Generation of Extended Surface Barrier Discharge on Dielectric Surface -Electrical Properties-

Transcription

1 Takashima et al. 14 Generation of Extended Surface Barrier Discharge on Dielectric Surface -Electrical Properties- K. Takashima 1,2, N. Zouzou 2, E. Moreau 2, A. Mizuno 1, and G. Touchard 2 1 Department of Ecological Engineering, Toyohashi University of Technology, Japan 2 Laboratoire d Etudes Aérodynamiques, Poitiers University, France Abstract Dielectric barrier discharge (DBD) with a grounded electrode (sliding electrode) on the same surface as the high voltage electrode (three electrode DBD) was experimentally studied to generate homogeneous surface discharge. Application of AC high voltage generated stable and homogeneous surface plasma between high voltage and grounded electrode without arcing. Electrical properties of the three electrode DBD was experimentally studied. With the same peak voltage, lower DC component generated more stable but less homogeneous plasma sheet. Lower DC/AC ratio reduced the electric field to initiate sliding discharge. With sliding electrode, longer surface plasma was generated compared with normal DBD. The effect of the sliding electrode was more significant when the gap between high voltage and the sliding electrode was increased. Keywords Dielectric barrier discharge (DBD), Surface discharge, Sliding discharge, Plasma actuator I. INTRODUCTION Electroaerodynamic systems utilizing corona discharge or discharge plasma have been widely studied for many years [1-3]. One of the most important possible applications of them is aeronautic devices or electroaerodynamic actuators that modify the airflow profile within boundary layer on the surface in order to reduce drag and turbulence and to control detachment and reattachment of the airflow [4]. The most important advantage of these devices is that electric energy can be converted to kinetic energy without moving parts because they are based on ionic wind induced by discharge plasma. Various types of electroaerodynamic actuators with different discharge plasma have been developed. The simplest one uses DC surface corona discharge, which takes place under non-uniform DC electric field. Application of a DC high voltage between wire and plane electrodes [] or two parallel wire electrodes [6] with different diameters that were flush mounted on the surface of PMMA plate generated airflow from the thinnest electrode toward thickest one. The maximum induced velocity was about 3.m/s and significant drag reduction was observed in an external airflow up to 2m/s. However, because surface DC corona discharge is significantly affected by the electrical properties of the ambient gas and dielectrics in which electrodes are mounted, it is difficult to generate stable discharge under some conditions such as high humidity, surface degradation and dust deposition. A second kind of plasma used for electroaerodynamic actuators is DBD, in which a dielectric is sandwiched by two electrodes between Corresponding author: Kazunori TAKASHIMA address: takashima@eco.tut.ac.jp Received; November 27, 26, Revised; January 1, 27, Accepted; February 28, 27 which AC high voltage was applied [7]. Because sufficiently high voltage can be applied without arcing owing to the dielectric barrier, stable discharge is generated on the dielectric surface. With an asymmetric DBD actuator, in which two electrodes are displaced to the center so that DBD takes place only on one edge of the electrodes, ionic wind up to 8m/s was observed [8]. DBD is stable and less sensitive to the humidity compared with surface corona discharge but plasma area was smaller than that of the surface DC corona discharge. The third one is a sliding discharge with three electrode configuration. This discharge was originally developed for laser pumping with fast-rising pulsed voltage and long electrode separation [9-17]. The first reports on the electroaerodynamic actuators employing the similar electrode configuration were published recently. It was shown that application of AC high voltage with DC offset voltage to the similar three electrode system resulted in a stable plasma sheet between two electrodes on the surface of a dielectric material. Although this kind of discharge is generated under quite different electrical condition, it was also named sliding discharge [18] and is getting more and more widely used now [4]. Sliding discharge is mentioned in this sense hereafter. This type of actuator is less sensitive to the condition of dielectric surface and can be operated even in humid air but ionic wind was not significantly enhanced compared with DBD [18,19]. Therefore, further investigations are necessary to know the nature of the plasma and improve ionic wind generation characteristics of this system. In this paper, the above mentioned three electrode system was experimentally examined in order to study fundamental characteristics of the plasma to generate homogeneous and stable two-dimensional surface

