1

Measurements of Temperature and Hydroxyl Radical Generation / Decay in Lean Fuel-Air Mixtures Excited by a Repetitively Pulsed Nanosecond Discharge1

Z. Yin2, A. Montello2, W.R. Lempert3, and I.V. Adamovich3

Michael A. Chaszeyka Nonequilibrium Thermodynamics Laboratories Department of Mechanical Engineering,

The Ohio State University, Columbus, OH 43210

Abstract

OH Laser Induced Fluorescence (LIF) and psec, broadband Coherent Anti-Stokes Raman Spectroscopy (CARS) are used for time-resolved temperature and time-resolved, absolute OH number density measurements in lean H2-air, CH4-air, C2H4-air, and C3H8-air mixtures in a nanosecond pulse discharge cell / plasma flow reactor. The premixed fuel-air flow in the reactor, initially at T0=500 K and P=100 torr, is excited by a repetitive nanosecond pulse discharge in a plane-to-plane geometry, operated in burst mode at 10 kHz pulse repetition rate. In most measurements, burst duration is limited to 50 pulses, to preclude plasma-assisted ignition. The discharge uniformity in air and fuel-air flows is verified using sub-nsec gate ICCD camera images. The results demonstrate that rotational temperature measured by OH LIF and CARS are close to each other. Temperatures measured at the end of the discharge burst are in the range of T=550-600 K, and remain essentially unchanged for up to 1 msec delay after the burst. Time-resolved temperature measured by CARS during plasma-assisted ignition of H2-air is in very good agreement with kinetic model predictions. Based on CARS measurement, vibrational disequilibrium is not a significant factor at the present conditions.

Time-resolved, absolute OH number density measurements after the discharge burst demonstrate that OH concentration in C2H4-air, C3H8-air, and CH4 is highest in lean mixtures. In H2-air, OH concentration is nearly independent of the equivalence ratio. In C2H4-air and C3H8-air, unlike in CH4-air and in H2-air, transient OH overshoot after the discharge is detected. In C2H4-air and C3H8-air, OH decays after the discharge on the time scale of ~20-100 μsec, suggesting little accumulation during the burst of pulses repeated at 10 kHz. In CH4-air and H2-air, OH decays within ~0.1-1.0 msec and 0.5-1.0 msec, respectively, showing that it may accumulate during the burst. In C2H4-air and C3H8-air, kinetic modeling predicts absolute OH concentrations vs. equivalence ratio but does not capture transient OH overshoot after the burst and overestimates the OH decay time. In CH4-air and H2-air, the model significantly underpredicts absolute OH number density but the rate of OH decay is predicted fairly accurately, especially in hydrogen. Identifying kinetic processes controlling the rate of OH generation / decay during and after the discharge requires additional kinetic sensitivity analysis. Reactions of radical species formation during energy transfer from electronically excited N2 molecules are likely to be among key processes affecting OH concentration balance.

1 Copyright, American Institute of Aeronautics and Astronautics. All rights reserved 2 Graduate Research Associate 3 Professor, Associate Fellow AIAA

43rd AIAA Plasmadynamics and Lasers Conference25 - 28 June 2012, New Orleans, Louisiana

AIAA 2012-3182

Copyright © 2012 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

2

1. Introduction

Nonequilibrium plasmas formed by nanosecond duration pulsed discharges have shown

significant effect on ignition delay reduction and flame stabilization in quiescent and flowing fuel-air mixtures [1-9]. Transient reduced field (E/N) in these discharges is of the order of 102-103 Td (1 Td=10-17 V·cm2), which enhances electronic excitation and dissociation of gas molecules, resulting in efficient generation of radical species. The radicals readily react with the fuel and oxidizer molecules, resulting in chain branching and generating heat, which can eventually lead to ignition. To analyze quantitatively the impact of radicals generated in the plasma on fuel oxidation and ignition, time-resolved measurements of gas temperature and radical species concentrations are required. In our previous work, various laser diagnostics method have been used for temporally and spatially resolved measurements in nanosecond discharges in fuel-air mixtures, including absolute NO and O atom concentrations by Laser Induced Fluorescence (LIF) and Two-Photon Absorption LIF (TALIF) [10, 11], rotational and N2 vibrational temperatures by Coherent Anti-Stokes Raman Spectroscopy (CARS) [12-14], rotational temperature and absolute OH concentration by LIF [15, 16].

In our previous measurements of ignition delay and OH concentration in H2-air mixtures, two discharge operation regimes have been used. In the first approach, nanosecond pulse discharge operating at a high pulse repetition rate was maintained through the ignition process [15, 17]. Kinetic modeling of the plasma chemical ignition process demonstrated that in this case ignition time had fairly weak dependence on the details of kinetic mechanism. This occurs due to continuous generation of radicals during the discharge, overlapping with ignition, which largely “masks” such dependence. In the second approach, the discharge burst was terminated shortly before the onset of ignition [16]. At these conditions, electronically excited molecules and atoms generated during the discharge decay rapidly, such that oxidation of fuel-air mixtures is controlled primarily by low-temperature reactions of species in their ground electronic states, with the pool of radicals (primarily H, O, and OH) generated during the discharge. In this case, ignition of H2-air mixtures, initially at T0=500 K, occurred when the temperature at the end of the discharge burst exceeded a threshold of approximately Tf ≈700 K, about 200 K lower than autoignition temperature at these conditions, predicted by kinetic modeling [17]. Based on ignition time, time-resolved temperature, and time-resolved absolute OH concentration measurements using OH LIF [15-17], insight into the effect of super-equilibrium concentrations of radical species generated in low-temperature hydrogen-air plasmas has been obtained.

However, quantitative study of plasma assisted ignition kinetics by a repetitive nanosecond pulse discharge could not be promptly extended to hydrocarbon fuels, due to non-uniform ignition caused even by moderate plasma nonuniformity, which generated a propagating flame front [16,17]. In the present work, this problem is circumvented by reducing the number of pulses in the discharge burst, as well as the equivalence ratio. This is done to study the mechanism of radical generation (with OH being the primary focus) in mildly preheated fuel-air mixtures excited by a repetitively pulsed nanosecond discharge, as well OH production and decay after the discharge, before ignition occurs. The objectives of the present work are: (i) to use both OH LIF thermometry and psec CARS for time-resolved rotational/translational temperature measurements of fuel-air mixtures after a nanosecond pulse discharge burst, and (ii) to measure absolute, time-resolved OH concentrations in H2-air and hydrocarbon-air mixtures excited by the discharge after it is turned off. These measurements are conducted for several

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

3

different fuels, over a wide range of equivalence ratios. The results are compared with kinetic modeling calculations, to help identify dominant processes controlling OH generation and decay in the discharge and in the afterglow, and to assess the accuracy of their rates. 2. Experimental

The experimental setup used in the present work, shown schematically in Fig. 1, is similar to the one used in our previous studies [16]. The discharge cell / plasma flow reactor consists of a 280 mm long, 22 mm x 10 mm rectangular cross section quartz channel with wall thickness of 1.75 mm. Two plane quartz windows are fused to the ends of the channel, providing optical access in the axial direction. A 60 mm long, right angle fused silica prism is placed along the channel to provide optical access from the side (see Fig. 1). The entire assembly is heated in a tube furnace (Thermcraft, Ltd., with 6 inch diameter, 12 inch long heated section), to improve plasma stability. A 1 m long quartz tube coil inlet (see Fig. 1) preheats the fuel-air flow to the furnace temperature, which was verified by thermocouple measurements. The furnace, mounted on rails, can be moved aside and replaced by an adiabatic Hencken burner used for absolute calibration of LIF diagnostics. Two ¼ inch diameter quartz-to-stainless-steel adaptors connect the reactor to the gas delivery system. Fuel and air flow rates through the reactor are controlled by MKS mass flow controllers. Fuel and air flows are premixed before entering the cell.

