IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 43 (2010) 015201 (14pp) doi:10.1088/0022-3727/43/1/015201
Gain and output power measurements inan electrically excited oxygen–iodine laserwith a scaled dischargeJ R Bruzzese, A Hicks, A Erofeev, A C Cole, M Nishihara andI V Adamovich
Michael A Chaszeyka Nonequilibrium Thermodynamics Laboratories,Department of Mechanical Engineering, The Ohio State University, Columbus, OH 43210, USA
Received 27 August 2009, in final form 5 November 2009Published 7 December 2009Online at stacks.iop.org/JPhysD/43/015201
AbstractSinglet delta oxygen (SDO) yield, small signal gain, and output power have been measured ina scaled electric discharge excited oxygen–iodine laser. Two different types of discharges havebeen used for SDO generation in O2–He–NO flows at pressures up to 90 Torr, crossednanosecond pulser/dc sustainer discharge and capacitively coupled transverse RF discharge.The total flow rate through the laser cavity with a 10 cm gain path is approximately0.5 mole s−1, with steady-state run time at a near-design Mach number of M = 2.9 of up to 5 s.The results demonstrate that SDO yields and flow temperatures obtained using thepulser-sustainer and the RF discharges are close. Gain and static temperature in the supersoniccavity remain nearly constant, γ = 0.10–0.12% cm−1 and T = 125–140 K, over the axialdistance of approximately 10 cm. The highest gain measured is 0.122% cm−1 at T = 140 K.Positive gain measured in the supersonic inviscid core extends over approximately one half toone third of the cavity height, with absorption measured in the boundary layers near top andbottom walls of the cavity. Laser power has been measured using (i) two 99.9% mirrors onboth sides of the resonator, 2.5 W, and (ii) 99.9% mirror on one side and 99% mirror on theother side, 3.1 W. Gain downstream of the resonator is moderately reduced during lasing (byup to 20–30%) and remains nearly independent of the axial distance, by up to 10 cm. Thissuggests that only a small fraction of power available for lasing is coupled out, and thatadditional power may be coupled in a second resonator. Preliminary laser power measurementsusing two transverse resonators operating at the same time (both using 99.9–99% mirrorcombinations) demonstrated lasing at both axial locations, with the total power of 3.8 W.
(Some figures in this article are in colour only in the electronic version)
1. Introduction
Scalability of electric discharge excited oxygen–iodine laser(e-COIL or DOIL) is a key issue which could determinewhether this laser can be used for practical applicationsrequiring high c.w. powers (in the multi-kilowatt range). Theother critical issue that needs to be resolved is improving thelaser efficiency. Although a significant increase in laser gainand output power has been reported over the last two years,γ = 0.10% cm−1 and 6.2 W [1], γ = 0.17% cm−1 and 12.3 W[2] and γ = 0.22% cm−1 and 28 W [3], the efficiency, definedas the ratio of the laser power to the electric discharge power,
remains fairly low, up to approximately 1% [3]. Scaling thelaser to high output powers requires increasing the power of anelectric discharge used to generate singlet delta oxygen (SDO)in the laser mixture, as well as the discharge pressure and theflow rate of the mixture through the discharge. Increasing laserefficiency requires optimizing SDO yield in the discharge, gainin the laser cavity and laser resonator parameters. In this paper,we present results of experiments with the scaled-up versionof the DOIL laser currently under development at the OhioState University. In the scaled-up laser, both the flow rate andthe gain path have been doubled compared with a small-scaleversion used in our previous work [4–7], and the discharge
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
Figure 1. Schematic of experimental setup.
volume has been increased by a factor of 4 to 8. In this paper,both the nanosecond pulser/dc sustainer discharge [4–7] andthe RF discharge [1–3] have been used for SDO generation.The main objectives of this work are to scale the dischargepower without producing plasma instabilities, to characterizethe flow in the supersonic cavity and to measure SDO yield inthe discharge, gain distribution in the cavity and output powerfor different laser mirror combinations.
2. Experimental
The schematic of the experimental setup is shown in figure 1.In this work, transverse electric discharges in a rectangulargeometry are used to generate SDO in an oxygen–helium flowusing two different approaches. In the first approach [4–7],ionization is produced by a high voltage, high repetition ratenanosecond pulse transverse discharge operated at a very lowduty cycle (∼1/1000), while transverse dc sustainer dischargefully overlapped with the pulsed discharge draws the currentand couples power to the plasma. The pulsed and the dc electricfields are perpendicular both to each other and to the directionof the flow (see figure 1). The dc voltage is kept too low toproduce ionization in the flow, which greatly improves thedischarge stability. The sustainer voltage can be varied tooperate at the reduced electric field (E/N) maximizing thedischarge energy fraction going to excitation of a 1� state ofoxygen by electron impact. This approach has been previouslyused to develop a 6 kW c.w. CO2 laser with a pulser-sustainerdischarge power of 70 kW [8]. In the second approach, similarto one used in [1–3], the dc electrodes are removed anda capacitively coupled transverse RF discharge is sustainedbetween the same electrodes that are used for the nanosecondpulse discharge. Note that this approach also provides apossibility of coupling additional dc power to the flow if thedc electrodes are left in place and powered. This method hasbeen used to develop a 27 kW c.w. CO2 laser, using a combinedRF/dc discharge operating at 160 kW [9].
In the scaled-up laser used in this work, the volume ofthe electric discharge section has been increased by a factorof 4 compared with the previous design [4–7], 200 cm3 versus50 cm3. Specifically, the new discharge section, made fromacrylic plastic, is 10 cm long, 10 cm wide and 2 cm high,compared with the 5 cm × 5 cm × 2 cm discharge section used
in our previous work [4–7]. Thus, the distance between thepulsed/RF electrodes is 2 cm and the distance between thedc electrodes is 10 cm. Preliminary experiments were alsoconducted using two identical discharge sections, the secondsection placed downstream of the first section, with the totaldischarge volume of 400 cm3. This was done to explore theeffect of longer flow residence time in the discharge on SDOyield and laser gain.
A premixed O2/He/NO flow is delivered to the dischargesection through a 1 inch diameter delivery line followed bya 16 cm long, 30◦ half-angle flow expansion section and a15 cm long flow conditioning section with a 2.5 cm long,3 mm cell diameter honeycomb flow straightener and twowire mesh screens with 1 mm mesh size. This arrangementgreatly reduces flow separation occurring in the flow expansionsection. Nitric oxide is added to the oxygen–helium mixtureto scavenge O atoms by a rapid recombination reaction,NO + O → NO∗
2 → NO2 + hν [10]. Otherwise, oxygen atomsgenerated in the discharge would rapidly deactivate excitediodine atoms in the laser cavity, I∗ + O → I + O [11]. Ourprevious experiments using a small-scale laser [6, 7] showedthat adding NO to the flow considerably increases small signalgain in the cavity. At baseline operation conditions, P0 =60 Torr and 15% O2 in He, the flow rates are 60 mmole s−1 ofO2 and 345 mmole s−1 of He, with the total mass flow rate of3.33 g s−1. The flow rates have been determined from O2 andHe partial pressures in the discharge section using the chokedflow equation. Raising the discharge pressure to P0 = 80 Torrin the same mixture increases the flow rates up to 80 mmole s−1
(O2), 460 mmole s−1 (He) and 4.44 g s−1 (mass flow). NO isadded to the main O2–He flow from a cylinder containing5% NO in helium. In most experiments, NO–He mixtureflow rate, measured by a mass flow controller, was kept thesame, 7 standard litres per minute (SLM) or 6.6 mmole s−1
(NO flow rate 330 µmole s−1). At P0 = 60 and 80 Torr, thiscorresponds to 815 ppm and 610 ppm NO mole fraction in theflow, respectively.
