22nd International Symposium on Plasma Chemistry
July 5-10, 2015; Antwerp, Belgium
The measurement of gas temperature in nanosecond pulsed discharge in liquids
via optical emission spectroscopy
Y. Seepersad1,2, A. Fridman1 and D. Dobrynin1
1
2
A.J. Drexel Plasma Institute, 08103 Camden, NJ, U.S.A.
Department of Electrical and Computer Engineering, Drexel University, 19104 Philadelphia, PA, U.S.A.
Abstract: We present the results of optical emission spectroscopy studies performed on
nanosecond pulsed discharges in liquids. This regime of microplasmas have been shown
previously to be formed directly in the liquid phase (no bubbles). The measurement of
temperature is an important plasma parameter needed to qualify the applicability of these
plasmas for applications purposes. We argue the non-thermal nature of the discharge and
present results of Boltzmann analysis of OH(A2 Σ − X 2 Π) rotational lines, its applicability
to the current regime, and challenges to the interpretation of these measurements
accurately.
Keywords: spectroscopy, liquids, diagnostics, microplasmas, non-equilibrium
1. Introduction
The possibility of forming plasma directly in the liquid
phase without an initial bubble nucleation phase has been
reported in previous works [1-4]. This unique plasma
regime has the potential to be utilized in applications
where very high densities of plasma and chemical species
are desirable without liquid vaporization associated with
pre-initiation bubbles, for instance as reported in [5-7].
Perhaps the most important parameter needed when
considering these micro-discharges for sensitive
applications are the associated temperatures – electron
temperature which will influence chemical activity, and
gas temperature which will determine the extent to which
the bulk liquid heats up. Certainly, in the context of recent
developments of plasmas in liquids employed in biology,
medicine, and material synthesis [8], a non-thermal
plasma would attract much attention.
The electron temperature of this regime of plasma in
liquids was previously estimated from optical emission
spectroscopy reported in [1] and found to be ~3eV. These
results were obtained through deconvolution of the Van
der Waal’s and Stark contributions of the measured Hα
and O I lines, based on integrated emission spectra. Time
resolved spectroscopy would later reveal that the emission
due to a single nanosecond pulse (direct liquid ignition) is
dominated by strong broadband emission with no
measureable atomic transitions. The integrated spectra
reported in [1] contained emission from phases associated
with voltage ringing (multiple reflections) on the HV line.
The measurement of gas temperature is even more
elusive, as the traditional low-budget techniques typically
rely on the assumption of thermal equilibrium between
the rotational temperature of ground-state molecules and
the gas molecules [9]. This is a problem for plasma in
liquids, especially water, where collisional quenching
significantly hinders equilibration of rotational states, and,
P-I-3-24
multiple distributions exist due to several mechanisms of
formation of the rotationally excited molecule OH [10].
In this work, we present the results of optical emission
spectroscopy performed on nanosecond pulsed discharges
in water. Both time integrated and time resolved spectra
were analysed, and a high-resolution analysis of the
OH(A-X) band was performed. Typical analytical
techniques were applied based on gross assumptions.
2. Experimental Setup
The plasma reactor setup is identical to that reported in
[3], except the anode (pin electrode) was an Iridium probe
with ~5nm radius of curvature tip. Discharge was ignited
in distilled, deionized water between submerged
electrodes in a 3mm gap and the pulse frequency was
limited to 10Hz to allow the liquid to recover between
subsequent pulses. The problem of voltage ringing
(reflections) due impedance mismatch of the load and the
HV cable still existed and contributed to re-illuminations
and liquid vaporization. High voltage pulses with +24kV
amplitude was delivered to the plasma reactor via 100ft
of coaxial cable, with a return current shunt used for
signal triggering and pulse monitoring. We will refer to
the ‘first pulse’ as the leading voltage pulse associated
with a single trigger of the power supply. Thus, in 1
second there will be 10 ‘first pulses’, each with a series of
reflection transients lasting < 4µs.
Spectra were collected using a fibre-optic cable with
200µm core, coupled to a Princeton Instruments SP-500i
monochromator system with two gratings: 750l/mm and
1800l/mm. A PI: Max 1K Gen II detector was used, with
a spectral response from 250-900nm and a detector pixel
width of 13µm. The instrument function of the entire
system was measured to be ~0.05nm FWHM for highresolution spectral measurements, with an entrance slit to
the monochromator set to 40µm.
1
Corrected Absolute Intensity
1
0.9
Corrected Intensity
(Normalized)
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
300
Intensity a.u.
1
450
500
550
600
700
650
Original Spectrum
Background Average
Background Fit
0.8
0.6
0.4
0.2
302
304
306
308
310
312
314
316
318
316
318
Data with background subtraction
1
0.5
0
-0.5
302
304
306
308
310
312
Wavelength λ (nm)
314
Fig. 3. OH Emission during first phase, before (top) and
after (bottom) subtraction of continuum baseline.
