Plasma Chemistry and Catalysis in Gases and Liquids E-Book
165,99 €
Mehr erfahren.
- Herausgeber: Wiley-VCH
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
Filling the gap for a book that covers not only plasma in gases but also in liquids, this is all set to become the standard reference for this topic. It provides a broad-based overview of plasma-chemical and plasmacatalytic processes generated by electrical discharges in gases, liquids and gas/liquid environments in both fundamental and applied aspects by focusing on their environmental and green applications and also taking into account their practical and economic viability.
With the topics addressed by an international group of major experts, this is a must-have for scientists, engineers, students and postdoctoral researchers specializing in this field.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 836
Veröffentlichungsjahr: 2013
Ähnliche
Table of Contents
Related Titles
Title Page
Copyright
Preface
List of Contributors
Chapter 1: An Introduction to Nonequilibrium Plasmas at Atmospheric Pressure
1.1 Introduction
1.2 Coronas and Streamers
1.3 Glow Discharges at Higher Pressures
1.4 Dielectric Barrier and Surface Discharges
1.5 Gliding Arcs
1.6 Concluding Remarks
References
Chapter 2: Catalysts Used in Plasma-Assisted Catalytic Processes: Preparation, Activation, and Regeneration
2.1 Introduction
2.2 Specific Features Generated by Plasma-Assisted Catalytic Applications
2.3 Chemical Composition and Texture
2.4 Methodologies Used for the Preparation of Catalysts for Plasma-Assisted Catalytic Reactions
2.5 Catalysts Forming
2.6 Regeneration of the Catalysts Used in Plasma Assisted Reactions
2.7 Plasma Produced Catalysts and Supports
2.8 Conclusions
References
Chapter 3: NOx Abatement by Plasma Catalysis
3.1 Introduction
3.2 General deNOx Model over Supported Metal Cations and Role of NTP Reactor: “Plasma-Assisted Catalytic deNOx Reaction”
3.3 About the Nonthermal Plasma for NOx Remediation
3.4 Special Application of NTP to Catalytic Oxidation of Methane on Alumina-Supported Noble Metal Catalysts
3.5 NTP-Assisted Catalytic NOx Remediation from Lean Model Exhausts Gases
3.6 Conclusions
Acknowledgments
References
Chapter 4: VOC Removal from Air by Plasma-Assisted Catalysis-Experimental Work
4.1 Introduction
4.2 Plasma-Catalytic Hybrid Systems for VOC Decomposition
4.3 VOC Decomposition in Plasma-Catalytic Systems
4.4 Concluding Remarks
References
Chapter 5: VOC Removal from Air by Plasma-Assisted Catalysis: Mechanisms, Interactions between Plasma and Catalysts
5.1 Introduction
5.2 Influence of the Catalyst in the Plasma Processes
5.3 Influence of the Plasma on the Catalytic Processes
5.4 Thermal Activation
5.5 Plasma-Mediated Activation of Photocatalysts
5.6 Plasma-Catalytic Mechanisms
References
Chapter 6: Elementary Chemical and Physical Phenomena in Electrical Discharge Plasma in Gas–Liquid Environments and in Liquids
6.1 Introduction
6.2 Physical Mechanisms of Generation of Plasma in Gas–Liquid Environments and Liquids
6.3 Formation of Primary Chemical Species by Discharge Plasma in Contact with Water
6.4 Chemical Processes Induced by Discharge Plasma Directly in Water
6.5 Concluding Remarks
Acknowledgments
References
Chapter 7: Aqueous-Phase Chemistry of Electrical Discharge Plasma in Water and in Gas–Liquid Environments
7.1 Introduction
7.2 Aqueous-Phase Plasmachemical Reactions
7.3 Plasmachemical Decontamination of Water
7.4 Aqueous-Phase Plasma-Catalytic Processes
7.5 Concluding Remarks
Acknowledgments
References
Chapter 8: Biological Effects of Electrical Discharge Plasma in Water and in Gas–Liquid Environments
8.1 Introduction
8.2 Microbial Inactivation by Nonthermal Plasma
8.3 Chemical Mechanisms of Electrical Discharge Plasma Interactions with Bacteria in Water
8.4 Physical Mechanisms of Electrical Discharge Plasma Interactions with Living Matter
8.5 Concluding Remarks
Acknowledgments
References
Chapter 9: Hydrogen and Syngas Production from Hydrocarbons
9.1 Introduction: Plasma Catalysis
9.2 Current State of Hydrogen Production, Applications, and Technical Requirements
9.3 Description and Evaluation of the Process
9.4 Plasma-Assisted Reforming
9.5 Summary of the Results and Outlook
References
Index
Related Titles
Rauscher, H., Perucca, M., Buyle, G. (eds.)
Plasma Technology for Hyperfunctional Surfaces
Food, Biomedical, and Textile Applications
2010
Hardcover
ISBN: 978-3-527-32654-9
Kawai, Y., Ikegami, H., Sato, N., Matsuda, A., Uchino, K., Kuzuya, M., Mizuno, A. (eds.)
Industrial Plasma Technology
Applications from Environmental to Energy Technologies
2010
Hardcover
ISBN: 978-3-527-32544-3
Heimann, R. B.
Plasma Spray Coating
Principles and Applications
2008
Hardcover
ISBN: 978-3-527-32050-9
Hippler, R., Kersten, H., Schmidt, M., Schoenbach, K. H. (eds.)
Low Temperature Plasmas
Fundamentals, Technologies and Techniques
2008
Hardcover
ISBN: 978-3-527-40673-9
d'Agostino, R., Favia, P., Kawai, Y., Ikegami, H., Sato, N., Arefi-Khonsari, F. (eds.)
Advanced Plasma Technology
2008
Hardcover
ISBN: 978-3-527-40591-6
The Editors
Prof. Dr. Vasile I. Parvulescu
University of Bucharest
Faculty of Chemistry
Regina Elisabetha Bld. 4-12
030016 Bucharest
Romania
Dr. Monica Magureanu
Nat. Inst. for Lasers,
Plasma and Radiation Physics
Atomistilor Str. 409
077125 Bucharest-Magurele
Romania
Dr. Petr Lukes
Institute of Plasma Physics AS CR, v.v.i.
Dept. of Pulse Plasma Systems
Za Slovankou 3
182 00 Prague
Czech Republic
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Library of Congress Card No.: applied for
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at <http://dnb.d-nb.de>.
© 2012 Wiley-VCH Verlag & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form — by photoprinting, microfilm, or any other means — nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law.
