NO and NO2's Role in Efficiency and Quenching of Electrically Pumped O2-I Laser Systems, Papers of Health sciences

Download NO and NO2's Role in Efficiency and Quenching of Electrically Pumped O2-I Laser Systems and more Papers Health sciences in PDF only on Docsity! IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS J. Phys. D: Appl. Phys. 40 (2007) 4793–4809 doi:10.1088/0022-3727/40/16/009 O2(1∆) production and gain in plasma pumped oxygen–iodine lasers: consequences of NO and NO2 additives Ramesh A Arakoni1, Natalia Y Babaeva2 and Mark J Kushner2,3 1 University of Illinois, Department of Aerospace Engineering, Urbana, IL 61801, USA 2 Iowa State University, Department of Electrical and Computer Engineering, Ames, IA, 50011, USA E-mail: arakoni@uiuc.edu, natalie5@iastate.edu and mjk@iastate.edu Received 27 May 2007, in final form 29 May 2007 Published 3 August 2007 Online at stacks.iop.org/JPhysD/40/4793 Abstract The 1.315 µm [I(2P1/2) → I(2P3/2)] transition of atomic iodine in the chemical oxygen–iodine laser (COIL) is pumped by sequential reactions of I2 and I with O2(1
). In electrically pumped systems (eCOILs), electron impact excitation of O2 produces the O2(1
) and also produces O atoms through dissociative excitation. The O atoms, through reactions with I2, I(2P1/2) and I(2P3/2), lead to dissociation of I2, quenching of the upper laser level and removal of the lower laser level. While dissociating I2 is potentially beneficial, quenching of the upper laser level is detrimental and so management of the O atom density is necessary to maximize laser gain. In this regard, NO and NO2 additives have been used to manage the O atom density by cyclically reacting with O and I. In this paper, results from a computational investigation of eCOIL systems using plug flow and two-dimensional models are discussed where NO and NO2 additives are used. The system is a flowing plasma sustained in He/O2/NO mixtures with downstream injection of NO2 followed by injection of I2. We found that addition of NO and NO2 is effective in managing the density of O atoms and maximizing gain by minimizing quenching of the upper laser level. We found that by optimizing the additives, laser gain can be maximized even though O2(1
) densities may be lower due to the management of quenching and dissociation reactions. 1. Introduction Chemical oxygen–iodine lasers (COIL) operating on the 1.315 µm [I(2P1/2) → I(2P3/2)] transition of atomic iodine are being investigated due to their high efficiency and potential for multi-kilowatt CW operation [1–3]. A series of collisional transfer reactions between O2(1
), I2 and ground state I(2P3/2) (to be referred to as I) result in excitation of the upper laser level I(2P1/2) (to be referred to as I∗). Typically O2(1
) is generated upstream of the laser cavity. I2 is injected immediately prior to the cavity upon which the flow is supersonically expanded to lower the gas temperature as 3 Author to whom any correspondence should be addressed. required to maximize the gain. In conventional COILs liquid- phase chemistries (reactions between basic H2O2 and Cl2) produce the O2(1
) with high yields [4], although the use of liquid peroxides and Cl2 gas to produce the O2(1
) creates challenges for storage and transport. Recently, efforts have focused on generating the O2(1
) in electrical discharges (eCOILs) due to the increased robustness and safety of the all gas-phase system [1–10]. Laser gain and oscillation have been demonstrated in eCOILs by Hicks et al [1] and Verdeyen et al [3]. A challenge in eCOILs is to produce sufficiently high yields of O2(1
), and hence laser gain, to enable the laser transition to be saturated and produce high power [5]. In 0022-3727/07/164793+17$30.00 © 2007 IOP Publishing Ltd Printed in the UK 4793 R A Arakoni et al this regard, recent research has focused on tailoring the discharge parameters [6, 7, 10] and using additives such as CO, H2, D2 [8] and NO [9] to improve the excitation efficiency of O2(1
). NO and NO2 are also used as additives to control the post-discharge chemistry [9]. In fact, all demonstrations of laser gain and oscillation to date have used flowing plasmas in He/O2 mixtures with NO as an additive. eCOILs differ from conventional COILs in that atomic oxygen is also produced in the electric discharge by electron impact dissociation of O2. The O atoms flow downstream where 3-body reactions produce O3 and the remaining O atoms may react with the injected I2. Beneficial reactions of O with I2 produce I atoms, thereby eliminating the expense of O2(1
) molecules for initiating the dissociating reactions. Detrimental reactions involving O atoms include quenching of I∗ by O which reduces gain. As such, management of the O atom density is important to optimize these opposing effects. Injection of NO or NO2 through or downstream of the discharge in eCOIL systems has two goals; improving the efficiency of direct production of O2(1
) in the discharge by electron impact and management of the O atom density downstream of the plasma. Including NO in the gas stream flowing through the discharge has, in part, the goal of improving production of O2(1
). NO, having a lower ionization potential (9.26 eV) than either O2 or He, is likely to provide more rapid ionization with the possibility of lowering the operating E/N (electric field/gas number density) and electron temperature, Te. Lowering Te from the values typical of self-sustained He/O2 mixtures is advantageous in more efficiently producing O2(1
