On the influence of iron on the reaction between silicon and nitrogen
- PDF / 839,393 Bytes
- 4 Pages / 593.28 x 841.68 pts Page_size
- 64 Downloads / 194 Views
I. INTRODUCTION Iron, often present as an impurity in silicon, influences the reaction between silicon and nitrogen. At least three possibilities have been suggested to explain its effect: (1) it could help remove an oxide layer that almost always exists on the surface of silicon prior to the reaction,1 thereby accelerating the reaction, particularly during the early stages; (2) at temperatures near the melting point of silicon, iron and silicon form an alloy that has a lower melting point than pure silicon (the resulting liquid silicon alloy increases the rate of reaction and promotes the formation of the silicon nitride beta phase) 2 ' 3 ; (3) finally, it has been suggested that iron catalyses the formation of active nitrogen that promotes the formation of the beta phase.3 This last mechanism is probably only important in the absence of liquid silicon, i.e., in the absence of a rapid reaction path leading to the beta phase. There is experimental evidence for the first two possibilities,1"3 but there is none for the third.1 If the activity of nitrogen is in some way increased by iron, the reaction carried out in its presence might be expected to be more rapid than a reaction in its absence. At temperatures below the melting point of silicon, an increased rate of reaction in the presence of iron has not, however, been observed.1 This might be because the reaction step, between nitrogen and silicon, is not the ratecontrolling step in the overall process. Iron does appear to reduce the length of an induction period and this has been attributed to the catalytic removal of an oxide layer. ' It could also be due to more rapid nucleation of the product on the silicon surface. Hydrogen in the nitriding atmosphere is usually considered to react with oxygen.2'3 An increased reaction rate is explained by a rapid removal of an oxide layer. It has been argued,3 however, that hydrogen also may influence the activity of nitrogen, but a direct experimental check of this possibility has not been reported. Active nitrogen has been considered to be atomic J. Mater. Res. 3 (5), Sep/Oct 1988
nitrogen,3 but there exist other states in which the activity of nitrogen might be increased. Thus, iron may facilitate dissociative chemisorption of the nitrogen,4'5 or it may activate the nitrogen by raising one or more of its bonding electrons to a higher energy state.6 In any case, hot iron appears to activate nitrogen in several systems.1-3 This paper discusses several steps that might be rate-controlling in the nitridation of silicon, and also presents new results that show that iron somehow activates nitrogen. II. REACTION MECHANISM Atkinson et a!.7 have suggested that the first step in the nitridation process is the adsorption of nitrogen onto the silicon surface, and that after nucleation Si3N4 grows as the silicon and nitrogen migrate to the reaction site, Fig. 1. Recent experimental results support this mechanism.8 Several morphologies may appear: (1) product may form along the surface, (2) needles may grow out from the surface, and (3
Data Loading...
