Home
Search
Collections
Journals
About
Contact us
My IOPscience
Modelling of NO destruction in a low-pressure reactor by an Ar plasma jet: species
abundances in the reactor
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2011 J. Phys. D: Appl. Phys. 44 105202
(http://iopscience.iop.org/0022-3727/44/10/105202)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 148.6.27.69
The article was downloaded on 22/02/2011 at 10:08
Please note that terms and conditions apply.
IOP PUBLISHING
JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 44 (2011) 105202 (9pp)
doi:10.1088/0022-3727/44/10/105202
Modelling of NO destruction in a
low-pressure reactor by an Ar plasma jet:
species abundances in the reactor
Kinga Kutasi
Research Institute for Solid State Physics and Optics, Hungarian Academy of Sciences, POB 49,
H-1525 Budapest, Hungary
E-mail: kutasi@sunserv.kfki.hu
Received 8 September 2010, in final form 10 December 2010
Published 21 February 2011
Online at stacks.iop.org/JPhysD/44/105202
Abstract
The destruction of NO molecules by an Ar plasma jet in a low-pressure (0.2 Torr) reactor is
investigated by means of a 3D hydrodynamic model. The density distribution of species
created through molecular kinetics triggered by the collision of Ar+ with NO is calculated,
showing that in the case of the most abundant species a quasi-homogeneous density
distribution builds up in a large part of the reactor. The conversion of NO into stable O2 and
N2 molecules is followed under different plasma jet conditions and NO gas flows, and the
effect of N2 addition on NO destruction is studied. It is shown that in the present system the
reproduction of NO molecules on the surface through surface-assisted recombination of N and
O atoms becomes impossible due to the fast disappearance of N atoms in the jet’s inlet vicinity.
injected [18, 19]. The large pressure difference between
the cascaded arc source (40 kPa) and the process chamber
(typically 20–100 Pa) causes a supersonic expansion of the
plasma from the nozzle of the cascaded arc into the chamber.
The high-velocity (≈2000 m s−1 ) Ar ions so introduced into the
vessel can dissociate the NO molecules [17, 20]. The reduction
of the NO molecules in the system under different discharge
conditions (arc current) and gas flow rates is determined
through mass spectrometry measurements. The NO reduction
is shown to increase with the arc current and with a decrease
in the NO flow.
In addition to the destruction of NO molecules the ETP
system can also be interesting for providing a plasma composed
of N2 , O2 and NO molecules (ground state and excited) and N
and O atoms, that can be used for different applications, e.g.
plasma sterilization [21, 22], etching [23–25] and oxidation
[26]. Since ETP is a system that is not driven by electroninduced ionization and excitation, the plasma composition
in the reactor can be varied through parameters that can be
more easily controlled than the electron density and energy
distribution, such as the molecular gas flow [27].
The aim of this work is to study theoretically the
destruction of NO molecules in a low-pressure reactor under
different conditions, and to determine the distribution of
species densities in the reactor—that are difficult to access
1. Introduction
Due to environmental importance, great attention is given to
the destruction of NOx molecules. Toxic nitrogen oxides are
present in the emission of combustion gases and are sources of
photochemical smog formation, acid rain and the greenhouse
effect. The decomposition of NO has been studied by various
types of plasmas in different mixtures mostly at atmospheric
pressures [1–16]. Many investigations have been carried out
in N2 –NO gas mixture discharges in order to study the effect of
N2 on the depletion efficiency, as well as the role of different
processes. It has been shown that in N2 –NO discharges a
small quantity of NO can be converted into N2 and O2 with
a high decomposition efficiency [3, 4, 6]. NO molecules are
also produced in low-pressure processing systems; they can be
the end product—through surface-assisted recombination—of
N and O atoms created in low-pressure N2 –O2 discharges.
Therefore the study of NO destruction under low-pressure
conditions can also be important.
For NO destruction at low pressures van Helden et al [17]
have conducted experimental investigations at p < 100 Pa
using an expanding thermal plasma (ETP) system. The ETP
consists of a high-pressure thermal plasma, here namely a
dc cascaded Ar arc discharge, and a low-pressure process
chamber, where the molecular gases to be dissociated are
0022-3727/11/105202+09$33.00
1
© 2011 IOP Publishing Ltd
Printed in the UK & the USA
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
the Soret and pressure diffusions are neglected, as well as the
Dufour effect [30]:
ρ v · n dS = 0,
(1)
S
ρyk v · n dS − ∇(Dk ρyk ) · n dS =
mk SkV dV
S
S
V
S
+ mk Sk dS,
(2)
S
ρui v · n dS =
S
S
ρT v · n dS =
S
µ grad ui · n dS −
S
p ii · n dS,
(3)
S
λ
grad T · n dS.