Print ISBN: 978-3-527-33006-5
ePDF ISBN: 978-3-527-64955-6
ePub ISBN: 978-3-527-64954-9
mobi ISBN: 978-3-527-64953-2
oBook ISBN: 978-3-527-64952-5
Preface
Plasma-chemical and plasma-catalytic processes associated with low-temperature plasma generated by electrical discharges in gas, liquid, and gas–liquid environments have recently generated considerable interest. Nonthermal plasmas offer a unique way to initiate chemical reactions in the gas phase as well as in liquids, which have potential for practical utilization in different environmental, biological, or medical applications, and also in energy topics or molecular synthesis. Since plasma-chemical processes are rather nonselective, combination with catalysis can provide improved selectivity, by steering the reactions in the desired direction. Catalyst activation by plasma is different from that in case of conventional heating, and therefore the knowledge of plasma-catalyst interaction represents a key issue both from the fundamental point of view, for the understanding of reaction mechanisms involved in the plasma-catalytic process, and obviously, from the point of view of applications. Promising results have been obtained in environmental applications, where it was found that nonequilibrium plasma generated in electrical discharges at atmospheric pressure and room temperature can be successful in destroying a wide range of air pollutants. Serious attention is also directed to plasma-catalytic applications for hydrogen production, which plays a key role in fuel cell technology, as well as for the conversion of natural gas into syngas or into higher hydrocarbons, which can be used as fuel for transportation and raw material in chemical industry. In this direction, the control of catalyst properties by preparation or treatment techniques, as well as their modifications during plasma-catalytic reactions, catalyst stability, and regeneration processes, are important issues. Another vital issue for environmental research is water pollution. During the past 20 years, promising results have been obtained for the degradation of water pollutants and inactivation of various microorganisms using nonequilibrium plasma generated by electrical discharges in liquids and gas–liquid environments. These discharges have been shown to initiate various chemical and physical processes that have potential for practical utilization in different environmental, biological, or medical applications. For example, electrical discharges were successfully applied to degrade and inactivate a number of organic compounds and microorganisms in water. There are also first successful biomedical applications of discharge plasma in liquids.
This book provides an overview of the basic principles of plasma-chemical and plasma-catalytic processes generated by electrical discharges in gas, liquid, and gas–liquid environments, which is addressed by experts in the fields of plasma physics, plasma chemistry, and plasma catalysis. The book is divided into four major sections containing altogether nine chapters that cover the state of the art of this topic in both fundamental and applied aspects.
The first section contains two introductory chapters (Chapters 1 and 2). The first chapter provides an introduction to the fundamental aspects of nonthermal plasma generated by various types of electrical discharges operating in gas at atmospheric pressure and its properties. Chapter 2 focuses on the analysis of the intrinsic characteristics of the catalysts used in plasma-catalytic processes. The control of catalyst properties by preparation and treatment techniques and factors controlling the catalyst stability and regeneration processes represent other issues analyzed in this chapter. All these aspects are important criteria for the selection of appropriate catalysts for the desired applications.
The Chapters 3–5 give an extensive overview of the plasma-catalytic processes associated with low-temperature electrical discharge plasma in gases and their application for air pollution abatement. Chapter 3 is devoted to nitrogen oxides remediation (deNOx) by plasma-assisted catalysis. Chapters 4 and 5 are dedicated to the decomposition of volatile organic compounds (VOCs) in air using plasma-catalytic systems. Results obtained in different plasma-catalytic systems are discussed, and the interactions between plasma and catalysts as well as the mechanisms responsible for NOx and VOC remediation are addressed.
The Chapters 6–8 present the state-of-art fundamental and applied knowledge on plasma-chemical processes associated with nonequilibrium plasma generated by electrical discharges in liquids and gas–liquid environments. In these chapters, for the first time, a comprehensive overview of the elementary chemical and physical phenomena in low-temperature plasma in liquid and gas–liquid environments is provided, including fundamental mechanisms of plasma generation by electrical discharges in water and gas–liquid environments, chemistry and reaction kinetics of primary and secondary species generated by plasma in water and gas–liquid interfaces, mechanisms of interaction of plasma with chemical and biological content in water, plasma-catalytic processes in water and gas-liquid environments, and environmental and biomedical applications of plasma in water and gas–liquid environments.
Chapter 9 focuses on applications of nonthermal plasma and plasma-catalytic processes in energy conversion. An overview of the current state of hydrogen and syngas production, applications, and technical requirements is presented. Detailed discussions are provided with respect to steam reforming, partial oxidation, and carbon dioxide dry reforming, including coupling to higher hydrocarbons and plasma pyrolysis, as well as combined processes, highlighting the key issues to determine practical and economic viability.
This book is equally addressed to scientists and engineers with research interests in the fields of plasma, chemistry, catalysis, pollution abatement, synthesis of new materials, or energy conversion techniques. It may also be a very good support for students and Ph.D. students performing research in one of these fields.
Monica Magureanu
Petr Lukes
Vasile I. Parvulescu
List of Contributors
Chapter 1
An Introduction to Nonequilibrium Plasmas at Atmospheric Pressure
Sander Nijdam, Eddie van Veldhuizen, Peter Bruggeman, and Ute Ebert
1.1 Introduction
1.1.1 Nonthermal Plasmas and Electron Energy Distributions
Plasmas are increasingly used for chemical processing of gases such as air, combustion exhaust, or biofuel; for treatment of water and surfaces; as well as for sterilization, plasma deposition, plasma medicine, plasma synthesis and conversion, cleaning, and so on. These plasmas are never in thermal equilibrium – actually, we know of no exemption – and this fact has two main reasons.
In this manner, the nonthermal nature of the plasmas that are created electromagnetically is made into an asset. By varying gas composition, electrode and wall configuration, and circuit characteristics more energy can be channeled into specific excitations and reactions. Recent examples include the optimization of the pulsed power source for ozone generation in streamer corona reactors [1], or dual frequency RF-generated plasmas [2].
To elaborate the physical understanding further, Mark Kushner has proposed a workshop at the Gaseous Electronics Conference (GEC) 2011 on how the electron energy distribution within a discharge can be tailored for a specific application. A joint approach to this question by theory and experiment now seems within reach because of the large progress of theory in recent years.
