AIAA JOURNAL

Vol. 42, No. 3, March 2004

Mechanisms and Responses of a Single Dielectric Barrier

Plasma Actuator: Plasma Morphology

C. L. Enloe,∗ Thomas E. McLaughlin,† Robert D. VanDyken,‡ and K. D. Kachner§

U.S. Air Force Academy, Colorado Springs, Colorado 80840-6222
and
Eric J. Jumper¶ and Thomas C. Corke∗∗
University of Notre Dame, Notre Dame, Indiana 46556-5684

We present simultaneous optical, electrical, and thrust measurements of an aerodynamic plasma actuator. These
measurements indicate that the plasma actuator is a form of the dielectric barrier discharge, whose behavior is
governed primarily by the buildup of charge on the dielectric-encapsulated electrode. Our measurements reveal
the temporal and macroscale spatial structure of the plasma. Correlating the morphology of the plasma and the
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

electrical characteristics of the discharge to the actuator performance as measured by the thrust produced indicates
a direct coupling between the interelectrode electric field (strongly modified by the presence of the plasma) and the
charges in the plasma. Our measurements discount bulk heating or asymmetries in the structure of the discharge as
mechanisms for the production of bulk motion of the surrounding neutral air, although such asymmetries clearly
exist and impact the effectiveness of the actuator.

Introduction: Morphology of the Plasma Actuator We take the light emissions from the plasma actuator as a surro-
gate for plasma density, assuming that the recombination time of the
T HE mechanical configuration of the aerodynamic plasma ac-
tuator is shown in Fig. 1. The plasma actuator consists of a set
of thin electrodes (in our case made of copper foil tape) arranged
plasma is short compared to the timescale of the discharge. (This
assumption is confirmed by our emissions measurements and is also
spanwise on an aerodynamic surface. One electrode is exposed to consistent with the calculations of Vidmar and Stalder,11 which in-
the air, while the other is encapsulated in dielectric (Kapton® poly- dicate that we should expect a plasma lifetime on the order of 10−8 s
imide tape, in our test articles). The electrodes are offset as shown for our atmospheric-pressure plasmas.) The first observation from
in the figure and are typically 80 mm wide. the data is that what appears as continuous discharge has consid-
When high ac voltage (5–10-kV amplitude, with frequency in erable temporal structure. Figure 5, for example, shows two cycles
the range of 1–10 kHz) is applied to the electrodes, the unaided eye of a plasma discharge that turns on and off four times during each
sees an apparently diffuse plasma discharge, as shown in Fig. 2. The cycle of the applied voltage.
appearance of the plasma is accompanied by a coupling of directed The temporal nature of the actuator indicates that this plasma is
momentum into the surrounding air. This momentum coupling can indeed (as one would have inferred from the electrode configuration)
be effective in substantially altering the flow of air over the actuator a dielectric barrier discharge (DBD), a configuration about which
surface,1−9 but, as Fig. 3 shows, it is equally effective in introducing there is considerable information in the literature12−23 (see, for ex-
a flow in initially still air.10 ample, the review paper by Kunhardt16 ) dating even from the turn of
Although the plasma appears as a relatively uniform diffuse dis- the 20th century.24 The plasma actuator differs from the most com-
charge to the unaided eye, optical measurements of the plasma in- mon DBD configuration used in plasma processing in that it employs
dicate that it is highly structured in both space and time. Figure 4 a single encapsulated electrode and an asymmetric electrode ar-
illustrates the experimental apparatus used to make these measure- rangement, but the principles of the discharge are the same. (Gibalov
ments. A photomultiplier tube (PMT) was used to observe the bulk and Pietsch17 have compared the development of this surface-
plasma with high time resolution. For most of the optical mea- discharge configuration of the DBD with the more common “volume
surements presented here, the PMT was arranged so as to observe discharge” configuration.) The most important feature of the DBD is
approximately one-third of the length of the plasma actuator. For that it can sustain a large-volume discharge at atmospheric pressure
some measurements, a thin slit aperture was interposed between the without the discharge’s collapsing into a constricted arc.
plasma and the PMT, so that the light observations could be limited The DBD can maintain such a discharge because the configuration
to approximately a 1-mm-wide region in the chordwise direction. is self-limiting, as illustrated in Fig. 6. To maintain a DBD discharge,
an ac applied voltage is required. Figure 6a illustrates the half-cycle
of the discharge for which the exposed electrode is more negative
Presented as Paper 2003-1021 at the AIAA 41st Aerospace Sciences Meet-
ing, Reno, NV, 6–9 January 2003; received 7 May 2003; accepted for pub-
than the surface of the dielectric and the insulated electrode, thus
lication 4 November 2003. This material is declared a work of the U.S. taking the role of the cathode in the discharge. In this case, assuming
Government and is not subject to copyright protection in the United States. the potential difference is high enough the exposed electrode can
Copies of this paper may be made for personal or internal use, on condi- emit electrons. Because the discharge terminates on a dielectric
tion that the copier pay the $10.00 per-copy fee to the Copyright Clearance surface, however (hence the term “dielectric barrier”), the buildup of
Center, Inc., 222 Rosewood Drive, Danvers, MA 01923; include the code surface charge opposes the applied voltage, and the discharge shuts
0001-1452/04 $10.00 in correspondence with the CCC. itself off unless the magnitude of the applied voltage is continually
∗ Professor, Department of Physics. Senior Member AIAA.
† Research Associate, Department of Aeronautics. Associate Fellow
increased. This is the explanation of the behavior shown in Fig. 5.
At point a in the figure, because of some impedance mismatch in
AIAA.
‡ Research Assistant, Department of Aeronautics. the driving circuit there is a momentary reversal in the slope of
§ Cadet First Class, Department of Physics. the applied waveform. Because the applied voltage is no longer
¶ Professor, Department of Aerospace and Mechanical Engineering. Fel- becoming more negative, the discharge shuts off. When, at point b,
low AIAA. the applied voltage again resumes its negative course, the discharge
∗∗ Clark Chair Professor, Department of Aerospace and Mechanical reignites and stays ignited until the slope of the voltage waveform
Engineering. Associate Fellow AIAA. goes to zero at approximately t = 0.4 ms.
589
 590 ENLOE ET AL.
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

