.. «a...
.2 in.

s 7?.
r.. . 7.40:... . .. .

.3:....).L.....n.u:: . .r . V I.
:3. .. .2
15...!

 

5!).

a“: :amwg.§amrp

 

.43....5

la.

 

.45.. that
, 11...:
fl (.5.
c .3. .{zn

.15 .

cub-A

m

A .I (I
. r .3. I, .
.nn.».‘.e€v}.fiuhmn . 1.1.1

2 .323...

:3 \ 3!. x1 K15
:tth..r§.v AL. up}!
2:3:10 I»?

33.353 ‘

:4

. L.
5...!)
.x z.

. 3%
am. V
. x.

J .vqluvi;
:11... .3.

3.8- ..,.
, , Irwm
4...“...th an
ramfixfl.” qfiih:
i.

.3.
a 1..
u I...

10:11.3:

E332! .35 1’.
23...»... . .. ..

.1
12...-..
:1...
it
(do. a:
.. t .9...
.i .
m... z
1.} 4
2-53:5: If!
.5}... ..i?.....2.!
: 1!...7912v
314.53,... 4..
3...... . 31......
. .36??an

A $2504
34.1: t 3m _
43.4...»

533.512.!
.lel-I..l
. .151}:
v ‘I I“ _
3. 1:.
5.5.1 13...}

imbaxdvddu

3.1..
.3.
. a...
a. m. 31%.
. .. . 3 .5;
:11}... .621. .4912.“ n
, a. .9: {51:52.23
1... #1.:

a
l.‘ ..
I .21!!!)
.5..sz 4.1,!
[.3 .
It...

..‘
u (312..-
LA u .)

. 3.11....3
I; v «I...:J~»

 

Z LIBRARY
W11; Michigan State
University

 

This is to certify that the
thesis entitled

An Experimental Study of Microwave Plasma-Enhanced
Combustion

presented by
Chandra Lynn Romel

has been accepted towards fulfillment
of the requirements for the

MS. degree in Mechahical EnfleerinL

MM. j

Major Professor’s Signature

Maj II) 200 6
O I #.
Date

 

 

 

MSU Is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:lClRC/DateDue.indd-p.1

 

AN EXPERIMENTAL STUDY OF MICROWAVE PLASMA-ENHANCED
COMBUSTION

By

Chandra Lynn Romel

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Mechanical Engineering

2006

ABSTRACT

AN EXPERIMENTAL STUDY OF MICROWAVE PLASMA-ENHANCED
COMBUSTION

By
Chandra Lynn Romel

Advantages of a hybrid flame, created by microwave power, include more
efficient and stable combustion and a more concentrated, higher temperature
flame/discharge, potentially useful for material synthesis, material cutting and
welding, and various surface treatments. A compact torch has been designed
and experimentally evaluated while operating in both plasma-only and plasma-
assisted combustion modes. Operation of the torch in a hydrocarbon/oxygen
combustion mode is investigated with microwave power applied to modify the
combustion process. The objective of this investigation is to quantify the changes
in the combustion process as microwave power is applied to intensify the
discharge.

Diagnostic measurements performed include (1) gas temperatures by
optical emission spectroscopy, (2) flame/discharge power densities calculated
from flame geometries, and (3) flame species to show how the presence of
microwaves affects the flame. These measurements are made using a brass

nozzle and a ceramic nozzle.

ACKNOWLEDGMENTS

I thank my mom for being my support for 26 years. I also thank the Man
above for helping me through the tough times and for His love. I would also like
to thank my advisor, Dr. lndrek Wichman, for challenging me and encouraging
me.

Special thanks goes to the Electrical Engineering professors, Dr. Grotjohn
and Dr. Asmussen, and Materials Science professor, Dr. Case, all of whom
advised me throughout this project.

I also thank the electrical engineering students, Kadek Hemawan, Stanley
Zuo, and Jeffri Narendra for their help and guidance. Thank you to Kadek who

has given me permission to use select figures created by him.

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vi
LIST OF FIGURES ......................................................................................................... vii
Chapter 1 ........................................................................................................................... 1
1.1 Introduction ................................................................................................................ 1
1.2 Literature on plasma-enhanced combustion ......................................................... 6
Chapter 2 ........................................................................................................................... 8
2.1 Nomenclature ............................................................................................................ 8
2.2 Single-step chemistry ............................................................................................... 9
2.3 Multi-step chemistry ............................................................................................... 27
Chapter 3 ......................................................................................................................... 43
3.1 Introduction to equipment ...................................................................................... 43
3.2 Overall experimental setup .................................................................................... 43
3.2.1 Torch .................................................................................................................. 43
3.2.1.1 Introduction to torch .................................................................................. 43
3.2.1.2 Torch design and dimensions ................................................................ 44
3.2.1.2.1 Ceramic nozzle design and fabrication ......................................... 47

3.2.2 Electrical network and cavity ......................................................................... 52
3.2.3 Gas system ....................................................................................................... 55
3.3 Diagnostics and measurement techniques ......................................................... 57
3.3.1 Optical emission spectroscopy (OES) ......................................................... 58
3.3.2 Infrared camera ............................................................................................... 60
3.3.3 Thermocouple ................................................................................................... 63
Chapter 4 ......................................................................................................................... 66
4.1 Introduction ............................................................................................................... 66
4.2 Rotational temperature theory .............................................................................. 66
4.3 Gas temperature of N2 ........................................................................................... 71
4.4 Gas temperature of C2 ............................................................................................ 74
Chapter 5 ......................................................................................................................... 78
5.1 Introduction ............................................................................................................... 78
5.2 Flame/discharge photographs .............................................................................. 78
5.3 Geometric analysis .................................................................................................. 88
5.4 Impact of microwaves on combustion .................................................................. 91
5.4.1 Optical emission spectroscopy (OES) .......................................................... 93
5.5 Flame/discharge temperatures and profiles ..................................................... 105
5.5.1 Data collected from infrared (IR) camera ................................................... 109
Chapter 6 ....................................................................................................................... 113
6.1 Summary of results .............................................................................................. 113

iv

6.2 Recommendations ............................................................................................... 115

APPENDICES ............................................................................................................... 117
Appendix A .................................................................................................................... 118
Appendix B .................................................................................................................... 119
BIBLIOGRAPHY ........................................................................................................... 121

LIST OF TABLES

Table 3.1Torch part dimensions in millimeters and inches .................................. 46
Table 3.2 Nozzle flow limits. Maximum flow rates for each type of nozzle, in sccm
and m/s. Note the ceramic nozzle allows a higher maximum flow rate than the
brass nozzle with the same diameter. .......................................................... 57
Table 4.1 Rotational constants for the electronic states of nitrogen and carbon. 67

Table 4.2 Values used for the N2 rotational temperature calculations for the R
branch of the (2,0) SPS. ...................................................................................... 73

Table 5.1 Gas flow limitations for three nozzles with units in sccm and m/s.......79

Table 5.2 Lengths of flames with varying flow rates, corresponding to Figure 5.1.
The shaded cells highlight stoichiometric flow rates ............................................ 81

Table 5.3 Flame speeds calculated for flames shown in Figure 5.7 as power

decreases from 39 W to 33 W. Calculations are based on Equation (5.1). ......... 88
Table A1 Values for thermocouple temperature measurements .................... 118
Table 8.1 Values used for flame extinction plot in Chapter 5. ........................... 119
Table 32 Values found in temperature plots in Chapter 5. ............................... 119

vi

LIST OF FIGURES

Figure 1.1 Schematic of experimental apparatus including the plasma/flame torch
inside microwave cavity. The torch houses a gas line and water line, and
microwave power is inputted midstream. A combustion flame is produced inside
a cavity that creates an electric field pattern exciting a TM012 (transverse
magnetic) mode. A window is present for observation and diagnostic
measurements (optical emission spectroscopy (OES), for example) [3]. .............. 2

Figure 1.2 Species and radical concentration curves in regions 6,, 6n. and 6m of
the combustion process. With the addition of microwaves, the electron/ion
concentration is possibly greater than with combustion alone. ............................. 4

Figure 2.1 Temperature distribution in relation to flame front for the simple
infinite-reaction-rate model described by Eq (2.3). Note the jump in temperature
gradient across the flame front. 14

Figure 2.2 Schematic of reactant mass fraction distribution for the infinitesimally

thin flame described by Eq (2.6) on either side of the flame, with f=0 on either
side. Note, once again as in Figure 2.1, the discontinuity in the gradient of YR,
[dYR/dX]=YRUUf/D. ............................................................................................... 16

Figure 2.3 Schematic of stretching transformation in nondimensional scale. The
inner coordinate E is large and negative near 6 = 0 ', large and positive near 6 =
0*, and of the order unity in the flame sheet ........................................................ 21

Figure 2.4 Plot of anticipated solution after applying stretching transformation. .24
Figure 2.5 Curve representing a two-step combustion mechanism ..................... 29

Figure 2.6 Curve representing fuel species behavior for a two-step mechanism.
............................................................................................................................ 35

Figure 3.1 Left: photograph of disassembled PTB. Right: photograph of
assembled torch demonstrating hand-held easiness. .............................................. 44

Figure 3.2 Internal structure schematic drawing of the PTB with the brass nozzle.
This unit is also called the “torch applicator.” Note the center conductor and the
water cooling system. The inner conductor is adjustable to create an optimal
electric field that is concentrated at or near the tip of the nozzle [18]. .................. 45

Figure 3.3 Left: cross section of nozzle with gas and water lines. The gas line is

soldered to the wall of the nozzle. Water cooling is necessary to keep the solder
from melting. Water enters the inside of the inner conductor near the tip of the

vii

nozzle and circulates throughout, and exits near the base of the conductor. Right:
cross-sectional drawing of the nozzle with dimensions in millimeters. .................. 47

Figure 3.4 Left: sketch of ceramic nozzle showing smoothly converging path from
nozzle inlet to exit, including diameter dimensions. Right: photograph of ceramic
nozzle attached to the inner conductor. Note the gold-sputtered coat on the
ceramic. ........................................................................................................................... 48

Figure 3.5 Schematic of ceramic nozzle. Left: ceramic powder being pressed
axially by two cylinders at 2000 psi inside a cylindrical metal die and shaped by a
parabolic graphite mold. Right: nozzle after pressing and drilling of exit pathway.
Note the outside corners of the nozzle are still square. Careful hand sanding
rounds these corners (represented by dotted lines) so the flame does not sit on a
radially-large surface. .................................................................................................... 50

Figure 3.6 Photographs of: (a) tumbler used for mixing powders and for shaping
graphite, (b) axial press used for pressing powder in the die, and (c) digitally
controlled furnace used for sintering specimens ....................................................... 51

Figure 3.7 Overall experimental setup, including electrical circuit, torch and
cavity, gas system, and diagnostic setup [3]. ............................................................ 53

Figure 3.8 PTB and microwave cavity setup, including details of the cavity and
the electrical field pattern and wave modes. The torch is placed vertically in the
cavity and tuned so the flame interacts with the microwaves, creating a plasma-
enhanced hybrid flame. The dimensions of the cavity-penetrated torch are given.
........................................................................................................................................... 55

Figure 3.9 Schematic of the gas setup. Gas flows from the compressed tank,
through MFCs. It is then premixed before entering the torch. Methane (CH4) and
oxygen (02) are the primary gases used in the experiments. The microwave
power is fed through the side of the torch applicator (Pmicro) and the water cooling
lines are also shown. ..................................................................................................... 56

Figure 3.10 Schematic of OES system. Light is emitted from discharge, analyzed
in the monochromator, and collected and organized in the computer ................... 60

Figure 3.11 Left: photograph of the data acquisition system used to measure
temperature with a Type K thermocouple. System is connected to a computer
with a GPIP cable. Right: device built to sweep thermocouple through flame at a
constant speed. Note the spherical thermocouple junction ..................................... 65

Figure 4.1 Energy level diagram for a band with P, Q, and R branches, relative to
wavelength (A), rotational energy levels (J), and vibrational levels (v) [27].........68

viii

Figure 4.2 Laboratory process for temperature calculations: light is emitted from
the flame/discharge and focused by a lens into the spectrometer, where
intensities and wavelengths are recorded. .......................................................... 71

Figure 4.3 Plot of intensity (or current since the units are arbitrary) versus
wavelength for the N2 spectrum, commonly used for rotational temperature
calculations. R20 - R30 list the order of the rotational bands. ................................ 72

Figure 4.4 Boltzmann plot for the bands R20 — R30 of the N2 spectrum. .............. 74
Figure 4.5 C2 spectrum of a Swan System, identifying the bands R25 — R45. ...... 75

Figure 4.6 Boltzmann plot for the bands R25 - R45 of the C2 spectrum, for the
180/60 sccm O2/CH4 pure flame. ........................................................................ 77

Figure 5.1 Photographs of pure combustion flames with varying flow rates,
measured in sccm. The 0.4 mm-diameter brass nozzle is used. ..................... 80

Figure 5.2 Photographs of a pure plasma discharge, with a flow rate of 200 sccm
argon and 30 W microwave power. Note that the torch is not confined within the
cavity. .................................................................................................................. 82

Figure 5.3 Photographs of flame/discharges inside the cavity as 40-100 W
microwave power is added. The 0.2 mm-diameter brass nozzle is used. The flow
rate for the top set of photos is 40/0 sccm O2/CH4, i.e. no fuel is present. The flow
rate for the bottom set is 40/20 sccm O2/CH4, a stoichiometric ratio. Scales are
provided for flame dimensions. Camera speed and aperture are 1.6 and 3.5,
respectively. ........................................................................................................ 83

Figure 5.4 Photographs of flame/discharges inside the cavity as microwave
power is added, 20-80 W. The 0.4 mm-diameter ceramic nozzle is used. The flow
rate is 200/100 sccm O2/CH4, a stoichiometric ratio. A scale is provided for flame
dimensions. Camera shutter speed and aperture are 1.6 s and 3.5, respectively.
............................................................................................................................ 84

Figure 5.5 Photographs of flame/discharges inside the cavity with a ceramic
nozzle comparing 0 W to 80 W of microwave power ........................................... 84

Figure 5.6 Photographs of flame/discharges inside the cavity as microwave
power is added, 10-100 W. The 0.4 mm-diameter brass nozzle is used. The flow
rate for the top set is 70/24 sccm O2/CH4. The flow rate for the bottom set is
153/70 sccm O2/CH4. A scale is provided for flame dimensions. Camera shutter
speed and aperture are 1/30 s and 3.5, respectively. ......................................... 86

Figure 5.7 Additional photographs from Figure 5.6. The flow rate is 70/24 sccm
O2/CH4 and all other conditions are the same as listed in Figure 5.6. The top

ix

sequence shows the flame/discharge as microwave power increases; the bottom
sequence is power decreasing. This intermediate behavior gives rise to a
hysteretic affect. .................................................................................................. 87

Figure 5.8 Power density as a function of non-dimensional power, evaluated from
flame/discharge geometries shown in Figure 5.6, with flow rates of 70/24 sccm
O2/CH4 and 153/70 sccm O2/CH4. The powers of combustion are provided: 15 W
for the first flow and 44 W for the second flow. ................................................... 90

Figure 5.9 Flame height as a function of non-dimensional power, evaluated from
flame/discharge geometries, shown in Figure 5.4, with flow rates of 70/24 sccm
O2/CH4 and 153/70 sccm O2/CH4 ........................................................................ 90

Figure 5.10 Flame extinction or blow-out plot of equivalence ratio versus total
volumetric flow rate, VTOT. Each data point is provided with the combustion power
of the flame at a particular flow rate, where fuel is present. Note that a discharge
is maintained with no fuel when 20-100 W of microwave power is added. A
spreadsheet of raw data can be seen in Appendix B. ......................................... 93

Figure 5.11 OES scan of C2 in a stoichiometric (200/100 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40 W microwave power)
discharge using a 0.4 mm-diameter ceramic nozzle. .......................................... 95

Figure 5.12 OES scan of N2 in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40, 50, 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” is a flow rate of
160/54 sccm O2/CH4. The last set of data is from a discharge of only 160 sccm
O2 and 50 W of microwave power (no fuel is present). Data from 30 W of
microwave power and below show no N2 signal, similar to the combustion-only
case, and therefore are excluded from the plot. .................................................. 96

Figure 5.13 Second set of data: OES scan of N2 in a hybrid (160/80 sccm O2/CH4
combustion plus 10-80 W microwave power) flame/discharge using the 0.4 mm-
diameter brass nozzle. ........................................................................................ 97

Figure 5.14 OES scan of N2 in a hybrid flame/discharge with a flow rate of
160/80/10 sccm O2/CH4/N2 with varying microwave power, 30-80 W. Note that N2
is injected upstream into the flow of gases, CH4 and O2. .................................... 97

Figure 5.15 OES scan of C2 in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 10-100 W microwave power)

discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” has a flow rate of
160/54 sccm O2/CH4. .......................................................................................... 98

Figure 5.16 Second set of data: OES scan of C2 in a combustion (160/80 sccm
O2/CH4) and hybrid (combustion plus 10-80 W microwave power)
flame/discharge using the 0.4 mm-diameter brass nozzle. ................................. 99

Figure 5.17 OES scan of O2 in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40, 50, 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” is a flow rate of
160/54 sccm O2/CH4. The last set of data is from a discharge of only 160 sccm
O2 and 50 W of microwave power (no fuel is present). Data from 30 W of
microwave power and below show no 02 signal, similar to the combustion-only
case, and therefore are excluded from the plot. ................................................ 100

Figure 5.18 OES scan of OH in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40, 50, 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” is a flow rate of
160/54 sccm O2/CH4. Data from 30 W of microwave power and below show no
OH signal, similar to the combustion-only case, and therefore are excluded from
the plot. ............................................................................................................. 101

Figure 5.19 OES scan of CH in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 10-100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” is a flow rate of
160/54 sccm O2/CH4. ........................................................................................ 101

Figure 5.20 Relative intensities of N2 emissions as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm, 160/54 sccm, and 160/0 sccm O2/CH4, and also a flow rate of 160/80/10
sccm O2/CH4/N2 with N2 injection. Intensities are recorded from four sets of
experiments to show consistency, and values are recorded from peak intensities,
at a wavelength of ~ 3769 A .............................................................................. 102

Figure 5.21 Relative intensities of C2 emission as a function of absorbed
microwave power, measured from a stoichiometric flame/discharge with a flow
rate of 160/80 sccm O2/CH4. Intensities are recorded from three sets of
experiments to show consistency, and values are recorded from peak intensities,
at a wavelength of ~ 5130 A .............................................................................. 103

Figure 5.22 Relative intensities of O2 emission as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm O2/CH4 and 160/54 sccm O2/CH4. Intensities are recorded from one set of
experiments, and values are recorded from peak intensities, at a wavelength of ~
3340 A ............................................................................................................... 103