2 1 International Journal of Plasma Environmental Science & Technology Vol.1, No.1, MARCH 27 PMMA tube φ 1mm 4mm Plasma High voltage electrode (a) V peak = 1kV (b) V peak = 17kV Copper sheet I sliding I total I DBD Sliding gap:g sliding Sliding electrode DBD electrode Fig. 1. Schematic illustration of the experimental setup. I DBD, I sliding, and I total represent DBD current, sliding current, and total current respectively. discharge plasma. It should be noted that homogeneous surface discharge plasma at atmospheric pressure is very important not only for the electroaerodynamic actuators but also for atmospheric pressure plasma surface process such as chemical vapor deposition (CVD), surface treatment, plasma induced catalytic reactions and so on. Fig. 2. Time averaged photographs of DBD and sliding discharge. Applied voltage: sine wave, 2kHz. Upper electrode: high voltage electrode, Lower electrode: grounded electrode. were placed at edge-to-edge position without overlapping. The high voltage power amplifier was driven by a function generator (HAMEG, HM813) to generate sine wave high voltage. High voltage up to 2kV was applied to the high voltage electrode with fixed frequency of 2kHz. DC high voltage was also superimposed to the sine wave in some experiments. Applied voltage was monitored with an oscilloscope (Tektronix, TDS314B) with a high voltage divider (LeCroy 1:1, 2kV). DBD current, sliding current and the total current were measured with a current transformer (Bergoz ACCT) of which bandwidth and accuracy are.16-16khz and some μa. Time averaged photographs of the discharge were taken by a digital camera (Canon EOS 3D) with exposure time of 1s. All the experiments were carried out in room air under atmospheric pressure at room temperature. The gas flow inside and outside of the tube was not controlled. II. EXPERIMENTAL Figure 1 shows a schematic illustration of the experimental setup. In this study, cylindrical reactors were employed in place of planer ones, which were commonly used for DBD and sliding discharge study, in order to eliminate edge effect and to generate uniform plasma along the electrodes. This is because uniform plasma is important for the analysis of fundamental discharge phenomena because it is difficult to separate the edge effect from the experimental data. It is also easier to observe radial profile of the plasma in this configuration. Plasma actuator used in this study consists of a PMMA tube and three copper electrodes. Outer diameter of the PMMA tube was 1mm and the thickness was 4mm. Two thin copper sheets with 4cm width were wrapped on the outer surface of the PMMA tube. One was connected to a high voltage power amplifier (Trek 2/2C) serving as a high voltage electrode. The other (sliding electrode) was grounded. These two electrodes were placed with a separation up to 3mm (sliding gap). A thin copper sheet with 8cm width was placed on the inner surface of the PMMA tube and grounded (DBD electrode). The high voltage electrode and the DBD electrode III. RESULTS AND DISCUSSION A. Generation of DBD and sliding discharge with three electrode configuration Figure 2 shows time averaged photographs at two different applied voltages. Separation between the high voltage electrode and the sliding electrode on the outer surface of the tube (sliding gap) was 2 mm in these cases. In this experiment, AC high voltage without DC component was used. When the applied voltage was 1kV peak, luminous part was observed only in the vicinity of the high voltage electrode showing a typical DBD. On the other hand, when the applied voltage was 17kV peak, plasma sheet looks observed between the high voltage and grounded electrode that was homogeneous both along the electrode and along the axis. Intense DBD were also observed near the high voltage electrode as well. The plasma was very stable and no arcing was observed although plasma sheet bridged the high voltage and the sliding electrode in this case. Figure 3 shows time evolution of total measured current and applied voltage corresponding to figure 2. Measured current includes fast component (pulse

