Determination of Diffusion Parameters for Arsenic
- PDF / 355,302 Bytes
- 7 Pages / 420.48 x 639 pts Page_size
- 110 Downloads / 210 Views
DETERMINATION OF DIFFUSION PARAMETERS FOR ARSENIC
MICHAEL HEINRICH, MATTHIAS BUDIL AND HANS W. POTZL TU-Wien, E3598, Guflhausstrafle 27-29, A-1040 Wien, Austria ABSTRACT This paper deals with an analysis of the oxynitridation and direct nitridation experiments of Fahey et al.(1984[1]). The inconsistencies in the evaluation of the fractional component of diffusion for arsenic are discussed. Using a recently developed diffusion model an explanation
for the diffusion data for arsenic is given. The standard equation
=
for fGtrv)
C(*IV) analyzing diffusion under nonequilibrium conditions is shown to leadC.A.(IV) to erroneous results.
A consistent treatment must account for the concentrations of arsenic, point defects and arsenic point defect pairs. Pair formation kinetics must be included. Values are presented for the diffusion constants DIA,, DVA. and the reaction constants kIA,, kVAo depending on the fractional interstitial component of diffusion ff'. INTRODUCTION Modeling the redistribution of impurities in silicon is still a major task. Diffusion modeling has advanced due to the implementation of point defect and pair formation kinetics (Budil[2], Orlowski(3], Mulvaney[4], Morehead[5]) putting the models on a sound physical basis, but increasing the number of parameters at the same time. Budil[21 introduced separate treatments of dopant, dopant point defect pairs and point defects. The model needs three equations to be solved for the dopant and dopant point defect pairs, and two equations for the point defects. Compared with earlier models diffusion and reaction constants have been introduced for dopant point defect pairs, which are not defined a priori. This paper deals with the analysis of the oxynitridation and nitridation experiments of Fahey[1] and the evaluation of the diffusion and reaction constants for arsenic at 11000C. NITRIDATION AND OXYNITRIDATION EXPERIMENTS It was found by Fahey[1] that nitridation of a SiO 2 layer enhances diffusion of interstitialmediated diffusion type dopants like P and As, while direct nitridation of the Si surface enhances Sb-diffusion. Diffusion of As is enhanced in both cases - significantly during interstitial injection and slightly during vacancy injection. Generally, those data are explained by an excess generation of interstitials at the surface during the oxynitridation process and an excess generation of vacancies during direct surface nitridation. Both kinds of point defects diffuse into the bulk, where they enhance or retard the diffusivity of the dopants. Fahey used FZ (100) oriented, n-type (p-doped), silicon. Arsenic, Phosphorus, Boron and Antimony were implanted to a dose of 5. 10' 3cm-2 . Wafers were annealed, a thin thermal oxide was grown, and 600A of Si 3N4 was deposited on the sample surface. Lithographic patterning produced three different regions on the same wafer corresponding to three different surface conditions during nitridation: an area in which the Si0 2/Si 3 N4 films remain, which act as a blocking layer to nitridation reactions at the surfac
Data Loading...