Figure 5.23 Relative intensities of CH emission as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm O2/CH4 and 160/54 sccm O2/CH4. Intensities are recorded from one set of

xi

experiments, and values are recorded from peak intensities, at a wavelength of ~
4270 A ............................................................................................................... 104

Figure 5.24 OES scan of combustion flame (160/80 sccm O2/CH4) and hybrid
flame/discharge (addition of 40 W microwave power) for an entire range of visible
emission spectra. The wavelength and relative intensities are given at the peak
(A = 4280 A) for each set of data. Molecules and radicals are listed next to their
visible spectra. A 0.4 mm-diameter ceramic nozzle is used for this data. ......... 105

Figure 5.25 Temperature profile of a stoichiometric CH4/O2 flame from a type K
thermocouple. Photograph of the flame shows where the thermocouple scans the
flame. The plot provides experimental data and also displays a qualitative
example of a theoretical temperature distribution with an adiabatic flame
temperature of 3000 K, taken as an average temperature. ............................... 106

Figure 5.26 N2 rotational temperature of a flame/discharge as a function of
absorbed microwave power for data for three different flow rates: 160/80/10
sccm, 100/50/3.6 sccm, and 60/30/2 sccm O2/CH4/N2. ..................................... 107

Figure 5.27 N2 and C2 rotational temperatures of flame/discharges as a function
of absorbed microwave power, for four sets of data (independent experiments).
The flow rate for the first three sets is 160/80 sccm O2/CH4, and 160/80/10
O2/CH4/N2 for the fourth set. Tables of raw data are found in Appendix B. ....... 108

Figure 5.28Temperature profile images from the IR camera of a flame as it is
modified by microwave power, 40 W and 60 W. The flame is not confined within
the cavity. The temperature scale is provided, but does not provide an accurate
temperature reading. The gas flow rate is 200/100 sccm O2/CH4. .................... 110

Figure 5.29 Temperature profiles sketched from the IR camera images (from
Figure 5.28) for both a combustion flame (represented by the solid lines) and a
hybrid flame/discharge with 40 W microwave power added (dashed lines). ..... 110

Figure 5.30 Temperature profile images from the IR camera for a plasma-only
discharge: 200 sccrn argon excited by 40 W microwave power. Each image is
recorded with a different filter. The image in the bottom right corner shows the
temperature profiles of the hot gases beyond the plasma ................................. 111

Figure 5.31 Temperature profile images from the IR camera for a stoichiometric
combustion flame formed by a ceramic nozzle. Different camera filters are used
to highlight different sections of the flame. The temperature profiles of the nozzle
are also illustrated. ............................................................................................ 112

xii

Chapter 1

Introduction and Literature Review
1.1 Introduction

Plasma-assisted combustion is an area of research that has been studied
for several decades. Thoroughly understanding the physics of this coupled
phenomenon and optimizing it have yet to be accomplished. The motivation
behind the current research at MSU is to improve the combustion process, to
conserve energy required for combustion, and to design and build an efficient,
small-scale, plasma-enhanced, premixed-flame, combustion torch. Increased
power, combustion efficiency, and flame stability are all advantages of combining
a non-thermal plasma (NTP) with combustible gases [1].

Experimental data are gathered and theoretical calculations are performed
in order to define the flame characteristics and behaviors;'flame temperature and
flame dimensions (height, width) are measured, relative specie concentrations
are observed, stoichiometric mass flow rates are calculated for three fuels, the
premixed flame theory is reviewed, and a ceramic-flow nozzle is fabricated.

The flame configuration employed in this project is, similar to other studies
[1, 2], a circular premixed flame burner in which the premixed gases are
combusted as they exit a nozzle into ambient air. The nozzle is enclosed in a
microwave cavity, which houses the microwaves, producing the flame/plasma
interaction. It is hoped that the experiments can demonstrate a quantitative
difference between a pure fuel/oxygen flame and a flame that has been

enhanced by microwave plasma. Varying levels of microwave power (measured

in Watts) and premixed reactant flow rates (and mixture ratios) are examined. A

schematic diagram of the experimental configuration is shown in Figure 1.1.

“my top
wall I plate
E-field V f

pattern

 

 

 

 

 

 

flame/plasma observation
window

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

water[

. .I
cooling gas leed

Figure 1.1 Schematic of experimental apparatus including the plasma/flame torch
inside microwave cavity. The torch houses a gas line and water line, and
microwave power is inputted midstream. A combustion flame is produced inside
a cavity that creates an electric field pattern exciting a TMo12 (transverse
magnetic) mode. A window is present for observation and diagnostic
measurements (optical emission spectroscopy (OES), for example) [3].

Flame stability, flame propagation speed, and combustion chemistry of
premixed, laminar flames are all affected by inducing an electric current in the

flow of gases. It is believed that the plasma assists in breaking down the

combustion fuel thereby producing free radicals (e.g. H, OH, O), which are
chemically unstable and therefore highly reactive. By using the plasma to convert
gaseous fuels into reactive species, the ensuing combustion process does not
rely on the self-generation of reactive species, and therefore hotter flames may
be produced. Plasma also yields a potentially more concentrated flame.

A possible mechanism for flame power augmentation is the acceleration of
the reaction by enhanced radical concentration levels. Thus, if the microwave
power is suitably and optimally “focused” it may produce, in the upstream gases
(prior to the flame sheet), a larger concentration of free radicals, H, OH, O, that
can accelerate the combustion reaction (see Figure 1.2). The accelerated
reaction may produce a narrow, more spatially concentrated flame, thereby
making a flame/plasma with a high power density. The practical goal is to employ
this enhanced power density to perform some kinds of cutting, machining,

material conditioning, localized heating, or other applications.

I Reaction I

l l
Fuel (CH4) Zone, Products
Oxygen 5' I
l l
| | Electrons/Ions
I | With Plasma
Intermediate I I
Species (H2. CO. GIG) I I Electrons/Ions
I Without Plasma

Upstream, 1". ° €L Downstream,
& an
............ ! .m...

Figure 1.2 Species and radical concentration curves in regions 6,, 6”, and 6”, of
the combustion process. With the addition of microwaves, the electron/ion
concentration is possibly greater than with combustion alone.

Microwaves are the type of electric field induced in the MSU torch.
Microwave power produces a clean discharge and has been studied by
researchers at MSU for many years [4]. Microwave plasmas are also referred to
as non-thermal plasmas (NTPs), meaning these plasmas are not produced by
extremely high temperatures but by some other means. NTPs have several
advantages over thermal plasmas. High gas and ion temperatures are produced
by thermal plasmas, whereas NTPs diminish the effects of high gas
temperatures, while still producing energetic electrons and other reactive
molecules (H, OH, 0, etc). As a result, with a microwave plasma it is possible to
consider a gas mixture with multiple temperatures. When only a select number of
molecules are excited, this results in little waste enthalpy associated with a gas

stream [1]. The MSU apparatus currently operates with a non-thermal plasma.

In addition to combustion and plasma considerations, there is the
important issue of nozzle design, materials, and construction. The nozzle must
channel the reactants into the flame; hence it must accommodate a controlled
flow of the gases in the design range of flow rate. The nozzle must also allow the
microwaves to concentrate in the appropriate locations so that the beneficial
flame/plasma interaction will in fact occur. Whether this concentration (of
microwaves) location is immediately in front of the flame or further upstream (in
the nozzle, for example) is a matter of research and study. Finally, the nozzle
material must allow the microwaves to concentrate without being absorbed, and
thereby rendered ineffective.

A plasma-enhanced torch is desirable for practical purposes. Industry is
interested in such a torch for cutting and metal forming. The small, concentrated
flame allows precision cutting. A plasma welder eliminates the mess that many
combustion-based welders create. Applications may even be designed for outer
space. Buoyancy does not play a role in the flame produced by the torch
because it is so small and concentrated. In addition, the torch is easy to handle
because of its small size.

To further develop this ongoing plasma-enhanced combustion research,
experiments are conducted and theoretical calculations are studied. A mini torch-
was built that allows combustible gases to flow through and be activated by an
electric field, producing free electrons, a current, and a plasma discharge.
Temperatures are measured using an IR camera, an optical emission

spectrometer (OES) for specific gas temperatures, and a thermocouple. Power

density is calculated based on flame/discharge geometry and absorbed
microwave power. In addition, a ceramic nozzle was designed and implemented
for experiments.

1.2 Literature on plasma-enhanced combustion

There has been no systematic and prolonged effort to study plasma-
enhanced combustion, but over the past several decades isolated attempts have
been made to enhance combustion with an electric field. A short account follows.

Plasma-induced combustion research was conducted at Los Alamos
National Laboratory in the past five years or so in an attempt to improve the
combustion process. More stable and efficient combustion was desired and
thought to be achieved by introducing a supplemented electric field [1].

The experimental setup consisted of a coaxial silent discharge reactor,
producing a dielectrical barrier discharge (DBD), or silent discharge. The propane
and air mix just before ignition and no converging nozzle is used to form a flame.
The gases are pre-treated by NTP prior to ignition. Rosocha et al. argue that
NTPs have advantages over thermal plasmas, as discussed in Section 1.1. NTPs
promote efficient combustion by increasing flame speed, decreasing flame
length, reducing soot formation, and reducing NOx and other combustion by-
product emissions.

The Los Alamos team found that with an NTP system, plasma increases
flame blow-out limits, but at a relatively low propane flow rate. The partial
pressures of common hydrocarbon combustion products (water and carbon

dioxide) increased while propane fragments decreased when the plasma was

added, which suggests that the propane is being burned more completely. The
flame propagation speed increased with an increase in electric field intensity.

Experimental work done by Chintala et al. at Ohio State University shows
that non-equilibrium RF plasma-assisted ignition and combustion achieves large-
volume ignition in premixed flows. At higher reactant flow rates and lower
temperatures, the RF discharge allowed ignition to occur, as compared with no
discharge. Leaner combusting fuel flows were achieved with the RF discharge.
Radical species such as CN, CH, C2, OH, and O atoms, measured by visible
emission spectroscopy, were detected in the hydrocarbon air flow [2].

Much experimental work has been performed by Starikovskii et al. in
Russia, quantifying pulsed nanosecond plasma-assisted combustion. They found
that the plasma discharge significantly decreases the ignition temperature and
that the flame’s blow-off velocity increased significantly for a low discharge
energy input. An increase in the flame propagation velocity can be explained by

the production of atomic oxygen in a discharge by the quenching of electronically

excited N2 and the dissociation of molecular oxygen on electron impact (02+e'—*

O+O+e'). A numerical model was developed to describe the influence of the

pulsed discharges on the combustion process [5, 6].

Experiments that were performed by Whitehair et al. at Michigan State
University use a high-temperature nozzle and discharge chamber materials for
the development of a microwave electrothermal thruster. It was shown that

regenerative cooling and flow stabilization improved thruster performance [4].

Chapter 2

Premixed Flame Theory

2.1 Nomenclature

“ [Ole-s c=§35r~xrmmmoogw>m

C

"(Denma-

N-<‘<><§C—I

Integration constant

Pre-exponential factor (cm3/gmol—s)

Species parameter, cpT/Q + YR

Constant-pressure specific heat (J/kg-K or J/kmol-K)
Arbitrary coefficients

Binary diffusion coefficient (m2/s)

Damkohler number

Activation energy (J/kmol)

Fuel

Thermal conductivity (W/m-K) or Rate constant (cm3/gmol-s)
Damkdhler number

Loss term (J/m3-s)

Normalized heat loss term

Lewis number

Flame thickness (m)

Third body

Reaction order or integer in gamma function
Products

Heat release rate (Jls)
Energy per unit mass (J/kg)
Normalized heat release
Reaction rate function

Mass generation rate (1/s)
Universal gas constant (J/kmoI-K)
Normalized flame speed

Flame speed (m/s)

Time (3)

Temperature (K)

Velocity (m/s)

Mean molecular weight (kg/moi)
Distance in dimensional system (rn)
Normalized mass fraction

Mass fraction (kg/kg)

Radical

Greek Symbols

Thermal diffusivity, k/pcp (m2/s) or exponential coefficient
Enthalpy ratio, (Tb-Tu)/Tb

Zeldovich number, -E(Tb-Tu)/(RuTb2)

Laminar flame thickness (m)

Normalized temperature, (T-Tu)/(Tb-Tu)
Normalized temperature after stretching, 8(1-6)
Burning rate eigenvalue

Exponential coefficient

Constant density (kg/m3)

Nondimensional length scaling, x’/(a/uf)
Normalized length after stretching, E/6
Normalized temperature

Molar consumption rate (moI/ma-s)

g «4 IIIW‘Dt >@a>ovm>o

Subscripts

Consumed
Burned gas
Burned

Flame

Fuel
Homogeneous
Laminar
Unbumed gas
Particular
Reactant
Burned reactant
Unbumed reactant
Radical

NZJJUIJ‘DC r‘21:T|"“I'DCTO
CCU

2.2 Single-step chemistry

One-dimensional, premixed-flame theory for single-step chemistry has
been studied for decades and is well understood [7]. Reactants combust and
form products and heat energy, as shown in expression (2.0). Equation (2.1) is
the governing energy equation for the transient propagation of a reaction front
into a traveling gas. The first term represents the rate of change of temperature

at position x; the second term represents convection in the gas moving with

velocity, u; the diffusion and reaction terms follow. The source/sink term, in the

simplified model of Equation (2.1), is replaced by boundary conditions on the

temperature and is redefined in Equation (2.5) as pq. Assumptions to be made

are: complete combustion; constant pressure and density; inviscid flow; and a
laminar and infinitesimally thin flame front. The analysis derived in this chapter is
referenced from [7] and also from notes of an independent study course taken
from the advisor of this thesis, Dr. lndrek Wichman.

Although the assumptions seem drastic, they provide simple mathematical
models while proving remarkable success in laying the foundations for
understanding the structure, properties, stability, and dynamics of many
combustion phenomena. For example, the density changes can be neglected
when considering the diffusive transport [1]. Explicit and asymptotic solutions are
found by making difficult problems tractable to mathematical analysis
[8,9,10,11,12,13,14,15]. provide examples of explorations of various types of
flames with a single-step chemistry model. An asymptotic approach, based on
high activation energy, is used in one-step chemistry analysis.

The temperature far upstream is given by a constant unburned value, Tu,
and the temperature far downstream is a constant burned value, Tb. Figure 2.1
schematically represents the temperature distribution for this idealized premixed

flame (PF).

Fuel + Oxidizer—-> Products + Heat (20)

10

2
fl+ufl=afl+sourcflsink (2-1)

at 3x 8x2

Equation (2.1) is invariant under the application of a Galilean

transformation according to the following definitions:

f=x+ut
f
w=u+u
f
t'=t

Essentially, the new x-coordinate follows the flame front since dx’/dt = dx/dt + Uf
and dx/dt = 'Uf at the flame; therefore, dx’/dt = 0 at the flame. The velocity, u, is
the flow speed entering the flame front from the upstream direction. Note also
that it is assumed that the flame front propagates at a constant speed, Uf. Without
this assumption a Galilean transformation is not useful. The transformation yields
the governing equation in (2.2) for the gas temperature on either side of the

flame.

3T 3T BZT
_ '_= _ 2.2
at' “ ex “3va I )

The location of the origin is arbitrary since x’—> x’ + constant produces exactly the

same equation as in (2.2); as a result, it is convenient to let the flame front sit at

11

x’= 0. The flame front is not affected by t = t’; therefore, in the coordinate system
attached to the steadily propagating flame front, the unsteady term in Equation

(2.2) vanishes (a(-)/at’ = 0). When the flow velocity is zero (u = 0), the

transformed velocity is constant (u’= Ur), and Equation (2.2) reduces to a simple

ODE, shown in Equation (2.3), in which the coefficients, Uf and a, are constant:

. dim—012T
f dx' (1er

(2.3)

The solution to Equation (2.3) is as follows:

/ /
T=C1+C2exm “f)

The physical domain is divided into two regions: one upstream of the flame sheet
(- °° < x < 0); the other downstream of the flame sheet (0 < x < °°). In the

upstream region, the two following boundary conditions are imposed:

T(—oo)=Tu and lim T=Tb

x—)0—

These boundary conditions yield Equation (2.4), also seen in Figure 2.1:

12

x/(a/uf)

T = Tu + (Tb — Tu)e (2.4)

The following boundary conditions are imposed for the downstream region to find

T=Tb as the downstream solution:
T<oo, as x—)oo and T=Tb, as x—-)0+

Across the PF the temperature gradient is discontinuous:

Ia—TI
ax flame

It will be shown later that this jump in temperature gradient is proportional to the

3T 3T
— — — =0— T —T / /
(ijxao— (axjxao- (b a) (a uf)

combustion heat release by the flame.

13

 

 

 

T
b
x/(a/uf)
T=Tu+(Tb—Tu)e
‘— Flame Front
Tu
x -——9 -m x, = 0 X —_’ "'m

Figure 2.1 Temperature distribution in relation to flame front for the simple
infinite-reaction-rate model described by Eq (2.3). Note the jump in temperature
gradient across the flame front.

In this work, consider that the coordinate system is fixed to the PF, and
that the flame has a steady speed, SL (SL = u,). In this case, the Galilean
transformation from the laboratory to flame-fixed coordinates works. In contrast
to the infinitesimally thin flame described on either side of the flame by Equation

(2.3), the energy equation is now written in the form shown in Equation (2.5).

This equation is valid everywhere in the gas! In the following expressions, Q is

the energy per unit mass, p is the density, the heat release rate (Jls) is 2, and i

is the mass generation rate (kg/s) :

14

CPE—I: E‘qu (2.5)
pquQr
r=AY£e_E/R“T

The species equation is given in Equation (2.6). The first term represents
convection, and the second and third terms represent diffusion and reaction,
respectively. Figure 2.2 schematically illustrates the mass fraction (YR)
distribution of reactants in the combustion process for the infinitesimally thin

flame.

dYR_ DdZYR .
Mfg-.0 TIE-’1‘” (2-6)

15

 

“'— Flame Front

x'/(D/u )

 

 

 

X—.-oo x=0 X—v+oo

Figure 2.2 Schematic of reactant mass fraction distribution for the infinitesimally

thin flame described by Eq (2.6) on either side of the flame, with i=0 on either
side. Note, once again as in Figure 2.1, the discontinuity in the gradient of YR,
[dYR/dx]=YRqu/D.

