See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/256196329

Water dissociation in a radio-frequency electromagnetic field with ex situ

electrodes - Process characterization

Article  in  Plasma Sources Science and Technology · February 2013

DOI: 10.1088/0963-0252/22/1/015010

CITATIONS READS

6 564

5 authors, including:

Jens Schneider Frank Holzer

Deutsches Biomasseforschungszentrum Helmholtz-Zentrum für Umweltforschung
12 PUBLICATIONS   194 CITATIONS    50 PUBLICATIONS   926 CITATIONS   

SEE PROFILE SEE PROFILE

Frank-Dieter Kopinke
Helmholtz-Zentrum für Umweltforschung
117 PUBLICATIONS   2,557 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Guidelines for in situ thermal treatment (ISTT) of contaminated soil and groundwater View project

CARBOWERT View project

All content following this page was uploaded by Frank Holzer on 10 June 2014.

The user has requested enhancement of the downloaded file.

IOP PUBLISHING PLASMA SOURCES SCIENCE AND TECHNOLOGY
Plasma Sources Sci. Technol. 22 (2013) 015010 (8pp) doi:10.1088/0963-0252/22/1/015010

Water dissociation in a radio-frequency

electromagnetic field with ex situ
electrodes—process characterization
Jens Schneider, Frank Holzer, Markus Kraus, Frank-Dieter Kopinke and
Ulf Roland
Helmholtz Centre for Environmental Research–UFZ, Department of Environmental Engineering,
Permoserstr. 15, 04318 Leipzig, Germany
E-mail: [email protected] and [email protected]

Received 9 August 2012, in final form 22 October 2012

Published 13 December 2012
Online at stacks.iop.org/PSST/22/015010

Abstract
A new type of water dissociation at ambient pressure initiated by the irradiation of aqueous
electrolytes using an electromagnetic field with a frequency of 13.56 MHz is described in this
study. A special reactor design allows the use of ex situ electrodes to form in situ electrical
discharges in water vapour bubbles. The observed formation of molecular hydrogen (H2 ) and
oxygen (O2 ) combined with the emission of light (‘burning water’ phenomenon) originates
from a non-thermal plasma in water vapour bubbles. The influences of type of electrolyte, its
concentration, pH value and external RF voltage on the gas formation rate as well as on the gas
composition are presented.
(Some figures may appear in colour only in the online journal)

1. Introduction The idea of using external electrodes for water dissociation

is rather interesting since the known problems of electrode
The use of radio-frequency (RF) electromagnetic fields in order corrosion could be overcome with such an approach.
to dissociate water has been recently described by several Furthermore, discharges in pure water vapour avoid energy
research groups [1–5]. It was shown that exposing distilled losses known from discharges in non-inert gases (e.g. air) and
water or solutions containing an electrolyte such as dissolved the formation of undesirable products (e.g. NOx in the case of
NaCl to an RF field can lead to the generation of H2 , O2 and air discharges).
H2 O2 . Thereby, the use of rods with internal gas flow [1, 2], This study in some respects represents a combination of
rod and spherical plate electrodes with dielectric barrier [3] as external electrodes with the capillary reactor design [9, 10] to
well as mesh and plate electrodes within or in direct contact realize a new kind of water scission. This concept allows
with the solution [4] are described as employed experimental us to create a gas phase consisting of pure water vapour
setups. However, even more widely used for the initiation in situ representing an alternative to discharge formation in
of discharges in water are pulsed direct- (dc) and alternating- water-saturated (inert) gases or bubbles formed by external gas
current (ac) voltages mostly in combination with needle or dosage [3, 11]. The solution surrounding the bubbles formed
rod electrodes. The review papers [6, 7] give a wide overview in situ practically acts as electrode assembly. The current
of experimental setups for electrical discharges in water and densities in the film between bubble and glass tube are strongly
above the water surface. increased in accordance to the capillary reactor design [9, 10]
A new concept is the application of specifically designed which is a prerequisite for the formation of discharges.
antennas placed outside of the solution which produce a This paper presents a novel glass reactor design allowing
focused RF field with a frequency of 13.56 MHz [8]. In this us to focus RF energy to a restricted volume of an aqueous
case, H2 and O2 are produced on the surface of the solution. electrolyte, enabling selective heating, in situ formation of
The basic mechanism of this new phenomenon (‘burning water vapour bubbles and initiation of electrical discharges in
water’) cannot yet be clarified. these bubbles leading to H2 and O2 generation. A water column

