High quality diamondlike a-C:H has been deposited, at low ion bombardment energies, from an expanding thermal argon/acetylene plasma at high growth rate. It is observed that quality improvement, in terms of hardness, is equivalent to maximization of the refractive index. The highest refractive indices are obtained when the admixed acetylene flow and the argon ion flux emanating from the plasma source are comparable in magnitude, which suggests critical loading. This also indicates that the acetylene has to be dissociated only once. Combination with the observed quality behavior at higher deposition rates suggests that there is one preferred hydrocarbon radical for deposition, probably C2H.

An improved plasma beam deposition set-up, based on an expanding thermal plasma, is presented. Amorphous hydrogenated carbon films have been deposited on glass and crystalline silicon, under variation of the arc current and admixed acetylene flow. The films have been analysed ex situ with infrared absorption spectroscopy, broadband visible light transmission and nano-indentation measurements. These techniques reveal the growth rate, refractive index, bonded C-H density, optical bandgap and hardness. The growth rate and refractive index are found to increase with decreasing arc current and increasing acetylene flow admixture. The quality of the films in terms of refractive index and hardness increases with increasing growth rate and inverse energy coefficient, whereas the bonded hydrogen concentration and optical bandgap then decrease. From comparison of the growth rate dependency with the inverse energy coefficient dependency, we conclude that the growth rate is the preferred parameter in terms of which to describe the film properties because it is directly related to the plasma composition.

No abstract

A study on the effect of substrate conditions was performed for the plasma beam deposition of amorphous hydrogenated carbon ( a-C:H) from an expanding thermal argon/acetylene plasma on glass and crystalline silicon. A new substrate holder was designed, which allows the control of the substrate temperature independent of the plasma settings with an accuracy of 2 K. This is obtained via a combination of a good control of the holder’s yoke temperature and the injection of helium gas between thermally ill connected parts of the substrate holder system. It is demonstrated that the substrate temperature influences both the a-C:H material quality and the deposition rate. The deposition rate and substrate temperature are presented as the two parameters which determine the material quality. In situ studies prove that the deposition process is constant in time and that thermally activated etching processes are unlikely to contribute significantly during deposition. Preliminary experiments with an additional substrate bias reveal that an energetic ion bombardment of the growing film surface does not influence the deposition process. A tentative deposition model is proposed based on the creation and destruction of active sites, which depend on the particle fluxes towards the substrate and the substrate temperature. This model allows the qualitative explanation of the observed deposition results.

Due to cataphoresis, axial segregation of mercury will occur when the gas discharge of a fluorescent lamp is operated by means of a direct current. A consequence of this is a non-uniform axial luminance distribution along the lamp. To determine the degree of axial mercury segregation experimentally, axial luminance distributions have been measured which are converted into axial mercury vapour pressure distributions by an appropriate calibration method. The mercury segregation has been investigated for variations in lamp tube radius (3.6–4.8 mm), argon buffer gas pressure (200–600 Pa) and lamp current (100–250 mA) at mercury vapour pressures set at the anode in the range from 0.2 to 9.0 Pa. From the experiments it has been concluded that the mercury vapour pressure gradient at any axial position for a certain lamp tube diameter, argon pressure and lamp current depends on the local mercury vapour pressure. This observation is in contrast to assumptions made in earlier modelling publications in which one mercury vapour pressure gradient is used for all axial positions. By applying a full factorial design, an empirical relation of the mercury segregation is found for any set of parameters inside the investigated parameter ranges.