3 Takashima et al Total discharge current DBD current Sliding current 2 Measured Current Current (ma) - Discharge Current 1-1 Current (ma) ,,2,4,6,8 1, (a) V peak = 1kV -8-2,,2,4,6,8 1, 1 2 Fig. 4. Time evolution of total discharge, DBD and sliding current. Measured Current 1 Current (ma) (a) V peak, AC = 7kV, V DC = 1kV (b) V peak, AC = 17kV - Discharge Current -1-2,,2,4,6,8 1, (b) V peak = 17kV Fig. 3. Time evolution of applied voltage and discharge current. Applied voltage: sine wave, 2kHz. (Capacitive current, which was not accompanied by the discharge, was subtracted from the measured current.) current) and slow component (capacitive current and pseudo continuous current). The capacitive current, which was estimated from the I-t curve with applied voltage low enough to generate any discharge, was subtracted from the measured current to obtain discharge current. All the currents hereafter were presented in this way. Although pulse component was affected due to the bandwidth of the current transformer, qualitative difference can be found in the discharge current. Amplitude of both pseudo continuous component and pulse component increased with applied voltage. DBD onset voltage decreased with increasing applied voltage, which can be accounted for the enhancement of electric field due to charges deposited on the dielectric surface in the previous half cycle. Current measurement with a shunt resistor showed that the number and the amplitude of current pulse increased with applied voltage (data not shown here). Figure 4 shows total, DBD, and sliding current when the applied voltage was 17kV peak. Sliding current was observed only in positive half cycles and -1 Fig.. Time averaged photographs of DBD and sliding discharge. Applied voltage: sine wave, 2kHz. Upper electrode: high voltage electrode, Lower electrode: grounded electrode. total current in negative half cycle was the same as DBD current. Widely spread luminous plasma sheet (figure 2b) was always accompanied by the sliding current. These results suggest that sliding discharge took place only when positive voltage was applied to the high voltage electrode while DBD took place with both positive and negative voltage. Sliding discharge can take place only on the outer surface of the tube while DBD is generated on both side of the tube because the third electrode was placed only on the outer surface. Positive surface discharge is generally more likely to develop compared with negative one [2, 21]. Therefore, in our case, discharge between high voltage electrode and the third electrode should start first when positive high voltage was applied to the high voltage electrode. As a result, sliding discharge takes place only in the positive half cycle under the intermediate applied voltage. This is why the sliding current was observed at one polarity only. Sliding current started after the DBD current reached the maximum and stopped when the applied voltage reached the peak value. It was also observed that DBD and sliding discharge stopped at the same time. These results suggest that not only the potential of the high voltage electrode but also DBD near the high voltage electrode play an important role on the generation of sliding discharge. These results agree well with the proposed mechanism of establishing sliding discharge in which charge deposited by DBD slides toward the third electrode [19].

4 17 International Journal of Plasma Environmental Science & Technology Vol.1, No.1, MARCH 27 2 DC Voltage_V DC Onset Voltage_V peak,ac +V DC 2 DC and Onset 1 1 Sliding OFF Sliding ON DBD OFF DBD ON AC voltage_v peak,ac (kv) Onset Voltage_V peak,ac + V DC (kv) 1 1 2mm 2mm 3mm AC voltage_v peak,ac (kv) g sliding =1mm Fig. 6. Onset voltage of the sliding discharge and its DC component for various AC component. Applied voltage: sine wave, 2kHz. Sliding gap: 2mm. Fig. 7. Onset voltage of the sliding discharge against AC component for various sliding gap. Applied voltage: sine wave, 2kHz. 7, B. Effect of DC bias voltage on generation of sliding discharge Results in the last section suggest the possibility of modifying or controlling the sliding discharge by changing AC voltage and thus changing DBD. In this section, effect of DC superimposed AC voltage application was examined in