0963-0252/13/015010+08$33.00 1 © 2013 IOP Publishing Ltd Printed in the UK & the USA
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

Figure 1. Reactor type I (left) and reactor type II (right) made of

duran and quartz, respectively, with emphasis on the restriction.
Figure 2. Scheme of the experimental setup emphasizing the
geometry of the restriction.
remaining above the discharge zone acts as transfer medium
for the unconsumed oxyhydrogen gas. This concept uses a (MS) (Shimadzu GC-17A/MS-QP5000; column Agilent DB-
plate electrode system outside of a glass reactor. Thus, the 1, 60 m × 0.32 mm × 5 µm) was used to identify additional
electrodes are not in direct contact with the aqueous solution. products in the gas and liquid phases.
For temperature measurements in the RF field, two
2. Methods different methods were used. Firstly, fibre optical temperature
probes (Nortech or Neoptix, Quebec) directly quantified
2.1. Basic setup the temperature at different positions within the electrolyte
solution. Secondly, an infrared (IR) thermo-camera
A water-cooled RF generator (PFG 1000 RF, Hüttinger (VarioTHERM, Infratech, Dresden) allowed measuring either
Elektronik, Freiburg/Brsg.) working at a fixed frequency of the outside surface temperature of the reactor wall or, using a
13.56 MHz and having a maximum power of 1 kW was used special ‘through-glass filter’, the temperature of the electrolyte
in combination with an impedance adjustment network (PFM solution in a defined region.
3000A, also from Hüttinger). This so-called matchbox con-
tains adjustable vacuum capacitors and coils to reduce energy
reflection from the load to the generator. The electrode sys- 2.3. Chemicals
tem consisted of two parallel copper plates. One electrode was Electrolyte solutions were prepared by adding the respective
connected to the RF matchbox output supplying a maximum chemicals with purities of more than 99% provided by Merck
voltage of 2 kV while the other was connected to the ground to deionized water with a residual specific conductivity of
and the Faraday cage utilized for electromagnetic shielding. <3×10−4 S m−1 . All conductivities were measured at 22.5 ◦ C.
Two glass reactors, both with an inner diameter of 12 mm
(outside of the restriction) and a wall thickness of 1.5 mm,
3. Results and discussion
one consisting of duran (type I, figure 1, left) and the other
consisting of quartz (type II, figure 1, right), were employed. 3.1. Phenomenological description
The lower cylindrical and U-shaped parts of the two reactors
shown in figure 1 were placed between the parallel plate When an electrolyte solution in the glass reactor was exposed
electrodes while at least one tube end left the shielding cage. to a sufficiently strong RF electromagnetic field, generally the
The reduced inner diameter in the restriction of the glass tubes following three phases could be observed. First, selective
was about 3 mm (magnification, figure 2). heating of the solution within the restriction, caused by the
On top of the reactor tube leaving the Faraday cage a water occurring high current densities, was observed. Subsequently,
cooler was installed followed by a glass vial for gas sampling. when local temperatures within the heated volume of the
Figure 2 shows a scheme of the whole experimental setup. restriction reached the boiling point of the solution, water
vapour bubbles were formed periodically with increasing
2.2. Analytical methods frequency. Finally, after establishing a quasi-stationary
vapour bubble positioned in the tube restriction, a continuous
A gas chromatograph (GC) coupled with a thermal- discharge accompanied by gas formation of molecular
conductivity detector (Agilent GC System 6850A; column hydrogen and oxygen and emission of light occurred. A
HP-PLOT, 30 m × 0.32 mm × 12 µm) was used to identify photograph of such a discharge obtained when using a NaCl
and quantify H2 , O2 , N2 , CO and CH4 concentrations in the solution with a concentration of 5 weight per cent (wt%) is
product gas. Another GC coupled with a mass spectrometer presented in figure 3.