Two 14 mm x 60 mm rectangular plate copper electrodes, rounded at the edges, are placed on the top and bottom of the quartz channel, as shown in Fig. 1, and held in place by ceramic clamps. A 1/16-inch thick high-temperature perfluoroelastomer sheet (Kalrez, DuPont) is placed between each electrode and the channel wall, to reduce air gaps and prevent corona discharge outside the cell. The electrodes are connected to an FID GmbH FPG 60-100MC4 pulse generator (peak voltage up to 30 kV, pulse duration 5 nsec, repetition rate up to 100 kHz). In the present work, the pulser is operated in repetitive burst mode, producing bursts of 50 pulses at a pulse repetition rate of 10 kHz and burst repetition rate of 5 Hz.

A schematic diagram of the OH LIF setup is also shown in Fig. 1. Briefly, 532 nm output from an Nd:YAG laser (Continuum, Model Powerlite 8010) is used to pump a tunable dye laser (Laser Analytical Systems, Model LDL 20505) to produce output at 566 nm, which is then frequency doubled by a β Barium Borate (BBO) crystal. While the flash lamp of the YAG laser is operated at 10 Hz, Q-switching controlling the laser output is maintained at 5 Hz, to match the burst repetition rate of the pulser and to probe OH after each discharge burst. A half wave plate and thin film polarizer pair is used to control the UV beam energy for LIF operation in the linear regime (below 5 μJ/pulse). The UV beam energy is monitored using a photodiode placed after a beam splitter (see Fig. 1).

Absolute OH concentration and rotational/translational temperature are measured by calibrated OH LIF in lean CH4-, C2H4-, C3H8-, and H2-air mixtures, each at four different equivalence ratios, at flow velocity through the reactor of u~40 cm/sec. Low discharge burst repetition rate, 5 Hz, is chosen to keep the delay between the bursts longer than the residence time of the flow in the plasma, ~0.1 sec, such that the flow in the reactor would experience only one burst. Slow flow velocity also improves flow preheating in the inlet coil and reduces pressure drop across the reactor.

For OH concentration measurements, Q1(3) line in the OH A-X (1,0) band, centered at 282.241 nm, is used to maximize signal-to-noise ratio at temperatures below 600 K, expected

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

4

after the discharge burst. For OH thermometry, four additional lines in the (1,0) band system are used to generate a more accurate Boltzmann fit for rotational temperature, including P1(1) at 282.170 nm, Q1(2) at 282.067 nm, Q1(4) at 282.439 nm, and Q1(5) at 282.667 nm. The choice of the excitation lines, as well as LIF signal processing was discussed in detail in Ref. [16]. More recent LIF diagnostics modifications made since will be discussed in Section 4. LIF signal is collected from the center of the discharge, a region approximately 15 mm long, via a right angle prism and fused silica lenses, into a photomultiplier tube (PMT) (Hamamatsu, Model R106) in combination with a small monochromator (Horiba, Model H-10). The monochromator, centered at 310 nm, has a band pass of approximately 20 nm to incorporate the entire (1,1) and (0,0) OH fluorescence bands. The LIF signal is integrated over a 100 nsec gate and analyzed in real time during the experiment using a programmable digital oscilloscope with 1 GHz bandwidth (LeCroy, Model WAVEPRO 7100A). Relative OH concentrations are put on an absolute scale using calibration in an atmospheric pressure, near-adiabatic C2H4-air flat flame generated by a Hencken burner (Technologies for Research, Model RD5X5) [16]. During calibration, the burner was positioned at the same location as the discharge cell to ensure the same LIF signal collection constant.

Rotational/translational temperature and vibrational temperature of nitrogen in the discharge cell / plasma flow reactor have also been measured by picosecond CARS diagnostics described in detail in Ref. [14]. Briefly, in this system a broadband Stokes beam is generated by an in-house fabricated modeless dye laser, with energy efficiency as high as 10%. The dye laser is pumped by an Ekspla SL-333 Nd:YAG laser, generating output pulses ~150 psec duration, with energy output of up to 120 mJ/pulse at 532 nm. The Nd:YAG laser also generates the pump / probe beams for the CARS mixing. The dye laser output is centered near 604 nm, with a full width at half maximum of approximately 5-6 nm. In the present work, the Unstable-resonator Spatially Enhanced Detection (USED) CARS phase matching geometry has been employed, with longitudinal spatial resolution of ~5 mm. For this geometry, the 532 nm pump/probe beam is enlarged and the center portion is removed, creating an annulus. This beam is then combined coaxially with the Stokes beam by a dichroic mirror and focused into the test section. After re-collimation, dichroic mirrors isolate the CARS signal beam which is then directed into a 0.75 m grating spectrometer. At the spectrometer exit plane a 2.3x relay lens magnification system is used to image the dispersed CARS signal onto a CCD (Andor Newton EM-CCD) camera. The combination of the 2.3x magnification and 3600 lpmm grating results in a spectral resolution of ~0.4 cm-1, which is sufficient to partially resolve the rotational structure in the Q-branch spectra of nitrogen. Details of data reduction, as well as rotational and vibrational temperature inference from the CARS spectra are discussed in Ref. [14]. The entire picosecond CARS system is placed on a custom built cart, allowing the entire setup to be easily transported between experimental facilities. 3. Plasma Chemistry Model and Nanosecond Pulse Discharge Model

To obtain insight into kinetics of plasma/chemical fuel oxidation and ignition, we use a kinetic model developed in our previous work [10-13, 15-17]. Briefly, the model incorporates nonequilibrium air plasma chemistry [18], expanded to include hydrocarbons and hydrogen dissociation processes in the plasma [11, 20], GRI Mech 3.0 hydrocarbon chemistry mechanism [19], and hydrogen-oxygen chemistry model developed by Popov [20]. In the present work, the

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

5

plasma chemical reaction list was expanded to incorporate propane dissociation reactions by electronically excited nitrogen molecules [21]. The list of air plasma chemistry processes and hydrogen-oxygen chemical reactions incorporated in the present model is given in Ref. [11] and Ref. [20], respectively. The dominant radical species generation processes in hydrocarbon-air and H2-air plasmas are listed in Table 1. The species concentration equations are coupled with the two-term expansion Boltzmann equation for plasma electrons. The model incorporates energy equation for the temperature on the discharge centerline [17], with heat transfer to the walls being the dominant energy loss mechanism, as well as quasi-one-dimensional flow equations. The model is validated using TALIF O atom concentration measurements [10, 11], rotational CARS temperature measurements [13,14], OH LIF and ignition time measurements [15-17] in nanosecond discharges in air, methane-air, ethylene-air, and hydrogen-air.

