LASER DIAGNOSTICS OF PLASMA ASSISTED COMBUSTION
By
Xing Rao

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Doctor of Philosophy
Mechanical Engineering
2010

ABSTRACT
LASER DIAGNOSTICS OF PLASMA ASSISTED COMBUSTION
By
Xing Rao
In this study, a microwave re-entrant cavity discharge system and a direct current
(DC) plasmatron are used to investigate flame enhancement and nitric oxide (NO)
formation using laser and optical diagnostics. The uniqueness of this study lies in the
direct coupling concept, a novel highly efficient strategy used here for the first time.
To investigate combustion dynamics of direct microwave coupled combustion, an
atmospheric high-Q re-entrant cavity applicator is used to couple microwave (2.45 GHz)
electromagnetic energy directly into the reaction zone of a premixed laminar
methane-oxygen flame using a compact torch. When microwave energy increases, a
transition from electric field enhancement to microwave plasma discharge is observed. At
6 to 10 Watts, ionization and eventually break-down occurs. 2-D laser induced
fluorescence (LIF) imaging of hydroxyl radicals (OH) and carbon monoxide (CO) is
conducted in the reaction zone over this transition, as well as spectrally resolved flame
emission measurements. These measurements serve to monitor excited state species and
derive rotational temperatures using OH chemiluminescence for a range of equivalence
ratios (both rich and lean) and total flow rates.
Combustion dynamics is also investigated for plasma enhanced methane-air flames

in premixed and nonpremixed configurations using a transient arc DC plasmatron.
Results for OH and CO PLIF also indicate the differences in stability mechanism, and
energy consumption for premixed and nonpremixed modes. It is shown that both
configurations are significantly influenced by in-situ fuel reforming at higher plasma
powers.
Parametric studies are conducted in a plasma assisted methane/air premixed flame
for quantitative NO production using a DC plasmatron with PLIF imaging. Quantitative
measurements of NO are reported as a function of gas flow rate (20 to 50 SCFH), plasma
power (100 to 900 mA, 150 to 750 W) and equivalence ratio (0.7 to 1.3). NO PLIF images
and single point NO concentrations are presented for both plasma discharge only and for
methane/air plasma enhanced combustion cases. NO formation occurs predominantly
through N2(v)+O NO+N for the pure plasma discharge without combustion. The NO
concentration for the plasma enhanced combustion case (500 to 3500 ppm) was an order
of magnitude smaller than the pure plasma discharge (8000 to 15000 ppm) due to the
reduction of nitrogen break up from plasma reactions by the methane. Experiments show
the linear decay of NO between the equivalence ratio range 0.8 to 1.2 under the same
flow condition and discharge current.
The diagnostic methods performed in this study include: (1) species concentration
by laser induced fluorescence (NO, OH and CO); (2) rotational temperature by optical
emission spectroscopy; (3) excited species emission by optical emission spectroscopy; (4)
IR thermometry; and (5) multi-line NO PLIF thermometry.

© Copyright by Xing Rao 2010
All Rights Reserved

To my wife, Shujian Shi
For her dedication and support

v

ACKNOWLEDGMENTS

First and foremost, I would like to thank my adviser, Dr. Tonghun Lee, for his
generous support during my Ph.D. study, and for his guidance and advising of my
research work. Without his wisdom and drive, this program would have been
significantly less. You will be my life-time mentor and advisor.
I want to express my sincere gratitude to Dr.Jes Asmussen, Dr. Timothy Grotjohn,
Dr. Indrek Wichman and Dr. Guoming Zhu for their helpful and insightful suggestions
throughout this program.
I would like to thank my colleagues, Casey Allen, Steve Hammack, Zach
Williams, Kadek Hemawan, Xiaojian Yang, Zhen Ren, Xuefei Chen and Xin Wang for
their support and assistance with the experiments. I am also grateful to all the staffs and
students in Energy and Automobile Research Lab, especially Gary and Jeff from the
machine shop were of great help. Special thanks to Dr. Harold Schrock for the nice
building and equipments.
I also would like to acknowledge the sponsorship of the U.S. Air Force Office of
Scientific Research with Dr. Julian Tishkoff as Program Manager, and Michigan State
University. I also want to thank Dr. Campbell Carter from WPAFB for his help with
experiments and guidance in publication writing.

vi

Finally, I am sincerely grateful to my parents and my sister Dan for their
understanding and encouragement for my whole life. I would also like to convey my
heartfelt gratitude to my wife, Shujian Shi, for her love, support and confidence in me.

vii

TABLE OF CONTENTS
LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES .......................................................................................................................... xi
LIST OF SYMBOLS...................................................................................................................... xvi
LIST OF ABBREVIATIONS ...................................................................................................... xvii
Chapter 1, Introduction ..................................................................................................................... 1
1.1 Background and Motivation .................................................................................. 1
1.2 Recent Progress on Plasma Assisted Combustion ................................................. 3
1.3 Scope of Research................................................................................................ 16
Chapter 2, Basic Knowledge of Spectroscopy ........................................................................... 19
2.1 Introduction to Laser Induced Fluorescence (LIF) .............................................. 19
2.2 NO LIF: Literature Survey................................................................................... 29
Chapter 3, Experimental Facility .................................................................................................. 33
3.1 Microwave Plasma System .................................................................................. 33
3.2 Direct Current Plasma System ............................................................................. 43
3.3 Laser and Optical Diagnostics Setup ................................................................... 48
Chapter 4, Combustion Dynamics with Microwave Discharge ............................................. 55
4.1 Impact on Flame Structure and Radical Concentration ....................................... 55
4.2 Flammability Limits and Flame Speed ................................................................ 62
4.3 Emission Measurements and Temperature .......................................................... 64
4.4 In-Situ Fuel Reforming ........................................................................................ 69
Chapter 5, Combustion Dynamics in a DC Plasmatron ........................................................... 73
5.1 Premixed and Non-Premixed Modes ................................................................... 73
5.2 Impact on Flame Structure................................................................................... 75
5.3 Impact on Radical Concentration......................................................................... 80
5.4 Flame Stability, CO Formation and Energy Efficiency ....................................... 84
Chapter 6, Nitric Oxide Formation in a DC Plasmatron .......................................................... 91
6.1 Nitric Oxide Formation in Flames ....................................................................... 91
6.2 Quantification of Nitric Oxide ............................................................................. 94
6.3 Temperature Profile Measurements ..................................................................... 96

viii

6.4 Planar Laser-Induced-Fluorescence (PLIF) Imaging of NO.............................. 100
6.5 Temperatures and NO Concentrations in Plasma Discharge Only..................... 102
6.6 Temperatures and NO Concentrations in a Plasma Enhanced Flame ................ 109
Chapter 7, Conclusion and Future Work ....................................................................................117
7.1 Microwave Discharge Results ............................................................................117
7.2 DC Discharge Results .........................................................................................119
7.3 Nitric Oxide Measurements Results .................................................................. 120
7.4 Some Suggestions for Future Work ................................................................... 121
Bibliography.....................................................................................................................................123

ix

LIST OF TABLES

Table 4.1: Transitions for OH rotational Temperature Method .............................................. 67
Table 5.1: Thermocouple Measurements and Calibrations ..................................................... 79
Table 6.1: Reactions for NO formation in methane/air flame in plasma .............................. 91
Table 6.2. Energy data of discharge only in air with different current.................................106
Table 6.3. NO Concentration in Air over in PAC ..............................................................................109

x

LIST OF FIGURES
Figure 1.1 Effect of different types of discharge on the maximum co-flow speed
(blow-off velocity) [8].............................................................................................. 5
Figure 1.2 Discharge-enhanced flame stabilization. (a) Without discharge. (b) With
discharge. The flame is lifted off in (a) and reattached by discharges in image (b).
The electrode position is 10 diameters (46 mm) above the nozzle. [8].................... 6
Figure 1. 3 Comparison of OH (A-X) emissions with and without discharge [30]............. 6
Figure 1.4 Schematic of the plasma-assisted combustion experiment in a supersonic flow
[5]. ............................................................................................................................ 7
Figure 1.5 Photograph of the discharge in air (top) and of the discharge/flame region in
ethylene-air (bottom) [4]. ......................................................................................... 8
Figure 1.6 Experimental and computational results of ignition temperatures with 20%
Methane in Nitrogen with preheated air in a MGA (I = 350 mA) [6]....................... 9
Figure 1.7 Lifetimes of plasma produced species as a function of pressure. [7].............. 10
Figure 1.8 Lifted flame speed enhancements by addition of O3 at 1 atmospheric pressure
[6]. ...........................................................................................................................11
Figure 1.9 Images of OH-number density in equivalence ratio of 0.76 CH4/air flame
without (a) and with (b) microwave radiation.[12]. ............................................... 12
Figure 1.10 Schematic of gliding arc and the electric components (Left). Photo image of
GA discharge (Right) [32]. ..................................................................................... 13
Figure 1.11 ICCD images of plasma discharge development (camera gate time is 200 ps.
One atmospheric pressure, U = 20 kV) [33]. .......................................................... 14
Figure 1.12 Flow temperatures (Left Y axis) and unreacted fuel fraction (Right Y axis)
downstream of the discharge vs. equivalence ratio in flow mixed with methane and
air. Pressure is 100 torr with no flame detected in the test section. [23] ............... 14

xi

Figure 1.13 Images of plasma combustion, 70 torr total pressure in air (left) and C2H4/
air at equivalence ratio of 0.9 (right). [23] ............................................................. 15
Figure 1.14 Percentage of fuel left which is unreacted in the flow in a transverse RF
discharge with shaded areas indicating conditions when flame occurs. [34]......... 15
Figure 1.15 Discharge temperature (a) and unreacted fuel fraction (b) as functions of flow
velocity. Ethylene/air, P=70 torr, φ=1.0, pulser rep rate 50 kHz. Closed symbols
indicate ignition is achieved [35]. .......................................................................... 16
Figure 2.1 Two-level system with pertinent energy transfer processes.. .......................... 22
Figure 2. 2 Transition from linear regime to saturation regime in the non-transient
two-level model in a 1 bar, equivalence ratio of 0.9 methane/air flame. Difference
for four small figures are different laser intensity scale [44].................................. 27
Figure 3.1: Picture of coaxial re-entrant cavity applicator with three fine adjustments
using unislides; picture of methane/oxygen flame shown in the inset...................... 34
Figure 3.2 Scheme of the structure of coaxial re-entrant cavity applicator with burner
centered in the middle............................................................................................... 35
Figure 3.3: Equivalent circuit for coaxial re-entrant cavity applicator............................. 39
Figure 3.4: Numerical simulation of electric field from (V/m) COMSOL MultiPhysics RF
Electromagnetic solver for current experiments ....................................................... 40
Figure 3.5 Components of microwave cavity plasma assisted flame system with E-field
pattern shown in the figure [53]................................................................................ 42
Figure 3.6 Flame extinction limits (lean) with 10 Watts, and 20–100 Watts microwave
power. [53]. ............................................................................................................... 43
Figure 3.7 Image and schematic of transient arc plasmatron with gas/liquid feeding (a);
Air plasma imaging at 20 SCFH and 300 mA (b); Plasma assisted stoichiometric
methane/air imaging at 20 SCFH and 300 mA (c). ................................................... 44
Figure 3.8 Schematic of transient arc plasmatron with gas/liquid feeding....................... 45
Figure 3.9 Typical DC discharge gap voltage and current passing through the gap [54]. 47

xii

Figure 3.10 Voltage and current plots for discharge in air/propane mixture and CCD
images of the discharge with exposure time of Δt shown in the voltage-current plot
[54]............................................................................................................................ 48
Figure 3.11 Overall experimental setup with synchronization of lasers, energy monitor,
ICCD camera and control system. ............................................................................ 49
Figure 3.12 Excitation Spectra of OH (A-X) with Q1(8) transition shown at 283.029 nm.
................................................................................................................................... 50
Figure 3.13 Excitation Spectra of CO (B-X) with highest intensity at 230.109 nm.......... 51
Figure 3. 14 . Spectrally resolved 1-D line imaging (inset) and the emission spectra of
NO-LIF with 226.03 nm excitation and selected filter function (solid red line). The
emission spectra has contributions from NO, O2 and Rayleigh scattering. .............. 53
Figure 4.1 Emission (top), OH PLIF (middle) and CO PLIF (bottom) intensity images of
CH4/O2 flame vs. microwave power with three stages noted (100 SCCM total
flowrate at φ=1.1)...................................................................................................... 55
Figure 4.2 OH number density in burnt gas of CH4/O2 flame vs. microwave powers from
averaged PLIF images. ............................................................................................. 58
Figure 4.3 OH number density vs. equivalence ratios (φ=0.8 to 1.2) with different
microwave powers (0 to 30 Watts). Total flow rate is 100 SCCM. ........................... 62
Figure 4.4 Flammability limits for methane/oxygen flame (lean and rich) with flamespeed
increase demonstrated by flame photographs with no microwave radiation and 10
Watts microwave radiation (right top). ..................................................................... 63
Figure 4.5 Emission spectra from CH4/O2 flame with no microwave radiation and 10
Watts microwave radiation (top) and OH rotational temperature at the same
conditions as in Figure 4.2 (bottom). ........................................................................ 65
Figure 4.6 Sample Image to show how to get OH rotational temperature at condition of
100 SCCM total flow rate and equivalence ration of 1.1. The OH rotational
temperature using this method is 2,900 K................................................................. 68
Figure 4.7 CO number density in burnt gas of CH4/O2 flame vs. microwave powers at the
same conditions as in Figure 4.2 using CO Two-Photon Laser Induced Fluorescence.
................................................................................................................................... 71

xiii

Figure 5.1: Schematic of transient arc plasmatron with cathode in the center and two
different inlets are used to get two different experimental setup(left); Images for
premixed (100 mA and 400 mA DC current) and nonpremixed (0 and 400 mA DC
current) flames of methane/air with equivalence ratio of 0.9................................. 73
Figure 5.2: Single and averaged 2-D OH Planar Laser-Induced Fluorescence (PLIF)
images as a function of increasing microwave power for both premixed flames (top)
and nonpremixed flames (bottom). ........................................................................ 76
Figure 5.3: OH number density is plotted for different DC currents, total flow rates (top,
equivalence ratio is fixed at 0.95 and for different DC current, equivalence ratios
(bottom, total flowrate is fixed at 15 LPM) in premixed mode.............................. 82
Figure 5.4: OH number density with the same flow conditions (total flow rate is 9 SLPM,
and φ=0.95) but different currents for both premixed (from data point the same as
in Figure 5.3) and nonpremixed (two data points, one outer flame front and the
other one inner flame front as shown) flames. ....................................................... 84
Figure 5.5: Averaged OH PLF images for all flowrates at premixed mode with phi= 0.7,
1.0 and 1.3 (top); Averaged OH PLF images for 6 LPM (middle)/15 LPM
(bottom) at nonpremixed mode with hi= 0.7, 1.0 and 1.3...................................... 85
Figure 5.6: Averaged CO PLF images for premixed (insert top) and nonpremixed (insert
bottom); CO number density is plotted for with equivalence ratio 0.7 (premixed
and nonpremixed) and 1.1 (nonpremixed). ............................................................ 88
Figure 5.7: DC Power input with different current for air plasma, premixed flames and
nonpremixed flames. .............................................................................................. 90
Figure 6. 1 NOx emission as f function of combustion temperature [30, 84].................... 93
Figure 6.2 NO addition method for determination of the baseline strength and the nascent
NO concentration. ...................................................................................................... 96
Figure 6.3 IR temperature profile with different equivalence ratio for premixed methane
and air flame with circle shown in the figure where the multi-line thermometry
calibration is used. ..................................................................................................... 97
Figure 6. 4 Multi-line NO LIF thermometry (126 transitions between 225.975 nm and
226.1 nm). .................................................................................................................. 98

xiv

Figure 6.5 Single shot NO concentration fields using NO PLIF of gas discharge of air and
plasma enhanced flames with different plasma power at same gas flow rate (a).
Averaged NO concentration fields from NO PLIF (200 images averaged) of plasma
enhanced flames with different plasma power at same gas flow rate (b). ............... 101
Figure 6.6 Temperature (a) and NO concentration (b) data at specific location of the
discharge for different power levels and air flow rates in discharge only case in air.
................................................................................................................................. 103
Figure 6.7 Temperature (a) and NO concentration (b) data as function of current and flow
rate when the equivalence ratio is held at 1 in plasma assisted methane/air flame. .111
Figure 6.8 Temperature (a) and NO concentration (b) data as function of current and
equivalence ratio when flow rate the same in plasma with methane/air combustion..
..................................................................................................................................113
Figure 6.9 Temperature (a) and NO concentration (b) data with equivalence ratio and flow
rate when the current is the same in plasma assisted methane/air combustion. .......114

xv

LIST OF SYMBOLS

Te

Electron Temperature

Tg

Gas Temperature

T

Temperature

K

Boltzmann Constant

N NO

Number Density of NO

X NO

Mole Fraction of NO

P

Pressure

ILIF

LIF Intensity

Cp

Specific Heat

Rf

Radio Frequency

λ

Wavelength

ω

Angular Frequency

U

Voltage

I

Current

xvi

LIST OF ABBREBIATIONS

PAC

Plasma Assisted Combustion

NO

Nitric Oxide

CO

Carbon Monoxide

LIF

Laser Induced Fluorescence

PLIF

Planar Laser Induced Fluorescence

TALIF

Two Photon Absorption LIF

SCCM

Standard Cubic Centimeter per Minute

SCFH

Standard Cubic Foot per Hour

ppm

particle per million

DC

direct current

AC

alternative current

mA

mili-ampere

UV

UltraViolet

IR

Infrared

eV

electron volt

TEM

Transverse Electromagnetic

CCD

Charge Coupled Device

ICCD

Intensified Charge Coupled Device

NOX

Nitrogen Oxide

xvii

SECD

Single Electrode Corona Discharge

USRD

Ultra-short Repetitive Discharge

DBD

Dielectric Barrier Discharge

MGA

Magnetic Gliding Arc

GA

Gliding Arc

PIV

Particle Image Velocimetry

E.U.

European Union

xviii

Chapter 1 Introduction
1.1 Background and Motivation
Plasma assisted combustion has emerged as an important research topic in recent
decades as a potential method to improve combustion efficiency in energy conversion
and propulsion systems while simultaneously reducing the pollution emissions from
advanced gas turbines and aircraft engines. The term ‘plasma’ is described by Bittencourt
as “a wide variety of macroscopically neutral substances containing many interacting free
electrons and ionized atoms or molecules, which exhibit collective behavior due to
8

long-range Coulomb forces.” [1]. The electron density in a typical flame is close to 10
-3

cm

11-12

while for plasma assisted combustion cases, it rises to 10

-3

cm . The process of

combining a plasma discharge with combustion is a sophisticated one, requiring
knowledge of physics, chemistry, engineering (electric and mechanical) et al. There are
still unknown subject areas in exact discharge coupling processes and chemical kinetics
simulations. The prospect of gaining insight into the systematic integration of
electrodynamics and combustion chemistry is a challenge, which this dissertation
explores.
The use of electromagnetic energy and nonequilibrium discharge in particular for
enhancing ignition and combustion stability is receiving great attention due to the
potential for enabling highly efficient thermal energy conversion, reliable ignition in

1

harsh operating conditions, flammability limit extension, reaction with alternative fuels of
different burning characteristics [2, 3], and for the development of next generation
propulsion systems using hypersonic flow [4, 5]. From a future application perspective, a
light-weight power supply and a discharge unit be integrated into a practical combustion
system, and the generated nonequilibrium discharge must target energy transfer to
efficiently boost the combustion chemistry.
The exact mechanisms for flame enhancement by plasma (either thermal or
nonthermal) are being actively studied. A number of studies have provided a glimpse into
various aspects of this process [6-12]. In the literature, the electromagnetic energy alters
the reaction chemistry by: (1) increased flame temperature by ohmic heating; (2)
radiation-induced excitation of electrons; (3) decomposition and breakup of the fuel
molecule to smaller hydrocarbon molecules, including in-situ fuel reforming; and (4)
creation of radicals via collision with electrons creating an increase in electrons, ions and
excited state species. Dr. Starikovskii has written an extensive review of the progress on
plasma assisted ignition and combustion in [2].
Another aspect of evaluating a combustion system is the level of emissions. A
practical system requires low or even zero emissions of pollutants while coupling high
temperature reactive flows (combustion) with a plasma discharge, especially for high
power,. This can lead to significant changes in the formation of NOx (95% is nitric oxide).
Nitric Oxide is one of the most hazardous combustion products and constitutes a critical
design parameter for practical combustion systems using air as the oxidizer. Generally,
2

NO in conventional combustion forms from the following four main path ways [13, 14]:
(1) thermal (Zeldovich) mechanism; (2) N2O mechanism; (3) prompt NO mechanisms;
and (4) fuel-bound nitrogen mechanism (This will be discussed in detail later in this
dissertation). Aside from formation of NO, reduction of NO in combustion can be
achieved by reacting with excess hydrocarbon radicals in a rich flame through a ‘NO
reburn’ process [15]. In a rich flame, small excess hydrocarbon molecules can generate a
large reduction of NO. In the case of plasma assisted combustion, atomic and molecular
excited states and radicals reactions also play a critical role in the formation and
reduction of NO [16, 17].

