Journal of Electrostatics 54 (2002) 207–214 Experimental characterization of multi-point corona discharge devices for direct ozonization of liquids Ilie Suarasana, Letitia Ghizdavub, Iustin Ghizdavuc, Sorin Budua, Lucian Dascalescua,* aTechnical University of Cluj-Napoca, 15 C. Daicoviciu Street, 3400 Cluj-Napoca, Romania bFaculty of Chemistry, Babes Bolyai University, 11 Arany Janos St, 3400 Cluj-Napoca, Romania cUniversity of Agricultural Sciences of Cluj-Napoca, 11 Calea Manastur Street, 3400 Cluj-Napoca, Romania Received 26 January 2001; received in revised form 15 May 2001; accepted 15 June 2001 Abstract Experimental investigations were carried out with the aim of evaluating the ozone generation characteristics of a class of multi-point corona electrodes employed for direct treatment of potable and industrial water. The corona electrodes were supplied from an AC high-voltage generator, 0,y, 25 kV, 5mA, 50Hz. The efficiency of ozonization depended on the following factors: (i) number and density of the discharge points of the corona electrode; (ii) gap length between the discharge points and the surface of the liquid; (iii) thickness of the liquid layer; (iv) duration of the exposure to the corona discharge and (v) presence of a dielectric barrier between the collecting electrode and the liquid. Typically, the multi-point corona electrode devices produces more than 9 g of ozone absorbed in 1000 l of water for each 1 kWh of energy provided by the high-voltage power supply. The conclusions of the present research work could guide the design of optimized devices for the direct ozonization of liquids. r 2002 Elsevier Science B.V. All rights reserved. Keywords: Corona discharge; Ozone generation; High-intensity electric field; Water treatment *Corresponding author. University Institute of Technology, Equipe Electron. and Electrostatique, LAII-ESIP, UPRES-EA 1219, 4 avenue de Varsovie, 16021 Angouleme Cedex, France. Tel.: +33-5-45-67- 32-40; fax: +33-5-45-67-32-49. E-mail address:
[email protected] (L. Dascalescu). 0304-3886/02/$ - see front matter r 2002 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 3 8 8 6 ( 0 1 ) 0 0 1 7 8 - 4 1. Introduction Ozonization is one of the methods extensively used for water purification [1,2]. At present, ozone (O3) is almost exclusively produced by dielectric barrier discharges, generated in devices similar to the one first proposed by Siemens in 1857 [3,4]. Many investigators studied the mechanisms and evaluated the efficiency of ozone generation in such installations. The advantages and the limits of this technology are now well documented [1,5]. During the last ten years, the investigations have focused on other types of electrical discharges with a potential use in the field of wastewater treatment. Thus, the production of active species such as OH, H2O2, and O3 in water and aqueous solution has been studied using the so-called pulsed-streamer corona discharge [6–9]. A combination of air stripping and pulsed corona has also been tested [10,11]. Other research groups have studied ozone generation in devices where DC corona discharge is created between an electrode and the surface of a liquid [12]. The work presented in the present paper aimed at evaluating AC corona discharges from multi-point high-voltage electrodes as means for direct ozonization of liquids. The obtained results will hopefully contribute to the development of novel, more effective ozone generators for wastewater treatment. 2. Experimental procedure The experimental devices designed for direct ozonization of liquids by means of AC corona discharges were composed of an ‘‘active’’ high-voltage electrode and a grounded plate electrode (Fig. 1). The electrode system was fed from a continuously adjustable 0,y, 25 kV, 5mA, 50Hz, high-voltage supply. The liquid to be treated (volume: 10–70ml) was taken in a Petri glass vessel of inner diameter+¼ 95mm. Fig. 1. Schematic representation of the experimental device; (a) test cell provided with a corona electrode disposing of one, two, or three discharge points; (b) test cell provided with a ‘‘brush-type’’ multi-point corona electrode. 