Content uploaded by Ulrich Kogelschatz
Author content
All content in this area was uploaded by Ulrich Kogelschatz
Content may be subject to copyright.
Plasma Chemistry and Plasma Processing, Vol. 23, No. 1, March 2003 ( 2003)
Inûited Reûiew
Dielectric-barrier Discharges: Their History,
Discharge Physics, and Industrial Applications
Ulrich Kogelschatz
1
Receiûed April 5, 2002; reûised May 7, 2002
Dielectric-barrier discharges (silent discharges) are used on a large industrial scale.
They combine the adûantages of non-equilibrium plasma properties with the ease
of atmospheric-pressure operation. A prominent feature is the simple scalability
from small laboratory reactors to large industrial installations with megawatt input
powers. Efficient and cost-effectiûe all-solid-state power supplies are aûailable. The
preferred frequency range lies between 1 kHz and 10 MHz, the preferred pressure
range between 10 kPa and 500 kPa. Industrial applications include ozone gener-
ation, pollution control, surface treatment, high power CO
2
lasers, ultraûiolet
excimer lamps, excimer based mercury-free fluorescent lamps, and flat large-area
plasma displays. Depending on the application and the operating conditions the
discharge can haûe pronounced filamentary structure or fairly diffuse appearance.
History, discharge physics, and plasma chemistry of dielectric-barrier discharges
and their applications are discussed in detail.
KEY WORDS: Dielectric-barrier discharges; silent discharges; non-equilibrium
plasmas; ozone synthesis; pollution control; surface treatment; CO
2
lasers;
excimer lamps; plasma displays.
1. HISTORY OF DIELECTRIC-BARRIER DISCHARGES
Dielectric-barrier discharges, or simply barrier discharges, have been
known for more than a century. First experimental investigations were
reported by Siemens
(1)
in 1857. They concentrated on the generation of
ozone. This was achieved by subjecting a flow of oxygen or air to the influ-
ence of a dielectric-barrier discharge (DBD) maintained in a narrow annular
gap between two coaxial glass tubes by an alternating electric field of suf-
ficient amplitude. The novel feature of this discharge apparatus was, that
the electrodes were positioned outside the discharge chamber and were not
in contact with the plasma (Fig. 1). In his later years Werner von Siemens
considered his discharge configuration for the generation of ozone as one
of his most important inventions. It is interesting to note, that he never
1
Retired from ABB Corporate Research, Switzerland.
1
0272-4324兾03兾0300-0001兾0 2003 Plenum Publishing Corporation
Kogelschatz2
Fig. 1. Historic ozone discharge tube of W. Siemens, 1857 (Ref. 1, natu
¨
rl. Gro
¨
sse means natural
size).
applied for a patent for this configuration, although he received many
patents on other subjects. A few years after Siemens’ original publication,
Andrews and Tait,
(2)
in 1860, proposed the name ‘‘silent discharge,’’ which
still is frequently used in the English, German, and French scientific litera-
ture (stille Entladung, de
´
charge silentieuse). Ozone and nitrogen oxide for-
mation in DBDs became an important research issue for many decades.
(3,4)
At the beginning of the 20th century Emil Warburg conducted extensive
laboratory investigations on the nature of the silent discharge.
(5,6)
Becker
(7,8)
in Germany and Otto
(9)
in France made important contributions to the
design of industrial ozone generators utilizing DBDs. An important step in
characterizing the discharge was made by the electrical engineer K. Buss,
(10)
who found out that breakdown of atmospheric-pressure air between planar
parallel electrodes covered by dielectrics always occurs in a large number of
tiny short-lived current filaments. He obtained the first photographic traces
(Lichtenberg figures) of these microdischarges and oscilloscope recordings
of current and voltage. More information about the nature of these current
filaments was collected by Klemenc,
(11)
Suzuki,
(12,13)
Honda and Naito,
(14)
and later by Gorbrecht et al.
(15)
and by Bagirov
(16)
and co-workers. In 1943
T. C. Manley
(17)
proposed a method for determining the dissipated power in
DBDs by using closed voltage兾charge Lissajous figures and derived an equa-
tion which became known as the power formula for ozonizers. Ozone forma-
tion in DBDs was further investigated by Briner and Susz
(18)
in Switzerland,
by Philippov
(19)
and his group in Russia, by Devins
(20)
in the United States,
by Lunt
(21)
in England and by Fuji
(22)
et al. in Japan. More recent references
can be found in the booklet ‘‘Physical Chemistry of the Barrier Discharge’’
by Samoilovich, Gibalov and Kozlov
(23)
and in a number of review papers
and handbook articles.
(24–32)
Until about ten years ago, ozone generation
was the major industrial application of DBDs with thousands of installed
ozone generating facilities used mainly in water treatment. For this reason
the dielectric-barrier discharge is sometimes also referred to as the ‘‘ozonizer
discharge.’’ Occasionally, also the term corona discharge is used in connec-
tion with DBDs, although most authors prefer to use this term only for
discharges between bare metal electrodes without dielectric. Both discharge
types have common features: the generation of ‘‘cold’’ non-equilibrium
Dielectric-barrier Discharges 3
plasmas at atmospheric pressure and the strong influence of the local field
distortions caused by space charge accumulation. Extensive research activi-
ties employing modern diagnostic and modeling tools started around 1970.
Originally aimed at a better understanding of the plasma physical and
plasma chemical processes in ozonizers, these research efforts resulted not
only in improved ozone generators, but also in a number of additional appli-
cations of dielectric-barrier discharges: surface modification, plasma chemi-
cal vapor deposition, pollution control, excitation of CO
2
lasers and excimer
lamps and, most recently, in large-area flat plasma display panels used in
wall-hung or ceiling attached television sets. These new applications of
DBDs have reached market values substantially larger than the original
ozone market. The annual market for plasma displays alone is expected to
surpass US$10 billion by the year 2005.
2. THE DIELECTRIC-BARRIER DISCHARGE
Typical planar DBD configurations are sketched in Fig. 2. As a conse-
quence of the presence of at least one dielectric barrier these discharges
require alternating voltages for their operation. The dielectric, being an insu-
lator, cannot pass a dc current. Its dielectric constant and thickness, in com-
bination with the time derivative of the applied voltage, dU兾dt, determine
the amount of displacement current that can be passed through the dielec-
tric(s). To transport current (other than capacitive) in the discharge gap the
electric field has to be high enough to cause breakdown in the gas. In most
applications the dielectric limits the average current density in the gas space.
It thus acts as a ballast which, in the ideal case, does not consume energy.
Preferred materials for the dielectric barrier are glass or silica glass, in spe-
cial cases also ceramic materials, and thin enamel or polymer layers. In
some applications additional protective or functional coatings are applied.
At very high frequencies the current limitation by the dielectric becomes less
effective. For this reason DBDs are normally operated between line fre-
quency and about 10 MHz. When the electric field in the discharge gap is
Fig. 2. Basic dielectric-barrier discharge configurations.
Kogelschatz4
high enough to cause breakdown, in most gases a large number of microdis-
charges are observed when the pressure is of the order of 10
5
Pa. This is a
preferred pressure range for ozone generation, excimer formation, as well
as for flue gas treatment and pollution control. Figure 3 shows microdis-
charges in a 1-mm gap containing atmospheric-pressure air, photographed
through a transparent electrode.
In this filamentary mode plasma formation resulting in electrical con-
ductivity are restricted to the microdischarges. The gas in between is not
ionized and serves as a background reservoir to absorb the energy dissipated
in the microdischarges and to collect and transport the long-lived species
Fig. 3. End-on view of microdischarges in atmospheric-pressure air (original size: 6 cmB6cm,
exposure time: 20 ms).
Dielectric-barrier Discharges 5
created. In most high-power applications liquid cooling of at least one of
the electrodes is used.
Besides the planar configuration sketched in Fig. 2 also annular dis-
charge gaps between cylindrical electrodes and dielectrics are used in many
technical applications. The discharge gap itself has a typical width ranging
from less than 0.1 mm to several centimeters, depending on the application.
To initiate a discharge in such a discharge gap filled with a gas at about
atmospheric pressure, voltages in the range of a few hundred V to several
kV are required. The gas can either flow through the DBD (ozone gener-
ation, surface treatment, pollution control) or it can be recirculated (CO
2
lasers) or fully encapsulated (excimer lamps, excimer based fluorescent
lamps and light panels, plasma display panels).
2.1. Breakdown Phenomena and Microdischarge Formation
At atmospheric pressure electrical breakdown in a large number of
microdischarges is the normal situation for most gases in DBD configur-
ations. Under certain circumstances also apparently homogeneous, diffuse
discharges
(33–35)
can obtained, or also regularly spaced glow discharge
patterns.
(36–39)
Such phenomena are easily obtained in pure He or He-rich
gas mixtures or, with special electrode configurations and operating con-
ditions, also in other gases. At lower pressure, typically below 100 Pa, dif-
fuse glow discharges can always be obtained. Such discharges are normally
referred to as RF glow discharges and have found widespread applications
in the semiconductor industry for plasma etching and plasma deposition
processes. These low-pressure discharges have different properties and are
not the subject of this article. The most common appearance of dielectric-
barrier discharges at elevated pressure, however, is that shown in Fig. 3.
It is characterized by a large number of short-lived microdischarges. Each
microdischarge has an almost cylindrical plasma channel, typically of about
100
µ
m radius, and spreads into a larger surface discharge at the dielectric
surface(s). Figure 4 shows a schematic diagram of a single microdischarge
and a simple equivalent circuit.
By applying an electric field larger than the breakdown field local
breakdown in the gap is initiated. In the equivalent circuit that this is sym-
bolized by closing a switch and forcing some of the current through the
plasma filament, whose resistance R(t) rapidly changes with time. In reality,
growing electron avalanches quickly produce such a high space charge that
self-propagating streamers are formed.
(25,39–46)
A space-charge induced field
enhancement at the streamer head, moving much faster than the electron
drift velocity, is reflected at the anode and travels back to the cathode where,
within a fraction of 1 ns, an extremely thin cathode fall layer is formed. At
Kogelschatz6
Fig. 4. Sketch of a microdischarge and a simple equivalent circuit.
this moment the current flow through the conductive channel bridging the
electrode gap peaks. Subsequently charge accumulation at the dielectric sur-
face(s) reduces the local electric field to such an extent that ionization stops
within a few nanoseconds and the microdischarge is choked.
Three stages of microdischarge development are shown in Fig. 5. The
metal cathode is at the top, and the anode at the bottom is covered with a
dielectric layer (
ε
G3, thickness: 0.8 mm). The electron density is shown by
Fig. 5. Two-dimensional numerical simulation of microdischarge development in an atmos-
pheric-pressure mixture of 80% H
2
and 20% CO
2
(computations by W. Egli, Ref. 45).
Dielectric-barrier Discharges 7
contour lines. Electric field lines are superimposed. We see the initial elec-
tron avalanche travelling towards the anode at 5 ns, the electron distribution
at peak current (27.25 ns), coinciding with the formation of the cathode fall
region, and a very late stage at 100 ns, long after the current has stopped.
At this stage the electrons are still present but the current is choked because
the field at this location has collapsed. A few millimeters away from the
location of the microdischarge we still find the unperturbed initial homo-
geneous field of 34 kV兾cm. The lateral extension of the surface charge (bot-
tom) is considerably larger than the diameter of the original microdischarge
channel. The described choking effect is more pronounced in electronegative
gases due to rapid electron attachment.
The plasma filament can be characterized as a transient glow discharge
with a developed cathode fall and a positive column. At atmospheric pres-
sure electron densities of 10
14
to 10
15
cm
−3
and current densities in the range
of 100 to 1000 A cm
−2
are reached.
(25,41)
Surprisingly, computations per-
formed for different gas mixtures always arrive at the same order of magni-
tude. The cathode layer with extremely high electrical fields extends only
about 10
µ
m into the gap.
(42,43,45,46)
The described choking effect due to the
local field reduction depends on the lateral extension of the surface dis-
charge and the properties of the dielectric barrier and, of course, on the gas
properties. For an isolated microdischarge the surface discharge covers an
area which is much larger than the channel diameter. The extension of the
surface discharge determines the capacitive coupling across the dielectric
barrier and the area of reduced field after the termination of the micro-
discharge. Outside this area of influence we still have undisturbed high-field
conditions reached at breakdown (see Fig. 5). As long as the external volt-
age keeps rising, subsequent microdischarges will therefore preferentially
strike at other locations outside this region. Thus, the dielectric serves a
dual purpose. It limits the amount of charge and energy imparted to an
individual microdischarge and, at the same time, distributes the microdis-
charges over the entire electrode area. An intact dielectric guarantees that
no spark or arc can occur in the discharge gap. Typical charges transported
by individual microdischarges are of the order of 100 pC, typical energies
are of the order
µ
J.
(25,41,42)
As a consequence of the minute energy dissi-
pation in a single microdischarge the local heating effect of the short current
pulse is low, in air typically less than 10°C in narrow discharge gaps.
Research on DBDs has focused on tailoring microdischarge character-
istics by making use of special gas properties, adjusting pressure and tem-
perature, and optimizing the electrode geometry as well as the properties of
the dielectric(s). The radius of a propagating streamer, for example, depends
on the gas density n and on the ionization properties of the gas. It is pro-
portional to the reciprocal value of the product of the gas density n and the
Kogelschatz8
derivative of the effective ionization coefficient
α
eff
with respect to the
reduced electric field, E兾n, taken at breakdown.
(25,28)
In many gases break-
down occurs at about 100–200 Td (1 Td ‘‘Townsend’’ corresponds to
10
−17
Vcm
2
). The effective ionization coefficient
α
eff
is mainly determined
by the ionization coefficient
α
and the attachment coefficient
η
. In electro-
negative gases its value practically corresponds to
α
eff
G
α
A
η
, with a small
correction for detachment processes. The slope of
α
eff
vs. E兾n at breakdown
is much steeper in electronegative gases due to electron attachment. Accord-
ingly, the radii of streamer channels in these gases can be grouped in the
following order:
oxygenFcarbon dioxideFairFnitrogenFxenonFhelium.
Diffuse discharges can be obtained when, at breakdown, there is sufficient
overlap of simultaneously propagating electron avalanches to cause smooth-
ing of transverse field gradients. This condition is met most easily in helium
with its relatively wide streamer channels.
The total charge Q transferred in a microdischarge depends on the gas
properties and can be influenced by the gap spacing and by the properties
of the dielectric. Q is proportional to the width of the discharge gap d, and
to the quantity
ε
兾g (
ε
: relative permittivity, g: thickness of dielectric). The
latter relation was experimentally checked to hold up to extreme
ε
-values
of about 1000.
(47–49)
Contrary to what one might expect, Q does not depend
on the gas density.
(25)
Recent advances in spectroscopic measuring techniques and in laser diag-
nostics provided important additional information by in situ determinations
of electron, atom, free radical and excited species concentrations in individ-
ual microdischarges.
(50–55)
The gas temperature inside microdischarge fila-
ments was derived from rotational band structures.
