A dual-cathode arc plasma source was combined with a computer-controlled bias amplifier such as to synchronize substrate bias with the pulsed production of plasma. In this way, bias can be applied in a material-selective way. The principle has been applied to the synthesis metal-doped diamond-like carbon films, where the bias was applied and adjusted when the carbon plasma was condensing, and the substrate was at ground when the metal was incorporated. In doing so, excessive sputtering by too-energetic metal ions can be avoided while the sp 3 /sp 2 ratio can be adjusted. It is shown that the resistivity of the film can be tuned by this species-selective bias. The principle can be extended to multiple-material plasma sources and complex materials.

Metal-containing tetrahedral amorphous carbon films were produced by dual filtered cathodic vacuum arc plasma sources operated in sequentially pulsed mode.Negatively pulsed bias was applied to the substrate when carbon plasma was generated, whereas it was absent when the molybdenum plasma was present. Film thickness was measured after deposition by profilometry. Glass slides with silver pads were used as substrates for the measurement of the sheet resistance. The microstructure and composition of the films were characterized by Raman spectroscopy and Rutherford backscattering, respectively. It was found that the electrical resistivity decreases with an increase of the Mo content, which can be ascribed to an increase of the sp 2 content and an increase of the sp 2 cluster size.

Molybdenum-containing amorphous carbon (a-C:Mo) thin films were prepared using a dual-cathode filtered cathodic arc plasma source with a molybdenum and a carbon (graphite) cathode. The Mo content in the films was controlled by varying the deposition pulse ratio of Mo and C. Film sheet resistance was measured in situ at process temperature, which was close to room temperature, as well as ex situ as a function of temperature (300-515 K) in ambient air. Film resistivity and electrical activation energy were derived for different Mo and C ratios and substrate bias. Film thickness was in the range 8-28 nm. Film resistivity varied from 3.55×10 -4 Ω m to 2.27×10 -6 Ω m when the Mo/C pulse ratio was increased from 0.05 to 0.4, with no substrate bias applied. With carbon-selective bias, the film resistivity was in the range of 4.59×10 -2 and 4.05 Ω m at a Mo/C pulse ratio of 0.05.The electrical activation energy decreased from 3.80×10 -2 to 3.36×10 -4 eV when the Mo/C pulse ratio was increased in the absence of bias, and from 0.19 to 0.14 eV for carbonselective bias conditions. The resistivity of the film shifts systematically with the amounts of Mo and upon application of substrate bias voltage. The intensity ratio of the Raman D-peak and G-peak (I D /I G ) correlated with the pre-exponential factor (σ 0 ) which included charge carrier density and density of states.

No abstract