Investigation of dielectric barrier discharge dependence on permittivity of barrier materials

Article abstract of DOI:10.1063/1.2716848

It has been evidenced by both theory and experiments that a barrier possessing a high permittivity can create a high energy for dielectric barrier discharge. However, this argument was challenged by the recent experiments of the authors. They found an optimum permittivity value to generate denser and stronger microdischarges, as well as a high reactivity of destruction CO ₂ in the entire gap space for plasma chemistry. When the barrier permittivity was higher than this optimum value, the microdischarges became sparser and even extinguished since the large number of charges resulted from higher permittivity accumulated on the dielectric surface to reduce the electric field.

Full text of DOI:10.1063/1.2716848

Investigation of dielectric barrier discharge dependence on permittivity of barrier
materials
Ruixing Li, Qing Tang, Shu Yin, and Tsugio Sato
Citation: Applied Physics Letters 90, 131502 (2007); doi: 10.1063/1.2716848
View online: http://dx.doi.org/10.1063/1.2716848
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/90/13?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Effect of packing material on methane activation in a dielectric barrier discharge reactor
Phys. Plasmas 20, 123507 (2013); 10.1063/1.4843875
Pattern formation in dielectric barrier discharges with different dielectric materials
Phys. Plasmas 18, 033506 (2011); 10.1063/1.3566003
Decomposition of methylene blue in water using a dielectric barrier discharge: Optimization of the operating
parameters
J. Appl. Phys. 104, 103306 (2008); 10.1063/1.3021452
Optimization of ultraviolet emission and chemical species generation from a pulsed dielectric barrier discharge at
atmospheric pressure
J. Appl. Phys. 98, 023301 (2005); 10.1063/1.1980530
New technique deducing plasma potential by a capacitive coupling method in spraying dielectric barrier
discharge plasmas
Rev. Sci. Instrum. 76, 013503 (2005); 10.1063/1.1832172
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.133.66.132 On: Wed, 26 Nov 2014 17:07:44