1FAC high-voltage power supply, 25 kV, 5mA, 50Hz; 2FPetri vessel containing the liquid to be treated; 3Factive electrode; 4Ftreated liquid; 5Fmetallic contact between the treated liquid and the grounded electrode and 6Fgrounded plate electrode. I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214208 Two types of ‘‘active’’ electrodes were tested. The first consisted of 1–3 metallic points, disposed at variable distances (p ¼ 122:52527:5210mm) on a cylindrical metallic support (Fig. 1a). The second is a multi-point (brush-type) arrangement, the distance between two adjacent spikes being 1mm (Fig. 1b). The tungsten wire segments 150 mm in diameter, the tips of which were employed as coronating points, were attached to a metallic plate. The distance (s) between the surface of the ‘‘treated’’ liquid and the tips of the tungsten wire segments could be continuously adjusted from 20 to 40mm. Two series of experiments were carried out with the ‘‘treated’’ liquid either connected to the earth (case I) or at floating potential (case II) with respect to the grounded electrode. The variables in the experiments were (i) the configuration of the electrode system, (ii) the duration of the discharge, (iii) the applied high-voltage and (iv) the thickness of the liquid layer. The quantity of ozone in the ‘‘treated’’ liquid was determined with the iodometric method. Thus, the generated ozone was absorbed by a neutral or alkaline 2% potassium iodide (KI) solution (70ml), which was then acidified by adding 5ml of H2SO4 solution (2N), for the obtainment of iodine from the complex of KI3. The obtained iodide atoms were visualized (blue coloration) by using 1ml of amidon solution (0.5%), and their quantity was determined by titrating with Na2S2O3 solution (0.001N). The involved chemical reactions, which occur in acid medium, were O3þ2KIþH2O-I2þ2KOHþO2; I2þ2Na2S2O3-Na2S4O6þ2NaI: For concentrations situated between 2 and 160mg O3/l, the error of the method is 1%. The quantity of absorbed ozone was estimated with the following formula: C ¼ 171:6� 10�3V2=V1 ðmgO3=l liter of 2%KI solutionÞ; where V1 is the volume of KI solution (2%), V2 the volume of Na2S2O3 solution employed in the titration process. 3. Results The quantity of ozone generated in an experimental cell provided with a single point corona electrode was found to depend on both the applied high-voltage and the duration of the exposure to the discharge (Fig. 2). The results of the experiments carried out with two or three points, located at a variable distance p from one another, are presented in Fig. 3. A comparison between the quantities of ozone generated in the case of grounded or floating liquid is shown in Fig. 4, for two discharge exposure durations. Fig. 5 shows the ozone quantity generated in a test cell provided with a multiple-point active electrode (brush-type), for different discharge gaps and discharge exposure durations. If the average electric field strength Eav ¼ U=s was maintained at a constant level of 10 kV/cm, the quantity of ozone I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214 209 Fig. 2. Ozone quantity as a function of high-voltage level (a) and discharge exposure duration (b). Experimental conditions: discharge gap s ¼ 20mm; discharge exposure duration 60 s (only for a), and applied high-voltage U ¼ 20 kV (only for b). Case I: liquid connected to the ground; Case II: liquid at floating potential. Fig. 3. Ozone quantity as a function of the inter-points distance, for a discharge gap s ¼ 20mm, an applied high voltage U ¼ 20 kV, and discharge exposure durations of 30 s (a) and 60 s (b), with the liquid grounded (Case I) or at floating potential (Case II). I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214210 varied with the applied high voltage as shown in Fig. 6. The curve in Fig. 7 represents the variation of the ozone quantity with the volume of the ab- sorbent liquid (KI 2% solution). The correspondence between the volume of the absorbent liquid and the thickness of the liquid layer present in the Petri vessel is given in Table 1. Fig. 4. Comparison between the ozone quantities generated by a 1-point corona electrode, with the liquid grounded (Case I) or at floating potential (Case II), for a discharge gap s ¼ 20mm, and an applied high voltage U ¼ 20 kV. Fig. 5. Comparison between the ozone quantities generated by a multi-point (brush type) corona electrode at an applied high voltage U ¼ 20 kV, for two discharge exposure durations and several discharge gaps, with the liquid grounded (Case I) or at floating potential (Case II). I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214 211 4. Discussion The quantity of ozone is proportional to the discharge exposure duration (Figs. 2b, 4 and 5) and to the levels of the high voltage applied to the electrodes Fig. 6. Ozone quantity generated by a 3-point corona electrode as a function of the high voltage applied to electrodes, at constant average electric field strength U=s ¼ 10 kV/cm, for an interval between adjacent points p ¼ 10mm, and a discharge exposure duration t ¼ 60 s. Fig. 7. Ozone quantity generated by a 3-point corona electrode as a function of the volume of absorbent liquid, for a discharge gap s ¼ 10mm, an interval between adjacent points p ¼ 10mm, and a discharge exposure duration t ¼ 60 s. I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214212 (Fig. 2a), even in the case of constant average electric field strength U=s (Fig. 6). In any case, the higher the applied voltage, the larger was the quantity of gener- ated