(56–58)
2.2. Microdischarge Plasma Chemistry
The early phases of microdischarge formation are characterized by elec-
tron multiplication, by excitation and dissociation processes initiated by
energetic electrons, and by ionization processes and space charge accumula-
tion. The ionic and excited atomic and molecular species initiate chemical
reactions that finally result in the synthesis of a desired species (e.g., ozone,
excimers) or the destruction of pollutants (e.g., volatile organic compounds
or VOCs, nerve gases, odours, NH
3
,H
2
S, NO
x
,SO
2
, etc.) If the major
reaction paths are dominated by charged particle reactions the term plasma
chemistry adequately describes the situation. This is the case in many low-
pressure discharges. In the majority of DBD applications, however, most
charged particles disappear before any major changes occur. In this case it
Dielectric-barrier Discharges 9
is most appropriate to speak of a free-radical chemistry primarily involving
neutral species like atoms, molecular fragments and excited molecules. In
any case, discharge activity and energy dissipation occur mainly within the
small volume fraction occupied by microdischarges. The generated active
species set the initial conditions for the ensuing chemical reactions. An
adequate picture of the physical processes during breakdown and microdis-
charge formation is prerequisite for a detailed understanding of DBD chem-
istry. Each individual microdischarge can be regarded as a miniature plasma
chemical reactor. Scaling up or increasing the power density just means that
more microdischarges are initiated per unit of time and per unit of electrode
area. In principle, individual microdischarge properties are not altered dur-
ing up-scaling.
Typically, the first step is a dissociation of the initially molecular species
by electron collisions. Electron impact dissociation of O
2
and N
2
(59–66)
has
been investigated in connection with ozone generation from oxygen and
from air. The properties of CO
2
(67–70)
have, for example, been studied with
the aim of optimizing CO
2
lasers. One important question influencing the
efficiency of DBD applications is the efficiency of the initial dissociation
process with respect to energy consumption. It depends on the fractional
energy losses of the different electron collision processes involved. The dis-
sociation of oxygen by electron impact can be highly efficient (up to 85% of
the electron energy can be utilized for the dissociation process), provided
that the reduced electric field is in the range of 100–300 Td, corresponding
to average electron energies of about 4 to 8 eV. This situation is unique in
oxygen. In pure nitrogen the fraction of discharge energy spent on dis-
sociation is much less due to energy lost for the excitation of vibrational
levels. In CO
2
under favourable conditions, at most 40% of the electron
energy can be utilized for the dissociation process. The efficiency of the
different processes involved depends on their cross sections for electron col-
lisions and the reduced field E兾n.
2.3. Numerical Modeling
Depending on the goal different modeling approaches have been taken.
To clarify the dynamics and kinetics of chemical changes rather extensive
reaction systems including free radical reactions have been treated. Since
the dominant reactions are fast, as a first step, this may be carried out
neglecting spatial variations. To simulate effects of re-striking micro-
discharges it is useful to investigate repetitive injection of electrons or free
radicals.
(28,71–74)
In many cases the final products depend in a complex way
on the assumed repetition rate and on intermediate concentrations of transi-
ent species.
Kogelschatz10
To model the formation of a microdischarge 2D models are required.
Early phases can be treated with equations also used to describe streamer
propagation in pulsed corona discharges. To include the choking action of
the dielectric barrier(s) introduction of additional boundary conditions is
required, which can treat charge accumulation at the dielectric surface and
the resulting local reduction of the electric field in the discharge gap. As
soon as the field falls below the value necessary to sustain ionization the
microdischarge decays. Clearly, this process is faster in gas mixtures show-
ing strong attachment.
For the 2D simulations of microdischarge development shown in Fig.
5 rotational symmetry was assumed. Two equation sets were used to
describe the evolution of a microdischarge. The continuity equations for the
charged species n
i
(electrons, positive and negative ions) with source and
sink terms S
i
(describing ionization and attachment) can be written as
∂n
i
∂t
C∇ · (
ν
n
i
CD
i
∇n
i
)GS
i
(1)
where
ν
G
µ
E is the drift velocity, defined as mobility
µ
times electric field
strength E and D
i
is the diffusion coefficient. E again is defined as the gradi-
ent of the potential
φ
, EG−∇
φ
. It is related to the charge density
ρ
by
Poisson’s equation
∇
2
φ
G
−
ρ
ε
(2)
The equations have to be solved simultaneously in the discharge gap and
inside the dielectric with appropriate boundary conditions at the interface.
Fast Poisson solvers and special grids were used to keep computation times
within reasonable limits.
To consider also the radial expansion of the microdischarge channel
during chemical changes rather complex 2D models treating breakdown,
plasma chemistry and hydrodynamic expansion have been developed.
(74,75)
Also the interaction of adjacent microdischarges has been attacked.
(76)
Rather sophisticated numerical models have recently been developed to
describe discharge initiation, excitation and the generation of VUV radi-
ation in rare gas mixtures used in miniature plasma display pixel cells.
(77–82)
At atmospheric pressure collision rates are high and electrons reach
equilibrium conditions corresponding to the local field within picoseconds.
Microdischarges develop at a nanosecond time scale, free radical reactions
may take microseconds to milliseconds, and ground state chemical reactions
may take much longer to reach equilibrium conditions. The rate coefficients
for electronic collisions primarily depend on the reduced electric field E兾n
Dielectric-barrier Discharges 11
or the mean electron energy. Rate coefficients for chemical reactions may
strongly depend on the gas temperature. The use of tabulated values for
the rate coefficients of electron-induced reactions is justified if steady state
conditions are reached faster than the relevant changes occur and if electric
field gradients are not too steep. The validity of this ‘‘local field approxi-
mation’’ can be checked by Monte Carlo calculations of individual electron
paths. From such calculations we know that in atmospheric-pressure gases
electrons approach steady state values in about 20 ps.
(83)
This is roughly one
thousand times shorter than the typical duration of a microdischarge.
(41)
2.4. Overall Discharge Parameters
Since microdischarge development occurs at a nanosecond time scale
and the operating cycle in most applications is of much longer duration,
DBDs are normally characterized by a large number of microdischarges per
unit of electrode area and per cycle. A typical value is about 10
6
micro-
discharges per cm
2
per second.
(84)
This number depends primarily on the
power density. It is also influenced by the presence of UV radiation. For a
given configuration and fixed operating parameters all microdischarges are
of similar nature. They are initiated at a well-defined breakdown voltage,
and they are terminated after a well-defined current flow or charge transfer.
For many purposes it is sufficient to describe the discharge by overall quan-
tities: the applied frequency f, the maximum applied voltage U
ˆ
and the
average discharge voltage U
Dis
in the gap, at which microdischarge activity
is observed. The interesting feature of the DBD power formula first pro-
posed by Manley
(17)
is, that only the peak value of the applied voltage enters
and not its form.
PG4 fC
D
U
Dis
{U
ˆ
AC
−1
D
(C
D
CC
G
)U
Dis
}, U
ˆ
¤ C
−1
D
(C
D
CC
G
)U
Dis
(3)
It relates the total power P to the operating frequency f, the peak voltage
U
ˆ
and the capacitances of the dielectric(s) (C
D
) and the discharge gap (C
G
),
quantities that characterize the electrode configuration. Since the (fictitious)
average discharge voltage U
Dis
cannot be determined directly it is sometimes
more convenient to use the quantity U
Min
instead, the minimum external
voltage at which microdischarges are observed in the discharge gap.
PG4 fC
2
D
C(C
D
C
G
)
−1
U
Min
(U
ˆ
AU
Min
), U
ˆ
¤ U
Min
(4)
The simple relation U
Min
G(C
D
CC
G
)C
−1
D
U
Dis
connects these quantities. As
a matter of fact, this relation can be used as a definition of U
Dis
. The power
formula has proved very useful for a large variety of dielectric-barrier dis-
charges. It is used in many technical developments. For a given configur-
ation and fixed peak voltage the power is directly proportional to the
Kogelschatz12
frequency. The reason is that we generate the same number of identical
microdischarges per period. The discharge voltage depends on nature of the
gas mixture, the gas density and the width of the discharge gap. Microdis-
charge activity, symbolically indicated in Fig. 6 by short peaks on top of
the capacitive current, is observed only when the voltage in the gap reaches
U
Dis
. During these phases (2 → 3 and 4 → 1) the slope in the Lassajous figure
corresponds to the capacitance of the dielectric(s). During the rest of the
cycle it corresponds to that of the electrode configuration in the absence of
a plasma.
(85)
Depending on the peak voltage microdischarge activity is
limited to a certain fraction of the cycle and occurs at twice the driving
frequency.
The experimental determination of the power dissipated in DBDs has
often proved to be difficult because in reality the power is consumed in a
large number of short-lived microdischarges. Following the original work
of Manley many authors have used voltage兾charge Lissajous figures to
determine the average power. The trick is to use the time-integrated current,
the charge, rather than trying to resolve individual microdischarge current
peaks. This can be achieved in a simple way by putting a capacitance in
series with the DBD experiment.
(85)
The voltage across this measuring
capacitor is proportional to the charge. It can be rigorously shown that the
area of a closed loop of the applied voltage vs. charge always represents the
energy consumed during one period.
(17,85,86)
The form of this voltage兾charge Lissajous figure contains important
information about the discharge. For an ideal capacitance the Lissajous
figure collapses into a straight line, a resistive load results in an ellipse. In
the majority of DBD applications (ozone generators, excimer lamps) the
figure is close to an ideal parallelogram, from which the discharge voltage
and the effective capacitances of the discharge gap and the dielectric(s) can
Fig. 6. Symbolic presentation of microdischarge activity and corresponding voltage兾charge
Lissajous figure.
Dielectric-barrier Discharges 13
Fig. 7. Voltage兾charge Lissajous figures of different dielectric-barrier discharge types. (a) ozone
tube at 1 kHz, (b) CO
2
laser at 50 kHz, (c) plasma display at 100 kHz square wave voltage,
Refs. 85, 242, 87).
be inferred (see Fig. 7). In the case of silent-discharge CO
2
lasers (4 cm
discharge gap at 13 kPa) the Lissajous figure resembles an ellipse, indicating
the presence of residual ions at all times. In the case of a plasma display
cell operated with a 100-kHz square wave voltage the figure is again a paral-
lelogram with two almost horizontal sections.
(87)
Also, plots of current vs.
voltage are used to obtain information about ignition and decay of the
discharge.
3. OZONE GENERATION
After Christian Friedrich Scho
¨
nbein,
(88)
in 1839, had identified ozone
as a new chemical compound, Werner Siemens, in 1857, proposed his
method for reliably generating ozone by passing air or oxygen through an
ac discharge bounded by at least one dielectric barrier.
(1)
Both gases, oxygen
or air, are still used today for industrial ozone generation, preferably at
pressures 0.1 and 0.3 MPa. As mentioned before, in oxygen or air at atmos-
pheric pressure or above, the discharge is of filamentary nature. The number
of microdischarges per unit of electrode area and time depends on the power
density. Their strength (energy-density, transferred charge), is determined
by the gap spacing, pressure and dielectric properties.
3.1. Reaction Kinetics of Ozone Formation from Oxygen and from Air
The control of the plasma conditions inside the microdischarge col-
umns is of eminent importance for optimizing the reaction kinetics of ozone
formation.
(23–32,40–44,85,89–92)
For a given feed gas composition and desired
power density this can be achieved by adjusting the operating parameters
pressure, and兾or gap width as well as the properties of the dielectric barrier
Kogelschatz14
(permittivity, thickness), and the feeding circuit. The plasma conditions in
the microdischarges have to be optimized for exciting and dissociating
oxygen and nitrogen molecules. Initially, the major fraction of the energy
gained by the electrons in the electric field is deposited in excited atomic
and molecular states. Starting from electron impact on ground state O
2
molecules two reaction paths leading to dissociation are available: via exci-
tation of the A
3
Σ
+
u
state with an energy threshold of about 6 eV and via
excitation of the B
3
Σ
−
u
state starting at 8.4 eV. Ozone is then formed in a
three-body reaction, involving O and O
2
, leading to the formation of the
O
3
molecule.
OCO
2
CM → O*
3
CM → O
3
CM (5)
M is a third collision partner: O
2
,O
3
, O or, in the case of air, also N
2
.O*
3
stands for a transient excited state in which the ozone molecule is initially
formed after the reaction of an O atom with an O
2
molecule. The time scale
for ozone formation in atmospheric pressure oxygen is a few microseconds.
Side reactions, also using O atoms, compete with ozone formation.
OCOCM → O
2
CM (6)
OCO
3
CM → 2O
2
CM (7)
OCO*
3
CM → 2O
2
CM (8)
The undesired side reactions (6)–(8) pose an upper limit on the atoms con-
centration, or the degree of dissociation, tolerable in the microdischarges.
Optimizing microdischarge properties for the ozone formation process
amounts to finding a compromise between excessive energy losses due to
ions in weak microdischarges and avoiding undesired chemical side reac-
tions when the microdischarges become too strong. Model calculations
using many additional reactions in oxygen suggest a reasonable compromise
at relative atom concentrations of the order 0.2% in a microdischarge
channel.
(41)
In air discharges the presence of the nitrogen ions N
+
,N
+
2
, nitrogen
atoms and excited atomic and molecular species add to the complexity of
the reaction system.
(42–44,61,93–98)
For mixtures of 20% oxygen and 80% nitro-
gen, simulating dry air, and also in humid air the chemical changes due
to a short discharge pulse were computed using fairly extended reaction
schemes.
(99,100)
In addition to ozone a variety of nitrogen oxide species are
generated: NO, N
2
O, NO
2
,NO
3
, and N
2
O
5
, all of which have been detected
by spectroscopic techniques
(75,95,101,102)
or by dynamic mass spectrometry.
(103)
The main aspects of ozone generation in dry air can be summarized as
follows. Excitation and dissociation of nitrogen molecules lead to a number
Dielectric-barrier Discharges 15
of additional reaction paths involving nitrogen atoms and the excited mol-
ecular states N
2
(A
3
Σ
+
u
) and N
2
(B
3
Π
g
) that can produce additional oxygen
atoms for ozone generation.
NCO
2
→ NOCO (9)
NCNO → N
2
CO (10)
N
2
(A)CO
2
→ N
2
OCO (11)
N
2
(A, B)CO
2
→ N
2
C2O (12)
About half of the ozone formed in air discharges results from these indirect
processes.
(42,61)
As a a result, ozone formation in air takes longer (about
100
µ
s) than in oxygen (about 10
µ
s), and a substantial fraction of the elec-
tron energy initially lost in collisions with nitrogen molecules can be re-
covered and utilized through reactions (9)–(12) for ozone generation.
At a certain level of NO兾NO
2
concentrations, which can easily be
reached in ozone generators by reducing the air flow or by applying excess-
ive power, ozone formation breaks down completely. This state is referred
to as discharge poisoning, a phenomenon already observed by Andrews and
Tait
(2)
in 1860. Under such conditions neither ozone nor N
2
O
5
appears in
the product gas. Instead, NO, NO
2
, and N
2
O are detected. This state is
characterized by rapid NO
x
reactions, consuming oxygen atoms at a faster
rate than the ozone formation reaction (5). The result is an accelerated
recombination of oxygen atoms, catalyzed by the presence of NO and NO
2
.
Previously formed ozone is removed in a catalytic ozone destruction process
also involving NO and NO
2
.
(27)
According to Crutzen
(104)
and Johnston
(105)
similar reactions also influ-
ence the stratospheric ozone concentration. Reactions involving NO and
NO
2
are extremely fast so that relatively small NO
x
concentrations on the
order of 0.1% can seriously interfere with ozone formation. These effects
can readily be demonstrated in ozone generators. If traces of NO or NO
2
are added to the feed gas oxygen or air at the intake of an ozone generator
ozone formation is inhibited. Elevated temperature in the discharge gap can
be another reason for reduced ozone generation or complete absence of
ozone at the exit. As Schultz and Wulf
(106)
demonstrated in 1940 an air-fed
ozonizer operated at the high temperature of 800°C produces mainly nitro-
gen oxides.