Equations (2.5) and (2.6), along with the specification of i: f(T,YR), serve
to complete the specification of the PF problem. The two unknown dependent
variables are T and YR. The number of unknown variables (and equations) can
be reduced to one by forming a coupling function defined by the sum, B = cpT/Q

+ YR, and by assuming p0 = k/cp (Lewis number of unity). As a result, Equations

(2.5) and (2.6) yield the following homogeneous equation and its solution:

16

2
u £12=D——d B
fdx aix2

/D
B=CO +Cle(ufx )
Since 8 must be bounded at x = + °°, it is required that C1 = 0. Thus, 8 = Co,
which is evaluated at x = - °°, where T = T00 and YR = YRU. This provides the

following expression:

 

The above relationship is exact, and it applies everywhere between -°° < x < +°°,

even inside the flame. At x = + °°, T = Tb and YR = 0 to find Tb = Tu + QYRu/cp as

the expression for the burned gas temperature. The relation for 8 gives YR in

terms of T, as follows:

C
YR =YRU —Ep(T—Tu) (2.7)

By substituting Equation (2.7) into r(T,YR), the mathematical problem can be

reduced to the solution of the energy equation only (Equation 2.8), subject to

T(°°) = Tu and T(+°°) = Tb:

17

 

2 c n
u .d_T=Dd T+ Q A YRU—J—(T—T) e'E/RuT (2.8)
fdx de pCp Q 11

Prior to solving for T(x), however, normalized dependent and independent

variables are introduced as follows:

 

T - T
Dependent: 6 E “
Tb _ Tu
x!
Independent: 6 E u f —
a

The following dimensionless parameters are also defined:

 

 

a _
Damkdhler number: B a —2 A YIgUI E / Ru Tb
u
f
. — E
Zeldovrch number: ,6 E (T — T )
R T2 b u
u b
' T b — T
Enthalpy ratio: A E u
T b

18

Substituting these definitions into Equation (2.8) leads to the following equation

for 9(f,), subject to 6(-°°) = 0 and 9(+°°) = 1:

d6 6129 ————_fl(l-9)
— = ——3 +D(1- a)" eI‘AII‘”) (29)
d5 d5

Equation (2.9) is solved in three zones that are subsequently matched.

The upstream-zone solution, already obtained in Figure 2.1, is shown in Equation
(2.4) and is also given by 9 = e‘5 for f < 0. The downstream solution in Figure 2.1

is T = Tb, or 9 = 1. The solution for 9(6) in the reaction zone remains to be found.
The reaction zone connects the upstream “preheat” zone to the downstream
“reacted gas” zone.

In the “inner” or reaction zone near 6 = 0, it can be shown that the diffusion
and reaction terms dominate over the convection term [15]. Applying a stretching
transformation demonstrates this.

A stretching transformation, or reaction zone scaling, utilizes the following

definitions:

Rescaled, nondimensional
temperature: 6 E fl(1 - l9)

Rescaled Damkdhler number: DE Afln+l

é
a

[II
III

Rescaled special coordinate:

19

Here 6 is an as-yet—undefined nondimensional scaling of the coordinate, if, in the
reaction zone, and is a measure of the nondimensional reaction zone thickness.

It is noted that the following expression is defined for convenience:

N I on
U
[u
M
1+

NIH

Integrating Equation (2.9) from -°° to +°°, or —6/2 to +6/2, suggests that it

is reasonable to define the quantity, 6043" as of the order of unity. Therefore, the

characteristic flame width is approximated by Equation (2.10) below. Note that
the nondimensional characteristic flame width is of the order 143, so that for large

[3 the reaction zone becomes very thin [15]. Thus 2 = [345 is defined.

~

z'Bn: '6" ~i (2.10)
D Afln+l fl

Equation (2.9) reduces to Equation (2.11) below, after applying reaction-zone

scaling factors, 9, D, and E:

fldE dEZ

 

2
1 d®:_d 9+A®ne—G/(l-AO/fl) (2.11)

20

The convection term is of the order 1/B which is significantly smaller that the
diffusion and reaction terms which are of the order unity. As a result, in the limit
of large B, the convection term is neglected to lowest order. Figure 2.3 illustrates

the normalized temperature distribution after the stretching transformation.

 

--v
-a“
-—"
,.
--
o
.0
-

 

:40

Figure 2.3 Schematic of stretching transformation in nondimensional scale. The
inner coordinate E is large and negative near 6 = 0 ', large and positive near § =
0", and of the order unity in the flame sheet.

It is presently impossible to analytically solve the nondimensional ODE in
Equation (2.11). Rather, an approximate solution is found. Perturbation
expansions of the normalized temperature and eigenvalue, 9 and A,

respectively, are performed. The following perturbation expansions are

substituted into Equation (2.11).
-l —2
O=Oo+fl ®l+fl (92+...
—1 —2
A=Ao+fl Al+,6 A2+...

21

This substitution results in the Equations (2.12) and (2.13) below for the zeroth

and first order problems:

a; = A093690 (2.12)

9" 9n n 1 9 - er 9" A1 A62
1‘ 0‘—_‘ 1‘ 0+ 0'_'_ o 91”
6 A0

The boundary conditions for Equations (2.12) and (2.13) are obtained from the

outer solution in the following way. First, the upstream solution is analyzed and

provides the following two expressions, as f —* 0':

2 3
6=e5=l+§+é—+é;+u.
2! 3!

d6? 2

——=1+ +——+”.
d5 62

Then, applying the relationships, 6 = 1 — (9/6 and f = E/B, yields the following:

22

From the perturbation expansion, letting O = 90 + 91/6 + gives the following

boundary conditions, and so on for the higher values:

dOO

H

90 —> —E and

 

—->—1 as E-—>—oo (2.14)

 

—-)—E as z —)—oo (2.15)

For the downstream solution, as § —+ 0", the following distributions are

found:

23

Once again, applying 9 = 1 - O/B and 6 = E/B gives 9 = 0 and dO/dE = 0, so that,

along with O = 90 + 91/[3 + the following boundary conditions are found and

are displayed in Figure 2.4:

 

 

dGO
(90 = O and = O as
d91
('91 = 0 and = O as
9

 

90(5)

‘
q
u
s
-
s
u.
5

 

Anticipated solution

‘-
‘Q
-
-
‘g
‘-
‘-
.s
-o
--
.....
---_
c
..............

(2.16)

(2.17)

 

 

Figure 2.4 Plot of anticipated solution after applying stretching transformation.

24

The lowest-order problem consists of Equation (2.12) and BCs (2.14) and
(2.16). The first-order problem consists of Equation (2.13) and BCs (2.15) and
(2.17). Both problems are over-determined and thus each solution will also
produce a result for the various eigenvalues, A0, A1, etc.

To solve the lowest-order problem, first the eigenvalue is calculated by
integrating Equation (2.12) from E = - °° to E = + °° and imposing the conditions

given by Equations (2.14) and (2.16). Therefore, the following computes:

 

 

=—..., e o 2

"if—W - .. ”I tigejzdefi _ 0f 1, d9,
_.,, daz “ E:_oo2d(-9 d3 d3 " 90%2 d3
__1_[d€-)0]2 _ [[deOT

2 a 90—20 2 d: 90%”

1 1
=—0—1=——

2[ I 2

O 3: oo
- 0( + )A 9" —90d@ -A De" ‘GOde — A r 1
T i 008 0‘ oioe 0—-0(n+)
90(52—00) oo

Thus, A0 = 1/2 l'(n+1) gives the lowest-order flame speed eigenvalue. The
procedure above is replicated, but integrated from E = - °° only, to a finite (and

unspecific) value, where 90 = OO(E). The result is as follows:

25

 

 

 

 

2
d9 n+1,9
(ca/"0] :1_°° : _Q(r( +1)O)EP( 1’90)
:. n
Isne—Sds
0
90
Isne—Sds
Q 0
P +1,@ =—= SI 2.18
(n 0) F 0° ( )

[5"?st
0

Equation (2.20) is obtained below by taking the square root of Equation (2.18)

and rewriting it as Equation (2.19). Equation (2.20) is solved numerically.

 

 

 

dOO
d3 =—\/P(n +1,€-)0) E—g(n,®o) (2.19)
90 .
ds
—E= azm
(i g(n,S)

The integration constant, a, is found by integrating Equation (2.20) while applying
upstream boundary conditions on Oo(E). A transcendental equation forces a
numerical solution for Oo(E), which is then used to deduce 6 and thus T, and

finally YR from Equation (2.7).

26

The expressions for O1 and A1 can be found by introducing a variable

change in the form of g, as follows:

 

dOO
d5
2
9 =—$ (221)
l 2 '

By a similar procedure (by integrating the first-order problem from the far
upstream (E = -°°) to the far downstream (E = +°°)). it is relatively straight forward

to find, for the first-order eigenvalues correction, the following expression in

Equation (2.22):

A1 °°
— = (n + 2)(n + 1) - 2 [(1— g)ds (2.22)
A0 0

The first-order analysis will not be pursued any further. The interested reader
may consult Chapter 5 of [16].
2.3 Multi-step chemistry

The analysis of a single-step chemical reaction does have limitations in

describing combustion behavior. For example, the Zeldovich number (B) for a

27

satisfactory description of stable flame balls must be about three times the actual
value found in typical hydrocarbon reactions [7].

Realistic Zeldovich numbers are invoked in the two-step mechanism
shown in expression (2.23). This expression describes how a radical species Z
attacks a fuel species F to produce more radical in the form of chain branching,
until all the fuel particles are consumed. A third body, denoted by M, represents
any type of molecule. In reaction 2 the radical collides with M to produce product
P and leftover M. The rate constant, kg, in the first reaction varies with
temperature; however, the rate constant, kg, for the second reaction does not.
The chain-branching reaction requires energy from the flame to generate
radicals, T3 = EB/Ru (heat release step). The second reaction has zero activation

energy, and is therefore independent of T. Thus, To = 0 (heat absorption step).

1) F+Z-—>Z+Z : k =A e'Tb’T

B B

2) Z+M—>P+M: kC=A (2.23)

C

schematically represents the first step in the above mechanism.

28

 

 

 

Figure 2.5 Curve representing a two-step combustion mechanism.

The governing equations for the two-step mechanism include the fuel

species equation, the radical species equation, and the energy equation, shown

in Equations (2.24), (2.25), and (2.26), respectively. The primes imply

differentiation with respect to x.

pSY1; =pDFY}: —W a)

F B

pSYZ =pDZYZ —WZa)F 'szc

pcpST =kT +QwC—r

29

(2.24)

(2.25)

(2.26)

The laminar flame speed is denoted by S, W is the mean molecular weight, and

we and we are defined as follows:

In Equation (2.26), the quantity I refers to a heat loss term that can be used to
represent the rate of bulk heat loss (through radiation, for example).
For simplification, the equations are normalized using the following

definitions for the independent (x —’ of) and dependent (Yz, YF, T) variables:

30

0 . a
x = JCS: (xs rs chosen for convenience, xs = —)
u

- . W Z LeZ
Y2 = yz Y ZS (YZS IS chosen for convenience, YZS = moi/7.1::
F

- S . O C .
s =— (u rs chosen for convenience later In the analyszs)
u

 

T — T
7: 0 (normalized temperature field)
T f — T 0

Y F = y F Y F0 (normalized fuel mass fraction)

The fuel, radical, and energy equations are now reduced to the following

nondimensional equations:

 

 

d d2

LeF‘ yF — yzF —Kr (2.27)
d: d;
dy dzy

Le a Z = Z +Kr—y (2.28)

31

2 _
§fl=d T+ Q y —L (2.29)
d6 6152 Le], Z

 

Note the similarity of these equations to Equations (2.5) and (2.6). Select

variables that are included in Equations (2.27), (2.28), and (2.29) are defined

below:
Lewis number: Le = 2
D
x2 pYFO (0st —TB /T
Damk6hler number: K S W A e f

 

=——_ F B
pDFYFO WF W2

£212
_ 7f T
Reaction rate function: r = y F yZe

QYFO
cp (Tf — T0)WF

 

Heat release: 6 =

(x2
L = S
k(Tf —TO)

 

Heat loss term:

The quantity r is redefined in terms of the Arrhenius rate expression k(T). as

follows:

32

 

1 1
TB[T T +r(T —T )]
k(T)=e f 0 f 0

The expression for k(T) is rearranged after some algebra to give the expression

as shown below, where a = 1 — Tom and p = T3(Tf-To)/T;2.

 

l—z'
k(T) = e-flil‘MHJ

To solve the fuel species equation, the reaction zone is assumed to be
very thin. Upstream (f < 0') the flame, the reaction rate r is zero. Therefore,

Equation (2.27) becomes Equation (2.30).

Ler’yF =yF (2.30)

Letting y;’ = exp(m§), it can be shown that the characteristic equation gives

L9F§ m = m2, so that the two roots are m = 0 and m = LeFE for the

homogeneous solution. The solution to Equation (2.27) is Written below:
yF = Cleo: + CzeILeFE)‘: (2.31)

The boundary conditions are as follows:

33

Far upstream: y F =1 as ,f —) —oo

On upstream side of
flame front: y F = 0 as f —> O—

The values for C1 and C2 in Equation (2.31) are computed, and the fuel species

upstream solution is given below:
y F =1 — 19”ng): for 5 < 0‘ (2.32)

The downstream (f > 0+) solution uses the same solution (Equation (2.31 )) as the

upstream solution. Equation (2.33) and the boundary conditions that follow lead

to the solution given below by Equation (2.34):

y F = C3 + C4eILeFEI5 (2.33)
Far downstream: y F <1 as 4‘ —> +00
At downstream side
of flame front: y F = O as 6 —> O+
Solution: y F = 0 for .5 > 0+ (2.34)

34

A curve demonstrating this behavior for the fuel species is shown in Figure 2.6.

Yr:

 

 

1 1
I
a=o-I b;0* 2

Figure 2.6 Curve representing fuel species behavior for a two-step mechanism.

Equation (2.27) can also be integrated, from a to b, to show that y): is continuous,
or that the jump of y): across the flame is zero [7].
The radical species equation, given in Equation (2.28), is solved by

implementing similar methods as the fuel equation. By integrating from a to b in
the above Figure 2.6, when a = 0' and b = 0+ are f—locations that straddle the

6
reaction front, the result, IKrdf = —§LeF , is found. Then following expression

a

is deduced, which yields the jump condition on the gradient of yz:

35

Iy'zlay'z

= —§LeF (2.35)

 

0“r yZIO‘

Letting y2 = exp{a§}, the characteristic equation obtained from Equation (2.28)

with r = 0 is Le2§ a = 02 — 1, whose solution gives the two roots 612, seen below:

 

a = LeZs
1,2 2

 

1 - 2
i§J(LeZS) +4

Solutions are found both upstream (§<0) and downstream (§>0) of the flame; the
unknown coefficients are set equal to each other because yz is continuous
across the flame front. The one unknown coefficient is then solved for using the

jump condition from Equation (2.35). The solutions are written below:

Le E

yz - F ea]: for §<O
“1‘0’2
Le E

yZ : F 802: for §>0
“1‘0’2

Equation (2.29), the energy equation, is solved using a coupled
homogeneous and non-homogeneous solution. Equation (2.29) is rearranged

into Equation (2.36), where L=a2 7, assuming I is linear in temperature.

36

2 __
d T 'd7 azrz—Q yz (2.36)

E7371?- LeF

Using the results from the species solution, Equation (2.36) can be written

as Equation (2.37):

r

e%,f<0
2 — — L -

d 2' E617 a2 Q y Q er

(152 d6 Le z LeF (a!1 —a2)

 

< (2.37)

 

Keaz, §>O

The upstream solution (§<0) is found from a homogeneous and particular solution

where r = r H + Tp. The homogenous equation for 1' is as follows:

+

\l4a2 +52

The following roots for a and p are redefined below, for convenience, noting that

.1.
2

11(1) and M2) are always positive:

 

1 _ 2 Lez§
a(1)':§\/(Lezs) +4 +——2—-= a]

37

 

1/ 2 _2 f
fl(2)=§ 4a +S —§=—fl2

These roots yield the homogenous solution for the normalized energy equation,

as seen in Equation in (2.38):

l1“); flag):

TH =Cle +C2e (2.38)

The particular solution is found from Equation (2.37) for the upstream (§<0) case:

2 —— _
i:_§£__027 = Q Ler can):
(1,52 d5 P Le F (am + am)

 

Letting To = Aexp{p(1)f), the following algebraic expression is found:

38

E Lops
Le}: (a(1)+a(2))

 

— 2
AW?” ‘ 50(1) — a )= 1401(1) — #(1) )(0!(1) + #(2)) = '-

Q Lela-S‘-

where A = — .
LeF (61(1) + “(2) )(0«'(1) —fl(1))(a(1)+/1(2))

The total solution for the upstream energy equation is written in Equation (2.39).

T=T +1 =

(2.39)
_ _ a 6
41(2): _ Q LeF s e (I)

“F (“(1) + “(2))(0‘01 ‘ ”(1) )(“m + ”(21)

 

,u
e (1)5

Cl +C2e

The coefficient C2 is zero because I < °° and 6 —+ - °°; therefore the second term

goes to zero. As 6 —> 0', the following expressions are obtained:

dr
2' _ =C -K a d — = C —a K
0 1 1 n [délo- #(1) 1 (1) 1
where K1 Q Ler

 

_ Le]: (“(1) +a'(2))(a’(1) —fl(1))(a’(1) +l’(2))

Downstream of the flame ({>0), 7 < °° as c,‘ —> + °°; therefore, C1 must be

zero and the first term in Equation (2.39) disappears. As 15 —> 0+, the following

expressions are found:

39

dr
2' 2C -K and — =- C +a K
0+ 2 2 [615]“ 111(2) 2 (2) 2
._ a LeF 5
LeF (0(1) + 0(2))(05(2) +fl(1))(0’(2) -#(2))

 

where K 2

Applying the jump condition leads to expressions for C1 and C2. The solutions are

shown below:

 

T : 5.? a1 —a2 e’ul‘f _ ea]; (6 < 0)
01-02 #1 "/12 (at ‘fllX-a’z +1111) (al-fll)(0!1—/12)

a1 _ 072 81126 8026

6';
r = -
0’1 ‘02 {#1 ‘1“2 (-02 ‘/12 )(0'1 #12) (—a2 HQ )(-(12 +#1)

 

1..

The definition for K was seen earlier as the following:

2
xs W A pYFO pYZS e—TB/Tf

K=———
pDFYFO F 8 WF W2

 

This expression can be simplified into the following:

—TB /Tf

40

The asymptotic analysis for a second order reaction shows that K = 32 [7].

Therefore, Tf is defined by Equation (2.40).

2 2
A —T /T T T
K=i£YFOe 3 f = .5. 1__(L (240)
AC WF Tf Tf

For various parameter choices the equations of this subsection can be
solved (and have been solved [7]). The solutions plot the flame speedj , versus

1/2a2

the parameters a2/(Q--1) for fixed Lez and Lez for fixed Q , showing that for

1/282

each parameter value (a2/(—Q— -1) and Lez ) there are either two values of the

flame speed 5 , or none (see Figure 3 of [7]). It is believed, consistent with
several decades of premixed flame stability analysis, that the lower branch is
unstable whereas the upper branch is stable.