Fig. 1 Plasma actuator is an asymmetric arrangement of electrodes,

one of which is insulated, on an aerodynamic surface. Fig. 4 PMT is used to take fast measurements of the light emission
from the plasma actuator.

Fig. 2 Plasma actuator in action appears as a diffuse plasma formed

on the surface of the dielectric.
Fig. 5 Light emission from the plasma actuator clearly establishes it
as a dielectric barrier discharge.

a)

Fig. 3 Plasma actuator can couple momentum into still air along the
aerodynamic surface, as illustrated by this flow visualization.

The behavior of the discharge is similar on the opposite half-cycle:

a positive slope in the applied voltage is necessary to maintain the
discharge. In this half-cycle, the charge available to the discharge
is limited to that deposited during the previous half-cycle on the
dielectric surface (which now plays the role of the cathode), as b)
shown in Fig. 6b. Fig. 6 Dielectric barrier discharge is self-limiting because charge
This self-limiting behavior caused by charge buildup on the di- buildup on the dielectric surface opposes the voltage applied across
electric surface impacts the spatial as well as the temporal structure the plasma, when the applied voltage is a) negative going. b) When the
of the plasma. (We note in passing that the rapid termination of op- voltage reverses, the charge transferred through the plasma is limited
tical emissions with the even momentary reversal of dV /dt in the to that deposited on the dielectric surface.
 ENLOE ET AL. 591