1.2 Recent Progress on Plasma Assisted Combustion
Early work on plasma interaction with combustion focused mainly on interactions
involving combustion and electric fields, not combustion with an actual discharge.
Combustion with a direct current (DC) electric field can lead up to 100% in the blowout
limit [18], which may came from an “ionic wind” between the “chemi-ionized” species
and the applied electric field in the flame reaction zone. A more significant increase in
the blowout limit was observed in [19] using 3 kV DC voltage. Investigations were also
made using alternating current (AC) [20] and the enhancements were marginal compared
to the results using DC power.
In recent decades (especially the past 10 years), several research groups (US, Russia,
EU , Japan et al) have made intense efforts to exam plasma enhanced combustion from

3

different angles using various plasma discharge systems. The exact mechanism for flame
enhancement using a discharge is still unclear and is being actively studied, but many
studies have provided insights into different aspects of this complicated process [6-12].
Different kinds of plasma discharges, including thermal plasma [21], nanosecond pulsed
discharge [22], dielectric barrier discharge [8], pulsed corona discharge [23], RF
discharge [24],

gliding arc [25], DC or low frequency AC discharges [26], plasmatron

[27] and microwave discharge [28] have been investigated for enhancing high
temperature thermal oxidation.
For this study, I will summarize some of the published results from various groups.
Please note that these only comprise a fraction of the researches on plasma assisted
combustion.
Cappelli et al. at Stanford University, investigated three different types of
nonequilibrium plasmas used for comparison of flame stabilization in a lifted jet diffusion
flame of methane-air: single electrode corona discharge (SECD); dielectric barrier
discharge (DBD); and ultra-short repetitively pulsed discharge (USRD, pulse duration of
about 10 ns) [8, 29]. It was shown that the flame can be stabilized using SECD discharge
at a 20% higher co-flow speed than without SECD The blow-off velocity with the DBD
discharge was increased by up to 50%, and the USRD discharge increases the blow-off
velocity by nearly a factor of ten, as shown in Figure 1.1.
It was shown by Kim et al. that significant increases in the critical co-flow velocity
(the velocity at which flame will extinguish) [8] and the rate of NO [29] production in
4

plasma assisted combustion flames were obtained from ethane-air using repetitively
pulsed nanosecond plasmas. The lifted jet diffusion flames were created in a vertical
wind tunnel with a nozzle oriented parallel to the flow direction producing a co-flow jet.
The lifted-off distance is seen to reduce from the case without discharge (shown in Figure
1.2(a)) to with the discharge (shown in Figure 1.2(b)), indicating greater flame stability in

Jet Velocity

the case with plasma discharge.

0

2

4

6 8 10 12 14 16 18 20 22
Normalized coflow speed

Figure 1.1 Effect of different types of discharge on the maximum co-flow speed
(blow-off velocity) [8]. For interpretation of the references to color in this and all other
figures, the reader is referred to the electronic version of this dissertation.
Pilla et al. [30] have shown an extension of the flame stability limit in the lean
premixed turbulent propane-air flame (φ =0.3). They compared flame stability using
ICCD images of OH (A X) emissions with and without the nanosecond repetitively
pulsed discharge. This is shown in Figure 1.3.

5

Figure 1.2 Discharge-enhanced flame stabilization. (a) Without discharge. (b) With
discharge. The flame is lifted off in (a) and reattached by discharges in image (b). The
electrode position is 10 diameters (46 mm) above the nozzle. [8]

OH emission
without discharge

OH emission with
discharge 75W

mask

Figure 1. 3 Comparison of OH (A-X) emissions with and without discharge [30].
Leonov et al. [4, 5] at the Institute of High Temperature of the Russian Academy of
Science investigated a DC discharge between separately ballasted pin electrodes using a

6

supersonic air flow to ignite the fuel (a hydrogen and ethylene mixture) which is injected
into the flow in a cavity. Tthe schematic of the experiment is shown in Figure 1.4. The
Mach number of air flow is M=2 with the initial static pressure ranging between 0.2 to
0.8 bars. The fuel injectors installed on the bottom wall of the cavity (the in Figure 1.4).
The discharge is a filamentary DC discharge with input power in range 1-10 kW.

Flow
Area of
Ignition

Electrodes
Ethylene
Figure 1.4 Schematic of plasma-assisted combustion experiment in a supersonic flow
[5].
Experiments show that at discharge power greater than 1 kW, ignition of hydrogen
and ethylene occurs. When the discharge power greater than 3 kW with the equivalence
ratio greater than 1, the flame extended above the cavity as shown in Figure 1.5. In all
these experiments the flow rates of the fuel are high, up to 4 g/sec.

7

Figure 1.5 Photograph of the discharge in air (top) and of the discharge/flame region in
ethylene-air (bottom) [4].
At Princeton University, Ju et al. tested a non-equilibrium Magnetic Gliding Arc
(MGA) discharge plasma for enhancement of hydrocarbon-air and hydrogen-air
counter-flow flames in diluted nitrogen [6, 31]. They showed that the strain rate at
extinction in methane-air flows was significantly increased with plasma sustained in the
air flow, up to 220 percent at a low discharge power of about 78 Watts. Ju et al. suggested
that this was predominantly from thermal effects. However, it is also mentioned that
excited species generated by the plasma contribute to the enhancement of combustion.
Reduction of the ignition temperature by the plasma was both achieved experimentally
and numerically with 20% Methane in Nitrogen for preheated air in a Magnetic Gliding
Arc as shown in Figure 1.6 [6].

8

Figure 1.6 Experimental and computational results of ignition temperatures with 20%
methane in nitrogen with preheated air in a MGA (I = 350 mA) [6].
Enhancement of the lifted flame speed by plasma excitation of oxygen (including
1

O3 and O2(a Δg) by was studied T. Ombrello and et al. [7]. They isolated two oxygen
containing species produced by plasma and investigated their influence on plasma
enhanced combustion in order to isolate various flame enhancement mechanisms.
Lifetimes of plasma produced species are shown in Figure. 1.7 as a function of pressure.
From the figure, they also obtain some rough pressure ranges for a practical system
(internal combustion engine, gas turbine, etc.). It is clear that stable species (NOX and O3)
have the highest lifetimes whereas electrons and ions have the lowest. The lifetimes of
1

the radicals and some metastables (O, OH, H, O2, a Δg, etc.) and excited species and
1

other metastable species (O( S), O2(v), N2(v), etc.) are in between, with the lifetimes of

9

1

the former being higher. The results show significant kinetic enhancement by O2(a Δg)
and O3 when comparing flame stabilization locations with or without the extra species
generated by the plasma. The enhancement of the lifted flame speed by O3 is shown in
Figure 1.8. With ~500 ppm O3 addition, ~2% enhancement is achieved while with ~1,300
ppm, it is 8%. The experiments were also conducted under different pressure ranges.

Figure 1.7 Lifetimes of plasma produced species as a function of pressure. [7]

10

Figure 1.8 Lifted flame speed enhancements by addition of O3 at 1 atmospheric
pressure [6].
Miles, et al. in Princeton studied microwave induced flame speed enhancement in a
laminar, premixed CH4/air wall stagnation flat flame in a high-Q microwave cavity [12].
When the flame is activated by a continuous wave microwave field of about 5 kV/cm, the
flame front approaches the burner exit and stabilizes at a standoff distance with a flame
speed increase of up to 20%. The temperature was measured using Rayleigh scattering
thermometry, the laminar flame speed was measured using Particle Image Velocimetry
(PIV), and the OH radical concentration was measured by laser induced fluorescence.
Images of OH-LIF signal intensity for an equivalence ratio of 0.76 CH4/air flame without
(a) and with (b) microwave radiation are shown in Figure 1.9. The authors indicate that

11

microwave radiation can be practical means to control and enhance flame stability
non-invasively.

Figure 1.9 Images of OH-number density in equivalence ratio of 0.76 CH4/air flame
without (a) and with (b) microwave radiation.[12].
A gliding arc (GA) discharge system was used by Fridman et al. for flame
enhancement research [32]. The study showed combined advantages of both thermal and
non-thermal discharges for certain applications in optimized GA regimes. This discharge,
as a source of intermediate plasma, is a promising way to modify important plasma
chemical process (a plasma partial oxidation of methane). The authors compared this
kind of discharge with a microwave discharge. The gliding arc discharge has unique
advantages in that it is not restricted by pressure, which has limited the actual application
of the microwave discharge.

12

Figure 1.10 Schematic of gliding arc and the electric components (Left). Photo image of
GA discharge (Right) [32].
Starikovskii et a; also in Drexel University studied plasma-assisted ignition below
the self-ignition threshold for different fuel-air mixtures (including methane, ethane,
propane and butane) [33]. A series of ICCD images is shown in Figure 1.11 with the
plasma developing between two electrodes over a short time period. The ambient
temperature for ignition is between 300 and 800 K which is well below the auto-ignition
temperature for most of the fuels. Laser-induced fluorescence was utilized to measure the
OH concentration after the discharge. The OH profile shows that the OH concentration
persists at a significant level after plasma discharge generation for a long time. The time
increases with increasing temperature.

13

0.4ns

9.0 ns

14.2 ns

Figure 1.11 ICCD images of plasma discharge development (camera gate time is 200
ps. One atmospheric pressure, U = 20 kV) [33].
Lou et al. [23] (shown in Figure 1.12), and Bao et al. [34] (Figure 1.14) investigated
ignition and almost-complete combustion in low temperature oxidation of methane and
ethylene in air at P=70 to 90 torr, at the conditions when the temperature for repetitively
0

pulsed air plasma before fuel was addition ranged from 100 - 200 C. Pictures of the
discharge in air, and in methane/air mixtures, are shown in Figure 1.13 with flow from
left to right. A flame through a rectangular window is observed in the picture on the right.

Figure 1.12 Flow temperatures (left Y axis) and unreacted fuel fraction (right Y axis)
downstream of the discharge versus equivalence ratio in flow mixed with methane and
air. Pressure is 100 torr with no flame detected in the test section. [23]

14

Figure 1.13 Images of plasma combustion, 70 torr total pressure in air (left) and C2H4/
air at equivalence ratio of 0.9 (right). [23]

Figure 1.14 Percentage of fuel left which is unreacted in the flow in a transverse RF
discharge with shaded areas indicating conditions when flame occurs. [34]

15

Temperatures of the air and air/ ethylene flow with respect to percentage un-reacted
flow and flow velocity are shown in Figure 1.15(left) and (right) [35], respectively.
Unburned percentage of fuel indicates the completeness of oxidation using the discharge
for different equivalence ratio for 70 torr and 90 torr with both methane and ethylene.
Please note here under some conditions there is flame while under others there isn’t.

Figure 1.15 Discharge temperature (a) and unreacted fuel fraction (b) as functions of
flow velocity. Ethylene/air, P=70 torr, φ=1.0, pulser rep rate 50 kHz. Closed symbols
indicate ignition is achieved [35].

1.3 Scope of Research
The goal of this dissertation is to examine energy-efficient and effective flame
enhancement with reduced pollutant emissions by a plasma discharge. Two discharge

16

systems are investigated: (a) direct current (DC) plasmatron; and (b) microwave (MW)
re-entrant cavity discharge system. Two combustion configurations, a premixed
methane/oxygen flame and a premixed methane/air flame are also tested and reported.
The scope of this research can be summarized by the following four fundamental
questions:
1.

By how much can plasma enhance combustion? What constitutes efficient kind

of plasma delivery system?
2.

How does a plasma discharge enhance a flame – i.e. what are the roles of

radicals, thermal effects, and non-equilibrium kinetics?
3.

What differences exist between premixed and non-premixed plasma assisted

flames?
4.

What are the NOX production mechanisms and how to reduce NOX production

using high electric power?
The dissertation is organized as follows: Chapter 2 introduces basic knowledge on
laser induced fluorescence and the computational simulation of Nitric Oxide. Chapter 3
describes the experimental facilities and diagnostics setup used. Chapter 4 discusses the
results of combustion dynamics using a microwave cavity applicator here. Chapter 5
presents the results of premixed and nonpremixed combustion in a DC plasmatron.
Chapter 6 presents NO formation and reduction in the DC plasmatron. Finally, Chapter 7

17

summarizes the results of the present work and offers some suggestions for future work
suggestions.

18

Chapter 2 Spectroscopy
2.1 Laser Induced Fluorescence (LIF)
Spectroscopy was originally used to study of the relations between radiation and
matter. Spectroscopy used to be referred to disperse visible light of its corresponding
wavelength, e.g. using a prism. Nowadays spectroscopy is expanded to include
measurements using either wavelength or frequency, thus it can be also referred as the
responses to varying frequency or an alternating field. After the relationship of energy
with wavelength was realized, E = hν, spectroscopy was also related to energy. The
spectrum is a plot with wavelength or more commonly frequency or energy.
Spectrometry is a spectroscopic technique used to investigate a specific species
(atoms, molecules, ions). The instrument used for spectrometry is called a spectrometer,
spectrograph, or spectrophotometer. Spectroscopy (spectrometry) is often used in
physical and analytical chemistry for the quantification of substances through their
absorption or emission spectra. This technology is heavily used in area of astronomy as
most large telescopes have a similar function as spectrometers. In these areas,
spectrometry is not only used to measure the physical and chemical properties of
astronomical targets but also used to measure their velocities using the Doppler shift of
the spectral lines.

19

Laser-Induced Fluorescence (LIF) is a well-established optical technique in many
areas of science. It is especially useful in combustion because it can withstand harsh
conditions and doesn’t disturb the chemical reactions or the flow field. As LIF is both
species and quantum state selective, this enables measurements of not only concentration,
but also temperature, density, pressure, velocities and reaction chemistry. In combustion,
LIF has also been widely used for the measurement of species (OH, CH, NO and CO
among others) in reacting flows and flames. Planar LIF imaging (PLIF) has become an
effective tool for 2-D imaging of chemical species concentrations and temperature fields.
The historical progress of LIF is described in these review articles [36-39].
LIF can be understood as a two-step process as in the name: Laser Excitation and
Fluorescence. First, molecules are excited to exact higher energy states by the absorption
of a photon from the laser with the same wavelength. After a certain length of time, the
excited molecules de-excite back to the ground state along both radiative and
non-radiative pathways. Among these two pathways, the LIF signal comes from radiative
relaxation with the excess energy released in the form of fluorescent light. Non-radiative
de-excitation always happens with LIF: its impact is important for deducing quantitative
results from the LIF signal.
To provide insight into such complicated process, a two-level model is introduced
[40, 41]. A schematic of the model is shown in Figure 2.1 with transitions from two
electronic energy states of a typical diatomic molecule.

20

This two-level model is based on following four assumptions to serve to the
complicated process [42]:
1. Laser beam for excitation is uniform and linear in polarization;
2. The entire population of molecules is in the electronic ground state before
excitation;
3. The fluorescence light is measured at the maximum of the emission pulse and the
upper excited level population is in the steady state.
4. The emission fluorescence signal is of one single wavelength which corresponds
to a single rovibranic transition.
With these assumptions, this model consists of four processes (absorption,
stimulated emission, spontaneous emission, and collisional quenching) with four rate
coefficients. A more accurate model with practical physics and energy distribution levels
of actual diatomic molecules used the approach with quantum-mechanical density-matrix
[41, 43]. In practice, the tow-level model can be used as a rough approximation in
quantification of LIF results.
There are some rules for electronic transitions in diatomic molecules:
J (quantum number of the total angular momentum): ΔJ = +1, 0, -1 and J = 0 to
J = 0 is forbidden
Symmetry of the molecular system: +
R branch: f f)

21

– (Q branches: e f; P branch: e e;

Λ (quantum number from the electron orbit angular momentum) ΔΛ= +1, 0, -1
S (Quantum Number for total Electronic Spin): ΔS = 0

Potential Energy

Excited State A*

V’=2
V’=1
V’=0

k21

B21
B12

B
B12
21

v”=2
v”=1
v”=0

Ground State X
Figure 2.1 Two-level system with pertinent energy transfer processes
Transitions of NO, through either absorption or emission, between the electronic
2 +

2 +

states A Σ −X Π are usually observed in the UV range, and are traditionally called the
NO γ bands.
All the following processes can happen during the excitation and de-excitation of
the target molecule. We will discuss their relations with LIF signals.

22

Stimulated Absorption of excitation laser photon: the molecule excited from the
ground state into the upper excited state. The absorption rate B12 is proportional to the
laser spectral irradiance and Einstein coefficient.
Stimulated Emission: the excited molecule relaxes from the upper excited state
into the lower ground state and emits a photon.
Spontaneous Emission (also Fluorescence): a photon emits with energy
decreasing from the excited state into the ground state.
Simultaneous Excitation: more than one ground state level is excited due to the
overlap function of broadened absorption. NO in particular has a dense absorption
spectrum. Moreover, under high pressures conditions, the widths of collisional lines can
expand to many wave numbers, causing different pairs of states to participate in the
overall LIF process.
Spontaneous Emission: fluorescence from the laser excited higher energy state
into other lower electronic energy states.
Photoionization. Photoionization happens when a molecule absorbs a second
photon on top of the excited state with its overall process main part of Resonantly
Enhanced Multi-photon Ionization (REMPI) spectroscopy.
Rotational Energy Transfer (RET): If the target molecule collides with another
one, its rotational state may alter. Since RET is a fast process, having the same time scale
as the laser induced fluorescence processes, it then has critical implications on LIF
dynamics. While molecules are continuously pumped from a lower energy state to a
23

higher state by the laser, and continuously de-excited, which perturbs the thermal
equilibrium, RET in both states keeps on refilling and depleting these two laser related
states to reestablish rotational equilibrium.
Vibrational Energy Transfer (VET). Similarly if the target molecule collides
with another molecule, its vibrational state may alter. VET establishes and maintains
vibrational equilibrium. VET, however is often less effective than RET.
Quenching or Electronic Energy Transfer (EET): Similarly if the target molecule
collides with another molecule, its electronic state may alter. The higher excited
electronic state is likely to relax to a ro-vibronic level of the ground electronic state after
collision.
RET and VET keep the molecule remain in the electronically excited state, therefore
leaving the possibility of fluorescence, while EET disables the fluorescence as it prohibits
the fluorescence of the molecule. Therefore Quenching/EET is a direct and usually most
important competitor to the LIF process.
Rate coefficients of absorption and stimulated emission rates are B12 and B21,
respectively. Quenching rates are k21 with spontaneous emission rates using A21. The
total molecule number density is N=N1+N2, with N1 and N2 representing the population
number density in states 1 and 2. The energy density for excitation laser is
population of each energy state can be written as

24

ρ v . The

dN1
= − ρ v B12 N1 + ( ρ v B21 + k21 + A21 ) N 2 ,
dt

(2.1)

dN 2
dN
= − 1 = ρ v B12 N1 − ( ρ v B21 + k21 + A21 ) N 2 ,
dt
dt

(2.2)

and

The number of fluorescence photons Nf can be obtained as an integral over the
duration

τ

of laser excitation:
τ

N f = A21 ∫ N 2 (t )dt ,
0

(2.3)

If the excitation laser pulse function is a perfect square pulse, then the exact
solution to Equations (2.1) to (2.3) is:

N f = A21τ ⋅ N ⋅

ρ v B12

ρ v ( B21 + B12 ) + k21 + A21
1 − exp{1 − [ ρ v ( B21 + B12 ) + k21 + A21 ]τ } ,
(1 −
)
[ ρ v ( B21 + B12 ) + k21 + A21 ]τ

(2.4)

Equation 2.4 is too complicated to analyze. To simplify the problem and quantify
for the laser fluorescence signal used here, let us consider the steady state case. In the
steady state, the population of both states remains the same, thus the derivative of the
number density on each energy state is zero. Consider that there is degeneracy in each
energy state, then Equation 2.4 can be simplified as follows (the degeneracy for each
state is g1 and g2):

25

Nf =

g2
g1 + g 2 ρ B +
v 12

ρ v B12

g2
( k21 + A21 )
g1 + g 2

⋅ A21τ N

,

(2.5)

The denominator has two terms. The first, ρ v B12 , is related to rate of molecule
excitation; the second term relates to quenching and emission rate coefficients.
Depending on excitation laser intensity, two different limit cases are possible:
(1). Linear Regime. When

ρ v B12 << k21 + A21 , which means the laser intensity

is low, one obtains

Nf =

A21
ρ v B12τ N = φρ v B12τ N ,
k21 + A21

(2.6)

The number of fluorescence photons is linearly proportional to the excitation laser
intensity, which means that in this case signals must be calibrated for fluctuating laser
intensities. The term A21/A21+Q21 in this equation is the fluorescence yield (φ), with the
definition of the ratio of the spontaneous LIF emission rate to all de-excitation from
excitation rates. Fluorescence yield is also a direct display the percentage of the excited
molecules which will actually fluoresce.
(2). Saturation Regime. When

ρ v B12 >> k21 + A21 ,

which means the laser

intensity is high,

Nf =

g2
A21τ N ,
g1 + g 2

26

(2.7)

In this case, the fluorescence signal is independent of the excitation laser intensity.
In practical diagnostics application, this case is called the laser induced saturated
fluorescence or saturation spectroscopy. Here the fluorescence signals are independent of
the laser intensity. But, the very high laser intensities for saturation LIF require careful
experimental attention. What is more, complete saturation is hard to achieve, especially at
elevated higher pressures.
0.3

LIF signal [a.u.]

0.04

0.2

0.03
0.02

0.1
0.01
0.00
0.00

0.0
0.02

0.04

0.06

0.08

0.10

0.2

0.4

0.6

0.8

0

20

40

60

1.0

80

1.0

0.8

LIF signal [a.u.]