2
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

Figure 3. Discharge in the restriction of the reactor using a 5 wt%

NaCl solution as electrolyte.

Figure 5. Temperatures at different positions in and close to the

plasma zone (conditions as specified in figure 4).

probes was limited to temperatures lower than 270 ◦ C such

that the true temperature could not be directly measured. The
distinct local temperature differences indicate the presence of
hot spots in the plasma zone.
After establishing a quasi-stationary water vapour bubble,
the two temperature sensors in the restriction were permanently
situated in the gas phase. The sensor in the lower part of the
restriction seemed to be in contact with discharges at the gas–
liquid interface explaining the high temperatures >250 ◦ C.
At the same time, the second sensor in the upper part of
the restriction did not show temperatures >100 ◦ C though it
shared the same gas phase as the probe no 3. Due to the
limited temporal and spatial resolution of the IR camera and
the fibre optical sensors, the measured temperatures have to
be considered as mean or effective values. Actual values can
Figure 4. IR image of the glass reactor (restriction region) during a even be higher. In spite of these locally high temperatures it is
steady-state plasma discharge (1.9 kV, 600 W RF power, 5 wt% postulated that the formation of hydrogen and oxygen as well
NaCl electrolyte). Positions of the fibre optical probes are marked as the emission of light originates from plasma which is in its
as white dots: 1—above restriction, 2—upper part of restriction, nature non-thermal but not necessarily ‘cold’ plasma. Electron
3—lower part of restriction and 4—below restriction. impact processes governing this plasma were at the centre of
a study that will be published in a forthcoming paper [12].
3.2. Selective heating Although relaxation processes connected with ‘thermalization’
lead to measured temperatures of at least about 300 ◦ C in the
The IR camera was used to demonstrate the selective heating discharge, the identified large electron energies above 10 eV
of the electrolyte solution in the restriction of the reactor tube. are a clear indication for the non-thermal nature of the plasma.
Figure 4 shows an example for such an IR image. These
measurements confirmed the occurrence of bulk temperatures
3.3. Vaporization of water
of more than 240 ◦ C when discharges were running. The IR
camera produced five pictures per second showing that this Two experiments applying the reactor type II were used to
high-temperature region occurred at different positions within study the frequency of bubble formation with 5 wt% NaCl
the restriction indicating local fluctuation of the plasma. In solutions that contained no surfactant or 1.8 g l−1 of sodium
addition to IR observations, four fibre optical probes were dodecyl sulfate (SDS) at RF voltages of 1.3 and 1.9 kV (300
placed in the glass reactor as indicated by white dots in figure 4. and 600 W RF power, respectively). The surfactant SDS was
Figure 5 gives an example of typical temperature evolution used to analyse if the surface tension of the solution has an
within the solution at these four measuring positions in influence on bubble formation and gas formation rates. An
the restriction range of the reactor. Three of the four optical camera recording 30 pictures a second was applied to
probes detected temperatures approaching 100 ◦ C. One sensor observe bubble formation in the restriction of the reactor and
(lower part of the restriction) detected temperatures >250 ◦ C. to calculate its frequency. RF power was applied to solutions
Unfortunately, the measuring range of the applied fibre optical at a temperature of 25 ◦ C before each measurement. Time

