- Email: [email protected]
Measurement of the surface charging of a plasma actuator using surface DBD
Journal of Electrostatics 71 (2013) 547e550
Contents lists available at SciVerse ScienceDirect
Journal of Electrostatics journal homepage: www.elsevier.com/locate/elstat
Measurement of the surface charging of a plasma actuator using surface DBD D. Hong a, *, H. Rabat a, Y.-K. Pu b, A. Leroy c a
GREMI, UMR 7344 CNRS/University of Orléans, 14 rue d’Issoudun, 45067 Orléans, France Tsinghua University, Department of Engineering Physics, Beijing 100084, China c PRISME, University of Orléans, 8 rue Léonard de Vinci, 45072 Orléans Cedex 2, France b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 August 2012 Accepted 18 November 2012 Available online 3 December 2012
In order to quantify the surface charging of a plasma actuator using surface Dielectric Barrier Discharge, we propose a new equivalent circuit for this surface DBD and a new simple method allowing the measurement of the surface charging during the first half cycle and the discharging during the other half cycle. Using this method, we observed the temporal evolution of the total charge on the dielectric surface during an operation of a SDBD starting with positive cycle. We also observed the same phenomenon during an operation starting with a negative cycle. The comparison between these two observations suggests that the high electro-negativity of oxygen plays an important role in these discharges. Finally, we compare the total amount of charge transferred over a cycle under different experimental conditions and we find that the transfer is the lowest in oxygen and the highest in nitrogen. Ó 2012 Elsevier B.V. All rights reserved.
Keywords: Surface dielectric barrier discharge Charges Plasma actuator
1. Introduction As described by Moreau in his review paper [1], non-thermal plasmas have been intensively investigated for active airflow control. After 2007, the number of teams investing on this research topic is growing. Most of these teams use a surface dielectric barrier discharge (SDBD) [2,3], since each micro-discharge (filament) in DBD discharge is self-extinguished, thus the discharge cannot become an electric arc. In case of the actuator using a DC Corona discharge [4,5], an arc can occur occasionally and it may damage the actuator and the neighboring body. The operation of a DBD discharge is well known, we recall here very briefly: mechanically, the discharge consists of two electrodes separated by a dielectric. An AC high voltage, often sinusoidal, is applied between these two electrodes, with one of them is generally connected to ground. With increasing voltage, the electric field near the high voltage electrode increases and microdischarges are initiated when the electric field exceeds the critical field. Each micro-discharge deposits charges on the dielectric surface which induce an electric field opposite to that generated by the applied high voltage. This opposition leads to the extinction of the microdischarge whose lifetime is several tens nanoseconds. If the applied voltage rises enough yet, the total electric field may again exceed the critical field and another micro-discharge begins. Moreover, a micro-discharge has a very small size, the deposition of
* Corresponding author. E-mail address: [email protected] (D. Hong). 0304-3886/$ e see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.elstat.2012.11.017
charge is local. Thus the electric field should not be spatially homogeneous. Therefore, a micro-discharge can be initiated independently of others. Generally, when the voltage reaches its maximum value, all the micro-discharges stop, then when the voltage decreases, the local field increases due to that generated by the charges deposited on the surface and the discharge restarts when the critical field strength is reached. Given the operation of the DBD, it is clear that the extent of the deposited charge is paramount. However, because of the difficulty of this measurement, there are still very few published papers on this subject. Recently, Font et al. [6] have measured the surface potential in a DBD actuator using a V-dot probe technique, while Opaits et al. [7] have measured the surface potential due to the residual charge after the operation of their actuator, using an electrostatic voltmeter. They did not calculate the residual charge, because a two dimensional scan of the surface potential is necessary for such a calculation [8,9]. In this paper, we present a new technique for measuring the total charge deposited on the surface of the dielectric as a function of time. In the next section, we present the experimental setup and experimental conditions; then we describe the proposed measurement method. In the third section, results and discussion will be presented. 2. Experimental setup and equivalent circuit In this study, the dielectric is composed of two Kapton sheets of 35 mm in thickness with a Mylar sheet of 0.5 mm placed in the middle. The dielectric thus constructed is called KMK. All the