order to modify DBD and sliding discharge. Positive DC voltage was superimposed to AC high voltage in this experiment because negative DC offset does not result in surface plasma widely spread between the electrodes [19]. Sliding gap was 2 mm and applied voltage was 2kHz sine wave. All the measurements were carried out in the same way as above. Figure shows time averaged photographs of surface discharge generated by applying AC high voltage with and without DC offset voltage. In figure (a), the applied voltage was AC component of 7kV peak with +1kV DC component resulting in a sine high voltage with minimum voltage of +3kV and maximum of +17kV. Therefore, these two discharges had the same maximum voltage. Clear difference is the luminosity of the plasma near the high voltage electrode. More intense plasma was observed there without DC voltage superimposition. This is simply because AC component, which is directly associated with DBD, was larger in this case. On the other hand, luminosity of the plasma sheet excepting the vicinity of the high voltage electrode was not affected by the DC voltage superimposition. Onset voltage of sliding discharge was measured with various combinations of AC and DC components to investigate how DC/AC ratio affects the sliding discharge. In this study, we define the onset voltage of the sliding discharge as the minimum value of the peak voltage necessary to observe persistent current pulses on the sliding current. This can be another / Gap (kv/cm) Onset Voltage 6, 6,,, g sliding =1mm 2mm 2mm 3mm DC/AC Voltage ratio ( kv/kv ) Fig. 8. Minimum electric field necessary to establish the sliding discharge against DC/AC ratio. Applied voltage: sine wave, 2kHz. criterion of establishment of the sliding discharge than luminosity. Figure 6 shows the onset voltage of the sliding discharge and its DC component against various AC component. Sliding gap was 2 mm in this experiment. With AC component of 14.9kV or higher, sliding discharge was established without DC component showing that the onset voltage was 14.9kV and DC component was. With AC component of 4.kV, minimum requirement for establishing the sliding discharge was +12kV DC resulting in the onset voltage of 16.kV. With AC component lower than 4.kV, there was no DBD and arcing occurred before observing the sliding discharge as we increased the DC component. We can see that the onset voltage increased with decreasing AC component and with increasing DC component. In addition, with higher AC component, arcing was less likely to take place. These results suggest that DBD seeds and stabilizes the sliding discharge. In other words, DBD plays the role of a ionizer around the HV electrode, and then this space charge slides along the surface [4].

5 Takashima et al mm 8 With sliding electrode Without sliding electrode 2 2mm (a) Without sliding electrode (b) With sliding electrode Fig. 9. Time averaged photographs of surface discharge with and without sliding electrode. Applied voltage: sine wave, 2kHz, 17kV peak. Figure 7 shows onset voltage of the sliding discharge against AC component for various sliding gap. DBD started at nearly the same voltage independently of the sliding gap examined in this experiment showing that DBD was localized to the vicinity of high voltage electrode and the effect of the sliding electrode on DBD initiation can be neglected. The onset voltage of the sliding increased with the sliding gap. The onset voltage decreased with increasing AC component. It is notable that the effect of AC component on the onset voltage was more significant in the case of shorter sliding gap. This can be associated with the extension of DBD. Luminous DBD plasma area in point to plane configuration increased with applied voltage in a way that the radius of the luminous area is a linear function of the applied voltage [2]. Plasma area due to DBD works as an extended electrode resulting in a shortened gap between the high voltage and the sliding electrodes. The effect of this is more significant in shorter sliding gap. The correlation between plasma length and applied voltage will be discussed in more detail in another section. Figure 8 shows minimum electric field necessary to establish the sliding discharge against DC/AC ratio. The electric field was calculated by dividing the onset voltage (V peak, AC + V DC ) by the sliding gap. The minimum electric field increased with DC/AC ratio. If we assume that data with the largest DC/AC ratio