1.1.2 Barrier and Corona Streamer Discharges – Discharges at Atmospheric Pressure
The past has mainly seen an experimental approach by trial and error, also guided by some physical understanding. Within the limited space available here, we will review some setups and their physical mode of operation. A common theme is the avoidance of plasma thermalization in the form of arcs and sparks. Variations over two basic approaches are used very commonly and will make the main theme of this review: the corona discharge and the barrier discharge. In a barrier discharge, large currents are suppressed by dielectric barriers on the electrodes. Basically, the discharge evolves only up to the moment when so much charge is deposited on the insulator surfaces that the field over the gas is screened. In a corona discharge, the discharge expands from a needle or wire electrode into outer space where the electric field decreases and finally does not support a discharge anymore. The discharge then has to feed its current into the high-ohmic region of the nonionized gas, which limits the current as well. These two basic principles have seen many variations in the past years and decades. For example, in corona discharges, short and highly ramped voltage pulses create much more efficient streamers that do not cease due to the spatial decrease of the electric field away from the curved electrode but due to the final duration of the voltage pulse.
Both discharge types can (but need not) operate at atmospheric pressure. This poses an advantage as well as a challenge. The advantage lies in the fact that no expensive and complex vacuum systems are required. This makes the design of any reactor a lot simpler, not only when the operating gas is air but also when other gases (such as argon or helium) are used. The challenge consists of the observation that characteristic length scales within the discharge can be much smaller than the discharge vessel and that the discharge can therefore form complex structures, rather than a more or less uniform plasma. These structures have to be understood and used appropriately. For instance, the initial evolution of streamer discharges follows similarity laws [3]: when the gas density is changed, the same voltage will create essentially the same type of streamer, but on different length and timescales. Therefore, streamer fingers and trees grow in a similar manner at 10 µbar as at 1000 mbar, but 10 µbar corresponds to an atmospheric altitude of 83 km where the so-called sprite streamers have a diameter of at least ∼10 m, while at 1000 mbar, the minimal streamer diameter is ∼ 100 µm and conveniently fits into typical experiments.
1.1.3 Other Nonthermal Discharge Types
There is a large variety of nonthermal plasmas. They can be classified into different discharge types, although definitions used by different authors vary significantly. The plasmas or discharges can be classified according to their time dependence (transient or stationary), importance of space charge effects or of heating of the neutral gas species, and presence of a surface close to the discharge. The most important nonthermal plasmas along with their energization method and typical applications are listed in Table 1.1.
Table 1.1 Overview of Nonthermal Discharge Types and Their Most Common Applications
This table is intended to give a general idea, but it is far from complete. A further complication is that definitions are used in different ways. For example, in Ref. 8, Braun et al. use what they call a microdischarge for ozone generation, whereas the microdischarges as intended in Table 1.1 are much smaller. The microwave discharge made by Hrycak et al. [28] qualifies much more for the term plasmajet than for microdischarge. More information on the different types of microdischarges is given in [29]; some examples of the use of microdischarges are given in Section 1.4.4.
In many transient discharges, the different discharge types can occur after each other. For example, a discharge can start as an avalanche and then become a streamer, which can develop into a glow and finally into an arc discharge. When applying a DC field between two metal electrodes, a discharge at high pressure will become a thermal arc if the power supply can deliver the current. Nonthermal discharges are, by definition, almost always transient.
An essential feature of a cold nonthermal discharge is its short duration. Therefore, the largely varying timescales of the processes inside the discharge must be considered. The excitation timescales, which often range from picoseconds to a few microseconds, are clearly not the timescale necessary for preventing thermalization as thermalization occurs in millisecond-order timescales. The critical timescale is basically the characteristic time of the glow-to-spark transition. This transition time can highly depend on conditions such as voltage amplitude and gas composition but is often in the order of a (few) hundred nanoseconds [30]. Dielectric barrier discharges (DBDs) are a well-known example of how (dielectric) barriers can reduce current density and ne to keep the gas temperature of the discharge low.
Like streamer and avalanche discharges, Townsend and glow discharges are cold discharges. They usually occur as a stationary discharge but have to be preceded by another discharge such as a streamer or avalanche discharge to ignite. In Townsend and glow discharges, electrons are emitted from the electrode and are then multiplied in the gap. In the case of a Townsend discharge, the electron multiplication takes place in the whole gap, while in a glow discharge, space charge concentrates the multiplication in the cathode sheath region. Electrons are freed from the cathode by the temperature of the cathode itself or by secondary emission either due to the impact of energetic positive ions or due to photons or heavy neutrals.
Several cold atmospheric pressure discharges operate in helium. This is not a coincidence as He has a thermal heat conductivity that is about 10 times larger than that of most other gases, which renders heat removal from the discharge to be more efficient. Other methods for efficient heat removal include strongly forced convection cooling in flow stabilized discharges and creation of discharge with a large area-to-volume ratio (microplasmas, see also further) to make the heat losses to the walls more efficient.
1.1.3.1 Transition to Sparks, Arcs, or Leaders
Avalanches, Townsend, streamer, and glow discharges are examples of cold discharges. This means that the heavy particle temperature is not much above room temperature and definitely far below the electron temperature (Te Ti ≈ Tn where e,i, and n stand for electron, ion, and neutral, respectively). At even higher currents, at higher pressures, or with longer pulse durations, these discharges can transform into spark, arc, or leader discharges. These are hot discharges, the heavy particle temperature is close to the electron temperature and can reach thousands of Kelvin (Te ∼ Ti ≈ Tn). In applications, heating of the gas is often unwanted, and therefore, cold discharges are preferred in many plasma treatment applications.
1.1.4 Microscopic Discharge Mechanisms
1.1.4.1 Bulk Ionization Mechanisms
The main ionization mechanism in electric discharges is impact ionization; in attaching gases such as air, impact ionization is counteracted by electron attachment. Other mechanisms that create free electrons such as photoionization or electron detachment from negative ions are discussed in Section 1.2.4.1. Impact ionization occurs when electrons are accelerated in a high local electric field. At a certain kinetic energy, they can ionize background gas atoms or molecules and create more electrons. In air, this occurs by the following reactions:
1.1
1.2
1.3
This notation illustrates that the Townsend coefficient is characterized by two parameters: E0 characterizes the electric field where impact ionization is important; this electric field is proportional to the gas density n0. α0 characterizes the inverse of the ionization length at these fields. More precisely, 1/αi(|E|) is the mean length that an electron drifts in the field E before it creates an electron–ion pair by impact. Therefore, in geometries smaller than this length, no gas discharge can occur. Both the electron mean free path, between any collision, and the ionization length scale with inverse gas density.
