Plasma Sources Science and Technology

PAPER Related content

- Afterglow chemistry of atmospheric-
An in situ and downstream study of non-thermal pressure helium–oxygen plasmas with
humid air impurity
plasma chemistry in an air fed dielectric barrier Tomoyuki Murakami, Kari Niemi, Timo
Gans et al.

discharge (DBD) - Reactive species output of a plasma jet

with a shielding gas device—combination
of FTIR absorption spectroscopy and gas
To cite this article: Abdullah Al-Abduly and Paul Christensen 2015 Plasma Sources Sci. Technol. 24 phase modelling
065006 A Schmidt-Bleker, J Winter, S Iseni et al.

- Kinetic modelling for an atmospheric

pressure argon plasma jet in humid air
W Van Gaens and A Bogaerts

View the article online for updates and enhancements.

Recent citations
- Microwave air plasmas in capillaries at low
pressure II. Experimental investigation
G D Stancu et al

This content was downloaded from IP address 149.150.236.147 on 16/01/2018 at 21:42

 Plasma Sources Science and Technology

Plasma Sources Sci. Technol. 24 (2015) 065006 (16pp) doi:10.1088/0963-0252/24/6/065006

An in situ and downstream study of

non-thermal plasma chemistry in an air fed
dielectric barrier discharge (DBD)
Abdullah Al-Abduly and Paul Christensen
School of Chemical Engineering and Advanced Materials, Bedson Building, Newcastle University,
Newcastle upon Tyne, NE1 7RU, UK

E-mail: [email protected]

Received 1 May 2015, revised 20 August 2015

Accepted for publication 22 September 2015
Published 22 October 2015

Abstract
This paper reports a spectroscopic study of non-thermal plasma chemistry in an air-fed
dielectric barrier discharge (DBD) plasma jet. In situ analysis (i.e. the analysis of the plasma
glow) and downstream analysis were carried out to identify and monitor species produced
in the plasma as they propagate from the plasma glow to downstream regions. The analyses
were carried out using Fourier Transform InfraRed (FTIR) and UV–Vis spectroscopies. The
species: O3, N2O5, N2O, HNO3, CO2, CO and, for the first time, a vibrationally excited form
of CO2 (i.e. CO*2 (v)) were identified in the plasma glow, while O3, N2O5, HNO3 and N2O
were detected in the downstream exhaust. The behaviour of these species was monitored as a
function of a range of experimental conditions including: input power, gas flow rate, relative
humidity, gas temperature and feed gas composition. In addition, the uncertainty associated
with UV–vis detection of ozone in the presence of N2O5 and/or HNO3 as interfering species
was determined.

Keywords: non-thermal plasma, FTIR spectroscopy, in situ

(Some figures may appear in colour only in the online journal)

1. Introduction typically determined by several parameters such as: feed gas

composition, flow rate, gas temperature and power input.
Non-thermal plasma (NTP) has received significant atten- Hence, determining the composition of the NTP is impor-
tion during the past few decades, and its application in dif- tant, particularly for biomedical applications [14, 15]. Pipa
ferent fields is increasing. One of the most common types of and Ropcke et al [14] have shown that Fourier Transform
NTP generator is dielectric barrier discharge (DBD) in which InfraRed spectroscopy (FTIR) can be employed to identify
the NTP is formed between powered and earthed electrodes the species within the exhaust stream of an argon-fed atmos-
separated by one or more layers of dielectric. In addition to pheric pressure plasma jet (APPJ) which was contaminated
the use of DBD to generate ozone, DBD finds wide applica- with 1.0% ambient air. The authors identified: NO2, NO, N2O,
tion in the industrial, environmental and medical fields e.g.: HNO2, HNO3, CO2 and CO, and suggested that the formation
polymer processing [1–3], the removal of air contaminants of the latter two may due to the interaction of the plasma jet
such as SO2, NOx, VOC [4–6] and the hydrogenation of CO2 with plastic components of the APPJ. The study carried out
[7]. Furthermore, DBD has proved to be useful for the ster- by Reuter et al [16] showed that the ozone formed by feeding
ilization and disinfection of contaminated surfaces [8–11] a mixture of pure oxygen and argon to an APPJ can be moni-
including living tissues [12, 13]. tored in and after the plasma glow region using mid-infrared
Within a NTP, a number of active species are formed quantum cascade laser absorption spectroscopy.
including: free radicals, ions, electrons and reactive neu- The formation of chemically and biologically active species
tral molecules. The amounts and nature of these species are through the interaction of a plasma jet with the surrounding air

0963-0252/15/065006+16$33.00 1 © 2015 IOP Publishing Ltd Printed in the UK

Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

Figure 1. Non-thermal plasma jet transmission cell: (1) CaF2 windows, (2) exhaust gas outlet, (3) Pyrex glass tube with a nozzle, (4) feed
gas flow gap, (5) and (6) high voltage electrodes and (7) in situ sampling zone of the NTP glow.

during the application of an argon fed APPJ was investigated output of the plasma but also the rates of species diffusion
by Bleker et al [15]. In their system, the plasma jet device from the plasma to afterglow regions [20]. Hence, the change
was mounted in a chamber in which humidified mixtures of in the plasma products as a function of runing time was invis-
oxygen and nitrogen were fed to the jet. The exhaust was then tigated in this work. The work reported in this paper seeks to
monitored downstream of the plasma using FTIR spectros- identify and quantify the concentrations of neutral species in
copy. Their results showed that ozone and nitrogen dioxide glow and downstream regions, under continuous gas flow con-
were the main products, with small amounts of H2O2 and ditions. The concentrations of these species were monitored
HNO3. Thus, in contrast to the work of Pipa and Ropcke et al using FTIR spectroscopy coupled with a UV spectrometer.
[14], it may be concluded that the chemical composition of the The DBD employed in this work was designed and fabri-
exhaust of NTP depends on the method of introducing air to cated in-house with a configuration similar to that reported
the plasma jet (i.e. in the feed gas, or by interaction of the air by Koinuma et al and co-workers [21] to allow the formation
surrounding the plasma jet). In addtion, these results suggest of a sufficient plasma jet for the direct in situ analyses. The
that the plasma effluent changes as it move downstream from individual and cumulative effects of: feed gas composition,
the discharge [17]. input power, running time, flow rate and gas temperature on
Despite the wide applications of DBD, there are few studies the chemical composition of NTP were investigated. In order
in which the output of these reactors near the glow region is to maintain the consistency of experimental conditions and
monitored, and no papers report the analysis of the plasma accuracy of measurements all the experiments were carried
glow itself. To the best of our knowledge, the only study in out such that surrounding ambient air was not in contact with
which the near glow region of a DBD was investigated (i.e. the plasma or the plasma exhaust. The effects of other parame-
ca. 2 cm post the plasma glow) was that by Sakiyama et al ters such as the physical and chemical properties of the dielec-
using FTIR spectroscopy [18]. In their work, a numerical tric on the NTP exhaust will be reported in a subsequent paper.
model including 600 chemical reactions was proposed to sim-
ulate the changes in the exhaust of a surface dielectric barrier
2. Experimental
discharge (SDBD) with time. Their model was based on the
assumption that the concentrations of the neutral species in 2.1. Gases
post plasma region are mainly determined by the rate of their
diffusion from the discharge region. In their experimental O2 (⩾95.5%), N2 (>99.99%) and 1.0% CO2 in argon were
work, FTIR analysis was carried out to identify the neutral obtained from BOC. Artificial air (i.e. 20% O2  +  80% N2) and
species in the exhaust of a SDBD ca. 2.0 cm after the plasma ambient air were also used.
glow in a closed system using a fixed volume of humidified,
artificial air (i.e. 80% N2  +  20% O2). The authors detected
2.2. Non-thermal plasma jet generator
a number of neutral species, i.e. O3, N2O, N2O5 and HNO3.
However, no quantitative analysis was conducted to compare A DBD with a coaxial wire-to-cylinder configuration was
the experimental data with those predicted by their model. In employed as the non-thermal plasma jet system. A schematic
the discussion below ‘in situ’ will refer to the monitoring of figure of the DBD configuration is shown in figure 1. As may
the plasma glow. be seen from the figure, the plasma was produced by passing
One of the properties associated with DBD is the contin- the desired gas composition over a central, stainless steel
uous change in the temperature of the dielectric during opera- rod, 1.5 mm thick, mounted in a borosilicate glass tube of
tion [11, 13, 19]; this may be expected not only to affect the 1.8 mm internal diameter, 1.0 mm thick and 6.3 cm in length.

2
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

Table 1. Summary of the experimental conditions employed in this work.

