Learning Objectives

To provide a plasma ignition without any additional igniters as well as a stable and continuous plasma operation a high quality coaxial resonator with an adjustable resonance frequency was combined with a low quality cylindrical resonator to a microwave plasma torch. The schematic of this plasma torch is presented in Figure 3. The plasma is confined into a microwave-transparent tube, here a quartz tube. This tube can act as a reaction chamber for volume plasma processes or a plasma brush for surface treatments can be formed by an orifice. The microwave power is guided via a rectangular waveguide from the magnetron to the microwave plasma torch. Different kinds of gases can be supplied via either the tangential gas supply or axially through the metallic nozzle of the coaxial resonator. The microwave plasma torch is equipped with a frontal slit, so that the plasma inside the torch and the ignition can be investigated in detail.

To guarantee an ignition of the plasma solely by the supplied microwave power a high electrical field of about 3 to 6 MV/m is needed. To get a better understanding of the electrical field distribution, simulations of the electric field distribution as well as Eigenmode analysis with the commercially available simulation software COMSOL Multiphysics were conducted. Modeling and simulations of electrical field distributions of atmospheric pressure microwave plasma torches provided already detailed insights and led to further developments and improvements regarding for example their ignition or operation behavior ^19–22.

The electrical field distribution of the coaxial mode as well as of the common cylindrical E₀₁₀ mode is depicted in Figure 4a and 4b, respectively. The electrical field is displayed in arbitrary units, since the electrical field in the coaxial resonator is many times higher compared to the electrical field in the cylindrical resonator. It can be seen that a high electrical field at the nozzle tip is reached with the coaxial resonator and the highest electrical field of the cylindrical resonator is in the center of the cylindrical resonator. The resonance frequency of the coaxial resonator can be varied by the position of the metallic nozzle. The simulation results for the resonance frequencies for different nozzle positions for a microwave plasma torch with a cylindrical resonator with a radius of 0.05 m and a height of 0.048 m are shown in the diagram in Figure 4C. It can be seen that the resonance frequency of the cylindrical mode is not affected by the position of the metallic nozzle. However, the resonance frequency of the coaxial mode is dependent on the nozzle position and decreases when the metallic nozzle is moved upwards into the cylindrical resonator.

To reach the required high electrical field in the coaxial resonator this resonance-frequency-adjustable coaxial resonator exhibits a high quality and a sharp and narrow resonance curve. However, a sharp and narrow resonance curve requires that the resonance frequency of the coaxial resonator matches perfectly the frequency of the supplied microwave. Since usually magnetrons do not emit the microwave at their nominal frequency and since the frequency of the microwave is dependent on the output power of the microwave, the frequency dependence of the magnetron has to be measured by means of a directional coupler and a spectrum analyzer. The experimental set-up to measure the frequency dependence of the magnetron with a spectrum analyzer is schematically given in Figure 1a. The measured frequency dependence of the utilized magnetron is shown in the diagram in Figure 1B. The center frequency was set to 2.45 GHz and the video bandwidth was 200 MHz. It can be seen that at a power of 200 W (10% of the maximum output power of the magnetron) the frequency of the microwave is at 2.44638 GHz and increases when the microwave power is increased. At the maximum output power of 2 kW the microwave frequency reaches a value of 2.45213 GHz.

