Characteristics of Atmospheric Microplasma in Small Discharge Gaps
Marius Blajan and Kazuo Shimizu
Organization for Innovation and Social Collaboration, Shizuoka University, Japan
Abstract: Emission spectroscopy analysis and imaging technique were used to study the microplasma phenomena.
The microplasma discharge in Ar and N2/Ar was analyzed along the discharge gap area and also for specific periods
of time during the discharge. Microplasma showed characteristics specific with an atmospheric glow discharge:
positive column, Faraday dark space and negative glow. Spatial and temporal distribution showed the propagation of
light emission from anode towards cathode. The light emission was measured up to 365 ns as long as the discharge
current was measured.
The intensity of N2 Second Positive Band System (N2 SPS) peak at 337.1 nm was measured along the discharge
gap and at various time intervals during the discharge using an effective shutter opening time of high-speed ICCD
camera of 3 ns. The peak of Ar I at 696.5 nm was used to estimate the value of electron density using Stark
broadening method. The calculated value of electron density ne=1.338x1015 /cm3 is specific for atmospheric
pressure nonthermal plasmas.
Keywords: dielectric barrier discharge, microplasma, emission spectroscopy
1. Introduction
Microplasma applications are ranging from NOx
removal to indoor air treatment, sterilization of
bacteria or surface treatment of polymers [1-3].
Among
nonthermal
plasma
technologies
microplasma has advantages due to its small size of
reactor and power supply and the operation at
atmospheric pressure. This makes it suitable to be
integrated easily with other devices and economical
to use. Our microplasma is a dielectric barrier
discharge at atmospheric pressure [4].
Analysis of microplasma could be performed by
various techniques but due to the small discharge
gaps emission spectroscopy and the use of high speed
imaging camera is suitable for microplasma analysis
[5-6]. This could lead to a better understanding of
phenomena and furthermore to a development and
optimization of the application processes.
2. Experimental Setup
Emission spectra were measured by an ICCD
camera (Ryoushi-giken, SMCP–ICCD 1024 HAMNDS/UV) and a spectrometer to which was attached
a fiber optic (Fig. 1). Photos of microdischarges were
taken using a high speed ICCD camera for imaging
(Princeton Instruments, PI-MAX 3). A fiber optic
with diameter of about 100 µm was used in order to
have an accurate measurement of a small part of the
microplasma discharge. A negative pulse Marx
Generator was used to energize the electrodes [7].
Experiments were carried out at atmospheric pressure
in Ar, N2/Ar. Discharge voltage was negative pulse,
rise time 100 ns, width 1 µs at 1 kHz. The pulse
width can be varied from 200 ns to 2 µs.
Microplasma electrodes are perforated metallic
plates covered with a dielectric layer. Due to small
discharge gaps (0~100 μm) and to the assumed
specific dielectric constant of εr = 104, a high
intensity electric field (107~108 V/m) could be
obtained with relatively low discharge voltages
around 1 kV [1].
The electrode size was 20 mm versus 40 mm. They
have holes to flow the gas, with a diameter of Ø 3mm
and an aperture ratio of 36%. The discharge gap was
set at about 100 μm in our study using a spacer.
Microplasma was measured in different points along
X axis by varying the position of the fiber optic in 20
µm incremental steps. Figure 2 shows side views of
the microplasma electrodes thus X axis was
considered to be along the discharge gap. The
position of fiber optic was perpendicular on X axis.
The position of fiber optic was perpendicular on X
axis. Relative intensity of emission peak of the N2
Second Positive Band System (N2 SPS) peak at 337.1
nm was plotted for each incremental step of 20 µm.
The opening time of ICCD camera for emission
spectroscopy measurements was 15 ns considering
rise time and fall time of 6 ns each thus it could be
considered that the effective opening time was 3 ns.
Fig. 1. An experimental setup.
Fig. 2. Electrode arrangement for measuring
spatial and temporal distribution of microplasma.
Measurements for time incremental steps were
performed up to 190 ns considering as the time origin
the rise of the discharge current.
Images of microplasma were taken using a high
speed ICCD camera with an opening gate time of 2
ns, for 200 gates per exposure and 5 ns time
incremental steps up to 365 ns.
The surface plot depicted in Figure 5 represents
only the relative intensities for each time step
measurement thus measurement corresponding to one
time step could not be compared in term of absolute
intensity with other one. The plot shows how the area
with highest intensity at a given moment varies inside
the discharge gap. Thus it could be observed that at
the beginning of discharge the highest intensity was
measured near anode and after 25 ns it started to shift
towards cathode and back to anode after 130 ns.
Dynamics of microdischarge was observed with
high speed ICCD camera for microplasma in 1% N 2
in Ar at 1 kV.