ozone. Nevertheless, it should be noted that the quantity of ozone fixed in the absorbent liquid arrives to a saturation level at long discharge exposure duration (Fig. 2b), or at very high values of the voltage applied to the electrodes (Fig. 6). Reducing the distance between the adjacent points is generally accompanied by an increase of the generated ozone quantity. Beyond a certain threshold, increasing the density of the discharge points has a rather insignificant effect (Fig. 3). This observation can be explained by the saturation of the corona current when the points are close to each other. As expected, at the same applied high-voltage levels, the ozone production is more important when the discharge gap is shorter and the ‘‘treated’’ liquid is grounded (Fig. 5). When the liquid is at a floating potential, the corona current has lower values, but the aspect of the discharge is more uniform. In otherwise similar conditions, the concentration of the adsorbed ozone is lower for larger quantities of ‘‘treated’’ liquid (Fig. 7). Therefore, the ozonization is more effective for thinner liquid layers. The previously published data on the V-A characteristics of the multi-point corona discharge devices [13] enable the estimation of the energy efficiency of proposed method of direct ozonization of liquids. For instance, the three-point corona discharge device operating for 60 s at a current I ¼ 1:5mA and a voltage U ¼ 20 kV (Case I, p ¼ 20mm, s ¼ 20mm) generates 0.63mg O3, absorbed in 70ml of water. The energy consumption can be expressed as 1.5 [mA]� 20 [kV]� (1/60) [h]=0.5VAh. Thus, the device can generate 9mg O3 in 1 l of water, for an input energy of 0.5 [VAh]� 1000 [ml]/70 [ml]=7.1VAh. For a largely over-estimated power factor cos f ¼ 0:1; an active energy of 1 kWh should be provided for the generation of 12.7 g of O3, absorbed in 1400 l of water. This result is difficult to compare to the energy efficiency of commercial ozonizers, for which the reported data refer to the quantity of O3 generated in air (typi- cally 200 g for 1 kWh), and not to the one absorbed in water. As O3 is known to have a very short life time in air, a significant part of the 200 g generated in such a device for 1 kWh will vanish before having the chance to enter in contact with the ‘‘treated’’ liquid. The device for direct ozonization eliminates this draw- back and could represent a promising alternative to present solutions, even if the difficulty of controlling the humidity of ambient air could limit its performances. Table 1 The correspondence between the volume and the layer thickness of the absorbent liquid occupying the Petri vessel QKI (ml) 10 20 30 40 50 60 70 h (mm) 1.5 3 4.6 6.1 7.7 9.2 10.7 I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214 213 5. Conclusions The generation of an AC corona discharge between an electrode and the surface of the ‘‘treated’’ liquid is an effective ozonization method. The efficiency of the procedure depends on the geometry of the electrode system, the applied high-voltage level, the discharge exposure duration, and the thickness of the liquid layer. Ozone concentrations greater than 9 g/1000 l of tap water can be obtained with an energy expense of about 7 kVAh (o1 kWh). Further research is in progress to investigate the possibility of increasing the ozone generation efficiency by rising the frequency of the high-voltage supply, by reducing the discharge gap, and by modifying the geometry of the corona electrode. References [1] U. Kogelschatz, B. Eliasson, Ozone generation and applications, in: J.S. Chang, A. Kelly, J.M. Crowley (Eds.), Handbook of Electrostatic Processes, Marcel Dekker, New York, 1995, pp. 585–605. [2] J. Staehelln, J. Holgne, Environ. Sci. Technol. 19 (1985) 1206. [3] B. Eliasson, M. Hirth, U. Kogelschatz, J. Phys. D 20 (1987) 1421. [4] J. Kitayama, M. Kuzumoto, J. Phys. D 30 (1997) 2453. [5] S. Yagi, M. Tanaka, J. Phys. D 12 (1979) 1509. [6] D.M. Wilberg, P.S. Lang, R.H. H .ochemer, A. Kratel, M.R. Hoffmann, Environ. Sci. Technol. 30 (1996) 2526. [7] J.S. Clements, M. Sato, R.H. Davis, IEEE Trans. Ind. Appl. 23 (1987) 224. [8] M. Sato, T. Ohgiyama, J.S. Clements, IEEE Trans. Ind. Appl. 32 (1996) 106. [9] B. Sun, M. Sato, A. Harano, J.S. Clements, J. Electrostat. 43 (1998) 115. [10] A.A. Ai-Arainy, S. Jayaram, J.D. Cross, Conference Record of ICDL, Rome, Italy, 1996, pp. 427– 431. [11] P. Lubucki, S. Jayaram, J.D. Cross, A.A. Al-Arainy, Annual Report CEIDP, San Francisco, 1996, pp. 730–733. [12] S.C. Goheen, G.M. Mong, G. Pillay, D.M. Camaioni, First International Conference on Advanced Oxidation Technology for Water and Air Remediation, London, Ont, 1994, pp. 83–84. [13] I. Suarasan, S. Budu, S. Biluca, L. Ghizdavu, I. Ghizdavu, Acta Universitatis Cibiniensis 41 (1999) 201. I. Suarasan et al. / Journal of Electrostatics 54 (2002) 207–214214