Key parameters controlling plasma chemistry in the nanosecond pulse discharge include the reduced electric field, E/N, and coupled pulse energy. Energy coupled to the plasma using two different nanosecond pulse generators, including the FID pulser used in the present work, was measured in air in a low-temperature cell outside of the furnace, over a wide range of pressures [22]. It has been shown that at P=10-250 torr, coupled pulse energy is nearly independent on pulse repetition rate (at ν=1-40 kHz), remains nearly constant during the pulse burst (up to ~100 pulses), and is proportional to discharge pressure, i.e. energy coupled per molecule remains constant. These experimental results are in good agreement with the analytic model of energy coupling in a nanosecond pulse discharge [23], which incorporates key effects of pulsed breakdown and sheath development on nanosecond time scale. The model predicts pulse energy coupled to the plasma vs. pulse voltage waveform, discharge geometry, pressure, and temperature. In the present work, reduced electric field in the plasma, E/N, is approximated as a Gaussian pulse with the same time constant as the experimental voltage waveform, with peak value corresponding to breakdown value predicted by the nanosecond discharge model [23].

Table 1. Dominant primary radical species generation processes in the plasma [11,12,17,21].

Process Rate, cm3/s

A1 N2 + e- = N2(A3Σ, B3Π, C3Π, a'1Σ) + e- σ1

A2 N2 + e- = N(4S) + N(4S) + e- σ A3 O2 + e- = O(3P) + O(3P,1D) + e- σ A4 N2(C

3Π) + O2 = N2 (B3Π ) + O2 3.0·10-10

A5 N2(a'1Σ) + O2 = N2 (B3Π) + O2 2.8·10-11

A6 N2(B3Π) + O2 = N2 (A

3Σ) + O2 3.0·10-10 A7 N2(A

3Σ) + O2 = N2 + O + O 2.5·10-12 H1 H2 + e- = H + H + e- σ2

H2 N2(a'1Σ) + H2 = N2 + H + H 2.6·10-11 H2 N2(B

3Π) + H2 = N2(A3Σ) + H2 2.5·10-11

H4 N2(A3Σ) + H2 = N2 + H + H 4.4·10-10 exp(–3170/T)

H5 O(1D) + H2 = H + OH 1.1·10-10 M1 CH4 + e- = CH3 + H + e- σ M2 N2(A

3Σ) + CH4 = N2 + CH3 + H 1.2·10-10 exp(–3500/T) M3 N2(B

3Π) + CH4 = N2 + CH3 + H 3.0·10-10

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

6

M4 N2(C3Π) + CH4 = N2 + CH3 + H 5.0·10-10

M5 N2(a'1Σ) + CH4 = N2 + CH3 + H 3.0·10-10 E1 C2H4 + e- = products3 + e- σ E2 N2(A

3Σ) + C2H4 = N2 + C2H3 + H 9.7·10-11 E3 N2(B

3Π) + C2H4 = N2 + C2H3 + H 3.0·10-10 E4 N2(C

3Π) + C2H4 = N2 + C2H3 + H 3.0·10-10 E5 N2(a'1Σ) + C2H4 = N2 + C2H3 + H 4.0·10-10

P1 N2(A

3Σ) + C3H8 = N2 + C3H6 + H2 1.2·10-12 P2 N2(B

3Π) + C3H8 = N2 + C3H6 + H2 3.0·10-10 P3 N2(C

3Π) + C3H8 = N2 + C3H6 + H2 3.0·10-10 P4 N2(a'1Σ) + C3H8 = N2 + C3H6 + H2 4.0·10-10

1 Calculated by the Boltzmann solver from the experimental cross sections 2 Sum of electronic excitation cross sections (b3Σ, b1Σ, c3Π, a3Σ, c1Π, and d3Π) 3 Three dissociation channels into C2H3 + H, C2H2 + H2, and C2H2 + H + H.

4. OH LIF and LIF Thermometry 4.1 Basics

In the present work, OH is excited by pumping a transition in the A2Σ+ ← X2Π (v'=1, v"=0) band. Operating in linear excitation regime, fluorescence signal, Sf, from the collection volume Vc can be expressed as follows [24]:

JBOHcf EBfnVS 12 , (1)

where η represents the overall efficiency of the optical setup in converting fluorescence photons into photoelectrons incident on the detector, nOH is OH number density, fB is the Boltzmann factor for OH molecules in the absorbing state, B12 is the Einstein absorption coefficient, and E is the laser energy fluence (energy per unit cross sectional area of the beam). The fluorescence quantum efficiency, ϕJ, can be derived under the assumption that the system reaches steady state during the laser pulse, by taking into account spontaneous emission (A1 and A0), collisional quenching (Q1 and Q0), and vibrational energy transfer (VET) from v'=1 to v'=0 (V10) :

1110

001001 )/(

QAV

QAVAAJ

, (2)

In Eq. (2), signal collection efficiency for (0,0) and (1,1) bands is assumed to be the same. The quenching rates in the plasma reactor are determined directly from fitting the fluorescence signal decay, and the VET rate at different conditions is estimated as a fraction of the corresponding quenching rate, following the relations given by Paul [25]. This estimate introduces up to ±25% uncertainty in the inferred OH number density. In the calibration flame, on the other hand,

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

7

quenching rates are calculated based on the cross sections and temperature scaling relations given by Tamura et al. [26], and VET rate coefficients are taken from Paul [25].

By measuring LIF signal from several pump transitions, rotational temperature can be obtained using a linear fit to natural log of normalized fluorescence signal plotted vs. rotational energy (Boltzmann plot), which can be obtained from Eq. (1),

kT

BVnη

kT

E

)J(φB

/EScOHJ

J

f ln12

ln12

, (3)

where B is the rotational constant of ground vibrational state and EJ is the rotational energy, taken from Dieke and Crosswhite [27].

