INVESTIGATION OF BORON-DIFFUSION IN POLYSILICON AND ITS APPLICATION TO THE DESIGN OF P-N-P POLYSILICON EMITTER BIPOLAR-TRANSISTORS WITH SHALLOW EMITTER JUNCTIONS

Boron implantation into polysilicon and subsequent diffusion is investigated with the aim of fabricating shallow (< 0.05-mu-m) p-n-p polysilicon emitter transistors. It is found that for implantation doses above 5 x 10(15) cm-2 precipitation of the boron at the as-implanted peak occurs. This behavior can render a large proportion of the implanted boron immobile, and limits the boron chemical concentration in the polysilicon to around 1-2 x 10(20) cm-3. The corresponding electrical concentration is found to be limited to a maximum of 2 x 10(19) cm-3. Subsequent modeling of the profiles show that the enhancement of boron diffusivity in polysilicon, over that in single-crystal silicon, is only a factor of 50-220, which compares with a value of 10(4) for arsenic. This low value of enhanced diffusion means that the simultaneous achievement of highly doped polysilicon (greater-than-or-equal-to 5 x 10(19) cm-3) and shallow (< 0.05-mu-m) emitter junctions, becomes extremely difficult. Base current and emitter resistance are measured on shallow p-n-p polysilicon emitter transistors, and it is shown that the use of a deliberately grown interfacial oxide layer can decrease the base current by a factor of 10, and increase the emitter resistance by a factor of around 2. Comparisons with identical n-p-n polysilicon emitter transistors show that the modeled interfacial oxide tunneling parameters for n-p-n and p-n-p devices are inconsistent.