The influence of the radicals enters through the pre-exponential factor, A.
As this increases (or decreases) the parameter containing A changes and the
influence of the radicals is modified. The influence of the plasma, it is believed,
can be examined by varying the reaction rate A, which is altered by the level of
the applied electric field. This influence is not examined in this thesis. Another
strategy, which also is not examined, is to postulate that in front of the main

reaction zone a radical formation region is produced by the plasma. Thus, at a

certain distance 6,, upstream of the reaction front the radical mass fraction takes

41

the fixed value yzp. The remainder of the solution is of course dictated by the

conservation equations. This solution, however, in preliminary calculations
becomes algebraically extremely complicated and it is not clear that a simpler

approach, based on scaling analysis, would not be more feasible.

42

Chapter 3

Equipment and Experimental Methods
3.1 Introduction to equipment

The test bed used to study the microwave/flame interaction consists of a
miniature microwave plasma torch/burner (PTB) that operates inside a cylindrical
cavity. The experimental conditions that occur can be divided into three distinct
groups of flames/discharges: 1) pure microwave plasma discharge; 2) pure
combustion; and 3) hybrid plasma/combustion discharge. Discussed in this
chapter are the overall setup, including the torch, cavity, electrical network, and
gas flow, along with the experimental methods employed for data collection and
analysis. Images in this thesis are presented in color.
3.2 Overall experimental setup
3.2.1 Torch
3.2.1.1 Introduction to torch

The PTB used for this research was designed and built by researchers in
the MSU Electrical and Computer Engineering (ECE) department 1. It was
originally designed for the formation of plasma (argon, helium) in order to study
micro-scale plasmas (plasma diameter and height of 1mm and 3mm,
respectively) [17, 18]. It operates under atmospheric pressure, but has the
capabilities of operating in low pressures down to O(10‘1 atm) also. This pure
microwave plasma discharge produces relatively low temperatures (1600 K-1800
K) which limits it to cutting paper and melting plastics. As a cutting tool, the
torch’s discharge must cut plastics and metals with precision; therefore,

1 Under review for US. Patent in 2005

43

combustion is introduced to produce a high-temperature, hybrid flame/discharge
that consists of a plasma discharge coupled with a flame [19].
3.2.1.2 Torch design and dimensions

The torch is considered “small-scale” because of the flame/discharge size
it produces, which is of the several millimeter scale. Photographs of the brass
PTB are shown in Figure 3.1. It is a simple hand-held torch comprising coaxial
inner and outer conductors, with inlet/outlets for microwave power, gases, and
water (used for cooling). The photograph to the left in Figure 3.1 shows the torch
before assembly. It includes, from bottom to top, the outer conductor, the fine-
tuning adjustor (which is placed inside the outer conductor and controls the
placement of the inner conductor, with the help of a teflon spacer used for
centering), the inner conductor, and the microwave coupling unit; the finger

stock, allen wrench, and bolts are also shown.

 

Figure 3.1 Left: photograph of disassembled PTB. Right: photograph of
assembled torch demonstrating hand-held easiness.

Figure 3.2 is a schematic of the torch with dimensions. The outer
conductor has an 11.1 mm inner diameter, and the inner conductor has an outer
diameter of 4.75 mm. All major dimensions of the torch are listed in Table 3.1.
The torch is designed so the inner conductor can slide through the outer
conductor in axial adjustment (useful for assembly and tuning of the discharge
inside the cylindrical cavity). It is centered with two teflon spacers. Fine tuning is

available at the base of the torch with an adjustable microwave tuning short.

Water Water Input

Microwave
output

power

   
   
   
 
 
 

‘Gases: fuel and oxygen

Adjustable microwave tuning short

Microwave coupling structure that allows inner
tube to slide through

Outer conductor

ID: 11.1 mm
Center conductor

k1. / OD:4.75mm

Nozzle exit diameter: 0.4 mm

Figure 3.2 Internal structure schematic drawing of the PTB with the brass nozzle.
This unit is also called the “torch applicator.” Note the center conductor and the
water cooling system. The inner conductor is adjustable to create an optimal
electric field that is concentrated at or near the tip of the nozzle [18].

A 1/16” gas line feeds premixed fuel and oxygen (mixed ~ 3 ft upstream of
torch base) into the nozzle entrance. Because the brass nozzle is soldered to the
inner conductor, flowing water prevents the solder from melting due to flame

temperatures above 2000 K. The water inlet and outlet are at the base of the

torch, and the water is allowed to circulate throughout the inner conductor. The

45

microwave current is fed into the torch by a 50-ohm coaxial cable with an N-type
connection. The inner and outer conductors create an electrical field that ionizes

the gases at or very near the tip exit of the nozzle.

Table 3.1Torch part dimensions in millimeters and inches.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Torch Dimension Units.

mm Inches
Inner Conductor Length 197.488 7.775
Inner Conductor OD 4.763 0.188 (3/16)
Inner Conductor Thickness 0.356 0.014
Outer Conductor Length 148.590 5.850
Outer Conductor OD 12.70 0.5 (1/2)
Outer Conductor Thickness 0.813 0.032
Nozzle Inlet Diameter 3.340 0.132
Nozzle Exit Diameter 0.40 0.016
Microwave @ut Coaxial Cable Diameter 6.350 0.25 (1/4)
Gas Line Diameter 4.234 0.167 (1/16)
Inlet Water Line Diameter 4.234 0.167 (1/16)
Outlet Water Line Diameter 4.234 0.167 (1 Mg

 

The nozzle, like the rest of the torch, is made of brass, and was
designed for relatively easy machining. Drawings of the nozzle can be seen in
Figure 3.3, which shows gas/water lines and dimensions. The exit diameter is 0.4
mm. A second nozzle was made with a diameter of 0.2 mm; however, most
experiments presented in this thesis are performed with the 0.4 mm-diameter

nozzle.

46

 

.__l
00
X
'0
$4.—
__f__
.0
.h

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

——+ 4——0.4 mm 1‘ 0.9
I /
A ’ __J/ /
/ .
20 5 p 1 6
/ / g g
,JSolder / 5 z;
/ r
: Gas // § 1‘ ; fife—()4
line " 1
:Water : I j i Q
d line / 3.3
1
4 mm 4 4'0 *—

 

 

 

 

Figure 3.3 Left: cross section of nozzle with gas and water lines. The gas line is
soldered to the wall of the nozzle. Water cooling is necessary to keep the solder
from melting. Water enters the inside of the inner conductor near the tip of the
nozzle and circulates throughout, and exits near the base of the conductor. Right:
cross-sectional drawing of the nozzle with dimensions in millimeters.

3.2.1.2.1 Ceramic nozzle design and fabrication

Innovative design and fabrication of a second, ceramic nozzle are
explored and experimentally tested. The original nozzle of the torch cannot
withstand temperatures that the combustion flame produces, and also reflects

poor flow geometry. This nozzle is made of brass and is soldered to the inner

47

conductor. To prevent the solder from melting and detaching from the torch, the
inner conductor is water cooled.

Water circulates through the inner conductor to keep the nozzle at a
relatively low temperature. A water-cooled nozzle causes the flame to stand off,
forming a gap between the flame base and the nozzle exit. A detached flame is
not desired because it is unstable and also loses heat energy to the cooled
nozzle. In addition, the brass nozzle reflects poor fluid flow. The previous Figure
3.3 shows a drawing of the brass nozzle. The sharp corners induce flow
separation and can cause considerable circulation patterns. To eliminate the
necessity for water-cooling and to develop improved flow conditions, a smoothly-
converging ceramic nozzle has been designed and fabricated. A sketch and

photograph of a ceramic nozzle are shown in Figure 3.4.

——>l F—O.4 mm

'1.

1' '1

1
1 |
“C )4

 

|
l
l
.1:—
_L

 

 

 

3/16”
or 0.5 cm

 

Figure 3.4 Left: sketch of ceramic nozzle showing smoothly converging path from
nozzle inlet to exit, including diameter dimensions. Right: photograph of ceramic
nozzle attached to the inner conductor. Note the gold-sputtered coat on the
ceramic.

48

The process for designing and fabricating a ceramic nozzle involved first shaping
graphite “molds” to form the smoothly converging flow region, then preparing the
powder used for pressing and sintering.

The design of the nozzle is developed from shaping graphite rods; no
specific equation was developed for the converging shape. The graphite (ordered
from McMaster Carr) is used to form the converging shape of the nozzle. Rods of
1/4” diameter and lengths varying from 1 - 4 inches are tumbled in bottles lined
with sand paper of various grits. The tumbler or rolling machine can be seen In
Figure 3.5. Rods are first tumbled inside a bottle (3" diameter) lined with 40-grit
paper for 24 - 48 hrs. The ends of the rods become rounded, closely resembling
a parabolic curve. This curve is manually plotted and curve-fitted using Excel.
Spotlight, a software program developed by NASA, is also useful for such
analysis, thus eliminating the time spent manually plotting points along the rods
edge. These procedures demonstrate that the rods are nearly parabolic at the
ends. These rods are then tumbled with 400 grit paper, and finally with 1200 grit
paper for 12 - 24 hrs at a time. The higher-grit paper creates a polished,
reflecting surface on the rod. The rods are then finely polished with a diamond
paste. The ends of the rods (noses) are cut off with a low-speed diamond saw.
The parabolic noses are used to press the ceramic powder into its nozzle shape.

The ceramic powder is made with 85% weight alumina (Al203) and 15%
weight zirconia. Alumina is affordable and offers a 99.9% purity level. It is the
close packing of aluminum and oxygen atoms within the structure that leads to its

good mechanical and thermal properties [20]. Zirconia is added for increased

49

strength. The powder is tumbled/mixed for 24 hrs to ensure a homogeneous
mixture. Approximately 10 g of powder is axially pressed under 2000 psi using a
cylindrical die after inserting the graphite tip. Figure 3.5 provides schematics of

the pressing process.

Axial pressure
2000 psi

    
  

Parabolic
graphite
mold

   
   

   
   
 

Inner surface
of nozzle

\

Inner die cylinders

Flow exit

 

 

 

 

 

 

Figure 3.5 Schematic of ceramic nozzle. Left: ceramic powder being pressed
axially by two cylinders at 2000 psi inside a cylindrical metal die and shaped by a
parabolic graphite mold. Right: nozzle after pressing and drilling of exit pathway.

Note the outside corners of the nozzle are still square. Careful hand sanding
rounds these corners (represented by dotted lines) so the flame does not sit on a

radially-large surface.
The specimens are partially sintered at 800°C for 4 hrs using a heating
rate and cooling rate of 5 °C min'1 in a digitally controlled Carbolite RHF 1500
furnace with SIC heating elements, shown in Figure 3.6. Partial sintering hardens

the material but also allows an opportunity for machining the ceramic. The exit

diameter (~ 0.4 mm) is drilled and the tip of the outer shell is sanded to form a

50

rounded tip. The specimens are then fully sintered at 1475°C for 4 hrs. During
this process the specimens decrease in size, approximately 17% less than the
original diameter. The ceramic particles exhibit “necking” or bridging between
surfaces of adjacent particles, creating a denser overall nozzle. After full sintering
no machining is possible for the ceramic nozzle. The final outer diameter of the

nozzle is ~8.5 mm.

 

(a) (b) (C)
Figure 3.6 Photographs of: (a) tumbler used for mixing powders and for shaping
graphite, (b) axial press used for pressing powder in the die, and (c) digitally
controlled furnace used for sintering specimens.

Several nozzles have been fabricated. The nature of the fabrication
process does not produce Identical nozzles; each nozzle may be slightly different
due to graphite shifting during pressing. Therefore, the best designs (axially

aligned convergence path) are chosen for experimental analysis. The nozzle fits

51

snug onto the end of the inner conductor and epoxy glue is used as a sealer to
prevent gas leakage. A photograph of a ceramic nozzle attached to the torch is
shown in Figure 3.4.

An optional step in the manufacturing process of the ceramic nozzle
includes gold coating or sputtering of the ceramic’s outer surface. The ceramic
seems to absorb the microwaves that the flame should absorb. Therefore,
sputtering the ceramic nozzle with a metal (gold in this case because of its high
temperature endurance) cause the microwaves to converge in the flame zone at
the exit of the nozzle. An approximately 20 um-thick gold coating is sputtered
onto the surface using a sputter machine. The ceramic nozzle shown in Figure
3.4 is sputtered with gold.

3.2.2 Electrical network and cavity

Figure 3.7 provides a schematic of the overall experimental setup. The
torch operates in plasma-only, combustion-only, and hybrid, plasma-enhanced
combustion modes. For plasma production, power is supplied by 2.45 GHz
microwave energy. The power absorbed by the system is measured by the
difference between incident and reflected power, and is defined at the coaxial
input line termination of the PTB. Both incident and reflected power are
measured by 50-dB and 30-dB directional couplers, respectively, and are read
using HP 435A power meters connected to HP 8481A thermistor power sensors.
The reflected signal is fed into a 50-ohm matched load or dummy load (coaxial

resistor model 8201 ).

52

   
 
    

    
   

Microwave I TOP
Cavity Plate
Incident Applicator - Camera/OES
Microwave Power I

GPIB

Computer

     

Reflected

Power . Water
Dumm Gas. ' .
Loaii fuel 8. Cooling
oxygen

Figure 3.7 Overall experimental setup, including electrical circuit, torch and
cavity, gas system, and diagnostic setup [3].

The microwave input power is side-fed into the torch applicator. The

system is optimally tuned by repositioning the torch so the reflected power reads

zero. Therefore, under these conditions, the incident power equals the power

absorbed inside the cavity. The outer conductor and the inner conductor are

independently adjustable. Tuning is also available by moving the top plate of the

cavity in the vertical direction (see Figure 3.8). The coaxial inner and outer brass

conductors allow an electric field to form in the pattern of a TM012 (transverse

magnetic) mode inside the cavity (also shown in Figure 3.8) [18]. As a result of

optimal tuning, the nozzle tip is positioned so the flame burns in a concentrated

area of the electric field while absorbing microwaves. The cavity is sealed for

safety to eliminate microwave radiation to the environment. The two viewing

windows are covered with metal screening.

53

Figure 3.8 shows a schematic of the torch inside the cavity. The torch is
placed vertically inside the cavity and held in place by finger stock. The
cylindrical cavity is 7 inches (17.78 cm) in diameter, with a maximum height of 23
cm. The optimal operating height for this project’s experiments is 14.2 cm.
Eleven millimeters of the outer conductor is inside the cavity and an additional 7
mm of the inner conductor is also inside the cavity, as shown in Figure 3.8.
However, constant tuning of the torch/cavity is necessary for optimal interaction
of microwaves and flame; as a result, these dimensions are not always exact for
optimized interaction (i.e. zero reflected power). This inconsistency is caused by

varying flame size, power input, and power-meter drifting.

54

\\\\

y

’/

\

 

 

 

 

 

7 5f? / /
- 5 Z /
Cav1ty Z Z Z Top plate

wall Z Z Z Z
Z WWW. 7//////////////////.’< Z n 7 mm
Z TM012 Z +
5 "‘09 Z 11 mm
2’ /
/ / .
6 % out3lde
5 LS Z .
Z E-field Z cav1ty
Z pattern Flame Z X
5 ’ , fl
5 1 Z
g I 'I .4 %
Zlmzxxxx’Z/zg . 7//.¢¢;¢7:¢¢2¢¢¢%|%

f”? Input

TORCH I I microwave

51 power

Water ..
C00"“9 Gas line

Figure 3.8 PTB and microwave cavity setup, including details of the cavity and
the electrical field pattern and wave modes. The torch is placed vertically in the
cavity and tuned so the flame interacts with the microwaves, creating a plasma-

enhanced hybrid flame. The dimensions of the cavity-penetrated torch are given.
3.2.3 Gas system

The gases used for combustion are methane (CH4) and oxygen (02).
Methane in 02 produces a relatively high theoretical adiabatic flame temperature

(Tab = 3000 K), and yet is safer than hydrogen. It is readily available as “house

gas” in the lab. Oxygen, as opposed to air, is used to produce a hotter, more

55

concentrated flame. The fuel and oxygen are premixed approximately 3 ft. before
the nozzle exit plane.

The gas flow is controlled by MKS digital mass flow controllers (MFC). The
MFCs are calibrated for each type of gas, and the readout is 10.001 sccm
accurate. A schematic of the gas system is shown in Figure 3.9. Gas flows from

pressurized tanks to the MFCs through 1/4” flexible metal gas lines using VCR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

connectors.
(1:331 _
Microwave
cavity
(V |Ml=c MFC NFC IMFC
n I I ._.L_.
A 03
C E) w
O2 CH4. N2 H2
P micro
Torch
Water—
coofing

 

 

Premixed gas line

Figure 3.9 Schematic of the gas setup. Gas flows from the compressed tank,
through MFCs. It is then premixed before entering the torch. Methane (CH4) and
oxygen (02) are the primary gases used in the experiments. The microwave

power is fed through the side of the torch applicator (Pmicm) and the water cooling
lines are also shown.

56

The gases are mixed downstream of the MFCs where a single 1/4" plastic tube
replaces the multiple gas lines. This single gas line is fitted to a 1/16” plastic tube
which is connected using epoxy glue to the torch gas inlet (1/16” brass tube) at
the base. Gas then flows through the inner conductor to the nozzle and is ignited
at the exit of the nozzle using a standard butane lighter. Flow rates from the
nozzle range from 15 sccm to 309 sccm (2 m/s - 41 m/s) at the nozzle exit for the
0.4 mm-diameter brass nozzle. Reynolds numbers range from 1 to 50, using a
constant 02 kinematic viscosity of 4 cm2/s (value interpolated at 2000 K). Table
3.2 lists the maximum flow rates in volumetric units (sccm) and length units (m/s)
for each nozzle type. The ceramic nozzle allows a slightly higher maximum flow
rate than the brass nozzle of the same diameter. This is most likely due to stand-

off and temperature gradients between the flame and the nozzle.

Table 3.2 Nozzle flow limits. Maximum flow rates for each type of nozzle, in sccm
and m/s. Note the ceramic nozzle allows a higher maximum flow rate than the
brass nozzle with the same diameter.

 

 

 

 

 

Brass Brass Ceramic
Nozzle Type 0.2mm 0.4mm 0.4mm
diameter diameter diameter
Max Total Flow Rates 75,10 309,41 321,43
(sccm or m/s)

 

 

 

3.3 Diagnostics and measurement techniques

Experimental techniques used to measure temperature and diagnose the
hybrid flame/discharge include: 1) optical emission spectroscopy (OES), used for
identifying species and for calculating average rotational temperatures; 2) an

infrared (IR) camera, used to examine relative temperature values and contours;

57

and 3) a type K thermocouple with a data acquisition system used to find local
temperatures across the flame. The first two methods are most useful for the
hybrid experiments because they do not perturb the flame since there is no
physical contact. Even though a thermocouple is useful for obtaining point
temperature measurements, it interferes with the electric field created by the
microwave system. The cavity has two windows that are used for viewing and
diagnostics.