3.2. Technical Aspects of Industrial Ozone Generation
The first major ozone installations in drinking water plants using ozone
for disinfection were built around the beginning of the last century in Paris
Kogelschatz16
(1897) and Nice (1904), France, and in St. Petersburg, Russia (1910). Di-
electric plates or tubular dielectrics in the form of glass tubes were used.
More recently, industrial ozone generation has profited substantially from
a better understanding of the ozone formation process in dielectric-barrier
discharges including ways of tailoring microdischarge properties.
(107,108)
Also, the replacement of line frequency and low frequency motor-generator
sets by static frequency converters using modern power electronics led to
important innovations in the power supply units and in process control.
The use of higher frequencies brought the advantage of lower operating
voltages for a given input power. This resulted in reduced strain on the
dielectrics. In addition the power densities were increased substantially. As
a result, higher ozone production rates and much higher ozone concen-
trations are reliably attained in much more compact ozone
generators.
(85,107,108)
In this context it may be of interest to point out, that
already in 1921 Starke
(109)
had demonstrated that the energy efficiency of
the ozone formation process does not depend on the applied frequency
between 50 Hz and 10 kHz.
Most technical ozone generators use cylindrical discharge tubes of
about 20–50 mm diameter and 1–3 m length. The favorite dielectric material
has for a long time been borosilicate glass. In many technical ozone gener-
ators Pyrex (Duran) glass tubes, closed at one side, are mounted inside
slightly wider stainless steel tubes to form annular discharge gaps of about
0.5–1 mm radial width (see Fig. 8). Geometrical tolerances of steel and glass
tubes as well as mounting procedures for the glass tubes have been import-
ant issues in technical ozone generators. Local deviations from the design
gap width have an important influence on microdischarge properties, on the
mass flow, the power deposition and on the heat flow to the cooled steel
Fig. 8. Configuration of discharge tubes in a technical ozone generator (not to scale).
Dielectric-barrier Discharges 17
electrode. Such deviations can result in drastic reductions of the ozone gen-
erating efficiency. Superior performance has been obtained with precise,
extremely narrow discharge gaps of 0.1 mm width.
(110)
Metal coatings, e.g., thin aluminum films, inside the glass tubes serve
as high-voltage electrodes, which are contacted by metal brushes. Modern
high-performance ozone generators use non-glass dielectrics in the form of
thin coatings on steel tubes. Typically, a number of layers with optimized
dielectric characteristics are deposited. These coated steel tubes are less
fragile than the traditional Pyrex tubes. Individual discharge tubes are pro-
tected by high-voltage fuses (Fig. 8). In case of dielectric failure the fuse
blows and disconnects the faulty element. This way, the other tubes can stay
in operation. Large ozone generators use several hundred, actually close to
one thousand, discharge tubes inside big steel tanks to provide the required
electrode area for mass ozone production. The outer steel tubes are welded
to two paraallel end flanges thus forming a sealed cooling compartment in
a conventional heat exchanger configuration. They are cooled by a trans-
verse water flow. Efficient cooling is essential for good performance because
the ozone forming reaction and also the low thermal stability of the ozone
molecule demand low operating temperatures. The design of the cooling
water flow requires major attention. Since the gas flow close to the exit of
the discharge tubes contains the highest ozone concentrations this section is
most sensitive to high temperatures. The inlet part of the tubes with low
ozone concentrations is much less sensitive to temperature. Therefore, the
cooling water flow has to be designed in such a way that the incoming cold
water first cools the exit section of discharge tubes. It is essential to avoid
re-circulating regions in the water flowing through the tank.
An alternative technology mainly employed in small ozone generators
is the use of surface discharges with thin parallel electrode strips deposited
on ceramic tubes. This electrode configuration was originally proposed by
the late S. Masuda
(111)
and co-workers in 1988. It had only a minor impact
on mass ozone production, but it is used to a large extent, in a slightly
modified version, in plasma displays today.
Modern high-power ozone generators utilize advanced semiconductor
power conditioning. They make use of thyristor- or transistor-controlled
frequency converters to impress square-wave currents or specially formed
pulse trains in the medium frequency range. Typical fundamental operating
frequencies are between 0.5 and 5 kHz. Using this technology, applied volt-
ages have been reduced to the range of about 5 kV, thus practically elimina-
ting the risk of dielectric failure. With large ozone generators power factor
compensation has become an important issue. Typical power densities now
reach 1–10 kW兾m
2
of electrode area. Large ozone installations reach input
Kogelschatz18
Fig. 9. Advanced ozone generator producing 60 kg兾h with non-glass dielectrics (Ozonia Ltd.)
powers of several MW. The ozone generating capacity of one big tank is
now in the range of 100 kg兾hr.
Ozone concentrations up to 5% (wt) from air and up to 18% can be
reached using oxygen feed. Modern water works utilize ozone at concen-
trations up to 12%. Considerable progress has been made in recent years
with respect to attainable ozone concentrations and energy consumption.
Ozone properties, ozone applications, and ozone technologies were reviewed
in detail by Wojtowicz,
(112)
the ozone chemistry in water was treated by
Hoigne
´
.
(113)
3.3. Feed Gas Preparation
Ozone generators are normally operated with dried clean air or dry
oxygen. In both cases the dew point should be kept below −60°C. This limits
the water vapor content to a few ppm. Humidity in the feed gas has two
adverse effects on ozone generation. It increases the surface conductivity of
the dielectric resulting in stronger microdischarges. Through the formation
of OH and HO
2
radicals it also interferes with the ozone formation reactions
by introducing additional catalytic ozone destruction mechanisms. In the
presence of OH radicals NO and NO
2
molecules are rapidly converted to
Dielectric-barrier Discharges 19
HNO
2
and HNO
3
, respectively. Negative effects on the ozone formation
efficiency are also observed when traces of H
2
or hydrocarbons are present
in the feed gas. Both impurities have an adverse influence on ozone
generation.
(108,114,115)
Most large ozone generators use pure oxygen or oxyen
blended with about 1% nitrogen as a feed gas. The presence of nitrogen has
a pronounced beneficial effect on ozone generation. Especially at high
specific energies, traces of nitrogen increase the ozone production efficiency
and the attainable ozone concentration compared to the values obtained in
pure oxygen.
(108)
4. DEPOLLUTION OF GAS STREAMS
Since dielectric-barrier discharges, based on the mature ozone-genera-
tion technology, can operate at high power levels and can treat large
atmospheric-pressure gas flows with negligible pressure drop, potential
applications in pollution control have been systematically exam-
ined.
(26–30,32,116–125)
The utilization of dielectric-barrier discharges for the
control of gaseous pollutants and the destruction of poisonous compounds
has been addressed by several researchers. Early investigations on the
destruction of H
2
S were carried out by Berthelot
(126)
in 1876 and by Schwarz
and Kunzer in 1929.
(127)
The subject was taken up again by Traus and
Suhr
(128)
in 1992. Some work on military toxic wastes was carried out by
Clothiaux et al.
(129)
and by Fraser and Sheinson.
(130)
Much work has been
devoted to the decomposition of nitrogen oxides and sulphur oxides in flue
gases
(131–134)
and of volatile organic compounds (VOCs), such as hydro-
carbons, chlorocarbons and chlorofluorocarbons (CFCs), perfluorocarbons
(PFCs), and other hazardous air pollutants (HAPs). Contamination of
exhaust air with VOCs occurs in many industrial processes, e.g., in semicon-
ductor manufacturing, in chemical processing, in dry cleaners, and in print
and paint shops. VOC contamination of soils and water can often be treated
in the gas phase after vapor extraction of the VOCs.
(119)
Odour control in
animal houses and fish factories is still another application of dielectric-
barrier discharges. In many cases the treatment involves the destruction of
H
2
S and NH
3
.
(128,135–138)
4.1. Principles of Non-Thermal Plasma Remediation of Gaseous Pollutants
Many hazardous organic molecules are readily attacked by free radicals,
electrons, or UV photons. In many cases the carrier gas is air. Dielectric-
barrier discharges are used to provide reactive species such as N*
2
(A
3
Σ
+
u
),
N*
2
(B
3
Π
g
), O*
2
(a
1
∆
g
), O(
1
D), O(
3
P), and N(
4
S). In exhaust and flue gases
water vapor may play an important role. In addition to direct dissociation
Kogelschatz20
of water molecules by electrons, species initially formed by electron colli-
sions in the microdischarge filaments, subsequently react with H
2
O to form
additional H, OH, or HO
2
radicals for decomposing pollutants.
N*
2
CH
2
O → OHCHCN
2
(13)
O(
1
D)CH
2
O → OHCOH (14)
O
3
COH → HO
2
CO
2
(15)
HCO
2
CM → HO
2
CM (16)
DBDs are normally run at conditions in which radical concentrations in
optimized microdischarges are so low that radical–radical recombination or
annihilation reactions between radicals can be neglected. The hydroxyl rad-
ical (OH) is an extremely efficient oxidant, with an oxidizing potential
(2.8 V) substantially higher than that of ozone (2.07 V). It is the dominant
oxidant in many applications, which also plays an eminent role in cleaning
the troposphere.
(139)
In industrial pollution control applications the aim is
to convert toxic compounds to form non-hazardous or less hazardous sub-
stances such as O
2
,O
3
, CO, CO
2
,H
2
O, simple acids or, upon addition of
ammonia, to form solid salt particles, that can easily be disposed of, or
can be used as fertilizers. In some applications additional selective catalytic
converters and dust collection devices follow the plasma treatment.
4.2. Treatment of Volatile Organic Compounds
As far as DBD plasma remediation is concerned several successful
laboratory and pilot investigations have been reported in the litera-
ture.The destruction of methane (CH
4
),
(140)
butane (C
4
H
10
),
(141)
pro-
pene (C
3
H
6
),
(142)
benzene (C
6
H
6
),
(143–145)
toluene (methylbenzene,
C
6
H
5
CH
3
),
(119,146,147)
styrene (vinylbenzene, C
6
H
5
CHCH
2
),
(148)
xylene
(dimethylbenzene, C
6
H
4
(CH
3
)
2
),
(147,149)
formaldehyde (HCHO),
(145,150)
acetaldehyde (CH
3
CHO),
(151)
methanol (CH
3
OH),
(120,152)
propanol
(C
3
H
7
OH)
(153)
carbon tetrachloride (CCl
4
),
(117,119–121,153–155)
dichloro-
methane,
(156)
trichloroethane (TCA, C
2
H
3
Cl
3
),
(119,121)
trichloroethylene
(TCE, ClHC=CCl
2
),
(117,119,121,147,152–155,157–159)
perchloroethylene (PCE,
C
2
Cl
4
),
(72,119,160)
methylene chloride (CH
2
Cl
2
),
(120)
chlorobenzene
(C
6
H
5
Cl),
(161)
and tetrafluoromethane (CF
4
),
(162)
was investigated.
In general it can be stated that non-equilibrium plasmas use most of
the discharge energy to produce and accelerate electrons. The electrons then
generate highly reactive free radicals which can selectively decompose toxic
compounds. This can be achieved at low gas temperatures and at atmos-
pheric pressure, conditions that are of utmost importance for flue gas or
Dielectric-barrier Discharges 21
off gas treatment. At dilute pollutant concentrations, typically less than
1000 ppm, non-equilibrium plasma treatment requires substantially less
energy than incineration or thermal-plasma treatment. In many cases it is
also more economical than raising the temperature of the entire gas flow to
temperatures where catalytic decomposition can be achieved (200–500°C).
In many laboratory investigations dielectric-barrier discharges achieve simi-
lar results to those obtained in pulsed corona discharges or with electron
beam injection. The main advantage of DBDs is their simplicity and the
availability of reliable, efficient and affordable power supplies. Compared
to pulsed corona discharges they require no sophisticated pulsing circuits
and compared to electron beam treatment they require no vacuum chambers
with delicate windows separating the acceleration chamber from the pol-
luted atmospheric-pressure environment. Even more important, contrary to
most other discharges, dielectric-barrier discharges can be scaled up without
additional difficulties.
Following the work of Foster and Butt
(163)
in 1972 a large number of
investigations have used pellets with different dielectric and catalytic proper-
ties inside the discharge gap of dielectric-barrier discharges
(164–166)
(see also
Refs. 135, 138, 140, 141, 143, 144, 147, 151, 153, 155, and 156). Such pellets
have a strong influence on the discharge, which is now forced to ignite in
the interstices between pellets and on their surfaces. Special pellet materials
and coatings lead to additional catalytic processes. A more recent suggestion
is to ignite barrier discharges inside the pores of reticulated ceramic
foams,
(167)
which also may have additional catalytic coatings to promote the
desired reactions.
4.3. Treatment of Diesel Exhaust Gases
The treatment of NO
x
in Diesel exhaust streams of passenger vehicles
and trucks has recently become an important issue with potentially large
markets for dielectric-barrier discharges.
(168–170)
Modern direct-injection
Diesel engines offer a fuel economy unmatched by any other automotive
engines. For this reason major automobile companies are investigating
Diesel engines to meet the CO
2
emission reductions required by the Kyoto
Protocol. Using Diesel engines on a large scale would require better NO
x
and particulate control of the exhaust streams. Initial experiments showed
that the reduction of NO in N
2
carrier streams is feasible.
(171)
In a real Diesel
exhaust stream, however, the presence of oxygen and water vapor leads to
a highly oxidizing environment under plasma conditions. As a consequence,
desired reductive reaction paths converting NO to N
2
and O
2
are only of
minor importance. Recent experiments show that with the aid of hydro-
carbon additives (unburned fuel) a selective oxidation of NO to NO
2
is
Kogelschatz22
possible without the undesired formation SO
3
from SO
2
. Also the formation
of HNO
2
and HNO
3
can be suppressed, even in the presence of water
vapor. The presence of some hydrocarbons, like ethene, propene, and
propane, dramatically changes the reaction kinetics of NO oxidation. In
such a plasma environment HO
2
and peroxy radicals (R–OO) are the
dominant oxidants (R–OO–
–
H
3
C–CH(OO)–CH
3
,H
3
C–CH
2
–CH
2
(OO), or
HOC
2
H
4
OO).
NOCHO
2
→ NO
2
COH (17)
NOCR–OO → NO
2
CR–O (18)
HO
2
is produced from a reaction of H with O
2
, the peroxy radical from
reactions of hydroxyalcyl and alcyl radicals, which also quickly react with
O
2
. Measurements as well as numerical simulations show that the energy
required to oxidize an NO molecule can be substantially reduced in the
presence of ethene,
(172,173)
and propene, or propane.
(73,170,172,174)
The most promising scheme for Diesel exhaust treatment is a hybrid sys-
tem using DBD plasma oxidation followed by a heterogeneous process that
can chemically reduce NO
2
to N
2
,O
2
, and H
2
O by selective
catalysis.
(168,17,0,175)
This process is referred to as plasma enhanced selective
catalytic reduction (PE–SCR). Extensive modeling efforts have been devoted
to the plasma kinetics of NO
x
in dry and wet air streams
(94–96,101,176–183)
and to
the role of unburned hydrocarbons (UHCs)
(170,171,184,185)
and particles
(soot)
(186)
in exhaust streams. For more detailed information the reader is
referred to the PhD theses of M. Klein,
(187)
S. Bro
¨
er
(188)
and the MS thesis of
R. Dorai.
(189)
DBD plasmas can also be used to eliminate submicron size soot
particles formed during fuel combustion.