548
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
measurements presented here were made with the KMK dielectric. Two parallel strip electrodes are placed on each side of the dielectric. An electrode is connected to the high voltage and the other is connected to ground via a capacitor of suitable capacitance. The second electrode is covered with a Kapton sheet to inhibit the discharge on this side. The width of the strips and their relative position can be easily changed from one experiment to another. Generally for actuator applications, we use the configuration called 6e3e6, which means that the width of the two strips was 6 mm and the gap between the two electrodes is 3 mm [10,11]. In this study, the electrode exposed to air has a width of 6 mm, while the buried electrode has a width of 50 mm (see Fig. 1). The area of the facing surface of the two electrodes SES is about 720 mm2. The capacitance of the corresponding capacitor is about 57 pF knowing the relative permittivities of Kapton and Mylar. Actually, the true value should be slightly less because the thickness of adhesive was not taken into account. To measure the active power, the method “Lissajous Figure” is generally used by using a capacitor placed between the buried electrode and ground [2]. In this study, such a capacitor is used to measure the deposited charge. The capacitance of this capacitor, named Cm, is 2 nF, and is much larger than the capacity CES due to the electrodes. So, the presence of this capacitor has a negligible influence on the discharge. With a relatively low voltage, for instance 450 V, there is no initiated discharge. The circuit of Fig. 2 is reduced to a voltage divider and we find a ratio of k ¼ 44.8 between UHV and Um. We deduce a value of 46 pF for CES because CES ¼ Cm/ (k1). This experimental value is quite close to the value previously estimated.
Fig. 2. The electrical setup and the equivalent circuit. Note that the component A represents the actuator part where the charges are localized on dielectric surface.
To analyze the operation of such a SDBD, similar equivalent circuits have been used by many authors. Pons et al. [12] considered the SDBD as a resistor R in parallel with a capacitor C3 (notation used in that paper) and the assembly is in series with a capacitor C2 (notation used in that paper). In this paper, we propose to consider the discharge as a variable current source with a resistor R in series to take into account the dissipated power. The current source deposits randomly the charges on the surface of the dielectric. Charge distribution is necessarily inhomogeneous and the total amount of charges is called QCS. This quantity increases when the current is positive, decreases when the current is negative. We assume, like all other authors, that the presence of this charge QCS induces a charge of same amount but of opposite sign, that is to say QCS. Therefore, regarding the capacitor Cm, the electrode connected to the DBD contains an electric charge of Qm ¼ QCS þ QES, where QES is the charge stored in the capacitor CES, that is to say, the product of UES with CES. Since UES ¼ UHV Um z UHV (k 1)/k and QCS ¼ Qm QES, one deduces: QCS ¼ Um Cm (UHV (k 1)/k) Cm/(k 1) ¼ (Um UHV/k) Cm Therefore, the measurement of Um and UHV as a function of time gives information about the temporal evolution of the total amount of charge on the surface of the dielectric. In this work, the experiments were performed with high voltage UHV of 6 kV in amplitude and of 1 kHz in frequency which are typical parameters used in our other studies [10,11]. The high voltage UHV is measured using a Tektronix P6015 probe and the voltage Um is measured with a voltage probe whose input impedance is 100 MU. The time constant calculated with this impedance and the value of capacitance Cm (2 nF) is 200 ms and measuring a voltage across the capacitor as a function of time allowed to obtain a time constant of about 240 ms. This means that the capacitor is systematically emptied between two series of measurements. The signals were stored using a LeCroy oscilloscope (HRO 66Z, 600 MHz, 12 bits) with in general 10,000 samples over 10 ms. Some measurements were made under a controlled atmosphere with pure nitrogen or pure oxygen gas. The device for theses latter experiments has been described in Tong et al. [13]. 3. Results and discussions
Fig. 1. The geometrical configuration of the two electrodes of the actuator. The narrower electrode is air exposed and connected to HV, while the larger electrode is buried and connected to ground via a capacitor for measurement.