The resonance frequency of the microwave plasma torch can be measured with a network analyzer and since the nozzle is movable the resonance frequency of the coaxial resonator can be adjusted. To do so, the microwave plasma torch assembly has to be connected to a network analyzer via a rectangular-to-coaxial waveguide transition as shown in the schematic in Figure 2A. By measuring the S11 parameter of the microwave plasma torch assembly the resonance frequency can be determined. The S11 parameter represents the ratio of the input power to the reflected power in dependence of the frequency. When a resonance is reached, an electrical field establishes in the resonator structure leading to a reduced reflected microwave power. However, the field strength inside the cavity is directly related to the fixed wave amplitude of the microwave provided by the network analyzer. A dip appears in the S11 spectrum which corresponds to the resonance frequency. A typical measurement of the S11 parameter is depicted in Figure 2B. Here a resonance is observed at a frequency of 2.846 GHz. By moving the metallic nozzle up and down, the resonance frequency of the coaxial resonator can be varied as the simulations depicted in Figure 4C showed. This dependence of the resonance frequency of the coaxial resonator on the metallic nozzle position can be measured by means of the S11 parameter. A measurement of the resonance frequency in dependence of the nozzle position and the appertaining simulation results are presented in the diagram in Figure 2C. This diagram shows that there is a good agreement between the simulation results and the measured values of the resonance frequency. The very small shift of the two curves can be explained by very small deviations of the geometry or dimension of the manufactured nozzle compared to the one used for the simulations. To adjust the resonance frequency of the coaxial resonator to the frequency of the supplied microwave, the metallic nozzle has to be iteratively moved up and down until the dip in the S11 parameter is located at the measured microwave frequency. Then the metallic nozzle has to be locked and the forward power can be maximized by iteratively adjusting the stubs of the three stub tuner so that the S11 parameter dip reaches its maximum depth. The high quality of the resonator and the maximized forward power lead to fewer microwave reflections and a high electrical field is established in the resonator which is why a deep dip in the S11 parameter results.

After the microwave plasma torch assembly is mounted to the magnetron and the gas supply is connected, the plasma torch can be ignited and operated. The ignition of the plasma can be investigated best by observing the ignition with a high speed camera. The ignition of the plasma was recorded at 1,000 fps. The presented plasma ignition was conducted at a microwave power of 1 kW and a supplied gas flow of 15 slm air. Images of each phase of the ignition are summarized in Figure 5. The image in Figure 5A shows the view from above, looking down on the nozzle at an angle through the diagnostic slit at the front of the inoperational plasma torch. The bottom of the cylindrical resonator is in the front. In the mid plane you can see the beginning of the coaxial resonator. The tip of the nozzle can also be seen. The bottom of the cylindrical resonator is located in the background again. Since the focus is on the nozzle tip, the bottom of the cylindrical resonator is somewhat blurry. The other images show the phases of the plasma ignition. When the microwave power is turned on at t = 0 msec, the plasma ignites somewhere in the coaxial resonator as can be seen in Figure 5B. Then, during 64 msec, the plasma winds up the metallic nozzle to its tip and then burns straight at the nozzle tip in the coaxial mode as the Figure 5C para 5E show. The intensity of the plasma grows for the following 692 msec as it is shown in Figure 5F. Then, due to the shift of the resonance frequency caused by the burning plasma in the coaxial resonator 1 msec later, the plasma starts to break away from the nozzle tip as shown in Figure 5G and 5H. The complete break away of the plasma from the nozzle tip is reached after 58 msec as depicted in Figure 5I. The plasma is now burning freely above the metallic nozzle in the cylindrical mode. During the last second, the three stub tuner is readjusted to maximize the forward microwave power. This leads to an increase of the plasma as the image in Figure 5J shows. However, the plasma is still burning freely above the nozzle tip with no contact to it. Due to the low quality of the cylindrical resonator the plasma can be operated continuously and stably in this cylindrical resonator mode.

The dimension of the plasma depends on the supplied microwave power and the gas flow. Photos of the plasma for microwave powers of 1 and 2 kW and gas flows of 10, 30 and 70 slm are presented in Figure 6. The resonator with its diagnostic slit at its front is located in the lower part of the photos. The plasma is confined into a quartz tube within and above the cylindrical resonator. UV light couples into the quartz tube which is why the quartz tube exhibits a bluish glowing. It can be seen that the dimensions — radial and also the axial extension — of the plasma increase with an increase of the supplied microwave power while an increase of the gas flow leads to a smaller plasma flame. However, measurements of the gas and electron temperature show the maximum temperatures of T_g = 3,600 K and electron temperature T_e = 5,800 K are independent of the outer parameters, supplied microwave power and gas flows, as well as of the plasma volume ¹⁹. The temperatures were obtained by means of optical emission spectroscopy. The A²Σ⁺ – X²Π_γ-transition of the free OH radical was used for the determination of the gas temperature while a Boltzmann-plot of atomic oxygen lines was conducted for the estimate of the electron temperature. A detailed description on how the temperatures have been measured and the complete temperature distributions can be found in References 23 and 24.