The electron loss rate due to electron attachment on attaching gas components has a similar functional dependence as the impact ionization rate, but different parameters. One needs to distinguish between dissociative attachment
1.4
and three-body attachment
1.5
where M is an arbitrary third-body collider, for example, N2 or O2. As a third body is required here to conserve energy and momentum, the importance of three-body attachment relative to dissociative attachment increases with density. Dissociative attachment scales with gas density in the same manner as the impact ionization reaction, while three-body attachment is favored at higher gas density. On the other hand, dissociative attachment becomes more important at higher electric fields, even at standard temperature and pressure. For detailed discussions of the derivation of these reaction coefficients, we refer to [33–36].
The breakdown field is defined as the field where impact ionization and electron attachment precisely balance; at higher electric fields, an ionization reaction sets in. The spatial and temporal evolution of the discharge depends on the distribution of electrons and electric fields; this is discussed in more detail below.
1.1.4.2 Surface Ionization Mechanisms
Next to the bulk gas, the presence of a dielectric or metallic surface can also affect the discharge significantly. It will modify the electric field configuration, and it is able to provide electrons. Dielectrics can also store surface charges [37] and prevent charge carrier flow through the surface.
Electrons can be freed from a surface by high fields or by secondary emission on impact of ions [38], fast neutrals, or (UV) photons [39]. Photons can be generated in the bulk of the discharge and then free an electron from the surface. Electron emission can be enhanced by the local electric field at the surface or by higher surface temperatures. The freed electrons can form the start of an avalanche, which enables the discharge to initiate or propagate (over the surface). See Section 1.4.3 for a more elaborate discussion on this topic.
1.1.5 Chemical Activity
The main advantage of nonthermal plasmas is their high chemical efficiency. As little or no heat is produced, nearly all input energy is converted to energetic electrons. This is in contrast to thermal plasmas in which the heating itself leads to higher thermal losses and thereby can be a waste of energy, which reduces the chemical efficiency of these hot plasmas [40] and can damage walls and other nearby surfaces (such as the substrate in a surface processing application). Furthermore, higher gas temperatures will change the reaction kinetics which, amongst others, may lead to breakdown of ozone and increased formation of NOx. Of course, the different reaction kinetics of higher gas temperatures can also be beneficial for some chemical reactions such as destruction of hydrocarbons.
The fast electrons produced in a nonthermal plasma can have energies of the order 10 eV or even higher and can therefore trigger many different chemical processes. Besides fast electrons, energetic photons can also play a role in the reactions in a nonthermal plasma. One important example of such a reaction is photoionization in air, which is discussed in detail in Section 1.2.4.1. However, the primary source of all reactions is electron impact on the bulk gas molecules, which leads to many reactive species that can than further react with more stable species. Examples of the reactive species are OH, O, and N radicals; excited N2 molecules; and atomic and molecular ions (e.g., O+, ).
One of the main paths of chemical activity in nonthermal plasmas in air is ozone production. This is generally believed to be a two-step process as described by Chang et al. [41] and Ono and Oda [42].
1.6
1.7
1.8
1.9
Ozone can be produced with a wide range of electrode and discharge topologies, many of which are treated below; the most popular are dielectric barrier discharges. An early example is the ozone generator of Siemens made in 1857. The most important application of this device was ozone production for disinfection of water. Even now, this device is used, with only minor modifications [43]. But corona discharges can create O radicals (and thereby ozone) with very high energy efficiency as well [1], as will be discussed in more detail further below. In commercial ionizers, pure oxygen is often used as the starting gas because the nitrogen that is present in air can lead to the formation of NOx (a general term used for NO and NO2 and sometimes other nitrogen–oxygen compounds) with the following reactions [44]:
1.10
1.11
where the O radicals come from Eqs. (1.6, 1.7, 1.8) and the N radicals are produced by [45]
1.12
The produced NO can further react with NO2 as described in [45, 46]
1.13
1.14
1.15
However, nonthermal plasmas can also remove NO from gas streams. The main path for the removal of NO from air at low NO concentrations is (Eq. (1.12)) followed by [47]
1.16
A second type of radical that is important in nonthermal plasmas is OH. This is produced in moist gases (e.g., moist air) by the following reaction [48]:
1.17
Note that apart from electron-induced dissociation, dissociative electron recombination of water containing ions can also efficiently produce OH.
1.18
The rate of this reaction for nonthermal discharges with Te in the range 1–2 eV is sometimes even faster than electron dissociation [49]. Several secondary reactions are also believed to play an important role in the production of OH
1.19
1.20
where O(1D) is an excited state of atomic oxygen, N2(A) is a metastable nitrogen molecule and N2(X) is a nitrogen molecule in the ground state. It is clear that Eq. (1.17) occurs only in the ionizing phase, while Eqs. (1.18, 1.19, 1.20) also occur in the recombining phase when the electron temperature is equal to the gas temperature.
Which reactions dominate depends on the electron energy (which is dependent on topology, voltage shape, and amplitude, etc.) and the composition of the gas. In general, thermal discharges mostly produce NOx, while nonthermal discharges produce ozone instead and can remove NOx when concentrations are high. At low NOx concentrations also, nonthermal discharges can lead to the net production of NOx. A comparison of NOx production by sparks and corona discharges was performed by Rehbein and Cooray [50]. They found that sparks produce about 2 orders of magnitude more NOx per Joule than corona discharges. Overviews of different reactive species and the conditions in which they are important are given by Eliasson and Kogelschatz [51] and Kim [43].
Besides NOx removal, which was discussed above, a host of other species can be removed from gas streams by nonthermal plasmas. Examples are volatile organic compounds (VOCs), chlorofluorocarbons (CFCs), SO2, odors, and living cells (in disinfection or sterilization).
Most charges in a nonthermal discharge in air are initially produced by the direct impact ionization of nitrogen
1.21
with a threshold ionization energy of 15.58 eV or of oxygen (Eq. (1.1)) with a threshold ionization energy of 12.07 eV. According to Aleksandrov and Bazelyan [52], and will quickly change to other species according to the following scheme (for dry air under standard conditions):
1.22
After some tens of nanoseconds, the positive ions are dominated by . Electrons are quickly attached to molecular oxygen by reactions given in Eqs. (1.4) and (1.5).
1.1.6 Diagnostics
In all nonthermal plasmas, fast electrons excite species. Many of the excited species can fall back to lower excited levels or the ground level and thereby emit a photon. These photon emissions are by far the most important property of cold discharges that are studied experimentally. They are used for imaging and for optical emission spectroscopy. Spectra of cold discharges in air are dominated by the emissions of the second positive systems of N2 (SPSs, upper states B3Πg and C3Πu). The SPS is often used to obtain the rotational temperature, which is mostly a good indication of the gas temperature [53].
