close

Вход

Забыли?

вход по аккаунту

?

Characterization and gas temperature measurements of a waveguide-based microwave plasma torch

код для вставкиСкачать
The Pennsylvania State University
The Graduate School
CHARACTERIZATION AND GAS TEMPERATURE MEASUREMENTS
OF A WAVEGUIDE-BASED MICROWAVE PLASMA TORCH
A Dissertation in
Aerospace Engineering
by
Peter J. Hammond
c 2013 Peter J. Hammond
Submitted in Partial Fulfillment
of the Requirements
for the Degree of
Doctor of Philosophy
August 2013
UMI Number: 3576533
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI 3576533
Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
The dissertation of Peter J. Hammond was reviewed and approved∗ by the following:
Michael M. Micci
Professor of Aerospace Engineering
Dissertation Advisor, Chair of Committee
Sven G. Bilén
Associate Professor of Engineering Design,
Electrical Engineering, and Aerospace Engineering
Deborah A. Levin
Professor of Aerospace Engineering
Victor P. Pasko
Professor of Electrical Engineering
George A. Lesieutre
Professor of Aerospace Engineering
Head of Department of Aerospace Engineering
∗
Signatures are on file in the Graduate School.
Abstract
Research to characterize a microwave plasma torch was initiated at Penn State University. Microwave power input into the device initiates and sustains plasma in an
argon gas jet issuing from a copper nozzle into the ambient atmosphere. Protruding through a rectangular waveguide, the nozzle acts to enhance the local electric
field when microwaves are excited in the waveguide. The plasma resembles a
small flame, approximately 2–4 cm in length and less than 1 cm in total diameter. The primary research interests which have driven experimental design and
characterization of the torch include (1) increasing plasma jet control via improved
impedance matching; (2) reducing the erosion of the nozzle tips; and (3) determining the viability of applying the Penn State Microwave Plasma Torch (PSMPT) to
the cutting and melting of materials via gas temperature measurements. Literature on the similar microwave torches—particularly, those of the single-electrode
plasma (SEP) type—was reviewed.
Several design issues were encountered during early testing with the torch.
Impedance matching and nozzle erosion presented the most significant obstacles.
Poor impedance matching was overcome most effectively with an automatic tuner
that could determine a match quickly. Nozzle erosion is not often addressed in the
literature on SEPs. However, significant erosion was a limiting factor in early tests
with the torch. More recent testing reveals that erosion can be mitigated by addition
of a secondary flow of argon around the primary nozzle gas flow. Gas temperature
in the plasma was deduced via OH rotational temperature measurements. Molecular nitrogen spectral interference with the OH spectra required fitting both the
OH and N2 second positive system in the region of 305–318 nm. The results of this
testing indicate an OH rotational temperature—and assumed gas temperature—
between 2700–3400 K. These results indicate that the torch should prove useful
in cutting and heat-treatment applications for some materials. Recommended areas of future study include examining the plasma for possible filamentation and
enhanced spectroscopic diagnostics.
iii
Table of Contents
List of Figures
vii
List of Tables
ix
Acknowledgments
x
Chapter 1
1
1.1
1.2
1.3
Description and Motivation . . . . . . . . . . . . . . . . . . . . . . . . 1
Classification of Plasma Devices . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Atmospheric Pressure Plasmas . . . . . . . . . . . . . . . . . . 3
1.2.2 Microwave Plasmas: General Classification . . . . . . . . . . . 4
1.2.3 Microwave Plasma Torches . . . . . . . . . . . . . . . . . . . . 5
1.2.3.1 Electrode Type Torches . . . . . . . . . . . . . . . . . 6
Development of Microwave Single-Electrode Plasmas at Atmospheric
Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.1 Early Work: Prior to Mid-1960s . . . . . . . . . . . . . . . . . . 8
1.3.2 SEPs as Spectrochemical Sources: Mid-1960s to 1990s . . . . . 10
1.3.3 Contemporary Research: 1990s to Present . . . . . . . . . . . . 11
Chapter 2
2.1
2.2
2.3
Introduction
Theory
Microwave Transmission . . . . . . . . . . . . . . .
2.1.1 Electromagnetic Wave Propagation . . . . .
2.1.2 Impedance Matching . . . . . . . . . . . . .
Microwave Plasmas . . . . . . . . . . . . . . . . . .
2.2.1 Breakdown and Plasma Initiation . . . . . .
2.2.2 Microwave Plasma Maintenance Processes
2.2.3 Surface-Wave Plasmas . . . . . . . . . . . .
Optical Emission Spectroscopy of Plasmas . . . . .
18
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
18
18
23
26
26
28
29
30
2.3.1
2.3.2
2.3.3
2.3.4
Chapter 3
3.1
3.2
3.3
4.2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Methods
Experimental Apparatus . . . . . . . . . . . . . . . . . . .
3.1.1 PSMPT-Waveguide Assembly . . . . . . . . . . . .
3.1.2 Microwave Transmission System . . . . . . . . . .
3.1.3 Gas Supply System . . . . . . . . . . . . . . . . . .
3.1.4 Optical Spectroscopy System . . . . . . . . . . . .
Experimental Procedures . . . . . . . . . . . . . . . . . . .
3.2.1 Nozzle Preparation . . . . . . . . . . . . . . . . . .
3.2.2 System Assembly . . . . . . . . . . . . . . . . . . .
3.2.3 Plasma Initiation and Data Collection . . . . . . .
Computational Procedures . . . . . . . . . . . . . . . . . .
3.3.1 Spectral Simulation and Fitting . . . . . . . . . . .
3.3.1.1 OH Without N2 SPS Overlap Correction
3.3.1.2 OH with N2 Band Overlap Correction . .
3.3.2 Electromagetic Modeling . . . . . . . . . . . . . . .
Chapter 4
4.1
Line Positions in Diatomic Rovibronic Spectra
Line Intensities in Diatomic Rovibronic Spectra
2.3.2.1 Level Populations . . . . . . . . . . .
2.3.2.2 Transition Probabilities . . . . . . . .
Line Shapes . . . . . . . . . . . . . . . . . . . .
Relative Intensity Spectral Profile . . . . . . . .
31
32
32
34
34
35
36
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Results and Discussion
Preliminary and Qualitative Results: Experimental and Operational
Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Impedance Matching and Power Coupling . . . . . . . . . . .
4.1.1.1 Sliding Short Position . . . . . . . . . . . . . . . . . .
4.1.1.2 Automatic Tuning . . . . . . . . . . . . . . . . . . . .
4.1.1.3 Nozzle Tip Position . . . . . . . . . . . . . . . . . . .
4.1.1.4 Power Losses . . . . . . . . . . . . . . . . . . . . . . .
4.1.1.5 Microwave Breakdown Control . . . . . . . . . . . .
4.1.2 Nozzle Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.2.1 Description of Problem . . . . . . . . . . . . . . . . .
4.1.2.2 Remedies for Erosion . . . . . . . . . . . . . . . . . .
Spectroscopic Determination of Gas Temperature . . . . . . . . . . . .
4.2.1 Rotational Temperature Without N2 Band Overlap Correction
4.2.2 Rotational Temperature With N2 Band Overlap Correction . .
36
36
38
40
40
41
41
42
43
44
44
45
46
50
51
51
51
51
57
57
58
61
62
62
64
66
66
66
v
4.2.3
Chapter 5
5.1
5.2
Analysis . . . . . . . . . . . . . . . . . . . . . . . .
4.2.3.1 Spectroscopic Limitations . . . . . . . . .
4.2.3.2 Species Concentration . . . . . . . . . . .
4.2.3.3 Rotational and Vibrational Temperatures
Conclusions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
73
73
74
75
82
Summary and Contributions . . . . . . . . . . . . . . . . . . . . . . . . 82
Recommendations for Future Research . . . . . . . . . . . . . . . . . . 85
References
89
vi
List of Figures
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Pressure dependence of electron and gas temperatures . .
Microwave plasma torch types . . . . . . . . . . . . . . . .
SEP torch types . . . . . . . . . . . . . . . . . . . . . . . . .
Coaxial electronic torch design by Cobine and Wilbur . . .
Coaxial torch designs by Swift . . . . . . . . . . . . . . . . .
Murayama and Jecht and Kessler microwave torch designs
TIA and TIAGO torch designs . . . . . . . . . . . . . . . . .
MPT and MPJ designs . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4
6
7
9
10
11
13
14
2.1
2.2
2.3
2.4
2.5
Rectangular waveguide cross-section dimensions . . .
WR–284 midline field strength . . . . . . . . . . . . . .
Terminated lossless transmission line . . . . . . . . . .
Microwave Paschen curve for argon . . . . . . . . . .
Axial variations along a surface-wave plasma column
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19
22
24
27
30
3.1
3.2
3.3
3.4
PSMPT assembly schematic . . .
Experimental assembly schematic
Optics positioning . . . . . . . . .
Simulated N2 SPS bands . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
37
42
44
48
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
Coupling efficiency vs. normalized sliding short position . . . . . . .
Nozzle tip E-field vs. sliding short position . . . . . . . . . . . . . . .
Full-height electric field strength contours . . . . . . . . . . . . . . . .
3/4-height electric field strength contours . . . . . . . . . . . . . . . .
Half-height electric field strength contours . . . . . . . . . . . . . . . .
Plasma jets with and without automatic tuning . . . . . . . . . . . . .
Coupling efficiency vs. nozzle tip position . . . . . . . . . . . . . . . .
Microwave shielding cage around plasma . . . . . . . . . . . . . . . .
Measured radiated power . . . . . . . . . . . . . . . . . . . . . . . . .
Radiated power vs. sliding short position . . . . . . . . . . . . . . . .
Nozzle erosion sample images . . . . . . . . . . . . . . . . . . . . . . .
OH Trot contours vs. NFR and SFR for various powers and positions
Experimental spectra, 290–390 nm . . . . . . . . . . . . . . . . . . . .
Simulated spectra fitted to experimental . . . . . . . . . . . . . . . . .
52
53
54
55
56
58
59
60
61
61
63
67
69
71
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4.15
4.16
4.17
4.18
4.19
4.20
OH and N2 SPS relative band head intensities vs. position and NFR
Trot vs. position, at various power levels, NFR=0.7 SLM . . . . . . .
OH & N2 rot. temperatures, pos.=11.1 mm, fwd. power=300 W . .
OH & N2 rot. temperatures, pos.=11.1 mm, fwd. power=400 W . .
OH & N2 Vib. temperatures, pos.=11.1 mm, fwd. power=300 W . .
OH & N2 Vib. temperatures, pos.=11.1 mm, fwd. power=400 W . .
.
.
.
.
.
.
75
76
78
79
80
81
5.1
MPT torch filamentation . . . . . . . . . . . . . . . . . . . . . . . . . . 86
viii
List of Tables
1.1
Microwave SEP torches . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1
2.2
WR–284 TE-mode cutoff frequencies . . . . . . . . . . . . . . . . . . . 20
WR–284 TE10 frequencies and wavelengths . . . . . . . . . . . . . . . 21
4.1
Average OH and N2 SPS rotational and vibrational temperatures. . . 72
Acknowledgments
I would like to acknowledge the following people at Penn State:
• my advisor throughout my graduate school experience, Dr. Michael Micci,
for his willingness to take me on as his student and all the opportunities he
provided for me, both in the lab and as a teaching assistant;
• Dr. Sven Bilén, for sitting on my committee, and his support and guidance in
many areas, especially electronic and hyphenating;
• Dr. Deborah Levin and Dr. Victor Pasko, who graciously agreed to sit on my
committee;
• all those in the Department of Aerospace Engineering who made my studies
possible, especially Dr. Lesieutre and the administrative staff;
• Dr. Melton and Dr. Spencer for taking me on as a teaching assistant;
• all my lab colleagues over the course of my Ph.D. studies—most especially Dr.
Silvio Chianese and Dr. Dan Clemens, for their mentor-ship and support—as
well Carl, Tom, Karl, Jake, Jeff, and Erica for their assistance;
• Mr. Doug Smith for his glass craft, cutting and forming quartz tubes for me;
and
• Mr. Bob Dillon for the countless design iterations he assisted me with and
machine hours he put in.
Dedication
Many family and friends deserve credit for bearing with me during my pursuit for
a Ph.D., and to each and every one I remain humbly grateful.
I fondly recall the many friends in State College who bore with me through the
early years of my candidacy, especially the fraternal encouragement and support
of Fr. Boniface Hicks.
It’s difficult to imagine how quickly my time in Ireland has gone by, but I am
truly grateful for the advice and support of the members of Castleville and Equipes
Notre–Dame.
A great deal of love and support and patience from my parents, Wayne and
Theresa, and extended family cannot ever be repaid.
My daughters, Clara and Cecilia, may have kept me from finishing sooner than
I would have liked, but I wouldn’t replace them with a million doctorates.
And to my wife, Michelle, whose wisdom, perseverance, and love brought this
work to its fruition: thank you and I love you.
Dedicated to the memory of
Roy David Nelson
(May 8, 1956–August 26, 2012),
my favorite “rocket scientist.”
Chapter 1
Introduction
1.1
Description and Motivation
Microwave plasma torches have been studied for several decades, in part to develop
a technology to cut and melt various materials. Plasma heat-treatment applications,
such as cutting and melting, require—as a first estimate—knowledge of the gas
temperature of the plasma. The purpose of this research program was to obtain
spectroscopic measurements of the gas temperature on such a torch at Penn State—
hereafter referred to as the Penn State Microwave Plasma Torch (PSMPT). Varying
ranges of temperatures are observed within each of the microwave plasma torch
devices mentioned in the literature (see Table 1.1), providing a difficult comparison
for our own work. A deeper understanding of the gas temperatures produced
in the PSMPT will help to shed insight into whether it is viable when compared
with similar torch-like devices, especially for heat-treatment applications. Such
comparisons may reveal a more appropriate course of action for future design of
or operational changes to the PSMPT.
Determining the viability of applying the PSMPT to the cutting and melting
of materials via gas temperature measurements was the primary goal of this research effort. Other research interests that have driven experimental design and
characterization of the PSMPT include (1) better impedance matching to initiate
and sustain a plasma jet for controlled parameters (flow rates, power levels, frequency, geometry, etc.) and (2) reducing the erosion of the nozzle tips through a
secondary assist gas that appears to reduce oxidation, or via nozzles with more
efficient heat transfer capabilities at the tips. Early testing on the PSMPT focused
on analyzing various device setups to determine their suitability for creating a
plasma jet. Adjustments to the size, geometry, and materials of the PSMPT device,
to the microwave components, and to the gas flow all had various effects on both
the plasma and the device. The most recent set of tests on the PSMPT initiate and
sustain plasma reliably without striking aids and with negligible erosion (relative
to earlier tests) of the nozzle tip. Much of the early testing sought only to obtain
an operational torch, not yet optimized for any specific application. So as to gain
further insight on the PSMPT, gas temperature measurements were obtained from
the emission spectra of the hydroxyl radical, OH. Analysis of these results and
issues surrounding the PSMPT are provided in Chapters 4 and 5.
Microwave energy deposited into the PSMPT initiates and sustains plasma in
an argon gas jet flowing from a nozzle into the ambient atmosphere. Protruding
through a waveguide, the nozzle acts to enhance the local electric field when
microwave power is input in the waveguide. The plasma resembles a small flame,
approximately 2–4 cm in length and less than 1 cm in total diameter. Further
descriptions of the device are found in Chapter 3. Reviewed in this dissertation
are the underlying theory aiding the current design and testing status, the current
results of testing and simulation, and a proposal for further study and analysis of
the PSMPT. In this chapter, an attempt will be made to place the PSMPT within
the context of low-temperature plasmas—in particular, similar microwave plasma
torches. Identifying similar torch designs (and the physical processes determined
to be responsible for the behavior of such torches) may indicate how to better
design the PSMPT in the future. This is seen as especially important in regards to
• understanding power absorption channels in the plasma;
• controlling the plasma jet structure, size, and stability;
• understanding thermal transfer to the nozzle tip; and
• examining gas heating within the plasma in order to verify its use in differ-
ent materials processing applications, particularly high-temperature applications.
1.2
Classification of Plasma Devices
Plasmas—especially those found in laboratory settings—are often referred to as
gas discharges. These discharges have been put to use throughout a wide body of
disciplines and applications. Plasmas have the ability to create multiple reactive
species, at different pressures, in varying environments and configurations, and
2
with a broad range of temperatures. Most important among the parameters controlling the applicability of a particular plasma device are (1) the pressure of the
environment in which the plasma is created and sustained and (2) the operating
frequency of the power source.
1.2.1
Atmospheric Pressure Plasmas
Low-pressure plasmas, especially, have seen wide applicability due to their ability to selectively create reactive species at or near room temperature. For some
applications that require this additional reactivity, the ability to operate at higher
pressures or temperatures is sought. High-pressure operation reduces the complexity of the overall system by removing the need for vacuum pumping systems.
To this end, research into atmospheric pressure discharges—those that operate in
ambient conditions or at elevated pressures—has increased dramatically over the
past two decades (a review of such devices was recently given [1]). Microwave
plasma torches and jets (of which the PSMPT is a type) hold no exception. The
versatility and range of applications for these plasmas lend seriously to the need
to study their properties and range of applicability.
Pressure has a significant effect on a plasma since the collision frequency between particles describes how well a power source is able to couple energy to
the plasma—whether operating at low frequency (i.e., direct-current (DC) or
alternating-current (AC)), radio frequency (RF), or microwave frequency. Lowtemperature plasmas—those for which the charge particle density is significantly
smaller than the neutral particle density, and which are employed in laboratory and
industrial applications—are often classified as either non-thermal or thermal. Nonthermal plasmas exhibit a high-degree of non-equilibrium between free electrons
and heavier particles (neutral atoms and molecules, ions, and excited species), between the various excitation levels among the various atomic and molecular energy
modes, and between the various ionization levels. These plasmas are capable of
operating with gas temperatures as low as room temperature. Thermal discharges,
on the other hand, are often said to be in local thermodynamic equilibrium (LTE),
such that the temperatures of the electrons and heavy particles are all in a state of
quasi-equilibrium. The PSMPT and other similar devices appear to lie somewhere
between both of these regimes, such that equilibrium exists between electrons and
3
some highly excited species, while the majority of heavy particles and lower excited
species are often out of equilibrium with the electrons, but still not at the typical
“cold” level of non-thermal plasmas. This is due to competing effects between the
high-pressure operating conditions, on the one hand (which typically tend towards
equilibrium), and the smaller plasma dimensions and higher electric-field strengths
on the other. Pulsed, transient fields also lead to non-thermal conditions, but were
not applied in the case of the PSMPT as the fields under study have always been
of continuous-wave (CW) type. Temperature deviations between plasma species
based on pressure and the degree of thermalization can be seen in Figure 1.1. The
PSMPT may be able to accomplish either non-thermal or thermal-type tasks depending on operating conditions and configurations, but it remains to be seen from
diagnostics and modeling what it may be capable of.
Figure 1.1. Pressure dependence of electron and gas temperatures (from [2]).
1.2.2
Microwave Plasmas: General Classification
Plasmas used in laboratories and industrial applications divide into three groups
by input electrical power frequency: (1) low frequency (i.e., DC or AC), (2) RF,
and (3) microwave frequency. While applications of microwave plasmas abound,
a review of the literature indicates that their low frequency and RF counterparts
have received much more intense study and extensive utilization. This result is
4
undoubtedly due in large part to the ease with which early researchers were able
to obtain low frequency and RF generators and amplifiers, and the subsequent
resistance to change. However, with the availability of generators, and increasing
interest and demand for plasmas operating at elevated pressures, microwave plasmas have been quickly making their presence felt. Compared to other torch and
plasma jet designs, microwave plasma torches are more compact. They generally
use lower flow rates and power levels than their RF and DC torch counterparts
while suffering from minimal nozzle erosion. Without entering into a discussion
on the physics of microwave plasmas (see Chapter 2), a brief overview of the
classification of the relevant devices is presented here.
Authors have lamented the inability to succinctly classify microwave plasma
discharges. It has been surmised the difficulty rests with the various ways by
which microwave energy can be coupled to the plasma—in particular, because of
the similarities in scale between the microwave wavelength, device dimensions,
and plasma size. Several classification schemes for microwave plasmas have been
proposed. Marec et al. [3] divided these sources into six different categories, while
Moisan et al. [4] limited the classification to two categories, localized discharges
and traveling-wave discharges. Several reviews over the years have made attempts
to classify each of the different coupling schemes used, generally within specific
disciplines. Our interest lies in those microwave plasmas that are often referred to
as torches, jets, or flames.
1.2.3
Microwave Plasma Torches
The literature on the class of gas discharges known as microwave plasma torches
encompasses many different types of discharge structures. Adding to the foray of
naming conventions is the apparent coupling mechanisms, often related to their
RF counterparts, and so we end up with terms such as “capacitively coupled
microwave plasmas” and “microwave induced plasmas”. No physical basis for
these names appears to have been put forth in the literature.
