Possible behavior of a diamond (111) surface in methane/hydrogen systems
- PDF / 833,186 Bytes
- 9 Pages / 593.28 x 841.68 pts Page_size
- 101 Downloads / 160 Views
A combined numerical and experimental investigation into the behavior of diamond (111) surfaces in plasma CVD reactors is presented. Numerically, semiempirical molecular orbital methods are used as a model of diamond (111) surfaces represented by a 20-atom carbon cluster plus surface species. The abstraction of hydrogen atoms by gas-phase hydrogen atoms, the coverage dependence of the heat of formation for submonolayers of CH3 and C2H groups coadsorbed with H, and the energy change for abstraction of H atoms from the surface by various radicals in the gas phase are examined. No barrier to abstraction is found, steric effects in achieving clusters of CH3 groups are large, and C2H and atomic oxygen are found to be the most energetically favored for removal of adsorbed H. Experimentally, relative concentrations of atomic H in the near-surface region as a function of added O2 mole fraction were measured. A weak dependence on O2 concentration is observed, but does not appear to be significant enough to account for observed changes in growth rate. This suggests that other radical species be investigated for their contribution to diamond film growth.
I. INTRODUCTION
The problem of manufacturing high quality diamond films by low-pressure, low-temperature methods continues, at least in part, due to a lack of understanding with regard to the nucleation and growth processes. Only the coarsest knowledge of the influences of temperature, gas concentrations, pressures, flow conditions, field strengths, additives, cycle times, surface conditions, and biases on the gas-surface chemistries is known. Many of our notions come from analogies with other systems, such as carbon pyrolysis1'2 and silicon systems.3"5 However, in spite of a considerable effort, little is known about the actual surface chemistry of diamond film deposition in plasma and filamentassisted methods. For the most part, ideas about surface chemistry are largely conjectural at this point. Significantly, though, Harris et al.16-21'22 reach the conclusion that diamond film growth is dominated by surface kinetics. Essentially the only guide has been and continues to be computer models of the sort used by Tsuda and coworkers6 (TNO) and by Huang and coworkers7 (HFM). The models are atomic descriptions using semiempirical molecular orbital concepts of the surface and the gas phase, the generic term we will use for both the plasma and the filament atmospheres. In the TNO study an MINDO/3 Hamiltonian8 is used, whereas in the HFM study, the somewhat more sophisticated and more reliable MNDO Hamiltonian9 is used. From there the two studies diverge in their assumptions about the system. In both studies simplifying as2296
http://journals.cambridge.org
sumptions are chosen based on some physical intuition or argument about the behavior of the system. It is essential to understand that some assumptions are necessary in order to make the problem tractable computationally. In the work of TNO, epitaxial growth is considered preferable and the system is viewed as a diamond (111)
Data Loading...