Two 10 cm × 10 cm copper plate electrodes are flush-mounted in the top and bottom walls of the discharge section,2 cm apart, as shown in figure 1. The electrodes are coveredby alumina ceramic plates 1.6 mm thick to prevent secondaryelectron emission from the electrode surface and to improve thenanosecond pulse or the RF discharge uniformity and stability.
2
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
When the crossed pulsed/dc discharge is used, 5 cm × 2 cmcopper dc electrode plates in rectangular alumina ceramicholders are placed along the side walls. The holders preventdirect contact of the dc electrodes with the plastic walls.Channel walls downstream of the discharge are covered byalumina plates, to prevent damaging the plastic test section inthe case of discharge instability development and arc filamentformation. The pulsed electrodes are powered by a highvoltage (up to 30–40 kV), short pulse duration (5 ns), highpulse repetition rate (up to 100 kHz) FPG 60-100MC4 plasmagenerator (FID GmbH). The dc electrodes are connected toa Glassman 5 kV, 4 A dc power supply through a R = 1 k�
ballast resistor. For the experiments using capacitively coupledtransverse RF discharge, the dc electrodes were removed andthe side walls of the discharge section were protected by Tefloninserts. In these experiments, pulsed electrodes in the top andbottom walls were connected to a Dressler 13.56 MHz, 5 kWRF plasma generator with an automatic impedance matchingnetwork. In experiments with two identical discharge sectionsthe electrode pairs were connected parallel to each other.An additional variable inductor, adjusted manually betweenthe runs, was connected in series with the load to improveimpedance matching.
A mixture of iodine vapour and helium (I2 flow rate of upto 350 µmole s−1) is injected into the flow downstream of thedischarge through two aluminium injector blocks (see figure 1).Each injector block has a single row of 28 injection holes1.3 mm diameter each. The flow channel height at the injectorlocation is reduced to 0.6 cm, to increase the flow velocity.Based on the flow area ratio (channel height to nozzle throatheight ratio of 1.9), the Mach number at the injector location isM ≈ 0.3. Iodine vapour is produced by flowing helium overa bed of heated iodine crystals. Iodine flow rate is determinedby measuring the iodine vapour number density (by moleculariodine continuum absorption at 488 nm in an absorption cell),iodine–helium mixture pressure and temperature and heliumflow rate through the iodine crystal cell. Typical pressure inthe iodine vapour delivery line is 900 Torr, much higher thanthe main flow static pressure at the injector location, a fewtens of Torr, so that the injection flow is choked. Injectinghelium/iodine vapour mixture into the main flow considerablyincreases the discharge pressure, from the baseline value ofP0 = 60 Torr up to P0 = 79–83 Torr. At these conditions,injection helium flow rate, measured by a mass flow controller,is 75–90 SLM (71–85 mmole s−1), or 18–21% of the main flow,increasing the total flow rate of up to 0.5 mole s−1 (3.75 g s−1).Two BK-7 glass circular windows 10 mm in diameter, onelocated upstream of the injector and one downstream of theinjector, as shown schematically in figure 1, provide opticalaccess to the flow for emission spectroscopy measurements.
The discharge section, which also serves as a nozzleplenum, and the injector are followed by a 2.4 cm long M = 3nozzle with throat dimensions of 0.32 cm × 10 cm and asupersonic cavity 14 cm long (see figures 1 and 2). Thedistance from the injector to the nozzle throat is 5 cm. Thecavity height at the nozzle exit is 1 cm. The top and bottomwalls of the cavity diverge at 1.5◦ angle each to provideboundary layer relief (see figures 1 and 2). Each side wall
Figure 2. Photographs of the M = 3 nozzle/laser cavity profile andof I2 emission from the flow during operation. Both photographs areshown in the same scale. Flow is left to right.
has nine 12.7 mm diameter circular apertures providing opticalaccess to the flow in the cavity at different axial locations,up to 12.7 cm downstream from the end of the nozzle, asshown schematically in figure 1. The centre of the upstreamwindow is located at the end of the M = 3 nozzle. Twostainless steel arms 12.7 cm long and 3.8 cm in diameter canbe attached to the side walls of the cavity made of aluminium,forming a transverse laser cavity approximately 45 cm long.The supersonic laser cavity is followed by a gradual flowexpansion section 36 cm long, with 3◦ expansion angle on thetop and the bottom walls, attached to a 6 inch diameter vacuumpipe and vacuum tanks with the total volume of approximately1000 ft3, evacuated by a 150 ft3 min−1 Stokes vacuum pump.
For small signal gain measurements in the laser cavitywith gain path of 10 cm, nine pairs of wedged and anti-reflection coated BK-7 glass windows 12.7 mm in diameterare either placed in the side wall apertures, shown in figure 2,or attached to flanges at the end of the resonator arms. Thus,gain can be measured at multiple axial and transverse locationsin the cavity. Small signal gain at 1315 nm in the cavityis measured by tunable diode laser absorption spectroscopy(TDLAS) using a PSI iodine scan probe [12]. Positive gain ismeasured when iodine atom population inversion is achievedin the flow. For laser power measurements, the windows at theends of the arms are replaced with laser mirrors (lattice optics,99.9 ± 0.075% or 99.0 ± 0.4% reflectivity at 1315 nm andcurvature radius of 1 m), forming a stable resonator. The laseroutput power has been measured by a 2 inch aperture ScientechAC5000 thermopile calorimeter connected to Scientech S310power meter, monitoring both its digital readout and the time-dependent calorimeter output signal on the oscilloscope.
SDO yield in the discharge is determined from infraredemission spectroscopy measurements (O2(a
1� → X 3�)
spectra), using an optical multichannel analyzer (OMA) with a0.5 m spectrometer, 600 lines mm−1 grating blazed at 1 µm anda Roper Scientific liquid nitrogen-cooled 1D array 1024-pixelInGaAs PDA camera and calibrated using a blackbody source(Infrared Systems IR-564). The emission signal was collectedusing a Thor Labs 1 m long optical fibre with a 1.3 inchdiameter collimator on the collection end. The collimatorwas positioned in front of a window in the discharge sectionwall and the opposite end of the optical fibre was placed infront of the spectrometer slit. For calibration, the test section
3
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
0 5 10 15 20 25
1
10
100
Time, s
Pressure, Torr
plenum
upstream
downstream
Figure 3. Static pressures in the laser cavity. P0 = 80 Torr, mainflow 15% O2 in helium, helium injection flow rate 78 mmole s−1
(12% O2 in the flow after injection).
side wall, with a window in place, was placed between ablackbody source and the collimator lens and the blackbodyaperture was set to be the same as the diameter of the window,10 mm. Calibration showed that the signal intensity collectedby the optical fibre collimator remains nearly independent ofthe distance between the blackbody source and the collimatorlens (from 5 to 15 cm), indicating nearly equal contributionsof different ‘slices’ of a nearly cylindrical SDO emissioncollection region in the flow into the total signal intensity.The same OMA system has also been used to take visibleO2(b
1� → X 3�) emission spectra used to infer rotationaltemperature of the subsonic flow downstream of the discharge.SDO yield and flow temperature measurements have beenconducted using an optical access window upstream of theinjector (see figure 1).
