Theoretical Investigation of Implanted Dopant Diffusion From a Silicide Layer to the Silicon Wafer for Ultra Shallow P-N
- PDF / 311,941 Bytes
- 7 Pages / 414.72 x 648 pts Page_size
- 37 Downloads / 211 Views
INTRODUCTION.
As a result of the decrease in device dimensions in ultra very large scale integration (UVLSI), ultra shallow p - n junction formation is becoming a challenging task in the submicron complementary metal oxide semiconductor (CMOS) technology [1]. A promising technique to achieve ultra shallow junctions was recently proposed [2 - 5] and it is based on using a silicide as a diffusion source (SADS). In brief, a thin silicide layer - 30 - 50 nm is formed by some technique on a silicon wafer followed by dopant implantation into this layer only. The p - n junction is formed as a result of dopant penetration from this layer into the silicon during diffusion anneal. One of the major advantages of SADS technique is the absence of implantation damage in the silicon wafer. Moreover, it has been reported [3 - 5] that p - n junction formation by this method leads to low leakage. Basically, in the SADS method after implantation, the diffusion process occurs in two regions, i. e., diffusion of the dopant in the silicide with high diffusivity [6 - 8] and diffusion of the dopant in the silicon with lower diffusivity [9]. One has to emphasize that the redistribution problem of the dopant should be analyzed simultaneously in the two regions and not sequentially as done in [5]. This requirement is important because of the dopant penetration into the silicon occurs already at an early stage of its redistribution in the silicide. The important parameters affecting the p - n junction depth are the energy of implantation, the dose of the implanted dopant atoms, the time and the temperature of annealing, and the possibility that evaporation of dopant from the silicide layer may occur. Evaporation of dopant from the silicide layer might have a pronounced effect on dopant distribution [4 - 5]. Therefore, this effect has to be considered explicitly in the analysis, so that the consequence of evaporation be understand. Moreover, because of the evaporation it is important to take into account the redistribution of the dopant in the silicide. The main objective of this work is to evaluate the influence of the evaporation on the p-n junction depth. A simple two layer model is used assuming constant diffusivities in these layers. The model describes approximately real dopant distribution. Our model is simple but sufficient to investigate the dependence of the p-n junction depth on process parameters.
233 Mat. Res. Soc. Symp. Proc. Vol. 389 ©1995 Materials Research Society
II. THE MODEL AND THE SOLUTIONS. The silicide layer and the silicon wafer are considered as a two phase system with constant dopant diffusivities D1 and D2 in the silicide and the silicon, respectively. The equations describing dopant redistribution in the silicide layer and in the silicon wafer are given by ac,
-t
D
aC-=D2 2 =
i2
-d!x
Data Loading...