applied waveform refutes the claim of Roth et al.25 that charge builds
up in the interelectrode gap because of ion trapping on timescales
for a long time compared to the period of the applied voltage. If that
were the case, then the ion population would reach an equilibrium
value and recombination/deexcitation, with associated light emis-
sion, would occur at a relatively uniform rate. In fact, the plasma
lifetime is shorter by several orders of magnitude than the timescale
of the applied voltage waveform.11 ) It is well established in the liter-
ature that although the dielectric barrier discharge consists in many
cases of a series of microdischarges12−23 the same discharge sup-
ports other, more diffuse modes,12,14,16 depending on a number of
factors of which the plasma chemistry in the discharge is the chief.14
Our optical measurements indicate that there is considerable
macroscopic structure spanwise in the plasma actuator discharge.
Figure 7 shows one discharge cycle of the plasma actuator with a
sinusoidal applied voltage waveform. Both the current through the
discharge and the emitted light are shown. The figure shows that the
discharge is much more irregular on the positive-going half-cycle
than the negative-going. (This behavior is consistent with data in
Fig. 9 Time to first light as a function of lateral (chordwise) distance
the literature for DBDs with a single dielectric barrier,17,23 although
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

shows that the plasma grows laterally at a constant rate as the discharge
it is not widely noted; for example, see comments by Gibalov and progresses. The propagation speed for the negative-going half-cycle
Pietsch.17 ) (•,
) is essentially the same as for the positive-going half-cycle ( e, ).
Zooming in on the same data on a finer timescale (see Fig. 8)
shows that each pulse of light observed on the PMT corresponds to
a pulse in the current signal. The reverse, however, is not true: not that there are discharge events (current pulses) that do not occur in
every current pulse corresponds to a light pulse. The explanation the field of view of the PMT. When the voltage on the exposed elec-
for this observation is straightforward. The PMT’s field of view is trode is negative-going, the discharge is relatively uniform across
approximately one-third of the plasma actuator. The current mon- the width of the actuator. When the same voltage is positive-going,
itor, however, “sees” the entire discharge. Therefore we conclude however, the discharge is “patchy,” akin to flashbulbs going off in
a stadium. (These measurements are consistent with those made
by Gibalov and Pietsch17 and Wilkinson1 using fast photography.)
This asymmetry in the discharge plays a role in the efficiency of
momentum coupling to the flow, as described in the next section.
Optical measurements also indicate that, as Gibalov and Pietsch
have noted,17 the lateral extent of the plasma develops in time. Fig-
ure 2 is essentially an open-shutter view of the plasma; the shutter
speed is longer than the period of the applied voltage waveform.
One would be tempted to interpret this photograph as showing a
density gradient in the plasma, with the maximum density nearest
the edge of the exposed electrode. This interpretation would be in
error, however, as measurements of light emission through a narrow
aperture shows. Figure 9 shows the relative time to first light as a
function of the lateral position of the aperture. The figure clearly
shows that the plasma grows in the lateral (chordwise) direction at
a constant rate. Therefore, the fact that the plasma appears brighter
nearer the electrode in Fig. 2 corresponds to that location’s having
emitted for a greater fraction of the discharge cycle, rather than to
the presence of a higher plasma density.
Fig. 7 Emission from the plasma indicate a much more irregular dis- From Fig. 9 it is also clear that the propagation speed of the dis-
charge on the positive-going part of the cycle (0.0 to 0.2 ms in this figure) charge is a function of the amplitude of the applied voltage. The
than on the negative-going part (0.2 to 0.4 ms). higher the voltage, the faster the discharge spreads along the dielec-
tric surface. Furthermore, the propagation speed of the discharge is
essentially the same for both half-cycles of the discharge (negative
and positive going) for a given voltage, and in both cases the dis-
charge ignites at the edge of the exposed dielectric and propagates
downstream along the dielectric surface. This level of symmetry in
the structure of the discharge refutes the model proposed by Shyy
et al.,26 which implies that electrons leaving the dielectric surface
would have energy sufficient to ionize the background only when
they near the exposed electrode and proceeds to attribute the action
of the actuator to such asymmetry in the discharge. On the contrary,
although there is a difference in the transverse (spanwise) struc-
ture of the plasma between half-cycles of the discharge the lateral
(chordwise) extent and development of the plasma is essentially the
same.

Electrical Circuit Analysis of the Plasma Actuator

Understanding that the plasma actuator is in fact a dielectric bar-
rier discharge makes it possible to analyze the discharge with a
Fig. 8 Detailed look at simultaneous current and light data indicate lumped-element circuit model. The simplest circuit model that one
that the discharge is “patchy” across the entire surface. can use is shown in Fig. 10. The key to building an applicable model
 592 ENLOE ET AL.

