Open Access
Add to BibSonomyAbstract
Dynamic processes in a gliding arc plasma generated between two diverging electrodes in ambient air driven by 31.25 kHz AC voltage were investigated using spatially and temporally resolved optical techniques. The life cycles of the gliding arc were tracked in fast movies using a high-speed camera with framing rates of tens to hundreds of kHz, showing details of ignition, motion, pulsation, short-cutting, and extinction of the plasma column. The ignition of a new discharge occurs before the extinction of the previous discharge. The developed, moving plasma column often short-cuts its current path triggered by Townsend breakdown between the two legs of the gliding arc. The emission from the plasma column is shown to pulsate at a frequency of 62.5 kHz, i.e., twice the frequency of the AC power supply. Optical emission spectra of the plasma radiation show the presence of excited N2, NO and OH radicals generated in the plasma and the dependence of their relative intensities on both the distance relative to the electrodes and the phase of the driving AC power. Planar laser-induced fluorescence of the ground-state OH radicals shows high intensity outside the plasma column rather than in the center suggesting that ground-state OH is not formed in the plasma column but in its vicinity.
©2013 Optical Society of America
1. Introduction
Atmospheric pressure non-equilibrium plasma has found many applications in different fields such as surface treatment [1–3], combustion enhancement [4–9], bacterial inactivation [10–14] and pollutant reduction [15–17]. Some other applications are reviewed in references [18–21]. As a non-equilibrium plasma generator, a gliding arc is technically of particular interest since it can be used to produce, with high efficiency, large plasma volumes at highly non-equilibrium conditions characterized by hot electrons and colder ions [18, 22, 23]. This opens up chemical pathways not available under equilibrium conditions [19]. A typical gliding arc is a plasma column that extends between two diverging electrodes in a turbulent gas flow. When applying high electric power in a gas flow, the generated gliding arc is visualized as a blurred glowing region, as shown in Fig. 1 . The gliding arc reveals its true instantaneous structure in photographs with short exposure time as the ones shown in Fig. 2 . The gliding arc is a thin, string-like plasma column that repeatedly ignites in the narrowest gap between the diverging electrodes, glides up along the electrodes, and extinguishes. The large-scale motion of the gliding arc is caused by convection in a turbulent free jet that stretches the cold plasma column to form an arc connecting the anchor points on the electrodes. The length of the plasma column increases until its extinction unless the plasma column short-cuts a long current path with a shorter one. Figure 2 shows an ignition event of a new plasma column between the electrodes. After the ignition of a new plasma column, the previous plasma column decays. At the ignition stage the plasma is hot but the ions are quickly cooled, creating non-equilibrium conditions that are attractive for industrial applications, as referred above.
Fig. 1 Photograph of an ambient-air gliding arc discharge taken with a digital camera (Cannon 350D) using an exposure time of 33 ms.
Fig. 2 Photographs of an ambient-air gliding arc discharge taken every 50 μs. The exposure time of the digital camera was 13.9 μs. A new plasma column ignites in the third frame after which the previous discharge extinguishes and the optical emissions decay.
In spite of the widely established applications, detailed mechanisms of plasma treatment especially with gliding arcs at atmospheric pressure are still unclear. Knowledge of the freely developing gliding arc discharge is a pre-requisite to understand more complicated practical situations, such as surface treatment where the gliding arc interacts with a surface. Free-developing gliding arcs, therefore, have been investigated in modeling [19, 22, 24–29] and in experiments [22, 24, 26–32]. Existing measurements of gliding arcs were usually done at temporal resolution that did not allow the fastest dynamic processes, such as ignition and small-scale movement, to be tracked. The temporal resolution of spontaneous emission measurements of species concentrations was typically longer than the period of the supplied voltage and hence dependences on the phase of the AC drive could not be demonstrated.
Recent advances in optical diagnostic make it possible to do spectrally resolved measurements of the gliding arc discharge with unprecedented temporal and spatial resolution even simultaneously. In the present work, we seek an understanding of the dynamics and chemistry of a gliding arc discharge in ambient air, and this work is focused on the dynamics of the gliding arc discharge on time scales accessible with the new optical diagnostic equipment. High speed photography reveals the dynamic processes and tracks the position of the gliding arc in time including ignition, short-cutting events triggered by Townsend breakdown, and extinction. The framing rate used in this work is up to 420 kHz giving a temporal resolution up to 2.1 μs, which essentially freezes the motion of the plasma column and resolves the changes of the emission intensities originating from the alternating current (AC) driving power varying at 31.25 kHz (i.e., the half cycle duration is 16 µs). High-speed Schlieren photography is employed to visualize the flow fields and the temperature gradient generated by the gliding arc. Optical emission spectroscopy is utilized to characterize the chemical composition of the plasma column, demonstrating the spatial distribution of the dominant radiating species OH, NO and N2 and the time-dependent emission during the AC voltage cycles. Instantaneous distributions of ground-state OH radials generated in the plasma column are visualized using planar laser-induced fluorescence (PLIF).
The setup of the gliding arc and the optical diagnostics techniques are described in section 2, measurement results are reported in section 3 and a summary and concluding remarks are given in section 4.
