Miniaturized Argon Plasma: Neutral Gas Characteristics in Dielectric Barrier Discharge

Transcription

1 Miniaturized Argon Plasma: Neutral Gas Characteristics in Dielectric Barrier Discharge Ashraf FARAHAT 1,2 1 Department of Prep Year Physics, College of Applied and Supporting Studies, King FahdUniversity of Petroleum and Minerals (KFUPM), Dhahran 31261, Saudi Arabia 2 Department of Physics, Faculty of Science, Moharam Beek, Alexandria University, Egypt Abstract Plasma-neutral gas dynamics is computationally investigated in a miniaturized microthruster that encloses Ar and contains dielectric material sandwiched between two metal plates using a two-dimensional plasma mode. Spatial and temporal plasma properties are investigated by solving the Poisson equation with the conservation equations of charged and excited neutral plasma species using the COMSOL Multiphysics 4.2b. The microthruster property is found to depend on the secondary electron emission coefficient. The electrohydrodynamic force (EHD) is calculated and found to be significant in the sheath area near the dielectric layer and is found to affect gas flow dynamics including the Ar excimer formation and density. The effects of pressure and secondary emission coefficient are discussed. The plasma characteristics are affected by small changes in the secondary electron emission coefficient, which could result from the dielectric erosion and aging, and is found to affect the electrohydrodynamic force produced when the microthruster is used to produce thrust for a small spacecraft. Keywords: plasma modeling, plasma species, dielectric barrier discharge PACS: Kn, Dg, Mq, s, y DOI: / /17/10/08 (Some figures may appear in colour only in the online journal) 1 Introduction Miniaturized microdischarger devices (MD) operate at atmospheric pressure with a pd 10 Torr cm (pressure interelectrode length scale). If a large enough potential difference is applied to a gas, it can develop a small ionization current that can be significantly amplified and can endure a gas transition from an electrically insulating state to a conducting state [1 3]. In its basic form, a microdischarger is composed of a few hundreds of micrometer gap confined with 100-1,000 Torr helium or argon gas [4,5]. The gas discharge can be initiated using a magnetic or an electric field using a current carrying wire coiled around the MD or opposite voltage plates with a potential difference of a few hundred volts. This article focuses on using electric fields where the incidence of positive ions on the cathode plays a major role in maintaining the discharge process. The initial charged active particles, normally electrons, disappear shortly after the initiation of the discharge [6 8]. Secondary electron emission due to the incidence of photons, metastable, and excited atoms on the cathode and the dielectric surface also plays a role in the development of the discharge process because the potential energy associated with excited and metastable atoms may be used to eject secondary electrons from metal and dielectric surfaces, or may cause ionization or excitation of other gas molecules. However, in the geometry used in this work, the cathode is totally covered with a dielectric material, and it does not contribute to the production of secondary electrons. The discharge process comprises the production and loss mechanisms of several particles species in pure gas, including ionization, excitation, recombination, attachment, detachment, and ionization by positive ions. For a mixed gas, other processes such as photoionization and penning effects take place [9]. The small dimensions of the microdischargers make it appropriate for space propulsion applications. Several researchers have investigated different possibilities of building microdischargers for small satellites [10]. The small geometric dimensions of plasma interactions at a dielectric boundary have limited detailed quantitative experimental diagnostic studies of their properties, though some examples of such studies have been reported [11 13]. Computational modeling studies of surface dielectric barrier discharge (DBD) have been recently reported [14 16]. The electronic excitation temperature of a surface DBD at atmospheric pressure has been experimentally investigated by optical emission spectroscopic measurements combined with numerical simulation for seven Ar atomic emission lines in Ref. [17]. A two-dimensional 853