3
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

Figure 6. Frequency of bubble formation at RF voltages of 1.3 and

1.9 kV for 5 wt% NaCl solutions with (open symbols  , ◦) and
without (filled symbols , •) 1.8 g l−1 SDS. Straight lines represent Figure 7. Gas formation rates and H2 /O2 ratios for several
linear fits. electrolytes used under similar experimental conditions
(conductivity of the solution: 5.8 S m−1 , RF voltage: 1.3 kV, reactor
type I).
recording started after the first bubble was observed. The
frequencies of bubble formation are shown in figure 6.
As expected, at 1.9 kV, bubbles in the restriction appeared observable gas formation rate was chosen as characteristic
faster after RF power activation and with a higher frequency criterion representing the intensity of the discharge. The
than at 1.3 kV. The frequency increased with progressing time influence of parameters such as electrolyte type, concentration
after switching on RF power as the mean temperature of the and pH value on gas formation rates and composition are
solution in the vicinity of the restriction increased. presented below. In this study where ex situ electrodes are
The addition of the surfactant SDS to the solution led to applied, high electrolyte concentrations are essential to create
a higher starting frequency of bubble formation, as shown discharges within the voltage range of the matchbox (<2 kV).
in figure 6. Due to reduction in the surface tension of the A sufficiently high voltage drop over the water vapour bubble
electrolyte solution, formation of phase interfaces is facilitated. necessary for discharge initiation could only be realized if the
The increase in frequency with time was also enhanced by overall resistivity of the electrolyte solution was low. This
addition of SDS. condition seems to be valid for discharges in capillaries where
Electrons of energies lower than 100 eV which are likely the solution itself acts as electrodes [9, 10]. In contrast, rather
present in the gas phase when discharges are initiated do not low conductivities are sufficient for discharge initiation over a
penetrate the liquid phase deeper than a few nanometres [13]. water surface or with gas dosing [6, 7, 11].
As surfactants are enriched in the gas–liquid interface one
would expect a decomposition of SDS by electron impact or 3.4.1. Type of electrolyte. Electrolyte solutions with a
thermal processes. And indeed, CO, CO2 and CH4 as products (low-frequency) specific conductivity of 5.8 S m−1 each were
of decomposition of SDS were detected in the product gas in prepared using one of the following chemicals: Na2 SO4 ,
the range of 1 vol% and 0.1 vol% for carbon oxides and CH4 , K2 SO4 , KCl, NaCl, Na2 CO3 , Mg(ClO4 )2 , KNO3 , NaNO3 ,
respectively. But due to the fact that in comparison with NaCl HCl, HNO3 , NaOH and KOH. To ensure comparability
solutions without SDS overall gas composition and formation between the results, the electrical conductivity instead of
rate (figure 7) did not change significantly, one can assume molar or mass concentrations was chosen as the parameter
that the gas–liquid interface does not play a significant role for that did not vary between the solutions. This is justified since
the hydrogen and oxygen formation process. In other words, firstly, ohmic losses are the dominant heating mechanism in
electron impact processes finally leading to the formation of concentrated electrolyte solutions at 13.56 MHz and secondly,
hydrogen and oxygen most likely take place in the bulk phase the conductivity determines the maximal field strength in the
of the water vapour bubble. There, the water partial pressure, restriction that is responsible for discharge initiation. Each
i.e. the concentration of water molecules, is not likely to be solution was filled in the reactor type I that was placed between
influenced by the presence of the surfactant. the electrodes and exposed to the RF field using an RF voltage
of 1.3 kV. The gas formation rates as well as the H2 /O2 ratios
3.4. Influences on gas formation and gas composition of the product gas are displayed in figure 7.
Note that electrolytes such as NaNO3 and KNO3 showed
One of the key features of the described plasma phenomenon neither discharges nor gas formation. Electrolytes such as
is the formation of a hydrogen–oxygen gas leaving the NaOH, KOH produced different kinds of discharges with
glass tube. Discharges were always accompanied by gas negligible gas formation. Due to very high pH values (>13)
formation in the restriction region of the glass reactor. The the quartz glass reactors were rapidly damaged in the plasma

