the driving frequency effects on the atmospheric pressure...
Post on 02-Sep-2020
1 Views
Preview:
TRANSCRIPT
The driving frequency effects on the atmospheric pressure corona jetplasmas from low frequency to radio frequency
Dan Bee Kim,1,a) H. Jung,2 B. Gweon,2 S. Y. Moon,3 J. K. Rhee,2 and W. Choe2
1Division of Physical Metrology, Korea Research Institute of Standards and Science, 209 Gajeongno,Yuseong-gu, Daejeon 305-340, Korea2Department of Physics, Korea Advanced Institute of Science and Technology, 335 Gwahangno, Yuseong-gu,Daejeon 305-701, Korea3LG Electronics Advanced Research Institute, 16 Woomyeon-Dong, Seocho-Gu, Seoul 137-724, Korea
(Received 21 February 2011; accepted 16 March 2011; published online 26 April 2011)
Lately, the atmospheric pressure jet type corona plasma, which has been typically driven by dc to low
frequency (LF: several tens of kHz), is often generated by using radio frequency of 13.56 MHz. Yet,
the relationship between the plasma and its driving frequency has seldom been investigated. Hence, in
this study, dependence of the atmospheric pressure corona plasma characteristics on the driving
frequency was explored experimentally from LF to rf (5 kHz–13.56 MHz). The plasmas generated by
the driving frequency under 2 MHz were cylindrical shape of several tens of millimeters long while
the 13.56 MHz plasma is spherical and a few millimeters long. As the driving frequency was
increased, the plasma length became shortened. At the lower driving frequencies (below 2 MHz), the
plasmas existed as positive streamer and negative glow for each half period of the applied voltage, but
the discharge was more continuous in time for the 13.56 MHz plasma. It was inferred from the
measured I–V curves that the higher driving frequency induced higher discharge currents, and the gas
temperature was increased as the driving frequency was increased. VC 2011 American Institute ofPhysics. [doi:10.1063/1.3574256]
I. INTRODUCTION
Compared to the large size atmospheric pressure plasmas,
the small size atmospheric pressure plasmas have their own
merits of better stability, high electron energy, high radical
density, etc., and they are being applied to various new fields,
such as light source and display.1–4 Particularly, the plasma
bio-applications using the small size atmospheric pressure
plasmas have intrigued many researchers. Accordingly, several
corona based (needle/pin to plane or coaxial electrode struc-
ture) small size plasmas have been developed and used to treat
many kinds of germs on culture grounds and living cells.5–8
The atmospheric pressure corona plasma is one of the most
important discharge types in the industry, and its study is also
important for the large area processing plasmas like capacitively
coupled plasmas to avoid the corona generation at the electrode
edge or at the electrode where the flatness is not perfect or lost
due to defect. Lately, unlike the conventional frequency range
of dc to low frequency (LF: several tens of kHz), 13.56 MHz rf
corona discharge has been studied, especially for the bio-appli-
cations.7 In doing so, most of the plasma sources employed rf,
because it is rather common but not necessarily better.
It has been hard to find thorough comparative study results
for various driving frequency ranges although the different
driving frequency means the distinctive plasma features. Most
of the studies have inclined towards the application itself, pay-
ing less attention to the fundamental plasma characteristics.
Such basic study cannot be ignored, because it is essentially
connected with the achievement of best application results.
Therefore, the study about the driving frequency effects on the
atmospheric pressure corona plasma characteristics has been
started by the authors in the previous works, and this work
introduces the additional experimental results.9,10
In general, rf is known to be favorable for its low break-
down voltage and high plasma density. Also, due to the short
duration of the rf field, the discharge is continuous in time
unlike the LF corona of which the discharge mode has pulse-
like nature due to the space charge accumulation. On the
other hand, the streamer, one of the corona discharge modes
containing very highly energetic ions and electrons due to its
fast moving nature, cannot be found in high rf corona
plasma, because it involves the ion motion.11
Because it is in the early stage of research, the LF and
the rf corona based jet type plasmas have been independently
investigated, each originated from different plasma sources.
However, as they evolve, their sources have ended up in sim-
ilar structures of the pin electrode with/without the coaxial
or the plane ground electrode. Hence, for now, it is required
to study them in the same system and compare their dis-
charge characteristics in order to define their distinctive fea-
tures and to clarify advantages and disadvantages of each for
certain application purposes. In this study, the atmospheric
pressure corona plasmas were produced in a single pin elec-
trode system using different driving frequencies (fin) ranged
from LF to rf (5 kHz–13.56 MHz), and their discharge char-
acteristics like physical appearances, electrical properties,
temperatures, and discharge modes were investigated.