To obtain an absolute scale for OH number density, the near-adiabatic flat flame generated by the Hencken burner is used. Having verified that the measurement location in the flame is near adiabatic, equilibrium OH number density in the flame can be calculated from the temperature measured by OH LIF thermometry, as described above, using an equilibrium thermochemical code STANJAN [28]. Then factor η in Eq. (1) can be readily inferred from the fluorescence signal, Sf, obtained from the flame. 4.2 Signal Collection and Processing

As described in greater detail in our previous work [16], the selection of excitation

transition is governed by maximizing signal-to-noise and temperature sensitivity. In the present work, since the discharge burst only contains 50 pulses, temperature after the burst is estimated to be near 600 K (~100 K increment from the initial temperature of T0=500 K), based on measurements and kinetic modeling in Ref. [16]. In this temperature range, Jʺ=3 is the most populated rotational level in the ground electronic state OH. Q1(3), which has the highest absorption coefficient among all pumping transitions from Jʺ=3 in the OH A-X (1,0) band, is chosen for measurements of relative OH concentrations in the plasma reactor. For temperature measurements, four additional excitation transitions are added, P1(1), Q1(2), Q1(4), and Q1(5), to reduce the rotational temperature inference uncertainty at low temperatures, compared to two-line OH LIF thermometry used in our previous work [16]. The energy separation between P1(1) and Q1(5) transitions is 782 K, which provides sufficient accuracy in temperature inference. Due to both low temperatures and low OH concentrations at the present conditions, both of these transitions are relatively weak. For this reason, Q1(2) and Q1(4) transitions are used to reduce systematic error in the Boltzmann fit for rotational temperature. For calibration in the C2H4-air flame, Q1(5) and Q1(14) transitions are used as a line pair to determine the temperature, as suggested by Kostka et al. [29], since the equilibrium temperature in the flame is close to 2400 K.

For accurate quantitative measurements, linear operation regime and absence of detector saturation during excitation and LIF signal collection must be verified. The maximum laser pulse energy to maintain linear regime is determined to be ~5 μJ. LIF signal is normalized by the laser pulse energy, monitored simultaneously by a photodiode. For signal collection from the calibration flame, a neutral density filter (Thorlabs ND10A, D=1.0, with signal attenuation by approximately a factor of 10) is placed in front of the monochromator entrance slit to avoid saturating the PMT. LIF excitation spectra are obtained by scanning the dye laser across each selected excitation transition. The tuning speed of the laser system is set to 1.25 pm (in the UV)

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

8

scanning step per 60 laser shots. Comparing to using single laser shots per smaller tuning increment, this method allowed taking the median of each group of data to represent statistically most probable signal at each scanning step, such that random noise, e.g. electromagnetic interference (EMI) from the pulser can be largely eliminated. Q-branch transitions usually have closely spaced satellite lines. In such cases, spectral overlap correction is needed to extract the actual lineshape of the transition in interest. This is done by least-square fitting of the experimental spectra using a Voigt profile, as discussed in detail in Ref. [16].

5. Results and Discussion 5.1 Discharge Uniformity

Plasma uniformity is critical for isolating nonequilibrium chemistry effects from thermal

heating occurring in filaments and hot spots, which is difficult to quantify. The approach used for sustaining diffuse, nearly zero-dimensional nanosecond pulse discharges in plane-to-plane geometry was discussed in detail in our previous work [16,17]. Before quantitative OH LIF measurements, plasma uniformity is verified qualitatively by taking ICCD images of an individual discharge pulse. Time-resolved, broadband, single-shot images taken during a high voltage nanosecond pulse discharge in air at T0=500 K and P=50 torr, using a picosecond gate PI-MAX ICCD camera and a UV lens (Objectif UV 100F/2.8), are shown in Figure 2. The camera gate is set at 490 psec, and is synchronized with the high voltage pulse generator to open during the last pulse in a 10-pulse discharge burst, with t=0 representing the beginning of the pulse. The timing of the gate is indicated on the top of each image. The images capture primarily nitrogen second positive band emission, N2(C

3Π→B3Π). From the images in Fig. 2, it is evident that the plasma is generated in the entire volume between the electrodes and remains nearly uniform until it decays completely. Total emission intensity (a sum of intensity counts over all pixels in each ICCD image) is plotted vs. time in Fig. 3, demonstrating that excited radiating species (mainly N2(C

3Π) state) decay very rapidly after each discharge pulse, on the time scale of a few nanoseconds, as expected.

As discussed in our previous work [17], addition of hydrogen or hydrocarbons to the air plasma makes the plasma less uniform. This effect is especially well pronounced in hydrocarbon-air mixtures at high pressures (P~100 torr and above), although mild preheating to T0=400-5000 reduces the nonuniformity considerably. However, nanosecond pulse plasma assisted ignition in CH4-air and C2H4-air mixtures still exhibits a well-defined propagating flame front, such that near-0-D ignition could not be achieved. In the present work, this problem is circumvented by employing short discharge bursts (50 pulses at the pulse repetition rate of ν=10 kHz, far less than needed for ignition) and operating in fuel-lean mixtures, with the main emphasis of measuring OH number density in the plasma prior to ignition. Plasma uniformity at these conditions has been verified for three representative fuel-air mixtures (CH4-, C2H4-, and H2-air mixtures at ϕ=0.3), at three different pressures (100, 300 and 500 torr), as shown in Figure 4. As discussed in Section 2, LIF signal is collected from the region in the center of the plasma approximately 1.5 cm long, about ¼ of the length of the discharge. In the absence of a propagating flame, this excludes the effect of discharge non-uniformities that may be formed near electrode edges at high pressures [17] on plasma chemistry in the center of the discharge.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

9

5.2 OH LIF Thermometry and Comparison to CARS Temperature Measurements

A typical excitation scan spectrum of the five OH A-X (1,0) transitions used in the present work is shown in Fig. 5. The spectrum is taken in an H2-air mixture at ϕ=0.12, T0=500 K, and P=100 torr, excited by a ν=10 kHz, 50 pulse discharge burst, 2 μsec after the last pulse. This is close to minimum time delay for which LIF signal is not affected by EMI noise from the discharge. In Fig. 5, each experimental data point represents a median of 60 laser shots taken at each scanning step, as described in Section 4.2. For the Voigt lineshape fit, relative distances between individual line centers, as well as line intensity ratios (based on line strengths), are determined according to the values reported by Luque and Crosley [30]. This is critical for isolating overlapping spectral lines, as occurs for all Q-branch transitions shown in Fig 5. After isolating the major transitions and integrating them over the same wavelength range, symmetric to the centerline wavelength, rotational temperature is inferred from the Boltzmann plot, such as shown in Fig. 6. Each of the excitation scans shown in Fig. 5 is repeated three times to improve accuracy. Satellite lines (Q21 transitions) are excluded from the fit due to large uncertainties in the Voigt fit, controlled by signal strength. At some conditions, including the one shown in Fig. 6, P1(1) and Q1(2) transitions somewhat deviate from the straight line formed by Q1(3), Q1(4) and Q1(5) transitions. If P1(1) and Q1(2) are removed from Fig. 6, the best fit rotational temperature would be Trot=557±25 K, which differs by 30 K from the best fit temperature obtained using a 5 rotational line fit (Trot=587±27 K). This relatively small discrepancy could be due to P1(1) line being relatively weak, as well as a strong overlap between Q1(2) with Q21(2) lines, which may introduce uncertainty in data processing and Voigt line fitting. Also note that the five pump transitions used are taken not simultaneously but during different runs, which is likely to be a major source of systematic error. Nonetheless, this difference in best fit rotational temperature values is close to the statistical error for the Boltzmann fit. In most cases, low J transitions (J=1 and J=2) are needed for better rotational temperature sensitivity, as discussed in Section 4.2.

