Phase Resolved Optical Emission Spectroscopic Measurements and Results of Pulsed Capacitively Coupled Radiofrequency Discharges Through Argon and Tetrafluoromethane-Argon
The experimental results and discussion presented in this dissertation contribute new findings for pulsed capacitively coupled radiofrequency (rf) discharges at low pressures during the re-ignition of the plasma using the phase resolved optical emission spectroscopy, or PROES, technique. The gases used in the experiments were Argon (Ar) and a 50-50 Tetrafluoromethane-Argon (CF₄-Ar). The start of the experiments used optical emission intensity (OEI) measurements to survey the re-ignition process to find and analyze transients with respect to the conditions of the plasma environment. Given the current understanding of the electropositive Ar pulsed plasmas, their OEI results exhibited some expected intensity profiles which have already been measured and modeled given measurements with time resolutions that are larger than the rf period. For instance, the OEI measurements of Ar 10 kHz pulsed plasma, exhibited little transients as the plasma reformed during re-ignition. This was due to the plasma environment not changing substantially for the 50 μs off times and only the emission intensity was seen to increase as the rf voltage amplitude grew. The relative intensity spatial profile not changing during the re-ignition was found to be consistent with the electron heating dynamics also not changing. This relationship was found as a result of the PROES measurements exhibiting the same electron heating profiles within different rf periods throughout the re-ignition. The PROES measurements were compared to PROES measurements made by other groups in continuous wave Ar plasmas. The electron heating mechanism that was found for the Ar 10 kHz PROES measurements was consistent with heating caused by sheath expansion, known as stochastic or α heating. Some profiles like the average sheath width and intensity profiles were also found in the OEI and were seen to translate to the data measured in the PROES. They were seen to be results of the plasma environment reacting to the electrical nature of the charging of the electrodes (rf voltage and DC self-bias) but were not seen to affect the electron heating mechanisms which are of main interest. (These transients and the ones from the other plasmas that do not directly affect the electron heating mechanisms are important and will be discussed in detail in the dissertation.) The next pulsing phenomenon expected to be seen in the OEI surveying measurements was an intensity overshoot as the Ar pulsing frequency was changed to 100 Hz. It was seen that an intensity overshoot occurred but so did other unexpected optical spatial profiles. The intensity overshoot which is consistent with an electron temperature, Tₑ, spike has been modeled and is understood to be a result of an electron density, nₑ, becoming so small that the Tₑ will rapidly rise and will gradually drop as the nₑ grows larger. The current theory does not, however, mention the mechanisms that allows the Tₑ to rise. This is where the PROES results begin to present new findings and further our understanding on what is occurring during re-ignition in pulsed capacitively coupled plasmas. After analyzing the PROES measurements for re-ignition of the Ar 100 Hz plasma, it was found that the OEI and the optical characteristic profiles changing were indications of electron heating-mode changes. The PROES results at the steady state times exhibited heating profiles within the rf period that were consistent with α heating, but when reviewing the PROES results for the intensity overshoot times, more excitations were transpiring for more of the rf period. Not only were there additional excitation events but there was also evidence that the additional excitations were increasing the effectiveness of the dominant α heating mode. The effectiveness of the α heating mode was seen to fade as the presence of the additional heating diminished. Due to this heating mode not being seen in continuous wave (CW) Ar plasmas at this operating conditions (pressure, powers, etc.), literature on other gases were reviewed and a similar heating profile was found. This heating mode is referred to as drift ambipolar (DA) heating. With the Ar pulsed data as a basis, the pulsing frequency effects of CF₄-Ar plasmas were studied next. (Due to the complexity of these plasmas, there were no pulsing models for this gas chemistry to be found to make predictions.) The first pulsing frequency used for CF₄-Ar plasmas was 10 kHz. Though the Ar 10 kHz exhibited no strong transient indications for changes of electron heating dynamics, the CF₄-Ar 10 kHz did with the same intensity overshoot and relative intensity spatial profile changes that were seen in the Ar 100 Hz. It was found that the addition of the CF₄ molecule added electron loss reactions which substantially changed the plasma environment during the 50 μs off times. The PROES measurements of the re-ignition showed that the α heating mode still dominated the electron dynamics but the electronegative nature of the plasma created was conducive for the presence of DA heating mode to appear briefly. This was not too surprising since PROES measurements of pure CW CF₄ plasmas are seen to be dominated by the DA heating mode where most of the excitation occurs near the electrode. This is consistent with the additional heating within the rf period in the PROES measurements during the intensity overshoot time. In order to further explore pulsing frequency effects on the CF₄-Ar plasmas, the pulsing frequency was changed to 1 kHz. Even though the frequency was only lowered by a factor of 10, the effects measured in the OEI survey were unlike anything measured in the previous experiments. The OEI did exhibit an intensity overshoot but with a different optical characteristic profiles occurring earlier in the re-ignition. The OEI also exhibited a weak and nearly instantaneous intensity streak across the viewable discharge gap. The thickness of the maximum intensity was also seen to change throughout the re-ignition; as well as, some intensity “streaks” were seen during the first 5 μs. Unfortunately, the intensity of the emission from the Ar sensing atom dropped substantially for the PROES measurements which made analysis difficult but some claims could still be made off the results. First, the weak intensity signal was an indication of the CF₄ (and its daughter species) reactions dominating the mechanisms by which the electrons were being heated. This is demonstrated by the noisy heating profiles from the PROES results where the excitation rate profiles were unlike the profiles seen in the other three pulsed plasmas. The heating profiles associated with α heating exhibited shallow heating depth (∼ millimeter). This may be indicative of a sheath that is not well formed with a less distinctive edge profile. The PROES also exhibits faint but substantial heating farther into the plasma bulk for the majority of the rf period. This type of heating was typically seen only when the DA mode was occurring in the rf period. This is consistent with electron being energized directly from the rf voltage which is not being effectively screened by the charged population. Therefore, if the sheath is not as well formed it is conceivable that it will not heat and not screen electric fields as effectively. During the process of making experimental measurements, a need was presented to acquire more points with an accurate data representation. Due to the measurement tools limitations already being met, a chapter is also presented for a technique that can temporally deconvolve measurements when the timing resolution of a tool is greater than the time steps between each data point.