0.0

0.8

0.6
0.6

0.4
0.4

0.2

0.2

0.0

0.0

0

2

4

6

8

10
2

100
2

Laser intensity [MW/cm ]

Laser intensity [MW/cm ]

Figure 2.2 Transition from linear regime to saturation regime in the non-transient
two-level model in a 1 bar, equivalence ratio 0.9 methane/air flame. A difference for
four figures is the different laser intensity scales [44].
Figure 2.2 shows LIF signals over a different ranges of laser intensity calculated
using the non-transient two-level model discussed below by Bessler [44]. It is clear that

27

the linear and saturation regimes are extreme cases for lowest and highest laser intensities.
The highest signal intensity in the full linear range is only about five percent of the signal
intensity at full saturation regime. In reality, transitions between these two cases were
observed. In this research, all data is presented in the linear description, in which
interpreting of the LIF signals is easier. It is also noted here that Rotational Energy
Transfer (RET) in both states might change the saturation regime dramatically, especially
during the extended linear ranges.
To sum, the LIF equation, from the key physics involved in the LIF process, can be
derived from a non-transient analysis of the two-level model. The LIF intensity can be
written as the following in Equation 2.7.

I LIF = c × Iν0 × n10 × f B × B21 × V × Γ ×
Here, V is the volume,

Iν0 is

Ω
A21
×
4π A21 + k21

,

(2.8)

the laser intensity, c is the experimental setup optical

collection efficiency and Γ is the dimensionless overlap function [45]. Equation 2.1 is
known as the LIF equation which is a fundamental one to solve for LIF signals between
2 +

practical experimentation and computational simulations. Values of A21 for NO A Σ
6 -1

(υ’=0) state are ~4.6 × 10 s

10 -1

while the quenching rate k21 is on the order of ~10
-4

which leads to a fluorescence yield of about 10 . Ω is the solid angle of collection.

28

s ,

2.2 NO LIF: Literature Survey
There are some unique advantages to use NO as a tracer for LIF measurements in
combustion. First and foremost, NO has a relatively good fluorescence efficiency and a
strong absorption cross-section, which lead to strong LIF signals for detection. Also, NO
is a molecule whose spectroscopy is well studied and known. Third, NO does not exhibit
J-dependent or so-called rotational-level dependent quenching rates or radiative LIF
times that result in systematic measurement errors. Fourth, NO emits a strong
nonresonant fluorescence signal, which can provide detection strategies without worrying
about Rayleigh scattering or radiative trapping when combined with spectral filtering. All
of these enable numerical modeling of the LIF signal. In addition the modeling has
reached a high level of maturity and could provide flexibility in design and interpretation
of the NO-LIF experiments. Also, NO is quite stable in harsh high temperature
(T≥2000K) and pressure (P≥30bar) combustion environments to be a good tracer. Finally,
NO can be easily seeded into the combustors to make LIF measurements in the entire
flowfield. This is different from molecules such as OH, which are only feasible in the gas
mixture reaction regions.
There have been extensive publications regarding NO-LIF for both experimental
research and practical applications. Perhaps the first report of NO-LIF use was by Grieser
and Barnes [46]. They used single-point NO fluorescence measurements. McKenzie and
Gross [47] later reported the use of two-photon NO LIF for single-point temperature

29

measurements. Kychakoff et al. [48] reported the first 2-D imaging of NO fluorescence of
NO in CH4/air flames and also suggested the potential use of 2-D imaging to get
temperature. Paul et al. first reported two-line PLIF thermometry [49]. These references
are a small fraction of the NO-LIF articles in the literature, and are chosen because of
their pioneering contribution.
With the development of new technologies, single laser pulse LIF of NO for
measurement of concentration, temperature, and velocity is well established. There is a
need to obtain higher speed images of NO LIF to improve image and dynamic behavior
of the reacting flow. This requires a high-speed laser system and a high-speed imaging
system. Kaminski used four double-pulsed Nd:YAG lasers to pump one single dye laser,
for OH PLIF imaging, resulting in a four times laser frequency and then a four times
signal collection rate [50]. Using 270 mJ per pulse at 532 nm out of Nd:YAG, Kaminski
and et al. generated pulses at 282 nm with a averaged energy of about 1 mJ/pulse 8 kHz
repetition rate [50]. Kittler and Dreizler investigated the turbulent flame using UV light at
283 nm of about 22 uJ/pulse from a dye laser which was excited by a 5 kHz Nd:YLF
pumping laser [51]. Lempert reported an ultrahigh (100 KHz) PLIF images of NO using
ultrahigh (100 KHz) burst pulses of deep UV of 226 nm by mixing the third harmonic
laser (355 nm) from a burst pulse laser with a 622 nm output from an optical oscillator
[52]. One important parameter is the energy of the ultrahigh pulse, which in [52] are
about 0.5 mJ at the 226 nm excitation laser, which is sufficient for instantaneous PLIF.
A simulation code, LIFSim [53], widely used to model NO-LIF, is also introduced in
30

this section. LIFSim was developed to conduct numerical simulations of excitation and
emission processes in LIF (for both NO and O2) to quantify species concentrations and
temperature measurements. Several basic criteria must be satisfied to perform
quantitative LIF measurements. First, the emission spectrum of the molecule for LIF
should be well known. Second, knowing the absorption spectrum from a practical
accessible laser is necessary. Third, non-radiative pathways (besides LIF including
quenching, predissociation and photoionization) should be understood. Fourth, the rate of
excited state radiative decay should be known because the fluorescence power is
proportional to this decay rate. LIFSim was developed by Bessler (2003, Heidelberg
University) which is based on a previous numerical code by Sick (1995). This software is
available in a DOS shell executable format and also via an interactive web interface
(www.lifsim.com) [53].
Here some examples of LIFSim are listed for different purposes.
1, Absorption Spectra Calculation: pressure, temperature, excitation wavelength
range and resolution are necessary parameters for calculation of absorption spectra.
2, LIF Excitation Spectra Calculation: excitation wavelength range, pressure and
resolution are necessary parameters for this calculation with detection bandpass being a
variable.
3, LIF Emission Spectra Calculation: the monochromater function is calculated
depending on parameters provided by the user.
4, Temperature and Pressure Dependence of the LIF Signal Calculation: pressure
31

range, excitation wavelength, and temperature range are necessary parameters. The LIF
signal can also be used for constant mole fraction or number density, which is on a per
volume or per molecule basis.
5, Simulated LIF Excitation Spectra Comparison with Experimental Data: this
comparison has an adjustment of analysis of fit precision, line positions and accuracy.
The relative LIF signal strength of the different transitions can show temperature
information in a LIF excitation spectrum. Temperature is obtained by comparing
experimental data with simulated spectra. To obtain correct temperatures from the
comparison, the line positions from the simulation are adjusted to data from the
experiments.

Parameters

used

in

fitting

are

intensity,

temperature

and

a

wavelength-independent baseline. There can be additional parameters used for the fitting,
such as laser Lorentzian and Gaussian line width contributions (to evaluate the spectral
shape of the laser) and a linear inclination of the baseline (to evaluate extended scans).

32

Chapter 3 Experimental Facilities
3.1 Microwave Plasma System
3.1.1 Coaxial Re-Entrant Cavity Applicator
A unique coaxial re-entrant cavity applicator was developed for coupling
microwave energy directly into the reaction zone. Using microwave power for plasma
generation offers a range of benefits including: (1) highly efficient production of radicals
and in sustaining a discharge in high-flow gases; (2) ability to operate in thermal and
non-thermal modes; (3) ability to sustain good plasma uniformity at atmospheric
pressures and higher; (4) electrodeless microwave plasma system; (5) ability to be
focused in a specific location and stabilized by gas flows [54]; and, (6) operation in
continuous-wave

modes

(partial

non-equilibrium)

and

pulsed

modes

(strong

non-equilibrium with pulse duration < 100ns). This torch also utilizes many of these
advantages including discharge uniformity in a glow mode in the reaction zone at
atmospheric pressure when viewed using an intensified CCD camera with gate times
down to several nanoseconds.
In Figures 3.1 and 3.2 we show the coaxial re-entrant cavity applicator. Figure 3.1
displays a photograph of the system along with an inset picture of a methane/oxygen
flame. Figure 3.2 displays the corresponding schematic. The applicator is a brass
re-entrant cavity excited in the TEM mode [55] (in the current study, the TEM 011 mode

33

[56]) and includes the following components: (1) brass gas tube burner; (2) adjustable
endplate; and (3) mono-pole antenna loop. All three adjustments are set by using a
micrometer unislide with an accuracy of 0.01 mm. This redesign of our previous system
[57] has a modified tip around the main chamber that enhances focus of the microwave
energy into the reaction zone of the flame.

Figure 3.1: Picture of coaxial re-entrant cavity applicator with three fine adjustments
using unislides; picture of methane/oxygen flame shown in the inset.

In the bottom portion of Figure 3.2, the setup and components for microwave power
supply and measurement system. It is consisted of a power supply, two directional
couplers, two power meters with a power sensor and a directional coupler. Two
directional couplers and two power meters are used because both incident and reflected
power have to be measured and calibrated. This system not only sets up for tuning exact
34

resonant modes in the cavity but also is used for the calibration of power delivered into
the system.

gap g is localized
coupling region

flame

brass plug
mono pole
antenna loop
sliding
short
adjustable
endplate
brass gas tube

Circulator

To MW Cavity System
Directional
Coupler

Load

Power
Supply

Power Meter

Control System

Figure 3.2 Scheme of the structure of coaxial re-entrant cavity applicator with burner
centered in the middle (Top). Components of microwave power supply and
measurements setup (Bottom).
2.45 GHz microwave energy is transmitted via a coaxial cable and emitted with a
mono-pole antenna (Figure. 3.2). When the relative location of the torch, antenna, and
baseplate are adjusted using the unislides, an optimized resonant mode can be found
when most of the energy is focused into the region between the edge of the outer cavity

35

and the end of the combustion torch. Further details regarding simulations to optimize the
orientation of the electrical field will be discuessed next. There are several nearby TEM
modes that can be accessed by changing the combination of the three way adjustments.
We select a mode that allows the center torch to be exposed outside the main chamber, so
that the flame is clearly visible and can accommodate laser and optical diagnostics. The
chamber itself is 35 mm in diameter and is generally optimized to a length of around 3 cm
(λ/4 for 2.45 GHz). The chamber is constructed of brass, which can be considered a
perfect electrical conductor for simulation purposes.

3.1.2 Direct Microwave Plasma Coupling
In the first part of this study, we investigated direct coupling of electromagnetic
energy into the reaction zone of an existing premixed flame for the energetic
enhancement of combustion chemistry. A practical extension of this concept could be a
scramjet engine where electromagnetic energy is coupled into the combustor to extend
operation limits during transient conditions or harsh loading. The plasma generation
process can be achieved with minimal power, since a high electric field is directly
coupled to the electrons created in a hydrocarbon-based flame through chemi-ionization.
It is expected that the combustion-produced electrons are localized to the flamefront
11

-3

and have densities much lower than the critical density (<10 /cm

for 2.45 GHz

excitation), where the microwave electric fields can freely penetrate the electron gas to
accelerate the electrons through a collision cascade process. In this atmospheric flame (Tg

36

= 2200 to 3300 K) the mean free path (MFP) for the electron–molecule/atom collisions is
of the order of 1 μm. The collision frequency for momentum transfer is of the order of
11

10 /s.

Thus, the microwave electric field will heat the electron gas via elastic

electron–neutral collisions, i.e., electron gas heating occurs by a collisional or Ohmic
heating process that is governed by the following equation:

υ
1 noe (r )e 2
( 2 m 2 ) | E (r ) | 2
> (r ) =
2 2meυ m ω + υ m
2

< Pabs

（3.1）

3
where < Pabs > (r ) is the absorbed microwave power density in W/cm , noe (r ) is
-3

the electron density in cm as a function of position

r

in the electron gas,

υm

is the

electron–neutral collision frequency for momentum transfer, ω = 2πf is the angular
frequency, and E (r ) is the impressed electric field strength. Typically for our gas
temperatures and pressures,

υm

is in the order of 10 –20 ω.

When the microwave energy is coupled into the applicator/flame load, electron
heating process takes place, and the electron temperature Te increases above the gas
temperature Tg (Te>Tg) creating a non-equilibrium plasma. As the microwave electric
field and microwave power are increased, the Te increases to the level where some
electron–neutral collisions are inelastic, producing new radical and excited species. If
electric field strengths and power levels are further increased, ionization collisions may
also occur. Eventually, the increased electric field strength is sufficient to cause
breakdown, resulting first in micro plasmas and then a partially ionized plasma plume in

37

the reaction zone. The de-excitation of this gas through three body collisions results in
the increase in Tg downstream.
3.1.3 Equivalence Circuit and Electric Field Simulation
Figure 3.3 shows an equivalent circuit for the coaxial re-entrant cavity applicator.
For simplicity, the microwave losses due to the surface currents on the coupling loop and
the applicator are neglected. The circuit consists of a variable length Ls−Lg, 50Ω
transmission line, which extends from the sliding short to the gap region. The coupling
loop is modeled as an ideal transformer with an n:1 turns ratio and the lumped circuit jYL.
When the flame is ignited, the gap capacitance is further modified by the presence of the
hot neutral gases and the low-density electron and ion gases produced by combustion.
Microwave energy is introduced into the applicator by adjusting the applicator to an
electromagnetic resonance at the 2.45 GHz excitation frequency, and then the microwave
energy is coupled into the flame via Ohmic heating of the flame electron gas. Detail
analysis is shown in [57].

38

ω/c, 50Ω

C

G
jY

LS-LG

L

ZY
n:1

VS

Figure 3.3: Equivalent circuit for coaxial re-entrant cavity applicator 
The overall size and geometry of the re-entrant cavity applicator was designed to
enable coupling of the microwave energy into the reaction zone of the flame and was
based on numerical simulation of the electric field propagation. Figure 3.4 shows a
numerical solution of Maxwell’s equations using “Rf Electromagnetic Solver” module of
the program COMSOL MultiPhysics 3.4.

This provided a relative 2-D electric field

distribution within the coaxial cavity structure. For our simulations, the medium inside
the cavity is air, and all boundaries are set to be perfect electrical conductors. In our
experiments, the focus of the electromagnetic energy into the tip of the torch is enhanced
as the radicals and the weak concentration of electrons in the flame amplify the coupling
process.

39

MAX

MIN

Figure 3.4: Numerical simulation of electric field from (V/m) COMSOL MultiPhysics
RF Electromagnetic solver for current experiments
The dimensions for simulation are from experimental conditions when plasma
discharge is observed with the cavity tuned to resonance at 2.45 GHz microwave
excitation frequency. The E-field simulation shows an intense electric field at the end of
the torch (reaction zone) and the sharp edge of the cavity. In practice, this type of mode
can be used to focus the microwave energy when a flame exists on top of the burner exit.
The tip of the burner can be raised almost flush with the cavity surface so that the flame
is fully exposed for study using optical and laser diagnostics.

3.1.4 Preview Study
A similar microwave cavity applicator with combustion torch burner was
investigated by Hemawan at Michigan State University [58]. The components and e-field
40

pattern are shown in Figure 3.5. The top plate can be moved to change the height of the
cavity where the exact mode is created. The sliding short is tuned to the length needed,
which is approximately 14.2 cm for TM012 mode. After putting microwave power into
the system and tuning the system directly to TM012 cavity mode fields, an atmospheric
microwave discharge will be created at the end of the center conductor. The mixture in
the torch is methane and oxygen. The torch needs a water cooling system due to the heat
release of combustion and the thermal effect of the microwave power. It is reported that
in this kind of experimental setup, the flame size increases with increased microwave
power, especially after 40 Watts, which is when the microwave discharge is initiated [59].
With 30 Watts microwave power, the flame slightly increases and maintains a premixed
combustion flame. When the microwave power exceeds 40 Watts, the visual size of the
flame and emission intensity both dramatically increase. Also flammability limits of the
microwave assisted flame with different microwave power levels were investigated.
These are shown in Figure 3.6. The lean flammability limit is achieved under the same
total flow condition and decreases the fuel to blow off, which misinterpreted as the lean
limit. With the addition of only10 Watts microwave power, the lean flammability limit is
extended by 5%–12% in equivalence ratio. With more than 20 Watts, the equivalence
ratio of lean flammability limit is zero indicating there isn’t any need for fuel to sustain
the plasma discharge at these plasma levels.

41

Figure 3.5 Components of microwave cavity plasma assisted flame system with E-field
pattern shown [58].

42

0.8
0.7
0.6
0.5
0.4

9% avg

0.3

Reduction

0.2
0.1
0.0
70

120

170

220

270

Figure 3.6 Flame extinction limits (lean) with 10 Watts, and 20–100 Watts microwave
power. [58].

3.2 Direct Current Plasma System
3.2.1 Transient Arc Plasmatron
The transient arc plasmatron used in Chapters 4 and 5 in this research, shown in
Figure 3.7, is a transient direct current (DC) discharge system that employs the high
voltage arc plasmatron [26]. The plasmatron has 5 mm inner and 10 mm outer diameters
with an arc chamber distance of about 5 cm. Figure 3.8 shows a schematic of the transient
arc plasmatron. The transient arc DC system operates on a repetitive glow to spark
transition mode, offering the advantages of a thermal plasmatron with low average power
output and temperature (Tg < 1200 K in an air plasma, ne~10

43

14-15

-3

cm ). During glow to

spark transition (~100 ns), a short duration spark or diffused channel arises. The torch
works mainly in the glow mode with voltage applied to the cathode (inner electrode) and
the surrounding anode. One advantage is that the electrode erosion is less than traditional
DC torches due to the lower electrode temperature. Therefore, intensive cooling is not
required. Classical DC arc plasmatrons operate using very high current while this new
plasma igniter/plasmatron uses a semi-glow discharge with random completed or
noncompleted transitions to spark the mixture, which results in highly efficient ignition
and flame control with relatively low current, and thus low average power input. This
spark exists for only about 100 nano second, also hence the operation described as being
‘transient’.

liquid fuel
atomizer

(b) discharge only

(a) plasma torch

gas fuel inlet

(c) discharge + flame

Figure 3.7 Image and schematic of transient arc plasmatron with gas/liquid feeding (a);
Air plasma imaging at 20 SCFH and 300 mA (b); Plasma assisted stoichiometric
methane/air imaging at 20 SCFH and 300 mA (c).

44

Gas Flow
RS

VG

Figure 3.8 Schematic of transient arc plasmatron with gas/liquid feeding.
This plasmatron has been shown to increase the flammabity over a wide range of
operating conditions. For example, the flame speed with an unburned gas temperature of
300 K in a traditional premixed methane/air flame is SL=0.40 m/s, which can be
increased more than three fold with a current of 300 mA. A fuel/air mixture can be ignited
and combustion stabilized at lean conditions with oxidation occurring with equivalence
ratios as low as 0.1, and possibly lower. The ignition mechanism is related to the glow to
spark transition mentioned above. At the low current glow mode, dissipated energy is not
high enough to initiate the fast oxidation process of the fuel, but it is enough to build a
pool of chemically excited species. When glow to spark transition occurs, short duration
spark or a diffused channel arises, initiating the ignition process. Even with the small
energy released by the transition spark or diffused channel, localized energy deposition
coupled with chemically active particles can lead to the rapid initiation of thermal
oxidation. Both liquid and gas fuels can be fed directly into the arc chamber, allowing for
45

high efficiency thermal oxidation without prior hydrocarbon decomposition. In practice,
this igniter will be attached to the main combustor to initiate ignition and to sustain
combustion.

3.2.1 Glow-to-Spark Transition
The uniqueness of this DC plasmatron lies in its glow-to-spark transition which was
investigated by Matveev et al. [60]. The average discharge current for current DC
plasmatron has been reduced to 0.1 A, much less than typical classical high-current DC
arc plasmatron. Even though the total power from power supply is low (around 200 Watts)
and the temperature output from the plasmatron is less than 500 K, this plasmatron has
proven to be an efficient device for reliable ignition and flame holding in various flow
conditions (flow rates, equivalence ratios).
A typical voltage and current plot is shown in Figure 3.9 in this type of plasmatron.
The condition for such voltage and current plot is that the discharge gap doesn’t recover
the dielectric strength completely during the time interval between two high current
pulses. From the figure, we can see that the situation in the electric circuit and the
discharge gap changes radically. But for this system, the fuel oxidation process is
initiated by the high current short time spark shown in Figure 3.9 while the flame cannot
be sustained during the pause between the high current pulses.

46

VS
iG
0

t1 t2

t3 t4

t5

t6

t7

Figure 3.9 Typical DC discharge gap voltage and current passing through the gap [60].
The main idea of glow-to-spark transition is to form a regime where the gap is not
fully recovered for its dielectric strength between each high current pulse. The most
attractive advantage of such a regime is that the fuel/oxidizer is ignited by the high
current spark discharges and is then additionally sustained because of a low-current
glow-type discharge. Overall we can call ignition of the discharge “non-steady” since
current and voltage fluctuate in this case. The transition of glow to spark can also be
divided into completed transition and non-completed transition, as shown in Figure 3.10
by the plot of current and voltage and CCD camera photograph of the discharge
between Δt . We can clearly see two types of glow-to-spark transition. Also the number of
the diffusion channels in the CCD images is exactly the same as shown number of
glow-to-spark transitions in the current-voltage plots.
To summarize the process of flame ignition in this type of DC transient arc
plasmatron: most of the time the discharge is a low current glow type with specific power
not high enough to ignite the mixture. But with the short duration high current, spark or

47

diffused channels appear which serve to initiate the ignition process. The initial ignition
kernel grows quickly and efficiently into the surrounding medium full of the low density
non-equilibrium glow discharge and large number of chemically active particles. In this
condition, even very small spark energy is sufficient to sustain combustion process.