) by direct electron impact [6, 8, 11]. Discharges in NO have also been known to produce O2(1
) in relatively large amounts even though O2 may not be present as a feedstock gas [12]. However, when flowing NO through the discharge, some of the power that would otherwise be available to excite O2 is dissipated by excitation and ionization of NO, an unwanted consequence. In addition to possibly improving the production of O2(1
), NO and NO2 are potentially effective in managing the inventory of O atoms through direct and cyclic reactions which have the effect of converting O atoms back into O2. In the context of eCOIL systems where I2 is injected, O atoms are beneficial by dissociating I2 (and product species IO) to form I atoms which are then pumped to I∗ by collisions with O2(1
). The O atoms are detrimental by quenching I∗. Totally eliminating O atoms is therefore not necessarily beneficial. NO and NO2 also have secondary effects in that they react with I atoms forming intermediary species such as INO, and INO2 which further react with I to reform I2 [13]. In this paper, we report on results from a computational investigation of the consequences of NO and NO2 additives on flowing He/O2 plasmas and their afterglows with I2 injection in the context of eCOIL systems. These investigations were conducted using plug flow and 2-dimensional (2D) plasma- hydrodynamics models. Although for most conditions the addition of NO to the inlet gas stream reduced the density of O2(1
), it ultimately increased the densities of I∗ downstream of the discharge through management of the O atom density. The inlet NO mole fraction also typically increased the extent of the region over which positive laser gain could be achieved. This can be particularly useful in high speed flows where mixing lengths are longer. Injection of NO2 in the post- discharge flow can help in rapidly scavenging O atoms in two body reactions (as compared with scavenging by NO which proceeds by a 3-body mechanism). In general, addition of NO2 in the post-discharge region improves laser gain [9]. The downside to NO2 injection is a rise in gas temperature due to the exothermicity of the reactions between NO2 and O. The models used in this investigation are described in section 2 followed by a discussion of the reaction mechanism in sections 3. In section 4 the consequences of NO flowing through the discharge and its influence on downstream kinetics when I2 is injected are described. In section 5, the post- discharge kinetics with NO2 addition upstream of I2 injection are discussed. Our concluding remarks are in section 6. 2. Description of models This investigation was conducted using a plug flow model, GlobalKIN, and a 2D plasma hydrodynamics model, nonPDPSIM. GlobalKIN has been previously described in [11,14] and so will only be briefly discussed here. GlobalKIN consists of a volume averaged plasma chemistry module and an electron energy transport module. The plasma chemistry module provides the time rate of change of species densities based on gas-phase chemistry and surface reactions. Electron temperature, Te, and average gas temperature, Tg, are also solved for by integrating their respective conservation equations. The electron energy transport module consists of a solution of Boltzmann’s equation for the electron energy distribution (EED) which provides electron impact rate coefficients based on the EEDs and fundamental cross- sections. For plasmas flowing through cylinders having large aspect ratios, transport to the radial surfaces is taken into account by using a diffusion length. By simultaneously calculating the axial speed of the flow based on constant pressure, change in enthalpy, species densities, conservation of mass and gas temperature, the integration in time is mapped to axial position. The resulting rate equations are integrated in time using a stiff ordinary differential equation solver. To address the use of additives, GlobalKIN was modified to enable the downstream injection of gases into the flow. Injection nozzles were treated as point sources of mass, axial momentum and enthalpy. Power deposition as a function of axial position must be specified in GlobalKIN. This distribution was estimated by averaging the power deposition obtained from the nonPDPSIM over the cross-section of the tube. nonPDPSIM has been discussed in detail in [15,16], and so will only be briefly described here along with pertinent updates to the model. Continuity equations for charged gas-phase species, surface charges and Poisson’s equation for the electric potential are simultaneously implicitly integrated in time. Updates of these quantities are followed, in a time splicing manner, with updates of Te and neutral species densities using a modified form of the compressible Navier–Stokes equations for continuity, momentum and energy (gas temperature) which accounts for interactions with the plasma. A circuit model was 4794 O2(1
) production and gain in plasma pumped oxygen–iodine lasers Table 1. Continued. Reaction Rate coefficienta Reference O2(1
) + I∗ → O2 + I 1.10 × 10−13 [9] O2 + I∗ → O∗2 + I 1.0 × 10−10 (Tg/300)−1.0 e−403/Tg [9] O + I∗ → O + I 8.0 × 10−12 [9] NO + I∗ → NO + I 1.2 × 10−13 [9] NO2 + I∗ → NO2 + I 8.5 × 10−14 [9] I2 + I∗ → I∗2+ I 1.40 × 10−13 e1600/Tg [9] IO + IO → O2 + I + I 8.2 × 10−11 [9] NO + IO → NO2 + I 4.3 × 10−12 e−397/Tg [32] O + IO → O2 + I 1.4 × 10−10 [9] O + IO → O2(1