3.3.1 Optical emission spectroscopy (OES)

OES is used to measure and analyze the photons or light emitted by the
flame/discharge. Its primary function for this project is to calculate the gas
temperature of a particular gas-phase species, nitrogen and carbon rotational
temperatures in this case, which has been performed in a number of previous
experiments [21,22,23,24]. The rotational temperature is essentially a measure of
the energetic level of the molecule at which all rotational modes of the molecule
are fully engaged. The theory of gas kinetic temperature is explained in Chapter
4. This temperature is a Iine-of-sight, average temperature. Flame species can
also be identified using OES since different atoms and molecules emit unique
wavelengths in the spectrum; however, a quantitative value (i.e. emission
intensity) is not known from this spectrometer.

The McPherson Model 216.5, 0.5 meter, f/8.7, plane grating
monochromator is the OES system used. The operating wavelength range is
1050 A — 5000 A, from a grating of 2400 grooves/mm. A schematic of the setup

is shown in Figure 3.10. Light is emitted from the flame/discharge and focused by

58

a biconvex lens (focal length of 5 cm) into the spectrometer. The adjustable
entrance slit is set to a width of 50 pm.

The monochromator diffracts the incoming light. Since the angle of
diffraction varies with wavelength, only specific wavelengths exit and are
measured. The monochromator is manually set to a specific wavelength by
rotating the diffraction grating about the axis at a specific angle. The signal of
species present (light based on wavelength) is amplified by a photomultiplier and
read by the picoammeter to certify that ample signal is being captured. The
system is covered by a black cloth to reduce any external light signal, thus
enhancing the signal—to—noise ratio.

The spectrometer is connected to a computer by a GPIB (general purpose
interface bus) cable and data is collected from a Quick Basic program written by
Jayakumaran Sivagnaname [22] to extract wavelengths of various

flame/discharge species.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cavity V Monochromator
(:23, _. A ...... 1---- ............ Entraneex
Focusing 3'”
Flame/ I lens
discharge
L :' Grating
High voltage
Computer input
. Exit J
Plcoammeter """ slit """"""
' .='= EPIB Photo Collimating

 

 

 

multiplier mirror

 

Figure 3.10 Schematic of OES system. Light is emitted from discharge, analyzed
in the monochromator, and collected and organized in the computer.

The system (particularly the torch and lens) is aligned so a maximum
signal is fed into the entrance slit. The entire image of the flame is focused
through the lens, supplying data for an average overall flame temperature.
Rotational temperatures calculated from the OES method come from C2 and N2
emissions.

3.3.2 Infrared camera

Infrared (IR) cameras are useful for nonintrusively measuring
temperatures without directly contacting the sample (i.e. flame), unlike intrusive
thermocouples that can cause flame perturbations and conduct heat away from
the flame. A major obstacle or limitation of the IR temperature analysis, however,
is the inability to measure the flame while inside the cavity. The cavity's viewing

window is protected with a metal screen, to prevent microwave leakage. The

60

camera cannot bypass this; because it measures the first object in sight, the
cooler metal screen is detected rather than the flame inside the cavity.

A metal stand or platform was built to prop the torch in either the vertical
or horizontal direction, for operation outside the cavity. With the torch in an open,
unprotected atmosphere, the disadvantage is it cannot be optimally tuned to
couple the microwaves with the flame due to leakage. IR measurements were
collected from the flame while the torch was outside the cavity.

The camera measures the radiation an object emits, since an object, not
at absolute zero temperature, will always emit or absorb energy In the form of
electromagnetic radiation. A blackbody, the ideal object with an emissivity (a) of
unity, absorbs and emits energy with no reflection. Examples include the sun, a
candle flame, and even the flame of a welding torch! Even though torch flames
have almost identical properties and characteristics as the flame used in this
research, the process for measuring temperature with the IR camera is not so
simple (1: #1). Because there are hot gases surrounding the flame, the emissivity
is no longer unity; it is a function of temperature, 5(7), and in addition the
temperature varies with location. Thus, IR measurements of gas temperatures
are much more complicated than surface temperature measurements.

The Stefan-Boltzmann law, defined in Equation (3.1) below, is used to
measure the energy, R( T), emitted from a surface relative to a black body, where
the energy is a function of temperature and proportionally related to the Stefan-

Boltzmann constant, 033, defined in Equation (3.2).

61

R(T) = £O'SBT4 12
m (3.1)

2 5k 4 _
053: ’r 332 =5.67x10 8 2W4
15h c m K (3.2)

 

 

The Stefan-Boltzmann constant is defined as a combination of three other
universal constants, the Boltzmann constant, k3, Planck’s constant, h, and the
speed of light, 0 [25].

IR detectors are commonly used to measure temperature because for
objects at temperatures of 10-5000 K, the radiation predominantly occurs
between wavelengths of 0.2-500 pm in the IR spectrum. IR cameras monitor the
temperature from an array of IR sensors, and each sensor represents a pixel in
an acquired thermal image [26].

The IR camera used for this research comes from Mid-range Merlin from
Indigo Systems Company, a division of Flear International. It has interchangeable
25 mm and 50 mm lenses, has focal length adapters for distant viewing, and
contains filters used for higher temperatures. Windows 2000 and Therrnagram

make up the computer’s operating system and software. Further specifications
are available at http://www.indigosystems.com/product/merlin.html.

The camera uses a cooled InSb detector array and is designed for use
with 3-5 pm wavelengths, corresponding to temperatures of approximately 273-
1000 K. The temperatures of the flame/discharge studied in this project produce

a temperature beyond that range (2000-3000 K); therefore, only relative

62

temperatures are extracted from the analysis, used to compare combustion-only
and hybrid flames. Temperature contours throughout the flame body are also
obtained.
3.3.3 Thermocouple

A type K thermocouple is used to measure flame temperature (combustion
only). Type K Chromel-Alumel thermocouples are rated for a long range capacity
of 173-1643 K. As mentioned previously, combustion produces a much hotter
flame than the thermocouple is capable of measuring. Therefore, the solution to
the transient heat conduction equation, assuming the lumped heat capacity
method (Biot number < 0.1), was used to calculate temperature based on time-
step iterations, and can been seen in Equation (3.3) below. Here T(t) is the
temperature as a function of time, To is the initial temperature (room temperature

in this case), Ts is the surface temperature, h is the heat transfer coefficient, A is
the cross-sectional area, p is the density, CD is the specific heat, and V is the

volume, all with respect to the thermocouple material or junction.

T(t)—TS
—= — / V
T0_TS exm (hA pcp )t (3.3)

The thermocouple is welded together to form a spherical junction. It must
be noted that the junction is not shielded (may account for inaccuracies in
results). Because the sphere is so small (1.7 mm diameter), the lumped capacity
method is implemented. The heat transfer coefficient is calculated based on the

Whitaker Nusselt (Nu) number correlation for a sphere, seen in Equation (3.4)

63

and a general definition in Equation (3.5) below. This correlation if valid for
Reynolds numbers between 3.5 and 7.6e4 and Prandtl numbers ranging from
0.71 to 380 [25]. The Reynolds number is denoted by Re, where Re = Ud/v and
U is flow velocity, dis diameter, and v is kinematic viscosity. The Prandtl number,

Pr, is defined as Pr = v/a where a is thermal diffusivity. Dynamic viscosity, (1, is

the viscosity of the fluid at Too, and us is the dynamic viscosity of the fluid at the

surface evaluated at the film temperature, T), the average temperature of the

solid and fluid, L is the characteristic length, and k; is the thermal conductivity of

the fluid (02 gas in this case). All parameters, except #3. are evaluated at Too.

Nuavg = 2 + (0.4 Re0-5+ 0.06 Re” 3) Pr0'4(,u / 1190-25 (3 4)

hL
Nuavg = —

kr (3.5)

An Excel spreadsheet is created with known constants and used to calculate the
temperature for a 0.5-second time step. Reynolds numbers ranged from 3-375,
based on a constant 3.5 m/s flame speed (calculated from cone angle) and
temperature-dependent O2 kinematic viscosities. This spreadsheet is found in

Appendix A.

64

A motorized device built enables the thermocouple to sweep through the
flame at a constant speed, while the data acquisition system reads the voltage
difference and converts to temperature. A Hewlett Packard (model # 3852A) data
acquisition system/control unit, connected to a computer with a GPIB cable, is
used. The time delay for this system is 0.1 3. Photographs of the control unit and

motorized device are shown in Figure 3.11.

 

Figure 3.11 Left: photograph of the data acquisition system used to measure
temperature with a Type K thermocouple. System is connected to a computer
with a GPIP cable. Right: device built to sweep thermocouple through flame at a
constant speed. Note the spherical thermocouple junction.

65

Chapter 4
Theory and Background of Gas Rotational Temperatures

4.1 Introduction

Flame/discharge temperatures can be approximated by measuring gas
temperatures. The gas temperatures are calculated by measuring the rotational
energy levels of various molecules present in the flame/discharge. The
homonuclear, diatomic molecules that are measured in this study are nitrogen
(N2) and carbon (C2). Rotational temperatures of these gases are calculated
based on signal (wavelength) measured by optical emission spectroscopy (OES).
4.2 Rotational temperature theory

The total energy of a given state of a diatomic molecule is the summation
of electronic energy (Te), translational energy (Tt), vibrational energy (G), and

rotational energy (F), as written in Equation (4.1) in wave-number units.

T=Te+T,+G+F (4.1)

In general, rotational energy changes in a given vibrational and electronic
state are small compared with the thermal translational energy, therefore making
F a small number. Conversely, gas molecule collisions produce changes in the
vibrational or electrical quantum numbers much less frequently than in rotational
quantum numbers precisely because the rotational energies are lower than the
others and changes are easier to excite [27]. The rotational temperature is a

reliable measure of the gas kinetic temperature because the rotational population

66

distribution in a sufficiently long-lived vibrational state has a Boltzmann
distribution [27,28,29,22].
The rotational energy listed in Equation (4.1) is defined in Equation (4.2)

below in terms of J, the rotational quantum number which takes the values 1, 2,
3,..., By, the rigid rotator rotational spacing or rotation radius, and By, the first

anharmonic correction to the rotational spacing.
F = BvJ(J+l)—D,,J2(J+l)2 +... (4.2)

In addition, there are nonrigid rotator corrections to both 3,, and DV. These
corrections are defined in Equations (4.3) and (4.4) below, where v is the

vibrational frequency, 8,, and 0., are constants that correspond to the equilibrium

separation, and 0,9 and [3,, are the first anharmonic corrections. Constants for the

first order rotational energy are listed in Table 4.1 below for N2 and C2 [27, 22].

B, = Be —ae(v+%)+... (4.3)

D, = De + ,Be(v+%) + (4.4)

Table 4.1 Rotational constants for the electronic states of nitrogen and carbon.

 

 

 

 

 

 

 

State B. a.
N2 (C3pu-B3pg) 1.8259 0.0197
02 (d3pg-a3pu) 1.7527 0.01808

 

67

Each molecule is denoted by a specific quantum state. For this research
the Second Positive System (Capu—tBapg) is used for N2 values and the Swan
System (danga3pu) is used for C2 values. These systems are now explained.
The upper (u) and lower (9) states may have different electronic angular
momenta, A. Therefore, two or three series of lines or branches may appear: the
P, Q, and R branches. Shown in Figure 4.1 are these three branches, which

result from transitions in vibrational and rotational energy levels.

 

 
   
 

Q Branches

v“
I

A .

Figure 4.1 Energy level diagram for a band with P, Q, and R branches, relative to
wavelength (A), rotational energy levels (J), and vibrational levels (v) [27].

68

For both N2 and C2, AA = 0 because the angular momentum at both the upper
and lower electronic states is zero; therefore, the AJ = 0 transition is forbidden
and it follows that AJ = :l: 1. The R branch is associated with AJ = +1 and the P
branch is associated with AJ = -1. The Q branch is not present for N2 and C2
because AA = 0. It is important to note, however, that all branches may be
present for a specific electronic transition with a different angular momentum, as
shown in Figure 4.1 [27].

Generally, emission methods, as opposed to absorption methods, are
suited for measuring the temperature in a Boltzmann distribution. The relative
emission intensity (I) of rotational lines is described as Equation (4.5) by [30].

This intensity is not related to size.

 

(4.5)

B. ' '+
1:16.48). J. exp[— VJ” 1W]

kT,

The intensity is defined by: K, the constant for all lines originating from the same
electronic and vibrational level; v, the frequency of the radiation; 8,1211, the
appropriate H6nI-London factor; 8.1, the molecular rotational constant for the
upper vibrational level; J, the rotational quantum number (J’ represents the upper
energy level and J” represents the lower level) ; h, Planck’s constant; c, the
speed of light; k, the Boltzmann constant; and T,, the rotational temperature.

It is seen from Equation (4.5) that the intensity is proportional to the
frequency, v, to the fourth power and also to the exponential ratio of rotational

temperature, T,. As the wavelength of an emitted species decreases, the

69

intensity increases to the fourth power; frequency is defined as the speed of light
divided by wavelength, v = cl). As temperature increases, so does the intensity
exponentially.

The H6nl-London factor describes the line strength of rotational spectra,
and it is dependent on J. The H6nI-London formula is defined in Equation (4.6)

for emission associated with the R branch, where AA = 0, [27].

= (J'+A'+1)(J'—A'+I)

 

S 4.6
J '+1 ( )
Since AA = 0, Equation (4.6) reduces to Equation (4.7) below.
S = J '+1 (4.7)

The experimental data is fitted to the expression / ~ SJuexp{-

BVJ’(J’+1)hc/(kT,)}, using several (11 or more) R-branch emission lines to

determine the rotational temperature, T,. The plot of In(IIS) is a linear function of
the upper rotational energy for the diatomic molecules used in this study. Chapter
3 describes the experimental procedure for measuring the spectra of various
flame/discharge species, which are used for the temperature calculation. Figure
4.2 pictorially shows how the light signal is emitted from the flame/discharge, and

then focused through a lens into the spectrometer.

7O

Flame/
Discharo e

  

 

 

Spectrometer

 

 

 

 

Figure 4.2 Laboratory process for temperature calculations: light is emitted from
the flame/discharge and focused by a lens into the spectrometer, where
intensities and wavelengths are recorded.

4.3 Gas temperature of N2

Rotational temperatures are first calculated based on the N2 spectrum,
since this molecule has been extensively measured [21]. Nitrogen is measured in
two different sources. In the first set of experiments, only methane and oxygen
serve as the feed gases; N2 is detected from the atmosphere (air) as the
flame/discharge reacts. In a second set of experiments, a small amount (~ 2%) of
N2 is fed into the incoming methane and oxygen streams. Parameters that are
varied in each experiment include microwave input power and gas flow rate. The
system currently operates at atmospheric pressure.

The Second Positive System (SPS) is used to determine the N2 rotational
temperatures. The SPS describes the energy level transition from Capu to Bapg.
Figure 4.3 is a plot of an N2 spectrum, measured in the laboratory, identifying 11
emission lines (R20 — R30) in the spectrum range 3758 A - 3783 A of the (2,0)

transition in the SPS.

71

 

 

 

Intensity (a.u.)

 

 

 

 

 

I r T

3753 3763 3773 3783 3793

Wavelength (A)

Figure 4.3 Plot of intensity (or current since the units are arbitrary) versus
wavelength for the N2 spectrum, commonly used for rotational temperature
calculations. R20 - R30 list the order of the rotational bands.

Table 4.2 provides values used in the rotational temperature calculations.
These values include the rotational lines/bands (R20 — R30), wavelengths
recorded from text (the wavelengths gathered from the experiments of this thesis

are compared to book values to ensure accuracy), relative upper-level energies

calculated from Equation (4.2), and the H6nI-London factor calculated from the

corresponding rotational band.

72

Table 4.2 Values used for the N2 rotational temperature calculations for the R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

branch of the (2,0) SPS.
Rotational Wavelength Relative upper-level S
Line (A) energy (cm-1)

R20 3780.44 837.76 19.80
R21 3778.58 917.42 20.81
R22 3776.66 1000.67 21 .82
R23 3774.68 1087.51 22.83
R24 3772.64 1177.95 23.83
R25 3770.53 1271 .97 24.84
R26 3768.37 1369.57 25.85
R27 3766.14 1470.76 26.85
R28 3763.86 1575.51 27.86
R29 3761.51 1683.84 28.86
R30 3759.1 1 1 795.74 29.87

 

 

The natural log of the ratio, l/S, is plotted against the relative upper-level
energy, and the gradient of a linearly-fitted line is used to calculate the rotational
temperature in Equation (4.5). Figure 4.4 displays an example of a Boltzmann
plot for the R20 — R30 bands. This plot is created from a set of experimental
flame/discharge data with a methane/oxygen flow rate of 80/160 sccm combined
with 40 W microwave input power. Rotational excitation N2 temperatures are
calculated even though N2 is not present in the gas flow; the flame reacts with N2

found in the surrounding air, similar to a diffusion flame.

73

 

R20

R30

 

 

 

 

-2.5 1 . 1
700 1050 1400 1750
Relative upper level energy (cm-1)

Figure 4.4 Boltzmann plot for the bands R20 — R30 of the N2 spectrum.

The accuracy of rotational temperature is within 1 100 K, as estimated from the
reproducibility of the data obtained.
4.4 Gas temperature of C2

Rotational temperatures of C2 are calculated in a fashion similar to N2
rotational temperatures. The carbon source is methane, CH4, the fuel used for
combustion in this thesis. The light emitted from the breakdown of CH4 eventually
to C2 and the subsequent C2 excitation is processed in the spectrometer and the
wavelengths and intensities of the C2 spectrum are extracted and plotted.

The Swan System (d3nga3pu) is used for calculating the C2 rotational
temperatures [24]. The emission intensity is measured using the (0,0) band head,

with an intensity of 513 nm. Figure 4.5 is a plot of a C2 spectrum, Identifying 21

74

emission lines (R25 — R45) in the spectrum range 5028 A — 5089 A of the (0,0)

transition in the Swan System.

 

 

(0.0)
i
3 R45 R25
'1 \ \
4999 5019 5099 5059 5079 5099 5119 5139
Wavelength(A)

Figure 4.5 C2 spectrum of a Swan System, identifying the bands R25 — R45.

Values used to calculate C2 rotational temperatures are listed in Table 4.3
and are obtained as described in Section 4.3 for N2 rotational temperature. Note
the difference in the H6nl-London factor, defined previously by Equation (4.6),

takes a new definition for the C2 analysis [24], written in Equation (4.8) below.