In Fig. 3, a typical result is reported. The values given in this paragraph may vary slightly from one shot to another. The first curve in dotted line shows the applied high voltage as a function of time, while the second curve gives the total charge on the surface as a function of time. The high voltage is applied at time t ¼ 0 ms, during the first 10 ms, the discharge is not initiated and there is no charge on the
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
549
Fig. 3. Time evolution of the electrical charge on dielectric surface, together with the applied high voltage.
surface. At time t ¼ 80 ms where the voltage is about 2.6 kV, the discharge starts and a charge of 31.4 nC is suddenly deposited on the surface. The microdischarges are initiated and extinguished successively until time of t ¼ 0.21 ms, where the high voltage reaches approximately its maximum value and the electric field is no longer sufficient for any discharge. At this time the surface is positively charged with a total charge of 130 nC. During this phase, the air exposed electrode acts as the anode and this phase is called “anodic phase” or “anodic discharge”. With the decrease of high voltage, the total field increases, and at time t ¼ 0.45 ms, the discharge restarts with a reverse current comparing to the first one, that is to say that the positive charges leave the surfaces. In reality, it is likely that the electrons from the electrode will gradually neutralize the positive charges. At time t ¼ 0.75 ms when the voltage is minimum (maximum in absolute value), the discharge stops and at that time, the surface is
negatively charged with a total charge of 13 nC. Similarly, this phase is called “cathodic phase or discharge” since the exposed electrode acts as the cathode. Then, with increasing voltage, field increases again and the discharge stops again at t ¼ 0.94 ms when the voltage is about 2.5 kV. All the discharges of new anodic phase stop when the voltage reaches to its maximum value and the total charge is again of about þ130 nC. Then the cycle repeats itself as long as the AC high-voltage is switched-on. In this case with positive cycle at first (case shown in Fig. 3), a quantity of positive charge is deposited on the surface during the anodic phase and then during cathodic phase, all deposited charges are virtually neutralized. In case with negative cycle at first, one may think that the phenomenon will be reversed. But in reality, this is not the case. Indeed, as shown in Fig. 4, we find that the amount of negative charge deposited during the first cathodic phase is barely half (in
Fig. 4. Time evolution of the electrical charge on dielectric surface, together with the applied high voltage beginning with a negative cycle.
550
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
Table 1 Approximate value of QPP at several experiment conditions. Case Case Case Case
1 2 3 4
In In In In
ambient gas, 20 C ambient gas, 90 C pure N2 gas, 20 C pure O2 gas, 20 C
130 150 220 110
nC nC nC nC
absolute value) of the one when the discharge starts with a positive cycle. Then, quickly, the transfer of the discharge is similar to that in the case where the discharge begins with a positive cycle. It seems that the high mobility of electrons and the high electronegativity of oxygen play a role in this observed phenomenon. When the discharge starts with an anodic phase, electrons can come from far and micro-discharges (filaments) grow easily. When the discharge starts with a cathodic phase, the oxygen molecules rapidly capture electrons and prevent microdischarges to develop, the charge deposition is then lower and close to the exposed electrode and the charges will be neutralized easily and quickly at the beginning of the anodic phase allowing the discharge to develop as if the first anodic phase is not preceded by a cathodic phase. This difference in expansion of the filaments during the two half cycles has been clearly shown by Audier et al. [14] using the time resolved discharge images taken with an ICCD camera. We find that the initial state of the surface can affect the transfer of the charge. Depending on the initial surface charging, the maximum value of the QCS changes, together with the minimum value. We note that the gap between these two values, called here QPP for peak-to-peak value, remains about the same when the initial surface charging changes. Since it is difficult to control the initial state of the surface, we opt to compare the QPP value for different experimental conditions. In the Table 1, we give this value for four different atmospheres while keeping the same dielectric (KMK) and the same sinusoidal voltage (6 kV in amplitude and 1 kHz in frequency). It can be seen with pure O2 (case 4), charge transfer is the weakest, certainly because the strong electro-negativity of oxygen as already mentioned. If we consider that this transfer was done during a half cycle, i.e. 0.5 ms, we yield a mean current of 0.22 mA. This value is in good agreement with previous results. With pure N2 (case 3), the charge transfer is the strongest, thanks to the absence of oxygen. The strong charge transfer means a higher mean current and therefore greater active power dissipation. This deduction is in perfect agreement with the power measurements published by Audier et al. [14]. The measurements were also made by heating the actuator with a halogen lamp up to 90 C (case 2), we find that charge transfer is stronger compared to that obtained at 20 C (case 1). This increase is certainly due to the improvement of thermionic emission of the electrode and electron avalanche in air.