Parameter
Feed gas compositions N2, O2, artificial air (80% N2  +  20% O2), ambient air and 1.0% of CO2 in Ar.
Input power/W 24, 30, 33, 36
Discharge voltage/kV (±0.2) 8.8, 9.2, 9.6, 10
Gas flow rate/dm3 min−1 0.2–0.5
Gas velocity within the DBD/cm s−1 1698–4244
Jet length/cm 0.1–0.6
Relative humidity/% 5.0–80

The external electrode was made of stainless steel foil 10 mm of 308.3  ±  4.0 atm−1 cm−1 for the 254 nm absorbance band
long and ca. 0.1 mm thick, this was wrapped around the glass [23]. The NO2 concentration was monitored at 1631.0 cm−1
tube near the tip and covered with PTFE as an external insu- employing an extinction coefficient of 1.81  ×  10−19 cm2 mol-
lator. The external and internal electrodes were connected to ecule−1 [24], and the concentration of N2O was quantified by
a sinusoidal type, high-voltage power supply (NP-10000-30, integrating the band from 2167 to 2268 cm−1 using an inte-
NeonPro, China) which was employed to provide a range of grated band intensity coefficient of 463 cm−2 atm−1, as deter-
high voltage outputs at a constant frequency of 24 kHz. A vari- mined during this work.
able transformer (Y16HM, Zenith) was used to control the
input power to the HV power supply. The discharge voltage
2.5. Relative humidity control
was monitored using an oscilloscope (PM3217, Philips) cou-
pled with a HV probe (Testec, RS 366-9141, UK) with a max- The relative humidity of the feed gas was controlled using a
imum load of 15 kV. The input power was monitored using humidity regulator which was a jacketed Dreschel bottle con-
a power meter (2000 MU, Pro-digit Electronics). The input taining supersaturated aqueous solution of NaCl (400 cm3).
power values quoted in the text are the total power consumed The temperature of the Dreschel was maintained constant
by the system after subtracting that consumed by the variable between 4.0 and 40.0 °C via a Grant LTC1 Water Recycler.
transformer. The relative humidity of the gas exiting the Dreschel bottle
was monitored using a calibrated Testo 605-H1 humidity
meter. Table 1 summarizes the ranges of the experimental
2.3. FTIR analysis
conditions employed in this work.
FTIR experiments were carried out using a Varian FTS7000
spectrometer equipped with a ceramic, air cooled infrared
2.6. Experimental procedures
source and a liquid nitrogen-cooled MCT detector. In each
experiment, a reference spectrum (SR, 8 cm−1 resolution, The FTIR analyses of the plasma composition was carried
100 co-added and averaged scans) was obtained with the out directly (in situ) by passing the infrared beam through the
gas flowing but with no discharge. A sample spectrum was plasma, or through the exhaust gas ca. 2 m downstream, see
then collected, immediately after which the high voltage figure 2. As can be seen from the figure, the feed gas with
power supply was switched on, and further sample spectra a fixed flow rate was passed through a gas mixing cylinder
(SS) collected at regular intervals up to 10 min. The sample to allow homogeneity when more than one gas was used.
spectra were ratioed to the reference spectrum according to The mixed gas was then passed, if needed, through a rela-
equation (1): tive humidity regulator placed prior to the dielectric barrier
discharge. Depending on the type of experiment, the feed
Absorbance, A = Log10(SS /SR )
(1)
gas was passed either through channel (I) or (II) for in situ or
This data manipulation results in difference spectra in which downstream FTIR monitoring, respectively. As may be seen in
peaks pointing up, to  +(Absorbance), arise from the gain of figures 1 and 2, the DBD was inserted into an in-house built
absorbing species in SS with respect to SR, and peaks pointing infrared (IR) transmission cell, with a pathlength of 5.1 cm
down, to -(Absorbance), to the loss of absorbing species. The made of a Pyrex glass cylinder capped by two 2.5 cm dia.,
FTIR resolution 8.0 cm−1 was chosen to maintain rapid scans, 3.0 mm thick CaF2 windows (Crystran). The total volume of
and the time required for each scan was 45 s. the transmission cell was ca. 47 cm3.
For the in situ experiments, the transmission cell was
placed into the sample compartment of the FTIR spectrom-
2.4. Quantitative analyses
eter such that the IR beam passed through the non-thermal
The ozone concentration was monitored by integrating the plasma jet, the length of which depended on the feed gas flow
FTIR absorbance band from 2070 to 2135 cm−1 using an inte- rate. The exhaust gases exited via a port opposite the DBD,
grated band intensity coefficient (IBI) of 6.04  ×  10−19 cm as can be seen from figure 1. The concentration of ozone in
molecule−1, as reported by Perner et al [22]. UV–vis detec- the exhaust stream was monitored using a UV–Vis flow cell
tion of ozone was carried out using an extinction coefficient (1.0 cm pathlength, Ocean Optics Systems) positioned ca.

3
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

Figure 2. The experimental setup employed for in situ and downstream analysis of NTP. The system consists of: (1) feed gases, (2) flow
meters, (3) mixing cylinder, (4) relative humidity regulator, (5) relative humidity meter, (6) non-thermal plasma jet transmission cell,
(7) in situ FTIR, (8) downstream FTIR and (9) Ocean Optics spectrometer flow cell. Channels I and II were employed for in situ and
downstream monitoring, respectively.

2.7 m after the transmission cell, connected via polyethylene 1055

tubing. The flow cell was connected via fibre optic cables to 0.04
an Ocean Optics USB2000 spectrometer. The time required (i)
(ii) 1030
for ozone to pass from the transmission cell to the UV flow 0.03 (iii)

cell at flow rate of 0.5 dm3 min−1 was ca. 9.0 s. For the down- (iv)
Absorbance

(v)
stream FTIR analyses, channel (II) was employed, where the 0.02 (vi)
(vii)
DBD was connected to an external transmission cell, similar
to that used for the in situ analysis, via 2.0 m 6.0 mm diameter 0.01
2121
polyethylene tubing. The time required for the effluent gas to 2098
pass from the outlet port of the external to the internal trans- 0.00
mission cell was ca. 7.0 s at 0.5 dm3 min−1.
4000 3500 3000 2500 2000 1500 1000
2.7. Plasma jet temperature measurement Wavenumber/ cm
-1

In separate experiments, the change in the plasma jet tempera- Figure 3. In situ FTIR spectra collected as a function of running
ture with time was monitored using a transmission cell similar time: (i) reference spectrum at 0 W, and after (ii) 1 min, (iii) 2 min,
to that employed for in situ and downstream analyses, but with (iv) 3 min, (v) 4 min, (vi) 5 min and (vii) 10 min. The input power
an additional port on the top of the cell. The port at the top was was 36 W, pure oxygen feed at a flow rate of 0.5 dm3 min−1.
used to pass a thermometer (Fisherbrand, FB58555, UK) into
the plasma jet 5.0 mm from the DBD nozzle. gases were tested first, and then gas mixtures. Figure 3 shows
spectra collected as a function of time using pure oxygen. As
can be seen from the figure, the spectra consist of two pairs
3. Results and discussion
of bands, at 2121 & 2098 cm−1 and 1055 & 1030 cm−1, which
may be unambiguosly attributed to the 1st overtone and fun-
3.1. In situ analyses
damental absorptions, respectively, of O3 [25]. Apart from the
3.1.1. The effect of feed gas composition. One of the most distinctive bands due to ozone, no other bands were observed.
important parameters that determine the chemical composition When oxygen was replaced by pure nitrogen, no absorption
of non-thermal plasma is the feed gas composition. Hence, the bands were observed, in agreement with the work of Schmidt-
species formed within the glow and near glow regions were Bleker et al [15]. In all cases, as soon as the DBD power sup-
monitored as a function of the feed gas composition. Pure ply was switched off, all absorptions disappeared.

4
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

N 2O 5 Table 3. List of initiating reactions induced by electron collisions

0.005 NO2
(i) HNO3 HNO3 with different molecules within ambient air [17, 29, 30].
(ii) N2O5
0.004 N2O
(iii) NO2 (R#) Reaction
O3
0.003 CO2* (1) e  +  N2  →  N*2   +  e
Absorbance

N 2O
(2) e  +  O2  →  O*2   +  e
0.002
(3) e  +  CO2  →  CO*2   +  e
0.001 (4) e  +  N2  →  N  +  N  +  e
CO2
(5) e  +  O2  →  O  +  O  +  e
0.000 (6) e  +  CO2  →  CO  +  O
(7) e  +  H2O  →  OH  +  H  +  e
-0.001
(8) e  +  N2  →  N+2   +  e  +  e
2600 2400 2200 2000 1800 1600 1400 1200
(9) +
e  +  O2  →  O2   +  e  +  e
Wavenumber/cm -1

(10) e  +  H2O  →  H2O+  +  e  +  e
Figure 4. In situ FTIR Spectra collected after 1 min at 36 W input (11) e  +  CO2  →  CO+ 2   +  e  +  e
power as a function of the feed gas composition: (i) ambient air, (ii)
artificial air  +40% relative humidity and (iii) dry artificial air. The
gas flow rate was 0.5 dm3 min−1. concentrations of ions were expected to be significantly lower
Table 2. Assignments of the various features observed in the than those of the neutral species (i.e. 4 orders of magnitude
spectra showed in figure 4. [18]) and hence their effects would be limited to the NTP glow
region, these reactions were not included.
Band/cm−1 Assignment Reference
2326, 2366 CO*2 (v) This work 3.1.2. Non-thermal plasma & CO2. From figure 4 it may be
1300, 2211, 2237 N2O [26] seen that two distinctive gain features were observed at 2366
1055, 1030, 2098, 2121 O3 [25] and 2326 cm−1 with air but not with artificial air, and hence
1247, 1720 N2O5 [26] these bands were assumed to be due to one or more forms
1313, 1341, 1700 (sh) HNO3 [26] of CO2. To test this hypothesis, experiments were conducted
1600, 1627 NO2 [26, 28] using CO2 mixed with various feed gases, and the results are
2360, 2340 CO2 [27] shown in figures 5(a) and (b). As can be seen from the figures,
2176, 2121 CO [27] the gain features at 2366 and 2326 cm−1 were superimposed
upon the CO2 loss features, suggesting (given the chemical
Figure 4 shows typical in situ spectra collected after 1 min simplicity of the system) that these bands are due to some
using: (i) ambient air, (ii) humidified artificial air at 40% RH form of CO2. In addition, the intensities of these absorbance
and (iii) dry artificial air at flow rate of 0.5 dm3 min−1 and features were found to be greater in plasma than in the after-
input power of 36 W. The bands observed in the figure are glow region, as will be shown below. We attribute the bands
summarized in table 2. The features at 2121 and 2098 cm−1 at 2366 and 2326 cm−1 to a vibrationally excited (linear) form
can be taken as indicative of ozone production, and hence the of CO2, (i.e. CO*2 (v)). It is highly unlikely that the CO*2 (v) is
features at 1055 and 1030 cm−1 are not shown in the spectra electronically excited, as the first anti-bonding orbital of CO2
presented below for clarity. From figure 4, it can be seen that (2πu) is strongly stabilized by reducing the angle of the mol-
bands due to N2O, O3, N2O5 and NO2 were observed when ecule from 180°, hence CO− 2 is bent [49]. Furthermore, the
the feed gas contained N2 and O2. An additional shoulder at nature of the operation of a Fourier Transform spectrometer
1700 cm−1, and bands at 1341 and 1314 cm−1 were observed rules out the 2366 and 2326 cm−1 features as being due to
when the feed gas was ambient air or humidified artificial air, emission, leaving only absorption by a high concentration of
and may be attributed to HNO3 [26]. The features at 2360 vibrationally excited CO2 molecules.
and 2340 cm−1 in spectra (ii) and (iii) may be unambiguously Modest and Bharadwaj [50] reported transmission spectra
assigned to CO2 [27], and are due to purge mismatch in the of 1% CO2 in dry N2 at temperatures up to 1550 K using long
FTIR sample compartment. The two features in the same pathlength cells. At temperatures  ⩾600 K they observed a
region as CO2 in spectrum (i) may be attributed to a vibra- highly asymmetric band around 4.3 μm (2325 cm−1) with a
tionally excited form of CO2, CO*2 (v), and this is discussed in tail to the low energy side, the band being broader than that
detail below. of CO2 at 298K. The band broadened significantly to the low
Table 3 shows the plasma initiating reactions induced by energy side with increasing temperature. The identity of the
electron collisions with ambient air molecules. The reactions species responsible for the 4.3 μm feature was not stated by
from (R1) to (R11) include electronic and vibrational exci- the authors. Toselli and Barker [51] attributed emission at 4.3
tation, ionization and dissociation. Table 4 lists the chemical μm as being from vibrationally excited levels of the asym-
reactions taking place in the plasma along with their rate coef- metric ν3 O=C=O vibration, ie. ν3  ⩾  1. Hence we attribute
ficients involving free radicals, free radicals  +  neutral spe- the absorptions at 2366 and 2326 cm−1 to vibrational excita-
cies and neutral  +  neutral species. Due to the fact that the tion of the ν3  =  1 excited state of CO2.