3. Results and discussion
Figure 3 plots discharge (stagnation) pressure and staticpressures at two axial locations in the supersonic cavity (3.2and 12.7 cm downstream of the nozzle exit) as functions oftime. Pressure traces plotted in figure 3 are measured at thebaseline plenum pressure of P0 = 60 Torr, 15% O2 in helium,with additional helium flow, 78 mmole s−1, injected into thebaseline O2–He flow, which increased the plenum pressure upto P0 = 80 Torr and reduced oxygen mole fraction in the flowto 12%. From figure 3, it can be seen that the steady-staterun time, at a constant static pressure, is approximately 5 s.One can also see that during steady-state operation, upstreamand downstream static pressures are very close to each other,indicating nearly constant Mach number along the cavity. Thesteady-state static pressures, P1 = P2 = 2.9 Torr, correspondto near-design flow conditions at both locations, M = 2.9 atγ = 1.62 (12% oxygen in the O2–He flow downstream of theinjector). These static pressures are consistent with predictionsof the 3D compressible Navier–Stokes flow code [13] atthe same flow conditions and nozzle/laser cavity geometry,
Figure 4. Mach number distribution in the supersonic laser cavitypredicted by a 3D compressible Navier–Stokes flow code.P0 = 80 Torr, T0 = 380 K, 12% O2 in helium. Axial distance (fromthe nozzle throat) is in mm.
0 40 80 120 160 200Time [s]
0.0
1.0
2.0
3.0
Current [A] and Voltage [kV]
Current
Voltage
Figure 5. Typical current and voltage oscillograms of the scaled-upsustainer discharge. P0 = 80 Torr, 15% O2 in helium, NO/He flowrate 7 SLM, ν = 35 kHz, UPS = 3.4 kV, discharge power 2.9 kW.
P1 = 3.8 Torr (M1 = 2.7) and P2 = 3.0 Torr (M2 = 2.9).Figure 4 shows Mach number distribution in the supersoniccavity in a 12% O2 in helium flow at P0 = 80 Torr, T0 = 380 Kand assuming adiabatic wall boundary conditions. Thesecalculations predict that the height of the supersonic core flowregion, with the Mach number of M = 2.7–2.9, remains nearlyconstant, h ≈ 6 mm (see figure 4).
3.1. Pulser-sustainer discharge experiments
All measurements discussed in this section have beenconducted using a single discharge section. The scaled-uppulser-sustainer discharge tests showed the discharge to bediffuse, uniform and stable. Although at high dc voltagesmultiple filaments have been detected near the cathode, wherethe discharge remains self-sustained [14], the discharge neverlost stability in the entire range of dc sustainer power supplyvoltages tested, UPS = 0.5–3.75 kV. Figure 5 shows typicaloscillograms of sustainer discharge voltage and current atP0 = 80 Torr, 15% O2 in helium, NO/He flow rate 7 SLM (NOflow rate 330 µmole s−1), high-voltage pulse repetition rateν = 35 kHz and UPS = 3.4 kV. Note that the voltage between
4
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
0 400 800 1200 1600
Voltage [V]
0.0
0.5
1.0
1.5
2.0
2.5
Time-averaged current [A]
10% O2-He
15% O2-He
20% O2-He
Figure 6. Current–voltage characteristics of the scaled-up sustainerdischarge for different O2–He mixtures. P0 = 60 Torr, ν = 35 kHz.Cathode voltage fall is Uc ≈ 260 V for 10% O2 and Uc ≈ 470 V for15% O2 in helium.
the dc electrodes, plotted in figure 5, is significantly lower thanthe dc power supply voltage because of the voltage drop on theR = 1 k�ballast resistor in the dc circuit, U(t) = UPS−I (t)R.At the conditions of figure 5, the time-averaged dischargepower is 2.9 kW. The highest dc sustainer discharge powerachieved is 3.45 kW at P0 = 80 Torr (at UPS = 3.75 kVand time-averaged voltage and current of 〈U〉 = 1.61 kV,〈I 〉 = 2.14 A), which is significantly higher than the powerachieved in the smaller scale (5 cm × 5 cm × 2 cm) pulser-sustainer discharge, 2.4 kW at P0 = 107 Torr [6]. At a lowerplenum pressure, P0 = 60 Torr, the sustainer discharge powerachieved so far is 2.7 kW (at UPS = 3.3 kV, 〈U〉 = 1.5 kVand 〈I 〉 = 1.8 A), compared wih 1.8 kW in the small-scalelaser [6].
Figure 6 shows current–voltage characteristics of pulser-sustainer discharges in three different O2–He mixtures at P0 =60 Torr. Qualitatively, the dependence of the sustainer currenton voltage is similar to our previous measurements in thesmall-scale pulser-sustainer discharge [15]. In particular, thesustainer discharge current increases linearly with the voltageafter it exceeds the cathode voltage fall, Uc, which depends onthe gas mixture and the electrode material [16] (see figure 5).The linear dependence of the current on voltage is typical fora non-self-sustained discharge and occurs because the plasmaconductivity does not depend on voltage, which is too lowto produce ionization in the flow, except in the self-sustainedcathode layer [14].
Figures 7 and 8 show typical O2(b1� → X 3�) visible
emission and O2(a1� → X 3�) infrared emission spectra in
a 15% O2–He mixture at P0 = 60 Torr and sustainer dischargepower of 1.94 kW. The spectra are taken in the nozzle plenum,downstream of the discharge and upstream of the injector (seefigure 1). The rotational temperature of the flow (stagnationtemperature) at these conditions, inferred from the syntheticspectrum shown in figure 7, is T = 380 ± 20 K. Assuming thecathode voltage fall of Uc ≈ 470 V (see figure 6), the estimated
755 760 765 770 775
Wavelength [nm]
Intensity [Arb.]
Experiment
Calculation
Figure 7. Experimental O2(b1� → X 3�) emission spectrum
taken upstream of the injector and synthetic spectrum. P0 = 60 Torr,15% O2 in helium, NO/He flow rate 7 SLM, ν = 35 kHz,UPS = 2.8 kV, 〈I 〉 = 1.55 A, discharge power 1.94 kW. Inferredrotational temperature is T = 380 ± 20 K.
1230 1240 1250 1260 1270 1280 1290 1300
0
1000
2000
3000
4000
5000
Wavelength [nm]
Intensity [counts]
blackbody
singlet delta
Figure 8. Typical O2(a1� → X 3�) emission spectrum taken
upstream of the injector and blackbody emission spectrum usedfor calibration. Spectral region used for signal integration,�λ = 40 nm, is indicated with dashed lines.
sustainer discharge reduced electric field at these conditions isE/N ≈ 5 Td (1 Td = 10−17 V cm2). SDO emission spectrumin figure 8 is plotted together with the blackbody emissionspectrum at TBB = 800 ◦C (1073 K), used for calibration. SDOyield was calculated from the spectra such as shown in figure 8using the following equation:
Y = nSDO
nO2
=(
kT
yO2P
) (SSDO
SBB
) (tBB
tSDO
)4π
L
×(
I (λ0, TBB) · �λ
εA
), (1)
where T , P and yO2 are temperature, pressure and oxygen molefraction in the flow downstream of the discharge, respectively,SSDO and SBB are SDO and blackbody emission intensities
5
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
0.5 1.0 1.5 2.0 2.5
0
1
2
3
4
SDO yield, %
Sustainer discharge power, kW
P0=60 torr, 15% O2 in He
without NO
with NO
Figure 9. SDO yield versus sustainer discharge power, with andwithout NO added to the flow. 15% O2 in helium, NO/He flow rate7 SLM, P0 = 60 Torr.