) + I 1.5 × 10−11 [9] I + INO → I2 + NO 1.6 × 10−10 [13] I + INO2 → I2 + NO2 8.3 × 10−11 [13] INO + INO → I2 + NO + NO 8.4 × 10−11 e−2620/Tg [13] INO2 + INO2 → I2 + NO2 + NO2 2.9 × 10−11 e−2600/Tg [13] I + I + I2 → I2 + I2 3.6 × 10−30 [9] I + I + He → I2 + He 3.8 × 10−33 [9] I + I + O2 → I2 + O2 3.3 × 10−32 [9] I + I∗ + I2 → I + I + I2 3.6 × 10−30 [9] I + I + O2 → I2 + O2(1
) 3.7 × 10−33 [9] I + NO + He → INO + He 6.0 × 10−33 (Tg/300)−1.0 [13] I + NO + O2 → INO + O2 1.6 × 10−32 [13] I + NO2 + He → INO2 + He 1.5 × 10−31 (Tg/300)−1.0 [13] I + NO2 + O2 → INO2 + O2 2.6 × 10−31 [13] I∗ → I 10 [33] a Rate coefficient in cm3 s−1 for 2-body reactions, and cm6 s−1for 3-body reactions and s−1 for radiation reactions. b Rate coefficients calculated using cross-section data from the indicated reference. c Estimated. d Where M is one of the cations O+, O+2 , He + or NO+. e Where M is one of the anions O−, O−2 or O − 3 . f Where M is one of the major neutral species He, O2, O2(v), O2(1
), O or O3. O3. The negative ions NO− and NO−2 were not included in the 2D model (to increase computational speed), whereas they were included in the plug flow model. Computational experiments were conducted using the plug flow model to quantify the effects of the NO− and NO−2 and are discussed below. One of the primary motivations of injecting NO and NO2 is in their potential for managing of the O atom inventory. Much of this chemistry is cyclic. NO reacts with O to form NO2 which then further reacts with O to regenerate O2, NO + O + M → NO2 + M, (k = 1.0 × 10−31(Tg/300)−1.6,
H = −3.18 eV), (18) NO2 + O → NO + O2, (k = 4.2 × 10−12 exp(273/Tg),
H = −2.0 eV). (19) This is an important reaction chain that can reduce the inventory of O atoms and so eliminate a quencher of the upper laser level, I∗. At low pressures where the 3-body density is low, the rate of reaction of O with NO2 occurs at a higher rate than with NO. Hence addition of NO2 in the downstream region may be preferred over that of NO if the residence time in the flow tube is a limiting factor. These reactions are, however, exothermic and so can increase the gas temperature which is generally not beneficial. Laser gain is ultimately achieved by injection of I2 downstream of the plasma zone, and by its reacting with O, O2(1
), and O2(1) to create I atoms and to pump the upper laser level. In a conventional COIL there is a negligible inventory of O atoms and O2(1), and so the dissociation of I2 is dominantly by O2(1
) in a two-step process, O2( 1
) + I2 → O2 + I∗2, (k = 7 × 10−15 cm3 s−1), (20) O2( 1
) + I∗2 → O2 + I + I, (k = 3 × 10−10 cm3 s−1). (21) In eCOIL, the presence of O2(1) and O helps in dissociating I2 and producing I atoms. A reaction intermediate IO is also helpful in this regard, O2( 1) + I2 → O2 + I + I, (k = 2.8 × 10−11 cm3 s−1), (22) O + I2 → IO + I, (k = 1.4 × 10−10 cm3 s−1), (23) O + IO → I + O2, (k = 1.4 × 10−10 cm3 s−1). (24) The laser pumping reaction by O2(1
) in collisions with I is favoured at lower temperatures over its endothermic back reaction, O2( 1
) + I → O2 + I∗, (k = 7.7 × 10−11(Tg/300)−1 cm3 s−1), (25) O2 + I ∗ → O2(1
) + I, (k = 1.0 × 10−10(Tg/300)−1e−403/Tg). (26) To suppress the back reaction, the laser cavity is typically placed in a supersonically expanded flow to lower the translational temperature. As such, the threshold yield, Yth, of O2(1
) required for positive optical gain in an undissociated flow of O2 is [17], Yth = [O2( 1
)] [O2] = 1 1 + 1.5 exp(401/Tg) . (27) 4797 R A Arakoni et al For example, in an undissociated flow of O2, the threshold yield at room temperature is 15% whereas at 180 K the threshold yield is 6%. Quenching of I∗ occurs dominantly by reactions with O and O2 (the backward reaction), and to a lesser extent with O2(1
) and NO, O + I∗ → O + I, (k = 8.0 × 10−12 cm3 s−1), (28) NO + I∗ → NO + I, (k = 1.2 × 10−13 cm3 s−1). (29) NO and NO2 also help in removing ground state I atoms and so aid in maintaining the inversion, I + NO + O2 → INO + O2, (k = 1.6 × 10−32 cm6 s−1), (30) I + NO + He → INO + He, (k = 6.0 × 10−33 cm6 s−1), (31) I + NO2 + O2 → INO2 + O2, (k = 2.6 × 10−31 cm6 s−1), (32) I + NO2 + He → INO2 + He, (k = 6.0 × 10−33 cm6 s−1). (33) The INOx species in turn abstract I atoms to release I2 and NOx back into the flow, INO + I → I2 + NO, (k = 1.6 × 10−10 cm3 s−1), (34) INO2 + I → I2 + NO2, (k = 8.3 × 10−11 cm3 s−1). (35) Typically, the injection of I2 occurs during or following a supersonic expansion of the gas to velocities equivalent to Mach 2 or Mach 3 to lower Tg to decrease the yields of O2(1