5
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

Table 4. List of chemical reactions taking place in humid artificial air non-thermal plasma and their rate coefficients.

(R#) Reaction Rate coefficient (k) Ref

(12) N*2   +  O2  →  O  +  O  +  N2 1.5  ×  10−12 cm3 s−1 [31]
(13) N*2   +  N2O  →  NO  +  N  +  N2 8  ×  10−17 cm3 s−1 [32]
(14) N*2   +  O2  →  N2O  +  O 7.8  ×  10−14 cm3 s−1 [33]
(15) O*2   +  O3  →  O  +  O2  +  O2 5.2  ×  10−11 exp(−2840/Ta) cm3 s−1 [34]
(16) O*2   +  N2  →  O2  +  N2 3.0  ×  10−18 exp (−200/T) cm3 s−1 [35]
(17) N  +  OH  →  NO  +  H 3.8  ×  10−11 exp (−85/T) cm3 s−1 [36]
(18) N  +  O  +  M  →  NO  +  M 6.3  ×  10−33 exp(140/T) cm6 s−1 [37]
(19) N  +  NO  →  N2 (V)  +  O 3.4  ×  10−11 exp (−24/T) cm3 s−1 [37]
(20) N  +  O2  →  NO  +  O 4.4  ×  10−12 exp (−3220/T) cm3 s−1 [36]
(21) N  +  O3  →  NO  +  O2 2.0  ×  10−16 cm3 s−1 [38]
(22) N  +  NO2  →  N2,+  O  +  O 9.1  ×  10−13 cm3 s−1 [38]
(23) N  +  NO2  →  NO  +  NO 6  ×  10−19 cm3 s−1 [39]
(24) N  +  NO2  →  N2O  +  O 5.8  ×  10−12 exp(220/T) cm3 s−1 [37]
(25) N  +  HO2  →  NO  +  OH 1.7  ×  10−17 exp(−1000/T) cm3 s−1 [40]
(26) N  +  HNO  →  H  +  N2O 1.44  ×  10−12 [41]
(27) N  +  N2O  →  N2  +  NO 1.5  ×  10−17 exp(−570/T) cm3 s−1 [37]
(28) N  +  NO  →  N2O 6  ×  10−17 cm3 s−1 [38]
(29) O  +  NO  +M  →  NO2  +M 1  ×  10−43 (300/T)1.6 [37]
(30) O  +  O2  +M  →  O3  +  M 5.6  ×  10−34 (T/ 300)−2.8 cm6 s−1 [35]
(31) O  +  HO2  →  OH  +  O2 2.9  ×  10−11 exp(−200/T) cm3 s−1 [36]
(32) O  +  OH  →  H  +  O2 2.3  ×  10−11 exp (110/T) cm3 s−1 [35]
(33) O  +  O3  →  O2  +  O2 8.0  ×  10−12 exp (−2060/T) cm3 s−1 [36]
(34) O  +  HO2  →  OH  +  O2 2.7  ×  10−11 exp (224/T) cm3 s−1 [35]
(35) O  +  NO2  →  NO  +  O2 6.5  ×  10−12 exp(120/T) cm3 s−1 [36]
(36) O  +  NO2  +  M  →  NO3  +  M 9.0  ×  10−32 (T/300)−2.0 cm6 s−1 [35]
(37) O  +  NO3  →  O2  +  NO2 1.7  ×  10−11 cm3 s−1 [38]
(38) O  +  N2O5  →  products 3  ×  10−16 cm3 s−1 [38]
(39) O  +  HNO  →  OH  +  NO 5.99  ×  10−17 cm3 s−1 [32]
(40) O  +  HNO2  →  NO2  +  OH 2  ×  10−11 exp(−3000/T) cm3 s−1 [37]
(41) O  +  N2O  →  N2  +  O2 4.4  ×  10−17 cm3 s−1 [37]
(42) H  +  HNO2  →  NO2  +  H2 2  ×  10−17 exp(−3700/T) cm3 s−1 [37]
(43) H  +  HNO3  →  NO2  +  H2O 1.39  ×  10−20 (T/298)3.29 exp(−3160/T) [42]
(44) H  +  O2  +  M  →  HO2  +  M, 1.8  ×  10−32 (T/298)−0.8 cm6 s−1 [43]
(45) H  +  O3  →  OH  +  O2 1.4  ×  10−10 exp (−470/T) cm3 s−1 [37]
(46) H  +  H  +  M  →  H2  +  M 1.8  ×  10−42/T [44]
(47) H  +  OH  +  M  →  H2O  +  M 6.1  ×  10−38/T [44]
(48) H  +  NO2  →  OH  +  NO 4.0  ×  10−10 exp(−340/T) cm3 s−1 [37]
(49) H  +  NO3  →  OH  +  NO2 5.8  ×  10−16 exp(−750/T) cm3 s−1 [40]
(50) H  +  H2O2  →  OH  +  H2O 1.69  ×  10−17 exp(−1800/T) cm3 s−1 [42]
(51) H  +  H2O2  →  HO2  +  H2 2.8  ×  10−18 exp(−1900/T) cm3 s−1 [40]
(52) H  +  HO2  →  H2  +  O2 5.6  ×  10−18 cm3 s−1 [35]
(53) H  +  HO2  →  O  +  H2O 2.4  ×  10−18 cm3 s−1 [35]
(54) H  +  HO2  →  OH  +  OH 4.2  ×  10−16 exp(−950/T) cm3 s−1 [40]
(55) H  +  HNO  →  NO  +  H2 3  ×  10−17 exp(−500/T) cm3 s−1 [37]
(56) OH  +  HO2  →  H2O  +  O2 4.8  ×  10−11 exp(250/T) cm3 s−1 [35]
(57) OH  +  O3  →  HO2  +  O2 1.9  ×  10−12 exp(−1000/T) cm3 s−1 [35]
(58) OH  +  NO2  +  M  →  HNO3  +  M 2.6  ×  10−30 (T/300)−2.9cm6 s−1 [35]
(59) OH  +  NO3  →  HO2  +  NO2 2  ×  10−17 cm3 s−1 [35]
(60) OH  +  OH  +  M  →  H2O2  +  M 6.9  ×  10−31 (T/298)−0.8 cm6 s−1 [35]
(61) OH  +  HNO3  →  NO3  +  H2O 1.5  ×  10−14 exp(650/T) cm3 s−1 [45]
(62) OH  +  H2O2  →  H2O  +  HO2 2.9  ×  10−12 exp(−160/T) cm6 s−1 [35]
(63) OH  +  H2  →  H  +  H2O 3.2  ×  10−17 exp(−2600/T) cm3 s−1 [40]

(Continued)

6
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

Table 4. (Continued )