4. Conclusion In this paper, we propose a new equivalent circuit for the DBD surface discharge used by most of the plasma actuators for airflow
control. In this circuit, a current source is used and we believe that it better reflects the reality of the electrical discharge. We also propose a new method to quantify the surface charging during the first half cycle and the discharging of this surface during the other half cycle. With this simple method, we observed the temporal evolution of the total charge on the dielectric surface during an operation of a SDBD starting with positive cycle. We also observed the same phenomenon during an operation starting with a negative cycle. The comparison between these two observations suggests that the high electro-negativity of oxygen plays an important role in these discharges. The initial state can influence the surface charge exchange on a half-cycle, but has little influence on the total amount of charge transferred over a full cycle. We compare this total amount of charge transferred under different experimental conditions and we find that the transfer is the lowest in oxygen and the highest in nitrogen. In the future, it is important to experimentally validate the proposed method for measuring the charging versus time. Once the method is validated, it will be interesting to make measurements by varying the frequency and amplitude of the voltage and changing the nature of the dielectric. References [1] E. Moreau, Airflow control by non-thermal plasma actuators, Journal of Physics. D: Applied. Physics vol. 40 (3) (2007) 605e636. [2] B. Dong, J.M. Bauchire, J.M. Pouvesle, P. Magnier, D. Hong, Experimental study of a DBD surface discharge for the active control of subsonic airflow, Journal of Physics D: Applied Physics vol. 41 (15) (2008) 155201. [3] T.C. Corke, M.L. Post, D.M. Orlov, Single dielectric barrier discharge plasma enhanced aerodynamics: physics, modeling and applications, Experiments in Fluids vol. 46 (1) (2009) 1e26. [4] L. Léger, E. Moreau, G. Artana, G. Touchard, Influence of a DC corona discharge on the airflow along an inclined flat plate, Journal of Electrostatics vol. 51e52 (0) (2001) 300e306. [5] P. Magnier, D. Hong, A. Leroy-Chesneau, J.-M. Bauchire, J. Hureau, Control of separated flows with the ionic wind generated by a DC corona discharge, Experiments in Fluids vol. 42 (5) (2007) 815e825. [6] G.I. Font, C.L. Enloe, T.E. McLaughlin, D. Orlov, Proceedings of the 45th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, 2007, AIAA, Washington, D.C., Paper No. AIAA-2007-188, 2007. [7] D.F. Opaits, M.N. Shneider, Richard B. Miles, A.V. Likhanskii, S.O. Macheret, Surface charge in dielectric barrier discharge plasma actuators, Physics of Plasmas 15 (2008). 073505 (5 pp). [8] A. Kumada, S. Okabe, K. Hidaka, Residual charge distribution of positive surface streamer, Journal of Physics D: Applied Physics 42 (2009). 095209 (8 pp). [9] J. Deng, S. Matsuoka, A. Kumada, K. Hidaka, The influence of residual charge on surface discharge propagation, Journal of Physics D: Applied Physics 43 (2010). 495203 (8 pp). [10] R. Joussot, R. Weber, V. Bouchina, A. Leroy, D. Hong, Modification of the Laminar-to-Turbulent Transition on a Flat Plate Using DBD Plasma Actuator, 5th AIAA Flow Control Conference, Chicago, USA, AIAA-2010-4708, 2010. [11] R. Joussot, R. Weber, A. Leroy, D. Hong, Transition control using a single plasma actuator, Int. J. Aerodyn. 3 (2013). Nos. 1/2/3. [12] J. Pons, E. Moreau, G. Touchard, Asymmetric surface dielectric barrier discharge in air at atmospheric pressure: electrical properties and induced airflow characteristics, Journal of Physics D: Applied Physics 38 (2005) 3636e3642. [13] B.-S. Tong, D. Hong, H. Rabat, H.-L. Chen, M.-B. Chang, Investigation on ozone formation with plasma actuator, in: Proceeding of 8th ISNTPT, Camaret, France, June 25e29, 2012. [14] P. Audier, A. Leroy, H. Rabat, D. Hong, Influence of the N2/O2 Volumetric Ratio on the behavior of a SDBD plasma actuator, 6th Flow Control Conference, AIAA-2012-3094, New Orleans, Louisiana, June 25e28, 2012.
