Electrical breakdown in a pulsed asymmetric dielectric barrier discharge between a glass-covered mesh electrode and a grounded metal electrode in the air at atmospheric pressure is investigated. Volume discharge forms between the metal tip and the dielectric surface and spreads over the dielectric surface. Breakdown and discharge behaviors depend on the polarity of the charged electrode covered with glass compared to the metal rod electrode. In the case of the dielectric cathode (covered mesh), volume discharge features a stronger and longer-lasting emission. Volume discharge is weaker with outstretched surface discharge developing on the opposite glass electrode sustained by the embedded mesh when the metal rod functions as a cathode. The development and spatial distribution of the surface discharge depend on the relative polarity of the dielectrics caused by the charge deposition of the preceding discharge and is independent of the polarity of the applied high voltage. The discharge emission is brighter for the metal cathode and dielectric anode than for the metal anode, with a branching discharge developing and spreading in a star-like structure along the embedded grid, while a ring-like structure was observed for the metal anode and dielectric cathode. The duty cycle influences the discharge development and properties through the effects of the gas phase and surface pre-ionization.
The aim of this study is to investigate the discharge behavior in the volume and on the surface of an asymmetric DBD with a thin dielectric layer covering a fine mesh electrode and analyze the coaction of volume and surface discharges. We concentrate on the influences of the electrode polarity on the discharge and on the plasma distribution on the surface, the embedded grid on the surface discharge, and the pre-ionization on the discharge morphology by variation of the duty cycle of pulsed DBDs.
However, fundamental questions on the role of the thin barrier, the influence of the surface and pre-ionization, 19–22 the morphology of single filaments, and the properties of microdischarges including basic quantities as number densities, electron and gas temperatures, and the electric filed strength for special configurations are current research objects. 23–31
Plasma medicine is an emerging field, and the first DBD-based devices for the wound and skin treatment have already been introduced in the market. 4–11 However, the surface of the human body is anything but homogeneous or regular, which has motivated the design of flexible DBDs, e.g., a surface DBD incorporating thin (50 μ m) polymer dielectric barriers. 12 Another problem is the filamentary and, thus, erratic structure of the DBD in most molecular gases, including the air. An attempt to improve the surface uniformity of filamentary DBDs is the use of thin mesh electrodes covered by a dielectric. 13–18 The thin dielectric layer, on the one hand, prevents the discharge from the transition to a spark and, on the other hand, does not blur the electric field structure of the embedded mesh that much. In combination with a thin dielectric barrier, distinct but uniformly distributed spots with a maximum electric field strength are present in the discharge gap near the surface. The use of metal and dielectric electrodes in one arrangement leads to asymmetric discharges with higher discharge intensity and streamer velocities. 19
Due to their operation at atmospheric pressure and their scalability, dielectric barrier discharges (DBDs) have many different applications. Plate-to-plate DBD arrangements are useful for large scale applications such as activation and treatment of sheets. Homogeneous plasma formation, especially on the surface, is of major importance for a smooth operation of the DBD. 1,2 Not only diffuse DBDs are interesting approaches for this issue but also filamented DBDs with “diffuse” foot points spread over the surface. Self-organization of filamented DBDs has been investigated in this context, especially the formation and the breakdown mechanism of the barrier discharge and the memory effects. 3
The SPS contributed to the major part of the emission, while the FNS was correlated to the region of high electric field strength (i.e., high energetic electrons) due to its high excitation energy.