(R#) Reaction Rate coefficient (k) Ref

(64) OH  +  OH  →  O  +  H2O 8.8  ×  10−18 exp(−503/T) cm3 s−1 [40]
(65) OH  +  HNO  →  NO  +  H2O 8  ×  10−17 exp(−500/T) cm3 s−1 [37]
(66) OH  +  HNO2  →  NO2  +  H2O 2.7  ×  10−12 exp(260/T) cm3 s−1 [35]
(67) OH  +  N2O  →  HO2  +  N2 3.69  ×  10−13 [46]
(68) OH  +  N2O  →  HNO  +  NO 3.80  ×  10−17 [47]
(69) OH  +  N2O5  →HNO3  +  NO3 2.85  ×  10−15 [37]
(70) OH  +  N2O5  →HO2  +  2NO2 2.85  ×  10−15 [37]
(71) NO  +  OH  +  M  →  HNO2  +  M 7.4  ×  10−31 (T/300)−2.4 cm6 s−1 [35]
(72) NO  +  HO2  →  NO2  +  OH 3.4  ×  10−12 exp (270/T) cm3 s−1 [35]
(73) NO  +  O3  →  NO2  +  O2 4.3  ×  10−12 exp (−1560/T) cm3 s−1 [38]
(74) NO  +  NO3  →  NO2  +  NO2 1.7  ×  10−11 cm3 s−1 [38]
(75) NO  +  NO2  +  M  →  N2O3  +  M 3.09  ×  10−46 (300/T)7.7 [42]
(76) NO  +  HO2  →  O2  +  HNO 3.3  ×  10−19 exp(−1000/T) cm3 s−1 [40]
(77) NO2  +  NO3  →  NO2  +  NO  +  O2 2.3  ×  10−13 exp(−1600/T) [38]
(78) NO2  +  NO2  +  M  →  N2O4  +  M 1.17  ×  10−45 (300/T)3.8 [42]
(79) NO2  +  NO3  +  M  →  N2O5  +  M 2.8  ×  10−42 (300/T)3.5 [35]
(80) NO3  +  HO2  →  NO2  +  OH  +  O2 4.0  ×  10−12 cm3 s−1 [35]
(81) NO3  +  HO2  →  HNO3  +  O2 9.2  ×  10−13 cm3 s−1 [42]
(82) NO3  +  NO3  →  2NO2  +  O2 5  ×  10−12 exp (−3000/T) cm3 s−1 [38]
(83) O3  +  NO2  →  NO3  +  O2 1.2  ×  10−13 exp(−2450/T) cm3 s−1 [38]
(84) O3  +  M  →  O  +  O2  +M 3.92  ×  10−16 exp(−11 400/T) cm3 s−1 [48]
(85) O3  +  HO2  →  OH  +  O2  +  O2 1.4  ×  10−20 exp(−600/T) cm3 s−1 [35]
(86) N2O3  +  M  →  NO  +  NO2  +  M 1.03  ×  10−16 exp(−2628/T) cm3 s−1 [42]
(87) N2O4  +  M  →  NO2  +  NO2  +  M 1.09  ×  10−13 exp(−4952/T) cm3 s−1 [42]
(88) N2O5  →  NO3  +  NO2 9.7  ×  1014 (T/298)0.1 exp(−11080/T) s−1 [35]
(89) HNO  +  O2  →  NO2  +  OH 1.7  ×  10−15 cm3 s−1 [32]
(90) HNO  +  O2  →  NO  +  HO2 3.3  ×  10−14 cm3 s−1 [32]
(91) HNO  +  HNO  →  N2O  +  H2O 1.4  ×  10−21 exp (−1600/T) cm3 s−1 [32]
(92) HNO2  +  HNO2  →NO  +  NO2  +  H2O 1  ×  10−26 cm3 s−1 [32]
(93) HNO2  +  HNO3  →  2 NO2  +  H2O 1.6  ×  10−23 cm3 s−1 [32]
(94) HO2  +  HO2  +M  →  H2O2  +O2  +  M 1.9  ×  10−33 exp(980/T) cm6 s−1 [35]
(95) N2O5  +  H2O  →  2HNO3 2.50  ×  10−22 cm3 s−1 [37]
(96) N2O5  +  M  →  NO2  +  NO3  +  M 1  ×  10−9 (300/T)3.5 exp(−11 000/T) cm3 s−1 [35]
a
T is the temperature in K.
Note: N*2 and O*2 correspond to the electronically excited states N2(A) and O2(1∆), respectively.

The formation of CO*2 (v) may take place through direct (2) excitation to the first excited triplet state of CO2 (3B2) via
electron collision [46, 52], or through the interaction with intersystem crossing which dissociates to the ground elec-
vibrationally excited species such as molecular nitrogen N*2 tronic states CO and ground state atomic oxygen (O3P). The
latter is more efficient due to the non-adiabatic intersystem
(v) [46]:
crossing between the ground electronic and triplet states of
CO2 + N*2 (v ) → CO*2 (v ) + N2
(R97) the CO2. Singlet oxygen is 1.48 eV higher in energy than O3P
[55], hence the energy required for dissociation is reduced
CO*2 (v) decay may take place through relaxation to CO2 or from 6.94 eV to 5.46 eV [56].
dissociation to CO and atomic oxygen [46, 52]: As can be seen in figure 5(a) (i), the only gain features
CO*2 (v ) → CO2 + hv
(R98) observed were due to CO*2 (v) and CO, suggesting that these
were the only detectable metastable and neutral products,
CO*2 (v ) → CO +O
(R99) respectively, from CO2. The loss feature due to CO2 at
2360 cm−1 has an absorbance of ca. 0.095: given that the
The dissociation of CO2 to CO and O can take place via two absorbance of this band in the absence of plasma is 0.65, this
possible direct routes depending on the level of excitation suggests ca. 15% conversion of CO2 to CO*2 (v) and CO, a sig-
[53, 54]: (1) excitation to vibrational levels of the ground excited nificant amount.
state with energies exceeding the dissociation threshold, this From figures 5(a) and (b) it can be seen that the amount of
results in ground electronic state CO and singlet oxygen, O1D; CO formed via CO2 dissociation varies with the composition

7
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

(a) 0.02 CO2* or dissociative attachment [52, 60]:

CO2* CO
CO2 + e → (CO−
(R101) 2 )* → CO +O
0.00
Thus, it does not seem unreasonable to postulate that, in the
O3
-0.02 presence of other species such as N2, water vapour or oxygen
Absorbance

(i) within the feed gas, a lower number of electrons with suffi-
-0.04 (ii)
(iii)
cient energy will be available to induce CO2 dissociation [61].
(iv) From figures 5(a) and (b) it may be seen that, in the absence
-0.06
of oxygen, the absorbance intensity of the CO*2 (v) band was
CO2 found to correlate inversely with the formation of CO. This
-0.08
suggested that additional excitation may be required to facili-
-0.10 tate the dissociation of CO*2 (v) to CO.
2500 2400 2300 2200 2100 2000 1900 1800
Wavenumber/ cm
-1 3.1.3. The effects of discharge power and gas flow rate. The
change in the plasma chemical composition was monitored as
(b) 0.015 a function of the input power and gas flow rate using ambient
air and dry artificial air as feed gases. The specific power den-
0.010 sity was employed to assess the combined effect of the input
(i) power and gas flow rate on the plasma composition:
(ii)
Absorbance

0.005 (iii) Specific power density (J cm−3) = dissipated power (W)/

(iv) 
0.000 gas flow rate (cm3 s−1) (2)
Figure 6 shows the partial pressures of NO2, O3, N2O and
-0.005 the absorbance of CO*2 (v) at 2326 cm−1 monitored in situ as
a function of power density collected after 1 min of the DBD
-0.010 being turned on. HNO3 and N2O5 produced in ambient air
2500 2400 2300 2200 2100 2000 1900 1800 were not monitored due to their interference with each other
Wavenumber/ cm
-1 and with other species, as shown in figure 4. Table 5 shows
the range of specific power densities at different input power
Figure 5. (a) In situ FTIR Spectra collected from the NTP glow and gas flow rates employed for the data presented in figure 6.
region after running the DBD for 10 min as a function of feed gas As can be seen from the table, power densities from 2.8 to
composition: (i) 0.2 dm3 min−1 1% CO2/Ar  +  0.2 dm3 min−1 O2;
4.3 J cm−3 were obtained at a constant gas flow rate of 0.5
(ii) 1% CO2/Ar; (iii) 0.2 dm3 min−1 1% CO2/Ar  +  0.2 dm3 min−1
N2 and (iv) 1% CO2/Ar  +  60% relative humidity, both (ii) and (iv) dm3 min−1 and varied input power. Higher power densities
at a flow rate of 0.5 dm3 min−1. Each spectrum was ratioed to its were observed at an input power of 36 W, increasing with
reference spectrum collected with no discharge power with the gas decreasing gas flow rate.
flowing at the specified rate. (b) The spectra in (a) replotted to show As can be seen from figure 6, the partial pressure of N2O
the gain features more clearly. The input power was 36 W.
was found to increase linearly with the power density when
ambient air was used. However, in dry artificial air the par-
of the feed gas. When the feed gas contained water vapour
tial pressure of N2O was found to increase rapidly with the
or N2 lower concentrations of CO were produced and higher
power density from 2.8 to 4.3 J cm−3, but the increase was
amount of CO*2 (v) was obtained, in contrast to that observed much less pronounced at higher power densities. In the case of
with pure CO2/Ar. Thus, it is clear from figures 5(a) and (b) ozone, similar behaviour was observed with ambient and dry
that, in the presence of molecules able to absorb vibrational artificial air; in both cases the ozone partial pressure increased
energy, the dissociation route is disfavored with respect to more rapidly with power density at lower power than at higher
vibrational relaxation to CO*2 (v). Furthermore, in the presence power densities. At all power densities, the partial pressures
of oxygen the absorbance due to CO was not observed, and of ozone were higher with the dry artificial air feed than with
the intensities of the CO2 loss features were found to be the ambient air. No steady state was observed. This behaviour
lowest of all the gases; this suggests that CO was oxidized, as may be attributed to the combined effect of relative humidity
soon as it was formed, back to CO2 or to CO*2 (v) most likely and gas temperature, as will be discussed below. The forma-
by ozone or atomic oxygen [57, 58]. tion of NO2 was found to vary with feed gas composition and
In addition to the direct dissociation of CO*2 (v), several flow rate: as can be seen from the figure, when ambient air
pathways were proposed for CO2 dissociation to CO such as was employed, two trends were observed: (I) from 2.8 to 4.3 J
electron impact ionization [59]: cm−1 and (II) from 5.4 to 10.8 J cm3, with the slope of the plot
in region (I) higher than in region (II).
CO2 + e → CO+
(R11) 2 + e + e The partial pressure of NO2 at 5.4 J cm−3 was found to be
24% lower than that obtained at 4.3 J cm−3, corresponding to a
 CO+
2 + e → CO + O (R100) reduction in flow rate from 0.5 to 0.4 dm3 min−1. Considering

