American Institute of Aeronautics and Astronautics

1

Mechanisms of Kinetic Combustion Enhancement by

O2(a1Δg)

Timothy Ombrello1

U.S. Air Force Research Laboratory, Propulsion Directorate, Wright-Patterson AFB, OH, 45433

Wenting Sun,2 Sang Hee Won,

3 Yiguang Ju

4

Princeton University, Princeton, NJ, 08544

and

Skip Williams5 and Campbell Carter

6

U.S. Air Force Research Laboratory, Propulsion Directorate, Wright-Patterson AFB, OH, 45433

The isolated enhancement effects of O2(a1Δg) on lifted flame propagation was investigated

experimentally and numerically at reduced pressures. Through quantitative absorption

measurements it was found that O2(a1Δg) was produced in excess of 5500 ppm and was

isolated from all other plasma-produced species via NO injection before transport to the

C2H4 lifted flame. Clear trends of increased flame propagation enhancement with increased

O2(a1Δg) concentrations were found, with up to 3% enhancement observed. Numerical

simulations with the current O2(a1Δg) kinetic mechanisms, containing only hydrogen species,

showed significant trend deviations from the experimental results, indicating errors in the

kinetics. The inclusion of new estimated temperature dependent rates of O2(a1Δg) collisional

quenching by hydrocarbon species mitigated the deviation, but require validation. Flow

reactor experiments with conditions mimicking the early stages of the flame showed that low

temperature oxidation of C2H4 in the range of 583 K to 728 K mimicked the low temperature

kinetic pathways of enhancement through NO sensitization with OH production. Therefore

the inclusion of O2(a1Δg) quenching by hydrocarbon species, as well as low temperature

chemistry is necessary for accurate modeling of the enhancement pathways by O2(a1Δg).

Nomenclature

eV = electron volts

NP w/ NO = plasma off with addition of NO

P w/ NO = plasma on with addition of NO

P w/o NO = plasma on without addition of NO

ppm = parts per million

1 NRC Post-Doctoral Research Associate, Aerospace Propulsion Division, 1950 Fifth Street, Member AIAA.

2 Graduate Student, Department of Mechanical and Aerospace Engineering, Engineering Quadrangle, Olden Street,

Student Member AIAA. 3 Senior Research Associate, Department of Mechanical and Aerospace Engineering, Engineering Quadrangle,

Olden Street, Member AIAA. 4 Associate Professor, Department of Mechanical and Aerospace Engineering, Engineering Quadrangle, Olden

Street, Associate Fellow AIAA. 5 Present Position and Address: Chief, Modeling and Analysis Section, Air Force Research Laboratory, Directed

Energy Directorate, Detachment 15, 535 Lipoa Parkway, Suite 200, Kihei, Maui, Hawaii 96753, Senior Member

AIAA. 6 Senior Aerospace Engineer, Aerospace Propulsion Division, 1950 Fifth Street, Associate Fellow AIAA.

48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition4 - 7 January 2010, Orlando, Florida

AIAA 2010-1586

Copyright © 2010 by the American Institute of Aeronautics and Astronautics, Inc.The U.S. Government has a royalty-free license to exercise all rights under the copyright claimed herein for Governmental purposes.All other rights are reserved by the copyright owner.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

2

I. Introduction

HERE has recently been considerable interest in trying to understanding the effects of individual plasma-

produced species on the enhancement of critical combustion phenomena such as ignition, flame propagation,

flame stabilization, and extinction. One of the main motivations for trying to understand the fundamental

interactions of plasma-assisted combustion systems comes from the problems of ignition and sustained combustion

in the harsh environments of high-speed air-breathing propulsion devices. The short residence times create limiting

conditions for the extraction of chemical heat release to sustain combustion in the system. Past research has shown

that when plasma is applied to combustion systems, it can enhance critical combustion phenomena,[1-20]

but an

understanding of the fundamental interaction remains a topic of much discussion and debate. In a combustion

system the two main governing parameters are the temperature and the radical pool concentration. Plasma activation

can couple into both parameters by providing elevated temperatures as well as new kinetic pathways to affect the

radical pool concentration and the rate of fuel oxidation. The challenge is that hundreds and thousands of species

and reaction pathways are created through plasma activation that are not present in typical combustion systems.

Therefore, in order to develop predictive modeling tools for plasma-enhanced combustion systems, at a minimum,

the most important plasma-produced species have to be identified and their contribution through reaction pathways

have to be understood. Once some of the most important species interactions are understood, steps can be taken to

develop plasma systems to target and produce key enhancement species and couple into combustion systems more

efficiently.

One of the possible key plasma-produced species that has been identified is O2(a1Δg). This electronically excited

oxygen species has a low excitation energy (0.98 eV) and therefore can be produced in most oxygen plasmas. It has

a very long lifetime in comparison to other excited species and has been proposed as having a higher oxidation

potential than ground state O2. Because of this, there here have been numerous computational studies aimed at

quantifying the enhancement of ignition and flame stabilization by O2(a1Δg). Starik and Titova showed that when

O2(a1Δg) was produced by laser radiation in a supersonic flow of H2-air, the induction time and temperatures

necessary for ignition behind a shockwave were reduced significantly.[21]

They attributed the enhancement to come

from new pathways with O2(a1Δg) to generate active species, such as O, H, and OH. Work was also performed

computationally using an electrical discharge to achieve similar results as the laser radiation.[22]

Detailed

investigations of the ignition kinetics with the presence of O2(a1Δg) by non-equilibrium excitation in a H2-O2 system

was reviewed in a paper by Popov.[23]

The review pointed out that if the quenching of O2(a1Δg) by H2 is not

considered, there will be a gross overestimate of the amount of enhancement because the collisional quenching rate

increases significantly with temperature. Popov also emphasized that there is a lack of experimental studies of the

effect of electronically excited species on combustion phenomena.

Starik et al. studied ignition enhancement in an H2-O2 mixture. Here, they created O2(a1Δg) along with O(

1D)

with a flash photolysis approach: O3 was the sensitizer species and the radiation was from a UV laser beam. The

collective effect of these two species led to decreased ignition delay times by several orders of magnitude.[24]

The

absorbed energy by UV laser radiation was very small compared to the energy required for the same amount of

enhancement by thermal means, indicating the distinct advantage of having strong non-equilibrium excitation for

combustion enhancement. Nevertheless, the results were not verified with experiments.

Kozlov et al. conducted numerical simulations to show the enhancement of H2-O2 flame speed with O2(a1Δg)

addition.[25]

The results showed that a concentration of 10% O2(a1Δg) gave more than a 50% increase in the laminar

flame velocity. Lean mixtures were enhanced more than stoichiometric and rich mixtures because of the chain

initiation and branching reactions involved. Bourig et al. extended the numerical modeling by investigating ignition

and flame propagation, as well as flame stabilization by O2(a1Δg).

[26]

Experimental validation of the enhancement of O2(a1Δg) on ignition was provided by a measurement of the

emission from OH* at 306.4 nm.[27]

Here, they monitored OH* behind a shockwave in an H2-O2 mixture that was

subjected to an electrical discharge. It was found that there was accurate reproduction of the experimental results

with the kinetic mechanism and that OH* might be a good indicator of enhancement by O2(a1Δg) because of the

elevated radical pool concentrations. Skrebkov and Karkach performed subsequent simulations using the ignition

and emission spectroscopy results from reference [27] and showed that there was reasonable agreement; however,

they emphasized that the main stumbling block was the “availability of experimentally evaluated amounts of

O2(a1Δg),” as well as other plasma excited species.