Fig. 10 Lumped-element circuit model of the plasma actuator takes

into account the charging of the dielectric surface as a virtual electrode.

is understanding that in addition to the two physical electrodes in

the actuator the exposed surface of the dielectric acts as a virtual
electrode as it collects charge. Therefore there are three capacitive
elements in the circuit model shown in Fig. 10. The capacitor C1
represents the capacitance between the exposed physical electrode
Fig. 11 Positive- and negative-sawtooth voltage waveforms were ap-
and the virtual electrode on the dielectric surface. The capacitor C2 plied to the plasma actuator to investigate the effects of discharge asym-
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

represents the capacitance between the virtual electrode and the en- metry.
capsulated physical electrode. Because the electrodes are offset, it is
also necessary to include a capacitance C3 because some field lines
connect the physical electrodes directly. (This capacitance provides
a parallel path for additional displacement current in the circuit, but
does not affect the discharge itself.)
Because, as we have shown in the preceding section, the chord-
wise extent of the plasma changes during the discharge, the values
of C1 and C2 will as well; hence, they are indicated in Fig. 10 as
variable elements. It is useful to consider the average capacitance
values for these elements and to realize that this average depends
on the amplitude of the applied voltage.
The plasma, shown as a resistance R1 in the circuit model, is the
single dissipative element in the circuit. The plasma does not exist
during the entire discharge, and so we indicate R1 as a variable re-
sistance value. When the absolute value of the potential difference
across C1 exceeds a threshold value, the plasma ignites, and the re-
sistance R1 drops from an effectively infinite, open-circuit value, to
a low value. When the absolute value of the potential difference falls
below another threshold, the discharge quenches, and R1 returns to
its open-circuit value. The voltage source VAC must be, by the nature Fig. 12 Positive- and negative-sawtooth current waveforms were ap-
of the DBD plasma, an ac source in order for the discharge to be plied to the plasma actuator to investigate the effects of discharge asym-
sustained. metry.
Knowing that the spatial structure of the plasma actuator dis-
charge is asymmetric, we investigated the importance of this asym-
metry by applying two different asymmetric voltage waveforms,
mirror images of each other, to the plasma. Both were sawtooth
waveforms, in one case the positive sawtooth, where the voltage
applied to the exposed electrode had a large positive slope and a
smaller negative slope. The negative sawtooth had its faster transi-
tion when negative-going and its slower when positive going. We
monitored voltage and current waveforms simultaneously and inte-
grated the power dissipated in the plasma directly from those wave-
forms. (Each averaged over a number of cycles to average out the
noise, as shown in Fig. 8.) Figures 11 and 12 show the voltage and
current waveforms, respectively. On the gross scale, the light emis-
sion from the plasma in each case (shown in Figs. 13 and 14) seem to
reflect the fact that the shape of the positive- and negative-sawtooth
waveforms are essentially the same. If we look in detail, however,
we see that the asymmetry of the discharge noted earlier also appears
in these measurements. For each waveform, the negative-going por-
tion of the waveform (Figs. 13b and 14a) produces the more uniform Fig. 13 Light emission from the plasma actuator for the case of the
discharge. The positive-going portion (Figs. 13a and 14b) produces positive sawtooth applied voltage waveform.
the more irregular discharge, consistent with the results shown in
Fig. 5. One often discussed (but rarely referenced) theory of the opera-
The importance of the difference in the structure of these two tion of the plasma flap attributes its effect to heating of the air. If
plasmas is evident when we measure the effect that each has on the this theory is correct, then either polarity of the sawtooth waveform
surrounding air. We gauge the actuator’s effectiveness by measur- should be equally effective, given the same average power dissipated
ing the thrust it produces when operated in initially still air. The by the plasma. In fact, this is not the case. Figure 16 shows thrust vs
arrangement used to make this measurement is shown in Fig. 15. dissipated power for both the positive- and negative-sawtooth wave-
The actuator is mounted on a lever arm, and the thrust it produces forms. As the figure shows, there is a considerable difference be-
is measured on a mass balance at the opposite end of the arm. tween the two waveforms. The positive-sawtooth waveform, which
 ENLOE ET AL. 593

Fig. 14 Light emission from the plasma actuator for the case of the Fig. 17 Power dissipated in the plasma goes as VAC 7/2 , indicating that
negative sawtooth applied voltage waveform.
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

the average value of the lumped circuit elements depends on the voltage
applied to the circuit.