5 = (J'+1)(J'—1)

J' (4.8)

75

Table 4.3 Values used for C2 rotational temperature calculations for the R branch
of the (0,0) Swan System. Note the change in the H6nl-London factor.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RotLaiggnal Wav(eAl?ngth Reéitggyuafiglfvel S = (J'+11(J"11’J'
R25 5089 564.40 24.96
R26 5086 609.55 25.96
R27 5084 656.44 26.96
R28 5081 705.07 27.96
R29 5079 755.43 28.97
R30 5076 807.53 29.97
R31 5074 861 .36 30.97
R32 5071 916.94 31.97
R33 5068 974.24 32.97
R34 5066 1 033.29 33.97
R35 5063 1 094.07 34.97
R36 5060 1 156.59 35.97
R37 5057 1220.84 36.97
R38 5054 1286.84 37.97
R39 5051 1354.56 38.97
R40 5047 1424.03 39.98
R41 5045 1495.23 40.98
R42 5042 1 568.1 7 41 .98
R43 5038 1642.84 42.98
R44 5035 1719.25 43.98
R45 5031 1 797.40 44.98

 

 

As with the N2 calculations, the natural logarithm of l/S is plotted against
the relative upper-level energy and is linearly fit to extract a rotational
temperature from the slope of the line. An example of this plot is shown in Figure
4.6, gathered from experimental data using the ceramic nozzle, having
methane/oxygen flow rates of 60/180 sccm, and no microwave power present.
This is a pure flame, not a flame/discharge. The uncertainty of the measured gas

temperature is approximately 100 K based on reproducibility.

76

 

-2.5
.26 1

.27 .

    

-2.8 - R25 -) R45

.29 -

In(l/S)

{3.05
(3.1 -

-3.2 ~

 

 

-3.3 1 I f 1 1 r
5(1) 700 fl) 11(1) 111) 1500 17(1) 191)

Mafive mper-Ievel energy (cm-1)

 

Figure 4.6 Boltzmann plot for the bands R25 — R45 of the C2 spectrum, for the
180/60 sccm O2/CH4 pure flame.

Results from OES measurements of the rotational temperatures for N2 and

C2 molecules are tabulated and discussed in Chapter 5 of this thesis.

77

Chapter 5

Results and Discussion
5.1 Introduction

The results in this thesis are collected primarily from experimental
procedures. Much preliminary work was done building an expandable foundation
for this research because plasma-enhanced combustion is not a fully developed
and clearly understood area of research. The research takes two approaches: 1)
studying the fundamental concepts of plasma-enhanced combustion phenomena
and understanding the underlying effects of such a hybrid system; and 2)
engineering a practical torch/burner to be used in industry.

Results collected and evaluated include flame/discharge photographs,
geometric measurements which account for flame speed and power density
calculations and non-dimensional power plots, extinction (blow-out) data, species
data from OES, and temperature measurements using different methods. Data
are collected from both brass and ceramic nozzles; however, focus is on data
from the 0.4 mm-diameter brass nozzle. Where the nozzle is not specified, it is
safe to assume the 0.4 mm-diameter brass nozzle is used. Images in this thesis
are presented in color.

5.2 Flame/discharge photographs

The torch has gas flow limitations it cannot exceed to produce a flame,
depending on which nozzle is used. Table 5.1 lists the maximum total flow rates
for three nozzles: 1) 0.2 mm-diameter brass nozzle; 2) 0.4 mm-dlameter brass

nozzle; and 3) 0.4 mm-diameter ceramic nozzle. These limitations are recorded

78

based on stoichiometric (2:1 sccm) methane/oxygen flow rates, stoichiometric
referring to the ideal combustion process during which a fuel is burned
completely. The minimum flow rates are not measurable due to limitations of the
mass flow controllers (MFCs) i.e. the flame is maintained at a low flow rate of
18/9 sccm O2/CH4, and when the rate falls to 9 sccm and below the MFCs do not

measure accurately.

Table 5.1 Gas flow limitations for three nozzles with units in sccm and m/s.

 

 

 

 

 

 

 

Brass Brass Ceramic
Nozzle Type 0.2mm 0.4mm 0.4mm
diameter diameter diameter
Max Total Flow Rates 75 ’ 10 309 ’ 41 321 ’ 43
(sccm or mls)

 

Several photographs are taken, using a digital Nikon D70 Outfit camera, of
flame configurations using the three nozzles mentioned earlier. Figure 5.1
includes photographs of pure combustion flames with varying flow rates, using
the 0.4 mm-diameter brass nozzle. As the O2 flow rate increases for a fixed fuel
rate, the flame decreases in length. The length is increased with a higher overall
CH4 flow rate, but still follows the trend that as 02 increases, flame length
decreases. Table 5.2 lists the flame lengths for each flow rate. The stoichiometric
flame parameters are highlighted in the table. As the overall stoichiometric flow
rate increases, the flame length also increases from 3.7 cm, to 5.0 cm, and finally
to 6.5 cm. As the total flow rate increases by 33% from 200 sccm to 300 sccm,
the length increases by 43% from 3.7 cm to 6.5 cm. The flame length increases
almost linearly with an increase in total stoichiometric flow rate, yielding a

coefficient of determination (R2 value) of 0.9903. Notice that as the fuel rate

79

exceeds the 02 rate, the flame front increases in size. The flame front is the
outline of the white inner core of the overall flame. The blue volume beyond the

flame front is the heat released from the chemical reaction.

 

   
 

O23 30 65 100 135 170 200
CH4: 5,

 

 

Figure 5.1 Photographs of pure combustion flames with varying flow rates,
measured in sccm. The 0.4 mm-diameter brass nozzle is used.

80

Table 5.2 Lengths of flames with varying flow rates, corresponding to Figure 5.1.
The shaded cells highlight stoichiometric flow rates.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mass flow rate (sccm) Length of Flame
Methane (C H.) Oxygen (02) (cm)
65 30 5.0
65 65 4.5
65 100 4.0
65 135 3.7
65 1 70 3.2
65 200 3.1
85 30 6.5
85 65 6.0
85 100 5.7
85 135 5.2
85 170 5.0
85 200 4.7
100 30 9.6
100 65 8.4
100 100 8.0
100 135 7.5
100 1 70 7.2
100 200 6.5

 

 

 

 

Figure 5.2 provides photographs of a pure plasma made from 200 sccm
argon (Ar2). The plasma discharge is much smaller in length and diameter than a
pure combustion flame. The temperature is also much lower than a flame
temperature (~1800 K vs. ~3000 K), as will be discussed later in this chapter.
Notice that the torch is not confined within the microwave cavity. Microwaves do
not escape the torch when the inner conductor is kept inside the outer conductor
and the microwave power is held at a relatively low value (10 — 40 W). The
plasma discharge, when inside the cavity, can become very long (> 10 cm) when

the torch is tuned and the microwave power is increased.

81

 

Figure 5.2 Photographs of a pure plasma discharge, with a flow rate of 200 sccm
argon and 30 W microwave power. Note that the torch is not confined within the
cavity.

Photographs in Figure 5.3 show how a pure combustion flame is affected
or altered by the presence of microwaves. A stoichiometric flow of 40/20 sccm
O2/CH4 is burned as 40 - 100 W microwave power is added, shown in the bottom
row of photographs. As 40 W microwave power is added to the flame, the torch is
adjusted and tuned inside the microwave-sealed cavity (the metal screen used to
prevent microwave leakage from the cavity’s window is visible in the
photographs) to trigger optimal flame/microwave interaction. The flame
transforms into a hybrid discharge as it changes color and increases in size (four
times its size from pure combustion to a 100 W-hybrid flame/discharge). The
position of this optimal tuning is referred to as the “sweet spot," where also the
reflected power of microwaves is zero.

The top row of photographs in Figure 5.3 shows flame/discharges with a
flow rate of 40/0 sccm O2/CH4 and with 40 — 100 W microwave power. In other
words, an oxygen plasma is present. Once the flame transforms from the pure

combustion form Into a hybrid flame/discharge, the fuel is reduced to zero flow.

82

The fuel-free flame/discharge is smaller (5 mm shorter in length at 100 W) than

that which burns fuel.

40/20 s

300

 

Figure 5.3 Photographs of flame/discharges inside the cavity as 40-100 W
microwave power is added. The 0.2 mm-diameter brass nozzle is used. The flow
rate for the top set of photos is 40/0 sccm O2/CH4, i.e. no fuel is present. The flow

rate for the bottom set is 40/20 sccm O2/CH4, a stoichiometric ratio. Scales are
provided for flame dimensions. Camera speed and aperture are 1.6 and 3.5,
respectively.

Photographs of a pure combustion flame and microwave-induced
flame/discharges formed by a ceramic nozzle are shown in Figures 5.4 and 5.5.
The flow rate of the flames in Figure 5.4 is 200/100 sccm O2/CH4, and 20 W, 40
W, and 80 W microwave power is added. The photographs do not show evidence
of a plasma-enhanced combustion flame; the color, shape, and size of the
original flame are not affected by microwaves even with fine adjustment (zero
reflected microwave power), from what the human eye can detect. However,

OES measurements show that the flame is indeed altered by microwaves and

will be discussed later in this chapter.

83

The ceramic nozzle may absorb the majority of the microwaves, rather
than the flame, or the gases might be excited by the microwaves upstream the
nozzle. At this point an explanation is uncertain. Figure 5.5 shows that even
though the flame Is not physically altered by microwave power, the flame

increases in length as the total flow rate Increases.

O

W 20 W 40 W 80 W
ZOO/100 S - l l l
C 3.55
c
m cm

Figure 5.4 Photographs of flame/discharges inside the cavity as microwave
power is added, 20-80 W. The 0.4 mm-diameter ceramic nozzle is used. The flow
rate is 200/100 sccm O2/CH4, a stoichiometric ratio. A scale is provided for flame
dimensions. Camera shutter speed and aperture are 1.6 s and 3.5, respectively.

 

80/40 80/40 140/70 140/70 200/100 200/100

Figure 5.5 Photographs of flame/discharges inside the cavity with a ceramic
nozzle comparing 0 W to 80 W of microwave power.

The sets of photographs in Figure 5.6 show very clearly how the
microwaves alter a pure combustion flame, formed with the 0.4 mm-diameter
brass nozzle. The flow rates are listed in the figure for each set, and a scale is
provided to estimate flame/discharge dimensions. It appears that the microwaves
start to dominate after 30 — 40 W microwave power is added to the flame.

To further investigate, photographs are taken at 1-Watt increments from
31 W to 40 W, and then again from 39 W down to 33 W, shown in Figure 5.7. As
power increases the flame transforms at 38 W; however, as power decreases, a
major change does not occur until 35 W. A hysteresis is present. There appears
to be a delay in the change of state (transition from hybrid form to combustion
form) when the power is decreased. Further theoretical investigation for an

explanation of this behavior is advised.

85

ANOJ-bmmV

 

10W20W 30w 40w 50w 60W 70w 80W 90w100w cm

70 sccm CH.
53 sccm O2

1
CW 10W 20W 30W 40W 50W 60W 70W 80W 90W 100W

        

Figure 5.6 Photographs of flame/discharges inside the cavity as microwave
power is added, 10-100 W. The 0.4 mm-diameter brass nozzle is used. The flow
rate for the top set is 70/24 sccm O2/CH4. The flow rate for the bottom set is
153/70 sccm O2/CH4. A scale is provided for flame dimensions. Camera shutter
speed and aperture are 1/30 s and 3.5, respectively.

86

 

31W 33W 35W 37W 38W 39W 40W

39W 38W 37W 36W 35W 34W 33W

Figure 5.7 Additional photographs from Figure 5.6. The flow rate is 70/24 sccm
O2/CH4 and all other conditions are the same as listed in Figure 5.6. The top
sequence shows the flame/discharge as microwave power increases; the bottom
sequence is power decreasing. This intermediate behavior gives rise to a
hysteretic affect.

 

Flame velocity, or flame propagation speed, is calculated from the
geometry of a flame. The cone angle, a, is the angle between the center of the
flame and an edge, when the flame is modeled as a two-dimensional cone and
assumed axisymmetric. This angle is measured, and along with the flow rate of

gases that is measured, u, these measurements are used to calculate flame

87

speed, 82, as defined in Equation (5.1). Flame speeds are calculated for the
flames in Figure 5.7, as the microwave power decreases from 39 W to 33 W, and
are listed in Table 5.3. As microwave power decreases, the flame speed also
decreases. Significant change occurs at 35 W when the flame/discharge
transforms from hybrid form to combustion form. Most probably microwaves
excite the species in the combustion process at a higher level and accelerate the
radicals/ions at higher rates. Typical book-value flame speeds for an O2/CH4
flame are ~ 100 cm/s.
SL=usina (5.1)

Table 5.3 Flame speeds calculated for flames shown in Figure 5.7 as power
decreases from 39 W to 33 W. Calculations are based on Equation (5n.

 

 

 

 

 

 

 

 

Cone Angle Microwave Power Flame Speed, SL
4099) (W) (cm/S)
12 39 259
12 38 259
9 37 195
9 36 195
3 35 65
3 34 65
2 33 44

 

 

 

 

 

5.3 Geometric analysis

Along with flame speeds, deduced from flame geometry are quantities
such as power density and non-dimensional power as a function of flame height.
Power density is calculated first by modeling the flame/discharge as a cone to
calculate a volume; it is defined here as the addition of microwave power and
combustion power of a flame/discharge divided by the volume of the

corresponding flame/discharge. Combustion power is approximated using the

88

higher heating value (HHV) tabulated in text for gaseous methane, based on
stoichiometric combustion with air, a value of 55,528 kJ/kg, and a CH4 density of
0.68 kg/m3 [31]. Values of combustion power in this research range from 10 - 40
W.

Power densities are plotted against non-dimensional power in Figure 5.8
for the two sets of data shown in Figure 5.6. Non-dimensional power is defined
as PP/(Pp + PC), where Pp is microwave (plasma) power and PC is combustion
power (PC is provided in the figure for each flow rate). With the absence of
microwaves, power density is greater for the flame with a lower total flow rate,
which is also slightly more fuel-lean. The flame volume is significantly larger for
the second set of data (153/70 sccm O2/CH4).

The power density is greatest just before the flame transforms in size and
color, suggesting the flame/discharge of highest intensity is formed when ~ 30 W
microwave power interacts with the flame. An intense and highly concentrated
discharge is desired, but temperature must also be considered. As microwave
power increases, temperature increases, as discussed in Section 5.5.

Figure 5.9 shows how the power (both microwave and combustion)
physically alters the flame in terms of geometry. Flame height is plotted against
non-dimensional power. An exponential growth in flame height occurs as

microwave power is added to the flame, for both sets of data.

89

 

—e— 70/24soom02/CH4 Flow
0.5 -~a~153/70$CUT102/CH4FIOW

 

 

‘7

0.0 0.2 0.4 0.6 0.8 1.0
PP/(PP'I'PC)

Figure 5.8 Power density as a function of non-dimensional power, evaluated from
flame/discharge geometries shown in Figure 5.6, with flow rates of 70/24 sccm
O2/CH4 and 153/70 sccm O2/CH4. The powers of combustion are provided: 15 W
for the first flow and 44 W for the second flow.

 

 

 

F 8
7 1 o 70/24soanO2/CH4 Flow 0
6 l a 153/70soon02/CH4Flow °
1
E 5 5 E] El El 0
E o o
.m 4,, 0
2 n °
3 I 0 °
5 21 D U 0
1 9‘” 0
o o o 0
o E , 7, ., - , ---. E- . , , .
0.0 0.2 0.4 0.6 0.8 1.0
PP/(PPI'PC)

 

 

 

Figure 5.9 Flame height as a function of non-dimensional power, evaluated from
flame/discharge geometries, shown in Figure 5.4, with flow rates of 70/24 sccm
O2/CH4 and 153/70 sccm O2/CH4.

90

It is possible to deduce that the combustion-like flame transforms to the
hybrid flame/discharge form as the microwave power exceeds the power
produced by the combustion process. The combustion power of the flame for the
set of data with a flow rate of 153/70 sccm O2/CH4 is calculated as 44 W. The
photographs in Figure 5.6 show that between 30 W and 40 W microwave power,
the flame dramatically changes in appearance, to what resembles more closely a
plasma than a flame. However, the combustion power for the second set of data
(70/24 sccm O2/CH4) is ~ 16 W, but the flame does not transform at a microwave
input power near this value.

It must also be noted that the torch is very sensitive to positioning/tuning.
The “sweet spot” is dependent on several factors: 1) cavity height; 2) inner
conductor position relative to outer conductor position; 3) outer conductor
position inside cavity; 4) gas flow rate; and 5) microwave input power. Because
all these factors must be considered, there are no exact measurements for
factors 1, 2, and 3. The experimentalist has a feel and general idea of how each
should be positioned; however, fine tuning is always required at the start of each
experiment to find the sweet spot.

5.4 Impact of microwaves on combustion

To further show how microwaves impact a combustion flame, a flame
extinction (or blow-out) plot is created and several OES scans of the
flame/discharge are made.

Flame extinction plots are powerful for analyzing combustion efficiency in

terms of fuel efficiency. An extinction plot is made to show that, with the addition

91

of microwaves, a flame is maintained burning less fuel than is required when no
microwaves are present. Figure 5.10 is an extinction plot of experimental data
showing this very effect. The figure plots the equivalence ratio versus total
volumetric flow rate. The equivalence ratio, rp, is a combustion term, defined as rp
= (VF/V0x)/(VF/Vo,.)s,o,-ch, where V is the volumetric flow rate in sccm; (p = 1 is the
case for stoichiometric flow, (p > 1 represents fuel-rich flow, and rp < 1 represents
fuel-lean flow, the range plotted in Figure 5.10. The figure includes values of
combustion power for each gas flow rate.

The data is collected by first creating a stoichiometric flame, lowering the
fuel slowly (considering delay of MFC responses) until the flame blows out, and
then recording the flow rates at the point of extinction. The data points in the
figure are the values of total flow rates at the point where the fuel flow rate
causes flame extinction, while holding O2 flow constant. The equivalence ratio is
calculated based on the fuel flow rate when blow out happens. The experiment is
run several times to ensure accuracy of results; accuracy ranges from :I: 1 sccm.

Three extinction curves are plotted: 1) combustion-only data; 2) hybrid
data with 10 W microwave power; and 3) hybrid data with 20 — 100 W microwave
power. With the addition of a minimal amount of microwaves (10 W), the flame
continues to burn at a lower fuel flow, a flow that in not possible without the
presence of microwaves. The average fuel flow rate reduction is 9%, ranging
from 5 - 12%, not including the first data point in the first two curves. The sweet
spot cannot be (easily) located at such a low flow rate; as a result, the

microwaves do not affect the blow-out curve. At 10 W microwave power, the

92

flame/discharge does not completely transform in size and color as seen in
previous data. Although no physical alterations are evident, the flame is affected
by microwaves.

When 20 - 100 W of microwave power is added to the flame, a physical
transformation takes place and a “discharge” is maintained with no fuel present.

The third curve in Figure 5.10 reflects this behavior.

 

0.8 5
0.7 .
0.6 a
0.5 1

Ratio

0.4 a
0.3 -
0.2 4

 

90/o a9 + Chrbustim ody
—--e—— 10 W
0-1 1 .13.. 20100 w

Eq
1.
8

 

00 r1 1:1 13 _m n
- w— 1: w a ‘El

70 12) 170 220 270

VTor (scam)

Figure 5.10 Flame extinction or blow-out plot of equivalence ratio versus total
volumetric flow rate, VTOT. Each data point is provided with the combustion power
of the flame at a particular flow rate, where fuel is present. Note that a discharge

is maintained with no fuel when 20-100 W of microwave power is added. A
spreadsheet of raw data can be seen in Appendix B.