Figure 7 compares time-resolved rotational temperature measured by OH LIF in a C2H4-air mixture at ϕ=0.09, T0=500 K, P=100 torr, after a ν=10 kHz, 50-pulse discharge burst with rotational temperature measured by psec CARS at the same conditions. Temperatures measured by these two techniques are close to each other, showing that the temperature in the discharge increases by ∆T=50-80 K (from the initial value of T0=500 K) and remains nearly constant for up to 1 msec after the burst. These results are consistent with temperature rise in the discharge predicted by the kinetic model at these conditions, also shown in Fig. 7, ∆T~100 K. A relatively modest temperature rise in the discharge, as well as its weak dependence on the equivalence ratio are also observed in other fuel-air mixtures at the same initial temperature, pressure, and discharge burst duration, as shown in Fig. 8. The temperature rise during the burst, averaged over different fuels and equivalence ratios, is Tf≈570 K. Again, OH LIF temperature measurements plotted in Fig. 8 are consistent with CARS results and fuel-air plasma chemistry model predictions. Based on these results, in the subsequent experiments, temperature was measured only immediately after the discharge burst (with 2 μsec delay), and the result was used to infer absolute, time-resolved OH number densities in reacting fuel-air mixtures after the discharge from LIF data.

Figure 9, taken from our recent work [16], plots temperature and OH concentration measured by OH LIF after a discharge burst (ν=10 kHz, 115 pulses) in a plasma-ignited ϕ=0.4

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

10

H2-air mixture at T0=500 K and P=80 torr. In Ref. [16], these results were compared with kinetic model predictions, also plotted in Fig. 9, showing good agreement. In Fig. 9, the results of Ref. [16] are overlapped with psec CARS temperature measurements taken in the present work at very similar conditions (H2-air, ϕ=0.4, T0=500 K, P=92 torr, ν=10 kHz, 120 pulses). It can be seen that temperature measured by CARS is in excellent agreement with the model predictions from Ref. [16], throughout the entire ignition process, except at short time delays after the burst (~1 msec), where the model overpredicts the temperature by about ∆T~50 K. The most likely reason for the difference between LIF and CARS temperatures during (but not prior to) ignition is lack of burst-to-burst ignition delay time reproducibility over long time periods. The advantage of use of broadband CARS for temperature measurements during at these conditions, compared to OH LIF, is apparent. Indeed, OH LIF measurements take significantly longer time compared to CARS, and thus are more susceptible to long-term variation in ignition delay.

Figure 10(left) plots temperatures measured by CARS after the discharge burst in H2-air at ϕ=0.4, T0=500 K, P=92 torr, and ν=10 kHz, for three different burst durations (50-120 pulses). Also plotted in Fig. 10(left) are the model predictions at these conditions, which are in very good agreement with CARS data up to fairly long delay times after the burst, 20 msec, although the model again systematically overpredicts that temperature at the end of the discharge burst, by ∆T~50 K. From the results shown in Fig. 10(left), it is apparent that threshold temperature at the end of the discharge burst, above which ignition occurs, is Tf ≈700 K, approximately 200 K below equilibrium autoignition temperature [17]. This result confirms our previous temperature measurements by N2 emission spectroscopy [17], although the results of Ref. [17] were not spatially resolved and had a more significant uncertainty. Below threshold, ignition does not occur and temperature gradually decays to pre-excitation baseline value of T0=500 K (see Fig. 10(left)). Figure 10(right) demonstrates that at these conditions, OH generated during the discharge burst (shown by dashed lines) also gradually decays and does not exhibit an overshoot typical for ignition (see Figs. 9 and 10(right)).

Although CARS measurements demonstrated that significant vibrational excitation of nitrogen occurs in a ν=10 kHz nanosecond pulse burst discharge sustained in the present cell in room temperature air [14], N2 vibrational temperature measured in preheated air (T0=500 K) was quite low, Tv(N2) ≈850 K for 50 pulses and Tv(N2) ≈1050 K for 100 pulses (see Fig. 11). In fuel-air mixtures, N2 vibrational temperature after a 50-pulse burst was even lower, near the detection limit, Tv(N2) ≈ 600-700 K. This demonstrates that at the present experimental conditions, vibrational disequilibrium of excited fuel-air mixtures is not a significant factor. Since estimated characteristic decay time of excited electronic states of nitrogen at the present conditions is quite short, τ~1 μsec for N2(A

3Σ), and even shorter for N2(B3Π), N2(C

3Π) and N2(a'1Σ) states, chemical kinetics of fuel-air mixtures on longer time scales is dominated by conventional chemical reactions among radical species generated in the discharge. This lends the present experiment for studies of radical reaction kinetics at relatively low temperatures, ranging from room temperature to ignition temperature.

5.3 OH Number Density after the Discharge.

Temperature measured in a ϕ=0.99 C2H4-air flat flame in the calibration Hencken burner,

determined using Q1(5) and Q1(14) line pair is T=2338±58 K, very close to the theoretical

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

11

equilibrium temperature of T=2366 K. As discussed in Section 4.1, equilibrium OH concentration in the flame calculated by STANJAN code was used to obtain the signal collection constant in Eq. (1), to put relative OH signal in the plasma reactor on an absolute scale.

Figure 12 plots absolute, time-resolved OH number densities measured in lean C2H4-air mixtures at T0=500 K and P=100 torr vs. time delay after a ν=10 kHz, 50-pulse discharge burst, at different equivalence ratios (ϕ=0.05-0.36). Rotational/ translational temperatures measured at the end of the burst (time delay 2 μsec after the burst) at these conditions are also shown in Fig. 12 (circled). It can be seen that in the entire range of equivalence ratios, OH number density after the burst first increases, by up to 60%, before eventually decaying. Reducing the equivalence ratio increases peak transient OH number density after the burst by approximately a factor of 2, from ≈4·1013 cm-3 at ϕ=0.36 to ≈8·1013 cm-3 at ϕ=0.05 (see Fig. 12). Finally, OH decay time is fairly short and increases considerably in lean mixtures, from ~30 μsec at ϕ=0.36 to ~100 μsec at ϕ=0.05. This suggests that no significant OH accumulation occurs during the discharge burst, with time interval between the pulses at the present conditions of 100 μsec (ν=10 kHz). As discussed above, temperature at the end of the discharge burst is not very sensitive to the equivalence ratio and delay time after the discharge (see Figs. 7,8), and remains within T=550-600 K (see Fig. 12).