) required to achieve positive gain. Accurately simulating a supersonic expansion is difficult in our modelling platform. In order to provide a best case estimate for laser gain we used for the temperature of the reactions in equations (25) and (26) a value appropriate for a Mach 2 flow. This temperature can be approximated from T0 Tg = ( 1 + γ − 1 2 M2 ) , (36) where T0 is the stagnation (or tank) temperature, γ is the ratio of specific heats and M is the Mach number. For M = 2 and T0 = 300 K, this ratio is ≈2.18, and would lead to a gas temperature of 137 K. The reaction mechanism was validated by comparing with experimental data from Carroll et al [34], as discussed below. 4. Consequences of NO in the inlet flow A schematic of the cylindrically symmetric, 6 cm diameter flow tube used in this study is shown in figure 1. In the base case, a power of 40 W was capacitively coupled using ring electrodes operated at 25 MHz. A mixture of He/O2/NO at 3 Torr entered through the inlet at a flow rate of 6 slpm which corresponds to an axial speed of ≈890 cm s−1. The flow consisted of 30% O2 with the balance divided between He and NO. The NO mole fraction was varied from 0–10%. Two injection nozzles are located downstream. (In the 2D cylindrically symmetric Figure 1. Geometry of the cylindrically symmetric reactor. The flow enters from the top and is pumped from the bottom. Discharge power at 25 MHz is capacitively coupled through ring electrodes. Two nozzles downstream of the discharge inject mixtures of He, NO, NO2 and/or I2. geometry these nozzles appear to be rings.) The first nozzle at 51.5 cm was used to inject a mixture of He/NO/NO2. The second nozzle at 64.5 cm injected a mixture of He/I2. At the outlet, axial gradients were assumed to be zero and the exit speeds were adjusted to maintain constant pressure and mass flux. Plasma characteristics (ne, Te, negative total negative ion density M−, total positive ion density M+ and power density) obtained with nonPDPSIM are shown in the vicinity of the electrodes in figure 2 for the base case having an inlet flow of He/O2/NO = 67/30/3. Te in the bulk of the discharge was 2.3 eV and the peak electron density was 9.0 × 109 cm−3. Since NO and O2 are attaching species, the negative ion density (maximum of 1.1×1010 cm−3) is commensurate with ne. The electron density is fairly symmetric between the electrodes and relatively uniform. The power deposition is moderately higher near the upstream electrode due to gas heating which reduces the neutral densities downstream. This is not a general result as higher power deposition and more rarefaction can produce regions of locally intense power deposition near the downstream electrode. The densities of the neutral species O, O2(1
), I and I∗ and Tg for the base case are shown in figure 3. A flow of 36 sccm of pure NO was injected through the first nozzle and a flow of 100 sccm of He/I2 = 99/1 was injected through the second nozzle. Having few quenchers at these pressures, the O2(1
) accumulates as the gas flows through the discharge reaching a maximum value of 1.35 × 1015 cm−3. The O2(1
) density remains nearly constant thereafter until the injection point for I2. Reactions of O2(1
) with I2 and I (the latter being the laser pumping reaction producing I∗) reduce its density by a factor of five to 2.7 × 1014 cm−3. 4798 O2(1
) production and gain in plasma pumped oxygen–iodine lasers Figure 2. Base case plasma characteristics from the 2D model for 3 Torr, 40 W, He/O2/NO = 67/30/3 and a flow rate of 6 slpm. (a) Electron density, (b) electron temperature, (c) sum of negative ion densities, (d) sum of positive ion densities and (e) power density. The scales are linear for Te and power (0 to maximum), and the densities are plotted on a 2-decade log scale. Similar to O2(1
), the density of O atoms accumulate from electron impact dissociation of O2 passing through the discharge (with a small amount of depletion of the O in forming O3) until the injection point of NO at the first nozzle. The O densities decrease from a peak value of 1.4 × 1015 to 4.0× 1014 cm−3 downstream of the NO injection due to its reaction with NO. The remaining O atoms flow to the injection point of I2 where they are further depleted in dissociating reactions with I2. The I2 is nearly completely dissociated by reactions with the O atoms to form I. Pumping reactions between O2(1
) and I produce I∗. The region of positive gain (where the inversion density G = [I ∗] − 1/2[I ] > 0) is a narrow band downstream of the I2 injection point, with a peak value of G = 2.6 × 1011 cm−3. Tg has two local maxima. The first is due to discharge Joule heating, Frank–Condon heating and exothermic reactions of NO with O reaching 340 K adjacent to the downstream electrode. As the walls are held at 300 K, the gas rapidly cools by thermal conduction. A second local maximum occurs downstream of the first injection nozzle (up to 348 K) due to additional exothermic reactions of O and the newly injected NO. To investigate the consequences of NO in the inlet flow over a wider parameter space, the plug flow model GlobalKIN was used. We first addressed the importance of NO− and NO−2 in the reaction mechanism and their effects on the densities of electrons, O2(1
) and O2(1). The densities of these species are shown along the axis of the discharge in figure 4 for a 3 Torr mixture of He/O2/NO = 60/30/10 with 40 W power deposition. A high mole fraction of NO was used as an extreme Figure 3. Neutral species densities and Tg for the base case. 36 sccm of pure NO was injected through the first nozzle and 100 sccm of He/I2 = 99/1 was injected through the second nozzle. Densities of (a) O2(1
), (b) O, (c) I, (d) I∗, (e) inversion density ([I∗]-0.5[I]) and (f ) gas temperature. The scales are linear except for O2(1
) and O which are plotted on a 2-decade log scale. Injection of NO decreases the flow of O and injection of I2 consumes O2(1
). Exothermic reactions at both injection points produce local maxima in Tg. case. When including NO− and NO−2 the peak electron density decreased approximately 10% from 1.4 × 1010 cm−3. The reduction in ne results in large part from the attachment of electrons to NO through 3-body reactions similar to those for O−2 formation (equation (7)) and dissociative attachment to NO. The changes in densities of O2(1