Cp
(4)
Here ρ denotes the total gas density (mass density), v the
gas velocity and n the unit vector orthogonal to the S surface
and directed outwards. Further, yk denotes the relative mass
density (yk = ρk /ρ), Dk and mk are the diffusion coefficient
and the mass of the species k, and SkV and SkS represent the
source terms associated with volume and surface reactions,
respectively. Since SkS represents a term taking into account
surface losses, this term is considered in (2) only on the last
grid point at the proximity of the surface. ui is the velocity in
the i direction, p the static pressure, µ the dynamic viscosity, T
is the gas temperature, Cp the specific heat at constant pressure
and λ the thermal conductivity.
One of the main aims of the model calculations is the
determination of the species density distributions in the reactor.
The density of species in the system is determined by the
chemical kinetics, which is incorporated into the model in the
continuity equation of different species through the sources
and losses of species resulting from the different chemical
reactions, denoted by SkV in equation (2). The neutral species
kinetics in the reactor starts up with the creation of N and
O atoms from the dissociation of NO molecules through the
following reactions: Ar + +NO → Ar+NO+ (2.7×10−16 m3 s−1
[31]) and NO+ + e → N + O (2 × 10−13 m3 s−1 for the
recombination of NO+ (v = 0) with 0.1 eV electrons [32]).
In order to simplify our model we do not follow the electrons,
which in fact are low-energy electrons with Te = 0.1–0.3 eV,
as reported in [17, 18]; thus these electrons do not play an
important role in the excitation and ionization kinetics, they
are involved only in the recombination processes. The electron
dissociative recombination of molecular ion, created in the
charge transfer reaction presented above, is very fast [32],
therefore we assume that the collision of the Ar+ ion with NO
results in the dissociation of the molecule (i.e. the molecular
dissociation occurs in one step Ar+ +NO → [Ar +NO+ , NO+ +
e] → N + O), producing ground-state N(4 S) and O(3 P) atoms.
With the O(3 P) and N(4 S) atoms in the reactor appear the
surface processes, which are the atomic recombination. The
recombination of atomic species on surfaces, characterized
by the surface recombination coefficient, depends on many
parameters that can change from one experimental condition to
another, e.g. surface material purity, cleanliness, morphology,
oxide or nitride type, surface temperature, surface coverage
and plasma environment. Therefore it is very difficult to
Figure 1. Structure of the 40 cm × 25 cm × 25 cm parallelepipedic
reactor. The reactor has three inlets as follows: (i) a 4 × 4 mm2
square inlet on the left side plate, (ii) a 4 × 4 cm2 inlet on the top
plate and (iii) a 4 × 4 cm2 inlet on the bottom plate. The 4 × 4 cm2
gas outlet is positioned on the top plate.
(This figure is in colour only in the electronic version)
experimentally—by means of a 3D hydrodynamic model
presented in section 3. As shown by Kaminska et al [28]—and
as will be discussed in detail in section 4—the velocity and the
ionization degree of a plasma jet generated by a dc arc depend
on the discharge current. Therefore, to simulate the different
discharge conditions, i.e. different arc currents, calculations
are performed at different inlet Ar–Ar+ jet mixtures and flows.
2. System set-up
The system investigated in this work has a similar structure to
that of van de Sanden et al [18] and van Helden et al [17]. Here
the plasma reactor is a parallelepipedic stainless steel chamber
with dimensions of 40 cm × 25 cm × 25 cm (x, y, z). The
4 × 4 mm2 square inlet, where the high-velocity Ar plasma jet
from the dc cascaded arc source enters the reactor, is located in
the middle on the left plate, with the 4×4 cm2 gas outlet on the
top plate, as shown in figure 1. Two more inlets of 4 × 4 cm2 ,
which serve as inlets for the molecular gases, are located on
the bottom and top plates, respectively, at about 2 cm from the
left plate.
3. Hydrodynamic model
The plasma generated at 0.2 Torr in a reactor by a high-velocity
Ar plasma jet produced in the external cascaded arc source
(not modelled here) is described with a hydrodynamic model,
that has been found feasible for the study of the expansion
of a supersonic cascaded arc plasma into a low-pressure
atmosphere [29]. The three-dimensional hydrodynamic model
developed by us is composed of (i) the total mass conservation,
(ii) the continuity equations for the different species, (iii) the
total momentum conservation equation and (iv) the total energy
conservation. The gas is assumed to be a Newtonian fluid. The
continuity equations can be written in the following form when
2
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
Table 1. The main reactions taken into account in the hydrodynamic model. The rate coefficients are taken from [22, 36–38]. The rate
coefficients for the two- and three-body reactions are in m3 s−1 and m6 s−1 , respectively, and the decay frequencies are in s−1 ; T is the
temperature in K.