Many of the industrial and research applications of microwave plasmas use
flowing gases, which pass through the transmission lines and escape into some
sort of atmosphere. We can place so-called microwave plasma torches into three
different categories, demonstrated in Figure 1.2:
5
1. Plasmas created within a resonant cavity but emanating as post-discharges
through some aperture, which would not allow for the propagation of microwaves outside of the cavity.
2. Plasmas created and emanating from a dielectric tube placed through a
microwave field-enhancing structure, usually a resonant cavity or a surface
wave (SW)-initiating applicator.
3. Free-expanding plasma jets, or flames, at the end of a single electrode-type
structure.
Due to the similarities with the PSMPT, this report will focus primarily on category
(3), though there are several examples of the torches in the other categories also
worth noting (see, e.g., [5]).
Plasma Jet
Microwave
Cavity
Dielectric Tube
Primary Discharge
Input gas(es)
Microwave Cavity or
Waveguiding Structure
“Electrode”
Microwave Input
Figure 1.2. Microwave plasma torches under general classification of three torch types:
(left) post-discharges, (center) discharges in tubes, and (right) single electrode-type.
1.2.3.1
Electrode Type Torches
Various devices are described in the literature where a single-electrode plasma
(SEP) is created at microwave frequencies in a free-expanding flowing gas, such
that the plasma appears flame-like. The term “electrode” refers not to the transfer
of current nor to the possibility of sheaths (as occurs RF discharges), but, rather, to
the highly-localized electric field generated by the microwaves near some protuberance. This is perhaps more akin to the electrode of a DC corona discharge (see,
e.g., ongoing research on breakdown in microwave corona discharges [6, 7]). More
6
confusion is caused when authors refer to both “electrode-less” and “electrode”
microwave plasmas when they appear to have similar field application methods.
For example, Ahmadi et al. [8] specify an SEP as electrodeless. It is perhaps more
appropriate to speak of the application method of the electric field generated by
microwaves. Within the context of microwave coupling elements between waveguides and coaxial lines, the elements are not referred to as electrodes. We can
think of these structures as a type of transmission line termination permitting a
high electric-field density to occur in a very localized region. These devices are
peculiar to microwave frequencies because of the similarity between wavelength,
device, and plasma length scales.
SEPs are typically implemented in different microwave feed structures, utilizing some level of cavity resonance (whether coaxial or waveguide) to create a
high-strength, localized electric field. The designs employ either (1) coaxial, (2)
waveguide-to-coaxial, or (3) waveguide resonant structures (see Figure 1.3), and
may either exhaust the plasma (1) open to the atmosphere or (2) into a closed,
possibly resonant, cavity. The PSMPT belongs to the waveguide type, open to the
atmosphere.
Figure 1.3. SEP torch types: waveguide (left), coaxial (center), and waveguide-to-coaxial
(right). Arrows indicate microwave input.
7
There are also hybrid types of free-expanding plasma torches, which are created
within resonators, as well (see, e.g., [9]). This may even be the case with the PSMPT
as shielding is provided to protect against microwave radiation, the shielding
providing a sort of cavity. This work focuses on the electrode-type torches because
of their apparent ability to create relatively-small plasmas due to highly localized
electric fields.
In some cases, an SEP open to the atmosphere (i.e., no metallic cavity surrounding the discharge) does not immediately expand into the atmosphere. The
plasma flow may be somewhat restricted, due to either an outer electrode wall (for
those employing a coaxial-type design [10]) or due to a long section of open-ended
dielectric tube (used for producing surface-wave plasmas [11]).
1.3
Development of Microwave Single-Electrode
Plasmas at Atmospheric Pressure
Research into microwave SEP torch-like devices has been ongoing for several
decades. Here we layout a history of these devices, describing some of their
properties and applications. There are several reviews of SEP-type devices in the
literature, particularly with respect to their use as element detection devices (see,
e.g., [12–15]).
1.3.1
Early Work: Prior to Mid-1960s
There is mention in the literature of several studies made (particularly in Germany
and the U.S.S.R.) with high-frequency flame-like discharges occurring as early as
the 1920s. A search of the available literature indicates that prior to the 1940s any
reference to a high-frequency torch discharge is in reference to inductive RF plasma
torches. With the invention of the resonant multi-cavity magnetron and other
high-power microwave amplifiers, microwaves came to be employed in plasma
applications. Though the invention of the magnetron is often first attributed to Hull
in 1920 [16], the necessary stability and power for radar applications was lacking.
Not until World War II did a successful design of the magnetron emerge to satisfy
Allied radar needs. After the war, research and manufacturing of magnetrons
proliferated and their use entered the scientific community.
8
Researchers at General Electric (GE) presented some of the earliest examples of
an SEP torch in the U.S. using a magnetron. Cobine and Wilbur [17–19] are often
cited for their study of coaxial-type torches operated at microwave frequencies (see
Figure 1.4). They examined at least two similar torches using different gases under
various powers and frequencies, primarily obtaining a qualitative description of
the operation of the torch. Melting and welding were their primary applications.
They appear to have successfully welded fernico using nitrogen.
Another researcher at GE, Jordan, presented a similar design in a patent in
1952—presumably the design used by Cobine and Wilbur, as their patent application was filed in 1947 [20]. Laroche presented a thesis in 1961 [21], and at least one
earlier paper [22], regarding a device very similar to the GE torch; the magnetron
power source operated at 1.3 GHz and 1 kW. Several European researchers also
worked on similar torch designs, including Mollwo [23], Scholz [24], Schmidt [25],
and Trunec̆ek [26].
Figure 1.4. Early coaxial electronic torch design by Cobine
and Wilbur (from [17]).
In the 1960s, Swift [27, 28] developed at least two different microwave plasma
torches. Both designs utilized a coaxial type of microwave coupling to the plasma,
though the earlier design employed a resonator-type construction and microwave
feeding mechanism (see Figure 1.5). Akin to the the previous studies, this work
seemed to be of a qualitative nature, examining the behavior of the plasma and
a lengthy discussion of the electronics setup for the plasma torch. Similarities
between combustion flames and the plasma torch “flame” were used to describe
the preliminary physical appearance of the plasma, though the author did note
that they were two fundamentally different physical processes. Obtaining stable
plasmas and efficient energy coupling to argon plasmas was particularly difficult.
9
Figure 1.5. Coaxial torch designs by Swift, from [27] (left) and [28] (right).
1.3.2
SEPs as Spectrochemical Sources: Mid-1960s to 1990s
The real advent of microwave SEPs appears to have occurred in the 1960s in Japan
and Germany when researchers applied these torches to multiple spectrochemical
analyses. These sources utilized coaxial-type designs.
Researchers at Hitachi Ltd. created a commercial ultra-high frequency (UHF)
plasma torch for spectrochemical analysis. Yamamoto and Murayama [29] developed a coaxial microwave torch by attempting to improve on their previous
RF torch design [30]. Their design was very similar to Swift’s second design,
but operating below 1 GHz. Excitation temperatures with and without a seeded
gas flow were obtained in an attempt to ascertain the suitability of the torch for
spectrochemical analysis of different solution samples. Murayama [10] modified
the design further for a higher frequency using a waveguide-to-coaxial type torch
where the coaxial section protruded through a waveguide cavity (see Figure 1.6).
Their designs inhibited any escape of microwave radiation to the environment
since the plasma filled the volume within the coaxial structure. Several reports
using their designs for spectrochemical analysis exist in the literature (see, e.g.,
[31–33]).
10
Figure 1.6. Murayama (left) [10] and Jecht and Kessler (right) [34] microwave torch
designs.
Jecht and Kessler [34] and Tappe and van Calker [35] investigated similar coaxial torch designs (see Figure 1.6). The work of Jecht and Kessler yielded gas
temperatures around 4000 K for 600 W input power in air. Both research groups
noted flashover1 from the plasma to the outer coaxial conductor when using argon.
Again, spectrochemical analysis seemed to be their motivation. A paper by Sermin
[36] continued research on their device.
1.3.3
Contemporary Research: 1990s to Present
Microwave torches (and microwave induced plasmas, in general) came into maturity and more widespread use in the 1980s–90s. Mitsuda et al. [37] developed
a coaxial microwave plasma torch that is reminiscent of a DC plasma torch, with
the plasma forming between the tip of an inner electrode and two outer electrodes.
Their first study examined diamond growth.
Moisan et al. [38] developed the Torche à Injection Axiale (TIA), as well as the
Torche à Plasma de Surfatron (TPS) [39], after much of their successful work studying microwave surface-wave plasmas (see Figure 1.7). A similar design is described
by Stańco [40]. Several studies of microwave plasma torch behavior using the TIA
exist in the literature, the earliest applying studies of its electrodynamic character1
Flashover in microwave plasma torches is the presence of (generally short-lived) plasma
“sparks” from the nozzle tip or plasma jet to the outer conductor. The outer conductor, in
waveguide-based torches, is the wall of the waveguide aperture.
11
istics. As the device has typically been applied to spectrochemical analysis, much
of the more recent work by other research groups using the TIA and its derivatives
has focused on understanding the particle kinetics, processes, and structure of the
plasma jet.
A design called the TIAGO (Torche à Injection Axiale sur Guide d’Ondes) was
developed out of the TIA concept by Moisan’s research group [41]. This design
deviated from the TIA by applying a waveguide-type design, based off of the
surfaguide concept. Their work appears to be continuing through a Polish research
group, attempting to examine the use of the TIAGO for treatment of hazardous
waste gases [42]. It is also interesting to note that the TIAGO concept was attempted
in a multiple-nozzle configuration as an array of torches from a single waveguide.
Describing the plasma jet, they also were aware of two distinct regions, although
of different configuration than the PSMPT (see Figure 3.1): an inner core and an
outer envelope. The TIAGO employs a reduced-height waveguide, which is at
full height (WR–430) at both the generator and sliding short ends, and the nozzle
is fully conical and protrudes significantly through the waveguide aperture (see
Figure 1.7). The system is operated at 2.45 GHz and an input power of 0.2–2.0 kW.
The authors present a method for determining the plasma admittance for use in
determining the equivalent circuit characteristics of the torch system. Central to the
study they present is the experimental determination of the input impedance and
the plasma conductance for single and multiple-torch systems. Their circuit model
matched closely with collected data. For a single nozzle, the reflected power
appears to be much more dependent upon operating conditions (e.g., stub and
sliding short positions) than when the authors tested with multiple nozzles. This
suggests that the interaction of several nozzles and plasma jets helps to stabilize
the system and match the TIAGO load to the generator much more easily. The TIA
and TIAGO designs and resulting plasmas present a strong basis for comparison
with the PSMPT and will be referenced further in this paper.
A capacitively-coupled microwave plasma (CMP) torch was further studied and
examined by Bings et al. [43]. Their torch is quite similar to the torches designed
by Murayama in the 1960s–70s, and is based on much work on CMPs through the
1980s, primarily dedicated to studying the use of such devices for spectrochemical
solution analysis. As early as 1990, Jin et al. [15] also began using what is formally
called the microwave plasma torch (MPT—not to be confused with the PSMPT)
12
Figure 1.7. TIA (left) [38] and TIAGO (right) [41] microwave torch designs.
for analysis purposes (see Figure 1.8). The inner conductor of their device was
composed of two concentric metal tubes; it was determined by other researchers
that this led to some instabilities in the operation of the MPT and, as such, has
been modified in later research [44, 45]. It should be noted that extensive analysis
has been done at Eindhoven Technological University (see, e.g., [46–48]), Cordoba
University (see, e.g., [49]), and the Instituto de Plasmas e Fusão Nuclear (see, e.g.,
[50]) on both the MPT and TIA over the past 15 years.
More recently, researchers in the Ukraine have been developing and studying a
microwave torch based on a waveguide-to-coaxial design [51,52]. These researchers
have been seeking to understand the microwave field structure in such microwave
torch plasmas. Similar to the MPT, they noted a constriction, or a “waist,” which
separated the torch plasma into two different regions. They determined that the
plasma was not necessarily a single coherent “flame,” but, rather, resulted from
the assemblage of multiple filaments rotating around the rim of the nozzle. From
measurements of the electric field profile along the axis of the discharge, they
came to the conclusion that the filaments allow propagation of the microwaves as
surface waves. A majority of the applied power (> 90%) was reported to have been
absorbed by the discharge.
13
Figure 1.8. MPT (left) [15] and MPJ (right) [53] torch designs.
A group from the University of Liverpool designed a waveguide-based microwave plasma jet (MPJ—see Figure 1.8), very similar to the PSMPT, particularly in
the nozzle design. They were able to create plasmas at two frequencies, 896 MHz
and 2.45 GHz, at upwards of 5 kW and 6 kW, respectively. Several studies were
performed to examine the viability of their microwave plasma jet (MPJ) to perform
cutting and industrial applications [53].
A summary of the different torches and some of their operating conditions
is provided, in an approximately-chronological order, in Table 1.1. Often these
devices are identified by very similar names (if any name is ascribed to them), as can
be seen from the table. It is for this reason that we will specify our own microwave
plasma torch as the PSMPT. There are also several other recent microwave torch
designs not mentioned here [54–56] including some at low power [57].
14
Table 1.1: A summary of microwave SEP torch development in the literature with
coaxial, waveguide-to-coaxial (WG-to-coax), and waveguide-only (WG) electrode
structures.
Designationa Structureb
Frequencyc
(GHz)
Powerd
(W)
Gas
Flow Rate
(SLMe)
Gas Temperaturef
(K)
Refs.
Electronic Torch
coaxial
0.5–1.1
0.915
≤ 1000
≤ 5000
Air, O2 , N2 , H2 , Ar,
He, CO2 , CCl4
7.9–31.5
1870–3740f2
[17–19]
Torche Électronique
coaxial
1.3
≤ 1000
Air, O2 , N2 , Cl2 , Ar
1.7–6.7
2000–2800f5
[21, 22]
High-frequency
torch discharge
coaxial
0.940, 1.5
(100)
Air, N2
—
4000f6
[23]
High-frequency
corona torch
discharge
coaxial
1000
—
—
—
3000
[26]
Plasma Flame
coaxial
0.461, 2.4
—
Air, O2 , N2 , Ar
—
3000–4000 (air)
[35]
UHF Plasma Torch
coaxial
0.460
≤ 240
Air, Ar
0.5
4000f3
[27]
Microwave
Plasma Torch
coaxial
2.45
500–2500
Air, Ar
1.4–7.1
—
[28]
UHF torch
discharge
coaxial
0.520
200
Ar
6.5
7500–8200f4
[29]
UHF discharge
WG-to-coax
2.469
50, 200, 400
Ar
11 (g/min)
4400–4700
[10]
Microwave
plasma beam
WG-to-coax
2.45
200–3000
N2 , Air, CO2 , Ar, H2
30–200
6700 (assumed LTE)
[58]
15
Continued on next page
Table 1.1: A summary of microwave SEP torch development in the literature with
coaxial, waveguide-to-coaxial (WG-to-coax), and waveguide-only (WG) electrode
structures.
Designationa Structureb
Frequencyc
(GHz)
Powerd
(W)
Gas
Flow Rate
(SLMe)
Gas Temperaturef
(K)
Refs.
coaxial
2.45
2000–5000
H2 , Ar, CH4
≤ 30
—
[37]
WG-to-coax
2.45
≤ 2000
Ar, He
3
(varies with source)
[38]
Torche à
Plasma de
Surfatron (TPS)
coaxial
2.45
10–180
Ar, He
0.2–1
—
[39]
CapacitivelyCoupled
Microwave
Plasma (CMP)
coaxial
2.45
600
Air, N2 , Ar
0.8
2500–4300
[43]
Microwave
Plasma
Torch (MPT)
coaxial
2.45
40–200
Ar, He, N2
1–5
1000–6000
[15]
Coaxial microwave
plasmatron
coaxial
10
≤ 10
Ar
≤2
—
[51]
WG
0.896, 2.45
2000–5000
Ar
2–7
—
[53]
Microwave
plasma jet
Torche à
Injection
Axiale (TIA)
Microwave
Plasma
Jet (MPJ)
Continued on next page
16
Table 1.1: A summary of microwave SEP torch development in the literature with
coaxial, waveguide-to-coaxial (WG-to-coax), and waveguide-only (WG) electrode
structures.
Designationa Structureb
Torche à
Injection
Axiale sur Guide
d’Ondes (TIAGO)
WG
Frequencyc
(GHz)
Powerd
(W)
Gas
Flow Rate
(SLMe)
Gas Temperaturef
(K)
Refs.
2.45
200–2000
Ar, N2
1–3
1500–3500
[41, 42]
Notes:
“—" indicates an unknown value or parameter; values in parentheses are uncertain.
a The designation for the torch is that specified most often in the literature when citing the given torch design.
b See Figure 1.3 on different base design structures.
c The operating frequency of the microwave generator.
d The power output of the microwave generator—not necessarily the power absorbed by the plasma.
e Various flow rate dimensions and units were described in the literature; an attempt was made here to specify all flow rates in terms of SLM.
f The source and method of the temperature measurements varies in each of the references.
f2 Via buoyancy effect estimations for N at 5 kW.
2
f3 Via unspecified spectroscopic method for air.
f4 Excitation temperature via absolute line intensity method assuming LTE for Ar.
f5 Via pyrometer.
f6 Via nitrogen bands at 337.1 nm.
17
Chapter 2
Theory
2.1
Microwave Transmission
The microwave energy used to breakdown the gas is coupled to the plasma using
waveguides, which are a class of electromagnetic transmission lines most often used
for high-power transmission of microwaves. They are distinguished from other
types of transmission lines (e.g., coaxial or microstrip) by the type of propagation
modes they permit, by their geometry, and by their power handling capability.
Waveguides are produced in standard sizes, often with a rectangular cross-section,
and, combined with their high-power capabilities, are ideally suited for use with
the PSMPT.
2.1.1
Electromagnetic Wave Propagation
In free space, electromagnetic waves propagate in the transverse electromagnetic
(TEM) mode. However, microwaves transmitted through a rectangular waveguide
are constrained to propagate in either transverse electric (TE) modes (i.e., with no
electric field component in the direction of propagation), transverse magnetic (TM)
modes (i.e., with no magnetic field component in the direction of propagation),
or a combination of TE and TM modes. Considering the waveguide geometry
presented in Figure 2.1, we can say that a TE mode propagates with Ez = 0.
The propagation constant, βz , for TE modes is given by (see, e.g., [59])
r
βz =
β2
mπ
−
a
2
nπ 2
−
,
b
(2.1)
y
x
b
a
z
Figure 2.1. Rectangular waveguide cross-section dimensions.
where m, n = 0, 1, 2, . . . , with m = 0, n , 0 and n = 0, m , 0, and
β = 2π f
√
µε
(2.2)
is the propagation constant for TEM waves. In Equation (2.2), f is the frequency
of the electromagnetic radiation, µ the magnetic permeability of the medium, and
ε the electric permittivity of the medium. The gaseous dielectric medium within
waveguides is generally assumed to have the properties of free space, µ = µ0 and
ε = ε0 .
The propagation constant must be real for the microwaves to operate in a propagating regime. Cutoff—the condition where microwaves can no longer propagate
in the waveguide—is then given by the frequency
1
fc =
√
2π µ0 ε0
r
mπ
a
2
nπ 2
+
.
b
(2.3)
Frequencies below fc are unable to propagate within the waveguide. Cutoff frequencies for several of the first TE modes for a WR–284 waveguide (with dimensions a = 7.214 cm and b = 3.404 cm) are listed in Table 2.1. TM modes only begin
propagating for m and n not equal to zero, so that TM11 is the lowest TM mode
allowed; therefore, the fundamental mode is the TE10 mode. Higher modes will see
an increase in cutoff frequency, however. Examining the propagation constant for
19
Table 2.1. Cutoff frequencies for the five lowest modes of TEmn propagation in a WR–284
full-height waveguide.
m
n
fc (GHz)
1
2
0
1
2
0
0
1
1
1
2.078
4.156
4.404
4.869
6.055
this mode, we can see that the propagation of the microwaves is unaffected by the
height of the waveguide, b (assuming we maintain b < a). This result implies that
the waveguide height may be varied while maintaining a single operating mode
with the same cutoff frequency.
Another parameter of interest is the guide wavelength, λg , of the waveguide.
If λ is the wavelength of a wave of a given frequency, f , propagating in an unbounded medium, then λg is the wavelength of the wave at the same frequency
but propagating inside the waveguide. For frequencies above the cutoff frequency,
its value is given by
λg = p
λ
λ
= p
,
1 − ( fc / f )2
1 − (λ/λc )2
(2.4)
where λc is the wavelength of the cutoff frequency. For the TE10 mode in an air-filled
waveguide, the guide wavelength is
−1/2
λg = µ0 ε0 f 2 − (2a)−2
.