Contents lists available at SciVerse ScienceDirect
Journal of Electrostatics journal homepage: www.elsevier.com/locate/elstat
Measurement of the surface charging of a plasma actuator using surface DBD D. Hong a, *, H. Rabat a, Y.-K. Pu b, A. Leroy c a
GREMI, UMR 7344 CNRS/University of Orléans, 14 rue d’Issoudun, 45067 Orléans, France Tsinghua University, Department of Engineering Physics, Beijing 100084, China c PRISME, University of Orléans, 8 rue Léonard de Vinci, 45072 Orléans Cedex 2, France b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 August 2012 Accepted 18 November 2012 Available online 3 December 2012
In order to quantify the surface charging of a plasma actuator using surface Dielectric Barrier Discharge, we propose a new equivalent circuit for this surface DBD and a new simple method allowing the measurement of the surface charging during the first half cycle and the discharging during the other half cycle. Using this method, we observed the temporal evolution of the total charge on the dielectric surface during an operation of a SDBD starting with positive cycle. We also observed the same phenomenon during an operation starting with a negative cycle. The comparison between these two observations suggests that the high electro-negativity of oxygen plays an important role in these discharges. Finally, we compare the total amount of charge transferred over a cycle under different experimental conditions and we find that the transfer is the lowest in oxygen and the highest in nitrogen. Ó 2012 Elsevier B.V. All rights reserved.
Keywords: Surface dielectric barrier discharge Charges Plasma actuator
1. Introduction As described by Moreau in his review paper [1], non-thermal plasmas have been intensively investigated for active airflow control. After 2007, the number of teams investing on this research topic is growing. Most of these teams use a surface dielectric barrier discharge (SDBD) [2,3], since each micro-discharge (filament) in DBD discharge is self-extinguished, thus the discharge cannot become an electric arc. In case of the actuator using a DC Corona discharge [4,5], an arc can occur occasionally and it may damage the actuator and the neighboring body. The operation of a DBD discharge is well known, we recall here very briefly: mechanically, the discharge consists of two electrodes separated by a dielectric. An AC high voltage, often sinusoidal, is applied between these two electrodes, with one of them is generally connected to ground. With increasing voltage, the electric field near the high voltage electrode increases and microdischarges are initiated when the electric field exceeds the critical field. Each micro-discharge deposits charges on the dielectric surface which induce an electric field opposite to that generated by the applied high voltage. This opposition leads to the extinction of the microdischarge whose lifetime is several tens nanoseconds. If the applied voltage rises enough yet, the total electric field may again exceed the critical field and another micro-discharge begins. Moreover, a micro-discharge has a very small size, the deposition of
* Corresponding author. E-mail address: [email protected] (D. Hong). 0304-3886/$ e see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.elstat.2012.11.017
charge is local. Thus the electric field should not be spatially homogeneous. Therefore, a micro-discharge can be initiated independently of others. Generally, when the voltage reaches its maximum value, all the micro-discharges stop, then when the voltage decreases, the local field increases due to that generated by the charges deposited on the surface and the discharge restarts when the critical field strength is reached. Given the operation of the DBD, it is clear that the extent of the deposited charge is paramount. However, because of the difficulty of this measurement, there are still very few published papers on this subject. Recently, Font et al. [6] have measured the surface potential in a DBD actuator using a V-dot probe technique, while Opaits et al. [7] have measured the surface potential due to the residual charge after the operation of their actuator, using an electrostatic voltmeter. They did not calculate the residual charge, because a two dimensional scan of the surface potential is necessary for such a calculation [8,9]. In this paper, we present a new technique for measuring the total charge deposited on the surface of the dielectric as a function of time. In the next section, we present the experimental setup and experimental conditions; then we describe the proposed measurement method. In the third section, results and discussion will be presented. 2. Experimental setup and equivalent circuit In this study, the dielectric is composed of two Kapton sheets of 35 mm in thickness with a Mylar sheet of 0.5 mm placed in the middle. The dielectric thus constructed is called KMK. All the
