Global (volume-averaged) models of high-density, low-pressure electropositive and electronegative discharges are described both for continuous wave (CW) and for pulsed-power excitation. Argon and chlorine discharges are treated. The particle and energy balance equations are applied to determine the charged particle and neutral dynamics. For argon just after the power has been turned on, the analysis shows an initial very sharp rise in electron temperature T e , followed by a decay of T e and an increase in the electron density n e to steady state values during the pulse-on time. Just after the power has been turned off, T e decays rapidly and n e decays more slowly. The time-average n e can be considerably higher than that for CW discharges for the same time-average power. For chlorine, a CW discharge is highly dissociated and the negative ion density n Cl − is lower than n e . For a pulsed discharge, the initial rise and subsequent decay of n Cl − just after the power has been turned off are determined analytically. A pulsed discharge can have the same neutral radical (Cl) flux to the walls for a reduced average power. The analytical models are compared to more complete global model simulations and to experimental observations. We find that global models can provide considerable insight into the discharge dynamics.

The behavior of argon plasmas driven by time modulated power in ‘‘high density’’ plasma reactors is investigated using a global model. The time evolution of the electron temperature and the plasma density is calculated by solving the particle and energy balance equations. In the first stage of power application during the ‘‘on’’ time, the electron temperature rapidly increases above the steady state value. In this region, charged particles accumulate in the plasma due to the relatively higher power applied than for the continuous wave (cw) case. In the first stage of the ‘‘off’’ time, the electron temperature drops quickly, yielding a smaller particle loss (Bohm) velocity. These effects give rise to higher time-average plasma densities than for the cw plasma driven by the same average power. The highest average plasma density obtained was more than twice the density of the cw plasma for a duty ratio of 25%. Even higher plasma densities were obtained for shorter duty ratios. The possibility of controlling chemical reactions in the plasma by changing the modulation period is also shown.

Properties of chlorine plasmas generated in a high-density, low pressure ( ∼5 mTorr) plasma source using a pulsed-power modulation method are calculated by a spatially averaged (global) model. Both time resolved and time average plasma properties are discussed. The recombination coefficient for chlorine atoms at the wall plays an important role in determining the electronegativity of the plasma, since the dissociation fraction of the gas greatly affects the generation of negative ions. Even for such discharges, which have substantial electronegativity in the bulk plasma, it is only possible to extract negative ions at the walls by using modulations having longer “off” times than a few tens of microseconds. This is the time scale required for the sheath voltage, which confines negative ions within the bulk plasma, to collapse, due to thermalization and loss of electrons.

The behavior of pulsed-power (square wave) modulated argon plasmas generated by an inductively coupled plasma (ICP) source is experimentally investigated. The apparatus is an ICP source with a flat coil geometry equipped with a cylindrical Langmuir probe on the axis of the cylindrical chamber. The evolution of the plasma density is determined from the wave forms of ion saturation currents. The rise and fall of the density had a time scale of a few tens of microseconds. The time average plasma density is also measured as a function of pulse frequency and duty ratio, holding the average absorbed power constant. When the plasma is modulated, the density is larger than that for a continuous wave excitation of the same average power. Larger densities are obtained for smaller duty ratios. The density increases monotonically as the period is decreased down to 100 μs. This agrees qualitatively with the modeling result, which accounts for the higher density by the difference of time scales for the generation and the loss of charged particles. The results for electron temperature measurements also show good agreement with the model quantitatively during the pulse ‘‘on’’ times. If the period is long enough, it is found that the electron temperature rises abruptly at the initial stage of power application, in agreement with the calculation.

No abstract