2 fluid simulation to model the plasma dynamics in surface DBD operating in air is discussed in Ref. [18]. The EHD force exerted on the gas molecules in a DBD has been studied by Boeuf and Pitchford [19] with the help of a 2-D fluid model of the discharge in a nitrogen-like gas at atmospheric pressure. The EHD force has many applications, including microactuators and flow control over airfoils for aerodynamic applications [20]. It is also used to modify the airflow profile over boundary layers of aerodynamic surfaces to reduce drag and to stabilize the flow. This paper discusses the development of plasma species in Ar discharge near a dielectric surface. The temporal development of the species number densities at various gas pressure is calculated. Electron temperature is calculated, and an estimation of the EHD is presented. The simulation is performed using the COMSOL Multiphysics 4.2b software and built on the DBD flow control model developed by COMSOL for the basic geometry and meshing. The DBD flow control model is also based on the models of Boeuf and Pitchford [19] and Punset et al [15]. This work investigates new neutral gas plasma parameters that have not been investigated in previous work on DBD flow control, including a detailed analysis of the development of plasma species such as electron number density, ion number density, electron temperature, metastable species Ar(4s), Ar(4p), and dimer metastable species Ar 2. In addition, an estimation of the electrohydrodynamic forces within the microthruster is provided. The effect of pressure within the thruster is investigated. An investigation into the optical and UV emission within the microthruster and the effect of changing the secondary emission coefficient on the potential, which could happen because of the aging of the microthruster, are also provided. The new neutral gas plasma parameters presented in this paper provide broader information with regard to microthrusters characteristics. The detailed investigation of charged and neutral species presented in this study is a significant step toward the understanding of possibility and constraints of DBDs flow control for EHD applications. 2 Quantitative method The model applied in this study is similar to the direct current surface discharge models described by Novak et al. [8], Thomas et al. [14], Jugroot [21], Farahat and Ramadan [22], and the numerical model of surface DBD [19,23 26]. The basic geometry and meshing are based on the surface DBD model developed by COM- SOL. The gas is assumed to be pure argon with an initial electron density of m 3. The relative permittivity of the dielectric material, ε r, is equal to 5, and the initial mean electron energy is 3 ev. Walls are assumed to have a zero reflection coefficient and a zero thermal emission flux. To initiate the discharge, a positive electric potential is applied at the anode. 2.1 Equations The conservation equations for electrons, ions, and excited species are presented as follows: n ( ) e t + n e v ie De n e = n e v i kn e n i, (1) n i t + ( n i µ i E Di n i ) = n e v i kn e n i, (2) n ex (D ex n ex ) = n e v ext k rm n m k 2m n m n 2 t 0, (3) where n s is the number density of species s, E is the electric field component normal to the anode, v i and v ext are the ionization and excitation frequencies, k is the recombination rate, µ s is the species mobility, D s is the species diffusion coefficient, n 0 is the neutral gas density, and k rm n m and k 2m n m n 2 0 represents excited atoms loss due to the escape via radiation and the conversion of metastable atom to metastable molecule. The applied electric field is described by the following Poisson equation [1] : E = 1 ε o ε r [e (n i n e ) + ρδ DB ], (4) where ρ, the surface charge density, is nonzero only on the dielectric surface as expressed by the onedimensional Dirac function δ DB concentrated at the dielectric surface. The neutral gas plasma mixture density is assumed to satisfy the ideal gas law, P = n n K B T g + n e K B T e + n m K B T g, (5) m n,m e where P is the discharge total pressure, n n, T g, T e, and K B are the neutral gas density, gas temperature, electron temperature, and the Boltzmann constant, respectively. 2.2 Limiting and wall conditions The discharge is initiated and sustained by an electric potential V applied to the anode boundaries, and the dielectric surface is charge-free before the discharge. The lower boundary of the dielectric material is grounded. The cathode is completely covered by the dielectric material, and the secondary emission at the dielectric surface is only responsible on sustaining the electron cloud within the thruster and is calculated according to the following equation: Γ s ˆn = γ e ( Γe ˆn ), (6) where γ e is the secondary emission coefficient, ˆn is the unit vector normal to the dielectric, and at the discharger s walls. Γ s ˆn = 0, (7) 854

