A Comparison of Low Energy BF 2 Implantation in Si and Ge Preamorphized Silicon
- PDF / 1,346,383 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 106 Downloads / 171 Views
A COMPARISON OF LOW ENERGY BF 2 IMPLANTATION IN Si AND Ge PREAMORPHIZED SILICON
Gary A. Ruggles, Shin-Nam Hong*, Jimmie J. Wortman*, Mehmet Ozturk*, Edward R. Myers*, John J. Hren*, and Richard B. Fair" North Carolina State University, Raleigh, NC 27695 **Microelectronics Center of North Carolina, Research Triangle Park, NC 27709
ABSTRACT Low energy (6 keV) BF 2 implantation was carried out using single crystal, Ge-preamorphized, and Si-preamorphized silicon substrates. Implanted substrates were rapid thermal annealed at temperatures from 600'C to 1050'C and boron channeling, diffusion, and activation were studied. Ge and Si preamorphization energies were chosen to produce nearly identical amorphous layer depths as determined by TEM micrographs (approximately 40 nm in both cases). Boron segregation to the end-of-range damage region was observed for 6 keV BF 2 implantation into crystalline silicon, although none was detected in preamorphized substrates. Junction depths as shallow as 50 nm were obtained. In this ultra-low energy regime for ion implantation, boron diffusion was found to be as important as boron channeling in determining the junction depth, and thus, preamorphization does not result in a significant reduction in junction depth. However, the formation of junctions shallower than 100 rmu appears to require RTA temperatures below 10000 C which can lead to incomplete activation unless the substrate has been preamorphized. In the case of preamorphized samples, Hall measurements revealed that nearly complete electrical activation can be obtained for preamorphized samples after a 10 second rapid thermal anneal at temperatures as low as 600'C.
INTRODUCTION Ion implantation has been the workhorse for establishing and controlling dopant profiles in VLSI technology. The demands on this technology continue to increase as device dimensions shrink. For future 0.25 •tm CMOS designs, source and drain junction depths may be as shallow as 70 nm [1]. The extension of ion implantation into the regime of very shallow junctions necessitates a reduction in ion energy. For p~n junctions, the reduction in B ion energy may be accomplished in two distinct ways. The incident B+ ion energy may be reduced (to less than approximately 5 keV for junctions shallower than about 150 nm [2]) or molecular implant species, such as BF 2 may be employed. The former approach suffers from the practical limitation of a required ion energy below that which is available on most commercial implanters. This limitation can be overcome, and is the subject of study reported elsewhere [3]. The later approach, which is the subject of the current study, works because the BF, ions dissociate upon impact with the Si substrate surface, dividing the total energy between B+ ions and F+ ions in a ratio of (11/49) to (38/49). This technique, however, is clearly more complex, involving the simultaneous implantation of F' ions into Si, as well as the desired B+ ions. Unfortunately, low energy B implantation into crystalline Si results in the creation of a chann
Data Loading...