8
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

the direct effect of the feed gas flow rate on the length of the increased with relative humidity and running time. However,
plasma jet, it may be concluded that a significant fraction of when using dry artificial air, the partial pressure of NO2 was
the NO2 was concentrated within the plasma region, in agree- low and remained essentially constant with running time.
ment with previous studies [17, 18]. Because the same amount After the first minute, the partial pressure of NO2 using dry air
of NO2 was obtained at power densities of 4.3 and 10.8 J cm−3, increased by ca. 50, 82 and 109  ±  2% as the relative humidity
i.e 9.44  ×  10−5 atm  ±  1%, it does not seem unreasonable was increased to 20, 40 and 60%, respectively. In humidified
to postulate that the reduction of NO2 partial pressure with artificial air the partial pressure of NO2 after 10 min was ca.
decreasing gas flow rate was compensated by increasing the 2.2 times higher than that obtained after the first minute, and
power density. However, by replacing the ambient air with this was observed at all humidity levels studied.
artificial dry air, the NO2 partial pressure declined with increas­ Taking into account the continuity of gas flow in these
­ing power density. Thus, it is clear that the presence of water experiments, it may be postulated that the gradual increase
vapour has a major effect upon the production of NO2. in NO2 concentration with time is mainly due to a change in
Unlike the case of N2O, the formation of CO*2 (v) from dielectric temperature [11, 13, 19]. As a result, the gas tem-
ambient air was found to be more sensitive to input power and perature may change due to heat transfer from the dielectric
flow rate than to specific power density. As can be seen from by convection and thermal radiation [19, 62]. The amount
figure 6, the absorbance of CO*2 (v) increased significantly of heat that can be transferred to the plasma is determined
with power density from 3.2 to 4.3 J cm−3. Within the same by several factors including: the feed gas composition, dis-
range of input power densities using ambient air, it can be charge current, the proximity to the discharge centre and the
seen that increase in the intensity of CO*2(v) was the highest of shape and configuration of the discharge electrodes [62–65].
the species shown in figure 6. But as the power density The gas temperature would be expected to be the highest near
increased from 4.3 to 10.8 J cm−3 (by reducing the flow rate the discharge electrodes and to decrease along the length of
the plasma jet [17, 60, 62, 66]. Thus, it can be assumed that the
from 0.5 to 0.2 dm3 min−1 at 36 W) the absorbance of CO*2 (v)
gas temperature profile along the jet is determined by the
decreased. By determining the CO*2 (v) absorbance at 10.8 J
temperature near the discharge electrodes. Based on this
cm−3 it was found to be ca. 65% lower than that at 4.3 J cm−3, assumption and because of the small discharge gap width of
corresponding to a 60% reduction in the flow rate from 0.5 to the DBD employed in this work, and the limited length of
0.2 dm3 min−1. Therefore, it can be concluded that the CO*2 the jet (i.e. 6.0 mm) it was found to be difficult to monitor the
(v) molecules were confined within the plasma glow region. It change in the gas temperature at different points along the
should be noted that heat generation and ionization would be jet axis. Instead, the change in gas temperature with time was
expected to be additional routes of power dissipation. monitored at a single point ca. 5.0 mm beyond the DBD nozzle
using a thermometer with its bulb immersed in the jet and cor-
3.1.4. Relative humidity and gas temperature effects. Due to related with the changes in concentration of different species
the apparently significant effect of water vapor on the chemi- observed with time. This point was chosen to ensure direct
cal composition of the NTP, it was decided to determine this monitoring of the gas temperature within the jet without
in a more measured fashion. Figure 7 shows spectra collected interfering significantly with the jet flow. Figure 9 shows the
after 1 min of NTP using: (i) dry and (ii) humidified artificial plasma gas temperature monitored 5.0 mm beyond the DBD
air at 20% RH. Spectrum (iii) in figure 7 shows spectrum (i) nozzle as a function of running time. As can be seen from
subtracted from (ii). As can be seen from the same figure, the the figure, the gas temperature increased rapidly during the
addition of water vapor to the feed gas led to a reduction in the first 5 min and less significantly at longer times. In general,
concentrations of O3, N2O5 and N2O, an increase in NO2 and this behaviour was found to agree with the work reported by
the appearance of HNO3. Akitsu et al [11] and Visser et al [19]. In addition, it can be
Due to interference between the bands of N2O5 and HNO3 seen from figure 9 that the change in gas temperature showed
and other products bands, as shown in figures 4 and 7, it was the same basic profile as the plots of NO2 partial pressure
difficult to monitor the variation of their concentrations with showed in figure 8.
relative humidity. N2O5 formation was expected to occur via Figure 10 shows the change in NO2 concentration as a func-
reaction (R79), as well as being involved in several decom- tion of relative humidity and gas temperature; the data points
position reactions in dry air such as (R38), (R88), and (R96). were extracted from figures 8 and 9. As can be seen from
However, the reduction of N2O5 observed with humidified the figure, the concentration of NO2 was found to increase
air may be attributed to the additional destructive reactions linearly with the gas temperature with the exception of dry
(R69), (R70) and (R95). In addition to reactions (R95) and air, where it varied very little. This suggests that the change
(R69), HNO3 can be produced through the reactions (R58) in NO2 with time using humid air was primarily due to the
and (R81). However, HNO3 can decompose to NO2 through increase in gas temperature. Forthermore, it can be seen from
the reactions (R43) & (R93), or to NO3 through the reaction the same figure that the effect of humidity on the formation
(R61). of NO2 was more pronounced at high temperatures. Hence,
Figure 8 shows the change in NO2 partial pressure as a it can be concluded that there is a synergic effect of humidity
function of operational time and relative humidity, measured and temperature.
using in situ FTIR spectroscopy at 36 W and 0.5 dm3 min−1. The marked difference in behaviour between humidi-
As can be seen from the figure, the partial pressure of NO2 fied and dry artificial air may be explained in terms of the

9
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

5
12
(i)
4 N2O
(ii)
10 O3
PO3 x 10-5/ atm

P N2 O x 10 / atm
3

-5
8
2
(i)
6 (ii) 1

4 0
2 3 4 5 6 7 8 9 10 11 2 3 4 5 6 7 8 9 10 11
Power density/ J cm-3 Power density/ J cm -3

1.2
10

(II) 1.0 (ii)

8 (I)

-3
P NO2 x 10 / atm

Absorbance x 10
0.8
6
-5

NO2 (i)
0.6

4 (ii)
0.4
CO2*(v)
2
0.2

0 0.0
2 3 4 5 6 7 8 9 10 11 2 3 4 5 6 7 8 9 10 11
-3 -3
Power density/ J cm Power density/ J cm

Figure 6. Plots of the partial pressures of O3, N2O, NO2 and CO* 2 (v) absorbance at 2326 cm
−1
as a function of discharge power density
and feed gas composition monitored using in situ FTIR spectroscopy 1 min after the plasma was turned on. The feed gases were: (i) dry
artificial air and (ii) ambient air. The input power and feed gas flow rates ranged from 27–36 W and 0.2–0.5 dm3 min−1, respectively.

Table 5. Input power and feed gas flow rates and their N2O5
0.003 O3
corresponding specific power densities for the data presented in N 2O 5N 2
(i) O
figure 6.
0.002 (ii) N2O
(iii) NO2 NO2
Gas flow Specific power
Input power/W rate/dm3 min−1 density/J cm−3 0.001 CO2
Absorbance

23 0.5 2.8
24 0.5 2.9 0.000
27 0.5 3.2
30 0.5 3.6 -0.001
33 0.5 4.0
36 0.5 4.3 -0.002
36 0.4 5.4 HNO3

36 0.3 7.2 -0.003

4000 3500 3000 2500 2000 1500
36 0.2 10.8
-1
Wavenumber/ cm
formation and decomposition reactions in each case. For Figure 7. In situ FTIR spectra collected after 1 min of DBD
instance, the reactions responsible for NO2 formation from operation at 36 W using: (i) dry artificial air, (ii) artificial air  +20%
dry air are: (R37), (R73), (R74), (R77), (R82), (R86)–(R88) RH and (iii) the difference between (i) and (ii). The input power
and (R96), while the direct decomposition reactions of NO2 was 36 W and gas flow rate 0.5 dm3 min−1.
are: (R23), (R24), (R35) and (R36). However, in the presence
of water vapour, additional formation reactions take place: (R48) and (R58). Thus, the number of reactions responsible
(R40), (R42), (R43), (R49), (R66), (R72), (R80), (R89), for NO2 formation increases significantly in the presence of
(R91)–(R93) and (R70) while the decomposition reactions are water vapour, in contrast to the decomposition reactions.

10
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

320
1.0
(i)
(ii)
(iii)
315
(iv)
0.8

Gas temperature/ K
310
PNO2 / P* NO2

0.6

305

0.4

300

0.2
295

0.0
0 1 2 3 4 5 6 7 8 9 10 290
0 1 2 3 4 5 6 7 8 9 10
Time/ minutes
Time/ minutes
Figure 8. NO2 partial pressure monitored in situ using FTIR
spectroscopy as a function of running time and relative humidity Figure 9. Plot of gas temperature monitored at 5.0 mm from the
using artificial air: (i) dry, (ii) 20% RH, (iii) 40% RH and (iv) 60% DBD nozzle as a function of running time using ambient air. The
RH. The input power was 36 W at 0.5 dm3 min−1. Maximum N2O input power was 36 W at 0.5 dm3 min−1.
partial pressure  =  P*NO2  =  7.7  ±  0.62  ×  10−5 atm.
1.0