C1 ) it forms part of a voltage divider. The impedance Z 2 of the

other element of the divider, the capacitance C2 , depends on the
frequency of the applied waveform Z 2 = −i/ωC2 , but for a fixed
2
frequency we would expect power dissipated to go as VAC if C2
is constant. As Fig. 17 shows, however, power dissipated in the
7/2
plasma goes approximately as VAC . This is consistent with one or
both of two situations: 1) an average capacitance C2 that increases
with increasing applied voltage or 2) an average resistance R1 that
decreases with increasing applied voltage. As we have seen from
the preceding section, the higher the voltage, the faster the plasma
sweeps over the surface of the dielectric, meaning that the average
area of the virtual electrode atop the dielectric, and hence the value
of the capacitance C2 , increases with increasing VAC . In numerical
Fig. 15 Effect of the plasma actuator on still air is determined by simulations of the circuit, to reproduce the output waveforms with
measuring the thrust it produces with a mass balance.
reasonable fidelity it is necessary to introduce a variable resistance
that decreases with voltage just as these power traces imply. We
intend to investigate possible techniques27 to directly diagnose the
temperature and density of the plasma to further confirm this issue.

Electric Forces on the Plasma Actuator

The direct measurement of thrust from the plasma actuator, al-
though a simple measurement, is instructive in terms of the mecha-
nism involved. To measure thrust on the mass balance, as shown in
Fig. 15, there must be a mechanical coupling between the moving
air and the actuator. Because this coupling only occurs when the
plasma is present, we can infer that the plasma is the intermedi-
ary. The way that the plasma can couple force into the actuator is
via the electric field interactions with the charged particles in the
plasma. Essentially, the charges in the plasma “push” on both the
background gas and the image charges in the electrodes, completing
the chain of forces leading to a measurement of thrust.
We have asserted that although the structure of the plasma is
different in each half-cycle of the DBD discharge in the plasma ac-
tuator it is not this asymmetry that appears to drive the direction of
Fig. 16 Thrust vs dissipated power for both positive- and negative- the induced airflow, as Shyy has suggested.26 To further test this,
sawtooth voltage waveforms applied to the plasma actuator.
we applied sawtooth waveforms to a different configuration of elec-
trodes, as shown in Fig. 18. In this case, the electrodes were made
has a higher negative-going duty cycle and therefore produces a of insulated magnet wire, so that unlike the configuration shown in
more diffuse plasma for a greater fraction of the discharge cycle, Fig. 1 neither electrode was the preferential electron emitter. The
produces the greater thrust. The negative-sawtooth waveform, in geometric asymmetry in the arrangement, however, was maintained.
contrast, produces a more irregular plasma for a greater fraction With both electrodes encapsulated, this arrangement of the plasma
of the discharge cycle and is less efficient in coupling momentum actuator was much less efficient at producing plasma and therefore
into the flow for a comparable dissipated power. Therefore, we can at producing thrust, and so it was not feasible to measure thrust
disregard bulk heating as the primary mechanism of the operation directly. Instead, we used smoke as a flow-visualization tool, with
of the plasma flap. Finally, we note that the power dissipation as the results shown in Fig. 19. As the figure shows, the direction of
a function of amplitude of the applied voltage is consistent with the induced airflow was the same, to the right, regardless of the
the morphology of the plasma previously noted. As Fig. 10 shows, polarity of the waveform. Therefore, it was clearly the geometry of
when the plasma ignites (effectively shorting out the capacitance the electrodes that determined the direction of the flow.
 594 ENLOE ET AL.