I!

 

 

 

5.4.1 Optical emission spectroscopy (OES)

OES is used to detect species present in the flame/discharge, leading to
gas temperature calculations, and is used as a tool to gain perspective on how
microwaves affect a flame. The molecules/radicals scanned in this research

include N2, C2, 02, OH, and CH, all commonly found in hydrocarbon combustion

93

processes. The OES equipment used in this research is not capable of
measuring quantitative values of intensity inferring the amount of CH, for
example, present (arbitrary units, a.u., are labeled in plots; the scale could be
labeled as current rather than intensity); however, relative intensities provide
insight.

Photographs in Figures 5.4 and 5.5 in 5.2 show that a combustion flame
formed by a ceramic nozzle is not physically altered by the presence of
microwaves. OES analysis provides evidence that the flame may in fact be
altered. Figure 5.11 is a plot of C2 emission spectra, detected in a combustion-
only flame and also in a hybrid flame/discharge with 40 W microwave power. All
variables but microwave power are held constant. After 40 W microwave power
is added to the flame, the C2 emission intensity increases. Figure 5.24 in this
section supports these results. The plot provides wide-range spectra for a
combustion flame and a hybrid flame/discharge. Again, the Intensity is greater
with the presence of microwaves, suggesting the microwaves excite these

molecules/radicals at higher levels.

94

 

6. 5% 2

5.E09« -———Corrbtslion
-~--~Hybfid(40V\0 2i

4. E-09 ~

3. E—09 1

Intensity (a.u.)

 

 

 

 

 

 

25095 , .; . .,

A ' I '. I I It

. 2,111,135 '3: 5 .1251 «5.1-05:54:39 ,"

: - lz';"=: , 19'". Eli" ‘- ‘33" -

15094 .‘ lr- ‘ 1511.54, '. .
0.3m T I T I I T I I 1
5070 am 509) 5100 5110 5120 5130 5140 5150 516)
VibvelengflHA)

 

Figure 5.11 OES scan of C2 in a stoichiometric (200/100 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40 W microwave power)
discharge using a 0.4 mm-diameter ceramic nozzle.

The next set of plots, Figures 5.12 — 5.14, show N2 emission spectra in
combustion flames and hybrid flame/discharges. The general trend follows that
as microwave power increases, N2 emission intensity increases. Three
independent sets of data are presented to show consistency within experimental
results. Figure 5.14 shows intensities resulting from an N2-injecti0n flame (N2 is
flowed into the gases upstream ignition). Peak intensities vary in all three
experiments; this is probably caused by alterations in the experimental setup, i.e.

the distance between the lens and flame may differ, or the torch may not be

tuned exactly the same each time.

95

 

 

2E—O71

2E—07 —

htensity (a.u.)
Fri
«2

5.E-(B<

 

 

0.E+(I) *7

 

 

WU»

Figure 5.12 OES scan of N2 in a stoichiometric (160/80 sccm 02/CH4)
combustion flame and hybrid (combustion plus 40, 50, 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean" is a flow rate of
160/54 sccm 02/CH4. The last set of data is from a discharge of only 160 sccm
02 and 50 W of microwave power (no fuel is present). Data from 30 W of
microwave power and below show no N2 signal, similar to the combustion-only
case, and therefore are excluded from the plot.

96

 

 

 

 

 

 

 

 

 

Figure 5.13 Second set of data: OES scan of N2 in a hybrid (160/80 sccm O2/CH4
combustion plus 10-80 W microwave power) flame/discharge using the 0.4 mm-
diameter brass nozzle.

 

6.E-(B

Intensity (a.u.)
P: so .A w
a U :3;

$5

 

iIO"
é

 

Figure 5.14 OES scan of N2 in a hybrid flame/disfigé with a flow rate? W
160/80/10 sccm O2/CH4/N2 with varying microwave power, 30-80 W. Note that N2
is injected upstream into the flow of gases, CH4 and O2.

‘3’—
: --‘1s—J

   
    

.‘igr

 

MW

97

    
  

   

  

«77 de
—~—— l'M’Jidl‘U/V
l: - WW
{‘9} -~---|'deN
it?“ “““ W070”
‘1 [15:55! ----Hytxidww
.1“.’Rln."vlii"’l,fl.r:; ‘5‘
115“” 5| ll‘flai}; ‘1‘ .."Vl' )1} 'I
523 ll, , ., ,. Ni"
. , .1. 5 :5 ‘ I
gfiéa‘f‘lhpml . “V“‘My‘l “dowry“ "“l‘
h . -.'.1 I . _ ’ ' I _ '5 :‘l 1N1 .‘. ’55 VJ‘f-‘i
1 m1? {‘3le' J Ix‘lhr~,_.\’,m',--:I;"' 1'..." w In

(343% ,, WMM‘W
*7 , v r

3775 37%

 

Figures 5.15 and 5.16 show C2 emission spectra in combustion flames
and hybrid flame/discharges. The general trend follows that as microwave power
increases, C2 emission intensity increases. Two sets of data are recorded to

show consistency in experimental results.

 

    

 

 

1.508— 1
,il
1.50m l. 1 :1“ 1" 1 1‘
11111171151 NW ‘1‘F4 115111r‘i1‘11‘1111l‘rl‘f‘1 11*”1‘1iiir1‘" 1% W515 i
54
A; aeoe- _.._.C‘arb(1m&)) ‘ma‘rw
s; ' " mam
,3 SEE wmdaow
g ....... wd4OW
9:3 4.509 “" de
_ _._..wmd1oow
25(95 \w’l Mil-1W 1. m1 5:13:31” 15“,«!~.‘,‘2 ‘11‘1‘1‘ “1111‘1‘l‘h‘1‘11‘
:mamwfm/ymlr‘ 414.4. 'rP-‘lgti ’ 1W1” 1’ ‘11“? H I “‘
' 5W Iri"
o.E+oo . ' ‘ ‘ 1 1 1
4995 5015 was m 5075 5095 5115 5135

W609”! (A)

Figure 5.15 OES scan of C2 in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 10-100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” has a flow rate of
160/54 sccm O2/CH4.

 

 

 

98

 

1.5001 ---~»7ow -———90w
9.5%
1 8.15419
7.E-(B<
6.5-(B
55(9-
4.E-(B‘
3519‘ ”WW“ EINVWYVWW‘ 111111 i11111*'1‘11I‘1 “113 15:511’1'1‘51‘151 1551; 111W"

2509- 1191111515111”- ‘ 1f”:
‘NMIT ;:y {11‘ iMWI :‘i11t1111111141'7n1‘1‘11‘ “111111111 1111 1.1-1'11

01“] [I
'1; L 44,“. s ‘-

o.E+oo fir . , . -_f_ gz *5 z_ .
4990 5010 5090 5050 5070 5090 5110 5130 51
WW

Figure 5. 16 Second set of data. OES scan of C2 In a combustion (160/80 sccm
O2/CH4) and hybrid (combustion plus 10- 80 W microwave power)
flame/discharge using the 0.4 mm-diameter brass nozzle.

 

Intensity (a.u.)

 

 

 

 

 

 

 

Emissions of 02, OH, and CH are also recorded in Figures 5.17 — 5.19. As
seen for N2 and C2, a general trend follows that as microwave power is added to

a stoichiometric flame, the emission intensity increases.

99

 

$507 bummedam
__wm'dww
~ an'dSOW
’5 5507 —-— andKDW
g: WbridSOW—fudlem

 

 

 

 

 

 

 

 

Figure 5.17 OES scan of O2 in a stoichiometric (160/80 sccm O2/CH4)
combustion flame and hybrid (combustion plus 40, 50. 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean" is a flow rate of
160/54 sccm O2/CH4. The last set of data is from a discharge of only 160 sccm
O2 and 50 W of microwave power (no fuel is present). Data from 30 W of
microwave power and below show no 02 signal, similar to the combustion—only
case, and therefore are excluded from the plot.

100

 

2507 —
Comb (16080)

"~de
2E—07~ ———I-ybn‘d1oow
—~-—-wbddwN—mellean

;::-_  M Hill”) W H

0.E+(X) -

Intensity (a.u. )

 

 

 

 

WWW

Figure 5.18 OES scan of OH in a stoichiometric (160/80 sccm 02/CH4)
combustion flame and hybrid (combustion plus 40, 50, 100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean” is a flow rate of
160/54 sccm 02/CH4. Data from 30 W of microwave power and below show no
OH signal, similar to the combustion-only case, and therefore are excluded from

 

 

 

 

  
    

  
 
 

 

theplot
- - -erb(1w&))
fiE-w —--—me
-4I-ymdaow
SE43} ' _~de
‘ "1 —l-Mxid50N
3‘ i ii i' q i ——— H/Utd1ww
d 4508] '1" 5 \ —|'M!id4OW—fuellem
v I“ ’lflfh " {I ---- deV-fifilai
' 3508+ ’- =""“ 'i E IW ll“: ~~~~~~ d1(DW—fuellea1
l I. i f in W“ M‘W‘W
_ 2543‘ .- ' 1 ; a- mum.“ kg?!“ W@
1
1.508
0.500 of, fl W7—
43!) 450 43!) 4:50 W
WW —

 

 

Figure 5.19 OES scan of CH in a stoichiometric (160/80 sccrn OleH4)
combustion flame and hybrid (combustion plus 10—100 W microwave power)
discharge using the 0.4 mm-diameter brass nozzle. “Fuel lean" is a flow rate of
160/54 sccm 02/CH4.

101

The peak intensities from each scan (N2, C2, 02, and CH) are plotted
against absorbed microwave power. Figures 5.20 - 5.23 contain data from
multiple sets of experiments, i.e. independent experiments, to show consistency
within data. Aside from a scattering of data points due to variations in each

independent experiment, the trend follows that intensity increases as microwave

power increases.

 

j

25“ x3et15toich
oSd1FudLm1

20, asazstoich <>
-Set2NoFueI x
ASdBStoioh

15~ +Set4Stoioh-Nerj

10~

N2Em'ssionlrMnsity(a.u.)x10-8

 

15 25 35 45 55 65 75 85 95 105
AbsorbendmwavePaweer)

 

 

 

Figure 5.20 Relative intensities of N2 emissions as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm, 160/54 sccm, and 160/0 sccm 02/CH4, and also a flow rate of 160/80/10
sccm 02/CH4/N2 with N2 injection. Intensities are recorded from four sets of
experiments to show consistency, and values are recorded from peak intensities,
at a wavelength of ~ 3769 A.

102

 

.s _s
N A

.5
0

Oz Em'ssion Intensity (a.u.) x10-9
O) on
D

 

 

A A 3
A 3 o A
4“ 0 ° 0 U 0861
OSetZ
2< A863
0 Y I 17 I V 1
0 Z) 40 a) a) 100 120

 

 

Absorbendrawavernrm

Figure 5.21 Relative intensities of Oz emission as a function of absorbed
microwave power, measured from a stoichiometric flame/discharge with a flow
rate of 160/80 sccm 02/CH4. Intensities are recorded from three sets of
experiments to show consistency, and values are recorded from peak intensities,
at a wavelength of ~ 5130 A.

 

 

oSd1Stdch
0861me

jg:

j

104

amunmns'ty (a.u.) x10-8
8 8

 

o o D
Y i T T l T l l

20 30 40 5o 60 7D 80 90 100
AbsorbendrowavePcmaer)

Figure 5.22 Relative intensities of 02 emission as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm 02/CH4 and 160/54 sccm 02/CH4. Intensities are recorded from one set of
experiments, and values are recorded fromfi‘ peak intensities, at a wavelength of ~
3340 .

 

 

 

 

103

 

i
l
|

CH Emission Intensity (a.u.) x10-8
w A
o a .
0
0
0
0
0

0 20 4o 60 80
AbsorbendmwavePowerM

 

oSd1Stdch
nSet1FudLea'1

1CD 120

 

 

Figure 5.23 Relative intensities of CH emission as a function of absorbed
microwave power, measured from flame/discharges with flow rates of 160/80
sccm O2/CH4 and 160/54 sccm O2/CH4. Intensities are recorded from one set of
experiments, and values are recorded from peak intensities, at a wavelength of ~

A wide-range scan including spectra of all species mentioned in this

section is shown in Figure 5.24. The plot provides a clear, overall summary of

how microwaves alter a combustion flame at a micro-scale level, not detectable

by the human eye. The spectra of N2 and C2 are used to calculate gas

temperatures which approximate flame/discharge temperatures, discussed in the

next section.

104

 

l 5.E-08-
—-Wb"'d(4°W) CH (4280479508)

55% q ---»- Cbntxstion

4.508 ~
4.E-08 l

 

3.E-08 J

 

3.5-08 J

Intensity (a.u.)

ZE-(B 4
2.5% -
1.E-08 ~
5.E-09 ~

 

 

 

O.E+m‘ l 7' I T
2500 3000 3500 4000 4500 5000 5500 6000 6500
VlhvalengtlflA)

Figure 5.24 OES scan of combustion flame (160/80 sccm O2/CH4) and hybrid
flame/discharge (addition of 40 W microwave power) for an entire range of visible
emission spectra. The wavelength and relative intensities are given at the peak
(A = 4280 A) for each set of data. Molecules and radicals are listed next to their
visible spectra. A 0.4 mm-diameter ceramic nozzle is used for this data.

 

 

 

5.5 Flame/discharge temperatures and profiles

Data for flame/discharge temperature are gathered with three different
diagnostic techniques using: 1) a thermocouple; 2) OES; and 3) an IR camera.
The experimental method for each technique is provided in Chapter 3.

As discussed previously, the thermocouple is capable of measuring a pure
combustion flame temperature only because the metal interferes with the
electrical field produced by microwaves. Figure 5.25 presents a temperature
profile through the width of a stoichiometric flame, using a type K thermocouple.
The photograph in the figure shows where the thermocouple collects data

through the flame. From a time-step iterative algorithm, the maximum flame

105

temperature calculated is 2253 K. Adiabatic flame temperature for
methane/oxygen flames, found in text [25], is approximately 3000 K. The
theoretical curve drawn in Figure 5.25 is not plotted with theoretical data; it is a
qualitative curve used for illustrative purposes. The 750 K difference in
temperatures is most likely due, in majority, to heat loss. The thermocouple is not
shielded, leading to radiation loss. Furthermore, the flame is not confined or
insulated; much heat is lost to the environment. It is also important to mention
that thermocouples are used to calculate point temperatures, whereas adiabatic

flame temperatures tabulated are average temperatures of the entire flame.

 

—E>perirrertdee
—TleoretiwCuve

 

Terrperature (K)

 

 

CD

4.0 -3.0 -20 -1.0 0.0 1.0 2.0 3.0 4.0
Wclh of Flam (mu)

 

 

 

Figure 5.25 Temperature profile of a stoichiometric CH4/O2 flame from a type K
thermocouple. Photograph of the flame shows where the thermocouple scans the
flame. The plot provides experimental data and also displays a qualitative
example of a theoretical temperature distribution with an adiabatic flame
temperature of 3000 K, taken as an average temperature.

106

Rotational temperatures of N2 and C2 are calculated for several sets of
experiments, i.e. independent experiments. Figure 5.26 plots N2 rotational
temperatures for three sets of stoichiometric flow rates. Note that N2 is added to
the gas flow upstream the nozzle. It is shown that temperature remains constant
for each stoichiometric flow rate, in agreement with theory. Possible outlier data
in the plot is found at 70 W microwave power; the temperature spread is 652 K.
Such a range may be caused by variations in the experimental setup, mainly
tuning the torch to find the sweet spot. In addition, as the volume of the flame
increases, so will the heat loss, causing a decrease in temperature. The
temperature increases as microwave power increases, ~ 1200 K from 30 W to 80

W.

 

NzRotationalTerrperatu'eM)
fié

 

 

2&1) j 0

A o A
2600 — 0 1666610

U
2400 — o A n 1065036
2200 4 D A 6630/2

A
m l T I T T T l
20 30 40 50 60 70 60 ed

 

 

AbsorbendmwavePowarMb

Figure 5.26 N2 rotational temperature of a flame/discharge as a function of
absorbed microwave power for data for three different flow rates: 160/80/10
sccm, 100/50/3.6 sccm, and 60/30/2 sccm O2/CH4/N2.

 

Figure 5.27 graphs rotational temperatures of both N2 and C2 gases. Four

independent experiments present data in this plot, the last set defined differently

107

because N2 is added to the gas flow. Data show a general increase in
temperature as microwave power increases. The greatest increase occurs for N2
rotational temperatures (Set 1), with a temperature increase of 1489 K from 20 W
to 80 W microwave power. The smallest increase occurs for C2 rotational
temperature (Set 1), with a temperature increase of 95 K from 0 W to 80 W
microwave power. Rotational temperatures of N2 and C2 in disagreement with
each other does not infer that the temperature is miscalculated. Further
investigation and research is required to know which rotational temperature, if
either, best represents the flame temperature. Raw data from Figure 5.27 is

listed in Appendix B.

 

WTerrperaue(
E 3
C;
..
Q‘DO
<> . mu-
0 ><+ D:
0 J +
0
0
..

 

 

 

 

200m
. 0
1500~ oNZSd1
OQSd1
1000 .msa2
0 .Qsa2
I: 5001 xmsas
0C2863
g +Nery'ectionSet4
o T r . . .
0 20 4O 60 8) 1(1)
i AbsorbendmwavePowerM

 

Figure 5.27 N2 and C2 rotational temperatures of flame/discharges as a function
of absorbed microwave power, for four sets of data (independent experiments).
The flow rate for the first three sets is 160/80 sccm O2/CH4, and 160/80/10
O2/CH4/N2 for the fourth set. Tables of raw data are found in Appendix B.

108

Flame and hybrid temperature greatly exceed those of pure plasma
temperatures. Plasma temperatures measure ~ 1700 K (measured from N2 gas
temperatures) while combustion temperatures, measured by the same technique
are close to double that.

5.5.1 Data collected from infrared (IR) camera

Although quantitative temperatures cannot be measured with the IR
camera at this time (another student’s ongoing Ph.D. research), qualitative
deductions are made. The images shown in Figure 5.28 are temperature profiles,
created by the IR camera, of a stoichiometric flame as microwave power
increases from O W to 40 W to 60 W. The emissivity is unity for all the results
shown.

The flame analyzed in these images is not confined within the microwave
cavity because the IR camera cannot bypass the metal screen in the cavity’s
windows. Therefore, the torch is not optimally tuned due to microwave leakage.
Results do show, however, that the microwaves do indeed alter the flame. The
center of the flame, the white temperature profile in Figure 5.28, is the hottest
part of the flame, relative to all the profiles shown. As microwave power is added,
this center section expands while the outermost sections contract. This behavior
suggests that microwaves create a more intense flame/discharge by decreasing
volume and increasing temperature. Figure 5.29 is a sketch of the first two IR

images in Figure 5.28 against each other, to clearly show the change in volume.