(190)
4.4. Gas Liquefaction and Greenhouse Gas Abatement
Early investigations aimed at the cracking of hydrocarbons in DBDs
date back to de Saint-Auney
(191)
in 1933. They were followed by investi-
gations of Zhitnev and Philippov
(192)
in 1967 and those by Rutkowsky
(193)
et al. in 1982. The mechanism of hydrocarbon decomposition in electrical
discharges was discussed by Slovetsky.
(194)
More recently the subject has
attracted much attention in the context of converting natural gas or hydro-
gen to a liquid fuel and coping with greenhouse gas emissions.
(195)
There is
a considerable interest in converting off-shore methane to middle distillates.
Different groups investigate the hydrogenation of CO
2
and the partial oxi-
dation of CH
4
in DBDs.
(28,71,196–202)
Methanol formation has been observed
in CH
4
兾O
2
mixtures and in CH
4
兾air mixtures.
(203,204)
Methanol, a liquid at
standard conditions, can be handled like gasoline and can be used in a
Dielectric-barrier Discharges 23
conventional car engine or, in connection with a reformer, to power a fuel
cell driving an electric motor. The difficulties met in storing, transporting,
and distributing hydrogen can be avoided when a liquid fuel is used instead.
High conversion rates of more than 60% have been obtained in DBDs
running on CO
2
兾CH
4
mixtures, the two most abundant greenhouse
gases.
(205,206)
The main product is syngas, a mixture of CO and H
2
, which
is considered a valuable feedstock for many chemical processes, including
methanol synthesis. So far these investigations have mainly been of aca-
demic interest, because the discharge energy is still too high to meet eco-
nomic requirements. Further improvements are expected from a better
understanding of the plasma kinetics, better discharge control, and the sim-
ultaneous use of catalysts.
(207)
5. SURFACE TREATMENT
With the advent of readily available and affordable plastic foils and
other polymer materials about 1960 it soon became apparent that, for a
number of applications, their surface properties required modification. Most
plastics have non-polar chemically inert surfaces making them non-receptive
to bonding, to printing inks, coatings, and adhesives. The main property
responsible for this behavior is the low surface energy. One possibility to
substantially increase the surface energy of different substrates is corona
treatment in atmospheric-pressure air. It developed into a reliable surface
treatment process that can match the production speed of foils.
(208–210)
The
term corona treatment stems from the early days when thin wires, threaded-
rod, or knife-edge electrodes where used as corona electrodes. In the mean-
time most commercial ‘‘corona treaters’’ in fact employ dielectric-barrier
discharges by covering the rotating drum or the electrodes with dielectrics.
Figure 10 shows a sketch of a modern version using water-cooled ceramic
tube electrodes.
(211)
Fig. 10. Continuous double-sided foil treatment with dielectric-barrier discharges (after
Ref. 211).
Kogelschatz24
The use of DBDs has the advantage that the presence of the dielectric
prevents electric puncture of the plastic foils in case of pinholes and that
also electrically conductive webs and metallic foils can be treated. Large
corona treaters handle foils up to 10 m width at speeds surpassing 10 m兾s.
Operating with discharge gaps of a few millimeters width in atmospheric-
pressure air leads to the filamentary silent discharge known from air-fed
ozone generators (Fig. 3). Typical operating frequencies are in the range
10–50 kHz, generated by solid-state power supplies now reaching output
powers of more than 50 kW. The development of HV power devices like
SCRs (silicon controlled rectifiers or thyristors) and IGBTs (insulated gate
bipolar transistors) led to new microcomputer-controlled power supplies
that automatically adjust the operating frequency to the resonance fre-
quency of the circuit formed by the capacitance of the electrode configur-
ation and the inductance of the secondary winding of the HV transformer.
Special power supplies were developed to generate repetitive pulse trains
resulting in improved statistical distribution of the microdischarges across
the surface, a prerequisite for more uniform treatment.
(32,212)
Further
improvements can be expected from new solid-state devices using IGCTs
(integrated gate commutated thyristors) that can switch up to 5 MVA at
10 kV. This way the use of bulky high-voltage transformers can be elimin-
ated altogether.
The result of this atmospheric-pressure plasma treatment is similar to
that obtained in low-pressure oxygen discharges, a substantial increase of
surface energy. For plastic foils changes from about 20–30 mJ兾m
2
, typical
values for untreated polymers, to 50–70 mJ兾m
2
have been reported. This is
a very desirable effect because the higher the surface energy, the more wet-
table the material. Plasma treatment thus enhances adhesion, printability
and dye uptake. Recent investigations also include the upgrading of wool
and textiles and the plasma treatment of insulated wires and cables. In this
special application pretreatment of the surface without damage to the insu-
lation is required before permanent marking becomes possible.
During surface activation radicals formed in the plasma disrupt chemi-
cal bonds in the surface layer causing the formation of new species with
different properties. This results in a modification of the near-surface region
without changing the desirable bulk properties of the material.
(213,214)
In air
plasmas the active species were identified as oxygen atoms resulting in the
buildup of oxygenated carbon centers in the surface layer. XPS spectra of
plasma treated LDPE (low density polyethylene) revealed an oxygen兾car-
bon ratio up 30% after plasma treatment.
(215)
If different gas environments are chosen also the incorporation of other
atoms like N, F, Si, etc. is possible. Traditionally this has been achieved in
the closed chamber of a low-pressure glow discharge.
(216–219)
Progress in
Dielectric-barrier Discharges 25
purging and sealing techniques (Fig. 11) made it possible to use such
environments and special gas mixtures also on installations with fast moving
webs operating at atmospheric pressure.
(32,212)
Special interest has also been devoted to reactive gas mixtures that lead
to the deposition of extremely thin coatings with improved wettability or
other desired properties. Polymer, and SiO
x
coatings
(219–223)
or even hard
diamond-like carbon films
(224,225)
have been obtained in dielectric-barrier
discharges.
Also the removal of photoresist layers in lithographic processes was
investigated.
(226)
The possibility to modify, etch or coat surfaces at low tem-
perature and close to atmospheric pressure is an important advantage for
large-scale industrial applications. The main advantage is that no expensive
evacuation systems are required. It is to be expected that coating techniques
using vapor or gas phase deposition in DBDs and also the annealing and
oxidation of sol-gel films subjected to such discharges will be further
developed.
A more recent development, still at the laboratory stage, is the modifi-
cation of surfaces and thin film deposition in spatially homogeneous, diffuse
atmospheric-pressure discharges.
(33–35,227–230)
The spatial homogeneity of
dielectric-barrier discharges can be influenced by using two dielectric bar-
riers and specially shaped metal electrodes,
(231)
by carefully selecting the gap
width and operating frequency, and by using large fractions of helium or
neon.
(232,233)
Also certain additives like traces of acetone or methane in
argon have an influence.
(234)
Investigations employing modern surface analysis testify improved
properties, similar to the best obained in low-pressure glow discharges, if
spatially homogeneous rather than filamentary dielectric-barrier discharges
Fig. 11. Online surface treatment with reactive gases (after Refs. 32, 212).
Kogelschatz26
are used.
(215,235,236)
Relative recent investigations on low-current-density, dif-
fuse dielectric-barrier discharges also suggest novel applications of mild
plasmas for sterilization and disinfection purposes
(237–239)
and for selectively
influencing biological cells.
(240)
If reliable control of these diffuse discharges
can be obtained in an industrial environment, further applications of dielec-
tric-barrier discharges for surface modification are to be expected.
6. SD CO
2
LASERS
Based on their experience with dielectric-barrier discharges in ozone
generators Yagi and co-workers
(241–244)
developed their SD CO
2
laser (SD
stands for silent discharge). This high-power infrared laser (
λ
G10.6
µ
m),
used for precision welding and cutting of thick metal plates, soon became a
commercial success. Contrary to other commercial high-power CO
2
lasers
that operate at 13.65 MHz or 27.3 MHz, the SD laser uses rather low fre-
quencies between 50 kHz and 200 kHz. The advantage of the low frequency
is that the dielectric layers on the electrodes (glass or alumina) effectively
limit the discharge current and stabilize the discharge. Another advantage
is the availability of inexpensive, highly efficient solid-state power supplies.
Operating at reduced pressure (5–20 kPa) in a mixture of N
2
兾He兾CO
2
兾
CO (60兾28兾8兾4) the discharge looks fairly homogeneous. Since there is not
enough time to remove residual ions between half waves, the plasma essen-
tially acts as a resistive load with modulated conductivity (ion dragging
mode). The laser performance was optimized by choosing a wide discharge
gap of 4–5 cm spacing and by using this unusually high nitrogen content.
(245)
A strong transverse gas flow at a velocity of 50–80 m兾s provides discharge
stabilization and cooling. Output powers reach 5–20 kW. The efficiency in
the TM
00
mode is about 10%. Within a few years this SD CO
2
laser has
gained a substantial market share in Japan. More recently, also sealed oper-
ation of CO
2
lasers pumped by dielectric-barrier discharges at 600 kHz has
been proposed.
(246)
7. ULTRAVIOLET EXCIMER LAMPS AND MERCURY-FREE
FLUORESCENT LAMPS
In dielectric-barrier discharges the plasma conditions in the micro-
discharges, resembling those of high-pressure glows, are ideally suited to
induce excimer formation. Appropriate gas mixtures can be sealed in silica
glass vessels to make fairly simply excimer lamps, efficient and powerful
sources of ultraviolet (UV) or vacuum ultraviolet (VUV) radiation (Fig. 12).
Dielectric-barrier Discharges 27
Fig. 12. Sealed cylindrical and planar excimer lamp configurations (Ref. 277).
Long lifetimes can be obtained since the electrodes are not in contact
with the plasma. Metal deposition on the walls and electrode erosion, pro-
cesses that normally limit the lifetime of lamps, are avoided. To extract as
much radiation from the plasma volume as possible the front electrode is
perforated, or a wire mesh is used. Both solutions provide up to 90% trans-
mission. The rear electrode can be used as a mirror. The configuration in
the right part of Fig. 12 uses co-planar electrodes of alternating polarity.
This way surface discharges are initiated that have similar properties to
those of the microdischarges bridging the discharge gap.
A large number of excimers have been obtained in dielectric-barrier
discharges.
(247)
Table I gives a selection of wavelengths obtained, covering
the range from 126–354 nm. The most important excimer lamps are those
based on rare gas dimers and the rare gas halides known from excimer
lasers.
7.1. Excimer Lamps Based on Rare Gas Dimers
The first excimer lamps mentioned in the literature were small VUV
sources used for spectroscopic purposes.
(248–250)
The relatively broad second
excimer continua of the rare gases were used as background radiation for
absorption measurements in the VUV range between 70 and 180 nm. During
the last decade powerful efficient VUV excimer lamps were developed that
have found novel important industrial applications. Dielectric-barrier dis-
charges in high density rare gases operated at about 0.1 MPa can efficiently
Table I. Selection of Excimers Obtained in Dielectric-barrier Discharges and their Respective
Peak Wavelengths in nm (Commercially Available Lamps in Thick Print)
Ar*
2
Kr*
2
F*
2
Xe*
2
ArCl* ArF* KrCl* KrF* XeI* Cl*
2
XeBr* Br*
2
XeCl* I*
2
XeF*
126 146 157 172 175 193 222 248 253 259 282 289 308 342 354
Kogelschatz28
convert electron kinetic energy in electronic excitation energy. Due to high
collision rates this excitation is rapidly funneled to a few low lying atomic
and dimer levels. The dimer Xe*
2
is formed in a three-body reaction involv-
ing excited and ground state Xe atoms. Computations indicate that under
favorable conditions 40–80% of the discharge power can be converted to
VUV radiation concentrated in the second excimer continua of Ar, Kr or
Xe.
(251–257)
The observed emission results from the lowest two dimer states
Xe*
2
(
1
Σ
+
u
,
3
Σ
+
u
) which decay predominantly by radiation concentrated in a
narrow wavelength region of 10–15 nm half width around the peak wave-
lengths given in Table I. Best results are obtained with short-pulse exci-
tation. Since the shorter wavelengths of Ar*
2
and Kr*
2
require special window
materials (MgF
2
, LiF, CaF
2
) to extract the VUV radiation the xenon
excimer lamp became the most important one. Its radiation, peaking at
172 nm, is transmitted by the walls of high purity silica (Suprasil) vessels.
The highest VUV efficiency of commercial xenon excimer lamps is quoted
to reach 40%. To overcome the problems with window materials at
short wavelengths also open or windowless excimer systems have been
proposed.
(258,259)
In these systems the (dielectric-coated) electrodes and the
sample to be irradiated are placed in a chamber with an excimer forming
gas mixture like, e.g., Ar. The sample is subjected to VUV radiation from
the discharge region without intermediate windows. Recent experiments
with a 0.3-m wide electrode system purged with atmospheric-pressure Ar
demonstrated VUV induced surface modification on a technical scale
through photo polymerization of thin acrylate films on moving polymer
foils.
(260)
7.2. Excimer Lamps Based on Rare Gas Halide Excimers
The halogen dimers and rare gas halide excimers listed in Table I show
much narrower emission features, typically a single emission peak of 2–
4 nm half width. The most important lamp of this category is the XeCl*
excimer lamps delivering practically monochromatic radiation at 308 nm.
The predominant reaction path leading to the formation of the exciplex
XeCl* under such conditions is the recombination of Xe
+
and Cl
−
ions.
Sealed lamps of several kW electrical power and narrow-band UV output
of about 1 kW are on the market. For UV curing applications on printing
machines tubular XeCl* lamps up to 2 m length are available. Since F atoms
attack silica surfaces, KrF*, ArF*, and F *
2
lamps have only been used in
through-flow systems in the laboratory.
(261–263)
For sealed lamps the devel-
opment of internal protective coatings (MgF
2
, LiF) would be required.
Powerful and efficient XeBr* and XeI* lamps, on the other hand, have
successfully been tested in the laboratory
(264,265)
and will soon be introduced
Dielectric-barrier Discharges 29
Fig. 13. Emission spectra of various excimer lamps (Ref. 25).
to the market. Using pulsed discharges also high-power excimer flash lamps
were investigated.
(266–272)
Since excimer formation works up to fairly high
power densities in the plasma (10
7
Wcm
−3
), UV peak powers of several MW
and average UV powers in the kW range have been obtained.
In general in can be stated that the use of dielectric-barrier discharges
to generate VUV or UV excimer radiation has proved to be a very useful
concept.
(25–30,32,247,251,273–277)
Excimer lamps produce high-intensity narrow-
band radiation at various UV and VUV wavelengths. Their efficiencies
reach 10–40%. The use of excimers has several advantages. High photon
fluxes can be extracted from the plasma without self-absorption. Many
properly optimized excimer forming gas mixtures exhibit only one dominant
narrow emission band. Commercial sealed lamps use external electrodes and
can be operated near room temperature if adequate cooling is used. Typical
filling pressures are in the range 10
4
to 10
5
Pa, typical discharge gap widths
in the range of 0.1 to 5 mm. The operating frequency can be between line
frequency and microwave frequencies, but is normally chosen to lie between
20 kHz and 500 kHz. For this frequency range compact and efficient switch-
mode power supplies are available. Since at about atmospheric pressure
excimer formation and decay are extremely fast processes and no preheating
of the gas mixture is required excimer lamps can be pulsed and gated at a
fast rate.
7.3. Applications of Excimer UV Lamps
Based on high-intensity UV radiation from excimer lamps a number of
industrial applications have emerged. A relative straightforward application
is the replacement of hot medium-pressure mercury lamps in UV curing
processes on printing machines.