II. EXPERIMENTAL
As illustrated in Fig. 1(a), the plasma source is quite
simple. It consists of a copper pin and a Pyrex tube. More
a)Author to whom correspondence should be addressed. Electronic mail:
danbeek@kriss.re.kr.
1070-664X/2011/18(4)/043503/5/$30.00 VC 2011 American Institute of Physics18, 043503-1
PHYSICS OF PLASMAS 18, 043503 (2011)
Downloaded 26 Apr 2011 to 210.98.16.31. Redistribution subject to AIP license or copyright; see http://pop.aip.org/about/rights_and_permissions
detailed information about the source can be found in the pre-
vious report.9 In this study, the pin electrode was powered by
different driving frequencies, and the plasmas were generated
in the ambient air with the helium gas supply. The corona
plasma was generated at LF (5–75 kHz) using a function gen-
erator (Marconi 2023, 9 kHz–1.2 GHz) and an amplifier
(Trek 20/20C) or a power source (FTLab HPSI200). Follow-
ing the plasma generation using LF, rf (2–13.56 MHz) was
employed with appropriate powers and matching network
systems (Marconi 2023, 9 kHz–1.2 GHz; AR500A100A,
10 kHz–100 MHz, 500 W; DresslerCesar 1312, 13.56 MHz,
1200 W). Then, the plasma characteristics were investigated
with various diagnostic tools like cameras, I–V probes, ther-
mometer, and spectrometer.
To reduce the experimental parameters, the helium gas
flow rate was fixed at 3 lpm, and the electrode head was fixed
at 5 mm inside from the Pyrex tube end. We are not report-
ing here in detail about the other experiments of controlling
these parameters. Briefly stating, the LF plasma length was
affected more than the rf plasma length by the gas flow.
Also, a decrease in the breakdown voltage was observed for
both LF and rf plasmas when the pin electrode was placed
deeper in the dielectric tube due to the higher helium gas pu-
rity and less air impurities.
III. CORONA DISCHARGE GENERATED BY fin
UNDER 2 MHz
Various atmospheric corona plasmas of different fin are
depicted in Fig. 1(b). The most prominent difference is the
plasma length. The plasma length decreased when the driv-
ing frequency was raised as shown in Fig. 2. First, in case of
the 5 kHz plasma, the plasma was confined near the pin elec-
trode at Vin < 2200 V as a glow mode, then it became longer
and a cylindrical shape on changing its mode to the streamer
at Vin > 2200 V. Its length was about 50 mm at the maxi-
mum. For the increasing current and voltage, the discharge
mode of corona plasma usually transforms from the glow to
the streamer. The abrupt increase in the plasma length is
resulted from the mode transition from the glow to the
streamer, because the latter propagates away from the pin
electrode.
When the fin was raised by ten times to 50 kHz, the
plasma developed into the streamer mode soon after the
breakdown. Its length varied between 5 and 30 mm for the
increasing input voltage as shown in Fig. 2 which is shorter
than that of 5 kHz plasma. The rather large decrease in the
plasma length by about 10 mm at around 1400 V is coming
from the change in the plasma. The change could be
observed with the bare eyes. As the input voltage is
increased, a bright thin core region starts to appear in the
center and grows in its length. Then, at a certain voltage,
only the bright core is remained, and this causes the drop in
the plasma length.
The asymmetric electrode configuration makes the dis-
charge mode for each half period of the applied voltage dif-
ferent. When the pin electrode is negatively biased, the
electrons are pulled away from it toward the low field region.
On the other hand, when it is positively biased, the electrons
are accelerated toward the high field of the pin electrode. As
a result, the positive corona plasma has higher energy elec-
trons, and thus, it can be developed into the streamer mode
at a relatively lower current. Using the time-resolved meas-
urements, it can be actually observed that the plasma is in
the negative glow and the positive streamer for each half pe-
riod at the same applied voltage, respectively.12
The 2 MHz plasma length plotted in Fig. 2 shows that
the plasma length got shortened further by more than one
third compared to the 50 kHz plasma. At this frequency
range of 2 MHz, evidence of the positive streamer propagat-
ing away from the pin electrode and the negative glow stay-
ing near the pin electrode is still seen in the time-resolved
images of Fig. 3 taken by using an intensified charge-
coupled device (ICCD) camera. Figure 3(a) illustrates two
FIG. 1. (Color online) (a) Schematic of the experimental setup and (b) the
generated plasma images of different fin. The diameter of the Pyrex glass
tube in which the pin electrode was placed was about 6 mm, and the tube is
shown or drawn for the scale reference.