To treat surfaces in the afterglow of the plasma, the plasma can be shaped with different kinds of orifices. Figure 7 depicts photos of differently shaped plasmas. The layout is similar to the photos of the plasma confined to a long quartz tube: the cylindrical resonator is at the bottom of the image; its diagnostic slit illuminated by the plasma. Differently shaped plasmas can be seen burning above the top-opening. On the photo in Figure 7A the confining quartz tube does not extend outside of the resonator. The plasma can burn freely above the resonator. An extended plasma brush can be formed with as slit orifice as depicted in Figure 7B. A plasma needle can be achieved by using an orifice with a hole in its center. This is shown in Figure 7C. Very small and smooth afterglow plasmas are formed by orifices which have a narrow slit or some small holes arranged in a circle as the photos in Figure 7D and 7E show.

Figure 1. Measurement of the magnetron. The schematic in (A) shows how the frequency dependence of a magnetron of the microwave output power can be measured by means of a spectrum analyzer. The frequency dependence of the used magnetron of the output power is depicted in (B). Please click here to view a larger version of this figure.

Figure 2. Measurement of the resonance frequency. The setup for the measurement and adjustment of the resonance frequency of the microwave plasma torch by means of a network analyzer is given in (A). (B) shows a typical measurement of the S11 parameter. The dip in the S11 parameter reflects the resonance frequency of the microwave plasma torch. The measured dependence of the resonance frequency on the metallic nozzle position and the results of the numerical simulations are summarized in c). Please click here to view a larger version of this figure.

Figure 3. Plasma torch setup. Schematic of the setup of the atmospheric microwave plasma torch. Please click here to view a larger version of this figure.

Figure 4. Coaxial and cylindrical mode. The distribution of the electrical field strength is depicted in (A) and (B). (A) shows the distribution for the coaxial mode while (B) shows the one for the cylindrical mode. The diagram in (C) shows the dependence of the resonance frequency of both the coaxial and the cylindrical mode on the position of the metallic nozzle in the plasma torch. The resonator has a diameter of 0.05 m and a height of 0.0482 m. Please click here to view a larger version of this figure.

Figure 5. Ignition of the plasma. Images of each phase of the ignition of the plasma recorded by a high speed camera at 1,000 fps and at a microwave power of 1 kW and a gas flow of 15 slm air. (A) View from above, looking down on the nozzle at an angle through the diagnostic slit at the front of the inoperational plasma torch. (B) Ignition of the plasma in the coaxial resonator. (C) – (E) Winding up of the plasma to the tip of the metallic nozzle until it burns in the coaxial mode. (F) The plasma increases. (G) – (I) The plasma breaks away from the metallic nozzle and burns freely above the nozzle tip in the cylindrical mode. (J) The plasma increases due to the readjustment of the three stub tuner to maximize the forward power. Please click here to view a larger version of this figure.

Figure 6. Photos of an air plasma for different supplied gas flows of 10, 30 and 70 slm and microwave powers of 1 and 2 kW. The extent of the plasma depends on the supplied microwave power and gas flow and it increases with an increase of the microwave power and a decrease of the gas flow. Please click here to view a larger version of this figure.

Figure 7. Different orifices. By using differently shaped orifices the plasma can be formed. (A) The confining quartz tube does not extend outside of the resonator and the plasma can burn freely above the resonator. (B) The plasma is shaped into a brush with a slit orifice. (C) A plasma needle is formed by a hole orifice. (D) A very smooth plasma brush can be achieved by using an orifice with a narrow slit and (E) a smooth plasma area is formed by an orifice with some small holes arranged in a circle. Please click here to view a larger version of this figure.