Trigger signal for ICCD camera was set at 2 ns as
shown in Figure 6(a). The trigger signal was set
initially at the time when the discharge current started
to rise and this was considered as time origin. The
time setting was then increased for each measurement.
Relative Intensity Scale
0.5-0.55
0.55-0.6
0.6-0.65
0.65-0.7
0.7-0.75
0.75-0.8
0.8-0.85
0.85-0.9
0.9-0.95
0.95-1
190
160
130
100
70
60
50
40
35
30
25
20
15
10
5
0
Cathode
Anode
Time (ns)
3. Results and Discussion
The emission spectrum of the microplasma discharge
generated at 1 kV in Ar and 1% N2 in Ar mixture
showed high intensity peaks of N2 second positive
band system (N2 SPS 315.9 nm, 337.1 nm, 357.7 nm,
380.4 nm) [8] and OH peaks (306 ~ 309 nm) [9] (Fig.
3).
For the discharge in pure Ar (purity 99.999%) high
intensity peaks corresponding to ArI were measured
in the violet and red regions (Fig.4) [10]. The
intensity if N2 SPS peak at 337.1 nm was measured
along X axis using experimental setup shown in Fig.
2 by varying across X axis with 20 µm incremental
steps the position of electrodes versus the fiber optic.
Microplasma was generated in 1% N2 in Ar at 1 kV.
Fig. 4. Emission spectrum of microplasma
discharge in Ar at -1.2 kV in red region.
0
40
80
120
160
200
240
280
Distance X (um)
Fig. 3. Emission spectrum of microplasma in Ar
and N2/Ar mixture at -1.2 kV.
.
Fig. 5. Spatial and temporal evolution of the
relative intensity for N2 SPS peak at 337.1 nm.
(a)
1 ns
2 ns
3 ns
4 ns
Anode
Cathode
100 µm
10 ns
20 ns
25 ns
35 ns
45 ns
55 ns
100 ns
365 ns
(b)
Fig. 6. Discharge voltage, corresponding discharge
current and ICCD camera trigger signal (a) and
spatial and temporal evolution of light intensity (b)
during microplasma discharge in 1% N2 in Ar.
Images show that with various incremental steps
the distribution of light emission in the gap was
different suggesting the waves of light intensity that
started directly from the anode (grounded) and
propagate towards the negative one (cathode) (Fig.
6(b)).
The gating of the ICCD camera was set at 2 ns and
it was placed in time for the first image at -1 ns thus
first image showed the phenomena up to 1 ns. Then
the cathode layer development, cathode layer
enhancement and cathode layer decay were observed.
The initial phenomena of electron avalanche from
cathode and the cathode directed streamer are
difficult to capture mainly due to the small discharge
gap characteristic to our microplasma. The cathode
layer development could be observed up to 35 ns.
The highest intensity was observed between 35 and
70 ns which corresponded to cathode layer
enhancement stage. Finally the post discharge
phenomena was observed starting from about 70 ns.
Weak light intensity was measured up to 365 ns. This
could be due to the long life active species that
remained during the afterglow. The discharge gap
was about 100 µm but due to the asperities on the
dielectric the discharge gap could vary. The asperities
on the dielectric could favor due to an increase of the
electric field in certain spots the apparition of
microdischarges.
According to N. Sewraj et al. the electron avalanche
from cathode has an average speed of 2.5 × 105 m/s
and the cathode directed streamer an average speed of
5 × 105 m/s [6]. This data was obtained at 50 torr in
pure nitrogen and for atmospheric pressure it is
expected that due to higher molecule density the
phenomena should be slower. In a 100 µm discharge
gap the electron avalanche from cathode needs 0.4 ns
to reach the anode and after the cathode directed
streamer reaches the cathode in 0.2 ns if we consider
the above mentioned speeds and slower for
atmospheric pressure. Considering this the
phenomena of electron avalanche towards anode and
first streamer reaching cathode was captured in the
first image of Figure 6(b).
The addition of N2 in Ar contributed to a
filamentary discharge due to N2 instead of a diffuse
one in the case of pure Ar discharge. This allowed us
to observe the phenomena of microdischarges. As
shown in Figure 6(b) the narrowest diameter of the
microdischarge is under 30 µm, thinner than other
researchers reported [11].
Stark-broadened spectra can be used to determine
microplasma density [12],[13],[15],[16]. Considering
the dimensions of microplasma only spectroscopy
techniques can be used to measure the plasma density.
In addition to their Stark profiles, the spectra are also
convolutions of instrument broadening and Doppler
broadening. Electric microfields of neighbouring
particles give rise to Stark broadened line profiles.