548
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
measurements presented here were made with the KMK dielectric. Two parallel strip electrodes are placed on each side of the dielectric. An electrode is connected to the high voltage and the other is connected to ground via a capacitor of suitable capacitance. The second electrode is covered with a Kapton sheet to inhibit the discharge on this side. The width of the strips and their relative position can be easily changed from one experiment to another. Generally for actuator applications, we use the configuration called 6e3e6, which means that the width of the two strips was 6 mm and the gap between the two electrodes is 3 mm [10,11]. In this study, the electrode exposed to air has a width of 6 mm, while the buried electrode has a width of 50 mm (see Fig. 1). The area of the facing surface of the two electrodes SES is about 720 mm2. The capacitance of the corresponding capacitor is about 57 pF knowing the relative permittivities of Kapton and Mylar. Actually, the true value should be slightly less because the thickness of adhesive was not taken into account. To measure the active power, the method “Lissajous Figure” is generally used by using a capacitor placed between the buried electrode and ground [2]. In this study, such a capacitor is used to measure the deposited charge. The capacitance of this capacitor, named Cm, is 2 nF, and is much larger than the capacity CES due to the electrodes. So, the presence of this capacitor has a negligible influence on the discharge. With a relatively low voltage, for instance 450 V, there is no initiated discharge. The circuit of Fig. 2 is reduced to a voltage divider and we find a ratio of k ¼ 44.8 between UHV and Um. We deduce a value of 46 pF for CES because CES ¼ Cm/ (k1). This experimental value is quite close to the value previously estimated.
Fig. 2. The electrical setup and the equivalent circuit. Note that the component A represents the actuator part where the charges are localized on dielectric surface.
To analyze the operation of such a SDBD, similar equivalent circuits have been used by many authors. Pons et al. [12] considered the SDBD as a resistor R in parallel with a capacitor C3 (notation used in that paper) and the assembly is in series with a capacitor C2 (notation used in that paper). In this paper, we propose to consider the discharge as a variable current source with a resistor R in series to take into account the dissipated power. The current source deposits randomly the charges on the surface of the dielectric. Charge distribution is necessarily inhomogeneous and the total amount of charges is called QCS. This quantity increases when the current is positive, decreases when the current is negative. We assume, like all other authors, that the presence of this charge QCS induces a charge of same amount but of opposite sign, that is to say QCS. Therefore, regarding the capacitor Cm, the electrode connected to the DBD contains an electric charge of Qm ¼ QCS þ QES, where QES is the charge stored in the capacitor CES, that is to say, the product of UES with CES. Since UES ¼ UHV Um z UHV (k 1)/k and QCS ¼ Qm QES, one deduces: QCS ¼ Um Cm (UHV (k 1)/k) Cm/(k 1) ¼ (Um UHV/k) Cm Therefore, the measurement of Um and UHV as a function of time gives information about the temporal evolution of the total amount of charge on the surface of the dielectric. In this work, the experiments were performed with high voltage UHV of 6 kV in amplitude and of 1 kHz in frequency which are typical parameters used in our other studies [10,11]. The high voltage UHV is measured using a Tektronix P6015 probe and the voltage Um is measured with a voltage probe whose input impedance is 100 MU. The time constant calculated with this impedance and the value of capacitance Cm (2 nF) is 200 ms and measuring a voltage across the capacitor as a function of time allowed to obtain a time constant of about 240 ms. This means that the capacitor is systematically emptied between two series of measurements. The signals were stored using a LeCroy oscilloscope (HRO 66Z, 600 MHz, 12 bits) with in general 10,000 samples over 10 ms. Some measurements were made under a controlled atmosphere with pure nitrogen or pure oxygen gas. The device for theses latter experiments has been described in Tong et al. [13]. 3. Results and discussions
Fig. 1. The geometrical configuration of the two electrodes of the actuator. The narrower electrode is air exposed and connected to HV, while the larger electrode is buried and connected to ground via a capacitor for measurement.