FIG. 2. (Color online) Length of plasmas generated by different fin against
the discharge voltage (Vrms).
043503-2 Kim et al. Phys. Plasmas 18, 043503 (2011)
Downloaded 26 Apr 2011 to 210.98.16.31. Redistribution subject to AIP license or copyright; see http://pop.aip.org/about/rights_and_permissions
different time frames for the positive discharge. As time
evolved from t1 to t2, the streamer head can be seen travel-
ing away from the electrode. It is the streamer propagation
distance that determines the plasma length. Increase in the
driving frequency means decrease in the time period, and
the short time gives the streamer less time to travel away
from the electrode. Unlike the 2 MHz plasma, of which
only one streamer head can be seen during the positive half
period [Fig. 3(a)], we could observe two streamer heads at
50 kHz. Experimental investigations showed that it is due to
the time gap between the prebreakdown and main break-
down, which gradually reduces to zero with the increase in
the driving frequency. The behavior of the positive streamer
including its relation to the driving frequency was studied in
detail and reported in another paper.12 It is worthwhile to
note here that the driving frequency can affect the time and
spatial distribution of the corona plasma by controlling the
discharge mode.
The experimental results about the corona discharge and
its driving frequency can be classified by the fin of under or
above 2 MHz, roughly. Two megahertz is around the plasma
ion frequency, and both electrons and ions can follow the
external electric field at fin under 2 MHz while only the elec-
trons can at fin above 2 MHz.13 Hence, the plasma properties
are different for these two driving frequency ranges with
respect to 2 MHz. The 13.56 MHz plasma exhibited differ-
ent plasma features, and it will be discussed in the next
section.
IV. 13.56 MHZ CORONA DISCHARGE
The 13.56 MHz plasma image shown in Fig. 1(b) is dif-
ferent from the plasmas driven by the lower fin. It is rather
spherical and confined near the pin electrode with a length of
only a few millimeters.
The main difference of the plasmas driven by different
fin is the plasma length. In the limiting case of very frequent
collisions (mm�x), as in the atmospheric pressure plasma,
the electron displacement can be expressed as below:
r ¼ eEa
mexmmcos xt ¼ leEa
xcos xt; (1)
where le is the electron mobility.14 The equation indicates
that the electrons oscillate with the larger amplitude for the
lower frequency. In other words, the electrons are less local-
ized near the electrode for the lower frequency. Because it is
mainly the electrons that play an important role in the plasma
ionization process, the electron oscillatory displacement is
related to the plasma physical size. The calculated electron
displacement and experimentally obtained plasma length
were agreed well with each other for the 13.56 MHz plasma
of which only the electron motion is involved, but for the
lower driving frequency plasmas the ion motion needs to be
considered for a better agreement.9
The 13.56 MHz plasma is rather continuous in time
wise, having a single glow mode throughout the whole pe-
riod. Such dissimilarity in the discharge characteristics with
the lower frequency discharges is attributed to the fact that at
this frequency range, only the electrons of much higher mo-
bility follow the field while the ions only experience the av-
erage field.13 Thus, for the 13.56 MHz plasma, it is always
the electrons of which the density profile instantaneously
changes with the electric field, making the discharge more
continuous.15 However, in the plasmas generated by fin under
2 MHz or the ion oscillation frequency, both electrons and
ions move along the field, and due to their mobility differen-
ces, different plasma modes are observed for each half pe-
riod of the applied voltage.16
Figure 4 shows the time-resolved (wavelength-unresolved)
13.56 MHz plasma light emission intensity for one period of
the applied voltage. Although the stationary ions make the dis-
charge having a single mode through the one voltage period at
this frequency range, the 13.56 MHz plasma still possesses an
asymmetrical nature due to the asymmetric source geometry.
The positive plasma has higher energy electrons as they accel-
erate toward the high field pin electrode. Thus, the plasma
emission intensity is higher during the positive half period.