) and O2(1) were relatively small (<5%) and could be attributed to the fact that ne decreases and Te increases modestly (0.05–0.1 eV from ≈2.0 eV) with the inclusion of NO− and NO−2 . (Recall that electron impact excitation of O2(1
) is maximum for Te = 1– 1.5 eV.) The density of O atoms was relatively independent on the inclusion of NO− and NO−2 . Based on these trends, we can expect results from nonPDPSIM (which do not include NO− and NO−2 in the reaction mechanism) to over-predict O2( 1
) densities by a few per cent. The consequences of NO in the inlet flow on the maximum densities of charged species, Tg, Te, and power deposition into different species by electron impact are shown in figure 5 for 3 Torr and 40 W. A modest increase in ne, from 1.1×1010 cm−3 to 1.25 × 1010 cm−3, occurs as the NO mole fraction is increased from 0 to 10% due to the higher rates of ionization with NO in spite of a decrease in the positive and negative 4799 R A Arakoni et al With the decrease in Te with addition of NO, the production of O(1D) by electron impact dissociation of O2 decreases and so these secondary sources of O2(1
) and O2(1) also decrease. Furthermore, both NO and NO2 are quenchers of O(1D), NO + O(1D) → NO + O, (k = 4.0 × 10−11 cm3 s−1), (39) NO2 + O( 1D) → NO + O2, (k = 3.0 × 10−10 cm3 s−1), (40) NO2 + O( 1D) → NO2 + O, (k = 3.2 × 10−10 cm3 s−1). (41) So the addition of NO reduces the production of O(1D) and increases its rate of quenching, thereby reducing the production of O2(1
) by excitation transfer. The importance of the quenching reactions of O(1D) is demonstrated by excluding the reactions in equations (39)–(41) from the mechanism. For 10% NO in the flow, the yield of O2(1
) improved from 4.6% to 5.3% when quenching of O(1D) is eliminated. The quenching of O2(1
) by NO has a small rate coefficient (3.5 × 10−17 cm3 s−1) and so does not significantly contribute to the loss of O2(1
). The density of O atoms increases monotonically through the discharge zone to a maximum value of 2.5 × 1015 cm−3 in the absence of NO. (The small increase in the density of O after the discharge is mainly due to gas cooling.) With the addition of NO, the density of O atoms decreases throughout the flow-tube and, in particular, downstream of the discharge. This decrease is due to both the reduction in the rate of electron impact dissociation of O2 by the decrease in Te and the formation of NO2 in reactions with NO. In the post- discharge region in the absence of NO, the majority of O2(1) is converted to O2(1
) through collisions with O and O3. Due to the reduction of O atoms with increasing NO, the rates of quenching of O2(1) to O2(1
) are also smaller, leading to higher densities of O2(1) in the afterglow. Higher O2(1) densities are not necessarily bad since they help in dissociating I2 but maintaining those densities does result in lower densities of O2(1
) that directly pump I∗. The densities of I and I∗, and the optical gain at 1.315 µ are shown in figure 7 for the conditions of figure 6 (1 sccm of I2 injected in a 100 sccm, He/I2 = 99/1 flow). The gain was given by σ([I∗] − 0.5[I]) where σ is the stimulated emission cross-section. At pressures of less than tens of Torr, Doppler broadening dominates over pressure broadening, and so the stimulated emission cross-section can be approximated by [35], σ = 1.33 × 10−16T −1/2g cm2. (42) In the absence of NO, the density of I increases from 1.75 × 1013 cm−3 at the I2 injection point to 3.2 × 1013 cm−3 downstream due in large part to the reaction of O atoms with I2. The density of I∗ is maximum at the injection point at 1.45×1013 cm−3 and decreases to negligible values by the end of the flow tube due to the depletion of O2(1
) (the species responsible for the pumping reaction) and quenching by O atoms. Since the lifetime of I∗ (125 µs for quenching by O) is short compared with flow times, the density of I∗ does not appreciably accumulate in the discharge and its density is a Figure 7. Densities of atomic iodine species as a function of inlet NO mole fraction for 3 Torr, 40 W, He/O2/NO = 70 − x/30/x and 6 slpm. 100 sccm of He/I2 = 99/1 is injected through the second nozzle. Densities of (a) I, (b) I∗ and (c) optical gain. These results are from the plug flow model. Low flows of NO produce the maximum peak gain whereas high flows of NO produce larger plumes of positive gain. reflection of instantaneous production and quenching rates. The end result is a peak gain of 4.2 × 10−5 cm−1 within a centimetre of the I2 injection point. Note that the rate of quenching of I∗ by O is faster than that due to spontaneous emission (0.1 s) [33]. With injection of I2 the density of O2(1
) decreases by virtue of excitation transfer and dissociative excitation of I2, and reactions with I which pumps I∗. In the absence of NO, the density of O2(1
) is fully depleted by the reactions. In the absence of NO there is also a large density of O atoms at the injection point which react with I2 producing IO and I. Since the rate of dissociation of I2 by O2(1
) and O2(1) is slower than by O atoms there is a larger density of I available for O2(1
) to react with, and so the O2(1
) is rapidly depleted. 4802 O2(1