Stark broadening is characterized by a Lorentzian
profile, whose full width at half max (FWHM) is
directly related to electron density. A Lorentzian and
Gaussian convolution is a Voigt function [12], [13]:
V( )
exp(
G
L( )
4 ln 2( '
G 2W
'
G( )L(
0)
2
)d
'
A*
L2W
)
4(
'
0)
2
L2W
d ' (1)
where A is amplitude of Voigt profile, Gw is FWHM
of the Gaussian function, Lw is FWHM of the
Lorentzian function, λ is the wavelength and λ0 is the
centroid wavelength.
A pseudo Voigt function was used to determine Lw
by applying least square fit and optimizing the
parameters to fit the Ar I peak at 696.5 nm spectra:
LW
2
V( ) V0 S(m u
2
4(
L2W
0)
4 ln 2
4 ln 2
2
(2)
(
0) )
GW
G 2W
where V0 is offset, S area and mu profile shape factor
[14].
After optimizing the parameters to fit the spectra
using Scilab as shown in Figure 7 to optimum value
of Lw = 0.0042564 nm. Considering van der Waals
broadening Ww= 0.003 nm and extracting it from Lw,
the value of Stark width resulted Sw=0.0012564 nm.
(1 m u )
exp(
Fig. 7. Measured and fitted spectra of Ar I peak at
696.5 nm.
The Stark width equation was approximated by
[13],[15]:
2(1 1.75 *10 4 * 4 ne * *
1
(1 0.0068 * 6 ne
)) *10 16 * w * ne
(3)
Te
where ne is the electron density, α is the static ion
broadening parameter, Te is the electron temperature
and w is electron impact half-width. Previously we
reported the electron temperature for microplasma
discharge in Ar at atmospheric pressure
Te=10000~23000 K [2]. Considering also α=0.032,
w=0.00537 nm/1016 cm−3 [13],[15],[16], and
Sw=0.0012564 nm in equation (3) electron density
was calculated ne=1.338x1015 /cm3.
SW
4. Conclusion
The analysis of the spatial and temporal behavior of
the dielectric barrier discharge microplasma showed
its
particularities
regarding
dynamics
and
characteristics of microdischarges:
1) Spatial and temporal distribution of light emission
from microplasma was observed for about 365 ns
corresponding to the duration of discharge current.
Light near the negative electrode was preceded by
the waves of light intensity that started directly
from the anode and propagated towards the
cathode.
2) Spatial and temporal distribution for the N2 SPS
peak at 337.1 nm inside the discharge gap shown a
maximum intensity towards the anode for the 25
ns.
3) Electron density measured a specific value for
atmospheric pressure nonthermal plasmas of
ne=1.338x1015 /cm3.
Acknowledgment
The authors would like to thank Prof. M. Nagatsu
from Shizuoka University for the fruitful discussions.
5. References
[1] K. Shimizu, M. Yamada, M. Kanamori, M.
Blajan, IEEE Trans. Ind. Appl., 46, 641 (2010).
[2] M. Blajan, A. Umeda, S. Muramatsu, and K.
Shimizu, IEEE Trans. Ind. Appl., 47, 1100
(2011).
[3] K. Shimizu, M. Blajan, and T. Kuwabara, IEEE
Trans. Ind. Appl., 47, 2351 (2011).
[4] K. Shimizu, A. Umeda, S. Muramatsu, M.
Blajan, IEEJ Trans. on Fundamentals and
Materials, 130, 858 (2010).
[5] F. Iza, J. L. Walsh, and M. G. Kong, IEEE
Trans. Plasma Sci., 37, 1289 (2009).
[6] N. Sewraj, R. J. Carman, N. Merbahi, F.
Marchal, and E. Leyssenne, IEEE Trans. Plasma
Sci., 39, 2128 (2011).
[7] K. Shimizu, T. Ishii, and M. Blajan, IEEE
Trans. Ind. Appl., 46, 1125 (2010).
[8] A. Lofthus and P. H. Krupenie, J. Phys. Chem.
Ref. Data, 6, 113 (1977).
[9] F. Liu, W. Wang, S. Wang, W. Zheng, and Y.
Wang, J. Electrostat., 65, 445 (2007).
[10] Göran Norlén, Physica Scripta, 8, 249 (1973).
[11] U. Kogelschatz, IEEE Trans. Plasma Sci. 30,
1400 (2002).
[12] G. V. Vogman and U. Shumlak, Rev. Sci.
Instrum. 82, 103504 (2011).
[13] Lifang Dong, Junxia Ran, and Zhiguo Mao,
Appl. Phys. Lett. 86, 161501 (2005).
[14] www.originlab.com.
[15] C. Penache, M. Miclea A. Bräuning-Demian, O.
Hohn, S. Schössler, T. Jahnke, K. Niemax, and
H. Schmidt-Böcking, Plasma Sources Sci.
Technol. 11, 476 (2002).
[16] C. S. Lee, D. M. Camm, and G. H. Copley, J.
Quant. Spectrosc. Radiat. Transf. 15, 211 (1975).