2372 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 42, NO. 10, OCTOBER 2014 Development of a Single Filament Pulsed Dielectric Barrier Discharge in Volume and on Surface Hans Höft, Manfred Kettlitz, Tomáš Hoder, Ronny Brandenburg, and Klaus-Dieter Weltmann Abstract— Pulsed-driven dielectric barrier discharges were recorded in a single filament arrangement with a 1-mm gap in aN 2 –O 2 gas mixture at atmospheric pressure with a fast ICCD camera. The comparison of spectrally-integrated and spectrally- resolved emissions for the second positive system (SPS) and the first negative system (FNS) of N 2 reveals different structures in the volume and on the dielectric surfaces. A propagation of surface discharges with approximately the same velocity is visible on both electrodes, but with a higher intensity of the FNS on the cathode and of the SPS on the anode. This indicates a higher electric field strength in the discharges on the cathode surface. Index Terms— Atmospheric pressure plasmas, electric breakdown, nitrogen, oxygen, plasma diagnostics. D IELECTRIC barrier discharges (DBDs) in small gaps at atmospheric pressure are used in several applications, such as surface modification or cleaning, as well as decon- tamination or purification of gases [1]. The reactive species needed for the chemical processes are generated during the transient microdischarge (MD) developing in the volume, as well as on the dielectric surface [1]. This paper focuses on the development of the MD emission for the 0–0 transition of the second positive system (SPS) at 337.1 nm and the first negative system (FNS) at 391.4 nm of N 2 , which have different excitation energies (the lowest excited state for SPS is 11.0 eV and for FNS 18.8 eV [2]). The DBD arrangement used for this experiment is shown schematically in Fig. 1(a). The configuration consisted of half- sphere alumina-covered stainless steel electrodes mounted in a Plexiglas discharge cell [3]. The working gas (0.1-vol% O 2 in N 2 ) flowed through the cell perpendicularly to the gap. The DBDs were generated by a positive unipolar HV pulse with 10-kV amplitude, 10-kHz repetition rate, and ∼250 V/ns slope steepness. All results presented here are for MDs gen- erated in the rising slope of the HV pulse. The MD emission was recorded with an Andor iStar ICCD camera connected to a far-field microscope. The spectral resolution was realized by means of interference filters (Melles Griot) with central wavelengths of 337.1 and 391.4 nm and a bandwidth full- width at half-maximum of 3 and 1 nm, respectively. For the Manuscript received November 13, 2013; revised April 7, 2014; accepted June 14, 2014. Date of publication July 14, 2014; date of current version October 21, 2014. The authors are with INP Greifswald, Greifswald 17489, Germany (e-mail: hans.hoeft@inp-greifswald.de; kettlitz@inp-greifswald.de; hoder@inp-greifswald.de; brandenburg@inp-greifswald.de; weltmann@ inp-greifswald.de). Digital Object Identifier 10.1109/TPS.2014.2332882 spectrally-integrated measurements, the filter was replaced by a quartz plate. The use of quartz optics and quartz windows in the DBD cell allowed for transmission of UV emissions from the discharge. The spatial resolution for this arrangement was ∼2 μm. The field of view and the scale can be found in Fig. 1(a). The intensity of the ICCD signals in Fig. 1(b)–(o) is scaled with respect to each maximum and displayed in pseudocolor. The spectrally-integrated single shot with 150-ns gate width in Fig. 1(b) shows the main characteristics of the MD structure; a thin constricted channel in the gap that broadens in front of the anode, while the surface discharge channels spread on both electrodes. Comparing the single shot result with that of 1000 accumulations in Fig. 1(c) shows the stability and reproducibility of the MD: 1) the channel in the gap; 2) the emission structure with its maximum in the volume and on the anode; and 3) the branches of the surface discharge can be found in both Fig. 1(b) and (c). The MD structure of the entire discharge for the 0–0 transition of the SPS [Fig. 1(d)] shows no significant difference from the spectrally-integrated signal, since the SPS emission is the dominant part of the optical emission spectrum of the discharge investigated [3]. Contrary to the SPS emission structure, the intensity maximum of the FNS emission [Fig. 1(e)] is located on the cathode sur- face, which can be correlated with the impact of the cathode- directed ionization front (positive streamer). A small local maximum is also visible in the gap near the anode [4], [5]. The emission of the FNS reveals the region of high electric field strength due to its high excitation energy. The time- resolved development of the MD emission in a 5-ns ICCD gate is shown in Fig. 1(f)–(j) for the SPS and in Fig. 1(k)–(o) for the FNS. The delay times were identical for the SPS and FNS. The first pair of images at t delay = t 0 shows the initial breakdown phase. For the FNS [Fig. 1(k)], there is a localized emission in the volume near the anode, which can be interpreted as the start of the cathode-directed propagation [4], [5]. The SPS emission in Fig. 1(f) shows a maximum in front of the anode, which is caused by electrons generated in the discharge channel and accelerated toward the anode. In the next time steps after 2.5 and 5.0 ns, the streamer has crossed the gap. For the SPS, there is an emission maximum in the gap in front of the anode [Fig. 1(g) and (h)], whereas for the FNS the max- imum occurs on the cathode surface [Fig. 1(l) and (m)], with only weak emission in the MD channel. This dif- ference between the SPS and FNS emission structure in 0093-3813 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.