4
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

zone. Other than for strong bases, nitrates seem to inhibit

plasma chemistry probably due to the formation of NOx gases.
Surprisingly, nitric acid, though also containing nitrate anions,
did not quench the plasma discharges completely.
The product gas was typically composed of molecular H2
and O2 in a molar ratio of 2 : 1. This ratio reflects exactly the
stoichiometry of water and, therefore, excludes the formation
of significant amounts of other long-living products containing
hydrogen or oxygen. As can be seen in figure 7, discharges in
electrolyte solutions exhibiting an H2 /O2 ratio of 2.0 showed
comparable gas formation rates between 6 and 8 ml min−1 .
For these electrolytes, given the term ‘stable’ electrolytes,
there seems to be no significant influence of the electrolyte
composition on the gas formation.
Discharges in electrolytes with a different H2 /O2 ratio,
namely Mg(ClO4 )2 , HCl and HNO3 , exposed lower gas
formation rates between 3 and 5 ml min−1 . An H2 /O2 ratio
deviating from 2.0 indicates the participation of the electrolyte Figure 8. Gas formation rates for three electrolytes in the reactor
or its reaction products in the plasma chemical reactions taking type II (1.7 kV, 500 W RF power).
place in the gas phase and the gas–liquid interface. These
electrolytes are termed as ‘reactive’ in the progress of this These observations prove the decomposition of the reactive
paper in contrast to stable electrolytes producing an H2 /O2 ammonium salts.
ratio of 2.0. Increasing the concentration of NH4 Cl to 10.5 wt%
The electrolytes HCl and HNO3 produced H2 /O2 ratios (conductivity >20 S m−1 ) enhanced further the rates of
only slightly deviating from 2.0. HCl has a significant partial formation of both H2 and N2 (figure 8). This behaviour is
pressure in the vapour phase equivalent to a concentration different from that of stable electrolytes whose gas formation
of a few 100 ppm at 100 ◦ C dependent on its concentration rates remained constant above a certain conductivity level.
[14]. Moreover, HNO3 exhibits partial pressures like HCl One can conclude that reactive electrolytes act as an educt
[15] and is subject to thermal or light-induced decomposition for plasma-chemical reactions whereas from stable electrolytes
leading primarily to the formation of NO2 , H2 O and O2 [16]. only water is decomposed.
Interactions, such as scavenging of electrons and/or scavenging
of oxygen and hydrogen species in the discharge zone by 3.4.2. Electrolyte concentration and RF voltage. Gas
electrolyte molecules and their conversion products, seem to be formation could be observed for a wide range of salt
responsible for lower gas formation rates and different H2 /O2 concentrations but did not occur when deionized water was
ratios. used. Figure 9 presents gas formation rates in dependence on
The change in the H2 /O2 ratio was even more pronounced NaCl concentration and RF voltage as measured with reactor
for the electrolyte Mg(ClO4 )2 as a much higher O2 type II. In this study, a minimum NaCl concentration of about
concentration was observed in its product gas. Although it 0.5 wt% was found to be necessary to initiate discharges. The
is known that Mg(ClO4 )2 starts to decompose between 150 gas formation rate increased almost linearly with increasing
and 200 ◦ C yielding additional O2 [17], this is a property of NaCl concentration between 0.5 and 4 wt% while it was
the dry salt rather than of dissolved perchlorate. Taking into constant within the limits of reproducibility above 5 wt%
account the decreased gas formation rate from perchlorate marking a ‘saturation point’. By increasing the RF voltage,
solutions, it is more likely that additional oxygen species from the electric potential in the solution is enhanced. This leads to
its decomposition scavenge molecular hydrogen or hydrogen higher voltage drops in the restriction zone resulting in higher
atoms finally yielding H2 O as product. plasma intensities and gas formation rates.
In addition to the substances summarized in figure 7, a The specific conductivity of the electrolyte solution
series of experiments was conducted with the two ammonium seems to be the appropriate parameter being relevant for
salts NH4 Cl and N(CH3 )4 Cl and compared with the stable discharge initiation and intensity rather than the electrolyte
electrolyte NaCl at the same concentration of 5 wt%. Note mass concentration. This could be demonstrated with solutions
that the gas formation rate of the stable electrolyte NaCl does of Na2 SO4 exhibiting a lower specific conductivity than NaCl
not depend on the concentration for high values >4 wt%. Both solutions at the same mass-based electrolyte concentration. As
ammonium salts exhibited markedly larger gas formation rates shown in figure 9 for Na2 SO4 , the gas formation rate increased
and a much higher H2 /O2 ratio as compared with NaCl at the between 2 and 10 wt% and was constant at about 13 ml min−1
same concentration (figure 8). This was due to an increased at an RF voltage of 1.7 kV for Na2 SO4 concentrations above
formation of H2 . At the same time, the O2 formation was 10 wt% (saturation point).
strongly reduced and additional products such as N2 for both Comparing the measurements with NaCl and Na2 SO4 ,
ammonium salts and CO, CH4 and CO2 for N(CH3 )4 Cl were it can therefore be concluded that minimum solution
observed. Traces of N2 O could be also detected for NH4 Cl. conductivities between 1 and 2 S m−1 are necessary for