109

  

Fahrenheit
420.0 i
410.0
400.0
390.0
380.0
370.0
360.0
350.0
340.0
330.0
320.0
310.0
300.0
290.0

280.0 I

(2:123 [54.0

 

Figure 5.28 Temperature profile images from the IR camera of a flame as it is
modified by microwave power, 40 W and 60 W. The flame is not confined within
the cavity. The temperature scale is provided, but does not provide an accurate

temperature reading. The gas flow rate is 200/100 sccm O2/CH4.

 

Figure 5.29 Temperature profiles sketched from the IR camera images (from
Figure 5.28) for both a combustion flame (represented by the solid lines) and a
hybrid flame/discharge with 40 W microwave power added (dashed lines).

110

IR images of a pure argon plasma are shown in Figure 5.30. Each image
is recorded with a different filter, so as to highlight various sections of the plasma.
For example, the bottom-right image highlights the heat produced downstream
the plasma discharge, while the top-left image highlights the temperature profiles

within the intense plasma itself.

 

Figure 5.30 Temperature profile images from the IR camera for a plasma-only

discharge: 200 sccm argon excited by 40 W microwave power. Each image is

recorded with a different filter. The image in the bottom right corner shows the
temperature profiles of the hot gases beyond the plasma.

Figure 5.31 shows IR images of a flame produced by a ceramic nozzle. An
interesting observation is the temperature profiles of the ceramic. The center of
the nozzle, where the flame sits, appears much hotter than the outer edges.
Further study of ceramic temperature profiles should be carried out. This IR
camera is capable of accurately measuring temperatures of solid surfaces, and it

is also capable of recording time-dependent temperatures.

111

Time-dependent temperature measurements of a gold-sputtered ceramic
nozzle are made with the IR camera. Unfortunately visual results of this
experiment are not reported in this thesis because the only current recorded
format is an avi file (video file). In summary, the gold-sputtered ceramic
generates higher temperature profiles than the ceramic with no gold coating and

does not cool down as fast as the ceramic with no coating.

 

Figure 5.31 Temperature profile images from the IR camera for a stoichiometric

combustion flame formed by a ceramic nozzle. Different camera filters are used

to highlight different sections of the flame. The temperature profiles of the nozzle
are also illustrated.

112

Chapter 6

Summary and Recommendations
6.1 Summary of results

An extensive experimental study of microwave plasma-enhanced
combustion concludes that microwaves do in fact alter a combustion flame.
Physical alterations of the flame are made in color and geometry, as color
changes from a blue flame to a white/yellow hybrid discharge with the addition of
microwaves, and flame lengths and diameters grow in size. Alterations at the
micro-scale level are also made in flame species. Molecules and radicals
detected in the flame/discharge become more intense, or more accelerated, by
the presence of microwaves. In addition, temperatures increase with the addition
of microwaves, as hypothesized.

One question that remains asks what exactly is desired by adding
microwaves to a flame. Is it desirable to create a completely transformed flame
that changes in color and increases in size, thus raising temperature but lowing
power density; or is it more desirable to add just enough microwave power (10 —
30 W) so the power density rises exponentially without physically changing the
appearance of the flame. The trade-off is temperature. More microwave power
yields higher flame temperatures.

These questions may be more easily answered when considering the
applications of such a hybrid technology. Industry may be interested in a small—
scale, easily handled, clean torch used for welding, surface treatment, or

localized heating. In these cases, a higher-temperature flame/discharge is

113

desired. When considering cutting comparable to laser technology, geometry is a
major factor considered. The more concentrated and more intense the discharge
is, the more precise the cut will be. Geometry also becomes a critical factor when
considering medical applications, such as surgery.

Other applications include engines, both jet engines and automotive
engines. Results in this thesis show that, with the presence of microwaves, a
flame is maintained with lower fuel consumption than would be possible without
microwaves. Fuel-efficient engines are driving the automotive industry today.
Both higher temperatures and larger geometries may be desired for this
application.

Motivation behind this research includes improving the combustion
process by creating more efficient combustion with the addition of microwaves.
Results support the hypothesis that microwaves do improve the combustion
process. Flame stability and fuel-efficient combustion are enhanced in this hybrid
system. It is yet to be shown in this research if cleaner combustion results from
the addition of microwaves. Researchers at the Los Alamos National Laboratory
have shown that a slightly different plasma-enhanced hybrid system increases
combustion efficiency by reducing the amount of unburned hydrocarbon
(propane) emissions and also reduces toxic NOx/SOx emissions [1].

It appears that the presence of microwaves accelerates the combustion
reaction, reflected in flame velocities calculated from measured flame angles.
Flame speed more than doubles with the addition of microwaves when the

physical transformation takes place.

114

6.2 Recommendations

Further investigation and study of the ceramic nozzle is encouraged. In
terms of manufacturing, an improved method for attaching the ceramic nozzle to
the brass inner conductor should be sought. Currently the nozzle is glued with
epoxy to the inner conductor. This becomes a problem when microwaves are
present. The electrical energy is drawn to this epoxy material, thus retracting
from the flame itself. Additional temperature measurements made with the IR
camera could provide insight as to whether the ceramic is or is not absorbing
microwaves.

An improved design of the microwave cavity would attract the torch to
industry. The cavity is big and bulky. Designing a much smaller and compact
cavity to allow an intense formation of an electric field is critical. Temperature
measurements of the flame/discharge inside the cavity using the IR camera
would also provide insight for what is happening to temperature as microwaves
greatly transform the flame.

If time and funds allow, more detailed diagnostic measurements should be
made including cavity ring-down spectroscopy (CRDS) and laser-induced
fluorescence (LIF) to determine the influence of plasma enhancement on the
distributions of CH, C2 and OH. OES used for the research in this thesis provides
profitable results in terms of qualitative analysis. For quantitative analysis,
however, the suggested CRDS and LIF techniques are preferred.

CRDS is an absorption technique where pulsed laser light is coupled

into an optical cavity formed by highly reflecting mirrors. It is a sensitive

115

technique because the optical path length is made long by the highly reflecting
mirrors, and it is convenient because the quantity being measured is the decay
time of the light intensity, not the absolute intensity of the light as needed for
direct absorption spectroscopy measurements. CRDS works at both atmospheric
pressure and low pressures (5 — 30 Torr), suggested working conditions for
future experimental work. LIF is a high-spatiaI-resolution technique but can be
more difficult to calibrate and work with.

One final suggestion considers a numerical study. A computational
simulation can be created by factoring into a full flame chemistry simulation the
electronic or plasma reaction component. Full flame chemistry simulations are
developed and available as commercial software. The suggested hybrid
simulation has yet to be fully developed. Such a simulation would provide
validation for experimental results and provide a deeper understanding of the

research.

116

APPENDICES

117

Appendix A
Table A1 Values for therrnocou le temperature measurements.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Width of Measured Measured
Time Flame T(t) PfGViOUS psolld pr|d@2oc [.1 [1.
Step (mm) 462g) T (degC) 459/1113) (JIkg-K) m-s/m’) (N-s/m’)
1 -3.158 24.4 8600 523 2.00E-05 2.00E-05
2 -2.842 26.5 24.4 8600 523 2.02E-05 2.02E-05
3 -2.526 28 26.5 8600 523 2225-05 2.03E-05
4 -2.211 31.1 28 8600 523 2.16E-05 2.05E-05
5 -1.895 51.6 31.1 8600 523 2.33E-05 2.18E-05
6 -1.579 101.2 51.6 8600 523 3.48E-05 2.45E-05
7 -1.263 167.4 101.2 8600 523 4.64E-05 2.75E-05
8 -0.947 274.2 167.4 8600 523 5.26E-05 3.23E-05
9 -0.316 382.9 274.2 8600 523 6.70E-05 3.62E-05
10 0.000 473.8 382.9 8600 523 6.62E-05 3.91 E-05
1 1 0.316 615.7 473.8 8600 523 6.46E-05 4.37E-05
13 0.947 1297.6 1285.9 8600 523 2.18E-04 6.40E-05
14 1.263 1196.2 1297.6 8600 523 6.51 E-05 6.11E-05
15 1.579 1096.2 1196.2 8600 523 2.81 E05 5.83E-05
16 1.895 985.2 1096.2 8600 523 3.93E-05 5.51 E-05
17 2.211 893.2 985.2 8600 523 4.03E-05 5.24E-05
18 2.526 818.5 893.2 8600 523 3.43E-05 5.03E-05
19 2.842 753.3 818.5 8600 523 2.93E-05 4.82E-05
20 3.158 712.5 753.3 8600 523 2.97E-05 4.68E-05
Calculat Calculat
ed T- ed T-
knmdm h flame flame
pip, (WIm-K) Pr mm’ls) Re (WlmzK) (degC) (K)
375.15
1.000 0.02532 0.74036 1.61 E-05 5 170.31 27.00 300.00
1.002 0.02553 0.74000 1.74E-05 348.54 166.14 58.25 331.25
1.094 0.02807 0.73563 1.98E-05 306.07 175.16 49.47 322.47
1.056 0.02736 0.73685 1.91 E-05 316.92 172.25 76.25 349.25
1.071 0.02954 0.73311 2.12E-05 285.99 177.89 340.37 613.37
1.423 0.04968 0.70553 5.65E-05 107.1 1 207.29 697.37 970.37
1.688 0.06884 0.72788 1.21 E-04 50.01 225.12 897.52 1170.52
1.626 0.08008 0.73307 1.65E-04 36.62 234.90 1400.86 1673.86
1.852; 0.11066 0.71217 3.14E-04 19.29 270.49 1371.83 1644.83
1.691 0.10889 0.71338 3.05E-04 19.83 265.17 1318.27 1591.27
1.478 0.10564 0.71560 2.89E-04 20.91 256.70 1979.69 2252.69
1.312 0.14582 0.68814 4.85E-04 12.49 303.49 6685.03 6958.03
3.411 0.43163 0.49279 1.87E-03 3.23 716.22 1334.39 1607.39
1.065 0.10662 0.71493 2.94E-04 20.57 247.25 182.38 455.38
0.483 0.03806 0.71882 3.45E-05 175.70 159.48 480.65 753.65
0.714 0.05721 0.71431 8.19E-05 73.94 185.62 511.00 784.00
0.769 0.05884 0.71621 8.74E-05 69.30 189.02 323.78 596.78
0.682 0.04879 0.70449 5.35E-05 113.11 182.42 206.73 479.73
0.607 0.04002 0.71555 3.75E-05 161.38 168.80 216.29 489.29
0.634 0.04079 0.71426 3.87E-05 156.38 171.15 114.35 387.35

 

118

 

Table B.1 Values used for flame extinction plot in Chapter 5.

Appendix B

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Combustion Only (0.4 mm
brass) Combustion+Microwaves
Blowout 10 W 20-100 W
V02 (sccm) Vm‘ (sccm) ch (sccmL fiflsccm) °/o Fuel Reduction
180 72 64 0 11
160 56 53 0 5
140 44 41 0 7
120 35 32 0 9
100 26 23 0 12
80 19 19 0 0
Combustion Only (brass,
0.4) Combustion-HOW Combustion+20-100W
Equivalence Equivalence VTOT Equivalence th
Ratio vim (sccm) Ratio (sccm) Ratio (sccm)
0.80 252 0.71 244 0.00 180
0.70 216 0.66 213 0.00 160
0.63 184 0.59 181 0.00 140
0.58 155 0.53 152 0.00 120
0.52 126 0.46 123 0.00 100
0.48 99 0.48 99 0.00 80
Table B.2 Values found in temperature plots in Ctflter 5.
Methane/0x1 en Mixture 160l80 sccm), Set 1
Microwave
1’0“" (W) o 10 20 30 4o 50 so 70 so
Rotational
Temperature
(K) N2 N/A N/A 1717 2682 2368 2223 2609 2991 3206
Rotational
Temperature
(KLCZ 3124 2994 2935 2441 3111 3049 3057 3350 3318
Methane/Oxygen Mixture (160l80 sccm), Set 2
Microwave
PW" (W1 0 1o 20 30 4o 50 100
Rotational
Temperature
(K) N2 N/A N/A N/A N/A 2524 2934 3612
Rotational
Temperature
(K) C2 3016 3115 2516 2584 2687 2886 3254

 

 

 

 

 

 

 

 

 

 

119

 

MethaneIOxygen Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

(160l80 sccmL Set 3
Microwave
Powerlwl o so 40 so
Rotational
Temperature
(K) N2 N/A N/A N/A 2589
Rotational
Temperature
(K) CZ 2637 2352 2738 2849
Methane/Oxygen Mixture (160l80/10 sccm), Set 4
Microwave
Poweriw) 30 40 so so 70 so
Rotational
Temperature
4KLN2 2368 2863 2674 3258 3329 3542

 

 

 

 

 

 

 

120

 

BIBLIOGRAPHY

[1] L. A. Rosocha, Y. Kim, S. Stange, V. Ferreri, D. M. Coates, and D. Platts,
“Plasma-enhanced combustion of propane using a silent discharge,” Plasma
Physics Research Highlights, Los Alamos National Laboratory.

[2] N. Chintala, R. Meyer, A. Hicks, B. Bystricky, J. W. Rich, W. R. Lempert, and
I. V. Adamovich, “Non-thermal ignition of premixed hydrocarbon-air and CO-air
flows by nonequilibrium RF plasma,” 42”” AIAA Aerospace Sciences Meeting and
Exhibit, Reno, NV, 2004.

[3] K. W. Hemawan, C. L. Romel, S. Zuo, I. S. Wichman, T. A. Grotjohn, J.
Asmussen, “Microwave plasma-assisted premixed flame combustion,”
International Workshop on Micro-plasma, Greifswald, Germany, 2005.

[4] S. Whitehair, L. L. Frasch, and J. Asmussen, “Experimental performance of a
microwave electrothermal thruster with high temperature nozzle materials,” 19’"
AIAA/DGLR/JSASS International Electric Propulsion Conference, Colorado
Springs, CO, 1987.

[5] A. Yu. Starikovskii, “Plasma supported combustion,” 30th International
Symposium on Combustion, Chicago, IL, 2004.

[6] S. M. Starikovskaia, I. N. Kosarev, A. V. Krasnochub, E. I. Mintoussov, A. Yu.
Starikovkii, “Control of combustion and ignition of hydrocarbon-containing
mixtures by nanosecond pulsed discharges,” 43m AIAA Aerospace Sciences
Meeting and Exhibit, Reno, NV, 2005.

[7] J. W. Dold, R. W. Thatcher, A. Omon-Arancibia, and J. Redman, “From One-
Step to Chain-Branching Premixed Flame Asymptotics,” Combustion Institute,
29:1519-1526 (2002).

[8] N. Peters, Turbulent Combustion, Cambridge University Press, 2000.

[9] A. Linan and F. A. Williams, Fundamental Aspects of Combustion, Oxford
University Press, New York, 1993.

[10] J. Buckmaster, Annu. Rev. Fluid Mech. 25:21 -53 (1993).
[11] P. Pelce, Dynamics of Curved Fronts, Academic Press, London, 1988.
[12] G. l. Sivashinsky, Annu. Rev. Fluid Mech. 15:179-199 (1983).

[13] F. A. Williams, Combustion Theory, Benjamin/Cummings, San Francisco,
1985.

121

[14] J. D. Buckmaster and G. S. S. Ludford, Theaty of Laminar Flames,
Cambridge University Press, 1982.

[15] Ya. B. Zeldovich, G. I. Barrenblatt, V. B. Librovich, and G. M. Makhviladze,
The Mathematical Theory of Combustion and Explosions, Consultants Bureau,
New York, 1985.

[16] I. K. Puri, Environmental Implications of Combustion Processes, CRC Press,
Boca Raton, Florida, 1993, Chapter 5 by I. S. Wichman.

[17] S. Zuo, K. Hemawan, J. J. Narendra, T. A. Grotjohn, and J. Asmussen,
“Miniature microwave plasma torch applicator and its characteristics,” IEEE
Conference on Plasma Science, 5A3, 2004.

[18] K. W. Hemawan, Numerical A_nalvsis and Experimental Measurements of
Material Loading in Cvlindricgl Microwave Cavitv Applicators, MS. Thesis,
Michigan State University, 2003.

[19] K. W. Hemawan, S. Zuo, C.L. Romel, T. A. Grotjohn, I. S. Wichman, E.
Case, and J. Asmussen, “Exploring microwave plasma-assisted combustion,”
International Conference on Plasma Systems (ICOPS), Monterey, CA, 2005.

[20] J. B. Wachtman, Jr., Mechanical Prcyerties of Ceramics, John Wiley and
Sons, Inc., New York, 1996.

 

[21] J. Zhang, L. Liu, T. Ma, and X. Deng, “Rotational temperature of nitrogen
glow discharge obtained by optical emission spectroscopy,” Spectrochism. Acta
A, 58, 1915-1922, 2002.

[22] J. J. Narendra, Characteristics and Modeling of Miniature Microwave Plasma
Discharge Created with Microstrioline Technology, MS. Thesis, Michigan State
University, 2004.

[23] E. Stoffels, A. J. Flikweert, W. W. Stoffels, And G. M. W. Kroesen, “Plasma
needle: a non-destructive atmospheric plasma source for fine surface treatment
of (bio) materials,” Plasma Sources Sci. Technol., 11, 383-388, 2002.

[24] W. Huang, Microwave Plasma Assisted Chemical Vapor Deposition of Ultra-
nanocrfitalline Diamond Films, Ph.D. Dissertation, Michigan State University,
2004.

[25] F. P. lncropera and D. P. Dewitt, Introduction to Heat Transfer, John Wiley
and Sons, New York, 2002.

122

[26] J. L. Short, A New Svstem Developed to Characterize Thermoelectric
Devices for Power Generation Applications, MS. Thesis, Michigan State
University, 2005.

[27] G. Herzberg, Molecular Spectra and Molecula_r Structure. I. Spectra of
Diatomic Molecules, D. Van Nostrand Co., Inc., Princeton, 1950.

[28] S. N. Suchard and J. E. Melzer, Spectroscopic Data 2: Homonuclear
Diatomic Molecules, lFl/Plenum, New York, 1976.

[29] G. L. King, Temperature and Concentration of Ionic and Neutral Species in
Resonant Microwave Cavitv Plasma Discharges, Ph.D. Dissertation, Michigan

State University, 1994.

[30] B. C. Wadell, Transmission LinejDesign Handbook, Artech House, Inc.,
NonNood, 1991.

[31] S. R. Turns, An Introduction to Combustion, The McGraw-Hill Companies,
Inc., Boston, 2000.

123

   

|111:1][Elllljflljlfllljflf|