Processes
(R1)
(R2)
(R3)
(R4)
(R5)
(R6)
(R7)
(R8)
(R9)
(R10)
(R11)
(R12)
(R13)
(R14)
(R15)
(R16)
(R17)
(R18)
(R19)
(R20)
(R21)
(R22)
(R23)
(R24)
(R25)
(R26)
(R27)
(R28)
(R29)
(R30)
(R31)
(R32)
(R33)
(R34)
(R35)
(R36)
(R37)
(R38)
Rate coefficients
Ar + NO(X) → Ar + NO
NO+ + e → N(4 S) + O(3 P)
N(4 S) + wall → 21 N2 (X) + wall
O(3 P ) + wall → 21 O2 (X) + wall
N(4 S) + O(3 P) + wall → NO(X) + wall
N(4 S) + NO(X) → O(3 P) + N2 (X, v = 3)
N(4 S) + N(4 S) + Ar → N2 (B) + Ar
N(4 S) + N(4 S) + N2 → N2 (B) + N2
N2 (B) + NO → N2 (A) + NO
N2 (B) + N2 (X) → N2 (A) + N2 (X)
N2 (B) → N2 (A) + hν
N2 (B) + O2 → N2 (X) + O(3 P) + O(3 P)
N2 (A) + NO(X) → N2 (X) + NO(A)
N(4 S) + O2 (X) → NO(X) + O(3 P)
O(3 P) + O(3 P) + Ar → O2 (X) + Ar
O(3 P) + O(3 P) + O2 → O2 (X, a, b) + O2
O(3 P) + O(3 P) + O(3 P) → O2 + O(3 P)
O(3 P) + O2 (X) + O2 → O3 + O2
O(3 P) + O(3 P) + O2 → O3 + O(3 P)
O2 (a) + O(3 P) → O2 (X) + O(3 P)
O3 + O(3 P) → O2 (a, b, X) + O2 (X)
O2 (b) + O(3 P) → O2 (X) + O(3 P)
O2 (b) + Ar → O2 (X) + Ar
O(3 P) + O2 + Ar → O3 + Ar
O2 (a) + Ar → O2 (X) + Ar
N(4 S) + O(3 P) + Ar(N2 , O2 ) → NO(X) + Ar(N2 , O2 )
N(4 S) + O(3 P) + Ar(N2 , O2 ) → NO(B) + Ar(N2 , O2 )
N(4 S) + O(3 P) + Ar(N2 , O2 ) → NO(A) + Ar(N2 , O2 )
N(4 S) + O(3 P) → NO(A)
NO(A) → NO(X) + hν
NO(B) → NO(X) + hν
O(3 P) + O2 (X) + N2 → O3 + N2
O(3 P) + NO(X) + N2 (Ar) → NO2 (X) + N2 (Ar)
O(3 P) + NO(X) + O2 → NO2 (X) + O2
NO(X) + O(3 P) + N2 → NO2 (A) + N2 → NO2 (X) + N2
NO(X) + O(3 P) + O2 → NO2 (A) + O2 → NO2 (X) + O2
N(4 S) + NO2 (X) → NO(X) + NO(X)
NO2 (X) + O(3 P) → NO(X) + O2 (X)
+
+
Table 2. The diffusion coefficients given in m2 s−1 are taken from
[39–42]. The temperature T is in K and the pressure p in Pa.
