Thermally induced defects in n-type and p-type silicon

Abstract

n‐type and p‐type silicon crystals were quenched from 900 to 1400 °C into a cooling liquid. From Hall coefficient and conductivity measurements a model was deduced, in which the changed electrical properties of the quenched samples are caused by an additional shallow donor and a more complex type of defect which acts – depending on Fermi level – as an acceptor or donor, respectively. An energy level nearer than 0.07 eV to the bottom of the conduction band was estimated for the shallow donor whereas an acceptor level EA = EC − 0.32 eV and a donor level ED = EV + 0.40 eV are connected to the more complicated defect. A migration energy of about 1 eV is ascribed mainly to the shallow donor, which seems also to be responsible for a distinct maximum of defect generation at 1110 °C quenching temperature. According to the given model a dependence of quenched‐in defects on the initial doping does not appear to be very important up to net acceptor or donor densities of about 5 × 1015 cm−3.

Self‐ and impurity diffusion in Ge and Si

“…Seeger and Chik [ l ] concluded that qnenching phenomena in Xi are more complex than in Ge and that the experimental information a t present is not, sufficient to attempt an identification of the site occupied by the donor. This view is substantiated by more recent work [43], [42].…”

“…4) has been used. This diffusivity range embraces the reported Au interstitial diffusivity of 2 3 x Quenching experiments in Si by different workers show general agreement on the existence of a donor defect [43]. Seeger and Chik [ l ] concluded that qnenching phenomena in Xi are more complex than in Ge and that the experimental information a t present is not, sufficient to attempt an identification of the site occupied by the donor.…”

“…I n both situations it is explicitly assumed (i) that the substitulional solubility of the impurity exceeds the vacancy concentration and (ii) that the diffusivity of theinterstitial impurity exceeds the vacancy diffusivity. Quenching experiments in Xi [43] suggest that the first assumption could be invalid for Ni and, also possibly, for Au at about 900" C.…”

“…The second assumption is probably valid for Ni, but is questionable in the case of Au. Quenching from -1100 "C introduces lo1* to 1015 centres c n r 3 [43], and if these are identified as vacancies, we arrive at an average vacancy diffusivity a t 1100 "C of 10-7 to em2 5-l at 1100 "C ( Fig. 4) has been used.…”

The distinction of different kinds of thermally induced defects in silicon

“…I n this way, the existence of different TID types can be proved, one of them also without the participation of iron. As regards the distinction of TID's, this paper exceeds presented results [3,4] and, thus, gives an explanation to some of the existing contradictions.…”