なかのにっき

A particle-in-cell/Monte Carlo numerical model has been developed to simulate a direct current discharge self-consistently with the motion and thermalization of both energetic charge-exchange neutrals and sputtered cathode atoms. In the model the charged particle motions are considered in a self-consistent electric field. A one-dimensional glow discharge in Ar has been simulated for the cases of Al and Cu cathode. The background argon gas heating has been predicted, with the temperature rise being larger for the case of Cu cathode which is characterized by a higher sputtering yield than Al. The balance of power input into the gas due to the energetic neutrals, sputtered atoms, and ions is analyzed. The dominant contribution is from energetic neutrals. Comparison of the calculated fluxes of these three species at the cathode surface shows a great contribution of the energetic neutrals into sputtering of the cathode material. The effect of applied voltage has been also investigated. Lastly, the influence of the gas heating on discharge characteristics is discussed.

The pressure dependence of the deposition rate for magnetron sputter deposition of various elemental semiconductors and metals was investigated by x-ray measurements on sputtered films and quartz monitor measurements. It was found that for all elements investigated the dependence of the rate on pressure-distance (pd) is well described by Φ=Φ0(1 - e^-cpd)/cpd. The value of c equals the inverse characteristic pressure-distance product (pd)0, which is the characteristics of the exponential decay of rate with pressure for low pressures. The experimental data of (pd)0 vary from 4.6 Pa cm for aluminum to 120 Pa cm for tungsten. It is shown that (pd)0 depend on both material specific properties and process parameters. The material specific properties are mainly the atomic mass and diameter, and the surface binding energy. The process parameters target voltage and power density act via the increase of the mean free path and the reduction of gas density, respectively, on (pd)0. As a first approximation, the characteristic pressure-distance product for argon as sputtering gas is proportional to the product of target atomic mass, average atomic energy and thermal mean free path.

Pressure measurements in the range of low and medium vacuum are conveniently and accurately performed by mechanical instruments which detect pressure-induced deformations. The stability of both calibration and zero setting are considerably improved by operating the transducers at a stabilized elevated temperature. Due to thermal transpiration the temperature difference between transducer and vacuum vessel may cause a pressure difference which depends on the pressure itself, the geometry and surface properties of the interconnection, and also the gas species. In order to determine the vessel pressure, the empirical correction formula of Takaishi and Sensui has proved to be successful. Our accurate measurements reveal for the first time a minor, but systematic failure of this correction procedure. This failure is attributed to an improper application of the formula, i.e., the disregard of the specific surface properties of the interconnection. To overcome this deficiency, we suggest using fitted values for the temperature and diameter which appear in the formula, instead of actual values. By this procedure the systematic deviations are removed, and the Takaishi and Sensui formula allows description of our vessel pressures better than 0.1%.