Figure 12 also compares kinetic model predictions with the experimental results in C2H4-air. As can be seen, the model predicts peak absolute OH number densities after the burst rather well and reproduces the trend of peak OH increasing in lean mixtures, as the equivalence ratio is reduced. However, transient OH rise after the burst, detected in the entire range of equivalence ratios tested, is completely missing from the model predictions. Finally, the model tends to overpredict OH decay time, up to a factor of 2-3.

The results in CH4-air mixtures are shown in Fig. 13. In this case, considerably more OH is generated for the same discharge burst parameters (ν=10 kHz, 50 pulses), up to (1.5-1.6)·1014 cm-3 at ϕ=0.03-0.06. Similar to C2H4-air, OH number density after the burst increases as the equivalence ratio is reduced, approximately doubling between ϕ=0.24 and ϕ=0.03 (see Fig. 13). However, unlike in ethylene-air mixtures, no significant transient OH increase is observed, except for a modest rise at ϕ=0.24. The OH decay time in CH4-air is considerably longer than in C2H4-air, ranging from ~200 μsec at ϕ=0.24 to ~1 msec at ϕ=0.03. This suggests significant OH accumulation during the discharge burst at the pulse repetition rate of ν=10 kHz. OH number densities after the discharge burst predicted by the kinetic model in CH4-air are considerably lower than measured in the experiment, by a factor of 2.5-3.5 (see Fig. 13, additional y-axis on the right is used for the model predictions). The model does predicts OH increase as the equivalence ratio is reduced, at ϕ=0.24-0.06, although on a more moderate scale than in the experiment (about 40% vs. a factor of 2). The OH decay time predicted by the model is in fairly good agreement with the model, especially in very lean mixtures (ϕ=0.03-0.06). The results in propane-air are very similar to ethylene-air (compare Figs. 12 and 14). In both fuel-air mixtures, OH increases as the equivalence ratio is reduced, exhibiting transient rise with time delay after the burst, and OH decay rates are close to each other. Absolute OH number densities in C3H8-air are approximately a factor of 2 lower compared to C2H4-air. Similar to C2H4-air, absolute OH number densities after the burst predicted by the kinetic model vs. the equivalence ratio (increasing in lean mixtures) are in agreement with the experiment (see Fig. 14). Again, transient OH rise after the burst, consistently observed in the experiments, is not reproduced by the model, and OH decay time is overpredicted by a factor of 2-3.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

12

Finally, the results obtained in lean H2-air mixtures are shown in Fig. 15. In this case, peak OH number density after the burst, (1.3-1.5)·1014 cm-3, is not very sensitive to the equivalence ratio. Transient OH rise, observed in ethylene-air and propane-air mixtures, is not detected. In fact, the results, including peak OH concentrations and OH decay time, are fairly similar to those obtained in methane-air mixtures at ϕ=0.03-0.12 (see Fig. 13). Fairly long OH decay time, ~0.5-1.0 msec shows that it accumulates during the burst, consistent with our previous results in H2-air taken at discharge pulse repetition rate of ν=40 kHz [15]. Again, similar to CH4-air, the kinetic model underpredicts absolute OH number densities after the burst by approximately a factor of 3, although both weak OH dependence on the equivalence ratio and OH decay rates are matched very well by the model (see Fig. 15).

Summarizing the results of kinetic modeling comparison with OH number density measurements, it can be seen that

(i) in C2H4-air and C3H8-air, the model predicts absolute OH amounts rather well but does not capture transient OH rise after the burst and overestimates the OH decay time, and

(ii) in CH4-air and H2-air, the model underpredicts absolute OH number density after the burst by a factor of ~3, but the rate of OH decay is predicted fairly accurately.

At this time, the reason for the difference between the model predictions and the experimental results has not been identified. This difference may well be due to uncertainties in the rates of dominant radical species formation, such as H, O, and CxHy radicals, during energy transfer from electronically excited nitrogen molecules at elevated temperatures, as well as in composition of reaction products. Compared to these processes, the effect of radical species generation by direct electron impact of fuel species is expected to be relatively weak, especially in lean mixtures of heavy hydrocarbon fuels such as propane, when the fuel mole fraction is very low. Identifying specific kinetic processes controlling the net rate of OH generation / decay during and after the burst requires additional kinetic sensitivity analysis.

6. Summary

In the present work, OH Laser Induced Fluorescence (LIF) diagnostics calibrated in an atmospheric pressure, near-adiabatic C2H4-air flame is used for time-resolved temperature and time-resolved, absolute OH number density measurements in fuel-air mixtures in a nanosecond pulse discharge cell / plasma flow reactor. Rotational temperature is inferred from OH LIF excitation spectra using 5 different excitation transitions. Rotational temperature and N2 vibrational temperature in the reactor have also been measured using broadband, picosecond Coherent Anti-Stokes Raman Spectroscopy (CARS) diagnostics, with spatial resolution of approximately 5 mm. The premixed fuel-air flow in the reactor is excited by a repetitive nanosecond pulse discharge in a plane-to-plane geometry, operated at a high pulse repetition rate, ν=10 kHz. The experiments are conducted in lean, slowly flowing H2-air, CH4-air, C2H4-air, and C3H8-air mixtures preheated to T0=500 K in a tube furnace, at a pressure of P=100 torr. In most of the present measurements, the discharge burst duration was limited to 50 pulses, insufficient to produce plasma-assisted ignition at the present conditions. The discharge uniformity in air and fuel-air flows has been verified using sub-nsec gate ICCD camera images, demonstrating that a diffuse, volume filling, uniform plasma is sustained in the entire range of operating conditions, at pressures of up to P=500 torr.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

13

The results demonstrate that rotational temperature inferred from OH LIF spectra and partially rotationally resolved CARS spectra are close to each other, providing additional validation to OH LIF thermometry diagnostics used in the present paper, as well as in our previous work [16]. Temperatures at the end of the discharge burst, measured by these two techniques, are in the range of T=550-580 K, indicating a modest temperature rise in the discharge burst of ∆T~50-80 K, consistent with kinetic modeling predictions. The temperature after the burst remains essentially unchanged, for up to 1 msec delay. Time-resolved temperature during plasma-assisted ignition of a ϕ=0.4 H2-air mixture, inferred from the present psec CARS measurements, is in very good agreement with kinetic modeling predictions published in our previous paper [16]. N2 vibrational temperatures inferred from CARS measurements are quite low, Tv≈850 K in air and Tv≈600-700 K in fuel-air after a 50-pulse burst, demonstrating that vibrational disequilibrium is not a significant factor at the present conditions.