in each curve should not be taken into account because discharge was very unstable, we can find clearer correlation: it increased with DC/AC ratio and then saturated. These results can be translated that DBD decreases the electric field necessary for the sliding discharge. C. Effect of sliding electrode on the length of plasma In this section, correlation between plasma length and applied voltage with and without the sliding electrode is presented. All the experiments in this section were carried out with AC high voltage without DC component. Figure 9 shows photographs of typical surface discharge plasma. In figure 9 (a), the sliding electrode was removed. In figure 9 (b), the sliding electrode was placed 2 mm apart from the high voltage electrode. Applied voltage was 2kHz sine wave with 17kV peak in both cases. As mentioned in DBD Current (ma) ,,2,4,6,8 1, Fig. 1. DBD current with and without sliding electrode. Applied voltage: sine wave, 2kHz, 17kV peak. above sections, homogeneous plasma sheet was observed between high voltage and sliding electrode (figure 9b). On the other hand, plasma was less homogeneous and less distributed without the sliding electrode (figure 9a). The length of the plasma in this case was 17mm, which was shorter than that of the case with the sliding electrode (2mm). Figure 1 shows DBD current with and without the sliding electrode. Experimental conditions were the same as figure 9. No significant difference was observed between them with our experimental setup. Correlation of plasma length and applied voltage with and without sliding electrode was examined experimentally. Plasma length in case without sliding electrode was determined by the length of luminous plasma sheet on photographs with various applied voltage. In case with sliding electrode, onset voltage of the sliding discharge was measured for various sliding gap. These plasma length identifications were not made under exactly the same criterion but the difference between optically measured onset voltage and that obtained from the sliding current measurement were small. The difference was less than.7kv, which was much smaller than the essential difference due to the sliding electrode. Therefore, the difference of the criterion can be ignored in the following discussion. Figure 11 shows the results. In both cases, plasma length increased with applied voltage. At the same applied voltage, three electrode configuration showed larger plasma length clearly showing that the sliding electrode enhances surface plasma. The difference in the plasma length increased with applied voltage and with the sliding gap showing that sliding electrode works better with longer gap. Length of DBD plasma follows a linear function of the applied voltage with a coefficient of 8kV/cm for AC voltage in point to plane configuration [2]. A linear correlation with a break was obtained for AC voltage in wire to plane configuration [21]. In this study, we obtained the 1-1

6 19 International Journal of Plasma Environmental Science & Technology Vol.1, No.1, MARCH 27 Applied With sliding electrode K T, DBD =9,1 kv.cm -1 V (kv)= K x L (cm) + Const T,, 1, 1, 2, 2, 3, 3, Plasma Length (cm) Fig. 11. Correlation of plasma length and applied voltage (V peak, AC ) with and without sliding electrode. Applied voltage: sine wave, 2kHz. similar linear correlations with a break at 6kV and 8kV for the sheet to sheet DBD and sliding discharge respectively. The coefficients above the break point were 9.1kV/cm and 6.4kV/cm for the BDB and sliding discharge respectively. These results agreed very well with those in the references[2,21]. It should be noted that the coefficient for the three electrode configuration was significantly smaller than that of DBD. This result clearly shows that surface discharge is more likely to develop in this configuration compared with normal DBD. IV. CONCLUSION Without sliding electrode K T, Sliding =6,4 kv.cm -1 In this study, we studied fundamental properties of surface discharge plasma generated by three electrode system. Major results obtained in this study are summarized as follows: (1) Stable and homogeneous discharge plasma was generated between high voltage and grounded electrodes without arcing with various combination of AC and DC high voltages. (2) With the same peak voltage, higher AC component and lower DC component generated more stable sliding discharge. (3) Lower DC/AC ratio reduced the electric field necessary to initiate