In Fig. 3, a typical result is reported. The values given in this paragraph may vary slightly from one shot to another. The first curve in dotted line shows the applied high voltage as a function of time, while the second curve gives the total charge on the surface as a function of time. The high voltage is applied at time t ¼ 0 ms, during the first 10 ms, the discharge is not initiated and there is no charge on the
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
549
Fig. 3. Time evolution of the electrical charge on dielectric surface, together with the applied high voltage.
surface. At time t ¼ 80 ms where the voltage is about 2.6 kV, the discharge starts and a charge of 31.4 nC is suddenly deposited on the surface. The microdischarges are initiated and extinguished successively until time of t ¼ 0.21 ms, where the high voltage reaches approximately its maximum value and the electric field is no longer sufficient for any discharge. At this time the surface is positively charged with a total charge of 130 nC. During this phase, the air exposed electrode acts as the anode and this phase is called “anodic phase” or “anodic discharge”. With the decrease of high voltage, the total field increases, and at time t ¼ 0.45 ms, the discharge restarts with a reverse current comparing to the first one, that is to say that the positive charges leave the surfaces. In reality, it is likely that the electrons from the electrode will gradually neutralize the positive charges. At time t ¼ 0.75 ms when the voltage is minimum (maximum in absolute value), the discharge stops and at that time, the surface is
negatively charged with a total charge of 13 nC. Similarly, this phase is called “cathodic phase or discharge” since the exposed electrode acts as the cathode. Then, with increasing voltage, field increases again and the discharge stops again at t ¼ 0.94 ms when the voltage is about 2.5 kV. All the discharges of new anodic phase stop when the voltage reaches to its maximum value and the total charge is again of about þ130 nC. Then the cycle repeats itself as long as the AC high-voltage is switched-on. In this case with positive cycle at first (case shown in Fig. 3), a quantity of positive charge is deposited on the surface during the anodic phase and then during cathodic phase, all deposited charges are virtually neutralized. In case with negative cycle at first, one may think that the phenomenon will be reversed. But in reality, this is not the case. Indeed, as shown in Fig. 4, we find that the amount of negative charge deposited during the first cathodic phase is barely half (in
Fig. 4. Time evolution of the electrical charge on dielectric surface, together with the applied high voltage beginning with a negative cycle.
550
D. Hong et al. / Journal of Electrostatics 71 (2013) 547e550
Table 1 Approximate value of QPP at several experiment conditions. Case Case Case Case
1 2 3 4
In In In In
ambient gas, 20 C ambient gas, 90 C pure N2 gas, 20 C pure O2 gas, 20 C
130 150 220 110
nC nC nC nC
absolute value) of the one when the discharge starts with a positive cycle. Then, quickly, the transfer of the discharge is similar to that in the case where the discharge begins with a positive cycle. It seems that the high mobility of electrons and the high electronegativity of oxygen play a role in this observed phenomenon. When the discharge starts with an anodic phase, electrons can come from far and micro-discharges (filaments) grow easily. When the discharge starts with a cathodic phase, the oxygen molecules rapidly capture electrons and prevent microdischarges to develop, the charge deposition is then lower and close to the exposed electrode and the charges will be neutralized easily and quickly at the beginning of the anodic phase allowing the discharge to develop as if the first anodic phase is not preceded by a cathodic phase. This difference in expansion of the filaments during the two half cycles has been clearly shown by Audier et al. [14] using the time resolved discharge images taken with an ICCD camera. We find that the initial state of the surface can affect the transfer of the charge. Depending on the initial surface charging, the maximum value of the QCS changes, together with the minimum value. We note that the gap between these two values, called here QPP for peak-to-peak value, remains about the same when the initial surface charging changes. Since it is difficult to control the initial state of the surface, we opt to compare the QPP value for different experimental conditions. In the Table 1, we give this value for four different atmospheres while keeping the same dielectric (KMK) and the same sinusoidal voltage (6 kV in amplitude and 1 kHz in frequency). It can be seen with pure O2 (case 4), charge transfer is the weakest, certainly because the strong electro-negativity of oxygen as already mentioned. If we consider that this transfer was done during a half cycle, i.e. 0.5 ms, we yield a mean current of 0.22 mA. This value is in good agreement with previous results. With pure N2 (case 3), the charge transfer is the strongest, thanks to the absence of oxygen. The strong charge transfer means a higher mean current and therefore greater active power dissipation. This deduction is in perfect agreement with the power measurements published by Audier et al. [14]. The measurements were also made by heating the actuator with a halogen lamp up to 90 C (case 2), we find that charge transfer is stronger compared to that obtained at 20 C (case 1). This increase is certainly due to the improvement of thermionic emission of the electrode and electron avalanche in air.