The spatiotemporal emission of the second positive system (SPS) of N 2 and of the first negative system (FNS) of N 2 + was recorded by cross-correlation spectroscopy (CCS) [or time-correlated single photon counting (TC-SPC)], 33 in a similar diagnostic setup as described in Refs. 19 and 34 . The CCS method allows the real-time measurement of the discharge event to be substituted by a statistically averaged determination of the cross-correlation function between two optical signals, the main signal and the synchronization signal, both originating from the same source. The main signal contains the data to be recorded, while the synchronization signal came from the light in the visible spectral range observed from the discharge. The time intervals between the detection of the signals are measured, resulting in a time histogram of counted photons of the discharge. Details of the CCS technique are described in Ref. 35 . The advantages of this method include its higher sensitivity compared to streak camera measurements and usefulness for discharges that have a statistical jitter. However, it requires data to be recorded for a considerable amount of time (stable conditions over hours). The CCS setup is pictured in Fig. 3 . It consisted of a time-correlated single photon counting module (Becker&Hickl SPS-150) and two sensitive photomultiplier tubes (PMTs) (Hamamatsu PMC-100-4). The main signal was imaged to the entrance slit of a 0.25 m-monochromator (Princeton Instruments SP-2300) with a grating of 2400 g/mm to record spectrally resolved emission from the discharge with PMT1. A flipping mirror scanned the discharge along the axis. The synchronization signal was collected by a light fiber, recorded with PMT2, and transmitted to the TC-SPC module via a delay box. The scanning mirror driver and the memory segments of the TC-SPC module were synchronized by a scan controller. The CCS measurements were performed for two transitions in nitrogen: the 0–3 transition of the second positive system of N 2 ,
Electrical measurements were performed with fast voltage (Tektronix P6015A) and current probes (Pearson 2877) at the HV and grounded side and recorded with a digital phosphor oscilloscope (Tektronix TDS 7254). Images of the DBDs were recorded in the visible spectral range with a fast ICCD camera (Andor iStar, Δ t ≥ 2 ns , Δ x < 10 μ m) with a commercial objective (Canon Compact-Macro 50 mm f/2.5), as shown in Fig. 2 . The camera was set either to 90 ° to observe the side-on emission of volume discharges or to 60 ° to observe both volume and surface discharges ( Fig. 2 ). Furthermore, the spatiotemporal development of the volume discharge was recorded side-on with a streak camera (Hamamatsu C5680). The streak camera measurements were performed spectrally integrated in the UV and VIS spectral range. A long-distance microscope (Questar QM100) was used to obtain a high spatial resolution of the volume discharge.
The discharge was driven by a high-voltage (HV) pulser producing pulses of 8 kV amplitude at a repetition frequency ( f rep ) of 4.88 kHz. The steepness of the rising and falling slopes of the high-voltage pulses was about 200 V/ns, specified by the pulse generator (DEI PVX-4110) powered by a DC HV source (FuG HCN 1400-12500), triggered by a pulse delay generator (SRS DG535) (see Fig. 2 ), similar to as described in Ref. 32 . Even though measurements were performed at a range of duty cycles between 5% and 95%, for the sake of clarity, this paper presents only the 10%, 50%, and 90% duty cycles, i.e., pulses of 20 μ s, 102.5 μ s, and 185 μ s duration (see Fig. 4 ). The discharges were generated in the open air at atmospheric pressure with humidity below 40%. The downward airflow of about 0.4 m/s was generated by a fan (Delta electronics, FFBO412VHN) close to the discharge.
The plasma source consists of a fine-meshed charged electrode covered by a 0.1 mm thick glass plate serving as a dielectric layer and a grounded rod. The mesh is slightly irregular, with an average spacing between wires of about 1 mm.
The plasma source consists of a fine-meshed charged electrode covered by a 0.1 mm thick glass plate serving as a dielectric layer and a grounded rod. The mesh is slightly irregular, with an average spacing between wires of about 1 mm.
The asymmetric dielectric barrier discharge arrangement for this study is shown in Fig. 1 . A fine rectangular mesh of 0.1 mm thickness wires spaced about 1 mm apart served as the charged electrode. The mesh was embedded in an epoxy resin and covered by a sheet of 0.1 mm thick glass. The glass cover served as the dielectric barrier and the epoxy resin prevented the mesh from coming in contact with the air and breaking down at unwanted places. The mesh was placed horizontally with the glass layer on top of it. A metal rod with a spherical tip and a diameter of 3 mm served as the ground electrode. To obtain spatially stable discharges, it was placed 0.5 mm away from the glass barrier and vertically aligned above a junction of the grid.