Figures 11(a) and (b) show the change in ozone partial (i)
(ii)
pressure monitored in situ as a function of running time 0.8 (iii)
(a) and gas temperature (b) at different humidities. As can (iv)

be seen from the figures, the partial pressure of ozone was

PNO2 / P* NO2

0.6
found to decrease significantly with relative humidity. From
figures 11(a) and (b) it can be seen that the partial pressure of
ozone using dry air decreased by 46, 64 and 70% at humidity 0.4

levels of 20, 40 and 60%, respectively. As can be seen from

figure 11(a), the ozone partial pressure was found to decrease 0.2

steadily, by ca. 35%, during the first 10 min, regardless of the

humidity of the feed gas. However, it can be clearly seen from 0.0
figure 11(b) that ozone decomposition was more significant 304 306 308 310 312 314 316 318

at high temperatures, regardless of the relative humidity. This Temperature/ K

would be expected on the basis of the fact that the primary

Figure 10. Plots of NO2 partial pressure as a function of relative
ozone generating step (R30) is exothermic, while the majority humidity and gas temperature monitored in situ using FTIR. The
of its decomposition reactions are endothermic. For instance, feed gas was artificial air: (i) dry, (ii)  +20% RH, (iii)  +40% RH
in dry air, the main reactions responsible for ozone decompo- and (iv)  +60% RH. The input power was 36 W at 0.5 dm3 min−1.
sition include: (R33), (R21), (R84), (R73) and (R83). While Maximum NO2 partial pressure  =  P*NO2  =  7.7  ±  0.62  ×  10−5 atm.
the significant loss of ozone observed in the presence of water
vapor in the feed gas may be attributed to the formation of additional reactions (R91) and (R26) in the presence of
ozone decomposing species such as H, OH and HO2 through water vapor. However, N2O can be consumed via reac-
reactions (R45), (R57) & (R85), respectively. In addition, the tions (R27), (R13) and (R41) in dry and humidified artifi-
formation of additional quantities of NO2 will increase ozone cial air. In addition, in the presences of water vapor, N2O
consumption via reaction (R83), which may be considered as may be consumed via its reactions with OH radicals (R67)
an indirect effect of relative humidity. Another indirect effect and (R68). Furthermore, the contributions of the reactions
is the dissociation of water molecules and concomitant con- (R28) and (R24) to N2O formation may be reduced due to
sumption of energetic electrons. the consumption of atomic nitrogen by OH and HO2 radi-
Figures 12(a) and (b) show the change in N2O partial cals through reactions (R17) and (R25). Hence, as in the
pressure produced from artificial air as a function of run- case with the other species, the partial pressure of N2O at
ning time and gas temperature at different humidity levels. any time represents an equilibrium between the formation
As can be seen from the figures, the partial pressure of N2O and decomposition reactions.
increased gradually with running time and temperature to
reach a steady state after ca. 5 min, regardless of the relative 3.2. Downstream analyses
humidity. However, the partial pressure of N2O decreased
with increasing humidity. In this section, the chemical composition of the exhaust from
The formation of N2O from dry air may take place the non-thermal plasma was monitored downstream to estab-
through reactions (R28), (R24) and (R14), with the lish a comparison with that monitored in situ.

11
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

(a)
1.0 1.0

0.8
(i) 0.9
(ii)

P N 2O / P* N 2O
PO3/ P*O3

(iii)
0.6 (iv)
0.8

0.4 (i)
(ii)
0.7
(iii)
(iv)
0.2
(a)
0.6
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 11
Time/ minutes Time/ minutes

1.05

1.0
(b) (i)
(ii)
1.00
(i) (iii)
(ii) (iv)
0.8 (iii) 0.95
(iv)

P N2O/ P* N2O
P O3 / P* O3

0.6 0.90

0.85
0.4

0.80

0.2
0.75
(b)
0.0
0.70
304 306 308 310 312 314 316 318
304 306 308 310 312 314 316 318
Temperature/ K
Temperature/ K

Figure 11. The change in O3 partial pressure monitored using

Figure 12. The change in N2O partial pressure monitored using
in situ FTIR spectroscopy as a function of relative humidity in situ FTIR spectroscopy as a function of relative humidity
versus: (a) running time, (b) gas temperature using artificial air: versus: (a) running time, (b) gas temperature using artificial
(i) dry, (ii)  +20% RH, (iii)  +40% RH and (iv)  +60% RH. The air: (i) dry, (ii) 20% RH, (iii) 40% RH and (iv) 60% RH. The
input power was 36 W at 0.5 dm3 min−1. Maximum ozone partial input power was 36 W at 0.5 dm3 min−1. Maximum N2O partial
pressure  =  P*O3  =  8.9  ±  0.08  ×  10−4 atm. pressure  =  P*N2O  =  2.07  ±  0.1  ×  10−5 atm.
3.2.1. FTIR analysis. Figure 13 shows the spectra collected
from the exhaust of the non-thermal plasma cell using FTIR 0.002 (i)
spectroscopy. As can be seen from the figure, the chemical (ii)
species observed downstream using dry artificial air were O3,
0.001
N2O5 and N2O. With ambient air, HNO3 was seen in addi-
Absorbance

tion to these species. In contrast to the chemical composition

observed in situ, neither NO2 nor CO*2 (v) were seen within the 0.000
downstream exhaust. Hence, it can be concluded that these
species are mainly concentrated in the plasma glow region. -0.001
In general, these data were in good agreement with the DBD
effluent analyses carried out downstream by Eliasson et al -0.002
[67], and in the afterglow region by Sakiyama et al [18].
2500 2000 1500 1000
3.2.2. Ozone. The analyses of ozone in the downstream
-1
exhaust from the NTP was carried out using FTIR and UV– Wavenumber/cm
Vis spectroscopies to cross check these methods, using pure
Figure 13. Downstream FTIR Spectra collected after 1 min of DBD
oxygen as the feed gas. In general, the variation between both operation as a function of the feed gas composition: (i) ambient air,
analytical approaches measurements was found to lie within (ii) dry artificial air. The input power was 36 W and gas flow rate
the level of uncertainty (i.e. 1.5%) based on the extinction 0.5 dm3 min−1.

12
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

1.0
(i)
3.0 (i)
(ii)
(iii)
(ii)
(iv) (iii)
0.8
(iv)
2.5

P N2O x 10 / atm
P O3 x 10 / atm

-5
0.6
-3

2.0

0.4

1.5

0.2

1.0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

Time/ Minutes Time/ minutes

Figure 14. Plots of ozone partial pressure as a function of running Figure 15. Plots of ozone partial pressure as a function of running
time and feed gas composition. Using dry (i)  +  (ii), and humidified time and feed gas composition, monitored in situ (ii)  +  (iv) and
artificial (iii)  +  (iv). Monitored using in situ (ii)  +  (iv) and downstream (ii)  +  (iv) using FTIR spectroscopy. Using dry
downstream FTIR spectroscopy (ii)  +  (iv). The input power was (i)  +  (ii), and 40% RH humidified artificial air (iii)  +  (iv). The
36 W at 0.5 dm3 min−1. input power was 36 W at 0.5 dm3 min−1.

coefficients employed [68]. However, when the feed gas artificial air indicated a rapid decomposition of ozone during
was replaced with dry or humidified artificial air, a signifi- the first 5 min, slowing down at longer times. This behaviour
cant variation between the UV and FTIR measurements were was not observed using humidified artificial air which may be
observed. Hence, the comparison between both techniques attributed to the lower amounts of atomic oxygen and nitrogen
will be discussed in a following section. being available for ozone decomposition as a result of: (i)
Figure 14 compares the FTIR measurements of ozone lower formation due to the consumption of energetic elec-
partial pressure collected from the plasma glow and down- trons by the dissociation of water molecules, and (ii) reaction
stream regions as a function of operation time using dry and with water dissociation products such as OH and HO2 through
humidified artificial air. From the figure, it can be seen that the reactions (R19), (R25), (R31) and (R32).
highest ozone concentrations were obtained using dry air in
both glow and downstream regions. In addition, regardless of 3.2.3. N2O. Figure 15 shows the change in N2O partial pres-
the humidity of the artificial air, the ozone concentration in the sure as a function of feed gas composition and running time
downstream exhaust was found to be 30–40% lower than that monitored in situ in the plasma and downstream in the exhaust.
in plasma glow. In fact, such a significant change suggests the As can be seen from the figure, in the case of humidified air,
occurrence of chemical reactions rather than a simple diffu- the partial pressure of N2O monitored downstream was lower
sion effect. In addition, the similar magnitude of ozone reduc- by ca.18% than found in the plasma. However, with dry artifi-
tion observed in dry and humidified artificial air suggested that cial dry air, the downstream partial pressure of N2O was found
the effect of feed gas humidity on ozone is more significant in to be higher than that monitored in situ and to increase sig-
the plasma region. Based on the supposition that 1 mole of nificantly with the running time. After 10 min, the downstream
NO2 reacts with one mole of O3, from figures 10 and 11, it partial pressure was found to be more than 30% higher than
would be expected that, at the highest partial pressure of NO2 that monitored in situ. Considering the fact that N2O genera-
produced in artificial air and 40% RH, no more than 2.2% of tion from dry air would be expected to take place through the
the ozone present would be decomposed. The fact that a 40% reactions of atomic nitrogen with NO2 (R24) or NO (R28), it
reduction in the partial pressure of O3 was observed between can be concluded that some of these reactions must be tak-
the plasma glow and downstream exhaust region strongly sug- ing place in afterglow region. This further indicates the major
gests that the decomposition of ozone occurs via other paths. role of atomic nitrogen in the afterglow region. In the case
Some of the species that would be expected to play roles in of humidified artificial air, the negative effects of water mol-
ozone decomposition post the plasma region are metastable ecules on atomic nitrogen, discussed in section 3.2.2, may be
atomic oxygen and nitrogen [17, 69–71]. expected to depress the formation of N2O in the downstream
Atomic nitrogen can react directly with ozone (R21) region. On the other hand, atomic oxygen which has diffused
or with oxygen (R20) to produce NO molecules which, in from the plasma or formed in the afterglow region through
turn, consume ozone via reaction (R73) to produce NO2. ozone dissociation can facilitate N2O decomposition through
Furthermore, atomic nitrogen can react with nitric oxide (R19) reaction (R41).
to produce vibrationally excited N2 with sufficient energy to
induce ozone dissociation [72]. Atomic oxygen can decom- 3.2.4. The uncertainty associated with downstream UV analy-
pose ozone through reaction (R33) or by reacting with atomic ses of ozone. UV spectroscopy is one of the most common
nitrogen (R18) to produce nitric oxide. From figure 14, it can methods that have been used widely for ozone monitoring
be seen that the downstream measurements obtained from dry [73], and a significant number of articles have employed this
13
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

0.8 4. Conclusions

0.7
The chemical composition of non-thermal plasma (NTP) pro-
(i) duced using an air-fed dielectric barrier discharge (DBD) jet
0.6 (ii)
(iii) was monitored in the plasma glow and downstream regions
P O3 x 10 / atm

