Self-bias

Non-thennal plasmas in contact with insulating walls (substrate) have an important property. The plasma with the hot electrons is positively charged relative to the wall (self-bias). A sheath with a positive space charge and an electric field is fonned between the wall and the plasma. The hot electrons travel faster to the wall than the heavy... 

FIG. 4. Schematic representation (a) of a parallel-plate, capacitively coupled RF-discharge reactor, with unequal-size electrodes. The potential distribution (b) shows the positive plasma potential Vp and the negative dc self-bias voltage... 

In most systems the substrate electrodes are larger than the powered electrodes. This asymmetric configuration results in a negative dc self-bias voltage Vdc on the powered electrode. Without that, the difference in electrode areas would result in a net electron current per RF period [134, 169]. It has been shown that the ratio of the time-averaged potential drops for the sheaths at the grounded (V g) and the powered electrode (Vsp) are inversely proportional to a power of the ratio of the areas of the two electrodes (Ag, Ap) [134, 170-172] ... 

Meijer and Goedheer [174] have developed a ID model with which, among other things, the dc self-bias and the sheath voltages can be calculated. It... 

FIG. 9. Relation between dc self-bias voltage Vj,. and applied RF voltage for different gases (Ar. H2. SiHa) and gas mixtures 0.1 < ISiHal/tlHTl-FlSiHal) < 0.9. Note that the slope is independent of the gas used. 

Goedheer et al. [148, 149] have attempted to formulate scaling laws. They varied the power P, pressure p, and frequency co and measured the dc self-bias Vdc. From their experimental results (frequency range 60-100 MHz, pressure range 20-60 Pa) a scaling law was formulated [148] ... 

Comparing published data on the relation between the dc self-bias and RF voltage between Rauf and Kushner [180] and our group [151] shows that the proportionality constant (slope) in the GECRC is nearly twice the one observed in the ASTER reactor. In other words, the ASTER reactor is less asymmetric than the... 

FIG. 10. Dc self-bias voltage as a function of electrode distance for two pressures and six power levels at 40 MHz. 

ID). This is due to the radial losses in the 2D model. The difference in ion fluxes to the grounded and powered electrode results in a dc self-bias voltage of —13.96 V, whereas the experimental Vdc is —22 V. 

FIG. 44. Plasma parameters as deduced from the lEDs and material properties as a function of power delivered to the SiHa-Ar discharge at an excitation frequency of 50 MHz and a pressure of 0.4 mbar (a) the plasma potential Vp (circles) and dc self bias (triangles), (b) the sheath thickness d, (c) the maximum ion flux r ax. (d) the growth rate r,/. (e) the microstructure parameter R. and (f) the refractive index ni ev- (Compiled from E. A. G. Hamers. Ph.D. Thesis, Universiteit Utrecht, Utrecht, the Netherlands. 1998.)... 

The plasma potential is about 25 V (Figure 63a). This value of the plasma potential is typical for the silane plasmas in the asymmetric capacitively coupled RF reactors as used in the ASTER deposition system, and is also commonly found in argon or hydrogen plasmas [170, 280, 327]. From the considerable decrease of the dc self-bias with increasing frequency (Figure 63a) it is inferred that the... 

The value of the self-bias potential was reported to strongly depend on the RF power and deposition pressure, following a (VF/p) /- dependence [30]. The... 

Since the main parameter influencing diamond-like carbon film structure is the energy of bombarding ions, it is expected that the same happens with a-C H films. In fact, it was found that in RFPECVD deposition of a-C H films, the variation of substrate self-bias results in strong changes of film growth, composition, structure, and properties. 

The optical gap of a-C H films was found to continuously decrease with increasing self-bias [42]. The gap shrinking was found to be strongly correlated to the variation of Raman spectra that is related to the increase of the graphitic clusters present in the a-C H films. Accordingly, the electrical resistivity of C H films was found to strongly decrease with substrate bias [43]. 

In Table I are shown representative data concerning plasma deposition of a-C(N) H films, produced by different methods and under different gaseous mixtures and deposition parameters. In this table are displayed data on chemical composition (maximum N content, and range of variation of H content), deposition details (deposition pressure, self-bias, and atmosphere composition), and the method used for chemical composition determination. 

FIG. 12. Growth rate of a-C(N) H films deposited by ECRRF as a function of RF self-bias, for several N2 partial pressures. (Reproduced from (62. )... 

This picture was found to be consistent with the comparison of Raman spectra and optical gap of a-C H films deposited by RFPECVD, with increasing self-bias [41], It was found that both, the band intensity ratio /d//g and the peak position (DQ increased upon increasing self-bias potential. At the same time, a decrease on the optical gap was observed. Within the cluster model for the electronic structure of amorphous carbon films, a decrease in the optical gap is expected for the increase of the sp -carbon clusters size. From this, one can admit that in a-C H films, the modifications mentioned earlier in the Raman spectra really correspond to an increase in the graphitic clusters size. 

The vacuum chambers were pumped down by means of an oil diffusion pump backed by a rotary vane vacuum pump. The base pressure achieved was 1 x 10 5 Torr (1.33 x 10 Pa). High-purity argon gas was bled into the chamber, the high-vacuum valve throttled, and the chamber pressure maintained as close as possible to 2 x 10 2 Torr (2.66Pa). For some of the experiments, the dc self-bias on the magnetron electrode was also measured. 

---