An estimate of the dissipated energy in the discharge gives ( 40 ± 10 ) μ J per pulse. From the discharge current integrals and the applied voltage of 8 kV, the amount of charge is estimated at ( 6 ± 2 ) nC.
When applying negative voltage pulses to the meshed electrode, the pulse amplitude was − 8 kV, while other settings remained the same. The current–voltage characteristics of the drop of the voltage from 0 to − 8 kV in the negative pulses and the drop of the voltage from + 8 kV to 0 in the positive pulses were very similar. Likewise, the increase of voltage in one polarity closely resembled the decrease of voltage in the other polarity. These effects can be explained by the conditions being actually set by the charging of the surface, 39 rather than the applied voltage. The electrons that were left on the dielectric by the previous discharge served as a rich source of pre-ionization. Consequently, the polarity applied was irrelevant, as long as the electric field was high enough to ensure discharge inception in the gap. Due to lower mobility, and thus slower recombination, the availability of electrons was much higher on the dielectric than in the gas volume. Therefore, the abundance of electrons on the glass surface to form electron avalanches made the difference between the positive and the negative streamers negligible.
The total current consisted of the discharge plus the displacement current. The displacement current at the HV side was higher due to the higher capacitance of the HV cord and of the grid including the resin of the charged electrode compared to the grounded rod electrode. The current peak was about 1 A for the rising slope for duty cycles of 1:9 and 1:1. The current dropped for a duty cycle of 9:1, i.e., for a decrease in the delay to the preceding pulse of 20 μ s due to the higher pre-ionization of the gap as explained in Ref. 37 . The residual ionization effect is similar to that in classical DBDs.
Screen shot (fast acquisition mode) of the current–voltage characteristics of the rising and falling slopes for duty cycles of 1:9 and 9:1; measured at grounded side and over approximately 10 000 discharges; the relative frequency is color-coded.
Screen shot (fast acquisition mode) of the current–voltage characteristics of the rising and falling slopes for duty cycles of 1:9 and 9:1; measured at grounded side and over approximately 10 000 discharges; the relative frequency is color-coded.
Figure 5 shows the current–voltage characteristics during the rising slope (RS) and the falling slope (FS) of + 8 kV for duty cycles of 1:9 and 9:1, measured at the grounded side. Displayed are about 10 000 subsequent discharges saved in one oscilloscope curve (fast acquisition mode of Tektronix TDS 7254), to show the statistics of discharges in both slopes and duty cycles. During the rising slope of the unipolar positive HV pulse, the dielectric surface above the mesh acted as the anode (D + ) and the metal rod electrode as the cathode (M − ). When the glass surface above the mesh was charged, the surface charge caused an electric field that countered the applied electric field imposed by the electrode system, extinguishing the discharge. During the falling slope, the applied voltage dropped to 0; hence, the residual charges on the glass generated an electric field sufficient for a breakdown. 39 In this case, the covered mesh electrode functioned as a cathode (D − ) (see the polarity of the current in Fig. 5 ).
Figure 4 shows the applied positive high-voltage pulses for three different duty cycles (pulse widths). During the rising slope and the falling slope of the applied voltage pulse, only one discharge occurred for each slope. Hence, the delay time between the two discharges (20 μ s–185 μ s) determined the precondition in terms of reactive species, residual charges in the gap, and on the dielectric surface for the following discharge event. This way, the discharge could be controlled by the pulse width. 32,37,38
The breakdown and discharge morphology depends on the polarity of the charged electrode (mesh) compared to the metal rod electrode. Figure 6 presents side-on ICCD images of the volume discharge of +8 kV at rising and falling slopes for all duty cycles at a gap of 0.5 mm recorded in the ultraviolet and visible spectral range. Voltage waveforms for all duty cycles are included for clarity. Displayed are single shots and accumulated images at both the rising and the falling slope, with an exposure time of 5 μs, tagged by gray bars in the voltage waveforms. The charged mesh is located on the left-hand side and the grounded metal rod on the right-hand side. The electrode polarity is indicated by M+/− (metal rod, grounded) and D−/+ (dielectric, charged mesh). Note the reflections of the volume discharge left on the glass surface above the mesh.
FIG. 6.