define a proper surface recombination coefficient for atoms
(γ ) when it comes to modelling of a given experimental
condition. The choice of γ has been discussed in detail in
our previous work, where the effect of the different γ values
has also been investigated [27]. Here we just recall, that for
the stainless steel surface in the case of N atoms we choose
γN = 7.5 × 10−2 [27, 33, 34], while in the case of O atoms we
choose γO = 7×10−2 [35]. The losses of atoms are calculated
according to the equation
vk
SkS = −γk nk ,
4
2.7 × 10−16 m3 s−1
2 × 10−13 m3 s−1
γ
γ
γ
1.05 × 10−18 T0.5
0.3 × 8.27 × 10−46 exp(T /300)
8.27 × 10−46 exp(T /300)
2.4 × 10−16
0.95 × 3 × 10−17
2 × 105
3 × 10−16
6.6 × 10−17
1.1 × 10−20 T exp(−3150/T )
5.21 × 10−47 exp(900/T )
3.8 × 10−42 exp(−170/T )/T
3.6 × 10−44 T −0.63
6.4 × 10−47 exp(663/T )
2.1 × 10−46 exp(345/T )
7 × 10−22
1.8 × 10−17 exp(−2300/T )
4 × 10−20
1.5 × 10−23
3.9 × 10−46 (300/T )1.9
1.5 × 10−26
1.76 × 10−43 T −0.5
3.09 × 10−46 (T /300)−1.4
2.12 × 10−46 (T /300)−1.24
1.18 × 10−23 (T /300)−0.35
4.5 × 106
3 × 105
5.7 × 10−46 (300/T )2.8
1 × 10−43
8.6 × 10−44
3.7 × 10−44
3.7 × 10−44
2.3 × 10−18
9.7 × 10−18
1.5
D(Ar) = 3.5 × 10−4 T p
D(O2 (a, b)) = 3.87 × 10−3 Tp
1.75
D(O2 , O3 ) = 8.8 × 10−5 T p
D(O) = 5.85 × 10−4
T 1.5
p
1.5
D(N2 (X, A, B)) = 3.99 × 10−4 T p
D(N) = 6 × 10−4
T 1.75
p
1.69
D(NO(X, A, B)) = 1.42 × 10−4 T p
1.5
D(N2 (X, A, B)) = 3.99 × 10−4 T p
(5)
D(NO2 ) = 6.24 × 10−6
√
where vk = 8kB T /πmk is the average velocity of k atoms
and γk is the corresponding atomic surface loss probability.
The N and O atoms on the surface can recombine either into N2
and O2 , respectively, or into NO molecules (table 1 (R3)–(R5)).
The experimentally determined γ surface recombination
coefficient includes all the possible surface reactions, thus
making it possible to describe the loss and creation of species
T 1.75
p
on the surface without a detailed surface kinetic model.
However, it gives no possibility to decide which recombination
process occurs, i.e. whether the N atoms recombine into N2 or
into NO together with the O atoms. The implementation of the
atomic surface losses and the choice of the surface elementary
processes were discussed in detail in [27].
3
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
Figure 2. The distribution of species abundances (%) in the x–z vertical plane at y = 12.5 cm for the following NO inlet velocities:
(a) 20 m s−1 , (b) 12 m s−1 and (c) 5 m s−1 when the Ar jet inlet velocity is 2000 m s−1 and the ionization degree is 15%. Columns 1 to 3 are
for different molecules.
With the appearance of N2 and O2 molecules in addition to
the O(3 P) and N(4 S) atoms further gas phase reactions can take
place in the reactor filled with NO, which give rise to excited
and newly formed molecules. First of all the N and O atoms
can also recombine in the gas phase through two-body, as well
as three-body processes. The most important process for the
N(4 S) atoms in a system with predominant NO molecules is
the dissociative collision with NO(X) giving rise to N2 and
O(3 P) (R6). The three-body recombination of N(4 S) can result
in excited N2 (B) molecules (R7) and (R8), which through
quenching by NO, N2 and radiative decay turn into metastable
N2 (A) molecules (R9)–(R11). The N(4 S) atoms can further
contribute to the formation of ground-state NO(X) molecules
through two-body collision with O2 (X) (R14). In the case
of O(3 P) atoms, their three-body recombination can result in
ground-state O2 and excited O2 (a) and O2 (b) molecules, as
well as O3 (R15)–(R19). The O atoms together with the N
atoms can also contribute to the formation of NO(X), NO(A)
and NO(B) molecules through the two-body and three-body reassociation processes in the presence of Ar, N2 and O2 (R29),
(R26)–(R28). And finally, the NO molecules participate in
the creation of NO2 (X) through the three-body re-association
with O(3 P) in the presence of N2 , O2 and Ar, respectively
(R33)–(R36). A full list of gas phase reactions that govern
the molecular kinetics of an Ar–NO–N2 –O2 system has been
given in our previous publication [27].
The distribution of species densities in the reactor in
addition to the chemical reactions is influenced by the gas
flow and the diffusion of different species. The diffusion
coefficients chosen are presented in table 2. The transport
data necessary as input data for the model are taken from
[38, 43]. The inlet and wall temperatures required for the
energy conservation equation are defined as follows: the inlet
temperature of the Ar jet is taken to be 12 000 K according to
Selezneva et al [29], while the inlet NO and wall temperatures
are chosen as 300 K [29].
As concerns the solution method, the model is solved
using the algorithm given by Ferziger and Perić [44]. The
equations are discretized using the finite volume method. The
linear algebraic equation system so obtained is then solved
with Stone’s method iteratively using the multigrid method.