2. Experimental setup
2.1 Gliding arc discharge setup
The gliding arc plasma is generated between two diverging electrodes constructed with stainless steel tubes with an outer diameter of 3 mm that are cooled internally with water. A similar configuration is presented in detail in [3, 33]. Air flow was fed in through a hole with 3 mm diameter between the electrodes (see Fig. 1) to convect the plasma column and create the gliding motion. The air flow rate can be adjusted, but in this work it was fixed at 17.5 standard liters per minute (SLM). The air velocity was calculated to be about 40 m/s at the nozzle exit providing turbulent flows with a Reynolds number of about Re = 8000. The plasma discharge was driven by an AC power supply at a frequency of 31.25 kHz (Generator 6030, SOFTAL Electronic GmbH, Germany). The electrical circuit driving the gliding arc plasma is presented in Fig. 3 . The input power to the gliding arc was approximately 800 W. The average input power was monitored by measuring the voltage and current with a high-voltage probe and a resistor, respectively. Typical waveforms of the voltage and the current are presented in Fig. 4 . Generally the current follows the voltage, but spikes in the current were occasionally observed. The gliding arc plasma was placed in a shielding metal box in order to reduce the electromagnetic noise radiated from the discharge and for safety reasons. The optical access was provided by three windows, one for the laser beam, one for the emission to the cameras or spectrometers, and the third for the Schlieren setup.
Fig. 3 The electrical scheme and the timing-controller system for the gliding arc discharge. HV, high-voltage power supply; SG, signal generator; DG, time delay generator and ICCD, intensified CCD camera. A digital oscilloscope was employed to monitor and record the current and voltage wave forms, the time-gating of the ICCD, laser pulses and the trigger signal from the SG.
The gliding arc investigated in this work can be run in two modes, i.e., the burst mode and the continuous mode. In most practical applications, the gliding arc was driven by a continuous AC voltage. However, in this mode it is difficult to synchronize the data acquisition with the plasma ignition events since they occur at seemingly random times. By programming the signal generator, the AC high voltage can be supplied to the electrodes with a predefined duration and repetition rates, like a burst. The gliding arc can be ignited and fully developed in each burst, provided the burst has a sufficiently long duration and the break time between consecutive bursts is enough to allow the plasma to extinguish fully. This arrangement gives the possibility to synchronize the optical detection system not only to the selected phase of the AC driving power but also to the ignition process, so that the exact location of the plasma column during the time of data acquisition can be predicted approximately. In the burst mode, the data acquisition is triggered after the delay time TD + Td (see Fig. 5 ). TD is counted by the number of full periods of the 31.25 kHz AC voltage, which can be used to predict the position of the plasma column downstream from the ignition point. An additional delay time, Td, is added to lock the phase of the AC voltage for data acquisition. The timing relation between the gliding arc plasma and the detection system (TD and Td) is controlled by using a time delay generator (DG). The exposure time of the intensified CCD (ICDD) camera is Tg. The duration of the plasma discharge burst duration Tp was set to 20 ms. The timing relations are sketched in Fig. 5, and the timing relation between plasma and ICCD gate was monitored by a 200 MHz digital oscilloscope.
Fig. 5 Electronic waveform and timing notations. In the burst mode, the gliding arc was driven by a 10 Hz repetitive burst of 31.25 kHz AC powder with a burst duration Tp of 20 ms. TD: delay time to the desired half cycle in which data ass taken; Td: the phase delay time of data acquisition; Tg: the ICCD exposure or gate time. In the case of continuous mode, Td and Tg keep the same meaning as in the burst mode.
2.2. Optical measurements
2.2.1 High-speed photography
High-speed photography has been applied to investigate gliding arc throughout recent years [22, 29, 32, 34, 35]. Usually the data acquisition was not synchronized with the plasma time-control in those works. Moreover, the relatively low framing rate adopted in the previous results was not sufficient to fully reveal the details propagation of the gliding arc plasma column or the emission evolution in an electrical cycle. In the present work, a high-speed CMOS camera (Photron Fastcam SA5) was used to capture the ignition, propagation and extinction of the gliding arc. The high-speed camera was operated at 20, 50 and 420 kHz in measurements to meet different requirements. The camera was equipped with a Nikon lens (f = 50 mm, F/N = 1.2) collecting radiation from 380 nm to 980 nm. A high-speed movie showing the dynamics of the gliding arc discharge is attached in Media 1 as supplementary material. Events captured in this movie will be discussed in section 3.1. Beside of direct view, Schlieren visualization of the propagating gliding arc was also captured using the same high-speed camera. Schlieren visualization is based on the deflection of light by a refractive index gradient. The plasma column has a higher temperature than the surroundings. Hence the flow field and the plasma column can be visualized by the Schlieren technique. An example Schlieren movie is shown in supplementary as Media 2.
2.2.2 Optical emission spectroscopy
Optical emission spectroscopy has been commonly used to find dominant excited species in gliding arcs [7, 25, 29, 30, 36, 37] and other plasmas [38]. Two different setups have been adopted in this work to monitor the optical emission. A fiber coupled spectrometer (OceanOptics, USB2000) was employed to collect the emission from the gliding arc plasma at different heights. A collimating lens (diameter = 5 mm) was installed before the fiber to enhance the signal collection. In the measurements, the gliding arc was imaged on the collimating lens which was adjusted vertically to achieve space-resolved measurements. For high temporal and spatial resolution, an ICCD based imaging spectrometer (SP-2300i) was employed to catch the plasma emission during its evolution in time and space. In this second setup, the input slit of the spectrometer was set parallel to the direction of the gas flow, and the spontaneous emission from plasma column can be time-resolved in a half cycle (16 μs). The position of the plasma column during the data acquisition could also be controlled as discussed in section 2.1. A 150 groves/mm grating was installed in the spectrometer and a spectral range of 170 nm was imaged by the ICCD camera. A Deuterium lamp was used to calibrate the spectral response of the system in the range of 220 – 390 nm.