) production and gain in plasma pumped oxygen–iodine lasers As the flow of NO is increased, the flow of O atoms decreases, resulting in a lower rate of dissociation of I2 and fewer I atoms. With the lower density of I the reactivity of O2(1
) is lower; and so its density decreases less rapidly after injection of I2. The production of I∗ results from reactions of O2(1
) with I while the quenching of I∗ is largely due to collisions with O atoms. Increasing the NO mole fraction decreases the O atom density downstream thereby lowering the rates of quenching of I∗ and extending its plume beyond the I2 injection point. However, having too large an NO mole fraction results in too low rates of dissociation of I2 and hence poor utilization of O2(1
). The end result is that the region over which positive gain can be sustained is maximum for an intermediate mole fraction of NO of 3%. The purpose of flowing NO (or injecting NO2) is largely to manage the O atom density and so control the quenching of I∗ by O. A sensitivity study was conducted of the rate coefficient for this quenching reaction from the nominal value of 8 × 10−12 cm3 s−1. The densities of O2(1
) and I∗, and gain are shown in figure 8 while varying this rate coefficient from 0 to 1.0×10−11 cm3 s−1. In the absence of quenching, the densities of I∗ and O2(1
) do not significantly decrease downstream of the I2 injection point. In the absence of quenching by O, the predominant quencher of I∗ is O2 and this quenching produces O2(1
). The forward and backward pumping reactions (equations (25) and (26)) reach an equilibrium where the I∗ and O2(1
) densities gradually decrease due to minor quenchers of I∗ (such as NO and O2(1
)) and radiative relaxation. When increasing the rate coefficient for quenching I∗, the density of I∗ decreases proportionately. By removing I∗ in this manner the rate of the backward reaction with O2 decreases and so does the density of O2(1
). The extent of positive gain is progressively limited to the few cm beyond the I2 injection point as the quenching of I∗ increases, though the peak value of gain is not particularly sensitive to the rate coefficient. The densities of O2(1
) and O, and Tg 2 cm upstream of the second nozzle are shown in figure 9 as a function of power deposition and NO mole fraction in the inlet flow. In general, Tg increases with power deposition and with NO addition reaching a maximum of 480 K with 400 W power deposition and 10% NO in the inlet flow. Increasing power deposition produces more electron impact dissociation of O2 and NO, producing larger densities of O atoms. However, increasing NO mole fractions decreases the density of O atoms by virtue of scavenging by NO and NO2 and decreasing Te. The density of O2(1
) decreases with NO addition, as discussed above, and increases with power deposition. The saturation in the density O2(1
) at 6.5 × 1015 cm−3 at higher powers is due in part to the depletion of O2 by electron impact dissociation and in part to gas heating. For example, with 400 W and 0% NO, the fractional dissociation of O2 is 68%. The addition of NO reduces the depletion of O2 by both reducing the rate of O2 dissociation and by recycling O atoms back to O2. For example, for 10% NO addition and 400 W, the fractional dissociation decreases to 36%. The yield of O2(1
) saturates with power at 18% due to the depletion of O2. The densities of I and I∗, and gain as function of power deposition and NO mole fraction are shown in figure 10 at a location of 2 cm downstream from the I2 injection point. This mixing length of 2 cm was chosen based on previous studies Figure 8. Sensitivities of the value of the rate coefficient for the quenching reaction between O and I∗ on the post-discharge kinetics for 3 Torr, 40 W, He/O2/NO = 69/30/1 and 6 slpm. 100 sccm of He/I2 = 99/1 is injected through the second nozzle. Densities of (a) O2(1
), (b) I∗ and (c) optical gain. These results are from the plug flow model. for similar flow conditions [36]. The inlet flow has 30% O2 and the flow injected through the second nozzle is 1 sccm of I2 in a 100 sccm flow of He/I2 = 99/1. At low power deposition, the density of I is large, (2.2–2.3) × 1013 cm−3, because the yields of O2(1
) are low enough that the pumping of I to I∗ is slow. As the yield of O2(1
) increases at higher powers, the pumping reactions reduce the density of I and increase that of I∗, leading to an increase in gain. However, at large power deposition, the density of O increases whereas that of O2(1
) saturates. This leads to an increased rate of quenching of I∗ by O which reduces the density of I∗ and increases that of I. The end result is a reduction in gain. Increasing NO in the inlet flow reduces the density of O atoms at the injection point, thereby reducing the rate of 4803 R A Arakoni et al Figure 9. Consequences of power deposition and NO mole fraction on the neutral gas properties for 3 Torr, He/O2/NO = 70 − x/30/x and 6 slpm. (a) Tg, (b) density of O, (c) density of O2(1
) and (d) yield of O2(1
). O2(1
) densities saturate with power due to a high degree of dissociation of O2. These results are from the plug flow model and the values are for 2 cm upstream of the second nozzle. dissociation of I2 and the density of I atoms. The reduction in O by addition of NO also reduces the quenching of I∗ by O, making the gain predominantly dependent on the yield of O2(1