sliding discharge. These results suggest that DBD generates charge deposition on the dielectric surface near the high voltage electrode and that sliding discharge is a motion of the deposited charge toward the grounded electrode. This is why stable plasma is generated between high voltage and grounded electrode without arcing because charge deposition is controlled by DBD. (4) Sliding electrode enhanced the length of surface discharge plasma. The effect of the sliding electrode was more significant with larger gap. ACKNOWLEDGEMENT This work has been carried out under the visiting researcher's program of Laboratoire d Etudes Aérodynamiques, Poitiers University, France. One of the authors (K. Takashima) is grateful for the financial and general support for the 6 month visit. REFERENCES [1] L. Loeb, Electrical Coronas, University of California Press, Berkeley, 196 [2] A. Yabe, Y. Mori, K. Hijikata, EHD study of the corona wind between wire and plate electrode, AIAA Journal, Vol. 16, pp , 1978 [3] M. Goldman, A. Goldman, and R.S. Sigmond, The corona discharge, its properties and specific uses, Pure and Appl. Chem., Vol. 7,pp , 198 [4] E. Moreau, Airflow control by non-thermal plasma actuators, J. Phys. D: Appl. Phys., Vol. 4, pp , 27 [] L. Léger, E. Moreau, G. Artana, and G. Touchard, Influence of a DC corona discharge on the airflow along an inclined flat plate, J. Electrostatics, Vol. 1-2, pp. 3-36, 21 [6] E. Moreau, L. Léger, G. Touchard, Effect of a DC surface non-thermal plasma on a flat plate boundary layer for airflow velocity up to 2 m/s, J. Electrostatics, Vol. 64, pp 21-22, 26. [7] J. Pons, E. Moreau, and G. Touchard, Asymmetric surface barrier discharge in air at atmospheric pressure: electrical properties and induced airflow characteristics, J. Phys. D: Appl. Phys., Vol. 38, pp , 2 [8] M. Forte, J. Jolibois, E. Moreau, and G. Touchard, Optimization of a Dielectric Barrier Discharge actuator with non-stationary measurements of the induced flow velocity Application to airflow control in AIAA paper, AIAA Flow Control Conference 3-8 June, San Francisco, CA, p [9] S. I. Andreev, I. M. Belousova, P. N. Dashuk, D. Yu, Zaroslov, E. A. Zobov, N. V. Karlov, G.P.Kuz min, S. M. Nikiforov, and A. M. Prokhorov, Plasma-sheet CO 2 laser, Sov. J. Quantum Electron., Vol. 6, pp , 1976 [1] K. Nakamura, N. Yukawa, T. Mochizuki, S. Horiguchi, and T. Nakaya, Optimization of the discharge characteristics of a laser device employing a plasma electrode, Appl. Phys. Lett., Vol. 49, pp , 1986 [11] B. Arad, Y. Gazit, and A. Ludmirsky, A sliding discharge device for producing cylindrical shock waves, J. Phys. D: Appl. Phys., Vol. 2, pp , 1987 [12] H. Laqua, H. Bluhm, L. Buth, and P. Hoppe, Properties of the nonequilibrium plasma from a pulsed sliding discharge in a hydrogen gas layer desorbed from a metal hydride film, J. Appl. Phys., Vol. 77, pp. 4-2, 199 [13] G. N. Tsikrikas and A. A. Serafetinides, The effect of voltage pulse polarity on the performance of a sliding discharge pumped HF laser, J. Phys. D: Appl. Phys., Vol. 29, pp , 1996 [14] G. N. Tsikrikas, A. A. Serafetinides and A. D. Papayannis, Development of a sliding discharge pumped HF laser, Optics Commun., Vol. 132, pp , 1996 [1] P. A. Atanasov, G. I. Furlinski and Z. Y. Peshev, Spectral and temporal behaviour of sliding-discharge excited UV-IR laser, Optics Commun., Vol. 139, pp , 1997 [16] A. A. Serafetinides, G. Chourdakis and P. Atanasov, Plasma cathode TEA Ar laser development, Optics & Laser Technol., Vol. 3, pp , 1998 [17] A. A. Serafetinides, G. Chourdakis and P. Atanasov, Quasi-simultaneous ultraviolet and infrared emission from a

7 Takashima et al. 2 plasma cathode TEA laser, Optics & Laser Technol., Vol. 33, pp. 8-9, 21 [18] C. Louste, G. Artana, E. Moreau, and G. Touchard, Sliding discharge in air at atmospheric pressure: electrical properties, J. Electrostatics, Vol. 63, pp , 2 [19] C. Louste, E. Moreau, and G. Touchard, Sliding discharge in air at atmospheric pressure: electrical behavior and measurements of the induced ionic wind, in Proceedings on ESA/IEJ/IEEE-IAS/SFE Joint Conference on Electrostatics 26, Berkley, CA, pp [2] E. Odic, A. Goldman, M. Goldman, M. Dhainaut, and R. Dussart, Current distribution of AC surface discharge and associated chemistry, J. Electrostatics, Vol. 64, pp , 26 [21] V.I. Gibalov, and G.J. Pietsch, The development of dielectric barrier discharges in gas gaps and on surfaces, J Phys. D: Appl. Phys., Vol.3 3, pp , 2