4 List, J., Byerley, A., McLaughlin, T., and VanDyken, R., “Using Plasma
Actuator Flaps to Control Laminar Separation on Turbine Blades in a Linear
Cascade,” AIAA Paper 2003-1026, Jan. 2003.
5 Huang, J., Corke, T., and Thomas, F., “Plasma Actuators for Separation
Control of Low Pressure Turbine Blades,” AIAA Paper 2003-1027, Jan.
2003.
6 Corke, T. C., Jumper, E. J., Post, M. L., Orlov, D., and McLaughlin, T. E.,
“Application of Weakly-Ionized Plasmas as Wing Flow-Control Devices,”
Fig. 18 Plasma actuator with both electrodes encapsulated was fab- AIAA Paper 2002-0350, Jan. 2002.
ricated to investigate whether the asymmetry in the discharge was the 7 Roth, J. R., Sherman, D. M., and Wilkinson, S. P., “Electrohydrody-
driver in the direction of the airflow. namic Flow Control with a Glow-Discharge Surface Plasma,” AIAA Journal,
Vol. 38, 2000, pp. 1166–1172.
8 Corke, T. C., and Matlis, E., “Phased Plasma Arrays for Unsteady Flow
Control,” AIAA Paper 2000-2323, Jan. 2000.
9 Roth, J. R., Sherman, D. M., and Wilkinson, S. P., “Boundary Layer Flow
Control with a One Atmosphere Uniform Glow Discharge Surface Plasma,”
AIAA Paper 98-0328, Jan. 1998.
10 Enloe, C. L., “Optical and Electrical Measurements of a Highly Asym-
metric Dielectric Barrier Discharge for Flow Control on Aerodynamic Sur-
faces,” American Physical Society, 55th Annual Gaseous Electronics Con-
ference, Paper QWP 87, Oct. 2002.
11 Vidmar, R. J., and Stalter, K. R., “Air Chemistry and Power to Generate
Downloaded by STANFORD UNIVERSITY on February 28, 2013 | http://arc.aiaa.org | DOI: 10.2514/1.2305

and Sustain Plasma: Plasma Lifetime Calculations,” AIAA Paper 2003-1189,

Jan. 2003.
12 Mangolini, L., Orlov, K., Kortshagen, U., Heberlein, J., and
Fig. 19 Airflow is induced in the same direction—as shown, in this
case, by smoke flowing to the right—irrespective of the polarity of the Kogelschatz, U., “Study of Different Discharge Regimes in a Dielectric Bar-
applied waveform in the case of the actuator with both electrodes en- rier Discharge: Electrical and Optical Characterization,” American Physical
capsulated. In this figure the electrodes are arranged vertically. Society, 55th Annual Gaseous Electronics Conference, Paper NR2.003, Oct.
2002.
13 Kang, W. S., Kim, Y., and Hong, S. E., “Spatio-Temporal Images of

With such an asymmetric arrangement of the electrodes, the elec- Single Streamer Propagation in Dielectric Barrier Discharge,” IEEE Trans-
actions on Plasma Science, Vol. 30, 2002, pp. 166, 167.
tric field will similarly be highly structured even in the absence of 14 Gherardi, N., and Massines, F., “Mechanisms Controlling the Transi-
plasma. Because of the mobility of the charges (ions and electrons) tion from Glow Silent Discharge to Streamer Discharge in Nitrogen,” IEEE
once the plasma ignites, the plasma will further enhance asymme- Transactions on Plasma Sciences, Vol. 29, 2001, pp. 536–544.
tries in the electric field structure. In general, the effect of having the 15 Liu, S., and Neiger, M., “Excitation of Dielectric Barrier Discharges by
plasma present requires a detailed calculation, but some insight can Unipolar Submicrosecond Square Pulses,” Journal of Physics D: Applied
be gleaned by considering a specific case of an asymmetric electrode Physics, Vol. 34, 2001, pp. 1632–1638.
16 Kunhardt, E. E., “Generation of Large-Volume, Atmospheric-Pressure,
arrangement in the presence of a plasma.
Nonequilibrium Plasmas,” IEEE Transactions on Plasma Science, Vol. 28,
2000, pp. 189–200.
Conclusions 17 Gibalov, V. I., and Pietsch, G. J., “The Development of Dielectric Barrier