(iv) employing FTIR and UV–Vis spectroscopies. Depending

-3

0.5
on the constituents of the feed gas, N2O, O3, N2O5, NO2 and
0.4
HNO3 were observed in the NTP glow region, while N2O, O3,
N2O5 and HNO3 were detected in the downstream exhaust.
0.3 However, when the feed gas included CO2, CO and a vibra-
tionally excited form of CO2 (i.e. CO*2 (v)) were also detected.
0.2 The latter has hitherto not been observed in NTP using FTIR
spectroscopy.
0 1 2 3 4 5 6 7 8 9 10
It was postulated that the formation of CO*2 (v) from CO2
Time/ minutes
may take place through direct electron collision and/or
Figure 16. Plots of ozone partial pressure as a function of running through interaction with other excited molecules. However,
time and feed gas composition monitored downstream using FTIR the relaxation to CO2 and/or dissociation to CO and atomic
spectroscopy (i)  +  (iii), and UV–Vis spectroscopy (ii)  +  (iv). Using oxygen were assumed to be the main pathways of CO*2 (v)
dry (i)  +  (ii), and 40% RH humidified artificial air (iii)  +  (iv). The
input power was 36 W at 0.5 dm3 min−1. Plots (ii) and (iv) were
loss. In 1% CO2/Ar, the conversion of CO2 to CO*2 (v) and CO
obtained from figure 15. appears to be ca. 15%. CO*2 (v) was found to be less stable at
high discharge power densities.
approach for ozone monitoring from various types of air- When using dry air, the partial pressures of N2O and
fed ozonisers. Recently, Moiseev and co-workers [74] have O3 increased with the specific power density, while NO2
employed UV–Vis spectroscopy to monitor the change in the decreased. However, in the presence of water vapour, the
chemical composition of humid air in a container after being partial pressures of NO2, O3 and N2O increased. In addition,
exposed to NTP for 15–120 s in a large gap DBD. From their while the presence of water vapour was essential for the
work, several species were assumed to interfere with the production of HNO3, the partial pressure of NO2 increased
ozone absorbance at 254 nm namely: NO2, N2O4, N2O5, H2O2, with the relative humidity and that of O3, N2O5 and N2O
HNO2, HNO3 and HNO4, and hence they have employed three decreased.
deconvolution methods to compute these interferences to The time-dependent behaviour of the partial pressure of
evaluate the concentrations of different species. Their results NO2 in the plasma glow region was found to correlate with
suggest that the ozone concentration can be overestimated by the time dependence of the gas temperature, i.e. increasing
up to 20% in the presence of the interfering species. However, with increasing temperature. In contrast, the ozone partial
since not all of these species were observed in our work, it pressure decreased with increasing temperature, as would be
was decided to evaluate the level of uncertainty in our UV–Vis expected on the basis of thermodynamic principles. Moreover,
measurements of ozone in the outputs of a DBD fed with dry in the presence of water vapour the effect of gas temperature
and humidified artificial air. on NO2, N2O and O3 was more significant, suggesting a syner-
Figure 16 shows a comparison between ozone partial pres- getic effect of relative humidity and gas temperature.
sure measurements carried out using FTIR and UV–Vis spec- Unexpectedly, the spectroscopic analyses of O3, NO2 and
troscopies in the downstream region as a function of running N2O in the plasma glow region and in the downstream exhaust
time using dry and humidified artificial air. As can be seen suggested the occurrence of chemical reactions in afterglow
from the figure, ozone measurements carried out using the UV region rather than simple diffusion. For instance, the higher
spectrometer were found to be higher than those obtained using partial pressure of N2O observed in the downstream exhaust
FTIR spectroscopy. Furthermore, the difference between the than in the plasma glow region when using dry air, suggested
UV and FTIR measurements was found to increase with time that active species such as atomic nitrogen play important
from 19% to 40% and from 33% to 42% when dry and humid- roles in the afterglow region.
ified artificial air were employed, respectively. The increase in Finally, it is clear from our research that the UV analyses of
the difference between the UV and FTIR measurements with ozone cannot employed with any accuracy in the presence of
time can be attributed to the increase in the concentrations of interfering species such as N2O5 and/ or HNO3.
N2O5 and HNO3 that can interfere with UV measurements of
ozone in dry and humidified artificial air, respectively. Thus,
the overestimation of ozone concentration obtained using UV Acknowledgments
is expected to increase with the number and concentration
of interfering species. Therefore, the determination of theses The authors would like to thank Newcastle University and
interfering species and the magnitude of their interference King Abdulaziz City for Science and Technology (KACST)
is important to maintain a reliable level of accuracy when for the financial and technical support received during this
employing UV spectroscopic analyses of ozone. work.

14
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

References microdischarge in humid air and dynamics of reactive

neutral species J. Phys. D: Appl. Phys. 45 425201
[19] Visser R J 1989 Determination of the power and current
[1] Borcia G, Anderson C A and Brown N M D 2003 Dielectric densities in argon and oxygen plasmas by in situ
barrier discharge for surface treatment: application to temperature measurements J. Vac. Sci. Technol. A 7 189–94
selected polymers in film and fibre form Plasma Sources [20] Chen N H and Othmer D F 1962 New generalized equation for
Sci. Technol. 12 335 gas diffusion coefficient J. Chem. Eng. Data 7 37–41
[2] Borcia G and Brown N M D 2007 Hydrophobic coatings on [21] Koinuma H, Ohkubo H, Hashimoto T, Inomata K, Shiraishi T,
selected polymers in an atmospheric pressure dielectric Miyanaga A and Hayashi S 1992 Development and
barrier discharge J. Phys. D: Appl. Phys. 40 1927 application of a microbeam plasma generator Appl. Phys.
[3] Kostov K G, dos Santos A L R, Honda R Y, Nascente P A P, Lett. 60 816–7
Kayama M E, Algatti M A and Mota R P 2010 Treatment [22] Perner D, Schmeltekopf A, Winkler R H, Johnston H S,
of PET and PU polymers by atmospheric pressure plasma Calvert J G, Cantrell C A and Stockwell W R 1985 A
generated in dielectric barrier discharge in air Surf. Coat. laboratory and field study of the equilibrium N2O5 yields
Technol. 204 3064–8 reversibly NO3 + NO2 J. Geophys. Res. 20 3807–12
[4] Chen Z and Mathur V K 2002 Nonthermal plasma for gaseous [23] Hearn A G 1961 The absorption of ozone in the ultra-violet
pollution control Ind. Eng. Chem. Res. 41 2082–9 and visible regions of the spectrum Proc. Phys. Soc. 78 932
[5] Gentile A C and Kushner M J 1995 Reaction chemistry and [24] Tuazon E C, Sanhueza E, Atkinson R, Carter W P L,
optimization of plasma remediation of NxOy from gas Winer A M and Pitts J N Jr 1984 Direct determination
streams J. Appl. Phys. 78 2074–85 of the equilibrium constant at 298 K for the nitrogen
[6] Liang P, Jiang W, Zhang L, Wu J, Zhang J and Yang D dioxide  +  nitrogen trioxide (NO3). dblarw. nitrogen
2015 Experimental studies of removing typical VOCs by pentoxide (N2O5) reactions J. Phys. Chem. 88 3095–8
dielectric barrier discharge reactor of different sizes Process [25] McCaa D J and Shaw J H 1968 The infrared spectrum of
Safety Environ. Protect. 94 380–4 ozone J. Mol. Spectrosc. 25 374–97
[7] Eliasson B, Kogelschatz U, Xue B and Zhou L-M 1998 [26] Nightingale R E, Downie A R, Rotenberg D L, Crawford B Jr
Hydrogenation of carbon dioxide to methanol with and Ogg R A Jr 1954 The preparation and infrared spectra
a discharge-activated catalyst Ind. Eng. Chem. Res. of the oxides of nitrogen J. Phys. Chem. 58 1047–50
37 3350–7 [27] Herzberg G 1945 Molecular spectra and molecular structure
[8] Kostov K G, Rocha V, Koga-Ito C Y, Matos B M, Algatti M A, Molecular Spectra and Molecular Structure (Infrared
Honda R Y, Kayama M E and Mota R P 2010 Bacterial and Raman Spectra of Polyatomic Molecules vol 2) ed G
sterilization by a dielectric barrier discharge (DBD) in air Herzberg (New York: Van Nostrand, Reinhold) p 1
Surf. Coat. Technol. 204 2954–9 [28] Haridass C, Aw-Musse A, Misra P and Jordan J 2000 Fourier
[9] Colagar A H, Sohbatzadeh F, Mirzanejhad S and Omran A V Transform infrared (FT-IR) spectroscopy of trace molecular
2010 Sterilization of Streptococcus pyogenes by afterglow species of importance for the elucidation of atmospheric
dielectric barrier discharge using O2 and CO2 working gases phenomena Comput. Electr. Eng. 26 47–65
Biochem. Eng. J. 51 189–93 [29] Lowke J J and Morrow R 1995 Theoretical analysis of removal
[10] Pavlovich M J, Chen Z, Sakiyama Y, Clark D S and of oxides of sulphur and nitrogen in pulsed operation
Graves D B 2013 Effect of discharge parameters and of electrostatic precipitators IEEE Trans. Plasma Sci.
surface characteristics on ambient-gas plasma disinfection 23 661–71
Plasma Process. Polym. 10 69–76 [30] Cravens T E, Kozyra J U, Nagy A F, Gombosi T I and
[11] Akitsu T, Ohkawa H, Tsuji M, Kimura H and Kogoma M 2005 Kurtz M 1987 Electron impact ionization in the vicinity of
Plasma sterilization using glow discharge at atmospheric comets J. Geophys. Res.: Space Phys. 92 7341–53
pressure Surf. Coat. Technol. 193 29–34 [31] Piper L G 1987 Quenching rate coefficients for N2(a’ 1Σ− u)
[12] Fridman G, Peddinghaus M, Balasubramanian M, Ayan H, J. Chem. Phys. 87 1625–9
Fridman A, Gutsol A and Brooks A 2006 Blood [32] Mätzing H 2007 Advances in Chemical Physics (New York:
Coagulation and living tissue sterilization by floating- Wiley) pp 315–402
electrode dielectric barrier discharge in air Plasma Chem. [33] Iannuzi M P, Jeffries J B and Kaufman F 1982 Product
Plasma Process. 26 425–42 channels of the N2 (A3Σ  +  u)  +  O2 interaction Chem. Phys.
[13] Sladek R E J, Stoffels E, Walraven R, Tielbeek P J A and Lett. 87 570–4
Koolhoven R A 2004 Plasma treatment of dental cavities: [34] Atkinson R, Baulch D L, Cox R A, Hampson R F,
a feasibility study IEEE Trans. Plasma Sci. 32 1540–3 Kerr J A, Rossi M J and Troe J 1997 Evaluated kinetic,
[14] Pipa A V and Ropcke J 2009 Analysis of the mid-infrared photochemical and heterogeneous data for atmospheric
spectrum of the exhaust gas from an atmospheric pressure chemistry: supplement V. IUPAC subcommittee on gas
plasma jet (APPJ) working with an argon-air mixture kinetic data evaluation for atmospheric chemistry J. Phys.
IEEE Trans. Plasma Sci. 37 1000–3 Chem. Ref Data 26 521–1011
[15] Schmidt-Bleker A, Winter J, Iseni S, Dunnbier M, [35] Atkinson R, Baulch D L, Cox R A, Hampson R F, Kerr J A,
Weltmann K D and Reuter S 2014 Reactive species output Rossi M J and Troe J 1997 Evaluated kinetic and
of a plasma jet with a shielding gas device-combination of photochemical data for atmospheric chemistry: supplement
FTIR absorption spectroscopy and gas phase modelling VI. IUPAC subcommittee on gas kinetic data evaluation
J. Phys. D: Appl. Phys. 47 145201 for atmospheric chemistry J. Phys. Chem. Ref. Data
[16] Reuter S, Winter J, Iseni S, Peters S, Schmidt-Bleker A, 26 1329–499
Dünnbier M, Schäfer J, Foest R and Weltmann K D 2012 [36] Atkinson R, Baulch D L, Cox R A, Hampson R F, Kerr J A
Detection of ozone in a MHz argon plasma bullet jet and Troe J 1989 Evaluated kinetic and photochemical
Plasma Sources Sci. Technol. 21 034015 data for atmospheric chemistry: supplement III.
[17] Van Gaens W and Bogaerts A 2014 Kinetic modelling for an IUPAC subcommittee on gas kinetic data evaluation
atmospheric pressure argon plasma jet in humid air (vol 46, for atmospheric chemistry J. Phys. Chem. Ref. Data
275201, 2013) J. Phys. D: Appl. Phys. 47 079502 18 881–1097
[18] Sakiyama Y, Graves D B, Chang H-W, Shimizu T and [37] Herron J and Green D 2001 Chemical kinetics database
Morfill G E 2012 Plasma chemistry model of surface and predictive schemes for nonthermal humid air plasma

