterogeneous or homogeneous nucleation of precipitates, the involvement of transition metals, the unfaulting of stacking fault loops. Answers to many of these questions arise out of our work which seeks to provide an over-view of the variety of oxygen precipitation routes in silicon.