Time-resolved, absolute OH number density measurements after the discharge burst demonstrate that OH concentration generated in C2H4-air and C3H8-air mixtures is highest in very lean mixtures, increasing by about a factor of 2 as the equivalence ratio is reduced by a factor of 8. In these two fuels, transient OH overshoot after the discharge burst, by up to a factor of 2, is detected. No such overshoot is observed in CH4-air and in H2-air mixtures. In CH4-air, OH concentration also increases as the equivalence ratio is reduced, but in H2-air OH concentration after the burst is nearly independent of the equivalence ratio. OH decay rate after the burst in C2H4-air and C3H8-air occurs on the time scale of ~20-100 μsec, suggesting little accumulation during the burst of pulses repeated at 10 kHz. In CH4-air and H2-air, OH decays within ~0.1-1.0 msec and ~0.5-1.0 msec, respectively, suggesting OH accumulation during the burst. In C2H4-air and C3H8-air, the kinetic model predicts absolute OH concentrations after the burst but does not capture transient OH overshoot and overestimates the OH decay time. In CH4-air and H2-air, the model significantly underpredicts absolute OH number density after the burst, but the rate of OH decay is predicted fairly accurately, especially in hydrogen. Identifying specific kinetic processes controlling the net rate of OH generation / decay during and after the burst requires additional kinetic sensitivity analysis. Reactions of primarily radical species formation, such as H, O, and CxHy radicals, during energy transfer from electronically excited nitrogen molecules are likely to be among key processes affecting OH concentration balance. 7. Acknowledgements

This work is supported by the U.S. Air Force Office of Scientific Research MURI “Fundamental Aspects of Plasma Assisted Combustion” Chiping Li ─ Technical Monitor. 8. References 1. F. Wang, J.B. Liu, J. Sinibaldi, C. Brophy, A. Kuthi, C. Jiang, P. Ronney, and M.A. Gundersen,

“Transient Plasma Ignition of Quiescent and Flowing Air/Fuel Mixtures”, IEEE Transactions on Plasma Science, vol. 33, 2005, pp. 844-849

2. S.V. Pancheshnyi, D.A. Lacoste, A. Bourdon, and C.O. Laux, “Ignition of Propane-Air Mixtures by a Repetitively Pulsed Nanosecond Discharge”, IEEE Transactions on Plasma Science, vol. 34, 2006, pp. 2478-2487

3. A. Bao, Yu.G. Utkin, S. Keshav, G. Lou, and I.V. Adamovich, “Ignition of Ethylene-Air and

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

14

Methane-Air Flows by Low-Temperature Repetitively Pulsed Nanosecond Discharge Plasma”, IEEE Transactions on Plasma Science, vol. 35, pp. 1628-1638, 2007

4. A. Dutta, Z. Yin, and I.V. Adamovich, “Cavity Ignition and Flameholding of Ethylene-Air and Hydrogen-Air Flows by a Repetitively Pulsed Nanosecond Discharge”, Combustion and Flame, vol. 158, 2011, pp. 1564-1576

5. H. Do, S.-K. Kim, M.G. Mungal and M.A. Cappelli, “Plasma Assisted Flame Ignition of Supersonic Flows over a Flat Wall ”, Combustion and Flame, vol. 157, 2010, pp. 2298-2305

6. S.A. Bozhenkov, S.M. Starikovskaia, and A.Yu. Starikovskii. "Nanosecond Gas Discharge Ignition of H2 and CH4 Containing Mixtures", Combustion and Flame, vol. 133, 2003, pp. 133-146

7. S.M. Starikovskaia, E.N. Kukaev, A.Yu. Kuksin, M.M. Nudnova, and A.Yu. Starikovskii, “Analysis of the Spatial Uniformity of the Combustion of a Gaseous Mixture Initiated by a Nanosecond Discharge”, Combustion and Flame, vol. 139, 2004, pp. 177–187

8. G. Pilla, D. Galley, D.A. Lacoste, F. Lacas, D. Veynante, and C.O. Laux, “Stabilization of a Turbulent Premixed Flame Using a Nanosecond Repetitively Pulsed Plasma”, IEEE Transactions on Plasma Science, vol. 34, 2006, pp. 2471-2477

9. W. Kim, H. Do, M.G. Mungal, and M.A. Cappelli, “Plasma-Discharge Stabilization of Jet Diffusion Flames”, IEEE Transactions on Plasma Science, vol. 34, 2006, pp. 2545-2551

10. M. Uddi, N. Jiang, E. Mintusov, I.V. Adamovich, and W.R. Lempert, “Atomic Oxygen Measurements in Air and Air/Fuel Nanosecond Pulse Discharges by Two Photon Laser Induced Fluorescence”, Proceedings of the Combustion Institute, vol. 32, 2009, pp. 929-936

11. M. Uddi, N. Jiang, I. V. Adamovich, and W. R. Lempert, “Nitric Oxide Density Measurements in Air and Air/Fuel Nanosecond Pulse Discharges by Laser Induced Fluorescence”, Journal of Physics D: Applied Physics, vol. 42, 2009, p. 075205

12. Y. Zuzeek, I. Choi, M. Uddi, I.V. Adamovich, and W.R. Lempert, “Pure Rotational CARS Thermometry Studies of Low Temperature Oxidation Kinetics in Air and Ethene-Air Nanosecond Pulse Discharge Plasmas”, Journal of Physics D: Applied Physics, vol. 43, 2010, p. 124001

13. Y. Zuzeek, S. Bowman, I. Choi, I.V. Adamovich, and W.R. Lempert, “Pure Rotational CARS Studies of Thermal Energy Release and Ignition in Nanosecond Repetitively Pulsed Hydrogen-Air Plasmas”, Proceedings of the Combustion Institute, vol. 33, Issue 2, 2011, pp. 3225-3232

14. A. Montello, Z. Yin, D. Burnette, I.V. Adamovich, and W.R. Lempert, “Picosecond CARS Measurements of Nitrogen Vibrational Loading and Rotational/Translational Temperature In Non-Equilibrium Discharges”, AIAA Paper 2012-, 43rd AIAA Plasmadynamics and Laser Conference, 25-28 June 2012, New Orleans, LA

15. I. Choi, Z. Yin, I.V. Adamovich, and W.R. Lempert, “Hydroxyl Radical Kinetics in Repetitively Pulsed Hydrogen-Air Nanosecond Plasmas”, IEEE Transactions on Plasma Science, vol. 39, 2011, pp. 3288-3299

16. Z. Yin, I.V. Adamovich, and W.R. Lempert, “OH Radical and Temperature Measurements During Ignition of H2-Air Mixtures Excited by a Repetitively Pulsed Nanosecond Discharge”, AIAA Paper 2012-0826, 50th AIAA Aerospace Sciences Meeting, 09-12 January 2012, Nashville, TN

17. Z. Yin, K. Takashima, and I.V. Adamovich, "Ignition Time Measurements in Repetitive Nanosecond Pulse Hydrogen-Air Plasmas at Elevated Initial Temperatures", IEEE Transactions on Plasma Science, vol. 39, 2011, pp. 3269-3282

18. I.A. Kossyi, A. Yu. Kostinsky, A. A. Matveyev, and V.P. Silakov, “Kinetic Scheme of the Nonequilibrium Discharge in Nitrogen-Oxygen Mixtures, Plasma Sources Science andTechnology, vol. 1, 1992, pp. 207-220

