The transport of ion-implanted F in amorphous Si is studied using secondary ion mass spectroscopy and transmission electron microscopy. Significant redistribution of F is observed at temperatures in the range 600–700 °C. The measured F depth profiles are modeled using a simple Gaussian solution to the diffusion equation, and the diffusion coefficient is deduced at each temperature. An activation energy of 2.2 eV±0.4 eV for F transport is extracted from an Arrhenius plot of the diffusion coefficients. It is shown that the F transport is influenced by implantation-induced defects.

A study is made of 1 noise in SiGe heterojunction bipolar transistors (HBTs) fabricated using selective growth (SEG) of the Si collector and nonselective growth (NSEG) of the SiGe base and Si emitter cap. The transistors incorporate a self-aligned link base formed by BF 2 implantation into the field oxide below the p + polysilicon extrinsic base. The influence of this BF 2 implant on the 1 noise is compared with that of a F implant into the polysilicon emitter. Increased base current noise S IB and base current are seen in transistors annealed at 975 C, compared with transistors annealed at 950 or 900 C. At a constant collector current, both the BF 2 and F implants reduce S IB , whereas at a constant base current, only the BF 2 implant reduces S IB. This result indicates that the BF 2 implant decreases the intensity of the base current noise source whereas the F implant decreases the base current. The proposed explanation for the increased 1 noise is degradation of the surface oxide by viscous flow at 975 C under the influence of stress introduced during selective Si epitaxy. The influence of the BF 2 implant on the noise is explained by the relief of the stress and hence the prevention of viscous oxide flow.

No abstract

Abstract-SiGe heterojunction bipolar transistors (HBTs) have been fabricated using selective epitaxy for the Si collector, followed in the same growth step by nonselective epitaxy for the p + SiGe base and n-Si emitter cap. DC electrical characteristics are compared with cross-section TEM images to identify the mechanisms and origins of leakage currents associated with the epitaxy in two different types of transistor. In the first type, the polysilicon emitter is smaller than the collector active area, so that the extrinsic base implant penetrates into the single-crystal Si and SiGe around the perimeter of the emitter and the polycrystalline Si and SiGe extrinsic base. In these transistors, the Gummel plots are near-ideal and there is no evidence of emitter/collector leakage. In the second type, the collector active area is smaller than the polysilicon emitter, so the extrinsic base implant only penetrates into the polysilicon extrinsic base. In these transistors, the leakage currents observed depend on the base doping level. In transistors with a low doped base, emitter/collector and emitter/base leakage is observed, whereas in transistors with a high doped base only emitter/base leakage is observed. The emitter/collector leakage is explained by punch through of the base caused by thinning of the SiGe base at the emitter perimeter. The emitter/base leakage is shown to be due to a Poole-Frenkel mechanism and is explained by penetration of the emitter/base depletion region into the p+ polysilicon extrinsic base at the emitter periphery. Variable collector/base reverse leakage currents are observed and a variety of mechanisms are observed, including Shockley-Read-Hall recombination, trap assisted tunneling, Poole-Frenkel and band to band tunneling. These results are explained by the presence of polysilicon grains on the sidewalls of the field oxide at the collector perimeter.

No abstract