(2.5)
Values for the wavelengths at an operating frequency of 2.45 GHz and with waveguide dimensions of a = 7.214 cm and b = 3.404 cm (WR–284) in the TE10 mode are
shown in Table 2.2.
20
Table 2.2. Frequencies and wavelengths of
standard WR–284 full-height waveguide operating in the TE10 mode at f = 2.45 GHz.
f
fc
2.450 GHz
2.078 GHz
λ
λc
λg
12.24 cm
14.43 cm
23.15 cm
Three different types of impedances—intrinsic, wave, and characteristic—can
be defined for microwave transmission lines of which microwave waveguides are
p
a type. The intrinsic impedance, η = µ/ε, depends upon the medium in which
the electromagnetic waves are propagating. In free space (and approximately so
in air), η = 377 Ω. The wave impedance for a TE-mode wave, ZTE , depends not
only on the material in the waveguide but also the geometry and, therefore, on the
frequency of the waves. It can be expressed as
η
ZTE = p
.
1 − ( fc / f )2
(2.6)
The characteristic impedance, Z0 , is a function of the voltage and current along a
transmission line and will be examined in the following section.
The power flow along the waveguide can then be obtained from Poynting’s
Theorem, which can be solved for the power flow of the TE10 mode in a rectangular
waveguide [60] as
ab 2 λc
E
P=
2ZTE 0 λg
!2
.
(2.7)
We can then solve for the root-mean-square electric field strength, E0 , for any power
input into a waveguide, which is found to be
s
E0 =
2η( fc / f )2
ab 1 − ( fc / f )2
3/2
√
P.
(2.8)
21
Figure 2.2 provides a plot of this relationship for two different waveguide heights.
Such a relationship demonstrates how reducing the waveguide height increases the
electric field strength within the waveguide, which is beneficial when attempting
to initiate a plasma. This can also be useful in the inverse situation, where one is
trying to eliminate the possibility of breakdown in the waveguide (e.g., when the
waveguide is used in radar or telecommunications systems).
TE10 -mode field structures yield a maximum of the electric field along the
waveguide’s x-directed center and in only a single field direction, parallel to the
y-axis (i.e., Ex , Ez = 0). For this reason, the PSMPT’s aperture and nozzle axis are
placed along the center of the waveguide. As a function of position and time, the
electric field strength can be expressed as
E y (x, z, t) =
√
2E0
λc
π
sin x sin ωt − βz z ,
λg
a
(2.9)
where we see that, for a given cross-section location, the maximum electric field
occurs at x = a/2.
Half height
Full height
Electric Field Strength (kV/cm)
700
600
500
400
300
200
100
0
0
500
1000
1500
Power (W)
Figure 2.2. Mid-line field strength in standard WR–284
rectangular waveguide for two different heights.
22
2.1.2
Impedance Matching
A waveguide network is a transmission line for microwaves. One intends for the
majority of microwave energy delivered to such a network to be absorbed by the
plasma, rather than through resistive dissipation in the walls and radiation losses.
This calls for proper impedance matching, which turns out to be an essential factor
for consistent and stable PSMPT operation. In any microwave transmission line,
we can define a characteristic impedance, Z0 . If that transmission line is terminated
by a circuit element with the same impedance, then the transmission line is able to
deposit all of its power (in theory) into the load. We can see this by examining the
following.
For the lossless transmission line terminated by a load impedance of ZL , as seen
in Figure 2.3, we can write the total voltage and total currents on the line as the sum
of forward and reflected waves. Assuming a propagation constant βz , we have
V (z) = V0fwd exp − jβz z + V0refl exp jβz z
(2.10)
V0refl
exp −jβz z +
exp jβz z .
Z0
(2.11)
and
I (z) =
V0fwd
Z0
Then at z = 0 we can write for the load impedance
refl
fwd
VL V0 + V0
ZL =
= fwd
Z0 .
IL
V0 − V0refl
(2.12)
We can then compare the strength of the reflected wave to that of forward wave to
determine how much of the wave is reflected through the reflection coefficient,
ΓL =
V0refl
V0fwd
=
ZL − Z0
Z0 .
ZL + Z0
(2.13)
23
Therefore, if the load impedance is matched to the impedance of the line, there
will be no reflections. The voltage and current along the transmission line can be
written as a superposition of the forward and reflected waves. The time-averaged
power flow can then be found as
fwd
1 |V0 |
1
(1 − |ΓL |2 ).
Pavg = Re[V(z)I(z)∗ ] =
2
2 Z0
(2.14)
The time-averaged power along is independent of position along the line and is
simply the difference between the forward and reflected power levels, or
Pavg = Pfwd − Prefl = Pabs ,
(2.15)
where Pabs is the power absorbed by the load. In reality, there will be losses
along any transmission line through resistive losses in the walls of the waveguide
components. However, typically these losses are ignored, especially for short
sections of waveguide operating away from cutoff and terminated by a resistive
load.
V(z), I(z)
IL
+
Z0 , βz
VL
ZL
−
l
z
0
Figure 2.3. Terminated lossless transmission line.
So that transmission line circuit theory can be applied to waveguides, it is important to understand how the characteristic impedance, Z0 , is defined in a waveguide.
When examining waveguides (as opposed to two-conductor, TEM-mode transmission lines, such as microstrip and coaxial), a unique voltage and current cannot be
defined between any two points. However, a consistent definition can be made be
made if a particular mode is chosen. Waveguides are typically operated in the TE10
mode, which simplifies the matter. The characteristic impedance of a line is given
24
by the ratio of voltage to current, i.e.,
Z0 =
V fwd V refl
= refl ,
Ifwd
I
(2.16)
and is typically taken to be equal to the wave impedance, ZTE . In order for these
voltages and currents to be applied in the case of a waveguide, they are typically
defined so as to yield the power flowing in the waveguide. Further details can be
found texts on microwave engineering (see, e.g., [60, 61]).
Ideally, we would like to have maximum power delivered to the load (Prefl = 0).
ΓL is not always zero, and so a means of matching the impedance of the line to
the load is required. Assuming that the generator side of the line is matched to
the line, with proper impedance matching from the matching network (effectively
terminating the waveguide line in Z0 ) maximum power can be delivered to the load.
Such a matching network for waveguides can consist of a series of “stubs” (typically
two or three in sequence) inserted partially into the waveguide at varying depths.
A three-stub tuner will effectively allow zero reflection back to the generator while
trapping the microwave energy in multiple reflections between tuner and load.
The interaction between the forward and reflected waves on the transmission
line set up standing waves along the length of the line, typically expressed through
the voltage standing wave ratio,
VSWR =
Vmax 1 + |ΓL |
.
=
Vmin
1 − |ΓL |
(2.17)
The maximum and minimum voltages are determined by the constructive and
deconstructive interference of the two waves. These standing waves within a
waveguide lead to a pattern of electric field strength maxima and minima along
the length of the waveguide. For a waveguide, the positions of these maxima and
minima can be changed by adjusting the position of a “short” at the end of the
waveguide.
25
2.2
Microwave Plasmas
Gaseous electrical breakdown can occur due to the imposition of fields of varying
frequency and strength. We will focus here on those properties of plasmas created
and sustained at microwave frequencies only. Microwave plasmas have the benefit
of having low electron loss due to diffusion (as compared to DC and other lowfrequency plasmas) and can operate without electrodes.
2.2.1
Breakdown and Plasma Initiation
Already present in any volume of gas are a small number of free electrons and ions.
When an electric field is applied to the gas, the charged particles gain energy and
a small current begins to flow in the gas between the two electrodes across which
the electric field is applied. In gaining energy, these charged particles can transfer
energy to other particles, or they may recombine with other molecules and ions,
or diffuse to the electrodes. Under the appropriate conditions, this energy transfer
can lead to the further ionization of atoms and molecules. Increasing the voltage
through a region of ion and electron density known as the Townsend regime (in
the DC case) brings the gas to the breakdown voltage where the current is able to
close the circuit between the two electrodes and a gas discharge is maintained. The
breakdown process continues until a steady state is reached in which the production
of electrons due to ionization is balanced by the loss of electrons through various
means (e.g., diffusion, attachment, recombination).
If an AC voltage is applied to a gas, a discharge can still be created. At low
frequencies and high pressures, the discharge will tend to behave similar to that of
a DC discharge. As the frequency of the field increases or the pressure decreases,
the frequency of collisions between electrons and ions eventually becomes less than
the frequency of the field. Increased electron density and the ensuing breakdown
tend to be localized, and the field helps to keep the electrons far from the edge of
the plasma from diffusing towards the edge [62].
Losses due to diffusion of electrons to the walls of the plasma container (for
a wall-stabilized plasma) or to the surrounding gas (for a size-stabilized plasma)
are often expressed in terms of the diffusion length, Λ [63]. This value describes
the typical length scale over which the electron density, ne , varies. The conditions
26
required for breakdown of the gas to create a discharge relating diffusion lengths
and electric field strengths to pressure can be determined experimentally from a
(microwave) Paschen curve for the gas. An example is shown in Figure 2.4 for
argon [64]. Higher pressure environments necessarily lead to a higher electric field
strength requirement in order to attain breakdown. This results from the fact that
the number of collisions between electrons and neutral molecules increases with
pressure, such that a larger proportion of the electron population will be unable to
attain significant energy from the field between collisions to initiate any ionization
avalanche.
Electric Field Strength (V/cm)
104
103
102
101
10−2
f
f
f
f
10−1
100
=0.994 GHz,Λ =0.631 cm
=2.800 GHz,Λ =0.151 cm
=2.950 GHz,Λ =0.135 cm
=2.800 GHz,Λ =0.051 cm
101
102
103
Pressure (Torr)
Figure 2.4. Microwave Paschen curve for argon for various
frequencies and characteristic diffusion lengths (from [64]).
Measurements were obtained within cylindrical resonant
cavities. The characteristic diffusion lengths depend upon
the dimensions of the cavity, most especially (for these measurements) the height. Larger lengths imply a taller cavity.
For homogeneous electric fields, relatively low pressure, and constant diffusion
coefficient (D), ionization frequency (νi ), and attachment frequency (νa ), this value
can be expressed as a characteristic diffusion length,
1
∇2 ne
=
−
.
Λ2
ne
(2.18)
27
In this situation, Λ is determined by the geometry of the plasma container, typically
a resonant cavity. However, for inhomogeneous electric fields and high pressure
(and, therefore, inhomogeneous parameters D, νi , νa ) it becomes important to
study an effective diffusion length [65]. Determining such a diffusion length can
only be accomplished analytically for simple geometries; otherwise, it is necessary
to determine this length experimentally [66]. As such, the knowledge of the exact
parameters (frequency, power, pressure, temperature, gas mixture composition,
etc.) for microwave breakdown is not typically available.
2.2.2
Microwave Plasma Maintenance Processes
Once breakdown is achieved, the microwave power is able to couple more easily
into the gas and sustain the plasma against electron and ion losses. During breakdown, the electrons are lost through free-diffusion, due to minimal interaction with
other charged particles. After breakdown and into steady-state, the electrons are
lost through ambipolar diffusion and (at elevated pressures) recombination. Free
electrons are excited and transfer energy from the microwaves to the heavy particles (neutrals and ions—being more massive than the electrons, ions are not able
to transfer energy from the microwave field). Power balance in the plasma is then
given by equating the rate of energy addition to electrons to the rate of energy lost
by electrons. The average power acquired by electrons between collisions is given
by
θabs =
ν2
e2 E2
,
2me ν ω2 + ν2
(2.19)
where ν is the collision frequency for momentum transfer and ω is the radial
frequency of the applied electric field. This power then corresponds to the power
lost by the electrons in colliding with other particles, through elastic and inelastic
collisions. The inelastic collisions lead to internal energy changes in the heavy
particles, forming excited states or additional ionized species.
28
2.2.3
Surface-Wave Plasmas
The propagation of microwaves in a plasma typically takes place in one of two ways,
either (1) within a microwave circuit (inside of waveguides or resonant cavities)
or (2) as traveling waves in which the electric field of the wave propagates along
the surface of the plasma [4]. For the second case, surface waves are able to travel
along the interface between the plasma and the surrounding media, typically a
dielectric tube. The electric field is evanescent (i.e., non-propagating) in directions
normal to the surface of the plasma, and much of the power is able to be absorbed
by the plasma. Microwave plasma torches, such as the PSMPT, do not produce
true surface-wave plasmas as they do not occur with a well defined boundary
between plasma and dielectric. However, there is evidence that these plasmas still
belong to this class (see, e.g., [51]). Typically, surface-wave plasmas are produced
in dielectric tubes, with relative permittivity εg . In order for the field to be able
to propagate along the surface of the plasma and tube, there is a constraint on the
microwave frequency, such that
ω ≤ ωpe (1 + εg ),
(2.20)
where the plasma frequency is
ωpe =
ne e2
.
me ε0
(2.21)
Should the electron number density decrease below a critical value, such that
ω > ωpe , then the wave will no longer be able to be sustained by the plasma.
As the wave travels along the plasma column, power is dissipated. The rate
at which power is dissipated in a direction along the axis of the plasma column
depends on the electron density and vice versa, and leads to both a decrease in the
power of the wave and the electron number density in the plasma column. This
can be seen, for example, in Figure 2.5.
29
×1010
Electron Density (cm−3 )
6
100
4
50
2
0
0
50
100
Wave Power (W)
Density
Power
0
150
Axial Position (cm)
Figure 2.5. Axial variation of electron number density and
power along a surface-wave plasma column (from [67],
where f = 300 MHz, P0 = 100 W, and tube diameter =
2.6 cm).
2.3
Optical Emission Spectroscopy of Plasmas
Determining plasma properties requires minimal impact on the plasma. Optical
emission spectroscopy techniques, using only the radiation emitted by the plasma,
provide a useful method for determining plasma parameters, including gas temperature. One of the more common methods assumes that the gas temperature
closely corresponds with the rotational temperature of diatomic molecular species
within the plasma. This assumption stems from the fact that rotational energy levels
are tightly spaced and typically adjust quickly to equilibrium values when compared with other plasma processes. Rotational temperature measurements arise
from studying transitions leading to rovibronic spectra in diatomic molecules. This
section introduces relationships that provide for an understanding of how these
transitions lead to different spectral emission profiles in the plasma. An estimate of
the temperature of the heavy particles within the plasma follows from comparing
these calculated profiles against experimental data.
30
2.3.1
Line Positions in Diatomic Rovibronic Spectra
Rovibronic transition spectral profiles result from combining (1) spectral lines, the
positions of which are due to the energy differences between electronic, vibrational,
and rotational energy states involved in a transition; (2) the intensity of the transition between these energy states; and (3) the line shape of a spectral line of the
transition due to physical broadening processes. We consider the calculation of the
line positions first.
A molecule can exist in multiple quantum states, all of which depend on the
electronic, vibrational, and rotational configuration of the molecule. The Hamiltonian, H, of one of these states provides a means to determine the energy, Em of the
state, via the eigenvalue equation
H |ψm i = Em |ψm i .
(2.22)
If the wavefunctions for each of the states is uncoupled, then the Hamiltonian for
the molecule can be determined from the Hamiltonian matrix elements,
Hmn = ψm H ψn ,
(2.23)
which leads to a diagonal matrix of energies (eigenvalues) for each of the states.
However, if coupling exists between states then the matrix is not diagonal (i.e., Hmn
is not necessarily equal to zero if m , n) and must be diagonalized to determine the
energies (eigenvalues) of the states. With these energies we can then calculate line
positions. For some molecules, effective Hamiltonian matrix elements for various
states have been calculated and can be found in the literature.
Within a diatomic molecule, internal energy state transitions can take place simultaneously between rotational, vibrational, and electronic energy modes, leading to rovibronic spectra. Assuming that the wave function for the molecule can
be written using the Born–Oppenheimer approximation, i.e.,
ψ = ψext ψint = ψtran ψrot ψvib ψelec ,
(2.24)
31
we can then write the change in energy between rovibronic states as a sum of
contributions from energy changes between states within each mode, i.e.,
∆ν̄ =
0 0 0
0
∆E
0 0
= (∆Te )nn00 + ∆G(v)nn00vv00 + ∆F(J)nn00vv00J J00 ,
hc
(2.25)
where each of the n, v, and J specify the electronic, vibrational, and rotational
levels. These transitions lead to systems of spectral bands of a specific periodic
structure and within certain spectral frequency regions, dependent upon the molecular species and nature of the transitions.
2.3.2
Line Intensities in Diatomic Rovibronic Spectra
The intensity of radiation emitted by a molecule undergoing a transition from an
upper rovibronic state n0 v0 J0 to a lower n00 v00 J00 is given by
0 0 0
0 0 0
0 0 0
I = Nn v J hνnn00vv00J J00 Ann00vv00J J00 ,
(2.26)
0 0 0
0 0 0
where Nn v J is the population of the upper state, hνnn00vv00J J00 is the energy of the
0 0 0
emitted photon between the two states, and Ann00vv00J J00 is the transition rate between
0 0 0
the two states. Calculation of the energy between states, ∆E = hνnn00vv00J J00 , results from
the line positions determined in the previous section. Here we will consider the
contributions due to upper-state populations and the transition rate.
2.3.2.1
Level Populations
We will assume that the individual energy modes—electronic, vibrational, and
rotational—can be described by a Boltzmann population distribution within each
respective mode in the gas. No attempt will be made to model the level populations
(e.g., using collisional–radiation models). Considering the electronic energy levels
first, the relative population of all molecules of a certain species in the gas excited
to a certain electronic level is given by
ge exp (−Te /kTelec )
Ne
=
.
N
Qe (Telec )
(2.27)
32
For the electronic levels in a diatomic molecule, the statistical weight is given by
ge = 2 − δ0,Λ (2S + 1) .
(2.28)
The relative population of vibrationally excited states is expressed as
gv exp (−G(v)/kTvib )
Ne,v
=
.
Ne
Qv (Tvib )
(2.29)
Vibrational levels have statistical weight gv = 1. Rotational level populations are
given by
Ne,v,J
g J exp (−F(J)/kTrot )
=
,
Ne,v
Q J (Trot )
(2.30)
g J = (2J + 1) .
(2.31)
where
From these expressions, rotational line intensities are
Ne,v,J Ne Ne,v Ne,v,J
=
N
N Ne Ne,v
2 − δ0,Λ (2S + 1) (2J + 1) exp (−Te /kTelec ) exp (−G(v)/kTvib ) exp (−F(J)/kTrot )
.
=
Qe (Telec )Qv (Tvib )Q J (Trot )
(2.32)
At any specific temperature for each of the energy level modes, the partition functions Qe , Qv , and Q J are constant. Also, for a set of rovibrational levels within the
same electronic level, the factor involving Te remains unchanged. Therefore, at a
specific set of temperatures Tvib and Trot , the relative population among rotational
levels for a single electronic level is given by
Ne,v,J
(Tvib , Trot ) ∝ 2 − δ0,Λ (2S + 1) (2J + 1) exp ((−G(v)/kTvib ) exp ((−F(J)/kTrot ).
N
(2.33)
33
2.3.2.2
Transition Probabilities
The transition probabilities (Einstein coefficient) can be expressed as
0 0 0
n0 v0 J0
An00 v00 J00 =
64π4 (νnn00vv00J J00 )3
3ε0
hc3
0
|Rve ,v” |2
S J0 ,J”
,
2J0 + 1
(2.34)
0
where |Rve ,v” |2 is the square of the electronic transition moment and S J0 ,J” are the
rotational line strength (Hönl–London) factors.
2.3.3
Line Shapes
For any transition between bound internal energy states, the expected line, as calculated from the energy of the transition, is not infinitesimally thin (i.e., purely
monochromatic). Rather, due to several interactions between the particles and
fields in the plasma, as well as the implications of the uncertainty principle, the
detected line intensity is spread over an energy range. The particular shape of a
spectral line depends upon the process leading to the broadening, whether it be
natural (or intrinsic), instrumental, Doppler, or collisional in origin. Typically, the
first three of these lead to line shapes following a Gaussian distribution about the
line center, while collisional broadening presents a Lorentzian distribution. Accumulating the effects of each of these requires a convolution of each of the process
profiles. So long as the effects are independent of each other, this convolution leads
to a Voigt profile distribution, in which the effects of Gaussian and Lorentzian process are combined. Simplifications to the calculation of the Voigt profile exist in the
form of pseudo-Voigt profiles (see, e.g., [68]).
Φ(ν − ν0 )p-V = η fL (ν − ν0 ) + (1 − η) fG (ν − ν0 ),
(2.35)
where fL and fG are the usual Lorentzian and Gaussian components, respectively,
and η measures the contribution from the Lorentzian profile.
34
2.3.4
Relative Intensity Spectral Profile
Putting together the contributions from line position, intensity, and shape, the
profile of any given rotational line in a rovibronic transition is
0 0 0
0
Irel (ν) = Φ(ν)(νnn00vv00J J00 )4 exp (−G(v)/kTvib ) exp (−F(J)/kTrot )|Rve ,v” |2 S J0 ,J” ,
(2.36)
where factors that are constant for a specified electronic transition at temperatures
Tvib and Trot have been neglected. With this result we can calculate the relative
intensities across a spectral band system.
