Atomistic simulation study of the interactions of SiH3 radicals with silicon surfaces

Abstract

SiH 3 radicals created by electron impact dissociation of SiH4 in reactive gas discharges are widely believed to be the dominant precursor for plasma deposition of amorphous and nanocrystalline silicon thin films. In this article, we present a systematic computational analysis of the interactions of SiH3 radicals with a variety of crystalline and amorphous silicon surfaces through atomistic simulations. The hydrogen coverage of the surface and, hence, the availability of surface dangling bonds has the strongest influence on the radical–surface reaction mechanisms and the corresponding reaction probabilities. The SiH3 radical reacts with unit probability on the pristine Si(001)-(2×1) surface which has one dangling bond per Si atom; upon reaction, the Si atom of the radical forms strong Si–Si bonds with either one or two surface Si atoms. On the H-terminated Si(001)-(2×1) surface, the radical is much less reactive; the SiH3 radical was reflected back into the gas phase in all but two of the 16 simulations of radical impingement designed to sample the high-symmetry adsorption sites on the surface. When SiH3 reacts on the H-terminated surface, it either inserts into the Si–Si dimer bond or returns to the gas phase as SiH4 after abstracting H from the surface. The insertion into the Si–Si bond occurs through a dissociative adsorption reaction mechanism that produces two surface SiH2 species after transfer of one of the H atoms from SiH3 to one of the dimer Si atoms. The energetics and dynamics of the surface reactions are analyzed in detail. During simulations of a-Si:H film growth, adsorption onto a dangling bond, dissociative insertion, and H abstraction reactions also were observed to occur with similar energetics as the corresponding reactions on crystalline surfaces. The radical is much more mobile on surfaces of a-Si:H films than crystalline surfaces, especially when the hydrogen concentration in the amorphous film and, thus, on the surface is high.

Time-resolved cavity ringdown study of the Si and SiH3 surface reaction probability during plasma deposition of a-Si:H at different substrate temperatures

“…Moreover, direct measurements of ␤ for the silane radicals have been carried out by means of the various diagnostics mentioned in Sec. I, revealing ␤ values of ϳ0.2-0.3, 6,8,[10][11][12][13]20,25 [40][41][42][43][44] and molecular dynamics simulations, 41,[45][46][47][48] which address the surface reactions at the atomistic scale. For SiH 3 , only indirect information is available about the dependence of ␤ on the substrate temperature from the early study of Matsuda et al 6 This study, though, suffers from the uncertainty that the measured ␤ values are averaged overall plasma species.…”

“…This trend is in agreement with the observations from molecular dynamics (MD) simulations. [45][46][47][48] To get an indication of the relative importance of the different radicals to a-Si: H film growth, the Si growth flux due to a specific radical, ⌫ Si (i.e., the number of Si atoms deposited per second by the specific radical) can be estimated from the value of ␤ and the density n in the plasma: 54 …”

“…9), consistent with the observations from MD simulations. 46 To obtain ␤ values in a rf plasma, Matsuda et al 6 employed the aforementioned indirect method of evaluating the conformality of deposition profiles. This method, however, yields only compound values of ␤ that are averaged overall plasma species coming to the surface.…”

“…There, it is argued that SiH 3 contributes to a-Si: H growth via a two-step reaction mechanism. First, a gaseous SiH 3 abstracts a surface hydrogen directly from the a-Si: H surface by means of the Eley-Rideal process, 41,46 thereby forming a dangling bond. From MD simulations it has been found that this process has a reasonably low activation energy ͑ϳ0.09 eV͒ (Ref.…”

“…Then, a second SiH 3 sticks on the dangling bond to contribute to a-Si: H film growth. This second step, which may or may not involve surface diffusion of SiH 3 , has no activation barrier, 46 corresponding to a substrate temperature independent s. As ␤ = ␥ + s, this two-step reaction mechanism also implies a substrate temperature independent ␤. Although this is only one possible growth model that complies with the boundary conditions, the model clearly demonstrates the far-reaching implications of our experimental findings on the growth mechanism of a-Si: H.…”