Figure 3.10 Voltage and current plots for discharge in air/propane mixture and CCD
images of the discharge with exposure time of Δt shown in the voltage-current plot
[60].

3.3 Laser and Optical Diagnostics Setup
A typical overall LIF experimental setup with synchronization of lasers, energy
monitor, ICCD camera and control system is shown in Figure 3.11. This general setup
applies to LIF of OH, CO and NO.

48

Energy Monitor

IR Camera

Laser

ICCD Camera
Spectrometer
Power Supply

Plasma
Torch
Gas Feed

Figure 3.11 Overall experimental setup with synchronization of lasers, energy monitor,
ICCD camera and control system.
For

OH

PLIF,

the

UV

laser

near

283

nm

was

used

for

2 +

planar
2

laser-induced-fluorescence (PLIF) using the Q1(8) transition from the A Σ −X Π (1,0)
[61]. The emission spectra from experiments are shown in Figure 3.12. The
measurements were conducted using a dye laser (Lumonics Hyperdye HD-300). R590
dye from Exciton was used for generating pulses at 566 nm, which is frequency doubled
through an Inrad Autotracker (ATIII) to a final output frequency of 283 nm (linewidth ~
-1

0.1 cm ). The 283-nm beam is isolated from the 566-nm beam within an Inrad Prism
Harmonic Separator (PHS). The laser was expanded into a sheet and the fluorescence
o

signal was collected at 90 using an intensified CCD camera with a WG305 filter. The
number density was derived by calibrating the signal to that of a Hencken Burner using a
methane-air mixture. Here it was assumed that the fluorescence was in the linear regime;

49

accordingly the OH number density was further corrected for rate differences for
electronic quenching between the methane-oxygen flames and Hencken calibration flame
using estimates of quenching rates from the OH A-X transition [62]. It should be noted
that the reported cross sections describe quenching from A(v′=0), whereas fluorescence is
derived from both the A-X (v′=1,v″=1) and (0,0) bands after excitation to A(v′=1).

100

283.029 nm

Intensity [a.u.]

80
60
40
20
0
2800

2825

2850

2875

2900

Wavelength [A]
Figure 3.12 Excitation Spectra of OH (A-X) with Q1(8) transition shown at 283.029 nm.
1 +

1 +

CO in the reaction was measured using two-photon PLIF, pumping the B Σ −X Σ

(0,0) transition at 230.10 nm [63]. The emission spectrum of CO with the maximum
signal at 230.10 nm is shown in Figure 3.13. The 532 nm beam from a Newport Nd:YAG
was used to pump a Sirah Precisionscan LG-2400 dye laser operating with Exciton DCM
dye. The 653-nm output of dye laser was sum-frequency-mixed with the third harmonic

50

beam (355 nm) from the injection seeded Nd:YAG laser to produce 230 nm radiation
-1

(linewidth ~ 0.1 cm ) needed for two-photon laser induced fluorescence.

The pulse

energy is recorded digitally using an energy monitor from LaVision.

Intensity [a.u.]

400

230.1nm

300
200
100
0
230.06

230.07

230.08

230.09

230.10

230.11

Laser Wavelength [nm]
Figure 3.13 Excitation Spectra of CO (B-X) with highest intensity at 230.109 nm.

The fluorescence signal was imaged with an interference filter centered at 486nm
with a bandwidth of 10 nm. The number density is calibrated from a Chemkin simulation
(mole fraction for 0 Watt microwave power) using GRI-mech 3.0 [64] with 1-D flame
modeling with methane and oxygen. Equilibrium values past the reaction zone are used
as a first order approximation of the Bunsen flame geometry of our flame. Here, it was
assumed that the rate for photo-ionization, from a third photon at 230 nm, dominated the
rate for electronic quenching and that no quenching correction was needed.
+

For NO LIF, the excitation of NO using A2Σ −X2Π (0, 0) transition requires
narrowband UV light near 226 nm. The measurements were conducted using a dye laser

51

(Sirah Precisionscan-LG-2400) with two stage amplification. Coumarin 450 from Exciton
was used for generating pulses at 452 nm, which is frequency doubled through a BBO
crystal to a final output frequency of 226.03 nm. Laser pulse energy ranged from 7~10
mJ/pulse with a 7 ns pulse duration. The laser is pulsed at 10 Hz with a final line width of
-1

0.2 cm

at 226.03 nm. For spectrally resolved 1-D imaging of the LIF emission, the

beam is weakly focused (diameter ~ 1.5 mm) along a line 20 mm above the top of the
torch. The pulse energy is recorded digitally using a fast photodiode (LaVision).
The fluorescence signals were collected at a right angle to the incident laser beam
and focused using a f=105 mm f/4.5 achromatic UV-lens (UV Nikorr 105), dispersed
spectrally through an imaging spectrometer (Acton SP-2300) and imaged onto an
intensified camera system composed of a intensified relay optic (LaVision IRO) and CCD
camera (LaVision Imager Intense Pro). The resulting image shows the laser path
propagating through the flame on the vertical axis and the spectral resolution on the X
axis as shown in Figure 3.14 (upper portion). It should be noted that the emission
wavelength region coincides with the emission from the Schumann Runge bands of O2.
In order to quantify this effect, the emission spectra of NO-LIF in Figure 3.14 have been
separated into various components using a nonlinear fitting scheme (using a numerical
emission model [53]) and show the potential contribution from O2. Therefore, selection
of an excitation line with a minimized O2 contribution should be considered.

52

900

Signal Intensity [a.u.]

800
700
600

1-D Line Imaging

Rayleigh

500

Filter Function

400
300
200

O2

100
0
220

240

260

280

300

Emission Wavelength [nm]

Figure 3.14 . Spectrally resolved 1-D line imaging (inset) and the emission spectra of
NO-LIF with 226.03 nm excitation and selected filter function (solid red line). The
emission spectra has contributions from NO, O2 and Rayleigh scattering.
For planar LIF imaging, the laser was optically stretched into a vertical sheet (0.5 X
35 mm) and imaged directed using the intensified camera with a set of filters for
suppression of Rayleigh scattering.
A high pass interference filter at 240 nm (Asahi Optics) was used for suppression of
the Rayleigh Scattering and a low pass filter at 300 nm (Asahi Optics) was superimposed
to cut off any red shift emission past 290 nm as shown in Figure 3.14. It should be noted
that the spectral shape of NO LIF emission spectra strongly depends on the NO rotational
state [65], with the upper excited state generally not reaching thermal equilibrium within
the fluorescence lifetime. Therefore, to avoid effects of nonequilibrium excited-state

53

population effects, broadband detection or one or more complete vibrational bands is
preferred.
The full resolution of our camera was 1024 X 1376, which was binned 2 X on the
vertical axis and 2 X on the horizontal axis to provide a resolution of 64 by 64
micrometers resolution per pixel. All the images were shot with a 50 ns gate to exclude
any stray emission and room light. To ensure the spectral accuracy of the laser, scans of
neighboring NO lines were conducted before and after each data set and compared to a
numerical simulations from LIFSim [53]. No perceivable drift of the wavelength was
detected. Additionally, an IR camera (Flir Phoenix PTS) is used for the imaging
temperature fields.
For the optical emission measurements, A McPherson 216.5 1-m spectrometer
(entrance slit of 20 µm) was used for flame emission measurements and rotational OH
thermometry. The emission was focused using a spherical lens (f=50 mm) through a
pin-hole filter to isolate the reaction zone. A wide-range scan was conducted over the
range 250 to 550 nm at an increment of 0.5 nm to monitor the qualitative change in the
emission of excited state species.

54

Chapter 4 Combustion Dynamics with Microwave Discharge
4.1 Impact on Flame Structure and Radical Concentration
Photographs, 2-D OH PLIF and CO PLIF images of the flame are shown in Figure
4.1 as a function of increasing microwave power. At this condition, the fuel combustion
power is 22 Watts using the lower heating value of methane.

MAX

MIN

0w

2w 4w
Stage I

6w

8w 10w
Stage II

15w 20w
Stage III

Figure 4.1 Emission (top), OH PLIF (middle) and CO PLIF (bottom) intensity images of
CH4/O2 flame vs. microwave power with three stages noted (100 SCCM total flowrate at

φ=1.1).

55

The plasma plume is established in a uniform glow mode and micro-arc transition
for certain conditions outside our operational conditions. Overall, the flame luminosity
and volume are increased as more microwave power is coupled into the flame region. At
low powers the flame length is about 2 mm and a typical light blue hue is observed from
the CH* emission. As the power is raised to about 6 Watts, the microwave energy
generates a distinctly different plasma plume with a purple emission that can be up to 10
mm in length (top row in Figure 4.1). From the OH images (middle row), it can be seen
that the location of the reaction zone does not change dramatically, and only a significant
rise of OH in the post combustion gas region is evident. It is expected that emission from
a dramatic increased ionization and excited state species dominates the normal visible
emission spectra at this point. The progression of the plasma discharge as a function of
microwave power can be divided into three distinct stages, which we will refer
throughout the rest of this dissertation. Note that the microwave power listed here
generates a volumetric discharge around the flame. Only a fraction of the energy is
actually coupled into the reaction zone.
•

Stage 1: “Electric Field Enhanced Stage”: The flame is influenced by only a weak
electric field. The input microwave power is below 6 Watts. No visual change can be
seen in either the photographs or the OH PLIF images as the power is increased.

•

Stage 2: “Transition Stage”: Between 6 and 10 Watts, the plasma discharge and flame
exist at the same time. The discharge volume is dramatically increased. Near the end
of this range, the system can sustain a plasma plume even if the combustion gas is
56

discontinued. The plume can re-ignite a flame if combustible mixtures are
reintroduced. The intensity of both OH and CO is increased. The reduced electric field
7-8

concentration is above the critical threshold required for breakdown in the range 10

V/m. The distribution of the electric field was shown in Figure 3.4 in the previous
Chapter.
•

Stage 3: “Full Plasma Stage”: If the microwave energy input exceeds 10 W, the system
enters a stage where the plasma plume dominates, and the discharge emits a pink hue.
As the power is further raised, the volumetric discharge of the plasma is increased and
the emission transitions to a pink-purple hue. The discharge has a uniform glow when
viewed using an intensified CCD camera with gate times down to several
nanoseconds.
Without a flame present, the microwave energy cannot couple itself into the air and

generate a plasma plume in the power ranges discussed here. An interesting phenomenon
was observed that air discharge can be sustained at microwave power higher than 10
Watts after oxygen/fuel line was shut off.
In all three stages microwave energy is coupled into the flame via energy transfer
from the input microwave energy source into the electron gas. This energy transfer then
raises the electron gas temperature, thereby creating a non-equilibrium chemistry
condition in all three stages. The electron gas transfers its energy via electron - heavy
molecule elastic and inelastic collisions processes into the heavier molecular gases. The
main differences for these three stages are that electrons in Stage 1 mostly come from
57

flame only whereas in Stages 2 & 3, the electrons are believed to increase by 1 to 2
orders of magnitude (from ~10

10

-3

cm to ~10

11-12

-3

cm ).

To quantify the change in the OH radical over the entire range of microwave powers
in more detail, the number density of OH in the reaction zone of the flame is shown in
Figure 4.2.

16 -3
OH Number Density [10 cm ]

100SCCM 0.9

15.0
100SCCM 1.1
60SCCM 0.9

12.5

60SCCM 1.1
60SCCM 0.9
100SCCM 0.9
60SCCM 1.1
100SCCM 1.1

10.0

Stage I

Stage II

Stage III

7.5
0

5

10

15

20

25

30

Microwave Power [watts]
Figure 4.2 OH number density in burnt gas of CH4/O2 flame vs. microwave powers from
averaged PLIF images.
The region of data extraction in the PLIF image is shown in the inset and is mainly
from the reaction zone of the flame. The OH number density is plotted as a function of
microwave power (0 to 30 W) for different total flow rates and equivalence ratios. The

58

data is extracted from OH 2-D PLIF images which have been averaged over 200 single
shots and calibrated for variations in laser energy. For each data set (0 to 30 W at the
same total flow rate and equivalence ratio), the same general trend over the three stages
can be observed. At the same equivalence ratio (φ=0.9 or 1.1), the OH number density
increases with total flow rate. Also at the same total flow rate (60 or 100 SCCM), the OH
number density for φ=0.9 is higher than that for φ=1.1.
•

Stage 1: “Electric Field Enhanced Stage”: In the first stage (0 to 6 W), the OH does not
show a noticeable increase when the measurement uncertainty is considered (~8%).
There is insufficient microwave energy to generate a cascade of ionization leading to a
plasma discharge. Nonetheless, both flame speed and stability is enhanced mainly
through temperature rise from elastic collisions and in-situ reforming.

•

Stage 2: “Transition Stage”: In this stage (6 to 10 W), as the microwave field intensity
increases, Te (electron temperature) exceeds Tg (gas temperature). High Te and high
reactivity lead to inelastic collisions for excitation, dissociation and ionization to
generate new species and eventually break down to generate a plasma discharge.

•

Stage 3: “Full Plasma Stage”: In the third stage (more than 10 W), the plasma volume
grows larger and OH production remains relatively constant even though microwave
power is raised.
It can be seen by quantifying the slope of OH increase with respect to power input,

for 60 SCCM total flow rate, the slope for Stage 2 is 4 times bigger than that in Stage 1

59

and 20 times bigger than that in Stage 3. For the 100 SCCM case, the change in slope is
not as dramatic although Stage 2 is still the region of maximum change rate.
It is believed that in plasma enhanced flames, free electrons are generated in the
+

-

flamefront through the reaction CH+O → HCO +e [66, 67]. At higher total flow rates,
more heat release from combustion with relatively less heat loss to the burner leads to
higher temperatures, which results in more CH and subsequently more electrons. This
results in more OH production for higher flowrates in Stage 2 of Figure 4.2. In Stage 1,
the microwave power is not high enough to generate more atomic oxygen.
-

-

The other important pathways for OH production in Stage 2 are H2O+ e →e +H
+OH, H2O +O → OH+OH, and H +O2 → OH+O [68]. Atomic oxygen is believed to
increase in a plasma coupled combustions more than in combustion only cases [11]
-5

(increased from the order of 10

-4

in mole fraction without plasma discharge to 10

with

pulsed plasma). A Chemkin simulation using GRI-mech 3.0 [64] with a premixed burner
-4

model of a flame with methane and oxygen show that, with the addition of 10

mole

fraction of atomic oxygen, the OH number density increases by 40%. A similar increase
was observed in our experiments for both φ= 0.9 and 1.1 conditions at a total flow rate of
60 SCCM.
There are three reasons for relatively stable OH number densities in Stage 3. First,
even though higher microwave power creates more electrons, the increased OH by the
electron impact reactions is cancelled by more space recombination as additional active
species are increased. Second, the microwave coupling efficiency is lower for higher
60

microwave powers, due to more recombination reactions. Third, even though the power
is increased, the plasma volume is also increased and the power/volume ratio remains
relatively constant. From a practical perspective, this would be the stage that can be
utilized for both flame stability enhancement and re-ignition.
The relative uncertainty associated with the beam energy fluctuation corrections and
wavelength shift of the laser are expected to be is less than ±4%. The camera readout
error was suppressed by subtracting a background image taken without the laser pulse
and can be neglected. Combined with the errors from the flowmeters and shot noise, the
overall relative uncertainty is expected to be less than 8%. The error in the absolute
accuracy is expected to be higher (10−20% range).
Figure 4.3 shows the OH number density versus equivalence ratio (0.8 to 1.2) for
various microwave powers (0 to 30 W). For the same equivalence ratio, the trend in the
OH number density is generally the same in the three stages with the highest number
density being exhibited around φ=0.9 and 1.0. For each of the lines in Figure 4.3, the OH
profile change as a function of the equivalence ratio is similar to that typically found in
regular premixed flames [69]. This profile will be influenced by the electron number
distribution as microwave energy is coupled into the flamefront. Coincidently, the
simulations of a microwave plasma enhanced combustion of methane and air [67] show
that the electron number density reaches a peak when the equivalence ratio is close to
stoichiometric and decreases towards both lean and rich conditions, and therefore, the
overall trend of the OH peaking near φ=1 is maintained. The results show that the OH
61

number density for powers above the break-down threshold (above stage 1: ~6 W) is
much higher than the two lines in stage 1 (0 and 5 W) where only the influence of the
electric field was altered due to the increase in electron-impact reactions.

16.0
0 watts
5 watts
10 watts
15 watts
20 watts
30 watts

16 -3
OH Number Density [10 cm ]

15.5
15.0
14.5
14.0
13.5

rich

lean

13.0
12.5
12.0
11.5
0.8

0.9

1.0

1.1

1.2

Equivalence Ratio
Figure 4.3 OH number density vs. equivalence ratios (φ=0.8 to 1.2) with different
microwave powers (0 to 30 Watts). Total flow rate is 100 SCCM.

4.2 Flammability Limits and Flame Speed
The addition of plasma has been shown to increase the flammability limits [29, 30].
Flammability limits (both lean and rich) for this plasma discharge system with and
without microwave enhancement are shown in Figure. 4.4.

62

Equivalence Ratio

4.0
3.8
3.6
3.4
3.2
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.6
0.6
0.5
0.4
0.3
0.2
0.1
0.0
-0.1

0W

NO MW-rich
2 watts MW-rich
4 watts MW-rich
NO MW-lean
2 watts MW-lean
4 watts MW-lean
6 watts MW-lean

40

50

60

10W

rich
lean

70

80

90

100

Oxygen Flow Rate [SCCM]
Figure 4.4 Flammability limits for methane/oxygen flame (lean and rich) with
flamespeed increase demonstrated by flame photographs with no microwave radiation
and 10 Watts microwave radiation (right top).
In these measurements, the total flowrate is not held constant in determining the
extended oxidation limits. In the upper portion of the figure, the oxygen flow rate is fixed
and methane is increased to achieve the rich flammability limit. Our goal in this section is
to show how much fuel can be put into and combusted with constant oxygen flow input.
In the bottom portion of the figure, the oxygen flow rate is fixed and methane is
reduced to obtain the lean flammability limits. All the data presented in Figure 4.4 are in
the low-power range (0 to 4 W), with a flowrate spanning from 50 to 250 SCCM. Clearly
we can see improvements in the flammability in this range, although the flame
enhancement is only through a weak electric field. At power levels of more than 6 W, the

63

discharge could be sustained with only oxygen or methane and without combustion. As
the ground state OH number density was not shown to increase in this stage, the
improvement in the flammability limit is mainly assumed to be from temperature increase.
This will be discussed further next.
Flamespeed enhancement was also shown in this region by observing the change in
the reaction zone contour as the flame velocity was changed. An example of this is shown
in the upper right inset of Figure 4.4, where a photograph of the flame is shown without
microwave input and with 10 W of microwave power. Here, the flamespeed increase is
demonstrated by a decrease in the Bunsen flame “cone” angle; this is observed in all
three stages of the microwave coupling. Flamespeed enhancement using microwave
fields has been observed previously in tunable microwave cavities, using free floating
laminar premixed flames [28, 70]. Note that in Figure 4.4 the flame images look different
from those in Figure 4.1 due to different exposure times. In Figure 4.1, a longer exposure
time (1/10 second) is used to integrate the emission from all the chemical species. In
Figure 4.4, a significantly shorter exposure time (1/30 second) was used for high contrast
of the flame front.

4.3 Emission Measurements and Temperature
In order to probe the changes in excited state species and temperature, spectrally
resolved emission of the reaction zone was obtained. Figure 4.5 shows the emission
spectra for the methane-oxygen flame with no microwave input and 10 W microwave

64

input (top) and OH rotational temperature for different total flow rates and equivalence
ratios (bottom) with the top portion of Figure 4.5 demonstrating a fast scan over the entire
spectral range.

Rel. Inten. [a.u.]

150
100
50

N2
CH

OH
10 watts MW

0

no MW power

3000

3200

3400

3600

3800

4000

4200

4400

Emssion Wavelength [A]

OH Rotational Temperature [K]

3 00

3300

3000

2700
60SCCM 0.9
100SCCM 0.9
60SCCM 1.1
100SCCM 1.1

2400
Stage I

2100
0

Stage II

5

Stage III

10

15

20

25

30

Microwave Power [watts]
Figure 4.5 Emission spectra from CH4/O2 flame with no microwave radiation and 10 W
microwave radiation (top) and OH rotational temperature at the same conditions as in
Figure 4.2 (bottom).

65

The OH rotational temperature is calculated using a more detailed scan over the OH
with a resolution of 1/30 nm and higher gain (not shown here). The general range and
trend of the temperatures reported here agree fairly well with a previous measurement
using CH emission in a similar torch configuration [71]. As the signal is from chemically
excited species, the emission intensity is a relative measure of concentration. Although,
temperature corrected concentrations of excited-state species are not reported here, it is
clear from Figure 4.5 that even without relevant temperature corrections, for 10 W of
microwave power, the emission is generally higher and excited state species, such as OH
and CH, have been shown previously to play an important role in the flame chemistry [29,
72]. Temperature measurements using CH emissions were also measured for comparison
and were shown to generally agree with the OH rotational temperatures (described below)
and in the same range. The main choice for using OH is due to the fact that distortion
from self absorption is lower (smaller cross-section), and more importantly, because OH
is more pervasively present past the reaction zone than CH which occupies a small
portion of the flame front (note that the flame is very small).
For the reported temperatures, the OH population density is determined under the
assumption of Boltzmann distributions. The emission intensity I can be expressed
according to the following equation [73]:

Ej
I
ln(
)=−
+ ln C ,
kT
λ 4S

66

(4.1)

where λ is the emission wavelength (nm), S is oscillator strength (Honl-London factor)
-1

(a.u.), Ei is the rotational energy of the initial level of the line (cm

or eV) , k is the

Boltzmann constant, C is the proportionality constant, and T is the rotational temperature.
Four transitions (R2(14) (307.1145 nm), R2(15) (307.3028 nm), R2(4) (307.4369 nm),
and R2(3) (307.7028 nm)) and associated constants (λ,S, EJ) are taken from [74, 75] for
the rotational thermometry.
Table 4.1 Transitions for OH rotational Temperature Method
-1

λ (nm)

S (a.u.)