1.0 1.5 2.0 2.5 3.0
0
1
2
3
4
SDO yield, %
Sustainer discharge power, kW
P0=80 torr, 15% O2 in He
without NO
with NO
Figure 10. SDO yield versus sustainer discharge power, with andwithout NO added to the flow. 15% O2 in helium, NO/He flow rate7 SLM, P0 = 80 Torr.
integrated over the same spectral range of �λ = 0.04 µm, asshown in figure 8, tSDO and tBB are camera gates used (typically,tSDO = 4 s and tBB = 4 ms), L = 0.1 m is the flow channelwidth, A = 2.2 × 10−4 s−1 is the Einstein coefficient for SDOspontaneous emission [17, 18], ε = 0.96 eV is the energy ofthe SDO emission quantum and I is the blackbody emissionspectral intensity (in W sr−1 µm−1 m−2) at λ0 = 1.268 µm andblackbody temperature of TBB = 1073 K, given by the Planckdistribution. The estimated uncertainty of the present SDOyield measurements is ±15%.
SDO yield versus discharge power in a 15% O2–Hemixture at two different discharge pressures, P0 = 60 and80 Torr, is plotted in figures 9 and 10. SDO yield hasbeen measured with and without NO added to the flow(at both plenum pressures, NO/He flow rate is 7 SLM). Itcan be seen that adding NO somewhat increases the yield,
0 3 6 9 12 15
0.0
1.0
2.0
3.0
SDO yield, %
NO flow rate, SLM
P0=60 torr, 1.0 kW
P0=80 torr, 2.1 kW
Figure 11. SDO yield versus NO/He flow rate at two differentdischarge pressures and sustainer discharge powers. 15% O2 inhelium.
by up to 30–35%. One can also see that although theyield increases with the sustainer discharge power (slightlyfaster than linearly), it remains fairly low, only 3.6–3.7%at the sustainer discharge power of 2.2–2.7 kW. This mostlikely occurs due to a fairly low energy loading per O2
molecule achieved in the pulser-sustainer discharge so far,0.35–0.38 eV/O2 molecule at the conditions of figures 9 and10. Measurements of SDO yield versus NO/He flow rate atdifferent discharge pressures and powers show that addingNO/He to the baseline O2–He flow (at 3–15 SLM) somewhatincreases the yield, up to 15–20% (see figure 11), consistentwith the results of figures 9 and 10.
Similar to the pulser-sustainer discharge, transverse RFdischarge sustained between two pulsed electrodes in the topand bottom walls of the test section remained diffuse, uniformand stable in the entire range of RF powers tested, up to 4.5 kW.Most of the runs were conducted in a relatively conservativeRF power range 0.5–3.0 kW. Measurements discussed inthis section have been conducted using a single dischargesections, unless specified otherwise. Flow temperatures inthe plenum downstream of the discharge, inferred from thevisible emission spectra such as shown in figure 7 and plottedin figure 12, are consistent with the temperatures in thepulser-sustainer discharge. At these energy loadings, 0.39–0.52 eV/O2 molecule at the conditions of figure 12, flowtemperature rise in the discharge is about 25 K kW−1 andtemperature does not exceed approximately 380 K at the RFdischarge power of 3 kW (see figure 12). Note that in theflow excited by the pulser-sustainer discharge, comparabletemperature is achieved at the dc sustainer discharge power ofapproximately 2 kW (see figure 7). Somewhat higher apparenttemperature in the pulser-sustainer discharge could be due toa larger discharge power fraction thermalized as heat or due tohigher temperature filament formation in the cathode layer ofthe dc discharge.
6
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
0.5 1.0 1.5 2.0 2.5 3.0
300
350
400
RF discharge power, kW
Temperature, K
P0=60 torr
P0=80 torr
Figure 12. Temperature downstream of the RF discharge versusdischarge power. 15% O2 in helium, NO/He flow rate 7 SLM.
0.5 1.0 1.5 2.0 2.5 3.0
0
1
2
3
4
SDO yield, %
RF discharge power, kW
P0=60 torr
P0=80 torr
Figure 13. SDO yield versus RF discharge power at two differentdischarge pressures, 60 and 80 Torr, both without helium injection.15% O2 in helium, NO/He flow rate 7 SLM.
Figures 13 and 14 plot SDO yield versus RF dischargepower. Yield data plotted in figure 13 are measured at dischargepressures of P0 = 60 and 80 Torr, 15% O2 in helium, withoutadditional helium flow through the injector downstream of thedischarge section. Yield in figure 14 is measured at dischargepressures of P0 = 60 Torr (without helium injection) andP0 = 86 Torr (while injecting helium to raise the plenumpressure). Comparing figures 13, 14 and figures 9, 10, it canbe seen that SDO yields in the pulser-sustainer discharge andin the transverse RF discharge are comparable, and appearto be controlled primarily by the energy loading per oxygenmolecule. In the RF discharge, raising the discharge pressureby increasing the flow rate of the main oxygen–helium mixturenoticeably reduces the yield (by 25–30%, see figure 13).On the other hand, when the discharge pressure is increasedby injecting helium downstream, SDO yield decreases onlyslightly, since this approach does not increase the oxygen flow
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
0
1
2
3
4
SDO yield, %
RF discharge power, kW
P0=60 torr, no He injection
P0=83 torr, with He injection
Figure 14. SDO yield versus RF discharge power at two differentdischarge pressures, 60 Torr (without helium injection) and 83 Torr(with helium injection). 15% O2 in helium, NO/He flow rate 7 SLM.
rate through the discharge (see figure 14 and discussion insection 2).
The fact that SDO yields in the pulser-sustainer and RFdischarges are comparable is not very surprising. Althoughthe theoretically predicted reduced electric field at which SDOyield in 10–20% O2–He mixtures peaks isE/N = 4–5 Td [15],this maximum is fairly broad and is reduced by only about30% at E/N = 10 Td. Therefore SDO yields in the non-self-sustained dc discharge, at E/N = 5 Td (see section 3.1), and inself-sustained dc and RF discharges, at a higher E/N , may befairly close. This is consistent with the present measurements.At P = 60 Torr and discharge power of 2 kW, SDO yield is3.2% in the pulser-sustainer discharge versus 2.7% in the RFdischarge (see figures 9 and 13). At P = 80 Torr and dischargepower of 2.75 kW the yield is 3.6% in the pulser-sustainerdischarge versus 2.8% in the RF discharge (see figure 10and 13).
From figure 14, it can be seen that at RF discharge powerof 3.0–4.5 kW, SDO yield nearly levels off at approximately3.5%. This might be caused by lower discharge power fractiongoing into O2(a
1�) excitation by electron impact at higherreduced electric fields (E/N) in the discharge as the poweris increased [15]. However, using two discharge sectionsconnected in parallel to increase the current and reduce thevoltage across the gap at the same RF power did not result inhigher SDO yield, which remained nearly the same. Figure 15shows that increasing oxygen percentage in the O2–He mixturefrom 10% to 20% steadily reduces SDO yield, which is highestat the lowest O2 fraction tested, 10%. Figure 16 demonstratesthat in the RF discharge the yield weakly depends on the NOmole fraction in a wide range of NO/He flow rates (0–15 SLM).