In our solution three grid levels are used, the finest grid has
80 × 40 × 80 control volumes.
4. Results and discussion
One of the recent modelling investigations of an argon plasma
jet generated by a dc arc is that of Kaminska et al presented
in [28]. They discuss the dc current dependence of jet’s
velocity in different parts of the divergent nozzle, as well as
the ionization degree of the plasma jet. They have shown
that in the 60–140 A current range at the end of the nozzle
4
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
Figure 3. The distribution of O atom abundance (%) in the x–z vertical plane at y = 12.5 cm for different NO inlet velocities from column 1
to 3, when the Ar jet inlet velocity is 2000 m s−1 and the ionization degree is 15%.
velocities between 3500 and 5500 m s−1 are obtained, while
the ionization degree can reach 60%. Regarding the ionization
degree, the calculations of Beulens et al [45] show that it is
about 15% at 60 A dc current and increases up to ≈75% at
200 A. Along with these results we have conducted systematic
calculations varying the jet velocity between 2000 m s−1
(estimated by Engeln et al [19]) and 3500 m s−1 , and the
ionization degree in the 15–35% range, as an attempt to
simulate the lower current conditions.
Figure 2 shows the distribution of relevant species—NO,
N2 and O2 molecules created through the recombination of
atoms provided by the dissociation of NO—in the reactor at
three different NO inlet velocities when the Ar jet inlet velocity
is set to 2000 m s−1 and the ionization degree to 15%. The
distributions are shown in the vertical symmetry plane of the
reactor, i.e. x–z vertical plane at y = 12.5 cm. The figures
show that in the case of these species at ≈10 cm from the jet
entrance quasi-homogeneous density distributions are built up.
On the other hand, along the same distance the density of N
atoms becomes negligible. The N atoms created from NO
dissociation cannot survive long, since they collide with the
non-dissociated NO molecules and recombine into N2 giving
rise to O atoms. As a result the reproduction of NO molecules
through surface processes, i.e. surface recombination of N and
O atoms, becomes impossible. The density of O atoms is
about one order of magnitude lower than that of O2 ; their
recombination into O2 occurs through a surface recombination
process, which results in a decrease in density distribution from
the jet’s entrance vicinity to the walls, as illustrated in figure 3.
With increasing NO inlet velocity a slight increase in O atom
abundance can be observed at the entrance vicinity; however, in
the big part of the reactor very similar abundances are obtained
for different NO inlet flows.
Figure 2 also clearly shows that by decreasing the
NO flow—from figure 2(a) to (c)—the dissociation of NO
molecules becomes more effective, i.e. the N2 and O2
abundances increase, while that of NO decreases. To have
a more clear idea about the depletion of NO molecules under
different conditions we present the reduced O2 and N2 mass
density fraction, ([N2 ] + [O2 ])/([N2 ] + [O2 ] + [NO]) [17], as
a function of NO inlet velocities as shown in figure 4. First of
all we investigate how the conversion of NO into N2 and O2
evolves with the ionization degree of the jet. The full symbols
Figure 4. Reduced O2 and N2 mass density fraction as a function of
NO inlet flow. The full symbols represent the case of 2000 m s−1
velocity plasma jet with different ionization degrees 15% ( ), 25%
(•) and 35% (). The open symbols represent two cases of higher
jet velocity as follows: 3000 m s−1 with 25% ionization degree (◦)
and 3500 m s−1 with 35% Ar+ ().
in figure 4 represent the reduced O2 and N2 density fraction for
three different ionization degrees: 15%, 25% and 35%, when
the jet velocity is kept constant at 2000 m s−1 . As already
suggested by the density distributions we can observe an
increase in the NO conversion with decreasing NO flow, shown
experimentally also by van Helden et al (figure 12 in [17]).
Similarly, the NO conversion increases also with the ionization
degree of the plasma jet, which is in concordance with the NO
conversion observed experimentally with the arc current (one
can refer to figure 13 in [17]). According to the calculations
of Kaminska et al [28], it is not only the ionization degree
that increases with the arc current, but also the jet velocity.
Accordingly, in the following when choosing the ionization
degree as 25% we increase the jet velocity to 3000 m s−1 , while
in the case of 35% to 3500 m s−1 . The results thus obtained
are illustrated by the open symbols presented in figure 4. As
expected, the increase in the jet velocity, which ensures a
higher Ar+ inlet mass flow, results in a more efficient NO
destruction. In this way the highest conversion rate obtained
is 75% in comparison with 50% resulting from the previous
conditions.