[28]

There have been only two experimental investigations of the isolated effects of O2(a1Δg) on combustion

phenomena. Smirnov et al. performed experiments aimed at isolating the effect of O2(a1Δg) on the ignition of H2-O2

mixtures at low pressure between 1.33 kPa and 2.67 kPa.[29]

The results showed decreased induction times. Here,

they noted that to achieve the same delay by heating the mixture, the energy input would be 4.4 times greater than

T

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

3

that needed to produce the O2(a1Δg). The experiments were designed to minimize all other plasma-produced species,

especially O and O3, with the injection of Hg and its oxide (HgO) into the flow. The results showed that there was a

hundred fold decrease in the emission from O, but no absolute measurement of its concentration was performed.

Furthermore, the O3 concentration was not measured while the O2(a1Δg) characterized was only through emission

measurements. Even though the experiments were the first of their kind in an attempt to isolate the effect of O2(a1Δg)

and the results agreed reasonably well with their previous calculations,[30, 31]

it was not certain whether O2(a1Δg) was

the only species causing the enhancement. In addition, the experiments were conducted only for an H2-O2 mixture.

More recently, the isolated effects of O2(a1Δg) on flame propagation was investigated experimentally.

[11] The

results showed the completely isolated effect of O2(a1Δg) and a clear correlation between O2(a

1Δg) addition and the

flame propagation speed, but the kinetic enhancement mechanisms involved were not understood. The goal of the

present work was to build upon the experimental data of our past investigation[11]

and develop an understanding of

the kinetic mechanisms of enhancement by O2(a1Δg) on C2H4 flames.

II. Experiments

A. Lifted Flame and Diagnostics System

To investigate the flame propagation enhancement by O2(a1Δg), a laminar lifted-flame burner was used in a

variable pressure vacuum chamber. The lifted-flame burner consisted of a central fuel jet with a diameter of 1.04

mm and was located in a 90 mm ID fused silica (quartz) tube to contain the co-flow of O2 and Ar in the chamber.

The fuel nozzle was aerodynamically shaped to produce a uniform vel ocity profile at the exit. To ensure that the co-

flow was uniform, two stainless steel meshes coated with quartz for chemical inertness were separated by 3 cm and

were located between the oxidizer inlet of the burner and the fuel jet exit. The gases used in the experiments were

C2H4 for the fuel and ultra-high purity O2 (99.99%) and Ar (99.95%) mixed for the oxidizer.

The high velocity fuel jet (approximately 20 m/s – 40 m/s) and low velocity co-flow (approximately 0.15 m/s –

0.2 m/s) created a mixing layer with a stoichiometric contour where the premixed flame head of a lifted flame was

located (shown in the top right insert in Fig. 1). The lifted flame, which is also called a tribrachial (triple) flame, had

a premixed flame head and diffusion flame tail where the flame was always anchored on the stoichiometric contour.

The lifted flame could be located at different distances from the fuel jet nozzle depending upon the local flow

velocity. For a fixed flow field, the flame is located in a stationary position where the lifted flame speed at the

premixed flame head is balanced with the local flow velocity. If the flame speed increases, the liftoff height

decreases to re-establish a local dynamic balance between the flame speed and flow velocity.

Due to the slow laminar boundary layer development and the velocity and concentration gradients created, the

lifted flame height was very sensitive to the changes in flame speed and therefore provided excellent flame geometry

for the direct observation of flame speed enhancement. Since the fuel and oxidizer were not mixed far upstream of

the flame, there was very short residence time for the fuel and oxidizer to react in the cold flow. The short residence

time helped to further decouple the enhancement effects to be directly from reactions in the flame zone and not far

upstream in the cold un-reacted flow.

To excite the co-flow in the low pressure experiments, an electrodeless microwave discharge (McCarroll cavity

driven by an Opthos MPG-4M microwave power supply) with up to 100 Watts of power was used external to the

chamber, upstream of the lifted flame burner to activate the O2. The co-flow composition was 15% O2 in 85% Ar

for 3.61 kPa and 11.9% O2 in 88.1% Ar for 6.73 kPa. The microwave discharge produced excited Ar, as well as

multiple oxygen-containing species including O, O3, O2(v), O(1D), O(

1S), O2(a

1Δg), O2(b

1Σg), etc. The only species

to survive in the system that would reach the flame would be O, O3, and O2(a1Δg) since all other species would

quench on the wall and with other gas species.

To isolate O2(a1Δg) from O and O3, NO was injected into the system downstream of plasma. The NO reacted

catalytically to remove O3 and O from the system while not affecting the O2(a1Δg) concentration. Therefore, when

NO was injected into the system and the plasma was turned on and off, the only species difference would be the

presence of O2(a1Δg).

To quantitatively measure the O3 and O2(a1Δg) in the system, absorption techniques were adopted. A one-pass

absorption measurement was used for O3 while O2(a1Δg) was measured by using highly sensitive integrated-cavity-

output spectroscopy by absorption at the (1,0) band of the b1Σg

+ - a

1Δg Noxon system. A more detailed description of

the experiments and diagnostic techniques can be found in the author’s previous work.[11]

B. Flow Reactor

To investigate the kinetic reaction pathways of O2(a1Δg), a heated flow reactor was developed and integrated

with the microwave plasma device, ICOS cavity, and FTIR instrument. A schematic of the system is shown in Fig.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

4

2. The flow reactor tube was comprised of a 2.2 cm ID by 230 cm long quartz tube with multiple sampling and

injection ports. At the entrance to the flow reactor was the same microwave plasma device used in the lifted flame

system; as before the O2(a1Δg) was isolated by the injection of NO. The flow then passed into small quartz capillary

tubes which were within a ceramic honeycomb where the fuel was introduced to the system. At the exit of the

capillary tubes, the fuel and oxidizer (containing O2(a1Δg)) merged and were also heated simultaneously. Therefore

the reactants mixed and were heated within a very short residence time, on the order of a few milliseconds. The

temperature along the flow reactor was maintained via external electrically resistive heat tape. The temperature was

monitored by thermocouples placed along the centerline at various positions down the flow reactor tube. The ICOS

cavity was used to sample the flow downstream of the fuel and oxidizer mixing system to ensure that O2(a1Δg)

existed in significant concentrations at the beginning of the reactor. Samples were continuously pulled from the flow

reactor to an FTIR instrument. In this way concentrations of C2H4, NO, NO2, and any stable fuel fragments from the

reactions with NO and O2(a1Δg) were measured.

Figure 1. Experimental set-up schematic of variable pressure lifted flame burner integrated with microwave

plasma discharge device and plasma flow diagnostics system. The abbreviation “PD” denotes photodetector.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

5

The conditions at the

entrance of the

microwave plasma device

were chosen to be similar

to the lifted flame

experiments at 6.73 kPa

with 11.9% O2 in 88.1%

Ar at 5.69 kPa. The

concentration of NO

added to the system was

2000 ppm, which

provided in excess of

3000 ppm of O2(a1Δg) for

the flow reactor. The

capillary tube and

honeycomb mixing

system provided a

pressure drop which

allowed the reactor to run

at a pressure of 1.8 kPa.

The temperatures in the

flow reactor were

between 293 K and 728

K and provided residence

times between 138 ms

and 849 ms.

III. Results and Discussion

A. Previous Experimental Results

Experiments were performed to examine the effects of O2(a1Δg) on flame propagation speed. Ethylene was used

as the fuel to produce stable lifted flames at low pressures (3.61 kPa and 6.73 kPa) with O2 and Ar as the oxidizer

co-flow. A fixed co-flow velocity was used with increasing fuel jet velocities. For each fuel jet velocity, the flame

was at a stationary liftoff height where photographs were taken of the flame while the pressure and temperature were

recorded simultaneously. The procedure was then repeated with the microwave plasma turned on. The change in

flame liftoff height was calculated from the photographs. A similar procedure was used again with NO addition just

downstream of the microwave plasma cavity, replicating the residence times in the plasma afterglow diagnostics

system. The results showed that there was a significant change in the flame liftoff height when the plasma was

turned on with and without NO addition. Therefore, the flame propagation enhancement came from a combination

of O3 and O2(a1Δg) without NO addition and only from O2(a

1Δg) with NO addition. Concentrations between 500

ppm and 2000 ppm of NO were added downstream of the plasma to give different concentrations of O2(a1Δg) at the

flame front. The more NO that was added, the faster the O3 and O were consumed catalytically before they reacted

with O2(a1Δg).