For strongly pulsed and high field discharges and also in discharges in, for example, He with air impurities, the first negative system of (FNS, upper state ) readily occurs. Relative intensity comparisons of the SPS and this FNS have been performed by many authors and are used to determine the electric field in nitrogen-containing discharges. This method is employed, for example, by Kozlov et al. [54] for laboratory scale discharges and by Liu et al. [55] for sprites.
There are many other rotational bands of different molecules that can be used to obtain rotational temperatures, which are mostly a good indication of the gas temperature. Especially popular is the UV emission band of OH(A–X) around 309 nm [53]. However, it has recently been found that the rotational population distribution is not always in equilibrium with the gas temperature and sometimes leads to overestimates [56].
Electron densities above 1020 m−3 can be determined by measuring the Stark broadening of the hydrogen Balmer lines. Especially the Balmer β line is very popular. It is important to note that it is necessary to carefully take into account all broadening mechanisms including van der Waals broadening, which can become quite important for low-temperature atmospheric pressure plasmas. A detailed description can be found in [53].
Besides (passive) optical emission spectroscopy, there are many other techniques to study nonthermal plasmas. Apart from standard voltage and current waveform measurements, several electrical probes exist, especially developed for low pressure plasmas, although it is often difficult and very complicated to apply them on atmospheric pressure plasmas. The active laser spectroscopy techniques have developed into a wide field. The techniques most commonly applied to atmospheric pressure plasmas include laser-induced fluorescence (LIF) and two-photon-absorption laser-induced fluorescence (TALIF), which are good ways to obtain information on the chemical composition of radicals. With proper calibration, even absolute densities can be obtained [57, 58]. Other well-known laser-based techniques are based on scattering of photons. Thomson scattering can give direct information on the electron density and temperature [59, 60]. Rayleigh and Raman scattering provide information on gas density and temperatures. The conceptually simplest active technique is absorption spectroscopy (often also performed with lasers). This technique is used to determine absolute densities of certain species, often in the ground state (e.g., OH). Radical density fluxes can also be obtained by appearance potential mass spectrometry [61]. Mass spectrometry also gives the possibility to measure the ion flux of one of the electrodes directly and determine the ion composition of the plasma [62].
1.2 Coronas and Streamers
1.2.1 Occurrence and Applications
Streamers are the earliest stage of electric breakdown of large nonionized regions. They precede sparks and create the path for lightning leaders; they also occur as enormous sprite discharges, far above thunderclouds. Streamers and the subsequent electric breakdown are a threat to most high-voltage technology.
However, streamers are also used in a variety of applications and are appreciated for their energy-efficient plasma processing. The following is an (incomplete) application list:
Gas and water cleaning: The chemical active species that are produced by streamers can break up unwanted molecules in industrially polluted gas and water streams. Contaminants that can be removed include organic compounds (including odors), NO
x
, SO
2
, and tar [3, 6, 63, 64].
Ozone generation: By simply applying a streamer discharge in air, first O* radicals and then ozone is created. The low temperature in a streamer discharge limits the destruction of the produced ozone. The ozone can be used for different purposes such as disinfection of medical equipment, sanitizing of swimming pools, manufacturing of chemical compounds, and more [4].
Particle charging: A negative DC corona discharge can charge dust particles in a gas flow. These charged dust particles can now be extracted from the gas by electrostatic attraction. Such a system is called an electrostatic precipitator (ESP) and is used in the utility, iron/steel, paper manufacturing, and cement and ore-processing industries. Similar charging methods are used in copying machines and laser printers [4, 65].
A corona discharge is (an often DC-driven) discharge in which many streamers are initiated from one electrode and, depending on the conditions, may or may not reach another electrode. The name corona comes from the crownlike appearance of the many streamer channels around the primary (driven) electrode.
Traditionally, DC corona discharges are classified in several different forms depending on the field polarity and electrode configuration [41]. In case of a positive point-plane discharge, one can recognize the burst pulse corona, streamer corona, glow corona, and spark for an increase in applied voltage. In a negative point-plane corona, this is replaced by a Trichel pulse corona, a pulseless corona, and again, a spark.
Since the 1980s, corona discharges are separated into two different categories: continuous and pulsed. Continuous corona discharges occur at DC or low-frequency AC voltages. If the circuit providing the voltage can support high currents, these will transform into a stationary glow or spark discharge. Therefore, continuous corona discharges can only occur if the current is limited. One example is a continuous corona discharge around high-voltage power lines, where the large gap to the ground limits the current. A recent example of work on DC-excited corona discharges is by Eichwald et al. [66].
The current of a continuously excited corona is often spiked because the discharge is not really continuous but is self-repetitive in nature. In such a self-repetitive corona, the discharge stops itself due to the buildup of space charge near the electrode tip. Only after this space charge has disappeared by diffusion and drift will a new discharge occur [67].
A pulsed corona is produced by applying a short (usually submicrosecond) voltage pulse to an electrode. Its practical advantages are that the short duration of the pulse ensures that no transition to spark takes place, therefore it can be used at voltages and currents higher than that at continuous corona can be used.
Shang and Wu [68] have shown that a positive-polarity-pulsed corona removes more NO than a negative polarity discharge. van Heesch et al. [1] show that negative coronas have a higher efficiency in the production of O* radicals (about a factor of 2 higher).
In laboratory studies of corona discharges, the most popular geometry is a point-plane geometry (Figure 1.1a), where a needle is placed above a grounded plane. The high voltage (pulse) is applied to the needle electrode. However, for industrial applications, this geometry is not sufficient, as it does not fill the whole gas volume with the discharge. The most popular geometries in industrial applications are the wire-cylinder, wire-plate, and the saw-blade geometries [41, 69]. See Figure 1.1b–e for schematic images of these geometries.
Figure 1.1 Schematic depictions of popular electrode geometries in corona reactors: (a) point-plane, (b) wire-cylinder, (c) double sawblade, (d) sawblade-plane, and (e) wire-plane. The high voltage is applied to the following parts: (a) top needle, (b) central wire, and (c–e) top sawblade/wires. The other parts are grounded.
The wire-cylinder geometry is probably used the most. It ensures a quite homogeneous distribution of the discharge and is easy to implement in a gas-flow system. Often, multiple wire-cylinder reactors are mounted in parallel with regard to the gas flow to enable high gas throughput.