Furthermore, three representative molecular/atomic emission
lines of N2þ (391.4 nm), He I (706.5 nm), and O I (777.5 nm)
were observed in detail [Fig. 4(b)]. N2þ is excited by the com-
bined action of He metastables (Penning ionization of N2) and
low energy electrons of 3 eV (excitation of N2þ), whereas the
helium excitation requires high energy electrons of 22 eV
through the ground state electron impact excitation.17 Although
it is a helium plasma, the O I intensity was larger, because of
its lower excitation energy (11 eV). Unlike the neutral species,
the positive ion N2þ shows the opposite behavior, because the
positive ion is lost at the relatively negative glass wall through
the recombination during the positive half period.18
V. ELECTRICAL CHARACTERISTICS AND Tgas
I–V characteristic curves are presented in Fig. 5 for the
different fin plasmas. It can be seen that the higher driving
FIG. 3. (Color online) Time-resolved 2 MHz plasma images corresponding
to (a) the positive voltage half period and (b) the negative voltage half pe-
riod. The discharge is in the streamer mode, traveling away from the elec-
trode, during the half period of the positive voltage (when the pin electrode
is positively biased), whereas it exists as the glow mode, being confined
near the electrode, during the negative voltage half period.
043503-3 The driving frequency effects … Phys. Plasmas 18, 043503 (2011)
Downloaded 26 Apr 2011 to 210.98.16.31. Redistribution subject to AIP license or copyright; see http://pop.aip.org/about/rights_and_permissions
frequency means the lower discharge voltage and the higher
discharge current. Considering that the plasma volume is
much smaller for the higher driving frequency, the discharge
current density difference is even much larger. The higher
discharge current indicates the higher electron density and/or
temperature.19 As a matter of fact, the plasma emission spec-
trum of the 13.56 MHz plasma had much higher intensity
and more helium atomic emission lines, indicating larger
plasma density and/or electron temperature.9 Also, the break-
down voltage decreased from 1100 to 350 V as the driving
frequency increased from 5 kHz to 13.56 MHz. When the
driving frequency is increased above a few megahertz, the
electron loss and thus the breakdown voltage are largely
reduced due to the great reduction in the oscillation ampli-
tude.20 Therefore, the breakdown and discharge voltage of
the rf plasma is much lower than that of LF plasma.
The gas temperature was measured using a fiber ther-
mometer (FISO FOT-H). The validity of the fiber thermome-
ter was confirmed by comparing the temperature values with
those from the optical emission spectroscopy based on of the
N2þ(391.4 nm) spectra.21 The gas temperature became
higher with the increase in the driving frequency as shown in
Fig. 6. The 50 kHz plasma gas temperature was only a little
higher than the room temperature (20 �C) and stayed about
the same even for the growing discharge voltage. On the
other hand, the 13.56 MHz plasma gas temperature was
increased with the discharge voltage. The higher gas temper-
ature is due to the fact that the 13.56 MHz discharge is rather
continuous while the 50 kHz discharge is pulselike and that
the plasma volume is smaller for the 13.56 MHz case, and
thus, the power density is higher. Also, the discharge current
and ohmic heating is larger for the higher frequency.22
FIG. 4. (Color online) (a) Applied voltage wave-
form and intensity of the time-resolved plasma
emission and (b) time- and wavelength-resolved
plasma emission intensities of various molecular
and atomic lines for the 13.56 MHz plasma.
FIG. 5. (Color online) I–V curves for plasmas generated by different fin. FIG. 6. (Color online) Tgas for plasmas generated by different fin.
043503-4 Kim et al. Phys. Plasmas 18, 043503 (2011)
Downloaded 26 Apr 2011 to 210.98.16.31. Redistribution subject to AIP license or copyright; see http://pop.aip.org/about/rights_and_permissions
VI. CONCLUSIONS
In conclusion, the atmospheric pressure corona plasmas
were generated in the single pin electrode structure for differ-
ent fin from 5 kHz to 13.56 MHz, and they showed several dis-
tinctive discharge characteristics. First, as fin was increased,
the plasma length was decreased due to the change in the dis-
charge mode and also by the decrease in the electron displace-
ment, which is inversely proportional to fin and proportional to
the external electric field. The longer plasma length can be ad-
vantageous for treating objects, because large distance
between the plasma source and the object means convenience
or less spark generation. Second, the discharge characteristics
were more continuous in time for the 13.56 MHz plasma as
seen from the current and the plasma emission measurements.