3 Ashraf FARAHAT: Miniaturized Argon Plasma: Neutral Gas Characteristics in DBD The potential boundary condition at the dielectric surface is determined using the total surface charge accumulated on the surface. The conservation equation for surface charge density ρ is given by ρ t = J e ˆn + J e ˆn, (8) where the right-hand side represents the normal component of the current densities of total electrons and ions. No-slip boundary conditions are applied on the solid walls. 2.3 Plasma chemistry Plasma reactions include electron elastic and inelastic collisions with the neutral gas. Reaction chemistry includes electrons, ions (Ar + ), and excited atoms Ar(4s) and Ar(4p) along with the neutral argon species (Ar). The chemical reactions considered in this study are listed in Table 1 together with their types, energy loss, and approximate momentum-transfer collision cross sections of the plasma species with the background gas [27]. These cross sections are derived from experimental mobility data for electrons and ions [28,29] and from the Leonard-Jones intermolecular potential data for neutrals [30]. The effect of impurities [31] is disregarded for their very small effects on microdischarges at high plasma species densities. Seed electrons are necessary to start the reaction, but they disappear after a few nanoseconds. 2.4 Numerical solution Using COMSOL Multiphysics 4.2b, Eqs. (1)-(8) are solved for a 2-D Cartesian geometry using the numerical method described by Punset et al [23]. A basic geometry and meshing developed by COMSOL for the DBD flow control model is used in the simulation where the continuity equations are solved using the Crank- Nicolson scheme with the Scharfetter-Gummel [32,33] exponential representation of the charged particle fluxes. The Poisson equation is solved with a successive over-relaxation method (SOR). A semi-implicit time integration method similar to the one applied by Punset et al. [15] is used, where the electric field in the integration of the continuity equation between t and (t + t) is not the field at time t but rather a prediction of the field at time (t+ t). The method is described in detail by Punset et al. [16], but it will be summarized below. The Poisson equation can be written as ( E tk+1) = e [ ( t n k t i n k) e + t (n ] i n e ), ε 0 ε r t (9) where t k+1 = t k + t. The third term at the right-hand side is presented to correct the electric field variation between t k and t k+1 and can be written using the drift-diffusion terms in Eqs. (1) and (2) as e t (n i n e ) t = e t [(n tk e µ tk e (D e tk n tk e D tk i + n tk i ) µ tk i E tk+1 )] n tk i. (10) The charged particle fluxes in Eq. (10) above are discretized using the Scharfetter-Gummel scheme and linearized assuming a small variation of the potential between t k and t k+1. After putting the terms depending on V k+1 on the left hand side of Eq. (9), one obtains a modified Poisson equation, which can be solved with an SOR method. A time step of ns was used in the simulation. A commercial code COMSOL Multiphysics 4.2b is used to solve the continuity and the field equations using the aforementioned numerical method. The model is set up to use a Maxwellian energy distribution function for all neutral species [28]. All ions and excited species are assumed to be returned to the gas as neutral species after colliding with walls. Table 1. Plasma chemistry Reaction Type Energy loss σ (Å2 ) ε (ev) 1 e + Ar e + Ar Elastic e + Ar e + Ar(4s) Excitation e + Ar(4s) e + Ar Quenching e + Ar e + Ar(4p) Excitation e + Ar(4p) e + Ar Quenching e + Ar 2e + Ar + Ionization e + Ar(4s) 2e + Ar + Ionization Ar + Ar(4s) Ar + Ar Reaction 7 Ar(4s)+ Ar(4s) Ar(4s) + Ar Reaction 8 Ar(4s) + Ar + Ar Ar 2*+Ar Reaction 9 Ar + + Ar + Ar Ar + 2 +Ar Reaction 855