EJ (CM )

R2(3)

307.7028

9.1

32643.83

R2(4)

307.4369

12.8

32778.99

R2(14)

307.1145

63.2

35913.21

R2(15)

307.3028

57.2

36394.95

The overall error is expected to be about ±100 K [74], including the fitting error.
The OH* temperatures reported here are meant to provide understanding of the change in
temperature through the various stages of plasma coupling, and we acknowledge that the
rotational temperatures reported here may vary from the actual translational temperature
due to effects such as self absorption [76] and chemical pumping mechanisms [77] which
may preferentially populate high lying rotational states in the electronically excited OH.
For microwave field enhanced combustion (Stage 1), there are different reports in
the literature about the temperature change: In some cases about ~100 K temperature

67

variations were observed [78] whereas in other cases measurements showed no increase
in rotational temperature [70].

We observed an OH rotational temperature increase of

around 100 to 150 K, which we believe is directly responsible for the observed increase
of both flame speed and flammability. Overall, the OH rotational temperature increases
with microwave power input, which confirms the results from a previous study using
CH* [71].

-25.0

T=2900K

-26.0

4

ln (I/r S)

-25.5

-26.5
-27.0
-27.5

34000

36000
-1

EJ [cm ]

Figure 4.6 Sample Image to show how to obtain OH rotational temperature at condition
of 100 SCCM total flow rate and equivalence ration of 1.1. The OH rotational
temperature using this method is 2,900 K.
The discharge power noted here is not the exact power coupled into the flame.
Much of it is lost along the transmission lines and volumetric spatial microwave field
distribution full of the whole cavity which is larger than the flame. We estimate that
about 40% of the power is coupled into the flame considering the electrical components
of the circuit and mode of microwave propagation. Assuming constant gas composition,

68

specific heats (products), and using rotational temperature as the overall global
temperature, about half of the coupled energy is going into the temperature increase, and
therefore a significant portion is coupled into the chemical energy conversion via kinetics
effects. The plasma energy is calculated as a fraction of the total combustion power (5%
to 10%). This fraction decreases with increasing power.
We acknowledge that the use of rotational temperatures over a wide range of
plasma powers is subject to error and unintended influence from non-equilibrium effects.
Other optical measurements such as Rayleigh scattering can be considered in the future to
corroborate these findings.

4.4 In-Situ Fuel Reforming
One mechanism to explain the enhancement by the plasma discharge is fuel
reforming. Fuel reforming, or fuel break up or fuel fragmenting is a process where
hydrocarbon molecules are broken into syn-gas (hydrogen and carbon monoxide). Fuel
reforming in PAC area was mostly conducted by Kim and et al at Stanford University.
They used a USRD discharge, finding that flame stability is most and only sensitive to
the concentration of CO and H2 in the system [9]. In another study, it is reported that the
depletion of H2 and CO leads to flame instability. Flame enhancement for the H2 flame is
not observed for the same conditions under which the hydrocarbon flame is enhanced
[79].

69

Figure 4.6 the shows CO number density in the reaction zone of the CH4/O2 flame
vs. microwave powers for total flow rates of 60 and 100 SCCM and for φ=0.9 and 1.1
from the two-photon absorption LIF (TALIF) images. The region of interest (displayed in
Figure 4.6) is in and around the reaction zone of the tip of the flame where the CO in
concentration is highest.
Overall CO number density increases with microwave power for all conditions. In
Stage 1, the in-situ fuel reforming (to hydrogen and carbon monoxide) as well as the
temperature increase is expected to contribute to the enhancement in flammability limits
and flame speed. It appears from Figure 4.7 that the increase in CO is most dramatic near
the intersection of stages 1 and 2, and tends to be linear towards the end of stage 2 and
into stage 3. This is because the fraction of microwave power that goes into fuel
reforming is larger near stages 1 and 2, whereas more power is directed towards
ionization of gases in stages 2 and 3.
CO production in rich flames is generally higher than in lean flames. This is
confirmed in the first data point (0 W microwave power) of all four series. But in Stages 2
and 3 when microwave power is higher, CO number density at 100 SCCM with φ=0.9 is
higher than that at 60 SCCM with φ=1.1. This is because in the higher microwave power
range, the effects of fuel breakdown from the plasma discharge surpass the thermal CO
production in the flame. With lower flow rates, the microwave energy distribution per
particle will be larger. Therefore, at lower microwave powers (Stage 1), CO production is
higher for rich conditions while at higher microwave powers (Stages 2 and 3), CO
70

production for lower total flow rate is higher than that at the higher flow rate. Also at
higher flow rate, the residence time (during which the mixture gas reacts with the
discharge) is shorter than that at a lower flow rate.

-3

CO Number Density [10 cm ]

4.5

17

4.0
3.5
3.0
60SCCM 0.9
100SCCM 0.9
60SCCM 1.1
100SCCM 1.1

2.5
2.0
Stage II

Stage I

Stage III

1.5
0

5

10

15

20

25

30

Microwave Power [watts]
Figure 4.7 CO number density in burnt gas of CH4/O2 flame vs. microwave powers at
the same conditions as in Figure 4.2 using CO Two-Photon Laser Induced Fluorescence.
It is of great interest to quantify of the degree of fuel reforming assisted by the
plasma. Syn-gas generation has other practical implications in terms of hydrogen
17

generation [80, 81]. The CO number density increases to 3.88 X 10
17

-3

microwave power of 30 W, from 2.57 X 10 cm

71

-3

cm

at the highest

without microwave at the same flow

condition (100 SCCM and φ=1.1). Mole fraction was calculated using the rotational
temperature in the previous section. CO mole fraction increases approximately to 15%
from 9% without microwave in the same condition. Assuming that the water is in vapor
state, in our configuration, 18% of the total methane undergoes reforming to syn-gas at
30 W of microwave power. This is significant considering the competing oxidation
reactions which subsequently lead to CO2.

72

Chapter 5 Combustion Dynamics in a DC Plasmatron
5.1 Premixed and Non-Premixed Modes
The discharge system used in this chapter is a transient arc direct current (DC)
plasmatron [26, 60, 82, 83] from ‘Applied Plasma Technologies’ and is shown in Figure

Main gas feed Co-flow feed
(premixed) (nonpremixed)

5.1.

Anode

100
mA

arc

Cathode

0
mA

400
mA

flame

400
mA

Premixed Nonpremixed

Figure 5.1: Schematic of transient arc plasmatron with cathode in the center and two
different inlets are used to get two different experimental setup(left); Images for
premixed (100 mA and 400 mA DC current) and nonpremixed (0 and 400 mA DC
current) flames of methane/air with equivalence ratio of 0.9.
The plasmatron exit has dimensions of 5 mm for the inner diameter, and the arc
chamber is approximately 50 mm in length. This discharge system operates on a

73

repetitive glow-to-spark transition mode, offering the advantages of a thermal
plasmatron with low average power output, low average temperatures (600−1300 K),
and high electron density. The torch operates primarily in the glow mode where high
voltage (~10 kV) is applied between the cathode (inner electrode) and the surrounding
anode. During the glow-to-spark transition, a short duration spark (~100 ns) [60] or
diffuse channel arises and propagates towards the torch tip. Since the overall
temperature is low, however, electrode erosion is significantly lower than with
traditional DC torches, and the torch can withstand prolonged operation without internal
cooling.
There are two gas inlets in this discharge system: a lower one is connected directly
to the arc chamber, and a top inlet provides co-flow around the main exit of the torch, as
shown in Figure 5.1. In the premixed mode, a mixture of fuel and air flows through the
lower inlet, while a nitrogen co-flow is sent through the top inlet to further stabilize the
flame and cut off entrainment of atmospheric oxygen. In the nonpremixed mode, air
flows through the lower inlet (to generate the discharge) and methane through the top
inlet. As this is an igniter for a combustor, it is designed for high flow with plasma
stabilization. Therefore, most of the flowrates used in this study are conditions that are
too high for unassisted stabilization; however, a few low power nonpremixed mode
settings where natural stabilization can be achieved are also employed, as shown in
Figure 5.1 (far right for the nonpremixed flames). Here, images for premixed (100 mA
and 400 mA DC current) and nonpremixed (0 and 400 mA DC current) flames of
74

CH4-air are shown. It can be seen that in both modes, the height of the illuminated
region is increased with higher DC current due to the presence of a plasma plume. For
premixed flames, the reaction mainly occurs inside the chamber at low DC currents,
while at higher currents the discharge plume is observed outside the torch as the plasma
is not fully quenched inside the cavity. For nonpremixed flames, the combustion occurs
entirely outside the chamber (since there is no fuel in the plasma chamber), and the
center region shows a bright white hue as the plasma plume interacts with the flame.
The anchor of the nonpremixed flame can be seen at the base of the plasma plume. Both
the size and luminosity of the plume increase as a function of the input power.

5.2 Impact on Flame Structure
Single and ensemble averaged OH Planar Laser-Induced Fluorescence (PLIF)
images are shown in Figure 5.2 as a function of increasing microwave power for both
premixed (top) and nonpremixed flames (bottom). Average images were based on
acquisition of 200 instantaneous images.

75

400
mA

600
mA

100
mA

200
mA

400
mA

600
mA

0
mA

200
mA

400
mA

600
mA

0
mA

200
mA

400
mA

600
mA

3*1016cm-3

200
mA

0

cm

3

0

Single
Average
Single
Average

Nonpremixed Flame

Premixed Flame

100
mA

Figure 5.2: Single and averaged 2-D OH Planar Laser-Induced Fluorescence (PLIF)
images as a function of increasing microwave power for both premixed flames (top)
and nonpremixed flames (bottom).

Here, the total flow rate for the reactants is 9 standard liters-per-minute (LPM,
referenced to STP conditions), and the equivalence ratio is φ=0.95. The equivalence
ratio in the nonpremixed case is defined here as the ratio between the fuel and oxygen in
the inner nozzle. At these conditions, combustion power from the oxidation of the fuel is
st

about 500 Ws using the lower heating value of methane. For 1 order comparisons,
measurements using a thermocouple show that temperatures for premixed flames are

76

about 100 to 200 K higher than in nonpremixed flames for the same fuel flowrates. For
premixed flames, the visible flame luminosity and plasma volume increase with
increased current/power. The range of current is between 100 and 900 mA in our
research, resulting in a discharge power of 150 to 750 W. At low powers (~100 mA), the
OH signal is lower than that in a typical premixed flame, presumably because a
significant portion of the oxidation reactions occur inside the arc chamber, where the
mixture initially comes in to contact with the discharge. When the DC current is higher,
the reaction rates are expected to increase, resulting in a more compact flame structure
especially as a larger portion of the fuel is oxidized inside the chamber. However, the
introduction of added species and charged particles with higher momentum (and new
reaction chemistry involving ions and electrons) provides new chemical pathways. This
ultimately results in a higher concentration of OH radicals, via reactions such as O
3

( P)+H2O

1

OH+ OH, O( D)+H2O

OH+ OH [68].

There are many kinds of thermocouple, each made from a unique metal or alloy to
achieve its temperature measurements range and application. Also there are different
diameters for the bead: on the one hand a bigger bead can last longer, on the other hand
a bigger bead leads to higher error and interference with the measured volume. Also,
different types of thermocouples have dramatically different measurements ranges, for
example the B type can measure a reading range of 300 K - 2000 K whereas it is 0 K 1500 K for a K type thermocouple. What is needed in this research is one that can
measure high temperature, with a medium size of bead diameter and a longer lifetime
77

for the thermocouple. For the readings of thermocouple measurements, calibration
should account for systematic errors from emission to the wall. Calibration for different
types with different bead diameters is shown in Table 5.1. In our experiments, the B type
thermocouple with a 0.02 inch diameter was used for its relatively long life and
moderate size bead.
As can be seen in Figure 5.2, nonpremixed flames show a unique two-cone flame
structure: within the inner cone the main oxidation of the fuel with the air from the
plasma discharge takes place; within the outer cone combustion with the ambient air
occurs. Overall, the height of the flame is greater for the nonpremixed mode than that
for the premixed case while the flow appears to be more turbulent in the premixed
flames due to the shorter main reaction zone. The outer cone flame in the nonpremixed
mode can be sustained without any DC plasma power input for some of the lower flow
rates used. In the nonpremixed flame, the reaction zone for the outer cone remains
relatively constant as power is increased. However, the inner cone is dramatically
increased as highly reactive air interacts with the fuel. At powers exceeding 400 mA, the
OH LIF intensity for the inner flame front is much higher than the outer one. Flame
stability is dominated by the plasma discharge.

78

Table 5.1 Thermocouple Measurements and Calibrations
T Measured

B 0.020”

B 0.032”

J 0.032”

K 0.032”

100 K

N/A

N/A

N/A

99.95 K

200 K

N/A

N/A

N/A

199.95 K

300 K

300 K

300 K

300 K

300 K

400 K

400.12 K

400.15 K

400.15 K

400.15 K

500 K

500.41 K

500.53 K

500.53 K

500.53 K

600 K

601.01 K

601.30 K

601.30 K

601.30 K

700 K

702.10 K

702.69 K

702.69 K

702.69 K

800 K

803.92 K

805.02 K

805.02 K

805.02 K

900 K

906.80 K

908.71 K

908.71 K

908.71 K

1000 K

1011.13 K

1014.24 K

1014.24 K

1014.24 K

1200 K

1226.16 K

1233.48 K

N/A

1233.48 K

1400 K

1454.09 K

1469.23 K

N/A

1469.23 K

1600 K

1701.82 K

1730.31 K

N/A

N/A

1800 K

1978.33 K

2028.24 K

N/A

N/A

1900 K

2130.82 K

2195.43 K

N/A

N/A

2000 K

2294.98 K

2377.54 K

N/A

N/A

2100 K

N/A

N/A

N/A

N/A

2200 K

N/A

N/A

N/A

N/A

79

5.3 Impact on Radical Concentration
The number density of the OH was calibrated from the LIF signal using a
Hencken Burner with a near-stoichiometric methane-air mixture in the weakly excited
OH fluorescence regime. The uncertainty of the resulting OH number densities is
estimated as ±20% of the indicated value. This accounts for the uncertainty of the
calibration number density (from the original absorption measurements) and for the
application of the calibration to these flames (accounting for differences in quenching
rates, temperatures, laser sheet intensity, etc.). In Figure 5.3, the OH number density is
plotted for different DC currents, total flow rates and equivalence ratios for the
premixed mode. In the left portion of Figure 5.3, the equivalence ratio is fixed while
DC current and total flow rate change. The data points are taken at a fixed height in the
center of the reaction zone (shown in the insert) where OH LIF is most intense. For low
total flow rates (6 LPM), the flow is not strong enough to propagate the nano-second arc
discharge to the end of the plasmatron chamber, and most of the fuel is partially
oxidized inside the chamber resulting in a low OH concentration above the tip. It is clear
that with higher DC currents (higher DC power), the OH number density increases
almost linearly with the input power. For different total flow rates, it is noticeable that
16

all OH number densities at 100 mA are less than 10

-3

cm , which is the typical OH

production level in a Bunsen flame without plasma enhancement. This is an indication

80

that the plasma is mostly acting to dissociate and preheat the fuel inside the arc chamber
and that combustion outside is generally not influenced by the discharge.
It can be seen from Figure 5.3 (top) that OH is highest with an equivalence ratio of
0.8 and is less at values of 0.9 and 1.0 [67, 69]. This is due to the fact that OH
production in this case is a combination of direct coupling of the mixture inside the
chamber and partial coupling after exiting the torch, and higher OH concentration in
lean conditions indicates that the oxidation of lean mixtures extends further out to the tip.
The increase of OH number density with DC power levels is a direct influence of the
plasma chemistry such as the additional oxidation of product water vapor (H2O) by
electron impact. H2O plays an active role in the production of active radicals through
1

direct electron impact disassociation: e+H2O→H+OH+e; and reaction with O( D)
1

1

disassociated from O2: O( D)+H2O→2OH, e+O2→O( D)+O+e [84]. These radicals in
turn result in increased OH production which can offset termination reactions when the
DC power is raised.

81

Data Point

16

-3

OH Number Density [10 ]
cm

2.5
2.0

6 LPM
9 LPM
12 LPM
15 LPM
18 LPM

1.5
1.0
0.5
0.0
200

16

-3

OH Number Density [10 ]
cm

1.4
1.2

400
600
DC Current [mA]

800

400
600
DC Current [mA]

800

phi=0.8
phi=0.9
phi=1.0
phi=1.1
phi=1.2
phi=1.3

1.0
0.8
0.6
0.4
200

Figure 5.3: OH number density is plotted for different DC currents, total flow rates
(top, equivalence ratio is fixed at 0.95 and for different DC current, equivalence ratios
(bottom, total flowrate is fixed at 15 LPM) in premixed mode.

82

A first order comparison of OH densities for both premixed and nonpremixed
flames is shown in Figure 5.4. In this plot, the OH number densities are shown at the
same flow rates, but with increasing levels of power input for both flame geometries.
For the premixed flame, the data is extracted from a central location identical to that
shown in Figure 5.3. For the nonpremixed flame, the data is extracted from two
locations, representing the inner and outer cones, as shown in the insert of Figure 5.4.
OH number density in the premixed mode increases by almost 100% from 100 mA
to 900 mA of DC current input. In contrast for nonpremixed flames, OH number density
remains relatively stable on the outer flame front and is close to the Burner flame OH
production rates, which indicates that outer flame has little influence from the inside air
discharge. However, the OH number density in the inner flame front shows an
incensement of about 750% percent and the highest OH number density reaches over 3 ×
16

10

-3

cm , which is more than twice that of the OH production of a premixed flame for

the same condition. This shows that much of the flame enhancement for the
nonpremixed flame is carried out at the inner flame front, as the power is enhanced. For
the inner flame, the significant increase in OH number density comes from the active
excited species in the air discharge exiting the plasma chamber, including atomic
oxygen, which has been previously documented as a significant factor for flame
enhancement [7] and a cause for increased OH concentration.

83

16

-3

OH Number Density [10 cm ]

3.2
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2

Premixed Flame
Nonpremixed (outer)
Nonpremixed (inner)
Inner Outer

0

200

400

600

800

DC Current [mA]
Figure 5.4: OH number density with the same flow conditions (total flow rate is 9
SLPM, and φ=0.95) but different currents for both premixed (from data point the same
as in Figure 5.3) and nonpremixed (two data points, one outer flame front and the
other one inner flame front as shown) flames.

5.4 Flame Stability, CO Formation and Energy Efficiency
The influence of geometry and other parameters on the flame stability can also be
observed from OH PLIF images. Figure 5.5 shows averaged OH PLF images for both
premixed and nonpremixed flames at various flowrates and equivalence ratios. The top
row shows a premixed flame with a rate of 12 SLPM, with equivalence ratios of φ= 0.7,

84

1.0 and 1.3. The bottom two rows show nonpremixed mode flowrates of 6 SLPM
(middle)/15 SLPM (bottom) with overall equivalence ratios of φ= 0.7, 1.0 and 1.3. In the
premixed mode, the figures represented here are generally reproduced for all the
flowrates where the flame is stable at the end of the plasmatron tip for leaner conditions;
however, the flame tends to lift off at higher equivalence ratios (φ=1.3, top row in Figure

φ=1.0

φ=1.0

φ=0.7

φ=1.0

cm

φ=1.3

0

Nonpremixed Nonpremixed
15 LPM
6 LPM

φ=1.3

3 0

φ=0.7

φ=1.3

Premixed
12 LPM

φ=0.7

3*1016cm-3

5.5).