The flow field, temperature, and pressure remained constant when the plasma was cycled off and on, so a direct

comparison between the plasma off and on conditions can be attributed to the enhancement by O3 and O2(a1Δg). The

temperature and pressure remained constant within an uncertainty of 0.1 K and 26.7 Pa respectively. Experiments

were performed to quantify that the uncertainty in temperature and pressure was not affecting the flame liftoff height

enough to mask the enhancement by O3 and O2(a1Δg) addition. For a temperature change of 0.1 K, the flame liftoff

height changed by 0.07 mm and for a pressure change of 26.7 Pa the flame liftoff height changed by 0.29 mm.

Therefore, for the change in flame liftoff height observed in the experiments with O3 and O2(a1Δg) addition, which

was on the order of millimeters, the uncertainty was more than an order of magnitude smaller.

Also of note was that the presence of NO did affect the flame structure by changing the stoichiometric contour

and flame speed slightly, but the comparison was between the plasma being off and on with constant NO addition.

Furthermore, since the flow diagnostics showed that the NO acted catalytically to reduce the O3 and did not change

in concentration, the conditions of the plasma being off and on could be compared directly to each other.[11]

The near

Figure 2. Experimental set-up schematic of flow reactor integrated with microwave

plasma discharge device.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

6

zero concentrations of NO2 would have a negligible effect on flame speed since the enhancement was calculated to

be more than three times lower than that of O3.

The same O2 loadings, flow rates, and pressures were then used on the plasma afterglow diagnostics system to

find the concentrations of O3 and O2(a1Δg) that were present at the lifted flame. The measured concentrations of O3

and O2(a1Δg) were corrected for residence time using the kinetic model described in our previous paper.

[11] The

results are shown in Fig. 3 with the trend of increasing change in flame liftoff height, and hence more flame speed

enhancement, with increasing O2(a1Δg) concentration for both 3.61 kPa and 6.73 kPa. Overall with NO addition,

there was no change in the flow-field, temperature, or any species other than O2(a1Δg) when the plasma was cycled

on and off. Therefore, by turning on the plasma, it was the equivalent of introducing a pure source of O2(a1Δg) and

reducing the ground state O2 by the same amount.

With the results showing clearly

the enhancement of flame

propagation speed by O2(a1Δg), it was

important to determine the

enhancement quantitatively. Unlike

the lifted flame at atmospheric

pressure described in our previous

work on O3,[20]

a cold flow similarity

solution does not correctly describe

the low pressure experiments.

Therefore, an indirect method was

used to find the amount of flame

speed enhancement quantitatively.

The change in flame liftoff height

shown in Fig. 3 indicated that

approximately ten times the amount

of O2(a1Δg) (approximately 5500

ppm) was needed to achieve the same

enhancement as O3 (approximately

500 ppm). Therefore, the

enhancement of flame propagation

with the addition of O3 can be used to

quantify the enhancement by O2(a1Δg). The C2H4 laminar and lifted flame speed enhancement was computed using

the C2H4 kinetic mechanism of Wang and Laskin[32]

with the addition of the O3 reactions shown in our previous

work.[20]

The laminar and lifted flame speed are related through the square root of the density ratio;[20]

therefore the

percent enhancement of the flame speeds were comparable. Nevertheless, the lifted flame speed was used to

describe the enhancement here. The results for the conditions of 500 ppm of O3 addition at 3.61 kPa and 6.73 kPa

showed approximately 1% enhancement of the lifted flame speed.[20]

Therefore, since the flame liftoff height change

for 500 ppm O3 is roughly equivalent to 5500 ppm O2(a1Δg), it was reasonable to assume that their effects on flame

speed enhancement were comparable. The overall lifted flame speed enhancement would be higher because of the

kinetic-induced hydrodynamic enhancement described in our previous work.[20]

If the effect of O3 at low pressure is

comparable to what was found at high pressure, then 5500 ppm of O2(a1Δg) will give approximately 2% to 3%

enhancement of the lifted flame speed.

B. Modeling Analysis with Current O2(a1Δg) Kinetics

Numerical simulations were performed in order to explain the enhancement mechanism with the addition of

O2(a1Δg). The C2H4 combustion mechanism

[32] with the O3 reactions added

[20] was used along with the inclusion of

O2(a1Δg) reactions. The reactive and collisional quenching rates of O2(a

1Δg) with oxygen and inert species as a

function of temperature have been well studied and were used to compute the cold flow transport of O2(a1Δg) in the

ICOS system and prior to the flame in the combustion system. The O2(a1Δg) reactions were added to the C2H4/O3

mechanism. The reactions of O2(a1Δg) with the fuel have also been well studied at 298 K, but there is little data at

intermediate and high temperatures. The most studied O2(a1Δg) reactions applicable to a combustion system are for

H2-O2 mixtures. These reactions of hydrogen and oxygen containing species with O2(a1Δg) are shown in Table 1;

these rates have been compiled specifically for H2-O2 mixtures activated by plasma.[21, 23, 33-35]

There were a few

differences that arose, specifically regarding the two reactions

Figure 3. Plot of experimental results of flame liftoff change with

O2(a1Δg) and O3 concentration for a plasma power of 80 Watts.

0

1

2

3

4

5

6

7

8

0 1000 2000 3000 4000 5000 6000 7000

ΔH

L[m

m]

Concentration [ppm]

O3 (exp.)

O2(a) (exp.)

O3 (exp.)

O2(a) (exp.)

3.61 kPa

6.73 kPa

O3

O2(a1Δg)

O3

O2(a1Δg)

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

7

1

2 gH O a O OH (1)

1

2 2 2 2gH O a H O . (2)

There was a small difference in reaction (1) between reference[23]

and[34]

, as well as in reaction (2) between

reference[23]

and,[35]

as shown in Table 1. The differences in the reported reaction rates will be discussed in relation

to the sensitivity of the computed flame speed enhancement shortly.

The reactions were added to the C2H4/O3 kinetic mechanism and the results are shown in Fig. 4 using the

concentrations of O3 and O2(a1Δg) from the experiments. The results of enhancement using the O2(a

1Δg)

concentrations found in the experiments showed flame speed enhancement of more than 5%. When compared to the

results of lifted flame speed enhancement by O3 (which was found to be approximately 1%) there was a large

discrepancy compared with the trends of the experimental results. For example, in the experiments at 3.61 kPa, the

change in flame liftoff height for 500 ppm O3 was approximately equal to 5500 ppm O2(a1Δg) (Fig. 3). According to

the numerical simulations, there was more than a factor of five difference in the enhancement of flame speed.

Considering that O3 and O2(a1Δg) should enhance the lifted flame speed similarly for lean and rich more than

stoichiometric,[25, 26]

there was a significant error in the O2(a1Δg) kinetic calculations by only including the current

kinetics with hydrogen containing species. With regard to the differences in the rates for reactions (1) and (2), the

deviations in the flame speed enhancement were no more than ± 4%. The vertical error bars in Fig. 4 show the

negligible enhancement differences when using the two published rates for reactions (1) and (2) and therefore the

sensitivity of the flame speed to the differences in reaction rates do not account for the significant deviations from

the experimental trends.