4 3 Results 3.1 Microdischarger geometry and plasma conditions The geometry used in this study, as schematically shown in Fig. 1, is a two-dimensional microdischarger, including a dielectric material separating parallel metallic electrodes. The top electrode is the anode, and the bottom one is the cathode. All the microdischarger boundaries are metal, except the dielectric material. The microdischarger length and height are 0.4 mm and 0.2 mm, respectively, with a 0.1 mm anode embedded in a dielectric material such that the anode and the dielectric material are directly exposed to the gas. The dielectric thickness is 0.05 mm, and the bottom boundary of the dielectric material is grounded as shown in Fig. 1. A nonuniform, triangular mesh with 4,508 elements, 267 edges, and 9 vertex elements is used and verified for solution grid convergence. The mesh area is 79,500 µm 2. Fig.1 Geometry used for the microdischarger simulations The laminar gas flow velocity is considered very small and, hence, disregarded in this study, with a gas density of 1.66 kg/m 3 (i.e., there is no gas inlet or outlet). All the solid surface temperatures and the inlet gas temperature are fixed at 300 K. A time-dependent voltage is applied to the anode, and gas pressure levels of 100 Torr, 300 Torr, 500 Torr, 760 Torr, and 1,000 Torr are considered. Considering the applied gas pressure and the interelectrode distance, the discharge operates in transient glow mode. The gas used is argon, and the species considered in the model are Ar(4s), Ar(4p), Ar +, Ar 2, and Ar + 2. Secondary electron emission caused by the ions impact is considered at the dielectric surface with a probability of The low coefficient was chosen to make the discharge operate in a transient rather than a streamer glow mode [34] but still in good agreement with the values used by Novak and Bartnikas [6,7] and Wang et al [35]. The mean energy of secondary electrons is set to be 5 ev. A time-dependent voltage scheme was chosen to match the practical experiment of DBD where a sinusoidal voltage wave is used with frequencies between 1 khz and 10 khz. The applied voltage then depends on the amplitude of the voltage and its frequency. At relatively low frequencies, the discharge operates at Townsend regime and at glow regime at a relatively high frequency-dependent voltage. In our simulation, we will be using a hyperbolic tangent frequencydependent voltage, [1200tanh( )t] V, for computational convenience and to make the microdischarger operate at almost a single constant voltage pulse after a few time steps such that it operates in a glow regime. The thickness of the dielectric layer used in the simulation is smaller than those normally used in experiments, but increasing the dielectric thickness will mainly increase the operating voltage, and this can be compensated by the value of the operating voltage used and also the electron secondary emission coefficient used (0.05). In a real experiment, the effect of charge accumulation on the dielectric layer will result in an increase of the operating voltage from one pulse to a previous pulse, but using the hyperbolic tangent frequency-dependent voltage allows us to study discharge conditions corresponding to a current pulse. 3.2 Incipient plasma discharge The electric potential of the microdischarger is shown in Fig. 2 for the base case of 760 Torr at 20 ns, 40 ns, and 80 ns. Plasma properties including electron density and electron temperature at 40 ns and 80 ns are shown in Fig. 3. The densities of the argon monomer ion Ar + and dimer ion Ar + 2 and the densities of the excited states of argon including Ar(4s), Ar(4p), and Ar 2 at 40 ns and 80 s are shown in Figs. 4 and 5. The applied potential on the anode ignites the discharge process. Initial electrons disappear after a few mean free paths because of collision with gas atoms and the discharger s walls, and more electrons are produced by secondary emission from the dielectric surface to sustain the reaction. Secondary emission is the major source of producing electrons. As ions move along the dielectric surface, more surface charges are accumulated on the surface, which results in a virtual anode extension. The microdischarger design with a grounded cathode at the bottom boundary allows a charged species cloud to move