5
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

Figure 11. Dependence of the gas formation rate on the pH value of

Figure 9. Dependence of the gas formation rate on the electrolyte a 5 wt% NaCl solution at three RF voltages.
concentration at three RF voltages in reactor type II.
composition (H2 : N2 = 75 : 22):

2NH3 → N2 + 3H2 . (1)

The small excess of H2 can be attributed to the water

decomposition.

3.4.3. pH value of the solution The influence of the pH value

was investigated using a 5 wt% NaCl solution in the reactor
type II and varying the pH value by adding small amounts of
HCl or NaOH. As displayed in figure 11, no significant effect of
the pH on the gas formation rates was observed. The measured
rates just scatter around an average value which depended on
the applied RF power. The molar H2 /O2 ratio was constant at
2.0 for all measurements.
Figure 10. Dependence of the gas composition (H2 , O2 and N2 This holds for an extremely broad pH range covering
concentrations) and the total gas formation rates on the about 26 orders of magnitude (!) in the H+ to OH− activity
concentration of NH4 Cl (1.7 kV, reactor type II). ratio. Therefore, it can be concluded that the concentrations
(activities) of H+ and OH− ions in the solution of a stable
discharge initiation in the experimental setup and that a electrolyte do not significantly affect the plasma chemistry.
constant gas formation rate is found for solutions of stable It has already been mentioned that pure hydroxide
electrolytes with conductivities >6–7 S m−1 . solutions (pH > 13) yielded very low gas formation rates,
possibly caused by modifying the surface of the reactor walls
Figure 10 shows the gas formation rates and composition
in the hot plasma zone. This may interrupt the electrolyte film
from NH4 Cl solutions in more detail. The concentration
between vapour phase and glass wall so that no current can
of H2 in the product gas increased and the concentration of
flow through the solution. After short plasma operation times
O2 decreased with increasing NH4 Cl concentration forming a
the reactor walls were damaged.
plateau at high electrolyte concentrations. At the same time,
the concentration of N2 in the product gas was nearly constant.
Obviously, the decomposition of NH4 Cl takes place in parallel 4. Summary and conclusions
to the water dissociation.
A combination of a new reactor type characterized by a
It can be supposed that thermal and electron stimulated restriction in cross section and plate electrodes situated outside
processes are responsible for the primary step of decomposi- of the reactor was presented. It can be demonstrated that this
tion of NH+4 ions at the gas–liquid interface resulting in the setup is suitable to focus RF energy to a restricted volume of
formation of ammonia gas. the reactor leading to the initiation of discharges in aqueous
NH3 can undergo various reactions with radicals and electrolyte solutions. Three subsequent phases—selective
electrons in the gas phase. Its decomposition into the heating, in situ formation of water vapour bubbles and initiation
elements (equation (1)) describes adequately the measured gas of discharges—were observed. As a consequence of the

