Plasma Deposition of Silicon Clusters: A Way to Produce Silicon Thin Films With Medium-Range Order ?
- PDF / 2,905,875 Bytes
- 11 Pages / 414.72 x 648 pts Page_size
- 67 Downloads / 214 Views
be the main film precursor [5,6], even though there is some controversy on this subject [7]. In fact, the two questions are related. The understanding of the growth mechanisms, sketched in Figure 1,can indeed throw some light on the film structure. Focusing on the plasma processes, one usually considers the primary reactions, i.e. the dissociation of the silane molecules after inelastic collisions with electrons [8]. Because of the complex plasma chemistry most of the models used to describe the a-Si:H deposition are limited to these primary reactions. However, even in this simple case the increase of atomic hydrogen flux towards the substrate achieved through: i) a high dissociation of silane [9], ii) a high dilution of silane in hydrogen [10], iii) the use of a layer-by-layer technique [11], results in the growth of an heterogeneous material with long-range order: microcrystalline silicon (jtc-Si:H). Many models have been proposed to explain the growth of this latter material, but most of them fail to explain the formation of crystallites and the long term evolution of the film properties [12]. Now, given these two materials (a-Si:H and ýtc-Si:H) which can be deposited in the same reactor, one can ask whether there is a sharp transition between the amorphous and the microcrystalline silicon materials. In other words, is it possible to grow silicon-thin-film materials with different degrees of medium-range order ? One might think that just by choosing plasma conditions at the border between a-Si:H and g.tc-Si deposition, silicon films with medium-range order should be obtained. However, thermodynamic considerations suggest a discontinuous disorder/order transition, taking place for crystallite sizes of about 3 nm [13]. Nevertheless, some experimental observations support the existence of ordered domains with sizes below 3 nm [14,15,16]. In order to progress in the understanding of the deposition and structure of silicon thin films, we like to stress the importance of secondary reactions. Indeed, the SiH 3 hypothesis which may be reasonable under "soft" plasma conditions, becomes less plausible under conditions of high 855 Mat. Res. Soc. Symp. Proc. Vol. 507 © 1998 Materials Research Society
silane dissociation (needed to achieve high deposition rates) where secondary reactions and powder formation take place [17]. This phenomenon, which is a real drawback in the microelectronics industry because it can lead to a marked decrease in production yield, has attracted much attention in the last five years [18]. Detailed studies of powder formation in silane plasmas have allowed to separate the two main stages of the process, namely the initial formation of a high density of nanosized silicon particles followed by their coalescence to produce powder, into five steps [19]. Moreover, studies based on TEM analysis of the particles produced in silaneargon discharges at room temperature before the coalescence stage have shown that the 2-nm silicon particles are crystalline [20]. Therefore, the incorporation of these crystal
Data Loading...
