EXPERIMENTAL INVESTIGATION AND COMPUTATIONAL MODELING OF HOT-FILAMENTDIAMOND CHEMICAL-VAPOR-DEPOSITION

Citation
V. Zumbach et al., EXPERIMENTAL INVESTIGATION AND COMPUTATIONAL MODELING OF HOT-FILAMENTDIAMOND CHEMICAL-VAPOR-DEPOSITION, The Journal of chemical physics, 107(15), 1997, pp. 5918-5928
Citations number
44
Categorie Soggetti
Physics, Atomic, Molecular & Chemical
ISSN journal
00219606
Volume
107
Issue
15
Year of publication
1997
Pages
5918 - 5928
Database
ISI
SICI code
0021-9606(1997)107:15<5918:EIACMO>2.0.ZU;2-X
Abstract
A joint investigation has been undertaken of the gas-phase chemistry t aking place in a hot-filament chemical vapor-deposition (HFCVD) proces s for diamond synthesis on silica surfaces by a detailed comparison of numerical modeling and experimental results, Molecular beam sampling using quadrupole mass spectroscopy and resonance-enhanced multiphoton ionization time of flight mass spectroscopy (REMPI-TOF-MS) has been us ed to determine absolute concentrations of stable hydrocarbons and rad icals, Resulting species of a CH4/H-2, a CH3/D-2 (both 0.5%/99.5%) and a C2H2/H-2 (0.25%/99.75%) feedgas mixture were investigated for varyi ng filament and substrate temperatures. Spatially resolved temperature profiles at various substrate temperatures, obtained from coherent an ti-Stokes Raman spectroscopy (CARS) of hydrogen, are used as input par ameters for the numerical code to reproduce hydrogen atom, methyl radi cal, methane, acetylene, and ethylene concentration profiles In the bo undary layer of the substrate. In addition, the concentration of vibra tionally excited hydrogen is determined by CARS. Results reveal only q ualitative agreement between measured data and simulations, concerning concentrations of stable species and radicals probed near the surface , on filament and substrate temperature dependence, respectively. Hydr ogen and deuterium experiments show similar behaviour for all species, In the case of CH4 as feedgas the model describes measured concentrat ion profiles of CH3, CH4, and C2H2 qualitatively well. Large differenc es between model and experiment occur for hydrogen atoms (factor of 2) and C2H4 (factor of 3). For acetylene as feedgas the model is not abl e to give any predictions because no conversion of C2H2 is seen in the model in contrast to the experiment. (C) 1997 American Institute of P hysics.