parallel to the dielectric material as ions accumulate on the dielectric surface, making it work as an extension to the anode and attracting more electrons to interact with the gas to produce more ions and so on. The plasma discharge parallel to the dielectric surface is pseudo-neutral with a non-neutral tip (ion sheath) moving along the surface and elongating the plasma column. More charges are accumulated on the surface as electrons and ions move toward the microdischarger s boundary. The plasma propagation characteristic is similar to an experimental study [36,37]. The maximum electron temperature is ev reducing to 1.03 near the anode at 40 ns and ev reducing to 1.63 near the anode at 80 ns. This high electron temperature is produced because of the large metastable Ar(4s) density ( m 3 ). The calculated metastable densities are in good agreement with experimental measurements [16]. 856

5 Ashraf FARAHAT: Miniaturized Argon Plasma: Neutral Gas Characteristics in DBD Fig.2 Plasma characteristics for a microdischarger at 760 Torr; development of electric potential at (a) 20 ns, (b) 40 ns, and (c) 80 ns (contour labels in volts). The microdischarger s dimensions are the same as in Fig. 1. Fig.3 Plasma characteristics for a microdischarger at 760 Torr; (a) electron density at 40 ns, (b) electron density at 80 ns, (c) electron temperature at 40 ns, and (d) electron temperature at 80 ns. The maximum value of each frame is denoted. Contour label is in cubic meters in frames a and b and in electron volts in frames c and d. The microdischarger s dimensions are the same as in Fig. 1 Fig.5 Plasma characteristics for a microdischarger at 760 Torr; (a) Ar(4s) at 40 ns, (b) Ar(4s) at 80 ns, (c) Ar(4p) at 40 ns, (d) Ar(4p) at 80 ns, (e) Ar 2 at 40 ns, and (f) Ar 2 at 80 ns. The maximum value of each frame is denoted. The contour label is in cubic meters in the frames. The microdischarger s dimensions are the same as in Fig. 1 The photoemission and the photoionization resulting from photon emission and transport from argon excited states are disregarded as we found that they will not be critical in our calculations; however, their effects should be taken into consideration if a large potential gradient is produced in the plasma and significantly altered the photoemission properties. These effects were not considered in this study. Initial electron and ion densities are assumed to be uniform in the volume and equal to m 3. The ion bombardment on the dielectric releases secondary electrons with a coefficient equal to The combined sources from initial and beam electrons produces a maximum electron density of m 3 at 80 ns near the dielectric. The density of visible photon emitters Ar(4p) and UV photon excimer emitter (Ar 2) are shown in Fig. 5. The emitters density develops near the dielectric surface in a similar style to the plasma density. The Ar 2 is formed by three-body collisions of slow quenched metastable Ar(4s), and it is more widely distributed near the anode where higher Ar(4s) density exists. The lifetimes of the visible emitter are a few hundred nanoseconds, and they are expected to quench faster after 80 ns. Fig.4 Plasma characteristics for a microdischarge at 760 Torr; (a) Ar + at 40 ns, (b) Ar + at 80 ns, (c) Ar + 2 at 40 ns, and (d) Ar + 2 at 80 ns. The maximum value of each frame is denoted. The contour label is in cubic meters in the frames. The microdischarger s dimensions are the same as in Fig Plasma-neutral gas interactions Plasma discharge in the microdischarger not only includes charged and excited species transport but also involves momentum and energy transfer to the gas through collision. The net force F from charged species 857

6 (electrons and ions) is calculated based on the method described by Pitchford [10] and is briefly described here, F = s n s m s v ds us, (11) where n s is the charged particle number density, m s their masses, v ds the momentum-transfer frequencies for charged particle-neutral collisions, and u s is their mean velocities. The net force per unit volume on the neutral gas molecules can also be written in terms of charged particle current densities j s and mobilities µ s as F = s j s µ s, (12) where the current densities of charged species are given by j s = en s µ s E eds n s, (13) where D s is the charged particles diffusion coefficients. Using Eq. (12) in (13), we obtain the net force per unit volume as F = e (n i n e ) E kt i ni kt e ne. (14) This force is produced by diffusion and the drift of the charged particles. This force described in Eq. (14) is not constant and can vary over the discharge region. In the case of perfectly neutral plasma, the force produced by electrons will be exactly balanced by the ions (n i = n e and n i = n e = 0). This does not contradict the fact that the force depends on the particles masses, because the rate of electron-neutral collisions is much larger than that of ion-neutral collision, and this produces a much higher electron flux in the positive column. This force equality will be true only if we assume that the force is produced by the drift-diffusion term in Eq. (13). Including other factors for electron-ion generation will result in a nonzero force; however, we consider it insignificant for this study. If the electron flux and the ion flux are equal, and electrons and ions are moving in the same direction, a pseudo-neutral, nonuniform plasma is formed. As ion mass is much larger than electrons, we can ignore the force on gas molecules due to electrons collisions. Another non-neutral region can be formed in the discharge in the case where the force produced by the electric field terms is much larger than that produced by the density gradient. In this region, the ion number density is much larger than the electron number density, and the force will be significant only if the ion current density is large. Fig. 6 shows the force produced by the ion interaction with the neutral gas at 760 Torr at 50 ns and 70 ns. The largest component of the force is parallel to the dielectric surface. The effect of this force on the gas molecules is based on the ion sheath speed (2 mm/µs), and with a sheath length of approximately 26.9 µm, the force will be acting on the gas molecules for approximately 13.4 ns. Fig.6 Calculated force field with force direction for the used geometry at 760 Torr at (a) 50 ns and (b) 70 ns; arrow lengths are normalized. The contour label is in Newtons per cubic meter in the frames. The microdischarger s dimensions are the same as in Fig. 1 This results in a time-integrated force per unit volume of approximately 8 N/(m 3 s). This force can be used as a thrust in many applications, e.g., for a small spacecraft. The results provide support to several recent applications and experimental studies [19,14,25]. 3.4 Effect of change in pressure Electron densities at pressure levels of 100 Torr, 400 Torr, 500 Torr, and 1,000 Torr are shown in Fig. 7, while keeping the same applied potential on the anode. The maximum electron density and anode voltage as function of pressure are shown in Fig. 8. The anode voltage is barely increasing from 100 Torr to 500 Torr and has a minimum value at 100 Torr, similar to the Paschen law behavior. Electrons are confined near the anode at 100 Torr, and they slowly progress at 400 Torr and become significant at pressure levels exceeding 500 Torr. Fig.7 Plasma characteristics for a microdischarge at different pressure levels; electron density at (a) 100 Torr, (b) 400 Torr, (c) 500 Torr, and (d) 1,000 Torr. The maximum value of each frame is denoted. Values are presented in cubic meters in the frames. The microdischarger s dimensions are the same as in Fig. 1 The maximum electron density is found to keep increasing from 100 Torr to 760 Torr and slightly decrease at 1,000 Torr because of the reduced range of secondary 858

7 Ashraf FARAHAT: Miniaturized Argon Plasma: Neutral Gas Characteristics in DBD electrons. The cathode fall represented by the gap between the dielectric and the electrons is found thicker at lower pressure because of the lower electron density and gets thinner at higher pressure levels. Fig.8 Plasma characteristics for a microdischarge at different pressure levels: (a) Maximum densities of Ar(4p) (solid) and Ar 2 (dotted), (b) Anode voltage (solid) and maximum electron density (dotted) The spatial distributions of Ar(4p) and Ar 2 for pressure levels of 100 Torr, 500 Torr, and 1,000 Torr are shown in Fig. 9. The maximum Ar(4p) density increases with pressure up to 500 Torr, then slightly decreases at 800 Torr and 1,000 Torr, which is a consequence of the electron density distribution shown in Fig. 7 and the reduced range of secondary electrons. The Ar 2 density