1
3us Exposure - Integrated over Reflections Only
Measured Intensity
600
0.8
500
0.7
400
0.6
Intensity
Measured Intensity a.u.
400
Fig. 2. Emission from first incident high voltage pulse
only. Exposure time 50 ns.
0.9
0.5
0.4
300
200
0.3
100
0.2
0
0.1
0
200
300
400
500
600
700
800
900
1000
Wavelength λ (nm)
Fig. 1. Time-integrated spectrum for dis discharge in
water. No correction made for detector sensitivity to
wavelength.
2
350
Wavelength λ (nm)
Intensity a.u.
3. Results
The measured integrated spectrum is shown in Fig. 1.
The exposure time was 2s and emission is an
accumulation of light from multiple re-illuminations per
applied voltage pulse (due to reflections), while, the
spectrum corresponding to emission form only the first
incident pulse is shown in Fig. 2. This emission profile
was also corrected for wavelength sensitivity of the CCD
detector in Fig. 2. Comparing the two spectra, one
observes that in the first 50ns there is no emission from
atomic hydrogen (Hα ) or oxygen (OI), or from rovibrationally excited OH. Time resolved studies of the
development of the spectrum show that these distinct
spectral features develop in the emission from reflected
pulses. In these experiments, these correspond to restrikes occurring in gas channels remnant of the plasma
from the first pulse, consequently associated with
different initiation and propagation mechanisms.
The emission from the OH band near 308nm is shown
in Fig. 3. Much of the structure associated with rotational
temperatures greater than 600K is absent, though a few
band heads could be discerned. In contrast, the OH
emission activity is clearly present after the initial phase
as shown in Fig. 4 with the R 2 band well developed.
Boltzmann analysis of the distribution is in progress for
the R 2 , P1 , Q1 , Q 2 lines as performed in [11]. Conclusions
based on this analysis however will only be valid for
plasma formed in gas phases channels are described
above.
Analysis of the FWHM of the measured Hα line was
performed previously in [1], and in this work, a timeresolved study of the development of this emission line
was performed. The results match those reported by
Marinov et al. at a previous meeting. Analysis of the
evolution of electron temperature and density as related to
the FWHM of this line is also in progress.
-100
300
305
310
wavelength λ nm
315
Fig. 4. Rotationally excited OH(A-X) bands occurring in
plasma formed during reflected pulses.
P-I-3-24
4. Discussion
The wideband spectrum recorded for this regime of
plasma in liquids is typical of what is observed with laser
induced breakdown plasma [12-14]. Careful analysis
shows that the emission does not fit Planck’s equation and
cannot be attributed to blackbody radiation. As discussed
in [15], this intense continuum radiation is most likely
due to radiative recombination and Bremsstrahlung of
high energy electrons. Both of these mechanisms are
plausible given the high density (n0 ~1022 cm−3 ) of
molecules in water, assuming initiation directly in the
liquid phase. This result indicates the existence of a large
population of high energy electrons in the initial stages of
the plasma development, a theory supported by the
electrostriction based initiation hypotheses discussed in
[2, 16, 17]. The absence of atomic hydrogen lines in the
early stages could be due to several reasons, including the
emission lines being dominated by the continuum
background and the plasma being optically thick. Nondissociative recombination of H2 O+ (or H3 O+ ) ions could
lead to the formation of non-Boltzmann distributed OH
radicals, as well as atomic hydrogen and oxygen
according to [10], though perhaps the rotationaltranslational equilibrium might be true for low rotational
quantum numbers. Boltzmann analysis of the OH
rotational structure is currently underway to determine the
deviation from equilibrium and the various temperatures
associated with each rotational distribution.
The measurement of gas temperature from the optical
emission spectroscopy of these types of discharges
requires many assumptions, the validity of which needs to
be properly assessed. Firstly, the optical thickness of the
plasma in the early stages needs to checked. Furthermore,
typical methods which are based on the measurement of
rotationally excited OH requires equilibrium with the gas
molecules, as well as thermalization of the OH states. The
significant effect of collisional quenching by water
vapour on the rotational energy transfer for discharges in
gases will be amplified for discharges directly in the
liquid phase. The increased density however, might allow
efficient rotational energy transfer and establishment of
equilibrium with the gas molecules. These factors are
currently being scrutinized.
The existence of the dark phase throughout the
emission lifetime from a single pulse supports that the
plasma itself is non-thermal. One can draw this
conclusion from comparison to laser induced breakdown
where the plasma development and ablation is sustained
by gas temperature[18]. Shadowgraphic studies
performed previously showed post-plasma gas channels
develop only along the filament channels and so is
probably due to dissociative recombination of water ions
and electrons into hydrogen gas.