35
Chapter 3
Methods
3.1
Experimental Apparatus
The experimental apparatus is composed of several elements: the PSMPT–waveguide
assembly, the microwave transmission system, the gas supply system, and the optical spectroscopy system.
3.1.1
PSMPT-Waveguide Assembly
Electrical components in typical microwave plasma torches should accomplish two
goals: (1) provide the proper conditions for a microwave field distribution that both
creates and sustains the plasma and (2) ensure the efficient power transfer to the
plasma. In the case of the PSMPT, an axial gas-injection nozzle primarily satisfies
the first goal, while providing the necessary gas flow. Machined from oxygen-free
high-conductivity (OFHC) copper, the nozzle protrudes through a reduced-height
waveguide. A schematic of the PSMPT assembly can be seen in Figure 3.1.
Chapter 2 explains that, when operating in the TE10 mode, the peak of the electric
field distribution occurs at the centerline of a waveguide. Therefore, the nozzle
was positioned with its axis aligned as such. The nozzle threads into a coupler
soldered to the outer surface of one of the waveguide’s wide walls. A small circular
aperture in that same wall allows the nozzle to protrude through the waveguide.
Nozzle axis alignment is made perpendicular to the wide wall, such that the nozzle
tip is centered with a larger circular aperture (diameter of 2 cm) in the opposing
wide wall. This second aperture permits microwave energy to radiate from the
waveguide. Depth alignment of the nozzle through the waveguide is such that the
nozzle tip rests approximately flush with the outer surface of the waveguide (n.b.:
there is a finite thickness to waveguide walls). That part of the nozzle found within
Waveguide
Plasma jet
Microwave power input
Quartz tube
to Sliding short
Nozzle
Injectors
Secondary gas
Nozzle gas
Figure 3.1. PSMPT assembly schematic.
the waveguide—and therefore exposed directly to the microwave radiation—has a
smooth surface. The segment of the nozzle nearest the tip reduces in diameter in a
conical manner. The nozzle and PSMPT assembly is typically oriented such that gas
flow is directed upward into the ambient atmosphere, though other orientations
are possible.
Unlike typical off-the-shelf waveguides, the PSMPT waveguide consists of a
WR–284 section tapered to a reduced-height—a decrease in narrow-wall height
to half of its standard value, approximately 1.7 cm. While maintaining the same
guide wavelength and cutoff frequency, such reduced-height waveguides are often
used in microwave plasma applications to increase power density (and, therefore,
electric field strength) in the plasma region (see, e.g., [69, 70]). No special surface
treatment was given to the brass reduced-height waveguide.
To mitigate the effects of nozzle erosion, a secondary assist gas flows through
the waveguide walls through two injectors. A small quartz tube, made to slip-fit
inside of the waveguide aperture and sit on the bottom waveguide wall, confines
this secondary flow out through the aperture in the waveguide. Soldered to the
outer waveguide wall, the two injectors for the secondary flow gas are aligned at
45◦ to the nozzle axis separated by 180◦ to each other about the nozzle axis. As
such, these injectors introduce flow tangentially to the inside of the quartz tube
37
wall, creating a swirl flow around the nozzle. The quartz tube extends slightly
beyond the outer waveguide wall, the length being such so as to balance (1) the
attempt to reduce the amount of air entrained from the ambient environment and
(2) to allow for any future possibility plasma interaction with a movable surface.
3.1.2
Microwave Transmission System
Early testing used a 2.45-GHz, 2.5-kW magnetron source (Model: Gerling Laboratories GL103A) operating with no active tuning. Output power control of the source
involved tuning a dial that controlled the output current level of the magnetron.
Power levels were measured using power meters (Model: Hewlett–Packard 432A)
connected to thermistor-mounted (Model: Hewlett–Packard 478A) bi-directional
couplers.
Testing to date primarily involved the use of a magnetron microwave power
generator (Model: Daihen SGM-15A) operating at a frequency of 2.45 GHz. Directcurrent, high-voltage power (Model: Daihen SGP-15A) provides a maximum output (of microwave power) of 1500 W, with a quoted output stability within ±1.0%
of output power. A dial on the power supply unit provides magnetron output
power control, which can be adjusted over a range of 0–1500 W. Magnetron output
power is delivered into a microwave waveguide circuit. The circuit components are
similar to those utilized in many other microwave plasma and materials processing
applications in research and industry.
The microwave energy leaving the magnetron first enters a three-port circulator,
the purpose of which is to control the direction of microwave transmission such
that reflected microwaves from the plasma source do not interfere with or damage
the magnetron. It is oriented such that microwave energy from the magnetron is
delivered to the microwave circuit and any reflected power is deposited into an aircooled dummy load. Immediately following the circulator is a combined detector
and impedance matching unit (Model: Daihen SMA-15A) with three independent
probes. These probes both (1) sample the forward and reflected power to provide
feedback to the power supply and (2) tune the impedance of the circuit using
38
an automated microwave tuner incorporating three movable stubs. The quoted
response time of the tuner is within ±2.0 s. A separate user-interface controller
(Model: Daihen CMC-10A) is used for setting the control levels and outputs for
the tuner.
A series of WR–284 aluminum waveguides are used to position the PSMPT
assembly on the workbench, the assembly being secured at its full-height terminal
to this series of waveguides. Termination of the circuit at the reduced-height
terminal of the PSMPT assembly is made using a reduced-height sliding short,
which is simply a movable shorting block inside of a reduced-height waveguide.
It is used to partly “tune” the system by attempting to position the standing wave
maximum electric field at the nozzle tip. Positioning the standing-wave maximum
at the nozzle tip would require the sliding short to be positioned at a distance d =
λg /4 + n(λg /2) (where n = 0, 1, 2...) from the nozzle axis, neglecting any impedance
effects from the nozzle, aperture, or plasma. In practice, however, the best position
has been found to deviate slightly from this, and, with the benefit of the three-stub
tuner, a precise placement is not necessary to initiate and maintain the plasma.
A Faraday cage is used to shield lab equipment and personnel from microwaves
radiating from the aperture. For the spectroscopic tests reported here, this is
simply an aluminum-wire mesh screen designed to fit on the waveguide around
the waveguide aperture. Earlier testing utilized a larger sheet-metal cage that
encompassed the entire PSMPT assembly. The literature from other researchers
makes mention of similar cages (see, e.g., [38]).
Limited preliminary testing at 16 GHz has been performed using a similar setup,
albeit with components at reduced dimensions. It is necessary at this frequency
to use waveguides with smaller cross-sections—WR–62, with typical inner dimen-
sions of 1.58×0.79 cm—in order to maintain the TE10 mode. Aside from a sliding
short, no active tuning was implemented, which made sustaining a plasma more
difficult at this frequency. Microwave power was delivered by a traveling-wave
tube amplifier (Model: EEV NA4703).
39
3.1.3
Gas Supply System
Argon, being the gas primarily used for all testing to date and, in particular, for
the spectroscopic study of the torch, is supplied to the nozzle and secondary-flow
injectors from separate standard compressed-gas cylinders with a rated purity
of 99.997%. Standard 1/4” copper tubing is used to interconnect the gas-supply
components. Flow to the nozzle is controlled using a mass flow controller (MFC)
(Model: UNIT UFC-8100), which has been calibrated to output 5 SLM (standard
liter per minute) of nitrogen at 100% of flow capacity. Therefore, assuming standard
temperature and pressure (STP), at 100% capacity this flow controller outputs
approximately 7.1 SLM of argon. In order to sustain flow rates of this magnitude
through the nozzle—the aperture of which was less than 1 mm in diameter—an
increased back pressure, controlled by the cylinder regulator, is necessary since the
flow was choked through the nozzle. Control of the secondary flow is provided
by a separate MFC (Model: UNIT UFC 2050A), which has been calibrated to
output 30 SLM of nitrogen at 100% of flow capacity. Again assuming STP, at 100%
capacity this flow controller outputs approximately 42.4 SLM of argon; increased
back pressure is not necessary for the secondary-flow injectors because they are not
choked. The level of mass flow for each MFC was controlled by separate meters
(Model: UNIT URS-20), both of which can be set to a user-specified percentage of
total flow rate capacity of the corresponding MFC or set to “purge’” the MFC.
3.1.4
Optical Spectroscopy System
Light from the plasma is collected from a high-resolution spectrometer (Model:
OceanOptics HR4000). The spectrometer has a spectral collection range from
approximately 290–390 nm using a 2400-line holographic grating. A full-width
half-maximum (FWHM) instrumental resolution of approximately 0.53 Å is obtained with the fixed entrance slit width of 5 µm, a height of 1 mm, and a linear,
3648-element, CCD-array sensor (Model: Toshiba TCD1304AP). The detector was
enhanced with a quartz window (Model: OceanOptics UV4) and lens (Model:
OceanOptics L4) to increase detector efficiency. A mount was fabricated to support
a collimating lens (Model: OceanOptics 74-UV) for collecting the spectrum from
the plasma, with a lens diameter of 5 mm and lens length of 10 mm. The lens was
connected to a fiber optic cable (2 m in length, with a core diameter of 100 µm)
40
for use in the ultraviolet and visible spectral region. An SMA connector on the
opposite end of the cable allowed it to be directly connected to the spectrometer
unit in front of the entrance slit. The mount for the collimating lens permitted
two-axis movement of the lens, along directions parallel and perpendicular to the
nozzle axis. Data from the spectrometer was output directly to a computer that
utilized the spectrometer’s software, SpectraSuite, to control the spectrometer and
acquire the spectral data in real-time. The software allowed the user to control,
among other options, the integration time of the spectrometer, which proved useful
for acquiring the low-intensity OH signals.
As was mentioned previously, the preliminary goal of using optical emission
spectroscopy was to collect OH spectra to ascertain the gas (heavy particle) kinetic
temperature of the plasma. In order to acquire a strong enough signal of the
rovibronic bands of OH, water vapor was introduced into the primary gas flow
of the nozzle. This was accomplished by placing a small amount of water in the
gas supply line upstream of the nozzle MFC. As the argon passed through the
reservoir the flow would introduce water vapor into the stream. Heating tape
wrapped around the line aided the conversion from water to water vapor. OH
would then form as a product of the dissociation of the water vapor once in the
plasma jet. It should be noted that, using this method, the actual concentration
levels of water vapor in the primary argon flow are not known. The integration
time on the spectrometer was adjusted to tradeoff between a strong OH signal
strength (indicated most easily by examining the intensity around 306.5 nm) and
good temporal resolution.
3.2
3.2.1
Experimental Procedures
Nozzle Preparation
Testing with the PSMPT begins by first cleaning and inspecting the nozzle. Cleaning
involves wiping the nozzle tip with a petroleum-distillates–based solvent (Eagle
One “Original Nevr-Dull Wadding Polish,” www.eagleone.com) and acetone. A
visual inspection of the nozzle is performed to ensure that (1) the nozzle tip has
not suffered any damage during previous tests and (2) the nozzle aperture is not
clogged with dust or debris. As mentioned elsewhere, earlier testing of the PSMPT
41
demonstrated tip erosion and pitting. Obstructions to the nozzle aperture can be
removed using either fine-gauge steel wire or compressed air. The nozzle can
then be inserted into the waveguide by threading into the attached coupler and
the tip positioned such that it is approximately flush with the outside face of the
waveguide wall. This position, which appeared to work well at initiating and
sustaining plasmas, permits simple and reliable re-alignment of the tip.
3.2.2
System Assembly
The water reservoir in the nozzle gas supply line is replenished before reconnecting
the primary gas line. Next, both gas lines are secured to the nozzle and injectors.
The quartz tube is inserted into the waveguide aperture and secured using Kapton
tape (providing a low dielectric constant, the tape is used to help the tube resist
spinning and lifting out of the waveguide due to the force from the secondary
flow). Each of the gas flows are set to the desired level and allowed to run for a few
minutes to clear the lines. The shielding cage is secured to the waveguide and the
lens placed the desired distance from the nozzle along the plasma’s axial direction.
A schematic of the full experimental assembly can be seen in Figure 3.2.
Fiber
Optic
Mount
Air-Cooled
Dummy Load
Magnetron
Power Supply
& Tuner
Control
Circulator
Bidirectional 3-Stub
Coupler
Autotuner
Gas Cylinder &
Pressure Regulator
MFC
Gas Cylinder &
Pressure Regulator
MFC
H2 O
Waveguide
Network
PSMPT
Sliding
Short
to Nozzle
to Secondary Injectors
Computer
Spectrometer
Figure 3.2. Experimental assembly schematic.
42
3.2.3
Plasma Initiation and Data Collection
Initiation of the plasma then takes place as the microwave power is applied and
increased (impedance matching by the automatic tuner begins as soon as power
is applied). Once the plasma has been initiated, any “leaky” radiation can be
detected with a microwave survey meter (Model: Holaday Industries HI 1501). If
and when any significant changes or re-assembly are performed on the setup, a
new reading is taken with the meter to investigate any possible microwave leaks.
Any radiation must meet OSHA standards, which specify that the radiation level
should be kept below 10 mW cm−2 averaged over any possible 0.1-hour period [71].
With the Faraday cage in place, the microwave radiation levels have always been
determined to be well below this threshold. Only locations very near (i.e., within
a few centimeters of) the PSMPT register radiation levels of any significance to the
OSHA standard, but generally only less than 50% of the permitted value.
As the microwave power is brought to the desired forward power level the
plasma jet will appear to stabilize. The spectrometer, having been zeroed prior
to the test, is allowed to scan (but not store) the spectral data from the plasma,
displaying real-time results on a computer monitor. Further lateral alignment of
the collimating lens is then made with respect to the axis of the jet. This is a
necessary operation as the device invariably repositions slightly between tests, and
on some occasions the plasma appears to shift slightly off-axis from the nozzle.
Alignment is accomplished by moving the lens horizontally while simultaneously
examining the OH spectral signal intensity on the computer monitor. The final
lateral position of the lens is where the OH spectral intensity appears strongest,
presumed to be the center line of the plasma. The axial position of the lens was
taken as the distance between the center line of the lens and the nozzle tip face
(see Figure 3.3). If the OH spectral signal is too high or too low, the integration
time of the spectrometer can be adjusted accordingly. Once a suitable integration
time is chosen and the system stabilizes (typcially about one minute from plasma
startup) the software is set to collect and store the spectral signal. Each test consists
of a number of scans (typically 100). Tests were completed for each corresponding
power level and gas flow rate at four distinct axial positions along the plasma jet.
43
Plasma Jet
Incoming
Radiation
Collimator
Axial
Position
to Spectrometer
Nozzle Tip
Figure 3.3. Positioning of the optics system, indicating the position of the collimator along
the plasma’s axis.
3.3
Computational Procedures
3.3.1
Spectral Simulation and Fitting
The relative intensities of the optical emission spectral data were used to determine an estimate of the gas temperature from calculated spectra with known rotational temperatures. Preliminary results were obtained by fitting the experimental
spectral data to simulated OH spectra alone. Emission from the OH transition
A2 Σ+ → X2 Π, ever-present in discharges containing H2 O, was simulated using the
spectral database and simulation software, LIFBASE [72]. A more detailed comparison between experimental spectral data sought to fit both the OH as well as the
nitrogen second positive system (N2 SPS) simulated spectra. This latter, a promi-
nent system in nitrogen-containing gases, comprises the transition C3 Π g → B3 Πu ,
and was simulated using a code developed in MATLAB.
44
3.3.1.1
OH Without N2 SPS Overlap Correction
Initial temperature estimation proceeded by simulating the OH (A→X)(0, 0) band.
The band head is located near 306.4 nm, with intensity degradation of the OH
spectra towards longer wavelengths. Four prominent peaks located between 306–
310 nm are associated with the R1 , R2 , Q2 branch heads and the Q1 (3) line. LIFBASE
provides the ability to visually inspect experimental spectral data alongside simulated spectra for this transition. Viewing several scans, a best-fit line shape profile
was attempted visually. Relative spectral intensities were simulated for rotational
temperatures from 500–6000 K in increments of 5 K. A comparison was then made
between the two sets of data (experimental and simulated) between 306.3–308.3 nm.
At wavelengths below this region, OH spectral signals from satellite branches are
very weak; at wavelengths above this region, the interference from N2 signals
appeared stronger for some tests. This negated the use of the Q2 branch head.
The simulated spectral data were interpolated at each of the experimental wavelengths (the spectrometer’s wavelength-per-pixel parameter remained fixed). Both
data sets were normalized to the OH spectral peak associated with the P1 branch
head and the Q1 (3) line, between 308.2–308.3 nm. Finding the minimum of the
summed-squared difference (method of least squares) over the entire spectral region was used to determine a best fit with regards to both rotational temperature
at wavelength shift, 1 i.e.,


2 
X exp
In − Insim (Trot , shift)  .
∆(Trot , shift)fit = minTrot ,shift 
(3.1)
n
Results obtained in such manner provide a lower-bound estimate to the rotational
temperature owing to the structure of the overlap between OH and N2 SPS.
1
Though recently factory-calibrated, the experimental spectra always appeared to be shifted by
a small amount. Typical calculated values throughout this report lead to the conclusion that the
spectrometer is off in the range of 0.3–0.5 Å. That is, a line with a true position at 300.04 nm would
be measured near 300.00 nm.
45
3.3.1.2
OH with N2 Band Overlap Correction
As elucidated in Chapter 4, merely assuming an that accurate temperature could
be deduced on the basis of the OH signal alone would not suffice. The most
prominent and likely source of spectral interference with OH was the N2 SPS, the
nitrogen having been drawn in from the surrounding atmosphere. Principally, the
interference involved the band sequence ∆v = −1, with the most prominent band
head located near 315.9 nm. As OH is degraded to longer and the N2 SPS to shorter
wavelengths, the two signals overlapped quite strongly between 306–316 nm.
Putting together the relevant theory presented in Chapter 2, a MATLAB code
was developed to simulate the N2 SPS. The methodology follows that presented
by several authors, most notably in a report by Laux [73]. Both main and satellite
branches were included in the calculations. N2 SPS involves a 3 Π g –3 Πu transition.
Therefore, three spin-splitting substates exist for each branch. Each of these substates is further separated into a Λ-doublet leading to rotational lines of differing
parity. Standard parity–symmetry selection rules were applied. Rotational levels
up to J0 = 170 were calculated. Line positions were calculated by diagonalizing the
effective Hamiltonian matrix elements with corresponding molecular parameters
provided by Roux et al. [74, 75].2 Owing to the refractive index of air, a correction
based on Edlén [76] was applied to each of the lines.
Hönl–London factors, S J0 ,J” , were calculated from the expressions located in
Table 3.8 of Kovács [77] for the case of transitions intermediate between Hund’s
cases (a) and (b). Whiting et al. [78] list several corrections to these factors, along
with the appropriate normalization factor. Schadee [79] provides expressions for
the Hönl–London factors of first rotational lines. Owing to the fact that N2 is
homonuclear, an alternation factor was applied based on the upper-state parity–
symmetry combination. Electronic transition moment values were taken from the
calculations of Laux [80]. A Boltzmann distribution was assumed between individual rotational and vibrational levels. Term values for the upper-state rotational
and vibrational contributions were drawn from the calculated eigenvalues of the
effective Hamiltonian.
2
An apparent typographical error in Table VI of [75] lists Teff = 1750.187—the 0 and 5 should be
switched.
46
At the time of testing, a line source was not available to accurately provide a
line shape profile for the spectrometer. A FWHM resolution was estimated for the
spectrometer to be approximately 0.53 Å using information available from the manufacturer [81]. Also, a single test operating at 16.15 GHz and using a stainless-steel
nozzle provided three chromium peaks at 357.87 nm, 359.35 nm, and 360.53 nm.3
A pseudo-Voigt fit to these peaks revealed a FWHM of 0.80 Å. Doppler and collisional broadening widths for OH and N2 are typically an order of magnitude or
more smaller than this instrumental width. Therefore, all broadening was treated
as instrumental in origin. Since an accurate instrumental line shape could not be
deduced, and accurate data for broadening parameters was not available, a best fit
was made using a pseudo-Voigt profile. Following Whiting [82] and Olivero [83],
the pseudo-Voigt profile parameters are defined as
wL
η=
wV
and
wL
wV =
+
2
q
w2L + 4w2G ,
(3.2)
where wL and wG are the FWHM of the Lorentzian and Gaussian components,
respectively. A single experimental test was used to identify a best fit to this
profile, using the manufactured-specified minimum of 0.53 Å as a lower bound
and the chromium peak FWHM of 0.80 Å as an upper bound. The results of this
gave a Lorentzian fraction of η = 0.35 and a pseudo-Voigt FWHM of wV = 0.60 Å.4
The form of the pseudo-Voigt function follows that used by Biloiu et al. [84].