15
Plasma Sources Sci. Technol. 24 (2015) 065006 A Al-Abduly and P Christensen

chemistry. Part II. Neutral species reactions Plasma Chem. [59] Wilde K A, Zwolinski B J and Parlin R B 1957 Decomposition
Plasma Process. 21 459–81 of carbon dioxide in a radiofrequency arc J. Am. Chem. Soc.
[38] Kossyi I A, Kostinsky A Y, Matveyev A A and Silakov V P 79 1323–9
1992 Kinetic scheme of the non-equilibrium discharge in [60] Schulz G J 1962 Cross sections and electron affinity for O−
nitrogen-oxygen mixtures Plasma Sources Sci. Technol. ions from O2, CO, and CO2 by electron impact Phys. Rev.
1 207–20 128 178–86
[39] Eichwald O, Yousfi M, Hennad A and Benabdessadok M D [61] Savinov S Y, Lee H, Song H K and Na B-K 1999
1997 Coupling of chemical kinetics, gas dynamics, and Decomposition of methane and carbon dioxide in
charged particle kinetics models for the analysis of NO a radio-frequency discharge Ind. Eng. Chem. Res.
reduction from flue gases J. Appl. Phys. 82 4781–94 38 2540–7
[40] Capitelli M, Ferreira C M, Gordiets B F and Osipov A I 2000 [62] Stepaniuk V P, Ioppolo T, Otugen M V and Sheverev V A
Plasma Kinetics in Atmospheric Gases (Berlin: Springer) 2007 Measurement of gas temperature and convection
[41] Woodall J, Agúndez M, Markwick-Kemper A J and Millar T J velocity profiles in a dc atmospheric glow discharge
2007 The UMIST database for astrochemistry 2006 Astron. J.Appl. Phys. 102 123302–5
Astrophys. 466 1197–204 [63] Kiyoto I, Nobuyuki A and Hideomi K 1994 Production of
[42] National Institute of Standards and Technology 2011 NIST fullerenes by low temperature plasma chemical vaper
Chemical Kinetics Database (http://kinetics.nist.gov) deposition under atmospheric pressure Japan. J. Appl. Phys.
[43] Baulch D L et al 1994 Evaluated kinetic data for combustion 33 L197
modeling. supplement I J. Phys. Chem. Ref. Data 23 847–8 [64] Guohua N, Peng Z, Cheng C, Ye S, Hirotaka T and Yuedong M
[44] Kushner M J 1993 Plasma chemistry of He/O2/SiH4 and 2012 Characterization of a steam plasma jet at atmospheric
He/N2O/SiH4 mixtures for remote plasma-activated pressure Plasma Sources Sci. Technol. 21 015009
chemical-vapor deposition of silicon dioxide J. Appl. Phys. [65] Colucci D W and Viskanta R 1996 Effect of nozzle geometry
74 6538–53 on local convective heat transfer to a confined impinging air
[45] Mukkavilli S, Lee C K, Varghese K and Tavlarides L L 1988 jet Exp. Therm. Fluid Sci. 13 71–80
Modeling of the electrostatic corona discharge reactor IEEE [66] Ni T L, Ding F, Zhu X D, Wen X H and Zhou H Y 2008
Trans. Plasma Sci. 16 652–60 Cold microplasma plume produced by a compact and
[46] Gordillo-Vázquez F J 2008 Air plasma kinetics under the flexible generator at atmospheric pressure Appl. Phys. Lett.
influence of sprites J. Phys. D: Appl. Phys. 41 234016 92 241503
[47] Person J C and Ham D O 1988 Removal of SO2 and NOx from [67] Eliasson B and Kogelschatz U 1987 Nitrogen oxide formation
stack gases by electron beam irradiation Int. J. Radiat. Appl. in ozonizers Proc. 8th Int. Symp. on Plasma Chemistry
Instrum. C 31 1–8 (Tokyo, 1987) pp 736–41
[48] Jeong J Y, Park J, Henins I, Babayan S E, Tu V J, Selwyn G S, [68] Dufour G, Valentin A, Henry A, Hurtmans D and
Ding G and Hicks R F 2000 Reaction chemistry in the Camy-Peyret C 2004 Concentration measurements of
afterglow of an oxygen–helium, atmospheric-pressure ozone in the 1200–300 ppbv range: an intercomparison
plasma J. Phys. Chem. A 104 8027–32 between the BNM ultraviolet standard and infrared methods
[49] Freund H J and Roberts M W 1996 Surface chemistry of Spectrochim. Acta A 60 3345–52
carbon dioxide Surf. Sci. Rep. 25 225–73 [69] Mozetič M, Vesel A, Gaillard M and Ricard A 2001 Atomic
[50] Modest M F and Bharadwaj S P 2002 Medium resolution oxygen concentration in a flowing post-discharge reactor
transmission measurements of CO2 at high temperature Plasmas Polym. 6 71–9
J. Quant. Spectrosc. Radiat. Transfer 73 329–38 [70] Babayan S E, Ding G, Nowling G R, Yang X and
[51] Toselli B M and Barker J R 1991 Excitation of CO2 by Hicks R F 2002 Characterization of the active
energy transfer from highly vibrationally excited benzene species in the afterglow of a nitrogen and helium
derivatives J. Chem. Phys. 95 8108–19 atmospheric-pressure plasma Plasma Chem. Plasma
[52] Rond C, Bultel A, Boubert P and Chéron B G 2008 Process. 22 255–69
Spectroscopic measurements of nonequilibrium CO2 [71] Kang W S, Kim H-S and Hong S H 2010 Atomic oxygen
plasma in RF torch Chem. Phys. 354 16–26 generation by in situ plasma and post-plasma in dielectric
[53] Richard J 2011 49th AIAA Aerospace Sciences Meeting barrier discharges for surface treatment Thin Solid Films
including the New Horizons Forum and Aerospace 518 6578–82
Exposition (Orlando, Florida, 4 January 2011) (American [72] Phillips L F and Schiff H I 1962 Mass spectrometric
Institute of Aeronautics and Astronautics) pp 2011–447 studies of atom reactions. II. Vibrationally excited N2
[54] Fridman A 2012 Plasma Chemistry (Cambridge: Cambridge formed by the reaction of N atoms with NO J. Chem. Phys.
University Press) p 263 36 3283–6
[55] Moore C E 1949 Atomic Energy Levels (NBS Circular vol I) [73] Rakness K, Gordon G, Langlais B, Masschelein W,
(Washington, DC: Department of Commerce, National Matsumoto N, Richard Y, Robson C M and Somiya I 1996
Bureau of Standards) p 476 Guideline for measurement of ozone concentration in the
[56] Chase M 1998 NIST-JANAF Thermochemical Tables 4th edn process gas from an ozone generator Ozone: Sci. Eng.
(New York: American Institute of Physics) p 1790 18 209–29
[57] Arin L M and Warneck P 1972 Reaction of ozone with carbon [74] Moiseev T, Misra N N, Patil S, Cullen P J, Bourke P,
monoxide J. Phys. Chem. 76 1514–6 Keener K M and Mosnier J P 2014 Post-discharge gas
[58] Toby S, Sheth S and Toby F S 1984 Reaction of carbon composition of a large-gap DBD in humid air by UV–Vis
monoxide with ozone and oxygen atoms Int. J. Chem. absorption spectroscopy Plasma Sources Sci. Technol.
Kinetics 16 149–57 23 065033

16