5
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
Figure 5. The distribution of (a) O atom and (b) N atom abundances (%) in the x–z vertical plane at y = 12.5 cm for different NO inlet
velocities from column 1 to 3, when the Ar jet inlet velocity is 3500 m s−1 and the ionization degree is 35%. Abundances lower than 0.01%
are omitted in the figure.
By increasing the ionization degree of the cascaded arc,
in addition to the more efficient NO destruction, higher
atomic densities are also expected. Figure 5 shows the
distribution of the N and O atom abundances for different
NO inlet velocities in the case of 3500 m s−1 jet velocity and
35% ionization degree. Indeed, considerably higher O atom
abundances (it decreases from 1.4% to 0.3% to the walls) are
obtained than in the case of lower ionization degree jet, for
comparison one should refer to figure 3. It is also interesting
to examine the density distribution of atoms in the reactor.
While the distribution of O atom abundances is very similar
for the different NO inlet flows, the N atoms show a different
behaviour with the flow. With increasing NO inlet flow the
N atoms are quenched much faster due to collision with NO
molecules (R6).
Assuming a reactor surface that has a much lower atomic
recombination probability, such as γ = 10−3 , we can examine
the influence of surface recombination on the atomic density
distributions. Figure 6 shows the N and O atom abundances
in the reactor for the lower NO flow investigated here. We
can notice that the surface recombination has no effect on the
N atom density, while in the case of O atoms we can observe
considerable increase in the density compared with the surface
with higher recombination probability, see figure 5(a) first
column.
Figure 6. The distribution of (a) O atom and (b) N atoms
abundances (%) in the x–z vertical plane at y = 12.5 cm when the
NO inlet velocity is 2 m s−1 , Ar jet inlet velocity is 3500 m s−1 , the
ionization degree is 35% and the atomic surface recombination
probabilities are chosen as 10−3 . Abundances lower than 0.01% are
omitted in the figure.
Addition of N2 . The destruction of NO at atmospheric
pressure has been widely investigated in N2 –NO gas mixture
discharges, achieving a better efficiency with N2 addition.
In the active discharge for this mixture the main NO
destruction mechanism is the collision of NO with N atoms
[4]. However, in the ETP both charge transfer and electroninduced dissociative recombination (Ar + + NO → Ar + NO+
and NO+ + e → N + O), and the N-atom-induced dissociation
(NO + N → N2 + O) processes play comparable roles.
In the present system N2 has been introduced by replacing
the NO gas flow with N2 gas flow at the bottom plane inlet
(see figure 1). With the addition of N2 to the system, the
dissociation reactions of N2 and NO through charge transfer
and electron-induced dissociative recombination become
competing reactions. In this case the N atoms can be created
6
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
Figure 7. The distribution of (a) N2 , (b) N, (c) O2 , (d) O and (e) NO abundances (%) in the x–z vertical plane at y = 12.5 cm for different
w(N2 ) = −w(NO) inlet velocities from column 1 to 3, when the Ar jet inlet velocity is 3500 m s−1 and the ionization degree is 35%.
through two different reaction paths: (i) Ar + + NO → Ar +
NO+ (2.7 × 10−16 m3 s−1 [31]), NO+ + e → N + O (2 ×
10−13 m3 s−1 [32]) and (ii) Ar + + N2 → Ar + N2+ (4.45 ×
10−16 m3 s−1 [46]), N+2 + e → N + N (2 × 10−13 m3 s−1
[47]), respectively. Again for simplicity (see section 3), in
order not to follow also the slow electrons, and since the
electron recombination dissociation reaction is very fast, the
dissociation of molecules is assumed to occur in one step as
follows: Ar+ + NO → [Ar + NO+ , NO+ + e] → N + O and
Ar+ + N2 → [Ar + N2+ , N+2 + e] → N + N.
Figure 7 represents the distribution of species abundances
in the reactor’s symmetry plane. In the case of N atoms,
compared with the pure NO case, considerably higher densities
are obtained, although still abundances less than 1% are
achieved, figure 7(b). In this case nevertheless there is an N
atom density built up at the reactor’s wall vicinity; however, the
contribution of N atoms to the surface formation of molecules
is negligible. With increasing molecular gas flows we can
observe faster quenching of N atoms similarly to the case of
pure NO inlet. For the O2 molecules the higher abundances
obtained at lower molecular flows predict the higher depletion
rate of NO molecules, figure 7(c). Contrary to the molecules,
the O atom density, (figure 7(d)) increases with the molecular
gas flow; however, for each case it reaches similar values at
the wall.