APPLIED PHYSICS LETTERS 90, 131502 ͑2007͒
Investigation of dielectric barrier discharge dependence on permittivity
of barrier materials
Ruixing Li,a͒ Qing Tang,b͒ Shu Yin, and Tsugio Sato
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan
͑Received 25 December 2006; accepted 21 February 2007; published online 26 March 2007͒
It has been evidenced by both theory and experiments that a barrier possessing a high permittivity
can create a high energy for dielectric barrier discharge. However, this argument was challenged by
the recent experiments of the authors. They found an optimum permittivity value to generate denser
and stronger microdischarges, as well as a high reactivity of destruction CO2 in the entire gap space
for plasma chemistry. When the barrier permittivity was higher than this optimum value, the
microdischarges became sparser and even extinguished since the large number of charges resulted
from higher permittivity accumulated on the dielectric surface to reduce the electric field.
© 2007 American Institute of Physics. ͓DOI: 10.1063/1.2716848͔
Any success in research and development of a feasible
CO2 utilization will signify the attainment of double objec-
tives of slowing down a buildup of greenhouse gases in the
atmosphere and better carbon resource recycling utilization.
Many investigations have been done, and a nonequilibrium
plasma rout is one of the most promising methods, including
corona discharge by a packed-bed reactor,1 dielectric barrier
discharge ͑DBD͒ plasma,2,3 etc., to depend on some dielec-
tric materials for both cases. However, a high permittivity
dielectric ceramic generally exhibits low fracture and low
dielectric strengths so that it tends to break down under a
strong current pulse in a DBD apparatus, although a high
permittivity can theoretically create a high energy.4 We have
developed a dielectric barrier material, Ca1−xSrxTiO3 with
Li2Si2O5 additive, which has a high relative density, a high
dielectric constant, and a high dielectric strength.5–7 This ce-
ramic was used as dielectric barriers to ignite dense and
strong plasma with high energy.5 Moreover, it was employed
to break down CO2 efficiently by a DBD plasma.6,7 Com-
parative studies showed that the CO2 reactivity increased
with increasing relative permittivity of the barrier materials
and both of them followed the sequence Ca1−xSrxTiO3
ӷalumina ͑Al2O3͒ജsilica glass ͑SiO2͒.6,7 However, our re-
cent in-depth study showed an existence of an optimum per-
mittivity for the barrier materials to generate dense and
strong current pulses.
Li2Si2O5 was prepared by a conventional solid state re-
action using powders of Li2CO3 and amorphous SiO2. They
were mixed, calcined, and remilled with zirconia balls in a
polyethylene bottle. CaTiO3, SrTiO3, and BaTiO3 powders
were supplied by the Sakai Chemical Industry Co. Appropri-
ate quantities of CaTiO3, SrTiO3, BaTiO3, and Li2Si2O5
were wet mixed with ZrO2 balls and ethanol. The mixtures
were dried and then ground by an agate mortar. The speci-
mens were pressed uniaxially at 20 MPa and cold isostati-
cally at 200 MPa, and then sintered in air. The sintered bod-
ies were sliced and ground to be used as the dielectric
barriers. The permittivity was measured by an impedance
analyzer ͑Agilent Tech., 4299A͒ using a pellet sample of
6 mm in diameter and 5 mm in thickness.
The experimental setup for the CO2 decomposition is
shown in Fig. 1. A planar DBD plasma reactor, whose shell
was made of Teflon, was used for CO2 decomposition. The
fictitious gas mixture ͑CO2:N2=10:90͒ was fed into two
parallel-plate electrodes ͑24ϫ12ϫ3 mm3͒ made of stainless
steel at a flow rate of 100 ml/min, where the ground
electrode was covered with a 1 mm thick dielectric barrier
͑30ϫ15 mm2͒. The gap space between the dielectric barrier
and the counterelectrode ͑high-voltage electrode͒ was 1 mm,
and the CO2 destruction was carried out under atmospheric
pressure by using a conventional flow system. The CO2
concentration was quantitatively analyzed by a CO2 meter
͑Shimadzu URA107͒. A sinusoidal voltage was applied to
the electrodes by an ac high-voltage amplifier with a function
generator ͑Trek 12193͒. A current probe ͑Iwatsu SS-240͒
with about 15 A peak and 0.1 V/A was employed. The ap-
plied voltage was increased stepwise with 1 kV/min in time
until the plasma was generated and then was held at this
discharge onset voltage. During the plasma generation, the
discharge processes were continuously monitored by a mul-
tichannel digital oscilloscope ͑Iwasaki DS-8812͒.
Characteristics of DBD plasmas generated by various
dielectric barriers with different permittivities are shown in
Fig. 2, where the sinusoidal curve represents the voltage ap-
plied to the electrodes to generate the plasma. It may be
found that the sinusoidal profiles were changed from smooth
a͒
Author to whom correspondence should be addressed. Electronic mail:
ruixingli@yahoo.com
b͒
Permanent address: Institute of Process Engineering, Chinese Academy of
Sciences.
FIG. 1. Schematic diagram of experimental apparatus.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2007/90͑13͒/131502/3/$23.00 90, 131502-1 © 2007 American Institute of Physics
130.133.66.132 On: Wed, 26 Nov 2014 17:07:44

