Cyclic deposition of diamond: Experimental testing of model predictions

Abstract

The vapor deposition of diamond by cycling growth and etchant mixtures was investigated as a function of cycle time, tcycle, the fraction of the cycle at which growth occurs, τgrowth, and the atomic fraction of carbon in the growth mixture xC. A kinetic model that incorporates diamond growth through a new interdependent methyl and acetylene reaction sequence was used to predict both deposition rates and the carbon deposit sp2 fractions. The important influence of aromatics on the predicted sp2 fractions is thoroughly discussed. The results from the kinetic model were tested experimentally with a microwave-activated deposition system equipped to allow reactant gas cycling between growth (CH4 in He) and etchant (O2 in H2) mixtures. The deposits were characterized by their Raman spectra and scanning electron micrographs. Both the kinetic model and the experimental results show an increase in deposit quality (lower sp2 fraction) with decreasing tcycle and with decreasing τgrowth. Linear growth rates estimated from deposit particle dimensions were typically on the order of 1 μm/h. The kinetic model does not address nucleation rates, but experimental results indicate a trend toward lower rates at shorter tcycle and decreasing τgrowth. The deposit character was less sensitive to xC changes in the growth gas, but renucleation was more apparent at higher carbon fractions. Both the model and experiments show a critical cycle time t*cycle; at shorter cycle times the deposit quality is always high, while at higher times significant fractions of sp2 carbon are present. The t*cycle value is a function of experimental conditions, and in the present study was found to be strongly dependent on the fraction of carbon used in the reactant gas and total pressure, and slightly less dependent on τgrowth. Previously reported cyclic deposition studies were found to be consistent with predictions expected from the kinetic model used in this study.