(278)
For this application, especially used for
Kogelschatz30
high-speed printing on heat sensitive substrates, XeCl* lamps or combi-
nations of XeCl* and KrCl* are used. Novel applications were established
for the VUV radiation of the Xe*
2
excimer lamp. The energy of its 172-nm
photons corresponds to 7.2 eV, which is high enough to break most molecu-
lar bonds. Examples are the generation of O(
3
P) and O(
1
D) atoms in oxygen
or air, of O(
1
D) from N
2
O and OH radicals in water and in humid gas
streams. Large numbers of xenon excimer lamps are now routinely used for
‘‘UV cleaning’’ of substrates in display and semiconductor manufacturing.
The cleaning action, used as a means to remove hydrocarbon residues after
chemical rinses, is attributed to the action of oxygen atoms. The atoms
participate in chemical reactions of VUV photons with the residues to form
volatile products like CO, CO
2
, and H
2
O. Novel applications of excimer
UV lamps have also been propagated for pollution control
(279)
and water
treatment.
(280)
Advanced oxidation processes resulting in oxidative degra-
dation following VUV photolysis of air or water as well as direct UV兾VUV
photocleavage of pollutants are possible destruction mechanisms. Many of
the common pollutants (CCl
4
,CH
3
Cl, ClHC–
–
CHCl, ClHC–
–
CCl
2
,Cl
2
C–
–
CCl
2
) strongly absorb at wavelengths around 200 nm. Their absorption
coefficients can be well in excess of 10
−17
cm
2
兾molecule, more than six orders
of magnitude larger than those of oxygen or water in this wavelength region.
Consequently, selective destruction of micropollutants at very low concen-
trations can be achieved.
(279)
By UV-induced decomposition of properly selected precursor sub-
stances photo-deposition of large-area or patterned thin metal
films,
(27,32,273,281)
insulating layers with extremely high
(282,283)
or low
(284,285)
dielectric constants, or also semiconductor films has been demon-
strated.
(27,277,286)
This has been achieved with gas phase precursors, liquid
films or sol gel films. Also the photo-etching and structuring of polymer
surfaces
(273,287–290)
and the photo-assisted low-temperature oxidation of Si,
SiGe, and Ge at 250–400°C has been reported.
(277,291,292)
Very recently also
rapid photo-oxidation of Si at room temperature was achieved with Ar*
2
excimer radiation peaking at 126 nm.
(293)
It is quite evident that the avail-
ability of reliable and affordable sources of VUV radiation provides a new
stimulus to photochemistry in general and especially to materials processing
in semiconductor manufacturing. Photons and free radicals generated by
UV radiation can be used for surface treatment in a similar way as free
radicals generated in a plasma. For example, irradiation of hydrophobic
PTFE (polytetrafluoroethylene, Teflon) with xenon or krypton excimer
lamps under ammonia atmosphere results in hydrophilic surfaces. The
changes are attributed to the incorporation of nitrogen and hydrogen atoms
into the surface.
(294)
In general it can be stated that photon induced chemical
reactions are much more selective than plasma attack and that patterning
Dielectric-barrier Discharges 31
is much simpler with optical radiation. Low cost, high-power excimer lamp
systems can thus provide an interesting alternative plasma treatment and to
expensive excimer lasers.
(295)
This is especially the case when large substrates
have to be modified.
7.4. Mercury-free Fluorescent Lamps and Flat Panels
UV photons can also be utilized to excite phosphors that convert UV
radiation to visible light. This principle is used on a large scale in fluorescent
lamps and energy saving lamps. In these light sources the mercury line radi-
ation (
λ
G254 nm) of a low-pressure glow discharge is used to generate
white light from a mixture of phosphors coated on the interior surface of
the lamp. The advent of excimer lamps provided the possibility to discard
the metal electrodes in the discharge space and to produce mercury-free
fluorescent lamps in a variety of shapes, including cylindrical lamps and
planar light panels. The highly efficient Xe*
2
radiation at 172 nm is used to
replace the Hg line radiation. A very successful product development
resulted in small cylindrical lamps (Fig. 14) used in a new generation of
scanners and copying machines. These information processing lamps oper-
ate with external electrodes in a mixture of Ne and Xe at a frequency of
about 20 kHz. Ne is used as a buffer gas to reduce the ignition voltage and,
of course, to reduce costs. About 10% Xe is enough to obtain the well-
known second excimer continuum of Xe. Since only visible or near UV
radiation is emitted from the lamp it can be made of glass. The main advan-
tage of these new fluorescent lamps is that they can be switched on fast and
Fig. 14. Fluorescent lamp based on Xe excimers (USH10).
Kogelschatz32
that no warming up or idling is required. In scanners and copiers the lamp
is only active during the scan, which results in very long periods between
lamp replacement.
In addition flat planar white-light panels have been proposed
(296,297)
and
developed.
(298)
They have found applications as background illumination in
LCDs (liquid crystal displays) and in industrial, medical, and transportation
displays. Extremely thin, flat commercial light panels, having a thickness of
only 10 mm, are now available in the size range 15–30 in. diagonal. They
are much brighter (up to 10,000 cd兾m
2
) as conventional backlighting sys-
tems using multiple mercury lamps and reach lifetimes up to 100,000 hr. Co-
planar electrode systems are used similar to that shown in the right part of
Fig. 12 in combination with an internal phosphor coating. A typical sealed
flat panel is made of two parallel glass substrates, a surrounding glass frame
and multiple spacers that support the structure during evacuation and
filling.
Electrodes, dielectric barriers, reflecting and phosphor layers are
deposited by thick film printing, relatively inexpensive production tech-
niques. The printed electrode lines are specially designed to facilitate
ignition and to fix the microdischarges at certain locations. The electrode
layer is covered by a thin dielectric barrier. To enhance efficiency, these flat
panels use repetitive short-pulse excitation and to enhance discharge stab-
ility the filling pressure is lowered to about 10 kPa. Contrary to mercury
lamps these excimer-based fluorescent lamps produce no hazardous waste.
In addition, they avoid an important ageing problem, the incorporation of
mercury into the phosphor layer.
8. PLASMA DISPLAY PANELS (PDPs)
AC plasma displays utilizing Xe VUV radiation to excite phosphors are
the most recent addition to dielectric-barrier discharge applications.
(299,300)
Within a few years they reached market values by far surpassing those of
all other DBD applications. The original idea to use dielectric-barrier dis-
charges for large-area displays originated at the University of Illinois at
Urbana-Champlaign in 1964.
(301)
Bitzer and Slottow, two professors at that
institution, were engaged in computer-based education and were looking for
a graphic display of their computer outputs. They took a matrix of individu-
ally addressable tiny dielectric-barrier discharge cells filled with a Penning
mixture of Ne and traces of Ar.
(302)
This first monochromatic plasma display
already achieved a resolution of 512B512 pixels and prudently utilized the
inherent memory effect due to surface charges on the walls. The physics of
this effect had previously been investigated and discussed by Loeb and El
Baccal
(303)
following earlier the work of Harries and von Engel.
(304)
Efforts
Dielectric-barrier Discharges 33
to find gases emitting additional colors to the orange Ne lines were not
very successful. Multicolor operation based on photo excited phosphors was
demonstrated as early as 1973 with a 1024B1024 panel.
(305)
Further pro-
gress towards a color plasma display was only made when Xe VUV radi-
ation was utilized to excite phosphor layers deposited inside the tiny gas
discharge cells used for individual pixels. To obtain sufficient visible radi-
ation from such a small volume it is essential to raise the pressure to values
where efficient excimer formation occurs, typically about 50–70 kPa. Higher
pressure favors excimer formation. Above 10 kPa the three-body reaction
of excited Xe* atoms with two ground state atoms to form Xe*
2
becomes
faster than radiative decay of Xe*. The formation of the Xe*
2
excimer occurs
even when up to 95% of He or Ne are used as buffer gases. In addition to
lower costs this has the advantage of lowering the ignition voltage to values
suitable for integrated transistor drivers. In the tiny gas cells of about
0.1 mm width and height the plasma emission consists of the 147 mm Xe
resonance line, of the first excimer continuum around 150 nm and, predomi-
nantly, of the second excimer continuum around 172 nm. While the reso-
nance line is subject to radiation absorption, the excimer radiation can
escape from the plasma with practically no absorption. Adjacent cells are
separated by barrier ribs to reduce cross talk (see Fig. 15c).
In every cell the VUV radiation excites red, green, or blue internal
phosphor coatings. Each cell acts as a miniature fluorescent lamp. The dis-
charge cells are grouped as RGB (red, green, blue) color triplets or quad-
ruplets (RGBG). A large display has up to 1280 by 1024 such image points.
Two sets of perpendicularly arranged thin electrode strips allow addressing
of individual cells. The electrodes are deposited on two parallel glass plates
on opposite sides of the gas space or in a co-planar configuration on one
glass plate. In this case a trigger electrode is used on the opposite side. In
ac plasma displays all electrodes are covered by dielectric layers and sputter
resistant MgO coatings. This material has also an extremely high coefficient
for secondary electron emission caused by impinging Ne
+
or He
+
ions, which
helps in lowering the breakdown voltage. The discharge is sustained by a
square-wave voltage of 50–100 kHz frequency. At every polarity change a
short current pulse of about 20 ns duration is initiated in the cell resulting
in a short VUV pulse. The average intensity of each cell can be adjusted by
duty cycle modulation.
With this technology bright displays with close to 17 million color
shades and contrast ratios up to 3000:1 can now be realized. The peak
luminance has reached 750 cd m
−2
. It is interesting to note that it took more
than 20 years between the invention of the plasma display and the techno-
logical breakthrough that allowed mass production of large flat wall-hung
color TV sets. With about 200,000 sets sold in the year 2000 and up to 7
Kogelschatz34
Fig. 15. Pixel cells with opposite (a) and co-planar (b) electrodes and complete plasma display
configuration (c). (RGBR: red, green, blue, red).
million sets anticipated for 2005, PDPs are expected to surpass an annual
market value of US$ 10 billion. Production facilities for flat plasma displays
have been built in Japan, South Korea, and Taiwan, more recently also in
China and India. Intrinsic advantages of PDPs include distortion-free crys-
tal clear pictures, large screen sizes, and wide viewing angles. They can dis-
play digital high definition television (HDTV) pictures and can be directly
linked to video, audio, and DVD equipment.
The manufacturing of plasma displays profits from relatively inexpen-
sive materials, moderate clean room requirements, and comparatively
simple, high throughput thick film processes. Recent technological advances
increased the resolution, luminance, and lifetime by alternatively displaying
even and odd lines. The cost of the drive circuitry could be substantially
reduced by limiting the drive voltage to about 80 V without reducing the
discharge voltage. Screen sizes between 0.6 m and 1.6 m diagonal are avail-
able. The thickness of the complete panel is between 4 and 12 cm, the thick-
ness of the sealed glass discharge cell only 0.6 cm. It is expected that this
Dielectric-barrier Discharges 35
technology will eventually replace bulky cathode ray tubes in many tele-
vision and large-area monitor applications.
Presently, the efficiency of converting electrical power to visible radi-
ation is rather low, in the range of 1%. Detailed measurements and modeling
efforts have led to a satisfactory understanding of the physical processes in
a pixel cell. As an example, the energy losses in a mixture of 10% Xe in
66 kPa Ne were determined by Revel et al.
(306)
using numerical modeling.
Only 15% of the power is spent on the excitation of Xe atoms, the rest on
ion acceleration, Ne excitation and Xe and Ne ionization. This results in a
10% UV efficiency of which 40% can be collected on the phosphor. From
this fraction about 30% is converted to visible light resulting in total lumi-
nous efficiency of only 1 lm兾W. Clearly such models will be used to further
improve PDP performance by identifying more favorable electrode con-
figurations, cell geometries, gas composition and excitation mechanisms.
The other hope is to develop more efficient phosphors for the conversion of
VUV radiation. From an energy standpoint it should be possible to convert
one VUV photon to two visible photons. Recent investigations indeed suc-
ceeded in converting VUV photons into red radiation with a quantum
efficiency approaching two.
(307)
If similar progress can be reached for green
and blue phosphors this could practically double the efficiency of present
PDPs.
9. CONCLUSION AND OUTLOOK
Industrial applications of dielectric-barrier discharges have come a long
way and, more recently, have substantially increased their market penetra-
tion. No doubt, this trend will continue. Several reasons can be given for this
renaissance of an in principle rather old gas discharge. Obviously, our under-
standing of the fundamental processes determining discharge initiation and
the ensuing plasma chemical reactions has profited from modern modeling
and diagnostic tools. An important aspect is the availability of reliable cross
sections on electron collisions in various gases and of recommended values
for rate coefficients of various transient species. The development and cost
reduction of modern power electronics and the engineering skills of matching
power supplies to DBD properties were of equal importance. Miniaturization
of electrode structures and discharge cells and finding affordable large-
throughput manufacturing technologies resulted in an explosive growth of
novel applications in display devices.
Clearly these recent achievements will open roads to additional new
applications. Microstructured DBDs can be used for large-area plasma elec-
trodes or as a laboratory on the microchip for diagnosic purposes.
(308)
Further progress is expected from using ultra short pulses,
(52,56,309)
dielectrics
with special properties and from a better understanding of the surface
Kogelschatz36
discharges which represent an integral aspect of all dielectric-barrier
discharges.
(310)
ACKNOWLEDGMENT
Thanks are due to the University of Minnesota at Minneapolis. Work
on this review article began while the author stayed at the Department of
Mechanical Engineering during spring 2001.
REFERENCES
1. W. Siemens, Poggendorff ’sAnn. Phys. Chem. 102, 66 (1857).
2. T. Andrews and P. G. Tait, Phil. Trans. Roy. Soc.(London) 150, 113 (1860).
3. P. Hautefeuille and J. Chappuis, Compt. Rend. Acad. Sce
´
.(Paris) 92, 80 (1881).
4. E. Warburg and G. Leitha
¨
user, Ann. Physik (4) 28, 313 (1909).
5. E. Warburg, Ann. Physik (4) 13, 464 (1904).
6. E. Warburg, Z. Tech. Phys. 6, 625 (1925).
7. H. Becker, Wiss. Vero
¨
ff. Siemens-Konzern 1, 76 (1920).
8. H. Becker, Wiss. Vero
¨
ff. Siemens-Konzern 3, 243 (1923).
9. M.-P. Otto, Bull. Soc. Franc
¸
. Electr. 9, 129 (1929).
10. K. Buss, Arch. Elektrotech. 26, 261 (1932).
11. A. Klemenc, H. Hinterberger, and H. Ho
¨
fer, Z. Elektrochem. 43, 708 (1937).
12. M. Suzuki, Proc. Jpn. Acad. 26, 20 (1950).
13. M. Suzuki and Y. Naito, Proc. Jpn. Acad. 28, 469 (1952).
14. K. Honda and Y. Naito, J. Phys. Soc. Jpn. 10, 1007 (1955).
15. H. Gobrecht, O. Meinhardt, and F. Hein, Ber. Bunsenges. f. phys. Chem. 68, 55 (1964).
16. M. A. Bagirov, M. A. Kurbanov, A. V. Shkilev, and N. E. Nuraliev, Soû. Phys.-Tech.
Phys. 16, 1011 (1971).
17. T. C. Manley, Trans. Electrochem. Soc. 84, 83 (1943).
18. E. Briner and B. Susz, Helû. Chim. Acta. 13, 678 (1930).
19. Yu. V. Filippov, V. A. Boblikova, and V. I. Panteleev, Electrosynthesis of Ozone, Moscow
State University (1987) (in Russian).