ranges from m 3 at 100 Torr to m 3 at 1,000 Torr. The change in the density range is directly related to the Ar density, which falls as the pressure is decreased. are found away from this point due to the small effect of diffusion. The density of Ar 2 is found to have a less uniform distribution than the Ar(4p) because it is mainly produced by three-body collisions that depend on the rate of production of the Ar(4s). Also, the very short lifetime of the Ar 2 constrains its diffusion pattern along the dielectric surface. Rarefaction by gas heating is not considered because its effect will be more significant at pressure levels lower than those considered in this study. The net force produced by the momentum transfer between the ions and the neutral gas shown in Fig. 6 also facilitates the excimer formation as the experimental work [38] proves that gas flowing through a microdischarger improves the efficiency of excimer emission. The Ar(4p) spatial density distribution representing the optical visible emission is shown in Fig. 10. The emission has a peak between 100 µm and 150 µm near the excimer formation region. Experimental results for similar conditions are discussed by Penache et al. [39] and Moselhy et al [38]. Fig.10 Density of Ar(4p) for different pressure levels, measured in Torr, to describe the optical visible emission at Y =65 µm. 3.5 Secondary emission coefficient The plasma-neutral gas interaction characteristics depend on ionization produced by secondary electrons, and one should expect that changing the secondary emission coefficient will have a direct impact on the plasma behavior. Fig. 11 shows the anode potential as a function of the secondary emission coefficient. Fig.9 Plasma characteristics for a microdischarge at different pressure levels. Densities of Ar(4p) (representing optical emission) and (Ar 2) (representing UV emission) at different pressure levels: (a) 100 Torr, (b) 400 Torr, and (c) 1,000 Torr. The maximum value of each frame is denoted. The contour label is in cubic meters in the frames. The microdischarger s dimensions are the same as in Fig. 1 The maximum density of Ar(4p) and Ar 2 is directly related to the pressure in a pattern similar to the electron density. For the short lifetime of the Ar(4p), most of the Ar(4p) is confined near the point of production (anode-dielectric boundary), and lower densities Fig.11 Anode potential as a function of the secondary emission coefficient 859

8 The potential decreases as the secondary emission coefficient increases at 760 Torr. This potential change is to compensate for the rate of ionization of the electron beam flux. Similar patterns for secondary emission coefficient are reported by Kothnur et al [40]. 4 Conclusions We have analyzed a wide range of argon species, including Ar(4s), Ar(4p), Ar +, Ar 2, and Ar + 2 near a dielectric surface in a miniaturized volume. We have also demonstrated the thrust generation in DBD devices of microscale dimensions. Plasma-neutral gas characteristics in a miniaturized microthruster are computationally investigated using a two-dimensional computational model. The model is built on the basic geometry and meshing developed by COMSOL Multiphysics 4.2b DBD flow control model. The objective is to investigate plasma species development and gas dynamics. The plasma is sustained by secondary electron emission from the dielectric layer, and the plasma characteristics are affected by small changes in the secondary electron emission coefficient. The change in the coefficient could result from the dielectric erosion and aging and its effects on the magnitude of the electrohydrodynamic force produced when the microthruster is used to produce thrust for a small spacecraft. The force is effective only approximately 20 µm above the dielectric surface and at a length of approximately 50 µm. The maximum intensity of force per unit volume is of the order of N/m 3. However, the force is effective on the gas molecules only for a short period of time because of the ion sheath motion. The mean time-averaged integrated force per unit volume is approximately 8 N/(m 3 s). This is comparable with the DC corona discharges. Plasma-neutral gas momentum transfer studies agree with the numerical results given in this work. The microdischarger has the feature to operate at transient glow mode where secondary electron