Both OH and N2 SPS were simulated at rotational temperatures ranging from
1000–7500 K (in 100 K increments) and vibrational temperatures ranging from
2000–12000 K (in increments of 500 K). The results of this modeling procedure
for N2 SPS for three representative cases are shown in Figure 3.4.
3
While the presence of such atomic lines with minimal background intereference could provide
an estimate of the electron temperature in theory, such could not be reliably determined. A Boltzmann plot did not reveal a strong linear slope, most likely due to the close spacing between the
energy levels of the peaks.
4
LIFBASE requires different parameters: η = wL /(wL +wG ) and total instrumental resolution, wI =
wL + wG . For the sake of calculations, all broadening was assumed instrumental (i.e., “unchecked”
broadening parameters in LIFBASE). The corresponding LIFBASE values are η = 0.3053 and
wI = 0.6878 Å.
47
N2 (C → B)(∆v = −2)
N2 (C → B)(∆v = −1)
N2 (C → B)(∆v = 0)
N2 (C → B)(∆v = 1)
N2 (C → B)(∆v = 2)
Relative Intensity
2000 K
4000 K
6000 K
300
310
320
330
340
350
360
370
380
Wavelength (nm)
Figure 3.4. Relative intensity of the simulated N2 SPS (C→B) band structure at Tvib = 12000 K for three different rotational
temperatures. Above are the band sequences: lines indicate the leading band head; arrows indicate direction of intensity
degradation.
48
The simulated spectra for both OH and N2 were interpolated at each of the
spectrometer wavelengths for various wavelength shifts in the spectral region
studied. This region, spanning 305–318 nm, included the OH (A→X) (0, 0) and
(1, 1) bands and the N2 (C→B) (∆v = −1) bands and the tail of the (∆v = 0) bands.
As with the OH-only procedure, the fitting routine sought to minimize the sum of
squared differences between the (adjusted) experimental spectral profile and the
simulated spectra profile. However, two separate fittings were needed, one for
each species. Since the relative contribution of either species was unknown, the
fitted profiles were iteratively deduced. The iteration proceeded by attempting to
fit each of the species to the experimental data separately. Each updated iteration
presented the current species with an adjusted experimental spectrum to be fitted
with. This adjusted spectra was formed by subtracting out a combination of the
previously-calculated opposing spectra’s profile and a residual from the actual
experimental data.
OH fits to determine the rotational and vibrational temperatures were computed
separately. When fitting, only specific wavelengths were chosen to compare the
experimental against the simulated spectra. These wavelengths were likely to provide the most reliable information on rotational and vibrational temperature based
on their relative isolation and variability with changing temperature. For the OH
rotational temperature, the (0, 0) band was fitted near the peaks associated with
the R1 (7–11), R1 (6, 12), R1 (5, 13), R1 (14)/R2 (9–11), Q1 (4), Q1 (6), and Q2 (1–4)/Q1 (7)
lines. Fitting the OH vibrational temperature used the peaks associated with the
R1 (7–11) lines in the (0, 0) band and the R1 (2, 13)/R2 (8–10), P1 (1)/Q1 (3), Q1 (5), Q2 (5),
and Q1 (9)/P2 (2) lines in the (1, 1) band. The N2 SPS was fitted to both the rotational
and vibrational temperatures simultaneously since the rotational lines were insufficiently resolved against the OH lines. Fits were made to the wavelengths near the
peaks associated with the (1, 0), (2, 1), and (3, 2) band heads, along with the region
between 305–306 nm (with minimal interference expected from OH there).
In order that a fit could be found by this method, the experimental spectra
was required to be free from any saturated intensities within the region studied
(305–318 nm). Typically, within this region, saturation would first appear near the
region (1, 0) band head of N2 SPS. If the spectrum was saturated in this region,
then a fit was only attempted to the OH (0, 0) band rotational temperature, similar
to the method presented in Section 3.3.1.1.
49
3.3.2
Electromagetic Modeling
A multi-physics, finite-element software package, COMSOL Multiphysics, was
used to model the electromagnetic fields of the PSMPT. The software enables the
user to set up a physical model as well as create and modify finite-element meshes
in an easy-to-use graphical user interface compatible with Windows platforms. The
benefit of the software is its ability to couple different sets of physical equations
using various built-in modules (e.g., electromagnetic, fluids, heat transfer). For
preliminary modeling, however, only the RF (electromagnetic) Module was used as
attempts were made to understand the field-shaping structure of the PSMPT. Using
this RF Module allows the user to define the boundary conditions and material
properties. For all of the models run, the propagating medium was assumed
to be a vacuum and the walls of the waveguide and the nozzle were simulated
as perfect electric conductors (no resistive power loss due to skin depth effects).
Power was introduced using standard “port” boundary condition, specifying a TE10
mode, which also effectively absorbs reflected power. Radiated power from the
waveguide aperture was intercepted by a perfectly matched layer and/or scattering
boundary a distance away from the waveguide aperture.
50
Chapter 4
Results and Discussion
4.1
Preliminary and Qualitative Results:
Experimental and Operational Issues
Early testing sought an assembly that reliably produced plasmas but encountered
several difficulties in the design and operation of the PSMPT. Broadly speaking,
these difficulties fall into two categories: (1) impedance matching and power coupling and (2) nozzle erosion. We discuss here some of these issues, while further
comments elaborating on these and other issues will be presented in Chapter 5.
4.1.1
Impedance Matching and Power Coupling
4.1.1.1
Sliding Short Position
In determining the best position for the sliding short, a range was found within
which the plasma could be initiated and sustained successfully, as demonstrated
in Figure 4.1. Coupling efficiencies—the ratio of total power absorbed (across the
PSMPT and microwave transmission network) to applied forward power—less
than 30% are indicative of situations where no plasma was present, and so the
power can be considered lost entirely to the waveguide components and radiation
leakage. Also, in earlier tests—and in the data seen in this figure—a different
sliding short design was used that often suffered from arcing damage, leading to
poor power coupling to the plasma. This problem was overcome with a different
sliding short design that eliminated arcing (see Section 4.1.1.5) and allowed for
better impedance matching and power coupling.
1.0
0.9
Coupling Efficiency
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.10
0.15
0.20
0.25
0.30
d/λ g
Figure 4.1. Coupling efficiency vs. normalized sliding short
distance from nozzle axis, d/λg . The markedly lower values
of coupling efficiency indicate a plasma that would not
initiate successfully.
Regardless of the shorting design used, or differences between microwave
source and tuning, the best sliding short position was not necessarily at d/λg = 0.25
(the location of a standing-wave maximum in a standard waveguide). Different
configurations required adapting the nozzle-to-sliding short distance. The effect
of sliding short position was simulated using COMSOL Multiphysics and compared to experimental data. Results from these simulations indicate the degree of
sensitivity to short position on the PSMPT. Figure 4.2 presents the electric field
strength at a position just above the center of the nozzle tip. The standing-wave
pattern maximum should repeat in multiples of d/λg = 0.50 and so it is assumed
the pattern in the graph would repeat, as well.
Note that the electric field strength presented in Figure 4.2 should not be taken
as the maximum field near the tip. Rather, it is representative of the strength of the
electric field near the tip. The maximum likely occurs near the edge of the nozzle tip;
however, solutions very near sharp edges in finite-element electromagnetic analysis
are often not valid. It is also worth noting that the “port” boundary condition for
the input of microwave power is designed to absorb all power reflected back to
the port. In the real scenario involving the PSMPT with automatic tuning, power
52
reflected from the plasma, nozzle, and sliding short to the source may be reflected
back again into the PSMPT. Even though these results don’t provide us with
knowledge of the true field strength, we can at least elucidate the relative behavior
when changing waveguide height and sliding short distance.
Figures 4.3–4.5 present contours of the electric field strength at the center plane
of the waveguide corresponding to the conditions in Figure 4.2. The four plots
in each figure represent positions near d/λg = 0.25, d/λg = 0.50, and the two
peaks located immediately in either direction from the d/λg = 0.50 postion, as
seen in Figure 4.2. Each of the contour plots show electric field strength within
the waveguide, around the nozzle, through the waveguide aperture, and into
the space above the waveguide. The waveguide walls and nozzle are the empty
(white) areas of the plots. Microwave power is input from the left side of each
plot, while the sliding short wall is positioned on the right side of each plot. These
results in Figure 4.2 and Figures 4.3–4.5 indicate that, when compared to full-height
(standard) and 3/4-height waveguides, using a half-height waveguide makes the
electric field strength near the nozzle tip less sensitive to sliding short position.
Electric field strength (V/m)
×104
4.0
1/2-height
3/4-height
Full-height
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
d/λ g
Figure 4.2. Simulated relative electric field strength near
the nozzle tip vs. normalized sliding short positions for
full-, 3/4-, and half-height waveguides.
53
E (V/m)
2500
2000
1500
1000
500
0
Figure 4.3. Contours of electric field strength at the waveguide centerline in a full-height waveguide. Microwave input (300 W)
from the left into waveguide. Sliding short right most side of each plot. Free space above waveguide. Nozzle and upper
waveguide wall seen as empty (white) areas. From the top, plots correspond to normalized nozzle-to-sliding short distances,
d/λg , of 0.249, 0.476, 0.498, and 0.519. Values higher than 2500 V/m are plotted at that same level.
54
E (V/m)
2500
2000
1500
1000
500
0
Figure 4.4. Contours of electric field strength at the waveguide centerline in a 3/4-height waveguide. Microwave input (300 W)
from the left into waveguide. Sliding short right most side of each plot. Free space above waveguide. Nozzle and upper
waveguide wall seen as empty (white) areas. From the top, plots correspond to normalized nozzle-to-sliding short distances,
d/λg , of 0.249, 0.465, 0.498, and 0.541. Values higher than 2500 V/m are plotted at that same level.
55
E (V/m)
2500
2000
1500
1000
500
0
Figure 4.5. Contours of electric field strength at the waveguide centerline in a half-height waveguide. Microwave input (300 W)
from the left into waveguide. Sliding short right most side of each plot. Free space above waveguide. Nozzle and upper
waveguide wall seen as empty (white) areas. From the top, plots correspond to normalized nozzle-to-sliding short distances,
d/λg , of 0.249, 0.432, 0.498, and 0.573. Values higher than 2500 V/m are plotted at that same level.
56
4.1.1.2
Automatic Tuning
Tuning—aside from the sliding short—was initially made possible only by means
of a manual three-stub tuner. Tuning speeds were inadequate since the stubs
were adjusted separately with dials, which were cumbersome and slow. With a
fluctuating load (i.e., the plasma), a method for accurately deploying the stubs to
an optimum depth was not obtainable, and tuning became a process of elimination
over multiple tests. Even if a match were to be found in this way, it would need
to be adjusted for different situations (e.g., plasma initiation versus sustainment)
and conditions (e.g., power and flow rate). Without appropriate tuning creating
an apparent Z0 -load at the tuner, the measured forward power would fluctuate
for the same input current to the magnetron power supply. This was particularly
troublesome prior to plasma initiation as the plasma load was not established yet.
Changing the stub positions quickly and accurately became a necessary capability.
Using automatic tuning (i.e., a computer-controlled three-stub tuner) the plasma
could be sustained and remain stable over the course of several minutes. Coupling efficiencies increased dramatically with automatic tuning. We take applied
forward power to be synonymous with power lost in the plasma and through
other mechanisms (e.g., waveguide heating, radiation leakage). In all of the recent
spectroscopic tests, the reflected power was quite low, with coupling efficiencies
fluctuating within 95–100%. These levels would often fluctuate between tests, even
with similar parameters, due to variations in automatic stub positioning. The use
of automatic tuning, in general, led to much less pronounced instabilities in the
plasma and allowed the plasma to operate at significantly reduced power levels.
Visible tuned plasma lengths were generally less than 3 cm, while untuned lengths
could easily exceed 6 cm. These effects can be seen in Figure 4.6.
4.1.1.3
Nozzle Tip Position
Tests were made to determine the optimum position of the nozzle tip with respect
to the waveguide wall with no active tuning. Testing this effect was done using an
unloaded (i.e., no plasma) PSMPT with both a network analyzer (very low power)
and using forward and reflected power measurements with the magnetron in operation (high power). Results indicate that positioning the nozzle tip slightly outside
of the waveguide may allow for a decrease in reflected power and, therefore, an in-
57
Figure 4.6. Demonstration of plasma jet size and stability
at high power (> 1 kW) without automatic tuning (two left
images) and at moderate power (< 500 W) with automatic
tuning (right). Length scales in the two images on the left
are about five times larger than in the image on the right.
crease in absorbed and radiated power. An example can be seen from the coupling
efficiency data presented in Figure 4.7. These data were obtained for a situation in
which the PSMPT was unable to initiate a plasma, with microwave power being
applied; the absorbed power is then assumed to be lost to the waveguide components and through radiation (hence the low coupling efficiency values). Note
in Figure 4.7 that the forward power levels change for the same magnetron input
current since the waveguide assembly is not properly matched to the source.
As presented in Section 3.2.1, however, aligning the tip with the outside face of
the waveguide wall allows for easier and more consistent alignment between tests.
Active tuning appeared to mitigate the effectiveness of changing the tip position,
as well. The performance of the PSMPT suffered primarily not from poor nozzle
tip placement, but from difficulty in both initiating and sustaining the plasma with
the earlier magnetron and tuner, as mentioned in Section 4.1.1.2. Once a more
efficient magnetron and tuner (Daihen models) were in place, concerns over the
precise position of the nozzle tip did not seem relevant in the overall operation of
the PSMPT.
4.1.1.4
Power Losses
Most of the power not registered as reflected back towards the source was assumed
to be absorbed by the plasma. In reality, microwave power is likely lost through
other mechanisms, as well.
58
0.55
0.50
Coupling Efficiency
0.45
0.40
0.35
0.30
0.25
0.20
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
Tip Distance (mm)
Figure 4.7. Coupling efficiency varying the nozzle tip position, for a single magnetron current level, no plasma
present. Distances are measured from the outer wall of
the waveguide to the end of the nozzle; positive values indicate the nozzle tip protruding outside of the waveguide.
For example, the sliding short does not have a perfect electrical seal nor a perfect
face finish, such that there are likely to be some losses in the device. This effect
was quite notable with the use of an earlier sliding short design that suffered from
arcing between the shorting block and the waveguide wall. After long tests, the
sliding short, as well as the PSMPT waveguide, is notably warm to the touch. This
is at least partly due to the size of the waveguide with respect to the microwave
frequency, with higher power density leading to an increase in resistive heating
losses as compared to larger waveguides.
As mentioned previously, a shielding cage was utilized to protect equipment
and bystanders from any microwave radiation emitted through the waveguide
aperture and not absorbed by the plasma. Originally, a large aluminum cage
was used that surrounded much of the PSMPT and waveguide assembly and that
allowed for measuring the radiation levels in the cage. However, the cage was
cumbersome to use, and therefore a smaller cage was employed. While cage size
had no noticeable effect on the plasma performance, the smaller cage may act as a
sort of cavity, and its ultimate effect on total power absorbed in the plasma needs
to be understood. An image of the small cage with the plasma in operation can be
seen in Figure 4.8.
59
Figure 4.8. Aluminum mesh microwave shielding cage
mounted to the waveguide around the plasma.
Measurement of radiation emanating from the waveguide aperture was attempted using a small Yagi-type microwave antenna, in tandem with a spectrum
analyzer. The antenna was contained within the larger shielding cage for these
measurements. While the net power flow out of aperture was not able to be determined with this setup, some indication of the relative amount of power absorbed
by the plasma was received. This was done by comparing the measured radiation
with a plasma to that of the situation with metal post used in place of the nozzle and plasma. The situation with the metal cylinder is assumed to indicate the
highest level of coupling possible for the aperture. From Figure 4.9, we see that
the metal cylinder doubled the radiated power for the same sliding short position.
Another curve demonstrates the effect of having a fixed short, improperly tuned,
on the radiation. For these tests, the applied forward power was at 600 W with
reflected power levels less than 10 W. Kirichenko et al. [52] presented similar results for their microwave torch discharge when using a metal cylinder in place of
the plasma. Figure 4.10 compares measured and simulated results for a cylindrical
post by varying sliding short position. The simulated results, also produced using COMSOL Multiphysics, show the same trend versus normalized sliding-short
position as seen in Figure 4.2.
60
1.0
Plasma
Post, w/ sliding short
Post, w/o sliding short
0.9
Normalized Power
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
2.435
2.436
2.437
2.438
2.439
2.440
2.441
2.442
2.443
2.444
2.445
Frequency (GHz)
Figure 4.9.
Normalized intensity measurements of
power radiated from the PSMPT using a spectrum analyzer for three different loadings. Forward power = 600 W.
Reflected power < 10 W.
1.0
Measured
Simulated
0.9
Normalzied Power
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
d/λ g
Figure 4.10. Measured and simulated radiated power vs.
normalized sliding short position, d/λg , for a cylindrical
metal post in place of a plasma.
4.1.1.5
Microwave Breakdown Control
Early testing necessitated use of a striking aid in order to initiate the plasma.
It consisted of a retractable copper wire connected to a solenoid, which, when
activated, extends near the tip of the nozzle during microwave operation, increasing
the local electric field there. The striking aid proved useful but cumbersome, as it
61
required quite good positioning of the wire near the tip but needed to fully retract
out from the waveguide aperture. On many occasions, initiating the plasma proved
more difficult—several passes of the wire were needed to support the breakdown
electric field. Later tests have avoided using this method with better impedance
matching in the microwave transmission system.
Five locations within the PSMPT have seen undesirable microwave breakdown:
• the nozzle–waveguide joint;
• the sliding short (already mentioned);
• the space between the nozzle tip and waveguide aperture wall;
• the secondary flow injector ports; and
• the plasma to exterior surfaces.
In order to facilitate the slide-fit for the nozzle, the smaller waveguide aperture did
not provide a perfect electrical seal between the aperture and the nozzle. There
remains the possibility of drawing “arcs” from the plasma or nozzle tip to other
surfaces (in particular, to the waveguide aperture wall) with sufficient incident
power and improper tuning of the circuit.
4.1.2
Nozzle Erosion
4.1.2.1
Description of Problem
The PSMPT is electrode-less in the proper sense of the term; i.e., there is no corresponding set of anodes and cathodes drawing a true current. Even so, nozzles
used with the current PSMPT appear to be no exception. Somewhat troubling is
the seemingly scant attention given to the erosion of SEP nozzles in the literature.
Most of the literature describing results with the TIA (see Section 1.3.3) make no
mention of erosion problems using argon. Iordanova et al. [85] do note that using
hydrogen with the TIA greatly increased erosion of the nozzle. Erosion of the
nozzle may be arising from a number of possible sources:
• proximity of the plasma jet to the nozzle and subsequent heating;
• additional heating from local surface currents on the nozzle from the microwave field;
• oxidation from air entrainment and increased surface reactivity; and/or
62
• ion sputtering of the surface.
Several unwanted effects due to erosion include:
• deterioation of the nozzle tip, such that:
– the rim of the nozzle was no longer circular with clean, sharp edges;
– the nozzle aperture was no longer circular, tending to open up to a larger
effective area;
– the nozzle face was no longer smooth but became pitted; and
– the nozzle face would not present square to the direction of flow.
• aberrations and distortions to the flow of the plasma jet, both in direction and
structure, due to the aforementioned nozzle tip geometry variations;
• limited individual testing duration, so as to preserve the lifetime of the nozzle;
• copper contamination of the jet, visible as intermittent green “flashes” in the
plasma; and
• nozzle inoperability, as the plasma would become more difficult to initiate
and sustain.
Two low-resolution images, seen in Figure 4.11, demonstrate the progression of
erosion at the nozzle tip. Pitting of the nozzle tip face, melting of the material,
discoloration, and a change in nozzle aperture shape are apparent.
Figure 4.11. Demonstration of erosion using a water-cooled
nozzle with no secondary flow—unused nozzle (left) and
used nozzle (right).
63
A majority of these issues have not surfaced with the use of secondary gas
flow and diligent cleaning of the nozzle tip. These combined procedures led to
a significant reduction in visible deterioration and discoloration of the nozzle tip,
and may have also contributed to more consistent plasma initiation. It would
be beneficial to have a better understanding of all of the mechanisms leading to
erosion of the nozzle, to help determine future design of the nozzle geometry,
nozzle material, and secondary assist gas type.
4.1.2.2
Remedies for Erosion
As with any high-temperature plasma torch, restrictions are placed upon the choice
of materials for electrodes. To date, OFHC copper has been primarily used with
the PSMPT due to its high conductivity, low work function, and machinability.
The melting temperature of copper is high, approximately 1350 K, though below
typical gas temperatures in the plasma downstream of the nozzle tip. Other alloys
used included beryllium copper and an OFHC copper with tungsten-coated nozzle
tip, but they provided no noticeable benefit. Thus far, OFHC copper resists erosion
well when used in conjunction with a secondary assist gas. Plasma initiation was
generally more consistent with the OFHC and beryllium copper nozzles. Only the
OFHC copper has been continued due to concerns regarding the safety of beryllium
(considered a safety hazard in particulate form and subsequently inhaled into the
lungs). The tungsten-coated nozzle also exhibited faster erosion than is typically
seen with simple secondary flow assist, perhaps due to poor bonding. It is thought
that the OFHC alloy ensures lower oxidation levels, which also would appear to
assist in plasma initiation.