1.2.2 Main Properties of Streamers
Streamers are rapidly extending ionized fingers that can appear in gases, liquids, and solids. They are generated by high electric fields but can penetrate into areas where the background electric field is below the ionization threshold due to the strong field enhancement at their tip. The mechanism of field enhancement is illustrated in Figure 1.2, which shows the simulation of a positive streamer in air at standard temperature and pressure; for details we refer the readers to [70, 71]. The plots show electron and ion density, space charge, and field distribution. The plots can be understood as follows. Panels (a) and (b) show that the interior of the streamer channel consists of a conducting plasma with roughly the same electron and ion densities. The electric field (panel d) in this ionized area is largely screened by the thin space charge layer shown in panel (c).1 In front of the ionized finger, the space charge layer is strongly curved, and therefore, it significantly enhances the electric field in the nonionized area ahead of it. This self-organization mechanism due to space charge effects makes the streamer a well-defined nonlinear structure; gas heating is negligible in most cases.
Figure 1.2 Structure of positive streamers shown by zooming into the relevant region of a simulation by Ratushnaya et al. The panels show (a) electron density ne, (b) ion density n+, (c) space charge density (n+ − ne), as well as (d) electric field strength E and equipotential lines φ. The letters in (c) indicate the streamer regions: H – streamer head, I – interior, and W – wall of the streamer channel.
(Source: Image from Ref. [71].)
As described in a previous streamer review for geophysicists [3], the electrons in the high-field zone at the streamer head are very far from equilibrium. The electron energy distribution can develop a long tail at high energies, and it is now known that electrons at the tip of negative streamers can even run away [31, 72–75], if the field enhancement is above 180 kV cm−1 in STP air, corresponding to 720 Td. This is the current explanation for the hard X-rays emitted during the early streamer-leader phase of MV-driven pulses [76]. How to optimize the electron energy distribution for a particular plasma processing purpose is a current research question.
The fact that streamer velocities and diameters can vary substantially between different electrode geometries and electric circuits is by now well established [5, 77, 78]. Simulations show that the maximum of the enhanced electric field also varies substantially, as reviewed recently in [79].
The maximal field determines the ionization rate inside the streamer [31, 70, 80] and, therefore, the excitation rates for gas processing purposes. The search for optimal processing conditions determined by both the electron energy distribution and the ionization rate is currently underway, both theoretically and through the development of optimized electric circuits. Here it should be mentioned that very short voltage rise times create much thicker [5, 77, 78] and more efficient streamers [1].
An important distinction is between positive and negative streamers, where the polarity refers to the net charge at their tips (Figure 1.3). They are also known as cathode- or anode-directed streamers. A negative streamer moves in the electron drift direction, and as the streamer velocity is frequently comparable to the local electron drift velocity,2 its motion can be explained by purely local mechanisms. On the contrary, a positive streamer moves, in most cases, even faster [78]. The reason for this counterintuitive behavior lies in the fact that the relative immobility of the ions in the space charge layer around the positive streamer keeps the streamer finger thin and focused; therefore the electric field at the tip can be much higher [81]. The mechanism allowing positive streamers to propagate is explained below.
Figure 1.3 Illustration of downward propagating positive (a) and negative (b) streamers. The plus symbols indicate positive ions, while the minus symbols indicate negative ions or free electrons.
Concerning theory and simulations, there are currently three models: (i) Monte Carlo and (ii) hybrid models that follow the single-electron dynamics within a streamer, but are still constrained to rather short streamers, fluid, or density models, which now also start to treat the interaction of streamers, but cannot resolve the electron energy distribution, and (iii) moving boundary models where the thin space charge layer around the streamer is treated as a moving boundary. Currently, reviews of all three model classes have been published or are under review; we refer the reader for details to [71, 82, 83].
1.2.3 Streamer Initiation or Homogeneous Breakdown
When a discharge starts to develop, there are only few free charge carriers present, and therefore the electric field is not modified by space charge effects yet. The discharge is then said to be in the avalanche phase where free charge carriers multiply in regions where the electric field is above the breakdown value.
The discharge can then evolve either in a more homogeneous or a more streamerlike manner. If the initial ionization seed is very localized (e.g., because it evolves out of a single electron or because a macroscopic seed is ejected form a pointed needle electrode), or if the electric field is above breakdown only in a small part of space (again, e.g., close to a needle electrode), a localized structure such as a streamer that carries a field enhancement forward at its tip can emerge. On the other hand, if there is a higher level of preionization and if the electric field is at most places above the breakdown value, a more homogeneous discharge will emerge [84].
If a single electron or a very localized seed is placed in a homogeneous field above the breakdown value, Raether and Meek estimated in the late 1930s, that space charge effects set in and a streamer initiates when the total number of free electrons reaches 108 − 109 in air at standard temperature and pressure [85, 86]. However, this estimate is independent of the electric field. Taking into account that an electron avalanche grows with a slower rate in a weaker field, but that their diffusive broadening is essentially the same, a correction to the so-called Raether–Meek criterion was developed by Montijn and Ebert [87].
However, in most streamer experiments and applications, streamers are generated from a tip- or wirelike structure and not in a homogeneous field. At such a (sharp) tip or wire, the electric field will be greatly enhanced, which makes it easier to initiate a streamer. After initiation, the streamer can propagate into the rest of the gap where the background field may be too low for streamer initiation, but high enough for streamer propagation (discussed in the next section). Such a geometry with field enhancement greatly reduces the required voltages for streamer initiation, which makes experiments and applications smaller, cheaper, and easier to operate.
The lowest voltage at which a streamer can initiate from a pointed electrode is called the inception voltage; it depends on electrode shape and material as well as on gas composition and density and (up to now) has no direct interpretation in terms of microscopic discharge properties yet.
1.2.4 Streamer Propagation
After initiation, a streamer will propagate under the influence of an external electric field augmented by its self-generated field, as already discussed in Section 1.2.2. To sustain the extension of the plasma channel by impact ionization in the high-field zone, enough free electrons need to be present there. In negative streamers, the electrons drift from the ionized region in the direction of streamer propagation and reach the high-field zone. However, in positive streamers, the electrons cannot come from the streamer itself. Therefore, for positive streamer propagation, “fresh” electrons are needed in front of the streamer head. The possible sources of these free electrons are discussed below.