In the present experiments, 3 kW transverse RF dischargewas operated for up to 10 s, with no sign of instabilitydevelopment of arc filament formation. After a few hundredruns, the discharge section did not exhibit any sign of damage.We believe that the discharge power can be scaled fromthe currently achieved value of 4.5 kW to ∼10 kW using
7
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
10 15 20
0.0
1.0
2.0
SDO yield, %
% O2 in the mixture
Figure 15. SDO yield versus oxygen fraction in helium.P0 = 60 Torr, RF discharge power 1.0 kW, NO/He flow rate 7 SLM.
0 3 6 9 12 15
0.0
0.5
1.0
1.5
2.0
SDO yield, %
NO flow rate, SLM
Figure 16. SDO yield versus NO/He flow rate. P0 = 60 Torr, RFdischarge power 1.0 kW. 15% O2 in helium.
another 5 kW RF power supply to power the second dischargesection. Note that in comparison with the pulser-sustainerdischarge, RF discharge generates significantly less EMI noiseand produces essentially no interference with diagnostics(temperature controllers, infrared camera, OMA spectrometerand gain probe). However, both types of discharges aresimple in operation, scalable to comparable powers withoutproducing plasma instabilities and generate comparable SDOyields (compare figures 9, 10 and figures 13, 14). Long-termreliability of nanosecond pulse power supplies has long beena concern. For example, the FID pulser used in the presentexperiments broke twice over a period of approximately 3years, while Dressler RF generator has not experienced asingle malfunction over more than 8 years. However, theOSU group has recently acquired capability of building robustand versatile nanosecond pulse generators in-house, whichessentially resolves this issue.
When iodine vapour in helium carrier flow was injectedinto the main flow downstream of the RF discharge, iodine
1400 1600 1800 2000 2200
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Frequency, MHz
Gain, %/cmExperiment
Doppler fit, T=140 K
Figure 17. 1315 nm gain line shape profiles measured at x = 3.2 cmfrom the end of the nozzle and Doppler fit indicating laser cavitytemperature of T = 140 K. P0 = 80 Torr, main flow 15% O2 inhelium, RF power 2.75 kW, NO/He flow rate 7 SLM, I2 flow rate175 µmole s−1.
atom population inversion was achieved and gain on an iodineatom line at 1315 nm was measured in the supersonic section.Figure 17 shows a typical gain line shape measured on the flowcentreline at x = 3.2 cm, at P0 = 80 Torr (main flow 15% O2
in helium, NO/He flow 7 SLM, I2 flow rate of 175 µmole s−1)
and RF power of 2.75 kW. During these and all subsequentgain and laser power measurements, the helium carrier flowrate through the injector was 80 mmole s−1, which raised thedischarge pressure from P0 = 60 Torr before the injection toapproximately P0 = 80 Torr. As discussed in section 2, thiswas done to increase the pressure ratio across the injector, tomake sure that the injection flow was choked, and to producerapid mixing of injected iodine vapour with the main flow. Thegain line shown in figure 17 is sampled and averaged by thegain probe over 0.2 s. In figure 17, gain at the line centre isγ = 0.122% cm−1, which corresponds to 2.4% gain per doublepass of 20 cm. This greatly exceeds losses on 99.99% outputcoupler mirrors used in our previous work [5–7] to measure theoutput laser power and suggests that a lower reflectivity outputcoupler, down to 99%, can be used to couple the laser powerout. Doppler fit to the line shape gives the cavity temperatureof T = 140 K (see figure 17).
Figure 18 shows that after the RF discharge is turned on,gain remains steady, within about 5%, during the entire runapproximately 6.5 s long. In figure 18, gain is sampled andaveraged every 0.1 s. Figure 18 represents a typical resultfor the present conditions, where the flow/plasma parameters,such as pressure, gas mixture, discharge power, temperatureand SDO yield are stable and well reproducible run-to-run.One parameter which critically affects gain and which wasrather difficult to control and reproduce was iodine vapourflow rate. We believe that this may be caused by solid phaseiodine coating the I2 vapour/helium delivery lines and injectorblocks. Figure 19 shows gain dependence on the iodine vapourflow rate at P0 = 86 Torr, 15% O2 in helium, RF dischargepower 3.0 kW and NO/He flow rate 7 SLM. It can be seenthat gain falls off on both sides of the optimum iodine flow
8
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
1 2 3 4 5 6 7 8
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Time, s
Gain, %/cm
Figure 18. Gain at the line centre measured at x = 3.2 cm from theend of the nozzle versus time. P0 = 86 Torr, 15% O2 in helium, RFpower 3.0 kW, NO/He flow rate 7 SLM, I2 flow rate 140 µmole s−1.
0 50 100 150 200 250 300 350
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Iodine flow rate, µmole/s
Gain, %/cm
Figure 19. Gain versus iodine flow rate. P0 = 86 Torr, 15% O2 inhelium, RF discharge power 3.0 kW, NO/He flow rate 7 SLM.
rate of 150–200 µmole s−1. This emphasizes the importanceof iodine flow control for gain reproducibility run-to-run.Figure 20 plots gain dependence on the NO flow rate at thesame flow conditions as in figure 19 and at the iodine flowrate of 140 µmole s−1. One can see that the effect of addingNO to the flow on gain is considerably stronger than on SDOyield (compare figures 16 and 20). Indeed, adding NO tothe baseline O2–He mixture increases yield only slightly (seefigure 16) but dramatically affects gain, which changes from0.01% absorption without NO in the flow to 0.11% cm−1 gainwith 7–10 SLM of NO/He added (see figure 20). This againdemonstrates critical importance of scavenging excess O atomsgenerated in the discharge by adding nitric oxide to the flow.
Figure 21 plots gain versus RF discharge power, measuredon the centreline of the supersonic flow 3.2 cm downstream ofthe M = 3 nozzle exit, for a single discharge section and fortwo consecutive sections. These measurements are conductedat P0 = 80 Torr (main flow 15% O2 in helium, NO/He flow
0 3 6 9 12 15
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
NO flow rate, SLM
Gain, %/cm
Figure 20. Gain versus NO/He flow rate. P0 = 86 Torr, 15% O2 inhelium, RF discharge power 3.0 kW, I2 flow rate 140 µmole s−1.
0.0 1.0 2.0 3.0 4.0
0.00
0.02
0.04
0.06
0.08
0.10
0.12
RF discharge power, kW
Gain, %/cm
One discharge section
Two discharge sections
Figure 21. Gain versus RF discharge power for one and twotransverse discharge sections. P0 = 80 Torr, main flow 15% O2
7 SLM, I2 flow rate of 130 µmole s−1). It can be seen that gainmeasured using a single discharge section increases with powerand shows a tendency of leveling off at approximately γ =0.11% cm−1 at 3.5 kW. At these conditions, flow temperaturein the supersonic section, inferred from the gain line shape,varied from T = 110 K at 1.0 kW to T = 125 K at 3.5 kW.In contrast, gain measured using two consecutive dischargesections peaks at near 2.75 kW and decreases at higher powers.This is likely due to somewhat higher flow temperaturesachieved in two consecutive discharges, T = 135 K at 3.5 kW.Longer flow residence time in the discharge may result inadditional relaxation of energy stored in the internal energymodes. This suggests that injection of pre-cooled helium ornitrogen into the main flow upstream of the nozzle may be usedto reduce the temperature in the laser cavity, which would resultin higher gain.