5. Conclusions
Nanosecond pulsed discharges in liquids have been
shown to develop directly in the liquid phase without
bubbles. The temperature associated with these discharges
P-I-3-24
is difficult to measure, but can be argued as non-thermal.
Boltzmann analysis of OH emission spectrum will reveal
some information about how far the rotational distribution
is from equilibrium, however the bigger challenge is in
determining
whether
the
rotational-translational
equilibrium can be attained in this environment. Electron
temperatures are likely to be very high, and have been
estimated to be ~3eV previously, but these results cannot
be interpreted as the electron temperature associated with
the direct liquid mode.
We remark finally that the interesting aspect to be
investigated in this work lies in the study of the plasma
formed due to a single nanosecond pulse (without
reflections). From what was observed so far, continuum
emission dominates this initial stage and temperature
characterization from traditional OES techniques require
many assumptions that may not be valid. Active
diagnostics such as LIF might offer a better chance at
quantifying the gas temperature in this regime.
6. References
[1] D. Dobrynin, et al. Non-equilibrium nanosecondpulsed plasma generation in the liquid phase (water,
PDMS) without bubbles: fast imaging, spectroscopy
and leader-type model. J. Phys. D: Appl. Phys., 46,
105201 (2013)
[2] M. Pekker, et al. Initiation stage of nanosecond
breakdown in liquid. J. Phys. D: Appl. Phys., 47,
025502 (2014)
[3] Y. Seepersad, et al. Invetigation of positive and
negative modes of nanosecond pulsed discharge in
water and electrostriction model of initiation.
J. Phys. D: Appl. Phys., 46, 35 (2013)
[4] A. Starikovskiy. Pulsed nanosecond discharge
development in liquids with various dielectric
permittivity constants.
Plasma Sources Sci.
Technol., 22, 012001 (2013)
[5] P. Bruggeman and C. Leys. Non-thermal plasmas
in and in contact with liquids. J. Phys. D: Appl.
Phys., 42, 053001 (2009)
[6] J.F. Kolb, et al. Streamers in water and other
dielectric liquids. J. Phys. D: Appl. Phys., 41,
234007 (2008)
[7] P. Ceccato. Filamentary Plasma discharger inside
water: initiation and propagation of a plasma in a
dense medium. in: LPP Laboratory. (Palaiseau,
France: Ecole Polytechnique) 203 (2009)
[8] W.G. Graham and K.R. Stalder. Plasmas in liquids
and some of their applications in nanoscience.
J. Phys. D: Appl. Phys., 44, 174037 (2011)
[9] P.J. Bruggeman, et al.
Gas temperature
determination from rotational lines in nonequilibrium plasmas: a review. Plasma Sources Sci.
Technol., 23, 023001 (2014)
[10] P. Bruggeman, et al. Is the Rotational Temperature
of OH(A-X) for Discharges in and in Contact with
Liquids a Good Diagnostic for Determining the Gas
3
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
4
Temperature?
Plasma Process. Polymers, 6,
751-762 (2009)
P. Bruggeman, et al. Characterization of a direct
dc-excited discharge in water by optical emission
spectroscopy. Plasma Sources Sci. Technol., 18,
025017 (2009)
A. De Giacomo, et al. The role of continuum
radiation in laser induced plasma spectroscopy.
Spectrochim. Acta B, 65, 385-394 (2010)
S.S. Harilal, et al. Dynamics of femtosecond laser
produced tungsten nanoparticle plumes. J. Appl.
Phys., 114, 203302 (2013)
V. Hohreiter, J.E. Carranza and D.W. Hahn.
Temporal analysis of laser-induced plasma
properties as related to laser-induced breakdown
spectroscopy. Spectrochim. Acta B, 59, 327-333
(2004)
T. Sakka, et al. Laser ablation at solid–liquid
interfaces: An approach from optical emission
spectra. J. Chem. Phys., 112, 8645 (2000)
M.N. Shneider, M. Pekker and A. Fridman.
Theoretical Study of the Initial Stage of Subnanosecond Pulsed Breakdown in Liquid
Dielectrics. IEEE Trans. Dielectrics Electrical
Insulation, 19, 4 (2012)
Y. Seepersad, et al.
On the electrostrictive
mechanism of nanosecond-pulsed breakdown in
liquid phase. J. Phys. D: Appl. Phys., 46, 162001
(2013)
F. Anabitarte, A. Cobo and J.M. Lopez-Higuera.
Laser-Induced
Breakdown
Spectroscopy:
Fundamentals, Applications, and Challenges. ISRN
Spectroscopy, 2012, 1-12 (2012)
P-I-3-24