emission from the dielectric surface is small and the electron multiplication is relatively low; as a result, plasma-neutral gas momentum exchange is significant in affecting the gas flow dynamics, in the formation of excimer species and in affecting the UV and visible emission characteristics of the device. Operation at higher pressure levels affects the species formation and produces higher density peaks. The results provided in this paper are significant for understanding the plasmaneutral gas reactions which affect flow acceleration in a DBD. It is also a major step toward a detailed understanding of the limitations in using DBDs for EHD applications. Acknowledgments The authors would like to acknowledge the support provided by the Deanship of Scientific Research (DSR) at the King Fahd University of Petroleum and Minerals (KFUPM) for funding this work through project No. IN References 1 Davies A J, Evans C J. 1973, The Theory of Ionization Growth in Gases under Pulsed and Static Fields. CERN, Geneva, Mason E A, McDaniel E W. 1988, Transport Properties of Ions in Gases. John Wiley and Sons, New York 3 Goldston R J, Rutherford P H. 1995, Introduction to Plasma Physics. Institute of Physics Publishing, London 4 Kothnur P S, Raja L L. 2007, Contributions to Plasma Physics, 47: 9 5 Hong Yi, Lu N, Pan J, et al. 2013, Plasma Science and Technology, 15: Novak J P, Bartnikas R. 1987, Journal of Applied Physics, 62: Novak J P, Bartnikas R. 1988, Journal of Applied Physics, 64: Novak J P, Bartnikas R. 1991, IEEE Transactions on Plasma Science, 19: 95 9 Mohamed G H. 1999, Journal of Fluids Engineering, 121: 5 10 Yoshinori T, Takeshi T, Koji E, et al. 2008, Pure and Applied Chemistry, 80: Kohelschatz U, Eliasson B, Egli W. 1997, Dielectric- Barrier Discharges. Principle and Spplications. Journal Physique IV, 07(C4), p.c4-47-c Meyer C, Franzke J, Gurevich E L. 2012, Journal of Physics D: Applied Physics, 45: Qu G, Liang D, Qu D, et al. 2014, Plasma Science and Technology, 16: Thomas D, Shankar M, Laxminarayan L R. 2007, IEEE Transactions on Plasma Science, 35: Punset C, Boeuf J P, Pitchford L C. 1998, Journal of Applied Physics, 83: Punset C, Cany S, Boeuf J P. 1999, Journal of Applied Physics, 86: Zhang Y, Li J, Lu N, et al. 2013, Plasma Science and Technology, 15: Hoskinson A R, Hershkowitz N. 2011, Journal of Physics D: Applied Physics, 44: Boeuf J P, Pitchford L C. 2005, Journal of Applied Physics, 97: Khabiry S E, Colver G. M. 1997, Phys. Fluids, 9: Jugroot M. 2009, Journal of Applied Physics, 105: Farahat A, Ramadan E. 2014, Plasma Physics Reports, 40: Punset C, Cany S, Boeuf J P. 1999, Journal of Applied Physics, 86: Corke T C, Post M L, Orlov D M. 2009, Experiments in Fluids, 46: 1 25 Gerardus H. 2000, Modeling of Microdischarges for Display Technology [Ph.D]. Eindhoven University of Technology, Eindhoven 26 Alexandre C. 2003, Self-Organization of Microdischarges in DBD Plasma [MS]. Drexel University, Philadelphia 860

9 Ashraf FARAHAT: Miniaturized Argon Plasma: Neutral Gas Characteristics in DBD 27 Raizer Yu P. 1991, Gas Discharge Physics. Springer, Berlin 28 Prashanth S, Kothnur, Raja L L. 2005, Journal of Applied Physics, 97: Ellis H W, Pai R Y, McDaniel E W et al. 1976, Atomic Data and Nuclear Data Tables, 17: Kee R J, Dixon-Lewis G, Warnatz J, et al. 1995, A FORTRAN Computer Code Package for the Evaluation of Gas-Phase, Multicomponent Transport Properties. Sandia National Laboratories, Albuquerque, Sandia Report SAND Yuan X, Raja L L. 2003, IEEE Transactions on Plasma Science, 31: Gummel H K. 1964, IEEE Transactions on Electron Devices, 11: Scharfetter D L, Gummel H K. 1969, IEEE Transactions on Electron Devices, 16: Kuffel E, Zaengl W S, Kuffel J. 2000, High Voltage Engineering Fundamentals. 2 nd ed., Newnes, Oxford 35 Wang Qiang, Economou D J, Donnelly V M. 2006, Journal of Applied Physics, 100: Alina S C, Gheorghe P. 2013, Journal of Applied Physics, 113: Andrea C, Carlo A B, Gabriele N. 2013, Journal of Applied Physics, 113: Moselhy M, Petzenhauser I, Frank K, et al. 2003, Journal of Physics D: Applied Physics, 36: Penache C, Miclea, et al. 2002, Plasma Sources Science and Technology, 11: Kothnur P S, Yuan X, Raja L L. 2003, Applied Physics Letters, 82: 529 (Manuscript received 31 August 2014) (Manuscript accepted 11 March 2015) address of Ashraf FARAHAT: farahata@kfupm.edu.sa 861