Such continuity is somehow related to the plasma homogene-
ity, and thus, the 13.56 MHz plasma may be more favorable
for its better uniformity. However, existence of the streamer
mode for fin below 2 MHz needs to be noted, because the
streamer which has the highly energetic ions and electrons
makes some applications efficient. Third, the I–V measure-
ments inferred that the higher fin means the higher discharge
current, and thus, the higher plasma density. Finally, the gas
temperature was increased as the driving frequency was
increased. The temperature tends to increase in the presence
of a target, so the lower frequency plasma may be preferred
for the temperature for the treatments of thermally sensitive
materials including the living organisms.
In this manner, the LF and rf plasmas exhibit distinct dis-
charge characteristics with both advantages and disadvantages,
which can be identified differently for various application pur-
poses. This comparative study can establish a foothold to clar-
ify the difference and widen each employment field, thus,
improve the application results. Furthermore, the two fre-
quency regimes can be simultaneously used to generate a dual
frequency plasma with maybe both discharge features.23
ACKNOWLEDGMENTS
This work was supported by KAIST.
1K. H. Becker, K. H. Schoenbach, and J. G. Eden, J. Phys. D: Appl. Phys.
39, R55 (2006).2J. Franzke, K. Kunze, M. Miclea, and K. Niemax, J. Anal. At. Spectrom.
18, 802 (2003).3K. Kelly-Wintenberg, A. Hodge, T. C. Montie, L. Deleanu, D. Sherman, J. R.
Roth, P. Tsai, and L. Wadsworth, J. Vac. Sci. Technol. A 17, 1539 (1999).4Y. Yang, Ind. Eng. Chem. Res. 41, 5918 (2002).5B. Gweon, D. B. Kim, S. Y. Moon, and W. Choe, Curr. Appl. Phys. 9, 625
(2009).6B. Gweon, D. Y. Kim, D. B. Kim, H. Jung, W. Choe, and J. H. Shin, Appl.
Phys. Lett. 96, 101501 (2010).7E. Stoffels, I. E. Kieft, and R. E. J. Sladek, J. Phys. D: Appl. Phys. 36,
2908 (2003).8A. Shashurin, M. Keidar, S. Bronnikov, R. A. Jurjus, and M. A. Stepp,
Appl. Phys. Lett. 93, 181501 (2008).9D. B. Kim, J. K. Rhee, B. Gweon, S. Y. Moon, and W. Choe, Appl. Phys.
Lett. 91, 151502 (2007).10D. B. Kim, B. Gweon, S. Y. Moon, and W. Choe, Thin Solid Films, 518,
6700 (2010).11M. Goldman, A. Goldman, and R. S. Sigmond, Pure Appl. Chem. 57, 1353
(1985).12D. B. Kim, H. Jung, B. Gweon, and W. Choe, Phys. Plasma 17, 073503
(2010).13C. Tendero, C. Tixier, P. Tristant, J. Desmaison, and P. Leprince, Spectro-
chim. Acta, Part B 61, 2 (2006).14Y. P. Raizer, M. N. Shneider, and N. A. Yatsenko, Raido-freqeuncy
Capacitive Discharges (CRC, Boca Raton, 1995), p. 3.15Y. Sakiyama and D. B. Graves, J. Phys. D: Appl. Phys. 39, 3451 (2006).16H. Ueno, K. Hata, and H. Nakayama, Jpn. J. Appl. Phys. 46, 1142 (2007).17D. W. Liu, F. Iza, and M. G. Kong, Appl. Phys. Lett. 95, 031501 (2009).18N. K. Bibinov, A. A. Fateev, and K. Wiesemann, J. Phys. D: Appl. Phys.
34, 1819 (2001).19E. Amanatides and D. Mataras, J. Appl. Phys. 89, 1556 (2001).20M. Radmiloviæ-Radjenoviæ and J. K. Lee, Phys. Plasmas 12, 063501
(2005).21S. Y. Moon and W. Choe, Spectrochim. Acta, Part B 58, 249 (2003).22S. J. You, S. K. Ahn, and H. Y. Chang, Appl. Phys. Lett. 89, 171502
(2006).23D. B. Kim, S. Y. Moon, H. Jung, B. Gweon, and W. Choe, Phys. Plasmas
17, 053508 (2010).
043503-5 The driving frequency effects … Phys. Plasmas 18, 043503 (2011)
Downloaded 26 Apr 2011 to 210.98.16.31. Redistribution subject to AIP license or copyright; see http://pop.aip.org/about/rights_and_permissions
top related