2.2.3 Planar laser-induced fluorescence
Planar laser-induced fluorescence (PLIF) was employed to investigate the distribution of the ground-state OH radicals. PLIF is a common technique for OH detection in gas-phase environments like in combustion [39] and in plasma [7, 9, 40]. The laser beam was generally shaped into a thin laser sheet to enable a two-dimensional measurement of OH with high resolution in time and space. In this work, the Q2(5) line of the OH X 2П – A 2Σ+ (0, 1) band at 283.268 nm was excited using the second harmonic radiation from a dye laser (Continuum ND60, with Rhodamine 590 dye) pumped by a pulsed Nd:YAG (Brilliant B) laser. The fluorescence emission from the X 2П – A 2Σ+ (0, 0) and (1, 1) bands around 308 nm was collected by an ICCD camera through a filter combination (UG5 + WG295) that has a band-pass of 290 – 405 nm. The laser energy was around 8 mJ/pulse. The ICCD camera was synchronized with the pumping laser, and the fluorescence was collected perpendicular to the laser sheet with an exposure time of 30 ns.
The plasma column was also directly imaged using the ICCD camera through filters (UG5 + WG295) in order to investigate the spatial distribution of the plasma column. In this case, an exposure time of 2 μs was adopted. The OH-PLIF signal and the spontaneous plasma emission can be simultaneously collected using the ICCD with an exposure time of 2 μs.
3. Results
3.1 High-speed photography of the gliding arc
The motion of the gliding arc plasma was recorded with the high-speed camera and the gliding motion of the plasma column was captured at high temporal resolution to reveal detailed information on ignition, evolution, short-cutting, pulsation and extinction. Media 1 is a video image of the gliding arc plasma at 17.5 SLM air flow. The frame rate of the high-speed camera was 50,000 frames/s, and the exposure time was 16.3 μs. This movie covers 21 ms in real time, during this period the phenomena of an ignition of the gliding arc, the evolution in the convection flow and especially a short-cut of the plasma column can be observed. Note that the images shown in Media 1 are line-of-sight signals. Phenomenologically, the following features can be observed:
- 1. At t = 1.74 ms, a new ignition is observed in the smallest gap between the two electrodes, where strong radiation is emitted. The emission from the previous plasma column is still visible for a short time (~500 μs). The plasma column is attached to both electrodes (bright points) and is convected in the gas flow while the anchor points of the plasma column glide up along the electrodes. Before t = 10 ms, the plasma column short-cuts several times leading to a sudden decrease of the plasma column length. Similar short-cut events were observed in [31]. At about 10 ms, the anchor points of plasma column become stationary on the electrodes (also see the photograph in Fig. 2). After t = 10 ms, the length of the plasma column increases while the anchor points remain approximately at the same locations.
- 2. At t = 12.64, 13.94 and 19.58 ms, short-cutting of the plasma column happens when suddenly a new, shorter current path forms after Townsend breakdown. Townsend breakdown is described for example in [41]. In every short-cutting process, before the real current path is formed, for example at t = 13.94 ms, a Townsend discharge is already evident at t = 13.68 ms.
- 3. When the plasma column is above the electrodes, there is a blurred emission signal around the plasma column. These emitting species help to form a new current path and then induce a short-cut event. (Especially see from t = 18 ms to 19.6 ms).
An example of the short-cut events captured in the fast movies is shown in Fig. 6 . As shown in the frame for t = 50 μs, the short-cut current path (indicated by the red arrow) exists simultaneously with the previous, longer current path. After the new current path is formed, emission is still observed (in the dashed box). Its intensity slowly decays with time, as shown by the curve in Fig. 6. The decay time scale is on the order of 500 μs. The brightness of the plasma column seems homogeneous apart from strongly emitting regions appearing when the plasma column is oriented in the direction along the line-of-sight. It is notable that there is a blurred signal around the thin bright plasma column visible in Fig. 6.
Fig. 6 An example of a short-cut event recorded at 20 kHz framing rate using an exposure time of 13.9 μs. The short-cut current path is indicated by the arrow in the frame of t = 50 μs where a Townsend breakdown occurs between the two legs of the plasma column, and a new current path forms. The integrated emission intensities of the plasma column (the part in the dashed box) are shown as a function of time to the right. A 50 kHz movie revealing the plasma evolution can be found in the supplements (Media 1).
Figure 7 shows the projected length of the plasma column in 10,000 continuous frames representing 200 ms recoding time. The number of CCD pixels illuminated by the plasma column was used to track the projected length of the plasma column. The shortest possible length of the plasma column is the gap width between the electrodes. This length is attained only at re-ignition which occurs three times in Fig. 7. The projected length of the plasma column generally increases with time but there are sudden drops at short-cutting events. In contrast to the re-ignition at the narrowest gap, the short-cutting events happen about 30 times. In other words, the period of new ignition was around tens to hundreds of milliseconds, while that of short-cutting events is less than 10 ms. The rates of new ignition and short-cutting events are strongly dependent on air flow rate, which will be reported elsewhere. It may be necessary to model short-cutting events if the ignition frequency of a gliding arc is to be predicted (e.g [42].).
Fig. 7 The projected length of the plasma column in 10,000 continuously recorded frames. New ignition happened in tens to hundreds of ms while the time between two subsequent short-cutting events (only part of which were marked by the red arrow) was always less than 10 ms.
