Radio frequency (RF) inductively coupled plasmas (ICP) when used in a pulsed mode (PICP) have tremendous advantages in comparison to the continuous wave mode such as, less damage, controllable radicals and ion energies.
The emphasis has been placed on the control of the plasma parameters to better understand the fundamentals of PICPs. However, an ICP is a nonlinear system which includes the RF power system (source, matching network, antenna and plasma). When the pulse is initially switched on there is a large mismatch between the source and the load due to the lower plasma density, which results in high reflected power. The temporal evolution of the input power, in the initial stage of the pulse, is therefore an important issue, as it may affect the electron density.
The experiment is shown in Figure 1, comprised of a planar-coil ICP reactor, a 13.56 MHz RF power source applied through a matching network and a grounded electrode. An Impedans Octiv Suite VI probe is fixed between the power source and the matching network to provide the time-resolved power output of the generator. The power is then calculated as P = V I Cos(θ), where V, I and θ are the RF voltage, RF current and the phase between them as measured by the VI probe. An Impedans Langmuir probe is used to measure the temporal evolution of the electron density.
Figure 1. Schematic diagram of the experimental setup.
The probe tip is made of a tungsten wire 5 mm in length and 0.2 mm in diameter. The probe is inserted into the chamber through a flange 3 cm above the grounded electrode. The Langmuir probe measurements are validated in a Ar/CF4 plasmas through comparison with a hairpin probe.
The temporal evolution of the input power along with the corresponding electron density is shown in figure 2 for different powers in 10 mTorr Ar discharges with pulse frequency of 1 kHz and 50% duty cycle. For the first 4 μs the input power is almost zero, before rising to an initial peak after around 8 μs. It should also be noted that the evolution of the input power is independent of the set power or pressure (figure 3) for the initial 10 μs, implying this is characteristic of the generator rather than the plasma parameters.
Figure 2. Temporal evolution of the input power and the electron density for various powers in an Ar, 10 mTorr discharge.
A second peak in the RF power is then seen between 20-40 μs as the electron density does not rise, causing a mismatch between the RF power source and the load. However, the electron density increases rapidly causing the power to increase as the matching improves. The electron density peaks around 200 μs. A phenomenon called over-shoot occurs for high power discharges before settling to their steady-state.
When the power is turned off, the input power initially has a sharp peak before rapidly dropping to zero. The electron density responds much slower with values for all powers at the end of the afterglow being (1-2)1016m-3. Figure 3 shows the temporal evolution of the input power and electron density for a variety of pressures ranging from 1-80 mTorr in a 300 W Ar discharge with a pulse frequency and duty cycle of 1 kHz and 50% respectively. As the pressure is increased the second peak in the power is gradually delayed with all pressures having a plateau between the first two peaks, the width of which is dependent on the pressure but not on the power. This is due to the rapid increase in the electron density being delayed as the pressure is increased.
As the pressure is increased the over-shoot behavior is gradually suppressed. When the pressure is over 50 mTorr the plasma never has the time to reach the steady-state stage, meaning the electron density is always increasing until the power is switched off. High pressures also cause the electron density to decay more gradually during the afterglow period. Figure 4 shows the temporal evolution of the input power and the electron density for various powers in a 10 mTorr Ar/CF4 discharge with a pulse frequency of 1 kHz and 50% duty cycle.
Figure 3. Temporal evolution of the input power and the electron density for various pressures in an Ar 300 W discharge.
Figure 4. Temporal evolution of the input power and the electron density for various powers in an Ar/CF4 (90/10) 10 mTorr discharge.
The behavior of the input power is similar to the pure Ar discharge except that the second peak of the input power is delayed due to the wider plateau caused by the lower electron density in the initial stage of the pulse in the Ar/CF4 discharges.
The behavior of the electron density is also similar to the pure Ar case except for the magnitude of the over-shoot phenomenon which is approximately half of the value in the pure Ar discharges under the same experimental conditions.
Figure 5 shows the temporal evolution of the input power and the electron density for various pressures in a 300 W Ar/CF4 discharge with a pulse frequency of 1 kHz and 50% duty cycle.
Figure 5. Temporal evolution of the input power and the electron density for various pressures in an Ar/CF4 (90/10) 300 W discharge.
The input power characteristic is the same as in the Ar discharge, however the second peak gradually appears earlier as the pressure is increased. The over-shoot behavior of the electron density also gradually disappears as the pressure increases with a non-monotonic change in the steady state level.
In both Ar or Ar/CF4 discharges the first peak in the delivered RF power profile is determined by the characteristics of the generator, while the second peak is determined by the evolution of the electron density. The over-shoot behavior in the electron density appears at high powers and low pressures. When the power is switched off there is a small spike in the RF power before it rapidly drops to zero. The electron density exhibits a small peak due to the power spike before gradually decaying in the afterglow.
Gao, F. et al, “Complex transients of input power and electron density in pulsed inductively coupled discharges” Journal of Applied Physics. Doi: 10.1063/1.5114661Published in 2019.
Click Here to download the PDF