The Application of Solid Phase Epitaxy for the Incorporation of Substitutional Carbon in Silicon
- PDF / 315,949 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 110 Downloads / 185 Views
INTRODUCTION Carbon in silicon alloys (Sil-yCy) are being investigated as band gap engineered materials with possible applications in silicon based heterojunction bipolar transistors (HBTs) [1-6]. It would be expected that increasing the carbon concentration in unstrained silicon would raise the band gap from the value for silicon (1.1 eV) to that of SiC (2.5 eV) and finally to that of diamond (5.5 eV) [7]. However, in strained Sil-yCy alloys the band gap initially decreases with carbon incorporation [5,8]. This interesting phenomenon could be exploited for fabrication of Si HBT devices using Sil.yCy alloys in the base. In order to achieve a band gap reduction, carbon must be incorporated into the silicon lattice substitutionally. The difference in the covalent radii of carbon and silicon strains the lattice around the carbon atoms. The presence of strain eliminates the degeneracy of energy levels associated with unstrained silicon, and thus reduces the band gap. Deposition of Sil-yCy alloys has been achieved by using remote plasma enhanced chemical vapor deposition [1], molecular 473 Mat. Res. Soc. Symp. Proc. Vol. 321. @1994 Materials Research Society
beam epitaxy (MBE) [2-51 and by carbon implantation followed by an anneal to induce solid phase epitaxial (SPE) regrowth [6]. In this study, carbon was implanted from 1.75 x 1015 to 1.05 x 1016/cm 2 through a 400A screen oxide into 16-20 fl-cm Si(100) substrates at 00 tilt. After removing the screen oxide, one wafer with each carbon dose was then implanted with 25 keV, 4.0xl015/cm 2 29 Si÷ to amorphize the wafer at 00. This process, which will be referred to as "post-amorphization", is thought to improve recrystallization. The screen oxide was removed to ensure that the silicon implant was deeper than the carbon implant. All of the wafers received a 7000 C, 30 minute anneal in N2 ambient to induce SPE regrowth of the implanted region. A rapid thermal anneal (RTA) was then performed at 1000°C for 30 seconds. Information on the microstructure of the wafers was obtained from TEM analysis and RBS channeling measurements. SIMS measurements provided the concentration profile of carbon within the silicon substrate. FTIR analysis determined the relative levels of carbon incorporated into substitutional lattice sites. Electrical measurements on n+/p+ diodes were used to estimate the band gaps of the silicon/carbon alloys. RESULTS AND DISCUSSION The RBS/ion channeling spectra of the non-amorphized wafers are displayed in figure 1. The large surface peaks of the 7.0 and 8.75 x 1015/cm 2 C-implanted samples relative to the starting material indicates that these samples remained heavily damaged after annealing. Thus at doses of 7.0 x 10S/cm 2 or greater, the carbon inhibited effective SPE regrowth. The surface peaks of the three samples implanted with 5.25 x 10tS/cm 2 C-implanted samples than for the lower doses. EOR damage as well as a high density of damage above the EOR band was present with dense thickets of microtwins and dislocations extending down from the surface of th
Data Loading...