Figure 8 shows the integrated signal intensity (for wavelengths above 380 nm) as a function of the projected lengths reported in Fig. 7. The blue bar B denotes the shortest lengths of the plasma columns anchored at the top stationary points of the electrodes. The shorter lengths, i.e. to the left of the blue bar B, belong to plasma columns that glide up the electrodes from the ignition points to the stationary anchor points on the electrodes. The averaged emission intensity per unit length (proportional to the slope of every point with respect to the origin) varies with the length of the plasma column, and it decreases from K1 to K2 when the plasma column length increases. The emission per unit length shortly after the ignition (K1) was much stronger than that when the plasma column was well developed in non-equilibrium conditions. After the anchor points of the plasma column stabilized on the top of the electrodes, the unit length emission decreased much slower as the length of the plasma column increased compared with the region before bar B.
Fig. 8 The relation between integrated emission signal intensity and the projected length of the plasma column. Before the bar B, the anchor points of the plasma column were gliding up along the electrodes, and after that the anchor points were stationary on the top of the electrodes.
Using a framing rate of 420 kHz and an exposure time of 2.1 μs, the emission intensity variance in a half cycle of the 31.25 kHz power supply was temporally resolved. Figure 9(a) shows sequential images of part of the plasma column. At the framing rate of 420 kHz, the camera observation region is limited to 64 × 128 pixels. Therefore only part of the plasma column was recorded. The emission intensity varied with time, i.e., with the supplied voltage within a half cycle while the shape of the plasma column was almost frozen during such a short period. The emission intensity of every frame as a function of time is shown in Fig. 9(b), where one can see a periodic change of the emission intensity with a quick increasing slope and a slower decaying slope. The intensity does not decay to zero, meaning the plasma column does always emit radiation. A Fourier transform of the emission intensity varied with time for 17,000 frames indicating that the dominant frequency was 62.5 kHz as the frequency of supplied power, i.e., twice the voltage driving frequency at 31.25 kHz. Since the high framing speed camera was not synchronized with the gliding arc timing system, the phase relation between the emission intensity and the supplied voltage was not measured.
Fig. 9 (a) Sequential images of part of the plasma column. (b) The integrated emission intensity varying with time. (c) Fast Fourier transform for the curve in (b) for 17,000 frames.
The spatial location of the plasma column was investigated by statistically analyzing 17,000 images taken with 20 kHz framing rate. The plasma column position for every frame was obtained through image binary-conversion, and then the probability of spatial distribution was found by averaging all the frames, as shown in Fig. 10(a) . The distribution probability is large near the top of the electrodes. The anchor points of the gliding arc on the electrodes move up to these positions and then stay there until the gliding arc extinguishes. Shown in Fig. 10(b) are examples of plasma column reaching this state in six consequent frames. The low part of the plasma column near the anchor points is rather stationary while the top part moves with the flow, which was also observed in the Schlieren movie, i.e., Media 2. Since sequential images of the plasma column were captured, projected velocities can be estimated. The velocity in the imaging plane was around 8 m/s for the top part of the plasma column, indicated by the red arrows, which is consistent with the nozzle exit velocity of about 40 m/s as the jet velocity has slowed down to such magnitudes at this location.
Fig. 10 (a) Distribution possibility of plasma column at 17.5 SLM air flow. (b) The estimated velocity of the plasma column using high frame rate images. The scaling velocity of 20 m/s is indicated showing that the upper part of the plasma column moves at about 8 m/s. High-speed Schlieren movie can be found in the supplements (Media 2).
High-speed Schlieren photography was intended for visualization of the flow fields and the temperature gradient generated by the gliding arc. The result is shown as Media 2. Compared with the thin and clearly structured filament in direct-view photographs (as shown in Media 1 and Fig. 6), Schlieren photography did not show any sharp structure of the gliding arc column. This indicates that the ions are not much hotter than the surrounding air leading to small temperature gradients. The gas flow along the central axis moves fast while the outside part is rather stationary, which is consistent with the results shown in Fig. 10(b).
3.2 Optical emission spectroscopy
Spatially resolved emission spectra of the gliding arc running in continuous mode are shown in Fig. 11 . The spectra were recorded by imaging different parts of the plasma plum to the input fiber tip of the spectrometer. The exposure time was 50 ms. The spectra at two different heights, a and b, are shown at the top of Fig. 11. Below H = 4.5 cm (the height of electrodes as shown in Fig. 1) N2* rotational bands dominate the spectra, while above this position OH* rotational bands dominate and decrease at larger heights. NO emission exists at all heights, and the maximum intensity appears at around H = 5 cm. The OH bands belonging to the X 2П – A 2Σ+ transition at around 308 nm are partially overlapped with some N2* bands, which hinders the possibility of imaging the OH* spontaneous emission with spectral filters. Note that a broadband background appears above the electrodes. Two spectral structures at 592 nm and 780.5 nm were not identified, but will be investigated.
Fig. 11 Optical emission spectra of the gliding arc plasma recorded at different heights using a fiber-coupled spectrometer. The spectra at heights (a) and (b) are shown in curves to visualize the corresponding species.
Time-resolved optical emission spectra were recorded using the ICCD based imaging spectrometer. In this measurement, the gliding arc plasma was operated in the burst mode (10 Hz and Tp = 20 ms). The delay time between the gliding arc and the ICCD gate was adjusted using the time delay generator (for the cycle delay time TD) and the delay function in the ICCD software (for the phase delay time Td). The exposure time Tg of the ICCD was 5.3 μs enabling the resolve of a half period of the AC power. Figure 12 shows the emission spectra averaged over 100 frames at two different height ranges. The signals imaged on the ICCD were binned along the direction of the entrance slit. Therefore the two spectra shown in Fig. 12 were actually accumulated in the ranges of 0.4 – 1.8 cm (the middle height H = 1.1 cm) and 4.8 – 6.2 cm (the middle height H = 5.5 cm), respectively.