To understand what caused the significant flame speed enhancement by O2(a1Δg), a rate of production plot of

O2(a1Δg) in the early stage of the flame is shown in Fig. 5. The major consumption pathways of O2(a

1Δg) was from

Reaction Reaction Constant

[cm3/molecule/s]

Temperature

Dependence

Activation Energy

[kJ]

Reactive Quenching of O2(a1Δg)

H+O2 → O+OH [33]

5.00x10-9

0 60.3

H+O2(a1Δg) → O+OH

[34] 1.82x10

-10 0 26.5

H+O2(a1Δg) → O+OH

[23] 6.55x10

-11 0 21.0

OH+O2(a1Δg) → O+HO2

[34] 2.16x10

-11 0 141.4

OH+O2(a1Δg) → H+O3

[34] 7.31x10

-17 1.44 226.4

H2+O2(a1Δg) → OH+OH

[34] 2.82x10

-9 0 141.4

H2+O2(a1Δg) → HO2+H

[34] 4.13x10

-12 0 151.6

H2O+O2(a1Δg) → OH+HO2

[21] 9.03x10

-8 0.5 209.7

H2O+O2(a1Δg) → O+H2O2

[21] 2.05x10

-12 0.5 283.5

Collisional Quenching of O2(a1Δg)

H2+O2(a1Δg) → H2+O2

[35] 2.16x10

-13 0 21.6

H2+O2(a1Δg) → H2+O2

[23] 1.68x10

-12 0 32.0

H+O2(a1Δg) → H+O2

[34] 6.97x10

-16 0 0

OH+O2(a1Δg) →OH+O2

[34] 5.65x10

-18 0 0

HO2+O2(a1Δg) → HO2+O2

[34] 5.65x10

-18 0 0

H2O+O2(a1Δg) → H2O+O2

[34] 5.65x10

-18 0 0

H2O2+O2(a1Δg) →H2O2+O2

[34] 5.65x10

-18 0 0

Other Reactions with O2(a1Δg)

H+HO2 → H2+O2(a1Δg)

[34] 3.32x10

-12 0 2.4

OH+O → H+O2(a1Δg)

[34] 9.63x10

-12 0 51.8

O3+OH → HO2+O2(a1Δg)

[34] 7.97x10

-13 0 8.3

O3+HO2 → OH+O2+O2(a1Δg)

[34] 1.66x10

-14 0 8.3

HO2+HO2 → H2O2+O2(a1Δg)

[34] 1.49x10

-11 0 4.2

H2O2+O → H2O+O2(a1Δg)

[34] 6.97x10

-13 0 17.7

Table 1. Reaction rates of O2(a1Δg) with hydrogen containing species.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

8

the branching reaction with H, with

some contribution of collisional

quenching by H2. The reaction of

O2(a1Δg) with H will significantly

enhance the flame since it is a

primary radical branching reaction.

Therefore, it was reasonable that the

enhancement was so large in the

numerical simulations.

To take one step deeper, the

changes in the radical pool in the

early stages of the flame were

plotted and are shown in Fig. 6. In

the early stages of the flame where

the temperature is slightly elevated,

the O2(a1Δg) was consumed, causing

an increase in the concentration of

O and OH and a subsequent

decrease in the C2H4 concentration.

The increase in concentration of OH

in the pre-heat zone of the flame

would provide chemical heat release

through subsequent reactions early

in the flame to enhance the overall flame propagation speed significantly. These results of higher concentrations of

OH in the early stages of the flame leading to chemical heat release and enhanced flame speed was shown in our

previous work through the study of O3 addition.[20]

The increase in the radical pool concentration of O from the

reactions with O2(a1Δg) were investigated in more depth through a rate of production analysis. The results showed

that the major pathway for O consumption was from the reaction with the parent fuel, C2H4, and its fragment, CH3.

C. Modeling with Inclusion of O2(a1Δg) Quenching by Hydrocarbons

With an understanding of how O2(a1Δg) enhanced the flame speed in the numerical simulations, it became

apparent that there were two possible explanations for the discrepancy shown between Fig. 3 and Fig. 4 with respect

to the enhancement by O3 and O2(a1Δg). First, a significant concentration of the O2(a

1Δg) could quench before

reacting with H. The collisional quenching reactions of O2(a1Δg) with the parent fuel or its fuel fragments could

decrease the concentration

significantly, therefore leading to

less enhancement. To achieve the

same enhancement of O2(a1Δg) as

the calculations with O3, there

would need to be approximately

900 ppm O2(a1Δg) (shown as gray

point in Fig. 4 for 3.61 kPa).

Therefore, approximately 4600 ppm

of O2(a1Δg) would have to

collisionally quench, allowing only

a small fraction of the original

concentration to react with H. The

enhancement would be less,

especially if the dominating

reactions of O2(a1Δg) were with the

parent fuel and not providing chain

branching as in the reaction of H

with O2(a1Δg). Second, there could

be a combination of reactive and

collisional quenching pathways for

O2(a1Δg) that could be responsible

Figure 5. Rate of production plot of O2(a1Δg) superimposed on the

temperature profile showing the major consumption pathways of

O2(a1Δg) with current published rate data of hydrogen containing species.

200

500

800

1100

1400

1700

2000

-2.5E-06

-2.0E-06

-1.5E-06

-1.0E-06

-5.0E-07

-2.0E-21

5.0E-07

0 0.2 0.4 0.6 0.8 1 1.2

Tem

pera

ture

[K

]

Ra

te o

f P

rod

ucti

on

O2(a

1Δ

g)

[mo

l/cm

3/s

]

Spatial Coordinate [cm]

H+O2(a)=>OH+O

H2+O2(a)=>H2+O2

HO2+HO2=>H2O2+O2(a)

H+HO2=>O2(a)+H2

OH+O=>O2(a)+H

Temperature

H+O2(a1Δg) → OH+O

H2+O2(a1Δg) → H2+O2

HO2+HO2 → H2O2+O2(a1Δg)

H+HO2 → H2+O2(a1Δg)

OH+O → H+O2(a1Δg)

Temperature

Figure 4. Plot of computational results of lifted flame speed enhancement

with O2(a1Δg) and O3. The gray circle indicates that to achieve similar

enhancement between O2(a1Δg) and O3, only 900 ppm of O2(a

1Δg) would

have to be present.

0

1

2

3

4

5

6

7

0 1000 2000 3000 4000 5000 6000

Sli

fted

En

ha

ncem

en

t [%

]

Concentration [ppm]

O3 Current Kinetics O3 Current Kinetics

O2(a) Current Kinetics O2(a) Current Kinetics

3.61 kPa 6.73 kPa

O3

O2(a1Δg) (Current Kinetics)

O3

O2(a1Δg) (Current Kinetics)

3.61 kPa O2(a1Δg)

(Current Kinetics)

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

9

for the enhancement observed in the experiments.

The reactive and collisional quenching rates for some of the hydrocarbon species were added to the kinetic

mechanism one at a time to test the sensitivity of flame speed enhancement. In Table 2 is a list of the reactions

along with their rates. Initially the reactions with CH4 of

1

4 2 4 2gCH O a CH O (3)

and

1

4 2 3 2gCH O a CH HO (4)

were added, but these did not change the flame speed, which was reasonable considering the low concentrations of

CH4 in the system. Next, noting that the inclusion of O2(a1Δg) collisional quenching by the parent fuel H2 in an H2-

O2 system via reaction (2) was found in previous numerical investigations to be significant and decreased the

effectiveness of O2(a1Δg) enhancement,

[23] it was reasonable to include collisional quenching by the parent fuel via

1

2 4 2 2 4 2gC H O a C H O (5)

for our hydrocarbon fueled combustion system. There are some collisional reaction rates for O2(a1Δg) with C2H4 and

other small hydrocarbon fuels, but for 298 K.[35, 37]

To the authors knowledge, there are no verified quenching rates

of hydrocarbon species (specifically C2H4) with O2(a1Δg) at elevated temperatures.