As was discussed in Section 1.1.5, the positive charges indicated in Figure 1.3 will mainly consist of positive molecular ions and the negative charges indicated in the streamer tails in air in Figure 1.3 will be negative molecular oxygen ions, limiting the total conductivity. Therefore, streamers in pure nitrogen can become longer than those in air under similar conditions as less electron attachment occurs if current flow from behind is required. The negative charges in the streamer head, as well as the moving charges in front of the streamer heads, will be mostly free electrons.
Owing to the electric screening layer around the curved streamer head, the electric field ahead of it is usually much higher than the external or background field.
1.2.4.1 Electron Sources for Positive Streamers
Positive streamers need a constant source of free electrons in front of them in order to propagate. Because of the electronegativity of molecular oxygen, free electrons in air quickly attach to oxygen by Eqs. (1.4) and (1.5) if the electric field is below ∼ 30 kV cm−1. If this is the case, a high field is needed to detach the electrons so that they can be accelerated. The exact level of the detachment field depends on the vibrational excitation of the molecule. According to Pancheshnyi [88] and Wormeester et al. [89], a good value of the instant detachment field under standard conditions in air is 38 kV cm−1.
Photoionization
In most streamer models, air is the medium and the major source of electrons in front of the streamer head is taken as photoionization. In air, photoionization occurs when a UV photon in the 98–102.5 nm range, emitted by an excited nitrogen molecule, ionizes an oxygen molecule, thereby producing a free electron.
1.23
1.24
As the emitted photon can ionize an oxygen molecule some distance away from its origin, this is a nonlocal effect, therefore, excited nitrogen molecules in the streamer head can create free electrons in front of the streamer head (as well as in other places around the streamer head). The average distance that a UV photon can travel depends on the density of the absorbing species, oxygen in this case. In atmospheric pressure air under standard conditions, this distance will be about 1.3 mm [90].
Background Ionization
Besides photoionization, there is another source that can provide free electrons in front of a positive streamer head: background ionization. Background ionization is ionization that is already present in the gas before the streamer starts, or at least, it is not produced by the streamer. It can have different sources. In ambient air, radioactive compounds (e.g., radon) from building materials and cosmic rays are the most important sources of background ionization. They lead to a natural background ionization level of 109 − 1010 m−3 at the ground level (Pancheshnyi [88]).
Another source of background ionization can be leftover ionization from previous discharges. This is especially important in repetitive discharge types such as DC corona discharges or repetitive pulsed discharges. Already at a slow repetition rate of about 1 Hz, leftover charges can lead to background ionization densities of the order of 1011 m−3. Background ionization can also be created by external UV radiation sources, X-ray sources, addition of radioactive compounds to the gas or surfaces, electron or ion beam injection, and more.
Independent of the source of background ionization, in air, the created electrons will always quickly be bound by oxygen. This means that they will have to be detached by the high field of the streamer before they can be accelerated and form avalanches.
1.2.5 Initiation Cloud, Primary, Secondary, and Late Streamers
Recent imaging with high spatial and temporal resolution has shown how a streamer tree starts from a needle electrode, which in most cases is positively charged [91–93]. The discharge starts with a small ball of light around the needle tip that was called the initiation cloud. This ball expands and forms a shell; this shell can be interpreted as a radially expanding ionization front, and in the case of a negative needle tip in air, its maximal radius fits the theoretical estimates well [93]. For positive voltages, it has been verified that the size l of the initiation cloud scales with gas density n0 according to the similarity laws (l∝1/n0) but it also depends on gas composition and, of course, on the applied voltage. For example, in air, the initiation cloud is much larger (up to a factor 10 or more) than in pure nitrogen [91]. In fact, what on time integrated images of the discharge seems like a light emitting cloud is in fact often a smaller cloud that transforms into a thin expanding shell.
Eventually, the expanding shell breaks up into multiple streamer channels, except when the gap is so small that the initiation cloud extends into roughly half the gap distance; in that case, it usually destabilizes into one channel only. These first streamers emerging from the initiation cloud are called primary streamers. Example of such streamers are shown in Figure 1.4a,b. For long gaps, low voltages, or short pulse durations, the primary streamers often do not reach the other side and extinguish somewhere between the electrodes.
Figure 1.4 Streamer discharges in a 40 mm gap in atmospheric air with a 54 kV pulse, 30 ns risetime, and half-width of about 70 ns. The images are acquired with short (a,b) and long (c) exposure times. The exact image start delay is varied between (a) and (b) (exact values unknown).
(Source: Images by Tanja Briels, originally published in Figure 6 of Ref. [77].)
Briels et al. [77, 94] characterize different streamer types with very different diameters and velocities, although they realize and later show [78] that there is no phase transition between these types. For voltages between 5 and 95 kV, the streamer diameters vary by more than an order of magnitude and the velocities by almost 2 orders of magnitude. The relation between velocity and diameter is discussed in [79, 81]. The streamers with minimal diameter (the so-called minimal streamers) are never seen to branch. This minimal diameter depends on density, roughly in agreement with the similarity laws [3], but it does not depend on the background field or other pulse parameters. This concept was proposed by Ebert et al. [95]. The thick streamers grow only if the voltage rises sufficiently fast. Only then there is sufficient voltage initially on the pointed electrode to develop a very wide ionization cloud that can eject fat streamers.
After the primary streamer, more light-emitting discharge phenomena can occur. If the same streamer channels reilluminate rather immediately, one speaks of a secondary streamer, while if a streamer follows a different track at some later time, one speaks of a late streamer.
Secondary streamers have been described, for example, by Marode [96], Sigmond [97], Ono and Oda [98], or Winands et al. [5]. Sigmond remarks that moving secondary streamer fronts in centimeter-scale gaps in atmospheric air does not perturb the smoothly decaying streamer current and that they are only reported in air. Ono and Oda [98] have compared primary and secondary streamers; they were created in air in a needles-to-plane geometry with gaps of 13 mm length and voltages of 13–37 kV (compare with the 37–77 mm, 25–45 kV wire plane discharge of Winands et al.). They observe that emission from the FNS of (391.4 nm) is only observed in primary streamers and not in secondary streamers. This is attributed to the fact that electron energies required for propagation of primary streamers are higher than those for secondary streamers as primary streamers have to create ionization, while secondary streamers propagate along the ionized channel created by the primary streamers. Furthermore, they find that secondary streamers only occur at higher voltages (15 kV in air and 20 kV in pure nitrogen). van Heesch et al. [1] found that the O* radical yield from primary steamers is up to two times higher than that from secondary streamers. They explain this by higher local electric fields and electron energies in the primary streamers.