Fig. 12 Phase-resolved optical emission spectra of the plasma column at height ranges (a) 4.8 – 6.2 cm and (b) 0.4 −1.8 cm with the center at H = 5.5 cm and 1.1 cm, respectively. The gliding arc was operated in burst mode at 10 Hz. The camera exposure time was 5.3 μs and three Td values (0, 5, and 10 μs) were adopted, i.e., the emission in a half cycle was temporally resolved in a resolution of 5.3 μs. The spectra shown were averaged over 100 frames, and the intensity had been modified by the instrumental spectral response. Note the different intensity scales adopted in (a) and (b).
Figure 12 verifies the conclusion that emissions from NO*, OH* and N2* vary with heights. N2* bands dominate the emission source at lower positions (below the electrode top) while NO* and OH emissions dominate the positions above the electrodes. Besides the variation in space, the emission also varies within the voltage half cycle. In the newly ignited plasma column (TD = 0.144 ms, H = 1.1 cm), OH* emission intensity decreased slightly with time as shown in the Fig. 12(b). Note that the magnitude and form of the line intensity variation depend on the particular band. The band emissions above 350 nm kept constant from Td = 0 μs to Td = 5 μs and decreased significantly from Td = 5 μs to Td = 10 μs. The band around 335 nm, however, acted in the opposite way. It might be explained by the changes of vibrational temperature of N2 in a half cycle. The spectral structure below 290 nm in Fig. 12(b) could not be clearly assigned to any species. Nevertheless, two band peaks of NO* around 225 and 235 nm were identified.
When the plasma column moved up to H = 5.5 cm (above the electrodes) at TD = 4.992 ms, the OH and NO spectral structures were clearly identified besides that of N2*. At this height, within half cycles the N2* emission decreased with time while the emissions from NO* and OH* increased first and then decreased. The spectral structure of NO* did not vary with time indicating that the ambient temperature did not change much over the half cycles at this height where the plasma column is in a non-equilibrium condition.
The above results of phase-resolved emissions indicate that the characteristics of the plasma column vary within a half cycle, i.e., they follow the supplied AC voltage. Also the variance of the plasma column characteristics within every half cycle depends on the height of plasma column. Lastly, we note that these dependencies also greatly depend on the flow rate of air, which will be reported elsewhere.
3.3 PLIF of ground-state OH radicals
The ground-state OH radicals were measured using PLIF. The laser beam sheet was sent vertically through the two electrodes containing the central line of the jet and crossed the plane defined by the two electrodes at 45 degrees. The laser pulse duration was about 6 ns, and the effective OH fluorescence lifetime is less than a couple of nanoseconds, meaning most OH-PLIF signal was captured by the ICCD using a short gating time (30 ns in this work). Using a longer gating time (2 μs in this work) the OH-PLIF and the optical emission from the plasma column can be simultaneously captured. Note that, due to the filters (UG5 + WG295) installed in front of the ICCD objective, the emission from the plasma column comes from OH* and part of N2* (see Fig. 11), which differs from the images recorded by the high-speed camera.
Figure 13 shows two PLIF images recorded at two different ICCD gate times, 30 ns and 2 μs, respectively. The OH-PLIF shows a hollow structure, indicating that ground-state OH surrounds the plasma column containing excited OH and N2. This is verified in the image (the red arrow) recorded with 2 μs gating time in Fig. 13 where the plasma column was recorded as well by optical emissions. The ground-state OH is transported from the plasma column and its spatial structure is thicker than the plasma column. As shown by the white arrow in Fig. 13, OH radicals still survive even after local extinction of the plasma column (the current path) during short-cutting events. A zoom-in of the OH hollow structure is shown in Fig. 14 . The thickness of ground-state OH layer was about 3 mm, and the hollow structure is clearly shown. At the center of the hollow structure, the PLIF signal intensity is almost at background level.
Fig. 13 Single-shot OH PLIF images captured using the ICCD with two gate times, (left) 30 ns for only capturing PLIF signal of ground-state OH and (right) 2 μs for capturing both OH-PLIF signal and plasma column emission. Ground-state OH is located around the plasma column (the red arrow), therefore a hollow structure appears in the OH-PLIF image to the right. Even when the current path is extinct, the ground-state OH still survives (the white arrow). The laser beam sheet contained the central line of the jet between the two electrodes and crossed the plane defined by the electrodes at an angle of 45 degrees.
Fig. 14 OH PLIF image collected above the electrodes. The laser beam sheet contained the central line the jet between the two electrodes and perpendicularly crossed with the plane defined by the two electrodes. Vertical and horizontal cross sections of OH PLIF distribution show the thickness of OH radical in this case.
One may speculate that the hollow structure in the OH PLIF was due to the high temperature in the center of the plasma column which caused the volume number density of OH to become too low to be imaged. This explanation would require hot ions on the order of 1eV, but the ions in our experiments were much colder. A reasonable explanation might be that the ground-state OH is formed from OH* by energy release and the formation happens during the species transport. In the single-shot spatially resolved spectra of the plasma column, as shown in Fig. 15 , OH* is located in the inner part of the plasma column with a thickness of around 4 mm. It is very likely that the OH* is mainly generated by electron impact dissociation of H2O [43]. However, the reasons for OH hollow profile should be further investigated using a kinetic model.
Fig. 15 (upper) Spectrally and spatially resolved, single-shot image of the plasma column in the VU spectral region recorded by an imaging spectrometer (gate time = 5.3 μs). (upper right) cross section in the area between two blue dash line. (lower) the spectra integrated in the area between two red dash lines.
Besides detailed structure investigation from single-shot OH PLIF, phase-averaged PLIF images can show ground-state OH global distribution in space and in time. The variance of OH PLIF signal intensity in a half cycle of the AC driving power was investigated when the gliding arc worked in continuous mode. The result based on the average of 100 frames is shown in Fig. 16(a) . OH PLIF signal intensity is clearly independent of time and hence the supplied voltage. This is explained by the long lifetime of the ground-state OH radicals. After the ground-state OH radical is formed in the outer layer of plasma column, it survives more than a half cycle (16μs). Since the ambient temperature was relatively constant, Fig. 16(a) also indicates the concentration of OH is constant with time. Compared with PLIF signal, emission intensities vary clearly with time, as shown in Fig. 16(b). Around the middle of the half cycle, i.e., at t = 8 µs, the emission intensity increases significantly which is consistent with Fig. 9(a). Note that emission from excited N2 also contributed to the intensity level in Fig. 16(b).
Fig. 16 (a) OH PLIF signal intensity as a function of time in a half cycle. (b) The emission intensity of plasma column (290 nm – 405 nm) in a half cycle. The averaged signals were based on 100 frames and the plasma column length had been counted. The zero point in the time scales indicated the beginning of a half cycle, when V(voltage) = 0. The gliding arc was synchronized with the laser and the ICCD camera.
The evolution of the ground-state OH global distribution was investigated by running the gliding arc plasma discharge in the burst mode. While keeping the phase delay time Td = 8 μs, the global distribution of OH radicals was captured by imaging accumulations, as shown in Fig. 17 . Besides the expansion of OH distribution with the delay time TD, the signal intensity of OH PLIF also increased. The integrated OH PLIF signal intensity and the top height of the plasma column are plotted in Fig. 18 as a function of the delay time TD. It shows that with increase of the delay time, OH radicals distribute to a wider and higher region and the integrated signal intensity increased significantly. The velocity of the plasma column is estimated by the differentiation of the polynomial fit of plasma height, and the result of around 10 m/s above the electrodes (i.e. TD = 2.5 - 5 ms) is consistent with that shown in Fig. 10(b). In the early stage (i.e. for small TD values, 0 – 1 ms), the OH radical distribution changed with a high speed of around 25 m/s, which is consistent with the nozzle exit velocity of about 40 m/s.
Fig. 17 The evolution of ground-state OH global distribution by accumulating 120 shots of OH-PLIF. The gliding arc discharge was run in the burst mode, and the phase delay time Td was kept at 8 μs. The plasma delay times TD are shown above the images.
Fig. 18 Integrated signal intensity of OH PLIF and the height of the GA plasma column as a function of time TD shown as filled circles and squares, respectively, and the solid lines are the polynomial fits. The velocity of the gliding arc plasma evolution is estimated by differentiating the polynomial fit (the blue solid line) of plasma height.
4. Conclusions
The evolution of a gliding arc discharge has been investigated using multi-kHz high-speed movies and Schlieren images, spatial and temporally resolved optical emission spectroscopy, and PLIF. Short-cut events in the plasma column were tracked and shown to occur about ten times more often than the re-ignition of the discharges in the narrowest gap between the electrodes. This type of short-cut events reduces the re-ignition of the gliding arc, hence improves the efficiency of the production of non-equilibrium plasma generation. Townsend discharges were observed at the beginning of the short-cut events. It was found that the re-ignitions between the electrodes occur while the previous plasma column still exists. After formation of the new plasma column, the previous plasma column decays. In the direct optical emission measurements, it was found that excited N2 dominated upstream near the electrodes whereas OH and NO were the main emitting species downstream. The emission intensity varied within a half cycle of the AC voltage (16 μs). Especially, N2 emission had a different variance compared with NO and OH. The emission intensity per unit length of plasma column becomes weaker during the evolution of the gliding arc discharge, but the changes become smaller when the gliding arc develops. From high-speed images, the emission intensity variance of the plasma column within a half cycle was clearly observed with a time resolution of 2 μs. Moreover, PLIF was applied to reveal the distribution of the ground-state OH. A hollow structure was found in the single-shot images of OH, which might be explained by the ground-state OH formation from the decay of excited OH. The exploration of further details of the complicated processes in a gliding arc will challenge more advanced techniques such as quantitative PLIF measurements of key radicals with up to multi-kHz framing rate.
Acknowledgment
This work was financially supported by the Swedish Energy Agency, CECOST (Center for Combustion Science and Technology), VR (Swedish Research Council) and European Research Council (Grant No. 246850). Financial support from the Ministry of Science, Technology and Innovation through the National Danish Proof of Concept Funding Scheme (grant number: 09-076196) was also acknowledged.
References and links
1. Y. Kusano, H. Mortensen, B. Stenum, S. Goutianos, S. Mitra, A. Ghanbari-Siahkali, P. Kingshott, B. F. Sorensen, and H. Bindslev, “Atmospheric pressure plasma treatment of glassy carbon for adhesion improvement,” Int. J. Adhes. Adhes. 27(5), 402–408 (2007). [CrossRef]
2. Y. Kusano, S. V. Singh, A. Bardenshtein, N. Krebs, and N. Rozlosnik, “Plasma surface modification of glass-fibre-reinforced polyester enhanced by ultrasonic irradiation,” J. Adhes. Sci. Technol. 24(11-12), 1831–1839 (2010). [CrossRef]
3. Y. Kusano, K. Norrman, J. Drews, F. Leipold, S. V. Singh, P. Morgen, A. Bardenshtein, and N. Krebs, “Gliding arc surface treatment of glass-fiber-reinforced polyester enhanced by ultrasonic irradiation,” Surf. Coat. Tech. 205, S490–S494 (2011). [CrossRef]
4. W. Sun, M. Uddi, T. Ombrello, S. H. Won, C. Carter, and Y. Ju, “Effects of non-equilibrium plasma discharge on counterflow diffusion flame extinction,” Proc. Combust. Inst. 33(2), 3211–3218 (2011). [CrossRef]
5. W. Sun, M. Uddi, S. H. Won, T. Ombrello, C. Carter, and Y. Ju, “Kinetic effects of non-equilibrium plasma-assisted methane oxidation on diffusion flame extinction limits,” Combust. Flame 159(1), 221–229 (2012). [CrossRef]
6. A. Fridman, A. Gutsol, S. Gangoli, Y. Ju, and T. Ombrello, “Combustion-assisted plasma in fuel conversion,” J. Phys. D Appl. Phys. 44(27), 274001 (2011). [CrossRef]
7. A. Fridman, A. Gutsol, S. Gangoli, Y. Ju, and T. Ombrellol, “Characteristics of gliding arc and its application in combustion enhancement,” J. Propul. Power 24(6), 1216–1228 (2008). [CrossRef]
8. T. Ombrello, Y. Ju, and A. Fridman, “Kinetic ignition enhancement of diffusion flames by nonequilibrium magnetic gliding arc plasma,” AIAA J. 46(10), 2424–2433 (2008). [CrossRef]
9. T. Ombrello, X. Qin, Y. Ju, A. Gutsol, A. Fridman, and C. Carter, “Combustion enhancement via stabilized piecewise nonequilibrium gliding arc plasma discharge,” AIAA J. 44(1), 142–150 (2006). [CrossRef]
10. C. M. Du, J. Wang, L. Zhang, H. X. Li, H. Liu, and Y. Xiong, “The application of a non-thermal plasma generated by gas-liquid gliding arc discharge in sterilization,” New J. Phys. 14(1), 013010 (2012). [CrossRef]
11. F. Leipold, Y. Kusano, F. Hansen, and T. Jacobsen, “Decontamination of a rotating cutting tool during operation by means of atmospheric pressure plasmas,” Food Contr. 21(8), 1194–1198 (2010). [CrossRef]
12. F. Leipold, N. Schultz-Jensen, Y. Kusano, H. Bindslev, and T. Jacobsen, “Decontamination of objects in a sealed container by means of atmospheric pressure plasmas,” Food Contr. 22(8), 1296–1301 (2011). [CrossRef]
13. M. Laroussi and F. Leipold, “Evaluation of the roles of reactive species, heat, and UV radiation in the inactivation of bacterial cells by air plasmas at atmospheric pressure,” Int. J. Mass Spectrom. 233(1-3), 81–86 (2004). [CrossRef]
14. M. Moreau, N. Orange, and M. G. J. Feuilloley, “Non-thermal plasma technologies: New tools for bio-decontamination,” Biotechnol. Adv. 26(6), 610–617 (2008). [CrossRef] [PubMed]
15. A. Fateev, F. Leipold, Y. Kusano, B. Stenum, E. Tsakadze, and H. Bindslev, “Plasma chemistry in an atmospheric pressure Ar/NH3 dielectric barrier discharge,” Plasma Process. Polym. 2(3), 193–200 (2005). [CrossRef]
16. Y. Kusano, F. Leipold, A. Fateev, B. Stenum, and H. Bindslev, “Production of ammonia-derived radicals in a dielectric barrier discharge and their injection for denitrification,” Surf. Coat. Tech. 200(1-4), 846–849 (2005). [CrossRef]
17. F. Leipold, A. Fateev, Y. Kusano, B. Stenum, and H. Bindslev, “Reduction of NO in the exhaust gas by reaction with N radicals,” Fuel 85(10-11), 1383–1388 (2006). [CrossRef]
18. A. Czernichowski, “Gliding arc - Applications to engineering and environment control,” Pure Appl. Chem. 66(6), 1301–1310 (1994). [CrossRef]
19. A. Fridman, S. Nester, L. A. Kennedy, A. Saveliev, and O. Mutaf-Yardimci, “Gliding arc gas discharge,” Prog. Energ. Combust. 25(2), 211–231 (1998). [CrossRef]
20. A. Bogaerts, E. Neyts, R. Gijbels, and J. van der Mullen, “Gas discharge plasmas and their applications,” Spectrochim. Acta B. 57(4), 609–658 (2002). [CrossRef]
21. C. Tendero, C. Tixier, P. Tristant, J. Desmaison, and P. Leprince, “Atmospheric pressure plasmas: A review,” Spectrochim. Acta B. 61(1), 2–30 (2006). [CrossRef]
22. O. Mutaf-Yardimci, A. V. Saveliev, A. A. Fridman, and L. A. Kennedy, “Thermal and nonthermal regimes of gliding arc discharge in air flow,” J. Appl. Phys. 87(4), 1632–1641 (2000). [CrossRef]
23. Y. Kusano, “Plasma surface modification at atmospheric pressure,” Surf. Eng. 25(6), 415–416 (2009). [CrossRef]
24. Y. D. Korolev, O. B. Frants, V. G. Geyman, N. V. Landl, and V. S. Kasyanov, “Low-current “gliding arc” in an air flow,” IEEE Trans. Plasma Sci. 39(12), 3319–3325 (2011). [CrossRef]
25. P. Bruggeman and D. C. Schram, “On OH production in water containing atmospheric pressure plasmas,” Plasma Sources Sci. Technol. 19(4), 045025 (2010). [CrossRef]
26. I. V. Kuznetsova, N. Y. Kalashnikov, A. F. Gutsol, A. A. Fridman, and L. A. Kennedy, “Effect of “overshooting” in the transitional regimes of the low-current gliding arc discharge,” J. Appl. Phys. 92(8), 4231–4237 (2002). [CrossRef]
27. S. Pellerin, J. M. Cormier, F. Richard, K. Musiol, and J. Chapelle, “Determination of the electrical parameters of a bi-dimensional dc Glidarc,” J. Phys. D Appl. Phys. 32(8), 891–897 (1999). [CrossRef]
28. S. Pellerin, F. Richard, J. Chapelle, J. M. Cormier, and K. Musiol, “Heat string model of bi-dimensional dc Glidarc,” J. Phys. D Appl. Phys. 33(19), 2407–2419 (2000). [CrossRef]
29. F. Richard, J. M. Cormier, S. Pellerin, and J. Chapelle, “Physical study of a gliding arc discharge,” J. Appl. Phys. 79(5), 2245–2250 (1996). [CrossRef]
30. B. Benstaali, P. Boubert, B. G. Cheron, A. Addou, and J. L. Brisset, “Density and rotational temperature measurements of the OH degrees and NO degrees radicals produced by a gliding arc in humid air,” Plasma Chem. Plasma Process. 22(4), 553–571 (2002). [CrossRef]
31. S. Pellerin, J. M. Cormier, K. Musiol, B. Pokrzywka, J. Koulidiati, F. Richard, and J. Chapelle, “Spatial fluctuations of 'gliding' arc,” High Temp. Mater. P-US. 2, 49–68 (1998).
32. X. Tu, H. J. Gallon, and J. C. Whitehead, “Dynamic behavior of an atmospheric argon gliding arc plasma,” IEEE Trans. Plasma Sci. 39(11), 2900–2901 (2011). [CrossRef]
33. Y. Kusano, S. Teodoru, F. Leipold, T. L. Andersen, B. F. Sorensen, N. Rozlosnik, and P. K. Michelsen, “Gliding arc discharge - Application for adhesion improvement of fibre reinforced polyester composites,” Surf. Coat. Tech. 202(22-23), 5579–5582 (2008). [CrossRef]
34. N. Balcon, N. Benard, P. Braud, A. Mizuno, G. Touchard, and E. Moreau, “Prospects of airflow control by a gliding arc in a static magnetic field,” J. Phys. D Appl. Phys. 41(20), 205204 (2008). [CrossRef]
35. S. P. Gangoli, A. F. Gutsol, and A. A. Fridman, “A non-equilibrium plasma source: magnetically stabilized gliding arc discharge: I. Design and diagnostics,” Plasma Sources Sci. Technol. 19(6), 065003 (2010). [CrossRef]
36. S. P. Gangoli, A. F. Gutsol, and A. A. Fridman, “A non-equilibrium plasma source: magnetically stabilized gliding arc discharge: II. Electrical characterization,” Plasma Sources Sci. Technol. 19(6), 065004 (2010). [CrossRef]
37. J. C. Sagas, A. H. Neto, A. C. Pereira Filho, H. S. Maciel, and P. T. Lacava, “Basic characteristics of gliding-arc discharges in air and natural gas,” IEEE Trans. Plasma Sci. 39(2), 775–780 (2011). [CrossRef]
38. C. O. Laux, T. G. Spence, C. H. Kruger, and R. N. Zare, “Optical diagnostics of atmospheric pressure air plasmas,” Plasma Sources Sci. Technol. 12(2), 125–138 (2003). [CrossRef]
39. R. K. Hanson, J. M. Seitzman, and P. H. Paul, “Planar laser fluorescence imaging of combustion gases,” Appl. Phys. B-Photo. 50, 441–454 (1990).
40. S. Hammack, X. Rao, T. Lee, and C. Carter, “Direct-coupled plasma-assisted combustion using a microwave waveguide torch,” IEEE Trans. Plasma Sci. 39(12), 3300–3306 (2011). [CrossRef]
41. A. Fridman, A. Chirokov, and A. Gutsol, “Non-thermal atmospheric pressure discharges,” J. Phys. D Appl. Phys. 38(2), R1–R24 (2005). [CrossRef]
42. A. Lebouvier, C. Delalondre, F. Fresnet, F. Cauneau, and L. Fulcheri, “3D MHD modelling of low current-high voltage dc plasma torch under restrike mode,” J. Phys. D Appl. Phys. 45(2), 025204 (2012). [CrossRef]
43. S. Teodoru, Y. Kusano, and A. Bogaerts, “The effect of O2 in a humid O2/N2/NOx gas mixture on NOx and N2O remediation by an atmospheric pressure dielectric barrier discharge,” Plasma Process. Polym. 9(7), 652–689 (2012). [CrossRef]