6
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

discharge processes, water molecules dissociate and H2 plus prevent discharge initiation in particular when comparing with
O2 are formed as products. experiments with HNO3 , where under similar experimental
Generally, electrolytes can be divided into stable and conditions discharges could be initiated and significant gas
reactive electrolytes. The first group (e.g. NaCl, K2 SO4 ) is formation rates were measured.
characterized by a molar H2 /O2 ratio of 2.0 in the product gas The gas formation rate seems to be directly proportional
and similar gas formation rates at the same conductivity. This to RF power and proportional to the square of RF voltage.
group comprises electrolytes that are thermally stable below The physical principles explaining the initiation of
temperatures of 500 ◦ C in their pure form. Stable electrolytes discharges in the presented experimental setup under the
were not converted by plasma chemical processes so that conditions defined here will be discussed in a forthcoming
mixtures of H2 and O2 were formed. publication [12].
Using reactive electrolytes (e.g. HCl, NH4 Cl, In principle, potential applications of the phenomenon and
Mg(ClO4 )2 ), molar H2 /O2 ratios deviated from a value of 2.0 the novel reactor type can be seen in the field of ‘electrodeless’
and typically higher or lower gas formation rates than in the water scission for hydrogen production and the decomposition
case of stable electrolytes at the same electrolyte conductivity of water pollutants, in particular, of surfactants. However,
were measured. Reactive electrolytes are either thermally these applications compete with established techniques and
sensitive compounds in their pure form (decomposition their practical use depends mainly on the energy efficiency.
temperatures below 500 ◦ C) or exhibit a significant vapour
pressure at 100 ◦ C. Additional products such as N2 in the case
of NH4 Cl and CO, CO2 and CH4 in the case of N(CH3 )4 Cl Acknowledgment
were formed.
The fact that thermally labile electrolytes have an impact The authors thank the graduate school BuildMoNa at the
on plasma-chemical interactions in the discharge zone can be Leipzig University for financial support of one of the
seen as a hint but not as proof for the hypothesis of thermally authors (JS).
initiated reactions. Nevertheless, it is not yet clear where
electrolyte-converting processes occur. Decomposition might
occur thermally but also by electron impact on the gas–liquid References
interface followed by subsequent transport of decomposition [1] Mukasa S, Nomura S, Toyota H, Maehara T and Yamashita H
products into the discharge zone. Alternatively, the formation 2011 Internal conditions of a bubble containing
of electrolyte aerosols that circulate through the discharge radio-frequency plasma in water Plasma Sources Sci.
zone could contribute to the plasma chemistry. Since electron Technol. 20 034020
impact on a surface can produce locally high temperatures [2] Maehara T, Honda S, Inokuchi C, Kuramoto M, Mukasa S,
Toyota H, Nomura S and Kawashima A 2011 Influence of
anyway, it is hard to discriminate between thermally stimulated conductivity on the generation of a radio frequency plasma
and electron-collision-stimulated processes. Nevertheless, surrounded by bubbles in water Plasma Sources Sci.
experiments with the surfactant SDS that is enriched in the Technol. 20 034016
gas–liquid interface showed no significant influence on gas [3] Kitano K, Aoki H and Hamaguchi S 2006
(H2 and O2 ) formation. Therefore, it can be concluded that Radio-frequency-driven atmospheric-pressure plasmas in
contact with liquid water Japan. J. Appl. Phys. 45 8294–7