Based on electrical and optical measurements of the plasma, the Discharges in Gas Gaps and on Surfaces,” Journal of Physics D: Applied
aerodynamic plasma actuator is clearly identified as a dielectric Physics, Vol. 33, 2000, pp. 2618–2636.
18 Steinle, G., Neundorf, D., Hiller, W., and Pietralla, M., “Two-
barrier discharge. The discharge exhibits gross structure both in
Dimensional Simulation of Filaments in Barrier Discharges,” Journal of
space and time. Because this structure clearly affects the efficiency Physics D: Applied Physics, Vol. 32, 1999, pp. 1350–1356.
of momentum coupling into the neutral air, bulk heating can be 19 Massines, F., Rabehi, A., Decomps, P., Ben Gadri, R., Segur, P., and
discounted as a mechanism for this interaction. The fact that the Mayoux, C., “Experimental and Theoretical Study of a Glow Discharge at
asymmetry in the discharge does not, however, control the direction Atmospheric Pressure Controlled by Dielectric Barrier,” Journal of Applied
of the momentum coupling indicates that an interaction of the plasma Physics, Vol. 83, 1998, pp. 2950–2957.
20 Xu, X., and Kushner, M. J., “Multiple Microdischarge Dynamics in
with the applied electric field in the discharge is responsible for
the body force and subsequent momentum transfer to the neutral Dielectric Barrier Discharges,” Journal of Applied Physics, Vol. 84, 1998,
fluid through plasma-neutral collisions. The strong influence that the pp. 4153–4160.
21 Li, J., and Dhali, S. K., “Simulation of Microdischarges in a
structure of the plasma has on the momentum transfer suggests that
Dielectric-Barrier Discharge,” Journal of Applied Physics, Vol. 82, 1997,
a detailed, multidimensional model of the actuator will be necessary pp. 4205–4210.
to unwrap the physics of the problem and enable accurate predictions 22 Falkenstein, Z., and Coogan, J. J., “Microdischarge Behavior in the
of its performance. Silent Discharge of Nitrogen-Oxygen and Water-Air Mixtures,” Journal of
Physics D: Applied Physics, Vol. 30, 1997, pp. 817–825.
23 Pashaie, B., Dhali, S. K., and Honea, F. I., “Electrical Characteristics
Acknowledgments
of a Coaxial Dielectric Barrier Discharge,” Journal of Physics D: Applied
The authors gratefully acknowledge the support provided by John Physics, Vol. 27, 1994, pp. 2107–2110.
Anttonen of the Munitions Directorate of the U.S. Air Force Re- 24 Wharburg, E., Annals of Physics, Vol. 13, 1904, pp. 464–476.
25 Roth, J. R., Tsai, P. P., Liu, C., Laroussi, M., and Spence, P. D., “One At-
search Laboratory, Eglin Air Force Base, Florida.
mosphere, Uniform Glow Discharge Plasma,” U.S. Patent 5,414,324, 1995.
26 Shyy, W., Jayaraman, B., and Anderson, A., “Modeling of Glow-
References Discharge-Induced Fluid Dynamics,” Journal of Applied Physics, Vol. 92,
1 Wilkinson, S., “Investigation of an Oscillating Surface Plasma for Tur- 2002, pp. 6434–6443.
27 Park, J., Henins, I., Herrmann, H. W., and Selwyn, G. S., “Neutral
bulent Drag Reduction,” AIAA Paper 2003-1023, Jan. 2003.
2 Post, M., and Corke, T., “Separation Control on High Angle Bremsstrahlung Measurement in an Atmospheric-Pressure Radio Frequency
of Attach Airfoil Using Plasma Actuators,” AIAA Paper 2003-1024, Discharge,” Physics of Plasmas, Vol. 7, 2000, pp. 3141–3144.
Jan. 2003.
3 Ashpis, D., and Hultgren, L., “Demonstration of Separation Delay with M. Sichel
Glow Discharge Plasma Actuators,” AIAA Paper 2003-1025, Jan. 2003. Associate Editor