Search

Substrate biasing is an emerging technique extensively used in diamond and cBN chemical vapor deposition (CVD) processes. Direct current bias voltages between -400 V and +500 V are here applied to a diamond deposition surface in an RF induction thermal plasma CVD system. This is made possible by introducing a high-impedance filter network, eliminating the time-varying voltage drop across the plasma-substrate junction and suppressing the radio-frequency interference. A traditional limitation of thermal RF systems is thereby overcome. Negative. bias conditions enhance the initial diamond nucleation density. Positive bias improves the diamond quality and augments the film growth rate. A threefold increase in linear growth rate is attained at +500 V as compared to the unbiased case. In conjunction with optical emission spectroscopic diagnostics, the substrate is used as an electrical and thermal probe. Contrary to do arcjet CVD, there is no secondary discharge created in the RF system at positive bias voltage. Also, the role of ion bombardment at negative bias is shown to be of little importance. It is inferred that the predominant mechanism leading here to changes in the diamond deposit under bias conditions is the promotion and suppression of electron emission from the growing diamond.

Diamond films produced by Ar/H/sub 2//CH/sub 4/ RF induction plasmas have high growth rates but often lack uniformity across the substrate. In this paper, the diamond growth on a probe-like molybdenum substrate, 4.76 mm in diameter, is correlated with the location of the probe in the plasma flame. Optical emission spectroscopy is used to characterize the plasma during deposition. The differences in gas temperature and electron density between the central and the outer region of the flame, as well as the variations in heat flux to the substrate at the different positions, are well reflected in the formed deposit. It is shown that uniform growth can only be expected within the homogeneous core of the plasma. The paper also addresses the gas phase boundary layer above the growing film. Spatially resolved emission spectroscopy measurements of concentration evolution inside the boundary layer region are made possible by the small dimensions of the substrate. A decrease in thermal boundary layer thickness from approximately 4.5 to 3 mm induced by an increase in plasma power enhances the lateral growth of the individual diamond crystallites and improves surface coverage. The typical film thickness growth rate of 70 /spl mu/m/h is, however, not seen to be affected.

In this work, the dependence of the cathode erosion rate on vacuum arc spot velocities and cathode pore size are investigated for four types of graphite under an external variable magnetic field. Different graphite cathode properties show varying erosion rates, indicating that each graphite type should be treated as a different material. Increasing the spot velocity through an increase of the magnetic field intensity on a given cathode material decreases the erosion rate. At given magnetic field values, cathodes having higher arc velocities show an increase in the erosion rate. Other studies indicating that the increase in the arc spot velocity leads to a reduction in macroparticles emission indicate a possibility of increasing. The emission of the carbon ions component. A decrease in the pore size of the cathode also shows an increase in the cathode erosion rate.

Experimental data on the dependence of titanium erosion rate on cathode temperature, Ar, and N/sub 2/ pressure from vacuum to 1 torr and arc motion are presented. Erosion rate is found to decrease with conditions that promote cathode poisoning/contamination. Higher cathode temperatures result in enhanced nitriding (poisoning), leading to a reduced erosion rate. A critical nitrogen pressure (0.001 torr) exists where a sharp drop in erosion is measured. Steered arcs show lower erosion rate values of 38 and 15 /spl mu/g/C for argon and nitrogen, when compared to random arc values of 45 and 35 /spl mu/g/C. Erosion rate studies on TiN-coated cathodes show a low value of around 22 /spl mu/g/C.

Measurements of the wall current collected from the expanding cathodic arc plasma are reported as a function of angle to cathode plane normal, with hydrogen gas pressure in the arc chamber varying between 10/sup -4/ torr and 10 torr and distance R to the cathode varying between 10 and 80 mm. The angular distribution of the collected current is not seen to be affected by the gas pressure in the chamber. This flux decreases with distance to the cathode as R/sup -(1.4)/ in vacuum and R/sup -4/ in the 1-10 torr pressure domain. Flux intensity evaluated as a function of NR, where N is the number density of hydrogen in the chamber, leads to an evaluation of an effective collision cross section, sigma /sub e/, of 0.2*10/sup -16/ cm/sup 2/ for the C/sup +/-ion/H/sub 2/ molecule interaction. This value can be attributed to the reaction C/sup +/+H/sub 2/ to CH/sup +/+H and agrees with a velocity of the C/sup +/ ions following a shock wave interaction model. >

Results on the vacuum arc deposition (VAD) of thin carbon films on silicon, tungsten, and nickel are presented for different hydrogen pressures in the arc chamber. Raman and Auger electron spectroscopy and SEM analysis show that a large variety of film structures can be obtained. Highly oriented pyrolytic graphite is formed at high hydrogen pressure and low substrate temperature. Vacuum deposition resulted in a diamondlike structure. Evidence of the presence of multicrystalline spheres of diamond in some films is also presented. >

Observations of the cathodic copper plasma expansion at low pressures of He, Ar, and SF/sub 6/ showed that, for background gas mass densities of rho /sub g/=1 to 4*10/sup -4/ kg/m/sup 3/ and higher, the plasma and gas are separated into two volumes. A shock wave acts as a boundary between the two volumes. The boundary attains a stationary position once its expansion velocity decreases to the velocity of sound in the background gas. This position corresponds to a distance R/sub c/ to the cathode that agrees with a snowplow expansion model, giving R/sub c//sup beta /f=E/sub r/, where f is a function of the arc current and background gas characteristics, E/sub r/ is the erosion rate of the cathode, and beta varies between 2.1 and 2.5. The interaction model is based on kinetic energy exchanges between two gas-like volumes without other energy losses. A maximum pressure limit for vacuum arc deposition is set for rho /sub g//I=2 to 9*10/sup -6/ kg/m/sup 3/ A. >

Vacuum arcs are generated on dense graphite materials with enhanced local plasma density leading to strong improvement in arc mobility and stability. Observations of cross sections of the arc canyon erosion trace reveals erosion craters covered with a porous columnar film having a diamond structure. Scanning electron microscope (SEM) images of this growth structure are presented at various resolutions.

The distribution of the current collected by a multiring anode is measured in a diffuse vacuum arc for several values of the arc current and the electrode separation. It is found that the distribution widens as the electrode separation increases. Comparison with previous measurements suggests an influence of the electrode geometry on the results.

Experimental data on the erosion rates of a copper cathode in He, Ar, and SF6 from 10-6 to 760 torr are presented. The work performed by the cathodic-erosion plasma expanding against the gas is found experimentally to be constant, the volume of the expanding plasma cloud being linked to the gas pressure by the relation R3p = constant. These data agree with a redeposition model based on condensation of the metallic vapor produced by the arc on the cathode surface. The redeposited mass on the cathode is found to be proportional to the cube root of the gas mass density.