Limited gain measurements conducted using a pulser-sustainer discharge (sustained in a single discharge section)produced results similar to the RF discharge data plotted
9
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
0 2 4 6 8 10 12
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Axial distance, cm
Gain, %/cm
Figure 22. Centreline gain versus axial distance (from the endof the nozzle) along the cavity. RF discharge power 2.75 kW,P0 = 80 Torr (main flow 15% O2 in helium, NO/He flow rate7 SLM, I2 flow rate 150 µmole s−1).
in figure 21. With pulser-sustainer, gain peaked at γ =0.07% cm−1 at the sustainer discharge power of 2.6 kW. Thisdemonstrates that the two types of discharges, in addition togenerating similar SDO yields, also produce comparable gainin the laser cavity.
Figure 22 shows gain dependence on axial distance alongthe cavity, measured on the flow centreline at the RF dischargepower of 2.75 kW and P0 = 80 Torr (main flow 15% O2 inhelium, NO/He flow 7 SLM, I2 flow rate of 150 µmole s−1).Gain was measured at nine axial locations, ranging from x = 0(through the window centred at the M = 3 nozzle exit) tox = 12.7 cm. From figure 22, it can be seen that gain inthe supersonic section first slightly increases and then remainsnearly constant, γ = 0.10–0.12% cm−1, through the restof the cavity (over approximately 10 cm). Gain dependenceon the transverse (vertical) distance, measured at two axiallocations, x = 3.2 cm and x = 12.7 cm, is shown in figure 23.The uncertainty in the transverse location of the gain probelaser beam is of the order of half the gain probe laser beamdiameter, ∼1 mm. It can be seen that positive gain region (upto γ = 0.11% cm−1) extends over about half of the cavityheight at the upstream location (i.e. over approximately 6 mm)and over about one third of the cavity height at the downstreamlocation (i.e. over approximately 5 mm). At both locations,gain falls off and becomes absorption in the boundary layersnear the top and bottom walls of the cavity.
Figures 24 and 25 compare axial and transverse distri-butions of line-of-sight integrated temperature distributionsin the cavity, inferred from gain/absorption line shapes andpredicted by the 3D compressible Navier–Stokes code [13]at the conditions of figures 22 and 23 (P0 = 80 Torr,T0 = 380 K, 12% in O2 in helium and assuming adiabaticwall boundary conditions). As expected, experimental axialtemperature distribution on the centreline remains nearlyuniform, T = 125–140 K (see figure 24). Transversetemperature distributions measured at x = 3.2 cm and x =12.7 cm show nearly uniform temperature in the supersonic
-8 -6 -4 -2 0 2 4 6 8
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Vertical distance, mm
Gain, %/cm
x=3 cm
x=13 cm
Figure 23. Gain versus vertical distance across the cavity at twodifferent axial locations, at the conditions of figure 22. Cavityheight at x = 3 cm h = 11.5 mm, at x = 13 cm h = 15.5 mm.
0 2 4 6 8 10 12
0
50
100
150
Axial distance, cm
Temperature, K
Figure 24. Comparison of the centreline temperature at theconditions of figure 22 with the 3D Navier–Stokes code prediction.
inviscid core approximately 6 mm thick, 120–140 K, withtemperature in the boundary layers being much higher, 280–340 K (see figure 25). The uncertainty of temperatureinference from gain/absorption line shape is ±5 K in theinviscid core and ±10 K in the boundary layer. The boundarylayer thickness predicted by the code is consistent with theexperimental results within the uncertainty of the probe beamlocation, approximately 1 mm. From figures 24 and 25, it canbe seen that the experimental temperatures in the boundarylayer are somewhat higher than those predicted by the code,which does not incorporate temperature rise due to O atomrecombination and SDO relaxation, as well as the effect ofI2–He injection on the flow field.
Laser power measurements have been conducted usingtwo mirror combinations, one with two 99.9% reflectivitymirrors on both resonator arms, and the other using a 99.9%reflectivity mirror on one arm and a 99% reflectivity mirroron the other. The resonator centreline was located 3.2 cm
10
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
-8 -6 -4 -2 0 2 4 6 8
0
50
100
150
200
250
300
350
Vertical distance, mm
Temperature, K
x=3 cm
x=3 cm, code
x=13 cm
x=13 cm, code
Figure 25. Comparison of the temperature versus vertical distanceacross the cavity at two different axial locations, at the conditions offigure 22, with the 3D Navier–Stokes code prediction.
downstream of the nozzle exit. As expected, in the first caselaser powers measured on both sides were close to each other.In the second case, power coupled out on the 99% mirror sidewas much higher than power measured on the 99.9% mirrorside, which was barely detectable.
During the laser power measurements, it was determinedthat the power meter digital readout considerably overpredictsthe true laser power (typically by about 50%) until thethermopile calorimeter reaches steady state, which occurs after∼10–30 s. At steady-state, the power shown by the readoutbecomes accurate. We believe that the power ‘overshoot’artefact is entirely due to the flawed input signal processingalgorithm used by the power meter readout, aimed at reducingthe apparent time lag of power displayed. As discussed above,the DOIL laser run time at steady-state design conditions isonly about 5 s (see figure 3), which means that the power meterreadout cannot be used for accurate power measurements atthese conditions. To resolve this difficulty, the calorimeter wascalibrated using a CO laser operating at a known c.w. power,by taking oscillograms of the time-varying calorimeter outputsignal (i.e. without using the readout) after the laser beam wasunblocked. Figure 26 plots typical thermopile calorimeteroutput signal oscillograms taken at three different CO laserpowers, 1.0 W, 3.0 W and 4.5 W. Also shown in figure 26 arebest fits using the following equation,
This equation, which provides a very good fit of the powermeter oscillograms in the CO laser power range from 0.75 Wto 5.0 W, has been used to infer the DOIL laser power from thecalorimeter oscillograms taken during its operation. Note thatthis approach is accurate only if the DOIL laser power remainsconstant in time, and may well underestimate it if there is a briefpower overshoot in the beginning of the run. To incorporatethis effect, the relative output laser power also needs to bemonitored during the run, using a photodiode detector.
0 2 4 6 8 10
0.00
0.05
0.10
0.15
0.20
0.25
Time, s
Voltage, V
1 W
3 W
4.5 W
Figure 26. Summary of thermopile calorimeter calibration using aCO laser operating at a constant c.w. power. Symbols, oscillogramsof time-varying calorimeter signal; lines, best fits given byequation (2).
0 2 4 6 8 10
0.00
0.05
0.10
Time, s
Voltage, V
99.9% - 99.9%
1.1 W fit
99.9% - 99%
3.1 W fit
Figure 27. Power meter oscillograms and best fits used for DOILlaser power measurements for two mirror combinations atx = 3.2 cm: 99.9–99.9% (single-side power 1.1 W) and 99.9–99%(total power 3.1 W). Deviation from the 3.1 W fit is due to laserpower falling during the run.