). As a result, for 3% NO, the gain decreases from 6.0 Figure 10. Consequences of power deposition and NO mole fraction on iodine species and gain for 3 Torr, He/O2/NO = 70 − x/30/x and 6 slpm. Densities of (a) I and (b) I∗ and (c) optical gain. These results are from the plug flow model and the values are for 2 cm downstream of the second nozzle. Gain is maximum at high power only with high flow rates of NO. × 10−5 to 4.3 × 10−5 cm−1 between 150 and 400 W, whereas in the absence of NO, the gain reduces by nearly a factor of 3. The eCOIL system may operate in either power limited or iodine limited modes. In the iodine limited mode, the flow of O2(1
) and O generated by the discharge fully utilizes the injected flow of I2, and so gain saturates with increasing power. To some degree (quenching of I∗ and depletion of O2 aside), this is the mode that applies to the results shown in figure 10. In the power limited mode, the flow of O2(1
) and O is insufficient to fully utilize the injected flow of I2, and so gain saturates with flow of I2. These modes of operation are demonstrated by varying power deposition and I2 flow rate. The densities of I and I∗, and gain are shown 2 cm downstream of the I2 injection point in figure 11 for a 6 slpm inlet flow of He/O2/NO = 67/30/3 4804 O2(1
) production and gain in plasma pumped oxygen–iodine lasers Figure 14. Gain in a subsonic flow for 10 Torr, 26.9 slpm of He/O2 = 80/20, 25–800 W, with 10.8 sccm of I2 injected through the second nozzle. (a) 0 sccm NO2, (b) 672 sccm NO2 and (c) 1344 sccm NO2 through the first nozzle. Experimental values are from [34]. (20 mmol s−1) inlet flow of He/O2 = 80/20 and 25–800 W followed by injection of 0–1344 sccm (0–1 mmol s−1) NO2 and injection of 10.7 sccm (0.008 mmol s−1) of I2, equivalent to few per cent of the O2(1
) flow rate. The diameter of the reactor is 4.9 cm and I2 injection is 20 cm downstream of the NO2 injection point. The power deposition spans nearly 25 cm due to a larger separation between electrodes in the experiment. The experimental measurements were made 10 cm downstream of the I2 injection point in a subsonic (high gas temperature flow) and so gains are negative. With the exception of low powers and low flow rates of NO2, the experimental trends are captured by GlobalKIN. Addition of NO2 prior to injection of I2 scavenges some of the O atoms in the flow and so reduces the quenching of I∗ by O atoms. Higher powers produce larger flows of O atoms as well as more O2(1
) but the quenching of I∗ dominates. Increasing the NO2 flow rate increases the scavenging of O atoms and extends the power prior to transitioning to large negative gain. Figure 15. Consequences of injection of NO2 through the first nozzle for 3 Torr, 40 W, He/O2/NO = 67/30/3 and 6 slpm. 36 sccm of He/NO2 mixture is injected through the first nozzle. Densities of (a) O and (b) O2(1
) and (c) yield of O2(1
). Addition of NO2 rapidly consumes the O atoms. These results are from the GlobalKIN, the plug flow model. Values from the 2D nonPDPSIM are shown without NO2 injection. Since NO2 is more effective than NO in scavenging of O atoms, we investigated NO2 injection through the first nozzle. The conditions are a 3 Torr, 6 slpm inlet flow of He/O2/NO = 67/30/3 and discharge power of 40 W. A He/NO2 flow of 36 sccm was injected through the first nozzle with the fraction of NO2 being varied. As before a 100 sccm flow of He/I2 = 99/1 was injected through the second nozzle. The consequences of NO2 flow rate through the first nozzle on the densities of O and O2(1
), and yield of O2(1
) are shown in figure 15. These results are from GlobalKIN with a result from nonPDPSIM without NO2 injection for comparison. The injection of NO2 produces a decrease in the O atom density due to the titration of O by NO2 and the 4807 R A Arakoni et al Figure 16. Gain when including NO in the inlet flow and injecting NO2 through the first nozzle for 3 Torr, 40 W, He/O2/NO = 70 − x/30/x and 6 slpm. 100 sccm of He/I2 = 99/1 is injected through the second nozzle. (a) Gain as a function of NO mole fraction and (b) gain as a function of NO2 flow rate. These results are from the plug flow model. Maximum gain is obtained at low NO flow rates and high NO2 injection. conversion of O to O2, respectively. As the NO2 flow rate increases to 36 sccm the O atom densities decrease by a factor of 3–4 just downstream of the injection point due to the more rapid rate of reaction with O (compared with that of NO). The heat of reaction between NO2 and O locally increases the gas temperature leading to a reduction in the density of O2(1
) near the first nozzle due to rarefaction. Note, however, that the yield of O2(1
) is not affected by NO2 injection because NO2 does not appreciably quench O2(1
). The reduction in the densities of O with NO2 injection implies that the quenching of I∗ by O is reduced. This produces a reduction in the amount of O2(1
) used in pumping the I∗. Hence, downstream of the second nozzle, the yield of O2(1
) is higher for larger NO2 mole fractions. Results from GlobalKIN and nonPDPSIM are in general agreement except downstream of the I2 injection point due to the artificially higher rates of reaction upon injection of I2. Gain is shown in figure 16 for 2 cm downstream from the second nozzle as a function of NO mole fraction in the inlet flow and flow rate of NO2 through the second nozzle. The flow conditions are 3 Torr and 40 W power deposition. Having NO in the inlet flow affects the production of O2(1