Attempts at reducing nozzle erosion due to possible overheating involved testing two water-cooled nozzle designs. These designs were both implemented prior
to the secondary assist gas implementation. Electrodes in thermal plasma applications often use some level of water cooling to reduce erosion (see, e.g., [86]).
However, neither nozzle demonstrated any significant effect on reducing erosion.
It may have been that the water flow was not high enough or close enough to the
nozzle tip; even still, the cumbersome and expensive task of constructing watercooled nozzles does not appear necessary. Design and manufacture of a nozzle
64
with water cooling supplied near the nozzle tip has proven difficult due to the
size of the nozzle and nozzle aperture and the need to have a pressurized gas line
running through the nozzle. With the appropriate gas flow and care for the nozzle,
erosion was greatly reduced without the need for a complicated nozzle design.
One of the approaches that more significantly reduced erosion is the introduction of secondary flow around the nozzle. Without a secondary assist gas, the
entire nozzle tip would quickly erode and testing had to be done for short durations. More recent testing used secondary assist gas and the nozzle tip was often
cleaned using a metal polish. A better knowledge of the level of heat transfer near
the tip is sought. What decrease in the influx of oxygen and nitrogen provided to
the nozzle tip is not clear. As shown later, nitrogen was clearly being entrained
within 1 mm of the nozzle tip. Even with the presence of a secondary assist gas
deterioration of the nozzle tip still occurs (although, at a much reduced rate). For
the time being, it would appear that a secondary gas flow is crucial to eliminating
erosion.
65
4.2
4.2.1
Spectroscopic Determination of Gas Temperature
Rotational Temperature Without N2 Band Overlap
Correction
Obtaining OH rotational temperatures was the preliminary goal in ascertaining the
gas temperature inside the plasma. The spectral data were first analyzed assuming
limited spectral interferences from other sources. Contour plots of these results are
presented in Figure 4.12. The plots show the variation of rotational temperature
with secondary flow rate and nozzle flow rate for three different power levels and
four different axial positions. Note that the spectra appeared too weak at 225 W
and 11.1 mm so that no data are available at this location in the figure.
Secondary flow rate demonstrated minimal effect on rotational temperature,
while nozzle flow rate (NFR) shows more influence. The strongest influence on
temperature based on this chart appears to be the axial position. Temperatures
reported near the nozzle were in the range 800–150 K—exceptionally lower than
those further downstream (typically > 2500 K). It was determined that as we move
upstream in the plasma, stronger spectral interferences were observed, as shown
in the next section. These temperatures, though not likely to be accurate, do give
some indication as to a lower bound on the actual OH rotational temperature.
4.2.2
Rotational Temperature With N2 Band Overlap Correction
A more thorough examination of the spectral emission data was required. Examining all emissions from 290–390 nm provided a preliminary understanding of
the PSMPT and a basis for comparison between the various torches. With the
known reactant gases available to the plasma (argon with water addition, flowing
into an open air atmosphere), the most likely candidates for molecular emission
over the spectrometer range include OH (A→X), NH (A→X), N2 SPS (C→B), and
the singly-ionized nitrogen first negative system (N+2 FNS) (B→X). Representative
spectra collected under similar operating conditions are presented in Figure 4.13.
66
Power (W)
300 W
225 W
Trot
(K)
400 W
42.4
6000
42.4
11.1 mm
21.2
21.2
42.4
6.4
0.7 1.8
42.4
6.4
0.7 1.8
42.4
21.2
21.2
5500
Secondary Flow Rate (SLM)
5000
3.5
5.3
7.1
3.5
5.3
7.1
4500
8.0 mm
6.4
0.7 1.8
42.4
3.5
5.3
7.1
4000
21.2
6.4
0.7 1.8
42.4
3.5
5.3
7.1
6.4
0.7 1.8
42.4
3.5
5.3
3500
7.1
4.8 mm
21.2
21.2
21.2
6.4
0.7 1.8
42.4
6.4
0.7 1.8
42.4
6.4
0.7 1.8
42.4
3000
2500
3.5
5.3
7.1
3.5
5.3
7.1
3.5
5.3
7.1
2000
0.8 mm
21.2
21.2
6.4
0.7 1.8
3.5
5.3
7.1
1500
21.2
6.4
0.7 1.8
3.5
5.3
7.1
Nozzle Flow Rate (SLM)
6.4
0.7 1.8
1000
3.5
5.3
7.1
500
Figure 4.12. Contour plots of OH Trot vs. NFR and SFR for various powers and positions. No accounting for N2 SPS interference
made here.
67
Without a complete picture of the reactant spatial concentrations or an accurate
collisional–radiative model, deciphering a complete detailed structure of sources
of emission was not possible. Each of the aforementioned molecular emission
species that were able to be identified have several overlapping bands throughout
the entire spectrometer range, such that no one of the emission species is free from
interference from any of the other. Additional difficulty arises when considering
the integration time used for each of the tests. The original intent, when the data
were collected, was to ensure that the strong OH signal was achieved while limiting
the overall integration time for better temporal resolution. For several of the test
conditions—notably, at locations closer to the nozzle aperture—the intensity range
was saturated over much of the spectrometer’s wavelength range. It is believed
that the majority of this saturated interference was from the N2 SPS. This effect can
be seen clearly in Figure 4.13. In general, the total emission intensity across the
entire spectrum decreased away from the nozzle tip.
A source for modeling the NH (A→X) spectrum was not available. That fact—
coupled with the much stronger N2 SPS ∆v = 0 bands in the NH vicinity—meant
that NH could not be used as a thermometric species. Also, the N2 SPS bands with
∆v = 1, 2 could not be compared to experimental data because of the lack of knowledge of contributions from the N+2 FNS spectrum overlapping these bands. In the
end, this left several of the test conditions unusable for temperature measurement.
Though N+2 FNS bands are likely to be present, they do not appear to be influential
in the region under examination.
When examining the experimental spectra, it appeared that there may have been
some background interference not accounted for by any of the indicated species.
This is after accounting for the reference background spectrum collected by the
spectrometer. Both the simulated spectra of OH and N2 SPS near 317 nm indicate
that the value of their combined relative emission intensity should be at or near
zero there over most of the temperature range. However, when examining the
experimental spectra there appears to be always some level of relative intensity
above zero at that wavelength position. If there is another source of interference,
it cannot be deduced from an inspection of the experimental spectra, though the
interference appears to be quite small. Following the work of others (see, e.g.,
68
N+2 (B→X)(∆v = −2)
N+2 (B→X)(∆v = −3)
N+2 (B→X)(∆v = 0)
N2 (C→B)(∆v = 1)
N2 (C→B)(∆v = 2)
NH(A→X)(∆v = 0)
OH(A→X)(∆v = 0)
N2 (C→B)(∆v = −1)
N2 (C→B)(∆v = 0)
Relative Intensity
N2 (C→B)(∆v = −2)
N+2 (B→X)(∆v = −1)
300
310
320
330
340
350
360
380
370
Wavelength (nm)
N+2 (B→X)(∆v = −3)
N+2 (B→X)(∆v = −2)
OH(A→X)(1, 1)
OH(A→X)(0, 0)
N2 (C→B)(3, 2)
N2 (C→B)(2, 1)
N2 (C→B)(1, 0)
Relative Intensity
N2 (C→B)(4, 3)
N2 (C→B)(∆v = 0)
306
308
310
312
314
316
Wavelength (nm)
Figure 4.13. Experimental spectrum from full spectrometer output (top) and within the
fitting region (bottom). NFR = 0.7 SLM. SFR = 6.4 SLM. Forward power = 300 W. Lens
center position above nozzle tip: (blue) = 0.8 mm, (green) = 4.8 mm, (yellow) = 8.0 mm,
(red) = 11.1 mm. Candidates for molecular systems in this range shown at top of graphs—
vertical lines show the most prominent band head of a sequence; arrows show primary
direction of intensity degradation.
69
Nassar et al. [87]) it was decided to subtract out an additional background spectra.
This spectra was calculated by locating the minimum intensity value from 305–
306 nm and at 317 nm and forming a linear background spectra based on these
minima within 305–318 nm.
Figure 4.14 presents an example of the results of fitting the OH and N2 SPS
spectra to the experimental data. The result of the fitting procedure demonstrates
the need to fit both OH and N2 SPS to obtain an accurate result. Residuals between
the experimental and simulated spectra show the most deviation between 310–
312 nm.
The test parameters primarily consisted of sets of the following values:
lens position (mm) = 0.8, 4.8, 8.0, 11.1
applied forward power (W) = 225, 300, 400
nozzle flow rate (SLM) = 0.7, 1.8, 3.5, 5.3, 7.1
secondary flow rate (SLM) = 6.4, 10.6, 21.2, 31.8, 42.4
Certain test parameter sets are not available for various reasons. Notably, those
involving the lens position at 11.1 mm and power of 225 W were not collected as the
OH spectra was too weak for these conditions. For studies involving simultaneous
OH and N2 SPS fits, no temperature data are available involving a lens position of
0.8 mm due to saturation, nor for some tests at higher nozzle flow rates at other
positions. The results of tests with successfully combined OH–N2 temperature
measurements (i.e., no saturated spectra) are presented in Table 4.1.
70
1.0
Experimental
0.9
Simulated
0.8
Relative Intensity
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
306
308
310
312
314
316
314
316
Wavelength (nm)
Residual
0.3
0.2
0.1
0.0
−0.1
OH
Relative Intensity
N2
306
308
310
312
Wavelength (nm)
Figure 4.14. Combined fitted simulated spectra with experimental spectra (top), residual
(middle), and individual OH and N2 SPS that make up the combined spectra (bottom).
NFR = 0.7 SLM. SFR = 6.4 SLM. Forward power = 300 W. Position = 8.0 mm.
71
Table 4.1. Averagea OH and N2 SPS rotational and vibrational temperatures for
various forward power levels, nozzle flow rates, and lens positions.b
Forward Nozzle Flow
Power (W) Rate (SLM)
4.8 mm
8.0 mm
11.1 mm
225
0.7
1.8
3.5
5.3
7.1
3131 ± 69
2779 ± 59
—
—
—
3104 ± 59
2860 ± 65
2583 ± 62
—
2466 ± 59
—
—
—
—
—
300
0.7
1.8
3.5
5.3
7.1
2851 ± 58
—
—
—
—
3276 ± 61
2980 ± 55
2718 ± 59
—
—
400
0.7
1.8
3.5
5.3
7.1
2877 ± 74
—
—
—
—
3398 ± 59
3052 ± 60
2700 ± 50
—
—
Forward Nozzle Flow
Power (W) Rate (SLM)
a
b
Nozzle Flow
Rate (SLM)
OH Trot (K)
N2 Trot (K)
4.8 mm
8.0 mm
11.1 mm
0.7
1.8
3.5
5.3
7.1
3243 ± 60
3095 ± 57
—
—
—
3833 ± 89
2945 ± 63
2657 ± 59
—
2591 ± 56
—
—
—
—
—
3099 ± 62
2985 ± 60
2770 ± 66
2722 ± 63
2691 ± 58
0.7
1.8
3.5
5.3
7.1
3282 ± 54
—
—
—
—
4393 ± 75
3084 ± 58
2787 ± 57
—
—
4663 ± 180
3643 ± 101
2761 ± 72
2547 ± 61
2447 ± 62
3294 ± 58
3183 ± 59
2802 ± 59
2825 ± 58
2713 ± 60
0.7
1.8
3.5
5.3
7.1
3389 ± 58
—
—
—
—
4621 ± 88
3358 ± 61
3011 ± 56
—
—
5075 ± 124
3954 ± 86
2889 ± 61
2795 ± 64
2698 ± 54
Nozzle Flow
Rate (SLM)
OH Tvib (K)
4.8 mm
8.0 mm
11.1 mm
225
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
5278 ± 282
5728 ± 358
9500 ± 597
—
10255 ± 332
—
—
—
—
—
300
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
400
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
N2 Tvib (K)
4.8 mm
8.0 mm
11.1 mm
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
3200 ± 285
5333 ± 333
10345 ± 484
—
11195 ± 324
—
—
—
—
—
6218 ± 314 3099 ± 282
7005 ± 472 2985 ± 283
9375 ± 308 2770 ± 292
—
2722 ± 318
—
2691 ± 352
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
4104 ± 365
7955 ± 509
11505 ± 323
—
—
3165 ± 309
3068 ± 269
4852 ± 324
6483 ± 342
7045 ± 348
7583 ± 413 5470 ± 293
8857 ± 505 5012 ± 295
—
5703 ± 313
—
5855 ± 292
—
8170 ± 353
0.7
1.8
3.5
5.3
7.1
—
—
—
—
—
5438 ± 398 3227 ± 290
9687 ± 432 3178 ± 279
—
5307 ± 341
—
5868 ± 287
—
8610 ± 348
Values are means with 95% confidence interval bounds on the mean.
Although secondary flow rate was another control parameter, it had no significant effect on the temperatures
obtained. Mean values are therefore across all secondary flow rates.
72
4.2.3
Analysis
Several points are to be addressed in order to interpret these temperature data.
Some of these are related to the spectral collection process itself, while other issues
depend on the the physical processes of the emission species.
4.2.3.1
Spectroscopic Limitations
At the time the data were collected neither a line or intensity calibration source
were available. Both the actual spectrometer wavelength shift and instrumental
line shape were not pre-determined. The line shape can have a large effect on a
spectral profile and, therefore, the reported temperatures. So that, while an estimate
to the line shape was made by an iterative fit, its accuracy has not been verified
against the true instrumental line shape. However, the line shape parameters used
fall within the range of both that quoted by the manufacturer and that fitted to an
available atomic line, as presented in Section 3.3.1.2. Though the spectrometer had
been manufacturer-calibrated immediately prior to use, there was still a deviation
in wavelength position.
An initial poor fit between OH and N2 SPS led to the decision to subtract out a
linear background spectra. The reliability of this choice rests on the source of the
interference, uncertain as it currently stands.
Often, plasma parameter spatial profiles obtained via optical emission methods
are determined by reducing radial data using an Abel inversion process. To eliminate nozzle erosion and limit testing duration, while still obtaining good temporal
resolution, integration times were kept as short as possible (generally 1–5 s). It
was for this reason, in part, that the plasma emission was only studied at four
separate axial locations along the plasma. The data presented can be considered an
average across the collimating lens viewing area. Such a process lends to a quicker
determination of the bulk plasma gas temperature. What the underlying spatial
emission profile actually is may be significantly different. Work reported by other
authors on similar microwave plasma torches indicate that strong gradients exist
in these plasmas [88]. Also, the spatially resolved concentration of the two species
is not known.
73
4.2.3.2
Species Concentration
No explicit measurements of species concentration were made during testing, as
emission intensities were not calibrated for absolute measurements. The only
control parameter that directly influenced concentration was the volumetric flow
rate of argon (both through the nozzle and secondary flow). Water in the tube
reservoir was the primary source of OH in the plasma jet, but its flow rate and
concentration were not controlled for. Entrainment of the surrounding air into the
plasma introduced nitrogen and oxygen. No account was taken for entrainment
of H2 O from the atmosphere surrounding the plasma. An accurate assessment of
the pressure, temperature, velocity, and species concentration at the nozzle exit
was not available. With these limitations, no definitive statement can be made on
the concentrations of the constituents of the plasma. Yet, comparing the data with
the work of others on similar devices (notably, the TIA and MPT) may offer some
insight.
Relative emission intensities provide some indication as to the relative change
in concentration of the molecular species. Figure 4.15 presents normalized relative
intensities of the band heads of the OH (0, 0) and N2 SPS (1, 0) transitions as a
function of position and NFR. The N2 SPS intensity clearly decreases with nozzle
flow rate and axial position. Similar trends have been noted for molecular species in
rare gas ICP and microwave plasmas (e.g., NH in Ar [89], and N2 SPS and N+2 FNS
in Ar and He [90]).
Increased emission intensity of the N2 SPS band head at higher flow rates may
indicate increased nitrogen concentrations due to stronger turbulent mixing with
increased flow rate, while decreased intensity with position may be due to a decrease in electron and argon metastable density. Timmermans [89] and van der
Mullen and Jonkers [91] note for the TIA that higher flow rates lead to a longer
active zone in the plasma, the region of primary ionization nearest to the nozzle tip.
According to their reasoning, a region exists close to the nozzle where emission
intensity stays relatively constant where higher electron temperatures—and increased ionization—exists. Further downstream from the active zone, the emission
intensity would decrease as electron density diminishes—and we can see from the
N2 SPS emission intensity that this intensity does decrease with axial postion. Unfortunately, saturated spectral emission data closest to the nozzle tip (0.8 mm above
74
the nozzle) negated intensity comparisons there, so that this cannot be generalized
yet for the PSMPT. This increased saturation may indicate that significantly more
nitrogen is entrained in the PSMPT over the TIA. Higher flow velocities seem to
exist in the PSMPT as the nozzle aperture diameter is generally much smaller than
that employed in other similar microwave torches, such as the TIA. Therefore,
enhanced entrainment of the surrounding air would increase N2 concentrations
Relative Intensity
Relative Intensity
further.
0.7 SLM
1.8 SLM
3.5 SLM
5.3 SLM
7.1 SLM
4.8
8.0
11.1
0.7 SLM
1.8 SLM
3.5 SLM
5.3 SLM
7.1 SLM
4.8
Position (mm)
8.0
11.1
Position (mm)
Figure 4.15. Relative intensities of the OH (left) and N2 SPS (right) calculated spectra as
a function of position and NFR. Values are average over all powers (225 W, 300 W, and
400 W) and secondary flow rates (6.4 SLM, 10.6 SLM, 21.2 SLM, 31.8 SLM, and 42.4 SLM).
4.2.3.3
Rotational and Vibrational Temperatures
Ar(3 P0,2 ) metastables have a long lifetime, (τ > 1 s [92]). As such, they can exist in
the entire length of the plasma. The second positive system of nitrogen is easily
excited from ground to C state by energy transfer with argon metastables via
N2 (X) + Ar∗ (3 P0,2 ) → N2 (C) + Ar(1 S0 ).
(4.1)
This reaction is cited several times in SEP literature, most importantly for work done
on the TIA at powers less than 1 kW [88] (see Section 1.3.3), but also others (see,
e.g., [93]). H2 O can also undergo a reaction chain starting with argon metastables
that leads to the formation of excited OH(A). We would therefore expect to see OH
and N2 SPS emission signals throughout the plasma.
75
One negative side effect of these reactions, for temperature measurement purposes, involves the overpopulation of high rotational levels of OH(A) and N2 (C).
This effect is noted by many authors—even at elevated pressures—and it leads to
several possible values for the rotational temperature for both OH and N2 (see, e.g.,
[94–97]). The lowest rotational levels, however, are more likely to be in equilibrium with the gas temperature. Without an accurate collisional–radiative model,
the spectra here were simulated assuming a Boltzmann distribution among energy
levels for both rotational and vibrational modes separately for each species. To
mitigate the effects of OH quenching and selective rotational excitation, only the
lower rotational levels (J0 ≤ 14) were used to fit the spectra. In the combined fitted spectral plot of Figure 4.14, evidence for this effect is apparent at wavelengths
longer than 310 nm. Several peaks that clearly belong to the OH (A→X) spectra
are undercalculated by the simulated spectra. For N2 SPS, the rotational levels are
more difficult to separate out, and therefore no account had been taken of this effect
in choosing spectral fitting peaks. This may indicate why, as Figure 4.16 indicates,
the rotational temperature of N2 is substantially higher than the OH rotational temperature. Such a result is also testified to by Sarani et al. [95] and Raud et al. [98],
and they used that as a basis for arguing that the N2 SPS rotational temperature
8000
8000
7500
7500
7000
7000
6500
6500
Temperature (K)
Temperature (K)
does not correspond to the gas temperature.
6000
5500
5000
4500
5500
5000
4500
4000
4000
3500
3500
3000
3000
4.8
8.0
Position (mm)
11.1
OH, 225 W
OH, 300 W
OH, 400 W
N2 , 225 W
N2 , 300 W
N2 , 400 W
6000
11.1
8.0
Position (mm)
Figure 4.16. OH and N2 SPS mean rotational (left) and vibrational (right) temperatures
vs. position at 225 W, 300 W, and 400 W. NFR = 0.7 SLM. Means are across all SFRs. Error
bars indicate 95% confidence interval bounds.