4. Conclusion In this paper, we propose a new equivalent circuit for the DBD surface discharge used by most of the plasma actuators for airflow
control. In this circuit, a current source is used and we believe that it better reflects the reality of the electrical discharge. We also propose a new method to quantify the surface charging during the first half cycle and the discharging of this surface during the other half cycle. With this simple method, we observed the temporal evolution of the total charge on the dielectric surface during an operation of a SDBD starting with positive cycle. We also observed the same phenomenon during an operation starting with a negative cycle. The comparison between these two observations suggests that the high electro-negativity of oxygen plays an important role in these discharges. The initial state can influence the surface charge exchange on a half-cycle, but has little influence on the total amount of charge transferred over a full cycle. We compare this total amount of charge transferred under different experimental conditions and we find that the transfer is the lowest in oxygen and the highest in nitrogen. In the future, it is important to experimentally validate the proposed method for measuring the charging versus time. Once the method is validated, it will be interesting to make measurements by varying the frequency and amplitude of the voltage and changing the nature of the dielectric. References [1] E. Moreau, Airflow control by non-thermal plasma actuators, Journal of Physics. D: Applied. Physics vol. 40 (3) (2007) 605e636. [2] B. Dong, J.M. Bauchire, J.M. Pouvesle, P. Magnier, D. Hong, Experimental study of a DBD surface discharge for the active control of subsonic airflow, Journal of Physics D: Applied Physics vol. 41 (15) (2008) 155201. [3] T.C. Corke, M.L. Post, D.M. Orlov, Single dielectric barrier discharge plasma enhanced aerodynamics: physics, modeling and applications, Experiments in Fluids vol. 46 (1) (2009) 1e26. [4] L. Léger, E. Moreau, G. Artana, G. Touchard, Influence of a DC corona discharge on the airflow along an inclined flat plate, Journal of Electrostatics vol. 51e52 (0) (2001) 300e306. [5] P. Magnier, D. Hong, A. Leroy-Chesneau, J.-M. Bauchire, J. Hureau, Control of separated flows with the ionic wind generated by a DC corona discharge, Experiments in Fluids vol. 42 (5) (2007) 815e825. [6] G.I. Font, C.L. Enloe, T.E. McLaughlin, D. Orlov, Proceedings of the 45th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, 2007, AIAA, Washington, D.C., Paper No. AIAA-2007-188, 2007. [7] D.F. Opaits, M.N. Shneider, Richard B. Miles, A.V. Likhanskii, S.O. Macheret, Surface charge in dielectric barrier discharge plasma actuators, Physics of Plasmas 15 (2008). 073505 (5 pp). [8] A. Kumada, S. Okabe, K. Hidaka, Residual charge distribution of positive surface streamer, Journal of Physics D: Applied Physics 42 (2009). 095209 (8 pp). [9] J. Deng, S. Matsuoka, A. Kumada, K. Hidaka, The influence of residual charge on surface discharge propagation, Journal of Physics D: Applied Physics 43 (2010). 495203 (8 pp). [10] R. Joussot, R. Weber, V. Bouchina, A. Leroy, D. Hong, Modification of the Laminar-to-Turbulent Transition on a Flat Plate Using DBD Plasma Actuator, 5th AIAA Flow Control Conference, Chicago, USA, AIAA-2010-4708, 2010. [11] R. Joussot, R. Weber, A. Leroy, D. Hong, Transition control using a single plasma actuator, Int. J. Aerodyn. 3 (2013). Nos. 1/2/3. [12] J. Pons, E. Moreau, G. Touchard, Asymmetric surface dielectric barrier discharge in air at atmospheric pressure: electrical properties and induced airflow characteristics, Journal of Physics D: Applied Physics 38 (2005) 3636e3642. [13] B.-S. Tong, D. Hong, H. Rabat, H.-L. Chen, M.-B. Chang, Investigation on ozone formation with plasma actuator, in: Proceeding of 8th ISNTPT, Camaret, France, June 25e29, 2012. [14] P. Audier, A. Leroy, H. Rabat, D. Hong, Influence of the N2/O2 Volumetric Ratio on the behavior of a SDBD plasma actuator, 6th Flow Control Conference, AIAA-2012-3094, New Orleans, Louisiana, June 25e28, 2012.