) as discussed above as well as managing the O atoms density. Injection of NO2 downstream of the discharge largely only affects the density of O atoms (and gas temperature). As such, at low values of NO flow, injection of NO2 is effective in managing the O atom density and larger flow rates tend to maximize gain by reducing quenching of I∗ by O atoms. At large flow rates of NO, the management of O atoms is dominated by reactions with NO, and so the injection of NO2 is less effective. Since there is a deleterious effect on O2(1
) production by having large flows of NO through the discharge, managing the O atom density with injection of NO2 is likely the optimum strategy. 6. Concluding remarks The consequences of NO in the inlet flow of a He/O2 plasma and its flowing afterglow, and NO2 and I2 injection on the post-discharge kinetics of the eCOIL were investigated using plug flow and 2D models. The addition of NO to the inlet flow through the discharge produces a reduction in Te and a modest increase in ne resulting in the densities and yields of O2(1
) being generally lower with NO. Including NO in the flow reduces the density of O atoms both by a reduction in the electron impact dissociation of O2 and by exothermic reactions of O with NO. This proves beneficial to improving optical gain by reducing the quenching of I∗ by O atoms. At higher power deposition, the dissociation of O2 saturates the yield of O2(1
). By virtue of adding NO to the inlet flow, the reduction in Te reduces the rate of dissociation of O2. Even though the yields of O2(1
) were generally lower, the optical gain was generally higher when the NO mole inlet mole fraction was between 1% and 3%. The eCOIL system can operate in power limited and I2 limited regimes. At low flow rates, I2 is nearly totally dissociated and so the densities of I depend largely on the I2 flow rate. Upon increasing the flow rate of I2, the system transitions to a power limited regime and higher powers are required to optimize gain. Small flows of NO2 in the post- discharge region can be used to fine tune the gain. The addition of NO2 rapidly consumes O atoms without significantly changing other parameters (other than Tg) and so increases the optical gain. The injection of NO2 was most effective at low flow rates of NO. In general, management of the O atom density is critical to optimizing gain due to its rapid rate of quenching of I∗. Acknowledgment This work was supported by the Air Force Office of Scientific Research and the National Science Foundation (CTS-0520368). References [1] Hicks A, Utkin Yu G, Lempert W R, Rich J W and Adamovich I V 2006 Appl. Phys. Lett. 89 241131 [2] Braginsky O V et al 2006 J. Phys. D: Appl. Phys. 39 5183 [3] Verdeyen J T et al 2006 Appl. Phys. Lett. 89 101115 [4] Kodymová J and Spalek O 1998 Japan. J. Appl. Phys. 37 117 [5] Fujii H, Yoshida S, Iizuka M and Atsuta T 1990 J. Appl. Phys. 67 3948 [6] Hill A 2000 The next generation of controlled avalanche discharge gas lasers Int. Conf. on Lasers (Albuquerque, NM) [7] Hicks A et al 2005 J. Phys. D: Appl. Phys. 38 3812 [8] Ionin A A et al 2003 J. Phys. D: Appl. Phys. 36 982 [9] Palla A D, Carroll D L, Verdeyen J T and Solomon W C 2006 Effects of mixing on post-discharge modeling of ElectricOIL experiments Proc. SPIE. 6101 610125 [10] Vasiljeva A N et al 2004 J. Phys. D: Appl. Phys. 37 2455 [11] Stafford D S and Kushner M J 2005 J. Appl. Phys. 98 073303 [12] Cook T J and Miller T A 1974 Chem. Phys. Lett. 25 396 4808 O2(1
) production and gain in plasma pumped oxygen–iodine lasers [13] Atkinson R et al 2000 J. Phys. Chem. Ref. Data 29 167 [14] Stafford D S and Kushner M J 2004 J. Appl. Phys. 96 2451 [15] Lay B, Moss R S, Rauf S and Kushner M J 2003 Plasma Sources Sci. Technol. 12 8 [16] Kushner M J 2004 J. Appl. Phys. 95 846 [17] Hon J et al 1994 A Heuristic Method for Evaluating COIL Performance AIAA Paper 94-2422 [18] Rapp D and Briglia D D 1965 J. Chem. Phys. 43 1480 [19] Josic L, Wroblewski T, Petrovic Z Lj, Mechlinska-Drewko J and Karwasz G P 2001 Chem. Phys. Lett. 350 318 [20] Lindsay B G, Mangan M A, Straub H C and Stebbings R F 2000 J. Chem. Phys. 112 9404 [21] Vejby-Christensen L, Kella D, Pedersen H B and Andersen L H 1998 Phys. Rev. A 57 3627 [22] Shimamori H and Hotta H 1988 J. Chem. Phys. 89 2938 [23] Ikezoe Y, Matsuoka S, Takebe M and Viggiano A 1987 Gas Phase Ion-Molecule Reaction Rate Constants Through 1986 (Ion Reaction Research Group of the Mass Spectroscopy Society of Japan, Tokyo, Japan) [24] Atkinson R et al 2004 Atmos. Chem. Phys. 4 1461 [25] Boodaghians R B, Borrell P M and Borrell P 1983 Chem. Phys. Lett. 97 193 [26] Blais N C 1985 J. Phys. Chem. 89 4156 [27] Doroshenko V M, Kudryavtsev V M and Smetanin V V 1992 High Energy Chem. 26 227 [28] Gauthier M J E and Snelling D R 1975 J. Photochem. 4 27 [29] Valli G S, Orru R, Clementi E, Lagana A and Crocchianti S 1995 J. Chem. Phys. 102 2825 [30] Phillips L F and Schiff H I 1965 J. Chem. Phys. 42 3171 [31] Campbell I M and Gray C N 1973 Chem. Phys. Lett. 18 607 [32] Holscher D and Zellner R 2002 Phys. Chem. Chem. Phys. 4 1839 [33] Husain D, Slater N K H and Weisenfeld J R 1977 Chem. Phys. Lett. 51 201 [34] Carroll D L et al 2005 IEEE J. Quantum Electron. 41 213 [35] Zagidullin M V, Nikolaev V D, Svistun M I, Khvatov N I and Hager G D 2005 Appl. Phys. A: Mater. Process. 81 311 [36] Muruganandam T M et al 2002 AIAA J. 40 1388 4809