20. J. C. Devins, J. Electrochem. Soc. 103, 460 (1956).
21. R. W. Lunt, Adû. Chem. Ser. 21, 286 (1959).
22. S. Fuji and N. Takemura, Adû. Chem. Series 21, 334 (1959).
23. V. G. Samoilovich, V. Gibalov, and K. V. Kozlov, Physical Chemistry of the Barrier
Discharge (J. P. F. Conrads and F. Leipold, eds.), DVS-Verlag, Du
¨
sseldorf (1997) (Orig-
inal Russian version: Moscow State University, 1989).
24. U. Kogelschatz, in Proc. 16th Int. Conf. on Phenomena in Ionized Gases (XVI ICPIG),
Du
¨
sseldorf (1983), Invited Papers, pp. 240–250.
25. B. Eliasson and U. Kogelschatz, IEEE Trans. Plasma Sci. 19, 309 (1991).
26. B. Eliasson and U. Kogelschatz, IEEE Trans. Plasma Sci. 19, 1063 (1991).
27. U. Kogelschatz, in Proc. 10th Int. Conf. on Gas Discharges and their Applications (GD-
92), Swansea, UK (1992), Vol. II, pp. 972–982.
28. B. Eliasson, W. Egli, and U. Kogelschatz, Pure Appl. Chem. 66, 1275 (1994).
29. U. Kogelschatz, B. Eliasson, and W. Egli, J. Phys. IV (France) 7,C4–47 (1997).
30. U. Kogelschatz, B. Eliasson, and W. Egli, Pure Appl. Chem. 71, 1819 (1999).
Dielectric-barrier Discharges 37
31. U. Kogelschatz, Ozone generation and dust collection, in Electrical Discharges for
Enûironmental Purposes: Fundamentals and Applications (E. M. van Veldhuizen, ed.),
Nova Science Publishers, Commack, New York (2000), Chapter 12, pp. 315–344.
32. U. Kogelschatz and J. Salge, High pressure plasmas. Dielectric-barrier and corona dis-
charges: Properties and technical applications, in Low Temperature Plasma Physics:
Fundamental Aspects and Applications (R. Hippler, S. Pfau, M. Schmidt, and K. H. Scho-
enbach, eds.), Wiley–VCH, Weinheim (2001), Chapter 13, pp. 331–357.
33. K. G. Donohoe and T. Wydeven, J. Appl. Polymer Sci. 23, 2591 (1979).
34. S. Kanazawa, M. Kogoma, T. Moriwaki, and S. Okazaki, J. Phys. D: Appl. Phys. 21,
838 (1988).
35. F. Massines, A. Rabehi, P. Decomps, R. B. Gadri, P. Se
´
gur, and C. Mayoux, J. Appl.
Phys. 83, 2950 (1998).
36. D. G. Boyers and W. A. Tiller, Appl. Phys. Lett. 41, 28 (1982).
37. E. Ammelt, D. Schweng, and H.-G. Purwins, Phys. Lett. A 179, 348 (1993).
38. W. Breazeal, K. M. Flynn, and E. G. Gwinn, Phys. Reû. E 52, 1503 (1995).
39. U. Kogelschatz, Filamentary, patterned and diffuse barrier discharges, IEEE Trans.
Plasma Sci. 30 (2002), accepted for publication.
40. K. Yoshida and H. Tagashira, Memoirs Kitami Inst. Technol. 18, 11 (1986).
41. B. Eliasson, M. Hirth, and U. Kogelschatz, J. Phys. D: Appl. Phys. 20, 1421 (1987).
42. D. Braun, U. Ku
¨
chler, and G. Pietsch, J. Phys. D: Appl. Phys. 24, 564 (1991).
43. D. Braun, V. Gibalov, and G. Pietsch, Plasma Sources Sci. Technol. 1, 166 (1992).
44. V. I. Gibalov and G. Pietsch, Russ. J. Phys. Chem. 68, 839 (1994).
45. W. Egli, J. M. Favre, and B. Eliasson, Crosscuts 7(3), 14 (1998).
46. G. Steinle, D. Neundorf, W. Hiller, and M. Pietralla, J. Phys. D: Appl. Phys. 32, 1350
(1999).
47. J. Dr
ˇ
ı
´
mal, V. I. Gibalov, and V. G. Samoilovich, Czech. J. Phys. B37, 1248 (1987).
48. J. Dr
ˇ
ı
´
mal, K. V. Kozlov, V. I. Gibalov, and V. G. Samoilovich, Czech. J. Phys. B 38,
159 (1988).
49. V. I. Gibalov, J. Dr
ˇ
ı
´
mal, M. Wronski, and V. G. Samoilovich, Contrib. Plasma Phys. 31,
89 (1991).
50. E. Gerova and S. Mu
¨
ller, in Proc. 23rd Int. Conf. on Phenomena in Ionized Gases (XXIII
ICPIG), Toulouse, France (1997), Vol. 4, pp. 120–121.
51. R. Wendt and H. Lange, J. Phys. D: Appl. Phys. 31, 3368 (1998).
52. C. Hilbert, I. Gaurand, O. Motret, and J. M. Pouvesle, J. Appl. Phys. 85, 7070 (1999).
53. C. Lukas, M. Spaan, V. Schulz-von der Gathen, M. Thomson, R. Wegst, H. F. Do
¨
bele,
and M. Neiger, Plasma Sources Sci. Technol. 10, 445 (2001).
54. K. V. Kozlov, H.-E. Wagner, R. Brandenburg, and P. Michel, J. Phys. D: Appl. Phys.
34, 3164 (2001).
55. K. Kunze, M. Miclea, G. Musa, J. Franzke, C. Vadla, and K. Niemax, Spectrochimica
Acta B57, 137 (2002).
56. O. Motret, C. Hibert, S. Pellerin, and J. M. Pouvesle, J. Phys. D: Appl. Phys. 33, 1483
(2000).
57. M. Spaan, J. Leistikow, V. Schulz-von der Gathen, and H. F. Do
¨
bele, Plasma Sources
Sci. Technol. 9, 146 (2000).
58. N. K. Bibinov, A. A. Fateev, and K. Wiesemann, Plasma Sources Sci. Technol. 10, 579
(2001).
59. A. V. Phelps and S. A. Lawton, J. Chem. Phys. 69, 1055 (1978).
60. N. P. Penkin, V. V. Smirnov, and O. D. Tsygir, Soû. Phys.: Techn. Phys. 27, 945 (1982).
61. B. Eliasson, U. Kogelschatz, and P. Baessler, J. Phys. B: At. Mol. Phys. 17, L797 (1984).
62. B. Eliasson and U. Kogelschatz, J. Phys. B: At. Mol. Phys. 19, 1241 (1986).
Kogelschatz38
63. S. Kajita, S. Ushirosa, and Y. Kodo, J. Appl. Phys. 67, 4015 (1990).
64. P. C. Cosby, J. Chem. Phys. 98, 9560 (1993).
65. A. V. Phelps and L. C. Pitchford, Phys. Reû. A 31, 2932 (1985).
66. P. C. Cosby, J. Chem. Phys. 98, 9544 (1993).
67. W. L. Nighan, Phys. Reû. A 2, 1989 (1970).
68. A. N. Lobanov and A. N. Suchkov, Soû. J. Quant. Electron. 4, 843 (1975).
69. H. N. Ku
¨
cu
¨
karpaci and J. Lucas, J. Phys. D: Appl. Phys. 12, 2123 (1979).
70. M. Braglia, R. Winkler, and J. Wilhelm, Contrib. Plasma Phys. 31, 463 (1991).
71. B. Eliasson, F.-G. Simon, and W. Egli, Hydrogenation of CO
2
in a silent discharge, in
Non-Thermal Plasma Techniques for Pollution Control, NATO ASI Series, Vol. G 34, Part
B (B. M. Penetrante and S. E. Schultheis, eds.), Springer, Berlin (1993), pp. 321–337.
72. A. C. Gentile, Kinetic Processes and Plasma Remediation of Toxic Gases, PhD Thesis,
University of Illinois at Urbana-Champaign (1995), ftp:兾兾uigelz.ece.uiuc.edu兾pub兾
theses兾gentile-thesis.pdf
73. R. Dorai and M. J. Kushner, J. Appl. Phys. 88, 3739 (2000).
74. X. Xu, Dynamics of high- and low-pressure plasma remediation, PhD Thesis, University
of Illinois at Urbana-Champaign (2000), ftp:兾兾uigelz.ece.uiuc.edu兾pub兾theses兾
xxu_thesis.pdf
75. I. Stefanovic
´
, N. K. Bibinov, A. A. Deryugin, I. P. Vionogradov, A. P. Napartovich, and
K. Wiesemann, Plasma Sources Sci. Technol. 10, 406 (2001).
76. X. P. Xu and M. J. Kushner, J. Appl. Phys. 84, 4153 (1998).
77. R. C. Campbell, R. Veerasingam, and R. T. McGrath, IEEE Trans. Plasma Sci. 23, 698
(1995).
78. J.-P. Boeuf and L. C. Pitchford, IEEE Trans. Plasma Sci. 24, 95 (1996).
79. J.-P. Boeuf and H. Doyeaux, Europhys. News 27, 46 (1996).
80. Y. Ikeda, J. P. Verboncoeur, P. J. Christenson, and C. K. Birdsall, J. Appl. Phys. 86,
2431 (1999).
81. H. S. Jeong, B.-J. Shin, and K.-W. Whang, IEEE Trans. Plasma Sci. 27, 171 (1999).
82. T. Tamida, S. J. Sanders, and M. Tanaka, Jpn. J. Appl. Phys. 39, 583 (2000).
83. B. Eliasson and W. Egli, Helû. Phys. Acta 60, 241 (1987).
84. J. J. Coogan and A. D. Sappey, IEEE Trans. Plasma Sci. 24, 91 (1996).
85. U. Kogelschatz, Advanced ozone generation, in Process Technologies for Water Treat-
ment (S. Stucki, ed.), Plenum Press, New York, NY (1988), pp. 87–120.
86. Z. Falkenstein and J. J. Coogan, J. Phys. D: Appl. Phys. 30, 817 (1997).
87. T. Tamida, A. Iwata, J. Nishimae, and M. Tanaka, in Proc. 12th Int. Conf. on Gas Dis-
charges and Their Appl. (GD 97), Greifswald (1997), Vol. II, pp. 641–644.
88. C. F. Scho
¨
nbein, Poggendorff ’sAnn. Phys. Chem. 50, 616 (1840).
89. U. Kogelschatz, B. Eliasson, and M. Hirth, Ozone Sci. Eng. 10, 367 (1988).
90. R. Peyrous, P. Pignolet, and B. Held, J. Phys. D: Appl. Phys. 22, 1658 (1989).
91. G. Pietsch and V. I. Gibalov, Pure Appl. Chem. 70, 1169 (1998).
92. J. Kitayama and M. Kuzumoto, J. Phys. D: Appl. Phys. 30, 2453 (1997).
93. S. Yagi and M. Tanaka, J. Phys. D: Appl. Phys. 12, 1509 (1979).
94. V. G. Samoilovich and V. I. Gibalov, Russ. J. Phys. Chem. 60, 1107 (1986).
95. D. Braun, U. Ku
¨
chler, and G. Pietsch, Pure Appl. Chem. 60, 741 (1988).
96. I. A. Kossyi, A. Yu. Kostinsky, A. A. Matveyev, and V. P. Silakov, Plasma Sources Sci.
Technol. 1, 207 (1992).
97. J. Kitayama and M. Kuzumoto, J. Phys. D: Appl. Phys. 32, 3032 (1999).
98. J. T. Herron, J. Phys. Chem. Ref. Data 28, 1453 (1999).
99. B. Eliasson and U. Kogelschatz, J. Chim. Phys. 83, 279 (1986).
100. R. Peyrous, Ozone Sci. Eng. 12, 19 (1990).
Dielectric-barrier Discharges 39
101. B. Eliasson and U. Kogelschatz, in Proc. 8th Int. Symp. on Plasma Chemistry (ISP-8),
Tokyo (1987), pp. 736–741.
102. U. Kogelschatz and P. Baessler, Ozone Sci. Eng. 9, 195 (1987).
103. V. I. Gibalov, V. G. Samoilovich, and M. Wronski, in Proc. 7th Int. Symp. on Plasma
Chemistry (ISP-7), Eindhoven (1985), pp. 401–406.
104. P. J. Crutzen, Quart. J. R. Met. Soc. 96, 320 (1970).
105. H. S. Johnston, Ann. Reû. Phys. Chem. 43, 1 (1992).
106. J. F. Schultz and O. R. Wulf, J. Am. Chem. Soc. 62, 2980 (1940).
107. U. Kogelschatz and B. Eliasson, Ozone Generation and Applications, in Handbook of
Electrostatic Processes (J. S. Chang, A. J. Kelly, and J. M. Crowley, eds.), Marcel Dekker,
New York (1995), Chapter 26, pp. 581–605.
108. U. Kogelschatz, in Proc. Int. Ozone Symp., Basle (1999), pp. 253–265.
109. A. Starke, Z. Elektrochem. 29, 358 (1923).
110. M. Kuzumoto, Y. Tabata, and S. Yagi, Trans. IEE of Japan 116A, 121 (1996) (in
Japanese).
111. S. Masuda, K. Akutsu, M. Kuroda, Y. Awatsu, and Y. Shibuya, IEEE Trans. Ind. Appl.
24, 223 (1988).
112. J. A. Wojtowicz, Ozone, in Kirk–Othmer Encyclopedia of Chemical Technology, John
Wiley & Sons, Inc. (1996), Fourth Edition, Vol. 17, pp. 953–994.
113. J. Hoigne
´
, Cemistry of aqueous ozone and transformation of pollutants by ozonation
and advanced oxidation process, in Handbook of Enûironmental Chemistry, (J. Hrubec,
ed.), Vol. 5, Part C, Springer, Berlin (1998), pp. 83–141.
114. W. E. Cromwell and T. C. Manley, Adû. Chem. Ser. 21, 304 (1959).
115. E. Inoue and K. Sugino, Adû. Chem. Ser. 21, 313 (1959).
116. B. M. Penetrante and S. E. Schultheis (eds.), Non-Thermal Plasma Techniques for Pol-
lution Control, Part A: Oûerûiew, Fundamentals and Supporting Technologies, Part B: Elec-
tron Beam and Electrical Discharge Processing, NATO ASI Series, Series G: Ecological
Sciences, Vol. 34, Springer, Berlin (1993).
117. L. A. Rosocha, G. A. Anderson, L. A. Bechtold, J. J. Coogan, H. G. Heck, M. Kang,
W. H. McCulla, R. A. Tennant, and J. P. Wantuck, Treatment of hazardous organic
wastes using silent discharge plasmas, in Non-Thermal Plasma Techniques for Pollution
Control (B. M. Penetrante and S. E. Schultheis, eds.), NATO ASI Series, Vol. G 34, Part
B, Springer, Berlin (1993), pp. 281–308.
118. U. Kogelschatz, in Proc. 8th Int. Conf. on Switching Arc Phenomena/Int. Symp. on Electri-
cal Technologies for Enûironmental Protection (SAP&ETEP’97), Lodz (1997), pp. 299–
303.