Figure 27 plots thermopile calorimeter oscillograms takenat P0 = 80 Torr (main flow 15% O2 in helium, NO/Heflow 7 SLM, I2 flow rate of 150 µmole s−1) and RF dischargepower of 2.75 kW, as well as exponential fits used for DOILlaser power inference, using 99.9–99.9% and 99.9–99% mirrorcombinations. It can be seen that the first mirror combinationproduces a constant in time 1.1 W laser power output, withthe uncertainty of approximately ±0.05 W. In this case, thepowers measured on the two sides of the resonator were closeto each other, and the highest power measured on both sidesof the resonator combined was 2.5 ± 0.1 W (1.2 W + 1.3 W).The second mirror combination produced power output of3.1 ± 0.1 W (see figure 27), nearly all of which was measuredon the 99% mirror side, as expected. From figure 27, one canalso see that in this case laser power starts falling off after
11
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
Table 1. Gain measured without lasing and during lasing.
Gain during lasing (%/cm)Gain without
Laser lasing Lasermirrors x = 3.2 cm (%/cm) power x = 6.3 cm x = 9.6 cm x = 12.7 cm
99.9–99.9% 0.112 2.5 W 0.085 0.090 0.08299.9–99% 0.106 3.1 W 0.100 0.100 —
approximately 2.5 s of steady-state operation. Note that peaksingle-side powers given by the power meter readout duringthese two runs were 2.5 W and 4.6 W, respectively, whichillustrates a significant error given by the readout on the shorttime scale (a few seconds).
To estimate the ratio of power coupled out and poweravailable for lasing, we compared gain measured at theresonator location at x = 3.2 cm before installing the mirrormounts (i.e. without lasing) and gain measured at severaldownstream axial locations during lasing. Note that gainmeasured without lasing remains approximately the samethroughout the entire cavity (see figure 22). The resultsare summarized in table 1. It can be seen that lasing withtwo 99.9% output couplers reduces gain downstream of theresonator by 20–30%. Lasing with 99.9% and 99% mirrorsresults in gain reduction by only about 6%. In both cases,gain downstream of the resonator remains nearly independentof the axial distance (see table 1). In the absence of detailedgain recovery measurements (i.e. gain versus axial distancedownstream of the laser resonator), these data are insufficientto conclude whether (a) gain recovers rapidly (over less than3 cm) after it is brought to near transparency by lasing inthe resonator or (b) gain reduction in the resonator is fairlyminor and gain recovery is delayed significantly (over morethan 10 cm). We believe that the first scenario (i.e. rapidgain recovery) is more likely since it is consistent with recente-COIL laser gain recovery length measurements at similarconditions, 2–3 cm [19]. In either case, the present resultsdemonstrate that additional power, at about the same level, maybe coupled using the second resonator placed a few centimetresdownstream of the first resonator location.
To confirm this, an additional series of laser powermeasurements was conducted using two laser resonatorsoperated together, one located at x = 3.2 cm and the otherat x = 12.7 cm. Both resonators had a 99.9% mirror (‘totalreflector’) on one side and a 99% mirror (‘output coupler’)on the other side. Prior to these power measurements,gain has been measured at both axial locations to verifythat it remains fairly constant along the cavity (such asshown in figure 22), γ = 0.105% cm−1 upstream and γ =0.080% cm−1 downstream. During the power measurements,gain was also measured between the resonators, at x = 9.6 cm,γ = 0.10% cm−1. The power measurements demonstratedsimultaneous lasing in both resonators, with the total powerof 3.8 ± 0.1 W (2.9 ± 0.05 W upstream and 0.9 ± 0.05 Wdownstream, see figure 28). Lower power measured at thedownstream location is likely due to somewhat lower gain andsmaller height of positive gain region (see figure 23). Wealso believe that difference in the reflectivity of the two outputcoupler mirrors, 99.0 ± 0.4%, used in the present experiments
0 2 4 6 8 10
0.00
0.05
0.10
0.15
0.20
Time, s
Voltage, V
2.9 W (upstream)
0.9 W (downstream)
3.8 W (total)
Figure 28. Power meter oscillograms and best fits used for DOILlaser power measurements at two different axial locations,x = 3.2 cm (2.9 W) and x = 12.7 cm (0.9 W), with the total powerof 3.8 W. Both resonators use 99.9–99% mirror combinations.
may be sufficient to affect the output laser power significantly.Finally, fairly low output power may well be affected by poormirror alignment, which is difficult to adjust during the shortrun time available. In the present experiments, laser mirrors arealigned only once prior to each series of power measurement.
Basically, the present results suggest that only a fairlysmall fraction of power available for lasing is coupledout. For comparison, total power stored in SDO at thepresent conditions, obtained from SDO yield measurements,is approximately 200 W (at SDO yield of 3.5% and SDO flowrate through the laser cavity of 2.0 mmol s−1). Power stored inexcited iodine atoms is proportional to iodine atom populationinversion,
(nI∗ − 1
2nI) = γ /σ ≈ 1014 cm−3, (3)
where nI∗ and nI are the excited and the ground state iodineatom number densities, γ is gain and σ is the stimulatedemission/absorption cross section, σ = 1.1 × 10−17 cm2 atT = 140 K [7]. For γ = 0.1% cm−1, equation (3) gives thedifference between [I∗] and [I]/2 flow rates through the testsection of approximately 0.2 mmole s−1 (assuming uniformflow). Remembering that only about half of the flow isoccupied by the gain region in the inviscid core (see figure 23),this gives estimated power stored in excited iodine atoms of∼10 W, which is comparable to the laser power measured inthe present experiments but more than an order of magnitudelower than the power stored in SDO. This result, along withgain measurements downstream of the resonator (see table 1)
12
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
demonstrates significant potential in further optimization ofthe laser power.
4. Summary
In this work, an electric discharge excited oxygen–iodinelaser apparatus has been successfully scaled to increase theelectric discharge volume and power, the laser mixture flowrate and the gain path in the M = 3 laser cavity. Specifically,SDO generator discharge power has been increased upto 4.5 kW, laser mixture flow rate up to approximately0.5 mole s−1, and gain path up to 10 cm. The steady-state run time of the new scaled-up laser apparatus at theseconditions is approximately 5 s, with the Mach numberalong the laser cavity nearly constant, M = 2.9. Twodifferent types of discharges have been used to generate SDO,crossed nanosecond pulser/transverse dc sustainer dischargeand capacitively coupled transverse RF discharge. Flowtemperature downstream of the discharge, SDO yield and gainin supersonic cavity have been measured in a wide rangeof discharge power, nitric oxide mole fraction in the mainoxygen–helium flow, oxygen percentage in the mixture andiodine vapour flow rate, at discharge pressures ranging from60 to 90 Torr.
The results demonstrate that SDO yields are comparablefor both types of discharges. Highest yields achievedso far remain rather low, 3.6–3.7%, due to fairly lowenergy loading per oxygen molecule in the discharge, 0.3–0.5 eV/O2 molecule. SDO yield moderately increases whena small amount of NO (a few hundred ppm) is addedto the flow. Reducing oxygen percentage in the O2–Hemixture also increases SDO yield. The results show thatadding a few hundred ppm of NO to the O2–He flow iscritical for producing gain in the laser cavity. WithoutNO, absorption or very weak gain have been detected whileadding NO produced gain exceeding 0.1% cm−1, at the samedischarge and flow conditions. Gain at high RF dischargepowers (above 3 kW) using a single discharge section nearlylevels off, approximately at 0.11% cm−1. The use of twoconsecutive identical discharge sections, with flow residencetime in the discharge doubled, resulted in slightly lowergain measured at the same discharge power. Also, gainmeasured with two discharge sections peaks at 2.75 kW andstarts decreasing at higher RF powers. Comparable gain hasbeen measured at the same flow conditions using a pulser-sustainer discharge, 0.07% cm−1 at sustainer discharge powerof 2.6 kW. This suggests that replacing the pulser-sustainerdischarge with transverse RF discharge does not result in amarked improvement of DOIL laser performance.