19. http://www.me.berkeley.edu/gri_mech/version30/text30.html, GRI-Mech 3.0

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

15

20. N.A. Popov, “Effect of a Pulsed High-Current Discharge on Hydrogen–Air Mixtures”, Plasma Physics Reports, 2008, vol. 34, No. 5, pp. 376–391

21. N. Moreau, S. Pasquiers, N. Blin-Simiand, L. Magne, F. Jorand, C. Postel and J.-R. Vacher, “Propane dissociation in a non-thermal high-pressure nitrogen plasma”, J. Phys. D: Appl. Phys., vol. 43, 2010, p. 285201

22. Takashima, K., Yin, Z., and Adamovich, I.V., “Measurements and Kinetic Modeling Analysis of Energy Coupling in Nanosecond Pulse Dielectric Barrier Discharge, AIAA Paper 2012-, 43rd AIAA Plasmadynamics and Lasers Conference, New Orleans, LA, June 25-28, 2012

23. I.V. Adamovich, M. Nishihara, I. Choi, M. Uddi, and W.R. Lempert, “Energy Coupling to the Plasma in Repetitive Nanosecond Pulse Discharges”, Physics of Plasmas, vol. 16, 2009, p. 113505

24. A.C. Eckbreth, “Laser Diagnostics for Combustion Temperature and Species”, Taylor and Francis, 1996

25. P.H. Paul, “Vibrational Energy Transfer and Quenching of OH A2Σ+(v’=1) Measured at High Temperature in a Shock Tube”, Journal of Physical Chemistry, vol. 99, 1995, pp. 8472-8476

26. M. Tamura, P.A. Berg, J.E. Harrington, J. Luque, J.B. Jeffries, G.P. Simith, and D.R. Crosley, “Collisional Quenching of CH(A), OH(A), and NO(A) in Low Pressure Hydrocarbon Flames”, Combustion and Flame, vol. 114, 1998, pp. 502-514

27. G.H. Dieke and H.M. Crosswhite, “The Ultraviolet Bands of OH”, Journal of Quantitative Spectroscopy and Radiative Transfer, vol. 2, pp. 97-199

28. B.J. McBride and S. Gordon, Computer Program for the Calculation of Complex Chemical Equilibrium: Compositions and Applications, NASA Reference Publication 1311, NASA Lewis Research Center, 1996

29. S. Kostka, S. Roy, P.J. Lakusta, T.R. Meyer, M.W. Renfro, J.R. Gord, and R. Branam, “Comparison of Line-peak and Line-scanning Excitation in Two-color Laser-induced-fluorescence thermometry of OH”, Applied Optics, vol. 48, No. 32, 2009, pp. 6332-6343

30. J. Luque and D.R. Crosley, “LIFBASE, database and spectral simulation for diatomic molecules”, SRI International Report MP-99-009, 1999

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

16

(b)

(a)

Figure 1: (a) Schematic diagram of the preheated plasma assisted ignition test cell and OH LIF apparatus; (b) Hencken burner used for calibration to obtain absolute OH number density.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

17

Figure 3: Time-resolved broadband intensity (sum of intensity counts for each pixel in ICCD images of the plasma such as shown in Fig. 2) during a single discharge pulse (pulse #10 in the burst) in air at different pressures. T0=500 K, ϕ=0.3, ν=10 kHz. Camera gate 490 psec.

Figure 2: Broadband ICCD images of the plasma during a single discharge pulse (pulse #10 in the burst) in air. T0=500 K, p=50 torr, ν=10 kHz. Camera gate 490 psec. Timing of the gate is shown on top of each image. Intensity of images taken at t<10 nsec is enhanced for illustrative purposes.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

18

Figure 4: ICCD images of the plasma in fuel-air mixtures illustrating discharge uniformity. T0=500 K, ϕ=0.3, ν=10 kHz, pulse #10. Camera gate 50 nsec.

Figure 5: Typical OH LIF excitation spectrum used for temperature measurement. Spectrum taken 2 μsec after the last pulse in the burst (ν=10 kHz, 50 pulses). H2-air, T0=500 K, P=100 torr, ϕ=0.12.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

19

Figure 6: Boltzmann plot obtained from the OH LIF excitation spectrum shown in Fig. 5 (H2-air, T0=500 K, P=100 torr, ϕ=0.12, ν=10 kHz, 50 pulses). Best fit rotational temperature is T=587±27 K.

Figure 7: Comparison of temperatures measured by OH LIF and by CARS after a ν=10 kHz, 50-pulse discharge burst in a C2H4-air mixture at T0=500 K, P=100 torr, ϕ=0.09.

Figure 8: Comparison of temperatures measured by OH LIF and by CARS in a ν=10 kHz, 50-pulse discharge burst (2 μsec after the last pulse) in different fuel-air mixtures at T0=500 K, P=100 torr.

Figure 9. Experimental and predicted temperature, OH concentration, and OH(A→X) emission traces after the discharge burst in a ϕ=0.4 H2-air mixture at T0=500 K (P=80 torr, ν=10 kHz, 115 pulses) [16], overlapped with the present psec CARS temperature measurements (P=92 torr, ν=10 kHz 120 pulses)

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

20

Figure 10. Comparison of psec CARS temperature measurements after the discharge burst in a ϕ=0.4 H2-air mixture at T0=500 K, P=92 torr and ν=10 kHz with kinetic modeling calculations, for different burst durations (50-120 pulses). Solid lines, temperature; dashed lines, OH number density.

Figure 11: Vibrational temperature of nitrogen measured by CARS after a ν=10 kHz discharge burst in air T0=500 K, P=100 torr (50 and 100 pulses).

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

21

Figure 12: Comparison of experimental and predicted time-resolved, absolute OH number densities after a ν=10 kHz, 50-pulse discharge burst in C2H4-air mixtures at T0=500 K, P=100 torr, and different equivalence ratios. Data points for temperatures measured at the end of the burst (time delay 2 μsec) are circled.

Figure 13: Comparison of experimental (y-axis on the left) and predicted (y-axis on the right) time-resolved, absolute OH number densities after a ν=10 kHz, 50-pulse discharge burst in CH4-air mixtures at T0=500 K, P=100 torr, and different equivalence ratios. Data points for temperatures measured at the end of the burst (time delay 2 μsec) are circled.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82

22

Figure 14: Comparison of experimental and predicted time-resolved, absolute OH number densities after a ν=10 kHz, 50-pulse discharge burst in C3H8-air mixtures at T0=500 K, P=100 torr, and different equivalence ratios. Data points for temperatures measured at the end of the burst (time delay 2 μsec) are circled.

Figure 15: Comparison of experimental (y-axis on the left) and predicted (y-axis on the right) time-resolved, absolute OH number densities after a ν=10 kHz, 50-pulse discharge burst in H2-air mixtures at T0=500 K, P=100 torr, and different equivalence ratios. Data points for temperatures measured at the end of the burst (time delay 2 μsec) are circled.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

2, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

2-31

82