119. L. A. Rosocha, Processing of hazardous chemicals using silent-discharge plasmas, in
Plasma Science and the Enûironment (W. Manheimer, L. E. Sugiyama, and T. H. Stix,
eds.), AIP Press, Woodbury, NY (1997), Chap. 11, pp. 261–298.
120. B. M. Penetrante, J. N. Bardsley, and M. C. Hsiao, Jpn. J. Appl. Phys. 36, 7B, 5007
(1997).
121. L. A. Rosocha, J. Adû. Oxid. Technol. 4, 247 (1999).
122. B. M. Penetrante, R. M. Brusasco, B. T. Merrit, and G. E. Vogtlin, Pure Appl. Chem.
71, 1829 (1999).
123. E. M. van Veldhuizen (ed.), Electrical Discharges for Enûironmental Purposes: Funda-
mentals and Applications, Nova Science Publ., Huntington NY (2000).
124. V. A. Bityurin, M. A. Deminsky, and B. V. Potapkin, Chemical activity of discharges, in
Electrical Discharges for Enûironmental Purposes: Fundamentals and Applications,E.M.
van Veldhuizen, ed., Nova Science Publ., Huntington, NY (2000), Chapter 4, pp. 49–117.
Kogelschatz40
125. L. A. Rosocha and R. A. Korzekwa, Removal of volatile organic compounds (VOCs) by
atmospheric-pressure dielectric-barrier and pulsed-corona electrical discharges, in Electri-
cal Discharges for Enûironmental Purposes: Fundamentals and Applications (E. M. van
Veldhuizen, ed.), Nova Science Publ., Huntington, NY (2000), Chapter 10, pp. 245–278.
126. M. Berthelot, Compt. Rend. Acad. Sce
´
. (Paris) 82, 1360 (1876).
127. R. Schwarz and W. Kunzer, Z. Anorg. Allgem. Chem. 183, 287 (1929).
128. I. Traus and H. Suhr, Plasma Chem. Plasma Process. 12, 275 (1992).
129. E. J. Clothiaux, J. A. Koropchak, and R. R. Moore, Plasma Chem. Plasma Process. 4,
15 (1984).
130. M. E. Fraser and R. Sheinson, Plasma Chem. Plasma Process. 6, 27 (1986).
131. I. Sardja and S. K. Dhali, Appl. Phys. Lett. 56, 21 (1989).
132. S. K. Dhali and I. Sardja, J. Appl. Phys. 69, 6319 (1991).
133. M. B. Chang, M. J. Kushner, and M. J. Rood, Plasma Chem. Plasma Process. 12, 565
(1992).
134. W. Sun, B. Pashai, S. K. Dhali, and F. I. Honea, J. Appl. Phys. 79, 3438 (1996).
135. R. H. Zhang, T. Yamamoto, and D. S. Bundy, IEEE Trans. Ind. Appl. 32, 113 (1996).
136. M. B. Chang and T. D. Tseng, J. Enûiron. Eng. 122, 41 (1996).
137. R. Ruan, W. Han, A. Ning, P. L. Chen, P. R. Goodrich, and R. Zhang, J. Adû. Oxid.
Technol. 4, 328 (1999).
138. H. Ma, P. Chen, and R. Ruan, Plasma Chem. Plasma Process. 21, 611 (2001).
139. J. Comes, Angew. Chem. 106, 1900 (1994).
140. A. Ogata, K. Mizuno, S. Kushiyama, and T. Yamamoto, Plasma Chem. Plasma Process.
18, 363 (1998).
141. S. Futamura, A. Zhang, G. Prieto, and T. Yamamoto, IEEE Trans. Ind. Appl. 34, 967
(1998).
142. M. C. Hsiao, B. T. Merrit, B. M. Penetrante, G. E. Vogtlin, P. H. Wallman, R. G.
Tonkyn, R. R. Shah, and T. M. Orlando, J. Adû. Oxid. Technol. 1, 79 (1996).
143. A. Ogata, K. Yamanouchi, K. Mizuno, S. Kushiyama, and T. Yamamoto, Plasma Chem.
Plasma Process. 19, 383 (1999).
144. A. Ogata, N. Shintani, K. Mizuno, S. Kushiyama, and T. Yamamoto, IEEE Trans. Ind.
Appl. 35, 753 (1999).
145. W. C. Neely, E. I. Newhouse, E. J. Clothiaux, and C. A. Gross, Decomposition of com-
plex molecules using silent discharge processing, in Non-Thermal Plasma Techniques for
Pollution Control (B. M. Penetrante and S. E. Schultheis, eds.), NATO ASI Series, Vol.
G 34, Part B, Springer, Berlin (1993), pp. 309–320.
146. A. Sjo
¨
berg, T. H. Teich, E. Heinzle, and K. Hungerbu
¨
hler, J. Adû. Oxid. Technol. 4, 319
(1999).
147. T. Yamamoto, J.-S. Chang, A. A. Berezin, H. Kohno, S. Honda, and A. Shibuya, J. Adû.
Oxid. Technol. 1, 67 (1996).
148. G. K. Anderson, H. Snyder, and J. Coogan, Plasma Chem. Plasma. Process. 19, 131
(1999).
149. S. P. Bugaev, V. A. Kushinov, N. S. Sochugov, and P. A. Khryapov, Plasma Chem.
Plasma Process. 16, 669 (1996).
150. D. G. Storch and M. J. Kushner, J. Appl. Phys. 73, 51 (1993).
151. H. M. Lee and M. B. Chang, Plasma Chem. Plasma Process. 21, 329 (2001).
152. M. C. Hsiao, B. T. Merrit, B. M. Penetrante, and G. E. Vogtlin, J. Appl. Phys. 78, 3451
(1995).
153. Z. Falkenstein, J. Adû. Oxid. Technol. 2, 223 (1997).
154. R. G. Tonkyn, S. E. Barlow, and T. M. Orlando, J. Appl. Phys. 80, 4877 (1996).
Dielectric-barrier Discharges 41
155. T. Yamamoto, K. Mizuno, I. Tamori, A. Ogata, M. Nifuku, M. Michalska, and G.
Prieto, IEEE Trans. Ind. Appl. 32, 100 (1996).
156. C. Fitzsimmons, F. Ismail, J. C. Whitehead, and J. J. Wilman, J. Phys. Chem. A104, 6032
(2000).
157. D. Evans, L. A. Rosocha, G. K. Anderson, J. J. Cogan, and M. J. Kushner, J. Appl.
Phys. 74, 5378 (1993).
158. T. Oda, R. Yamashita, T. Takahashi, and S. Masuda, IEEE Trans. Ind. Appl. 32, 227
(1966).
159. S. Futamura and T. Yamamoto, IEEE Trans. Ind. Appl. 33, 447 (1997).
160. A. C. Gentile and M. J. Kushner, J. Appl. Phys. 78, 2977 (1995).
161. H. Snyder and G. K. Anderson, IEEE Trans. Plasma Sci. 26, 1695 (1998).
162. S. J. Yu and M. B. Chang, Plasma Chem. Plasma Process. 21, 311 (2001).
163. R. N. Foster and J. B. Butt, Enhancing reaction rates, US Patent No. 3,674,666 (1972).
164. J.-S. Chang, P. A. Lawless, and T. Yamamoto, IEEE Trans. Plasma Sci. 19, 1152 (1991).
165. C. M. Nunez, G. H. Ramsey, W. H. Ponder, J. H. Abbott, L. E. Mammel, and P. H.
Kariher, Air & Waste 43, 242 (1993).
166. T. Yamamoto, P. A. Lawless, M. K. Owen, D. S. Ensor, and C. Boss, Control of volatile
organic compounds by a pulsed corona reactor and a packed bed reactor, in Non-Thermal
Plasma Techniques for Pollution Control, NATO ASI Series, Vol. G 34, Part B (B. M.
Penetrante and S. E. Schultheis, eds.), Springer, Berlin (1993), pp. 223–237.
167. M. Kraus, B. Eliasson, U. Kogelschatz, and A. Wokaun, Phys. Chem. Chem. Phys. 3,
294 (2001).
168. M. Higashi, S. Uchida, N. Suzuki, and K. Fujii, IEEE Trans. Plasma Sci. 20, 1 (1992).
169. M. Klein, and R. Seebo
¨
ck, Phys. Bla
¨
tter 52, 886 (1996) (in German).
170. B. M. Penetrante, R. M. Brusasco, B. T. Merrit, and G. E. Vogtlin, Pure Appl. Chem.
71, 1829 (1999).
171. B. M. Penetrante, M. C. Hsiao, B. T. Merrit, G. E. Vogtlin, P. H. Wallman, M. Neiger,
O. Wolf, T. Hammer, and S. Bro
¨
er, Appl. Phys. 68, 3719 (1996).
172. W. O. Niessen, O. Wolf, R. Schruft, and M. Neiger, J. Phys. D: Appl. Phys. 31, 542
(1998).
173. T. Hammer, in Proc. 7th Int. Symp. on High Pressure Low Temperature Plasma Chemistry
(HAKONE VII), Greifswald (2000), Vol. 2, 234–241.
174. J. W. Hoard, T. J. Wallington, J. C. Ball, M. D. Hurley, K. Wodzisz, and M. L. Balmer,
Enûiron. Sci. Technol. 33, 3427 (1999).
175. T. Hammer, S. Bro
¨
er, and T. Kishimoto, J. Adû. Oxid. Technol. 4, 368 (1999).
176. A. C. Gentile and M. J. Kushner, J. Appl. Phys. 78, 2074 (1995).
177. J. J. Lowke and R. Morrow, IEEE Trans. Plasma Sci. 23, 661 (1995).
178. O. Eichwald, M. Yousfi, A. Hennad, and M. D. Bennabdessadok, J. Appl. Phys. 82, 4781
(1997).
179. Y. S. Mok, S. W. Ham, and I.-S. Nam, IEEE Trans. Plasma Sci. 26, 1566 (1998).
180. R. Dorai and M. J. Kushner, J. Phys. D: Appl. Phys. 34, 574 (2001).
181. L. W. Sieck, J. T. Herron, and D. S. Green, Plasma Chem. Plasma Process. 20, 235 (2000).
182. J. T. Herron and D. S. Green, Plasma Chem. Plasma Process. 21, 459 (2001).
183. J. T. Herron, Plasma Chem. Plasma Process. 21, 581 (2001).
184. R. Dorai and M. J. Kushner, J. Appl. Phys. 88, 3739 (2000).
185. I. Orlandini and U. Riedel, J. Phys. D: Appl. Phys. 33, 2467 (2000).
186. R. Dorai, K. Hassouni, and M. J. Kushner, J. Appl. Phys. 88, 6060 (2000).
187. M. Klein, Barrierenentladungen zur Entstickung motorischer Abgase, PhD Thesis, Univer-
sita
¨
t Karlsruhe (1995) (in German).
Kogelschatz42
188. S. Bro
¨
er, Plasmainduzierte Entstickung dieselmotorischer Abgase-Der Einfluss gasfo
¨
rmiger
Additiûe sowie die Kombination mit katalytischen und reaktiûen Materialien, PhD Thesis,
Technische Universita
¨
tMu
¨
nchen (1997) (in German).
189. R. Dorai, Modeling of plasma remediation of NO
x
using global kinetic models accounting
for hydrocarbons, MS Thesis, University of Illinois at Urbana-Champlaign, 2000, ftp:兾兾
uigelz.ece.uiuc,.edu兾pub兾theses兾rajesh_ms_thesis.pdf.
190. S. Mu
¨
ller, J. Conrads, and W. Best, in Proc. 7th Int. Symp. on High Pressure Low Temp.
Plasma Chem. (HAKONE VII), Greifswald (2000), Vol. 2, pp. 340–344.
191. R. V. de Saint-Auney, Chimie et Industrie 29, 1011 (1933).
192. Yu. N. Zhitnev and Yu. V. Filippov, Moscow Uniûersity Chemistry Bull. 22, 5 (1967).
193. J. Rutkowsky, H. Drost, and R. Mach, Beitr. Plasmaphys. 23, 181 (1983).
194. D. I. Slovetsky, Pure Appl. Chem. 60, 753 (1988).
195. B. Eliasson, CO
2
Chemistry: An Option for CO
2
Emission Control? in Carbon Dioxide
Chemistry: Enûironmental Issues (J. Paul and C.-M. Pradier, eds.), The Royal Soc. Chem.,
Cambridge (1994), pp. 5–15.
196. S. S. Shepelev, H. D. Gesser, and N. R. Hunter, Plasma Chem. Plasma Process. 13, 479
(1993).
197. K. Okazaki, in Proc. Int. Symp. on CO
2
Fixation and Efficient Utilization of Energy,
Tokyo, Japan (1993), pp. 37–42.
198. R. Bhatnagar and R. G. Mallinson, Methane conversion in an ac electric discharges at
ambient conditions, in Methane and Alkane Conûersion Chemistry (M. M. Bhasin and
D. W. Slocum, eds.), Plenum Press, New York (1995), pp. 249–264.
199. O. Motret, S. Pellerin, M. Nikravech, V. Massereau, and J.-M. Pouvesle, Plasma Chem.
Plasma Process. 17, 393 (1997).
200. L.-M. Zhou, B. Xue, U. Kogelschatz, and B. Eliasson, Plasma Chem. Plasma Process.
18, 375 (1998).
201. L. M. Zhou, B. Xue, U. Kogelschatz, and B. Eliasson, Energy and Fuels 12, 1191 (1998).
202. K. V. Kozlov, P. Michel, and H.-E. Wagner, in Proc. 7th Int. Symp. on High Pressure
Low Temp. Plasma Chem. (HAKONE VII), Greifswald (2000), Vol. 2, pp. 262–266.
203. B. Eliasson, U. Kogelschatz, B. Xue, and L.-M. Zhou, Ind. Eng. Chem. Res. 37, 3350
(1998).
204. B. Eliasson, C. Liu, and U. Kogelschatz, Ind. Eng. Chem. Res. 39, 1221 (2000).
205. K. Zhang, B. Eliasson, and U. Kogelschatz, Energy and Fuels 15, 395 (2001).
206. U. Kogelschatz, L.-M. Zhou, B. Xue, and B. Eliasson, Production of synthesis gas
through plasma-assisted reforming of greenhouse gases, in Greenhouse Gas Control Tech-
nologies (B. Eliasson, P. F. W. Riemer, and A. Wokaun, eds.), Pergamon, Amsterdam
(1999), pp. 385–390.
207. M. Kraus, W. Egli, K. Haffner, B. Eliasson, U. Kogelschatz, and A. Wokaum, Phys.
Chem. Chem. Phys. 4, 668 (2002).
208. D. A. Markgraf, Corona Treatment: An Oûerûiew, http:兾兾www.enerconind.com兾treating兾
papers兾overview兾overview.pdf, (2001).
209. J. L. Linsley Hood, in Proc. 6th Int. Conf. on Gas Discharges and Their Applications (GD-
80), Edinburgh (1980), pp. 86–90.
210. T. Uehara, Corona discharge treatment of polymers, in Adhesion Promotion Techniques:
Technological Application (K. L. Mittal and A. Pizzi, eds.), Marcel Dekker, New York,
Basel (1999), Chapter 7, pp. 139–174.
211. K. Pochner, W. Neff, and R. Leber, Surf. Coat. Technol. 74/75, 394 (1995).
212. S. Meiners, J. G. H. Salge, E. Prinz, and F. Fo
¨
rster, Surf. Coat. Technol. 98, 1121 (1998).
213. M. Hudis, Plasma Treatment of Solid Materials, in Techniques and Applications of Plasma
Chemistry (J. R. Hollahan and A. T. Bell, eds.), John Wiley & Sons, New York (1974),
Chapter 3, pp. 113–147.