131502-2
Li et al.
Appl. Phys. Lett. 90, 131502 ͑2007͒
FIG. 2. Profiles of microdischarges using barriers of ͑a͒ commercial silica
glass, and as-prepared ͑b͒ Ca0.8Sr0.2TiO3 and ͑c͒ Ca0.6Sr0.4TiO3 with
0.5 wt % Li2Si2O5 sintered at 1200 °C for 2 h and ͑d͒ BaTiO3 with
1.5 wt % Li2Si2O5 sintered at 1100 °C for 15 min at an input ac frequency
of 2 kHz ͑permittivity ␧: 25 °C and 10 MHz͒.
FIG. 3. ͑Color online͒ Photographs of carbon deposition on the surfaces of
͑a͒ Ca0.8Sr0.2TiO3 ͑␧=208.1͒ and ͑b͒ Ca0.6Sr0.4TiO3 ͑␧=252.4͒ dielectric
barriers after a CO2 destruction plasma reaction ͑permittivity ␧: 25 °C and
10 MHz͒.
ones ͓Figs. 2͑a͒ and 2͑b͔͒ to zigzaglike ones ͓Figs. 2͑c͒ and
2͑d͔͒ with increasing the permittivities of the barriers. The
microdischarges became sparser and even extinguished when
clear and deep zigzaglike profiles were exhibited. Addition-
ally, it could be found that those zigzags always tended to
decrease the voltage value. That is to say, the ac voltage of
the reactor was not stable and tended to drop during the burst
process of the DBD plasma. This might be attributed to the
large number of charges resulted from the higher permittivity
of the barriers accumulated on the dielectric surface to re-
duce the electric field.8 The key factor, here, is the decrease
of the reactor voltage. It is known that the plasma may be
generated just because the voltage is higher than the dis-
charge onset voltage of the background gas. Otherwise, a
current pulse will choke as soon as the voltage is lower than
this onset voltage. This is why sparser and even extinguished
current pulses are exhibited in Figs. 2͑c͒ and 2͑d͒.
charges deposited on the surface and new microdischarges of
opposite polarity occur. These discharges prefer the residual
channels of previous microdischarges. Consequently, the less
residual channels, the less microdischarges of opposite polar-
ity were ignited. On the other hand, from a perspective of
plasma chemistry, each individual microdischarge can be re-
garded as a miniature reactor. Thus, the plasma reactions
were only limited at the positions of channels and could not
distribute the entire gap space due to the difficulty of devel-
oping new microdischarges after the old channels choked if
the barrier permittivity was too high. This argument is more
pronounced by observing the behavior of carbon deposition
accompanied with the CO2 conversion in Fig. 3. Carbon
powders uniformly deposited on the surface of the
Ca0.8Sr0.2TiO3 barrier with a permittivity of 208.1 ͓Fig. 3͑a͔͒,
whereas an obvious agglomeration was exhibited when
Ca0.6Sr0.4TiO3 barrier with a permittivity of 252.4 was used
͓Fig. 3͑b͔͒. The nonuniform accumulation of carbon powder
might be attributable to a heterogeneous reaction related to
It is known that dielectric barrier serves two functions:4,8
it limits the amount of charge transported by a single micro-
discharge and distributes the microdischarges over the entire
electrode/barrier area. If the electric field in the gas gap is
sufficiently high to initiate avalanches, microdischarges are
established simultaneously and surface discharges described
above cover the dielectric barrier. As long as the external
sinusoidal voltage is rising, additional microdischarges will
occur at new positions because the presence of residual
charges on the surface of dielectric has reduced the electric
fields at the positions where microdischarges have already
occurred. Generally, this process goes along circularly to
generate more and more filament discharges until the voltage
is reversed. However, there were too much residual charges
on the surface of dielectric barrier so that a great decrease of
the electric fields was caused and a compensation of this
downturn by external voltage to exceed the discharge onset
voltage needed time to generate new microdischarges. Thus,
the amount of microdischarges was limited and could not
distribute on the entire barrier surface if the permittivity of
the barrier was too high. Additionally, the residual channels
of the extinguished microdischarges remain in the gap if the
channels have not recovered sufficiently in the meantime.
Otherwise, they tend to disappear. These residual channels
represent privileged locations for the ignition of new micro-
discharges, if the applied voltage is reversed. At this time,
FIG. 4. Logarithmic permittivity
␧ dependence of CO2 conversion
using commercial silica glass and alumina, as-prepared Ca SrxTiO3
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. D1o−wx nloaded to IP:
͑0.1ഛxഛ1͒, and BaTiO3 dielectric barriers.
130.133.66.132 On: Wed, 26 Nov 2014 17:07:44
the voltage across the gas is dominated by the memory