76
There appears to be a strong inverse dependence of rotational temperature on
NFR in Figures 4.17–4.18. In these figures are presented histograms of scan counts
that are counted at each rotational temperature (in increments of 100 K) at a position
of 11.14 mm and power levels of 300 W and 400 W. Each count indicates a spectral
scan whose experimental data was fitted to a simulated spectrum at the specified
temperature. Mean values and associated 95% confidence interval error bars are
shown, as well. A clear inverse dependence of rotational temperature on NFR can
be seen. A similar trend has been noted by other authors (see, e.g., [48, 99]).
Figure 4.16 demonstrates variation of rotational and vibrational temperature
with axial position and power. The mean values at each position span all sets of
nozzle and secondary flow rates.
The following is offered in order to explain these variations of rotational temperature with NFR and position. One of the primary processes involved with Ar–N2
plasmas is the charge transfer process
Ar+ + N2 → Ar + N+2 ,
(4.2)
N+2 + e → N + N + (energy).
(4.3)
which is followed by
Re-association of the nitrogen atoms is also exothermic, so that, combined with
the dissociative–recombination in Equation (4.3), heating can take place even after
the primary ionization processes are completed in the active zone of the plasma
[91]. As was discussed in the previous section, a higher flow rate appears to delay
the termination of the active (ionizing) zone of the plasma in the TIA. This would
imply that higher flow rates would also delay the heating of the gas. The fact that
the gas temperature is higher downstream is a consequence of this heating process,
if the results between the PSMPT and TIA are comparable.
As to the vibrational temperatures, there is a clear decrease with respect to position while the temperature increases with applied power (see Figure 4.16). Also an
increase with respect to nozzle flow rate can be seen in Figures 4.19–4.20. The variation with respect to position in Figure 4.16 indicates that vibrational–translational
equilibrium may be reached around 11.1 mm downstream of the nozzle tip.
77
Scan Counts
200
150
100
50
0
3500
7.1
3300
5.3
3100
3.5
Nozzle Flow
Rate (SLM)
2900
1.8
Temperature (K)
2700
0.7
2500
Scan Counts
200
150
100
50
0
7800
7.1
7000
5.3
6200
5400
3.5
Nozzle Flow
Rate (SLM)
4600
3800
1.8
Temperature (K)
3000
0.7
2200
Figure 4.17. OH (top) and N2 SPS (bottom) rotational temperature histograms vs. NFR.
Forward power = 300 W. Position = 11.1 mm. Vertical posts indicate mean with error
bars using 95% confidence interval bounds.
78
Scan Counts
150
100
50
0
3500
7.1
3400
3300
5.3
3200
3100
3.5
3000
2900
Nozzle Flow
Rate (SLM)
Temperature (K)
2800
1.8
2700
0.7
2600
Scan Counts
150
100
50
0
7.1
5800
5400
5.3
5000
4600
3.5
4200
3800
Nozzle Flow
Rate (SLM)
1.8
3400
Temperature (K)
3000
0.7
2600
Figure 4.18. OH (top) and N2 SPS (bottom) rotational temperature histograms vs. NFR.
Forward power = 400 W. Position = 11.1 mm. Vertical posts indicate mean with error
bars using 95% confidence interval bounds.
79
Scan Counts
200
150
100
50
0
7.1
6500
6000
5.3
5500
5000
3.5
4500
Nozzle Flow
Rate (SLM)
1.8
Temperature (K)
4000
0.7
3500
300
Scan Counts
250
200
150
100
50
0
7.1
8000
7000
5.3
6000
3.5
5000
Nozzle Flow
Rate (SLM)
1.8
4000
0.7
Temperature (K)
3000
Figure 4.19. OH (top) and N2 SPS (bottom) vibrational temperature histograms vs. NFR.
Forward power = 300 W. Position = 11.1 mm. Vertical posts indicate mean with error
bars using 95% confidence interval bounds.
80
Scan Counts
150
100
50
0
9000
7.1
8500
8000
5.3
7500
7000
3.5
6500
6000
Nozzle Flow
Rate (SLM)
Temperature (K)
5500
1.8
5000
0.7
4500
Scan Counts
200
150
100
50
0
7.1
9000
8000
5.3
7000
3.5
Nozzle Flow
Rate (SLM)
6000
5000
1.8
4000
0.7
Temperature (K)
3000
Figure 4.20. OH (top) and N2 SPS (bottom) vibrational temperature histograms vs. NFR.
Forward power = 400 W. Position = 11.1 mm. Vertical posts indicate mean with error
bars using 95% confidence interval bounds.
81
Chapter 5
Conclusions
5.1
Summary and Contributions
The Penn State Microwave Plasma Torch (PSMPT) is a device which utilizes microwave energy to initiate and sustain a plasma in an argon gas jet issuing from a
copper nozzle into the ambient atmosphere. Protruding through a rectangular
waveguide, the nozzle acts to enhance the local electric field when microwaves
are excited in the waveguide. The plasma resembles a small flame, approximately
2–4 cm in length and less than 1 cm in total diameter. The primary goals that have
driven experimental design and characterization of the torch include (1) increasing
plasma jet control via improved impedance matching; (2) reducing the erosion
of the nozzle tips; and (3) determining the viability of applying the PSMPT to
the cutting and melting of materials via gas temperature measurements. Literature on the similar microwave torches—particularly, those of the single-electrode
plasma (SEP) type—was reviewed. Through the realization of these goals and comparison with the available literature, several practical contributions can be made
as will be discussed in turn.
Plasma Control Via Impedance Matching:
In characterizing the PSMPT, several
parameters could be varied, including nozzle and secondary flow rates, power
levels, nozzle and waveguide aperture geometries, nozzle position, and sliding
short position. In order to initiate and sustain a stable plasma jet for various sets
of these parameters, workable solutions were generated for the most significant
obstacles: impedance matching and nozzle erosion.
Many of the other SEP designs rely on several tuning elements. Tuning with
the PSMPT originally relied only on the positioning of the sliding short and nozzle
tip. These factors alone were not sufficient for reliable and efficient PSMPT operation. Poor impedance matching was overcome most effectively with an automatic
three-stub tuner that could determine a match quickly. Without the use of said
tuner, plasma initiation and sustainment were significantly more difficult. A torch
with no three-stub tuner, or a three-stub tuner with manual or slowly adjusting
tuning, led to difficulties in initiating a plasma. For example, the use a striking aid
was no longer necessary with the current tuner. A plasma could be sustained at
applied powers down to approximately 200 W with the current tuner, as opposed
to requiring powers greater than 1 kW without this tuner. The response of the
plasma was no longer as sensitive to the position of the sliding short and nozzle
tip. Several other tests were made observing the power flow and loss through the
waveguide aperture. The PSMPT plasma appears to absorb a significant fraction
of all power not reflected back to the source, especially with the tuner in place.
Any future work on the PSMPT would clearly benefit by incorporating some level
of active-tuning with a fast response time. This solution may have practical value
for future research on similar plasma torches, as many other SEP designs rely on
the addition of several complicated tuning elements.
Nozzle Erosion Reduction:
The presence of nozzle erosion and how to mitigate
erosion is not addressed in detail in the literature on SEPs. Work on the MPJ
noted that smaller nozzle aperture sizes increased erosion of their copper nozzles
[100]. Significant erosion was a limiting factor in early tests with the PSMPT.
Water-cooled nozzles and nozzles with tungsten-coated tips did not provide any
noticeable reduction in erosion rates. More recent testing reveals that erosion can
be mitigated by addition of a secondary flow of argon around the primary nozzle
gas flow and regular cleaning of the nozzle surface. Though the exact mechanism
for this reduction in erosion is uncertain, tests now could be run continuously for
several minutes with no noticeable damage to the nozzle.
Viability From Temperature Measurements:
Very few SEPs are waveguide
based (see Figure 1.3). The MPJ [53] and the TIAGO [41] are perhaps the only
such devices comparable to the PSMPT in their design. Such devices are simpler
to construct than their counterpart designs (waveguide-to-coaxial and coaxial),
83
generally allowing higher power levels as well. Both the MPJ and TIAGO use only
a primary flow through the nozzle without secondary flow. As was mentioned
previously, the nozzle aperture diameters can be up to an order of magnitude larger
for these two torches compared to the PSMPT. Power levels in these two torches
are comparable to the PSMPT. Only recently have temperature measurements been
made on the TIAGO [101]; such data on the MPJ appears to be unavailable.
Gas temperatures in the PSMPT were deduced via OH rotational temperature
measurements. Significant emission from the N2 SPS (1, 0), (2, 1), (3, 2), and (4, 3)
bands overlapped the OH (0, 0) and (1, 1) bands. This interference required simultaneously fitting both OH and N2 SPS in the region of 305–318 nm. The commerical
software LIFBASE was used to simulate OH; an independent MATLAB code was
developed to accurately simulate N2 SPS.
Increased N2 SPS emission intensity, especially at higher flow rates, is believed
to be due to increased turbulent mixing of the surrounding air. Nitrogen concentrations increase in the plasma due to this entrainment. This result is contrary to
the numerical study performed on the MPJ by the Liverpool group [102], where
they claim that air entrainment is not pertinent to their plasma jet. As such, the
results here correspond more with the work done on the TIA, TIAGO, and MPT,
where account has been made for entrainment. PSMPT results indicate an OH rotational temperature—and assumed gas temperature—between 2700–3400 K. These
results correspond to N+2 FNS rotational temperature measurements performed on
the TIAGO, where they reported temperatures ranging from roughly 2500–3500 K
[101]. However, the results on the TIAGO were only performed at one location
along the plasma’s axis; whereas, the work here presents measurements at four separate axial locations, providing a more detailed spatial description of the PSMPT.
The rotational temperature measured in the PSMPT tends to
• increase with distance from the nozzle tip;
• decrease with nozzle flow rate;
• increase slightly with applied microwave power; and
• show no significant effect from secondary flow rate.
84
Rotational temperature measurements of the N2 SPS indicate it is not suitable for
measuring the temperature in the PSMPT due to overpopulation of higher rotational levels from excitation by argon metastables. Vibrational temperatures
appear to relax to rotational–translational temperatures near the end of the visible
discharge.
Though they only provide a spatially-averaged estimate to the gas temperature,
these results indicate the PSMPT could significantly melt, and possibly cut, many
materials. However, further studies are required to determine how well heat could
be transfered to a surface for cutting and/or melting applications.
5.2
Recommendations for Future Research
The goals and contributions of the present study were outlined above. Several areas
can be identified to further the understanding and design of the PSMPT. These
include (1) assessing whether plasma filamentation is occuring in the PSMPT;
(2) using a more suitable thermometric species; (3) determining the flow structure
of the PSMPT; (4) increasing diagnostics; and (5) modifying the nozzle material or
geometry.
Plasma Filamentation:
Microwave plasmas demonstrate a tendency to form
highly constricted filaments at high pressure. Several SEPs presented in the literature give indication that this class of plasma may frequently assume a filamentary
nature. In each case, the filamentation remains undetectable by eyesight or normal
photograph techniques. Detecting the filamentation has relied on high-speed photography, as well as inferences from the results of Thomson scattering, Schlieren
imaging, and microwave electric-field measurements. Though not visibly important, calculated plasma parameters nonetheless can depend significantly on the
state of the plasma—whether diffuse or filamentary [48]. Does the PSMPT exhibit
this filamentary nature? To answer that question would require employing one of
the detection methods just mentioned. Figure 5.1 presents an example of a torch
plasma structure created exhibiting filamentation.
85
Figure 5.1. MPT torch filamentation demonstrated through different
exposure times of 10 ms (left), 1 ms
(center), 100 µs (right) (from [48]).
Intense, localized electric field gradients near the nozzle tip would factor
strongly in filament initiation. Such fields tend to create regions of high electron
density. High power levels were needed to sustain the plasma in the PSMPT—
any existing discharge would extinguish below approximately 200 W. However,
increasing the power levels—and, therefore, the electron density—would further
contribute to filament formation, as noted previously. The presence of such filamentation may be partially responsible for nozzle erosion. Perhaps one factor
mitigating the development of filaments is the velocity of the gas exiting the nozzle. If we view the formation of filaments as a form of plasma instability, one way to
eliminate such instabilities is to increase the flow rate through a plasma. Based on
device dimensions presented in the literature, the nozzle aperture in the PSMPT is
substantially smaller than other designs. How many similar SEPs present choked
sonic flow through the nozzle (as the PSMPT does) is not apparent. Other authors have noted a stronger tendency to reveal filamentation through calculated
parameters when the flow rate through the plasma decreases.
Thermometric Species:
For the purposes of studying the plasma through opti-
cal emission spectroscopy, water was added to the primary nozzle flow so that an
OH rotational temperature might be deduced. From the results presented in the
previous chapter, it is apparent that there are several issues to address regarding
this decision. First, the addition of water to the plasma likely increases the gas temperature above its normal value: H2 O has a fast vibration–translation relaxation
time, so that energy stored in vibrational levels can quickly be converted into thermal energy. Such an increase in temperature may or may not be a desired property,
86
depending upon application. Second, the concentration level of water entering the
plasma could not be accounted for using the bubbling reservoir method. A more
controlled method of introducing water into the nozzle gas would allow for better
precision in measuring the rotational temperature. Lastly, the level of interference
by nitrogen (both molecular and ionic) makes the choice of OH as a thermometric species less appealing. The accuracy of the fit between multiple species relies
heavily on the assumptions regarding the radiative processes of those species. This
includes the need for a detailed knowledge of any quenching channels for each of
the rotational levels of the species involved. Some authors have noted their lack of
confidence in OH as a thermometric species [94,99]. A better option, if suited to the
spectroscopic system available, would be to utilize those species already available
in the plasma (e.g., N2 SPS and N+2 FNS). Such a method has been presented by
Nassar et al. [87].
Plasma Control and Variation of Gas Flow:
Part of the original design criteria
included determining if the plasma could form a very narrow, high-temperature
gas stream. Based on the results to date, it would appear that the current setup will
not permit that level of control. And this is not without good reason. The turbulent
mixing induced at the jet boundary entrains nitrogen into the plasma. Results from
work on the TIA and others indicate that this helps to increase the gas temperature
through exothermic charge-transfer and dissociation reactions. Reducing the level
of entrainment would seem to counteract these positive features.
Rather than entraining the nitrogen from the atmosphere, nitrogen could be
mixed with the argon and then the flow reduced accordingly. However, reducing
the flow rate further does not seem viable. Initiating plasmas at lower flow rates
was difficult and once initiated, such plasmas tended to be more erratic. Also,
based on the spectroscopic results, it appears that higher flow rates are desirable
assuming they shift the higher gas temperatures further downstream of the nozzle
tip. This likely helps limit erosion of the nozzle tip. In any case, an area of future
research could include detailed studies of the flow field structure around the nozzle.
Plasma size was more easily controlled by decreasing the power and employing
active tuning. With the current setup, the plasma could not be sustained at powers less than 200 W. Perhaps with further study of the microwave transmission
properties PSMPT, a more optimal design could permit a lower threshold.
87
Increased Diagnostics:
Already mentioned in the above recommendations is
the need for increased diagnostics of the PSMPT. The spectral region studied
encompassed 290–390 nm, but both the N+2 FNS and N2 SPS have several strong
bands above 390 nm. Strong lines of neutral argon exist primarily out of this range,
as well. In order to make stronger inferences on gas temperature, as well as including electron density and temperature measurements, a larger spectral region is
needed (without sacrificing resolution). Absolute intensity measurements would
aid heavy particle concentration measurements. Determining actual flow field parameters (velocity, pressure, etc.) at the nozzle exit and downstream would give
better insight into the entrainment process, the effect of flow velocity on gas temperature, and would aid any future modeling. And, as already mentioned above,
high-speed photography would help determine if the plasma exhibits filamentary
behavior.
Nozzle Geometry and Material:
Nozzle erosion still remains an issue with the
torch, albeit much diminished in short term effects. Cooling of the nozzle using
water has not proven beneficial yet, though a better interior design of the nozzle
may change that. Resistive power losses near the nozzle–waveguide joint should
be addressed. The joint should avoid gaps between components, lest these losses
(including arcing) decrease the overall efficiency of the PSMPT.
88
References
[1] C. Tendero, C. Tixier, P. Tristant, J. Desmaison, and P. Leprince, “Atmospheric pressure
plasmas: A review,” Spectrochimica Acta Part B, vol. 61, pp. 2–30, 2006.
[2] K. Tachibana, “Current status of microplasma research,” IEEJ Transactions on Electrical and
Electronic Engineering, vol. 1, pp. 145–155, 2006.
[3] J. Marec, E. Bloyet, M. Chaker, P. Leprince, and P. Nghiem, Electrical Breakdown and Discharges
in Gases (Part 2): Macroscopic Processes and Discharges, ser. NATO ASI: Physics. Plenum Press,
1983, ch. Microwave Discharges, pp. 347–382.
[4] C. Ferreira, M. Moisan, and Z. Zakrzewski, Microwave Excited Plasmas, ser. Plasma Technology,
V. 4. Elsevier Science and Technology Books, 1992, ch. 2: Physical Principles of Microwave
Plasma Generation, pp. 11–52.
[5] Y. Arata, S. Miyake, S. Takeuchi, and A. Kobayashi, “High power microwave plasma beam
as a heat source (Report II): Characteristics of 30 kW plasmatron,” Transactions of JWRI, vol. 3,
no. 1, pp. 21–26, 1974.
[6] U. Jordan, D. Dorozhkina, V. Semenov, T. Olsson, D. Anderson, M. Lisak, J. Puech, I. Nefedov,
and I. Shereshevskii, “Microwave corona breakdown around metal corners and wedges,”
IEEE Transactions on Plasma Science, vol. 35, no. 3, pp. 542–550, 2007.
[7] D. Koufogiannis, E. Sorolla, J. Mosig, and M. Mattes, “Analyzing corona breakdown with a
finite element-based electromagnetic solver,” 7th European Conference on Antennas and Propagation (EuCAP), pp. 2261–2263, 2013.
[8] Z. Ahmadi, M. Khani, S. Kooshki, F. Mirzajani, and B. Shokri, “Investigation of variation
power and additive gas effect on the SF6 destruction using atmospheric microwave plasma
torch,” IEEE Transactions on Plasma Science, vol. 39, no. 9, pp. 1834–1841, September 2011.
[9] Y. Arata, S. Miyake, T. Innami, and Y. Yoshioka, “High power microwave plasma beam as a
heat source (Report IV): Application to cutting,” Transactions of JWRI, vol. 4, no. 2, pp. 105–110,
1975.
[10] S. Murayama, “Ultrahigh-frequency discharge in atmospheric-pressure gas,” Journal of Applied Physics, vol. 39, no. 12, pp. 5478–5484, 1968.
[11] E. Tatarova, F. M. Dias, E. Felizardo, J. Henriques, C. M. Ferreira, and B. Gordiets, “Microwave
plasma torches driven by surface waves,” Plasma Sources Science and Technology, vol. 17, no. 2,
p. 024004, 2008.
[12] Q. Jin, Y. Duan, and J. A. Olivares, “Development and investigation of microwave plasma
techniques in analytical atomic spectrometry,” Spectrochimica Acta Part B: Atomic Spectroscopy,
vol. 52, no. 2, pp. 131–161, 1997.
[13] A. E. Croslyn, B. W. Smith, and J. D. Winefordner, “A review of microwave plasma sources
in atomic emission spectrometry: Literature from 1985 to the present,” Critical Reviews in
Analytical Chemistry, vol. 27, no. 3, pp. 199–255, 1997.
[14] J. D. Winefordner, E. P. Wagner, and B. W. Smith, “Status of and perspectives on microwave
and glow discharges for spectrochemical analysis. Plenary lecture,” J. Anal. At. Spectrom.,
vol. 11, pp. 689–702, 1996.
[15] Q. Jin, W. Yang, F. Liang, H. Zhang, A. Yu, Y. Cao, J. Zhou, and B. Xu, “Recent advances in
microwave plasma torch atomic emission spectrometry,” J. Anal. At. Spectrom., vol. 13, pp.
377–384, 1998.
[16] R. L. Wathen, “Genesis of a generator: The early history of the magnetron,” Journal of the
Franklin Institute, vol. 255, no. 4, pp. 271–287, 1953.
[17] J. D. Cobine and D. A. Wilbur, “The electronic torch and related high frequency phenomena,”
Journal of Applied Physics, vol. 22, no. 6, pp. 835–841, June 1951.
[18] ——, “Electronic torch,” Electronics, vol. 24, no. 6, pp. 92–92, June 1951.
[19] ——, High-Temperature Technology.
John Wiley and Sons, 1956, pp. 327–330, 332.
[20] J. P. Jordan, “High-frequency electrical heating method and apparatus,” U.S. Patent 2 587 331,
February 26, 1952.
[21] J. Laroche, “Recherches physiques et chimques sur l’excitation des gaz dans la torche électronique. Étude de la formation du bioxyde d’azote,” Ph.D. dissertation, University of Paris,
1961.
[22] ——, “Mesures physiques sur la torche électronique,” Appl. sci. Res.: Section B, vol. 5, pp.