the liquid phase which might be penetrated by electrons has no [4] Sugama C, Tochikubo F and Uchida S 2006 Glow discharge
significant influence on the oxyhydrogen formation processes. formation over water surface at saturated water vapor
The results presented in this work lead to the conclusion pressure and its application to wastewater treatment Japan.
that the electrical conductivity of the solution is the crucial J. Appl. Phys. 45 8858–63
parameter that determines discharge initiation. A minimum [5] Maehara T, Toyota H, Kuramoto M, Iwamae A, Tadokoro A,
Mukasa S, Yamashita H, Kawashima A and Nomura S 2006
conductivity between 1 and 2 S m−1 (equivalent to 0.5 wt% Radio frequency plasma in water Japan. J. Appl. Phys.
NaCl and 2 wt% Na2 SO4 , respectively) was required to initiate 45 8864–8
discharges and gas formation with any electrolyte in the [6] Bruggeman P and Leys C 2009 Non-thermal plasmas in and in
experimental setup. The gas formation rate enhanced almost contact with liquids J. Phys. D: Appl. Phys. 42 053001
linearly with increasing conductivity to a value between 6 and [7] Locke B R, Sato M, Sunka P, Hoffmann M R and Chang J-S
2006 Electrohydraulic discharge and nonthermal plasma for
7 S m−1 (4.5 wt% NaCl and 10 wt% Na2 SO4 , respectively). water treatment Indust. Eng. Chem. Res. 45 882–905
Further increase in conductivity had no influence on the gas [8] Roy R, Rao M and Kanzius J 2008 Observations of polarised
formation rate for stable electrolytes. Gas formation seems rf radiation catalysis of dissociation of H2 O–NaCl solutions
to further depend on the concentration of reactive electrolytes Mater. Res. Innov. 12 3–6
as found for NH4 Cl since these electrolytes act as educts for [9] Bruggeman P, Degroote J, Vierendeels J and Leys C 2008
Dc-excited discharges in vapor bubbles in capillaries
reactions initiated by discharge processes. As no dependence Plasma Sources Sci. Technol. 17 025008
of gas formation on pH values up to 13 was observed, it can [10] Bruggeman P, Leys C and Vierendeels J 2007 Experimental
be concluded that water molecules rather than H+ or OH− are investigation of dc electrical breakdown of long vapour
the dominant precursors for H2 and O2 formation. bubbles in capillaries J. Phys. D: Appl. Phys. 40 1937–43
While most of the tested electrolytes could be used to [11] Miichi T, Hayashi N, Ihara S, Satoh S and Yababe C 2002
Generation of radicals using discharge inside bubbles in
initiate discharges, this was not possible with nitrates of sodium water for water treatment Rep. Fac. Sci. Eng. Saga Univ. 31
and potassium. It is not yet understood why these nitrates 49–52

7
Plasma Sources Sci. Technol. 22 (2013) 015010 J Schneider et al

[12] Schneider J, Holzer F, Rabe C, Häupl T, Kopinke F-D and [15] Taylor G B 1925 Vapor pressure of aqueous solutions of nitric
Roland U 2012 Water dissociation in a radio-frequency acid Indust. Eng. Chem. 17 633–5
electromagnetic field with ex situ electrodes—modelling of [16] Stern S A, Mullhaupt J T and Kay W B 1960 The
discharge initiation Plasma Sources Sci. Technol. submitted physicochemical properties of pure nitric acid Chem. Rev.
[13] Pimblott S M, Laverne J A and Mozumder A 1996 Monte carlo 60 185–207
simulation of range and energy deposition by electrons in [17] El-Awad A M, Gabr R M and Girgis M M 1991
gaseous and liquid water J. Phys. Chem. 100 8595–606 Kinetics and mechanism of thermal decomposition
[14] Marsh A R W and Mcelroy W J 1985 The dissociation of magnesium perchlorate catalysed by metal
constant and henry’s law constant of HCl in aqueous metavanadate additives Thermochim. Acta
solution Atmos. Environ. 19 1075–80 184 205–12

View publication stats