The work of Borrell and Richards found that the temperature dependence of the O2(a1Δg) quenching rate by H2

was approximately Arrhenius [38]

in

nature and that rates of other species

also follow an Arrhenius temperature

dependence. Therefore, it was

assumed that electronic quenching

rate for C2H4-O2(a1Δg) collisions

might also follow this temperature

dependence. Starting with the rate for

298 K,[35, 37]

an Arrhenius temperature

dependent rate was estimated to

explain the trend discrepancy shown

between the results in Fig. 3 and Fig.

4. Three different temperature

dependencies were chosen with a

range of activation energies.

Previously published temperature

dependent collisional quenching rates

of O2(a1Δg) have shown that the

activation energy range is typically

between 15 kJ/mole and 20

kJ/mole.[39]

Furthermore, the

Reaction Reaction Constant

[cm3/molecule/s]

Temperature

Dependence

Activation Energy

[kJ]

CH4+O2(a1Δg) → CH4+O2

[36] 1.40x10

-18 0 0

CH4+O2(a1Δg) → CH3+HO2

[37] 6.14x10

-12 0 149.0

C2H4+O2(a1Δg) → C2H4+O2

[*low] 7.71x10

-16 0 15.0

C2H4+O2(a1Δg) → C2H4+O2

[*high] 3.12x10

-13 0 30.0

C2H4+O2(a1Δg) → C2H4+O2

[* exp. fit] 5.46x10

-10 0 48.6

C2H4+O2(a1Δg) → C2H3+HO2

[*] 7.01x10

-11 0 146.5

Table 2. Reaction rates of O2(a1Δg) with hydrocarbon species.

[*low] = estimated rate with low activation

energy, [*high]

= estimated rate with high activation energy, [*exp. fit]

= estimated rate with activation energy to fit

trend of experimental results, [*]

= estimated rate.

Figure 6. Temperature and mole fraction profiles of C2H4 flame,

dashed lines = w/o O2(a1Δg), solid lines = w/ O2(a

1Δg).

200

500

800

1100

1400

1700

2000

-1.0E-03

2.0E-03

5.0E-03

8.0E-03

1.1E-02

1.4E-02

1.7E-02

2.0E-02

-0.5 0 0.5 1 1.5T

em

pera

ture

[K

]

Mo

le F

ra

cti

on

Spatial Coordinate [cm]

O2(a1Δg)

C2H4

O

OH

T

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

10

activation energy for the collisional quenching of O2(a1Δg) by H2 is as high as 32 kJ/mole.

[23] Therefore, the range of

15-30 kJ/mole was chosen for the activation energies of reaction (5), with a reaction constant chosen in order to

agree with published rates at 298 K.[35, 37]

The envelope of rates for reaction (5) using activation energies from 15

kJ/mole to 30 kJ/mole is shown in Fig. 7 and is labeled as “Quenching A.”

Computations were performed

using the rates within the envelope

of “Quenching A” in Fig. 7 and the

results are shown in Fig. 8. The

flame speed enhancement decreased

slightly, but not enough to explain

the discrepancy. In an attempt to

explain the disagreement, a rate for

reaction (5) was chosen in order to

bring the enhancement by O2(a1Δg)

down to the enhancement by O3. To

accomplish this, a reaction constant

of 5.46 x 10-10

cm3/molecule/s and

an activation energy of 48.6 kJ/mole

were chosen with the temperature

dependence shown in Fig. 7 as

“Quenching B.” By using the

“Quenching B” rate for reaction (5),

the results of enhancement by O3

and O2(a1Δg) were approximately

equal, therefore agreeing with the

trends of the experimental results

(Fig. 8). A rate of production

analysis of O2(a1Δg) was performed

with the “Quenching B” rate and the

major consumption pathway shifted

to reaction (5) with negligible

consumption by reaction (1). The

concentration profile of O2(a1Δg)

showed a more rapid decrease in the

early pre-heat zone of the flame, but

no appreciable increase in C2H4

consumption or O and OH

production. Furthermore, the

quenching pathway involves

electronic-to-vibrational-

translational energy transfer[40]

that

releases too little energy to

appreciably change the temperature.

The computed temperature profiles

confirmed the negligible increase in

temperature and therefore that a

significant concentration of O2(a1Δg)

was collisionally quenched and did

not affect the flame in the process.

The high activation energy and

hence strong temperature

dependence of reaction (5) given by the estimated “Quenching B” rate mitigated the discrepancy between the

experimental and computed enhancement trends, but the rate is higher than would normally be expected.

Therefore, the last possible explanation of the trend discrepancy in computed enhancement lies in the reactive

quenching of O2(a1Δg) by C2H4 and its fragments. The first assumption was to decrease the activation energy of the

reaction

Figure 7. Arrhenius temperature dependent collisional quenching rates

of O2(a1Δg) by H2 and C2H4. The rates with C2H4 are estimates.

9

11

13

15

17

19

0.5 1.5 2.5 3.5

-lo

g(R

ea

cti

on

Ra

te [

cm

3/m

ole

cu

le/s

])

1000 K/Temperature

Borrell & Richards [39]

Popov [23]

Quenching A

Quenching B

H2+O2(a1Δg) → H2+O2

C2H4+O2(a1Δg) → C2H4+O2

Rate from Becker et al. [37]

& Ackerman et al. [38]

Figure 8. Plot of computational results of lifted flame speed enhancement

with O2(a1Δg) and O3 using the estimated collisional quenching rate of

C2H4 with O2(a1Δg) from Fig. 14. “Quenching A” = inclusion of

temperature dependent quenching of O2(a1Δg) by C2H4 with Ea=30

kJ/mol, “Quenching B” = inclusion of temperature dependent quenching

of O2(a1Δg) by C2H4 to fit trend of experimental results.

0

1

2

3

4

5

6

7

0 1000 2000 3000 4000 5000 6000

Sli

fted

En

ha

ncem

en

t [%

]

Concentration [ppm]

O3 O3

O2(a) Current Kinetics O2(a) Current Kinetics

O2(a) Quenching A O2(a) Quenching A

O2(a) Quenching B O2(a) Quenching B

3.61 kPa 6.73 kPa

O3

O2(a1Δg) (Current Kinetics)

O2(a1Δg) (Quenching A)

O2(a1Δg) (Quenching B)

O3

O2(a1Δg) (Current Kinetics)

O2(a1Δg) (Quenching A)

O2(a1Δg) (Quenching B)

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

11

1

2 4 2 2 3 2gC H O a C H HO (6)

by the energy contained within O2(a1Δg). This equates to decreasing the activation energy by 0.98 eV (94.5 kJ/mole),

and the rate is shown in Table 2. The inclusion of this reaction in the kinetic mechanism did not result in any change

in flame speed enhancement because the rate is slow in comparison to other reaction pathways with C2H4 and

O2(a1Δg). Beyond reaction (6) there could be other possible product pathways which have been investigated through

quantum calculations with C2H4.[41, 42]

These pathways show that O2(a1Δg) can attack the double carbon bond to split

the parent fuel molecule, possibly providing significant enhancement of fuel oxidation rates by producing CH2O and

other hydrocarbon fragments, but the rates are not known.

D. Flow Reactor Results

In an attempt to understand the

reaction quenching pathways and

products of the reaction of C2H4 and

O2(a1Δg), flow reactor experiments

were performed. For a baseline, the

flow reactor was run at 293 K and the

FTIR results showed that there was no

change in the concentration of the

reactants in the system within the 849

ms residence time of the reactor. Then

for each temperature of 583 K, 651 K,

693 K, and 728 K, the reactor was run

with three different reactant

conditions of plasma off with NO

addition (NP w/ NO), plasma on with

NO addition (P w/ NO), and plasma

on without NO addition (NP w/o NO).