Gain and static temperature measurements at multipleaxial locations in the supersonic section demonstrated nearuniform gain and temperature distributions along the cavity,0.10–0.12% cm and T = 125–140 K, over the distance ofapproximately 10 cm. Highest gain measured is 0.122% cm−1
at T = 140 K (2.4% per double pass). Gain dependence on thetransverse (vertical) distance, measured at two axial locations,x = 3.2 cm and x = 12.7 cm from the nozzle exit, showthat positive gain region (up to 0.11% cm−1) extends over one
third to one half of the cavity height, i.e. over approximately5–6 mm, with absorption measured in the boundary layersnear top and bottom walls of the cavity. Static temperaturedistributions in the supersonic inviscid core approximately6 mm thick are nearly uniform, T = 120–140 K. In theboundary layers, static temperature increases up to T = 280–340 K.
Laser power has been measured using two differentresonator configurations, (i) 99.9% output couplers on bothresonator sides, 2.5 W, and (ii) a 99.9% mirror on one side anda 99% output coupler on the other side, 3.1 W. The latter resultdemonstrates that at the present conditions (double pass gainup to 2.4%) laser power can be coupled out without using high-reflectivity output couplers. Gain measurements downstreamof the resonator during lasing demonstrate moderate gainreduction at these conditions, 20–30% for the first mirrorcombination and only 6% for the second mirror combination.Gain downstream of the resonator remains nearly independentof the axial distance, by up to 10 cm. This demonstrates thatonly a small fraction of power stored in the flow and availablefor lasing is coupled out, and that additional power may becoupled using the second resonator placed downstream ofthe first resonator location. Indeed, preliminary laser powermeasurements using two transverse resonators operating atthe same time (both using 99.9%–99% mirror combinations)demonstrated lasing at both locations, with the total power of3.8 W. Further laser power and gain recovery measurementsare currently underway.
Acknowledgments
This work was supported by the Joint Technology Office. Theauthors would like to express our gratitude to Drs T Maddenand D Hostutler of AFRL.
References
[1] Benavides G F, Zimmerman J W, Woodard B S, Carroll D L,Verdeyen J T, Field T H, Palla A D and Solomon W C 2008Gain and continuous-wave laser power enhancement with asecondary discharge to predissociate molecular iodine in anelectric oxygen–iodine laser Appl. Phys. Lett. 92 041116
[2] Zimmerman J W, Woodard B S, Benavides G F, Carroll D L,Verdeyen J T, Palla A D and Solomon W C 2008 Gain andcontinuous-wave laser power enhancement with a multipledischarge electric oxygen–iodine laser Appl. Phys. Lett.92 241115
[3] Zimmerman J W, Benavides G F, Woodard B S, Carroll D L,Palla A D, Verdeyen J T and Solomon W C 2009Measurements of Improved electricOIL performance, gain,and laser power 40th AIAA Plasmadynamics and LasersConf. (San Antonio, TX, 22–25 June 2009) AIAA Paper2009-4059
[4] Hicks A, Tirupathi S, Jiang N, Utkin Yu, Lempert W R,Rich J W and Adamovich I V 2007 Design and operation ofa supersonic flow cavity for a non-self-sustained electricdischarge pumped oxygen–iodine laser J. Phys. D: Appl.Phys. 40 1408–15
[5] Hicks A, Utkin Yu G, Lempert W R, Rich J W andAdamovich I V 2006 Continuous wave operation of anon-self-sustained electric discharge pumpedoxygen–iodine laser Appl. Phys. Lett. 89 241131
J. Phys. D: Appl. Phys. 43 (2010) 015201 J R Bruzzese et al
[6] Hicks A, Bruzzese J, Lempert W R, Rich J W andAdamovich I V 2007 Effect of nitric oxide on gainand output power of a non-self-sustained electricdischarge pumped oxygen–iodine laser Appl. Phys. Lett.91 071116
[7] Hicks A, Bruzzese J, Nishihara M, Lempert W R, Rich J Wand Adamovich I V 2008 Effect of iodine dissociation in anauxiliary discharge on gain in a pulser-sustainer dischargeexcited oxygen–iodine laser 46th Aerospace SciencesMeeting and Exhibit (Reno, NV, 7–10 January 2008) AIAAPaper 2008-1078
[8] Generalov N A, Zimakov V P, Kosynkin V D, Raizer Yu P andRoitenburg D I 1975 Tech. Phys. Lett. 1 431
Generalov N A, Zimakov V P, Kosynkin V D, Raizer Yu P andRoitenburg D I 1977 Sov. Plasma Phys. 3 626
Generalov N A, Zimakov V P, Kosynkin V D, Raizer Yu P andRoitenburg D I 1980 Sov. Plasma Phys. 6 1152
[9] Brown C O and Davis J W 1972 Appl. Phys. Lett.21 480
[10] King D M, Carroll D L, Verdeyen J T, Laystrom J K,Benavides G F, Palla A D, Zimmerman J W, Woodard B Sand Solomon W C 2006 Power enhancement of the hybridElectriCOIL laser 37th AIAA Plasmadynamics and LasersConf. (San Francisco, CA, 5–8 June 2006) AIAA Paper2006-3756
[11] Carroll D L, Verdeyen J T, King D M, Zimmerman J W,Laystrom J K, Woodard B S, Benavides G F, Kittell K Wand Solomon W C 2005 IEEE J. Quantum Electron.41 213
[12] Davis S J, Allen M G, Kessler W J, McManus K R, Miller M Fand Mulhall P A 1996 Diode laser-based sensors forchemical oxygen iodine laser Proc. SPIE 2702 195
[13] Gaitonde D V 2001 High-order solution procedure forthree-dimensional nonideal magnetogasdynamics AIAA J.39 2111–20
[14] Velikhov E P, Kovalev A S and Rakhimov A T 1987Physical Phenomena in Gas Discharge Plasmas(Moscow: Nauka)
[15] Hicks A, Norberg S, Shawcross P, Lempert W R, Rich J W andAdamovich I V 2005 Singlet oxygen generation in a highpressure non-self-sustained electric discharge J. Phys. D:Appl. Phys. 38 3812–24
[16] Raizer Yu P 1991 Gas Discharge Physics (Berlin: Springer)chapter 13
[17] Newman S M, Lane I C, Orr-Ewing A J, Newnham D A andBallard J 1999 Integrated absorption intensity and Enisteincoefficients for the O2(a
1�g → X 3�−g ) (0,0) transition: a
comparison of cavity ringdown and high resolution Fouriertransform spectroscopy with a long-path absorption cellJ. Chem. Phys. 110 10749
[18] Lafferty J, Solodov A M, Lugez C L and Fraser G T 1998Rotational line strength and self-pressure-broadeningcoefficients for the 1.27 µm a 1�g → X 3�−
g v = 0 → 0band of O2 Appl. Opt. 37 2264
[19] Zimmerman J W, Benavides G F, Palla A D, Woodard B S,Carroll D L, Verdeyen J T and Solomon W C 2009 Gainrecovery in an electric oxygen–iodine laser Appl. Phys. Lett.94 021109