131502-3
Li et al.
Appl. Phys. Lett. 90, 131502 ͑2007͒
FIG. 5. Various magnified profiles of microdischarges using the dielectric
barrier of as-prepared Ca0.8Sr0.2TiO3 with 0.5 wt % Li2Si2O5 ͑␧=208.1 at
25 °C and 10 MHz͒ sintered at 1200 °C for 2 h at an input ac frequency of
8 kHz.
the limitation of microdischarge channels. Furthermore, the
ability of CO2 destruction was decreased with increasing the
permittivity of the barrier after achieving a maximum CO2
conversion, as shown in Fig. 4. This is another evidence of
less microdischarge channels by a high permittivity barrier in
the gap space. The optimum value of the permittivity was
around 208 ͑measured at 25 °C and 10 MHz͒, i.e., the bar-
rier of Ca0.8Sr0.2TiO3, at input reactor frequency between 2
and 4 kHz. The characteristics of the current pulse caused by
Ca0.8Sr0.2TiO3 at ac frequency of 8 kHz are shown in Fig. 5,
and it may be concluded that each pulse width was about
1 ␮s.
FIG. 6. Plasma energy and efficiency for as-prepared Ca0.8Sr0.2TiO3 with
0.5 wt % Li2Si2O5 ͑␧=208.1͒, commercial alumina ͑␧=10.3͒, and silica
glass ͑␧=4.5͒ ͑permittivity ␧: 25 °C and 10 MHz͒.
Additionally, the fraction of CO2 conversion utilizes unit
plasma power is an important parameter for a plasma chem-
istry. It is known that the average plasma power P is given
by4
version attained to maximum by DBD plasma at input ac
frequency between 2 and 4 kHz when Ca0.8Sr0.2TiO3 with
permittivity of 208 ͑25 °C and 10 MHz͒ was used as a bar-
rier material. On the other hand, when Ca0.6Sr0.4TiO3 and
BaTiO3 that possessed higher permittivities were used as the
barrier materials, nonuniform microdischarge channels
tended to occur so that the chemical reaction was not able to
take place in the entire gap space to cause a decrease of
reaction ability.
2
D
−1
P = 4fC ͑CD + Cg͒ V ͑Vmax − Vmin͒,
V
max ജ Vmin
,
min
͑1͒
where f is the applied frequency, CD is the capacitance of the
dielectric, Cg is the capacitance of the discharge gap, Vmin is
the minimum external voltage required to maintain a dis-
charge, and Vmax is the maximum value of the applied sinu-
soidal voltage wave. Furthermore, the capacitance C is given
by
This research was carried out as one of the projects in
the MSTEC Research Center at IMRAM, Tohoku University.
It was partially supported by the JSPS Asian Core Program
“Interdisciplinary Science of Nanomaterials.” The authors
are indebted to B. Liu in the IMRAM, Tohoku University for
his help rendered in photography and the management of the
Sakai Chemical Industry Co. for supplying the starting pow-
ders of CaTiO3, SrTiO3, and BaTiO3 used in the present
study.
C = ␧ ␧S/d,
͑2͒
0
where ␧0=8.85ϫ10−12 F/m, ␧ is the relative permittivity,
and S and d are the area of the electrode and the distance
between two parallel-plate electrodes, i.e., either 1 mm gap
of air or 1 mm thickness of the barrier in the present study,
respectively. The efficiency of CO2 decomposition was de-
fined as
1K. Jogan, A. Mizuno, T. Yamamoto, and J. Chang, IEEE Trans. Ind. Appl.
29, 876 ͑1993͒.
Efficiency͑%/W͒ = CO2conversion͑%͒/P͑W͒ ϫ 100 % ,
͑3͒
2G. Zheng, J. Jiang, Y. Wu, R. Zhang, and H. Hou, Plasma Chem. Plasma
Process. 23, 59 ͑2003͒.
3S. L. Suib, S. L. Brock, M. Marquez, J. Luo, H. Matsumoto, and Y.
Hayashi, J. Phys. Chem. B 102, 9661 ͑1998͒.
where P is the plasma energy calculated by Eq. ͑1͒. As
shown in Fig. 6, the efficiency is not very high for the
Ca0.8Sr0.2TiO3 barrier with a high permittivity. This is prob-
ably attributed to plentiful “hot” electrons could dissociate
both forward and reverse reactions in the gap space.
Consequently, from the present results, it seems that
there is an optimum permittivity of the barrier material for
generation of dense and strong DBD plasma. The CO2 con-
4U. Kogelschatz, Plasma Chem. Plasma Process. 23, 1 ͑2003͒.
5R. Li, Q. Tang, S. Yin, Y. Yamaguchi, and T. Sato, Phys. Plasmas 11, 3715
͑2004͒.
6R. Li, Q. Tang, S. Yin, and T. Sato, Plasma Chem. Plasma Process. 26,
267 ͑2006͒.
7R. Li, Q. Tang, S. Yin, and T. Sato, Fuel Process. Technol. 87, 617 ͑2006͒.
8Low Temperature Plasma Physics, edited by R. Hippler, S. Pfau, M.
Schmidt, and K. H. Schoenbach ͑Wiley-VCH, Berlin, 2001͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.133.66.132 On: Wed, 26 Nov 2014 17:07:44