Figure 5.5: Averaged OH PLF images for all flowrates at premixed mode with phi=
0.7, 1.0 and 1.3 (top); Averaged OH PLF images for 6 LPM (middle)/15 LPM
(bottom) at nonpremixed mode with hi= 0.7, 1.0 and 1.3.
At high equivalence ratios, the radical pool is poorly established and fuel/air
mixture tends to react more with the ambient air resulting in combustion downstream of
the exit. Nonpremixed flames at low (6 SLPM, middle row in Figure 5.5) do not reflect
this phenomenon, and the flame is stable over a wide range of equivalence ratios. For

85

higher total flowrates (15 SLPM, bottom row in Figure 5.5), the stability of the
nonpremixed flame is compromised; the outer flame is no longer visible in the turbulent
structure and the two flame zones merge into a single structure. The stability for these
two configurations is different in that the premixed mode appears to be more chemistry
driven and dependent on the equivalence ratio whereas the nonpremixed mode is more
dependent on the flowrates and the physical interaction of the fuel and the plasma
discharge. Overall, the premixed mode can stabilize higher flowrates, but the smaller
dimensions of the combustion and sensitivity on the fuel and air ratios can render the
premixed mode more difficult to use.
The CO number density is measured using two-photon-absorption LIF (TALIF) to
investigate the in-situ reforming into carbon monoxide (CO) and hydrogen (H2), which is
believed to be a key mechanism for flame stabilization [9]. CO was measured with
1 +

1 +

two-photon PLIF, using the B Σ −X Σ

(0,0) transition at 230.10 nm [63]. A Sirah

Precisionscan LG-2400 dye laser was pumped by the 532 nm beam from a Q-switched
Nd:YAG laser. The 653-nm output of the dye laser was sum-frequency-mixed with the
third harmonic beam (355 nm) from the injection seeded Nd:YAG laser to produce an 18
-1

mJ/pulse of 230-nm radiation (linewidth of ~ 0.1 cm ). CO fluorescence is then detected
at 486 nm using an intensified CCD camera fitted with an interference filter (10 nm
FWHM).
In Figure 5.6, averaged CO PLF images for premixed (insert top row) and
nonpremixed (insert bottom row) are shown, and CO PLIF intensity are plotted for both
86

modes with equivalence ratios of φ=0.7 and 1.1. Due to the two-photon excitation of the
CO, the quality of the LIF signal for the non-linear process is not as high as that of OH.
Coupled with a more diffused profile of CO, the LIF signal did not resolve the two cones
of the nonpremixed flame. It should be noted that because of the weak absorption, the
laser sheet is smaller and more focused near the center, therefore a dramatic drop of CO
downstream is an artifact of the laser sheet.
For the lower portion of Figure 5.6, the CO PLIF signal intensity is corrected for
the quadratic dependence [85] on the laser power to reproduce a relative CO number
density (a.u.). since the irradiance is not high enough for photoionization to exceed
quenching in our case to influence the quadratic dependence [86]. We assume that the
increase of CO in the reaction zone is indicative of in-situ fuel reforming, which may
not be reflected in the overall byproduct as conversion to CO2 occurs. For both modes,
the CO LIF intensity is shown to increase with increased DC current input. It is clear
that for nonpremixed and premixed flames, CO LIF intensity for rich flames (φ=1.1) is
higher than lean flames (φ=0.7).
Comparison between CO number densities of premixed and nonpremixed flames
with an equivalence ratio of 0.7 indicates that higher fuel reforming is observed for the
nonpremixed flame over the entire range of powers used. This is expected as a very rich
interface is established in the nonpremixed flame between the plasma discharge and the
fuel.

For both flame configurations, we see a gradual rise in CO as the power is

increased. This additional CO generated from the in-situ reforming of the fuel will
87

contribute as a major mechanism for flame stabilization by increasing the mixture
hydrogen content. We assume that the CO increase also indicates temporary increase of
hydrogen. Increase of CO is about 30 a.u. in the premixed flame and 40 to 50 a.u. in the

160
140
120
100
80

600
mA

200
mA

400
mA

HIGH

400
mA

600
mA

LOW

180

200
mA

cm 1.5

CO Number Density [a.u.]

200

0

Nonpremixed Premixed

nonpremixed mode for the same conditions over the power range 0 - 800 mA.

Nonpremixed φ =1.1
Nonpremixed φ =0.7

60
40

Premixed φ =0.7

20
0

200

400

600

800

DC Current [mA]
Figure 5.6: Averaged CO PLF images for premixed (insert top) and nonpremixed
(insert bottom); CO number density is plotted for with equivalence ratio 0.7 (premixed
and nonpremixed) and 1.1 (nonpremixed).

For plasma assisted combustion, energy consumption (energy needed for the same
amount of air with the same current) during the coupling process is a critical parameter.
This parameter can be analyzed by comparing of both modes with the air discharge case.
Figure 5.7 shows the actual input power as a function of input current under identical
flow conditions. It is clear that most power is required for the air-only plasma because

88

significant energy is required for dissociation, ionization, and vibration excitation of
nitrogen, resulting in significant NOx generation [87]. When combustion is present, the
breakup of the nitrogen is assisted by the exothermic reactions involving the fuel. Thus,
the ionization threshold is lowered. Therefore, the conversion efficiency to H2 and CO is
increased.
In the premixed mode at lower currents, power consumption is lower as fuel is
injected into the discharge chamber and in-situ breakup of CH4 molecules and the
plasma impact reactions further accelerate the reaction chemistry. The study in the next
chapter shows that with combustion, NO production is reduced by an order of magnitude
less, meaning much less energy is used to break the N-N bond in Nitrogen. At higher DC
currents, the power consumption for premixed and nonpremixed flames is seen to
converge as plasma impact and fuel breakup are eventually saturated for both cases.

89

DC Power Input [Watts]

700
600

Premixed Flame 21 LPM
Non-Premixed 21 LPM
Air Discharge 21 LPM
Premixed Flame 15 LPM
Non-Premixed 15 LPM
Air Discharge 15 LPM

500
400
300
200
100
200

400

600

800

DC Current [mA]
Figure 5.7: DC Power input with different current for air plasma, premixed flames and
nonpremixed flames.

90

Chapter 6, Nitric Oxide Formation in a DC Plasmatron
6.1 Nitric Oxide Formation in Flames
Generally, NO formation in conventional combustion is due to four main processes
[13, 14]: (1) thermal (Zeldovich) mechanism which occurs at high temperatures because
of high energy required for dissociation of the nitrogen molecule (reactions 1,2 and 3);
(2) prompt mechanisms with important intermediate species HCN and NH (reactions 4
and 5); (3) N2O mechanism which involves third body collision (reactions 6, 7 and 9);
and (4) fuel-bound nitrogen mechanism where nitrogen is supplied from the fuel itself.
NO can also be removed by reacting with hydrocarbon radicals in rich flame through a
process known as ‘NO reburn’ [15], of which some of the key reactions are also listed in
Table 6.1. NOx emission in a conventional combustion is highly dependent on
temperature as shown in Figure 6.1. When temperature is high (higher than 1,800 K),
most Nitric Oxide from traditional combustion is from Thermal (Zeldovich) mechanism.
Coupling high temperature reactive flows with a nonequilibrium plasma discharge
can lead to significant changes in the formation of nitric oxide (NO), one of the most
problematic combustion effluents and a critical design parameter for all practical
combustion systems. The key mechanisms involved in the NO chemistry for plasma
discharges and methane flames are shown in Table 6.1.

91

Table 6.1: Reactions for NO formation in methane/air flame in plasma
Thermal
(Zeldovich)mechanism

Prompt mechanisms

O+N2 ↔ NO+N
N+O2↔NO+O
OH+N ↔ NO+H
NH+O ↔ NO+H
CN+O2 ↔ CO+NO

Mechanism involving N2O

O+N2+M ↔ N2O+M
N2O+O ↔ NO+NO

Other reactions to form NO

NNH+O ↔ NO+NH
N2+e ↔ N+N+ e

Electron and Ion Reaction

N2O+e ↔ NO+N+e
NO+e ↔ N+O+e
N2O+e ↔ N2+O+e
NO2+e ↔ NO+O+e

1
2
3
4
5
6
7
8
9
10
11
12
13

O+N2* ↔ NO+N

14

Excited State Species
Reaction (*: excited state)

NO2+N2* ↔ NO+N2+O

15

2

N( D)+O2↔NO+O
HCCO + NO ↔ HCN + CO2
HCCO + NO ↔ HCNO + CO
CH + NO ↔ HCN + O
NO Reburn mechanism

CH3 + NO ↔ HCN + H2O
CH3 + NO ↔ H2CN + OH
NH2+NO ↔ N2+H2O

92

16
17
18
19
20
21
22

60

NOx Emission [ppm]

50
40
30
20
10
0
1500

1600

1700

1800

1900

2000

2100

Flame Temperature [k]
Figure 6. 1 NOx emission as f function of combustion temperature [30, 88]
In the case where no combustion is involved and there is only a plasma discharge,
formation of NO from air varies depending on the discharge system. For example,
reaction 14 can be the principal NO formation pathway in pulsed discharge with shorter
-8

-6

time scales (10 −10 seconds) [89] while reaction 16 is shown to be dominant in corona
discharges [90]. For variety of reduced electric fields, different percentage of energy
deposition into air goes to elastic losses, ionization, N2/O2 electronic states, N2/O2
vibration [34], resulting in different pathways for NO formation.
In plasma assisted combustion, analysis of the NO formation mechanism becomes
difficult to determine as energy distribution of individual molecules can deviate from

93

thermal equilibrium and the relationship between the NO production and deposition of
heat is unclear. More importantly, high electron density generated by the electric field
gives rise to new electron and ion impact processes which can enhance the propagation
and branching of radicals and ultimately accelerate NO production. NO formation in
plasma assisted combustion could be formed both from the thermal mechanism due to
high combustion temperature and plasma particle reactions, especially reactions 14, 15
and 16 due to high energy for dissociation of nitrogen. As NO formation is a major
concern for all practical combustion systems, accurate methods to measure NO formation
in the reaction zone is important to the continued development of plasma enhanced
combustion systems, and a task which can benefit from the extensive research already
available on NO formation of both combustion and plasma systems, respectively.

6.2 Quantification of Nitric Oxide
Imaging of NO using LIF is a well established tool in combustion diagnostics [91].
Therefore, a detailed discussion of the relevant spectroscopy is omitted but it should be
noted that we are using an excitation at 226.03 nm which has been previous noted as
being well isolated from interference of O2 [92, 93] even in high-pressure flames where
the problem is most pronounced. This excitation line, composed of P1(23.5),
Q1+P21(14.5) and Q2+R12(20.5) transitions, has a moderate temperature dependence
over typical combustion temperatures between 1700 and 2300 K. In this section, we focus

94

on three other issues of practical importance: quantification of NO concentration, signal
distortion due to thermal nonequilibrium and temperature field imaging.
Quantification of the LIF signal can be a difficult task as it is not only dependent on
NO number density, but also on the optical setup and potential interference from
alternative species. The approach used in this study here is to calibrate the signal using a
well defined flat flame torch using a variable NO seeding method [94]. The calibration
torch is made by fitting a 50 micron sintered plate into a 1 cm diameter stain steel pipe in
order to stabilize a flat 1-D flame and seed various levels of nitric oxide. The LIF is
measured at two different excitation wavelengths, which have minimal and maximum NO
signal strengths, respectively as illustrated in Figure 6.2. The plot shows a magnified
view of the small box in the inserted graph. The NO addition is varied from 200 to 600
ppm and NO-LIF is detected with excitation at two different wavelengths. Excitation 1
and 2 refer to two different excitation wavelengths (λ1: 226.030 nm, λ2: 226.042 nm).
These wavelengths are selected relatively close together in the target scan range
where the baseline intensity, if present, can be assumed to be constant for a given
pressure. After measuring the total signal at two different wavelengths at several different
NO seeding concentrations, the overall signal is linearly fit as a function of NO seeding
for each of the excitation wavelengths. These two linear fits are extrapolated to their
intersection yielding the nascent NO concentration and the value of the baseline signal.
Determination of the nascent NO concentration (and LIF baseline) by this method
assumes the NO reburn mechanism is first order in NO concentration, as evidenced by the
95

linearity in Figure 6.2. In the stoichiometric and lean flames studied here with less than
100 ppm NO addition, the NO reburn is expected to be linear in NO concentration and of
no significant amount [95, 96]. It has also been shown that temperature in typical flames
does not vary (≤2%) for NO additions as large as 1000 ppm [94]. In all the cases, NO is
being weakly excited and no saturation effects are observed with our laser power.

60
250
200

Total Intensity [a.u.]

150
100

40

50
0
-200

0

200

400

600

20

Excitation λ1
Excitation λ2
0
-200

-100

0

Seeded NO concentration ppm
Figure 6.2 NO addition method for determination of the baseline strength and the
nascent NO concentration.

6.3 Temperature Profile Measurements
A 2-D temperature profile is measured using an IR camera and shown in Figure 6.3.
From the temperature profile, we can see a rise of reaction zone when equivalence ratio is

96

increased from lean (0.7) to rich (1.2). Also from the figure, we also can see the
turbulence of the combustion. The IR thermometry is a method for relative 2-D profile
and we need some more accurate way to calibrate for higher accuracy. In this research, I
used multi-line NO LIF thermometry. The data point for calibration is shown in Figure
6.3 with a circle on the upper right.
300K

Ф=0.8

Ф=0.9

Ф=1.0

Ф=1.1

Ф=1.2

2200K

Ф=0.7

Figure 6.3 IR temperature profile with different equivalence ratio for premixed
methane and air flame with circle shown in the figure where the multi-line
thermometry calibration is used.
Another effect which can significantly influence the LIF signal is the nonequilibrium
distribution of the rotational states in the NO. Generally, for most plasma enhanced
combustion systems at atmospheric pressure, it is assumed that the NO in the flame
region will follow a Boltzmann distribution of the rotational states and therefore in
thermal equilibrium. In our studies, this was tentatively confirmed by comparing signal
strengths from a number of varying rotational states, where the intensity generally

97

followed the nitric oxide concentration within 2%. However, for systems where highly
nonequilibrium plasma energy is coupled at lower pressures, the impact of deviation from
a thermal equilibrium of the rotational energy Erot on the LIF signal should be
considered.

Signal Intensity [a.u.]

800
Experimental Data
Multi-line fit

600
400
200
0
226.00

226.04

226.08

Excitation Wavelength [nm]
Figure 6. 4 Multi-line NO LIF thermometry (126 transitions between 225.975 nm and
226.1 nm).
To accurately calibrate the signal intensity for both concentration and spatial
variation, we need to acquire information regarding the temperature. It is needed to
correct for the temperature variation of the NO LIF signal via the temperature
dependence of the laser-excited ground state population, the spectral overlap between the
laser-spectral profile and NO absorption spectrum, and the fluorescence yield. In this
study, we use IR thermometry for determination of temperature fields, which measure
temperature using gray body radiation (infrared) emitted from reacting gases over the

98

whole viewing area. IR emission from excited species also adds to the intensity of the
figure, but it would be lower than gray body radiation and also an rough indication of
temperature. The images are then calibrated at a single point using a multi-line NO LIF
thermometry [97] which is based on excitation of multiple transitions by scanning the
wavelength of the laser over a wide spectral region and then fitting the excitation profile
with a numerical simulation. An example of the multi-line NO LIF thermometry method
is shown in Figure 6.4. This provides highly accurate temperatures at a single point which
can be used to correct for the less accurate temperature images obtained through the IR
thermometry.
The model used for the fitting is LIFSim [53]. LIFSim is a multi-level spectroscopic
simulation of nitric oxide for the A–X transition during laser absorption and spontaneous
+

fluorescence. It is particularly well calibrated for the A2Σ −X2Π (0,0) transition where
date for term energies, transition energies, transition strengths, collisional line broadening
and shifting, quenching, and energy transfer terms are well known. For use of LIFSim in
our studies, we must assume thermal equilibrium (or understand subsequent internal
energy modes) of the flame, which can be made as the reaction zone outside the torch is
far from the arc chamber where all the non-equilibrium effects are dominant. LIFSim is
also used to account for temperature dependence of the LIF signal as will be discussed in
the following sections.

99

6.4 Planar Laser-Induced-Fluorescence (PLIF) Imaging of NO
2-D NO PLIF images are taken for discharge only in air without combustion over a
range of flow rates and discharge powers. For plasma enhanced methane/air combustion,
images are taken as a function of flow rate, discharge power and equivalence ratio. The
laser sheet propagates through the center of the plasma torch exit and the emission is
synchronized with the intensified CCD camera. In order to correct for the laser energy
fluctuations, every image is calibrated by the corresponding laser pulse energy, collected
using a fast speed photodiode (LaVision). As the gas flow is turbulent, 200 images were
averaged when comparing the NO formation for different operating conditions. Averaged
images also allow us to accurately calibrate the signal over the imaging field as a function
of temperature by using averaged IR thermometry.
2-D NO PLIF signal intensity is not only affected by the laser pulse power, but also
the nonuniformities in the laser sheet profile and from the spherical aberration of the lens.
Emission from a fluorescent UV glass was used for the laser sheet corrections, and the
spherical aberration is corrected using an algorithm from the Davis software (LaVision)
based on a pre-shot grid. Both single shot and averaged 2-D NO concentration images
from NO PLIF of discharge only in air as a function of different plasma power and
stoichiometric methane/air flame with different plasma power are shown in Figure 6.4. In
both figures, you can observe the two plasma plumes coming out from either side of the
torch.

100

18000

2.0
1.5

ppm

Discharge Only
Distance [cm]

2.5

1.0
0.5

0

0
25SCFH 400mA

25SCFH 200mA

25SCFH 300mA

25SCFH 400mA

3000

25SCFH 300mA

2.5
2.0

ppm

1.5
1.0
0.5

0

0

18000

(a)

2.5

20SCFH 400mA

2.0
1.5

ppm

Discharge Only
Distance [cm]

25SCFH 200mA

25SCFH 100mA

Discharge + Flame
Distance [cm]

25SCFH 100mA

1.0
0.5

0

0
25SCFH 300mA

25SCFH 400mA

25SCFH 200mA

25SCFH 300mA

25SCFH 400mA

3000

25SCFH 200mA

25SCFH 100mA

2.5
2.0

ppm

1.5
1.0
0.5

0

Discharge + Flame
Distance [cm]

25SCFH 100mA

0

(b)

Figure 6.5 Single shot NO concentration fields using NO PLIF of gas discharge of air
and plasma enhanced flames with different plasma power at same gas flow rate (a).
Averaged NO concentration fields from NO PLIF (200 images averaged) of plasma
enhanced flames with different plasma power at same gas flow rate (b).

101

It can be seen from the Figure 6.5 that turbulent flows exists in both discharge-only
and plasma enhanced combustions for our plasmatron. The degree of instability is
observed to be more severe in plasma discharge only cases for the same conditions, as the
plasma assisted combustion is more energetically driven by the additional chemical heat
release. For the single shot images, it can be seen that the plasma only case (upper) row
has images which are smoother in quality due to the higher levels of the NO signal which
results in higher Signal to Noise Ratio (SNR). Additionally, it should be noted that NO
formed near the flame front is diffused along with the flow and is reduced about 30% to
40% from the highest concentration to a distance of 2.5 cm from the plasmatron exit.

6.5 Temperatures and NO Concentrations in Plasma Discharge Only
In order to compare the NO formation between different operational conditions, we
present both temperature and NO concentration at a specific location in our image as
shown in Figure 6.6 for a plasma discharge only case. The location was chosen as a
single point in the center of the flame where a stable and high temperature region exists
due to the convergence of the two plasma plumes from either side of the torch. In order to
obtain the temperature at this location, multi-line NO LIF thermometry is used for higher
accuracy.

102

1400

20SCFH
30SCFH
40SCFH
50SCFH

Temperature [K]

1200

1000

800

600

Data Point Location

0

200

400

600

800

1000

(a)

Current mA

NO concentration [ppm]

17500

20SCFH
30SCFH
40SCFH
50SCFH

15000

12500

10000

7500
0

200

400

600

800

1000

Current [mA]
Figure 6.6 Temperature (a) and NO concentration (b) data at specific location of the
discharge for different power levels and air flow rates in discharge only case in air.

103

NO concentration in the discharge plasma is quantified by the following process.
First the NO LIF intensity value (calibrated for both laser sheet nonuniformity and laser
pulse energy) is taken from an averaged image of 200 laser shots. Using the temperature
information provided from NO-LIF thermometry, the signal is calibrated for temperature
dependence of the NO excitation by using a multi-level excitation and fluorescence
model, LIFSim [53]. Note that we have verified that we are operated in the linear (weak)
excitation region and free from saturation effects. The LIF signal at this point is linear to
3

the number density of NO (molecules/cm ), and we apply the following relationship to
convert it into a mole fraction based concentration (molecules/total molecules). The final
signal is then quantified using a NO addition method previously described.