These three different run conditions

can be described in terms of the

expected reactants entering the flow

reactor. The plasma off with NO

addition provided a baseline

absorption spectrum on the FTIR with

which the other two run conditions

could be compared. For the plasma on

with NO addition, the only additional

reactant was O2(a1Δg), while for the

plasma on without NO addition, there

was less O2(a1Δg), as well as some O

and O3. FTIR spectra were taken for

all of the temperature and run

conditions and some representative

spectra at 651 K are shown in Fig. 9.

For the conditions of the plasma off

with NO addition, the many

absorption features of C2H4 were

apparent. When the plasma was

turned on with NO addition, there was

a noticeable increase in CH2O

concentration, as well as in NO2 and

CO concentrations, therefore

indicating consumption of some C2H4

and the conversion of NO to NO2.

Furthermore, when the plasma

Figure 10. Absorption spectra from FTIR for the conditions of:

Treactor=693 K, Preactor=1.8 kPa, treactor=81 ms, PFTIR=1.4 kPa, TFTIR=373

K, O2=10.4%, Ar=76.5%, C2H4=13.1%, (NO=2000 ppm).

0

0.25

0.5

0.75

1

1.25

1.5

1300 1700 2100 2500 2900 3300

Ab

sorp

tio

n [

Arb

itra

ry

Un

its]

Wavenumber [cm-1]

P w/o NO

P w/ NO

NP w/ NO

CH2OCH2O

C2H4

NO2 CO

Figure 9. Absorption spectra from FTIR for the conditions of:

Treactor=651 K, Preactor=1.8 kPa, treactor=92 ms, PFTIR=1.4 kPa, TFTIR=373

K, O2=10.4%, Ar=76.5%, C2H4=13.1%, (NO=2000 ppm).

0

0.25

0.5

0.75

1

1.25

1.5

1300 1700 2100 2500 2900 3300

Ab

sorp

tio

n [

Arb

itra

ry

Un

its]

Wavenumber [cm-1]

P w/o NO

P w/ NO

NP w/ NO

CH2OCH2O

C2H4

NO2 CO

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

12

remained on without NO addition, the CH2O and CO diminished in concentration. Note that the only difference in

the reactants without NO addition was a lower concentration of O2(a1Δg), as well as the emergence of O and O3 since

they were not removed catalytically by NO. Therefore, the results indicated that the CH2O and CO were being

produced from C2H4 through reactions either directly or indirectly with O2(a1Δg). The results at 583 K were similar

to those of 651 K, but the results at 693 K (see Fig. 10) and 728 K differed significantly where there was substantial

signals for CH2O and CO for the baseline condition of plasma off with NO addition.

The explanation for the presence of CH2O and CO for the plasma off with NO addition in Fig. 10 came from the

sensitized oxidation of C2H4, also known as the catalytic NOx effect. Referring to the papers by Doughty et al.,[43]

Hori et al.,[44]

and Dagaut et al.,[45]

the oxidation of C2H4 in the presence of NOx occurs at temperatures as low as

650 K – 700 K. The main mechanism for the NO sensitized oxidation of C2H4 comes predominately from the OH

produced catalytically via

2 2HO NO OH NO (7)

and

2H NO OH NO . (8)

The OH that is produced then reacts via two different pathways, either by OH adduction or H abstraction.[43]

For

lower temperatures the adduction of OH is dominant and there are subsequent rearrangement and decomposition

reactions to give the overall reaction of

2 4 3 2C H OH CH CH O . (9)

Reaction (9) is the primary source of CH2O when NO is present. The subsequent reactions of CH3 with oxygen

produces more CH2O, where the CH2O can readily decompose via

2 2CH O OH HCO H O (10)

to produce the HCO necessary for the production of HO2 for coupling back into the catalytic cycle of reactions (7)

and (8), as well as to produce CO.

The H abstraction reaction is

2 4 2 3 2C H OH C H H O (11)

and is more dominant at higher temperatures. The C2H3 that is produced then reacts via the overall reactions of

2 3 2 2 23 2 2C H O OH CO H O HO (12)

and

2 3 2 2 23 2C H O CO H O HO O , (13)

which both produce the HO2 necessary to feed back into the catalytic NOx system of reactions (7) and (8).[44, 45]

Reaction (9) and (10) are both active within the temperature range used in the flow reactor experiments and

therefore need to both be considered.

The emergence of NO sensitized oxidation of C2H4 created a problem of trying to observe the effect of O2(a1Δg)

at higher temperatures, but it was still possible to extract a trend from the data. Figures 11 and 12 show plots of a

preliminary analysis of the CH2O and CO concentrations, respectively, extracted from the FTIR spectra. For the

plasma off with NO addition, the concentration of CH2O and CO remained negligible until approximately 650 K –

700 K where the NO sensitized oxidation of C2H4 became active. For the plasma on with NO addition condition,

there was a clear trend of increasing CH2O and CO concentration with increasing temperature, indicating the effect

of O2(a1Δg). Lastly, for plasma on without NO addition, an increase in CH2O and CO concentration with increasing

temperature was also observed, but this increase was much less than that with NO addition. This result could be

attributed to the O that was produced by the plasma; the O would then reach the location of where the fuel and

oxidizer mixed. Since the reaction of C2H4 with O is fairly rapid at all of the temperatures involved, the sensitivity to

temperature is small. Therefore, the best representation of the effect of O2(a1Δg) on C2H4 consumption was at 651 K

where the NO sensitized oxidation of C2H4 was not active and there was the largest difference in CH2O

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

13

concentration with the plasma on and plasma off conditions. Furthermore, there remains a clear trend of elevated

concentrations of CH2O for 693 K and 728 K beyond what the NO sensitization causes.

E. Sensitized Oxidation of C2H4 by O2(a1Δg)

The emergence of CH2O and CO in the presence of O2(a1Δg) can be attributed to several possible reaction

pathways. For CH2O, one pathway could be the reaction

1

2 4 2 2 2gC H O a CH O CH O . (14)

This reaction of breaking the double

carbon bond would be extremely

beneficial to the combustion system

by initiating the radical pool with two

critical and highly reactive radicals.

The CH2O could then decompose via

reaction (10) to produce HCO and

hence the CO observed in the flow

reactor experiments. To confirm

reaction (14), the exact concentrations

of CH2O and O2(a1Δg) would need to

be known, as well as the collisional

quenching rates of O2(a1Δg) with

C2H4.

The other possible reaction

pathway could be similar to part of

the reactive scheme of NO sensitized

oxidation of C2H4. The main product

of the NO catalytic cycle of reaction

(7) and (8) is OH, which rapidly

reacts with C2H4 through different

pathways to produce CH2O and CO.

It has already been shown that a

dominant pathway of O2(a1Δg)

consumption is via reaction (1) to

produce OH. Therefore, instead of the

catalytic cycle of NOx to produce OH,

O2(a1Δg) can produce OH through

reaction (1). The reaction pathways

would then proceed through reaction

(9) and (11) and would eventually

produce the CH2O and CO that were

observed as products in the flow

reactor with the presence of O2(a1Δg)

(see Fig. 9 and Fig. 10). There is also

the question as to the origin of NO2

when O2(a1Δg) was present and the

temperature was too low for the NOx

pathways to be active (Fig. 9). Since

O2(a1Δg) does not react rapidly with

NO within the residence times of the

flow reactor, the probable conversion

could proceed through the reaction of

hydrocarbon fragments and HO2

produced from the sensitized oxidation of C2H4 by O2(a1Δg) with NO such as reaction (7) and

2 3 2 2 3 2C H O NO C H O NO . (15)

Figure 12. Preliminary experimental results of CO concentrations from

FTIR with and without plasma on and NO injection.

0

4000

8000

12000

16000

550 600 650 700 750

Co

ncen

tra

tio

n o

f C

O [

pp

m]

Flow Reactor Temperature [K]

NP w/ NO

P w/ NO

P w/o NO

Figure 11. Preliminary experimental results of CH2O concentrations

from FTIR with and without plasma on and NO injection.

0

1000

2000

3000

4000

5000

550 600 650 700 750

Co

ncen

tra

tio

n o

f C

H2O

[p

pm

]

Flow Reactor Temperature [K]

NP w/NO

P w/ NO

P w/o NO

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

14

IV. Summary and Conclusions

A platform to study quantitatively the enhancement effects of plasma-produced O2(a1Δg) on C2H4 lifted flame

propagation speeds was developed. It was found, for the first time, the isolated and definitive effect of O2(a1Δg) on

flame propagation. The addition of NO to the plasma afterglow allowed for an order of magnitude increase in the

O2(a1Δg) concentration at a given residence time by removing the quenching species of O3 and O. The NO was

extremely effective because of the catalytic cycle that removes O3 and O, and because it has negligible effect on the

O2(a1Δg) concentration or flame speed. The O2(a

1Δg) was produced in concentrations of over 5000 ppm to enhance

flame propagation of C2H4 lifted flames by several percent at 3.61 kPa and 6.73 kPa.

Numerical simulations using published collisional and reactive quenching reactions have shown that there is a

significant discrepancy in enhancement compared with the trends found in the experiments. The pathways of

enhancement found in the simulations showed that the branching reaction of O2(a1Δg) with H provided O and OH

early in the reaction zone to allow for chemical heat release and flame speed enhancement. The lack of temperature

dependent quenching rate data for O2(a1Δg) by hydrocarbon species was suspect for the discrepancy. Estimations of

the temperature dependent collisional quenching rate of O2(a1Δg) by C2H4 has shown good agreement with the

experimental results trends, but the suggested rates seem unreasonable.

Flow reactor experiments were performed to improve our understanding of O2(a1Δg) reactions within the C2H4

system. From the results it appears that there are two different reaction pathways; one of breaking the double carbon

bond of C2H4 to produce CH2O and the other by OH production to feed into a mechanism similar to NO sensitized

oxidation. Both oxidation pathways produce the CH2O and CO found in the flow reactor experiments and therefore

further investigations are needed to quantify the extent of each contribution to the reactive system.

Therefore, a combination of both the collisional and reactive quenching rates of O2(a1Δg) with hydrocarbon

species are required in order to accurately predict combustion enhancement. More detailed flow reactor experiments

are needed to quantify the concentration of all species present in the system at specific residence times in order to

establish collisional quenching rates, as well as the product channels and reaction pathways involved in combustion

enhancement by O2(a1Δg).

Acknowledgments

This work was supported by the Air Force Office of Scientific Research (award number FA9550-07-1-0136)

under program manager Dr. Julian Tishkoff and Dr. Michael Berman. Part of the research was performed while the

corresponding author held a National Research Council Research Associateship Award at the United States Air

Force Research Laboratory, Wright-Patterson Air Force Base. The authors wish to thank Mr. Francis Haas for the

insightful discussions.

References

1Kimura, I., Aoki, H., and Kato M., “The Use of a Plasma Jet for Flame Stabilization and Promotion of Combustion in

Supersonic Flows,” Combustion and Flame, Vol. 42, 1981, pp. 297-305. 2Takita, K., Moriwaki, A., Kitagawa, T., and Masuya, G., “Ignition and Flame-Holding of H2 and CH4 in High Temperature

Airflow by a Plasma Torch,” Combustion and Flame, Vol. 132, No. 4, 2003, pp. 679-689. 3Wagner, T., O'Brien, W., and Northam, G. E., “Plasma Torch Igniters for Scramjets,” Journal of Propulsion and Power,

Vol. 5, No. 5, 1989, pp. 548-554. 4Esakov, I., Grachev, L., Khodatev, K., Vinogradov, V., and Van Wie, D., “Efficiency of Propane-Air Mixture Combustion

Assisted by Deeply Undercritical MW Discharge in Cold High-Speed Airflow,” 44th AIAA Aerospace Sciences Meeting and

Exhibit (2006) AIAA-2006-1212. 5Ju, Y., Macheret, S., Miles, and R., Sullivan, D., “Numerical Study of the Effect of Microwave Discharge on the Premixed

Methane-Air Flame,” 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit (2004) AIAA-2004-2721. 6Zaidi, S., Macheret, S., Vasilyak, L., Miles, R., Ju, Y., and Sullivan, D., “Increased Speed of Premixed Laminar Flames in a

Microwave Resonator,” 35th AIAA Plasmadynamics and Lasers Conference (2004) AIAA-2004-3707. 7Ombrello, T. and Ju, Y., “Kinetic Ignition Enhancement of H2 Versus Fuel-Blended Air Diffusion Flames Using

Nonequilibrium Plasma,” IEEE Transactions on Plasma Science, Vol. 36, No. 6, 2008, pp. 2924-2932. 8Ombrello, T., Ju, Y., and Fridman, A., “Kinetic Ignition Enhancement of Diffusion Flames by Nonequilibrium Magnetic

Gliding Arc Plasma,” AIAA Journal, Vol. 46, No. 10, 2008, pp. 2424-2433. 9Ombrello, T., Qin, X., Ju, Y., Gutsol, A., Fridman, A., and Carter, C., “Combustion Enhancement via Stabilized Piecewise

Nonequilibrium Gliding Arc Plasma Discharge,” AIAA Journal, Vol. 44, No. 1, 2006, pp. 142-150. 10Bozhenkov, S., Starikovskaia, S., and Starikovskii, A. Y., “Nanosecond Gas Discharge Ignition of H2 and CH4 Containing

Mixtures,” Combustion and Flame, Vol. 133, 2003, pp. 133-146.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

15

11Ombrello, T., Won. S. H., Ju, Y., and Williams, S., “Lifted Flame Speed Enhancement by Plasma Excitation of Oxygen,”

47th AIAA Aerospace Sciences Meeting and Exhibit (2009) AIAA-2009-689. 12Starikovskii, A., “Plasma Supported Combustion,” Proceedings of the Combustion Institute, Vol. 30, 2005, pp. 2405-2417. 13Lou, G., Bao, A., Nishihara, M., Keshav, S., Utkin, Y., Rich, J., Lempert, W., and Adamovich, I., “Ignition of Premixed

Hydrocarbon-Air Flows by Repetitively Pulsed, Nanosecond Pulse Duration Plasma,” Proceedings of the Combustion Institute,

Vol. 31, 2007, pp. 3327-3334. 14Pancheshnyi, S., Lacoste, D., Bourbon, A., and Laux, C., “Ignition of Propane-Air Mixtures by a Repetitively Pulsed

Nanosecond Discharge,” IEEE Transactions on Plasma Science, Vol. 34, No. 6, 2006, pp. 2478-2487. 15Cathey, C., Cain, J., Wang, H., Gunderson, M., Carter, C., and Ryan, M., “OH Production by Transient Plasma and

Mechanism of Flame Ignition and Propagation in Quiescent Methane–Air Mixtures,” Combustion and Flame, Vol. 154, 2008, pp.