7
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
NO molecules on the wall through the surface recombination
of N and O atoms becomes impossible.
We have followed the conversion of NO into stable O2 and
N2 molecules under different plasma jet conditions and NO gas
flows, and have shown that the NO reduction becomes more
efficient with decreasing NO inlet flow and increasing plasma
jet ionization degree and velocity, which is equivalent to an
increase in the dc arc current. By adding N2 to NO—replacing
one of the NO inlet flows with N2 —we have observed a more
efficient NO destruction. Although in this way more nitrogen
has been added to the system, the N atom abundances do not
exceed 1%.
Acknowledgments
The work has been supported by the Hungarian Science
Foundation OTKA through project F-67556 and by the Janos
Bolyai Research Scholarship of the Hungarian Academy of
Sciences.
Figure 8. Reduced O2 mass density fraction as a function of equally
set NO and N2 inlet flows in the case of 3500 m s−1 velocity plasma
jet with 35% ionization degree.
We follow the conversion of NO in this case through
the reduced O2 density defined as ([O2 ])/([O2 ] + [NO])
presented in figure 8 in the case of Ar plasma jet velocity
3500 m s−1 and ionization degree 35%. Although, a one to
one comparison cannot be made with the reduced densities
defined in the case of pure NO (figure 4), obviously with N2
addition the depletion of NO becomes more efficient. With
N2 addition although the dissociation reactions of N2 and
NO through charge transfer and electron-induced dissociative
recombination are competing reactions, the N atoms resulting
from N2 dissociation contribute strongly to the NO depletion
through the (R6) process.
References
[1] Clements J C, Mizuno A, Finney W C and Davis R H 1989
IEEE Trans. Indust. Appl. 25 62
[2] Masuda S and Nakao H 1990 IEEE Trans. Indust. Appl. 26 374
[3] Penetrante B M, Hsiao M C, Merritt B T, Vogtlin G E and
Wallman P H 1995 Appl. Phys. Lett. 67 3096
[4] Penetrante B M, Hsiao M C, Merritt B T, Voglin G E and
Wallman P H 1996 Appl. Phys. Lett. 68 3719
[5] Kiyokawa K, Matsuoka H, Itou A, Hasegawa K and
Sugiyama K 1999 Surf. Coat. Technol. 112 25
[6] Gal A, Kurahashi M and Kuzumoto M 1999 J. Phys. D: Appl.
Phys. 32 1163
[7] Wojtowicz M A, Miknis F P, Grimes R W, Smith W W and
Serio M A 2000 J. Hazard. Mater. 74 81
[8] Baeva M, Gier H, Pott A, Uhlenbusch J, Höschele J and
Steinwandel J 2001 Plasma Chem. Plasma Process. 21 225
[9] Tsuji M, Nakano K, Kumagae J, Matzuzaki T and Tsuji T
2003 Surf. Coat. Technol. 165 296
[10] Chang M B, Kushner M J and Rood M J 1992 Env. Sci.
Technol. 26 777
[11] Gentile A C and Kushner M J 1995 J. Appl. Phys. 78 2074
[12] Dorai R and Kushner M J 2003 J. Phys. D: Appl. Phys. 36 1075
[13] Nusca M J, Rosocha L A and Herron J T 2000 38th Aerospace
Sciences Meeting and Exhibit (Reno, NV, 10–13 January
2000) AIAA 2000-0719
[14] Eichwald O, Guntoro N A, Yousfi M and Benhenni M 2002
J. Phys. D: Appl. Phys. 35 439
[15] Orlandini I and Riedel U 2000 J. Phys. D: Appl. Phys. 33 2467
[16] Filimonova E A, Kim Y, Hong S H and Song Y H 2002
J. Phys. D: Appl. Phys. 35 2795
[17] van Helden J H, Zijlmans R A B, Schram D C and Engeln R
2009 Plasma Sources Sci. Technol. 18 025020
[18] van de Sanden M C M, Severens R J, Kessels W M M,
Meulenbroeks R F G and Schram D C 1998 J. Appl. Phys.
84 2426
[19] Engeln R, Mazouffre S, Vankan P, Schram D C and Sadeghi N
2001 Plasma Sources Sci. Technol. 10 595
[20] Zijlmans R A B, Welzel S, Gabriel O, Yagci G, van Helden J
H, Röpcke J, Schram D C and Engeln R 2010 J. Phys. D:
Appl. Phys. 43 115204
[21] Ricard A and Monna V 2002 Plasma Sources Sci. Technol.
