Through the ages, mankind has been fascinated by diamonds because of their high monetary value and physical properties. In 1792 Antoine Lavoiser and Smithon Tennand discovered that diamond is a variant of carbon. The research for mechanisms of diamond synthesis began, however it would be another 157 years until science has its first success. This came in 1954 when H. Tracy Hall, an employee of General Electric, showed that it was possible to transform graphite into diamond using an iron catalyst at an ambient temperature of 1700℃ and 95kbar pressure

This high-pressure, high temperature process is still the most common way to produce synthetic single-crystal diamond outside of their thermodynamically stable area.

Spitsyn and Derjagin in Russia followed a completely different idea. In 1956 they patented a new method of growing diamond in its thermodynamically unstable state from a gas phase at comparatively low temperatures and pressures (3, 4). Further research work revealed that it was possible to grow diamond in a variety of materials such as copper, tungsten, chromium and silicon using different methods of gas activation including the hot-filament (HF), microwave and combustion-flame methods.

 

These results were not published in English until 1981, when they generated a great deal of worldwide interest and research activity in the field of gas phase diamond synthesis.

These techniques are currently used to produce diamond coating and layers for a variety of applications. Industrially upscaled processes in this field are state-of-the art. Manufacturing diamond volume crystals by CVD processes, though, has so far merely been pursued resulting in dissatisfying quality or on a very small scale. However, more recently there had been some news about high-quality diamond single-crystal growth with high growth rate by microwave CVD and even attempts to start commercial production of CVD single crystals.

 

Despite the widespread use of CVD techniques, the exact growth mechanism of diamond is not yet fully understood. Some of the most promising theories to date are explained below to highlight the differences between the growth mechanism on diamond crystal planes.

The aim of the work presented here is to define the main factors which lead to stable single-crystal growth and to characterize the surface morphology, stress states and quality of the planes with regard to those models and theories already outlined.

Experimental detail

All experiments were conducted in a HF-CVD reactor as it is common for coating processes. AKS-doped tungsten wires were used as filaments, positioned in front of tungsten carbide substrates. To obtain a broad range of surface temperature without any changes in filament temperature the samples were placed on a heated stage. Table 1gives an overview of the applied process parameters.

 

The surface temperature of the substrate was measured using a thermocouple in a dummy substrate close to the real one. To measure the temperature of the filaments a two color pyrometer was used.

Results

Seed density



All the experiments showed that there is a specific array of parameter values at which single-crystal growth is possible. Only high temperatures in combination with low methane content in the gas phase stabilizes the growth of single crystals. High temperatures are necessary to control the seed density on the surface and the low methane content influences the desired morphology. Crystals that grow at these conditions obtain a higher growth rate as diamond layers in similar conditions but at lower temperatures. The correlation between the resulting crystal density and the surface temperature

 

Growth rates

The various experiments showed the growth rates of the single crystals to be highly dependent on the process parameters. Lowering substrate temperature after the final nucleation density is reached leads to faster growth but can also accompany undesired secondary nucleation. In an appropriate parameter window average growth rates of 0.46um/h were observed. By increasing the process pressure as well as by employing higher methane contents the crystals average growth rates increased to 0.94 or 0.72um/h, respectively. Comparing these values to layer growth it always has to be taken into account that the growth here proceeds in 3D. Some of the crystals showed growth rates as high as 1.20um/h growing to sizes of over 100um in face-to-face diameter. However, at such conditions the formation of the growth spirals on the facets became clearly more dominant. This inhomogeneous growth proceeded still stable in our experiments, but could lead to increased defect formation in the crystals at longer growth times.

 

Outlook

After all the experiments were conducted it was found that the most important parameters to stabilize the single crystals are the substrate temperature and the methane content in the gas phase. Only the combination of high temperatures and low methane content permit single-crystal growth within reasonable time limits.

The best results were obtained at substrate temperatures between 900 and 950℃ and methane contents of a maximum of 1% to prohibit secondary crystal growth.

The exact growth mechanisms of the excrescences are not yet fully understood at a macroscopic level. A possible solution might be preventing the multiple enwinding of the growth spiral. A temporary increase in the methane content could lead to an adjustment of the altitude differences between the spiral planes.

Although there is still a lot of research work to do, the results confirm that there is a process window in which diamond single crystals can be grown by hot-filament CVD without using mono-crystalline substrates. The possibility to upscale hot-filament reactors offers the outlook of growing large diamond crystals of controlled quality within reasonable time.