I LIF ~ N NO = X NO ×

p
kT

（6.1）

ILIF is NO LIF signal intensity, NNO is the number density and XNO is the mole
fraction of NO.
The uncertainty in the temperature measurements affects the concentration
measurements through the mole fraction/number density correction in LIF equation, and
also through calibration of the LIF signal temperature dependence using LIFSim. The
temperature uncertainty is within 1% for our measurement techniques at atmospheric
pressure. This is because the quenching terms in LIFSim for the low temperatures as that
in the plasma discharge has larger errors. The uncertainty associated with beam energy
fluctuation corrections and wavelength shift of the laser is expected to be is less than 1%.
104

Camera readout error was suppressed by subtracting a background image taken without
the laser pulse and can be neglected. Another component of uncertainty comes from the
fact that the quenching coefficient used to determine the LIF strength in LIFSim lacks
accuracy at low temperatures in discharge only case (<5%). Combined with the errors
from the flowmeters and shot noise (due to statistical fluctuation of finite number of
signal photos reaching the CCD array), the overall uncertainty is expected to be ~7.4% in
the NO concentration measurements for plasma discharge and ~4.7% for plasma assisted
combustion.
Figure 6.6 shows temperature and NO concentration plots with different gas flow (20
to 50 SCFH) and average current (100 to 900 mA). The temperature and NO
concentration data is extracted from a region 10 mm from the top of the plasma torch
with an area of 2.56 mm X 2.56 mm (40 pixel X 40 pixel with resolution of 64 um X 64 um
per pixel). The amount of NO production is significantly large due to the large energy
input (200 to 700 Watts). As the temperature range for all conditions lies within 550 to
1300 K, NO production from the thermal mechanism is considered to be small in this low
temperature range. Moreover, the system lacks carbon and hydrogen compounds required
for prompt NO production. Therefore, we can conclude that the possible dominant NO
production mechanism is the electron and excited state species (i.e., oxygen) impact
reactions which lead to dissociation of N2 and also possible NO formation reactions from
N2O. In Figure 6.6 (b), NO production increases with current/power and inversely with
air flow. The same general trend is shown in the temperature plot. As formation of NO
105

from N2O is expected to increase as temperature is decreased, it seems to be of little
impact here indicating that the most dominant mechanism for NO production is the
electron, ion, and excited state species impact reactions.
It is apparent in Figure 6.6 (a) that temperature increases with current, as a result of
more energy per volume of gas. At the same time, temperature drops when the air flow is
raised as the gas resident time in the discharge is reduced, leading to fewer collisions
between excited particles and eventual lowering of the temperature even though the
discharge power is actually higher at higher flow rates due to an increase in the average
voltage. The voltage is shown to increase with increased flow rate (not plotted here) in
our experimental setup where the flow rate is increased from 20 SCFH to 50 SCFH,
corresponding to a gas speed increase from 8 m/s to 20 m/s. For all the gas velocities used
here, the plasma plume is injected outside the torch and visually accessible although the
initial breakdown is expected to occur inside the torch near the cathode.
-5

For a typical DC discharge, a range between 10

to 1 A is the normal glow regime,

where the voltage will increase as the current increases [98]. Although it is not shown
here, we have noticed that the voltage drops in our plasmatron when current is increased
for the same flow rate, which leads us to assume weaker electric field but more spark
channels to maintain the current. Discharge energy input to the system is not only from
the electric field, but also thermal deposition from the higher temperatures. Table 6.2
shows the percentage of energy which goes to heating the gas flow at 20 SCFH in a
current of 100 to 900 mA. Energy for heating the plasma gas is calculated using specific
106

heat Cp of air and assumes that temperature of the input gas is room temperature of 300
K. It can be seen that the percentage for heating the plasma gas increases with increased
current which also means an increase total power input. The increase in the kinetic
energy of the molecules and increased collisions further support that NO formation is
predominantly from electron and ion impact reactions.
Table 6.2. Energy data of discharge only in air with different current
Current

Final
Temperature

Voltage

Discharge Power for Heating
Power

heating Percentage

mA

K

kV

J/s

J/s

%

100

540

1.95

195

48

24.5

150

630

1.55

232.5

65

28

200

710

1.32

264

83

31.3

300

830

0.99

297

112

37.6

400

920

0.8

320

132

41.3

500

990

0.68

340

148

43.5

600

1050

0.6

360

163

45.3

700

1080

0.54

378

171

45.2

800

1170

0.49

392

193

49.2

900

1210

0.44

396

203

51.2

107

Voltage drop across the plasma column (from the end of the inner cathode) is about
500 to 1000 voltage for all the combustion conditions in this study and the length of the
plasma column is about 4 mm. Therefore, the reduced electric field (E/n) is calculated to
-17

be 50~85 Td (1 Td =10

2

V·cm ) for 500 to 1000 voltage, 1 atm and 4 mm. Reduced

electric field is a measure of kinetic energy transfer cross-section from electrons to
internal degrees-of-freedom molecules. For certain reduced electric field, appropriate
molecular energy states is excited. For example, vibrational levels of the ground
electronic state nitrogen, or low-energy electronic states of oxygen are more likely to be
excited at lower E/n level, while high-energy electronic levels of nitrogen and oxygen, as
well as in their dissociation and ionization for higher E/n. For this experiment, majority
of the energy deposition into the air results in N2 molecule vibration (90% to 95%)
excitation along with some N2 electronic states (~5%) while there exists little ionization,
rotation or O2 vibrational/electric excited states [34]. Therefore, the main pathway
responsible for large amount of NO production is N2(v)+O NO+N in this case. This is
consistent with kinetic model study of similar DC glow discharge with the same order in
magnitude reduced electric field in [99].
We find that NO production in a DC plasmatron can be significantly high near the
reaction zone and that the glow to spark transition strongly affects the NO production via
increase in the electron density. This NO formation channel will be interrupted by the
methane when combustion is involved as will be shown in the next section.

108

6.6 Temperatures and NO Concentrations in a Plasma Enhanced Flame
In this section, results in a plasma enhanced flame in a similar format as the previous
section will be presented. The NO concentration is shown as a function of current,
equivalence ratio, and total flow rate.
Glow to spark transition with even little energy dissipation can significantly
contribute to the ignition of fuel during plasma assisted combustion due to the abundance
of chemically actively particles already present in the glow discharge. This random and
continuous transition is then able to sustain the flame. Ignition in methane/air mixture can
be achieved with equivalence ratios as low as 0.1. As stated in the description of
plasmatron, the flame speed of premixed flames in this type of discharge can increase
three fold or more at 300 mA, with further increase at higher currents. Also note that in
the plasma energy added is on the order of 15-95% of the thermal energy from the
combustion, which is generally reasonable in an ignition system of this kind. Keep in
mind that the power level will be orders of magnitude lower than the actual combustion
system which the igniter will be a part of in a practical system.
In Figure 6.7, we show the temperature and NO concentration as function of current
and flow rate when the equivalence ratio is held at 1. It is apparent in Figure 6.6 (a) that
the gas temperature increases with current, an indication that the deposition of heat
increased when more energy is added to the flow. Unlike the discharge only case, the
temperature increases with the flow rate. This is due to the heat release of the combustion

109

process which increases the temperature of the gas, and although the stoichiometry is
maintained, the increased density of the burnt fuel in the confined arc chamber and
through the exit of the torch results in a higher temperature. The plasma input power
(current X voltage) to the flame is similar in magnitude to that of the plasma discharge
only case. It can be seen that the NO production is reduced to a much lower level of 500
to 3500 ppm rather than the 8000 to 15000 ppm shown in the discharge only case. The
relation of NO concentration in air over in PAC is shown in Table 6.3. Number for lower
to higher current decreases, indicating a NO saturation towards higher current range. This
reduction of NO has also been previously observed by Kim et al. [100, 101].
Table 6.3. NO Concentration in Air over in PAC
Times

20 SCFH

30 SCFH

40 SCFH

50 SCFH

100

19.64637

18.30008

18.50497

22.59199

200

9.050061

8.894798

8.58164

10.42774

300

7.534822

6.445892

6.047817

8.121705

400

5.94333

5.776397

5.223709

6.623402

500

5.796336

5.01545

4.525725

5.876072

600

4.976757

4.651474

4.129818

5.482375

700

4.697449

4.26576

3.948711

4.832408

800

4.201341

4.127193

3.662822

4.225948

900

4.509511

3.978688

3.463385

3.997834

110

4000

NO concentration [ppm]

3500
3000
2500
2000
1500

20SCFH
30SCFH
40SCFH
50SCFH

1000
500
0
0

200

400

600

800

1000

Current [mA]

2700
2600
Temperature [K]

2500
2400
2300
2200

20SCFH
30SCFH
40SCFH
50SCFH

2100
2000
1900
1800
0

200

400

600

800

1000

Current [mA]
Figure 6.7 Temperature (a) and NO concentration (b) data as function of current and
flow rate when the equivalence ratio is held at 1 in plasma assisted methane/air flame.

111

The main cause of this effect is the addition of methane which is a discharge
interruption species (especially compared to other compositions O2 and N2) and
alternative electron impact reactions which leads to reduce the NO formation. Note that
temperature is high in the plasma assisted methane/air combustion (1850 ~ 2600 K),
where thermal NO formation is most likely the dominant mechanism. In particular, we
notice from Figure 6.7 that NO concentration surprisingly decreases with the flow rate
although this leads to in increase in temperature and injection density of plasma energy
into the flow. Upon further inspection, we note that the plasma energy input per molecule
is actually decreased as the flow rate is increased from 20 SCFH to 50 SCFH. Energy
input for each molecule is 0.203, 0.162, 0.159 and 0.154 (eV/molecule) for 20 SCFH, 30
SCFH, 40 SCFH and 50 SCFH, respectively at 100 mA and stoichiometric conditions.
This lower power input per molecule at higher flow rate results in less electrons and
collisions needed to initiate reactions that can results in the increase of NO. This is a clear
example that although thermal NO is expected to be dominant in the high temperature
range here, it is not the only mechanism.
In Figure 6.8, the temperature and NO concentration as a function of current and
equivalence ratio when flow rate is held at 25 SCFH are shown. It can be seen that the
NO increases with the current/power, and also with the total flow rate. Once again, it is
interesting to note that NO formation is higher for leaner flames although the temperature
is lower. Therefore, in addition to what can be expected from thermal formation of NO,
we need to consider the impact of increased methane. This is addressed below.
112

2600
φ=0.8
φ=0.9
φ=1.0
φ=1.1
φ=1.2

2500

Temperature [K]

2400
2300
2200
2100
2000
1900
1800
0

200

400

600

800

1000

800

1000

Current [mA]
5000
φ=0.8
φ=0.9
φ=1.0
φ=1.1
φ=1.2

4500
NO concentraion [ppm]

4000
3500
3000
2500
2000
1500
1000
500
0
0

200

400

600

Current [mA]
Figure 6.8 Temperature (a) and NO concentration (b) data as function of current and
equivalence ratio when flow rate the same in plasma with methane/air combustion.

113

2600

Temperature [K]

2400

2200

2000
20SCFH
25SCFH
30SCFH
35SCFH

1800
0.8

0.9

1.0

1.1

40SCFH
45SCFH
50SCFH

1.2

Equivalence Ratio [φ]

NO concentration [ppm]

2000

1500

1000
20SCFH
25SCFH
30SCFH
35SCFH

500
0.8

40SCFH
45SCFH
50SCFH

0.9

1.0

1.1

1.2

Equivalence Ratio [ φ]
Figure 6.9 Temperature (a) and NO concentration (b) data with equivalence ratio and
flow rate when the current is the same in plasma assisted methane/air combustion.

114

In Figure 6.9, we show the temperature and NO concentration as function of
equivalence ratio and the flow rate when the current is held at 300 mA. The temperature
profile as a function of equivalence ratio is similar to that of a typical case of combustion
using methane without any plasma enhancements. The peak temperature is reached at an
equivalence ratio close to 1, and higher current/power of the plasma increases the
absolute temperature. The temperature for lower flow rate of methane/air mixture is
lower than that of higher flow rate due to the less chemical heat release from combustion.
It is interesting to observe the linear decay of NO as the equivalence ratio is increased.
Note that NO quenching cross section with methane is extremely low [102], and therefore,
direction between the two species is negligible. On the other hand, methane will act as a
discharge interruption species (especially compared to other composition O2 and N2), and
the indirect interaction between methane and discharge will significantly impact the NO
formation. First, generation of vibrational nitrogen, a key factor for NO formation, is
reduced even though the same amount of energy is added to the system as greater portion
of it is allocated to alternative electron impact reactions, including breaking of CH4 into
smaller molecules. Second, as the flame becomes richer, unburnt hydrocarbons can also
contribute to the reduction of NO through reburn reactions. Third, atomic oxygen and
ozone generated from hydrocarbon help bring NO level down through reverse Zeldovich
reaction NO+O

N+O2 and through NO interactions with ozone. Atomic oxygen and

ozone can be abundant in hydrocarbon/air mixtures when a plasma discharge is present
and have been reported to increase flame propagation speed [3] and enhance flame speed
115

[103]. NO decay mechanism through atomic oxygen and ozone is confirmed
experimentally and by using kinetic models in [104]. The main production channel of
atomic oxygen is believed to be the dissociative quenching of O2 [105, 106] by N2.
Another potential explanation can be from the fact that as fuel reforming is increased in
so called ‘cold pre-flame’ regions, it competes with NO formation reactions in the use of
atomic oxygen, thus leading to the reduction of NO [79, 100]. In actual practice where
this torch will be a part of a larger combustor, the additional chemistry and the fluid
dynamics of the combustor will determine the overall NO formation, and the conditions
of the igniter should be determined according to the conditions of the specific application.

116

Chapter 7 Conclusions and Future Work
7.1 Microwave Discharge Results
A newly designed microwave re-entrant cavity applicator discharge system is used
for plasma enhanced combustion to probe the concept of direct coupling of microwave
power into the reaction zone of methane/oxygen mixtures. In the research, I have
conducted optical and laser diagnostics to probe the physicochemical characteristics of
coupling electromagnetic energy into the premixed flame. The flame was subjected to a
wide range of microwave power and displayed three distinct stages, depending on the
input microwave power as described below.
•

Stage 1: “Electric Field Only Stage”: In this stage, the flame is influenced by only a
weak electric field under 6 W. No visual change can be seen in either the photographs
or the OH PLIF as the power is increased. Most collisions in this stage are elastic
leading to only ohmic heating, which is responsible for flammability limits expansion
in this stage.

•

Stage 2: “Transition Stage”: Between 6 and 10 W, the plasma discharge and flame
exist at the same time and the discharge volume is dramatically increased. Also, high
electron temperatures and high reactivity lead to inelastic collisions for dissociate into
new species, excitation and eventually ionization.

117

•

Stage 3: “Full Plasma Stage” If microwave energy input exceeds 10 W, the system
enters a stage where the plasma plume dominates and the discharge emits a pink hue.
The number density of OH does not vary during Stage 1 but proceeds to rise rapidly

through the rest of the stages as a plasma plume develops. Through Stage 1 of the
coupling process, increase in the flammability limit, the flamespeed, and the rotational
temperature (from OH Chemiluminescence) is observed from ohmic heating as a result of
inelastic collisions. As the microwave power is raised, the electron temperature increases;
generation of new radicals and breakdown of the gas results in a plasma plume. The
plasma plume can be self sustained at this point without combustion and can re-ignite
fuel and oxidizer. This plasma discharge grows in volume and luminosity as the power is
further increased.
As the microwave power is increased further, the reaction zone is overlapped by a
full plasma plume; the OH number density is relatively constant, but the CO
concentration increases further. The importance of in-situ fuel reforming in plasma
coupled flames is shown through the concentration of CO, which increases ~18% with 30
W microwave power. In this stage, the plasma discharge can be sustained without
combustion and can the ignite methane/oxygen mixture. The direct coupling of
electromagnetic energy can be utilized for various levels of combustion enhancement and
ignition for practical combustion systems including flame holding for harsh transient
operating conditions.

118

7.2 DC Discharge Results
A DC plasmatron characterized by a nano-second glow-to-arc transient is used to
investigate premixed and nonpremixed flames using OH Laser Induced Fluorescence and
CO Two Photon Absorption Laser Induced Fluorescence (TALIF). Unique two cone flame
geometry is seen for the nonpremixed mode, where the inner cone is produced by the
plasma enhanced air mainly interacting with the flame and the outer one is from
oxidation between input fuel and ambient air. Partial oxidation inside the chamber for
premixed flames generally resulted in higher energy efficiencies compared to air
discharge and diffusion mode as activation energy of methane is much lower than that of
nitrogen. For nonpremixed flames, the flame is mainly stabilized by the reactive
discharge of air in the inner core and the stability of nonpremixed flames is significantly
influenced by the flowrates of the gases. By comparison of OH PLIF images in both
modes for different total flow rates and equivalence ration, stability of premixed flames is
more dependent on chemistry. OH number density for inner and outer cone indicates the
increase of OH with increased discharge power, while stability for nonpremixed case
depends on inner cone from discharge. Two photon LIF measurements of relative CO
number density showed CO number density increases with high discharge power
indicating that both configurations are significantly influenced by in-situ fuel reforming
at higher plasma powers.

119

7.3 Nitric Oxide Measurements Results
To investigate Nitric Oxide formation in DC plasma assisted combustion, planar
laser-induced fluorescence (PLIF) is used to acquire spatially resolved nitric oxide (NO)
concentrations in plasma-enhanced flames of a transient arc plasmatron. Quantitative
concentration of NO is calibrated by comparing the LIF signal to a well defined
flat-flame calibration torch and using a NO addition method.
IR thermometry and a more accurate multi-line fitting technique of multiple NO
transitions were used to obtain temperature field for two purposes: correction of LIF
signal due to spectroscopic temperature dependence and also to investigate the changes in
NO production as a variation of temperature.
Planar LIF images of NO were presented to show relative NO concentration in both
cases of plasma only and plasma enhanced combustion. In the case of plasma only, NO
production is high in the region of 8000 to 15000 ppm with DC power input of 200 to 750
Watts. Temperature range at these conditions is 550 to 1250 K where thermal NO
production is minor and NO formation is predominantly from N2(v)+O NO+N.
For plasma enhanced combustion case, NO production in methane/air flame for
investigated equivalence ratio range is about one order lower in magnitude. In this case
temperature is so high that thermal NO mechanism is still dominant. Plasma power and
equivalence ratio are the two main operational parameters in producing NO in the
combustion system here. For lower NO concentration in plasma enhanced flames of this

120

type, we need to avoid lean and high current conditions. Linear decay from equivalence
ratio 0.8 to 1.2 may caused by quenching of vibrational nitrogen, interactions with atomic
oxygen and ozone ,and reaction with unburned hydrocarbon species at rich conditions.

7.4 Some Suggestions for Future Work
The plasma assisted combustion research is a challenge field with numerous
unsolved problems but the ultimate goal is to identify and quantify every mechanism and
apply the concept in real life for better fuel efficiency and less emission. Plasma assisted
combustion may contribute to solution of current energy crisis as well as the problem of
global warming.
The current research uses several thermometries: thermocouple, IR thermometry,
NO multi-line thermometry and OH rotational thermometry. However, a more advanced
measurement (e.g. Rayleigh scattering thermometry) could offer a more accurate
temperature field, and thus better validate any simulation work and therefore, the
robustness and accuracy of future plasma assisted combustion detailed mechanisms.
One important and significant aspect of current study lies in that all discharge
systems are under atmospheric pressure, which is more practical than a low pressure
discharge and also cost efficient without extensive pumping systems; and meanwhile,
discharge is achieved at low input electric power. But the use of pulsed discharge might
be a crucial aspect to further investigate the combustion dynamics for more detailed
studies and comparisons. The use of a pulsed discharge would provide a more

121

equilibrium with higher efficiency, even though the cost is double or more.
Implementation of a discharge system with a real-life combustor (e.g. IC engines) will
offer more insights on future plasma assisted combustion applications under real work
environments (higher pressure, requirements for lower total weight, etc.)
To extend current research, different enhancement mechanisms (thermal, kinetics,
in-situ fuel reforming and etc.) should be quantified and compared with different systems
for their roles in plasma assisted combustion. For now, these mechanisms are believed to
contribute to enhancements when plasma is coupled into flames, but how much each
mechanism influences enhancement for different systems is still unknown and would be
interesting.

122

BIBLIOGRAPHY

123

BIBLIOGRAPHY
[1]

Bittencourt, Fundamentals of plasma physics, 3rd ed: the author, 2003.

[2]

S. M. Starikovskaia, "Plasma assisted ignition and combustion," Journal of
Physics D: Applied Physics, vol. 39, pp. R265-R299, 2006.

[3]

A. Y. Starikovskii, "Plasma supported combustion," Proceedings of the
Combustion Institute, vol. 30, pp. 2405-2417, 2005.

[4]

A. B. Leonov, D. A. Yarantsev, A. P. Napartovich, and I. V. Kochetov,
"Plasma-assisted ignition and flameholding in high-speed flow," presented at 44th
AIAA Aerospace Sciences Meeting, AIAA-2006-563, Reno, Nevada, 2006.

[5]

S. B. Leonov and D. A. Yarantsev, "Plasma-induced ignition and plasma-assisted
combustion in high-speed flow," Plasma Sources Science and Technology, vol. 16,
pp. 132-138, 2007.

[6]

T. Ombrello, Y. Ju, and A. Fridman, "Kinetic Ignition Enhancement of Diffusion
Flames by Nonequilibrium Magnetic Gliding Arc Plasma," AIAA J., vol. 46, pp.
2424-2433, 2008.

[7]

T. Ombrello, S. H. Won, and Y. Ju, "Lifted Flame Speed Enhancement by Plasma
Excitation of Oxygen," presented at 47th AIAA Aerospace Sciences Meeting and
Exhibit, AIAA-2009-689, 2009.

[8]

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, pp. 2545-2551, 2006.

[9]

W. Kim, M. Godfrey Mungal, and M. A. Cappelli, "The role of in situ reforming
in plasma enhanced ultra lean premixed methane/air flames," Combustion and
Flame, pp. doi:10.1016/j.combustflame.2009.06.016, 2009.

[10]

E. Mintusov, A. Serdyuchenko, I. Choi, W. R. Lempert, and I. V. Adamovich,
"Mechanism of Plasma Assisted Oxidation and Ignition of Ethylene-Air Flows by
a Repetitively Pulsed Nanosecond Discharge," presented at 46th AIAA Aerospace
Sciences Meeting and Exhibit, AIAA 2008-1106, 2008.

[11]

M. Uddi, N. Jiang, E. Mintusov, I. V. Adamovich, and W. R. Lempert, "Atomic
124

Oxygen Measurements in Air and Air/Fuel Nanosecond Pulse Discharges by Two
Photon Laser Induced Fluorescence," presented at 46th AIAA Aerospace Sciences
Meeting and Exhibit, AIAA 2008-1110, Orlando, Fl, 2008.
[12]

E. S. Stockman, S. H. Zaidi, R. B. Miles, C. D. Carter, and M. D. Ryan,
"Measurements of combustion properties in a microwave enhanced flame,"
Combustion and Flame, vol. 156, pp. 1453-1461, 2009.

[13]

C. T. Bowman, "Gas-phase reaction mechanisms for nitrogen oxide formation and
removal in combustion," in Pollutants from Combustion, C. Vovelle, Ed.
Netherlands: Kluwer Academic Publishers, 2000, pp. 123-144.

[14]

J. Warnatz, U. Maas, and R. W. Dibble, Combustion, 3 ed. Berlin-Heidelberg-New
York: Springer, 1997.

[15]

B. Atakan and A. T. Hartlieb, "Laser Diagnostics of NO reburning in fuel-rich
propene flames," Applied Physics B-Lasers and Optics, vol. 71, pp. 697-702,
2000.