View largeDownload slideSide-on ICCD images of the volume discharge (single shots and accumulations) at rising (RS) and falling (FS) slopes for all duty cycles; electrode polarity indicated by M+/− and D−/+; electrode and mesh positions are marked by dotted lines; gap 0.5 mm; gate width 5 μs.
FIG. 6.
View largeDownload slideSide-on ICCD images of the volume discharge (single shots and accumulations) at rising (RS) and falling (FS) slopes for all duty cycles; electrode polarity indicated by M+/− and D−/+; electrode and mesh positions are marked by dotted lines; gap 0.5 mm; gate width 5 μs.
Close modal
The rod was adjusted opposite to a junction of the grid and the discharge always formed between the metal rod and the dielectric surface on a junction of the mesh under the glass. A channel was generated in the volume, which subsequently spread over the dielectric surface onto a much larger area. The surface discharge always expanded from the grid junction directly under the grounded electrode and followed the geometry of the mesh, for both polarities (to be further discussed in Sec. III C). A pronounced structure in the volume could be recognized during the falling slope, when the metal rod was the anode (M+D−) with the highest emission intensity in front of the anode. The emission intensity at the surface of the electrodes was weaker for the FS than the RS. In the case of the rising slope, the discharge emission was less intense in the volume, while the highest intensity was obtained at the electrodes. The single shots show one or multiple constricted discharge channels in the gap (see Fig. 6 RS 1:9; FS 9:1). The discharge diameter in the volume was about 100 μm.
The pre-ionization had a visible effect on the discharge behavior. The pre-ionization increased with a shorter delay (20–185 μs) to the preceding discharge,37–39 causing a little decrease in the discharge diameter in the center of the discharge (see single shots in Fig. 6). For the rising slope, this decrease occurred with an increase in the duty cycle, while for the falling slope with a decrease in the duty cycle. For comparison, during a relatively long pre-breakdown phase for sinusoidal driven DBDs, Hoder et al.37 found that residual surface charges on the dielectrics caused local field enhancement and promoted electron emission, leading to higher pre-ionization. Although the experiments presented here did not involve such a long period due to the steep gradient of the applied HV pulse,38,39 the influence of residual surface charges was still visible.
To obtain the temporal behavior of the discharge development in the volume, streak camera recordings were made along the central gap axis with a radial extension of about 20 μm of the discharge center, as described in, e.g., Ref. 32. The subsequent development of the surface discharge visible in Fig. 6 could not be detected by the streak camera because of the limited field of view of the center region of the discharge and at the surface foot point.
Figures 7 and 8 show streak images for the rising and falling slopes of a positive 9:1 pulse. Displayed are single shot and accumulated recordings. Figure 7 shows that the discharge in the rising slope was initiated in the volume and streamer development toward both electrodes within about 0.5 ns could be observed. A long-lasting surface discharge could be recognized on the glass surface. During the falling slope (Fig. 8), the discharge started in the volume closer to the rod and a cathode directed streamer to the mesh could be noticed. No pronounced long-lasting discharge on the glass surface (D−) was detected. This effect was due to the grounded metal electrode, where no charge collection took place, because of the instant withdrawal of electrons by the circuit. This behavior is different from symmetrical DBDs, which display symmetrical pictures for 1:9 (RS) and 9:1 (FS).32 A similar behavior was observed in Ref. 19 for asymmetrical DBDs, where the highest velocity of the ionization front was found for the metallic cathode (M−D+). It is assumed in Ref. 19 that for a dielectric cathode, the positive head of the ionization front is decelerated by charges on the glass surface. However, due to pulsed high-voltage,39 streamer velocities in Ref. 19 were not as high as in the current experiment with 2×106 ms−1 for the rising slope and 4×106 ms−1 for the falling slope (duty cycle 9:1).
FIG. 7.
View largeDownload slideStreak images for positive voltage at a duty cycle of 9:1 during the rising slope (M−D+); single shot record (on top) and accumulated ones (below).
FIG. 7.
View largeDownload slideStreak images for positive voltage at a duty cycle of 9:1 during the rising slope (M−D+); single shot record (on top) and accumulated ones (below).