294–296, 1955.
[23] L. Mollwo, “Elektronentemperatur und Eletronenraucschen in der hochfrequenten Fackelentladung,” Annalen der Physik, vol. 2, no. 7, pp. 97–129, 1958.
[24] O. Scholz, “Die hochfrequenz-plasma-flamme als neue Wärmequelle (The high frequency
plasma flame as a new source of heat),” Schweissen und Schneiden, vol. 11, pp. 497–498, 1959.
[25] W. Schmidt, “Der mikrowellen-plasmabrenner,” Elektronische Runhschau, vol. 13, no. 11, pp.
404–406, 1969.
[26] V. Trunec̆ek, “Die Umwandlung der Fackelentladung in die Hochfrequenz-korona,” Cechoslovackij fiziceskij zurnal B, vol. 10, no. 12, pp. 960–965, 1960.
[27] J. Swift, “A UHF plasma torch,” Proceedings of the IREE Australia, vol. 25, pp. 779–782, November 1964.
[28] ——, “A microwave plasma torch,” Electronic Engineering, vol. 38, no. 457, pp. 152–154, March
1966.
[29] M. Yamamoto and S. Murayama, “UHF torch discharge as an excitation source,” Spectrochimica Acta Part A: Molecular Spectroscopy, vol. 23, no. 4, pp. 773–776, 1967.
[30] ——, “Electrode erosion in a radio-frequency torch discharge,” Japanese Journal of Applied
Physics, vol. 2, pp. 65–66, 1963.
[31] S. Murayama, H. Matsuno, and M. Yamamoto, “Excitation of solutions in a 2450 MHz discharge,” Spectrochimica Acta, vol. 23B, pp. 513–520, 1968.
[32] R. Nakashima, S. Sasaki, and S. Shibata, “Determination of niobium, titanium and zirconium by high-frequency plasma torch emission spectronmetry and its application to steel,”
Analytica Chemica Acta, vol. 70, pp. 265–274, 1973.
[33] G. Larson and V. Fassel, “Comparison of interelement effects in a microwave single electrode
plasma and in a radiofrequency inductively coupled plasma,” Analytical Chemistry, vol. 48,
no. 8, pp. 1161–1166, 1976.
[34] U. Jecht and W. Kessler, “Beobachtungen und Untersuchungen an HF-Plasmaflammen,” Z.
analyt. Chem., vol. 198, pp. 27–35, 1963.
90
[35] W. Tappe and J. van Calker, “Quantitative spektrochemische Untersuchungen mit hochfrequenten Plasmaflammen,” Fresnius Z. Anal. Chem, vol. 198, no. 1, pp. 13–20, 1963.
[36] M. Sermin, “Analyse de solutions par spectrométrie d’émission a lecture direct avec une
source d’excitation plasma,” Analusis, vol. 2, pp. 186–189, 1973.
[37] Y. Mitsuda, T. Yoshida, and K. Akashi, “Development of a new microwave plasma torch
and its application to diamond synthesis,” Review of Scientific Instruments, vol. 60, no. 2, pp.
249–252, feb 1989.
[38] M. Moisan, G. Sauvé, Z. Zakrzewski, and J. Hubert, “An atmospheric pressure waveguidefed microwave plasma torch: The TIA design,” Plasma Sources Science and Technology, vol. 3,
no. 4, pp. 584–592, 1994.
[39] M. Moisan, R. Grenier, and Z. Zakrzewski, “The electromagnetic performance of a surfatronbased coaxial microwave plasma torch,” Spectrochimica Acta Part B: Atomic Spectroscopy,
vol. 50, no. 8, pp. 781–789, 1995.
[40] J. Stańco, D. Czylkowski, J. Podliński, M. Jasiński, Z. Zakrzewski, and J. Mizeraczyk, “Structure of the plasma jet of a pulsed microwave plasma torch of nozzle type,” Czechoslovak Journal
of Physics, vol. 56, pp. B1068–B1073, 2006.
[41] M. Moisan, Z. Zakrzewski, and J. C. Rostaing, “Waveguide-based single and multiple nozzle
plasma torches: The TIAGO concept,” Plasma Sources Science and Technology, vol. 10, no. 3, p.
387, 2001.
[42] J. Mizeraczyk, M. Jasiński, and Z. Zakrzewski, “Hazardous gas treatment using atmospheric
pressure microwave discharges,” Plasma Physics and Controlled Fusion, vol. 47, no. 12B, p.
B589, 2005.
[43] N. H. Bings, M. Olschewski, and J. A. C. Broekart, “Two-dimensional spatially resolved
excitation and rotational temperatures as well as electron number density measurements in
capactivitely coupled microwave plasmas using argon, nitrogen and air as working gases by
spectroscopic methods,” Spectrochimica Acta Part B, vol. 52, pp. 1965–1981, 1997.
[44] A. Bilgic, C. Prokisch, J. Broekaert, and E. Voges, “Design and modelling of a modified
2.45 GHz coaxial plasma torch for atomic spectrometry,” Spectrochimica Acta Part B: Atomic
Spectroscopy, vol. 53, no. 5, pp. 773–777, 1998.
[45] B. W. Pack and G. M. Hieftje, “An improved microwave plasma torch for atomic spectrometry,” Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 52, no. 14, pp. 2163–2168, 1997.
[46] E. Timmermans, J. Jonkers, I. Thomas, A. Rodero, M. Quintero, A. Sola, A. Gamero, and
J. van der Mullen, “The behavior of molecules in microwave-induced plasmas studied by
optical emission spectroscopy. 1. Plasmas at atmospheric pressure,” Spectrochimica Acta Part
B: Atomic Spectroscopy, vol. 53, no. 11, pp. 1553–1566, 1998.
[47] E. A. H. Timmermans, I. A. J. Thomas, J. Jonkers, E. Hartgers, J. A. M. van der Mullen, and
D. C. Schram, “The influence of molecular gases and analytes on excitation mechanisms in
atmospheric microwave sustained argon plasmas,” Fresenius’ Journal of Analytical Chemistry,
vol. 362, pp. 440–446, 1998.
[48] J. van der Mullen, M. van de Sande, N. de Vries, B. Broks, E. Iordanova, A. Gamero, J. Torres,
and A. Sola, “Single-shot Thomson scattering on argon plasmas created by the Microwave
Plasma Torch; evidence for a new plasma class,” Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 62, no. 10, pp. 1135–1146, 2007.
[49] L. L. Alves, R. Álvarez, L. Marques, S. J. Rubio, A. Rodero, and M. C. Quintero, “Modeling of
an axial injection torch,” The European Physical Journal - Applied Physics, vol. 46, no. 02, 2009.
91
[50] R. Álvarez and L. L. Alves, “Two-dimensional electromagnetic model of a microwave plasma
reactor operated by an axial injection torch,” Journal of Applied Physics, vol. 101, no. 10, p.
103303, 2007.
[51] A. Kirichenko, A. Motornenko, and O. Suvorova, “Structure of a discharge induced by a
coaxial microwave plasmatron with a gas-supply channel in the inner electrode,” Plasma
Physics Reports, vol. 29, pp. 528–533, 2003.
[52] A. Kirichenko and O. Suvorova, “Electromagnetic fields and structure of microwave torch
discharge,” Czechoslovak Journal of Physics, vol. 56, pp. B1079–B1085, 2006.
[53] A. I. Al-Shamma´a, S. R. Wylie, J. Lucas, and J. D. Yan, “Atmospheric microwave plasma jet
for material processing,” IEEE Transactions on Plasma Science, vol. 30, no. 5, pp. 1863–1871,
October 2002.
[54] M. Leins, K.-M. Baumgärtner, M. Walker, A. Schulz, U. Schumacher, and U. Stroth, “Studies
on a microwave-heated atmospheric plasma torch,” Plasma Processes and Polymers, vol. 4,
no. S1, pp. S493–S497, 2007.
[55] D. Shin, C. Bang, J. Kim, Y. Hong, H. Uhm, D. Park, and K. Kim, “Treatment of metal surface
by atmospheric microwave plasma jet,” Plasma Science, IEEE Transactions on, vol. 34, no. 4,
pp. 1241–1246, August 2006.
[56] E. Barkhudarov, S. Gritsinin, V. Knyazev, V. Kop’ev, I. Kossyi, N. Malykh, M. Misakyan, and
V. Silakov, “Microwave torch: Description of construction, results of investigation, possibility of application,” in Aerospace Sciences Meetings. American Institute of Aeronautics and
Astronautics, 2005.
[57] R. Stonies, S. Schermer, E. Voges, and J. A. C. Broekaert, “A new small microwave plasma
torch,” Plasma Sources Science and Technology, vol. 13, no. 4, pp. 604–611, 2004.
[58] Y. Arata, S. Miyake, and S. Takeuchi, “High power microwave plasma beam as a heat source
(Report I) : Microwave plasmotron in nitrogen gas,” Transactions of JWRI, vol. 2, no. 1, pp.
27–39, 1973.
[59] C. Balanis, Advanced Engineering Electromagnetics.
[60] P. Rizzi, Microwave Engineering Passive Circuits.
[61] D. Pozar, Microwave Engineering, 3rd ed.
Wiley, 1989.
New Jersey: Prentice Hall, 1988.
John Wiley and Sons, 2005.
[62] M. A. Herlin and S. C. Brown, “Breakdown of a gas at microwave frequencies,” Phys. Rev.,
vol. 74, pp. 291–296, August 1948.
[63] J. Jonkers, M. van de Sande, A. Sola, A. Gamero, and J. van der Mullen, “On the differences
between ionizing helium and argon plasmas at atmospheric pressure,” Plasma Sources Science
and Technology, vol. 12, no. 1, pp. 30–38, 2003.
[64] A. MacDonald, Microwave Breakdown in Gases.
New York: Wiley, 1966.
[65] U. Jordan, D. Anderson, L. Lapierre, M. Lisak, T. Olsson, J. Puech, V. Semenov, J. Sombrin, and
R. Tomala, “On the effective diffusion length for microwave breakdown,” IEEE Transactions
on Plasma Science, vol. 34, no. 2, pp. 421–430, 2006.
[66] D. G. Anderson, U. Jordan, M. Lisak, and T. Olsson, “Microwave breakdown in resonators and
filters,” IEEE Transactions on Microwave Theory and Techniques, vol. 47, no. 12, pp. 2547–2556,
December 1999.
92
[67] V. M. M. Glaude, M. Moisan, R. Pantel, P. Leprince, and J. Marec, “Axial electron density and
wave power distributions along a plasma column sustained by the propagation of a surface
microwave,” Journal of Applied Physics, vol. 51, no. 11, pp. 5693–5698, 1980.
[68] T. Ida, M. Ando, and H. Toraya, “Extended pseudo-Voigt function for approximating the
Voigt profile,” Journal of Applied Crystallography, vol. 33, no. 6, pp. 1311–1316, December 2000.
[69] S. Kuo, D. Bivolaru, H. Lai, W. Lai, S. Popovic, and P. Kessaratikoon, “Characteristics of an
arc-seeded microwave plasma torch,” IEEE Transactions on Plasma Science, vol. 32, no. 4, pp.
1734–1741, 2004.
[70] M. Moisan and Z. Zakrzewski, “Plasma sources based on the propagation of electromagnetic
surface waves,” Journal of Physics D: Applied Physics, vol. 24, no. 7, p. 1025, 1991.
[71] Nonionizing radiation, OSHA Std. 1910.97, 1996.
[72] L. Luque and D. Crosley, “LIFBASE: Database and spectral simulation program (Version
1.5),” SRI International Report MP 99-009, 1999.
[73] C. O. Laux, “Radiation and nonequilibrium collisional-radiative models,” in von Karman
Institute for Fluid Dynamics, Special Course on Physico-Chemical Models for High Enthalpy and
Plasma Flows June 4–7, 2002.
[74] F. Roux, F. Michaud, and M. Vervloet, “High-resolution fourier spectrometry of 14 N2 : Analysis
of the (0–0), (0–1), (0–2), (0–3) bands of the C3 Πu –B3 Π g system,” Canadian Journal of Physics,
vol. 67, no. 2–3, pp. 143–147, 1989.
[75] ——, “High-resolution fourier spectrometry of 14 N2 violet emission spectrum: Extensive
analysis of the C3 Πu –B3 Π g system,” Journal of Molecular Spectroscopy, vol. 158, no. 2, pp.
270–277, 1993.
[76] B. Edlén, “The refractive index of air,” Metrologia, vol. 2, no. 2, p. 71, 1966.
[77] I. Kovács, Rotational structure in the spectra of diatomic molecules. American Elsevier Publishing
Company, 1969.
[78] E. E. Whiting, J. A. Paterson, I. Kovács, and R. W. Nicholls, “Computer checking of rotational
line intensity factors for diatomic transitions,” Journal of Molecular Spectroscopy, vol. 47, no. 1,
pp. 84–98, 1973.
[79] A. Schadee, “Theory of first rotational lline in transitions of diatomic molecules,” Astronomy
and Astrophysics, vol. 41, pp. 203–212, 1975.
[80] C. O. Laux and C. Kruger, “Arrays of radiative transition probabilities for the N2 first and
second positive, NO beta and gamma, N+2 first negative, and O2 Schumann-Runge band
systems,” Journal of Quantitative Spectroscopy & Radiative Transfer, vol. 48, no. 1, pp. 9–24, 1992.
[81] (2013) Optical resolution. http://www.oceanoptics.com/technical/opticalresolution.asp.
[82] E. E. Whiting, “An empirical approximation to the Voigt profile,” Journal of Quantitative,
vol. 8, pp. 1379–1384, 1968.
[83] J. J. Olivero and R. L. Longbothum, “Empirical fits to the voigt line width: A brief review,”
Journal of Quantitative Spectroscopy & Radiative Transfer, vol. 17, pp. 233–236, 1977.
[84] C. Biloiu, X. Sun, Z. Harvey, and E. Scime, “An alternative method for gas temperature
determination in nitrogen plasmas: Fits of the bands of the first positive system (B3 Π g →
A3 Σ+u ),” Journal of Applied Physics, vol. 101, no. 7, p. 073303, 2007.
93
[85] E. Iordanova, J. Torres, E. Benova, A. Gamero, A. Sola, B. H. P. Broks, and J. J. A. M. van der
Mullen, “Characterization of a high-pressure hydrogen microwave plasma torch using the
method of dBR,” Journal of Physics: Conference Series, vol. 44, no. 1, pp. 185–190, 2006.
[86] P. Fauchais and A. Vardelle, “Thermal plasmas,” Plasma Science, IEEE Transactions on, vol. 25,
no. 6, pp. 1258–1280, 1997.
[87] H. Nassar, S. Pellerin, K. Musiol, O. Martinie, N. Pellerin, and J.-M. Cormier, “N+2 /N2 ratio and
temperature measurements based on the first negative N+2 and second positive N2 overlapped
molecular emission spectra,” Journal of Physics D: Applied Physics, vol. 37, no. 14, p. 1904, 2004.
[88] J. Jonkers, A. Hartgers, L. J. M. Selen, J. A. M. van der Mullen, and D. C. Schram, “The
influence of nitrogen entrainment on argon plasmas created by the ‘Torche á Injection Axiale’
(TIA),” Plasma Sources Science and Technology, vol. 8, no. 1, p. 49, 1999.
[89] E. A. H. Timmermans, “Atomic and molecular excitation processes in microwave induced
plasmas,” Ph.D. dissertation, Eindhoven University of Technology, 1999.
[90] M. Abdallah and J. Mermet, “Comparison of temperature measurements in ICP and MIP
with Ar and He as plasma gas,” Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 37, no. 5,
pp. 391–397, 1982.
[91] J. van der Mullen and J. Jonkers, “Fundamental comparison between non-equilibrium aspects
of ICP and MIP discharges,” Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 54, no. 7, pp.
1017–1044, 1999.
[92] R. S. Van Dyck, C. E. Johnson, and H. A. Shugart, “Lifetime lower limits for the 3 P0 and 3 P2
metastable states of neon, argon, and krypton,” Phys. Rev. A, vol. 5, pp. 991–993, February
1972.
[93] J. Henriques, E. Tatarova, and C. Ferreira, “Microwave N2 –Ar plasma torch. I. Modeling,”
Journal of Applied Physics, vol. 109, no. 2, p. 023301, 2011.
[94] P. Bruggeman, D. C. Schram, M. G. Kong, and C. Leys, “Is the rotational temperature of
OH(A → X) for discharges in and in contact with liquids a good diagnostic for determining
the gas temperature?” Plasma Processes and Polymers, vol. 6, no. 11, pp. 751–762, 2009.
[95] A. Sarani, A. Y. Nikiforov, and C. Leys, “Atmospheric pressure plasma jet in Ar and Ar/H2 O
mixtures: Optical emission spectroscopy and temperature measurements,” Physics of Plasmas,
vol. 17, no. 6, p. 063504, 2010.
[96] D. Venzke, J. Mizeraczyk, H. Szczepanska, and E. Korn, “On the rotational distribution of
N2 in the Ar–N2 -discharge. The 0–0-band of the second positive system,” Beiträge aus der
Plasmaphysik, vol. 23, no. 2, pp. 145–151, 1983.
[97] M. A. A. Clyne, J. A. Coxon, D. W. Setser, and D. H. Stedman, “Electronic excitation of OH
and OD following collision of H2 , HCOOH and D2 O with metastable Ar (3 P2,0 ) atoms,” Trans.
Faraday Soc., vol. 65, pp. 1177–1187, 1969.
[98] J. Raud, M. Laan, and I. Jõgi, “Rotational temperatures of N2 (C, 0) and OH(A, 0) as gas
temperature estimates in the middle pressure Ar/O2 discharge,” Journal of Physics D: Applied
Physics, vol. 44, no. 34, p. 345201, 2011.
[99] J. M. Workman, P. A. Fleitz, H. B. Fannin, J. A. Caruso, and C. J. Seliskar, “A comparative study
of rotational temperatures in a microwave plasma: OH radical versus N+2 ,” Appl. Spectrosc.,
vol. 42, no. 1, pp. 96–100, January 1988.
94
[100] A. I. Al-Shamma´a, S. R. Wylie, J. Lucas, and C. F. Pau, “Design and construction of a 2.45 GHz
waveguide-based microwave plasma jet at atmospheric pressure for material processing,”
Journal of Physics D: Applied Physics, vol. 34, pp. 2734–2741, 2001.
[101] R. Rincón, J. Muñoz, M. Sáz, and M. Calzada, “Spectroscopic characterization of atmospheric
pressure argon plasmas sustained with the Torche à Injection Axiale sur Guide d’Ondes,”
Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 81, no. 0, pp. 26–35, 2013.
[102] V. K. Liau, M. T. C. Fang, J. D. Yan, and A. I. Al-Shamma’a, “A two-temperature model for a
microwave generated argon plasma jet at atmospheric pressure,” Journal of Physics D: Applied
Physics, vol. 36, no. 22, p. 2774, 2003.
95
Vita
Peter J. Hammond
Education
Ph.D. in Aerospace Engineering
Penn State University
(Sept. 2006–Aug. 2013)
• Dissertation title: Characterization and Gas
Temperature Measurements of a Waveguide-Based
Microwave Plasma Torch
• Advisor: Dr. Michael M. Micci
M.S. in Aerospace Engineering
Penn State University
(Sept. 2004–Aug. 2006)
• Thesis title: Investigation of Microwave Air Plasmas
Intended for Enhancing Supersonic Hydrocarbon
Combustion
• Advisor: Dr. Michael M. Micci
B.S. in Aerospace Engineering
Penn State University
(Sept. 2000–May 2004)
• Graduated with High Distinction
• Minored in Physics
Additional Professional Study
• 49th Culham Plasma Physics Summer School, Oxfordshire, UK, 16–27 July 2012.
Publications and Presentations
• “Investigation of a Microwave Plasma Torch,” Poster presentation, Culham Plasma Physics
Summer School, July 2012.
• “Characterization and Gas Temperature Measurements of a Waveguide-Based Microwave
Plasma Torch,” Plasma Sources Science and Technology (in preparation).
• “Review of Microwave Single-Electrode Plasma (SEP) Torches,” Plasma Sources Science and
Technology (in preparation).
Professional Experience
Advanced Propulsion Lab, Propulsion
Engineering Research Center, Penn
State University
Graduate Research Assistant
(Sept. 2004–July 2009)
• Investigated microwave plasma sources for
combustion and materials applications
• Assisted in studies of microwave and RF
electrothermal thrusters and various other
propulsion devices
Department of Aerospace Engineering,
Penn State University
Teaching Assistant
(Jan. 2005–May 2007)
• Assisted course professors during lectures and
assessing students
• Supported students through formal and informal
meetings
Learning Center, Penn State
McKeesport, Penn State University
Peer Tutor
(Sept. 2001–May 2002)
• Provided college-level instruction in mathematics
and physics in a professional one-on-one
environment
Документ
Категория
Без категории
Просмотров
0
Размер файла
14 719 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа