Abstract

A phenomenon leading to locally distorted breakdown characteristics of MOS power transistors in the range of 0.1 to 1 mA, e.g. very far away from real operating conditions, has been observed. The characteristics might have suggested a reduced second breakdown of the parasitic bipolar transistor, but with the subtle difference that the voltage was normal again at larger currents. The attempt to understand the phenomenon in terms of ordinary transistor theory led to the assumption that small-area n-doping contaminations were present in the silicon substrate. Laborious experiments revealed the fact that the instabilities of the breakdown characteristics were a unique function of the plasma etch equipment used for polysilicon definition. The question of interest was now how is it possible for contaminating atoms to penetrate through a rather thick residual oxide during an isotropic plasma etch step. Therefore, the assumption was made that the contaminants were knocked through the residual oxide during the boron implant immediately after the polysilicon etch. TXRF-analysis and C/ V-measurements have shown that sulfur is the most probable candidate for the n-type contamination. The amount of knock-on by boron implants is demonstrated by C/ V and doping profile analysis. Finally, the phenomenon was eliminated by changing the seals of the etch chamber as well as the electrode defining ceramic rings to those of more suitable materials in order to avoid the contaminations a priori.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call