Close modal
FIG. 8.
View largeDownload slideStreak images for positive voltage at a duty cycle of 9:1 during the falling slope (M+D−); single shot record (on top) and accumulated ones (below).
FIG. 8.
View largeDownload slideStreak images for positive voltage at a duty cycle of 9:1 during the falling slope (M+D−); single shot record (on top) and accumulated ones (below).
Close modal
To obtain a better insight into the breakdown behavior of the volume discharge, CCS measurements were conducted for the SPS of N2 and the FNS of N2+. The SPS contributes to the major part of the emission and can be directly compared to the streak camera measurements. However, the FNS is dominantly excited by direct electron collisions with a threshold energy of about 19 eV in such microdischarges. Only electrons in high electric field regions can gain such energies. Furthermore, the effective live time of the N2+(B) state is in the range of 0.1 ns due to collisional quenching. Thus, the FNS emission is correlated to the propagation of the electric field strength (Ref. 19).
The CCS measurements were conducted for the same gap of 0.5 mm but for a lower voltage of +5.6 kV and a duty cycle of 9:1 (185 μs). It was not possible to conduct CCS measurements for the same voltage used for all other measurements (8 kV) due to interfering signals from the discharge arising from the high displacement current during the slopes. The interference could be avoided when operating at lower voltages, but this led to an inception delay and a shift of the discharge current to the plateaus of the pulse away from the voltage slopes with a high disturbing displacement current. In addition, a discharge did not occur at 5.6 kV at every voltage cycle, which led to higher currents at the next discharge. Nevertheless, the voltage rise time, as well as the basic roles of the grounded electrode, the charged mesh and the glass dielectric, remained the same for both higher and lower voltage settings. Therefore, we will use the additional information from the emission from the FNS and the SPS of nitrogen to substantiate the discussion on the role of the electric field in this discharge.
Figures 9 and 10 show CCS images for +5.6 kV at a duty cycle of 9:1 during the rising slope and the falling slope. Displayed are the SPS and the FNS. The SPS data confirm the results for the volume discharges obtained with the streak camera recordings with +8 kV for both slopes regarding volume discharge behavior and duration and surface discharge development. Here again, the surface discharge was observed to last longer during the rising slope than during the falling slope. The FNS, representing the high electric field component of the discharge, behaved differently. The duration of the FNS emission was much shorter than that of the SPS emission. High signal intensity, comparable to high electric field strength, was observed at the cathode in both cases, regardless whether this was the metallic rod or the dielectric surface but, lasted longer in the case of the dielectric cathode (D−). It can be concluded that during the rising slope, the highest E-field strength and the highest energy electrons occurred at the grounded electrode, while the discharge on the surface was driven by lower-energy electrons in a lower E-field. During the falling slope, the highest E-field region was on the mesh, resulting in a more widespread surface discharge. The duration of the discharge emission was calculated from streak camera recordings. It was different at rising and falling slopes, and also depended on the duty cycle, as summarized in Fig. 11. The discharge in the volume during the falling slope (M+D−) lasted about 2.5 ns (FWHM), while the filament on the rising slope (M−D+) emitted light in the visible spectral range for less than 1.6 ns for a duty cycle of 1:9. The emission duration for a duty cycle of 9:1 was equal for both slopes and lasted about 1.8 ns (Figs. 7, 8, and 11). Similar behavior was observed by Hoder et al.19 They found the decay of the discharge of the configuration with the metallic anode (M+D−) to be the slowest one. Calculations of Braun et al.40 and Gibalov et al.41 for similar arrangements with a larger gap confirm the longer emission for the metallic anode.
FIG. 11.
View largeDownload slideDuration of spectrally integrated discharge emission of rising and falling slopes of +8 kV in dependence of the duty cycle, obtained from streak images.
FIG. 11.
View largeDownload slideDuration of spectrally integrated discharge emission of rising and falling slopes of +8 kV in dependence of the duty cycle, obtained from streak images.
Close modal
For more information Primary Effect Full Metal Mesh Air Filter, please get in touch with us!