11 A150
[22] Kutasi K, Saoudi B, Pintassilgo C D , Loureiro J and
Moisan M 2008 Plasma Process. Polym. 5 840
5. Concluding remarks
We have studied the destruction of NO molecules by an Ar
plasma jet in a low-pressure p = 0.2 Torr reactor. The Ar
plasma jet produced by a dc cascaded arc (not studied here) is
introduced through a 4 × 4 mm2 orifice into the low-pressure
reactor. NO gas is fed into the reactor through two inlets placed
symmetrically on the top and on the bottom planes. Through
the molecular kinetics triggered by the collision of Ar+ with
NO several species are created in the reactor. To calculate
the density distributions of these species we have used a 3D
hydrodynamic model. Depending on the dc current of the arc
discharge the ionization degree of the plasma jet, as well as its
velocity, can vary in a wide range. With an attempt to simulate
the lower arc dc current situations we varied the jet velocity
between 2000 and 3500 m s−1 , and the ionization degree in the
15–35% range.
Calculations have shown that in the case of the most
abundant species, such as N2 , O2 and NO, a quasihomogeneous density distribution builds up in a large part
of the reactor, while the N atoms disappear within a few
centimetres. In the case of O atoms we obtain a decreasing
density profile from the vicinity of jet entrance to the walls.
We have shown that due to the fast disappearance of N atoms
through their collision with NO molecules, the reproduction of
8
J. Phys. D: Appl. Phys. 44 (2011) 105202
K Kutasi
[34] Adams S F and Miller T A 2000 Plasma Sources Sci. Technol.
9 248
[35] Mozetič M and Zalar A 2000 Appl. Surf. Sci. 158 263
[36] Guerra V and Loureiro J 1999 Plasma Sources Sci. Technol.
8 110
[37] Pintassilgo C D, Loureiro J and Guerra V 2005 J. Phys. D:
Appl. Phys. 38 417
[38] Belmonte T, Czerwiec T, Gavillet J and Michel H 1997 Surf.
Coat. Technol. 97 642
[39] Tachibana K and Phelps A V 1981 J. Phys. Chem.
75 3315
[40] Lide D R 1994 Handbook of Chemistry and Physics 75th edn
(Boca Raton, FL: CRC Press)
[41] Morgan J E and Schiff H I 1964 Can. J. Chem. 42 2300
[42] Bockel S, Belmonte T, Michel H and Ablitzer D 1997 Surf.
Coat. Technol. 97 618
[43] Kutasi K, Pintassilgo C D, Coelho P J and Loureiro J 2006
J. Phys. D: Appl. Phys. 39 3978
[44] Ferziger J H and Peric M 2002 Comput. Methods Fluid
Dynamics 3rd rev. edn (Berlin: Springer)
[45] Beulens J J, Milojevic D, Schram D and Vallinga P M 1991
Phys. Fluids B 3 2548
[46] Viaggiano A A and Morris R A 1993 J. Chem. Phys.
99 3526
[47] Peterson J R et al 1998 J. Chem. Phys. 108 1978
[23] Lin T H, Belser M and Tzeng Y 1988 IEEE Trans. Plasma Sci.
16 631
[24] Brussaard G J H, Letourneur K G Y, Schaepkens M,
van de Sanden M C M and Schram D C 2003 J. Vac. Sci.
Technol. B 21 61
[25] Hody V, Belmonte T, Pintassilgo C D, Poncin-Epaillard T,
Czerwiec T, Henrion G, Segui Y and Loureiro J 2006
Plasma Chem. Plasma Process. 26 251
[26] Yasuda Y, Zaima S, Kaida T and Koide Y 1990 J. Appl. Phys.
67 2603
[27] Kutasi K 2010 J. Phys. D: Appl. Phys. 43 055201
[28] Kaminska A, Lopez B, Izrar B and Dudeck M 2008 Plasma
Sources Sci. Technol. 17 035018
[29] Selezneva S E, Boulos M I, van de Sanden M C M, Engeln R
and Schram D C 2002 J. Phys. D: Appl. Phys.
35 1362
[30] Kutasi K, Pintassilgo C D and Loureiro J 2009 2nd Int.
Workshop on Non-equilibrium Processes in Plasmas and
Environmental Science (Belgrade, Serbia) J. Phys.: Conf.
Ser. 162 012008
[31] Shul R J, Upschulte B L, Passarella R, Keesee R G and
Castleman A W Jr 1987 J. Phys. Chem. 91 2556
[32] Florescu-Mitchell A I and Mitchell J B A 2006 Phys. Rep.
430 277
[33] Marković V Lj, Petrović Z Lj and Pejović M M 1995 Japan. J.
Appl. Phys. 34 2466
9