Dielectric-barrier Discharges 43
214. D. Briggs, Applications of XPS in Polymer Technology, in Practical Surface Analysis by
Auger and X-ray Photoelectron Spectroscopy (D. Briggs and M. P. Seah, eds.), John Wiley
& Sons, Chichester (1983), Chapter 9, pp. 359–396.
215. X. J. Dai and L. Kviz, Study of atmospheric and low pressure plasma modification on
the surface properties of synthetic and natural fibres, Textile Institute 81st World Conf.,
Melbourne, Australia, April 2001 (www.tft.csiro.au).
216. F. D. Egitto and L. J. Matienzo, IBM J. Res. Deû. 38, 423 (1994).
217. C. M. Chan, T. M. Ko, and H. Hiraoka, Surf. Sci. Rep. 24, 1 (1996).
218. M. R. Wertheimer, L. Martinu, and J. E. Klemberg-Sapieha, Plasma Treatment of poly-
mers to improve adhesion, in Adhesion Promotion Techniques: Technological Application
(K. L. Mittal and A. Pizzi, eds.), Marcel Dekker, New York, Basel (1999), Chapter 5,
pp. 191–204.
219. O. D. Greenwood, R. D. Boyd, J. Hopkins, and J. P. S. Badyal, J. Adhesion Sci. Technol.
9, 311 (1995).
220. J. Salge, J. de Physique IV 5, C5-583 (1995).
221. R. Thyen, A. Weber, and C.-P. Klages, Surf. Coat. Technol. 97, 426 (1997).
222. F. Denes, Z. Q. Hua, W. J. Simonsick, and D. J. Aaserud, J. Appl. Polymer Sci. 71, 1627
(1999).
223. P. Cocolios, F. Coeuret, J.-L. Gelot, A. Villermet, E. Prinz, and F. Fo
¨
rster, Coating 32,
314 (1999).
224. S. P. Bugaev, A. D. Korotaev, K. V. Oskomov, and N. S. Sochugov, Surf. Coat. Technol.
96, 123 (1997).
225. D. Liu, S. Yu, T. Ma, Z. Song, and X. Yang, Jpn. J. Appl. Phys. 39, 3359 (2000).
226. Z. Falkenstein and J. J. Coogan, J. Appl. Phys. 82, 6273 (1997).
227. F. Massines, C. Mayoux, R. Messaoudi, A. Rabehi, and P. Se
´
gur, in Proc. 10th Int. Conf.
on Gas Discharges and Their Applications (GD-92), Swansea (1992), pp. 730–733.
228. J. R. Roth, M. Laroussi, and C. Liu, in Proc. 27th IEEE ICOPS, Tampa (1992), Paper
#P21.
229. P. P. Tsai, L. C. Wadsworth, and J. R. Roth, Textile Res. J. 67, 359 (1997).
230. N. Gherardi, S. Martin, and F. Massines, J. Phys. D: Appl. Phys. 33, L104 (2000).
231. M. Kogoma and S. Okazaki, J. Phys. D: Appl. Phys. 27, 1985 (1994).
232. T. Yokoyama, M. Kogoma, T. Moriwaki, and S. Okazaki, J. Phys. D: Appl. Phys. 23,
1125 (1990).
233. S. Okazaki, S. M. Kogoma, M. Uehara, and Y. Kimura, J. Phys. D: Appl. Phys. 26, 889
(1993).
234. M. Kogoma, S. Okazaki, N. Kanda, H. Uchiyama, and H. Jinno, in Proc. Jpn. Symp.
Plasma Chem. 4 (1991), pp. 345–350.
235. F. Massines and G. Gouda, J. Phys. D: Appl. Phys. 31, 3411 (1998).
236. F. Massines, R. Messaoudi, and C. Mayoux, Plasmas and Polymers 3, 43 (1998).
237. M. Laroussi, IEEE Trans. Plasma Sci. 24, 1188 (1996).
238. T. C. Montie, K. Kelly-Wintenberg, and J. R. Roth, IEEE Trans. Plasma Sci. 28,41
(2000).
239. M. Laroussi, I. Alexeff, and W. L. Kang, IEEE Trans. Plasma Sci. 28, 184 (2000).
240. M. Laroussi, G. S. Sayler, B. B. Glascock, B. McCurdy, M. R. Pearce, N. G. Bright, and
C. M. Malott, IEEE Trans. Plasma Sci. 27, 34 (1999).
241. S. Yagi, M. Hishii, N. Tabata, H. Nagai, and A. Nagai, Laser Eng. 5, 171 (1977) (in
Japanese).
242. M. Tanaka, S. Yagi, and N. Tabata, in Proc. 8th Int. Conf. on Gas Discharges and Their
Applications (GD-85), Oxford (1985), pp. 551–554.
Kogelschatz44
243. K. Yasui, M. Kuzumoto, S. Ogawa, M. Tanaka, and S. Yagi, IEEE J. Quantum Electron.
25, 836 (1989).
244. S. Yagi and M. Kuzumoto, Australian J. Phys. 48, 411 (1995).
245. M. Kuzumoto, S. Ogawa, and S. Yagi, J. Phys. D: Appl. Phys. 22, 1835 (1989).
246. S. Wienecke, S. Born, and W. Vio
¨
l, J. Phys. D: Appl. Phys. 33, 1282 (2000).
247. B. Gellert and U. Kogelschatz, Appl. Phys. B52, 14 (1991).
248. Y. Tanaka, J. Opt. Soc. Am. 45, 710 (1955).
249. E. N. Pavlovskaya and A. V. Yakovleva, Opt. Spectros.(USSR) 54, 132 (1983).
250. G. A. Volkova, N. N. Krillova, E. N. Pavlovskaya, and A. V. Yakovleva, J. Appl. Spec-
trosc. 41, 1194 (1984).
251. B. Eliasson and U. Kogelschatz, Appl. Phys. B46, 299 (1988).
252. A. P. Gochelashvili, A. V. Dem’yanov, J. V. Kochetov, and L. R. Yangurazova, Laser
Phys. 3, 140 (1993).
253. V. V. Ivanov, K. S. Klopovskii, Yu. A. Mankelevich, A. T. Rakhimov, T. V. Rakhimova,
G. B. Rulev, and V. B. Saenko, Laser Phys. 6, 654 (1996).
254. F. Vollkommer and L. Hitzschke, in Proc. 8th Int. Symp. on the Science and Technology
of Light Sources (LS-8), Greifswals (1998), pp. 51–60.
255. A. Oda, H. Sugarawa, Y. Sakai, and H. Akashi, J. Phys. D: Appl. Phys. 33, 1507 (2000).
256. A. M. Boichenko, S. I. Yakovlenko, and V. F. Tarasenko, Laser and Particle Beams 18,
655 (2000).
257. R. J. Carman, B. K. Ward, and R. P. Mildren, in Proc. 25th Int. Conf. on Phenomena in
Ionized Gases (XXV ICPIG), Nagoya 2001, Vol. 4, pp. 331–332.
258. U. Kogelschatz, in Proc. 20th Int. Conf. on Phenomena in Ionized Gases (XX ICPIG),
Pisa (1991), Invited Papers, pp. 218–227.
259. H. Esrom and U. Kogelschatz, Appl. Surf. Sci. 54, 440 (1992).
260. M. Lenk and R. Mehnert, in Proc. RadTech Europe, Basle (2001), pp. 153–158.
261. H. Kumagai and M. Obara, Appl. Phys. Lett. 54, 2619 (1989).
262. H. Kumagai and M. Obara, Appl. Phys. Lett. 55, 1583 (1989).
263. H. Kumagai and K. Toyoda, Appl. Phys. Lett. 59, 2811 (1991).
264. Z. Falkenstein and J. J. Coogan, J. Phys. D: Appl. Phys. 30, 2704 (1997).
265. J.-Y. Zhang and I. W. Boyd, J. Appl. Phys. 84, 1174 (1998).
266. T. Gerber, W. Luethy, and P. Burkhardt, Opt. Comm. 35, 242 (1980).
267. V. M. Borisov, A. M. Davidovskii, and O. B. Kristoforov, Soû. J. Quantum Electron. 12,
1403 (1982).
268. A. M. Boichenko, V. S. Skakun, V. F. Tarasenko, E. A. Fomin, and S. I. Yakovlenko,
Laser Phys. 4, 635 (1994).
269. V. F. Tarasenko, M. I. Lomaev, A. N. Pachenko, V. S. Skakun, and E. A. Sosnin, High-
power UV excilamps, in High Power Lasers—Science and Engineering (R. Kossowsky,
M. Jelinek, and R. F. Walter, eds.), Kluwer Academic Publishers, Dordrecht (1996),
pp. 331–345.
270. J. Kawanaka, T. Shirai, S. Kubodera, and W. Sasaki, Proc. SPIE 3574, 466 (1998).
271. S. Bollanti, G. Clementi, P. Di Lazzaro, F. Flora, G. Giordano, T. Letardi, F. Muzzi,
G. Shina, and Z. E. Zheng, IEEE Trans. Plasma Sci. 27, 211 (1999).
272. A. M. Boichenko, V. S. Skakun, E. A. Sosnin, V. F. Tarasenko, and S. I. Yakovlenko,
Laser Phys. 10, 540 (2000).
273. U. Kogelschatz, Appl. Surf. Sci. 54, 410 (1992).
274. K. Stockwald and M. Neiger, Contrib. Plasma Phys. 35, 15 (1995).
275. J.-Y. Zhang and I. W. Boyd, J. Appl. Phys. 80, 633 (1996).
276. I. W. Boyd and J.-Y. Zhang, Nucl. Instrum. Methods B121, 349 (1997).
277. U. Kogelschatz, H. Esrom, J.-Y. Zhang, and I. W. Boyd, Appl. Surf. Sci. 168, 29 (2000).
Dielectric-barrier Discharges 45
278. R. Mehnert, UV Curing Equipment—Monochromatic UV Lamps, in UV and EB Curing
Technology and Equipment (R. Mehnert, A. Pincus, I. Janorski, R. Stowe, A. Bereika,
eds.), John Wiley兾SITA (1999), Chapter 4, pp. 83–105.
279. U. Kogelschatz, UV production in dielectric barrier discharges for pollution control, in
Non-Thermal Plasma Techniques for Pollution Control (B. M. Penetrante and S. E. Schul-
theis, eds.), NATO ASI Series, Vol. G 34, Part B, Springer, Berlin (1993), pp. 339–354.
280. O. Legrini, E. Oliveros, and A. M. Braun, Chem. Reû. 93, 671 (1993).
281. H. Esrom, J. Demny, and U. Kogelschatz, Chemtronics 4, 202 (1989).
282. J.-Y. Zhang, L.-J. Bie, and I. W. Boyd, Jpn. J. Appl. Phys. 37, L27 (1998).
283. J.-Y. Zhang, B. Lim, and I. W. Boyd, Thin Solid Films 336, 340 (1998).
284. J.-Y. Zhang and I. W. Boyd, Opt. Mater. 9, 251 (1998).
285. J.-Y. Zhang and I. W. Boyd, Mat. Sci. Semicond. Process. 3, 345 (2000).
286. I. W. Boyd and J.-Y. Zhang, Mat. Res. Soc. Symp. 617, J4.1.1. (2000).
287. H. Esrom, J.-Y. Zhang, and U. Kogelschatz, Mat. Res. Symp. Proc. 236, 39 (1992).
288. J.-Y. Zhang, H. Esrom, U. Kogelschatz, and G. Emig, Modification of polymers with
UV excimer radiation from lasers and lamps, J. Adhesion Sci. Technol. 8, 1179 (1994).
289. J.-Y. Zhang, H. Esrom, G. Emig, and U. Kogelschatz, Modification of polymers with
UV excimer radiation from lasers and lamps, in Polymer Surface Modification: Releûance
to Adhesion (K. L. Mittal, ed.), VSP International Science Publishers, Utrecht (1996)
pp. 153–185.
290. H. Esrom, Y.-J. Zhang, and U. Kogelschatz, Photochemical modification and etching of
PTFE with excimer VUV兾UV radiation in Polymer Surfaces and Interfaces: Characteriz-
ation, Modification and Application (K. L. Mittal and K.-W. Lee, eds.), VSP International
Science Publishers (1997), pp. 27–35.
291. I. W. Boyd, V. Craciun, and A. Kazor, Jpn. J. Appl. Phys. 32, 6141 (1993).
292. V. Craciun, B. Hutton, D. E. Williams, and I. W. Boyd, Electron. Lett. 34, 71 (1998).
293. J.-Y. Zhang and I. W. Boyd, Appl. Surf. Sci. 186, 64 (2002).
294. J. Heitz, H. Niino, and A. Yabe, Appl. Phys. Lett. 68, 2648 (1966).
295. H. Esrom and U. Kogelschatz, Thin Solid Films 218, 231 (1992).
296. C. Beneking, H. Dannert, M. Neiger, V. Schorpp, K. Stockwald, and H. Mu
¨
ller, Neuar-
tige UV-Lampen auf der Basis stiller Entladungen, BMFT (Bundesministerium fu
¨
r For-
schung und Technologie), Forschungsbericht FKZ 12N5695 (1992) (in German).
297. T. Urakabe, S. Harada, T. Saikatsu, and M. Karino, in Proc. 7th Int. Symp. on the Science
and Technology of Light Sources (LS7), Kyoto (1995), pp. 159–160.
298. M. Ilmer, R. Lecheler, H. Schweizer, and M. Seibold, in Proc. SID Int. Symp., Long
Beach, CA (2000), Digest of Technical Papers, Vol. XXXI, pp. 931–933.
299. S. Mikoshiba, Gas-discharge displays, in Wiley Encyclopedia of Electrical and Electronic
Engineering (J. G. Webster, ed.), Wiley-Interscience, New York (1999), vol. 8, pp. 233–
238.
300. J. K. Lee and J. P. Verboncoeur, Plasma display panel in Low Temperature Plasma Phys-
ics: Fundamental Aspects and Applications (R. Hippler, S. Pfau, M. Schmidt, and
K. H. Schoenbach, eds.), Wiley-VCH: Weinheim (2001), pp. 367–385.
301. H. G. Slottow, IEEE Trans. Plasma Sci. 23, 760 (1976).
302. D. L. Bitzer and H. G. Slottow, in Proc. 29th AFIPS Conf., Washington (1966), pp. 541–
547.
303. L. Loeb and J. El Baccal, J. Appl. Phys. 33, 1567 (1962).
304. W. L. Harries and A. von Engel, Proc. Phys. Soc.(London) B64, 916 (1951).
305. H. J. Hoehn and R. A. Martel, IEEE Trans. Electron. Deû. 20, 1078 (1973).
306. I. Revel, Ph. Belenguer, J. P. Boeuf, and L. C. Pitchford, Pure Appl. Chem. 71, 1837
(1999).
Kogelschatz46
307. R. T. Wegh, H. Donker, E. V. D. van Loef, K. D. Oskam, and A. Meijering, J. Luminesc.
87–89, 1017 (2000).
308. M. Miclea, K. Kunze, G. Musa, J. Franzke, and K. Niemax, Spectrochimica Acta B56,
37 (2001).
309. K. Okazaki and T. Nozaki, Ultrashort pulsed barrier discharges and applications, Pure
Appl. Chem. 74, 317 (2002).
310. V. I. Gibalov and G. J. Pietsch, J. Phys. D: Appl. Phys. 33, 2618 (2000).
