Mechanism of a low pressure synthesis of diamond has been investigated through a thermodynamic calculation and experiments.
Nucleation of a metastable phase focusing on diamond was studied, based on the classical nucleation theory. It was shown that when the specific surface energy multiplied by the molar volume to a power of two-thirds is larger for the stable allotrope than for the metastable one, the thermodynamic stability between two allotropes can be reversed below a certain size of the particle. The nucleation intensity ratio of the stable to the metastable phase is shown to be critically affected by the variation of the specific surface energy ratio of two phases. Within uncertainties of the surface energy data of diamond and graphite, the dominance of nucleation changes from graphite to diamond. A low pressure synthesis of diamond can be analysed also by the chemical potential of carbon, which is the criteria for the transfer of carbon from one phase to another.
The atomic hydrogen hypothesis was critisized based on the analysis of the chemical potential of carbon. The atomic hydrogen hypothesis for preferential etching of graphite over diamond is equivalent to saying that the chemical potential of carbon in the gas phase is larger than that in diamond and smaller than that in graphite, which is contradictory to the well-established stability between diamond and graphite. The chemical potential of carbon between diamond and graphite was derived to be reversed when the size is sufficiently small. The number of atoms in the carbon clusters that make diamond more stable than graphite was estimated to be ＝104 at 1200 K and 2700 Pa by the reported values of the surface energy data. This number is much smaller than that for the other system where the formation of the metastable phases is the rule rather than exception. Increase in this number, which will be advantageous to the low pressure synthesis of diamond, can be achieved by decreasing the surface energy ratio of diamond to graphite. Surface energy modification of diamond or graphite was suggested to be responsible for nucleation and growth of diamond in the CVD process.
The embryonic clusters generated by the supersaturated vapor species are known to have the electron affinity and the ionization energy approaching the work function of the condensed phase of the species. In a hot filament method, electrons are continuously emitted from the hot filament. Thus, on the presence of these electrons the embryonic clusters would become charged. Thermodynamic analysis shows that singly of multiply charged embryonic clusters tend to become stable nuclei because the presence of charges decreases the nucleation barrier markedly. The surface of the charged clusters is known to make an electrical double layer, modifiying the interfacial energy, which might reverse the stability between diamond and graphite for a sufficiently smaller cluster. Involvement of the charged clusters is supported by the report on the growth of the diamond film by adding halogen elements without hydrogen and gas activation. Involvement of the charged clusters is further supported by the detailed analysis of the high growth rate of the diamond film compared to its very low vapor pressure at the substrate temperature and of the evolution of the cauliflower structure for the high supersaturation and of the formation of the soot on the steel surface. The tendency of diamond nucleation on the high curvature area such as an edge instead of a valley is against the well-established concept of nucleation because the former provides the higher nucleation barrier. This behavior also supports the involvement of charged clusters because the high curvature area induces the high electric field gradient, which can attract the charged ones.
And three experiments have been performed with hot filament CVD system to prove the involvement of charged clusters in the diamond CVD process. First, by varying the thickness of a quartz plate, which was used as a substrate holder, it has been found that when the charge transfer rate is relatively high only soot is deposited on a Fe substrate, and as the charge transfer rate is decreased by increasing the thickness of a quartz substrate holder, the number and the size of diamond deposited on the initial soot are increased. This result is the evidence of charged cluster in the diamond CVD process and the fact that diamond instead of macro crystalline graphite is deposited, when the charge transfer rate was low, means that the charge alters the stability between diamond and graphite. In the second experiment, deposition was performed applying bias voltage between filament and substrate. Positive bias facilitated diamond deposition. And also it has been confirmed that under the high negative bias voltage, cauliflower instead of diamond is deposited. These results imply that negatively charged clusters are responsible for the formation of diamond by the CVD diamond. Finally, by observing the SEM images of soot deposited for 5 min, 10 min, 20 min, and 1 hour each, it has been found that the particles of soot grow from 10~30 nm to 100~300 nm.