The literature presents different suggestions for the physical mechanism of secondary streamers. Marode [96] suggests that secondary streamers correspond to a moving equivalent of the positive column of a glow discharge. Sigmond [97] suggests that the ionized column created after the primary streamer has crossed the gap decays into one region with high and another region with low electric field due to an attachment instability. The electrodynamic consistency of these calculations is under examination at present. A different mechanism is suggested by recent simulations of Liu [99] and Luque and Ebert [80]. They find that inside a streamer that requires a growing charge in its tip – because it accelerates and expands or because it propagates into a region with higher gas density – a secondary ionization wave can set in, and that the electric field inside this wave reaches approximately the breakdown field. This process can set in before the primary streamer has reached an electrode. We note that in the experiments of Winands et al. [5] where long secondary streamers were observed, the primary streamers were accelerating and expanding as well, just like in the simulations of Liu.
A third streamer category, besides primary or secondary streamers, is the so-called late streamers. They occur only for long enough pulses and are, in fact, the primary streamers that either start later than the dominant streamers or are so slow that they seem to have started later. Late streamers propagate along completely different paths than the other (primary) streamers before them. They are often very thin, which is related to their slow propagation velocity (see, e.g., Briels et al. [78]). In most cases, they do not appear from the sharp electrode tip itself but instead from the (less sharp) edges of the electrode or electrode holder because the tip is already screened by a glow region and therefore no longer enhances the electric field sufficiently. Examples of these late streamers are visible in Figure 1.4b,c. In Figure 1.4b, the late streamers have just started and are visible on the top of the image. In Figure 1.4c, a much longer camera exposure is used. Therefore the primary streamers are now overexposed as their secondary and glow phase is also included in this exposure. However, many (thin) late streamers are clearly visible crisscrossing all corners of the image.
1.2.6 Streamer Branching and Interaction
Most streamer discharges contain more than one streamer channel. Therefore, interactions between streamers are important when studying streamer behavior. One important aspect is streamer branching where one streamer channel splits into two (or more) channels. Other interactions are attraction and repulsion of streamer channels. Furthermore, neighboring channels influence each others field configuration. If attraction occurs, this may lead to streamer merging or (re-)connection. Discussion and measurements regarding streamer merging and (re-)connection are given by Nijdam et al. [100, 101].
Branching is observed in most streamer discharges, except when the gap is so short that the streamer has reached the other side before it has branched. Furthermore, streamers of minimal diameter (so-called minimal streamers, see below) also do not branch but eventually extinguish. This is the main argument why streamer discharges are never real fractals.
The mechanism of streamer branching has been under investigation for quite a long time now. It is certainly due to a Laplacian instability of the thin space charge layer visible in Figures 1.2 and 1.3; this instability bears strong mathematical similarities with viscous fingering [102]. For a recent review of the analytical, numerical, and experimental results, we refer to [71]. The Laplacian instability can actually set in without any stochastic effects [102, 103]. However, the branching instability can be accelerated by electron density fluctuations in the lowly ionized region ahead of the streamer [104]; these fluctuations are due to the discrete quantum nature of the electrons. Indeed, these fully three-dimensional recent simulations for positive streamers in air (with the standard photo-ionization model) show a ratio of streamer branching length to streamer diameter similar to that obtained in experiments [91, 100].
The acceleration of branching through electron density fluctuations is consistent with older concepts, which can be traced back to Raether [86] and Loeb and Meek in 1940 [105]. However, in these older sketches, the fact that the streamer has to develop a thin space charge layer before it can destabilize was missed. The older concept that can be found in many books emphasizes the spatially well-separated avalanches ahead of the streamer as direct precursors of different branches. Such avalanches have now indeed been seen in very pure gases [89, 106]. However, the photoionization density in air is much too large to create individual avalanches [107].
van Veldhuizen and Rutgers [108] have experimentally investigated streamer branching in argon and ambient air for different discharge geometries and pulse characteristics. They find that streamers in a point-wire discharge branch about 10 times more often (in the middle of the gap) than in a discharge between a plane with a protrusion and another plane.
A very different branching mechanism is branching at macroscopic inhomogeneities such as bubbles (for streamers in liquids). This mechanism was recently described in detail by Babaeva and Kushner [109].
A proper understanding of streamer branching, on the one hand, and streamer thickness and efficiency, on the other hand, is required to understand which volume fraction of gas is being processed in a streamer corona reactor. The streamer interaction mechanisms discussed above are an important ingredient for building models of a complete streamer discharge. However, a complete model based on measurements or theoretical understanding of the microscopic processes is not yet available. There are a number of models for streamer trees that start from phenomenological assumptions of streamer channel properties as a whole. All currently available models neglect the large variation of streamer diameters and velocities in pulsed corona reactors. The first phenomenological model for a complete discharge tree was proposed by Niemeyer et al. [110]; it approximates sliding surface discharges and creates fractal structures. This model includes streamer branching in a purely phenomenological manner and assumes that all streamers are equal and that the interior is completely screened from the electric field. Since then, a number of authors have developed this model further, in chemical physics, geophysics [111], and electrical engineering [112]. At present, the challenge lies in extending such models to all recently identified microscopic ingredients such as branching statistics, streamer diameters and velocities, and interior electric fields coupled to the external circuit.
1.3 Glow Discharges at Higher Pressures
1.3.1 Introduction
The classic low-pressure glow discharge has been studied extensively for several decades. The discharge is typically produced in a low-pressure (order of 1 mbar) noble gas between two electrodes that are separated from 1 cm up to 1 m. The light emission pattern of a low-pressure glow discharge is described in all standard books [113] and includes a cathode glow, cathode dark space, negative glow, Faraday dark space, the positive column, the anode dark space, and the anode glow.
The sheath region of a glow discharge has a high electric field because of charge separation between fast electrons and slow positive ions (creating the so-called cathode fall). The fast electrons emitted by the cathode and accelerated by the high field multiply by impact ionization on the sheath edge. In many glow discharges, most space between the electrodes is occupied by the positive column, a region with a relatively low, constant electric field. See also Šijacic and Ebert [114] for a detailed description and numerical model of the Townsend to glow discharge transition. In their one-dimensional model (equivalent to a plate–plate discharge), they found that depending on p · d (pressure times distance) and the secondary emission coefficient of the cathode γ, the transition can occur according to the subcritical behavior described in books (with a negative current–voltage characteristic (CVC) from Townsend to glow) or for smaller values of p · d, it can also behave supercritical or have some intermediate “mixed” behavior.