715-727. 16Kosarev, I., Aleksandrov, N., Kindysheva, S., Starikovskaia, S., and Starikovskii, A., “Kinetics of Ignition of Saturated

Hydrocarbons by Nonequilibrium Plasma: CH4-Containing Mixtures,” Journal of Propulsion and Power, Vol. 24, No. 6, 2008,

pp. 1182-1197. 17Starikovskaia, S., “Plasma Assisted Ignition and Combustion,” Journal of Physics D: Applied Physics, Vol. 39, 2006, R265-

R299. 18Uddi, M., Jiang, N., Mintusov, E., Adamovich, I., and Lempert, W., “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. 19Stockman, E. S., Zaidi, S. H., Miles, R. B., Carter, C. D., Ryan, M. D., “Measurements of Combustion Properties in a

Microwave Enhanced Flame,” Combustion and Flame, Vol. 156, No. 7, 2009, pp. 1453-1461. 20Ombrello, T., Won, S. H., Ju, Y., and Williams, S., “Flame Propagation Enhancement by Plasma Excitation of Oxygen Part

I: Effects of O3,” Combustion and Flame, submitted for review, 2009. 21Starik, A. M. and Titova, N. S., “Kinetics of Detonation Initiation in the Supersonic Flow of the H2+O2 (Air) Mixture in O2

Molecule Excitation by Resonance Laser Radiation,” Kinetics and Catalysis, Vol. 44, No. 1, 2003, pp. 28-39. 22Starik, A. M., Titova, N. S., Bezgin, L. V., Kopchenov, V. I., and Naumov, V. V., “Control of Combustion by Generation

of Singlet Oxygen Molecules in Electrical Discharge,” Czechoslovak Journal of Physics, Vol. 56, 2006, pp. 1357-1363. 23Popov, N.A., “The Effect of Non-Equilibrium Excitation on the Ignition of Hydrogen-Oxygen Mixtures,” High

Temperature, Vol. 45, No. 2, 2007, pp. 261-279. 24Starik, A., Kuleshov, P., and Titova, N., “Initiation of Combustion of a Hydrogen–Air Mixture with Ozone Impurity by UV

Laser Radiation,” Technical Physics, Vol. 53, No. 2, 2008, pp. 235-243. 25Kozlov, V., Starik, A., and Titova, N., “Enhancement of Combustion of a Hydrogen–Air Mixture by Excitation of O2

Molecules to the a1Δg State,” Combustion Explosions and Shock Waves, Vol. 44, No. 4, 2008, pp. 371-379. 26Bourig, A., Thevenin, D., Martin, J., Janiga, G., and Zahringer, K., “Numerical Modeling of H2–O2 Flames Involving

Electronically-Excited Species O2(a1Δg), O(1D) and OH(2Σ+),” Proceedings of the Combustion Institute, Vol. 32, 2009, pp. 3171-

3179. 27Smekhov, G., Ibraguimova, L., Karkach, S., Skrebkov, O., and Shatalov, O., “Numerical Simulation of Ignition of a

Hydrogen–Oxygen Mixture in View of Electronically Excited Components,” High Temperature, Vol. 45, No. 3, 2007, pp. 395-

407. 28Skrebkov, O. and Karkach, S., “Vibrational Nonequilibrium and Electronic Excitation in the Reaction of Hydrogen with

Oxygen behind a Shock Wave,” Kinetics and Catalysis, Vol. 48, No. 3, 2007, pp. 367-375. 29Smirnov, V.V., Stelmakh, O.M., Fabelinsky, V.I., Kozlov, D.N., Starik, A.M., and Titova, N.S., “On the Influence of

Electronically Excited Oxygen Molecules on Combustion of Hydrogen–Oxygen Mixture,” Journal of Physics D: Applied

Physics, Vol. 41, No. 19, 2008, paper # 192008. 30Starik, A., Lukhovitskii, B., Naumov, V., and Titova, N., “On Combustion Enhancement Mechanisms in the Case of

Electrical-Discharge-Excited Oxygen Molecules,” Technical Physics, Vol. 52, 2007, pp. 1281-1290. 31Starik, A. and Titova, N., “Low-Temperature Initiation of the Detonation Combustion of Gas Mixtures in a Supersonic

Flow Under Excitation of the O2(a1Δg) State of Molecular Oxygen,” Doklady Physics, Vol. 46, 2001, pp. 627-632.

32Wang, H. and Laskin, A., Internal Report, 1998. 33NIST Chemical Kinetics Database on the Web, Standard Reference Database 17, Version 7.0 (Web Version), Release 1.4.2,

http://www.kinetics.nist.gov/kinetics/, last accessed on 28 November 2008. 34Ibraguimova, L., Smekhov, G., and Shatalov, O., “Recommended Rate Constants of Chemical Reactions in an H2-O2 Gas

Mixture with Electronically Excited Species O2(1Δ), O(1D), OH(2Σ) Involved,” Institute of Mechanics of Lomonosov, Moscow

State University (2003). 35Ackerman, R. A., Pitts, J. N., and Steer, R. P., “Singlet Oxygen in the Environmental Sciences. VIII. Absolute Rates of

Deactivation of O2(a1Δg) by Terminal Olefins, Tetramethylethylene, and Methyl Chloride,” Journal of Chemical Physics, Vol.

52, 1970, pp. 1603-1604. 36Mayer, S. W. and Schieler, L., “Activation Energies and Rate Constants Computed for Reactions of Oxygen with

Hydrocarbons,” Journal of Physical Chemistry, Vol. 72, 1968, pp. 2628-2631. 37Becker, K. H., Groth, W., and Schurath, U. “The Quenching of Metastable O2(

1Δg) and O2(1Σg

+) Molecules,” Chemical

Physics Letters, Vol. 8, 1971, pp. 259-262.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86

American Institute of Aeronautics and Astronautics

16

38Borrell. P. and Richards, D. S., “Quenching of Singlet Molecular Oxygen, O2(a1Δg) and O2(b

1Σg), by H2, D2, HCl and HBr,”

Journal of the Chemistry Society, Faraday Transactions 2, Vol. 85, No. 9, 1989, pp. 1401-1411. 39Borrell, P., Borrell, P. M., Richards, D. S., and Boodaghians, R. B., “High Temperature Quenching Reactions of Singlet

Molecular Oxygen,” Journal of Photochemistry, Vol. 25, 1984, pp. 399-407. 40Plötz, J. and Maier, M., “Temperature Dependence of the Relaxation Rate Constants of the a1Δg and b1Σg

+ States of Oxygen

Isotopes,” Chemical Physics Letters, Vol. 138, No. 5, 1987, pp. 419-424. 41Hotokka, M., Roos, B., and Siegbahn, P., “CASSCF Study of Reaction of Singlet Molecular Oxygen with Ethylene.

Reaction Paths with C2(v) and Cs Symmetries,” Journal of the American Chemical Society, Vol. 105, No. 16, 1983, pp. 5263-

5269. 42Yoshioka, Y., Tsunesada, T., Yamaguchi, K., and Saito. I., “CASSCF, MP2, and CASMP2 Studies on Addition Reaction of

Singlet Molecular Oxygen to Ethylene Molecule,” International Journal of Quantum Chemistry, Vol. 65, 1997, pp. 787-801. 43Doughty, A., Barnes, F. J., Bromly, J. H., and Haynes, B. S., “The Mutually Sensitized Oxidation of Ethylene and NO: An

Experimental and Kinetic Modeling Study,” Proceedings of the Combustion Institute, Vol. 26, 1996, pp. 589-596. 44Hori, M., Matsunaga, N., Marinov, N., Pitz, W., and Westbrook, C., “An Experimental and Kinetic Calculation of the

Promotion Effect of Hydrocarbons on the NO-NO2 Conversion in a Flow Reactor,” Proceedings of the Combustion Institute, Vol.

26, 1998, pp. 389-396. 45Dagaut, P., Mathieu, O., Nicolle, A., and Dayma, G., “Experimental Study and Detailed Kinetic Modeling of the Mutual

Sensitization of the Oxidation of Nitric Oxide, Ethylene and Ethane,” Combustion Science and Technology, Vol. 177, 2005, pp.

1767-1791.

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n O

ctob

er 1

3, 2

012

| http

://ar

c.ai

aa.o

rg |

DO

I: 1

0.25

14/6

.201

0-15

86