[16]

I. Orlandini and U. Riedel, "Modelling of NO and HC removal by non-thermal
plamsas," Combustion Theory and Modelling, vol. 5, pp. 447-462, 2001.

[17]

L. Magne, S. Pasquiers, N. Blin-Simiand, and C. Postel, "Production and
reactivity of the hydroxyl radical in homogeneous high pressure plasmas of
atmospheric gases containing traces of light olefins," Journal of Physics D:
Applied Physics, vol. 40, pp. 3112-3127, 2007.

[18]

H. F. Calcote and R. N. Pease, "Electrical properties of flames," Industrial and
Engineering Chemistry, vol. 43, pp. 2726-2731, 1951.

[19]

H. F. Calcote and C. H. Berman, "Increased methane-air stabililty limits by a DC
electric field," presented at Proceedings of the ASME Fossil Fuels Combustion
Symposium, I1989.

[20]

G. P. Tweari and J. R. Wilson, Combustion and Flame, vol. 24, 1975.

[21]

T. Kitagawa, A. Moriwaki, K. Murakami, K. Takita, and G. Masuya, "Ignition
characteristics of methane and hydrogen using a plasma torch in supersonic flow,"
Journal of Propulsion and Power, vol. 19, pp. 853-858, 2003.

[22]

A. Y. Starikovskii, N. B. Anikin, I. N. Kosarev, E. I. Mintoussov, M. M. Nudnova,

125

A. E. Rakitin, D. V. Roupassov, S. M. Starikovskaia, and V. P. Zhukov,
"Nanosecond-pulsed Discharge for Plasma-Assisted Combustion and
Aerodynamics," Journal of Propulsion and Power, vol. 24, pp. 1182-1197, 2008.
[23]

G. Lou, A. Bao, M. Nishihara, S. Keshav, Y. G. Utkin, J. W. Rich, W. R. Lempert,
and I. V. Adamovich, "Ignition of Premixed Hydrocarbon-Air Flows by
Repetitively Pulsed, Nanosecond Pulse Duration Plasma," Proceedings of the
Combustion Institute, vol. 31, pp. 3327-3334, 2007.

[24]

N. Chintala, A. Bao, G. Lou, and I. V. Adamovich, "Measurements of combustion
efficiency in nonequilibrium RF plasma-ignited flows," Combustion and Flame,
vol. 144, pp. 744-756, 2006.

[25]

I. V. Kuznetsova, N. Y. Kalashnikov, A. F. Gutsol, A. F. Fridman, and L. A.
Kennedy, "Effect of "overshooting" in the transitional regimes of the low-current
gliding arc discharge," J. Appl. Phys, vol. 92, pp. 4231-4237, 2002.

[26]

Y. D. Korolev and I. B. Matveev, "Nonsteady-State Processes in a Plasma Pilot
for Ignition and Flame Control," Plasma Science, IEEE Transactions on, vol. 34,
pp. 2507-2513, 2006.

[27]

L. Bromberg, D. R. Cohn, A. Rabinovitch, and J. Heywood, "Emissions
Reductions Using Hydrogen from Plasmatron Fuel Coverters," Inter. J. of
Hydrogen Energy, vol. 26, pp. 1115-1121, 2001.

[28]

S. H. Zaidi, E. Stockman, X. Qin, Z. Zhao, S. Macheret, Y. Ju, R. B. Miles, D. J.
Sullivan, and J. F. Kline, "Measurements of hydrocarbon flame speed
enhancement in high-Q microwave cavity," presented at 44th AIAA Aerospace
Sciences Meeting and Exhibit, AIAA-2006-1217, 2006.

[29]

W. Kim, H. Do, M. G. Mungal, and M. A. Cappelli, "Investigation of NO
production and flame structure in plasma enhanced premixed combustion," Proc.
Combust. Inst, vol. 31, pp. 3319-3326, 2007.

[30]

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

[31]

T. Ombrello, X. Qin, Y. Ju, A. Gutsol, A. Fridman, and C. Carter, "Combustion
Enhancement via Stabilized Piecewise Nonequilibrium Gliding Arc Plasma
Discharge," AIAA J., vol. 44, pp. 142-150, 2006.

126

[32]

C. S. Kalra, A. Gutsol, and A. Fridman, "Gliding arc discharge as a source of
intermediate plasma for methane partial oxidation," IEEE Transactions on Plasma
Science, vol. 33, pp. 32-41, 2005.

[33]

L. Wu, J. Lane, N. P. Cernansky, D. L. Miller, A. A. Fridman, and A. Y.
Starikovskii, "Plasma-assisted ignition below self-ignition threshold in methane,
ethane, propane and butane- air mixtures," Proceedings of Combustion Institute,
vol. doi:10.1016/j.proci.2010.06.003, 2010.

[34]

A. Bao, "Ignition of hydrocarbon fuels by a repetitively pulsed nanosecond pulse
duration plasma," in Department of Mechanical Engineering, vol. Doctor Of
Philosophy: Ohio State University, 2008.

[35]

N. Chintala, R. Meyer, A. Hicks, A. Bao, J. W. Rich, W. R. Lempert, and I. V.
Adamovich, "Non-Thermal Ignition of Premixed Hydrocarbon-Air Flows by
Nonequilibrium RF Plasma," Journal of Propulsion and Power, vol. 21, pp.
583-590, 2005.

[36]

A. C. Eckbreth, Laser Diagnostics for Combustion Temperature and Species.
Amsterdam: Gordon and Breach, 1996.

[37]

D. R. Crosley, "Laser Probes for Combustion Chemistry," American Chemical
Society, Washington D.C. 1980.

[38]

K. Kohse-Höinghaus, "Laser Techniques for the Quantitative Detection of
Reactive Intermediates in Combustion Systems," Progr. Energy Combust. Sci., vol.
20, pp. 203-279, 1994.

[39]

J. W. Daily, "Laser Induced Fluorescence Spectroscopy in Flames," Progr. Energy
Combust. Sci., vol. 23, pp. 133-199, 1997.

[40]

C. N. Banwell and E. M. McCash, Fundamentals of Molecular Spectroscopy, 4 ed:
McGraw-Hill, 1994.

[41]

W. Demtröder, Laser Spectroscopy. Berlin: Springer-Verlag, 1982.

[42]

N. M. Laurendeau and J. E. M. Goldsmith, "Comparison of hydroxyl
concentration profiles using five laser-induced fluorescence methods in a lean
subatmospheric-pressure H2/O2/Ar flame," Combust. Sci. Technol., vol. 63, pp.
139-152, 1989.

127

[43]

A. Yariv, Quantum Electronics. New York: Wiley and Sons, 1975.

[44]

W. G. Bessler, "Quantitative nitric oxide concentration and temperature imaging in
flames over a wide pressure range with laser-induced fluorescence," vol. Ph.D.:
Heidelberg University, 2003.

[45]

W. P. Patridge and N. M. Laurendeau, "Formulation of a dimensionless overlap
fraction to account for spectrally distributed interactions in fluorescence studies,"
Applied Optics, vol. 34, pp. 2645, 1995.

[46]

D. R. Grieser and R. H. Barnes, "Nitric Oxide Measurements in a Flame by Laser
Fluorescence," Applied Optics, vol. 19, pp. 741-743, 1980.

[47]

R. L. McKenzie and K. P. Gross, "Two-photon excitation of nitric oxide
fluorescence as a temperature indicator in unsteady gasdynamic processes,"
Applied Optics, vol. 20, pp. 2153-2165, 1981.

[48]

G. Kychakoff, R. D. Howe, and R. K. Hanson, "Quantitative Flow Visualization
Technique for Measurements in Combustion Gases," Applied Optics, vol. 23, pp.
704-712, 1984.

[49]

P. H. Paul, M. P. Lee, B. K. McMillin, J. M. Seitzman, and R. K. Hanson,
"Application of Planar Laser-Induced Fluorescence Imaging Diagnostics to
Supersonic Reacting Flow," 28th AIAA/SAE/ASME/ASEE Joint Propulsion Conf.,
Orlando, FL, pp. paper 90-1844, 1990.

[50]

C. F. Kaminski, J. Holt, and M. Alden, "High repetition rate planar laser induced
fluorescence of OH in a turbulent non-premixed flame," Appl. Phys. B, vol. 68, pp.
757, 1999.

[51]

C. Kittler and A. Dreizler, "Cinematographic imaging of hydroxyl radicals in
turbulent flames by planar laser-induced fluorescence up to 5 kHz repetition rate,"
Appl. Phys. B., vol. 89, pp. 163-166, 2007.

[52]

N. Jiang and W. R. Lempert, "Ultrahigh-frame-rate nitric oxide planar
laser-induced fluorescence imaging," OPTICS LETTERS, vol. 33, pp. 2236-2238,
2008.

[53]

W. G. Bessler, C. Schulz, V. Sick, and J. W. Daily, "A versatile modeling tool for
nitric oxide LIF spectra," presented at 3rd Joint Meeting US Sec. Combust. Inst.,
Chicago (www.lifsim.com), 2003.

128

[54]

Y. P. Raizer, "Propagation of a high-pressure microwave discharge," Soviet
Physics JETP-USSR, vol. 34, pp. 114, 1972.

[55]

R. M. Fredericks and J. Asmussen, "Retuning and hysteresis effects of a rf plasma
in a variable size microwave cavity," J. Appl. Physics, vol. 42, pp. 3647-3649,
1971.

[56]

X. Rao, K. Hemawan, I. Wichman, C. Carter, T. Grotjohn, J. Asmussen, and T.
Lee, "Combustion dynamics for energetically enhanced flames using direct
microwave energy coupling," Proceedings of the Combustion Institute, vol.
doi:10.1016/j.proci.2010.06.024, 2010.

[57]

K. W. Hemawan, I. S. Wichman, T. Lee, T. A. Grojohn, and J. Asmussen,
"Compact Microwave Reentrant Cavity Applicator for Plasma Assisted
Combustion," Submitted, 2009.

[58]

K. Hamawan, C. Romel, S. Zuo, I. Wichman, T. Grotjohn, and J. Asmussen,
"Microwave plasma assisted premixed flame combustion," Appl. Phys. Lett., vol.
89, pp. 141501, 2006.

[59]

K. W. Hemawan, C. L. Romel, S. Zuo, I. S. Wichman, T. A. Grotjohn, and J.
Asmussen, "Microwave plasma-assisted premixed flame combustion," Appl. Phys.
Lett., vol. 89, pp. 141501, 2006.

[60]

Y. D. Korolev, O. B. Frants, N. V. Landl, V. G. Geyman, and I. B. Matveev,
"Glow-to-Spark Transitions in a Plasma System for Ignition and Combustion
Control," Plasma Science, IEEE Transactions on, vol. 35, pp. 1651-1657, 2007.

[61]

K. Kohse-Hoinghaus and J. B. Jeffries, Applied Combustion Diagnostics: Taylor
& Francis, 2002.

[62]

P. H. Paul, "A Model For Temmperature-Dependent Collisional Quenching of OH
A2Σ+," J. Quant. Spectrasc. Radiat. Transfer, vol. 51, pp. 511-524, 1994.

[63]

J. M. Seitzman, J. Haumann, and R. K. Hanson, "Quantitative two photon LIF
imaging of carbon monoxide in combustion gases," Applied Optics, vol. 26, pp.
2892-2899, 1987.

[64]

"GRI-Mech 3.0 http://www.me.berkeley.edu/gri-mech/version30."

[65]

C. Schulz, V. Sick, U. E. Meier, J. Heinze, and W. Stricker, "Quantification of NO

129

A-X(0,2) laser-induced fluorescence: investigation of calibration and collisional
influences in high-pressure flames," Applied Optics, vol. 38, pp. 1434-1443, 1999.
[66]

T. Pedersen and R. C. Brown, "Simulation of electric field effects in premixed
methane flames," Combustion and Flame, vol. 94, pp. 433-448, 1993.

[67]

Y. Ju, S. O. Macheret, M. N. Shneider, R. B. Miles, and D. J. Sullivan,
"Numerical study of the effect of microwave discharge on the premixed
methane–air flame," presented at 40th AIAA/ASME/SAE/ASEE Joint Propulsion
Conference and Exhibit, Fort Lauderdale, FL, 2004.

[68]

R. Ono and T. Oda, "OH radical measurement in a pulsed arc discharge plasma
observed by a LIF method," IEEE Transaction on Industry Applications, vol. 37,
pp. 709-714, 2001.

[69]

Q. V. Nguyen, R. W. Dibble, C. D. Carter, G. J. Fiechtner, and R. S. Barlow,
"Raman-LIF Measurements of Temperature, Major Species, OH, and NO in a
Methane-Air Bunsen Flame," Combustion and Flame, vol. 105, pp. 499-510,
1996.

[70]

K. Shinohara, N. Takada, and K. Sasaki, "Enhancement of burning velocity in
premixed burner flame by irradiating microwave power," J. Phys. D: Appl. Phys.,
vol. 42, pp. 182008 (4pp), 2009.

[71]

K. W. Hemawan, I. S. Wichman, T. Lee, T. A. Grotjohn, and J. Asmussen,
"Compact Microwave Reentrant Cavity Applicator for Plasma Assisted
Combustion," Review of Scientific Instruments, vol. 80, pp. 053507, 2009.

[72]

M. Uddi, "Non-equilibrium kinetic studies of repetitively pulsed nanosecond
discharge plasma assisted combustion," in Department of Mechanical
Engineering, vol. Ph.D.: Ohio State University, 2008.

[73]

P. W. J. M. Boumans, Inductively Coupled Plasma Emission Spectroscopy, vol.
Part 2. New York, Chichester, Brisbane, Toronto, Singapore: John Wiley & Sons,
1987.

[74]

J. Happold, P. Lindner, and B. Roth, "Spatially resolved temperature
measurements in an atmospheric plasma touch using the A2,v=0 --> X2, v''=0 OH
band," J. Phys. D: Appl. Phys., vol. 39, pp. 3615-3620, 2006.

[75]

S. Pellerin, J. M. Cormier, F. Richard, K. Musiol, and J. Chapelle, "A

130

spectroscopic Diagnostic Method using UV OH band spectrum," J. Phys. D: Appl.
Phys., vol. 29, pp. 726, 1996.
[76]

H. P. Broida, J. of Chem. Phys., vol. 21, pp. 1165-1169, 1953.

[77]

A. G. Gaydon and H. G. Wolfhard, Flames, their structure, radiation and
temperature, 3rd ed, 1970.

[78]

E. S. Stockman, J. B. Michael, A. Fuller, S. H. Zaidi, and R. B. Miles, "Toward
High Q Evanescent Coupled Microwave Controlled Combustion," presented at
47th AIAA Aerospace Sciences Meeting Including The New Horizons Forum and
Aerospace Exposition, 2009.

[79]

W. Kim, H. Do, M. G. Mungal, and M. A. Cappelli, "A Study of Plasma Stabilized
Diffusion Flames at Elevated Ambient Temperatures," IEEE Transactions on
Plasma Science, vol. 36, pp. 2898-2904, 2008.

[80]

L. Bromberg, D. R. Cohn, A. Rabinovich, and J. Heywood, "Emission reductions
using hydrogen from plasmatron fuel converters," International Journal of
Hydrogen Energy, vol. 26, pp. 1161-1121, 2001.

[81]

L. Bromberg, D. R. Cohn, A. Rabinovich, and N.Alexeev, "Plasma catalytic
reforming of methane," International Journal of Hydrogen Energy, vol. 24, pp.
1131-1137, 1999.

[82]

Y. D. Korolev, O. B. Frants, N. V. Landl, V. G. Geyman, and I. B. Matveev,
"Nonsteady-State Gas-Discharge Processes in Plasmatron for Combustion
Sustaining and Hydrocarbon Decomposition," IEEE Transaction on Plasma
Science, vol. 37, pp. 586-592, 2009.

[83]

X. Rao, I. Matveev, and T. Lee, "Nitric Oxide Formation in a Premixed Flame
With High-Level Plasma Energy Coupling," IEEE Transaction on Plasma Science,
vol. 37, pp. 2303-2313, 2009.

[84]

J. J. Lowke and R. Morrow, "Theoretical analysis of removal of oxides of sulphur
and nitrogen in pulsed operation of electrostatic precipitators," IEEE Transaction
on Plasma Science, vol. 23, pp. 661-671, 1995.

[85]

J. B. Kelman, D. A. Greanhalgh, E.Ramsay, D. Xiao, and D. T. Reid, "Flow
imaging by use of femtosecond-laser-induced two-photon fluorescence," Optics
Letters, vol. 29, pp. 1973-1975, 2004.

131

[86]

A. V. Mokhov, H. B. Levinsky, C. E. v. d. Meij, and R. A. A. M. Jacobs, "Analysis
of laser-induced-fluorescence carbon monoxide measurements in turbulent
nonpremixed flames," Applied Optics, vol. 34, pp. 7074-7082, 1995.

[87]

X. Rao, T. Lee, and I. Matveev, "Nitric Oxide Formation during Ignition and
Combustion of a Transient Arc Plasmatron," presented at 47th AIAA Aerospace
Sciences Meeting, 2009.

[88]

A. H. Lefebvre, Gas Turbine Combustion. Philadelphia, PA: Taylor & Francis,
1999.

[89]

J. T. Herron, "Modeling Studies of the Formation and Destruction of NO in pulsed
barrier discharges in nitrogen and air," Plasma Chem. Plasma Process, vol. 21, pp.
581-609, 2001.

[90]

G. Zhao, S. V. B. Janardhan, X. Hu, M. D. Argyle, and M. Radosz, "Effect of
oxygen on nonthermal plasma reactions of nitrogen oxides in nitrogen,"
Enviromental and energy engineering, vol. 51, pp. 1800-1812, 2005.

[91]

K. Kohse-Höinghaus and J. B. Jeffries, Applied Combustion Diagnostics: Taylor
and Francis, 2002.

[92]

M. D. DiRosa, K. G. Klavuhn, and R. K. Hanson, "LIF Spectroscopy of NO and
O2 in High-Pressure Flames," Combustion Science and Technology, vol. 118, pp.
257-283, 1996.

[93]

W. G. Bessler, C. Schulz, T. Lee, J. B. Jeffries, and R. K. Hanson, "Strategies for
laser-induced fluorescence detection of nitric oxide in high-pressure flames. III.
Comparison of A-X excitation schemes," Applied Optics, vol. 42, pp. 4922-4936,
2003.

[94]

T. Lee, W. G. Bessler, H. Kronemayer, C. Schulz, and J. B. Jeffries, "Quantitative
temperature measurements in high-pressure flames with multiline NO-LIF
thermometry," Applied Optics, vol. 44, pp. 6718-6728, 2005.

[95]

P. A. Berg, G. P. Smith, J. B. Jeffries, and D. R. Crosley, "Nitric oxide formation
and reburn in low-pressure methane flames," Proceedings of the Combustion
Institute, vol. 27, pp. 1377-1384, 1998.

[96]

J. H. Bromly, F.J.Barnes, S.Muris, X.You, and B.S.Haynes, "Kinetic and
Thermodynamic Sensitivity Analysis of the NO-Sensitised Oxidation of

132

Methane," Combustion Science and Technology, vol. 115, pp. 259-296, 1996.
[97]

W. G. Bessler and C. Schulz, "Quantitative multi-line NO-LIF temperature
imaging," Applied Physics B, vol. 78, pp. 519-533, 2004.

[98]

A. Fridman and L. Kennedy, Plasma Physics and Engineering. New Yourk:
Taylor and Fancis, 2004.

[99]

B. F. Gordiets, C. M. Ferreira, V. L. Guerra, J. M. A. H. Loureiro, J. Nahorny, D.
Pagnon, M. Touzeau, and M. Vialle, "Kinetic model of a low pressure N2 O2
flowing glow discharge," IEEE Transactions on Plasma Science, vol. 4, pp.
750-768, 1995.

[100] W. Kim, H. Do, M. G. Mungal, and M. A. Cappelli, "Investigation of NO
production and flame structure in plasma enhanced premixed combustion,"
Proceedings of the Combustion Institute, vol. 31, pp. 3319-3326, 2007.
[101] W. Kim, M. G. Mungal, and M. A. Cappelli, "The role of in situ reforming in
plasma enhanced ultra lean premixed methane/air flames," 2009.
[102] M. C. Drake and J. W. Ratcliffe, "High Temperature Quenching cross sections for
nitric oxide laser-induced fluorescence measurements," J. Chem. Phys., vol. 98,
pp. 3850-3865, 1993.
[103] T. Ombrello, S. H. Won, Y. Ju, and S. Williams, "Lifted Flame Speed
Enhancement by Ozone," presented at 4th International Workshop and Exhibition
on Plasma Assisted Combustion, 2008.
[104] 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," J. Phys, D: Appl. Phys, vol. 42, pp. 075205, 2009.
[105] G. D. Stancu, F. Kaddouri, D. A. Lacoste, and C. O. Laux, "Investigations of rapid
plasma chemistry generated by nanosecond discharges in air at atmospheric
pressure using advanced optical diagnostics," presented at 47th AIAA Aerospace
Sciences Meeting, Orlando, FL, 2009.
[106] M. Uddi, N. Jiang, E. Mintusov, I. V. Adamovich, and W. R. Lempert, "Atomic
oxygen measurements in air and air/fuel nanosecond pulse discharge by two
photon laser induced fluorescence," Proc. Combust. Inst, vol. 32, pp. 929-936,
2009.

133